id
stringlengths 20
20
| year
int64 2.01k
2.02k
| title
stringlengths 21
249
| sections
stringlengths 4.88k
32.5k
| headings
stringclasses 217
values | abstract
stringlengths 32
4.12k
| summary
stringlengths 227
2.45k
| keywords
stringlengths 7
1.31k
⌀ | toc
stringclasses 267
values |
|---|---|---|---|---|---|---|---|---|
journal.pgen.1006646 | 2,017 | Integrating molecular QTL data into genome-wide genetic association analysis: Probabilistic assessment of enrichment and colocalization | Genome-wide association studies ( GWAS ) have successfully identified many genomic loci that impact complex diseases and complex traits ., Nevertheless , the molecular pathways that connect genetic variants to complex traits are still poorly understood , primarily because a considerable proportion of trait-associated signals are located in the non-coding region of the genome ., With recent advancements in high-throughput sequencing technology , systematic investigations of cellular phenotypes have revealed an abundance of non-coding molecular quantitative trait loci ( QTLs ) 1–4 ., Integrating molecular QTL data into GWAS analyses has shown great potential in unveiling the missing links between trait-associated genetic variants and organismal phenotypes 5–7 ., In this paper , we focus on a specific type of integrative analysis that aims to assess the overlapping/colocalization of causal GWAS hits and causal molecular QTLs ( also known as quantitative trait nucleotides , or QTNs ) ., Following Giambartolomei et al 8 , we define a GWAS hit and a molecular QTN as being colocalized if a single genetic variant is causally associated with both the complex and molecular traits of interest ., Colocalizing genetic variants that jointly affect both molecular and organismal phenotypes provides an intuitive starting point for exploring the role of genetic variants in disease etiology ., Taking expression quantitative trait loci ( eQTL ) mapping as an example , colocalizing an eQTL signal with a GWAS hit naturally suggests that the target gene of the eQTL may play an important role in the molecular pathway of the complex traits ., Additionally , other types of available integrative analysis approaches , e . g . , Sherlock 9 , PrediXcan 5 and other similar approaches 10 , 11 , can also benefit from accurate colocalization analysis , either for improved power ( as in the case of Sherlock ) or better interpretation of the inference results ( as in the case of PrediXcan ) ., Considering the most common practical setting in which GWAS and molecular QTL data are obtained from two non-overlapping sets of samples , we propose a natural Bayesian hierarchical model for integrating the two types of association data ., Specifically , we regard the ( latent ) association status of each candidate SNP with respect to the molecular phenotype of interest as an SNP-level annotation , and we attempt to quantify the odds of an annotated SNP being causally associated with the complex trait of interest , which is statistically equivalent to evaluating the enrichment level of annotated SNPs in the causal GWAS hits ., Subsequently , the resulting enrichment estimates are utilized in the downstream fine-mapping ( of GWAS hits ) and colocalization analyses ., Within our Bayesian hierarchical model , we show that the problems of enrichment estimation , fine-mapping and colocalization testing can be seamlessly solved in a unified inference framework ., In addition , our approach is computationally efficient and requires only summary-level data from both molecular QTL mapping and GWAS ., Our proposed method is most similar to the probabilistic model-based approaches coloc 8 and eCAVIAR 12 , which represent the state-of-the-art in the current literature ., The advantages of the model-based colocalization analysis methods over the empirical methodologies ( e . g . , Nica et al 6 ) have been fully demonstrated through both rigorous theoretical arguments 8 , 13 and carefully constructed simulation studies 12 ., In this paper , we show that both coloc and eCAVIAR can be viewed as special cases of the proposed approach ., In particular , both approaches bypass the enrichment analysis by making subjective assumptions on the enrichment levels of molecular QTLs in GWAS signals ., In comparison , our approach shares the advantages of both existing approaches , but it enjoys additional flexibility and improved statistical rigor ., Most importantly , our approach provides calibrated statistical quantification on colocalized association signals ., Without loss of generality , we consider a GWAS of a quantitative trait and describe its associations with p candidate SNPs and n unrelated samples using a multiple linear regression model ,, y = ∑ i = 1 p β i g i + e , e ∼ N ( 0 , τ - 1 I ) , ( 1 ), where we assume that both the phenotype and genotypes are centered ( the intercept term is therefore exactly 0 ) and denote the complete collection of genotypes as G ≔ g1 , … , gp ., We further denote the latent binary association status of each SNP i by dichotomizing its genetic effect βi , i . e . , γi = 1 indicates that SNP i is genuinely associated ( thus , βi ≠ 0 ) , and γi = 0 otherwise ., It can be argued that the aim of the GWAS is to make inference of the binary vector γ ≔ ( γ1 , … , γp ) ., In addition , we assign the standard spike-and-slab prior for each regression coefficient βi and a flat gamma prior for the residual error variance parameter τ ., Suppose that a single quantitative annotation ( categorical or continuous ) is available for each candidate genetic variant ., We integrate the SNP-level annotation into the association analysis by specifying a natural logistic prior for each candidate SNP i , i . e . ,, log Pr ( γ i = 1 ) Pr ( γ i = 0 ) = α 0 + α 1 d i ., ( 2 ), In particular , we denote the complete collection of the SNP annotation data as d ≔ ( d1 , … , dp ) , and we refer to α ≔ ( α0 , α1 ) as the enrichment parameter: for a binary annotation , a positive α1 value indicates that SNPs with the feature have increased odds of being associated with the trait of interest , i . e . , the annotated feature is enriched in the trait-associated genetic variants ., In this paper , we consider a special setting in which the annotation is derived from the association analysis of molecular QTL data , namely , ( Y qtl , Gqtl ) ., Intuitively , the true association status of each SNP with the molecular phenotype can be naturally incorporated as annotations in Eq ( 2 ) for GWAS analysis ., However , due to the intrinsic limitations in the molecular QTL mapping , e . g . , imperfect power and complication of LD among SNPs , the precise binary association status of each SNP with respect to the molecular phenotype of interest , d , is practically impossible to obtain ., Consequently , there is considerable uncertainty in annotating any causal molecular QTN ., To fully characterize the uncertainty of the molecular QTL annotation and carry it over into the proposed integrative analysis , we propose embedding a latent covariate model for d in the prior model ( 2 ) ., Specifically , we consider d to be an unobserved random vector whose realization is drawn from the following probability distribution:, d ∼ Pr ( d ∣ Y q t l , G q t l ) ., ( 3 ), In particular , we obtain the desired posterior distribution Pr ( d ∣ Y qtl , Gqtl ) from a Bayesian multi-SNP association analysis of molecular QTL data 14 ., Henceforth , we refer to the distribution Pr ( d ∣ Y qtl , Gqtl ) as the “fuzzy” annotation for molecular QTLs ., Based on the proposed Bayesian hierarchical model , we perform statistical inference to address three related problems ., First , we aim to estimate the enrichment parameter α to quantify the enrichment level of molecular QTNs in the causal GWAS hits ., Second , we perform Bayesian fine-mapping analysis of GWAS hits accounting for the molecular QTL annotations , and we summarize the results in form of the posterior probability Pr ( γ ∣ y , G , Y qtl , Gqtl ) ., Third , we attempt to evaluate the colocalization of the molecular QTNs and the causal GWAS hits , i . e . , for each SNP i , we examine whether γi = di = 1 ., Within our proposed modeling framework , the colocalization at the single SNP-level is naturally quantified by the posterior probability Pr ( γi = 1 , di = 1 ∣ y , G , Y qtl , Gqtl ) ., A distinct feature of our proposed integrative analysis framework is the integration of the enrichment estimation in the colocalization analysis ., In this section , we illustrate the critical impact of enrichment estimates on the quantitative results of colocalization analysis ., LD is one of the primary factors that complicate the colocalization analysis ., This is mainly because of the increasing difficulty in identifying causal SNPs from the association data as the LD between candidate SNPs becomes stronger ., Consider a hypothetical example of two perfectly correlated SNPs and assume that they are in complete linkage equilibrium with the remaining candidate SNPs ., Suppose that one of the two SNPs is genuinely associated with the molecular phenotype ., A well-powered QTL mapping analysis should identify that one of the SNPs is a causal QTN , but there is no further information to distinguish the two ., The exact same situation arises if one of the two SNPs ( not necessarily the QTN ) is genuinely associated with the complex trait ., Because of the complete symmetry , the two candidate SNPs also carry identical SNP-level colocalization probabilities and are not identifiable based only on the association data ., Nevertheless , a statistical statement can be made regarding the genomic region harboring these two SNPs , and the quantification of such probability can be notably different depending on the enrichment information ., If the molecular QTNs are completely irrelevant to the causal GWAS hits , or statistically speaking , γ and d are independent ( hence , α1 = 0 in our prior model ) , we should conclude that there is a 50% chance that the two types of causal associations are overlapped in one of the two SNPs , i . e . , the probability that the genomic region harboring a colocalized signal is 0 . 50 ., Conversely , if ( almost ) all the molecular QTNs are indeed causal GWAS hits ( hence , α1 → ∞ in our prior model ) , we would conclude that , with near certainty , one of the two SNPs is responsible for both genuine associations , i . e . , the probability that the region harboring a colocalized signal is approaching 1 . 0 ., We would like to note two points from the above hypothetical example: first , in the presence of LD , a regional colocalization probability ( RCP ) has better practical interpretation than the SNP-level colocalization probability ( SCP ) ; second , the enrichment information characterized by α1 has a profound impact on quantifying RCPs ., Next , we show that the quantified enrichment estimate can be used to calculate the expected number of colocalized association signals based on the proposed prior model without delving into the detailed analysis of individual loci ., We denote the marginal ( prior ) probabilities pγ ≔ Pr ( γi = 1 ) and pd ≔ Pr ( di = 1 ) ., Based on Eq ( 2 ) , it follows that, Pr ( γ i = 1 , d i = 1 ) = p γ 1 + 1 - p d p d e - α 1 ., ( 4 ), Note that the quantity, ρ ≔ Pr ( d i = 1 ∣ γ i = 1 ) = 1 1 + 1 - p d p d e - α 1 ( 5 ), represents the fraction of causal GWAS hits overlapping causal molecular QTNs ., The interplay of pd , pγ and α1 with respect to ρ can be intuitively understood in some extreme scenarios ., For example , if the vast majority of the genome is annotated as molecular QTNs , i . e . , if pd → 1 , then ρ → 1 and Pr ( γi = 1 , di = 1 ) → pγ ., This is because if every SNP in the genome is likely a molecular QTN , then every causal GWAS SNP is also likely a molecular QTN ., More generally , the colocalization probability is affected by the enrichment level of molecular QTNs in the GWAS hits ., Specifically , if α1 → ∞ , ρ → 1 and Pr ( γi = 1 , di = 1 ) → pγ , i . e . , all GWAS hits are expected to be molecular QTNs ., Alternatively , if α1 = 0 , it follows that ρ = pd and Pr ( γi = 1 , di = 1 ) = pγ pd , i . e . , the two types of associations are mutually independent ., Moreover , if molecular QTLs are depleted in the GWAS hits , i . e . , α1 < 0 , ρ is expected to be < pd ., Furthermore , the prior expected number of colocalized association signals can be simply computed by, E Number of colocalized causal variants = M p γ 1 + 1 - p d p d e - α 1 , ( 6 ), where M represents the total number of SNPs interrogated ., The exact computation to fit the proposed hierarchical model is intractable ., Although approximate computation is theoretically possible using the Markov Chain Monte Carlo ( MCMC ) algorithm , it does not scale well to genome-wide GWAS and molecular QTL data ., Here , we provide the necessary background on the existing computational work and outline the computational procedures to achieve our three inference goals for the integrative analysis ., Assuming that the annotation d is observed , our previous work 14 proposes a two-stage empirical Bayes procedure to perform accurate and efficient approximate Bayesian inference in the GWAS setting ., Briefly , in the first stage , we obtain the maximum likelihood estimate of the enrichment parameter , α ^ , using an EM algorithm by treating γ as missing data ., Subsequently , in the second stage , we approximate the desired posterior probability Pr ( γ ∣ y , G , d ) in GWAS analysis by Pr ( γ ∣ y , G , d , α ^ ) ., In addition , and particularly for analyzing GWAS data , we divide the genome into K roughly independent LD blocks using the approach described in 15 , i . e . , γ = γ1 ⊕ γ2 ⊕ ⋯ ⊕ γK , and further approximate Pr ( γ ∣ y , G , d , α ^ ) by ∏ i = 1 K Pr ( γ i ∣ y , G , d , α ^ ) ., Within each LD block i , Pr ( γ i ∣ y , G , d , α ^ ) is then computed using the deterministic approximation of posteriors ( DAP ) algorithm ., Among the two variants of the DAP algorithm described in 14 , the adaptive DAP algorithm implements a fully automated Bayesian multi-SNP analysis procedure ., Conversely , the DAP-1 algorithm further assumes at most a single causal association within the LD block of interest , but it achieves even more efficient computation and requires only summary-level statistics from the GWAS data ., With the added latent covariate model ( 3 ) , the computational challenge becomes even greater ., We extend our existing empirical Bayes framework into a three-stage procedure to explicitly account for the fuzzy annotation of d ., The first stage focuses on finding the MLE α ^ in the presence of missing data d ., In the second stage , we approximate Pr ( γ ∣ y , G , Y qtl , Gqtl ) by Pr ( γ ∣ y , G , Y q t l , G q t l , α ^ ) to conduct fine-mapping of GWAS signals incorporating the annotation of molecular QTNs ., The particular emphasis in this step is to construct the SNP-level priors accounting for the uncertainties of molecular QTLs ., In the last stage , we use the results from the previous stages to approximate the SNP-level posterior probability Pr ( γi = 1 , di = 1 ∣ y , G , Y qtl , Gqtl ) by Pr ( γ i = 1 , d i = 1 ∣ y , G , Y q t l , G q t l , α ^ ) and the corresponding RCPs for colocalization analysis ., ( As a notational footnote , conditional on α ^ , the SNP-level γi and di depend only on one relevant molecular phenotype and its corresponding genotypes rather than the full collection of the molecular phenotypes . We keep the current notation for the consistency of the presentation . ), The subsequent sections provide the statistical and computational details within each stage ., We implement the computational procedure outlined above in the software package enloc ( Enrichment estimation aided colocalization analysis ) , which is freely available at https://github . com/xqwen/integrative ., Note that the computational procedure requires only summary-level information from both the molecular QTL data and GWAS data ., The primary objective of the enrichment analysis is to estimate the hyper-parameter α given the observed summary statistics from GWAS and the fuzzy annotation of molecular QTLs ., Recall that if the binary molecular QTL annotation is indeed known , then the EM algorithm that we previously described 14 , 16 can be directly applied to obtain the maximum likelihood estimate of α ., With incomplete information on annotation data , we adopt a principled statistical strategy in missing data inference known as multiple imputation 17 , 18 ., Specifically , the multiple imputation procedure creates m complete data sets by filling in , i . e . , imputing , the missing entries of the binary annotation data ., The imputed data sets are then individually analyzed using the existing EM algorithm , and the distinct estimates of α ^ from multiple imputed data sets are combined into a final estimate using a set of rather simple rules ( section S . 1 in S1 Text ) ., The key to implementing this strategy is to impute the annotations , which , in our case , is achieved by sampling from the posterior distribution Pr ( d ∣ Y qtl , Gqtl ) ., According to the missing data theory , the ideal probability distribution to impute d is Pr ( d ∣ Y qtl , Gqtl , y , G ) , i . e . , the imputation of d should also be conditioned on the observed GWAS data ., The proposed imputation distribution represents a simplified approximation and essentially assumes the independence between d and GWAS data , which is because Pr ( d ∣ Y qtl , Gqtl ) = Pr ( d ∣ Y qtl , Gqtl , y , G ) if and only if α1 = 0 ., Consequently , imputing from this simplified distribution ( or more generally , imputing without the consideration of GWAS data ) leads to conservative point estimates that are shrunk toward 0 ., ( This is because each imputed data set is generated as if α1 is set to 0 a priori . ), In practice , the underestimation of the true α1 under the simplified imputation distribution can be noticeable if the true α1 is much larger than 0 ( which is evident in some of our simulation scenarios ) ., Despite this shortcoming , we choose to work with the simplified imputation distribution , Pr ( d ∣ Y qtl , Gqtl ) , mainly because of its attractive computational property ., For example , it can be obtained by a single run of genetic association analysis based solely on the molecular QTL data and applied in the integrative analysis of any GWAS data ., In comparison , Pr ( d ∣ Y qtl , Gqtl , y , G ) is specific to each GWAS-molecular QTL data set pair , and its computation is considerably more expensive if not practically impossible ., Importantly , the empirical evidence from the simulation studies suggests that the bias of the enrichment estimate due to the use of the simplified imputation distribution has non-significant impacts on the results of downstream fine-mapping and colocalization analyses ., The number of imputed data sets ( m ) necessary for reliable estimation has been systematically studied in the missing data theory ., The common consensus in the statistical literature is that m should be determined by the percentage of missingness , and various theoretical and empirical studies 19 , 20 roughly agree that 20 imputations are required for 10% to 30% missing information and that 40 imputations are required for 50% missing information ., Although the true annotation d is completely unobserved in our context , we are certain that di = 0 for the vast majority of the candidate SNPs based on inspection of the posterior distribution Pr ( d ∣ Y qtl , Gqtl ) ., In fact , by examining the analysis results of cis-eQTLs from the GTEx whole blood data , we find that there are only ∼ 1 . 5% cis candidate SNPs with a posterior inclusion probability ≥ 0 . 01 ., Guided by this empirical evidence , we choose to impute m = 25 QTL data sets for each analysis ., ( We have also experimented with 50 and more imputed data sets in the simulations , and the inference results are virtually unchanged . ), Additionally , we observed that detectable GWAS hits and eQTLs ( with currently available sample sizes ) are both relatively sparse in practice , which can lead to large variances for the estimated enrichment parameter α1 ., To illustrate this point , we consider that both γ and d are observed; it is then trivial to estimate α ^ 1 using a 2 × 2 contingency table ., Because each binary vector contains only very few non-zero entries , the resulting contingency table is extremely imbalanced ., Consequently , the variance of α ^ 1 ( approximately equal to the inverse of the smallest cell count ) can be large , and the point estimate can be unstable ., To stabilize the estimate of the enrichment parameter , we modify the original EM algorithm and apply an l2 penalty with a shrinkage parameter λ in the M-step to shrink the estimate toward 0 ., This strategy is informed by the statistical principle of “variance-bias trade-off” ., Alternatively , this can be viewed as assigning a N ( 0 , 1/λ ) prior to α1 ., In practice , we select λ in a data-driven manner by assessing the degree of imbalance of the unobserved contingency table ( section S . 2 of S1 Text ) , which assigns stronger penalties for larger degrees of imbalance ., Given the point estimate of the enrichment parameter , we adopt an empirical Bayes procedure to infer the true association status , γ , for all SNPs in GWAS ., Specifically , we compute Pr ( γ ∣ y , G , Y q t l , G q t l , α ^ ) as an approximation of the desired quantity Pr ( γ ∣ y , G , Y qtl , Gqtl ) 21 ., In addition , we apply the same divide-and-conquer strategy described in 14 by decomposing the genome into K non-overlapping LD blocks 15 and performing independent Bayesian fine-mapping analysis within each LD block ., Finally , we summarize the evidence of association for each SNP by its posterior inclusion probability ( PIP ) , i . e . , Pr ( γ i = 1 ∣ y , G , Y q t l , G q t l , α ^ ) ., To account for the uncertainty of the association status of molecular eQTLs , we construct a two-component mixture prior for each SNP , i . e . ,, Pr ( γ i = 1 ∣ Y q t l , G q t l , α ^ ) = e α ^ 0 1 + e α ^ 0 · ( 1 - δ i ) + e α ^ 0 + α ^ 1 1 + e α ^ 0 + α ^ 1 · δ i , ( 7 ), where δi ≔ Pr ( di = 1 ∣ Y qtl , Gqtl ) denotes the PIP of SNP i being a causal molecular QTN ., Because the vast majority of the LD blocks harbor no noteworthy association signals for any given complex trait , we follow the common practice in the GWAS analysis and adopt a pre-screening procedure to identify LD regions that are potentially interesting for fine-mapping analysis ., Specifically , we use a rigorous Bayesian false discovery rate ( FDR ) control procedure 22 to screen and select LD blocks for the subsequent fine-mapping analysis ., This procedure is typically less conservative ( and hence more powerful ) than the commonly applied empirical procedures based on the combination of single-SNP testing and the Bonferroni correction ., For each identified LD block , we then proceed to perform fine-mapping analysis using the DAP algorithm ., We find that the DAP-1 algorithm is practically adequate for fine-mapping most LD blocks in GWAS data , as we observe that the vast majority of the selected LD blocks harbor no more than a single association signal ., Even if multiple GWAS signals co-exist in a single LD block , the DAP-1 algorithm can still be applied when aided by the conditional analysis approach proposed by 23 ., Alternatively , the adaptive DAP algorithm , which enables fully automated multi-SNP analysis , can be conveniently applied in this context , even with summary-level statistics ( section S . 5 of S1 Text ) ., However , there is an increased computational cost ., Our simulation study shows that the adaptive DAP algorithm slightly outperforms the DAP-1 algorithm , which confirms the benefit of multi-SNP analysis ., Nevertheless , we conclude that the results obtained from the two variants of the DAP algorithm are quite comparable in our simulation studies using realistically generated GWAS data ., By default , in this paper , we apply the DAP-1 algorithm for the fine-mapping procedure , and we only re-examine the noticeable loci ( e . g . , those identified in the subsequent colocalization analysis ) using the adaptive DAP algorithm ., Given the PIP from the fine-mapping analysis , the SNP-level colocalization probability ( SCP ) for SNP i can be obtained as, Pr ( γ i = 1 , δ i = 1 ∣ y , G , Y q t l , G q t l , α ^ ) = Pr ( γ i = 1 ∣ y , G , Y q t l , G q t l , α ^ ) / 1 + 1 - δ i δ i · 1 + e α ^ 0 + α ^ 1 e α ^ 1 + e α ^ 0 + α ^ 1 ( 8 ), by solving a simple linear system ( section S . 3 of S1 Text ) ., Based on the discussion in the previous sections and following Gaun and Stephens 24 and Wen et al 16 , we propose computing a regional colocalization probability , or RCP , by summing up the SNP-level colocalization probabilities ( SCPs ) of correlated SNPs within an LD block that harbors a single GWAS association signal ., RCP is naturally interpreted as the probability of a genomic region harboring a colocalized signal ., We recommend reporting both RCPs and SCPs in colocalization analysis ., In practice , we only compute RCPs for the same LD blocks that are identified by the pre-screening step in the fine-mapping analysis ., The rationale is simple: we do not expect an LD block to harbor a colocalized signal if it is unlikely to harbor a GWAS signal ., To demonstrate , we apply Eq ( 8 ) in our previously stated hypothetical example of two perfectly linked candidate SNPs ., Under the assumption , it follows that at the SNP level , δ1 = δ2 = 0 . 5 and Pr ( γ 1 = 1 ∣ y , G , Y q t l , G q t l , α ^ ) = Pr ( γ 2 = 1 ∣ y , G , Y q t l , G q t l , α ^ ) = 0 ., 50 ., From Eq ( 8 ) , it is evident that the SCPs of the two SNPs are also identical with the actual value depending on α ^ 1: as α ^ 1 → 0 , both take a value of 0 . 25 ( hence , RCP = 0 . 50 ) , whereas when α ^ 1 → ∞ , both take a value of 0 . 50 ( hence , RCP = 1 . 0 ) ., More generally , we show the functional relationship between RCP and the α1 values in Fig 1 , which illustrates the quantitative impact of the enrichment estimation on the probabilistic assessment of colocalized signals ., This study uses third party datasets and no additional ethics approval was needed ., First , we perform simulation studies to benchmark the performances of the proposed enrichment and colocalization analysis approaches ., We design the simulation scheme to generate realistic single SNP association z-statistics that are similar to the observed GWAS results ., Specifically , we select real genotypes of 2 . 7 million overlapping SNPs used by both Wood et al 27 and the GTEx project from the European samples from the 1000 Genomes Project ., For each SNP , we obtain its binary eQTL annotation by drawing from the posterior distribution of GTEx whole blood cis-eQTLs the GTEx ., This particular posterior distribution is obtained by performing multi-SNP fine-mapping of the GTEx whole blood data via the adaptive DAP algorithm 14 ., In total , we roughly annotate ∼ 6 , 000 SNPs per simulation ., We then simulate the association status of each SNP i ( γi ) by drawing from a Bernoulli distribution whose success rate is determined by the logistic model ( 2 ) with pre-determined α0 and α1 values ., Subsequently , a quantitative trait is simulated using a standard multiple linear regression model for which the residual error variance is set to 1 , and the effect size of each causal SNP is drawn from a N ( 0 , ϕ2 ) distribution ., Finally , we compute the single SNP association z-statistic for each SNP as the input for both the enrichment and the colocalization analyses ., Although the sample size in the 1000 Genomes Project European panel is limited , we are able to adjust the values of α0 ( which determines the prevalence of the causal associations ) and ϕ ( which determines the signal-to-noise ratio of the genetic effects ) to roughly match the z-value distributions from the available large-scale GWAS meta-analysis ., In particular , we estimate α0 and ϕ by analyzing the height data reported in Wood et al 27 , and we set α0 = −8 . 4 and ϕ = 0 . 4 ., Consequently , the distributions of the simulated z-statistics closely resemble the actual observed GWAS height data ( S1 Fig ) ., We vary the value of α1 across simulations for different levels of enrichment ., To demonstrate the proposed computational approach in a practical setting , we perform an integrative analysis of the eQTL data from the GTEx project 1 and the blood lipid data originally reported in Teslovich et al 7 ., The blood lipid data consist of meta-analysis results of four quantitative traits , namely , low-density lipoprotein ( LDL ) cholesterol , high-density lipoprotein ( HDL ) cholesterol , triglycerides ( TG ) and total cholesterol ( TC ) , with an aggregated sample size of ∼ 100 , 000 ., We obtain the version of single-SNP association z-statistics for the four traits re-analyzed by Pickrell 28 , where additional z-statistics for untyped SNPs are imputed according to the 1000 Genomes Project phase I panel ., In total , the complete data set contains z-scores of ∼ 6 . 1 million SNPs per trait ., For most of our analysis , we focus on the cis-eQTL data from the whole blood in the recent release ( version 6 ) of the GTEx project ., The selection of the whole blood is informed by the consensus of multiple independent enrichment analysis approaches ( GTEx consortium , manuscript in prep . ) to determine the relevant tissues for the blood lipid traits ., In addition to biological relevance , we suspect that one of the driving factors is that the whole blood is one of the GTEx tissues with the largest sample size ( 338 ) in the current release of the data; it therefore has better power to detect cis-eQTLs with small to modest effects ., The SNPs that are not directly genotyped are also imputed according to the same 1000 Genomes panel by the GTEx consortium ., We perform the Bayesian multi-SNP fine-mapping analysis for the GTEx whole blood data using the adaptive DAP algorithm and generate the joint posterior distribution Pr ( d ∣ Y qtl , Gqtl ) while controlling for the SNP distance to the transcription start site ( TSS ) of the corresponding target gene ., As shown in our previous results 14 , 22 , this approach significantly improves the eQTL discovery ., In this paper , we have proposed a statistically rigorous and computationally efficient analytic framework for performing integrative analyses of GWAS and molecular QTL data and providing quantitative assessments of enrichment and colocalization of their association signals ., One of the intrinsic challenges in genetic association analysis is that the resolution of identified association signals is always limited by LD ., Consequently , it is generally impossible to pinpoint the causal variants based solely on genetic association analysis , and it imposes a formidable challenge for assessing enrichment and colocalization in the integrative analysis ., To address this problem , we formulate a missing data problem and adopt a well-established statistical strategy , i . e . , multiple imputation , to fully account for the uncertainty in identifying causal genetic variants for complex traits and molecular phenotypes due to LD ., These efforts result in not only more accurate point estimates but also appropriate characterizations of uncertainties of our inference results in the enrichment and colocalization analyses ., Particularly , in the colocalization analysis , our theoretical demonstration and the real data example both clearly illustrate that individual SCPs can be unimpressive in high LD regions even if we are confident that the region does harbor a colocalized signal ., In light of these findings , we propose and recommend reporting RCPs rather than placing emphasis on colocalization probabilities of individual SNPs ., Compared to the existing methods for colocalization analysis , the most important distinction of our proposed approach is the natural integration of the enrichment estimation ., Throughout the paper , we have illustrated the importance of obtaining accurate enrichment estimates on the downstream quantitative evaluations of colocalization ., Our main conclusion is that the accurate enrichment estimates based on currently available data may not have an overall large effect on altering the ranking of potential colocalization signals; however , it is critically important for the calibration of the corresponding colocalization probabilities and has a profound impact on the outcome of formal statist | Introduction, Method, Results, Discussion | We propose a novel statistical framework for integrating the result from molecular quantitative trait loci ( QTL ) mapping into genome-wide genetic association analysis of complex traits , with the primary objectives of quantitatively assessing the enrichment of the molecular QTLs in complex trait-associated genetic variants and the colocalizations of the two types of association signals ., We introduce a natural Bayesian hierarchical model that treats the latent association status of molecular QTLs as SNP-level annotations for candidate SNPs of complex traits ., We detail a computational procedure to seamlessly perform enrichment , fine-mapping and colocalization analyses , which is a distinct feature compared to the existing colocalization analysis procedures in the literature ., The proposed approach is computationally efficient and requires only summary-level statistics ., We evaluate and demonstrate the proposed computational approach through extensive simulation studies and analyses of blood lipid data and the whole blood eQTL data from the GTEx project ., In addition , a useful utility from our proposed method enables the computation of expected colocalization signals using simple characteristics of the association data ., Using this utility , we further illustrate the importance of enrichment analysis on the ability to discover colocalized signals and the potential limitations of currently available molecular QTL data ., The software pipeline that implements the proposed computation procedures , enloc , is freely available at https://github . com/xqwen/integrative . | Genome-wide association studies ( GWAS ) have been tremendously successful in identifying genetic variants that impact complex diseases ., However , the roles of such studies in disease etiology remain poorly understood , primarily because a large proportion of the GWAS findings are located in the non-coding region of the genome ., Recent advancements in high-throughput sequencing technology enable the systematic investigation of molecular quantitative trait loci ( QTLs ) , which are genetic variants that directly affect molecular phenotypes ( e . g . , gene expression , transcription factor binding and DNA methylation ) ., Linking molecular QTLs to GWAS findings intuitively represents an important step for interpreting the biological and clinical relevance of the GWAS results ., In this paper , we describe a rigorous and efficient computational approach that assesses the enrichment and overlap between the GWAS findings and molecular QTLs ., Importantly , we illustrate that the accurate quantification of overlapping between molecular QTL and GWAS signals requires reliable enrichment estimation ., Our proposed approach fully accounts for the intrinsic uncertainty embedded in the association analyses of GWAS and molecular QTL mapping , and it outperforms the existing state-of-the-art approaches ., Applying the proposed approach to the GWAS data of blood lipid traits and the whole blood expression QTLs ( eQTLs ) yields some novel biological insights and also illustrates the potential limitations of the currently available molecular QTL data . | genome-wide association studies, medicine and health sciences, body fluids, quantitative trait loci, applied mathematics, simulation and modeling, algorithms, mathematics, genome analysis, lipid analysis, research and analysis methods, genomic signal processing, molecular biology, genetic loci, signal transduction, macromolecular structure analysis, blood, anatomy, cell biology, heredity, physiology, genetics, biology and life sciences, physical sciences, genomics, cell signaling, computational biology, complex traits, human genetics | null |
journal.pcbi.1004465 | 2,015 | Network-Based Isoform Quantification with RNA-Seq Data for Cancer Transcriptome Analysis | Application of next generation sequencing technologies to mRNA sequencing ( RNA-Seq ) is a widely used approach in transcriptome study 1–3 ., Compared with microarray technologies , RNA-Seq provides information for expression analysis at transcript level and avoids the limitations of cross-hybridization and restricted range of the measured expression levels ., Thus , RNA-Seq is particularly useful for quantification of isoform transcript expressions and identification of novel isoforms ., Accurate RNA-Seq-based transcript quantification is a crucial step in other downstream transcriptome analyses such as isoform function prediction in the pioneer work in 4 , and differential gene expression analysis 5 or transcript expression analysis 6 ., Detecting biomarkers from transcript quantifications by RNA-Seq is also a frequent common practice in biomedical research ., However , transcript quantification is challenging since a variety of systematical sampling biases have been observed in RNA-Seq data as a result of library preparation protocols 7–10 ., Moreover , in the aligned RNA-Seq short reads , most reads mapped to a gene are potentially originated by more than one transcript ., The ambiguous mapping could result in hardly identifiable patterns of transcript variants 10 , 11 ., A useful prior knowledge that has been largely ignored in RNA-Seq transcriptome quantification is the relation among the isoform transcripts by the interactions between their protein products ., The protein products of different isoforms coded by the same gene may contain different domains interacting with the protein products of the transcripts in other genes ., Previous studies suggested that alternative splicing events tend to insert or delete complete protein domains/functional motifs 12 to mediate key linkages in protein interaction networks by removal of protein domain-domain interactions 13 ., The work in 4 , 14 also suggested unique patterns in isoform co-expressions ., Thus , the abundance of an isoform transcript in a gene can significantly impact the quantification of the transcripts in other genes when their protein products interact with each other to accomplish a common function as illustrated by a real subnetwork in Fig 1 , which is constructed based on domain-domain interaction databases 15 , 16 and Pfam 17 ., Motivated by our observation that the protein products of highly co-expressed transcripts are more likely to interact with each other by protein domain-domain binding in four TCGA RNA-Seq datasets ( see the section Results ) , we constructed two human transcript interaction networks of different sizes based on protein domain-domain interactions to improve transcript quantification ., Based on the constructed transcript network , we propose a network-based transcript quantification model called Net-RSTQ to explore domain-domain interaction information for estimating transcript abundance ., In the Net-RSTQ model , Dirichlet prior representing prior information in the transcript interaction network is introduced into the likelihood function of observing the short read alignments ., The new likelihood function of Net-RSTQ can be alternating-optimized over each gene with expectation maximization ( EM ) ., It is important to note that the Dirichlet prior from the neighboring isoforms play two possible roles ., On one hand , for the isoforms in the same gene but with different interacting partners , the different prior information will help differentiate their expressions to reflect their different functional roles ., On the other hand , for the isoforms in the same gene with the same interacting partners , the uniform prior assumes no difference in their functional roles and thus , promotes a smoother expression patterns across the isoforms ., In both cases , the Dirichlet prior captures the functional variations/similarities across the isoforms in each gene as prior information for estimation of their abundance ., The paper is organized as following ., In the section Materials and Methods , we describe the procedure to construct protein domain-domain interaction networks , the mathematic description of the probabilistic model and the Net-RSTQ algorithm , qRT-PCR experiment design , and RNA-Seq data preparation ., In the section Results , we first demonstrate the correlation between protein domain-domain interactions and isoform transcript co-expressions across samples in four cancer RNA-Seq datasets from The Cancer Genome Atlas ( TCGA ) to justify using domain-domain interactions as prior knowledge ., We then compared the predicted isoform proportions with qRT-PCR experiments on 25 multi-isoform genes in three cell lines , H9 stem cell line , OVCAR8 ovarian cancer cell line and MCF7 breast cancer cell line ., Net-RSTQ was also applied to four cancer RNA-Seq datasets to quantify isoform expressions to classify patient samples by the survival or relapse outcomes ., In addition , simulations were also performed to measure the statistical robustness of Net-RSTQ over randomized networks ., Two binary transcript networks were constructed by measuring the protein domain-domain interactions ( DDI ) between the domains in each pair of transcripts in four steps ., First , the translated transcript sequences of all human genes were obtained from RefSeq 18 ., Second , Pfam-Scan was used to search Pfam databases for the matched Pfam domains on each transcript with 1e-5 e-value cutoff 17 ., Note that only high quality , manually curated Pfam-A entries in the database were used in the search ., Third , domain-domain interactions were obtained from several domain-domain interaction databases , and if any domain-domain interaction exists between a pair of transcripts , the two transcripts are connected in the transcript network ., Specifically , 6634 interactions between 4346 Pfam domain families from two 3D structure-based DDI datasets ( iPfam 15 and 3did 16 ) inferred from the protein structures in Protein Data Bank ( PDB ) 19 were used in the experiments ., Besides these highly confident structure-based DDIs , transcript interactions constructed from 2989 predicted high-confidence DDIs and 2537 predicted medium-confidence DDIs in DOMINE 20 were also included if the transcript interaction agrees with protein-protein interactions ( PPI ) in HPRD 21 ., In the experiments , we focused on the transcripts from two cancer gene lists from the literature for better reliability in annotations ., The first smaller transcript network consists of 11736 interactions constructed from the 3D structure-based DDIs and 421 interactions constructed from the predicted DDIs among the 898 transcripts in 397 genes from the first gene list 22 ., The second larger transcript network contains 711 , 516 interactions constructed from the 3D structure-based DDIs among 5599 transcripts in 2551 genes in a larger gene list 23 ., Since inclusion of the predicted DDIs results in a much higher density in the large network , the large network does not include predicted DDIs to prevent too many potential false positive interactions ., The characteristics of the two transcript networks are summarized in Table 2 . The density of the two networks are 3 . 02% and 4 . 54% respectively , which are in similar scale with the PPI network ., Both networks show high clustering coefficients , suggesting modularity of subnetworks ., Note that self-interactions ( interactions between transcript, ( s ) in the same gene ) are not considered since Net-RSTQ only utilizes positive correlation between the expressions of neighboring transcripts in different genes ., For simplicity , Net-RSTQ assumes that self-interactions will not change the transcript quantification of an individual gene in the model ., In Fig 1 ( A ) a subnetwork of the transcripts in gene CD79A and CD79B with their direct neighbors in the small transcript network is shown ., The RefSeq transcript annotations of CD79A and CD79B are shown in Fig 1 ( B ) ., In CD79A transcript NM_001783 contains an extra domain pfam07686 while transcript NM_021601 only contains a shorter hit pfam02189 ., Note that pfam02189 also has the same hit in NM_001783 with an e-value larger than 1e-5 ., In CD79B transcripts NM_001039933 and NM_000626 contain a domain pfam07686 , which is removed in alternative splicing of NM_021602 ., In the transcript subnetwork shown in Fig 1 ( A ) , the transcripts in CD79A or CD79B have different interaction partners in the network ., In the transcripts in CD79A , the expression of NM_021601 will correlate with the transcripts in LCK and SYK , and NM_001783 will correlate with two transcripts in CD79B ., The isoform transcripts in LCK and SYK show no different DDIs suggesting there is no functional variation by protein bindings and more similar expression patterns are potentially expected as prior knowledge ., We first consider the method proposed in 24 , 25 as the base model for quantification of the transcripts in a single gene ., Let Ti denote the set of the transcripts in the ith gene and Tik be the kth transcript in Ti ., The probability of a read being generated by the transcripts in Ti is modeled by a categorical distribution specified by parameters pik , where ∑ k = 1 | T i | p i k = 1 and 0 ≤ pik ≤ 1 . For the set of the reads ri aligned to gene i , we consider the likelihood of that each of the |ri| short reads is sampled from one of the transcripts to which the read aligns ., Specifically , for each read rij aligned to transcript Tik , the probability of obtaining rij by sampling from Tik , namely Pr ( rij|Tik ) is q i j k = 1 l i k - l r + 18 , 26 , 27 , where lr is the length of the read ., Assuming each read is independently sampled from one transcript , the uncommitted likelihood function 24 to estimate the parameters Pi from the observed read alignments against gene i is, L ( P i ; r i ) = P r ( r i | P i ) = ∏ j = 1 | r i | P r ( r i j | P i ) = ∏ j = 1 | r i | ∑ k = 1 | T i | P r ( T i k | P i ) P r ( r i j | T i k ) = ∏ j = 1 | r i | ∑ k = 1 | T i | p i k q i j k ., ( 1 ), This likelihood function is concave but it may contain plateau in the likelihood surface ., Therefore , Expectation Maximization ( EM ) is then applied to obtain the optimal Pi ., In the EM algorithm , the expectation of read assignments to transcripts were estimated in the E-step and the likelihood function with the expected assignments can be maximized in the M-step to estimate Pi ., The relative abundance of the transcript Tik in gene i , ρik , can be derived from, ρ i k = p i k l i k ∑ k = 1 | T i | p i k l i k , ( 2 ), and the transcript expressions in gene i , πik , can be calculated by, π i k = | r i | p i k l i k ., ( 3 ), The base model is applied independently to each individual gene and no relation among the transcripts is considered ., In the Net-RSTQ model , the transcript interaction network S based on protein domain-domain interactions is introduced to calculate a prior distribution for estimating P jointly across all the genes and all the transcripts ., The model assumes that the prior distribution of Pi is a Dirichlet distribution specified by parameters αi and each αik is proportional to the read count by average expression of the transcript Tik’s neighbors in the transcript network S . The prior read count ϕik is defined as follows ,, ϕ i k = l i k ( π ′ S * , ( i , k ) ∑ ( S * , ( i , k ) ) ) , ( 4 ), where S* , ( i , k ) is a binary vector represents the neighborhood of transcript Tik in transcript network S and ∑ ( S* , ( i , k ) ) is the size of the neighborhood ., The calculation of each ϕik is illustrated in Fig 2 . The Dirichlet parameter αi is defined as a function of ϕik as, α i k = λ ϕ i k + 1 , ( 5 ), where λ > 0 is a tuning parameter balancing the belief between the prior-read count and the aligned-read count ., To obtain the optimal P jointly for all genes , we introduce a pseudo-likelihood model to estimate P iteratively in each iteration ., Assuming uniform Pr ( ri ) , the pseudo-likelihood function is defined as ,, L ( P , α ; r ) = ∏ i = 1 N L ( P i , α i ; r i ) = ∏ i = 1 N P r ( P i | α i ) P r ( r i | P i ) P r ( r i ) ∝ ∏ i = 1 N P r ( P i | α i ) P r ( r i | P i ) ., ( 6 ), Note that the pseudo-likelihood model relies on the independence assumption among the likelihood functions of each individual gene when the α parameters of the Dirichlet priors are pre-computed ., Thus , the model simply takes the product of the likelihood function from each gene ., Each prior distribution Pr ( Pi|αi ) follows the Dirichlet distribution ,, P r ( P i | α i ) = C ( α i ) ∏ k = 1 | T i | p i k α i k - 1 , where C ( α i ) = Γ ( ∑ k α i k ) ∏ k Γ ( α i k ) ., ( 7 ) Integrating eqs ( 1 ) and ( 7 ) , the pseudo-likelihood function in eq ( 6 ) can be rewritten with Dirichlet prior as, L ( P ; r ) = ∏ i = 1 N C ( α i ) ∏ k = 1 | T i | p i k α i k - 1 ∏ j = 1 | r i | ∑ k = 1 | T i | p i k q i j k = ∏ i = 1 N C ( λ ϕ i + 1 ) ∏ k = 1 | T i | p i k λ ϕ i k ∏ j = 1 | r i | ∑ k = 1 | T i | p i k q i j k ., ( 8 ) In the pseudo-likelihood function in eq ( 8 ) , the only hyper-parameter λ balances the proportion between the Dirichlet priors and the observed read counts of each transcript ., The larger the λ , the more belief put on the priors ., The Net-RSTQ algorithm optimizes eq ( 8 ) by dividing the optimization into sub-optimization problems of sequentially estimating each Pi ., Specifically , we fix all Pc , c ≠ i , and thus ϕi when estimating Pi with EM in each iteration and repeat the process multiple rounds throughout all the genes ., In each step , the neighborhood expression ϕ is recomputed with new Pi for computing the quantification of the next gene ., For each sub-optimization problem , we estimate Pi with a fixed ϕ , the part of the likelihood function in eq ( 8 ) involved with the current variables Pi is, L ¯ ( P i ; r i ) = ∏ g ∈ n b ( i ) C ( λ ϕ g + 1 ) ∏ k = 1 | T g | p g k λ ϕ g k C ( λ ϕ i + 1 ) ∏ k = 1 | T i | p i k λ ϕ i k ∏ j = 1 | r i | ∑ k = 1 | T i | p i k q i j k , ( 9 ), where nb ( i ) is the set of the genes containing transcripts that are neighbors of the transcripts in gene i in the transcript network ., Eq ( 9 ) consists of three terms separated by the braces ., The second and the third terms are the Dirichlet prior and the likelihood of the observed counts in the data for gene i ., The first term is the Dirichlet priors of the neighbor transcripts of each Tik ., These prior probabilities are involved since ϕg are functions of the current variable Pi ( eqs ( 3 ) – ( 5 ) ) ., Eq ( 9 ) cannot be easily solved with standard techniques ., We adopt a heuristic approach to only take steps that will increase the whole pseudo-likelihood function in eq ( 8 ) ., The Net-RSTQ algorithm is outlined below Algorithm 1 Net-RSTQ 1: Initialization: random initialization or base EM ( eq ( 1 ) ) estimation of P ( 0 ) 2: for round t = 1 , … do 3: P ( t ) = P ( t − 1 ) 4: for gene i = 1 , … , N do 5: compute ϕi based on P ( t ) with eqs ( 3 ) and ( 4 ) 6: estimate Pi with EM algorithm ( see next section ) 7: if L ¯ ( P i ) > L ¯ ( P i ( t ) ) then 8: P i ( t ) = P i 9: end if 10: end for 11: if max ( abs ( P ( t ) − P ( t − 1 ) ) ) <1e-6 then 12: break 13: end if 14: end for 15: return P In the algorithm , the outer for-loop between line 2–14 performs multiple passes of updating P . The inner for-loop between line 4–10 scans through each gene to update each Pi ., Line 7 checks the the difference in the likelihood L ¯ of gene i before and after the estimated Pi is applied ., The newly estimated Pi is kept in line 8 only if the likelihood L ¯ in eq ( 9 ) is higher ., The convergence of P is checked at line 11 ., In each sub-optimization problem , EM algorithm ( described in the next section ) is applied to estimate Pi ., After convergence , the transcripts expression π can be learned by eq ( 3 ) with the optimal P . In line 6 of Algorithm 1 , we maximize the likelihood function of the sub-optimization problem in eq ( 9 ) to learn Pi as, L ( P i ; r i ) = C ( λ ϕ i + 1 ) ∏ k = 1 | T i | p i k λ ϕ i k ∏ j = 1 | r i | ∑ k = 1 | T i | p i k q i j k ., ( 10 ), Note that eq ( 10 ) is the part of eq ( 9 ) without the Dirichlet priors of the neighboring genes ., In line 7 of Algorithm 1 , the ignored Dirichlet priors are combined with the likelihood in eq ( 10 ) , when L ¯ ( P i ) is computed , to evaluate the whole likelihood in eq ( 9 ) ., The likelihood function in eq ( 10 ) is defined on a categorical variable with Dirichlet prior , which can be solved with EM algorithm ., Following EM formulation in 26 , the expectation aijk , a soft assignment of read j to transcript k in gene i , is first estimated in the expectation step and Pi is then learned in the maximization step ., When ϕi is given , by taking log of eq ( 10 ) we can write the EM steps to find Pi below ., Three qRT-PCR experiments are designed to measure the isoform proportions of 25 multi-isoform genes in three cell lines , H9 stem cell line , OVCAR8 ovarian cancer cell line and MCF7 breast cancer cell line ., The cell lines were selected based on the available of both RNA-Seq data and cell culture in our labs ., The qRT-PCR experiments focused on the gene with most different quantification results reported by Net-RSTQ and other compared methods ., Due to the limitations in time and cost of running qRT-PCR experiments , only the 25 genes in the three cell lines were tested with all the results reported in the experiments ., Quantitation of the real-time PCR results was done on the data from H9 human embryonic stem cells to obtain the absolute expressions for comparing more than two transcripts and comparative Ct method was done on the data from OVCAR8 ovarian cancer cells and MCF7 breast cancer cells to obtain the ratio between a pair of transcripts ., Three cell line RNA-Seq datasets were used for evaluating the accuracy of transcript quantification by comparison with qRT-PCR results ., The first dataset is the H9 embryonic stem cell line data from 28 , downloaded from SRA ., The second dataset is an in-house dataset from the ovarian cancer cell line OVCAR8 prepared at University of Kansas Medical Center ., The third dataset is the MCF7 breast cancer cell line data from 29 , downloaded from SRA ., There are 23 , 397 , 325 single-end 34bp reads in the stem cell line dataset , 19 , 892 , 473 paired-end 100bp reads in the OVCAR8 , and 21 , 855 , 632 paired-end 76bp reads in the MCF7 mapped to the human hg19 reference genome by TopHat2 . 0 . 9 30 with up to 2 mismatches allowed ., Exon coverages and read counts of exon-exon junctions were generated by SAMtools 31 to be utilized with Net-RSTQ and base EM ( eq ( 1 ) ) ., Cufflinks 32 directly infers transcript expressions based on the alignment by TopHat with the min isoform fraction set to 0 for better sensitivity ., TCGA RNA-Seq datasets of Ovarian serous cystadenocarcinoma ( OV ) , Breast invasive carcinoma ( BRCA ) , Lung adenocarcinoma ( LUAD ) and Lung squamous cell carcinoma ( LUSC ) were analyzed for patient outcome prediction with transcript expressions estimated by Net-RSTQ , base EM ( eq ( 1 ) ) , RSEM 33 and Cufflinks 32 ., Both the gene expression and transcript expression data reported by RSEM 33 in TCGA ( level 3 data ) were utilized as two baselines for cancer outcome prediction ., The raw RNA-Seq fastq files ( level 1 data ) were downloaded from Cancer Genomics Hub ( CGHub ) and processed by TopHat for use with Net-RSTQ , base EM and Cufflinks ., The patient samples in each dataset were classified into cases and controls based on the survival and relapse outcomes as shown in Table 3 ., The command lines for preparing the data with RSEM and Cufflinks are available in the S3 Text ., To investigate the correlation between protein domain-domain interactions and isofrom transcript co-expressions , we calculated the number of transcript pairs that are both nearby ( being neighbors or having a distance up to 2 ) in the transcript network and highly co-expressed in the TCGA samples ., The transcript co-expressions were calculated by Pearson’s correlation coefficients of each pair of transcripts across all the samples in each dataset with the isoform transcript quantification by Cufflinks ., The transcript pairs were then sorted by the correlation coefficients from the largest to the smallest and grouped into bins of size 1000 ., The number of transcript pairs that are nearby in the transcript networks out of 1000 pairs are calculated within each bin and plotted in Fig 3 ( A ) and 3 ( B ) for the two cancer gene lists , respectively ., In both Fig 3 ( A ) and 3 ( B ) , the left column shows the plots of the number of pairs that are neighbors in the transcript network , and the right column shows the plots of the number of transcript pairs with a distance up to 2 in the transcript network , among the 1000 pairs in each bin ., In all the plots , similar trends are observed in all the four cancer datasets: there are more interacting isoform pairs in the bins with higher co-expressions ., For example , among the 1000 transcript pairs with the highest correlation coefficients , there are 73 interactions in the transcript network in OV dataset and thus , 73 interactions ( y-axis ) for bin index 1 ( x-axis ) is plotted in the left column of Fig 3 ( A ) ., In all the plots , there is a clear pattern that the numbers of matched nearby transcripts in the transcript network among the 1000 pairs in the first few bins are higher than the expected average of 30 in the small network of density 3 . 02% , 114 in the small network of density 11 . 41% ( with distance up to 2 ) , 45 in the larger network of density 4 . 54% , and 203 in the larger network of density 20 . 33% ( with distance up to 2 ) ., Moreover , the 2-step walk clearly promoted the number of overlaps with the pairs of higher co-expressions in the small network ., For example , the significant overlap is extended from the first 25 bins to approximately the first 50 bins or more in the four datasets ., The observation suggests that higher co-expressions exist not only in the direct neighbors in the transcript network but also the nearby nodes by a small distance ., By exploring the network structure with prior information through neighbors by many steps in iterations , Net-RSTQ model is expected to propagate the expression values from each transcript to its nearby nodes in the network to capture the co-expressions ., Note that considering the neighboring pairs with distance up to 2 in the larger network will result in a graph of density 20 . 33% , which is likely to contain too many false relations by the two-step walk ., Thus , the plots of the larger network of distance-2 pairs are only included for the completeness of the analysis ., The canonical 2x2 chi-square test was also applied to compare the number of the domain-domain interactions in the first 10 , 000 transcript pairs ( first 10 bins ) with the number in the rest of the pairs ., In all the four datasets in both Fig 3 ( A ) and 3 ( B ) with one exception in the LUSC dataset on the large network of distance-2 relation , there is a significant difference that the highly co-expressed transcripts are more likely to interact with each other in the transcript network , confirmed by the significant p-values ., As explained previously , the exception is likely due to the large number of false-positive pairs in the dense network ., The observation further support the hypothesis that protein domain-domain interactions correlate transcript co-expressions reported in previous studies 12 , 13 ., To further understand the specificity of the domain-domain interactions in the highly co-expressed transcripts , we calculated the number of domain-domain pairs that construct the DDIs in the top 10 , 000 co-expressed transcript pairs ., The statistics suggest high diversity of the type of DDIs ., For example , there are 547 interacting transcript pairs among the 201 out of 898 transcripts in the top 10 , 000 co-expressed transcript pairs in OV dataset for small network ., The 547 interacting transcript pairs represent 770 different domain-domain interactions ( There might be more than one DDIs between a pair of transcripts ) ., There are 739 interacting transcript pairs among the 538 out of 5599 transcripts in the top 10 , 000 co-expressed transcript pairs in OV dataset for large network ., The 739 interacting transcript pairs represent 1277 different domain-domain interactions ., The statistics suggest that the correlation between protein domain-domain interactions and transcript co-expressions is not a bias due to a few highly spurious DDIs ., It is a general correlation in many different DDIs and co-expressed transcripts ., Very similar statistics were observed in all the datasets and both networks ., To further demonstrate the co-expression relations in the transcript network , two examples are shown in S1 Fig . In S1 ( A ) Fig , WHSC1L1 contains two isoforms connected with different interactions in the transcript network ., Isoform NM_017778 interacts with 12 transcripts with average correlation coefficients 0 . 22 and the other isoform NM_023034 interacts with 13 more transcripts with average correlation coefficients 0 . 30 compared with the average correlation coefficient 0 . 188 against the other unconnected isoforms across the samples in the OV dataset ., In S1 ( B ) Fig , gene BRD4 contains two isoforms both of which are connected with the same 14 neighbors in the network ., The average correlation coefficients between these two isoforms and the 14 neighboring isoforms are both above 0 . 26 compared with the average correlation coefficient less than 0 . 15 against the other unconnected isoforms across the samples on the BRCA dataset ., In both examples , we observed high degree of agreement between co-expressions and DDIs ., To further understand the transcript networks , we overlapped the DDIs between genes in the two networks with the 294 human KEGG pathways 34 ., Among the 397 genes in the small network , 10 . 97% ( 17284 ) of the pairs are co-members in at least one KEGG pathway ., The 10 . 97% KEGG co-member pairs covers 42 . 70% ( 2122 ) of the DDIs among the genes while the other 89 . 03% ( 140352 ) non-co-member pairs covers 57 . 30% ( 2748 ) of the DDIs ., By these numbers , there is about 6-fold enrichment of DDIs in the KEGG co-member genes in the small network ., Among the 2551 genes in the large network , the 5 . 15% ( 335372 ) KEGG co-member pairs covers 12 . 45% ( 40812 ) of the DDIs among genes while the other 94 . 85% ( 6172229 ) non-co-member pairs covers 87 . 55% ( 287090 ) of the DDIs ., By these numbers , there is about 2 . 6-fold enrichment of DDIs in the KEGG co-member genes in the large network ., We also list the KEGG pathways that are highly enriched with DDIs in the large network in S4 Table ., Specifically , we consider the subnetwork of genes that are members of one KEGG pathway and calculated the density of DDIs in the subnetwork to compare to the overall density of 5 . 04% in the whole network ., Interestingly , most of the enriched pathways are signaling pathways and disease pathways with very high DDI densities ., In the simulations , we applied flux-simulator 35 to generate paired-end short reads simulating real RNA-Seq experiment in silico based on a ground truth transcript expression profile , using hg19 reference human genome and RefSeq annotations downloaded from UCSC Genome Browser ., To generate the ground-truth expression profiles , the gene expressions were sampled from a poisson distribution and the proportions of the isoforms in each gene were derived based on a neighbor average expression in the small transcript network and an initial mixed power law expression profile with gaussian noise ., A sequential updating was used to compute the proportion of each isoform by adding the neighbors’ average expressions to the initial expression ., The update procedure can be found in the S2 Text ., At last , flux-simulator was applied to simulate the short reads based on the ground truth transcript expression file ., 15 million 76-bp paired reads were generated by Flux Simulator and mapped to the reference genome by TopHat 30 with up to two mismatches allowed ., To account for the large dynamic range of abundances , the expressions were normalized by log2 ( expression+1 ) ., The correlation coefficients between the transcript abundances estimated by Net-RSTQ under various λ , base EM ( eq ( 1 ) ) , Cufflinks and RSEM , and the ground truth transcript abundances are reported in Fig, 4 . Furthermore , Net-RSTQ was also tested with 100 randomized networks with permuted indexes of transcripts in the transcript network ., To assess the impact of the network prior , two cases are shown ., Fig 4 ( A ) reports the correlation between the transcripts in which isoforms coded by the same gene are connected with different neighbors ( 109 out of 898 transcripts in 29 genes ) ., Fig 4 ( B ) reports the results from all the genes with more than one isoform ( 712 out of 898 transcripts in 211 genes ) ., In both comparisons , the transcript expressions estimated by Net-RSTQ achieve higher correlation with the ground truth compared with base EM , Cufflinks and RSEM ., Slightly higher improvement was observed in the first case than in the second case since the network prior plays more significant role in differentiating the isoform expressions by their different neighbors ., When randomized networks are used , Net-RSTQ leads to similar or worse results due to the wrong prior information ., Note that since the datasets were generated to partially conform to the network prior , the isoform expressions are relatively “smooth” among the neighboring isoforms ., Net-RSTQ tends to generate smoother expressions than base EM , Cufflinks and RSEM ., When applying Net-RSTQ with small λs and randomized network priors , slight improvement was also observed due to the smoothness assumption on the data ., To evaluate the effect of missing edges in the transcript network due to the undetected protein domain-domain interactions , we randomly removed certain percentages of the edges in the transcript network and then run Net-RSTQ with λ = 0 . 1 on the incomplete networks ., The results are shown in Fig 4 ( C ) and 4 ( D ) for the 109 transcripts with different neighbors and the 712 transcripts in the gene with more than one transcript , respectively ., It is intriguing to observe that only when a large percentage of the edges are removed , the performance of Net-RSTQ is affected ., Intuitively , the observation can be explained by the fact that the Dirichlet prior parameter is proportional to the average of the neighbors’ expressions ., As long as some of the neighbors are still connected to the target transcript in the network , the prior information is still useful ., The result suggests that Net-RSTQ is relatively robust to utilize transcript networks potentially constructed with a large percentage of undetected protein domain-domain interactions ., The isoform proportions estimated by Net-RSTQ , base EM , RSEM , and Cufflinks were compared to the qRT-PCR results on the three cell lines ., Parameter λ = 0 . 1 was fixed in all the Net-RSTQ experiments ., Among the genes that Net-RSTQ , base EM , RSEM , and Cufflinks report most different quantification results , qRT-PCR experiments were performed to test the genes with relatively higher coverage of RNA-Seq data , coding two to three isoforms , and the feasibility of designing isoform-specific primers in the qRT-PCR products ( see S1 , S2 and S3 Tables ) ., Twenty-five genes in total were tested in the three cell lines: seven in H9 stem cell line , five in OVCAR8 ovarian cancer cell line , and thirteen in MCF7 breast cancer cell line ., The scatter plots of the relative abundance of the first transcript in each gene estimated by Net-RSTQ , base EM , Cufflinks and RSEM were compared to the qRT-PCR results in Fig 5 ( A ) and 5 ( E ) ., In the scatter plot , the estimated relative abundance by Net-RSTQ were closer to qRT-PCR results measured by the accuracy of various thresholds and Root Mean Square Errors ., Ne | Introduction, Materials and Methods, Results, Discussion | High-throughput mRNA sequencing ( RNA-Seq ) is widely used for transcript quantification of gene isoforms ., Since RNA-Seq data alone is often not sufficient to accurately identify the read origins from the isoforms for quantification , we propose to explore protein domain-domain interactions as prior knowledge for integrative analysis with RNA-Seq data ., We introduce a Network-based method for RNA-Seq-based Transcript Quantification ( Net-RSTQ ) to integrate protein domain-domain interaction network with short read alignments for transcript abundance estimation ., Based on our observation that the abundances of the neighboring isoforms by domain-domain interactions in the network are positively correlated , Net-RSTQ models the expression of the neighboring transcripts as Dirichlet priors on the likelihood of the observed read alignments against the transcripts in one gene ., The transcript abundances of all the genes are then jointly estimated with alternating optimization of multiple EM problems ., In simulation Net-RSTQ effectively improved isoform transcript quantifications when isoform co-expressions correlate with their interactions ., qRT-PCR results on 25 multi-isoform genes in a stem cell line , an ovarian cancer cell line , and a breast cancer cell line also showed that Net-RSTQ estimated more consistent isoform proportions with RNA-Seq data ., In the experiments on the RNA-Seq data in The Cancer Genome Atlas ( TCGA ) , the transcript abundances estimated by Net-RSTQ are more informative for patient sample classification of ovarian cancer , breast cancer and lung cancer ., All experimental results collectively support that Net-RSTQ is a promising approach for isoform quantification ., Net-RSTQ toolbox is available at http://compbio . cs . umn . edu/Net-RSTQ/ . | New sequencing technologies for transcriptome-wide profiling of RNAs have greatly promoted the interest in isoform-based functional characterizations of a cellular system ., Elucidation of gene expressions at the isoform resolution could lead to new molecular mechanisms such as gene-regulations and alternative splicings , and potentially better molecular signals for phenotype predictions ., However , it could be overly optimistic to derive the proportion of the isoforms of a gene solely based on short read alignments ., Inherently , systematical sampling biases from RNA library preparation and ambiguity of read origins in overlapping isoforms pose a problem in reliability ., The work in this paper exams the possibility of using protein domain-domain interactions as prior knowledge in isoform transcript quantification ., We first made the observation that protein domain-domain interactions positively correlate with isoform co-expressions in TCGA data and then designed a probabilistic EM approach to integrate domain-domain interactions with short read alignments for estimation of isoform proportions ., Validated by qRT-PCR experiments on three cell lines , simulations and classifications of TCGA patient samples in several cancer types , Net-RSTQ is proven a useful tool for isoform-based analysis in functional genomes and systems biology . | null | null |
journal.ppat.1000659 | 2,009 | Microglia Are Mediators of Borrelia burgdorferi–Induced Apoptosis in SH-SY5Y Neuronal Cells | Lyme borreliosis is the most prevalent vector-borne illness in the northern hemisphere 1 , 2 ., Transmission from the animal reservoir to the human host occurs via an Ixodes tick bite into the skin , where B . burgdorferi , the spirochete that causes Lyme disease , can then disseminate hematogenously to various organs , including the heart , joints and both the peripheral and central nervous systems 2–4 ., A prominent , recurring and yet only partially understood feature of Lyme disease is the presence of inflammatory infiltrates within infected tissues 1–5 ., Evidence of neurological involvement occurs to varying degrees in both the central and peripheral nervous systems of Lyme disease patients , and inflammation is often associated with the neurological manifestations that define neuroborreliosis 2 , 5–8 , 9 , 10 ., Upon gaining access to the central nervous system ( CNS ) , B . burgdorferi may induce cerebrospinal fluid pleocytosis , meningoradiculitis and cranial neuritis as well as encephalopathies with neurocognitive abnormalities 3 , 5 , 6 , 11 ., This complex process of B . burgdorferi-induced pathology in the CNS has many aspects as yet to be clarified , with reactive inflammation potentially being one of the principal contributors to neuronal dysfunction ., Although B . burgdorferi lacks lipopolysaccharide ( LPS ) , the organism does produce spirochetal lipoproteins that can induce inflammation 11–17 ., By signaling through CD14 , binding Toll-like receptors ( TLR ) 2 and 1 and subsequent activation of NFκβ , lipoproteins have been shown to generate inflammatory mediators in a variety of cell types 14 , 18–21 ., Recent studies with TLR2 and/or MyD88-deficient mice in which B . burgdorferi-induced inflammatory infiltrates were greater than that of wild type controls , indicated however , that there are alternative pathways regulating the inflammatory response to infection with B . burgdorferi 11 , 22 , 23 ., As neuronal dysfunction is further analyzed in the light of these findings , it becomes necessary to consider which cells of the CNS milieu can function most aggressively/efficiently in creating an inflammatory environment that will contribute to clearance of the pathogen , but may in the process , harm nearby neurons ., Using in vitro and ex vivo experiments , investigators have shown that B . burgdorferi can induce potent production of inflammatory cytokines and chemokines in microglia , the resident macrophage cells of the CNS 2 , 6 , 12 , 24 ., The secretion of these mediators is likely a vital step in the development of inflammatory reactions , as chemokines , depending on their CC/CXC family sequence , may recruit distinct immune-effector cells , including monocytes , lymphocytes or neutrophils to sites of inflammation ., Additionally , many cytokines and chemokines become involved in apoptosis , cell cycle regulation and angiogenesis 3 , 6 , 23 , 25–28 ., While a similar response is observed in astrocytes , the repertoire range and scale of concentration is typically much lower than that of activated microglia 2 , 12 , 29 ., Importantly , although astrocytes are considered to be the primary CNS support cells , activated microglia can also secrete a host of soluble agents , such as glia-derived neurotrophic factor , that are potentially neuroprotective 30–32 ., As the majority of molecules produced by activated microglia are however considered to be pro-inflammatory and neurotoxic 2 , 30 , 31 , 33 , the microglial response to CNS infection with B . burgdorferi could tip the scales toward neuronal damage and death rather than survival in Lyme neuroborreliosis ., We have argued that the associated neurocognitive symptoms of Lyme neuroborreliosis may be the result of neuronal dysfunction resulting from inflammatory mediators released in response to infection with B . burgdorferi ., Serving as an in vivo proof of concept to this hypothesis , our laboratory has recently shown the production of IL-6 by astrocytes , as well as the induction of oligodendrocyte and neuronal apoptosis in brain tissues taken from rhesus macaques that received intraparenchymal stereotaxic inoculations of live B . burgdorferi 6 ., With the goal of beginning to establish cause and effect relationships between glial cell responses to B . burgdorferi and neuronal apoptosis , we quantified the production/secretion of inflammatory cytokines and chemokines in purified rhesus brain cortex microglia and astrocytes , in human neuronal cells from the SH-SY5Y ( SY ) neuroblastoma cell line , and in combinations of the above cells co-cultured with live B . burgdorferi or recombinant purified lipidated outer-surface protein A ( L-OspA ) ., SY cells were cultured in three dimensions as opposed to monolayer culture , as we had shown that that mode of cultivation significantly narrows the phenotypic gap between neuronal cell lines and primary neurons 34 ., We also determined the extent of B . burgdorferi-induced apoptosis in each of the above cellular combinations ., Using microarray analysis , we further examined the principal inflammation and apoptosis pathways affected by B . burgdorferi in these cells ., Our findings suggest a bystander effect in which the neurotoxic surroundings generated by microglia may contribute to neuronal cell damage ., Brain tissues used in this study were collected from rhesus macaques ( Macaca mulatta ) ., These animals were not experimentally infected with B . burgdorferi and were culled from the breeding colony because of chronic diarrhea or injury ., The procedure used for euthanasia was consistent with the recommendations of the American Veterinary Medical Associations Panel on Euthanasia and was approved by Tulane Universitys Institutional Animal Care and Use Committee ., Dulbeccos modified Eagles medium ( D-MEM ) -F-12 with L-glutamine and 15 mM HEPES buffer , D-MEM high glucose with L-glutamine , F-12 ( Ham ) with glutamax , penicillin ( 100 units/ml ) , streptomycin ( 100 units/ml ) , amphotericin B ( 0 . 25 µg/ml ) , non essential amino acids ( NEAA ) ( 100 units/ml ) , sodium pyruvate ( 100 units/ml ) , sodium bicarbonate ( 7 . 5% solution ) , trypsin ( 0 . 25% ) /EDTA ( 0 . 38 g/ml ) , trypan blue™ , normal goat serum ( NGS ) , Alexa-562 ( red ) -conjugated secondary antibody , and the ToPro ( blue ) nuclear stain were each from Invitrogen ., Primocin was from Invivogen ., Fetal bovine serum was from Hyclone/Thermo Scientific and Cytodex-3™ micro-carrier beads were from Amersham Biosciences ., Granulocyte-macrophage colony-stimulating factor ( GM-CSF ) , L-leucine methyl ester ( LME ) , Barbour-Stoenner-Kelly-H medium with rabbit serum , rifampicin , amphotericin , fish skin gelatin ( FSG ) , propidium iodide ( PI ) ( red ) and anti-GFAP antibody were from Sigma-Aldrich ., Lipidated ( L-OspA ) and unlipidated ( U-OspA ) outer surface protein A were a kind gift from GlaxoSmithKline ., Paraformaldehyde ( PFA , 2% ) was from USB Corporation , and Triton X-100 from ICN Biochemicals ., HuD primary antibody was from Santa Cruz Biotechnology ., IBA-1 antibody was from WAKO ., Three-micrometer pore diameter polyester transwell culture inserts ( Becton Dickinson , Falcon ) were incorporated into our co-culture models to physically separate SY cells from microglia ., The SY cells were seeded directly into 24 well dishes ( Costar ) with the microglia then seeded onto culture inserts that had been placed into the same wells as the SY cells ., The cell density in these experiments was 1×105 cells per ml of culture medium with an initial microglia to neuronal cell ratio of 4∶1 ., Live B . burgdorferi was added to the culture medium in both chambers for a 5-day stimulation , at an MOI of 10∶1 , in relation to the entire cell population ., In this way , it was possible to maintain a generally homogeneous exposure of each cell type to the culture medium while still allowing for bidirectional transfer of secreted molecules between the cell types ., Supernatants collected from primary glia and/or SY cells co-cultured with B . burgdorferi ( MOI 10∶1 ) or in medium alone for either 24 hours or 5 days were used for quantification of secreted cytokines and chemokines ., All of the mammalian cells were seeded at a density of 5×104 cells per 500 µl of culture medium with a glial cell to SY cell ratio of 4∶1 ., The 27-cytokine bioplex assays ( Bio-Rad ) were performed according to manufacturers directions as were the individual antigen-capture ELISAs for IL-6 , IL-8 , TNF , CCL3 , CCL4 , and MCP-1 ( CCL2 ) ., Sandwich ELISA capture and detection antibody pairs for human IL-6 , IL-8 and TNF , along with recombinant human IL-6 , IL-8 , TNF and horseradish peroxidase ( HRP ) were from BD Biosciences ., We obtained CCL3 and CCL4 ELISA DuoSet kits and the MCP-1 ( CCL2 ) kit from R&D Systems ., Supernatants collected from primary glia and/or SY cells co-cultured with B . burgdorferi ( MOI 10∶1 ) or in medium alone for either 24 hours or 5 days were used for quantitative determination of total secreted nitric oxide ., The mammalian cells were seeded at a density of 5×104 cells per 500 µl of culture medium with an initial glial cell to SY cell ratio of 4∶1 ., The Total Nitric Oxide Assay Kit ( catalogue #917–020 ) from Assay Designs was used , and all analyses were performed according to directions from the manufacturer ., Microglia , astrocytes , SY cells or combinations thereof were cultured in the presence of B . burgdorferi ( MOI 10∶1 ) or medium alone for either 24 hours or 5 days ., After removal of the supernatant , the cells were harvested using trypsin , washed in phosphate buffered saline ( PBS ) , and fixed for 5–10 minutes in 2% PFA ., The fixed cells were permeabilized in PBS/FSG/Triton and blocked with 10% NGS ., Apoptosis was evaluated using the Apoptag TUNEL assay kit ( Chemicon/Millipore ) as per manufacturers instructions and the results were analyzed using a Leica TCS SP2 confocal microscope equipped with 3 lasers ., Briefly , 6–18 0 . 2-µm optical slices per image were collected at 512×512 pixel resolution ., In order to distinguish SY cells in the co-cultures containing glia , the cells were additionally stained with primary anti-HuD antibody for 1 hour , washed 3 times in PBS and then stained with Alexa 562-labeled secondary antibody for 45 minutes ., The To-Pro nuclear stain was combined with the secondary antibody at a concentration equal to 0 . 05 µg/ml ., The identities of microglia and astrocytes were additionally confirmed using the above protocol , substituting anti-IBA-1 and anti-GFAP respectively , for the anti-HuD ., Cell morphology consistent with apoptosis including cell shrinkage , nuclear condensation and membrane blebbing was assessed along with the fluorescein staining for TUNEL ., The number of apoptotic cells counted was divided by the total ( 500 minimum ) number of cells counted ., When the assays included co-culture of SY cells with glia , the number of cells that were double- stained for apoptosis and neuron specificity were divided by the total number of cells displaying the neuronal marker stain ., Statistical significance was evaluated by One Way analysis of variance ( ANOVA ) followed by Bonferroni , Tukey and Levines tests ., RNA was isolated from approximately 5×106 isolated SY cells or microglia using the RNeasy kit ( Qiagen ) plus DNA-free ( Ambion ) to eliminate DNA contamination ., Five hundred nanogram of total RNA was amplified and used to synthesize Cy-labeled cDNA with the Low RNA Input Linear Amplification Kit ( LRILAK , Agilent Technologies ) ., Cy3 ( control ) and Cy5 ( experimental ) labeled cDNA were mixed in equimolar quantities and hybridized overnight at 55°C , to Agilent whole-genome microarrays ( 4×44 k format ) ., While the samples derived from macaque microglia were hybridized to rhesus macaque arrays ( Agilent # G2519F ) with over 44 , 000 rhesus macaque probes , representing approximately 18 , 000 individual annotated genes , the samples derived from the neuronal cells were hybridized to human genome arrays ( Agilent # G4112F ) with over 41 , 000 oligonucleotides , representing approximately 22 , 000 unique human genes ., The slides were scanned on a GenePix 4000B scanner , and data were extracted from the resulting 16-bit TIFF images using GenePix Pro 6 . 1 software ., Data were analyzed using Spotfire DecisionSite for Microarray Analysis ., Values were log2 transformed and normalized using a Locally Weighted ScatterPlot Smoothing ( LOWESS ) script within S+ ArrayAnalyzer ., RNA was isolated from approximately 5×106 isolated microglia or neuronal cells using the RNeasy kit ( Qiagen ) plus DNA-free ( Ambion ) to eliminate DNA contamination ., The RNA was reverse-transcribed into DNA using a OneStep RT-PCR kit ( Qiagen ) and the QuantiFast™ SYBR® Green PCR kit ( Qiagen ) was then used for the quantitative real-time ( QRT ) -PCR ., All assays were performed according to directions from the manufacturer and using Qiagen Quantitect® primer pairs in a 96-well block Applied Biosystems 7900 HT fast RT PCR System ., PCR efficiencies , average fold change and statistical significance were evaluated using REST© software ., In experiments where B . burgdorferi or recombinant L-OspA was co-cultured with isolated rhesus cortex glia , we observed robust expression and release of IL-6 and IL-8 ( Figure 1A ) ., We also observed the production of TNF , although at a much lower level , and only by microglial cells and cells in the aggregate cultures , which themselves contained microglia ( Figure 1A insert ) ., SY cell production of IL-6 and IL-8 in response to B . burgdorferi and L-OspA were below the limit of detection for the assay used ., Cytokine/chemokine expression levels in response to the same stimuli often varied significantly in glial cells obtained from different animals ., The patterns of expression , however , were reproducible in experiments where the glial cells were isolated from tissue originating from a single individual animal ., As such , the highest response was always that of microglia , followed by that of aggregate cultures and then astrocytes , regardless of the animal from whom the cells had been obtained ., With the exception of TNF , whose expression peaked at 24 hours and then declined , levels of cytokine/chemokine expression also increased with the time of stimulation ( Figure 1B ) ., As the inherent difficulties in culturing primary neurons 35 , 36 often render their use in experiments impractical , the neuroblastoma cell line SH-SY5Y was employed to assess neuronal responses ., We used a three-dimensional ( 3D ) rather than traditional monolayer ( 2D ) culture for the SY cells , as this culture method has been shown to promote a more normal , untransformed phenotype as compared to that of transformed cells grown in 2D 34 , 35 , 37–45 ., In order to assess endpoint damage to the glia and SY cells responding to co-culture with B . burgdorferi or L-OspA , we employed the terminal deoxynuclease dUTP nick end labeled ( TUNEL ) assay as a tool for visualization of apoptosis ., When microglia isolated either alone or in aggregate with astrocyte cells , were co-cultured with both B . burgdorferi and SY for at least 5 days , increases in cellular apoptosis consistently occurred ., Apoptosis in glial cells was minimal as compared to the un-stimulated controls and there were no remarkable changes in apoptotic levels with regard to individual animals ., SY cells cultured for 5 days with B . burgdorferi alone , or in combination with astrocytes and B . burgdorferi , showed only baseline levels of apoptosis ( Table 1 ) ., No significant increase in astrocyte apoptosis was observed whether these cells were incubated with other cell types , L-OspA or B . burgdorferi ( Table 1 ) ., Confocal microscopy images of mixed cultures stained for TUNEL and with cell-specific markers indicated that the majority of cells dying in response to B . burgdorferi were SY cells ( Figure 2A ) ., In consideration of these findings , we focused our next experiments on SY cells cultured in the presence of microglial cells and B . burgdorferi for 5 days ., Significant increases in SY cell apoptosis occurred consistently in cell cultures from each of the 4 animals sampled when both microglia and B . burgdorferi were included in the culture conditions ( Figure 2B ) ., In order to determine whether stimulation of the culture medium with IL-6 , IL-8 , TNF or combinations thereof would be sufficient to induce apoptosis in the SY cells , we conducted experiments in which concentrations of each cytokine ( recombinant-human ) were added to the neuronal medium that were comparable to the highest expression values obtained during our in vitro assays with microglia and SY cells ., After a 5 day stimulation , we did not find increases in neuronal apoptosis comparable to those observed in the co-culture of SY cells with microglia and B . burgdorferi , or significantly above the baseline levels of SY cells cultured in medium alone ( data not shown ) ., In view of the evidence indicating that microglia were the more robust responders to the inflammatory stimuli of B . burgdorferi ( Figure 1 ) , and that molecules other than , or in addition to IL-6 , IL-8 and TNF were required to elicit SY cell apoptosis , we further explored the diversity of mediators produced by microglia in response to B . burgdorferi ., As several investigators had reported that the activated glia seen in many CNS pathologies were able to kill neurons by the release of nitric oxide ( NO ) into their surrounding environments 46–48 , and knowing that B . burgdorferi could elicit the release of NO from exposed macrophages 49 , we explored this possibility in our models ., Using an adaptation of the Greiss reaction , we were not able to detect any significant release of NO into the supernatant of microglia , astrocytes , SY cells or any combinations of the three cell types co-cultured with B . burgdorferi for either 24 hours or for 5 days ( data not shown ) ., Using a 27 cytokine Bio-Plex assay to expand on our previous data , we found along with IL-6 and IL-8 , significant B . burgdorferi-induced upregulation in expression of the pro-inflammatory chemokines CCL2 , CCL3 , CCL4 and CCL5 by microglial cells ( Table 2 ) ., L-OspA ( 0 . 25 µg/ml ) was observed to elicit a comparable upregulation of cytokines in both microglia and astrocytes , indicating that B . burgdorferi lipoproteins may elicit the lions share of the spirochetal stimulus ., However , when microglia were present in the cultures , alone or with astrocytes ( aggregate cultures ) , it appeared that spirochetes provided , overall , the stronger stimulus ., SY cell responses were absent or minimal in comparison to that of the microglia , except for the production of vascular endothelial growth factor ( VEGF ) ( Table 2 ) , which has been increasingly implicated as a contributing factor in neuronal protection and survival 50–53 ., B . burgdorferi induced significant cytokine/chemokine expression in astrocytes , but expression levels were generally orders of magnitude lower than those of microglia that had been derived from the same tissue ( Table 2 ) ., These results , combined with our previous cytokine and apoptosis assays , prompted us to focus on interactions of microglia with SY cells in the presence of B . burgdorferi , and to broaden our exploration of pathways involved in microglial activation and neuronal apoptosis ., We used microarray analysis to further investigate how exposure to B . burgdorferi might affect global gene expression in microglial cells ., As changes in the expression and activity of multiple genes often work in concert to affect responses to many cellular pathogens , including B . burgdorferi 10 , 54–56 , we used Ingenuity Pathways Analysis software to compare transcript levels in 18 , 000 annotated rhesus genes ., When microglia were cultured in the presence of B . burgdorferi , as compared to medium alone , five of the ten most altered canonical pathways , as well as the chemokine signaling pathway ( number 19 on the list of 232 affected pathways ) , showed significant upregulation in inflammatory signaling ( Figure 3A ) ., Many of the transcript changes that occurred within the triggering receptor expressed on myeloid cells-1 ( TREM1 ) , pattern recognition receptors ( PRR ) , IL-10 , IL-6 and chemokine signaling ( Chem . Sig . ) pathways were proinflammatory and they were repeated in several pathways that included both innate and adaptive immune responses ( Table 3 ) ., Interestingly , the pathway that exhibited the most profound changes in regard to inflammatory signaling was TREM1 ( Figure 3B , Table 4 ) ., TREM1 cell surface receptors associate with the adaptor molecule DAP12 for signaling and function and have been classified as immune/inflammatory response amplifiers ., Although the receptor has thus far not been found to be expressed on microglial cells 55 , 57–61 , our array data indicate that it is expressed in rhesus microglia , or a subpopulation thereof , and is also upregulated during infection with B . burgdorferi ., A low level of TREM1 expression in rhesus microglia was confirmed by RT-PCR ( data not shown ) ., RT-PCR was further used to confirm significant transcript up-regulation in the following three selected molecules: CCL2 ( TREM1 and Chem . Sig . pathways ) , IL-6 ( TREM1 , PRR , IL-10 and IL-6 pathways ) and CCL5 ( PRR and Chem . Sig . pathways ) , ( Table 5 ) ., Individual transcript changes for both animals sampled can be accessed online through the supplemental data portion of this report ( Tables S1 and S2 ) ., In addition to affecting the environment surrounding neurons exposed to B . burgdorferi , activated microglia may also affect gene expression of neurons themselves ., Microarray analysis for the p53 Signaling pathway revealed a microglia-induced shift in the gene expression of SY cells co-cultured with B . burgdorferi from one of cell survival and proliferation to one more in line with cell cycle arrest and apoptosis ( Figure 4A and 4B ) ., In cultures containing only SY cells and B . burgdorferi , there were prominent increases in transcript for the cell survival and anti-apoptotic molecules protein kinase B ( AKT1 ) and BCL2-like ( Bcl-XL ) genes , combined with a decrease in transcript for pro-apoptotic damage-regulated autophagy modulator ( DRAM ) ( Table 6-left , Figure 4A ) ., This scenario , however , changed dramatically in experiments when microglia were included in the co-culture of SY cells with B . burgdorferi ( Figure 4B , Table 6 , right ) ., In order to study microglia-induced gene expression changes in the SY cells , transwell inserts were incorporated into the culture system providing physical separation between the SY cells and the microglial cells , yet allowing for bidirectional transfer of secreted molecules between the cell types ., In this model , gene expression for phosphatase and tensin homolog ( PTEN ) , which functions to modulate cell survival and proliferation primarily through its downstream effects on AKT1 , was shown to be significantly increased ( Table 6 , right ) ., SY cell anti-apoptotic Bcl-XL expression was observed to completely swing from up to down-regulation in the presence of microglia , while pro-apoptotic BCL2-associated X protein ( Bax ) and phorbol-12-myristate-13-acetate-induced protein 1 ( NOXA-1 ) transcripts were increased ., Additionally , the transcript for cyclin-dependent kinase 2 ( CDK2 ) , which typically drives the cell cycle , was down-regulated , and the G1/S cell cycle checkpoint inhibitors , cyclin-dependent kinase inhibitor 1A ( p21 ) and glycogen synthase kinase 3 beta ( GSK3β ) , were both increased ( Figure 4 , Table 6 right ) ., The observed changes for individual gene transcript levels may be accessed online in the supplemental data portion of this report ( Tables S3 and S4 ) ., RT-PCR was used to confirm transcript changes in selected molecules from co-cultures of the SY neurons with B . burgdorferi both in the presence and absence of microglia ( Table 7 ) ., Several investigators have shown that when in the presence of certain stimuli , not only can macrophages/microglia induce apoptosis in neighboring cells through a bystander effect , but that there might be an additional requirement for cell-cell contact between the macrophages/microglia and the affected nearby cells 62–64 ., To address the question of whether the microglia-secreted mediators in our system were enough to stimulate SY cell apoptosis on their own , we again incorporated transwell inserts into our co-culture models to physically separate the microglia from the SY cells while at the same time , maintaining a generally homogeneous exposure of each cell type to the culture medium ., Parallel assays for B . burgdorferi-induced apoptosis were set up where one model followed the same experimental design that was described in Figure 2 , and the second model included the addition of transwell inserts into the co-culture system for the duration of the experiment ., In each case , primary microglia or aggregate cell isolates from three separate non-human primates were combined in co-culture for 5 days with SY neurons and B . burgdorferi ., Using the original co-culture format , we found an average 7 . 7-fold increase in neuronal apoptosis when SY cells were co-cultured with microglia and B . burgdorferi as compared to the SY cells cultured in medium alone ., In contrast , no significant changes in apoptosis levels were observed when the SY cells were co-cultured with B . burgdorferi , with microglia , or with microglia and B . burgdorferi using the transwell inserts ( Figure 5 ) ., Neuroinflammation is thought to be a contributing factor in a number of neurodegenerative disorders including Alzheimers disease , Parkinsons disease and multiple sclerosis , as well as in Lyme neuroborreliosis 6 , 7 , 31 , 65 ., Inflammatory mediator levels are often elevated in these disorders suggesting that they are actively involved in the disease process 2 , 6 , 31 , 66 ., Because neurological symptoms do occur in many patients with Lyme disease , and cognitive impairment is often a part of this scenario , it was important to discover which mediators likely caused the effects of B . burgdorferi-induced damage in neurons ., We hypothesized that the inflammatory environment generated during a possible in vivo exposure of glial cells to B . burgdorferi could harm neurons through a bystander effect ., To address this hypothesis , we used in vitro experiments to demonstrate that with regard to microglia , astrocytes , and neurons , the fundamental triumvirate of cells in the CNS , it was the microglia that most aggressively responded to interaction with B . burgdorferi ., In addition to inducing a pronounced and sustained production of cytokines and chemokines in microglial cells , B . burgdorferi also activated important inflammatory signaling pathways in these cells ., Together , these responses potentially contributed to creating a reactive environment that was toxic to the SY cells ., Interestingly , we also found that in addition to inducing SY cell apoptosis through a bystander effect , there might be a requirement for direct cell-cell contact between microglia and neurons for end-stage damage to occur ., In early experiments , we determined that when SY cells were co-cultured in the presence of microglia and B . burgdorferi for at least 5 days ( MOI 10∶1 ) , significant increases in SY , but not glial cell apoptosis occurred ., Our experiments also ruled out any B . burgdorferi related upregulation in the production of nitric oxide ., While previous reports had indicated that L-OspA would induce significant levels of apoptosis in primary rhesus astrocytes 67 , the L-OspA concentration of 1 µg/ml used in those experiments was 4 times greater than the one used in the current experiments ., Using a formula developed by Norgard , et al . 68 we approximated the total amount of outer surface lipoproteins correlating to the number of spirochetes used in our MOI ., An L-OspA concentration of 0 . 25 µg/ml provided a quantitatively more realistic representation of lipoproteins present in B . burgdorferi at the MOI used in our studies ., We further discovered that although B . burgdorferi did induce a potent expression and secretion of the inflammatory cytokines/chemokines IL-6 , IL-8 and TNF in microglia , additional mediators were required to trigger neuronal cell apoptosis ., These results prompted us to broaden our study of potential inflammatory mediators , and at the same time , to explore B . burgdorferi-activated pathways in microglia and SY cells that might be further contributing to the neurotoxicity demonstrated in the apoptosis assays ., By expanding our search for the expression of cytokines and chemokines that were potentially relevant to B . burgdorferi-induced neurotoxicity we found that in addition to IL-6 , IL-8 and TNF , the four well known proinflammatory chemokines CCL2 , CCL3 , CCL4 and CCL5 were significantly upregulated in microglial cells co-cultured with B . burgdorferi ., Upregulation of CC chemokines has been described both in murine models of neurocysticerosis and multiple sclerosis ( experimental autoimmune encephalomyelitis , EAE ) 69 , 70 ., CCL2 in particular has been shown to play a role in the pathogenesis of EAE 71 ., Perhaps most interesting were the microarray results showing that each of these molecules , with the exception of CCL3 , was represented in more than one of the microglial signaling pathways that were most affected by B . burgdorferi ., Considering the high number of inflammatory genes involved in these pathways and their potential for cross-talk and downstream regulation , we believe that the TREM1 pathway , together with those involving PRR , IL-10 , IL-6 and Chem ., Sig ., , contribute to B . burgdorferi-induced inflammation in the CNS ., Even though some of the remaining affected pathways may have provided checks and balances to our findings , many of the transcript changes that occurred in the pathways that we focused on were proinflammatory and included both innate and adaptive immune response molecules ., As such , there seems to be a consistent pattern of inflammatory signaling in microglial cells that is directly associated with the presence of B . burgdorferi , and that might adversely affect nearby neuronal cells ., When glial cells were co-cultured with B . burgdorferi in the presence of SY cells the concentration of inflammatory mediators was often reduced as compared to that elicited in the absence of SY cells ( Table 2 ) ., Since we saw no evidence of glial cell apoptosis , this decrease is most likely due to the reduced number of glial cells , the main producers of pro-inflammatory mediators that were included in these cultures ., There are reports that apoptotic neurons co-cultured with microglia may down-regulate microglial synthesis of pro-inflammatory molecules 32 ., Such a mechanism also may have contributed to the observed reduction , at least in the case of microglia/SY cell co-cultures ., Using microarray analysis , we also showed that B . burgdorferi-activated microglia could invoke changes in their cellular environment that affected gene expression in proximal SY cells ., When SY cells were co-cultured with B . burgdorferi alone , SY gene expression for molecules in the p53 signaling pathway indicated a mode of cell survival and proliferation ., This conclusion was founded on the observed upregulation in transcripts for AKT1 and Bcl-XL , paired with decreased gene expression for DRAM ., AKT1 , which is activated through the phosphatidylinositol 3-kinase ( PI3K ) pathway , is known to stimulate cell cycle progression ( via p21 phosphorylation and release from CDK2 ) and to play an important role in promoting cell survival through the suppression of apoptosis 72-74 ., Bcl-XL is likewise a strong promoter of cell survival , but as one of the major anti-apoptotic members of the conserved Bcl-2 family of proteins , it functions to inhibit programmed cell death through the cont | Introduction, Materials and Methods, Results, Discussion | Inflammation has long been implicated as a contributor to pathogenesis in many CNS illnesses , including Lyme neuroborreliosis ., Borrelia burgdorferi is the spirochete that causes Lyme disease and it is known to potently induce the production of inflammatory mediators in a variety of cells ., In experiments where B . burgdorferi was co-cultured in vitro with primary microglia , we observed robust expression and release of IL-6 and IL-8 , CCL2 ( MCP-1 ) , CCL3 ( MIP-1α ) , CCL4 ( MIP-1β ) and CCL5 ( RANTES ) , but we detected no induction of microglial apoptosis ., In contrast , SH-SY5Y ( SY ) neuroblastoma cells co-cultured with B . burgdorferi expressed negligible amounts of inflammatory mediators and also remained resistant to apoptosis ., When SY cells were co-cultured with microglia and B . burgdorferi , significant neuronal apoptosis consistently occurred ., Confocal microscopy imaging of these cell cultures stained for apoptosis and with cell type-specific markers confirmed that it was predominantly the SY cells that were dying ., Microarray analysis demonstrated an intense microglia-mediated inflammatory response to B . burgdorferi including up-regulation in gene transcripts for TLR-2 and NFκβ ., Surprisingly , a pathway that exhibited profound changes in regard to inflammatory signaling was triggering receptor expressed on myeloid cells-1 ( TREM1 ) ., Significant transcript alterations in essential p53 pathway genes also occurred in SY cells cultured in the presence of microglia and B . burgdorferi , which indicated a shift from cell survival to preparation for apoptosis when compared to SY cells cultured in the presence of B . burgdorferi alone ., Taken together , these findings indicate that B . burgdorferi is not directly toxic to SY cells; rather , these cells become distressed and die in the inflammatory surroundings generated by microglia through a bystander effect ., If , as we hypothesized , neuronal apoptosis is the key pathogenic event in Lyme neuroborreliosis , then targeting microglial responses may be a significant therapeutic approach for the treatment of this form of Lyme disease . | Lyme disease , which is transmitted to humans through the bite of a tick , is currently the most frequently reported vector-borne illness in the northern hemisphere ., Borrelia burgdorferi is the bacterium that causes Lyme disease and it is known to readily induce inflammation within a variety of infected tissues ., Many of the neurological signs and symptoms that may affect patients with Lyme disease have been associated with B . burgdorferi-induced inflammation in the central nervous system ( CNS ) ., In this report we investigated which of the primary cell types residing in the CNS might be functioning to create the inflammatory environment that , in addition to helping clear the pathogen , could simultaneously be harming nearby neurons ., We report findings that implicate microglia , a macrophage cell type in the CNS , as the key responders to infection with B . burgdorferi ., We also present evidence indicating that this organism is not directly toxic to neurons; rather , a bystander effect is generated whereby the inflammatory surroundings created by microglia in response to B . burgdorferi may themselves be toxic to neuronal cells . | infectious diseases/bacterial infections, infectious diseases/infectious diseases of the nervous system, immunology/innate immunity | null |
journal.pgen.1006430 | 2,016 | CLOCKWORK ORANGE Enhances PERIOD Mediated Rhythms in Transcriptional Repression by Antagonizing E-box Binding by CLOCK-CYCLE | Almost all organisms from Cyanobacteria to humans have internal circadian clocks that drive daily rhythms in physiology , metabolism and behavior , thereby synchronizing internal processes with the external environment ., In eukaryotes , the circadian clock keeps time via one or more transcriptional feedback loops 1 ., In Drosophila , a heterodimer formed by CLOCK ( CLK ) and CYCLE ( CYC ) binds E-box sequence activates transcription to initiate clock function ., In the core loop , CLK-CYC activates period ( per ) and timeless ( tim ) transcription during mid-day , effecting a rise in per and tim mRNA levels that peaks during the early evening ., PER and TIM proteins then accumulate , form a dimer , and move into the nucleus to bind CLK-CYC during the night , thereby inhibiting their transcriptional activity until PER and TIM are degraded early in the morning 2 , 3 ., Another interlocked transcriptional feedback loop is also regulated by the core feedback loop ., In this loop , CLK-CYC activates transcription of vrille ( vri ) and PAR-domain protein 1ɛ ( Pdp1ɛ ) , which bind D-boxes to repress and activate transcription , respectively , and drive RNA cycling of Clk and other output genes in the opposite phase as per , tim , vri and Pdp1ɛ 4–6 ., PER was previously found inhibit CLK-CYC binding to E-boxes in vitro 7 , which suggests that the rhythmic transcription of CLK target genes are mediated by PER-dependent rhythms in E-box binding by CLK-CYC ., Chromatin immunoprecipitation ( ChIP ) experiments using fly heads support this model , showing that CLK-CYC rhythmically bind E-boxes in the per circadian regulatory sequence ( CRS ) and the tim upstream sequence 8 ., However , the mechanism by which CLK-CYC heterodimers are removed from E-boxes during repression is not well understood ., PER is required for the rhythmic binding of CLK complexes , as CLK constantly binds to per and tim promoters in per01 flies 8 , indicating that PER inhibits transcription by removing CLK-CYC from E-boxes ., Interestingly , co-expression of another transcription factor , CLOCKWORK ORANGE ( CWO ) , strongly enhanced PER-mediated repression in cultured Drosophila Schneider 2 ( S2 ) cells 9 , suggesting that PER is unable to efficiently remove CLK from DNA in the absence of other transcription repressors ., Previous studies demonstrated that CWO , a basic helix-loop-helix ( bHLH ) -ORANGE transcriptional factor 10 , is a direct target of CLK-CYC 9 , 11 , 12 ., In Drosophila Schneider 2 ( S2 ) cells , overexpression of CWO reduces the basal transcription of per , tim , vri and Pdp1ɛ promoter-driven luciferase reporter genes 9 , 12 , 13 ., Furthermore , in the presence of PER , CWO repress CLK mediated transcription 5–10 fold in S2 cells , indicating that CWO is a strong transcription repressor that can cooperate with PER to repress CLK-CYC mediated transcription 9 ., In cwo mutants or cwo RNAi knockdown flies , the levels of per , tim , vri and Pdp1ɛ mRNAs are increased during the early to mid-morning 9 , 12 ., These results suggest that CWO co-represses CLK-CYC activity along with PER during the end of a cycle 9 , 12 ., However , the mechanism through which CWO represses CLK-CYC mediated gene transcription remains unknown ., In this study we demonstrate that CWO and CLK bind core clock gene E-boxes in a reciprocal pattern across the circadian cycle in vivo , which suggests that CWO competes with CLK to bind E-boxes ., We also show that CWO acts to decrease CLK binding to tim E-boxes during early morning , when PER binds CLK-CYC to reduce its binding to DNA 8 , but not during early night when CLK-CYC strongly binds E-boxes in the absence of PER ., These results suggest a model for CWO function where CWO has low DNA binding affinity compared to CLK-CYC complexes during the activation phase , but has higher affinity compared to CLK-CYC-PER complexes , and is thus capable of removing CLK-CYC-PER complexes from E-boxes to consolidate and maintain repression ., Constant high CWO binding to the tim promoter in Clkout flies ( i . e . comparable to binding at ZT2 in wild-type ) and constant low CWO binding in per01 flies ( i . e . comparable to binding at ZT14 in wild-type ) supports our model for CWO repression ., As a whole , these results suggest that CWO co-represses CLK-CYC activity with PER by competing with CLK-CYC-PER complexes for E-box binding , therefore promoting the transition to off-DNA repression ., Earlier studies demonstrated that cwo mRNA cycles in phase with per , tim , vri , and Pdp1 , but with a higher basal level , and thus lower amplitude 9 , 12–14 ., To determine whether CWO protein levels also cycle , western analysis was carried out using head extracts from wild-type flies collected every 4 hours in a 12-h light/12-h dark ( LD ) cycle ., We find that the levels of CWO do not change throughout an LD cycle ( Fig 1 ) , consistent with previous results 15 ., Given that cwo mRNA levels cycle , it is possible that constant CWO levels result from post-transcriptional regulation or a long half-life ., CWO contains a bHLH domain , a structural motif that characterizes a family of E-box binding transcription factors 16–19 , which suggests that CWO may regulate CLK-CYC target gene transcription via E-box binding ., Previous ChIP-on-chip and gel-shift analyses in S2 cells demonstrated that CWO specifically binds to the E-box of core clock genes 12 , 13 , however it is still unknown whether CWO binds those core clock genes in vivo , and whether the binding intensity changes throughout the day ., To test these possibilities , ChIP assays were carried out on wild-type flies collected in the early morning ( ZT2 ) and in the early night ( ZT14 ) using CWO and CLK antisera ., Fragments containing upstream E-boxes from tim , per , Pdp1 and vri , which are necessary for high-amplitude mRNA cycling in vitro or in vivo 4 , 5 , 20–25 , were amplified from the immunoprecipitates and then quantified ., In CWO immunoprecipitates , the tim , vri and Pdp1 E-box containing fragments were two to threefold more abundant at ZT2 than at ZT14 ( Fig 2A ) , suggesting that CWO binding is time-dependent , though the dynamic binding of CWO on the per E-box fragment is less robust than the others ., Importantly , this temporal binding pattern is antiphase to CLK binding , as CLK shows high binding intensity during the night at ZT14 and low binding during the daytime at ZT2 ( Fig 2B ) , consistent with previous results 8 , 11 ., The reciprocal binding pattern of CLK and CWO implies that these transcription factors compete for E-box binding ., If so , both CLK and CWO must occupy the same E-boxes ., To test this possibility , we determined how mutating E-boxes upstream of tim affected CLK and CWO binding ., The circadian enhancer upstream of tim is comprised of two tandem E-boxes that are spaced seven nucleotides apart 24 , 26 , a structure that is conserved among core clock genes in various species 27 ., Both of these E-boxes were indispensable for tim mRNA expression in S2 cells 24 , suggesting that these tandem E-box motifs are binding sites for both CLK and CWO ., To determine if this is the case , a series of 136bp fragments from the tim promoter containing an E-box1 ( E1 ) mutant ( mE1-E2 ) , an E-box 2 ( E2 ) mutant ( E1-mE2 ) , an E1 and E2 double mutant ( mE1-mE2 ) or a control with wild-type E-boxes ( E1-E2 ) were generated , inserted into the pHPdestGFP vector 28 , and targeted to the attP18 genomic site ( Fig 3A ) ., To confirm that this promoter fragment is sufficient to drive rhythmic expression , we carried out quantitative reverse transcription-PCR ( qRT-PCR ) to monitor GFP mRNA levels in flies collected every 4-h during an LD cycle ., Quantification of GFP mRNA levels in flies with WT tim promoter shows a ~10-fold diurnal rhythm with a peak at ZT14 and a trough at ZT2 to ZT6 ( S1 Fig ) , consistent with timing and amplitude of per and tim mRNA cycling in wild-type flies 29 , 30 ., However , even at the normal tim mRNA peak ( ZT14 ) , mE1-E2 , E1-mE2 and mE1-mE2 flies express little or no eGFP mRNA ( S1 Fig ) , indicating that both E1 and E2 are indispensable for expression of tim mRNA in vivo ., This result is consistent with previous tim-luciferase reporter results in S2 cells 24 ., We next carried out ChIP assays using CWO and CLK antisera on the same fly strains to test whether E1 and E2 are required for CWO and CLK binding ., At ZT2 , when CWO strongly binds to the tim promoter , CWO binding intensity was drastically reduced in mE1-E2 , E1-mE2 and mE1-mE2 flies compared to WT ( Fig 3B ) ., Likewise , CLK binding intensity was drastically reduced in mE1-E2 , E1-mE2 and mE1-mE2 flies compared to WT at ZT14 , when CLK binding is strongest ( Fig 3B ) ., These results indicate that both E1 and E2 are indispensable for both CWO and CLK binding to the tim circadian enhancer ., Given that CWO specifically targets E-boxes in S2 cells by Gel-shift analyses 13 , we conclude that both CLK and CWO bind intact tandem E1-E2 motifs in vivo ., In mice , CLK-BMAL1 dimers cooperatively bind tandem E-boxes in vitro 27 , 31 , and this may be the case for CWO given the requirement for both E1 and E2 E-boxes ., Previous studies showed that increasing the level of CWO expression reduces per , tim , vri and Pdp1ɛ mRNA levels in S2 cells and that their trough mRNA levels are higher in cwo mutant or knockdown flies , indicating that CWO acts to repress CLK-mediated gene transcription in vitro and in vivo 9 , 12–14 ., Given that CWO and CLK bind to the same E-box motif , we wondered whether CWO represses CLK-mediated transcription by inhibiting CLK binding ., To test this possibility , ChIP assays were carried out using CLK antiserum on wild-type and cwo5703 flies at the trough ( ZT2 ) and peak ( ZT14 ) times of CLK-CYC target gene transcription and mRNA abundance in LD ., Although cwo5703 mutants lengthen the period of activity rhythms by 2–3h in DD 9 , 13 , the peak and trough phases of CLK-CYC target gene transcription and mRNA abundance are comparable in cwo5703 mutants and wild-type flies in LD 9 , 13 ., We find that CLK binds tim E-boxes with a robust rhythm in wild-type flies and a lower amplitude rhythm in the cwo5703 mutant ( Fig 4A ) ., However , the intensity of CLK binding in cwo5703 is significantly increased at ZT2 compared to wild-type , indicating that CWO acts to reduce CLK-CYC binding at the trough of its binding cycle ( Fig 4B ) ., Given that CWO strongly binds tim E-boxes at ZT2 ( Fig 2A ) , we propose that CWO inhibits CLK-CYC binding during the repression phase by antagonizing PER-CLK-CYC complexes to maintain off-DNA repression ., There was no significant difference in CLK binding between cwo5703 and wild-type at ZT14 ( Fig 4B ) , despite decreased peak levels of per , tim , vri and Pdp1ɛ mRNA at ZT14 in cwo mutant and RNAi knockdown flies 9 , 12–14 , suggesting that CWO has little impact on CLK-CYC binding in the absence of PER ., Given that CWO suppresses CLK binding at ZT2 in the early morning but not at ZT14 during the early evening ( Fig 4B ) , it is possible that PER is necessary for CWO to antagonize CLK E-box binding since PER accumulates to high levels in the nucleus around dawn and is at low levels in the cytoplasm around dusk 32 ., Indeed , our results support a model developed previously to explain cooperation between CWO and PER to repress CLK-CYC mediated transcription in S2 cells 9 ., In this model , CWO is proposed to compete with CLK-CYC heterodimers for E-box binding only when PER binds CLK-CYC , thereby reducing their affinity for E-box binding ., To test this model , we performed ChIP assays using CWO antiserum on wild-type , Clkout and per01 flies collected at ZT2 and ZT14 in LD ., In Clkout flies , which necessarily lack CLK-CYC heterodimers 33 , CWO is bound to tim E-boxes at both ZT2 and ZT14 with binding signals comparable to the strong CWO binding in wild-type flies at ZT14 ( Fig 5A ) ., In contrast , in per01 flies , which lack PER-dependent repression of CLK-CYC activation 34 , low binding signals of CWO were detected at ZT2 and ZT14 , indicating that PER is indeed required for CWO to bind E-boxes ( Fig 5A ) ., Moreover , CWO binding was significantly increased in Clkout versus wild-type flies at ZT14 , indicating that CLK-CYC binding at ZT14 reduces CWO binding ., Likewise , a significant increase in CWO binding was also seen in wild-type versus per01 flies at ZT2 , indicating that PER enhances CWO binding ( Fig 5A ) ., To determine whether differences in CWO binding in Clkout and per01 flies were due to differences in CWO protein levels , we carried out western analysis using head extracts from these mutants collected at ZT2 and ZT14 ., Since cwo transcription is regulated in part by the transcriptional feedback loop , CWO protein levels are slightly lower in Clkout flies and slightly higher in per01 flies ( Fig 5B and 5C ) ., However , the lower levels of CWO in Clkout resulted in higher E-box binding , and higher CWO protein levels in per01 resulted in lower E-box binding ., This result suggests that the differences in CWO-E-box binding are not due to altered CWO protein levels , but due to the relative DNA binding affinities of CWO and CLK in these mutants ., These results , taken together , strongly support and extend the model described by Kadener et al . , 2007 , for CWO binding as it relates to CLK-CYC repression ., When CLK-CYC targets are activated , CLK-CYC binds DNA with higher affinity than CWO , thus CLK binding is not altered in the presence or absence of CWO ., When CLK-CYC targets are repressed , PER binds CLK-CYC complexes and decreases their DNA binding affinity , thereby favoring CWO binding to E-boxes and enhancing PER mediated removal of CLK-CYC-PER complexes from the DNA ( Fig 6 ) ., Although we can’t exclude the possibility that PER enables CWO E-box binding independent of its interaction with CLK-CYC , the available evidence strongly supports the model proposed ., Rhythmic binding of CLK-CYC to E-boxes is essential for rhythmic transcription of the core circadian oscillator genes per and tim in Drosophila ., CLK-CYC bind E-boxes upstream of per and tim in the late day and early night to activate transcription; and is released from these binding sites during late night 8 , 35 , 36 ., Previous work demonstrated that CLK constitutively binds per and tim E-boxes in per01 flies , indicating that PER is essential for rhythmic binding of CLK-CYC , and is key to removing CLK-CYC from E-boxes 8 ., In this study we report that CWO also contributes to removing CLK-CYC from E-boxes ., In cwo5703 mutant flies , CLK binding intensity is significantly increased at the trough of its binding cycle , suggesting that repression is incomplete in the absence of CWO ( Fig 4 ) ., We find that CWO and CLK bind E-boxes upstream of tim in a reciprocal manner during a daily cycle , and that CLK shows significantly increased binding intensity at the trough of its binding cycle in cwo mutant flies , indicating that CWO acts to antagonize CLK-CYC binding ., Given that both CWO and CLK are constitutively expressed ( Fig 1; 8 ) , we believe that the key driver for the transition between dynamic CLK-CYC and CWO binding is the accumulation of PER , which alters the relative affinity of E-box binding by CLK-CYC ., CWO shows low levels of tim E-box binding in per01 flies , in which CLK-CYC constantly bind E-boxes , but shows high levels of tim E-box binding in Clkout flies that lack CLK expression and E-box occupancy ., These results suggest that CWO E-box binding affinity is lower than the CLK-CYC heterodimer and higher than the CLK-CYC-PER complex , which could account for the PER-dependent rhythms in CLK-CYC and CWO binding ( Fig 6 ) ., During late day and early night , CLK-CYC binds E-boxes to activate transcription in the presence of CWO because CLK-CYC has higher DNA binding affinity ., PER starts to accumulate in the nucleus during the night and interacts with CLK-CYC , which decreases CLK-CYC DNA interaction via reduced DNA binding affinity ., Consequently , CWO displaces CLK-CYC-PER from E-boxes by binding with comparatively higher affinity ., Once CLK-CYC-PER is removed , CWO occupancy on E-boxes prohibits CLK-CYC-PER from re-binding , thus maintaining transcriptional repression ( Fig 6 ) ., Unlike the constitutive CLK-CYC E-box binding in per01 flies 8 , CLK-CYC binding is rhythmic in cwo5703 flies , but with a dampened amplitude due to elevated CLK binding at the trough ( Fig 4A ) ., This low amplitude rhythm in CLK binding may explain why a large proportion of cwo5703 flies show long period rhythms rather than losing rhythmicity entirely like per01 mutants 9 , 12–14 ., We speculate that the long period phenotype is caused in part by a prolonged repression process ., Based on the current model for repression of CLK-CYC transcription , PER-TIM complexes first bind CLK-CYC , thereby removing CLK-CYC from the E-boxes and inhibiting per and tim transcription , then PER and TIM degradation enables CLK-CYC binding to start another cycle of transcription 3 ., Both of these steps could be delayed in a cwo mutant ., In the absence of CWO it takes longer to remove CLK-CYC from the DNA; PER alone can repress CLK-CYC binding to some degree , but CLK-CYC-PER complexes still weakly bind E-boxes if CWO is absent , thus reducing CLK-CYC repression compared to wild-type flies ., The outcome of incomplete repression of CLK-CYC E-box binding would be an increase in the trough levels of per and tim mRNAs , which is exactly what was observed in cwo mutant and RNAi knockdown strains 9 , 12–14 ., Higher per and tim mRNA levels would in turn increase PER and TIM expression during the repression phase 14 ., Higher levels of PER and TIM would not repress CLK-CYC binding efficiently in the absence of CWO , but would take longer to be degraded , thereby delaying the next cycle of transcriptional activation ., In addition to the increased trough levels of core clock gene mRNAs in cwo mutant and RNAi knockdown flies , the peak levels of these mRNAs are lower , particularly during DD 9 , 12–14 ., Decreasing per mRNA levels also lengthen circadian period 37 , thus making it difficult to determine the extent to which a lower mRNA peak or increased mRNA trough contributes to period lengthening in cwo mutant and RNAi knockdown flies ., CLK binding at the peak of transcription is not significantly lower in cwo5073 than wild-type during LD ( Fig 4B ) , which argues that CWO enhances CLK-CYC transcriptional activity independent of CLK-CYC E-box binding ., Additional experiments will be needed to decipher the mechanism underlying this CWO dependent increase in CLK-CYC transcription ., PER dependent repression of CLK-CYC transcription is thought to occur in two stages ., First , PER is recruited to circadian promoters by interacting with CLK to form PER-CLK-CYC complexes “on-DNA” , which inhibit CLK-CYC dependent transcription via an unknown mechanism ., Subsequently , a decrease in the DNA binding affinity of PER-CLK-CYC complexes results in their release from DNA to initiate ‘‘off-DNA” phase of repression 35 ., According to our model , CWO is critical for the transition to , and maintenance of , off-DNA repression ., When PER-CLK-CYC complexes with low DNA affinity are formed , CWO promotes off-DNA repression by competing with CLK-CYC-PER complex for E-box binding ., CWO occupancy on E-boxes then prevents PER-CLK-CYC from re-binding , thereby maintaining off-DNA repression ., In mammals , a similar pattern of antagonistic binding on E-boxes between transcription factors was recently reported; USF1 and a mutant form of CLOCK , CLOCKΔ19 , bind to the same tandem E-boxes in a reciprocal manner ., Wild-type CLOCK-BMAL1 complex binds E-boxes with much higher affinity than USF1 , but CLOCKΔ19-BMAL1 binds E-boxes with a similar affinity to USF1 , thus allowing USF1 to bind E-boxes 31 ., Although this competitive binding is not thought to impact feedback loop function under normal circumstances , it demonstrates that other transcription factors can out-compete CLOCK-BMAL1 for E-box binding if the DNA binding affinity of CLOCK-BMAL1 is reduced ., In this case CLOCK-BMAL1 binding is compromised by the ClockΔ19 mutation , but other mechanisms such as interactions with repressors and protein modifications could also reduce the binding affinity of CLOCK-BMAL1 or its orthologs ., As in Drosophila , rhythmic binding of CLOCK-BMAL1 to E-boxes drives circadian transcription in mammals ( reviewed in 38 ) ., Recent ChIP-seq analyses in mouse liver revealed time-dependent binding of CLOCK , BMAL1 and key negative feedback components including PER1 , PER2 , CRY1 and CRY2 27 , 39–41 ., The mechanism underlying the dynamic DNA occupancy of these transcription factors is not known , but previous work shows that the PER2-CLOCK interaction is required to initiate repression of CLOCK-BMAL1 dependent transcription 42 , which suggests that CLOCK-BMAL1 may be removed from E-boxes by the same mechanism as CLK-CYC in Drosophila ., A recent genome-wide nucleosome analysis in mouse liver revealed that rhythmic E-box binding by CLOCK-BMAL1 removes nucleosomes 43 ., However , despite rhythmic CLOCK-BMAL1 binding , nucleosome occupancy on E-boxes is always well below surrounding sequences , even in Bmal1-/- mutant livers 43 ., This result indicates that chromatin at CLOCK-BMAL1 target sites is not closed even when there is no CLOCK-BMAL1 binding , suggesting that other transcription factors may occupy these E-boxes when CLOCK-BMAL1 is absent ., These results , taken together , suggest that rhythms in activator binding may be controlled by a common mechanism in Drosophila and mammals ., The mammalian orthologs of CWO , called DEC1 and DEC2 ( and also SHARP2 and SHARP1 , respectively ) , suppress CLOCK-BMAL1-induced activation 44–50 ., Gel mobility shift and ChIP assays in vitro revealed that both DEC1 and DEC2 bind to E-box motifs targeted by CLK-BMAL1 45–49 , and the DNA-binding domain is required for DEC1 to regulate CLK-BMAL1-induced transactivation 48 ., In addition , DEC1/2 shows synergistic activity to PER1 in the regulation of clock gene mRNA levels in the SCN , as exemplified by significant changes in the period of circadian activity rhythms when null mutants for Dec1 , Dec2 or both Dec1 and Dec2 are combined with that for Per1 44 ., In contrast to the constant levels of CWO , DEC1 protein is rhythmically expressed in mouse liver , where DEC1 levels are high when PER-CRY complexes repress CLK-BLMAL1 transcription 51 ., Taken together , these results raise the possibility that DEC1 and DEC2 may be a functional counterpart of CWO in competing with CLOCK-BMAL1 for E-box binding to repress CLOCK-BMAL1-mediated transcription ., DNA fragments containing wild-type or mutant E-boxes from the upstream tim circadian enhancer were used to construct GFP-reporter transgenes ., These 136bp fragments extend from -578 to -714 relative to the tim transcription start site , and contain “E1-E2” E-box motifs that are wild-type ( E1-E2 ) , E1 mutant ( mE1-E2 ) , E2 mutant ( E1-mE2 ) or E1-E2 mutant ( mE1-mE2 ) ., These wild-type and mutant E-box fragments were generated by PCR amplification using the following primer sets: E1-E2 , 5’-CACCTTTGGCAAATAAACGTGCGGCA-3’ and 5’-TGCCGGCGTTTGTGCGAA-3’; mE1-E2 , 5’-CACCTTTGGCAAATAAACGTGCGGCACGTTGTGATTAAGATCTAGCCGAT-3’ and 5’-TGCCGGCGTTTGTGCGAA-3’; E1-mE2 , 5’-CACCTTTGGCAAATAAGATCTCGGAGATTTGTGATTACACGTGAGCCGAT-3’ and 5’-TGCCGGCGTTTGTGCGAA-3’; mE1-mE2 , 5’-CACCTTTGGCAAATAAGATCTCGGAGATTTGTGATTAAGATCTAGCCGAT-3’ and 5’-TGCCGGCGTTTGTGCGAA-3’ ., The PCR products were inserted into the pENTR/D-TOPO vector using pENTR Directional TOPO cloning kit ( Invitrogen ) , and then subcloned into the pHPdesteGFP vector , which expresses Green Fluorescent Protein ( GFP ) according to the enhancer sequence inserted 28 , using Gateway LR-Clonase System ( Invitrogen ) ., The nucleotide sequences of all transgenes were confirmed by sequencing ., The resulting transgenes were injected into embryos ( BestGene ) for recombination into the attp18 genomic site via PhiC31-mediated transgenesis to yield tim circadian enhancer GFP ( tim-CEG ) flies 52–54 ., Flies were entrained in a 12-h light/12-h dark ( LD ) incubator for at least 3 days , collected at the indicated time points , and frozen ., Isolated frozen fly heads were homogenized in radioimmunoprecipitation assay ( RIPA ) buffer ( 20 mM Tris at pH 7 . 5 , 150 mM NaCl , 1 mM EDTA , 0 . 05 mM EGTA , 10% glycerol , 1% Triton X-100 , 0 . 4% sodium deoxycholate ) containing 0 . 5 mM PMSF ( phenylmethylsulfonyl fluoride ) , 10 μg/ml aprotinin , 10 μg/ml leupeptin , 2 μg/ml pepstatin A , 1 mM Na3VO4 , and 1 mM NaF ., This homogenate was sonicated 3 to 5 times for 10 s each time , using a Misonix XL2000 model sonicator at a setting of 3 and then centrifuged at 20 , 000 g for 10 min ., The supernatant was collected as RIPA S extract , and protein concentration was determined by the Bradford assay ., Equal amounts of RIPA S extract were run , transferred , and probed with guinea pig anti-CWO ( GP-27 ) , 1:5 , 000 and mouse anti-beta-actin ( Abcom ) , 1:20 , 000 ., Horseradish peroxidase-conjugated secondary antibodies ( Sigma ) against guinea pig and mouse were diluted 1:5 , 000 ., Immunoblots were visualized using ECL plus ( GE ) reagent ., Protein levels were measured by placing a rectangle of the same size over each CWO , ß-Actin or non-specific ( NS ) protein band on films used to visualize the immunoblots , and quantifying the signal within each rectangle via densitometric analysis using the ImageJ program ., The levels of CWO were calculated as a CWO:ß-Actin or CWO:NS ratio , and CWO abundance at each time point was plotted relative to wild-type at ZT2 ., Chromatin IP ( ChIP ) assays and qPCR quantification were performed as previously described 55 ., CLK and CWO binding to E-boxes in the circadian enhancers upstream of tim , per , vri , and Pdp1 in wild-type flies and the circadian enhancer in tim-CEG flies were first quantified via qPCR , and the resulting values were corrected for nonspecific binding to an intergenic region on chromosome 3R ( nucleotides 29576172 to 29576303 ) ., The primers used for qPCR were as follows: for tim E-boxes , 5’-ACACTGACCGAAACACCCACTC-3’ and 5’-GCGGCACGTTGTGATTACACG-3’; for per E-boxes , 5’-GGGTGAGTAATGCCGTTGCGAAAT-3’ and 5’-ATTTGCTGGCCAAGTCACGCAGTT-3’; for vri E-boxes , 5’-CTGGTGCCTCACATTCCACG-3’ and 5’- CAGCAGTCAAGTTATAGCAGCGC-3’; for Pdp1 E-boxes , 5’-GCACTCTCATTCTCTCTGTCGC-3’ and 5’-ACTTGGGGGACTGGAACTG-3’; for tim-CEG , 5’-GCCCCCTTCACCTTTGGCAAATA-3’ and 5’-TACAAGAAAGCTGGGTCGGCG-3’; and for the intergenic region , 5’-CAGGAGTCGVAGGACCAACC-3’ and 5’-GTCCTGAGAGGCTGAGAGGC-3’ ., PCR amplification using each pair of primers produced a single band of the expected size ., The tim-CEG primers target vector sequences that flank the genomic tim E-box insert , and thus do not amplify endogenous tim genomic sequences ., Quantitative RT-PCR was performed as described 55 , 56 , with some modifications , to measure GFP mRNA levels ., Total RNA was isolated from frozen fly heads using Trizol ( Invitrogen ) , and treated with a Turbo DNase DNA-free kit ( Ambion ) to eliminate genomic DNA contamination ., DNA-free total RNA ( 1 . 0 μg ) was reverse transcribed using oligo ( dT ) 12–28 primers ( Invitrogen ) and Superscript II ( Invitrogen ) ., The reverse transcription ( RT ) product was amplified with SsoFast qPCR Supermix ( Bio-Rad ) in a Bio-Rad CFX96 Real-Time PCR System using primers to GFP ( 5’-TACGGCAAGCTGACCCTGAAGT-3’ and 5’-CGCACCATCTTCTTCAAGGACG-3’ ) and ribosomal protein 49 ( rp49 ) ( 5’-TACAGGCCCAAGATCGTGAA-3’ and 5’-GCACTCTGTTGTCGATACCC-3’ ) ., For each sample , mRNA quantity was determined using the standard curve for each gene analyzed ., To determine the relative levels of GFP mRNA over a diurnal cycle , GFP mRNA levels were divided by rp49 mRNA levels for each time point and plotted as the GFP/rp49 mRNA ratio ., To quantify GFP mRNA in different tim-CEG strains at the wild-type ( E1-E2 ) peak , GFP/rp49 values were normalized to the E1-E2 value at ZT14 . | Introduction, Results, Discussion, Materials and Methods | The Drosophila circadian oscillator controls daily rhythms in physiology , metabolism and behavior via transcriptional feedback loops ., CLOCK-CYCLE ( CLK-CYC ) heterodimers initiate feedback loop function by binding E-box elements to activate per and tim transcription ., PER-TIM heterodimers then accumulate , bind CLK-CYC to inhibit transcription , and are ultimately degraded to enable the next round of transcription ., The timing of transcriptional events in this feedback loop coincide with , and are controlled by , rhythms in CLK-CYC binding to E-boxes ., PER rhythmically binds CLK-CYC to initiate transcriptional repression , and subsequently promotes the removal of CLK-CYC from E-boxes ., However , little is known about the mechanism by which CLK-CYC is removed from DNA ., Previous studies demonstrated that the transcription repressor CLOCKWORK ORANGE ( CWO ) contributes to core feedback loop function by repressing per and tim transcription in cultured S2 cells and in flies ., Here we show that CWO rhythmically binds E-boxes upstream of core clock genes in a reciprocal manner to CLK , thereby promoting PER-dependent removal of CLK-CYC from E-boxes , and maintaining repression until PER is degraded and CLK-CYC displaces CWO from E-boxes to initiate transcription ., These results suggest a model in which CWO co-represses CLK-CYC transcriptional activity in conjunction with PER by competing for E-box binding once CLK-CYC-PER complexes have formed ., Given that CWO orthologs DEC1 and DEC2 also target E-boxes bound by CLOCK-BMAL1 , a similar mechanism may operate in the mammalian clock . | Circadian clocks control daily rhythms in animal , plant and fungal physiology , metabolism and behavior via transcriptional feedback loops ., In Drosophila , the CLOCK-CYCLE ( CLK-CYC ) activator complex binds E-box regulatory sequences to initiate transcription of hundreds of effector genes including their own repressors , PERIOD ( PER ) and TIMELESS ( TIM ) , which feed back to repress CLK-CYC until they are degraded , thus allowing another cycle of CLK-CYC activation ., Although the repression process is critical for the stability and accuracy of circadian timekeeping , how PER-TIM complexes maintain a transcriptionally repressed state for many hours is not well understood ., Here we demonstrate that the transcription factor CLOCKWORK ORANGE ( CWO ) antagonizes CLK-CYC E-box binding , thus enhancing the removal of CLK-CYC from E-boxes to maintain transcriptional repression ., This process requires PER , which suggests that PER-TIM and CWO cooperate to maintain a transcriptionally repressed state by removing CLK-CYC from E-boxes ., These results demonstrate that PER-TIM requires CWO to effectively repress circadian transcription , and given that circadian transcriptional regulators are well conserved , this mechanism may function to repress transcription in other animals including humans . | invertebrates, gene regulation, regulatory proteins, messenger rna, dna-binding proteins, animals, dna transcription, circadian oscillators, animal models, drosophila melanogaster, model organisms, transcription factors, chronobiology, drosophila, research and analysis methods, proteins, gene expression, insects, arthropoda, biochemistry, rna, circadian rhythms, nucleic acids, genetics, biology and life sciences, organisms | null |
journal.pgen.1003579 | 2,013 | Evidence for Two Different Regulatory Mechanisms Linking Replication and Segregation of Vibrio cholerae Chromosome II | Studies in bacteria as well as in eukaryotes have shown that processes that maintain chromosomes , such as replication , recombination and repair , although able to occur independently of each other , often influence each other ., Chromosome segregation is a major maintenance process but our knowledge of it in bacteria is relatively recent ., This is because in well-studied bacteria such as Escherichia coli , genes dedicated to the segregation process have not been evident ., Such genes were discovered in bacterial plasmids , called parA and parB ., Subsequently , their homologs were found in a majority of sequenced bacterial chromosomes 1 , 2 ., Wherever tested , the chromosomal parAB genes were capable of conferring segregational stability on unstable plasmids bearing parS ( “centromere” analogous ) sites 3 , 4 , 5 , 6 , and made at least some contribution to chromosome segregation 7 , 8 ., In spite of limited study , it is becoming clear that chromosomal segregation systems can influence and be influenced by other chromosome maintenance processes ., In bacteria , replication and transcription have been proposed to provide motive force in chromosome segregation 9 , 10 , 11 ., Coupled transcription-translation of membrane proteins is also thought to play an important role in chromosome segregation 12 , 13 ., One of the segregation proteins , ParB , can also spread and silence transcription of genes in its path 14 , 15 , 16 ., The influence of segregation proteins in replication was suggested when ParB was found to load a condensin protein in the vicinity of the replication origin in Bacillus subtilis and in Streptococcus pneumoniae 17 , 18 , 19 ., A more direct role was evident when ParA was found to influence the activity of the initiator DnaA in B . subtilis chromosome replication 20 , 21 and in replication of Vibrio cholerae chromosome I ( chrI ) 22 ., Recently , ParB encoded by V . cholerae chromosome II ( chrII ) was also found to influence chrII replication 23 ., Here we report two distinct mechanisms for this ParB-mediated effect ., V . cholerae chrII replication is primarily controlled by its specific initiator protein , RctB 24 , 25 ., RctB binds to two kinds of site in the replication origin of chrII ., One kind , the 11- or 12-mers , plays both essential and regulatory roles 26 ., The other kind , two 39-mers and a 29-mer ( a truncated 39-mer ) , plays only an inhibitory role in replication 26 , 27 ., One of the 39-mers is situated at a locus called rctA , at one end of the origin , and the other is more centrally located in the origin ( Figure 1 , top ) ., The 29-mer is located in front of the rctB gene and is involved primarily in autorepression of the gene 27 ., The rctA locus contains , in addition to a 39-mer , one of the ParB2 binding sites , parS2-B 28 ., It has been reported recently that parS2-B alleviates some of the replication inhibitory activity of rctA in a ParB2-dependent fashion , but the mechanism is unknown 23 ., Here we show that ParB2 spreads from parS2-B into the rctA 39-mer , and suggest that the spreading likely interferes with RctB binding to the 39-mer and thereby restrains the inhibitory activity of rctA ., Unexpectedly , we also found ParB2 promotes replication by directly binding to the central 39-mer , without requiring spreading from parS2-B ., We provide evidence that ParB2 competes with RctB for binding to the central 39-mer specifically and could thereby restrain its activity ., In addition to revealing new ways by which a Par protein might influence replication , our results are significant in demonstrating that a segregation protein can bind specifically outside of centromeric sites ., The origin region of chrII comprises three functional units: A region required for controlling initiation ( incII ) , a region minimally required for initiation ( oriII ) and a gene required for synthesizing the initiator protein ( RctB ) ( Figure 1 , top ) ., The region covering the first two units , which consists mostly of sites for initiator binding , will be referred to as the origin ., The rctA locus of incII exerts a strong inhibitory effect on chrII replication because of the 39-mer it contains 26 ., The locus also has a site , parS2-B , for binding to the segregation protein ParB2 , and the inhibitory effect of rctA is reduced in the presence of ParB2 23 ., Knowing that ParB proteins can spread out from their binding sites to neighboring sequences 14 , we asked whether ParB2 spreading from parS2-B over the 39-mer might be a mechanism to control its inhibitory function ., The extent of ParB2 spreading was tested by ChIP-chip analysis using antibody against ParB2 ( Figure 1 , bottom ) ., Spreading was evident on either side of parS2-B ( grey profile ) ., In contrast , when the antibody was against RctB , the immunoprecipitated DNA in the origin region was restricted to where RctB has specific binding sites ( black profile ) ., These results suggest that ParB2 has the potential to modulate replication initiation activity by interfering with RctB binding ., Spreading of proteins across genes can silence them 14 ., For example , spreading into plasmid replication genes suitably close to a parS site can be lethal when selective pressure for plasmid retention is applied ., By such an experimental test , we found that ParB2 can spread from parS2-B ( Figure S1A ) ., Spreading is also suggested by the formation of ParB2-GFP fluorescent foci in parS2-B carrying plasmids ( Figure S1B ) ., ParB2 could also silence two promoters , PrctA and PrctB , in the origin of chrII ( Figure 1 , top ) ., PrctA is proximal to parS2-B whereas PrctB is located about 1 kb away at the other end of the origin ., The activity of the promoters was assayed by fusing them to a promoter-less lacZ gene present in a multicopy plasmid in E . coli ( Figure 2 ) ., The promoter fragments fused to lacZ carried either the entire origin including parS2-B ( 1A and 1B ) , or the origin lacking parS2-B ( 2A and 2B ) or no additional DNA ( 3A and 3B ) ., ParB2 was supplied constitutively in trans at about an order of magnitude higher than the physiological level ( monitored by Western blotting; Figure S2 ) , using Ptrc promoter without an intact lac repressor binding-site ( lacO1 ) , which makes the promoter unresponsive to IPTG ., The presence of ParB2 reduced the activities of PrctA and PrctB significantly , only when the parS2-B site was present ( Figure 2 , 1A and 1B ) ., These results suggest that ParB2 can spread over the entire origin in the presence of parS2-B , and does not have a significant effect on either promoter in the absence of parS2-B ., Since RctB has numerous binding sites in the origin , it appeared possible that RctB binding to them could counteract ParB2 spreading and reduce silencing of the promoters ., This possibility was addressed by supplying RctB from an arabinose-inducible promoter , PBAD 24 , 26 , 27 , 29 ., The induction of RctB alone ( at about two-fold the physiological level ) repressed PrctA marginally and PrctB about two fold ( Figure S3 , lanes 1 vs . 5 ) ., Silencing by ParB2 exceeded 90% for both the promoters ( Figure S3 , lanes 1 vs . 4 ) ., When RctB and ParB2 were supplied together , the repression of both the promoters was reduced about two fold compared to the level achieved with ParB2 alone ( Figure S3 , lanes 4 vs . 6; also inset ) ., These results indicate that RctB can counteract the ParB2-mediated silencing ., In the results presented above ( Figure 2 and Figure S3 ) , the level of ParB2 was about 14-fold the level normally present in V . cholerae ( Figure S2 ) ., When the concentration was reduced to about 10-fold , ParB2 could silence only the parS2-B proximal PrctA , but not the distal PrctB promoter ( Figure S3 , lanes 1 vs . 3 ) ., This reduced level of ParB2 was used in all subsequent experiments ., In order to determine how far ParB2 can spread beyond PrctA , progressively increasing lengths of incII were fused to a foreign reporter promoter , PrepA , itself fused to lacZ 30 ( Figure 3 ) ., In these experiments , in addition to a plasmid supplying ParB2 , another plasmid was used to supply RctB ., The two proteins were expressed from inducible promoters , Plac and PBAD , respectively ( Figure 3 , cartoon at the top right corner ) ., As expected , neither ParB2 nor RctB influenced the activity of PrepA itself ( Figure 3 , top panel ) ., In contrast , when rctA was present , ParB2 reduced the activity of PrepA by two-fold ( second panel ) ., We believe this is due to ParB2 spreading from parS2-B into PrepA , whose −35 box was only 167 bp away ., RctB alone was ineffective , most likely because it does not spread , and its specific binding site , the rctA 39-mer , is well separated ( by 85 bp ) from the −35 box of PrepA ., Supplying RctB was marginally effective in relieving the silencing by ParB2 ., The next extension of the incII fragment included the 3x11-mers ( third panel ) ., Neither ParB2 nor RctB could silence the reporter promoter in this case ., This result suggests that the spreading may not extend too far beyond PrctA ., A further extension of the incII fragment by only 74 bp that included the central 39-mer , restored ParB2-mediated repression of the reporter promoter ( fourth panel ) ., This result was surprising since parS2 sites were not found within incII 28 ., The effect of ParB2 was significantly reduced when RctB was supplied , which is to be expected since RctB binds strongly to the central 39-mer 26 ., This result suggests that ParB2 and RctB can compete for binding to the central 39-mer ., The largest fragment ( bottom panel ) did not show a significant ParB2 effect on the reporter promoter , suggesting that ParB2 may not spread significantly from the central 39-mer ., Together , the results suggest that under the conditions tested , ParB2 affects the origin primarily through interactions near rctA and the central 39-mer ., The interaction near the central 39-mer suggests that ParB2 might bind there directly ., The possibility of site-specific binding of ParB2 within the origin but outside of parS2-B was tested by EMSA ., Several fragments covering the origin were used ., Fragments 1 and 2 , carrying the parS2-B site ( positive controls ) , showed maximal ParB2 binding ( Figure 4 ) ., The next significant binding was with the fragment containing the central 39-mer ( fragment 5 ) ., This fragment contained natural flanking sequences of only 3 bp and 32 bp beyond the central 39-mer ., The sequences ( 3+39+32 ) are exactly those that were added to the incII fragment of the third panel to generate the silencing-proficient fragment of the fourth panel ( Figure 3 ) ., We found that the flanking sequences do not contribute to the central 39-mer binding ( Figure S4 , fragment #1 ) ., This result supports the inference from in vivo studies that ParB2 can directly bind to the central 39-mer without requiring parS2-B ., Binding to the rctA 39-mer ( fragment 3 ) was considerably weaker , possibly because the two 39-mers have several mismatches between them ( Figure 1 of 26; discussion related to Figure 5 below ) ., The level of binding seen with the rctA 39-mer was comparable to the levels seen with fragments 4 , 6 , 8 and 9 , and the level was marginally above that of the negative control that lacks any chrII sequences , suggesting that ParB2 has significant non-specific DNA binding activity ., The sequence requirement for specific binding of ParB2 to the central 39-mer was tested by variously mutating the sequence ., The 39-mer has two conserved 9 bp direct repeats ( called A and B boxes ) flanking a 19 bp AT-rich spacer ( Figure S4 ) ., The presence of both of the repeats and their proper phasing are important for RctB binding 26 ., The AT richness of the spacer is also important but not the sequence per se ., The parS2 sites are AT-rich inverted repeats , only 15 bp long ., Notably , the 39-mer spacer also contains an inverted repeat , which has some similarity to the consensus parS2 site ., However , the 39-mer spacer by itself was not sufficient for ParB2 binding; the presence of one of the direct repeats was necessary ( Figure S4 , fragments 4–6 ) ., Either of the direct repeats alone was also not sufficient ( fragments 2–3 ) ., The inverted repeat feature could also be destroyed without compromising the binding efficiency ( fragment 12 ) ., When the same fragments were tested for RctB binding , only the ones with the intact 39-mer and 10 bp deletion or addition ( fragments 9–11 ) showed significant binding , as was also found earlier ( data not shown; 27 ) ., It appears that while both ParB2 and RctB bind to the 39-mer , the presence of one of the direct repeats is not obligatory for ParB2 binding ., Specific binding of ParB2 to the 39-mer was also verified by DNase I footprinting ( Figure S5 ) ., Protection by ParB2 was conspicuous at the junction of the first direct repeat ( A-box ) and the AT-rich spacer ., At this junction an intact parS-2B half site , 5′-TGTAAA , is present ., This sequence is fully conserved in all 10 parS2 sites that were competent in ParB2 binding 28 ., In Figure S4 , this half-site sequence was intact in all the binding positive fragments and mutant in all the fragments that failed to show specific binding ., The half-site is also mutated to 5′-TTAAAC in the ParB2 binding-negative 39-mer in rctA ( Figure 4 , fragment #3 ) ., The half site thus appears to be necessary for 39-mer binding of ParB2 ., In further support of this inference , when we restored the original bases to some of the binding-negative 39-mer mutants to regenerate the half-site , binding proficiency was regained ( Figure 5 , fragments #3–6 ) ., Although necessary , the half site was not sufficient for binding ParB2 ( fragment #7 ) ., We conclude that extension of the half site either to the left or right is necessary ., This is not surprising since the affinity drops by orders of magnitude when one half of a dyad symmetric site is mutated 31 , 32 ., The minimal size of the extensions needed to regain binding activity of ParB2 remains to be determined ., ParB2 and RctB can bind to rctA simultaneously 23 ( Figure 6 , top panel ) ., This is not surprising since there are 34 bp of spacer sequence between the binding sites of the two proteins , and that ParB2 does not spread in vitro ., The sites also remain functional when isolated from each other 33 ., On the other hand , at the central 39-mer , the binding sites for the two proteins appeared to be largely overlapping , suggesting that they could not bind simultaneously ., This was indeed the result ( Figure 6 , bottom panel ) ., Even at the higher protein concentrations ( ++ ) , no new discrete species representative of dual binding was detected ., The results indicate that ParB2 and RctB compete for binding to 39-mer , unlike the simultaneous binding that can occur on rctA ., We previously showed that the central 39-mer is the most potent replication inhibitory site in incII and it functions through RctB binding 26 ., If ParB2 competes with RctB for binding to the central 39-mer , this competition appeared likely to influence oriII activity without requiring the parS2-B site ., This prediction was tested by determining the copy number of oriII-driven plasmids ( Figure 7 ) ., The copy number of oriII plasmids depends on the extent of the incII sequences present 26 ., Although the 39-mers are always inhibitory to replication , the 11- and 12-mers can either promote or inhibit replication depending upon whether the 39-mers are present or not ., In the present experiments also , the oriII plasmid copy number first decreased and then increased with increasing deletion of incII ( Figure 7 , − ParB2 column ) ., When ParB2 was additionally present , the copy number increased significantly in the first two 39-mer-carrying plasmids , the increase being maximal for the plasmid with the lowest copy number ( pTVC25 ) ., In this plasmid , we suggest that the 39-mer was unencumbered by the 3x11-mers , and was maximally available for binding to ParB2 ., Together , these results indicate that ParB2 has the potential to facilitate chrII replication by restraining the inhibitory activity of the incII sequences , and can do so whether parS2-B is present or not ., If ParB2 spreading is one of the mechanisms by which the protein stimulates chrII replication , it might be possible to restrain this activity by placing a roadblock in the path of spreading ., To this end , we inserted an array of five P1 RepA binding sites ( iterons ) between parS2-B and the 39-mer in rctA ( Figure 8 ) ., The effectiveness of the roadblock in preventing the spreading of P1 ParB protein was demonstrated earlier 34 ., Comparison of the top two rows of the Table in Figure 8 shows that in the absence of RepA ( that is in the absence of a roadblock ) , ParB2 was equally efficient in promoting cell growth that depended on the functioning of oriII plasmids ., In other words , the P1 iterons in pBJH218 did not compromise ParB2 spreading in the absence of the roadblock ., The same two plasmid-carrying cells behaved differently in the presence of RepA ( the last two rows ) ., Upon induction of ParB2 production by IPTG , cell growth improved more in the case of pTVC20 than in the case of pBJH218 ., In other words , ParB2 effect was compromised under the condition the roadblock was expected to be effective ., These results are consistent with ParB2 spreading as a mechanism for stimulating chrII replication initiation ., Note that some increase of growth rate was seen even when ParB spreading was inhibited by a roadblock ( generation time decreased 7% for cells in row #4 ) ., This result is not surprising because ParB2 can bind to the central 39-mer without requiring spreading from parS2-B ., Overall , the ParB2 effects were modest , which is to be expected because of the existence of multiple controls on chrII replication ., In chromosome and plasmid segregation , ParB proteins serve to couple centromeres to ParA proteins ( NTPases ) and modulate the NTPase activity that is believed to provide the movement required for segregation 8 ., The binding ParB2 to a 39-mer raises the possibility of an inherent centromeric function of the site ., This was tested by cloning the central 39-mer into a miniF plasmid , which is unstable due to deletion of its own segregation genes 35 ., The stability of the miniF plasmid improved with the inclusion of the parS2-B site but not with the 39-mer , when ParA and ParB proteins were supplied in trans ( Figure S6 ) ., This result suggests that ParB2 binding to parS2-B and the central 39-mer is different in an important respect ., The spreading of ParB2 from a centromeric site into the origin of chrII was evident from in vivo cross-linking experiments ( Figures 1 , S7 ) , from silencing of promoters within the origin ( Figure, 2 ) and from the reduction of reporter promoter activity when natural initiator ( RctB ) binding sites were present between the centromeric site and the promoter ( Figure 3 , third panel; Figure S3 , insets ) ., This latter result indicates that RctB binding could create a natural roadblock to ParB2 spreading ., The fact that the span of silencing lengthened with increased ParB2 concentration ( Figure S3 ) also supports the idea that the underlying mechanism involves spreading along the DNA ., Finally , the results of placing an artificial roadblock were also consistent with the spreading mechanism ( Figure 8 ) ., When a powerful replication inhibitory site ( the rctA 39-mer ) was present within the span of spreading , growth of cells dependent upon the functioning of the chrII origin improved ., It was also reported earlier that ParB2 could increase replication of chrII origin carrying plasmids when they included the adjacent rctA region 23 ., This increase was shown to be dependent on the presence of parS2-B ., Together with the finding that ParB2 does not directly bind to the rctA 39-mer ( Figure 4 ) , and cannot spread from its binding site in the central 39-mer as discussed below ( Figures 3 ( last panel ) , S1A , S1B ) , the simplest explanation of these results is that by spreading from parS2-B , ParB2 compromises the inhibitory activity of the rctA 39-mer by interfering with RctB binding ., ParB2 was also found to reduce the activity of another potent replication inhibitor ( the central 39-mer ) without requiring the centromeric site and spreading ( Figure 3 , S1 ) ., The latter effect appears to be due to direct binding of ParB2 to the central 39-mer ., This mode of ParB2 interaction with the 39-mer most likely also causes interference with RctB binding to this site ( Figure 6 ) ., Since the 39-mers are the two sites most inhibitory to chrII replication and their activities are mediated through RctB binding , the reduction in binding suffices to explain how ParB2 could promote replication of chrII ( Figures 7 , 8 ) ., Interference with binding of regulatory proteins to DNA by the spreading of a competing protein along DNA has also been invoked to explain transcriptional silencing , inhibition of DNA methylation and of DNA gyrase binding , and resistance to DNase I cleavage 37 , 38 , 39 , 40 ., Although the only model we have entertained so far to explain the ParB2 effect is interference with specific binding of RctB , we have also tested whether ParB2 and RctB could interact directly ., This possibility was suggested by the finding that ParA influences replication by protein-protein interaction rather than DNA-protein interaction 20 , 21 , 22 , 36 ., However , ParB2 did not show any detectable interaction with RctB , as was also reported earlier ( Figure S8 ) 23 ., ParA participates in a number of processes involving ParB 8 ., Here , we asked whether binding of ParB2 to the central 39-mer is also influenced by ParA2 ., The binding was assayed indirectly by fusing a foreign promoter close enough to the 39-mer that ParB2 binding to the site could interfere with the promoter activity ., The promoter activity did not change significantly upon supply of ParA2 ( Figure S9; data with PrepA ) ., This suggests ParB2 binding to the 39-mer is not influenced by ParA2 ., We did find a minor influence of ParA2 on PrctA silencing by ParB2 spreading , the basis of which was not explored ., In the case of P1 plasmid and B . subtilis chromosome , no ParA effect on ParB spreading was evident 14 , 41 ., Whereas deletion of parS2-B in V . cholerae was easily tolerated ( Figure S7; 23 ) , deletion of the parB2 gene was essentially lethal 42 ., In the absence of ParB2 , chrII loss is evident at every cell division that causes a severe growth defect ., We were therefore unable to test conveniently the role of ParB2 on replication of chrII in the native host ., On the other hand , there was no obvious effect of ParB2 on the growth of E . coli , in which we did most of our experiments ., The validity of extrapolating the observations in E . coli to the native host appears warranted by the observation that ParB2 effects were seen in the context of the entire origin ( Figure 8; 23 ) , and by the finding that some of the inferences from the E . coli results were valid when tested in vitro ( Figure 6 ) ., In the past , wherever chrII replication control was studied in both E . coli and V . cholerae , the results agreed 25 , 26 , 28 , 29 ., Nonetheless , the ParB2 concentration required for spreading to proceed just over rctA was an order of magnitude higher than that is normally present in the native host ., The reason for this discrepancy is not understood but a possibility is that ParB2 when supplied from a trans source is much less effective ., A discrepancy in the amount of protein required from a cis vs . trans source has been noted in the case V . cholerae ParA1 22 and ParA of Pseudomonas aeruginosa 43 ., The production of one of the Par proteins without its partner could also have altered the protein activity and stability ., The importance of maintaining the stoichiometry of Par proteins has also been indicated in studies of B . subtilis 20 , 44 ., Another possibility is that higher protein concentration may be required to bind to a single parS site , as we have used here , than when there exists neighboring sites , as in the native host , that might allow cooperative binding ., We show that in wild type V . cholerae cells ParB2 can bind and spread over the entire origin ( Figure 1 ) ., We detected considerable ParB2 spreading , even with the deletion of the origin proximal centromeric site ( parS2-B ) ( Figure S7 ) ., Most likely , this spreading originates from the neighboring parS2-A site 5 . 7 kb away ( Figure 1 ) ., The spreading could add an additional layer of control over PrctA by silencing the promoter , which is independently repressed by RctB ( Figure S3 ) 24 , 45 ., The PrctA activity in turn controls RctB binding to the rctA 39-mer 29 ., The multiple feedback loops that operate to control the initiation of replication from the origin of chrII appear securely interlocked with the specific segregation system of this chromosome ., The presence of multiple layers of control could compensate for a deficiency in any one of the regulators , and help in homeostasis of origin copy number ., The finding that ParB2 could spread over the entire origin might suggest that it could be a mechanism to promote chromosome segregation ., It might increase the effective size of the kinetochore , which might facilitate its interaction with ParA , the essential partner of ParB in chromosome segregation ., However , this role has yet to be established 16 , 34 , 46 ., ParB proteins of plasmids are known to be plasmid-specific and to bind to their cognate sites 47 ., This helps to avoid segregation-mediated incompatibility if different plasmids happen to be present in the same host ., By the same token , in multichromosome bacteria , the segregation systems should be chromosome-specific ., Such is clearly the case in Burkholderia cenocepacia 5 and in V . cholerae 4 , 48 ., The same ParB protein has been found to bind to variant parS sites but the sites are believed to be descendants of a common ancestor 49 ., In this context , it is noteworthy that although the central 39-mer is largely non-homologous to parS2-B , the region of the 39-mer crucial for ParB2 binding shares six bp of perfect identity with parS2-B ( Figures 5 , S4 ) , suggesting the possibility of an evolutionary link between the sites here also ., Chromosome segregation begins soon after replication initiation , thereby compressing the total time for the completion of these two processes ., Their close coordination also allows segregation to proceed in a more orderly fashion than if the substrate for segregation were a pair of completed and entangled sister chromosomes ., Here , we have described interactions that might assist in coordinating replication initiation and segregation ., In V . cholerae , following replication initiation , the majority of the RctB binding sites ( 11- and 12-mers ) stay hemi-methylated and are unable to bind the initiator 50 ., This stage of the cell cycle should favor spread of ParB2 into the origin ( Figure S7 ) , which is likely to favor origin segregation and at the same time discourage premature reinitiation ., Spreading of ParB2 towards the origin is apparently prevented later in the cell cycle when the origin is remethylated , allowing RctB binding to 11- and 12-mers that eventually leads to initiation ( Figure 3 , 3rd panel ) ., At these latter stages , when spreading is blocked , direct binding of ParB2 to the central 39-mer should favor initiation ., Thus , depending upon the stage of the cell cycle , ParB2 appears to play opposite roles in controlling chrII replication , but in such a way as to promote the orderly sequence of chromosome replication followed by segregation ., In the case of plasmids , which can complete replication in a tiny fraction of the cell division cycle , such coordination is neither necessary nor evident ., We suggest that the acquisition of interactions such as we describe are a feature of the putative adaptation of an acquired plasmid to permanent residency as a second chromosome ., V . cholerae and E . coli strains , and plasmids used in this study are listed in Table S1 ., ChrII fragments were amplified from N16961 ( CVC209 ) DNA by PCR using Phusion High-Fidelity polymerase ( NEB , Beverly , MA ) ., The sequences of primers used for PCR are shown in Table S2 ., For cloning sequences up to 100 bp , complementary oligonucleotides ( IDT , Skokie , IL ) were used after annealing the two 29 ., The exact chrII coordinates of each cloned fragment are given in Table S1 ., This was done in L broth cultures of E . coli strain BR8706 at OD600 between 0 . 4–0 . 5 , as described 24 ., To account for any effect that ParB2 might have on the replication of the lacZ-reporter plasmid , β-galactosidase activities were normalized for the plasmid copy number in all cases ., The copy number variation was small; one standard deviation was within 20% of the mean ., The copy numbers were measured ( see below ) from aliquots of the same cultures that were used for β-galactosidase measurements ., Some of the cultures were simultaneously monitored for ParB2 amounts by Western blotting ( Figure S2 ) ., Note that +/− ParB2 refer to cells carrying pTVC501 ( that carries parB2 under IPTG control ) with and without IPTG induction , respectively ., In Figure 2 and Figure S3 ( lanes 4 , 6 ) , + refers to cells carrying pTVC236 , which supplies ParB2 from a constitutive promoter and − refers to cells carrying the empty vector , pACYC184 ., The copy number of lacZ-carrying plasmids ( Figures 2 , 3 , S3 , S9 ) were measured exactly as described 32 ., Briefly , different experimental cultures were grown to log phase and mixed with separately grown cells carrying pNEB193 before plasmid isolation ., The latter plasmid helped to account for plasmid loss , if any , during plasmid isolation steps ., The copy number of oriII plasmids ( Figure 7 ) was determined similarly except that cells instead of growing in liquid cultures were obtained by washing out colonies from transformation plates directly , to avoid mutant accumulation , as described 26 ., The origin fragments were first cloned in a plasmid vector driven by the γ-origin of plasmid R6K , and the clones were maintained in cells that supplied the cognate initiator ( π ) protein ., The clones were electroporated into E . coli ( BR8706 ) carrying pTVC499 that supplied RctB ( but no π protein ) and pTVC501 that supplied parB2 ., The DNA probes were made from plasmids by PCR using oligonucleotides TVC286 ( 5′-TCCGATTACGGCACCAAATCGA-3′ ) and TVC287 ( 5′- AACGTGGATAAACTTCCTGTAAT-3′ ) , which allowed amplification of extra 100 bp of vector sequences from each flank of the region of interest ., The PCR products were labeled using 30 units T4 Polynucleotide Kinase ( NEB ) and 50 µCi of γ32-ATP ( Perkin-Elmer ) and purified by passing through G-50 columns ( Roche diagnostics ) ., Binding was done in the presence of 300 ng poly dI-dC ., Other details are as described 25 , except that the binding reactions were run in 0 . 5×TBE , which improved ParB2 binding ., In Figure 5 , the probe was non-radioactive and was visualized with SYBR Gold nucleic acid gel stain ( Molecular Probes ) at 0 . 5 mg/ml for 30 min ., As non-specific competitor , supercoiled pUC19 DNA was used instead of poly dI-dC , as the former stayed at the top of gels and did not interfere with visualization of probe bands ., The images were recorded using Fuji LAS-3000 imaging system ., ChIP assay was performed as described 50 ., Briefly , cells of V . cholerae CVC209 were cultivated in L broth at 37°C to exponential phase and cross-linked with 1% formaldehyde ., After cell lysis and sonication , RctB-DNA or ParB2-DNA complexes were immunoprecipitated using RctB or ParB2 antibody , respectively ., The precipitat | Introduction, Results, Discussion, Materials and Methods | Understanding the mechanisms that coordinate replication initiation with subsequent segregation of chromosomes is an important biological problem ., Here we report two replication-control mechanisms mediated by a chromosome segregation protein , ParB2 , encoded by chromosome II of the model multichromosome bacterium , Vibrio cholerae ., We find by the ChIP-chip assay that ParB2 , a centromere binding protein , spreads beyond the centromere and covers a replication inhibitory site ( a 39-mer ) ., Unexpectedly , without nucleation at the centromere , ParB2 could also bind directly to a related 39-mer ., The 39-mers are the strongest inhibitors of chromosome II replication and they mediate inhibition by binding the replication initiator protein ., ParB2 thus appears to promote replication by out-competing initiator binding to the 39-mers using two mechanisms: spreading into one and direct binding to the other ., We suggest that both these are novel mechanisms to coordinate replication initiation with segregation of chromosomes . | Replication and segregation are the two main processes that maintain chromosomes in growing cells ., In eukaryotes , the two processes are restricted to distinct phases of the cell cycle ., In bacteria , segregation follows replication initiation with a modest lag ., Influences of one process on the other have been postulated ., The act of replication has been suggested to provide a motive force in chromosome segregation ., Moreover , segregation proteins ( ParA ) have been found to interact with and control the replication initiator , DnaA ., Here we show that in V . cholerae chromosome II , which is believed to have originated from a plasmid , a centromere binding protein ( ParB ) could control replication by two distinct mechanisms: spreading from a centromeric site into the replication-control region , and direct binding to the primary replication-control site , which has limited homology to the centromeric site ., These studies establish that Par proteins can influence replication by at least three mechanisms ., Homologous Par proteins participate in plasmid segregation but they are not known to influence plasmid replication ., The expanded role of Par proteins appears likely to have been warranted to coordinate chromosomal replication and segregation with the cell cycle , which appears less of an issue in plasmid maintenance . | biology | null |
journal.pbio.1000596 | 2,011 | The Formation of the Bicoid Morphogen Gradient Requires Protein Movement from Anteriorly Localized mRNA | Accurate development of metazoan embryos requires precise production , reception , and interpretation of patterning cues along appropriate spatial axes on realistic timescales ., Many embryonic patterning events utilize graded spatial distributions of patterning molecules , or morphogens , whose activities rely fundamentally on molecular interactions susceptible to environmental fluctuations and stochasticity in gene expression 1 ., Despite the widespread occurrence of morphogen-mediated tissue patterning , in most cases little or no quantitative data exist regarding the dynamics of gradient establishment or the spatio-temporal regulation of morphogen production ., The wealth of molecular and genetic tools available in Drosophila melanogaster offers an optimal context in which to study the basis of developmental accuracy in embryonic patterning by morphogens ., In the Drosophila embryo , anterior-posterior ( AP ) axial patterning originates with maternal cues deposited into the developing egg 2 ., Among these cues is the transcription factor Bicoid ( Bcd ) , the mRNA of which localizes at the anterior cortex of the oocyte 3–6 ., Translation of bcd mRNA is believed to commence upon fertilization , after which the embryo undergoes 13 rapid nuclear mitotic cycles ( n . c . ) without cytokinesis ., By the start of interphase 14 about 2 h after egg deposition ( AED ) , the embryo consists of a syncytial blastoderm layer of about 6 , 000 nuclei at the cortical surface , surrounding the interior core of yolk and vitellogenic nuclei ., During the blastoderm stage Bcd protein distributes along the AP axis as an exponentially decaying gradient 7–9 , and nuclear Bcd activates target genes in a dosage-dependent manner ( reviewed in 10 ) ., Recently , quantitative analysis of living embryos expressing Bcd-GFP revealed that the protein gradient remains nearly unchanged subsequent to the arrival of nuclei to the cortex at n . c . 10 ( about 85 min AED ) , and therefore that the nuclear gradient achieves stability within only about 80 min at 25°C 9 ., Moreover , nuclear gradients at n . c . 14 exhibit remarkable reproducibility between embryos: along the AP axis , similarly positioned nuclei in different embryos contain Bcd-GFP concentrations that differ by only about 10% 9 ., In principle , this level of precision would be sufficient for nuclei to distinguish their AP positions with an error of only a single nuclear diameter 11 ., These observations raise fundamental questions regarding the underlying cell-biologic mechanisms responsible for the rapid , precise establishment of nearly equivalent gradients in essentially every embryo ., One such question concerns the dependence of protein gradient formation on the strength , localization , and dynamics of the underlying bcd mRNA ., Females carrying altered genetic dosages of bcd produce gradients of altered amplitude , resulting in mispatterned embryos 12–14 ., Moreover , the packaging of bcd mRNAs into discrete ribonucleoprotein ( RNP ) complexes , and their subsequent localization to the anterior oocyte cortex , requires a suite of maternal factors ( reviewed in 15–17 ) ., Ooctyes lacking these factors produce embryos with distorted protein gradients resulting from bcd mRNA translation at inappropriately posterior locations 12 ., The reliable formation of the Bcd protein gradient must depend , at least in part , on the spatial distribution and the number of bcd mRNA molecules ., While prior work has utilized in situ hybridization to document the spatial distribution of bcd mRNA during embryogenesis 3–5 , 18 , 19 , no work has yet determined bcd mRNA particle numbers or examined their global localization in the developing embryo ., A second question regards the biophysical processes affecting the Bcd protein , such as its transport and degradation ( reviewed in 20 ) ., The observed exponentially decaying nuclear steady state Bcd protein distribution is consistent with analytical models of gradient formation via constant anterior protein synthesis , coupled with diffusion and uniform degradation throughout the embryo ( the SDD model ) ., Within this model , the effective diffusion constant of the Bcd protein D and the protein lifetime τ determine the dynamics of gradient formation ., A relatively short τ would allow the gradient to reach an equilibrium distribution ( i . e . when protein degradation is matched by new synthesis ) within the available developmental time of ∼2 h ., Conversely , a relatively long protein lifetime will prohibit the achievement of such an equilibrated state within that time ., Likewise , larger or smaller values of D will increase or decrease , respectively , the predicted distance a protein can move away from its source mRNA ., A direct measurement of D for cytoplasmic Bcd-GFP at n . c . 13 9 yielded a substantially slower diffusion constant than for comparably sized , biologically inert molecules 21 , and about an order of magnitude too low to account for the rapid achievement of a steady state gradient with the appropriate length constant ., More recent work has suggested the presence of multiple populations of Bcd-GFP , where a fraction of Bcd-GFP may diffuse rapidly enough to establish stability in the time available 22 , 23 ., Slow diffusivity , however , would be consistent with an alternate model in which the protein gradient arises from graded bcd mRNA distribution 19 ., Low values of D coupled with rapid protein degradation would minimize protein movement away from source molecules , so that the local mRNA amount would predominately determine protein concentration along the AP axis ., Under these conditions , sufficiently mobile mRNA would greatly impact the dynamics of gradient formation , and protein diffusion contributes only minimally ., However , at present it is difficult to evaluate the legitimacy of these models , or to determine what values of D and τ , if any , might accurately describe the dynamics of Bcd gradient formation , because no quantitative data exist which span the first 80–90 min of development ., To address these questions , we developed two novel quantitative measurement approaches ., First , we adapted a method of fluorescent in situ hybridization to the Drosophila embryo employing fluorescently labeled DNA oligonucleotides 24 , 25 ., This method allowed us to identify individual bcd mRNA particles and determine their positions and intensities ., We found that bcd mRNA particles undergo movement into the core of the embryo , away from their initial site of localization at the ooctye cortex , by the end of the third cleavage division ( within 30 min AED ) ., Subsequently , bcd mRNA is relocalized to the egg cortex during cortical nuclear migration ( mitoses 6 to 9 , 55 to 75 min AED ) , during which bcd particles tend to dissociate without bcd mRNA degradation ., Despite the dynamic behavior of bcd particles , at all times >90% of mRNA is localized within the anterior 20% of the embryo , a distribution which is insufficiently extended to produce the observed protein gradient in the absence of large-scale protein redistribution along the anterior-posterior axis ., Second , we measured nuclear Bcd-GFP levels during presyncytial stages using GFP fluorescence in fixed embryos ., We discovered that Bcd-GFP does not begin accumulating in nuclei until interphase 6 , about 45 min AED , reaching peak levels 50 min later in n . c . 11 and 12 ., Unexpectedly , in fixed samples we found that the nuclear gradient declines between n . c . 13 and n . c . 14 , a result which differs from prior analysis of living embryos ., We present evidence that attributes this difference in part to delayed fluorescence maturation of GFP in living embryos ., Finally , to examine whether the bcd mRNA distribution impacts protein gradient formation , we incorporated its observed spatio-temporal dynamics into numerical simulations of protein production and movement ., We found that models utilizing the actual mRNA distribution result in improved predictions of protein gradient dynamics , compared to models employing a single point source at the anterior pole ., Therefore , we conclude that although the spatially extended mRNA localization contributes to the protein gradient , mRNA localization alone cannot account for protein gradient dynamics ., Our results demonstrate that protein movement , whether active or passive , is a necessary component of Bcd protein gradient formation ., To examine bcd mRNA distribution in fixed embryos , we adapted a fluorescent in situ hybridization ( FISH ) protocol 24 using a set of 48 20-mer DNA oligonucleotides complementary to the bcd open reading frame ( Table S1 ) directly conjugated to AlexaFluor fluorophores 25 ., This approach has been used previously to detect single mRNA molecules in mammalian cell culture and C . elegans embryos 25 ., By employing directly labeled oligonucleotides as probes , we bypassed the use of antibodies and enzymes for detection 26–28 , and thus minimized nonspecific background , nonlinear signal response , variable tissue penetration by reagents , or other potential difficulties associated with quantification of conventional FISH signal 29 ., Using standard laser scanning confocal microscopy , we can readily detect the anterior localization of bcd mRNA at low magnification ( Figure S1 ) and distinguish individual bcd RNP complexes in high resolution images ( Figures 1 and 2 ) ., To obtain the complete three-dimensional ( 3d ) structure of bcd mRNA distribution , we generated high resolution confocal stacks at high magnification , such that a spatial unit voxel represents a volume of 75 nm×75 nm×420 nm ., These stacks span the entire left or right half of each embryo from the midsagittal plane to the embryo surface , representative of the entire embryo due to left-right symmetry ., Figures 1 and 2 show typical images of embryos labeled during interphase 4 and interphase 11 , respectively ., Corresponding 3d stacks are provided in Movies S1 and S2 ., Images taken at the anterior of labeled embryos reveal readily identifiable individual bcd RNP particles above background noise ( Figures 1A–E and 2A–D ) , whereas the embryo posterior is essentially devoid of bcd mRNA particles ( Figures 1F–G and 2E–F ) ., Individual particles appear on multiple neighboring z-slices ( Figure 1H ) , owing to the objectives spatially extended point spread function ( PSF ) ., To obtain counts of discrete bcd mRNA particles , their localization , and fluorescence intensities , we designed custom image analysis software to determine the spatial positions , radii , and fluorescence intensities of bcd mRNA particles ( see Materials and Methods ) ., As illustrated in Figure S2 , the analysis algorithm discriminates between overlapping particles which densely populate the anterior and readily detects dim particles ., A large majority of detected bcd particles are circular in shape ( Figure 3A ) , confirming that the algorithm resolves closely neighboring particles ., The particles have a spatial extent of about 3 pixels on average , corresponding to a physical distance of ∼200 nm ( Figure 3B ) ., This is identical to the PSF in the confocal slices ( Figure S3 ) , indicating that the particles are smaller than the diffraction limit of our microscopy ., The extended PSF dictates that each mRNA particle must be detected simultaneously on consecutive z-slices , and the algorithm uses this property to discriminate true particles from local background noise ( see Materials and Methods ) ., We identified an optimal threshold for distinguishing between candidate particles and random fluorescence fluctuations by examining posterior stacks which contain few particles ( see Materials and Methods and Figure S4 ) ., As additional controls , we examined background fluorescence in cleavage stage embryos processed without any probes ( Figure S5A–C ) or exposed to probes against the purely zygotically expressed gene giant ( Figure S5D–F ) ., Both show a low level of fluorescent signal , but no detected particles ., These controls allow us to exclude falsely identified particles ., bcd particles coalesce in vivo during bcd mRNA anterior localization in late oogenesis , and subsequently disperse upon egg activation 18 , 30 , 31 , suggesting that bcd RNPs may remain at least partially intact in fertilized embryos ., The fluorescent intensities of particles observed by our method span about a 10-fold range ( Figure 3C ) ., This broad distribution suggests the presence of multiple bcd mRNAs in a single detected particle , consistent with previous biochemical and fluorescence microscopy observations 32–34 ., bcd mRNA is not degraded until the onset of cellularization at n . c . 14 3 ., Therefore , it is highly likely that the broad intensity distribution reflects variation in the number of RNAs per particle ., Fluorescence variability is reflected in the axial ( z ) diameter of particles: we observe that bright particles occupy up to 5 z-slices , whereas dim particles are detectable above noise level in fewer slices ( Figure 1H ) ., Our ability to distinguish discrete particles affords an unprecedented quantitative view of mRNA particles and allows us to separate their dynamics from the ultimate dynamics of Bcd protein ., To quantify the spatial distributions of bcd mRNA , we examined embryos ranging from n . c . 3 to n . c . 14 ., During all stages of development , bcd particles are found in a graded distribution peaking at around 7% egg length ( EL , as measured from the anterior pole ) and leveling off at nearly zero particle density by 40% EL ( Figure S6 ) ., Confocal images taken at both early cleavage ( Figure 1 ) and blastoderm ( n . c . 11 , Figure 2 , and n . c . 13 , Figure S7 ) stages reveal a higher density of bcd particles in anterior compared to posterior regions ., Moreover , anteriorly localized particles are brighter than those toward the posterior ( Figure 3D ) ., The higher density of brighter anterior particles results in a sharp , steep gradient of total bcd mRNA distribution within the anterior third of the embryo ( Figure 3E ) ., At all times , 90% or more of total bcd mRNA is found within the anterior 20% of the embryo ( Figure 3F ) , as previously reported 35 ., The quantitative difference between Figures 3E and S4A arises from both the increased number and intensity of anterior particles ., Although the majority of mRNA particles are detected in the anterior 20% , we nevertheless observe faint discrete particles scattered along the entire AP axis ., During both cleavage and blastoderm , very few or no particles are detected in the posterior embryo core ( Figures 1F–G , 2E–F , S7E–F ) ., Conversely , a small number of particles exhibiting weak fluorescence sparsely populate the posterior surface ( Figures 1I–K , 2G–L , S7I–L ) ., A handful of particles can be detected even at almost 100% EL near the posterior pole ( Figure 2K–L ) ., Thus , we can document even the small fraction of particles that either fail to localize during oogenesis or which are transported an unusually large distance away from the anterior cortex after egg activation ., To our knowledge , bcd RNP complexes have not been detected in the posterior of wild-type embryos ., These observations demonstrate the exquisite sensitivity of our method to the presence of dimly fluorescent bcd mRNA particles ., However , we emphasize that the particles found in the entire posterior 60% of the embryo constitute less than 1% of the total bcd mRNA ( Figure 3F ) ., These particles are uniformly dimly fluorescent ( Figure 3D ) and are found at a sparse density which does not change significantly during development ., Translation of bcd mRNA is repressed by Nanos activity 36 , silencing posteriorly localized bcd mRNA ., Based on these observations , the fraction of bcd mRNA present in the posterior 60% of the embryo likely contributes only negligibly to protein gradient formation ., Previous studies have demonstrated that egg activation triggers release of bcd mRNA from its initial tight localization at the anterior egg cortex , resulting in a posterior dispersion of bcd mRNA within 25 min at 25°C 30 ., This corresponds to the interphase of n . c . 3 ( Figure S8 ) , by which time bcd mRNA has already reached its most posterior extent , where it remains from n . c . 4 through 6 ( Figure 3E and 3F ) ., The AP particle distribution at n . c . 4 therefore results from this early release of bcd mRNA from the cortex ., Prior to and during n . c . 6 , 97% of all bcd mRNA is found in the anterior 20% ( Figure 3F ) , with the remainder forming a gradient which drops to nearly zero before 40% EL ., In midsagittal planes at n . c . 4 , regions near the embryo surface contain fewer particles than the center of the embryo ( compare upper and lower panels in Figure 1B–D ) ., Consistent with this , confocal planes taken near the embryo surface contain fewer particles compared to the same AP position at the midsagittal plane ( compare Figure 1A and 1I ) ., Therefore , much of bcd mRNA is not near the cortex but is found in the embryo interior ( Figures 1A , S1A–B , S8 ) , as observed previously 5 ., In low magnification images , we often observe mRNA localization in a wedge or cone-shaped distribution jutting in toward the interior core of the embryo in midsagittal planes ( Figures S8 , S9A–D ) , suggesting the presence of uncharacterized structures along which bcd mRNA particles might translocate upon egg activation ., The observation that particles do not progress further into the posterior after n . c . 3 supports the idea that bcd mRNA is tethered to underlying cytoskeletal structure ( s ) throughout embryogenesis 32 and is not free to diffuse ., We observed a marked change in bcd mRNA spatial localization after interphase 6 ., As the nuclei undergo expansion toward the axial poles beginning at the sixth mitosis and continuing at n . c . 7 , bcd mRNA moves ahead of the expanding nuclei ( Figures 4A , S1C–E ) ., bcd mRNA progresses to the cortex during nuclear cortical migration between n . c . 8 to 10 , such that in midsagittal planes at n . c . 11 , bcd mRNA is highly enriched near the embryo cortex , with the majority of fluorescence found within about 25 µm of the embryo surface ( Figures 2A–G , 4A ) ., This enrichment is also apparent in z-slices collected at the cortical nuclear layer ( Figures 2G–L , 4B , S7G–L ) ., Compared to earlier cycles , bright bcd particles are now present on the surface at a greater distance from the anterior pole ( compare Figure 1I to Figure 2G ) ., Thus , with the nuclei penetrating the cloud of bcd mRNA during their cortical migration , by the start of the syncytial blastoderm stage , bcd mRNA fluorescence resembles a cup covering the anterior end of the embryo ( Figure S9E–H ) , which remains in place through n . c . 14 ( Figures 3E , 3F , S1F–G , and S6 ) ., Concomitant with bcd particle translocation from core to cortex , we observe a mild posterior shift of the bcd mRNA distribution along the AP axis by the onset of the blastoderm stage at n . c . 10 ( Figures 3E , 3F , and S6 ) , as previously observed 5 ., Embryos at all blastoderm stages appear similar , with bcd mRNA particles surrounding the nuclei on both the basal and apical surfaces ( compare n . c . 11 in Figure 2 and n . c . 13 in Figure S7 ) ., Despite dynamic cortical relocalization , more than 90% of bcd is found in the anterior 20% of the embryo ( Figure 3F ) , with the remaining 10% falling to zero particle density by 40% EL ., Concomitant with the particle relocalization , the distribution of fluorescence intensities shifts from bright to dim spot intensities ( Figure 3C ) , and the number of detected particles tends to increase with developmental age from about 70 , 000 per embryo in cycles prior to n . c . 7 , to about 110 , 000 particles in blastodermal stages ( Figure 4C ) ., Despite these changes , the total fluorescent signal from bcd mRNA remains constant across all developmental times through early n . c . 14 ( Figure 4D ) , arguing that our method is capable of detecting the constant maternal bcd mRNA pool ., Given previous work indicating the presence of multiple bcd mRNA molecules per particle 32–34 , these observations indicate that such particles disassemble during their relocalization to the cortex and suggest that mRNA degradation likely plays little role in particle dissolution ., bcd RNA is degraded with other maternal mRNAs at the midblastula transition during n . c . 14; prior to this time , no degradation of bcd mRNA occurs 4 , 37 ., To contrast bcd mRNA particle disassembly with mRNA degradation , we documented the loss of bcd mRNA during n . c . 14 ( Figure 5 ) , which spans about 1 h at 25°C ., During this time , wholesale zygotic gene expression is activated , maternal mRNAs are degraded , and cellular membranes form between the cortically positioned nuclei ., We gauged the approximate age of fixed embryos in n . c . 14 on the presence of the cellularization front and the morphology of nuclei which elongate during cellularization ., Immediately following the 13th mitosis , embryos exhibit particles with intensity and spatial distributions similar to the earlier blastoderm stages ( Figure 5A , 5D , 5G ) ., During the next ∼10 min , the cell surface protrudes above each nucleus and membrane invagination begins 38 , 39 , during which we observe a dramatic reduction in overall fluorescence and a sharp decrease in the number of particles and fluorescence of particles in all regions of the embryo ( Figure 5B , 5E , 5H , 5J–L ) ., As nuclei undergo elongation , bcd mRNA is almost completely lost , and we detect only a handful of dim particles in the anterior , usually at the cortical or lateral surfaces between nuclei ( Figure 5F , 5L ) ., At this time , the total particle count has dropped to essentially zero ( see the final data point on x-axis in Figure 4D ) ., Thus , degradation of bcd can be readily distinguished from the separate earlier event of particle dissolution ., In summary , we have observed that bcd mRNPs behave dynamically in the embryo: after fertilization , bcd mRNA establishes a wedge-shaped distribution , which is subsequently reformed into a cup-shaped geometry during n . c . 6–10 ( Figure S9 ) , followed by mRNA decay at the onset of cellularization ., Despite these dynamic rearrangements , more than 99% of all mRNA particles occupy the anterior 40% of the egg at all times ., These findings demonstrate that an exponentially graded protein distribution , which is detectable as far as the posterior 85% of the egg 8 , 9 , 40 , can be established only if the protein moves away from the anterior source ., Understanding the degree to which a graded mRNA source influences the formation of the Bcd protein gradient requires a comparison of mRNA and protein dynamics ., Live imaging of transgenic embryos suggests that nuclear Bcd-GFP levels approach a nearly stable concentration gradient during the 10th interphase , approximately 80–90 min AED 9 ., Gradient formation must occur prior to this time; however , no previous study has quantified Bcd distributions in living cleavage stage embryos , due to challenges to optical measurements created by the greater opacity and autofluorescence of presyncytial embryos ., To minimize these complications , we devised a specific fixation protocol ( see Materials and Methods ) , which preserves GFP fluorescence of Bcd-GFP expressing embryos , allowing us to detect and measure nuclear Bcd-GFP with conventional confocal microscopy ( Figure 6 ) ., The earliest time at which we can detect nuclear Bcd-GFP is about 45 min AED , when Bcd-GFP begins to accumulate in the most anteriorly positioned nuclei during the first 2 min of the sixth interphase ( Figure 6A ) , coinciding with the expansion of nuclei into the cloud of bcd mRNA and with the onset of bcd mRNA relocalization ., We detect Bcd-GFP fluorescence above background autofluorescence in nuclei positioned within the anterior third of the embryo when the nuclei extend from about 25% to 60% EL ( Figure 6A ) ., In contrast , during interphase 5 nuclear Bcd-GFP appeared no greater than background ( Figure S10; n\u200a=\u200a12 embryos ) , showing that Bcd-GFP has not accumulated to high enough levels for visualization in nuclei ., Tissue autofluorescence may preclude detection of lower levels of nuclear localized Bcd-GFP; therefore , Bcd-GFP accumulation begins during n . c . 6 at the latest ., Delayed nuclear accumulation would be consistent with previous work suggesting that the translation rate of bcd may be relatively low during the first hour of development before polyadenylation of bcd mRNA 41 ., Between the sixth mitosis and interphase 10 , nuclei migrate toward the egg cortex as an expanding , elliptical shell , while continuing to accumulate Bcd-GFP ( Figure 6A–D ) ., A gradient of nuclear Bcd-GFP can be readily distinguished within the anterior third of the embryo from n . c . 7 and 8 ( Figure 6B , C ) ., By n . c . 9 , Bcd-GFP is apparent in nuclei throughout the anterior half of the embryo ( Figure 6D ) , demonstrating the continued heightening of Bcd-GFP levels along the AP axis ., The nuclei arrive at the egg cortex and form the syncytial blastoderm upon completing the ninth mitosis ., From n . c . 10 onward , the Bcd-GFP gradient is evident in single confocal slices of the midsagittal plane ( Figure 6E–I ) , reminiscent of images of Bcd-GFP obtained previously by live imaging 9 ., To quantify the Bcd gradient at these early stages , we extracted nuclear fluorescence intensities from embryos expressing both Bcd-GFP and Histone-RFP ., The latter provides a nominally uniform fluorescence signal in all nuclei and serves as a reference to normalize Bcd-GFP values at different optical depths in the sample ( see Materials and Methods ) ., Corrected nuclear Bcd-GFP gradients between n . c . 6 and n . c . 11 show that at all positions along the AP axis , nuclear Bcd-GFP rises continuously ( Figure 7A and 7C ) ., The shape of the gradient , however , does not change ( see slopes on log-linear inset to Figure 7A ) , whereas its amplitude continues to rise and increases by a factor of ∼3 over the course of n . c . 6–11 ., The embryo-wide increase in Bcd-GFP is abruptly halted along the entire AP axis at n . c . 11–12 , by which time nuclei have attained Bcd concentrations >90% of their maximal values ., Therefore , the nuclear gradient forms over approximately 50 min between n . c . 6 and n . c ., 11 . During n . c . 11 and 12 nuclear Bcd-GFP concentration remains within ∼95% of its maximum along the whole AP extent ., After n . c . 12 , nuclear concentrations drop , starting at the anterior; by n . c . 13 , values in the anterior 20% of the embryo are approximately 80%–90% of their maxima at n . c ., 12 . At the beginning of n . c . 14 , all positions along the AP axis decay to around 70%–85% of the maximum ( Figure 8D ) ., Despite this decrease in nuclear Bcd , the total amount of Bcd protein still rises through n . c . 13 ( Figure S11 ) 9 , consistent with continued protein production until mRNA degradation ., These observations of gradient dynamics differ sharply from the bcd mRNA distribution , which at no time resembles an exponentially decreasing gradient reaching to >75% EL , such as observed for Bcd-GFP from n . c . 9 onward ., Bcd-GFP is visible in nuclei positioned at 50% EL and beyond from n . c . 8 onward , yet at these positions we observe essentially no mRNA particles ., Instead , the protein gradient can only be accounted for if protein moves towards the posterior from anteriorly localized mRNA ., By live imaging , nuclear Bcd-GFP levels appear relatively stable after the 10th mitosis 9 , in contrast to our observations in fixed tissue ., Additionally , we observe that the decay length of fixed gradients is approximately 0 . 15 EL ( Figure 8A inset ) , compared to a decay length of around 0 . 2 EL in live embryos 9 ., These differences do not result from the use of different microscopy methods ( Figure S12 ) and therefore must arise from the fixation procedure ., To determine the effect of fixation on Bcd-GFP fluorescence , we imaged living Bcd-GFP and Histone-RFP embryos at blastoderm stages , then fixed the embryos within 3 min of live imaging , and subsequently re-imaged under the same microscopy settings ( Figure 8A–D ) ., We found that fixation increases the overall brightness by about 3-fold ( Figure 8E–F ) , possibly as a result of increased transparency of fixed material mounted in a glycerol-based medium ., However , rescaling live or fixed Bcd-GFP gradients by a factor of 3 reveals qualitatively different gradient shapes in the anterior 25% of the egg , wherein fixed gradients exhibit an additional increase in relative intensity ( Figure 8G ) ., This remaining difference could be accounted for if newly synthesized Bcd-GFP is not immediately visible but becomes fluorescent after fixation ., This phenomenon can be attributed to the well-known process of GFP maturation ( 42 and refs therein ) ., To test whether delayed maturation can affect the appearance of Bcd gradients detected by fluorescent protein fusions , we imaged transgenic embryos expressing protein fusions of Bcd with either the rapidly maturing GFP derivative Venus 43 or the more slowly maturing mRFP 44 ., We found that after 3-fold rescaling , live and fixed Bcd-Venus gradients exhibited similar appearances , consistent with the rapid attainment of fluorescence by Venus ( Figures 8H , S13A–C ) ., In contrast , rescaled live Bcd-mRFP gradients showed poor correspondence with fixed gradients along the majority of the AP axis ( Figures 8H , S13D–F ) ., This result supports the view that delayed maturation can alter the appearance of the Bcd protein gradient ., In addition , by incorporating delayed maturation into a simulation of gradient formation ( see below ) , we could qualitatively recapture the stable nuclear gradients observed in live embryos at late blastoderm ( Figure S14 ) ., We note that previous work has determined that gradients of equivalent shape are observed upon quantification of GFP fluorescence and anti-GFP immunostaining in the same embryo 45 ., Moreover , the distribution of Bcd protein detected by anti-Bcd antibodies does not differ between wild-type and embryos in which Bcd-GFP is the only source of Bcd protein 8 ., A maturation-based explanation supports the idea of protein movement away from anterior mRNA ., With delayed maturation , immature fluorescent protein synthesized at the anterior would tend to attain fluorescence while en route toward the posterior ., As a result , in a live embryo , the fraction of immature eGFP would be higher in the anterior than the posterior , and allowing additional time for eGFP to attain fluorescence would then disproportionately increase anterior fluorescence compared to posterior , as we observe by fixation ., Diffusion represents a simple plausible mechanism of protein transport away from the anterior mRNA source , and diffusion-based models such as the SDD model provide a straightforward mathematical framework for describing gradient establishment ., Previous modeling efforts have treated the site of synthesis either as an anteriorly localized point source 9 , 46–49 or as conjectured anterior domains 11 , 50–53 ., Having characterized the actual distribution of bcd RNA , we asked whether that distribution is a necessary component of gradient establishment by performing numerical reaction-diffusion simulations of gradient formation ., We compared our measured nuclear gradients to protein distributions predicted to arise from either a realistic mRNA distribution or an anterior point source ., We find that the combination of a non-monotonic time course for nuclear gradient amplitude ( Figure 7D ) and a relatively stable length constant between n . c . 8 and 14 ( Figure 7B ) precludes a classical SDD-type model in which the biophysical parameters of translation rate S , degradation rate τ , and diffusion D are kept constant at all times ., Therefor | Introduction, Results, Discussion, Materials and Methods | The Bicoid morphogen gradient directs the patterning of cell fates along the anterior-posterior axis of the syncytial Drosophila embryo and serves as a paradigm of morphogen-mediated patterning ., The simplest models of gradient formation rely on constant protein synthesis and diffusion from anteriorly localized source mRNA , coupled with uniform protein degradation ., However , currently such models cannot account for all known gradient characteristics ., Recent work has proposed that bicoid mRNA spatial distribution is sufficient to produce the observed protein gradient , minimizing the role of protein transport ., Here , we adapt a novel method of fluorescent in situ hybridization to quantify the global spatio-temporal dynamics of bicoid mRNA particles ., We determine that >90% of all bicoid mRNA is continuously present within the anterior 20% of the embryo ., bicoid mRNA distribution along the body axis remains nearly unchanged despite dynamic mRNA translocation from the embryo core to the cortex ., To evaluate the impact of mRNA distribution on protein gradient dynamics , we provide detailed quantitative measurements of nuclear Bicoid levels during the formation of the protein gradient ., We find that gradient establishment begins 45 minutes after fertilization and that the gradient requires about 50 minutes to reach peak levels ., In numerical simulations of gradient formation , we find that incorporating the actual bicoid mRNA distribution yields a closer prediction of the observed protein dynamics compared to modeling protein production from a point source at the anterior pole ., We conclude that the spatial distribution of bicoid mRNA contributes to , but cannot account for , protein gradient formation , and therefore that protein movement , either active or passive , is required for gradient formation . | The Bicoid protein gradient plays a crucial role in determining the anterior body pattern of Drosophila embryos ., This gradient is the classic example of morphogen-mediated patterning of a developing metazoan and serves as a major topic for mathematical modeling ., Accurate modeling of the gradient requires a detailed account of the underlying bicoid mRNA distribution ., The classic model holds that mRNA protein gradient arises via protein diffusion from mRNA localized at the anterior of the developing egg ., In contrast , recent proposals suggest that an mRNA gradient generates the protein gradient without protein movement ., In this study , we introduce a novel mRNA quantification method for Drosophila embryos , which allows us to visualize each individual mRNA particle accurately in whole embryos ., We demonstrate that all but a few mRNA particles are confined to the anterior 20% of the egg , and consequently that the protein must move in order to establish a gradient ., We further report that the mRNA distribution is highly dynamic during the time of protein synthesis ., In numerical simulations , we show that incorporating realistic spatial locations of the individual source mRNA molecules throughout the developmental period is necessary to accurately model the experimentally observed protein gradient dynamics . | developmental biology/embryology, developmental biology/morphogenesis and cell biology, biophysics/experimental biophysical methods, biophysics/theory and simulation, cell biology/developmental molecular mechanisms, developmental biology/pattern formation, molecular biology/mrna transport and localization, developmental biology/developmental molecular mechanisms | New quantitative data show that the Bicoid morphogen gradient is generated from a dynamic localized source and that protein gradient formation requires protein movement along the anterior-posterior axis. |
journal.pgen.1005553 | 2,015 | Metabolomic Quantitative Trait Loci (mQTL) Mapping Implicates the Ubiquitin Proteasome System in Cardiovascular Disease Pathogenesis | Despite the strong heritability of cardiovascular disease ( CVD ) , its underlying genetic architecture remains incompletely characterized ., Genomewide association studies ( GWAS ) have converged on association of CVD with a locus on chromosome 9p21 1 , but the variants confer modest risk and are of unclear functional significance ., One limitation of GWAS studies for complex diseases is the search for association with disease as a binary endpoint , rather than with molecular markers that define risk ., An alternative approach is to search for variations in the genome that associate with variation in complex traits ., In fact , many diseases can be defined by an underlying quantitative scale , and these “intermediate” traits may have a stronger functional relationship to the causative gene , thereby providing a stronger signal for the disease process ., Metabolite levels measured by the emerging tools of metabolomics may be particularly useful for such studies ., Indeed , integration of GWAS with metabolomic profiles in population-based cohorts 2 has demonstrated that as much as 12% of variance in metabolite levels is determined by single nucleotide polymorphisms ( SNPs ) ., However , most studies of this type performed to date have not used disease-burdened cohorts , so clear linkages between genetic signals , intermediate phenotypes and disease remain to be discovered ., Metabolomic profiling has identified novel biomarkers for CVD risk 3–5 ., For example , a cluster of heritable 6 short-chain dicarboxylacylcarnitine ( SCDA ) metabolites measured in plasma ( comprised of the mono-carnitine esters of short-chain , alpha- , omega-diacids ) , a cluster of long-chain dicarboxylacylcarnitines ( LCDA ) , and a cluster of medium-chain acylcarnitines ( MCA ) predict CVD events in cardiovascular cohorts 4 , 5 , in patients undergoing coronary artery bypass grafting 3 , and add incremental risk prediction to robust clinical models inclusive of >20 variables 5 ., Little is known about the biological pathways represented by these metabolites and how they may predispose to CVD ., Thus , we hypothesized that integration of metabolomics with genetics , epigenetics , and transcriptomics could define novel mechanisms of CVD pathogenesis by identifying metabolic quantitative trait loci ( mQTL ) that are CVD risk factors ., We performed a GWAS of metabolite levels in a large cardiovascular cohort referred for cardiac catheterization ( CATHGEN , N = 1490 ) and validated our findings in a second cohort ( CATHGEN , N = 2022 ) ., A proportion of study subjects ( 44% ) did not have clinically significant atherosclerotic coronary artery disease at time of catheterization; regardless , all individuals were analyzed given that metabolites predict risk of CVD events even in individuals without coronary artery disease , and because these individuals are still at risk for these events ., We found that genetic loci that strongly associate with SCDA levels also predict incident CVD events , and are linked to ER stress ., Genes differentially methylated in subjects at the extremes of SCDA levels also report on ER stress ., Gene expression quantitative trait loci ( eQTL ) pathway analysis identified ER stress as an expression module associated with disease risk , particularly highlighting the ubiquitin proteasome system ( UPS ) arm of ER stress ., Thus , this multi-platform “omics” approach identified a molecular pathway ( ER stress and dysregulation of the UPS ) associated with a prevalent complex disease ., Factor 1 , factor 2 and factor 3 scores were used as the quantitative traits in GWAS analysis to identify mQTL ., Q-Q plots suggested the presence of SNPs associated with levels of each of the three metabolite factors ( S2 , S3 and S4 Figs ) ., Several SNPs were significantly associated with metabolite factor levels at genomewide significance ( p≤10−6 ) in additive models in the discovery cohort ( Fig 1A–1F ) and confirmed ( p≤0 . 05 ) in the validation cohort ( Table 2 ) ., Specifically , eight SNPs were associated with factor 1 ( MCA ) levels in any race , but with only two of these SNPs showing more than nominal significance in the validation cohort ( Table 2 ) : rs10987728 ( in cyclin dependent kinase 9 CDK9 ) and rs6738286 ( intergenic between transition protein 1 TNP1 and disrupted in renal carcinoma 3 DIRC3 ) ., Twelve SNPs were associated with factor 2 ( LCDA ) levels in any race ( Table 2 ) , with only two of them showing more than nominal significance in the validation cohort ( rs12129555 just downstream from polymeric immunoglobulin receptor PIGR and rs17025690 in Usher syndrome 2A USH2a ) ., Factor 3 ( SCDA ) showed the strongest mQTL with twelve SNPs being associated with SCDA levels in any race ( Table 2 ) , and four of these SNPs showing more than nominal significance in the validation cohort: rs2228513 in HERC1 HECT and RLD domain containing E3 ubiquitin protein ligase family member 1 ( HERC1 ) , rs10450989 in ubiquitin specific protease 3 ( USP3 ) , rs11771619 in round spermatid basic protein 1-like ( RSBN1L ) , and rs1869075 ( intergenic between F-box protein 25 FBXO25 and glutamate rich 1 ERICH1 ) ., Effect sizes ( β , i . e . per 1 unit change in factor levels ) ranged from to -0 . 38 to 2 . 17 ( factor 1 ) , -0 . 19 to 1 . 16 ( factor 2 ) , and -0 . 43 to 1 . 72 ( factor 3 ) ., In meta-analyses combining the race-stratified results , eleven SNPs were associated with factor 1 ( MCA ) levels , with three of these SNPs showing more than nominal association ( Table 3 ) ; one of these ( rs10987728 in CDK9 ) was also identified from race-stratified results and two ( rs16990949 in PDX1 C-terminal inhibiting factor 1 PCIF1 ) and rs543129 intergenic between cutaneous T-cell lymphoma-associated antigen 1 ( CTAGE1 ) and retinoblastoma binding protein 8 ( RBBP8 ) ) were new mQTL identified in these race meta-analyses ., Eight SNPs were associated with factor 2 ( LCDA ) levels ( Table 3 ) ; one gene had been identified in race-stratified analyses ( ZNF521 ) but showed stronger association in the validation cohort in these analyses , and rs352216 near frizzled class receptor 3 ( FZD3 ) was a new mQTL ., Factor 3 ( SCDA ) again had the largest number and strongest mQTL with fourteen SNPs associated with SCDA levels , with eight SNPs showing more than nominal significance in the validation cohort ( Table 3 ) ., SNPs in USP3 , HERC1 and OLFM4|SUGT1 ( intergenic between olfactomedin 4 and SGT1 , suppressor of G2 allele of SKP1 S . cerevisiae ) had already been identified in race-stratified analyses; additional mQTL identified in these race meta-analyses included rs12589750 and rs3853422 ( in or near stonin 2 STON2 and sel-1 suppressor of lin-12-like ( C . elegans ) SEL1L ) , rs930491 and rs11827377 ( both intergenic between ribonucleotide reductase M1 RRM1 and stromal interaction molecule 1 STIM1 ) , rs11242866 ( between solute carrier family 22 ( organic cation transporter ) , member 3 SLC22A23 and PX domain containing 1 PXDC1 ) , and rs4544127 ( near FRAS1-related extracellular matrix protein 2 FREM2 and stomatin-like protein 3 STOML3 ) ., Thus , to summarize , the most robust results overall were for mQTL associated with SCDA metabolite levels ( factor 3 ) including an mQTL composed of USP3 ( rs10450989 ) and HERC1 ( rs2228513 ) ; and a locus composed of STON2 ( rs12589750 ) and SEL1L ( rs3853422 ) , with loci meeting genomewide significance in the discovery cohort ( p≤10−6 ) , strong significance in the validation cohort ( p = 2 . 4x10-3–7 . 7x10-7 , except rs3853422 which only showed borderline significance p = 0 . 01 ) , and stronger association in the meta-analyses ( p = 1 . 6x10-6–7 . 2x10-12 ) ., The next strongest overall results for SCDA mQTL ( based on race-stratified or race-combined meta-analysis p-values ) in descending order of significance were for RRM1|STIM1 , OLFM4|SUGT1 , SLC22A23|PXDC1 , RSBN1L , FBXO25|ERICH1 , and FREM2|STOML3 ., The next strongest results overall were for mQTL associated with LCDA ( factor 2 ) levels with SNPs in PIGR , ZNF521 , USH2A and FZD3 showing more than nominal significance in the validation cohort ., Finally , mQTL associated with MCA ( factor 1 ) levels included CDK9 , DIRC3 , CTAGE1|RBBP8 , and PCIF1 ., We have previously shown that all three metabolite factors predict risk of incident CVD events , however the results from those studies were most robust for the SCDA metabolites 5 ., Given these prior results , and the strength and consistency of findings for the SCDA metabolite factor in these GWAS analyses , we chose to focus the remainder of our analyses on this factor ., Fig 2 and S5 Fig display Locus Zoom plots for these eight mQTL most strongly associated with SCDA metabolite factor levels ., Interestingly , the majority of these ( i . e . HERC1 , USP3 , STIM1 , SUGT1 , FBXO25 and SEL1L ) encode proteins reporting on endoplasmic reticulum ( ER ) stress ., SCDA mQTL were tested for association with incident CVD events using Cox proportional hazards time-to-event analyses in the combined discovery and validation datasets , using meta-analysis of race- and dataset-stratified results , unadjusted for multiple comparisons ., Of these eight mQTL ( 15 SNPs ) loci , four SNPs predicted mortality in additive models: HERC1 rs2228513 ( p = 0 . 05 in race combined , p = 0 . 04 in whites only ) , RRM1 rs11826962 ( p = 0 . 03 ) , and FBOX025 rs1869075 ( p = 2 . 5x10-4 for blacks only , not significant in race combined analyses ) , with USP3 rs10450989 showing a trend for association ( p = 0 . 06 in race combined , p = 0 . 05 in whites only ) ., FREM2|STOML3 rs4544127 showed a trend for association ( p = 0 . 06 ) ., We observed for the HERC1 SNP a 33% event rate for non-carriers and a 36% event rate for carriers of at least one copy of the minor G allele ( the same allele associated with higher SCDA levels , S3 Table ) ., Adjustment for SCDA levels in these models resulted in attenuation of the association between mQTL and CVD event ( S3 Table ) , suggesting that the relationship between these mQTL and CVD events is in part mediated through SCDA metabolite levels ., To ensure that the relationships between SNPs and SCDA levels were not confounded by renal disease , we further adjusted for glomerular filtration rate ., This adjustment caused no or minimal attenuation of the association for our strongest SNPs ( S3 Table ) ., In multivariable models , we found minimal attenuation of the association between most SNPs and SCDA levels ( S3 Table ) , suggesting that these SNPs have effects on SCDA levels unrelated to other comorbidities ., There was attenuation of association of SNPs near RRM1|STIM1 and STON2|SEL1L after adjustment ( although still significant at p<0 . 05 , unadjusted for multiple comparisons ) , suggesting that these SNPs have effects on SCDA levels mediated through these clinical factors , in particular renal disease ., Visual comparison of the distribution of methylated probes revealed similar distributions in individuals with high and low SCDA levels ( N = 46 , combined methylation discovery and validation datasets , S6 Fig ) ., After filtering based on Δβ values , the presence of multiple correlated probes in a gene , and adjustment for estimated cell type proportions , sex , age and race , probes in 28 genes showed differential methylation in SCDA extremes ( i . e . |Δβ|≥0 . 10 in ≥2 probes within a gene ) ., Differential methylation in three of these genes was confirmed in the validation set based on |Δβ|≥0 . 10 ( BRSK2 , Hook2 and LMTK3 , Table 4 ) ., Two of these genes , including the most significant one , report on ER stress: Hook2 ( four probes , Δβ 0 . 25–0 . 30 ) and BRSK2 ( four probes , Δβ 0 . 11–0 . 20 ) ., Hook2 may be involved in pathways contributing to the ubiquitin proteasome system ( UPS ) arm of ER stress via its role in establishment and maintenance of pericentrosomal localization of aggresomes ( complexes of misfolded proteins , chaperones and proteasomes ) 8 ., BRSK2 encodes brain selective kinase 2 , a serine/threonine kinase of the AMPK family that acts as a checkpoint kinase in response to DNA damage induced by UV irradiation ., BRSK2 protein levels are down-regulated in response to ER stress and ER stress promotes localization of BRSK2 to the ER 9 ., Knockdown of endogenous BRSK2 expression enhances ER stress-mediated apoptosis in human pancreatic carcinoma and HeLa cells 9 ., Blood RNA microarray data were generated for N = 1204 CATHGEN individuals ., We began by examining cis effects for the identified SNPs; however , many of the top SNPs did not have available cis-transcripts after extensive QC ., Rs9591507 , rs17573278 , rs894840 , and rs9285184 ( all in OLFM4|SUGT1 ) , rs11771619 ( RSBN1L ) , rs1869075 ( FBXO25 ) , and rs1886848 ( SULF2 ) showed evidence of cis-regulation ( S4 Table ) ., HERC1 and USP3 are not well-represented on the microarray ( one probe per gene ) ; there was only a minimal trend toward association between the HERC1 and USP3 SNPs with HERC1 expression ( p = 0 . 16 and 0 . 19 , respectively ) and no association with the USP3 transcript ., We then performed eQTL analyses to find evidence of trans-acting pathways ( S4 Table ) ., When analyzed as single transcripts , among the top ten transcripts associated with HERC1 rs2228513 and USP3 rs10450989 were USP39 ( p = 0 . 0002 and p = 0 . 0004 , respectively ) and CYLD ( p = 0 . 00015 and p = 0 . 0007 ) , suggesting that these SNPs show functional relationships with expression of trans-acting pathways related to the UPS arm of ER stress ., USP39 has a role in pre-mRNA splicing and is essential for recruitment of the U4/U6 . U5 tri-snRNP to the prespliceosome ., The tumor suppressor CYLD is a deubiquitinating enzyme , acts as a negative regulator of NF-kappa-B signaling , and plays a pro-inflammatory role in vascular smooth muscle cells 10 ., Cis- and trans-eQTL analyses were not adjusted for multiple comparisons , as we were looking for focused functional effects for each SNP ., Using GSEA 11 , we then identified KEGG pathways of transcripts associated with each SNP; nominal p-values are reported ., The most significant pathway associated with HERC1 rs2228513 was “ubiquitin mediated proteolysis” ( p = 0 . 01; p = 0 . 12 for USP3 rs10450989 ) ., The most significant pathway for rs10450989 was “RNA degradation” ( p = 0 . 03 ) ., Pathways associating with the other SNPs reported on various cellular processes: rs930491 and rs11827377 ( RRM1|STIM1 ) with RNA polymerase pathway ( both p = 0 . 001 ) ; rs11826962 ( RRM1|STIM1 ) with JAK-STAT signaling pathway ( p<0 . 0002 ) ; rs17573278 ( OLFM4|SUGT1 ) with Alzheimer’s disease pathway ( p = 0 . 008 ) ; rs894840 ( OLFM4|SUGT1 ) with glycosaminoglycan biosynthesis ( p<0 . 0002 ) ; rs12589750 and rs3853422 ( STON2|SEL1L ) with ribosome pathway ( p<0 . 0001 and p = 0 . 001 , respectively ) and FC Gamma R mediated phagocytosis pathway ( p = 0 . 001 for both ) ., The Alzheimer’s disease pathway includes components of ER stress and there is evidence that neuronal death in Alzheimer’s disease may arise from ER dysfunction ., The ER is also thought to play an important structural role in phagocytosis ., Finally , we performed GSEA for the correlation between SCDA levels with genomewide RNA expression; nominal p-values are reported ., The most significant KEGG pathways were oxidative phosphorylation ( p<0 . 0002 ) , Parkinson’s disease ( p<0 . 0002 ) , cardiac muscle contraction ( p<0 . 0002 ) , porphyrin and chlorophyll metabolism ( p = 0 . 002 ) , and the proteasome pathway ( p = 0 . 008 ) ., The proteasome is an integral component of the UPS arm of ER stress , degrading cellular proteins that are modified by ubiquitin ., Also , an integral part of the Parkinson’s disease pathway includes components of the UPS ., In this and prior studies 4–6 , SCDA were measured using a flow-injection-MS/MS method that is ideal for rapid profiling of samples , but full resolution of isomeric species comprising each SCDA metabolite peak is not achieved ., C6-DC represents a SCDA that loads heavily on the PCA-derived SCDA factor in our studies , which can be comprised of either the branched-chain methylglutaryl acylcarnitine or the straight chain adipoyl acylcarnitine isomers ., To resolve these metabolites , we adapted a liquid chromatography ( LC ) -MS/MS method 12 ., Peak identification was facilitated by in-house chemical synthesis of internal standards for the two targeted analytes 13 ., Using this method , we re-analyzed 29 human plasma samples from our original studies 5 that contained the highest C6-DC levels ., We found that in the majority of individuals ( 19 of 29 ) , the clearly predominant C6-DC isomer was the branched-chain 3-methylglutaryl carnitine metabolite , and in in 23 of the 29 individuals levels of the branched chain isomer were higher than the straight chain isomer ( S7 Fig ) ., The correlation between the C6-DC measured by flow injection-MS/MS with each of these LC-MS/MS measured isomers further confirms that it is primarily the branched-chain isomer accounting for the signal ( r2 = -0 . 06 , p = 0 . 8 for straight chain isomer; r2 = 0 . 67 , p = 1 . 8x10-4 for branched-chain isomer ) ., Interestingly , one potential source of the branched-chain 3-methylglutaryl carnitine metabolite is the branched-chain amino acid leucine ., Our previous studies have shown an association of branched-chain amino acid metabolites with coronary artery disease 4 , 7 ., The above findings linking ER stress to SCDA metabolites led us to question whether nutrient-induced accumulation of dicarboxylacylcarnitines would be accompanied by ER stress in cultured cells ., Exposure of human HEK293 kidney cells to 500 uM fatty acids for 24 hours ( a condition designed to mimic elevated fatty acid levels observed in human obesity ) increased cellular production and efflux of several long , medium and short-chain dicarboxylacylcarnitines ( Fig 3A and 3B ) ., Interestingly , fatty acid-induced production of dicarboxylacylcarnitines was accompanied by elevated expression of the molecular chaperone protein BiP ( Fig 3C ) , a well-recognized marker of ER stress ., At low doses of the ER stress agent tunicamycin ( lower than required to cause cytotoxicity ) , fatty acid exposure also augmented BiP expression ( Fig 3C ) ., Together , these results point to an intriguing connection between cellular carbon load , dicarboxylic acylcarnitines and proteotoxicity ., We have analyzed metabolomics , genetics , epigenetics and transcriptomics together to establish genomewide associations between a cluster of SCDA metabolites that predict CVD events and specific genetic loci ., Our findings implicate the UPS arm of ER stress as a factor influencing SCDA levels and CVD event pathogenesis ., Several previous studies have successfully mapped metabolites to genetic loci 2 , but primarily have not triangulated such genetic variation with disease endpoints and functional studies ., Key findings of the current study include: ( 1 ) SNPs and CpG probes in genes reporting on components of ER stress were associated with levels of SCDA metabolites previously shown to predict CVD events 3–5; ( 2 ) several of these SNPs themselves also predicted CVD events; ( 3 ) some of the SNPs/genes were linked with SCDA metabolites and ER stress through eQTL analyses; ( 4 ) the isomeric composition of the peak containing the major SCDA metabolite C6-DC was clarified; and ( 5 ) in cultured cells , nutrient-induced accumulation of SCDA metabolites occurred in parallel with increases in the ER stress marker BiP ., Subjects in the CATHGEN cohort have a high prevalence of obesity , hyperlipidemia and diabetes ( Table 1 ) ., Thus , our in vitro experiment may be viewed as a mimetic of the metabolic environment to which CATHGEN subjects are commonly exposed ., Our strongest finding was for two SNPs ( HERC1 rs2228513 and USP3 rs10450989 ) that are in LD ( r2 = 0 . 99 ) despite being separated by 104 kB ., Rs2228513 is a missense variant ( serine to phenylalanine ) that is predicted to be “probably damaging” by PolyPhen , but no functional evaluation has been reported ., Rs10450989 is an intronic SNP ., The HERC gene family encodes a group of large proteins that contain multiple structural domains including a C-terminal HECT domain found in a number of E3 ubiquitin protein ligases ., HERC1 is involved in membrane trafficking and may also act as an E3 ubiquitin-protein ligase , a protein that accepts ubiquitin from an E2 ubiquitin-conjugating enzyme and then directly transfers the ubiquitin to targeted substrates ., Rs2228513 corresponds to residue 3152 , which does not map to a specific domain in the protein ., Our eQTL results suggest that this SNP is associated with differential expression of genes within a pathway reporting on the UPS ., USP3 encodes ubiquitin-specific protease 3 which mediates release of ubiquitin from degraded proteins by disassembly of the polyubiquitin chains in the ER ., Deubiquitination has been implicated in cell cycle regulation , proteasome-dependent protein degradation , and DNA repair 14 ., Interestingly , an intergenic SNP 58 kB upstream from USP3 ( rs10519210 ) was the strongest SNP associated with heart failure in a GWAS from the CHARGE consortium 15 ., Rs10519210 not associated with SCDA levels in our study ( p = 0 . 16 ) and is not in LD with rs10450989 ( r2 = 0 . 002 ) ., Our next strongest finding was for a locus in/near STON2 and SEL1L ., Rs12589750 is an intronic SNP within STON2 and rs3853422 is intergenic between STON2 and SEL1L ., SEL1L plays a role in the ER-associated protein degradation ( ERAD ) machinery , and is part of a complex necessary for the retrotranslocation of misfolded proteins from the ER lumen to the cytosol where they are then degraded by the proteasome in a ubiquitin-dependent manner ., Dysfunctional protein degradation causes ER stress ., Other mQTL included SNPs near RRM1 and STIM1; STIM1 encodes a calcium sensor in the ER that translocates to the plasma membrane upon calcium store depletion to activate calcium release-activated calcium channels ., STIM1 induction , redistribution and clustering are important during ER stress when calcium stores are depleted 16 ., FBXO25 is one of 68 human F-box proteins that serve as specificity factors for a complex composed of s-phase-kinase associated protein 1 ( Skp1 ) and cullin1 ( SCF ) , that act as protein-ubiquitin ligases , targeting proteins for destruction across the UPS ., FBXO25 is cardiac specific and acts as a ubiquitin E3 ligase for cardiac transcription factors 17 ., Rs17573278 and rs9591507 are intergenic SNPs >400 kB downstream from OLFM4 and SUGT1 ., SUGT1 is required cell cycle transitions and encodes a novel subunit of the SCF ubiquitin ligase complex 18 ., OLFM4 encodes an anti-apoptotic protein that promotes tumor growth ., The functions of the other SCDA mQTL loci are unclear ., Given the strength of association of SCDA metabolites ( factor, 3 ) with CVD and their particular strength of association in the current GWAS analyses , we chose to focus our subsequent analyses on SCDA ., However , we did also identify mQTL for LCDA and MCA , both of which have also been shown to predict CVD events ., LCDA are metabolic intermediates of long chain fatty acid oxidation in the mitochondria or peroxisomes ., The most significant mQTL for LCDA metabolite levels included PIGR , USH2a , ZNF521 and FZD3 ., PIGR is a member of the immunoglobulin superfamily and ZNF521 is involved in regulation of early B-cell factor , suggesting a potential relationship between LCDA levels and immune and/or inflammatory pathways as a link to CVD ., MCA are byproducts of mitochondrial fatty acid oxidation ., The most significant mQTL for MCA show no obvious potential biologic relationship to mitochondrial function and/or CVD ., More epidemiologic and functional work is necessary to clarify these links ., Importantly , and unique to this study , we have observed an association of mQTL and disease phenotypes ., The SNPs most significantly associated with SCDA levels ( HERC1 and USP3 ) were also associated with CVD events , with a consistent direction of effect ( G allele associating with higher SCDA levels and events ) ., STIM1|SEL1L SNPS were not associated with CVD events despite their strong association with SCDA levels; this may be due to limited power related to the low MAF in racial subsets ., Adjustment for SCDA levels in these models resulted in attenuation of the association between SNP and CVD event suggesting that the relationship between underlying mQTL and CVD events is in part or in full mediated through SCDA metabolites and not through a different biological pathway ., In combination , these results suggest potential functional and pathway relationships between SCDA metabolites and CVD events ., We also integrated transcriptomics and whole genome methylation with SNP and metabolomic data sets ., eQTL identified ER stress pathways , and specifically those reporting on the ubiquitin proteasome pathway , as associated with the SNPs linked to SCDA via GWAS , and with SCDA metabolites themselves ., Whole genome methylation identified epigenetic regulation of genes in ER stress pathways to be associated with extreme SCDA levels ., These results support the concept that these polymorphisms and ER stress underlie the relation between SCDA metabolites and CVD events ., Finally , we clarified the biochemical structure of the metabolite most strongly accounting for the C6-DC SCDA peak; these results will enable more accurate identification of the source pathways for C6-DC and other SCDA in future studies ., Many SCDAs result from the catabolism of amino acids , ω-oxidation of fatty acids or perhaps represent products of microbial metabolism 19 , but the reasons for their accumulation in plasma in at-risk subjects , and how they may be related to CVD pathogenesis remain uncertain ., Based on the convergence of GWAS , transcriptomic , metabolomic and functional data presented herein , we hypothesize that genetic and epigenetic variation predisposes to increased susceptibility to ER stress through proteasome dysfunction ( reflected by the observation of upregulation of expression of ER stress genes ) , with ER stress in turn contributing to increased production of SCDA metabolites ., This pathway of increased ER stress then leads to increased risk of CVD events , with SCDA metabolites and the genetic variants themselves predicting increased risk by reporting on this pathway ( Fig 4 ) ., Epigenetic variation could be the influence of environmental or lifestyle factors inducing methylation changes; in this working model , diet and lifestyle-induced dyslipidemia and hyperglycemia could result in methylation changes as a regulatory mechanism to handle nutrient overload , thus predisposing to dysregulated ER stress which then leads to subsequent CVD events ., The UPS arm of the ER is responsible for the removal of misfolded proteins but is sometimes insufficient , for example , in the setting of increased production of misfolded proteins ., The associated proteasome functional insufficiency can lead to cellular dysfunction and cell death , with cardiomyocytes being particularly vulnerable due to limited regenerative capability 20 ., The UPS has been hypothesized to be involved in atherosclerosis based on the recognized roles of inflammation , oxidative stress , and endothelial dysfunction in this condition , and the intertwined relationships between the UPS and those pathways 21 ., Preclinical evidence of the role of the UPS in atherosclerosis includes studies showing that oxidized LDL inhibits proteasomal activity in macrophages leading to apoptosis 22 , and data suggesting that the UPS may contribute to foam cell formation by suppression of apoptosis of lipid-bearing macrophages by aggregated LDL in in vitro models 23 ., Studies of proteasome inhibition have shown conflicting data; Hermann et al . found aggravation of atherosclerosis 24 and myocardial dysfunction 25 in pigs treated with proteasome inhibition , whereas a recent study showed reversal of uremia-induced atherosclerosis with proteasome inhibition in rabbits 26 ., Human studies suggesting the role of the UPS in atherosclerosis are limited ., Very small studies have shown greater amounts of ubiquitin conjugates in carotid endarterectomy tissues with unstable as compared with stable plaque morphologies 27 and increased UPS activity in carotid tissue from patients with symptomatic compared with asymptomatic carotid disease 28 ., While preclinical studies have suggested the role of UPS in atherosclerosis as secondary to oxidative stress or other pathophysiologies , our identification of genetic variants in UPS/ER stress genes using unbiased analyses in our human cohorts provides strong support for the direct etiologic role of the UPS in promoting long-term cardiovascular risk ., Importantly , we note that while ER stress is a common pathway in several disorders , we believe that the convergence of results on the UPS highlights its unique relationship to SCDA metabolism ., Our findings could have significant translational implications beyond CVD ., Our primary objective of discovery of novel genetic risk variants using an mQTL approach was successful; the unexpected finding of genetic variation predisposing to ER stress could have much broader importance to human disease ., Indeed , the response to ER stress is a trait that is known to be heritable in humans 29 , but the genetic architecture has not been characterized ., Equally as important , our data suggest the presence of easily quantifiable circulating biomarkers of ER stress , traditionally measureable only in tissue through ER stress-responsive gene expression studies ., Thus , these results could have more wide-reaching implications for ER stress research in humans ., Our prior work solidified the role of SCDA metabolites as predictors of CVD events 4 , 5; the current study has implications for clinical translation using SCDA metabolites for improved risk stratification even beyond CVD given the central role of normal and dysfunctional ER stress in health and disease ., The strengths of this study are the use of a priori defined discovery and validation cohorts; integration of genetics , epigenetics , metabolomics , transcriptomics in large cohorts; and careful biochemical refinement of the most strongly associated SCDA metabolite ., Importantly , this represents one of the first studies to successfully identify genetic variants through mapping of intermediate metabolomic traits that themselves associate with disease endpoints ., Our prior work had consistently identified SCDA metabolites as incremental predictors of CVD events , but little was known about the biological pathways underlying that association; the genomewide , multiple platform molecular approach taken in our study facilitated identification of the UPS more rapidly than other scientific methods ., This work also adds an important finding to the metabolomics literature , namely that SCDA metabolites may be reporting on increased or dysregulated ER stress and specifically to proteasome functional insufficiency or dysregulation ., There are limitations to the study; the study population was comprised of individuals referred due to a suspicion of CVD and thus represents a disease-prone population ., However , we note that 44% of study participants did not have significant coronary artery disease , highlighting the importance of the detailed angiographic phenotype to ensure that coronary artery disease is not confounding the relationship between genetic factors and outcome ., Further , the high burden of CVD risk factors mirrors that of the general population , enabling generalizability of the study findings ., Some of the results were isolated to a racial subset because the identified SNPs were either monomorphic or extremely rare in other races , underscoring the potential importance of including non-Caucasian races in such studies ., Race-stratified sequencing of these genomic regions may identify different variants in these genes present in other ra | Introduction, Results, Discussion, Materials and Methods | Levels of certain circulating short-chain dicarboxylacylcarnitine ( SCDA ) , long-chain dicarboxylacylcarnitine ( LCDA ) and medium chain acylcarnitine ( MCA ) metabolites are heritable and predict cardiovascular disease ( CVD ) events ., Little is known about the biological pathways that influence levels of most of these metabolites ., Here , we analyzed genetics , epigenetics , and transcriptomics with metabolomics in samples from a large CVD cohort to identify novel genetic markers for CVD and to better understand the role of metabolites in CVD pathogenesis ., Using genomewide association in the CATHGEN cohort ( N = 1490 ) , we observed associations of several metabolites with genetic loci ., Our strongest findings were for SCDA metabolite levels with variants in genes that regulate components of endoplasmic reticulum ( ER ) stress ( USP3 , HERC1 , STIM1 , SEL1L , FBXO25 , SUGT1 ) These findings were validated in a second cohort of CATHGEN subjects ( N = 2022 , combined p = 8 . 4x10-6–2 . 3x10-10 ) ., Importantly , variants in these genes independently predicted CVD events ., Association of genomewide methylation profiles with SCDA metabolites identified two ER stress genes as differentially methylated ( BRSK2 and HOOK2 ) ., Expression quantitative trait loci ( eQTL ) pathway analyses driven by gene variants and SCDA metabolites corroborated perturbations in ER stress and highlighted the ubiquitin proteasome system ( UPS ) arm ., Moreover , culture of human kidney cells in the presence of levels of fatty acids found in individuals with cardiometabolic disease , induced accumulation of SCDA metabolites in parallel with increases in the ER stress marker BiP ., Thus , our integrative strategy implicates the UPS arm of the ER stress pathway in CVD pathogenesis , and identifies novel genetic loci associated with CVD event risk . | Cardiovascular disease is a strongly heritable trait ., Despite application of the latest genomic technologies , the genetic architecture of disease risk remains poorly defined , and mechanisms underlying this susceptibility are incompletely understood ., In this study , we performed genome-wide mapping of heart disease-related metabolites measured in the blood as the genetic traits of interest ( instead of the disease itself ) , in a large cohort of 3512 patients at risk of heart disease from the CATHGEN study ., Our goal was to discover new cardiovascular disease genes and thereby mechanisms of disease pathogenesis by understanding the genes that regulate levels of these metabolites ., These analyses identified novel genetic variants associated with metabolite levels and with cardiovascular disease itself ., Importantly , by utilizing an unbiased systems-based approach integrating genetics , gene expression , epigenetics and metabolomics , we uncovered a novel pathway of heart disease pathogenesis , that of endoplasmic reticulum ( ER ) stress , represented by elevated levels of circulating short-chain dicarboxylacylcarnitine ( SCDA ) metabolites . | null | null |
journal.pntd.0000553 | 2,009 | Mycolactone Gene Expression Is Controlled by Strong SigA-Like Promoters with Utility in Studies of Mycobacterium ulcerans and Buruli Ulcer | Mycobacterium ulcerans is the causative agent of Buruli ulcer ( BU ) an emerging but neglected disease found predominantly in tropical regions of the world and with an increasing incidence in West and Central Africa 1 , 2 ., BU is a chronic infection of subcutaneous tissue that can result in high morbidity such as permanent scarring and functional disabilities ., The combination of rifampin and an aminoglycoside for four to eight weeks leads to the healing of early lesions without radical surgery and is now the recommended standard regimen 3 ., However , substantial tissue damage often necessitates surgery 4 ., The social and economic burden of BU can be severe , particularly in impoverished rural regions of West Africa where the prevalence of BU is sometimes higher than that of the two most significant mycobacterial diseases , leprosy and tuberculosis ., Cases of BU are usually clustered around swamps and slow-flowing water and while the mode of transmission of M . ulcerans is unknown , evidence to date suggests , fish 5 , snails 6 and certain carnivorous aquatic insects 7 , 8 can all harbour the bacterium ., Recent studies in Australia suggest mosquitoes may play a role in transmission 9 , 10 ., A major factor influencing the pathology of Buruli ulcer is the production by M . ulcerans of a secondary metabolite called mycolactone 11 ., Mycolactone is an immunosuppressive and cytotoxic macrocyclic polyketide , characterised by a 12-membered macrolactone core appended to a highly unsaturated acyl side chain 11 , 12 ., Polyketides are a class of naturally occurring compounds , some of which have potent pharmaceutical activity such as the immune suppressor rapamycin , the antibiotic erythromycin A , and the antiparasitic agent avermectin 13–15 ., Why M . ulcerans produces mycolactone is unknown ., However , studies on the effect of the molecule in cell culture and animal models have shown that in the microgram range it has cytotoxic properties , while at sub-cytotoxic concentrations it has immunomodulatory properties , most strikingly the inhibition of tumour necrosis factor production by monocytes and macrophages 16–18 ., In mice , mycolactone has been shown to concentrate within peripheral blood monocytes 19 ., Mycolactone synthesis is dependent on the pMUM megaplasmid found in M . ulcerans and closely related mycobacteria ( Figure 1 ) 20–23 ., This plasmid contains three , very large genes ( mlsA1: 51 kb , mlsA2: 7 kb , and mlsB: 42 kb ) that encode type I polyketide synthases ( PKS ) ., MlsA1 and MlsA2 synthesize the upper side chain and macrolactone core , whilst MlsB synthesizes the acyl side chain 22 ., A putative beta-ketoacyl transferase encoded by another pMUM gene , mup045 , is thought to catalyse the ester linkage between the acyl side chain and the macrolactone core whilst a P450 hydroxylase , encoded by mup053 , oxidizes the side chain at C12′ ( Figure 1 ) 22–25 ., A third gene , mup038 , is predicted to encode a type II thioesterase that might be required for removing aberrant polyketide extension products from the Mls PKS that form during synthesis ., An unusual feature of the mycolactone PKS is the very high level of sequence identity between domains of the same function ( 98 . 7–100% nt identity and 98 . 3–100% aa identity ) ., This observation suggested that the evolution of the locus may be recent and also prone to rearrangements that result in either loss of mycolactone production or production of new mycolactones ., These hypotheses have recently gained support by studies that have shown, ( i ) all mycolactone producing mycobacteria ( which includes M . ulcerans and some closely related fish and frog pathogens ) have recently evolved from a common Mycobacterium marinum ancestor by pMUM plasmid acquisition 23–28 ,, ( ii ) laboratory passaging leads to mycolactone negative mutants through spontaneous deletion of mls gene fragments 29 , and, ( iii ) natural swapping of particular acyltransferase and ketoreductase domains and loss or gain of entire extension modules in some strains of M . ulcerans has led to the production of new mycolactones 30 , 31 ., However , there have been very few studies of gene expression in M . ulcerans ., Therefore , in this study we began by investigation of the mycolactone-associated genes mlsA1/mlsA2 , mlsB , mup045 , mup053 and mup038 ., Promoter regions were mapped upstream of the above genes using a GFP reporter ., Putative transcriptional start sites and promoter sequences were then identified by primer extension analysis and site-directed mutagenesis ., The GFP reporter containing the promoter region of the mls genes was then used to transform M . ulcerans ., This recombinant GFP M . ulcerans fluoresced brightly and was used to follow infection in both mice and mosquito larvae ., The bacterial strains and plasmids used in this study are described in Table S1 ., All cloning experiments were performed in Escherichia coli DH10B , cultivated in Luria-Bertani ( LB ) broth at 37°C or on LB agar containing 50 µg kanamycin , for 16 hours at 37°C ., Mycobacterial strains were grown in Middlebrook 7H9 medium ( Difco ) supplemented with albumin ( 6 . 25% ( w/v ) ) , dextrose ( 2 . 5% ( w/v ) ) , sodium chloride ( 1 . 1% ( w/v ) ) , catalase ( 5×10−4% ( w/v ) ) and 0 . 05% ( v/v ) Tween−80 at 37°C for Mycobacterium smegmatis and 30°C for M . marinum and M . ulcerans ., Mycobacteria were also cultured on 7H10 agar supplemented with OADC ( Difco ) ., Recombinant mycobacteria were cultivated with kanamycin at a final concentration of 25 µg ml−1 ., Standard methods were used for cloning , PCR and DNA sequencing ., The oligonucleotides used in this study for PCR , RT-PCR and DNA sequencing are listed in Table S2 ., Genomic DNA was extracted from mycobacteria as described 23 ., The broad host range , promoterless , GFP ( gfpmut3 ) vector pSM20 , that replicates in E . coli , Corynebacterium sp ., and Mycobacterium sp , was used for all promoter cloning experiments 32 ., PCR products derived from the upstream regions were modified using oligonucleotides described in Table S2 and ligated into the unique restriction enzyme sites immediately upstream of the gfp gene in pSM20 ( Figure 1 ) ., Constructs were confirmed to be correct by DNA sequencing and then transformed into M . smegmatis mc2155 as described 33 ., Electrocompetent M . marinum and M . ulcerans were prepared as described 34 and these cells were transformed with 10 µg of DNA from plasmid pJKD2893 ( Table S1 ) ., The constructs were confirmed to be correct in mycobacteria by Southern hybridization and back transformation to E . coli ., Acetone soluble lipids were extracted from recombinant M . ulcerans and analysed by LC-MS for the presence of mycolactones as previously described 35 ., GFP expression in pSM20 and derivatives ( Table S1 ) was measured using a FLUOstar OPTIMA plate scanner ( BMG Lab Technologies ) ., M . smegmatis strains were grown to an OD of 1 . 0 using a WPA CO8000 cell density meter ( Isogen Life Science ) ., For each strain , 30 µl of starter culture was added to each of 16 wells of a 96-well flat-bottomed clear plate containing 150 µl of fresh 7H9 medium ., Plates were incubated at 37°C for 30 mins ., Each well was scanned using an excitation filter of 485 nm and an emission filter of 520 nm ., Fluorescence readings were taken every 10 minutes and the average of 20 flashes per well was taken to be the measure of fluorescence ., Prior to each reading , the plates were shaken for 5 minutes in an orbital motion ., Replicates were averaged for each experiment and the average value for the vector-only control was taken as background and subtracted from the average at each time point ., Total RNA was prepared from E . coli using the RNeasy mini kit as described and per the manufacturers instructions ( Qiagen ) 36 ., For M . ulcerans , a 0 . 5 volume of RNAlater ( Qiagen ) was added to 100 ml of late log-phase culture and allowed to stand at room temperature for 10 minutes prior to centrifugation at 4 , 600 g , for 10 minutes ., The resultant cell pellet was washed in 1 ml of 0 . 5% ( v/v ) Tween-80 per 50 mg of cells ( wet weight ) , resuspended in 800 µl of RNA lysis buffer ( 0 . 12 M sodium acetate ( pH 4 . 0 ) , 9 . 6% ( v/v ) liquid Pyroneg ( Diversey ) , pH 4 . 0 ) and then added to 250 µg of glass beads ( Sigma Aldrich ) , with 600 µl of acidified phenol:chloroform ( pH 4 . 0 ) ( Sigma Aldrich ) ., Cells were disrupted with a FastPrep tissue homogenizer ( Savant Instruments ) for 45 seconds , at speed 6 and chilled on ice for 5 minutes ., The aqueous phase was then re-extracted with chloroform:isoamylalcohol ( 24:1 ) and precipitated with isopropanol , and 3 M sodium acetate ( pH 4 . 6 ) ., Two 70% ( v/v ) ethanol washes were performed and the pellet was dried briefly under vacuum and resuspended in 100 µl of DEPC water ., RNA in this preparation was then further purified using an RNeasy extraction kit , including an on-column DNase treatment , following the manufacturers recommendations ( Qiagen ) ., For RNA extraction from M . smegmatis and M . marinum the following modifications to the above method were used ., The cell pellet was first resuspended in 2 ml of lysis solution ( 20 mM potassium acetate ( pH 4 . 8 ) , 1 mM EDTA , 0 . 5% ( v/v ) SDS , 100 µg proteinase K ml−1 ) ., One millilitre was added to 250 µg of glass beads with 700 µl acidified phenol:chloroform pH 4 . 0 ., Cells were disrupted by three cycles in a FastPrep instrument at speed 5 , for 30 seconds , and then centrifuged at 17 , 900 g for 10 minutes ., The aqueous phase was recovered and extracted once with 500 µl phenol:chloroform ( pH 4 . 0 ) followed by a chloroform only extraction ., Nucleic acids were precipitated as above and RNA extraction proceeded as for M . ulcerans using the RNeasy extraction kit ., The primer extension protocol used was modified from Lloyd et al . , 37 ., Two reverse transcription reactions were performed ., To the RNA-primer mix , 6 µl of 5x first strand buffer ( Invitrogen ) , 15 mM DTT ( Invitrogen ) , 1 mM dNTPs ( Promega ) , 1 U RNasin ( Promega ) and 100 U of Superscript II RNase H- reverse transcriptase ( Invitrogen ) were added ., After one hour at 42°C , 2 µl of 5x first strand buffer , 1 . 5 mM dNTPs , 1 U RNasin , 15 mM DTT and 100 U of Superscript II was added and incubated for a further hour at 42°C ., Ten nanograms of RNaseA ( Sigma ) was then added and allowed to incubate at 37°C for 30 minutes ., The resultant cDNA was precipitated and washed once with 70% ( v/v ) ethanol , dried and stored at −20°C until analysis ., Capillary electrophoresis was performed on an Applied Biosystems 3730 DNA analyzer using Liz™ 500 size standards to generate a standard curve ( Applied Biosystems ) ., Genemapper® version 3 . 7 ( Applied Biosystems ) was used to analyze the sample files with automated allele calling verified by manual inspection ., The sized cDNA fragments were then mapped to their respective first strand synthesis primer binding sites to identify the putative transcription start site ., From the alignment of each of the SigA , C , D , E , F , H & L promoters 38 , nucleotide frequency counts were derived and used to construct a library of 110 position specific scoring matrices ( PSSMs ) for each sigma factor ( PSSMs available upon request ) ., This allowed the gap between the -35 and -10 signals to vary between 14 and 23 residues , and the gap between the -10 signal and the TSP to vary between 3 and 13 residues ., PoSSuM software 39 was used to scan the pMUM001 genome for high scoring hits to these PSSM libraries 40 , using a background model consistent with the G+C biased nucleotide distribution of pMUM001 ., A p-value significance cutoff of 0 . 0001 was used ., Splice overlap extension PCR 41 was used to alter the sequence of putative promoter motifs with oligonucleotides 1075-F and 1074-R for mlsA1/mlsB , 1667-F and 1668-R for mup045 , and 1669-F and 1670-R for mup053 ( Table S2 ) ., Each PCR reaction consisted of 20 cycles of 94°C for 1 minute , 50°C for 1 minute and 72°C for 3 minutes then 94°C for 1 minute , 72°C for 10 minutes and held at 4°C ., Two overlapping PCR products were obtained and 2 µl ( ∼50 ng DNA ) of each were used in a subsequent reaction using the outermost primers for each product to yield a complete fragment incorporating both products ., Each product was then ligated into pSM20 as described above ., Mutations were confirmed by DNA sequencing ., Ten , six-week-old female BALB/c mice ( Charles River France , http://www . criver . com/ ) were injected subcutaneously into the tail with 30 µl of a suspension containing 5×104 bacteria ., To favour the growth of the GFP-expressing bacilli , animals received 0 . 1 ml of a solution containing 80 mg/ml of kanamycin ( 1% w/v ) , administered by oral gavage every day ., The mice were killed and their tails were collected fifty days after inoculation ., Mice were maintained in the animal house facility of the Centre Hospitalier Universitaire , Angers , France ( Animal Ethics Committee , Centre Hospitalier Universitaire , Agreement A 49 007 002 ) , adhering to the institutions guidelines for animal husbandry ., The tissue specimens from mice were minced with disposable scalpels in a Petri dish and ground with a Potter–Elvehjem homogeniser , size 22 , ( Kimble/Kontes , Vineland , NJ ) , in 0 . 15 M NaCl to obtain a tenfold dilution ., The suspension was decontaminated to remove other bacteria using an equal volume of N-acetyl-L-cysteine sodium hydroxide ( 2% ) 42 and inoculated on 7H10 agar supplemented with OADC ( Difco ) , containing 25 µg/ml of kanamycin ., For histological examination , tissues were fixed in 4% paraformaldehyde in phosphate buffer ( pH 7 . 4 ) ., Decalcification of the tissue was performed for 7 days in 0 . 1 M of EDTA solution in PBS ., Samples were frozen in isopentane cooled to −140°C in liquid nitrogen and stored at −80°C for subsequent histochemical analysis ., Eight-micron thick transverse sections were cut at −30°C on a cryostat ( Jung-Reichert Cryocut 1800 , Cambridge Instruments , Germany ) and kept at −80°C until histochemical processing , which was done within 1 week of sectioning ., For detection of GFP-expressing bacilli , tissues were counterstained with DAPI , with endogenous phosphatase activity first detected using alkaline phosphatase substrate kit I ( Vector Laboratories ) ., The preparation was then mounted in Vectashield mounting medium containing DAPI ( Vector Laboratories ) and the samples were visualized using fluorescence microscopy ( Leica DM5000B ) ., Hematoxylin phloxine saffron and Ziehl Nielsen staining were performed according to standard procedures ., Mosquito larvae ( Aedes camptorhynchus ) between first and second instar were distributed into 4×50 ml plastic tubes ( 10 larvae per tube ) , containing 20 ml of sterile tap water ., To three groups of four tubes were added 1 . 5 ml of an aqueous slurry of possum faecal material containing either 5×106 colony forming units ( cfu ) M . ulcerans-GFP , 5×106 cfu M . marinum-GFP , or possum faecal material alone ., The larvae were left to feed on the material for one week at 24°C ., At the end of one week and also at the end of every subsequent week up to week five , all larvae were transferred to new tubes containing 20 ml of sterilized tap water ., The original tubes spiked with possum faecal material were kept at room temperature and at the commencement of each week 500 µl of water from each of these tubes was tested by IS2404 and ppk qPCR to estimate the residual quantity in the water of M . ulcerans and M . marinum respectively 43 ., Results were reported as cfu by reference to standard curves for each PCR and bacterial species , correlating qPCR Ct values with cfu 43 ., From weeks 2–5 the larvae were sustained with small quantities of fish food added to each tube ., A larva was taken from each tube as it progressed through each instar and tested by IS2404 PCR for the presence of M . ulcerans as described 43 ., A selection of 4th instar larvae were also fixed overnight in 10% formaldehyde in PBS ( v/v ) then mounted in cedarwood oil ( Matheson , Coleman and Bell ) on a glass slide for examination by fluorescence microscopy with an Olympus BX51 microscope ( Olympus , Tokyo , Japan ) with the following filter sets: DAPI ( Blue ) ex: 360–70 nm , em: 420–60 nm , FITC ( Green ) ex: 450–80 nm , em: 535 nm , TRITC ( red ) ex: 535 nm , em: 635 nm ., Images were acquired using an Olympus DP-70 digital camera and merged using DP controller software ( version 1 . 1 . 1 . 71 ) or Adobe Photoshop ( version 8 ) These experiments were terminated before the insects progressed to pupal and adult developmental stages ., By cloning DNA fragments ranging from 229 bp–1646 bp located immediately upstream of mlsA1/mlsB ( these genes have a duplicated start and upstream sequence so one cloned fragment was sufficient to analyse both genes ) , mlsA2 , mup038 , mup045 and mup053 in the promoterless GFP E . coli/Mycobacterium reporter vector pSM20 ( Table S1 , Figure 1 ) we were able to discover regions containing promoter activities ., The resulting plasmids were used to transform E . coli , M . smegmatis and , for the mlsA1/mlsB construct , M . marinum and M . ulcerans were also transformed ., Bacteria were cultured in 96-well plates for 2 hours at 37°C and expression of GFP for each strain was assessed by continuous fluorescence measurements ., E . coli expressing GFP from the strong , constitutive promoter srp ( pSM22 ) 32 and M . smegmatis expressing GFP from the sigA promoter from Mycobacterium bovis BCG ( pJKD3042 ) 44 were used as positive controls for each genus ., Results were expressed as fold changes in fluorescence above the levels detected in bacteria containing the empty vector pSM20 ., The results for mlsA1 and mlsB are summarized in Figure 2 and show that strains containing the construct pJKD2893 with the region 1646 bp upstream of mlsA1/mlsB , led to detectable GFP expression in E . coli , and high levels of GFP expression in M . smegmatis and M . marinum ( Figure 2A ) ., A single copy version of pJKD2893 was also created where a DNA fragment spanning the 1646 bp mls upstream region and gfp gene from pJKD2893 was subcloned into the mycobacterial integrating shuttle vector , pJKD8003 resulting in pJKD3111 ., M . marinum transformed with pJKD3111 expressed GFP 40-fold less than the same strain containing pJKD2893 ( Figure 2A ) ., To further localize the region conferring promoter activity within the 1646 bp upstream of mlsA1/mlsB , four overlapping sub-clones of this region were prepared by PCR , cloned into pSM20 and used to transform E . coli and M . smegmatis ., Comparison of GFP expression in these constructs in another time course experiment , comparing fluorescence with the full-length 1646 bp fragment and controls , clearly showed that promoter activity was restricted to a 413 bp fragment located between nucleotide positions 35996-36409 for mlsA1 and 100821-101234 for mlsB in pMUM001 ( Figure 2B ) ., The region 1440 bp upstream of mup045 and 1466 bp upstream of mup053 also led to significant GFP expression in M . smegmatis , 8–15 fold above background , but these regions showed little transcriptional activity in E . coli ( Figure S1 ) ., No fluorescence was observed in either E . coli or M . smegmatis for strains containing the 229 bp region upstream of mup038 ( pJKD3269 ) or the 1096 bp region upstream of mlsA2 ( pJKD3041 ) ( data not shown ) ., These experiments demonstrate that the regions upstream from mlsA1/mlsB , mup045 and mup053 all harbour at least one strong promoter ., To identify TSPs upstream of each gene primer extension ( PE ) analysis was performed using RNA extracted from M . ulcerans Agy99 ., For mlsA1/mlsB , RNA was also extracted from M . marinum harbouring the GFP expression construct pJKD2893 ., One or more 5′ 6-FAM-labeled antisense oligonucleotides were used to prime cDNA synthesis to determine the TSP for mlsA1/mlsB , mup045 , and mup053 ( Table S2 ) ., Single , distinct PE products were identified for all three regions using multiple RNA preparations ( Figure S2 , Figure S3 ) ., Size fragment analysis of the PE products suggested single TSPs at 533 bp ( T533 ) upstream of the mlsA1/mlsB translational start ( Figure S2 ) , 207 bp upstream of mup045 ( T207 ) , and 68 bp upstream of mup053 ( T068 ) ( Figure S3 ) ., Primer extension analysis of mup038 and mlsA2 was not attempted due to the lack of promoter activity observed with the wild type sequences in the GFP reporter assays ., Several studies of promoters in mycobacteria facilitated the construction of position-specific scoring matrices ( PSSMs ) to perform in silico searches for potential regulatory regions in DNA sequences 38 ., We used sigma factor-specific libraries of PSSMs to scan the regions upstream of the three TSPs identified by our PE analysis ., High-probability SigA-like promoter motifs were predicted in the regions upstream of mlsA1/mlsB and mup045 and a SigD-like motif was predicted upstream of mup053 ( Table 1 ) ., To confirm the in silico promoter predictions , the GFP expression constructs spanning the putative -10 sequences from mlsA1/mlsB ( pJKD2893 ) , mup045 ( pJKD3040 ) and mup053 ( pJKD3039 ) were mutated by PCR ( Table 1 ) ., GFP production by E . coli and M . smegmatis harbouring these constructs was assayed as before by continuous fluorescence measurements over 2 hours at 37°C ., Fluorescence production was compared with the same strains containing the wild-type putative promoter sequences ., Mutation of the proposed −10 boxes for both the mlsA1/mlsB and mup045 reduced fluorescence in M . smegmatis to less than 4% of the wild-type sequences , strongly suggesting these sequences are functional −10 motifs , required for proper binding of the sigma factor and RNA polymerase to initiate transcription ( Figure 2B ) ., Mutation of the proposed −10 box from mup053 had no impact on GFP expression ., The strength of the mls promoter led us to develop a GFP-producing strain of M . ulcerans that might be useful in studies of pathogenesis or transmission ., We transformed M . ulcerans JKD8049 with plasmid pJKD2893 , resulting in highly green fluorescent M . ulcerans ( JKD8083 or M . ulcerans-GFP ) , with fluorescence expression more than 100 fold above empty vector ( Figure 2A ) ., To ensure that GFP expression did not stop mycolactone production we performed cell LC-MS analysis of acetone-soluble lipids from cultures of JKD8083 and confirmed the presence of mycolactone A/B and C ( Figure S4 ) ., Mouse-tail infection is a well-established animal model for studying M . ulcerans ., Forty days after subcutaneous inoculation of 105 M . ulcerans-GFP oedema was observed and on the 50th day the lesion became ulcerated and the mice were killed ., Histological study of the ulcerated region showed an area of necrosis consistent with wild type M . ulcerans infection ( Figure 3A , Figure 3B ) ., Granulomatous inflammation was not observed ., Acid-fast bacilli were localized in clumps in necrotic areas ( Figure 3C ) and expressed green fluorescent protein ( Figure 3D ) ., The viability of these bacteria was demonstrated by re-isolating them in bacterial culture media ., These results demonstrate that M . ulcerans-GFP is virulent in the mouse model and provokes lesions typical of M . ulcerans infection ., Adult mosquitoes in some Buruli ulcer endemic regions of Australia have tested PCR positive for M . ulcerans and epidemiological evidence suggests a role for biting insects in the disease ecology of M . ulcerans 45 , 46 ., These data and the presence of M . ulcerans in possum faecal material from the same endemic regions has led to the hypothesis that larval stages of mosquitoes may and probably do ingest M . ulcerans as well as other bacteria via filter feeding activity on decomposing , faecally contaminated environments 47 ., We mimicked this environment by establishing simple aquatic microcosms , seeded with 1 or 2 instar Aedes camptorhynchus larvae that were then transiently fed with possum faecal material , spiked with either M . ulcerans-GFP or M . marinum-GFP ( Figure 4A ) ., M . ulcerans and M . marinum were initially liberated into the water from the food source but neither bacterial species were detectable in water by week 4 ( Figure 4B ) ., Analysis of 4th instar larvae at week 4 by fluorescence microscopy revealed an accumulation of M . ulcerans primarily within the larval midgut and around the mouthpart ( Figure 4C ) ., Fourth instar larvae assayed by PCR for M . ulcerans had a mean bacterial load of 27 , 300±15 , 200 cfu ( n\u200a=\u200a4 ) ., The same pattern of accumulation within the insect was not seen with M . marinum-GFP with very few fluorescent bacteria observed in association with larvae ( Figure 4C ) ., Neither M . ulcerans or M . marinum were detected in the microcosms containing mosquito larvae only ( Figure 4C ) ., These data show that mosquito larvae in contaminated aquatic environments were able to ingest and maintain M . ulcerans within regions of the digestive tract over a significant time period ., In this study we have explored gene expression of six pMUM001 genes required or implicated in mycolactone synthesis and attempted to identify their transcriptional start sites and promoter motifs ., Using a combination of primer extension and in silico analysis together with a GFP reporter system , we were able to identify a SigA-like promoter that drives expression of the mycolactone polyketide megasynthases mlsA and mlsB in M . ulcerans ., Primer extension analyses with mRNA extracted from E . coli , M . smegmatis and M . marinum bearing the GFP reporter construct pJKD2893 and from wild-type M . ulcerans Agy99 all consistently demonstrated a transcription start point ( TSP ) 533 bp upstream of the mlsA1/mlsB initiation codons ., The primer extension analysis was fully supported by the GFP expression data , wherein only strains containing expression constructs that spanned the TSP at T533 produced fluorescence ., These results indicate the presence of a strong promoter preceding position T533 ., Sequence scanning using PoSSuM of the region immediately upstream of T533 for mycobacterial consensus promoter sequences predicted a high probability SigA-like promoter ( Table 1 ) ., Site-directed mutagenesis of the putative −10 box by substitution of three nucleotides completely abolished GFP expression ( Table 1 , Figure 2B ) , implicating this sequence in RNA polymerase ( RNAP ) binding ., The mlsA/mlsB promoter lies between two pseudogenes that once encoded transposases ., These CDS appear to be remnants of two distinct insertion sequence elements ( ISE ) as the partial transposase sequences display similarity to two different IS families ( IS3 family for MUP034/MUP042 and the IS6 family for MUP033/MUP041 ) 25 ., These vestigial ISE are quite distinct to the two high copy number elements , IS2404 and IS2606 present in M . ulcerans ., It is possible that the T533 promoter was once a component of an ISE ., A role for ISE in altering gene expression in mycobacteria has been reported 48 ., Similarly , we investigated DNA sequences upstream of mup045 and found a TSP at T207 with a potential SigA promoter element predicted by PoSSuM and confirmed by a loss of GFP expression in M . smegmatis following mutagenesis of the proposed −10 box ., The principal mycobacterial sigma factor sigA is utilized by genes expressed during exponential growth 49 , thus the data from mlsA/B and mup045 fit well with our previous report that show these genes are constantly expressed during exponential growth in the heterologous host , M . marinum 35 ., PoSSuM sequence scanning predicted a SigA-like promoter upstream of mup045 , a finding confirmed by mutagenesis of its putative −10 motif ( Table 1 ) ., The same in silico search suggested a SigD-like promoter element upstream of mup053 ., However , mutation of the putative −10 motif for this gene resulted in no significant difference in GFP production in either E . coli and M . smegmatis backgrounds compared to wild type sequence , indicating that this was not the promoter region or that the introduced mutations were not sufficiently different to the wild type sequence to alter transcription ., The latter scenario seems more likely given the low complexity of SigD −10 consensus sequences ( Table 1 ) ., The discovery in this study of the strong SigA-like promoter , active in diverse bacterial genera , and driving expression of the mycolactone mls PKS genes prompted us to transform M . ulcerans with a reporter plasmid with GFP under the control of the T533 mls promoter , resulting in the highly green fluorescent strain M . ulcerans JKD8049 ., M . ulcerans-GFP still produced mycolactone and was capable of causing disease in a mouse-tail infection model ., Interestingly , GFP expression was more than 2-fold higher in M . ulcerans than in M . marinum harbouring the same plasmid ( Figure 2A ) , suggesting additional regulatory factors might augment mls expression in M . ulcerans ( or conversely , repress gene expression from the same promoter in other mycobacteria ) ., The high level of mls promoter activity and the presence of viable M . ulcerans-GFP in the ulcerated tail tissue 50 days post inoculation implies that there was sustained expression of the mycolactone PKS and presumably sustained mycolactone production by the bacteria within necrotic tissue ( Figure 3 ) ., These observations demonstrate the utility of this M . ulcerans-GFP strain as a tool for following the dynamics of mls gene expression during infection and understanding the role of mycolactone in pathogenesis ., We also used M . ulcerans-GFP to explore the previously reported association of M . ulcerans with Aedes camptorhynhcus mosquitoes 45 ., Here , we addressed the specific question of whether or not A . camptorhynchus larvae could ingest M . ulcerans via feeding on possum faecal material and whether the bacteria could persist through the larval growth stages ., Many larval mosquito species filter feed on microbial particles and detritus where they aggregate at air-water interfaces near plant stems and algal mats in lentic waters 50 , 51 and a recent report has also suggested that M . ulcerans can persist within the gut of Ochlerotatus triseriatus mosquito larvae 52 ., We were also able to observe the presence of M . ulcerans within the gut contents of mosquito larvae in laboratory experiments ., However , the mode of larval ingestion via possum faecal pellets that we have employed in this study , presents a natural and viable pathway that A . camptorhynchus larvae as well as other filter-feeding macroinvertebrates might become infected for a long period of time with M . ulcerans ., The peritrophic matrix is a proteoglycan ‘sleeve’ that separates food sources from the gut epithelium in insects 53 and our data suggests an accumulation of M . ulcerans within this matrix through each instar ( Figure 4C ) ., The significantly greater mean bacterial load of M . ulcerans-GFP found in fourth instar larvae compared to M . marinum-GFP may indicate that A . camptorhynchus larvae are able to digest and assimilate M . marinum-GFP better than M . ulcerans or that M . ulcerans is able to persist and perhaps multiply within the peritrophic matrix ., Production of GFP in the mosquito larvae also indicates that the mycolactone mls genes are likely to be expressed and producing mycolactone under these conditions ., Whether or not M . ulcerans can be transferred through larval , pupal and then adult insects remains to be tested ., Experiments are now underway to examine vertical transmission of M . ulcerans within mosquitoes ., The data presented in this study provide the first insights into gene expression within the mycolactone biosynthesis locus and the development of M . ulcerans-GFP , a strain where fluorescence and toxin gene expression are linked thus providing a tool for studying Buruli ulcer pathogenesis and potential transmission to humans . | Introduction, Materials and Methods, Results, Discussion | Mycolactone A/B is a lipophilic macrocyclic polyketide that is the primary virulence factor produced by Mycobacterium ulcerans , a human pathogen and the causative agent of Buruli ulcer ., In M . ulcerans strain Agy99 the mycolactone polyketide synthase ( PKS ) locus spans a 120 kb region of a 174 kb megaplasmid ., Here we have identified promoter regions of this PKS locus using GFP reporter assays , in silico analysis , primer extension , and site-directed mutagenesis ., Transcription of the large PKS genes mlsA1 ( 51 kb ) , mlsA2 ( 7 kb ) and mlsB ( 42 kb ) is driven by a novel and powerful SigA-like promoter sequence situated 533 bp upstream of both the mlsA1 and mlsB initiation codons , which is also functional in Escherichia coli , Mycobacterium smegmatis and Mycobacterium marinum ., Promoter regions were also identified upstream of the putative mycolactone accessory genes mup045 and mup053 ., We transformed M . ulcerans with a GFP-reporter plasmid under the control of the mls promoter to produce a highly green-fluorescent bacterium ., The strain remained virulent , producing both GFP and mycolactone and causing ulcerative disease in mice ., Mosquitoes have been proposed as a potential vector of M . ulcerans so we utilized M . ulcerans-GFP in microcosm feeding experiments with captured mosquito larvae ., M . ulcerans-GFP accumulated within the mouth and midgut of the insect over four instars , whereas the closely related , non-mycolactone-producing species M . marinum harbouring the same GFP reporter system did not ., This is the first report to identify M . ulcerans toxin gene promoters , and we have used our findings to develop M . ulcerans-GFP , a strain in which fluorescence and toxin gene expression are linked , thus providing a tool for studying Buruli ulcer pathogenesis and potential transmission to humans . | Buruli ulcer ( BU ) is a serious skin infection of humans predominantly occurring in West and Central Africa ., The disease is caused by infection with Mycobacterium ulcerans , a bacterium that produces an unusual toxin called mycolactone ., There are many unanswered questions surrounding BU , particularly regarding the role of mycolactone in disease and how M . ulcerans is transmitted to humans ., Here , we have partly addressed these questions by identifying genetic factors controlling the transcription of the mycolactone genes ., Using a variety of experimental approaches , including green fluorescent protein ( GFP ) as a reporter of gene expression , we have identified strong promoters that drive transcription of the mycolactone genes in M . ulcerans ., We then used our GFP reporters to produce highly fluorescent M . ulcerans-GFP that were readily visualized by microscopy ., Mosquitoes have been proposed as a potential vector of M . ulcerans so we used M . ulcerans-GFP in feeding experiments with mosquito larvae ., M . ulcerans-GFP accumulated within the insects , whereas other mycobacteria did not ., This is the first report of the mycolactone gene promoters , and we have used our findings to develop M . ulcerans-GFP , a strain in which fluorescence and toxin gene expression are linked , thus providing a powerful tool for studying Buruli ulcer . | genetics and genomics/gene expression | null |
journal.ppat.1000922 | 2,010 | The Early Stage of Bacterial Genome-Reductive Evolution in the Host | The genomes of host-adapted bacteria , including endosymbionts and obligatory intracellular pathogens , go through reductive evolution 1 , 2 , 3 ., Such changes are partly due to a reduced pressure to maintain genes that are not essential for survival in the host ., Similarly , decreased efficiency of purifying selection , resulting from the reduced population size from a restricted life , results in inactivated genes , including beneficial genes , through genetic drift 3 ., During the early stage of the genome reduction process , the majority of genes are lost as large chromosomal fragments spanning multiple genes ., Such genome reduction has been documented in diverse bacterial groups , including Firmicutes , Chlamydiae , Spirochetes , and γ-Proteobacteria 1 , 3 , 4 , 5 , 6 , 7 ., Most of these bacteria have large expansion of IS elements ( insertion sequences ) , and thus it has been suggested that the IS elements may play an essential role during the genome reduction process 1 , 3 , 8 , 9 , 10 ., Burkholderia pseudomallei and Burkholderia mallei belong to the ß-Proteobacteria family , and are the causative agents of melioidosis and glanders , respectively 11 , 12 , 13 , 14 , 15 , 16 , 17 ., B . mallei has very recently ( ∼3 . 5 myr ) evolved from a clone of B . pseudomallei through extensive genome reduction 18 , 19 , accounting for as much as 1 . 41 Mb or 20% of the genome , as estimated by the size difference between the genomes of B . mallei ATCC 23344 and B . pseudomallei K96243 18 , 20 , 21 ., Concomitant with this process , B . mallei became constantly associated with mammalian hosts , specifically equines 22 , 23 , while B . pseudomallei maintains an opportunistic pathogenic lifestyle 17 ., Preliminary analyses of the two type strains , B . mallei ATCC 23344 and B . pseudomallei K96243 , have suggested that genome reduction and rearrangement in B . mallei were mediated by IS elements that are widely spread throughout the genome 20 , 21 ., Genes that have been deleted from the B . mallei genome but are maintained in B . pseudomallei include genes that are required for environmental survival ., Many of these genes encode metabolic functions for the synthesis of metabolites or the utilization of various sugars and amino acids , without which bacterial propagation in the environment could be significantly hindered 20 ., While the genomic reduction during bacterial restriction to their hosts has been well documented 1 , 8 , 10 , most of the stepwise processes have not yet been elucidated ., The B . mallei genome has unique significance , as it is much younger than the other genomes in which the genome-reductive evolutionary processes have been most studied to date , including Buchnera ( >150 million years ) and other much older groups 1 , 3 , 4 , 5 , 6 , 7 ., The studies with these older genomes have been challenging due to the subsequent genomic- and nucleotide-level mutations that accumulated over a long evolutionary history ., In this study , we dissected 10 genomes each of B . pseudomallei and B . mallei to understand the early-stage processes that drive genome-reductive evolution in host-associated bacteria ., It is well known that bacteria specialized to a ( host ) niche , often have a large number of IS elements compared to their free-living relatives 1 , 3 , 8 , 9 , 10 ., Likewise , by comparing genome sequences , we found that three types of IS elements , ISBma1 , ISBma2 , and IS407A , were significantly increased in B . mallei compared to B . pseudomallei ( Fig . 1A ) ., By contrast , other types of IS , including IS1356 , ISBma3 , ISBma4 , and ISBma5 were found in low copy number in both species of bacteria ., These elements appeared to be mostly degenerate evolutionary remnants ( i . e . , part of the IS disrupted or deleted ) of the Burkholderia lineage ., ISBma1 , ISBma2 , and IS407A also had degenerate elements in each species; the ISBma1 elements had the highest levels of degeneration ( 44% ) , followed by ISBma2 ( 20% ) , and by IS407A ( 5% ) ( Fig . 1A ) ., Intriguingly , up to almost 90% of ISBma1 , ISBma2 , and IS407A ( 88 . 5% , 86 . 1% , and 89 . 6% , respectively ) were found to be present at the corresponding loci in all 10 B . mallei strains , when examined after the rearranged genomic fragments in each strain were aligned against a reference genome of B . pseudomallei K96243 ( Fig . 1B; for a scaled map with the IS insertion sites in all B . mallei and B . pseudomallei strains , see Fig . S1; for the patterns of genomic rearrangements in the strains of each species , see Fig . S2; for the actual comparative blast data , see Tables S1 and S2 ) ., In contrast to these “core” elements , those elements that were not present in all ( singletons and those found in a few strains ) , collectively referred to as “accessory” elements , were much less common ., That the core elements , expected to be associated with the speciation of B . mallei from B . pseudomallei , accounted for most of the elements clearly reflects the common origin of B . mallei strains from a clone of B . pseudomallei 18 , 20 ., More importantly , it also suggests that further transpositions were significantly slowed after subsequent geographical segregation of the bacteria ., There are 13 core elements in B . mallei that have matching IS elements located at the same sites in B . pseudomallei strains ( Table 1 ) ., These elements were found to be composed of elements of ISBma1 and ISBma2 but not of IS407A ., This finding suggests that ISBma1 and ISBma2 have a longer history of association with B . pseudomallei than IS407A does ., Among the three largely expanded elements , we found that IS407A and ISBma2 were associated with almost all of the large genomic deletions and rearrangements in the B . mallei strains ( Fig . 2; Figs . S1 and S2; Tables S1 and S2 ) ., The only exception to this was a large deletion found in the strain ATCC 23344 and its direct derivatives , FMH , JHU , and GB8 horse 4 24 , between the 43rd and the 44th elements in chromosome 2 ( Fig . 2; Table S1 ) ., No genomic rearrangement was mediated by features other than the two IS elements ., ISBma1 , which was significantly increased in B . mallei , was not directly involved in any of the genomic deletions or rearrangements , however as many as 35% of it served as secondary entry points for IS407A ., The majority of the core elements of IS407A , 71 . 8% and 63 . 3% in chromosomes 1 and 2 , respectively , mediated rearrangements , deletions , or both ( Fig . 3A ) ., By contrast , accessory elements of IS407A contributed less , but were more active in chromosome 2 than in chromosome 1 ., By contrast , 50 . 4% and 53 . 2% of the core elements of ISBma2 in chromosomes 1 and 2 , respectively , contributed to rearrangements and/or deletions , and the accessory elements in both chromosomes were very rarely involved ( Figs . 2 and 3 ) ., We identified 59 and 28 genomic fragments in chromosomes 1 and 2 , respectively , which were encompassed by core elements of IS407A or ISBma2; these core elements mediated genomic rearrangements in at least one strain ( Figs . 2; Table S1 ) ., We referred to these genomic fragments as BRUs ( basic rearrangement units ) , a set of basic units for genomic reduction and rearrangement in B . mallei ., The BRUs formed various rearrangement patterns in the B . mallei strains ( Fig . S2A ) ., By contrast , B . pseudomallei strains had little variation in genome arrangement among one another due to low levels of IS elements- a few rearrangements were found but were around non-IS repeat sequences ( Fig . S2B ) ., When the pattern of the IS insertions and their involvement in genome-reductive and rearrangement processes in strains were used to construct a phylogenetic tree , strains sharing a recent common ancestry ( e . g . , ATCC 23344 and its immediate derivative isolates , FMH , JHU , and GB8 horse 4 ) or common recent geographical origins ( e . g . , strains NCTC 10257 , NCTC 10229 , and 2002721280 from European countries ) were grouped together ( Fig . 3B ) ., This phylogenic relationship supports the hypothesis that the accessory IS elements , which provided the major determinants for the tree rather than the common core elements , occurred following the speciation and geographical segregation of the B . mallei strains ., By contrast , such patterns were not obvious among the B . pseudomallei strains which did not go through IS element expansions; Australian strains 1655 and 668 did not branch separately from the South Asian strains ., The deletions and rearrangements that were mediated by accessory elements were most frequently noted in strains SAVP1 and 2002721280 , which lost virulence after successive passages in laboratory cultures 25 ( Figs . 2 and S1 ) ., Most of the extra deletions in these strains were more prominent in chromosome 2 than in chromosome 1 ., In SAVP1 , an IS407A-mediated deletion removed a major group of virulence genes encoding the animal-type type III secretion system in the BRU B22 ( Figs . 2 and S1 ) ; this deletion may be a major cause of the avirulence of that strain ., By contrast , there is no obvious deletion that may be responsible for the loss of virulence in strain 2002721280 ., That the strains SAVP1 and 2002721280 obtained deleterious mutations from in vitro culturing suggests that maintenance of the genomic contents in B . mallei requires selective pressure for survival in the host environment ., By contrast , the fully virulent strain PRL-20 showed more frequent deletions and rearrangements mediated by accessory elements than other virulent strains ., This strain may represent one of the more evolved ( more genome-reduced ) strains of B . mallei ., Although extra deletions and rearrangements were noted , the actual number of the accessory IS elements was not significantly increased in PRL-20 , SAVP1 or 2002721280 ., Furthermore , none of the direct derivatives of the strain ATCC 23344 ( i . e . FMH , JHU , and GB8 horse 4 ) had new IS insertions ( Fig . 2 and Table S1 ) ., These ATCC 23344 derivatives also did not have genomic rearrangements; the only change found was a single IS407A-mediated deletion located within the BRU B17 in the strain JHU ( Fig . 2 and Table S1 ) ., These lines of evidence suggest that B . mallei genomes are structurally flexible with regard to deletions , however perhaps not as much anymore for additional IS transpositions or genomic rearrangements ., IS407A elements are known to generate 4-bp target region duplications as direct repeats around them when they transpose 26 ., We found that ISBma1 generates 8-bp target region duplications , and that ISBma2 generates longer repeats of various lengths ( 18–26 bp ) ( Table 2; for the entire data , see Table S3 ) ., In addition to the various lengths of duplications , these target regions of the three types of IS had different nucleotide compositions and patterns ., Most notably , the sequences of ISBma1 contained homopolymers of A and/or T in up to 8-bp stretches of nucleotides ( Fig . 4A ) ., The target sequences of ISBma2 had a loose pattern in which the GC-rich central region was encompassed by strands of As and Ts on either side ., Target sequences of IS407A had the least characteristic composition ., It is intriguing to note that each IS element showed different levels of copy number expansion , ISBma1 with the lowest ( 3 . 3× ) , ISBma2 with an increased level ( 9 . 5× ) , and IS407A with the highest ( 16 . 7× ) ( Fig . 1A ) ., Perhaps this difference , at least in part , resulted from the availability of genomic sites suited for insertion targets ., There were concordant patterns of disruption of the core elements of one type by another , in that ISBma1 and ISBma2 were intersected by transposed IS407A ( Fig . 4B ) , while the reverse ( IS407A disrupted by ISBma1 or ISBma2 ) was not found ., A possible explanation for these insertion patterns may be that ISBma1 and ISBma2 could not transpose into IS407A due to the lack of sites suited for their rather uncommon target preferences , while IS407A did not have this problem ., Consistent with this hypothesis , ISBma1 and ISBma2 also did not have self-disrupted elements , while there were several self-disrupted IS407A elements ., The involvement of the three IS elements with different target sites increased the total number of IS insertions in the genome ., Furthermore , this increase led to further spread of IS407A , because ISBma1 and ISBma2 provided neutral insertion points for the element ., This in turn directly improved the efficiency of IS407A-mediated recombination in the genome , resulting in more sophisticated deletions and rearrangements ., We estimated that 83 . 7% of IS407A and 65 . 6% of ISBma2 elements in the B . mallei genomes lost their matching target duplicates , while all of the elements from intact ISBma1 elements were maintained ( Table S3 ) ., Almost all of the IS407A ( see Table S3 for details ) and all of the ISBma2 elements that contained matching repeats were not involved in genomic rearrangements in B . mallei ., This indicates that recombination among the elements were the major cause of the loss of the matching target duplicates ., B . mallei still has a high nucleotide-level identity ( 99% ) to B . pseudomallei ., Consistent with this , there was no AT-biased genome deviation in B . mallei , unlike that seen in many old symbionts or obligatory host-associated pathogens 1 , 3 ., Although the overall identity is still very high , significant nucleotide-level divergence exists , especially at the SSRs ( simple sequence repeats ) , where there are intrinsically high mutation rates 27 ., These SSRs were abundant in both B . mallei and B . pseudomallei at corresponding sites in the genomes ., However , there were more genes that were disrupted by frameshift mutations in B . mallei compared to B . pseudomallei ( Table S4 ) ., Most of these disrupted genes were commonly present in all B . mallei strains , reflecting the clonal origin of the strains ., Some of these gene disruptions may have contributed to better adaptation of the bacteria ( increased persistence ) in the host environment or simply became obsolete 28 ., One of the most characteristic loss of function or of surface structure in B . mallei is the loss of flagella 20 ., A gene essential for flagellum biogenesis , fliP , 29 in the strain ATCC23344 was disrupted by a 65-kb fragment flanked by IS407A elements , and this mutation completely turned B . mallei flagella-less ., This disruption in fliP is present in all B . mallei strains ( Table S1; Fig . 2 , between BRUs A2 and A3 ) , implying the significance of losing flagella in the evolution of host-restricted B . mallei ., The loss of flagella has been noted in other bacteria , including Bordetella pertussis and Bordetella parapertussis during their host specialization , derived from the strains of Bordetella bronchiseptica , 9 and Yersinia pestis during its conversion from a gut to a systemic pathogen 30 ., Additional disrupted genes not present in all strains were found at approximately the same levels as in B . pseudomallei , suggesting that there were no significant increases in mutation rates in B . mallei after geographical segregation ., There also was no significant level of erosion of these , so called , pseudogenes by purifying selection at levels high enough to contribute to the actual genome size reduction ( data not shown ) ., The extensiveness of the genome-wide reduction and rearrangements as well as additional nucleotide-level mutations may suggest that there is a potential for altered gene expression patterns in B . mallei ., A total of 341 potential regulatory genes survived the general IS-mediated genomic reduction in B . mallei ( not taking into account the diverse strain-specific deletions that occurred after speciation ) ., Among these genes , only a small fraction ( about 10 ) in each strain had deleterious ( e . g . frameshift , null , or IS-insertion ) mutations ( for the list of the genes , see Table S5 ) ., In addition , none of the predicted operons in B . mallei , which correspond to the putative operons previously found in B . pseudomallei K96243 31 , were disrupted by IS elements ( data not shown ) ., We also estimated the potential for changes in promoters ., There were 2 , 473 upstream sequences of genes , many of which may overlap or contain promoters , in the reference genome of B . pseudomallei K96243 that have homologous sequences ( with at least 95% identity over at least 95% of their lengths ) in all other strains of B . pseudomallei ., We found that up to 99% of these sequences also matched the corresponding regions in B . mallei ATTC23344 at the same homology levels ( see Table S6 for the list of the 2 , 473 upstream sequences , associated gene information , and the blast data ) ., Together , all these data from the analyses of the conserved genomic regions suggest that there is only a low potential for the genes in B . mallei to have significantly divergent gene expression patterns from B . pseudomallei ., By contrast , there were 56 genes with putative regulatory functions that were lost along with the commonly deleted genomic fragments of the B . mallei genome ., These genes include potential global regulatory genes , such as those encoding a quorum-sensing system ( genes BPSS1176 and BPSS1180 in the reference genome of B . pseudomallei K96243 ) , a two-component regulatory system ( the pair BPSS1994 and BpSS1995 in B . pseudomallei K96243 ) , and a number of regulators of various families ( Table S7 ) ., Whether the loss of any of these 56 regulatory genes affects the expression of the remaining genes in the B . mallei genome was yet to be examined ., To experimentally estimate the possible transcriptomic divergence between B . pseudomallei and B . mallei , we infected female BALB/c mice with B . mallei ATCC 23344 or B . pseudomallei K96243 , employing the previously established aerosol models of acute glanders and melioidosis 32 ., Gene expression was compared in the bacteria that colonized the lungs and the spleens of the mice ., Both B . mallei- and B . pseudomallei-challenged animals showed increases in the bacterial loads within these organs over time , with B . pseudomallei having slightly faster growth rates ( Fig . 5 ) ., In our experience , B . pseudomallei also grew faster than B . mallei in vitro ( data not shown ) ., Unlike the mice infected by B . mallei , sampling the B . pseudomallei-challenged animals after 72 hr was not possible due to animal mortality from the more rapid disease progression ., When gene expression profiles in the spleens and lungs were compared between B . mallei and B . pseudomallei at middle- ( i . e . , 24 hr for both bacteria ) and late stages ( i . e . , 48 hr for B . pseudomallei and 72 hr for B . mallei ) of infection ( a total of four comparison pairs ) , conserved B . mallei and B . pseudomallei orthologs showed nearly identical patterns with high Pearson correlation coefficient ( R ) values ranging from 0 . 94 to 0 . 97 , regardless of the host tissue type ( Fig . 5 ) ., Therefore , there was no indication of significant modifications of the expression schemes in the genes required by B . mallei to thrive in BALB/c mice compared to those in B . pseudomallei ., This is consistent with the findings of our previous gene expression studies in culture and in vivo , which also showed similar gene expression patterns in B . mallei and B . pseudomallei 20 , 33 , 34 , 35 ., These data suggest that , during the early stage , genomic reduction proceeds conservatively , not seriously affecting the indigenous gene expression patterns ., In contrast to B . mallei , most of the transcription units in the insect symbiont Buchnera were altered , most likely due to complex genomic alterations accumulated over a long period of time 2 ., In this study , we unveiled the mechanics of genomic deletions and rearrangements that occur in the early stage of bacterial specialization in the host , by conducting comparative analyses of B . mallei and its parental species , B . pseudomallei ., It became clear that stepwise IS intervention was the main driving force mediating a large genomic reduction in B . mallei ., Expansion of ISBma1 and ISBma2 in a clone of B . pseudomallei set the stage for the wide spread of IS407A , allowing its proliferation to sites , to which the element itself may rarely target ., Actual genomic deletions and rearrangements occurred through recombination reactions mainly among IS407A and also among ISBma2 ( Fig . 2 ) ., These processes achieved highly efficient deletions of dispensable genomic regions , causing only small disruptions to the portions of the genome that were maintained ., This was possible due to the guidance by selective forces in the host and via the intrinsic flexibility of the compactly IS-blended genome ., The B . mallei genome currently appears to still be structurally flexible with regard to deletions but is now less flexible with regard to genomic rearrangements and additional transpositions ., This may indicate that the genomic evolution in B . mallei has been moving into a second stage , in which large-scale genomic alterations are reduced and nucleotide-level erosion has become more important ., On the other hand , a large number of genes disrupted by frameshift mutations in SSRs were found in the B . mallei genome ., The loss of function encoded by these genes and of flagella via disruption in fliP by IS407A ( Table S1 ) , could be part of the adaptive evolution for survival in the host environment , which will eventually lead to genome size reduction by erosion over time ., Widespread relics of IS elements found in diverse symbionts and obligate pathogens 1 , 3 , 8 clearly suggest that a similar sequential IS intervention , modeled in Figure 6 , may illustrate a general mechanism , by which elaborate genome transition occurs during early bacterial evolution after establishing constant association with the host ., All research involving live animals was conducted in compliance with the Animal Welfare Act and other federal statutes and regulations relating to animals and experiments involving animals and adhered to the principles stated in the Guide for the Care and Use of Laboratory Animals , National Research Council , 1996 ., All mouse experiments conducted in the USAMRIID ( US Army Medical Research Institute of Infectious Diseases ) were approved by the Association for Assessment and Accreditation of Laboratory Animal Care International ., The type strains for B . mallei ( ATCC23344 ) 20 and B . pseudomallei ( K96243 ) 21 were previously sequenced ., Strains FMH , JHU , and GB8 horse 4 were direct derivatives of strain ATCC 23344 after passages in the human or horse , and these strains were also sequenced previously 24 ., B . mallei strains NCTC10229 , NCTC10247 , and SAVP1 were sequenced with full closure and manually annotated as previously described 20 ., The remaining three strains ( 2002721280 , ATCC10399 , and PRL-20 ) were sequenced to 8× Sanger sequence coverage by the whole genome shotgun method 36 without closure , and assembled using the Celera Assembler 37 , and contigs were oriented by alignment to the reference strain ATCC23344 using PROMER 38 ., ORFs were predicted and annotated automatically using GLIMMER 39 , 40 ., Pseudo-chromosomes were constructed from the ordered scaffolds , using manual examination where necessary ., Similarly , B . pseudomallei strains 1106a , 1710b , and 668 were sequenced with full closure and manual annotation , while 1655 , 406e , S13 , and Pasteur 6068 were sequenced without closure and annotated automatically ., For the analyses of genomic deletions and rearrangements in B . mallei and B . pseudomallei , 5 , 799 predicted protein sequences from the B . pseudomallei type strain K96243 were compared with the nucleotide sequences of the genomes of B . mallei ( ATCC 23344 , 2002721280 , ATCC 10399 , FMH , JHU , GB8 horse 4 , PRL-20 , NCTC 10229 , NCTC 10247 , SAVP1 ) and the other strains of B . pseudomallei ( 1106a , 1106b , 1655 , 1710a , 1710b , 406e , 668 , Pasteur , S13 ) using tblastn ( http://blast . wustl . edu ) ., For the mapping of the insertions of ISBma1 , ISBma2 , and IS407A in the genomes of B . mallei and B . pseudomallei , the entire sequences of the IS elements were searched against the 20 genomes using blastn ( http://blast . wustl . edu ) ., For the analysis of association of the IS elements with genomic deletions and rearrangements in B . mallei and of the target sequences in the genomes , strain ATCC 23344 represented all of its immediate derivatives , FMH , JHU , and GB8 horse 4 , to avoid redundancy in the data , because the three strains showed identical patterns ., To compare the patterns of genome rearrangements in the B . mallei strains , the positions of the BRUs in each strain of B . mallei relative to B . pseudomallei K96243 were visualized using a genome-comparative software tool ACT ( 41; http://www . sanger . ac . uk/Software/ACT ) , and the displays were compared in parallel among the strains ., We also examined B . mallei and B . pseudomallei for intergenic regions that potentially containing promoters , putative regulatory genes , and disruptions of putative operons to estimate the possibility of causing gene expression divergence ., For intergenic region comparisons , up to 100 bp upstream of the start codon , or up to as much as available if the neighboring gene was closer , of the genes that contain at least 50 bp of an untranslated upstream region were retrieved from the genome of B . pseudomallei K96243 ., Then , these sequences ( 2 , 268 and 1 , 566 from chromosomes 1 and 2 , respectively ) were searched against the genomes of B . mallei and B . pseudomallei using blastn ( http://blast . wustl . edu ) , and the length-match as well as the identity values of the orthologous regions were calculated ., Putative operons reported by Rodrigues et al . from the genome of B . pseudomallei K6243 31 were used to match the orthologous gene clusters in the genome of B . mallei ATCC 23344 , and these gene clusters were examined for any disruptions caused by IS elements ., All the genome sequences of B . mallei and B . pseudomallei used in this study are available through the Pathema web site ( http://pathema . jcvi . org/cgi-bin/Burkholderia/PathemaHomePage . cgi ) at the J . Craig Venter Institute ( http://www . jcvi . org/ ) ., A phylogenetic tree was constructed with the strains of B . mallei and B . pseudomallei based on the insertion patterns of and the role played in the genomic deletions and rearrangements by the three major IS elements , ISBma1 , ISBma2 , and IS407A ., All the data used are shown in Tables S1 and S2 and Figure 2 ., Bootstrapped maximum parsimony trees were calculated using the PAUP package with default parameters , and a consensus tree was produced from the bootstrap replicates ., Branches with bootstrap scores of less than 50 were collapsed in the tree ., Among the duplicated target regions encompassing the IS elements ISBma1 , ISBma2 , and IS407A , those regions that had perfectly matching sequences were first collected ., Then , among the sequences from unmatched pairs , those that occurred in more than two strains were assumed to be un-mutated valid sequences and , therefore , were added to the data pool for the analysis ., Strain ATCC 23344 represented all its direct derivatives ( FMH , JHU , and GB8 horse 4 ) in this analysis to avoid redundancy in the data ., The collected sequences were aligned with Clustal X , and the alignments were graphically visualized using Sequence logos 42 ., Exposure of mice to bacterial aerosol was performed as described by Roy et al . 43 ., Fresh overnight cultures of B . pseudomallei DD503 44 and B . mallei ATCC 23344 were prepared in LB or in LBG ( LB supplemented with 4% glycerol ) , respectively , at 37°C with aeration ( 250 rpm ) ., Thirty female BALB/c mice six to eight weeks old ( National Cancer Institute , Frederick , MD , USA ) were infected with these bacteria: nine mice each with B . pseudomallei and B . mallei for the gene expression studies , and six mice each for the bacterial load assays ., The mice infected with B . mallei received an inhaled dose of 7 . 2×103 cfu ( 7 . 2×LD50 ) , and those mice infected with B . pseudomallei received 1 . 8×104 cfu ( 18×LD50 ) , as estimated by colony counting on agar plates ., The infected mice were provided with rodent feed and water ad libitum and maintained on a 12-hr light cycle ., After 24 and 48 hr ( for both B . mallei and B . pseudomallei ) or 72 hr ( for B . mallei ) of infection , five mice from each point in time were euthanized in a CO2 chamber , and their spleens and lungs were removed ., Due to animal mortality , a 72 hr point in time was not possible for B . pseudomallei ., The organs from two randomly picked mice were saved for bacterial load estimations , and the rest were homogenized in 1 ml of Trizol ( Invitrogen Corp . , Carlsbad , CA , USA ) using a Tissue-Tearor ( BioSpec Products , Bartlesville , OK , USA ) ., Total RNA was purified according to the manufacturers recommendations ( Invitrogen Corp . , Carlsbad , CA , USA ) ., The bacterial load in the mouse organs was estimated as described by Ulrich and DeShazer 32 ., Total RNA , both bacterial and mouse , from the same organ types from three mice was pooled to compensate for potential individual variation ., These pooled RNA samples were used for the experiments without further purification of the bacterial RNA because RNA from mice does not cross-hybridize to the B . mallei microarray at a level affecting the legitimate interactions between the B . mallei array and the Burkholderia transcriptome 35 ., The B . mallei whole genome array used in this study for both B . mallei and the closely related B . pseudomallei ( average gene identity at the nucleotide level of 99% ) was described in detail previously 33 ., The B . mallei- and B . pseudomallei-infected organ samples were paired for the hybridization reactions based on early and late pathological states ., A total of eight hybridization reactions or four different comparisons were performed , each of which was replicated in flip-dye pairs and the final ratios were calculated as log2 ( B . pseudomallei gene expression intensity/B . mallei gene expression intensity ) ., Labeling of the probes , slide hybridization , and slide scanning were carried out as previously described 35 ., The independent TIFF slide images from each channel were analyzed using TIGR Spotfinder to assess the relative expression levels , and the data were normalized using a local regression technique LOWESS ( LOcally WEighted Scatterplot Smoothing ) with the MIDAS software ( <http://www . jcvi . org/cms/research/software> , The J . Craig Venter Institute , Rockville , MD , USA ) ., The resulting data were averaged from triplicate genes on each microarray and from duplicate flip-dye arrays for each experiment . | Introduction, Results/Discussion, Materials and Methods | The equine-associated obligate pathogen Burkholderia mallei was developed by reductive evolution involving a substantial portion of the genome from Burkholderia pseudomallei , a free-living opportunistic pathogen ., With its short history of divergence ( ∼3 . 5 myr ) , B . mallei provides an excellent resource to study the early steps in bacterial genome reductive evolution in the host ., By examining 20 genomes of B . mallei and B . pseudomallei , we found that stepwise massive expansion of IS ( insertion sequence ) elements ISBma1 , ISBma2 , and IS407A occurred during the evolution of B . mallei ., Each element proliferated through the sites where its target selection preference was met ., Then , ISBma1 and ISBma2 contributed to the further spread of IS407A by providing secondary insertion sites ., This spread increased genomic deletions and rearrangements , which were predominantly mediated by IS407A ., There were also nucleotide-level disruptions in a large number of genes ., However , no significant signs of erosion were yet noted in these genes ., Intriguingly , all these genomic modifications did not seriously alter the gene expression patterns inherited from B . pseudomallei ., This efficient and elaborate genomic transition was enabled largely through the formation of the highly flexible IS-blended genome and the guidance by selective forces in the host ., The detailed IS intervention , unveiled for the first time in this study , may represent the key component of a general mechanism for early bacterial evolution in the host . | It has been known for some time that bacteria undergo genome-reduction when they transition from a free-living state to a constantly host-restricted state ., High levels of IS element expansion were also found in these bacteria , and the IS elements were suggested to play a role in genome reductive evolution ., Here we provide evidence for stepwise IS actions as the exclusive mechanism that mediates bacterial genomic changes during the early stage of constant host-bacterial association , by unveiling the processes that resulted in the development of B . mallei genome ., We show the details of the multi-level interplay of IS elements , which facilitate the wide spread of the IS copies , and the overall mechanics in genome reduction and rearrangement ., These processes appeared to operate as chain reactions mediating elaborate genomic transition , without seriously affecting the original gene expression patterns ., The absence of differential gene expression in the resulting genome suggests that changes in transcriptional regulation that are often observed in other old bacterial genomes may take place subsequent to the IS-mediated steps , along with gradual nucleotide-level changes . | evolutionary biology/microbial evolution and genomics, evolutionary biology/genomics, evolutionary biology/bioinformatics, evolutionary biology/evolutionary and comparative genetics | null |
journal.pgen.1006933 | 2,017 | Systematic tissue-specific functional annotation of the human genome highlights immune-related DNA elements for late-onset Alzheimer’s disease | Large consortia such as ENCODE 1 and Epigenomics Roadmap Project 2 have generated a rich collection of high-throughput genomic and epigenomic data , providing unprecedented opportunities to delineate functional structures in the human genome ., As complex disease research rapidly advances , evidence has emerged that disease-associated variants are enriched in regulatory DNA elements 3 , 4 ., Therefore , functional annotation of the non-coding genome is critical for understanding the genetic basis of human complex diseases ., Unfortunately , categorizing the complex regulatory machinery of the genome requires integration of diverse types of annotation data as no single annotation captures all types of functional elements 5 ., Recently , we have developed GenoSkyline 6 , a principled framework to identify tissue-specific functional regions in the human genome through integrative analysis of various chromatin modifications ., In this work , we introduce GenoSkyline-Plus , a comprehensive update of GenoSkyline that incorporates RNA sequencing and DNA methylation data into the framework and extends to 127 integrated annotation tracks covering a spectrum of human tissue and cell types ., To demonstrate the ability of GenoSkyline-Plus to systematically provide novel insights into complex disease etiology , we jointly analyzed summary statistics from 45 genome-wide association studies ( GWAS; Ntotal≈3 . 8M ) and identified biologically relevant tissues for a broad spectrum of complex traits ., We next performed an in-depth , annotation-driven investigation of Alzheimer’s disease ( AD ) , a neurodegenerative disease characterized by deposition of amyloid-β ( Aβ ) plaques and neurofibrillary tangles in the brain ., Late-onset AD ( LOAD ) includes patients with onset after 65 years of age and has a complex mode of inheritance 7 ., Around 20 risk-associated genetic loci have been identified in LOAD GWAS 8 ., However , our understanding of LOAD’s genetic architecture and disease etiology is still far from complete ., Through integrative analysis of GWAS summary data and GenoSkyline-Plus annotations , we identified strong enrichment for LOAD associations in immune cell-related DNA elements , consistent with other data suggesting a crucial role for the immune system in AD etiology 9–11 ., Jointly analyzing GWAS summary data for LOAD and Parkinson’s disease ( PD ) , we identified substantial enrichment for pleiotropic associations in the monocyte functional genome ., Our findings provide support for the critical involvement of the immune system in the etiology of neurodegenerative diseases , and suggest a previously unsuspected role for an immune-mediated pleiotropic effect between LOAD and PD ., We use our previously established statistical framework to calculate the posterior probability of functionality for each nucleotide in the human genome 12 ., Integrating tissue and cell-specific genomic functional data available through Epigenomics Roadmap Project 2 , we make available GenoSkyline-Plus scores for 127 individual tissue annotation tracks ( Methods; S1 Table ) ., H3K4me3 and H3K9ac , known markers of open chromatin and active transcription 13 , are shown to have the largest odds ratios of predicting functionality across the genome ( Fig 1A ) ., Identifying H3K4me3 and H3K9ac as strong indicators of genomic functionality is a finding consistent with previous studies of gene regulation through chromatin marks 14 ., In contrast , H3K9me3 , a well established repressive mark 13 , has a reversed effect on genome functionality ., The bimodal pattern of GenoSkyline scores 6 allows us to impose a score cutoff to robustly define the functional genome ., Using a cutoff of 0 . 5 , 3% of the genome is considered functional on average across all annotation tracks ( Fig 1B ) ., This functionality percentage varies from 1% in pancreatic islet cells to 8% in PMA-I stimulated T-helper cells ., Our findings on functionality across all tracks are consistent with previous findings 12; 34% of the intergenic human genome is predicted to be functional in at least one annotation track ( Fig 1C ) ., Additionally , coding regions of the genome are predicted to have much greater proportions of functionality in multiple tissues than intronic and intergenic regions ., To assess the ability of GenoSkyline-Plus to capture tissue and cell-specific , non-coding functionality in the human genome , we consider a diverse set of known non-coding regulatory elements studied across the genome ., To start , we examined microRNAs ( miRNA ) , which are known to regulate a variety of cellular processes through the translational repression and degradation signaling of transcripts 15 ., Recent work by Ludwig et al . profiled miRNA expression in 61 different human tissues and identified miRNAs with functionality unique to single tissues through a tissue specific index 16 , 17 ( TSI; Methods ) ., We applied GenoSkyline-Plus scores to miRNA with tissue-specific functionality by calculating the total proportion of nucleotides predicted to be functional in each tissue ., We next looked for which annotation tracks are able to predict the highest proportion of functionality for these known functional regions ., The best predictors of high functionality for the three tissues with the largest sample sizes ( i . e . brain , liver , and muscle ) are tracks for brain structures , the liver track , and the muscle track , respectively ( Fig 2A ) ., We next examined long non-coding RNAs ( lncRNA ) , another non-coding element known for its tissue-specific regulatory action 18 ., Using a custom-designed microarray targeting GENCODE lncRNA , Derrien et al . profiled the activity of 9 , 747 lncRNA transcripts 19 ., In order to reidentify and validate the set of lncRNA transcripts that are specific to their respective tissues , we calculated the previously described TSI and selected lncRNAs with expression specific to only a few cell types ., Physiologically matching tracks show a higher proportion of predicted functionality than unmatched tracks in complex , heterogeneous tissue structures like the midbrain ., More functionally uniform tissues , such as the thymus or placenta , show the highest functional proportion in matching annotation tracks ( Fig 2B ) ., We also assessed enhancers , non-coding elements that can remotely regulate transcription of an associated promoter elsewhere on the genome with important roles in cell-type specificity 20 ., We extracted tissue and cell type-specific enhancer facets identified through the FANTOM5 cap analysis of gene expression ( CAGE ) atlas and positive differential expression when compared against other defined facets 21 ., To determine the utility of the large library of immune cells available in the Epigenomics Roadmap Project for which we developed annotation tracks , we focused on enhancer facets with differential CAGE expression in immune cells ., While the method by which enhancers are defined to be differential in a facet is liberal ( Methods ) and does not imply facet-specific expression , GenoSkyline-Plus still showed outstanding ability to identify matching cell types ., Indeed , matched annotation tracks for T-cells , natural killer cells , and monocytes show consistently higher functional proportions than other , non-matched immune cell annotation tracks ( Fig 2C ) ., Finally , we present a case study of the IL17A-IL17F locus control region ( LCR ) in humans , a ~200kb regulatory region surrounding the IL17A gene locus ., IL17A encodes the primary secreted cytokine effector molecule IL-17 of T helper 17 ( Th17 ) cells 22 ., The LCR has been studied in mouse models and is found to contain many potential human-conserved intergenic regulatory elements that bind transcription factors that are essential for Th17 cell differentiation and effector function 23 , 24 ., Experimentally , these conserved noncoding sequences ( CNS ) acquire functionally permissive H3 acetylation marks at much greater magnitudes under Th17-inducing conditions than naïve or combined Th1 and Th2 populations 25 ., Comparing annotation tracks for naïve CD4+ T-cells , differentiated Th17 cells , and differentiated Th1/Th2 cell populations , we identified highly Th17-specific functionality in the conserved regions of the human genome corresponding to known murine CNS regions ( Fig 2D and 2E ) ., CNS sites and their flanking regions showed substantially higher functional proportion in Th17 cells than in naïve CD4+ T-cells or Th1/Th2 cell subsets ., We jointly analyzed three tiers of annotation tracks that respectively represent the overall functional genome , 7 broad tissue clusters , and 66 tissue and cell types ( Methods; S2 Table ) , with summary statistics from 45 GWAS covering a variety of human complex traits ( S3 Table ) ., We applied LD score regression 26 to stratify trait heritability by tissue and cell type , and identified a total of 226 significantly enriched annotation tracks for 34 traits after correcting for multiple testing ( S4–S7 Tables ) ., In general , GWAS with a large number of significant SNP-level associations showed stronger heritability enrichment in the predicted functional genome ( Fig 3A and 3B ) ., Tissue and cell tracks refined the resolution of heritability stratification and provided additional insights into the genetic basis of complex traits ( Fig 3C and 3D ) ., The immune annotation track was significantly enriched for 7 immune diseases , namely celiac disease ( CEL ) , Crohn’s disease ( CD ) , ulcerative colitis ( UC ) , primary biliary cirrhosis ( PBC ) , rheumatoid arthritis ( RA ) , systemic lupus erythematosus ( SLE ) , and multiple sclerosis ( MS ) ., Using tracks for cell types , we identified several significant enrichments , including monocytes for CD ( p = 2 . 9e-11 ) and B cells for PBC ( p = 2 . 3e-6 ) , RA ( p = 1 . 2e-5 ) , and MS ( p = 2 . 2e-6 ) ., Inflammatory bowel diseases showed significant enrichment in the gastrointestinal ( GI ) annotation track ( CD: p = 1 . 4e-4; UC: p = 5 . 6e-5 ) ., Another autoimmune disease with a well-established GI component , CEL , also showed nominal enrichment in the GI annotation track ( p = 3 . 7e-4 ) ., Several brain annotation tracks were significantly enriched for associations of schizophrenia ( SCZ ) , education years ( EDU ) , and cognitive performance ( IQ ) ., Bipolar disorder ( BIP ) , neuroticism ( NEU ) , and chronotype ( CHT ) all showed nominally significant enrichment in the anterior caudate annotation track ., Body mass index ( BMI ) and age at menarche ( AAM ) were significantly enriched in multiple brain annotation tracks ., Compared to other brain regions , the substantia nigra annotation track showed weaker enrichment for these brain-based traits , which is consistent with its primary function of controlling movement ., Hundreds of height-associated loci have been identified in GWAS 27 ., Such a highly polygenic genetic architecture is also reflected in our analysis ., 59 of 66 tier-3 tissue and cell annotation tracks were significantly enriched for height associations , with breast myoepithelial cell ( p = 6 . 2e-14 ) and osteoblast ( p = 8 . 5e-14 ) being the most significant ., Waist-hip ratio ( WHR ) , birth weight ( BW ) , and three blood pressure traits showed significant enrichment in the adipose annotation track ., Overall , cardiovascular ( CV ) annotation tracks showed strong enrichment for blood pressure and coronary artery disease ( CAD ) ., Interestingly , the aorta annotation track is significantly enriched for pulse pressure ( PP ) but not systolic or diastolic blood pressure ( SBP and DBP ) ., CAD and 4 lipid traits , i . e . high and low density lipoprotein ( HDL and LDL ) , total cholesterol ( TC ) , and triglycerides ( TG ) , shared a similar enrichment pattern in liver , adipose , and monocyte annotation tracks , which is consistent with the causal relationship among these traits 28 ., Our results demonstrated that annotations with refined specificity could provide insights into disease etiology while broader annotations have greater statistical power ., Age-related macular degeneration ( AMD ) was significantly enriched in broadly defined annotation tracks including immune , brain , CV , and GI , despite the non-significant enrichment results using tier-3 annotation tracks ., Analyses based on all three tiers of annotations could systematically provide the most interpretable results for most traits ., Importantly , we note that greater GWAS sample sizes will effectively increase statistical power in the enrichment analysis while leaving the overall enrichment pattern stable ( S1 Fig ) ., Therefore , many more suggestive enrichment results are likely to become significant as GWAS sample sizes grow ., Finally , some traits , e . g . type-II diabetes ( T2D ) and age at natural menopause ( AANM ) , showed strong enrichment in the general functional genome but not in specific tissues , suggesting that we may be able to gain a better understanding of these traits when annotation data for tissues or cell types more relevant to these traits are made available ., Next , we performed an integrative analysis of stage-I GWAS summary statistics from the International Genomics of Alzheimer’s Project 8 ( IGAP; n = 54 , 162 ) with GenoSkyline-Plus annotations ( Methods ) ., SNPs located in the broadly defined immune annotation track , which account for 24 . 4% of the variants in the IGAP data , could explain 98 . 7% of the LOAD heritability estimated using LD score regression ( enrichment = 4 . 0; p = 1 . 5e-4 ) ., Somewhat surprisingly , the signal enrichment in DNA elements functional in immune cells was substantially stronger than the enrichment in brain and other tissue types ( Fig 4A ) ., To investigate if immune-related DNA elements are also enriched for associations of other neurodegenerative diseases , we analyzed a publicly accessible GWAS summary dataset for PD 29 ( n = 5 , 691; Methods ) ., Again , the immune annotation track was the most significantly enriched annotation ( enrichment = 6 . 3; p = 7 . 5e-6 ) , followed by epithelium and CV ( Fig 4A ) ., Analysis based on 66 tissue and cell tracks further refined the resolution of our enrichment study ., Monocyte ( enrichment = 10 . 9; p = 2 . 0e-5 ) and liver ( enrichment = 16 . 6; p = 4 . 1e-4 ) annotation tracks were significantly enriched for LOAD associations ( Fig 4B ) ., In fact , the combined functional regions in monocyte and liver covered 8 . 8% of the SNPs in the IGAP data , but could account for 99 . 6% of the LOAD heritability currently captured in the IGAP stage-I GWAS ( Fig 4C ) ., In PD GWAS , signal enrichment in liver was absent , but monocyte-functional regions remained strongly enriched ( enrichment = 16 . 3; p = 8 . 5e-7 ) ., Our findings support the critical role of innate immunity in neurodegenerative diseases 10 ., Significant enrichment for LOAD associations in liver-specific DNA elements also provides additional support for the possible involvement of cholesterol metabolism in LOAD etiology 30 , 31 ., LOAD signal enrichment in liver remained significant after removing the APOE region ( chr19: 45 , 147 , 340–45 , 594 , 595; hg19 ) from the analysis ( S2 Fig ) , suggesting a polygenic architecture in this pathway ., Finally , some adaptive immune cells also showed enrichment for AD and PD associations ., LOAD signal enrichment in the B cell annotation track was nominally significant , while multiple T cell annotation tracks were significantly enriched for PD associations ., These results not only suggest the involvement of adaptive immunity in neurodegenerative diseases , but also hint at distinct mechanisms of such involvement between AD and PD ., Finally , for comparison , we applied several other annotations including CADD 32 , GWAVA 33 , and EIGEN 34 to the LOAD GWAS data ., GenoCanyon and GenoSkyline annotations for seven tissues were also included in the comparison ., Our annotations outperformed these methods , showing stronger fold enrichment and more significant p-values ( S8 Table ) ., Our results showed strong enrichment for both AD and PD in the monocyte functional genome ., Next , we investigate if the enrichment for both diseases is through shared or distinct genetic components ., Recent studies have failed to identify statistically significant genome-wide pleiotropic effects between AD and PD 35 ., We instead hypothesize that the same set of immune-related genetic components are involved in both diseases ., Therefore , we aim to identify enrichment for pleiotropic effects in the genome localized to regions of monocyte functionality ., We first partitioned AD and PD heritability by chromosome ., Chromosome-wide heritability showed moderate correlation between the two diseases ( correlation = 0 . 65; Fig 5A ) ., When focusing on monocyte functional elements , chromosome-wide heritability showed high concordance between AD and PD ( correlation = 0 . 96; Fig 5B ) ., Interestingly , such high concordance cannot be fully explained by chromosome size ., In fact , the correlation between chromosome size and per-chromosome heritability estimates is 0 . 56 for AD and 0 . 59 for PD , both lower than the correlation between AD and PD’s per-chromosome heritability estimates , especially in the monocyte functional genome ., The percentage of explained LOAD heritability on chromosome 19 is lower than previous estimation 36 due to removal of SNPs with large effects in the APOE region ( Methods ) ., Next , to quantify the shared genetics between AD and PD , we identified significant enrichment for pleiotropic effects in monocyte functional regions ( enrichment = 1 . 8; p = 9 . 4e-4 ) using a window-based approach ( Methods ) ., To account for potential bias due to the moderate sample overlap between the two GWAS as well as other confounding factors , we applied a permutation-based testing approach ( Methods ) ., Enrichment for pleiotropic effects in the monocyte functional genome remained significant ( p = 4 . 6e-3 ) ., In addition , these results were robust with the choice of window size ., We identified 15 candidate loci for pleiotropic effects ( Methods; S9 Table ) , among which signals at SLC9A9 and AIM1 are the clearest ( Fig 5C and 5D ) ., SLC9A9 , whose encoded protein localizes to the late recycling endosomes and plays an important role in maintaining cation homeostasis ( RefSeq , Mar 2012 ) , is associated with multiple pharmacogenomic traits related to neurological diseases , including response to cholinesterase inhibitor in AD 37 , response to interferon beta in MS 38 , response to angiotensin II receptor blockade therapy 39 , and multiple complex diseases including attention-deficit/hyperactivity disorder 40 , autism 41 , and non-alcoholic fatty liver 42 ., Gene AIM1 is associated with stroke 43 , human longevity 44 , and immune diseases including RA 45 and SLE 46 ., A few candidate loci pointed to clear gene candidates but showed unclear or distinct peaks of association ( S3 Fig ) ., These include an inflammatory bowel disease risk gene ANKRD33B 47 ., PRUNE2 is a gene associated with response to amphetamine 48 and hippocampal atrophy which is a quantitative trait for AD 49 ., HBEGF is associated with AD in APOE ε4- population 50 and involved in Aβ clearance 51 ., PROK2 is a gene involved in Aβ-induced neurotoxicity 52 ., Additionally , the protein product of AXIN1 negatively affects phosphorylation of tau protein 53 ., Other gene candidates include CCDC158 , PRSS16 , and ZNF615 , which are previously identified risk genes for PD , SCZ , and BIP , respectively 54–56 ., Some other windows showed complex structures of linkage disequilibrium ( LD ) and contained large association peaks spanning a number of genes ( S4 Fig ) , which include the region near PD risk gene PRSS8 54 and the HLA region ., Interestingly , we also identified the surrounding region of MAPT , a gene that encodes the tau protein which is a critical component of both AD and PD pathologies 50 , 54 , 57 , 58 ., Pathway enrichment analysis for genes in 15 pleiotropic candidate loci identified significant enrichment in immune-related pathways staphylococcus aureus infection ( KEGG:05150; p = 1 . 9e-5 ) and systemic lupus erythematosus ( KEGG:05322; p = 3 . 7e-04; Methods ) ., Both pathways remained significant after removing two HLA loci from our analysis ., Finally , we reprioritize AD risk loci using monocyte and liver annotation tracks ., We integrated IGAP stage-I summary statistics with GenoSkyline-Plus using genome-wide association prioritizer ( GenoWAP 59 ) , and ranked all SNPs based on their GenoWAP posterior scores ( Methods ) ., Under a posterior cutoff of 0 . 95 , we identified 8 loci that were not reported in the IGAP GWAS meta-analysis using monocyte annotation and 4 loci using the liver annotation track ( S10 Table ) ., We then sought replication for SNPs with the highest posterior score at each of these loci using inferred IGAP stage-II z-scores ( Methods ) ., After removing shared SNPs between monocyte- and liver-based analyses , 10 SNPs remained in the analysis , 7 of which showed consistent effect directions between the discovery and the replication cohorts ( Fig 6A ) ., One SNP was successfully replicated in the inferred IGAP stage-II dataset , i . e . rs4456560 ( p = 0 . 013 ) ., SNP rs4456560 is located in SCIMP ( Fig 6B ) , a gene that encodes a lipid tetraspanin-associated transmembrane adaptor protein that is expressed in antigen-presenting cells and localized in the immunological synapse 60 ., A moderate replication rate in the IGAP stage-II cohort was expected since we focused on loci that did not reach genome-wide significance in the IGAP meta-analysis and the IGAP stage-II cohort is relatively small ( n = 19 , 884 ) compared to the data in the discovery stage ., Furthermore , data from IGAP stage-II cohort are not publicly available and we were limited to the inverse inference approach shown here ., It is possible additional loci will replicate when IGAP stage-II summary or individual-level data are made available ., However , all identified loci have been linked to AD or relevant phenotypes in the literature ., RPN1 was linked to AD through a network-based technique 61 ., Association between ECHDC3 and AD risk was established through a joint analysis of AD and lipid traits 62 ., Association between DLST and AD has also been previously reported 63 ., BZRAP1 and MINK1 were shown to be associated with cognitive function and blood metabolites , respectively 64 , 65 ., A pleiotropic effect candidate gene HBEGF showed up again in the SNP reprioritization analysis ., Multiple genes in the sorting nexin family have been found to participate in APP metabolism and Aβ generation 66 ., Association between SNX1 and AD has also been previously identified using gene-based tests 67 ., Finally , during the peer review process of this paper , three new genome-wide significant loci ( i . e . PFDN1/HBEGF , USP6NL/ECHDC3 , and BZRAP1-AS1 ) were reported in a trans-ethnic GWAS meta-analysis for AD 68 , all of which were among our reprioritized list of risk loci ., Further , the most significant SNPs at loci PFDN1/HBEGF ( rs11168036 , p = 7 . 1e-9 ) and BZRAP1-AS1 ( rs2632516 , p = 4 . 4e-8 ) matched with our top reprioritized SNPs ( Fig 6A ) ., Increasing evidence suggests that non-coding regulatory DNA elements may be the primary regions harboring risk variants in human complex diseases ., In this work , we have substantially expanded our previously established GenoSkyline annotation by incorporating RNA-seq and DNA methylation into its framework , imputing incomplete epigenomic and transcriptomic annotation tracks , and extending it to more than 100 human tissue and cell types ., With the help of integrative functional annotations , we identified strong enrichment for LOAD heritability in functional DNA elements related to innate immunity and liver tissue using hypothesis-free tissue-specific enrichment analysis ., This enrichment was also found in immune-related DNA elements using PD data ., Our analysis also clearly indicated that monocyte functional elements in particular appear to be highly relevant in explaining AD and PD heritability ., Of note , we analyzed 45 complex diseases and traits in addition to AD and PD ., The substantial enrichment for multiple psychiatric and neurological traits in the brain functional genome shows that the lack of brain enrichment in neurodegeneration is not due to poor quality of brain annotations ., Further , the monocytes annotation track was the most significantly enriched for Crohn’s disease among the 45 GWAS , and was not ubiquitously enriched for a large number of traits ., Consistent and biologically interpretable enrichment results on a large collection of complex traits demonstrate the effectiveness of our approach and increase the validity of novel findings ., It is worth noting that multiple studies have highlighted the role of myeloid cells in the genetic susceptibility of neurodegenerative diseases 11 ., Several genes expressed in myeloid cells ( e . g . ABCA7 , CD33 , and TREM2 ) have been identified in GWAS and sequencing-based association studies for AD 8 , 69 , 70 ., Further , AD risk alleles identified in GWASs have been shown to enrich for cis-eQTLs in monocytes 9 ., In addition , two recent papers identified enrichment for AD heritability in active genome regions in myeloid cells 71 , 72 , which suggested a polygenic genetic architecture for immune-related DNA elements in AD etiology and hinted at a large number of unidentified , immune-related genes for AD ., Compared to the aforementioned work , our study utilizes a better set of tissue-specific genome annotations and explicitly accounts for the similarity between different cell types through a multiple regression model ., One major limitation in our analysis is lack of data for other potentially AD-relevant cell types such as microglia ., Whether our findings correctly reflected the direct involvement of peripheral immune cells in neurodegenerative diseases rather than the detection of epigenomic similarities between monocytes and microglia remains to be carefully investigated in the future ., Furthermore , we successfully identified enrichment for shared genetic components between AD and PD in the monocyte functional genome , which hints at a shared neuroinflammation pathway between these two neurodegenerative diseases ., We note that several candidate loci with potential pleiotropic effects showed fairly marginal associations with AD and PD , which explains why they have been missed in traditional SNP-based association analysis ., Importantly , SNPs in immune-related DNA elements explain a large proportion of AD and PD heritability in total ., These results suggest that weak but pervasive associations related with immunity still remain unidentified ., Further evaluations of these relationships using GWAS with larger sample sizes may provide insights into the shared biology of these neurodegenerative conditions ., Through multi-tier enrichment analyses on 45 GWAS , an in-depth case study of neurodegenerative diseases , and validation of known non-coding tissue-specific regulatory machinery , we have demonstrated the ability of GenoSkyline-Plus to provide unbiased , genome-wide insights into the genetic basis of human complex diseases ., The analyzed GWAS represent a variety of human complex diseases and traits , highlighting the effectiveness of our method in different contexts and genetic architecture ., However , while our non-coding validation study demonstrated that GenoSkyline-Plus annotations indeed captured tissue-specific activity in a variety of intergenic machinery , there is a need to develop a more statistically robust framework to identify new non-coding elements rather than validate existing ones ., Our approach of identifying the functionally active proportion of all elements in aggregate is only able to identify tissue specificity while considering large groups of highly specific non-coding elements ., The availability of over 100 different annotation tracks introduces many multiple-testing issues that should be addressed in the case of a statistically sound analysis for tissue-specificity ., We have also demonstrated how GenoSkyline-Plus and its explanatory power improve with the addition of more data ., Currently , functionality in 28% of exonic regions still remains to be identified ., As the quantity and quality of high-throughput epigenomic data continue to grow , GenoSkyline-Plus has the potential to further evolve and provide even more comprehensive annotations of tissue-specific functionality in the human genome ., We will update our annotations when data for new tissue and cell types from the Roadmap consortium become available ., Finally , several recent papers have introduced novel models to integrate functional annotations in tissue-specific enrichment analysis 73 , 74 ., Many models that do not explicitly incorporate functional annotation information have also emerged in transcriptome-wide association studies and other closely-related applications in human genetics research 75–78 ., Our annotations , in conjunction with rapidly advancing statistical techniques and steadily increasing sample sizes in genetics studies , may potentially benefit a variety of human genetics applications and promise a bright future for complex disease genetics research ., Chromatin data were extracted from the Epigenomics Roadmap Project’s consolidated reference epigenomes database ( http://egg2 . wustl . edu/roadmap/ ) ., Specifically , ChIP-seq peak calls were collected for each epigenetic mark ( H3k4me1 , H3k4me3 , H3k36me3 , H3k27me3 , H3k9me3 , H3k27ac , H3k9ac , and DNase I Hypersensitivity ) in each Roadmap consolidated epigenome where available ., Peak calls imputed using ChromImpute 79 were used in place of missing data ., Next , peak files were reduced to a per-nucleotide binary encoding of presence or absence of contiguous regions of strong ChIP-seq signal enrichment compared to input ( Poisson p-value threshold of 0 . 01 ) ., DNA methylation data were also collected from the Roadmap’s reference epigenomes database ., CpG islands were identified in each sample using the CpG Islands Track of the UCSC Genome Browser ( http://genome . ucsc . edu/ ) , and unmethylated islands were those CpG islands with less than 0 . 5 fractionated methylation based on imputed methylation signal tracks in the Roadmap reference epigenomes database ., Presence of an unmethylated CpG island was then encoded for each nucleotide as a binary variable ., Finally , Roadmap’s RNA-seq data were dichotomized using an rpkm cutoff of 0 . 5 at 25-bp resolution and included in our annotations ., We adapt the existing framework established by Lu et al . to a broader set of genomic data 12 ., Briefly , given a set of Annotations A and a binary indicator of genomic functionality Z , the joint distribution of A along the genome is assumed to be a mixture of annotations at functional nucleotides and non-functional nucleotides ., Assuming that each of the annotations in A is conditionally independent given Z , we factorize the conditional joint density of A given Z as:, f ( A|Z=c ) =∏i=110fi ( Ai|Z=c ) , c=0 , 1, ( 1 ), All annotations have been preprocessed into binary classifiers , and the marginal functional likelihood given each individual annotation can be modeled with a Bernoulli distribution, fi ( Ai|Z=c ) =picAi ( 1−pic ) 1−Ai , i=1 , … , 10;\xa0c=0 , 1, ( 2 ), With an assumed prior probability π of functionality , the parameter pic of each individual annotation can be estimated with the Expectation-Maximization ( EM ) algorithm ., The posterior probability of functionality at a nucleotide , known as the GenoSkyline-Plus score , is then:, P ( Z=1|A ) =π∏i=110fi ( Ai|Z=1 ) π∏i=110fi ( Ai|Z=1 ) + ( 1−π ) ∏i=110fi ( Ai|Z=0 ), ( 3 ), Giving us with 21 parameters for each annotation track:, Θ= ( π , p1 , 0 , p2 , 0 , … , p10 , 0 , p1 , 1 , p2 , 1 , … , p10 , 1 ), ( 4 ), These parameters | Introduction, Results, Discussion, Methods | Continuing efforts from large international consortia have made genome-wide epigenomic and transcriptomic annotation data publicly available for a variety of cell and tissue types ., However , synthesis of these datasets into effective summary metrics to characterize the functional non-coding genome remains a challenge ., Here , we present GenoSkyline-Plus , an extension of our previous work through integration of an expanded set of epigenomic and transcriptomic annotations to produce high-resolution , single tissue annotations ., After validating our annotations with a catalog of tissue-specific non-coding elements previously identified in the literature , we apply our method using data from 127 different cell and tissue types to present an atlas of heritability enrichment across 45 different GWAS traits ., We show that broader organ system categories ( e . g . immune system ) increase statistical power in identifying biologically relevant tissue types for complex diseases while annotations of individual cell types ( e . g . monocytes or B-cells ) provide deeper insights into disease etiology ., Additionally , we use our GenoSkyline-Plus annotations in an in-depth case study of late-onset Alzheimer’s disease ( LOAD ) ., Our analyses suggest a strong connection between LOAD heritability and genetic variants contained in regions of the genome functional in monocytes ., Furthermore , we show that LOAD shares a similar localization of SNPs to monocyte-functional regions with Parkinson’s disease ., Overall , we demonstrate that integrated genome annotations at the single tissue level provide a valuable tool for understanding the etiology of complex human diseases ., Our GenoSkyline-Plus annotations are freely available at http://genocanyon . med . yale . edu/GenoSkyline . | After years of community efforts , many experimental and computational approaches have been developed and applied for functional annotation of the human genome , yet proper annotation still remains challenging , especially in non-coding regions ., As complex disease research rapidly advances , increasing evidence suggests that non-coding regulatory DNA elements may be the primary regions harboring risk variants in human complex diseases ., In this paper , we introduce GenoSkyline-Plus , a principled annotation framework to identify tissue and cell type-specific functional regions in the human genome through integration of diverse high-throughput epigenomic and transcriptomic data ., Through validation of known non-coding tissue-specific regulatory regions , enrichment analyses on 45 complex traits , and an in-depth case study of neurodegenerative diseases , we demonstrate the ability of GenoSkyline-Plus to accurately identify tissue-specific functionality in the human genome and provide unbiased , genome-wide insights into the genetic basis of human complex diseases . | blood cells, genome-wide association studies, medicine and health sciences, functional genomics, immune cells, neurodegenerative diseases, immunology, human genomics, genome analysis, genome annotation, alzheimer disease, white blood cells, animal cells, dementia, mental health and psychiatry, genetic loci, cell biology, monocytes, neurology, genetics, biology and life sciences, cellular types, genomics, computational biology, human genetics | null |
journal.pgen.1008013 | 2,019 | Identification of the master sex determining gene in Northern pike (Esox lucius) reveals restricted sex chromosome differentiation | The evolution of sex determination ( SD ) systems and sex chromosomes has sparked the interest of evolutionary biologists for decades ., While initial insights on sex chromosome evolution came from detailed studies in Drosophila and in mammals 1–5 , recent research on other vertebrate groups , such as avian 6 , 7 , non-avian reptiles 8 , 9 , amphibians 10–13 , and teleost fishes 14–17 , has provided new information that helps us understand the evolution of SD systems and sex chromosomes ., Teleosts display the highest diversity of genetic sex determination systems in vertebrates , including several types of monofactorial and polygenic systems 14 , 18 ., In addition , in some species , genetic factors can interact with environmental factors , most commonly temperature , i . e . in Odontesthes bonariensis 19 , generating intricate sex determination mechanisms ., Moreover , sex determination systems in fish can differ between very closely related species , as illustrated by the group of Asian ricefish ( genus Oryzias ) 20–25 , and sometimes even among different populations of one species , as in the Southern platyfish , Xiphophorus maculatus 26 ., Beside this remarkable dynamic of sex determination systems , the rapid turnover of sex chromosomes in teleosts provides many opportunities to examine sex chromosome pairs at different stages of differentiation ., Finally , recent studies on fish sex determination have revealed a dozen new master sex determining ( MSD ) genes 14 , 16 , 27 , providing additional insight to the forces driving the turnover of SD systems and the formation of sex chromosomes ., The birth of new MSD genes drives the formation of sex chromosomes and transitions of SD systems ., The origin of new MSDs falls into two categories: either gene duplication followed by sub- or neo-functionalization , or allelic diversification 14 ., To date , teleosts are the only group where examples of both gene duplication and allelic diversification mechanisms have been found 14–17 ., Yet , among teleost species with a known MSD gene , sex chromosomes have only been sequenced and characterized in two species: the Japanese medaka ( Oryzias latipes ) , whose MSD gene originated from gene duplication / insertion of the duplicated copy 28–31 , and the Chinese tongue sole ( Cynoglossus semilaevis ) , whose MSD gene originated from allelic diversification 32 ., The duplication / insertion mechanism could suppress recombination more readily than the allelic diversification mechanism , as the newly inserted genomic segment immediately lacks homologous regions for recombination ., Furthermore , the duplicated MSD gene could potentially be inserted into different chromosomes and thus provide further flexibility in sex chromosome turnover ., To understand how the mechanism of MSD origin could impact the evolution of sex chromosomes , additional empirical studies identifying the origin of MSDs and sex chromosomes are urgently needed ., Such studies will form a rich knowledge base allowing advances of theories of sex chromosome evolution ., Among identified teleost MSD genes , the salmonid MSD gene , named sdY , is the most intriguing because it revealed a previously unexpected flexibility in SD pathways in teleosts ., While all other currently identified MSDs belong to one of three protein families ( SOX , DMRT and TGF-β and its signaling pathway ) that were known to be implicated in the SD pathways , the salmonid MSD gene sdY arose from duplication of an immune-related gene 33 ., Despite sdY being conserved in the majority of salmonid species 34 , it was not found in Esox lucius , the most studied member of the salmonid’s sister order the Esociformes 34 ., The restriction of sdY to the salmonids raised the question of what was the ancestral MSD before the emergence of the “unusual” sdY ., The first step to answer this question was to identify the genetic component responsible for sex determination in E . lucius ., E . lucius , commonly known as the Northern pike , is a large and long-lived keystone predatory teleost species found in freshwater and brackish coastal water systems in Europe , North America , and Asia 35 ., It has emerged as an important model species for ecology and conservation because of its pivotal role as a top predator that shapes the structure of local fish communities , and also as a valuable food and sport fish 36 ., Consequently , genomic resources have recently been generated for E . lucius , including a whole genome assembly anchored on chromosomes 37 and a tissue-specific transcriptome 38 ., Yet , little is known about genetic sex determination in E . lucius beyond males being the heterogametic sex 39 , and its sex locus and MSD gene remain elusive ., In this study , we identified a duplicate of anti-Müllerian hormone ( amh ) with a testis- specific expression pattern as a candidate male MSD gene for E . lucius ., Using pooled sequencing ( pool-seq ) reads from a wild population and a new draft genome sequenced with Nanopore long reads , we found limited differentiation between the homomorphic sex chromosomes and that this male-specific duplicate of amh , which we call amhby , is located within the Y-specific sequence ., Using RAD-sequencing of a family panel , we identified Linkage Group 24 as the sex chromosome and positioned the sex locus in its sub-telomeric region ., In addition , we showed that amhby has an expression profile characteristic of a male MSD gene and is functionally both sufficient and necessary to trigger testis development , providing robust support for amhby as the MSD gene in this species ., Finally , through phylogenetic and synteny analyses , we showed that this amh duplication occurred around 40 million years ago and that amhby was translocated after its duplication , which likely initiated the formation of the proto-Y chromosome ., Taking advantage of recent advances in functional genomics and sequencing technologies , our study combines the location and characterization of the sex locus and the identification of a master sex determining ( MSD ) gene with substantial functional validation in a non-model species ., Our results expand the knowledge of sex determination genes and provide insight on the evolution of sex chromosomes in teleosts ., As many currently characterized MSD genes in teleosts belong to a few ‘usual suspect’ protein families , sex-specific allelic variants or sex-specific duplicate members of genes from these ‘usual suspect’ families are strong candidates for being potential MSD ., By searching tissue-specific transcriptomes of E . lucius ( 38 , phylofish . sigenae . org ) , we identified duplication of such a ‘usual suspect’ gene; an anti-Mullerian hormone ( amh ) gene ., The two amh transcripts share 78 . 9% nucleotide identity with one transcript being predominantly expressed in adult testis and expressed at a low level in adult ovary and adult muscle , and the other transcript being exclusively expressed in adult testis ( S1 Fig ) ., PCR amplification on genomic DNA from 221 wild-caught individuals , whose phenotypic sex was determined by gonadal inspection , showed that the genomic sequence of one amh copy was present in all phenotypic males and females , while the genomic sequence of the testis-specific amh was present in 98% of phenotypic males ( 157/161 ) and 0% of phenotypic females ( 0/60 ) ( Fig 1A ) ., The significant association between this testis-specific copy of amh and male phenotype ( Chi-squared test , p< 2 . 2e-16 ) indicates that the genomic sequence of this testis-specific copy of amh transcript is Y chromosome specific ., This male-specific amh was named amhby ( Y-chromosome-specific anti-Müllerian hormone paralog, b ) and the autosomal gene was named amha ( amh paralog, a ) ., To compare the genomic regions containing amha and amhby , clones were isolated from a phenotypic male genomic fosmid library and sequenced ., The amha-containing fosmid included the entire 5’ intergenic region of amha up to the closest gene , dot1l , and the amhby-containing fosmid included a 22 kb region upstream of amhby which contained no coding sequences for other proteins ( blastx search against Teleostei , taxid:32443 ) ., Nucleotide identity between amha and amhby exon sequences ranges from 74% to 95% ( Fig 1B and 1C ) and the only gross structural difference between the two genes is a 396 bp specific insertion of a repeated region in amhby intron 1 ., Little sequence similarity was found between the proximal sequences of the two genes ( Fig 1D ) , except for a 1020 bp repetitive element ( transposase with conserved domain HTH_Tnp_Tc3_2 ) , which was present in many copies in the genome of E . lucius and was not specifically enriched around the genomic regions containing either amh or amhby ( S2 Fig ) ., Predicted proteins contain 580 amino-acids ( AA ) for amha and 560 AA for amhby , sharing 68 . 7% identity and 78 . 4% similarity ., Both proteins have a complete 95 AA C-terminal TGF-β domain with seven canonical cysteines ( S3 Fig , S4 Fig ) , sharing 62 . 5% identity and 74 . 0% similarity ., To identify the sex chromosome in the genome of E . lucius , we generated RAD-Seq data from a single full-sib family including 37 phenotypic male offspring , 41 phenotypic female offspring and the two parents ., In total , 6 , 922 polymorphic markers were aligned to the 25 Northern pike linkage groups and 40 polymorphic markers were aligned to unplaced scaffolds ( GenBank assembly accession: GCA_004634155 . 1 ) ., Genome-wide average FST between male and female offspring was 0 . 00085 and the linkage group with the highest average FST was LG24 with an FST of 0 . 052 while other chromosomes had an average FST of 0 . 0067 ., Furthermore , only markers mapped to LG24 showed genome-wide association with sex phenotype ( Fig 2A ) , indicating that LG24 is the sex chromosome of E . lucius ., To locate the sex locus and to investigate the pattern of recombination between the X and Y chromosomes in the paternal genome , we compared LG24 in the male and female RAD-Seq genetic maps ., Overall , the order of markers in both male and female genetic maps agreed with their order when aligned to the LG24 assembly ( R2 = 0 . 94 for the female map , and R2 = 0 . 89 for the male map ) with little difference between the male and female total map lengths ( 64 . 1 cM in female and 61 . 7 cM in male ) ., In the female map , recombination was observed along the entire length of LG24 ., In the male map , while recombination with the sex locus ( 0 cM in the genetic map ) was observed for the majority of the LG24 markers , 11 markers aligned to a small 400 kb region between 1 . 1 and 1 . 5 Mb showed no recombination with the sex locus , suggesting that the region with restricted recombination is small and located only in the sub-telomeric region of the LG24 sex chromosome ., Additionally , 20 markers showed no recombination with the sex locus and could not be aligned to the female reference genome , suggesting that they are Y-specific sequences ., Two of these 20 markers aligned to the amhby sequence , indicating that amhby is passed down strictly from father to sons ( S3 Table ) ., Furthermore , in a separate analysis , we identified the parental origin of all markers and plotted the distribution of these markers between male and female offspring along LG24 ( S5 Fig ) ., While maternal markers showed no sex bias in their distribution among offspring , paternal markers displayed an important sex-bias at the proximal end of LG24 , again locating the sex locus to this region ., This recombination pattern of RAD-Seq markers aligned to LG24 inferred from a single family panel represents the X and Y recombination pattern of the sire ., In order to characterize the differentiation between the X and Y chromosomes at the population level , we sequenced a pool of 30 phenotypic males and a pool of 30 phenotypic females ., Reads were aligned to the female Northern pike reference assembly ( GCA_004634155 . 1 ) , and we computed the number of male-specific SNPs ( MSS ) in a 50 kb non-overlapping window across the whole genome ., The genome average was 0 . 55 MSS per 50 kb window ( 2 . 6 MSS per window when excluding windows without MSS ) ., A single window located around 750 , 000 bp on LG24 contained 298 MSS ( Fig 2C ) , close to three times the number of MSS in the next highest window ( 111 MSS on LG07 ) , indicating that the sex locus is located near the proximal end of LG24 ., A further analysis with a higher 2 . 5 kb window resolution revealed that this enrichment of MSS on LG24 is restricted to a 300 kb region located between ~0 . 72 Mb and ~1 . 02 Mb containing 442 MSS ( Fig 2C ) ., However , as no marker from the family panel aligned to the region between 0 Mb to 1 Mb on LG24 , we do not have information regarding the suppression of X/Y recombination in the sire of the family panel for this 300 kb ., These results confirm that the sex locus is located in the proximal of LG24 , and indicate that at the population level , there is strong differentiation between the X and the Y sequences restricted to a small 300 kb sub-telomeric region of the LG24 sex chromosome ., The fosmid clone sequence containing amhby was not found in the female reference assembly ( GCA_004634155 . 1 ) , suggesting that the Northern pike Y chromosome contains unique sequences that lack homology to the X chromosome ., Therefore , to better characterize the Northern pike sex locus , we sequenced and assembled the genome of a genetic male with Nanopore long reads ., Results from BUSCO show that this new assembly has comparable completeness to that of the female reference assembly ( S1 Table ) ., In this Nanopore assembly , the entire sequence of the amhby-containing fosmid was included in a 99 kb scaffold ( tig00003316 ) , from 24 , 050 bp to 60 , 989 bp , with amhby located from 27 , 062 bp to 30 , 224 bp ., To identify these potential Y-specific sequences absent in XX females , we aligned the pool-seq reads to the XY Nanopore assembly and searched for non-overlapping 1 kb regions covered only by male reads ( MR1k ) ., In total , we found 94 MR1k-containing regions located on only three Nanopore contigs ( tig00003316 = 53 MR1k , tig00003988 = 22 MR1k , tig00009868 = 19 MR1k ) ., In contrast , only four non-overlapping 1 kb regions were covered only by female reads , each on different contigs , and we found no MR1k from the same analysis on the female reference assembly ( GCA_004634155 . 1 ) , indicating a low false discovery rate for sex-specific regions with our method ., Blasting results on the three MR1k-containing Nanopore contigs showed that amhby is the only protein coding gene apart from transposable elements ( TEs ) associated proteins in the E . lucius sex locus ( S2 Table ) ., To quantify the length of Y-specific sequences , we then looked for regions with few or no female reads mapped and male reads mapped at a depth close to half of the genome average on the three MR1k-containing contigs from the Nanopore assembly ( S6 Fig ) ., In total , we identified ~ 180 kb of Y-specific sequences in these three contigs ., One of these contigs showed strong homology to a region on LG24 spanning from ~0 . 72 Mb to ~0 . 80 Mb ( megablast , e-value = 0 , identity = 95% ) , which indicates that the male specific region is located adjacent to the ~300 kb region enriched with male specific SNPs from 0 . 72 Mb to 1 . 02 Mb on LG24 ., Together , these results indicate that the size of the sex locus of E . lucius is ~480 kb and that amhby is the only non-TE , protein-coding gene in this locus ., To characterize the temporal and spatial expression of amhby in relation to the molecular and morphological differentiation between male and female gonads , both quantitative PCR ( qPCR ) and in-situ hybridization ( ISH ) were performed ., Expression of amhby , amha , three other genes ( drmt1 , cyp19a1a , and gsdf ) known for their role in gonadal sex differentiation , and amhrII , the putative receptor for the canonical amh , was measured by qPCR at four time points from 54 days post-fertilization ( dpf ) to 125 dpf , prior to the onset of gametogenesis ., The entire trunks were used for the first three time points when the gonads were too small to be isolated , and only gonads were used at 125 dpf for both males and females ., Expression of amha was detected in both males and females starting from 75 days post-fertilization ( dpf ) , with a significantly higher expression in males than in females at 100 dpf ( Wilcoxon signed-rank test , p = 0 . 043 ) ( Fig 3A ) ., In contrast , expression of amhby was detected only in males starting from 54 dpf and increasing exponentially thereafter till 125 dpf ( R2 = 0 . 79 ) ( Fig 3B ) ., Expression of drmt1 , cyp19a1a , and gsdf was only detected from 100 dpf onwards , i . e . much later than the first detected expression of amhby ( S7 Fig ) ., Moreover , among these three genes , only cyp19a1a showed significantly different expression between sexes with a higher expression in females at 100 dpf ( Wilcoxon signed-rank test , p = 0 . 014 ) ., Expression of amhrII was not detected until 100 dpf and did not differ significantly between sexes at any stage ( S7 Fig ) ., Expression of amha and amhby was also characterized by in-situ hybridization ( ISH ) performed on histological sections of the entire trunk of male and female E . lucius sampled at 80 dpf ., Expression of amha was detected in the gonads of both female ( Fig 3C ) and male ( Fig 3D ) samples , but the signal was much stronger in male gonads ., In contrast , expression of amhby was strong in male gonads ( Fig 3F ) but not detected in female gonads ( Fig 3E ) , confirming the specificity of the probe for amhby ., In addition , no morphological differences were observed between male and female gonads at 80 dpf , even though expression of amhby was already detected by qPCR and by ISH at this stage ., Our ISH results show that the expression of both amha and amhby is high in male gonads before the first signs of histological differentiation between male and female gonads ., Collectively , these results show that amhby is expressed in the male gonads prior to both molecular and morphological sexual-dimorphic differentiation of gonads in E . lucius ., To further investigate the functional role of amhby in initiating testicular development , we performed both loss-of-function and gain-of-function experiments ., We knocked out amhby using three pairs of TALENs targeting exon 1 and exon 2 of amhby ( S8A Fig ) ., Only the T2 TALEN pair targeting exon 1 was effective in inducing deletions in the amhby sequence ., Overall , 12 of 36 ( 33 . 3% ) surviving G0 males possessed a disrupted amhby that resulted in truncated proteins ( S8B Fig ) ., G1 XY offspring obtained from three amhby mosaic G0 males crossed with wild-type females were maintained until the beginning of testicular gametogenesis at 153 dpf and then processed for histology ., Gonads from 23 G1 amhby positive XY mutants were compared with those of wild-type control XY males ( N = 4 ) and control XX females ( N = 4 ) of the same age ., Control animals developed normal ovaries and testes ( Fig 4A and 4B ) , but all 23 XY F1 amhby mutants failed to develop a normal testes ., Among these 23 G1 XY mutants , 20 ( 87% ) showed complete gonadal sex reversal , characterized by the formation of an ovarian cavity and the appearance of previtellogenetic oocytes ( Fig 4C ) ; the three ( 13% ) remaining mutants developed potentially sterile gonads with no clear ovarian nor testicular structure ., None of these 23 G1 XY mutants display a potential off-target mutation in the corresponding region targeted by the T2 TALEN pair in the amha gene ( S9 Fig ) ., To investigate whether amhby alone is sufficient to trigger testicular development , we overexpressed amhby in XX genetic females ., Two G0 XY mosaic transgenic males possessing the amhby fosmid were crossed with wild-type females , and 10 G1 XX offspring carrying the amhby fosmid were maintained along with control wild-type siblings until the beginning of testicular gametogenesis at 155 dpf ., Upon histological analysis of the gonads , all ten ( 100% ) XX transgenics carrying amhby fosmid developed testis with testicular lumen and clusters of spermatozoids ( Fig 4D ) , while all 12 control genetic males developed testis and 19 of 24 ( 79% ) control genetic females developed ovaries ., The other five control genetic females ( 21% ) developed testes ., Such sex reversal was also observed in the natural population at a rate of 2% , and this might have been exacerbated by culture conditions , a phenomenon previously documented in other teleosts 40 ., Despite this effect , the XX transgenics with amhby fosmid had a significantly higher rate of sex reversal than their control female siblings raised in identical conditions ( Chi-squared test , p = 0 . 0001148 ) ., Taken together , these results show that amhby is both necessary and sufficient to trigger testicular development in E . lucius , and further support the functional role of amhby as the MSD gene in this species ., To determine the origin of the two E . lucius amh paralogs , we generated a map of conserved syntenies for amh in several teleost species , including the spotted gar ( Lepisosteus oculatus ) as outgroup ( Fig 5A ) ., Genes located upstream ( i . e . dot1l , ell , and fkbp8 ) and downstream ( i . e . oazla ) of amha on E . lucius LG08 showed conserved synteny in all teleost species included in the analysis , indicating that LG08 is the conserved location of the E . lucius ancestral amh , now called amha , and that amhby evolved from a duplication of amha that was later translocated to the sub-telomeric region of the future sex chromosome , LG24 ., We estimated the duplication event occurred ~ 38 and ~50 million years ago ( S1 File , S7 Table ) , and found no homology between the ~ 180 Kb Y-specific sequence identified in the Nanopore assembly and the sequence of LG08 from the reference assembly , besides the two amh genes , suggesting that the translocation is also likely to be ancient ., Prior to the discovery of amhby in E . lucius , male-specific duplications of amh were identified in Patagonian pejerrey ( Odontesthes hatcheri ) 41 and Nile tilapia ( Oreochromis niloticus ) 42 ., To test whether these duplications have a shared origin , we constructed a phylogeny of Amh from nine teleost species , including these three species with male-specific amh duplications , and spotted gar Amh as outgroup ( Fig 5B ) ., In this protein phylogeny , each sex-specific Amh paralog clusters as a sister clade to its own species’ ‘canonical’ Amh with significant bootstrap values , indicating that these three pairs of amh paralogs were derived from three independent and lineage-specific duplication events ., Since the discovery of dmrt1bY in the Japanese rice fish 28 , 29 , the first identified teleost master sex determination gene , studies in teleosts have unveiled a dozen novel genes as master regulators for sex determination 14–16 , 27 ., Interestingly , many of these master sex determining genes belong to the TGF-β superfamily ., To date , this finding has been mostly restricted to teleosts , highlighting the crucial role of TGF-β signaling in the sex determination pathway in this vertebrate group ., In the present study , we identified an old duplicate of amh , a member of the TGF-β superfamily , as the male MSD gene in E . lucius ., Results from genotyping demonstrate a strong and significant association of amhby with male phenotype ., RNA-seq , qPCR and ISH showed that amhby is expressed in the male gonadal primordium before histological testis differentiation , thus fulfilling another criterion for being an MSD ., Furthermore , knockout of amhby leads to complete gonadal sex reversal of XY mutants , while overexpression of amhby in XX animals leads to the development of testis , demonstrating that amhby is both necessary and sufficient for testicular differentiation ., Together , these independent lines of evidence provide strong support that amhby is the MSD in E . lucius ., This work provides a third functionally validated case of an amh duplicate evolving into the MSD gene in a teleost species , along with the Patagonian pejerrey , Odontesthes hatcheri , 41 , and the Nile tilapia , Oreochromis niloticus , 42 ., Besides these three examples , association of amh duplicates with phenotypic sex was also found in other teleosts: in O . bonariensis , the sister species of the Patagonian pejerrey , a male-specific amhy was found to interact with temperature in determining sex 19 , and in the ling cod , Ophiodon elongatus , a male-specific duplicate of amh was also identified using molecular marker sequences 43 ., More recently , a duplicated copy of amh was found in an Atheriniformes species , Hypoatherina tsurugae , and is suspected to be involved in male sex determination 44 ., Besides amh , its canonical receptor , amhrII , has also been shown to play a pivotal role as a MSD gene in the Tiger Pufferfish , Takifugu rubripes 45 ., Our phylogenetic analysis on Amh sequences from several teleost species revealed that the three confirmed male-specific Amh duplications are independent , lineage-specific events rather than the product of shared ancestry ., This finding supports the “limited option” hypothesis for master sex determining genes 46 and makes Amh pathway members the most frequently and independently recruited master sex determining genes identified in any animal group so far ., Among teleost species with amh duplicated MSD genes , E . lucius displays the highest degree of sequence divergence between paralogs , with an average of ~ 79 . 6% genomic sequence identity ., In the Nile tilapia , amhy is almost identical to the autosomal amh , differing by only one SNP 42 ., In the Patagonian pejerrey , the shared identity between the two paralogs ranges from 89 . 1% to 100% depending on the exon 41 ., Because of the low divergence between the two paralogous sequences in the Patagonian pejerrey and the Nile tilapia , the new MSD function of amhy was attributed to their novel expression patterns 45 ., Yet , low sequence divergence , as little as one amino-acid changing SNP , was shown to be sufficient to impact the signal transduction function of the human AMH protein , leading to persistent Müllerian duct syndrome 47 , 48 ., In E . lucius , amhby also has an expression profile different from amha , likely due to a completely different promoter region ., However , because of the relatively high level of divergence between amha and amhby sequences in E . lucius , especially in the C-terminal bioactive domain of the proteins 47 , it is tempting to hypothesize that the two proteins could have also diverged in their function ., For instance , they may have a different affinity for their canonical AmhrII receptor or even the ability to bind to different receptors , leading to divergence in their downstream signaling pathways ., Because we estimated the duplication of amh in E . lucius to be between ~ 38 and ~ 50 million years old , the long divergence time between the two paralogous amh genes potentially provided opportunities for the accumulation of these sequence differences ., Further functional studies would be required to unravel the downstream signaling pathways of Amha and Amhby in E . lucius to better understand the mechanisms leading to the novel function of amhby as the MSD gene in this species 37 ., The analysis of the genomic neighborhood of both amh duplicates showed that amha is located on LG08 in a cluster of genes regulating sexual development and cell cycling with conserved synteny in teleosts 49 , 50 , while amhby is located near the telomeric region of LG24 with no other identified gene besides transposable elements within at least 99 kb in its close vicinity ., These results indicate that amha is likely to be the ancestral amh copy in E . lucius , and that amhby was translocated near the telomere of the ancestor of LG24 after its duplication ., This scenario fits the description of a proposed mechanism of sex chromosome emergence and turnover through gene duplication , translocation and neofunctionalization 31 , 51 ., Following this model , the translocation of a single copy of amh into another autosome triggered the formation of proto sex chromosomes , possibly because the newly translocated genomic segment containing the amh copy halted recombination with the X chromosome ab initio due to a complete lack of homology ., This mechanism came to light after the discovery of dmrt1bY in the Japanese medaka , which acquired a pre-existing cis-regulatory element in its promoter through a transposable element 52 ., Besides dmrt1bY , the only other well-described case of sex chromosome turnover via gene duplication , translocation and neofunctionalization is the salmonid sdY gene , which maps to different linkage groups in different salmonid species ., Our study provides a third empirical example of gene duplication and translocation giving rise to new MSD genes and bolsters the importance of this mechanism in the birth and turnover of sex chromosomes ., Theories of sex chromosome evolution predict suppression of recombination around the sex determining locus , eventually leading to its degeneration because of its lack of ability to effectively purge deleterious mutations and repeated elements 53 , 54 ., Here , we found that only a small part ( around 480 kb ) of the sex chromosome shows differentiation between X and Y chromosome , suggesting that this sex locus encompasses less than 1% of LG24 in E . lucius ., Sex chromosome differentiation has been characterized in a few other teleost species , including Chinese tongue sole 32 , Japanese medaka 30 , 55 , 56 , stickleback 57 , Trinidad guppy 58 and a few cichlid species 59 , 60 ., Compared to these examples , we found that the Northern pike displays a very limited region of suppressed recombination between the sex chromosomes ., A usual explanation for a small sex locus is that the rapid transition of sex chromosomes frequently observed in teleosts , facilitated by duplication and translocation , can readily produce neo-sex chromosomes showing little differentiation ., This scenario was demonstrated in the salmonids with the vagabond MSD gene Sdy 34 , 61 , 62 ., However , the ~ 40 million years of divergence time between amhby and amha , and the lack of homology between the sequences suggest that the sex locus of E . lucius is not nascent ., Further comparative studies that include sister species in the same clade will be needed to better estimate the age of this MSD gene and the sex chromosome , but a nascent sex locus is likely not the explanation for such a restricted region suppression of recombination between the X and Y chromosome of E . lucius ., On the other hand , old yet homomorphic sex chromosomes have been observed , for instance in ratite birds 63 ., Furthermore , in Takifugu , a single SNP conserved for 30 million years determines sex and the rest of the sex chromosomes do not show evidence of suppressed recombination , raising the possibility that decay is not the only possible fate for sex chromosomes 45 ., One mechanism for the maintenance of a small sex locus has been proposed in the Japanese rice fish , where long repeats flanking the sex locus on the Y chromosome may recombine with the same repeats on the X chromosomes , thus hindering the spread of suppression of recombination around the MSD gene 51 ., We found a slight enrichment of repeated elements on the sequences from the sex locus of E . lucius , however a better assembly of the sex locus would be needed to investigate whether a similar mechanism could have contributed to the very limited differentiation between X and Y chromosomes in E . lucius ., A widely accepted model of sex chromosome evolution postulates that the presence of sexually antagonistic alleles near the SD locus would favor the repression of recombination along the sex chromosome 64 , 65 ., This process could eventually create ‘evolutionary strata’ of different ages and different levels of differentiati | Introduction, Results, Discussion, Material and methods | Teleost fishes , thanks to their rapid evolution of sex determination mechanisms , provide remarkable opportunities to study the formation of sex chromosomes and the mechanisms driving the birth of new master sex determining ( MSD ) genes ., However , the evolutionary interplay between the sex chromosomes and the MSD genes they harbor is rather unexplored ., We characterized a male-specific duplicate of the anti-Müllerian hormone ( amh ) as the MSD gene in Northern Pike ( Esox lucius ) , using genomic and expression evidence as well as by loss-of-function and gain-of-function experiments ., Using RAD-Sequencing from a family panel , we identified Linkage Group ( LG ) 24 as the sex chromosome and positioned the sex locus in its sub-telomeric region ., Furthermore , we demonstrated that this MSD originated from an ancient duplication of the autosomal amh gene , which was subsequently translocated to LG24 ., Using sex-specific pooled genome sequencing and a new male genome sequence assembled using Nanopore long reads , we also characterized the differentiation of the X and Y chromosomes , revealing a small male-specific insertion containing the MSD gene and a limited region with reduced recombination ., Our study reveals an unexpectedly low level of differentiation between a pair of sex chromosomes harboring an old MSD gene in a wild teleost fish population , and highlights both the pivotal role of genes from the amh pathway in sex determination , as well as the importance of gene duplication as a mechanism driving the turnover of sex chromosomes in this clade . | In stark contrast to mammals and birds , a high proportion of teleosts have homomorphic sex chromosomes and display a high diversity of sex determining genes ., Yet , population level knowledge of both the sex chromosome and the master sex determining gene is only available for the Japanese medaka , a model species ., Here we identified and provided functional proofs of an old duplicate of anti-Müllerian hormone ( Amh ) , a member of the Tgf- β family , as the male master sex determining gene in the Northern pike , Esox lucius ., We found that this duplicate , named amhby ( Y-chromosome-specific anti-Müllerian hormone paralog b ) , was translocated to the sub-telomeric region of the new sex chromosome , and now amhby shows strong sequence divergence as well as substantial expression pattern differences from its autosomal paralog , amha ., We assembled a male genome sequence using Nanopore long reads and identified a restricted region of differentiation within the sex chromosome pair in a wild population ., Our results provide insight on the conserved players in sex determination pathways , the mechanisms of sex chromosome turnover , and the diversity of levels of differentiation between homomorphic sex chromosomes in teleosts . | medicine and health sciences, reproductive system, gonads, sequence assembly tools, developmental biology, genome analysis, molecular biology techniques, morphogenesis, research and analysis methods, sequence analysis, sex chromosomes, sequence alignment, chromosome biology, bioinformatics, gene mapping, molecular biology, genetic loci, sex determination, cell biology, anatomy, sexual differentiation, database and informatics methods, genetics, biology and life sciences, genomics, computational biology, chromosomes, genital anatomy | null |
journal.ppat.1004048 | 2,014 | IFITM3 Restricts Influenza A Virus Entry by Blocking the Formation of Fusion Pores following Virus-Endosome Hemifusion | The recently identified interferon-induced transmembrane proteins ( IFITMs ) inhibit infection of diverse enveloped viruses 1–3 ., Ectopic expression of IFITM1 , -2 and -3 restricts a growing number of unrelated viruses , including IAV 1 , 2 , 4–7 ., IFITM3 has been shown to potently restrict infection by IAV and the Respiratory Syncytial Virus in vivo 8–10 ., In contrast , arenaviruses and some retroviruses , such as murine leukemia virus ( MLV ) , are resistant to IFITM restriction 2 , 6 ., The IFITMs have been reported to inhibit HIV-1 entry , albeit less potently than IAV and apparently in a cell type-dependent manner 11–13 ., The mechanism by which IFITMs inhibit infection of diverse viruses is not fully understood ., IFITM2 and -3 are predominantly found in late endosomes ( LE ) and lysosomes 13 , 14 , whereas IFITM1 is also found at the cell periphery 4 , 15 ., Different membrane topologies of IFITMs have been proposed 16 , but recent data suggests that IFITM3 is a type II transmembrane protein 17 ., Accumulating evidence implies that IFITMs may interfere with virus-endosome fusion 1 , 2 , 5 , 13 , 14 ., The fact that IFITMs seem to expand acidic intracellular compartments 13 indicates that the fusion block is downstream of the low pH trigger ., Effective restriction of viruses that enter from the LE , such as IAV , Ebola virus ( EBOV ) and SARS coronavirus seems consistent with the cellular localization of IFITM2 and -3 proteins ., However , these proteins also restrict Vesicular Stomatitis Virus ( VSV ) that appears to fuse with early endosomes 18 ., IFITMs have been reported to curtail viral infection by modifying properties of cellular membranes , such as fluidity and spontaneous curvature 3 , 5 , 14 ., These effects could be related , in part , to the accumulation of cholesterol in LE as a result of IFITM-mediated disruption of the interaction between the vesicle-membrane-protein-associated protein A ( VAPA ) and oxysterol-binding protein ( OSBP ) 14 ., Since lipids play an important role in membrane fusion , these findings offer an attractive paradigm for a broad antiviral defense mechanism that involves altering the lipid composition of cellular membranes ., The recent finding that amphotericin B , which forms complexes with sterols 19 , rescues IAV infection in IFITM2- and IFITM3-expressing cells 20 is in line with the notion that cholesterol may be directly or indirectly involved in IAV restriction ., However , lipid composition-based models do not readily explain the lack of restriction of amphotropic MLV and arenaviruses , which enter cells via distinct endocytic routes 21 , 22 ., These findings indicate that IFITMs may restrict virus entry from a subset of intracellular compartments ., In order to define the mechanism of IFITM restriction , it is important to identify the viral entry step ( s ) targeted by these proteins , define compartments in which restriction occurs , and elucidate potential changes in intracellular membranes that may be responsible for this phenotype ., Here , we examined the mechanism of IFITM3 restriction of IAV using single particle imaging and a direct virus-cell fusion assay ., Our results show that IFITM3 does not inhibit the lipid mixing stage of IAV fusion but blocks the release of viral contents into the cytosol , and that this phenotype does not correlate with cholesterol accumulation in intracellular compartments ., Specifically , IFITM3 inhibits the conversion of hemifusion to fusion through a mechanism that does not rely on cholesterol accumulation ., Together these findings reveal a previously unappreciated view of IFITM-mediated restriction and suggest new avenues of investigation to delineate the mechanism by which these proteins block infection ., We chose to focus on IFITM3 to study the mechanism of IAV restriction because this protein potently inhibits infection in vitro and in vivo 8–10 ., Since published data suggest that IFITM3 likely inhibits the viral fusion step , a direct virus-cell fusion assay was employed to evaluate the extent of restriction in different cell lines 23 ., HIV-1 particles carrying the β-lactamase-Vpr ( BlaM-Vpr ) chimera and pseudotyped with the influenza HA and NA proteins from the H1N1 A/WSN/33 strain ( referred to as IAVpp ) were allowed to fuse with cells transduced with an empty vector or with an IFITM3-expressing vector ., The resulting cytosolic BlaM activity was measured as previously described 24 ., Out of several cell lines tested , A549 and MDCK cells over-expressing IFITM3 were least permissive to IAVpp fusion ( Fig . 1A ) ., In agreement with the previous reports 2 , 13 , we found that IFITM3 over-expression partially inhibited VSV G glycoprotein-mediated fusion of pseudoviruses ( VSVpp ) carrying the BlaM-Vpr chimera ( Fig . 1A ) ., Similar to inhibition of IAVpp fusion , the IFITM3-mediated restriction of VSVpp was most potent in A549 and MDCK cells ., As expected , fusion of particles pseudotyped with the Lassa fever virus glycoprotein ( LASVpp ) , which directs virus entry through an IFITM3-resistant pathway 2 , 6 , was not considerably affected by IFITM3 over-expression ., We next checked if the strong suppression of virus fusion in A549 and MDCK cells was related to the level of IFITM3 expression ., Immunostaining for IFITM3 in these and CHO cells which exhibited modest restriction of viral fusion ( Fig . 1A ) did not reveal a clear correlation between IFITM3 expression and inhibition of IAVpp or VSVpp fusion ( Fig . 1B ) ., Of note , potent IAV restriction in A549 and MDCK cells was not related to the usage of HIV-1 core-based pseudoviruses ., Influenza virus-like particles containing the IAV BlaM-M1 chimera 25 also failed to efficiently fuse with A549-IFITM3 and MDCK-IFITM3 cells while fusing well with vector-transduced cells ( Fig . 1C ) ., We also found that both vector-transduced A549 and MDCK cells were highly susceptible to IAV infection , as determined by virus titration ( see Materials and Methods ) ., These two cell lines were therefore chosen for studies of IFITM3-mediated restriction described below ., IFITM-based restriction has been studied using a cell-cell fusion model , as well as by forcing viral fusion with the plasma membrane by lowering the pH 5 , 20 ., Since fusion with the plasma membrane is more amenable to mechanistic studies than endocytic entry , we asked whether IFITM3 can restrict forced IAV fusion ., Exposure to acidic buffer induced IAVpp fusion with A549-Vector cells pretreated with Bafilomycin A1 ( BafA1 ) , which blocked low pH-dependent entry from endosomes ( Fig . 1D ) ., The extent of forced fusion was lower compared to the conventional entry route ., By contrast , forced IAVpp fusion with A549-IFITM3 cells was ∼3-fold more efficient than endocytic fusion with cells not treated with low pH or BafA1 , showing that IFITM3 does not restrict IAVpp fusion at the cell surface ., Interestingly , IFITM1 suppressed IAVpp-plasma membrane fusion at low pH ( Fig . 1D ) , in agreement with the Jaagsiekte sheep retrovirus ( JSRV ) and IAV fusion data 5 , 20 ., The inability of IFITM3 to block IAV fusion with the plasma membrane is consistent with its lower abundance at the cell surface 13 , 14 , 20 and shows that the mechanism of restriction must be studied in intracellular compartments ., Preponderance of evidence implies that hemifusion is a universal intermediate ( reviewed in 26 , 27 ) that precedes the formation of a fusion pore ., Having shown that IFITM3 over-expression inhibits viral fusion ( Fig . 1A , C ) , we asked whether this protein also blocks the upstream hemifusion step ., This was accomplished by labeling the A/PR/8/34 virus membrane with a self-quenching concentration of vybrant DiD ( vDiD ) , using a modification of the previously published protocol 28 ., Incorporation of self-quenching quantities of a lipophilic dye enables the visualization of single lipid mixing events based on the marked increase in fluorescence upon dye redistribution to an endosomal membrane ( see for example 28 , 29 ) ., Significantly , to control for fluctuations in the vDiD fluorescence caused by deviation from a focal plane , the viral surface proteins were labeled with the amine-reactive AlexaFluor-488 ( AF488 ) dye ., The relatively steady AF488 signal before and after hemifusion is allowed correcting for the vDiD intensity fluctuations due to moving in and out of focus ., The vDiD/AF488 co-labeling protocol only modestly ( <2-fold ) reduced IAV infectivity compared to the mock-labeled viruses ( Fig . S1A ) ., Immunofluorescence staining of AF488-labeled virions with anti-HA antibodies revealed an excellent co-localization of the two signals ( Fig . S1B , C ) , thus supporting the notion that AF488/vDiD-labeled particles are bona fide virions ., Labeled viruses were allowed to enter A549-Vector cells , and the resulting lipid mixing activity was examined by single particle tracking ., A fraction of virions exhibited a marked increase in the vDiD signal ( Fig . 2A , B ) ., Redistribution of vDiD was mediated by low pH-dependent conformational changes in the IAV HA glycoprotein , as evidenced by potent inhibition of lipid mixing by anti-HA antibodies ( Fig . 2C ) and by NH4Cl ( Fig . 3A ) ., Without simultaneous monitoring of the viral content release into the cytoplasm , vDiD dequenching does not discriminate between hemifusion ( operationally defined as lipid mixing without content transfer 30 ) and full fusion ., To avoid over-interpreting dequenching events , we will refer to these events as lipid mixing or hemifusion ., A similar vDiD dequenching pattern was observed in MDCK cells transduced with an empty vector ( data not shown ) ., Analysis of lipid mixing showed that 2 . 2±0 . 4% and 5 . 6±0 . 6% of cell-bound particles released vDiD in A549 and MDCK cells , respectively ( Fig . 3A ) ., By comparison , a much greater fraction of virions ( 38 . 3±0 . 6% ) hemifused with CHO cells ( data not shown ) , in agreement with the previously reported data 28 ., Importantly , IAV lipid mixing was readily detected in IFITM3+ A549 and MDCK cells ( Figs . 2D–G and 3A ) ., Not only was lipid mixing not inhibited in A549-IFITM3 cells , but a >3-fold greater fraction of particles released vDiD in these cells compared to control cells ( Fig . 3A , P<0 . 001 ) ., By comparison , IFITM3 over-expression in MDCK cells did not significantly promote vDiD dequenching ( Fig . 3A ) ., Thus , contrary to the cell-cell fusion results 5 , IFITM3 does not inhibit and can even promote IAV lipid mixing , consistent with the block of virus entry at a post-hemifusion stage ., Accordingly , the addition of oleic acid , which augments hemifusion by altering spontaneous membrane curvature , did not rescue IAVpp or VSVpp fusion with A549-IFITM3 cells ( Fig . S2 ) ., This is in agreement with the recent infectivity results 20 , but in contrast with the rescue of fusion between JSRV Env- and IFITM-expressing cells by this fatty acid 5 ., The higher frequency of vDiD dequenching in A549-IFITM3 cells could be caused by the increased endosome acidity compared to control cells 13 ., However , the distribution of waiting times to the onset of lipid mixing was independent of IFITM3 expression or the type of target cells ( A549 vs . MDCK , Fig . 3B , P\u200a=\u200a0 . 37 ) ., The fact that the kinetic curves do not reach plateau indicates that IAV entry into A549 and MDCK cells is not completed within the first hour ., Our results thus demonstrate that IFITM3 restricts the IAV fusion at a post-hemifusion step , most likely at the point of fusion pore opening , as evidenced by the dramatic decrease of the BlaM signal in A549 and MDCK cells expressing this protein ( Fig . 1A ) ., Under our conditions , vDiD dequenching was typically completed within a few minutes for both control and IFITM3+ cells ( Fig . 2 ) ., This dequenching rate is much slower than sudden increases in fluorescence of the IAV membrane markers described previously 28 , 31 ., While a portion of vDiD dequenching could be completed within seconds ( Fig . S3 ) , these fast events were not common ., Slow dequenching was also typical with the vDiD/AF488-labeled X31 virus , as well as with the X31 virus labeled with a 15-fold excess of DiD , using the published protocol for single virus imaging 28 ( data not shown ) ., Slow vDiD dequenching during the first hour of virus-cell co-incubation did not appear to result from IAV degradation in LE/lysosomes , since the surface-exposed AF488 label persisted long after vDiD dequenching was completed and because anti-HA antibodies blocked vDiD dequenching ( Fig . 2 ) ., In addition , we did not detect any correlation between the lag before the onset of lipid mixing and the vDiD dequenching slope ( Fig . S4A ) ., This result reinforces the notion that late lipid mixing events are mediated by HA and not by virus degradation ., Control experiments , in which samples were not exposed to laser light during the first 30 min at 37°C , did not reveal fast dequenching events reaching completion in less than 1 min ( data not shown ) ., This control argues against phototoxicity-related attenuation of virus fusogenicity as the cause for sluggish lipid redistribution ., Since free vDiD diffusion between a virus and a small endosome should be completed in less than a second 32 , 33 , an initial membrane connection between IAV and an endosome must severely impair lipid movement ., To assess whether early fusion intermediates in control and IFITM3+ cells restrict vDiD diffusion to the same extent , we examined the rate of vDiD dequenching ., Single particle analysis revealed that , in A549 cells , the average vDiD dequenching profile ( Fig . 3C ) was independent of IFITM3 expression , as were the initial slopes of vDiD dequenching ( Fig . S4B , P>0 . 5 ) ., These results indicate that IFITM3 over-expression does not affect the properties of fusion intermediates responsible for vDiD redistribution , such as the size and/or architecture of a hemifusion site ( e . g . , 34 , 35 ) ., We then asked whether the rate of vDiD dequenching varied depending on the cell type ., The average rate of vDiD fluorescence increase in MDCK cells was ∼2-fold greater than in A549 cells ( Figs . 3C and S4B , P<0 . 02 ) ., This demonstrates our ability to detect changes in the rate of vDiD transfer and shows that lipid transfer lasts several minutes irrespective of the cell type ., We also examined the final extent of vDiD dequenching , which is proportional to the surface area of a target membrane over which it redistributes ., This parameter was not significantly affected by IFITM3 expression in A549 cells or by the cell type ( MDCK vs . A549 cells , Fig . 3D ) ., Together , similar kinetics and extents of viral lipid dilution in control and IFITM3+ cells suggest that neither the size/architecture of early fusion intermediates nor the surface area of target endosomes is considerably affected by IFITM3 expression ., To investigate the relationship between lipid mixing and productive IAV infection , we compared the fraction of cells “receiving” at least one vDiD dequenching event in live cell imaging experiments to the fraction of cells that got infected under the same conditions ., The only difference was that virus imaging was not continued beyond 1 h after initiation of fusion , whereas infection proceeded overnight ., We found that one or more vDiD dequenching events occurred in 15% of A549 cells while 44% of cells got infected ( Fig . S5 ) ., Under the same conditions , 20% of MDCK cells “hosted” one or more dequenching events and 36% were infected ., The greater fraction of infected cells compared to those permissive to hemifusion is likely due to the shorter time widow for single virus imaging , which is likely to miss late vDiD dequencing events ( Fig . 3B ) ., The lower apparent fraction of cells supporting vDiD dequenching could also be caused by the presence of viruses that did not incorporate self-quenching amounts of vDiD ., Importantly , the comparable efficiencies of lipid mixing and infection , indicate that the former events likely culminate in productive infection ., To determine whether IFITM3 impairs the IAVs ability to form small fusion pores , we attempted to load the virus with a content marker by soaking in a concentrated solution of sulforhodamine B , as described in 36 ., However , only a small fraction of AF488-labeled particles stained with sulforhodamine , and the retained dye was lost in live cell experiments under conditions that blocked IAV fusion ( data not shown ) ., We therefore resorted to using HIV pseudoviruses bearing A/WSN/33 HA and NA glycoproteins and co-labeled with the capsid marker , YFP-Vpr , and the content marker , Gag-iCherry 24 , 37 ., Upon virus maturation , the “internal” mCherry is proteolytically cleaved off the HIV-1 Gag-iCherry and released through a fusion pore , as manifested by the loss of the red signal ( Fig . 4 and 37 ) ., The YFP-Vpr signal , which remained associated with the viral core after fusion , provided a reference signal for single particle tracking ., Under our conditions ∼1% of double-labeled pseudoviruses entering A549-Vector cells lost their content marker , while approximately 2% fused with MDCK-Vector cells ., In sharp contrast , the mCherry release in IFITM3+ A549 and MDCK cells or in vector-transduced cells in the presence of NH4Cl could not be detected ( Fig . 4E , P<0 . 001 ) ., Thus , IFITM3 does not adversely affect IAV hemifusion but severely inhibits viral content release into the cytoplasm ., Together these findings suggest that the mechanism of IFITM3-mediated restriction arises from the entrapment of viruses at a hemifusion intermediate prior to fusion pore formation ., A recent study has shown that , through disrupting the interaction between VAPA and OSBP , IFITM3 causes cholesterol accumulation in LE 14 ., Based on this finding , the authors proposed that high levels of endosomal cholesterol may inhibit IAV fusion and/or the release of nucleocapsid ., Staining with filipin revealed that IFITM3+ A549 cells exhibited increased levels of intracellular cholesterol ( Fig . 5A ) ., However , the filipin signal was still primarily associated with the plasma membrane and the total cellular cholesterol was not elevated in IFITM3+ cells ( Fig . S6 ) ., In addition , the overall intensity of intracellular cholesterol poorly correlated with the level of IFITM3 expression ( Fig . 5C ) ., By comparison , pretreatment of A549-Vector cells with U18666A , which inhibits transport of LDL-derived cholesterol from LE/lysosomes ( reviewed in 38 ) , resulted in a dramatic shift in the filipin staining pattern from the plasma membrane to endosomes ( Fig . 5B ) ., Aberrant accumulation of cholesterol in LE is also known to occur in cells lacking the functional NPC1 cholesterol transporter 39 ., We therefore knocked down NPC1 expression in A549 cells using shRNA ( shNPC1 , Fig . 5D ) and examined the resulting cholesterol distribution ( Fig . 5B ) ., Reduced NPC1 expression correlated with excess cholesterol in intracellular compartments , which was also much more pronounced than endosomal filipin staining in A549-IFITM3 cells ., We next asked whether the cholesterol accumulation induced by U18666A pretreatment or by down regulation of NPC1 can phenocopy the IFITM3-mediated restriction of viral fusion ., Neither IAV lipid mixing ( vDiD dequenching ) nor fusion ( BlaM signal ) was inhibited by silencing NPC1 in A549 cells ( Fig . 5E , F ) ., VSVpp also fused with shNPC1-transduced cells as efficiently as with control cells ( Fig . 5E ) ., These results show that excess cholesterol does not inhibit viral fusion or hemifusion ., In control experiments , silencing the NPC1 expression potently suppressed fusion of Ebola GP-pseudotyped particles ( EBOVpp , Fig . 5E ) , which use NPC1 as a receptor 40 , 41 ., Similar to the NPC1 knockdown phenotype , pretreatment of A549 cells with 10 µM U18666A , which caused cholesterol buildup in endosomes ( Fig . 5B ) , did not inhibit fusion of IAVpp or VSVpp ( Fig . 5G ) ., As will be shown below for MDCK cells , higher doses of U18666A can inhibit viral fusion ( Fig . 5G ) , but this effect is due to elevation of endosomal pH as opposed to cholesterol accumulation in endosomes ., To generalize the effects of excess cholesterol in A549 cells , we tested whether endosomal cholesterol can inhibit viral fusion in MDCK cells ., As in A549 cells , IFITM3 over-expression in MDCK cells caused moderate accumulation of cholesterol in endosomes ( Fig . 6A ) , while pre-treatment with U18666A caused a much more dramatic buildup of intracellular cholesterol ( Fig . 6B ) ., However , unlike A549 cells , IAVpp and VSVpp fusion was significantly inhibited in U18666A-treated MDCK cells ( Fig . 6C ) ., Since prolonged exposure to U18666A has been reported to raise endosomal pH 42 , we sought to determine if insufficiently acidic pH could prevent IAV hemifusion/fusion with pretreated MDCK cells ., The pH in IAV-carrying endosomes was measured using virions co-labeled with the pH-insensitive AF488 ( green ) and CypHer5E ( red ) , which fluoresces brighter at acidic pH 28 ( Fig . S7A ) ., Cells were incubated with viruses for 45 min , and the red/green signal ratio from individual particles was measured ( Fig . S7B ) ., The average pH in virus-containing endosomes of MDCK-IFITM3 cells was slightly less acidic than in control cells: 5 . 38±0 . 03 ( n\u200a=\u200a498 ) vs . 4 . 98±0 . 04 ( n\u200a=\u200a242 ) , respectively ( Fig . 6D and F , P<0 . 001 ) ., Interestingly , as shown in Figure 6E , endosomal pH in U18666A-treated MDCK cells was markedly shifted to neutral values ( 6 . 44±0 . 05 , n\u200a=\u200a160 , P<0 . 001 ) ., Since the pH threshold for triggering A/PR/8/34 fusion is reported to be around 5 . 6 43 , elevation of endosomal pH in U18666A-treated MDCK cells is the likely cause of inhibition of viral fusion ., Together our results imply that U18666A most likely attenuates IAV fusion with MDCK cells by raising endosomal pH and not through inducing cholesterol accumulation ., We also took advantage of the available CHO cell line that does not express NPC1 44 to further ascertain the role of endosomal cholesterol in IAV fusion ., These cells ( designated CHO-NPC1− ) exhibited exaggerated endosomal cholesterol staining , in sharp contrast to a peripheral staining pattern in parental CHO cells ( Fig . 7A ) ., In spite of the high endosomal cholesterol content in CHO-NPC1− cells and of the elevated level of total cholesterol ( Fig . S6 ) , IAVpp fused with these cells as efficiently as with parental cells ( Fig . 7C ) ., The NPC1-null cells also supported IAV lipid mixing , albeit at somewhat reduced level compared to control ( Figs . 7D and S8 ) ., Pretreatment of CHO cells with U18666A also trapped cholesterol in endosomes and raised the total cholesterol content ( Figs . 7B and S6 ) , but only modestly diminished the extent of IAVpp or VSVpp fusion ( Fig . 7E ) ., Interestingly , in contrast to the decreased endosome acidity in MDCK cells , endosomes in U18666A-treated CHO cells were more acidic than in control cells ( Fig . S9 ) ., In control experiments , both the lack of NPC1 expression and U18666A pretreatment blocked EBOVpp fusion ( Fig . 7C , E ) , consistent with its reliance on NPC1 receptor and high sensitivity to disruptions of cholesterol transport 45 ., Together , our results show that the cholesterol accumulation achieved through two different interventions – U18666A pretreatment and NPC1 silencing – does not phenocopy IFITM3-mediated restriction of viral fusion ., This implies that, ( i ) elevated levels of endosomal cholesterol do not generally confer resistance to viral fusion , and, ( ii ) the mechanism by which IFITM3 blocks transition from hemifusion to full fusion is not through the mislocalization of cholesterol ., The IFITMs restrict the cellular entry of multiple pathogenic enveloped viruses ., Recent studies lead to a model that IFITMs inhibit virus-host hemifusion 5 and that the membrane-rigidifying properties of cholesterol may contribute to antiviral actions 14 ., In contrast to these studies , our results now demonstrate that IFITM3 prevents the release of viral genomes into the cytosol by inhibiting viral entry after hemifusion but prior to fusion pore formation ( Fig . 8 ) ., Moreover , we found that IFITM3 can promote hemifusion in some cells , perhaps secondary to its acidifying the endosomal pathway ., IFITM3 therefore does not negatively regulate the properties of contacting leaflets involved in hemifusion , but stabilizes the cytoplasmic leaflet of the endosomal membrane , thereby disfavoring the formation of fusion pores 35 ., In one potential scenario IFITM3 is located directly at the site of arrested hemifusion , perhaps “toughening” the endosomal membrane to create a barrier to viral entry ( Pathway 1 ) ., A considerable colocalization of IFITM3 with internalized IAV ( 3 and Fig . S10 ) is consistent with Pathway 1s direct mechanism of inhibition ., Alternatively , IFITM3 might arrest hemifusion through an indirect mechanism , perhaps involving modulation of lipid and/or protein composition of the cytoplasmic leaflet ( Pathway 2 ) ., Recent findings that changes in global membrane properties interfere with productive entry would appear to support an indirect mechanism 5 , 14 ., Lipids , such as unsaturated fatty acids and cholesterol that confer negative spontaneous curvature to membranes can promote hemifusion ( a net negative curvature structure ) and disfavor a fusion pore ( a net positive curvature intermediate ) , as has been previously shown for oleic acid 35 ., Although this prediction is consistent with efficient lipid mixing in endosomes of IFITM3+ cells observed in our imaging experiments , several studies 20 , 46–48 and our own results do not support cholesterol accumulation as playing a role in fusion inhibition ., We found that cholesterol-laden endosomes in cells pretreated with U18666A or expressing undetectable/low levels of NPC1 supported efficient viral fusion ., It is thus possible that IFITM3 interferes with cellular functions of VAPA other than the interaction with OSBP , such as regulation of SNAREs and modulation of lateral mobility of membrane proteins ( reviewed in 49 ) ., IFITM3 appears to induce the formation multivesicular bodies and increase the number of ILVs 13 , 14 ., One can therefore envision that IFITM3 may inhibit infection by redirecting viruses to a non-productive pathway , perhaps involving fusion with ILVs instead of the limiting membrane of LE ( Fig . 8 , Pathway 3 ) ., If , as suggested in 14 , IFITM3 disallows back fusion of ILVs with the limiting membrane , then virus-ILV fusion products will likely be degraded ., Indeed , back fusion has been implicated in the VSV core release into the cytosol following the virus-ILV fusion 50 ., It should be stressed that this “fusion decoy” model does not explain the ability of IFITM1 to interfere with fusion at the cell surface ( 5 and Fig . 1D ) ., It is also not clear why the Old World arenaviruses , which have been reported to enter from MVBs 51 , are not restricted by IFITMs ., The indistinguishable extents of vDiD dequenching in control and IFITM3+ cells ( Fig . 3D ) indicate that target endosomes have similar sizes ., While this appears to argue against redirection of IAV fusion to small ILVs , the lack of a post-hemifusion decay of vDiD fluorescence in A549 and MDCK cells ( Figs . 2 and S3 ) is consistent with IAV fusion with abundant ILVs in endosomes of IFITM3+ cells ., This is because a lipophilic dye in the limiting membrane of an endosome should be quickly removed through membrane trafficking 24 , 31 , 52 ., Because post-dequenching decay was not observed irrespective of the level of IFITM3 expression , it is possible that IAV may infect several cell lines by fusing with small intralumenal vesicles followed by the nucleocapsid release through back fusion ( Fig . 8 , dashed black arrows ) ., This pathway could explain the similar extents and rates of vDiD dequenching in control and IFITM3-expressing cells , which are indicative of similar lipid intermediates and of the size of a target membrane , respectively ., As discussed above , slow vDiD dequenching observed by single IAV imaging can be rationalized in the context of fusion with the limiting membrane of endosomes ( Pathways 1 and 2 ) , as well as in the context of fusion with ILVs ( Pathway 3 ) ., Slow dilution of this dye in Pathway 3 could occur through multiple rounds of IAV fusion with small ILVs , whereas Pathways 1 and 2 would predict restricted lipid diffusion through early fusion intermediates formed at the limiting membrane ., Although the latter notion is in agreement with the reported restriction of lipid movement through hemifusion sites and small fusion pores 34 , 35 , 53 , 54 , these intermediates are usually short-lived under physiological conditions and tend to resolve into larger structures that do not impair lipid movement 28 , 32 , 35 ., Clearly , more detailed studies of virus-endosome hemifusion and fusion are needed to understand the nature of slow lipid redistribution between IAV and endosomes ., The IFITMs may now arguably be one of the most broadly acting and clinically relevant restriction factor families 1 , 3 ., While both IFITM3s membrane-associated topology and its localization to the site of viral attenuation suggest it acts to restrict viral entry via a direct mechanism , additional work remains to be done to fully elucidate its actions ., Nonetheless , as the primary effector of IFNs anti-IAV actions , IFITM3 represents a previously unappreciated class of restriction factor that prevents viral entry by stabilizing a hemifusion intermediate , likely comprised of an invading virus fatally tethered to the interior of the endosomes limiting membrane ., Future single virus experiments combining the detection of both viral lipid and content release events ( see for example 52 ) should provide further insights into IAV entry pathways and the mechanism of IFITM3-mediated restriction ., Indeed , such efforts may also bring to light unknown viral countermeasures , which are perhaps employed by the IFITM-resistant New and Old World arenaviruses ., HEK 293T/17 cells and human lung epithelial A549 cells were obtained from ATCC ( Manassas , VA ) and grown as previously described 55 ., Wild-type CHO cells and CHO-NPC1− cells , a gift from Dr . L . Liscum ( Tufts University ) 44 , were grown in Alpha-MEM ( Quality Biological Inc , Gaithersburg , MD ) supplemented with 10% FBS and penicillin-streptomycin ., The A549 , MDCK , HeLaH1 and CHO cells stably expressing IFITM3 or IFITM1 were obtained by transducing with VSV-G-pseudotyped viruses encoding wild-type IFITM3 and IFITM1 or with the vector pQCXIP ( Clontech ) and selecting with puromycin , as described previously 2 ., The pR8ΔEnv , BlaM-Vpr , pcRev , HIV-1 Gag-iCherryΔEnv and pMDG VSV G expression vectors were described previously 37 , 55 ., The YFP-Vpr was a gift from Dr . T . Hope ( Northwestern University ) ., The pCAGGS vectors encoding influenza H1N1 WSN HA and NA were provided by Donna Tscerne and Peter Palese , and the pCAGGS BlaM1 ( WSN ) plasmid was a gift from Dr . A . Garcia-Sastre ( Mount Sinai ) ., Vectors expressing phCMV-GPc Lassa and pcDNA3 . 1-Ebola GP ( Zaire ) were gifts from Dr . F ., -L ., Cosset ( Université de Lyon , France ) 56 and Dr . L . Rong ( University of Illinois ) 57 , respectively ., U18666A was from Tocris Bioscience ( Bristol , UK ) ., Poly-L-lysine , filipin , sulphorhodamine B Bafilomycin A1 and the Cholesterol Kit were from Sigma-Aldrich ., AlexaFluor-488 amine-reactive carboxylic acid , vybrant-DiD ( vDiD , 1 , 1′-dioctadecyl-3 , 3 , 3′ , 3′-tetramethylindodicarbocyanine , 4-chlorobenzenesulfonate salt ) , Hoechst-33342 and Live Cell Imaging buffer were purchased from Life Technologies ( Grand I | Introduction, Results, Discussion, Materials and Methods | Interferon-induced transmembrane proteins ( IFITMs ) inhibit infection of diverse enveloped viruses , including the influenza A virus ( IAV ) which is thought to enter from late endosomes ., Recent evidence suggests that IFITMs block virus hemifusion ( lipid mixing in the absence of viral content release ) by altering the properties of cell membranes ., Consistent with this mechanism , excess cholesterol in late endosomes of IFITM-expressing cells has been reported to inhibit IAV entry ., Here , we examined IAV restriction by IFITM3 protein using direct virus-cell fusion assay and single virus imaging in live cells ., IFITM3 over-expression did not inhibit lipid mixing , but abrogated the release of viral content into the cytoplasm ., Although late endosomes of IFITM3-expressing cells accumulated cholesterol , other interventions leading to aberrantly high levels of this lipid did not inhibit virus fusion ., These results imply that excess cholesterol in late endosomes is not the mechanism by which IFITM3 inhibits the transition from hemifusion to full fusion ., The IFITM3s ability to block fusion pore formation at a post-hemifusion stage shows that this protein stabilizes the cytoplasmic leaflet of endosomal membranes without adversely affecting the lumenal leaflet ., We propose that IFITM3 interferes with pore formation either directly , through partitioning into the cytoplasmic leaflet of a hemifusion intermediate , or indirectly , by modulating the lipid/protein composition of this leaflet ., Alternatively , IFITM3 may redirect IAV fusion to a non-productive pathway , perhaps by promoting fusion with intralumenal vesicles within multivesicular bodies/late endosomes . | Interferon-induced transmembrane proteins ( IFITMs ) block infection of many enveloped viruses , including the influenza A virus ( IAV ) that enters from late endosomes ., IFITMs are thought to prevent virus hemifusion ( merger of contacting leaflets without formation of a fusion pore ) by altering the properties of cell membranes ., Here we performed single IAV imaging and found that IFITM3 did not interfere with hemifusion , but prevented complete fusion ., Also , contrary to a current view that excess cholesterol in late endosomes of IFITM3-expressing cells inhibits IAV entry , we show that cholesterol-laden endosomes are permissive for virus fusion ., The ability of IFITM3 to block the formation of fusion pores implies that this protein stabilizes the cytoplasmic leaflet of endosomal membranes , either directly or indirectly , through altering its physical properties ., IFITM3 may also redirect IAV to a non-productive pathway by promoting fusion with intralumenal vesicles of late endosomes instead of their limiting membrane . | cell biology, biology and life sciences, immunology, microbiology, molecular cell biology, biophysics | null |
journal.pcbi.1004850 | 2,016 | A Robust Feedforward Model of the Olfactory System | Although it is still debated how many different odorants humans can perceive , the most commonly cited number is on the order of 104 1–3 , much greater than the 500 olfactory receptor neuron ( ORNs ) types ., Many other species , including both vertebrates and insects , have the same order of magnitude of ORN types or even fewer ( around 1000 in mice , 50 in Drosophila ) ., The order of magnitude difference between the number of odorants and ORN types implies that humans as well as other species rely on compressed representations , potentially following the principles of compressed sensing 4–7 ., In the compressed sensing framework 4 , sparse high dimensional signals can be accurately reconstructed using a small number of measurements provided that the input signals are sparse ., Natural odors are sparse in the sense that they are dominated by a few molecular components 8–10 ., The relevance of compressed sensing algorithms to olfactory coding is reinforced by the anatomical organization of the olfactory system ., High dimensional odor signals are compressed into a low-dimensional representation in terms of the activity of a relatively small number of glomeruli in the olfactory bulb , in the case of vertebrates , or the antennal lobe in the case of invertebrates ., The standard compressed sensing algorithm performs signal reconstruction as a constrained ℓ1 minimization 4 ., Such optimization can be solved through neural dynamics 5 , 6 , but the resulting reconstructions were considerably less fault tolerant than observed experimentally ., For example , mice olfactory discrimination remains essentially intact when half of glomeruli are disabled 11 whereas theoretical reconstructions fail at this level of signal interference 5 ., Furthermore , signal reconstruction based on dynamical optimization by construction requires more time for signal recognition compared to feedforward reconstruction schemes ., Here we describe a feedforward reconstruction scheme based on compressed sensing ideas that is both fault tolerant and matches the main features of the organization of the olfactory system ., The results demonstrate that a purely feedforward network is capable of robustly compressing/decompressing binary signal without dynamical optimization ., We begin by reviewing the main results from compressed sensing literature as they pertain to olfactory coding ., The odor signal s0 can be described as a binary vector of length N where each element is either 1 or 0 depending upon whether a given molecular component is present or not in the odor ., We refer to the number K of nonzero components in the odor as the odor sparsity ., The main premise of compressed sensing is that a sparse signal s0 can be compressed into a vector x = As0 of length M < N and then recovered with high reconstruction quality provided K ≪ N . The encoding matrix A has dimensions M × N; its matrix elements can be chosen randomly ., With this setup , the original signal s0 can be recovered exactly from the convex ℓ1 optimization problem 4, s ^ = arg min | | s | | 1 subject to x = A s 0 ., ( 1 ) Although the ℓ1 minimization problem can be solved in polynomial time , it is not straightforward to implement such optimization algorithms in a neural circuit ., One solution involves a two-layer neural network that perform similar ℓ1 minimization through neural dynamics 6 ., However , this imposes certain requirements on the structure of recurrent connections in the second layer together with a static nonlinear activation function ., Another alternative implementation relies on ℓ2 minimization instead of ℓ1 ., In this case , the reconstruction is obtained simply as s ^ = ( A T A ) - 1 A T x where the −1 represents a pseudo-inverse relation ., However , such an approach does not produce exact signal reconstruction 7 and would predict much larger errors than observed in olfactory experiments ., We now propose a model for the olfactory system , which can compress and robustly recover sparse binary signal with high probability , without using any dynamical optimization ., The solution is based on a nonlinear binary encoding model instead of the linear encoding model used in the conventional compressed sensing approach ., Specifically , the compressed vector x has the form of a threshold function x i = H ( x i l - θ c ) where xl = As0 and H is the Heaviside step function with H ( 0 ) = 1 ., We assume that the measurement matrix ( affinity matrix ) A is a M×N random binary matrix where each element is chosen independently to be either 1 or 0 with equal probability p and 1 − p , respectively ., It is worth mentioning that while we use a random connectivity matrix in our model , we do not assume that this matrix differs across individuals ., Rather , the randomness is meant to characterize how well the system works in the absence of specificity between odorants and glomeruli identity ., By extending the definition of H to vectors , the measurement vector x can be compactly written as, x = H ( A s 0 - θ c ) , ( 2 ), where θc = 1 , reflects that all measurements larger than 1 are set to 1 so that x is binary ., This corresponds to a binary model of glomeruli activity described by the binary vector x ., The threshold value of θc = 1 corresponds to a logical OR operation , so that glomerulus k will be activated if any of the odor components that are associated with inputs to this glomerulus are activated ., To reconstruct the original signal , the glomeruli activity x are projected to another layer of neurons ( neurons in the olfactory cortex of vertebrates or Kenyon cells in the mushroom body of insects ) which has the same dimension as the original signal s0 ., The activity of neurons in this layer is denoted by vector s ^ which has the same dimensionality N as the original signal s0 ., The reconstructed signal can be computed as, s ^ = H ( W T x - θ r ) , ( 3 ), where θr is the activation threshold for neurons in the reconstruction layer ., The reconstruction matrix W equals the measurement matrix A normalized to 1 by column , i . e . Wki = Aki/∑k Aki ., With this normalization , the reconstruction threshold θr = 1 corresponds to logical AND operation ., That is , odor component i will be detected as present if all glomeruli that feed signals to node i in the reconstruction layer are activated ., Below we will present most of the results for θr = 1 and then analyze how the reconstruction quality and recovery robustness depend on this threshold ., We will also determine the optimal connectivity ratio from the compression to the reconstruction layer that maximizes the fidelity of reconstructions ., Our feedforward model can be thought of as an information transmission channel that compresses , transmits , and decompresses a sparse binary signal ., To find the optimal network configuration , we seek to maximize mutual information between the input and output of the channel as has been done to characterize performance in the visual and other sensory systems ., The mutual information between s0 and s ^ is given by, I ( s 0 , s ^ ) = ∑ s 0 ∑ s ^ P ( s ^ | s 0 ) P ( s 0 ) log 2 P ( s ^ | s 0 ) P ( s ^ ) ., ( 4 ), For a given signal sparsity K , the conditional probability P ( s ^ | s 0 ) of the reconstructed signal s ^ given the original signal s0 can be computed as:, P ( s ^ | s 0 ) = p false N e r r ( 1 - p false ) ( N - K - N e r r ) , ( 5 ), where p false ≡ P ( s ^ i = 1 | s i 0 = 0 ) is the probability of false detection for an odor component and N e r r = | | s ^ | | 0 - K is the number of false detection events for the odor s0 ., We note that for θr = 1 , the probability to miss an odor component is zero provided this odor component activates at least one of the glomeruli ., In this regime , the information is fully determined by the false detection rate pfalse , and as we show below decreases proportionally with pfalse ., Assuming a uniform prior over individual odor components P ( s 0 ) = 1 / ( N K ) , one can also compute the probability distribution of reconstructed signals:, P ( s ^ ) = ∑ s 0 P ( s ^ | s 0 ) P ( s 0 ) = ( K + N e r r K ) ( N K ) p false N e r r ( 1 − p false ) ( N − K − N e r r ) ( 6 ), Putting together Eqs ( 4 ) – ( 6 ) , the mutual information can be written as, I ( s 0 , s ^ ) = log 2 ( N K ) − ∑ N e r r = 0 N − K ( N − K N e r r ) p false N e r r ( 1 − p false ) ( N − K − N e r r ) log 2 ( K + N e r r K ) ., When ( N − K ) pfalse ≪ 1 , the summation above can be well approximated by its leading nonzero term, ∑ N e r r = 0 N − K ( N − K N e r r ) p false N e r r ( 1 − p false ) ( N − K − N e r r ) log 2 ( K + N e r r K ) ≈ ( N − K ) p false log 2 ( K + 1 ) , ( 7 ), so that the expression for the mutual information becomes:, I ( s 0 , s ^ ) ≈ log 2 ( N K ) − ( N − K ) p false log 2 ( K + 1 ) ., ( 8 ), Thus , for given N and K , maximizing I ( s 0 , s ^ ) can be approximated by minimizing the probability of false detection pfalse ., The false detection rate that appears in Eq 8 can be computed as, p false ≡ P ( s ^ i = 1 | s i 0 = 0 ) = ∑ k = 1 M P ( s ^ i = 1 | | | T i | | 0 = k ) P ( | | T i | | 0 = k | s i 0 = 0 ) = ∑ k = 1 M 1 - ( 1 - p ) K k ( M k ) p k ( 1 - p ) M - k 1 - ( 1 - p ) M = 1 1 - ( 1 - p ) M ∑ k = 0 M 1 - ( 1 - p ) K k ( M k ) p k ( 1 - p ) M - k - ( 1 - p ) M 1 - ( 1 - p ) M = 1 - p ( 1 - p ) K M - ( 1 - p ) M 1 - ( 1 - p ) M , ( 9 ), where Ti ≡ {xk ∈ x|Aki = 1} , and p is the average connectivity rate from the compression to the reconstruction layer ., In the last line above we use the binomial expansion ., Because we are interested in the regime where M is large , we have ( 1 − p ) M ≪ 1 − p ( 1 − p ) KM ≪ 1 as long as p is not too small ., Thus , Eq 9 can be approximated with great accuracy by the following simple equation:, p false = 1 - p ( 1 - p ) K M ., ( 10 ), As shown in the inset of Fig 1B , Eq 10 provides an accurate approximation when the connectivity p is not too sparse ., Since our main interest is near the optimal connectivity rate ( see below ) where Eq 10 is very accurate , we will use Eq 10 unless specified ., As expected , the false detection rate pfalse decreases as the number of glomeruli M increases and as the signal sparseness K decreases ., Importantly , for a given M and K , there is an optimal p , which we refer to as pm , that minimizes pfalse , as shown in Fig 1B ., Taking ∂pfalse/∂p = 0 leads to, p m = 1 K + 1 ., ( 11 ), It is worth noticing that the optimal connectivity pm is independent of the number of glomeruli M , and depends only on the signal sparseness K . Thus , optimal connectivity depends exclusively on the level of sparseness of signals in the environment and can be determined prior to any measurements on neural circuits ., For an optimal connectivity p = pm , the probability of fault activation decreases exponentially as M increases and thus can be very small ., This indicates that the proposed feedforward compression-reconstruction scheme from Fig 1A can achieve exact recovery with high probability ., To test the reconstruction quality , we compute the signal-to-noise-ratio ( SNR ) of the recovered signal ., Since all nonzero components in the original will be recovered , the only source of errors in the reconstructed signal are due to false detection rates ., Therefore , we can define the SNR of recovered signal as, SNR = | | s 0 | | 0 < | | s ^ | | 0 > - | | s 0 | | 0 = K ( N - K ) p false , ( 12 ), as shown in Fig 2A–2C , where < ⋅ > denotes the expectation value ., We can see from Fig 2B that the SNR increases exponentially with M . For our case where K ≪ N , we can achieve a high SNR for a number of glomeruli M much smaller than the number of odor components N or , equivalently , the number of third-order neurons ., A key characteristic of a compression algorithm is the compression ratio α ≡ M/N ., In previous compressed sensing frameworks , the critical compression ratio αc above which the signal can be perfectly recovered was shown to only depend on the relative signal sparsity f ≡ K/N ., As f → 0 , αc ( f ) ∼ −f log f 12 ., To compute the critical compression ratio for our reconstruction algorithm , we note that from Eq 12 , log pfalse = log f − log ( 1 − f ) − log SNR ., In the strong compression limit where f ≡ K/N is small , this yields, log p false ≈ log f - log SNR ., ( 13 ), On the other hand , for the optimal connectivity rate pm and large K , log pfalse can also be simplified using Eq 10 as follows:, log p false = M log 1 - 1 K + 1 1 + 1 K - K ≈ M log 1 - 1 K + 1 e - 1 ≈ - M e K = - α SNR e f ., ( 14 ), where αSNR is defined as the compression rate to achieve a certain SNR ., Combining Eqs 13 and 14 , in the limit of strong compression where f → 0 , the critical compression ratio behaves as αSNR ∼ −f log f ., We note that care should be taken when the SNR becomes comparable to or larger than N because 1/f = N/K ≤ N , so that log SNR cannot be neglected when f → 0 ., The obtained critical compression rate can be compared to its theoretical limit ., The latter corresponds to the minimal number of bits required to encode a sparse signal:, M m i n = ⌈ log 2 ( N K ) ⌉ , ( 15 ), where ⌈x⌉ is the smallest integer not less than x ., When N and K are large but f ≡ K/N is small , using Stirling’s approximation , we obtain that, M m i n × log 2 ≈ N log N - K log K - ( N - K ) log ( N - K ) ≈ K log N - K log K + K = K - K log f , ( 16 ), This yields that the theoretically possible compression ratio αmin in the strong compression limit of f → 0 as, α m i n → f log 2 e / f , ( 17 ), which also yields αmin ∼ −f log f as f → 0 ., Notice that although both αSNR and αmin behave as −f log f for f → 0 , they have different proportionality coefficients ., To be more specific , αSNR ∼ ef log 1/f while αmin ∼ ( log 2 ) −1 f log 1/f ., As a result , αSNR/αmin → e log 2 ≈ 1 . 88 as f → 0 ., Thus , the number of glomeruli needed in our model is about twice the theoretical limit but is achieved here with an extremely simple feedforward encoding model ., As shown in Fig 2D , the number of required glomeruli increases sub-linearly with K , and logarithmically with SNR ., In practice , with only a few times more glomeruli than the theoretical limit , a very high SNR can be achieved ., Advances in experimental techniques provide opportunities to test our theory under the circumstances of extreme genetic manipulations ., For example , following a genetic manipulation that caused most olfactory receptor neurons to express a single odorant receptor M71 , the M71 ligand acetophenone activates half of the glomeruli ., Despite this drastic manipulation , mice can still readily detect other odors in the presence of acetophenone , while their discrimination performance is only moderately compromised 11 ., This result is consistent with our model ., Assume there are M glomeruli in our model and half of them are always turned on ( corrupted ) ., Such a system is equivalent to a model with only M/2 glomeruli , since the anomalously activated glomeruli will not affect signal recovery ., Thus , the odor signal can still be recovered , but the SNR is decreased , which is in agreement with the experimental result ., As a comparison , in previous compressed sensing framework , one can only allow a small percentage of corrupted glomeruli even when M > N 4 ., In another set of experimental studies , part of the glomeruli in mice are removed or disabled 13–15 ., It is shown that the ability to discriminate odors and simple odor mixtures is not impaired even when most of the glomeruli are removed or disabled ., This seemingly surprising finding is also consistent with our model ., From previous results , one can see that decreasing M will only lead to larger noise in the recovered odor signal but not to a failure of the system if the activation threshold for neurons in the reconstruction layer can be properly adapted to the new M . Assume the mice need SNR > ν to discriminate odors ., When K is small , the minimal M needed for discrimination is, M l o w = log K N ν log 1 - p ( 1 - p ) K ., ( 18 ), From experiment data , p ≈ 0 . 05 ( although this is a very rough estimation , see 11 , 16–18 ) ., One can check that the equation above is insensitive to variations in K and Nν over a broad range ., If we assume K < 10 ( as in the experiments ) and Nν is within the range of 104 ∼ 105 , then Mlow is roughly between 200 and 300 , or around 20% of the glomeruli , which is in good agreement with the data in those experiments ., On the other hand , our model can tolerate negative gloleruli noise ( false negative ) by changing its recovery threshold θr ., Although we use θr = 1 in our results for analytical solution , it is very likely that real biological systems would use a lower threshold θr ., With θr < 1 , the SNR is somewhat lower , as shown in Fig 3 , yet the system is more robust to noise in the reconstruction stage since the activation of a third-order neuron doesn’t require all of its connected gloleruli to be active and it also leaves room for odor generalization and pattern completion 19 ., Indeed , when the threshold at the reconstruction stage is less than 1 , the reconstruction can tolerate some incompleteness in the glomeruli activation patterns ., Real biological systems likely have the ability to adaptively change the activation threshold in order to balance the needs of high quality reconstruction and pattern completion ., Our model is shown to be very robust and fault tolerant , and this robustness is achieved with accuracy ., As one can see , each glomerulus in the model only contains part of the information about the original signal ., Because the measurement matrix A is random , no single glomerulus or cluster contains more or unique information , so any subset of the glomeruli could recover the original signal ., The more glomeruli there are , the better recovery quality ( SNR ) can be achieved ., Thus , removing or disabling part of the glomeruli will not change the system qualitatively , but will make the recovered signal more noisy , up to a point where noise becomes comparable to the true signal at which point the reconstruction fails ., For a real biological system , it is reasonable to assume that the recovered signal has very high SNR , which also means high redundancy , as is observed experimentally ., From our analysis we observed that for a given level of signal sparseness K , there is an optimal connectivity rate pm that maximizes SNR as well as the mutual information ., Assuming that the biological system is adapted to a given value of odor sparseness in its environmental niche , one can essentially make predictions on the connectivity rate of matrix A . This is followed by another prediction that the percentage of glomeruli activated by a single odorant should be close to the percentage of glomeruli that could activate a neuron in olfactory cortex or a Kenyon cell , and this number should be similar among species which operate in similar olfactory environments ., The latter prediction should be easier to test , since the number of coexisting odorants in the environment is hard to measure ., Fortunately , previous experiments have gathered sufficient data to test our prediction indirectly ., It has been shown that in Drosophila , 9% of the glomeruli have a strong response to an odorant 20 , while the connectivity rate between glomeruli and Kenyon Cells is 6 . 5% 21 to 12 . 5% 22 ., ( The latter number is obtained based on the average number of claws per Kenyon cell measured in 22 ), These estimates are consistent with model predictions ., Furthermore , in the locust , a typical projection neuron responds to about half of the odorants 23 , while the connectivity rate between projection neurons and Kenyon Cell is also around 50% 24 , which is also consistent with our prediction ., We can see that the connectivity rate is very different between species ., Such differences can be unified in our model as the adaptation to different environmental niches ., The locust has an anomalously high connectivity rate ( 50% ) , which in our model implies that its olfactory system is adapted to extreme odor sparseness tuned to odors with primarily a single component ( pm = 0 . 5 when K = 1 ) ., Similarly , Drosophila is adapted to sense odors composed of a mixture of about 10 odor components , while mice are tuned to detect a mixture of about 20 mono-molecular odors ., In general , our model predicts that species with sparse connectivity will behave better in environments with complex odor mixtures , while species with dense connectivity have better performance in detecting simple odor mixtures ., In addition to the predictions above , further experimental evidence supports the structure of our model , in particular the approximate logical OR/AND operations associated with the compression/reconstruction stages , respectively ., For example , it has been observed experimentally that Kenyon Cells in Drosophila receive convergent input from different glomeruli and require several inputs to be co-active to spike 25 ., This is consistent with our threshold activation function which at the reconstruction stage uses a logical AND operation ., Functionally , experiments have shown that locust Kenyon cells are individually much better than projection neurons from glomeruli at detecting a single odorant; Kenyon cells that respond to an odorant also often respond to odor mixtures containing it 26 ., This observation agrees with our assumption that each Kenyon cell only responds to one odorant and it will respond when an odor mixture contains that odorant ., Since the affinity matrix A is determined genetically , all the connections in our model are predetermined before birth ., There is some debate about such stereotypy versus random connectivity , and a compressed sensing model of olfaction based on random connections from glomeruli to mushroom body has been proposed 27 ., Yet , our model supports both stereotyped and non-stereotyped projection from glomeruli to the mushroom body/olfactory cortex because the model is invariant under the exchange of neurons within the same layer ., In order to verify such predetermination , one needs to obtain a detailed connectivity map from glomeruli to the mushroom body/olfactory cortex for different individuals , which is experimentally very challenging ., An indirect approach to verify the predetermined connectivity hypothesis could be through an examination of innate behaviors that should depend primarily on predetermined connections ., If one could relate innate behaviors to projections between glomeruli and the mushroom body/olfactory cortex , it would then provide additional supporting evidence for the genetically predetermined structural connectivity of the feedforward model ., The feedforward structure of our model is an effective approximation to the more complicated structure of biological olfactory system where recurrent and feedforward-feedback connections exist ., For example , it has been observed that inhibitory interneurons modulate neuronal responses in the olfactory bulb 28 , 29 ., In linear dynamic systems , such feedforward-feedback structure could be mathematically modeled as a pure feedforward system with different effective feedforward connectivity ., Suppose that we add a layer of interneurons z in Fig 1 that is connected to the glomeruli layer x by feedforward-feedback connectivity B . Then the linear dynamics of the system are x ˙ = - x + A s 0 - B T z and z ˙ = - z + B x , where we assume B is feedforward excitatory and feedback inhibitory ., The steady state solution is x = ( I + BT B ) −1 As0 , which is the same for a pure feedforward system , except that connectivity A is replaced by ( I + BT B ) −1 A . This analysis is not exact if the activation function is nonlinear ., In general , the feedforward-feedback system in steady state with a nonlinear activation function does not have an equivalent feedforward system , but one can still write the linear perturbation when neurons receive only weak inputs , which allows a feedforward approximation ., Such a feedforward approximation is supported by experimental observations that the representations of odor mixtures in mouse glomeruli can be explained well by the summation of the glomeruli responses to their components 30 ., One advantange of the effective feedforward model is that it enables an adaptive affinity matrix even with pre-determined connectivity ., In the feedforward-feedback architecture mentioned above , the effective affinity matrix is ( I + BT B ) −1 A , where A is the pre-determined affinity matrix encoded in the genes , while B could be a learned matrix adapted to the environment ., From this perspective , the existence of interneurons in both insects and vertebrates 31 , 32 , as well as adult neurogenesis in the olfactory bulb of mammals 33 , could play the role of adjusting the effective affinity matrix for the purpose of adaptation ., We compare the performance of our feedforward architecture with the often-used LASSO ℓ1 minimization algorithm 34 provided by the Python scikit-learn library, min s ^ 1 2 M | | A s ^ - x | | 2 2 + β | | s ^ | | 1 , ( 19 ), where N = 1000 , M = 500 , β = 0 . 001 are used ., Linear measurement x = As0 is used for LASSO ., For each K , we conduct 100 experiments with different random measurement matrices and signals , and compute the average of the reconstruction errors | | s ^ - s 0 | | 1 as well as the number of iterations used in LASSO ., We also compute the mean reconstruction error when only 5 iterations are used in LASSO as a comparison ., The results are shown in Fig 4 ., As shown in the figure , the feedforward architecture has a lower reconstruction error when the signal is very sparse , while LASSO has a lower reconstruction error than the feedforward architecture when K becomes larger ., However , the number of iterations also increases as the signal becomes denser ., If we restrict the number of iterations to 5 in the LASSO ( equivalent to setting a maximum response time ) , LASSO performs much worse when the signal is very sparse ., But as K increases , it still has a lower reconstruction error than the feedforward architecture ., One drawback of this feedforward architecture is that it may not be able to achieve both compression and high-quality reconstruction simultaneously when the signal is not sparse ., Unlike the ℓ1 minimization method where the number of measurements required to reconstruct the signal will never exceed signal length N ( N/2 for binary signal ) 35 , 36 , the feedforward architecture may need more measurements than the signal length to accurately reconstruct the signal ., This can be seen by restoring the term in Eq 13 that we have previously neglected assuming that f is small, log p false = log f - log ( 1 - f ) - log SNR ., ( 20 ), Combining this with Eq 14 that remains the same when f is not small , we obtain:, α SNR = e f log SNR + e f log ( f - 1 - 1 ) , ( 21 ), which could be larger than 1 when f is not small ., Thus , the feedforward computation may require number of measurements that are larger than the input dimensionality to achieve reliable reconstruction ., From another perspective , we can compute the upper bound on the reconstruction SNR that can be achieved for a given compression level ., From Eq 21 and αSNR < 1 we get, log SNR < 1 e f - log ( f - 1 - 1 ) , ( 22 ), which only depends on signal sparsity ., For example , if f = 0 . 1 , then SNR < 4 . 4 , and the reconstructed signal will not be accurate ., Although our analysis above is based on a binary signal / measurement matrix / glomeruli activity and threshold activation function , our results can be extended to positive real-valued signal / measurement matrix / glomeruli activity and any monotonically increasing activation function ., Consider the case where the signal s0 and the element of measurement matrix Aij could take any positive value rather than just 0 and 1 ., Denoting xl = As0 , and letting the activation function g be any monotonically increasing function , the output at the glomerulus stage can be written as x i = g ( x i l ) ., Now , signal reconstruction can proceed based on the evaluation of a minimum function ( rather than the logical AND function that was used in the case of binary inputs and binary measurement matrices ) ., Indeed , when the ith component of the reconstructed signal s ^ i is computed as the smallest value {g−1 ( xj ) /Aji} across the set of its inputs ( i . e . where Aji ≠ 0 ) , then our analysis remains valid ., The only modification is that now the distribution of the signal and the measurement matrix elements are both required to compute the noise magnitude ., This procedure ensures that the recovered components are still recovered exactly , while corrupted components are still corrupted ., As a practical aside , we note that the minimum function can be implemented by short-term synaptic plasticity , see S1 Text and S1 and S2 Figs . | Introduction, Models and Methods, Results, Discussion | Most natural odors have sparse molecular composition ., This makes the principles of compressed sensing potentially relevant to the structure of the olfactory code ., Yet , the largely feedforward organization of the olfactory system precludes reconstruction using standard compressed sensing algorithms ., To resolve this problem , recent theoretical work has shown that signal reconstruction could take place as a result of a low dimensional dynamical system converging to one of its attractor states ., However , the dynamical aspects of optimization slowed down odor recognition and were also found to be susceptible to noise ., Here we describe a feedforward model of the olfactory system that achieves both strong compression and fast reconstruction that is also robust to noise ., A key feature of the proposed model is a specific relationship between how odors are represented at the glomeruli stage , which corresponds to a compression , and the connections from glomeruli to third-order neurons ( neurons in the olfactory cortex of vertebrates or Kenyon cells in the mushroom body of insects ) , which in the model corresponds to reconstruction ., We show that should this specific relationship hold true , the reconstruction will be both fast and robust to noise , and in particular to the false activation of glomeruli ., The predicted connectivity rate from glomeruli to third-order neurons can be tested experimentally . | Many olfactory systems are capable of accurately sensing a minimum of thousands of different odorants using as few as hundreds of different receptors ., This compression raises the possibility that the mathematical properties of compressed sensing might be relevant to olfaction , similar to how these properties were found relevant to other sensory systems ., In olfaction , previous applications of compressed sensing algorithms relied on the dynamics of neural circuits to reconstruct high dimensional signals ., Such approaches are relatively temporally inefficient and sensitive to noise ., To overcome these problems , we propose a purely feedforward compressed sensing model of the olfactory system where high dimensional signals can be recovered with a single feedforward layer of neural processing ., The reconstructions are shown to be robust to noise , account for a number of experimental observations , and because of the feedforward structure are temporally efficient ., Using the model , we make predictions that can be tested in future experiments with respect to optimal connectivity within the olfactory system ., Our results indicate that feedforward neural architectures can provide an efficient way to implement compressed sensing in neural systems . | invertebrates, medicine and health sciences, brain, vertebrates, neuroscience, animals, animal models, optimization, drosophila melanogaster, model organisms, mathematics, odorants, materials science, drosophila, research and analysis methods, sensory physiology, animal cells, materials by attribute, olfactory receptor neurons, insects, arthropoda, piriform cortex, cellular neuroscience, cell biology, anatomy, physiology, olfactory system, neurons, biology and life sciences, sensory systems, physical sciences, cellular types, afferent neurons, organisms | null |
journal.pcbi.1000691 | 2,010 | Temporal Sensitivity of Protein Kinase A Activation in Late-Phase Long Term Potentiation | Synaptic plasticity , the activity-dependent change in the strength of neuronal connections , is a cellular mechanism proposed to underlie memory storage ., One type of synaptic plasticity is long term potentiation ( LTP ) , which typically is induced by brief periods of high-frequency synaptic stimulation ., LTP displays physiological properties suggestive of information storage and has been found in all excitatory pathways in the hippocampus , as well as other brain regions ., Late-phase LTP ( L-LTP ) is induced by 4 trains of stimulation separated by either 3–20 sec ( massed ) or 300–600 sec ( spaced ) , lasts more than 3 hours , and requires protein synthesis 1 ., Interestingly , the temporal spacing between successive trains regulates the PKA-dependence of L-LTP 2 , 3 ., A spaced protocol ( using a 300 sec inter-train interval ) requires PKA , whereas massed protocols ( using 20 sec and 3 sec intervals ) induce L-LTP that is independent of PKA ., The mechanisms underlying this temporal sensitivity of PKA dependence are not understood ., PKA is composed of two regulatory subunits bound to two catalytic subunits that form a tetrameric holoenzyme ., Sequential and co-operative binding of four cAMP to these regulatory subunits results in the release of two catalytic subunits 4 , 5 ., In the hippocampus , cAMP is produced by adenylyl cyclase types 1 and 8 , which are activated by calcium and Gsα coupled receptors 6 ., Consistent with this pathway of reactions leading to PKA , activation of dopaminergic and glutamatergic pathways is required for the induction of L-LTP in hippocampal CA1 pyramidal neurons 7–11 ., NMDA receptor activation also leads to stimulation of the calcium sensitive isoform of adenylyl cyclase 12 ., Because the induction of L-LTP involves complex networks of intracellular signaling pathways , computational models have been developed to gain an understanding of LTP 13–17 ., Several of these studies , which specify the model using ordinary differential equations , explain the requirement for high frequency stimulation ( e . g . 100 Hz for LTP ) versus low frequency stimulation ( e . g . 1 Hz for long term depression ) in terms of the characteristics of CaMKII 18–21 ., Even though PKA has been incorporated in some of these models , PKA activation is typically described using simplified algebraic equations 21–23 ., These models do not include the role of dopamine or β-adrenergic receptors in PKA activation nor adequately describe the temporal dynamics of PKA activation ., Consequently , these models do not evaluate the temporal sensitivity of PKA , and cannot accurately explain why PKA is required for spaced stimulation ., In contrast , several models by Bhalla 24 , 25 include not only the signaling pathways leading to PKA activation , but also those for mitogen activated protein kinase ( MAPK ) activation ., However , Bhalla did not explore the role of dopamine or PKA in late-phase LTP , and we have utilized more recent experimental data to update several of the reactions , especially those involved in PKA activation ., To evaluate the biochemical mechanisms underlying the temporal sensitivity of PKA dependence of L-LTP and the role of dopamine , we developed a single compartment model of postsynaptic signaling pathways underlying L- LTP in CA1 pyramidal neurons of the hippocampus ., Reaction rates and pathways are based on published biochemical measurements ., Simulations explore the mechanisms underlying temporal sensitivity of LTP to PKA and complementary experiments test the model predictions of the critical temporal interval separating PKA-dependent and PKA-independent LTP ., Simulation results show that the activation of PKA is greater with spaced as compared to massed stimulation ( Fig 2A1 ) ., These results are consistent with experimental results 3 showing that PKA is required for spaced , but not massed stimulation ., The cumulative activity of PKA with spaced stimulation ( 2321 nM-sec ) is 60% greater than with massed stimulation ( 1455 nM-sec ) ., Although the massed protocol produces a higher peak PKA activity , it is not 4 times higher than the peak produced from a single spaced train of stimulation because of sub-linear summation: the PKA peak activity for massed stimulation is only 1 . 4 times higher than the peak activity in response to spaced stimulation ( Fig 2A2 ) ., Subsequent trains do not increase the peak activity of PKA , but do contribute to cumulative PKA activity over time by linear summation; therefore more PKA activity is available with spaced stimuli ., Simulations are repeated for a range of inter-train intervals to further explore the temporal sensitivity of PKA dependence ., Fig 2C shows that cumulative PKA activity increases with temporal interval , with a time constant , τ , of 8 . 5 sec ., PKA activity reaches 95% of maximal value within 3 time constant , i . e . , at 25 . 5 sec ., This temporal sensitivity is not observed if peak activity is evaluated ., Activity at a single time point , such as 10 minutes after stimulation , is often used to compare with experimental measurements that measure enzyme activity at a single time point ., Nonetheless , cumulative activity better indicates the ability of an enzyme to act on downstream targets ., Using single time point measures of activity may explain why a previous study did not observe temporal sensitivity of PKA ., This increase in PKA activity with increasing inter-train interval can partly explain the mechanism of temporal sensitivity of PKA dependence of L-LTP , but the other part of the explanation is likely a deficit in some other molecule , such as CaMKII , which is known to be sensitive to higher frequency stimuli and plays a major role in LTP ., Thus , levels of phosphorylated CaMKII were examined for 3 sec and 300 sec inter-train intervals to assess whether PKA dependence was related to a decline in phosphorylated CaMKII with longer inter-train intervals ., This peak was evaluated because experiments suggest that phosphoCaMKII anchors at the post-synaptic density ( PSD ) and is not accessible to dephosphorylation by protein phosphatase 1 28 ., This would imply that activity would be proportional to peak value , and the resulting slow decay of phosphoCaMKII precludes a reasonable calculation of the area under the curve ., Fig 2B shows that peak activity of phosphorylated CaMKII with 300 sec intervals is lower than with 3 sec intervals , which is opposite to the temporal sensitivity of PKA , suggesting that PKA activity is compensating for a frequency-dependent deficit in CaMKII ., To further compare the CaMKII temporal sensitivity with the PKA temporal sensitivity , Fig 2C explores the phosphorylated activity of CaMKII for a range of inter-train intervals ., PhosphoCaMKII decreases as temporal interval increases ( beyond 3 sec ) , in agreement with experiments 29 ., The time constant of this decrease is 20 . 8 sec , and phosphoCaMKII drops to 95% of its peak value with a 62 sec inter-train interval ., The sum of ( normalized ) phosphoCaMKII and PKA activity is independent of interval for all but the very shortest intervals suggesting that PKA is required for spaced stimulation to compensate for a decrease in CaMKII ., This result leads to the prediction that PKA will be required for inter-train intervals greater than ∼62 sec ., The prediction that PKA is required for intervals greater than ∼62 sec was tested by inducing L-LTP at Schaffer collateral-CA1 synapses in mouse hippocampal slices using 4 trains of high frequency stimulation , with either 40 sec or 80 sec inter-train intervals , in the presence of either KT5720 or vehicle as control ., As shown in Fig 3A , LTP induced by stimulation trains delivered at 80 sec inter-train intervals was attenuated in KT5720-treated slices compared to vehicle controls ., At 120 min after LTP induction , the average fEPSP slopes were significantly different: 196±11% for vehicle-treated slices and 112±7% for KT5720-treated slices ( Mann-Whitney U test , p<0 . 05 ) ., This demonstrates that LTP induced by 4 trains of high frequency stimulation delivered at 80 sec inter-train intervals requires PKA ., In contrast , fEPSP slopes are not significantly different between KT5720 and control slices using 40 sec inter-train intervals ( Fig 3B ) ., At 120 min after LTP induction , the average fEPSP slopes were 167±14% for vehicle-treated slices and 167±13% for KT5720-treated slices ( Mann-Whitney U test , p>0 . 05 ) ., This indicates that LTP stimulated by 4 trains of high frequency stimulation delivered at 40 sec inter-train interval is PKA-independent ., These results , and previous experimental results on PKA dependence 3 , are summarized in Fig 3C , which demonstrates that L-LTP induced with temporal intervals of 3 sec to 40 sec are PKA-independent , whereas L-LTP induced by temporal intervals of 80 sec and 300 sec are PKA-dependent ., These experiments support the model prediction , thus verifying the model and its explanation of mechanisms underlying PKA dependence ., In the hippocampus , adenylyl cyclase type 1 is synergistically activated by both calcium-calmodulin and dopamine , which is released during 100 Hz stimulation 30 from fibers innervating hippocampal area CA1 31 ., Further support for the role of dopamine is provided by experiments that show that L-LTP induced using a 10–12 min inter-train interval is reduced when dopamine receptors are blocked 8 , 30 , 32 ., Thus , simulations were repeated with the dopamine receptor blocked , to evaluate the contribution of dopamine to L-LTP ., Fig 4 shows that cumulative PKA activity is reduced significantly with both massed and spaced stimulation intervals when dopamine receptor function is blocked ., The PKA activity for a 300 sec inter-train interval with no dopamine is similar to the PKA activity for the 3 sec inter-train interval with dopamine present , suggesting that L-LTP induction with spaced stimuli requires the higher PKA produced by spaced stimuli ., Though the lack of dopamine reduces PKA activity for the 3 sec inter-train interval , this is not functionally significant because L-LTP with massed stimulation is PKA-independent ., In other words , a 300 sec inter-train interval activates insufficient quantities of CaMKII , and additional dopamine stimulated PKA activity is required for the 300 sec interval only ., Stimulation with a 3 sec interval activates sufficient CaMKII , and thus , the model predicts that blocking dopamine receptors would not block L-LTP for this interval ., The sensitivity of cumulative PKA activity to different temporal intervals follows that of adenylyl cyclase ( Fig 5A ) and cAMP ( Fig 5B ) ., The first 100 Hz train produces a 600 nM increase in adenylyl cyclase activity from binding to calmodulin and Gsα ( Fig 5A2 ) ., With the massed protocol , the second 100 Hz train only produces an additional 300 nM increase in adenylyl cyclase activity , because free adenylyl cyclase is depleted with massed trains to a significant degree ., More than 80% of unbound adenylyl cyclase 1 is available for activation by the first train of stimulation ( Fig 5C ) ; unbound adenylyl cyclase 1 decreases by 20% for massed ( Fig 5C1 ) , but remains at more than 80% for spaced stimulation ( Fig 5C2 ) ., Calmodulin , which activates adenylyl cyclase 1 , also exhibits a small degree of depletion , in part because it binds to other molecules , such as protein phosphatase 2B and phosphodiesterase 1B , with extremely high affinity ., Thus , subsequent stimulation trains produce smaller increments in activated adenylyl cyclase for massed , but not for spaced stimulation ., These lower adenylyl cyclase activity increments result in lower cAMP increments with subsequent trains using massed stimulation: 300 nM for the first train and 150 nM for the second train ( Fig 5B2 ) ; thus the total cAMP produced from four trains of stimulation is less than four times the cAMP produced for one train ., Note that the temporal pattern of cAMP , which decays within 40 sec to basal levels , agrees with measurements using a fluorescent Epac-1 probe 33 , 34 , verifying this aspect of the model ., Therefore , the activation of PKA is greater with spaced as compared to massed stimulation because adenylyl cyclase activity is greater with spaced as compared to massed stimulation ., PKA is important in LTP because it phosphorylates AMPA receptors and inhibitor-1 , as well as other plasticity related proteins 35–38 , not all of which have been identified ., Because rates of AMPA receptor phosphorylation have not been directly measured , we chose to evaluate the effect of PKA activity on a different target , namely inhibitor-1 ., Furthermore , inhibition of protein phosphatase 1 by phosphorylated inhibitor-1 will enhance phosphorylation of many PKA targets via inhibition of dephosphorylation ., Thus , examination of the phosphorylation state of inhibitor-1 in these simulations both represents the ability of PKA to phosphorylate downstream targets , and also indicates whether free protein phosphatase 1 will be sensitive to temporal interval ., As seen in Fig 6A , the amount of phosphorylated inhibitor-1 is 50% greater for spaced than massed stimulation ., Similar to that observed with PKA activity , the peak value is higher for massed stimuli , but total phosphorylated inhibitor-1 is greater for spaced stimuli ., This shows that the temporal sensitivity of PKA activity propagates to downstream targets ., The phosphorylated inhibitor-1 binds to protein phosphatase 1 with high affinity , inhibiting its activity ., Thus , the 50% increase in phosphorylated inhibitor-1 produces a 50% decrease in protein phosphatase 1 ( Fig 6B ) ., This suggests that the enhanced activity of PKA with spaced stimulation will suppress protein phosphatase 1 activity , reinforcing the phosphorylation of plasticity related proteins ., To test whether the enhanced inhibitor-1 phosphorylation increased CaMKII phosphorylation , simulations were repeated with PKA phosphorylation of inhibitor-1 blocked ., The decrease in CaMKII phosphorylation was small ( Fig S2A ) , suggesting other mechanisms to enhance PKA activity are important ( discussed below ) ., To investigate the robustness of results ( i . e . , whether the results are sensitive to variation in parameters ) , simulations are repeated using parameter values 2 to 10 times larger or smaller than the control values , for parameters that are least constrained by biochemical data ., For instance , though the quantity of PKA has been estimated to be 1 . 2 µM in brain tissue , assuming the protein distributes in 70% of intercellular space 39 , the existence of localized pools of PKA suggest that the effective quantity of this enzyme in the synapse could be higher than the estimated quantity ., Similar arguments can be made for protein phosphatase 1 ., Thus , simulations are repeated using both higher and lower quantities of PKA , protein phosphatase 1 , protein phosphatase 2B , as well as Ca2+ influx ., As shown in Fig S3 , the main results from this model are qualitatively robust ., Though the PKA activity increases when enzyme quantities are increased , spaced stimulation still produces ∼60% more total activity than massed stimulation ( Fig S3A ) ., The quantity of protein phosphatase 1 has no effect on PKA activity , but does modify the decay rate of phosphoCaMKII ., Regardless of protein phosphatase 1 quantity or dephosphorylation rate , spaced stimulation produces lower phosphorylated CaMKII than massed stimulation ( Fig S3B ) ., Peak Ca2+ has a different effect on phosphoCaMKII: it changes the peak value with no change in decay , and no change in frequency sensitivity ( Fig S3D ) ., PKA was minimally affected by variation of peak Ca2+ ( Fig S3C ) ., A recent FRET imaging experiment suggests that CaMKII activity in spines is transient in response to synaptic stimulation 40 ., Thus , additional simulations evaluated whether the results are sensitive to persistence of CaMKII ., Transient phosphoCaMKII was produced by allowing protein phosphatase 1 to dephosphorylate the calmodulin bound form of phosphoCaMKII ( Fig S2A ) ., Using this more transient phosphoCaMKII in simulations , CaMKII activity is quantified as area under the curve ( instead of peak ) ., Fig S2B shows that area under the curve increases for PKA and decreases for phosphoCaMKII with increasing inter-train interval , the latter with a time constant of 17 . 8 sec – close to the time constant for the persistent model of CaMKII ., Thus , the prediction that PKA is required to compensate for a decrease in phosphoCaMKII is robust to this variation in CaMKII dynamics ., To better understand the complex intracellular signaling networks underlying the temporal sensitivity of PKA dependence of L-LTP , we developed a computational model of the calcium and cAMP signaling pathways involved in PKA and CaMKII activation in hippocampal CA1 neurons ., The model is based on published biochemical measurements of many key signaling molecules , most notably PKA and CaMKII ., Simulations of four trains of 100 Hz stimuli separated by 300 sec or 3 sec revealed that spaced stimulation activates more PKA and less CaMKII than massed stimulation ., Thus , PKA activity may be required for spaced stimulation because more of it is active , and less phosphoCaMKII is available ., Simulations were repeated for a range of inter-train intervals , to further explore the PKA dependence of L-LTP induction ., PKA activity increases exponentially with increasing inter-train interval , compensating for the decrease in phosphoCaMKII with increasing inter-train interval ., The time constant of phosphoCaMKII decrease was 20 . 8 sec; thus , the model predicts that L-LTP induced with an inter-train interval greater than 62 sec ( 3τ ) will be dependent on PKA , and L-LTP induced with an interval less than 62 sec will be independent of PKA ., Experiments confirm this prediction , showing that a 40 sec inter-train interval is PKA-independent and an 80 sec inter-train interval requires PKA ., The temporal sensitivity of PKA differs from that in a previously published model 29 mainly due to the different method of quantifying PKA activity ., The present study measured cumulative PKA activity as area under the curve and found an increase with temporal interval ., In the previously published model , PKA activity was quantified as the peak activity at 600 sec after the last tetanus , to compare with experimental measurements which also measured activity at 600 sec after the last tetanus ., In that study , PKA peak activity did not exhibit temporal sensitivity , and thus could not explain the temporal sensitivity of PKA dependence of LTP ., To compare the present model results with that previous model , PKA activity was quantified as activity at 600 sec after the last tetanus ., Using this quantification , temporal sensitivity of PKA activation in the present model is minimal , in agreement with Ajay and Bhalla 29 ., Nonetheless , cumulative activity is a better measure of the ability of a kinase to phosphorylate downstream substrates such as AMPA receptors or inhibitor-1 , because cumulative activity is proportional to average enzyme activity over the time course of the enzyme ., With regards to CaMKII activity , both cumulative , when CaMKII phosphorylation is transient , and peak when CaMKII phosphorylation is persistent , were good predictors of the critical inter-train interval ., Another PKA-dependent form of L-LTP is induced by theta-burst stimuli 41 , which uses short bursts of 100 Hz stimulation ( e . g . , 4 pulses ) repeated at 200 msec intervals ., A typical experimental induction protocol uses fifteen repetitions of 4 bursts yielding 60 pulses total , far less than provided with 4 bursts of 100 Hz ., Model simulations show that both CaMKII and PKA are lower with this stimulation due to the lower number of pulses ., These simulations of post-synaptic mechanisms cannot explain the PKA-dependence of theta-burst L-LTP because theta-burst L-LTP involves pre-synaptic mechanisms 42 ., In our model , activated PKA is represented as the cumulative quantity of the free catalytic subunit ., Although stimulation produces about a 60% increase in free catalytic subunit , the peak quantity of free catalytic subunit is relatively small ( less than 50 nM for massed stimulation and less than 35 nM for spaced stimulation ) ., This may suggest that the quantity of free catalytic subunit would be insufficient for the PKA-dependent L-LTP ( i . e . , both the increase in inhibitor-1 phosphorylation , and the inhibition in CaMKII dephosphorylation were small ) , especially given the number of PKA targets ., The small quantity of PKA free catalytic subunit produced is due to the high affinity ( 9 nM ) of the regulatory subunit for the catalytic subunit even when all four cAMP molecules are bound ., One possible solution to the low quantity of PKA catalytic subunit is that the cAMP-saturated holoenzyme is catalytically active toward its substrates ., Binding of four cAMP to the linker region of the regulatory subunit causes a conformational change , exposing the catalytic site without complete dissociation 43 , 44 ., The L-LTP induction paradigms produce a significant amount of cAMP-saturated holoenzyme ( twice as much as free catalytic subunit ) ., If this form is active , the quantity of active PKA would be three times higher ., In addition , the actions of anchoring also increase local PKA activity in the synapse ., A kinase anchoring proteins ( AKAPs ) bind to the regulatory subunit of the PKA holoenzyme 45 ., By tethering the PKA holoenzyme near a preferred substrate at a particular subcellular location , a small number of molecules could produce significant phosphorylation of its substrate ., In support of this concept , experiment shows that hippocampal synaptic plasticity requires not only PKA activation , but also the activation of an appropriately anchored pool of PKA 42 , 46 ., As previously mentioned , the conceptual model of CaMKII activation predicts a positive feedback loop in which increased phosphorylation leads to an increased rate of subsequent phosphorylation ., Therefore , subsequent stimulus trains should produce increasing increments in CaMKII activity ., Similar to other single compartment models 19 , 20 , 47 , this positive feedback response is not observed in the model unless additional calmodulin is provided ( Fig S4A ) ., Calmodulin binds with high affinity to protein phosphatase 2B and phosphodiesterase 1B , and with intermediate affinity to adenylyl cyclase as well as CaMKII ., This binding causes a decrease in Ca4-calmodulin with subsequent trains due to competition for calmodulin between the CaMKII pathway and other pathways ., Calmodulin is a diffusible protein; thus , in a dendritic spine free calmodulin would diffuse into the spine from the dendrite to replace the bound calmodulin ., In addition , neurogranin is a calmodulin binding protein that releases calmodulin upon Ca2+ stimulation; in essence neurogranin acts as a calmodulin reservoir 48–50 ., Simulations in which additional calmodulin is provided yields a frequency sensitivity of phosphoCaMKII that agrees with experimental measurements 29 ., Not only CaMKII , but also PKA activation is limited by free available calmodulin , since the predominant adenylyl cyclases ( 1/8 ) in hippocampus are activated by calmodulin ., Calmodulin depletion results in decreasing increments of adenylyl cyclase activity , cAMP production , and PKA activation with massed stimulation , causing sublinear summation ., Providing additional calmodulin reduces the degree of sublinear summation , though the limited quantity of adenylyl cyclase 1 and adenylyl cyclase 8 also contributes to sublinear summation ., Thus , as illustrated in Fig S4B , the incorporation of additional calmodulin does not change the main result , namely that PKA cumulative activity is higher with spaced stimulation ., One way in which LTP is expressed post-synaptically is as enhanced phosphorylation of AMPA receptors leading to insertion of new AMPA receptors ., The phosphorylation state of AMPA receptors depends on the balance of kinases and phosphatases including PKA , CaMKII and protein phosphatase 1 51–53 ., Active PKA directly phosphorylates the AMPA receptor GluR1 subunit at Ser845 , enhancing AMPA channel function 54 and leading to increased AMPA channel expression ., PKA indirectly governs the dephosphorylation activity of protein phosphatase 1 by phosphorylating inhibitor-1 with very high affinity allowing it to bind protein phosphatase 1 ., Other substrates of PKA are implicated in hippocampal synaptic plasticity , including phosphodiesterase type 4D3 and inositol triphosphate receptor channels 55 , 56 ., AMPA channel phosphorylation modulates expression of LTP , but transcription and translation are required for L-LTP 57 ., A target of phosphorylation by active PKA involved in transcription is the cAMP Response Element Binding Protein ( CREB ) in the nucleus ., Phosphorylated CREB increases activation of transcription and protein translation ., Members of the mitogen activated protein kinase ( MAPK ) family are targets of PKA that plays a role in transcription , translation , and synaptic plasticity 58 ., One member of the MAPK family is extracellular signal-regulated kinase type II , which is phosphorylated by several signaling pathway kinases , such as PKA and also CaMKII through synGAP 59 , 60 ., Ajay and Bhalla 29 demonstrate that both extracellular signal-regulated kinase type II activity and the magnitude of LTP induction are maximal using inter-train intervals of 300–600 sec; in this context , our results suggest that part of the temporal dependence of extracellular signal-regulated kinase type II is due to PKA ., Yet another target of PKA involved in maintenance of LTP is the atypical protein kinase C , type Mζ , which is phosphorylated at a site of convergence of both PKA and CaMKII 61 ., Thus , our hypothesis that the combination of CaMKII plus PKA is critical for L-LTP is consistent with several of these target proteins whose activity integrates multiple kinases ., Additional evidence suggests that PKA is critical for synaptic tagging 46 , 62 , 63 , which provides the synaptic specificity important for information processing ., The synaptic tag theory proposes that L-LTP associated gene products can only be captured and utilized at synapses that have been tagged by previous activity 64 ., Both CaMKII and PKA have been implicated in phosphorylation of an unidentified synaptic substrate , which appears necessary to set a tag at activated synapses to allow capture of plasticity factors ( i . e . CRE-driven gene products , newly synthesized AMPA receptors or mRNAs ) ., One possibility is that phosphorylation of the tag can be provided by either CaMKII , PKA , or both , depending on the temporal interval of stimulation ., To further evaluate L-LTP , it will be necessary to include some of these signaling events downstream of PKA , such as activation of extracellular signal-regulated kinase type II ., Furthermore , anchoring of proteins in spines , communication with the larger dendrites , and other spatial details all suggest that single compartment models are not sufficiently accurate ., Thus , multi-compartmental models will be critical for evaluating issues such as the distribution of synaptic inputs underlying the spread of biochemical signals from synapses to dendrites 25 or the diffusion of biochemical signals between spines 65 ., For example , preliminary simulations using a multi-compartmental stochastic model suggest that localization of dopamine receptors and PKA leads to larger phosphorylation of inhibitor-1 , and inhibition of protein phosphatase 1 , as experimentally observed 35 ., Given the complexity of non-linear interactions among signaling pathways , simulations using these novel multi-compartmental models promise to enhance understanding of the mechanisms underlying synaptic plasticity ., All research with animals was consistent with NIH guidelines and approved by the IACUC at the University of Pennsylvania ., The single compartment , computational model , illustrated in Fig 1A , consists of signaling pathways known to underlie synaptic plasticity in hippocampal CA1 pyramidal neurons ., Calcium influx through the NMDA receptor leads to calcium-calmodulin activation of adenylyl cyclase types 1 and 8 66 , phosphodiesterase type 1B , protein phosphatase 2B ( PP2B or calcineurin ) and CaMKII ., In addition , CA1 is innervated by dopamine fibers67 , and dopamine type D1/D5 receptors , coupled to Gsα , are expressed in CA168 ., Dopamine levels increase in response to 100 Hz stimulation 30 , leading to enhanced adenylyl cyclase ( type, 1 ) activity 69 , 70 , and increases in cAMP , which activate PKA 71 , 72 ., The phosphorylation of inhibitor-1 by PKA transforms inhibitor-1 into a potent inhibitor of protein phosphatase 1 73 , 74 , thereby decreasing CaMKII dephosphorylation ., Though not included in the model , the phosphorylation state of the AMPA receptor is controlled by CaMKII , PKA and protein phosphatase 1 75 , 76 ., All reactions in the model are listed in Tables 1 and 2 and are described as bimolecular chemical reactions or as enzymatic reactions except for PKA ( described below ) and CaMKII reactions ( Text S1 ) ., A set of rate equations is constructed to describe the biochemical reactions of the models pathways ., These rate equations are nonlinear ordinary differential equations with concentrations of chemical species as variables ., Equations are derived assuming all reactions are in a single compartment and the number of molecules is sufficient for mass action kinetics , as follows: For a bimolecular chemical reaction: ( 1 ) in which substrates , A and B , are consumed to create products , C and D , the rate of reaction is represented by a differential equation of the form ( 2 ) where kf and kb are the forward and backward rate constants of the reaction , and Kd\u200a=\u200akf /kb is the affinity ., For an enzyme-catalyzed reaction: ( 3 ) where E , S , ES and P denote enzyme , substrate , enzyme-substrate complex and product , the rate of production of P , dP/dt , is given by: ( 4 ) For enzymatic reactions kcat defining the last , catalytic step , is the rate at which product appears , and the affinity Km is defined as ., When kb is not known explicitly , kb is defined as 4 times kcat 14 ., PKA ( cAMP-dependent protein kinase ) is activated by the cooperative binding 77 of cAMP to two tandem cAMP-binding sites ( called A and B sites ) on each of the two regulatory subunits ., The binding of four cAMP leads to the dissociation of the active catalytic subunits , allowing them to phosphorylate their protein targets 4 ., In the model pairs of cAMP bind with first order kinetics as measured by the fraction of free catalytic subunit as a function of cAMP concentration 78 , 79 ., The affinity of site A relative to the affinity of site B is obtained from Herberg et al . 77 , 80 ., Keeping this ratio , the affinity of these sites was adjusted to match the overall affinity of the holoenzyme 72 ( Fig S1A ) ., The only exception to the single compartment approximation is that additional calmodulin was provided to prevent calmodulin from decreasing significantly during stimulation ., Calmodulin is a diffusible protein , thus in a dendritic spine , free calmodulin would diffuse into the spine from the dendrite to replace the bound calmodulin ., In addition , neurogranin is a calmodulin binding protein that releases calmodulin upon Ca2+ stimulation; in essence neurogranin acts as a calmodulin reservoir 48–50 ., The increase in calmodulin was made proportional to the difference between initial calmodulin and free calmodulin ., The rationale is that without additional calmodulin , subsequent stimulation trains produce a smaller increment in phosphoCaMKII ( 1st: 166 nM , 4th: 154 nM ) , which is inconsistent with the positive | Introduction, Results, Discussion, Methods | Protein kinases play critical roles in learning and memory and in long term potentiation ( LTP ) , a form of synaptic plasticity ., The induction of late-phase LTP ( L-LTP ) in the CA1 region of the hippocampus requires several kinases , including CaMKII and PKA , which are activated by calcium-dependent signaling processes and other intracellular signaling pathways ., The requirement for PKA is limited to L-LTP induced using spaced stimuli , but not massed stimuli ., To investigate this temporal sensitivity of PKA , a computational biochemical model of L-LTP induction in CA1 pyramidal neurons was developed ., The model describes the interactions of calcium and cAMP signaling pathways and is based on published biochemical measurements of two key synaptic signaling molecules , PKA and CaMKII ., The model is stimulated using four 100 Hz tetani separated by 3 sec ( massed ) or 300 sec ( spaced ) , identical to experimental L-LTP induction protocols ., Simulations show that spaced stimulation activates more PKA than massed stimulation , and makes a key experimental prediction , that L-LTP is PKA-dependent for intervals larger than 60 sec ., Experimental measurements of L-LTP demonstrate that intervals of 80 sec , but not 40 sec , produce PKA-dependent L-LTP , thereby confirming the model prediction ., Examination of CaMKII reveals that its temporal sensitivity is opposite that of PKA , suggesting that PKA is required after spaced stimulation to compensate for a decrease in CaMKII ., In addition to explaining the temporal sensitivity of PKA , these simulations suggest that the use of several kinases for memory storage allows each to respond optimally to different temporal patterns . | The hippocampus is a part of the cerebral cortex intimately involved in learning and memory behavior ., A common cellular model of learning is a long lasting form of long term potentiation ( L-LTP ) in the hippocampus , because it shares several characteristics with learning ., For example , both learning and long term potentiation exhibit sensitivity to temporal patterns of synaptic inputs and share common intracellular events such as activation of specific intracellular signaling pathways ., Therefore , understanding the pivotal molecules in the intracellular signaling pathways underlying temporal sensitivity of L-LTP in the hippocampus may illuminate mechanisms underlying learning ., We developed a computational model to evaluate whether the signaling pathways leading to activation of the two critical enzymes: protein kinase A and calcium-calmodulin-dependent kinase II are sufficient to explain the experimentally observed temporal sensitivity ., Indeed , the simulations demonstrate that these enzymes exhibit different temporal sensitivities , and make a key experimental prediction , that L-LTP is dependent on protein kinase A for intervals larger than 60 sec ., Measurements of hippocampal L-LTP confirm this prediction , demonstrating the value of a systems biology approach to computational neuroscience . | computational biology/computational neuroscience | null |
journal.pcbi.1004673 | 2,015 | Clustered Desynchronization from High-Frequency Deep Brain Stimulation | High frequency deep brain stimulation ( DBS ) , a medical treatment in which high-frequency , pulsatile current is injected into an appropriate brain region , is a well established technique for alleviating tremors , rigidity , and bradykinesia in patients with Parkinson’s disease 1 , 2 ., While the underlying mechanisms of deep brain stimulation remain unknown , it is well documented that local field potential recordings recorded in the subthalamic nucleus of patients with Parkinson’s disease display increased power in the beta range ( approximately 13–35 Hz ) 3–5 ., These findings have led to the hypothesis that pathological synchronization among neurons in the basal ganglia-cortical loop contribute to the motor symptoms of Parkinson’s disease 6–8 ., This hypothesis has been supported by findings that when DBS is applied to the STN , abatement of motor symptoms is correlated with a decrease the power in the beta band of the local field potential recorded from STN 9–11 ., This line of thinking has led researchers to develop new strategies for desynchronizing populations of pathologically synchronized oscillators , 12–14 , some of which have shown promise as new treatment options for Parkinson’s disease in animal and human studies 15 , 16 ., While many factors including the location of the DBS probe , stimulus duration , and stimulus magnitude influence the efficacy of DBS , one factor that is difficult to explain is the strong dependency on stimulus frequency ., Low-frequency stimulation ( ≤ 50 Hz ) is generally ineffective at reducing symptoms of Parkinson’s disease while high-frequency stimulation from 70 to 1000 Hz and beyond has been shown to be therapeutically effective 17–19 ., However , not all high frequency stimulation is equally effective , and clinicians have generally settled on a therapeutic range at about 130–180 Hz ., 20 , 21 ., In an effort to provide insight into the frequency dependent effects of DBS , the authors of 22 postulated that specifically tuned pulse parameters might yield chaotic desynchronization in a network of neurons ., If desynchronization is the goal of DBS , then achieving it chaotically is a worthwhile objective ., However , this can generally only be seen in a small window of pulse parameters and frequencies which may make it difficult to observe in real neurons ., Furthermore , in both brain slices and in vivo recordings , individual neuronal spikes have been found to be time-locked to the external high-frequency stimulation 23–28 which would be unlikely if the spike times were chaotic ., Here we present a different viewpoint showing that with high frequency pulsatile stimulation , in the presence of a small amount of noise , a population of neurons can split into distinct clusters , each containing a nearly identical proportion of the overall population ., We find that the number of clusters , and hence desynchronization , is highly dependent the pulsing frequency and strength ., We provide theoretical insight into this phenomenon and show that it can be observed over a wide range of pulsing frequencies and pulsing strengths ., This viewpoint merges two seemingly contradictory hypotheses , showing that the therapeutic effect of the periodic pulsing could be to replace the pathological behavior with a less synchronous pattern of activity , even if individual neuronal spikes are phase locked to the DBS pulses ., Consider a noisy , periodically oscillating population of thalamic neurons from 29:, C V i ˙ = f V ( V i , h i , r i ) + I b + u ( t ) + ϵ η i ( t ) , h i ˙ = f h ( V i , h i ) , r i ˙ = f r ( V i , r i ) , i = 1 , … , N ., ( 1 ), Here Vi , hi , and ri represent the transmembrane voltage and gating variables of neuron i , respectively , with all functions and parameters taken to be identical to those found in 29 , DBS pulses are represented by an external current u ( t ) , taken to be identical for each neuron , ηi ( t ) is a Gaussian white noise process , C = 1μF/cm2 is the constant neural membrane capacitance , Ib = 1 . 93μA/μF is a baseline current chosen so that in the absence of external perturbations and noise the firing rate is 60 Hz , and N is the total number of neurons ., Using phase reduction , 30 , 31 , we can study Eq ( 1 ) in a more convenient form:, θ i ˙ = ω + f ( θ i ) δ ( mod ( t , τ ) ) + ϵ η i ( t ) Z ( θ i ) + O ( ϵ 2 ) , i = 1 , … , N , ( 2 ), where θ ∈ 0 , 1 ) is the phase of the neuron with θ = 0 defined to be the time the neuron fires an action potential , ω is the natural frequency of oscillation , f ( θ ) is a continuous function which describes the effect of the DBS pulse , τ is a positive constant that determines the period of the DBS input , and Z ( θ ) is the neuron’s phase response curve to an infinitesimal perturbation . Here we assume that ϵ is small enough so that higher order noise terms are negligible ( c . f . 32 , 33 ) ., Fig 1 shows an example charge-balanced pulsatile stimulus ., We take the positive portion to be five times larger than the negative portion , with the positive current applied for 100 μs ., The bottom panel shows the function f ( θ ) for a given stimulus intensity , calculated using the direct method 34: a pulsatile perturbation is applied to a neuron at a known phase θp so that f ( θp ) can be inferred by measuring the timing of the next spike ., We note that even though the DBS pulse itself is not a δ-function , it is of short enough duration that Eq ( 2 ) is an accurate approximation to Eq ( 1 ) ., We simulate Eq ( 1 ) with 1000 neurons , taking a pulse strength S = 110μA/μF , and noise strength ϵ = 0 ., 05 , for various pulsing frequencies , with results shown in Fig 2 . After some initial transients , we find the network tends to settle to a state with different numbers of clusters for different pulsing frequencies ., From the probability distributions of neural phases ρ ( θ ) , the bottom panels show somewhat surprisingly that once the network settles to a clustered state , each cluster contains a nearly identical portion of the overall population ., Also , upon reaching the steady distribution , neurons can still transition between clusters , but on average , the amount that leave a given cluster must be identical to the amount that enter ., Fig 3 shows individual voltage traces for 50 sample neurons from this population after the network settles to a clustered state ., Highlighted traces represent neurons from each cluster ., In general , increasing the number of clusters will decrease synchrony in the population ., Furthermore , neurons are more likely to transition between clusters as the overall number of clusters becomes larger ., For simplicity of notation , we will take ω = 1 for Eq ( 2 ) in the theoretical analysis , but note that any other value could be considered to obtain qualitatively similar results ., In the absence of noise , one may integrate Eq ( 2 ) for a single neuron θ to yield, θ ( t ) = θ ( 0 ) + t , for t < τ , θ ( t ) = θ ( 0 ) + f ( θ ( 0 ) + τ ) + t , for τ ≤ t < 2 τ ., ( 3 ), In this work , we are interested in the state of the system immediately after each pulsatile input ., By integrating Eq ( 2 ) , the system dynamics can be understood in terms of compositions of a map, θ ( n τ ) = g n ( θ 0 ) , n = 1 , 2 , … , ( 4 ), where g, ( s ) = s + f ( s + τ ) + τ and g ( n ) denotes the composition of g with itself n times , and θ0 is the initial state of a neuron ., In Eqs ( 3 ) and ( 4 ) , θ ( t ) and the arguments of f and g are always evaluated modulo 1 . If g ( n ) has a stable fixed point , then any oscillator which starts within its basin of attraction will approach that fixed point as time approaches infinity 35 ., With noise , the dynamics are more complicated ., In this case , the phase of each neuron cannot be determined exactly from Eq ( 2 ) , but rather , follows a probability distribution ., For a neuron with known initial phase θ0 , after mτ has elapsed , the corresponding δ-function distribution δ ( θ − θ0 ) will be mapped to the Gaussian distribution, ρ ( θ ) = N ( μ , ν ) , ( 5 ), with mean μ = g, ( m ) ( θ0 ) given by Eq ( 23 ) and variance ν given by Eq ( 24 ) ., In order to study the infinite time behavior of Eq ( 2 ) it can be useful to calculate steady state probability distributions for the population of neurons ., To simplify the analysis , we will study Eq ( 2 ) as a series of stochastic maps applied to an initial phase density ( c . f . 22 , 33 ) ,, ρ ( θ , t + m τ ) = P m τ ρ ( θ , t ) , ( 6 ), where Pmτ is the linear Frobenius Perron operator corresponding to evolution for the time mτ , and ρ ( θ , t ) is the probability distribution of phases at time t ., We can approximate Pmτ by the matrix P m τ ∈ R M × M by using eq ( 5 ) to determine each column of the matrix for a set of discretized phases , Δθ = 1/M ., In Fig 4 for instance , the map g ( 2 ) yields the stochastic matrix P 2 τ , shown in the panel on the right ., The delta function distribution ( arrow ) is mapped to a Gaussian distribution ( solid line ) ., The matrix P m τ has all positive entries , and since probability is conserved , the matrix is column stochastic ( i . e . the columns of P m τ sum to 1 ) ., For this class of matrices , the Perron-Frobenius theorem allows us to write 36 , 37 ,, lim k → ∞ P m τ k = v w T , ( 7 ), where v and w are the right and left eigenvectors associated with the unique eigenvalue of 1 , and normalized so that wT v = 1 . Recalling that P m τ is column stochastic , its left eigenvector associated with λ = 1 is 1T ., Therefore , as the map is applied repeatedly , any initial distribution will approach a steady state distribution determined by v . We find that the existence of m fixed points of the underlying map g, ( m ) ( θ ) provide the basis for the clustered desynchronization seen in Fig 2 , with a more formal main theoretical result given below ., Consider the map g, ( m ) with the following properties: Then for a given choice of ϵ1 ≪ 1 , we may choose ϵ small enough in eq ( 2 ) so that the distribution of phases will asymptotically approach a state with m distinct clusters , each containing 1 / m + O ( ϵ 1 ) of the total probability density ., A proof of this statement is given in the Methods Section ., In this detailed proof , we find that desynchronization can still be guaranteed even when g, ( m ) is not monotonic as long as a more general set of conditions is satisfied ., Note that because Eq ( 2 ) does not contain any coupling terms , noise will drive the system to a uniform , desynchronized , distribution in the absence of DBS input ., In the sections to follow , we give numerical and theoretical evidence that clustered desynchronization can emerge in a population of pathologically synchronized neurons when the DBS pulses overwhelm the terms responsible for the synchronization ., Using the main theoretical result , we can calculate regions of parameter space where we expect clustered desynchronization ., The top-left panel of Fig 5 gives regions of parameter space where clustering is expected , giving the appearance of Arnold tongues 35 ., White regions in the graph represent regions where either clustering is not guaranteed , or where we expect more than five clusters ., However , we do not include these regions in the figure because they only exist for very narrow regions of parameter space ., At around 60 Hz , the natural unforced period of the neural population , exactly one cluster is guaranteed , corresponding to 1:1 locking ( one DBS pulse per neural spike ) ., This locking corresponds to a highly synchronous state , which we found when forcing the population at 63 Hz in Fig 2 . For pulsing frequencies between 80 and 120 Hz , we see prominent tongues corresponding to states with three , four and five clusters , which correspond to the states in Fig 2 where we force at 83 Hz and 94 Hz ., A very wide tongue corresponding to 2:1 locking ( two DBS pulses per neural spike ) exists at frequencies ranging from 120 to 200 Hz , which is where DBS is often seen to be effective ., Pulsing in this region manifests in the behavior seen with 120 Hz forcing in Fig 2 . To make comparisons with 22 we calculate the average Lyapunov exponent of the resulting steady state distributions using Eq ( 12 ) ., For Lyapunov exponents greater than zero ( resp . , less than zero ) , the pulsatile stimulus will , on average , desynchronize ( resp . , synchronize ) neurons which are close in phase , and this has been proposed as an indicator of the overall desynchronization that might be observed in a population of neurons receiving periodic DBS pulses ., The Lyapunov exponent is calculated for multiple parameter values for a system with a noise strength ϵ = 0 . 1 ., Results are given in the bottom-left panel of Fig 5 . We note that results are not qualitatively different for different noise strengths ., Compared with the Arnold tongues in the top-left panel , we find very narrow regions where the Lyapunov exponent is positive at relatively high stimulus strengths ., The top-right panels show the steady state distribution for a population with pulses of strength S = 50μA/μF at a rate of 119 Hz ( resp . , 180 Hz ) corresponding to a two ( resp . , three ) cluster steady state ., The bottom-right panels show the steady state distribution for a pulsing strength of S = 208μA/μF at 120 Hz and S = 206μA/μF at 290 Hz corresponding to regions with positive Lyapunov exponents ., Even though the clustered states have very negative Lyapunov exponents , they show similar clustering behavior to the states with a positive Lyapunov exponent ., However , the clustered desynchronization in the top-right panels can be accomplished using a significantly weaker stimulus and can be observed at a much wider range of pulsing parameters ., Results thus far have focused on populations of neurons receiving homogeneous pulsatile inputs ., However , it is well established that voltage fields vary significantly with distance from an external voltage probe 38 ., In computational models such heterogeneity has been shown to create complicated patterns of phase locking that are dependent on the stimulation strength 39 and can improve methods designed to desynchronize large populations of neurons 14 ., To understand the emergence of clustered synchronization when external inputs are different among neurons , we can modify the stochastic differential eq ( 2 ) as follows, θ i ˙ = ω + f i ( θ i ) δ ( mod ( t , τ ) ) + ϵ η i ( t ) Z ( θ i ) + O ( ϵ 2 ) , i = 1 , … , N ., ( 8 ), Here , fi ( θ ) is calculated based on the pulsatile input to each neuron ., For each neuron , we use Eq ( 7 ) to calculate its steady state probability distribution ., The state of each neuron is independent of the others , so that the average of the individual distributions gives an overall probability distribution for the population ., As an illustrative example , we model 1000 neurons of the form Eq ( 1 ) receiving a charge balanced input of the same shape as in Fig 1 with τ = 1/ ( 140 Hz ) and S drawn from a normal distribution with mean 168 μA/μF and standard deviation 20 , giving values of S between approximately 100 and 240 ., From the top-left panel of Fig 5 , this range of stimulus parameters is mostly , but not completely , contained in a two cluster region ., g ( 2 ) ( θ ) is plotted in black in the top left panel of Fig 6 for a randomly chosen subset of these neurons with the identity line plotted in red for reference ., The top-right panel shows each neuron’s steady state probability distribution ( calculated from its associated stochastic matrix ) in black for a noise strength of ϵ = 0 ., 4 . While the main clustering results are guaranteed when the noise strength is small enough , we find that clustering can still occur at higher levels of noise ., The steady state probability distribution in corresponding simulations of Eq ( 1 ) with heterogeneous pulsing strengths ( blue dashed curve ) agrees well with the theoretical probability density ( red dashed curve ) calculated from the average of each black curve in the top-right panel ., The bottom panel shows corresponding cumulative distributions for the theoretical ( red ) and computationally observed ( blue ) probability densities highlighting that similar numbers of neurons are contained in each cluster ., Similar clustering results can be obtained for different heterogeneous stimulus parameters ., For instance , from Fig 5 , three-cluster behavior will emerge for pulsing frequencies of 200 Hz and stimulus strengths between approximately 90 and 170 μA/μF ., Our main clustering results are for single population of neurons which do not explicitly take interactions between multiple populations of neurons into account , as is the case for the brain circuit responsible for Parkinsonian tremor ., Here , we provide evidence that clustered desynchronization can still emerge when additional forcing terms are much smaller in magnitude than the external DBS pulses ., Populations of coupled oscillators subject to common external forcing have been widely studied in the form of the forced Kuramoto model 40–42 ., Synchronization can be observed when either external forcing or intrinsic coupling dominate the system dynamics ., For intermediate coupling and external forcing strengths , a complicated bifurcation structure emerges and the macroscopic order parameter , describing the overall synchronization of the population , can oscillate ., These behaviors have been observed in chemical oscillator systems 43 and have implications to externally forced biological rhythms such as circadian oscillations and neural oscillations 44 ., We simulate Eq ( 1 ) with an additional external sinusoidal forcing frequency which could represent an aggregate input from a separate , unperturbed neural population ., We note that this is not meant to represent a physiologically accurate model of DBS , but instead is meant to illustrate clustered desynchronization in the presence of a common periodic perturbation ., Here , we take u ( t ) = 0 . 1 sin ( ωext t ) + uDBS ( t ) , where ωext is chosen so that the frequency of oscillation is the same as the natural period of the unperturbed neurons ( 60 Hz ) and uDBS ( t ) represents the common pulsatile input ., For these simulations , N = 1000 ., As we show in the Methods Section we may write this system in an identical form as Eq ( 2 ) , for which the main theoretical result still holds ., For this particular example , clustering results are identical to those in the top left panel of Fig 5 . Results are shown with a pulsing strength S = 110μA/μF and noise strength of ϵ = 0 ., 02 in Fig 7 . We find that the presence of a sinusoidal external stimulus is sufficient to synchronize the network in the absence of DBS forcing ., When the DBS is turned on at both 83 and 94 Hz , we see three and four cluster states , respectively , just as we observed in the simulation without external forcing ., However , in this simulation , the mean phase of each cluster varies with the external sinusoidal stimulation ., Note that 120 Hz stimulation in this network also leads to two cluster desynchronization but is not shown ., When neurons are synchronized through forcing that is not periodic , clustered desynchronization may still emerge when the DBS pulsing overwhelms the stimulation responsible for synchronization ., As a second example , we model a network of neurons Eq ( 1 ) with an additional synaptic current , with each neuron’s transmembrane voltage dynamics taking the form C V i ˙ = f V ( V i , h i , r i ) + I b + u ( t ) + ϵ η i ( t ) + I i syn ( t ) ., Here ,, I i syn ( t ) = K N ∑ k = 1 N ( V i - V G ) s k ( t ) ( 9 ), where K determines the magnitude of the synaptic current , VG is the reversal potential of a given neurotransmitter , and sk an additional synaptic variable associated with neuron k that evolves according to ( c . f . 30 ) s ˙ k = α 2 ( 1 - s k ) ( 1 / ( 1 + exp ( - ( V k - V T ) / σ T ) ) ) - β 2 s k , where α2 = 2 , VT = -37 , σT = 2 , and β2 = 0 . 1 ., We simulate the resulting network withVg = 60mV , K = 0 . 015 and a noise strength of ϵ = 0 ., 02; neurons form a single synchronized cluster in the absence of DBS input shown in panel B of Fig 8 . starting at t = 0 . 5 ms , we apply 180 Hz stimulation with S = 200μA/μF , the pulsing quickly overwhelms the synchronizing influence of the coupling , and the population splits into two separate clusters as shown by the probability densities in Panels A and individual voltage traces in panel C . When DBS is applied , we see from the average probability distributions and cumulative distributions in panels F and E , respectively that there are nearly equal proportions of neurons in each cluster ., Other computational modeling 29 has suggested that pulsatile DBS may help regulate neural firing patterns , and help alleviate strongly oscillatory synaptic inputs ., Panel D shows a similar phenemenon , when DBS is on , high amplitude oscillations in synaptic current are replaced by oscillations with a higher frequency but smaller amplitude ., The desynchronization results here can be observed for many choices of parameters provided the pulsatile stimulation is significantly stronger than the synchronizing stimulation and that clustering behavior is expected in the absence of coupling ., Consider a two dimensional reduction of the classic Hodgkin-Huxley equations 45 which reproduce the essential dynamical behavior 46:, C V i ˙ = f V H ( V i , n i ) + I b + u ( t ) + ϵ η i ( t ) , n i ˙ = f n ( V i , n i ) , i = 1 , … , N ., ( 10 ), Here Vi and ni represent the transmembrane voltage and gating variables , respectively ., All functions and parameters are identical to those given in 47 ., DBS pulses are represented by the external current u ( t ) , which is given identically to each neuron , ηi ( t ) is a white noise process , C = 1μF/cm2 is the constant neural membrane capacitance , Ib = 10μA/μF is a baseline current yielding a firing rate of 84 . 7 Hz in the absence of external perturbation , and N is the total number of neurons ., Unlike the model for thalamic neurons used in the main text , the Hodgkin-Huxley neuron displays Type II phase response properties , i . e . , a monophasic pulsatile input can act to either significantly increase or decrease the phase of the neuron ., The top panel of Fig 9 shows an example monophasic stimulus which will be applied to the network Eq ( 10 ) at different strengths , S and periods τ ., For this example , the pulse duration will be 100 μs , approximately , one percent of the neural firing rate ., For this model , using our main theoretical results , we can also calculate regions of parameter space in which we expect to observe clustered desynchronization , with results shown in the middle panel of Fig 10 . The Arnold tongues for clustering greater than five become quite narrow and are not included in this figure ., We also calculate the average Lyapunov exponent for the steady state distribution using eq ( 12 ) from the main text for a noise strength of ϵ = 0 . 15 , with results shown in the bottom panel ., We note that unlike for the thalamic neurons , the Lyapunov exponent for the Hodgkin-Huxley network is never positive ., We find that regions with the lowest Lyapunov exponents tend to correlate with regions where clustered desynchronization is guaranteed ., Even though the Lyapunov exponent might be quite negative , the steady state distribution can still be relatively desynchronized if there are a large number of clusters , as evidenced by the four cluster state in the top panel ., Finally , we simulate Eq ( 10 ) with N = 1000 neurons with a pulse strength S = 10μA/μF and ϵ = 0 ., 3 for pulsing frequencies that are expected to yield clustered desynchronization determined from Fig 10 . Results are shown in Fig 11 . We find clustered states begin to form almost immediately , and in the bottom panel , after the system has approached the steady state distribution , each cluster contains an approximately identical proportion of the population ., While deep brain stimulation is an important treatment for patients with medically intractable Parkinson’s disease , its fundamental mechanisms remain unknown ., Making matters more complicated , experimental studies have shown that the symptoms of Parkinson’s can be alleviated using strategies that seek to desynchronize a population of pathologically synchronized oscillators 15 , 16 , while other seemingly contradictory studies have shown that neurons have a tendency to time-lock to external high-frequency pulses 23–27 , supporting the hypothesis that entrainment is necessary to replace the pathological neural activity in order to alleviate the symptoms of Parkinson’s disease ., In this work we have have shown that these two phenomenon may be happening in concert: in the presence of a small amount of noise , high frequency pulsing at a wide range of frequencies is expected to split a larger population of neurons into subpopulations , each with a nearly equal proportion of the overall population ., The number of subpopulations , and hence the level of desynchronization , is determined by phase locking relationships which can be found by analyzing the phase reduced system in the absence of noise ., We note that other theoretical 12 and experimental 1615 work has yielded control strategies that are specifically designed to split a pathologically synchronized neural population into distinct clusters ., The theory presented in this paper suggests that clinical DBS may be accomplishing the same task with a single probe ., The conditions we have developed guarantee clustered desynchronization for small enough noise , but we do not give any a priori estimate of how small the noise needs to be so that distinct clusters can be observed ., If the noise is too large , the clusters may start to merge into one another , particularly when there are a large number of clusters ( see the bottom left panel of Fig 2 ) ., Even in this case , however , we still have discernible clusters , throughout which the overall population of neurons is spread relatively evenly ., We also note that this theory does not give any estimates on the time it takes for the system to achieve its steady state population distribution , but this can be calculated for a specific population by examining the second smallest eigenvalue in magnitude , λ2 , of a given stochastic matrix P m τ ( c . f . 33 ) ., As λ2 becomes closer to 1 , more iterations of the map P m τ will be required for the transient dynamics to die out , and it will take longer for the system to approach the steady state distribution ., In general , we find that for a given map , λ2 becomes closer to one as noise strength decreases , which is consistent with the notion that the average escape time between clusters will increase as the strength of the external noise decreases 48 ., For the networks that we have investigated , regions of parameter space which are associated with either clustered desynchronization or positive Lyapunov exponents can display similar levels of desynchronization ., However , numerics show that the regions with positive Lyapunov exponents are quite small and may be difficult to find without explicit calculation ., In contrast , the regions of parameter space associated with clustered desynchronization are fairly large and are likely to be observed without knowledge of the system properties ., If desynchronization of the overall population is an important mechanism of high-frequency DBS , doing so chaotically may be an overly restrictive objective if clustered desynchronization is sufficient to alleviate the motor symptoms of Parkinson’s disease ., This study is certainly not without limitations ., For instance , the computational neurons considered in this study are based on simple , low-dimensional models of neural spiking behavior ., However , we have developed the theory to understand the clustered desynchronization phenomenon in such a way that it can easily be extended to more complicated neural models with more physiologically detailed dynamics provided the neural phase response properties can be measured experimentally in vivo49 ., Furthermore , while we only considered homogeneous populations in this study , the phase response properties and natural frequencies of a living population of neurons will surely have a heterogeneous distribution ., In this context , we could still show that clustered desynchronization is expected by applying the theory developed in this work to a family of neurons with different phase response properties and natural frequencies ., The expected steady state population could then be obtained as a weighted average of the individual steady state distributions ., Numerical results presented here apply to networks for which external DBS perturbations overwhelm the intrinsic coupling between neurons ., In this work , we have not considered the complicated interplay between multiple populations of neurons which give rise to the symptoms of Parkinson’s disease; more detailed modelling studies would be required to determine the effect of clustered desynchronization on the overall network circuit ., Others have studied synchrony and clustering behavior in coupled populations of neurons 50–52 and it is possible that our results could be extended to describe clustering for weaker pulsatile stimuli when coupling cannot be neglected ., Our results suggest that high-frequency external pulsing could have the effect of separating a neural population into equal subpopulations in the presence of noise ., This viewpoint could help explain the frequency dependent nature of therapeutically effective DBS and could help merge competing hypotheses , as desynchronization and entrainment are not mutually exclusive when even small amounts of noise are present ., If clustered desynchronization does provide a mechanism by which the motor symptoms associated with Parkinson’s disease can be mitigated , it could provide a useful control objective for designing better open-loop DBS stimuli in order to prolong battery life of the implantable device and to mitigate potential side effects of this therapy ., To make comparisons with 22 we calculate the average Lyapunov exponent of the resulting steady state distributions , giving a sense of whether , on average , the orbits of the trajectories oscillators from Eq ( 2 ) are converging or diverging ., For instance , let ϕ denote the phase difference between oscillators θ1 and θ2 which are close in phase , i . e . ϕ ( t ) ≡|θ1 ( t ) − θ2 ( t ) | ., Then from Eq ( 3 ) , immediately after a DBS pulse occuring at time τ ,, ϕ ( τ + ) = |f ( θ 1 ( τ - ) ) + θ 1 ( τ - ) - f ( θ 2 ( τ - ) ) - θ 2 ( τ - ) | , = |f ( θ 2 ( τ - ) ) + f ′ ( θ 2 ( τ - ) ) ϕ ( τ - ) + O ( ϕ ( τ - ) 2 ) + θ 1 ( τ - ) - f ( θ 2 ( τ - ) ) - θ 2 ( τ - ) | , = ϕ ( τ - ) |1 + f ′ ( θ 2 ( τ - ) ) | + O ( ϕ ( τ - ) 2 ) , ( 11 ), where ′ ≡ d/dθ and θ ( τ− ) ( resp ., θ ( τ+ ) ) denotes the limit of θ ( t ) as t approaches τ from below ( resp . above ) ., Note that in the second line , we have used a Taylor expansion of f about θ2 for small values of ϕ ( τ− ) ., Therefore , the oscillators converge or diverge locally depending upon the derivative of f ., For a population of neurons , the stochastic matrix P τ for a given pulsing rate can be used to determine the steady state distribution ρ* ( θ ) before each pulse , with an average Lyapunov exponent taken to be ( c . f . 22 ) :, LE = ∫ 0 1 ρ * ( θ ) log 1 + f ′ ( θ ) d θ ., ( 12 ), For LE > 0 ( resp . , LE < 0 ) , the pulsatile stimulus will , on average , desynchronize ( resp . , synchronize ) neurons which are close in phase , and this has been proposed as an indicator of the overall desynchronization that might be observed in a population of neurons receiving periodic DBS pulses ., For a single neuron with a known initial phase θ evolving according to the stochastic differential Eq ( 2 ) , we calculate the expected value and variance with a strategy that is similar to the one employed in 33 ., We first asymptotically expand θ ( t ) in orders of ϵ ,, θ ( t ) = θ 0 ( t ) + ϵ θ 1 ( t | Introduction, Results, Discussion, Methods | While high-frequency deep brain stimulation is a well established treatment for Parkinson’s disease , its underlying mechanisms remain elusive ., Here , we show that two competing hypotheses , desynchronization and entrainment in a population of model neurons , may not be mutually exclusive ., We find that in a noisy group of phase oscillators , high frequency perturbations can separate the population into multiple clusters , each with a nearly identical proportion of the overall population ., This phenomenon can be understood by studying maps of the underlying deterministic system and is guaranteed to be observed for small noise strengths ., When we apply this framework to populations of Type I and Type II neurons , we observe clustered desynchronization at many pulsing frequencies . | While high-frequency deep brain stimulation ( DBS ) is a decades old treatment for alleviating the motor symptoms Parkinsons disease , the way in which it alleviates these symptoms is not well understood ., Making matters more complicated , some experimental results suggest that DBS works by making neurons fire more regularly , while other seemingly contradictory results suggest that DBS works by making neural firing patterns less synchronized ., Here we present theoretical and numerical results with the potential to merge these competing hypotheses ., For predictable DBS pulsing rates , in the presence of a small amount of noise , a population of neurons will split into distinct clusters , each containing a nearly identical proportion of the overall population ., When we observe this clustering phenomenon , on a short time scale , neurons are entrained to high-frequency DBS pulsing , but on a long time scale , they desynchronize predictably . | null | null |
journal.pcbi.1004697 | 2,016 | Binding Site Identification and Flexible Docking of Single Stranded RNA to Proteins Using a Fragment-Based Approach | RNA participates in most processes leading to genome expression and its regulation 1 , 2 , mainly in association with proteins 3 , 4 ., Protein-RNA interactions are also involved in several neurodegenerative diseases 5 and cancers 6 ., Understanding such interaction and the design of drug molecules requires the three-dimensional structure of protein-RNA complexes 7 ., In many cases the protein bound RNA molecule is able to adopt a great variety of conformations ., In particular , the structure determination of complexes containing flexible single-stranded ( ss ) RNA is a major challenge ., Protein-RNA docking methods could help to generate at least models of such interactions ., The first task in docking is to sufficiently sample the space of possible conformations and relative orientations ( i . e . poses ) of the components so as to include near-native structures ., Similar to existing protein-protein docking methods 8 , 9 , most current protein-RNA docking methods consist of docking rigid structures of unbound RNAs or their domains 10 , 11 , with no or very limited conformational sampling of the RNA conformations prior to docking ., Recently , some efforts have been made to model RNA flexibility , by use of, ( i ) coarse-grained models to account for atomic-scale inaccuracies 12 ,, ( ii ) normal modes analyses and elastic network models 13 , 14 to explore large linear global motions ,, ( iii ) local backbone perturbations modeling non-linear deformation 14 , P . Setny , I . Chauvot de Beauchêne and M . Zacharias , in prep . , or, ( iv ) comparison to a template library of protein-RNA complex structures 15 ., Such semi-rigid-body methods can perform well when moderate or predictable conformational changes occur 8 , 9 ., However , RNA conformational changes upon association with protein can involve global rearrangements , changes of secondary structure elements and/or flipping-out of bases from intra- to extra-helical position , which most docking methods fail to model 16 ., The limits of current methods have been illustrated by the 15th round of the Critical Assessment of PRedicted Interactions ( CAPRI ) experiment 17 ., The first and so far unique CAPRI target consisting of a protein-RNA complex to be modeled from unbound structures has been largely unsuccessful 16 ., The accuracy of current protein-RNA docking methods is limited especially when some single-stranded loops participate in the binding 13 ., Even RNA molecules that are otherwise well-structured contain such single-stranded regions , which are highly flexible or even disordered in the unbound form but carry the specificity of most protein-RNA recognition processes 18 , 19 ., Moreover , many RNA-binding proteins bind only single-stranded RNA ( ssRNA ) , for which no unbound structures are available 20 , 21 ., For docking a highly flexible ligand , fragment-based docking forms an alternative approach to docking ., It consists of cutting the ligand into fragments , docking them separately on the receptor followed by assembling the compatible poses ., The main advantage is that no unbound structure of the whole ligand is required ., However , a structure of the fragments themselves is still needed: if the fragments are themselves flexible , a structural library that samples the possible conformations of each fragment is required ., The second limitation is that all fragments must participate in binding within the native complex , making enough favorable contacts with the receptor for this position to be sampled ., In contrast , in rigid-body docking , if the unbound conformation of the ligand is accurate enough , only part of the ligand needs to make specific contacts with the receptor , the position of the rest of the ligand being fully determined by the position of the interacting part ., Fragment-based docking has been successfully applied to protein-ligand docking , especially for drug design 22 ., In this case , the fragments are typically small ( ring , linker or side-chain 23 ) and the number of fragments to be joined is small ( 2–5 fragments ) 24 ., A first attempt to model protein-bound ssRNA has recently been made by the RNA-Lim method 25 ., Tested on one 6-nucleotide ssRNA–protein complex , and restricting the search on the known binding site , RNA-Lim achieved only limited success , with a 5 Å precision on the nucleotide placement in ~ 10% of the proposed solutions ., RNA-Lim does not predict the orientation of nucleotides , and so far , no fragment-based docking/modeling method has been developed that allows modeling ssRNA bound to a protein with high precision ., This highlights the highly challenging difficulty of this biologically relevant problem ., In the present study , we present a proof-of-principle of a fragment-based approach for ab initio modeling of a protein-bound ssRNA at an unprecedented level of detail ., Assuming that all nucleotides bind the protein , our method does not require any structural information on the ssRNA , assembling it from the sequence alone ., Moreover , in contrast to previous methods , no prior knowledge on the binding site is required ., Each fragment is approximated by a conformational ensemble generated by exhaustive docking of all conformers in a structural library of trinucleotides , built from all the existing experimental structures of protein-ssRNA complexes ., Ensembles corresponding to each trinucleotide sequence present in the RNA sequence are docked all around the protein , and spatially overlapping poses corresponding to overlapping sequences are selected to build the RNA chains ., In the current study , the scope is limited to complexes containing the most abundant RNA-binding domain in proteins: the “RNA recognition motif” ( RRM ) 20 , 21 ., RRMs are present in 2% of all proteins in the human genome , and 44% of RRM-containing proteins contain two or more RRMs 21 ., In particular , we consider complexes where the RNA binds two RRMs , each nucleotide interacting with the protein , and nucleotides outside the binding site are discarded ., The exhaustive docking of a single trinucleotide sequence results in a very large number of poses ( 35–40 million ) ., To reduce computational costs , we focus here on homopolymer RNAs , allowing us to perform only one docking predicting the structure of all RNA fragments simultaneously ., The PDB contains two complexes corresponding to those criteria ( two RRMs + homopolymer ssRNA ) : one poly ( U ) and one poly ( A ) 8-nucleotide ssRNA , bound to two different proteins ., For those two test-cases , we perform thorough analyses of different docking regimes with different amounts of structural knowledge ., We first validate the fragment-based approach by docking and assembling the bound RNA fragments on the bound protein ., Then we assess the cost on the sampling accuracy of the use of sub-optimal fragments conformations , by docking and assembling only the closest-to-bound conformers in our library ., The problem of combinatorial explosion due to usage of the whole library is then addressed and its impact on the results evaluated ., Finally , we explore and discuss the correlation between the precision of the best docking solution and the number of incorrect decoys ., For each complex , the native structure of seven consecutive nucleotides was sampled in close agreement with the published crystal structure ( ~ 1 . 5 Å RMSD ) , a precision never reached so far ., Such a limited benchmark does not allow us to claim any generality of our method ., However , it provide a convincing proof-of-principle for fragment-based docking of protein-ssRNA complexes , pushing farther the limits of modeling RNA flexibility in docking ., To validate the fragment-based approach , we took the ssRNA conformation from each complex and cut it into 6 overlapping trinucleotides ., We used those fragments as a “bound library” to perform fragment-based docking ., This was compared to a standard rigid-body docking of the bound ssRNA as a whole onto the protein ., For bound docking , we would not expect a fragment-based approach to outperform traditional rigid-body docking ., In general , assuming a reasonably accurate scoring function , a large number of favorable contacts highly favors the native pose ., In fragment-based docking , however , the favorable contacts are split among all fragments , making the docking of each fragment more difficult in terms of scoring ., Even more limiting is the possibility that some fragments establish no or few favorable contacts with the protein , their position in the native complex being only constrained by the favorable interactions established by the adjacent fragments ., In such cases , the sampling of these fragments is not possible within a reasonable number of poses ., Standard rigid-body bound docking reproduced the complex structure with 0 . 2 and 0 . 9 Å RMSD for 1B7F and 1CVJ , respectively ., To assess the effect of switching to a fragment-based approach , we performed separate docking for each fragment with the same protocol as for rigid-body bound docking ., After discarding the redundant poses , we selected the 20% top-ranked poses in ATTRACT force-field ., The fragment docking proved of comparable efficiency to the rigid-body docking , with very little loss in accuracy of the sampling: all fragments were sampled with a precision of 0 . 5–1 . 2 Å ( Table 1 ) ., Therefore , the process of cutting the RNA into fragments does not lead to a significant loss in accuracy or precision , at least in our two test-cases ., Also , the scoring was accurate enough to keep almost all generated hits ( RMSD < 2 Å ) in the 20% top-ranked poses ., These results suggest that , in the native form of our test complexes , each fragment establishes enough favorable contacts with the protein to participate in the positioning of the whole RNA , independent of the constraints applied by the adjacent fragment ., However , the number of hits and near-hits ( RMSD < 5 Å ) formed only a small fraction within the pool of selected poses ., To enrich this fraction , we performed a position-specific filtering based on the propensity of each pose to form ssRNA chains ., With six different pools , we tested all possible chain connections between the poses in pool n and in pool n+1 ., Connectivity was defined by an overlap criterion , based on a strict upper distance limit between the atoms in the last nucleotides of n and the first nucleotides of n+1 ., Based on these connectivities , all possible six-fragment chains were enumerated ., By selecting only chain-forming poses , the large majority of hits was kept and essentially all wrong poses were eliminated ( Table 1 ) ., For each complex , the complete RNA chain was modeled with sub-angstrom resolution ( geometric mean of CG RMSD over the 6 fragments ) ., This method proved highly selective: more than 99% of the 6 . 105–4 . 105 chains built for 1B7F - 1CVJ respectively had an RMSD under 2 Å , and 75–96% under 1 . 5 Å ., More importantly , we found that position-specificity is not a requirement for the chain propensity filter ., In a second chain assembling test , all six bound-docking fragment pools were merged into a single pool , equivalent to performing a single docking run with a library of six ( bound ) conformers ., Connectivity was considered between all poses within the pool ( a pose from the conformer corresponding to fragment 1 could thus be placed at any position in the chain , not only 1st position ) , and all poses with a propensity to form chains of at least five fragments were kept ., As shown in Table 1 , this chain propensity filter , used in all subsequent experiments , performs as well in terms of selectivity as the position-specific filter ., More than 81% of the hits are kept by the chain propensity filter , whereas only ~1% of the total poses are kept ., Cutting the RNA into fragments allows us to model flexibility at the fragment level ., To do so at the trinucleotide level , the conformational space for each possible trinucleotide sequence ( in our case , AAA and UUU ) must be sampled ., In the absence of a bound structure , conformational sampling can be provided by using a generic library for single-stranded , protein-bound ssRNA trinucleotides that occur in nature ., However , to the best of our knowledge , no such library exists ., The RNA fragment library used by FARNA for de novo prediction of RNA was built “from a single crystal structure containing just over 2 , 700 ribonucleotides from the large ribosomal subunit from Haloarcula marismortui 1FFK” 29 , which is mainly double-stranded ., The libraries used by MC-Fold/MC-Sym 30 ModeRNA 31 or RNA-MoIP 32 represent only fragments that are partially or fully double-stranded fragments ( “Nucleotide Cyclic Motifs” ) 33 or internal loops ( which limit the backbone conformations sampling by a loop closure constraint ) ., Therefore , for the current approach we extracted all trinucleotide structures from ~500 ssRNA-protein complexes available in the PDB ( July 2014 ) and built exhaustive non-redundant libraries of 1305/1140 UUU/AAA protein-bound fragments ., The two test-case complexes ( 1B7F and 1CVJ ) were excluded from the library building process ., We computed the RMSD of the best-fitting conformer of the library with respect to each fragment in our test cases ., Our library proved exhaustive enough to approximate each bound trinucleotide fragment within 2 Å , and in the great majority of cases ( 75% ) within 1 Å ( Table 2 ) ., In the future , to further increase the accuracy of the docking , one should regularly update the library with new resolved structures of protein-RNA complexes ., As a next step , we evaluated the effect of the inaccuracy of even the best conformations ( closest to the bound fragments ) in our library on the docking results ., When docking the whole UUU/AAA libraries and assembling the poses into chains , the best solutions ( smallest RMSD toward native form ) are likely to be formed by a chain of poses of the library conformers that are similar to the bound form ., For a first evaluation of the capacity of our library to sample near-native solutions , with a reduced computational cost , we performed a biased docking test for each complex: prior to docking , we selected for each bound fragment the best fitting conformer in our library , resulting in six UUU/AAA conformers out of 1305/1140 ., After docking , we retained the 20% best poses for each conformer and merged them into a unique pool , ending up with a total of 19 , 293 and 17 , 345 non-redundant poses for 1B7F and 1CVJ respectively ., For each complex , all poses were compared to each of the bound fragments , and the number of poses close to each fragment was assessed ., Hits were found for 75% of the fragments , and near-hits for all fragments ( Table 2 col . I ) ., As expected , the RMSD of the best pose is linearly correlated to the accuracy of the best-fitting conformer ( Pearson coeff 0 . 72 , p-val 0 . 008 ) ., The most inaccurately docked fragments are frag1 in 1B7F , and frag6 in both 1B7F and 1CVJ ( 2 . 3 Å , 2 . 0 Å and 2 . 9 Å respectively ) ., The first one corresponds to the most deeply buried fragment in the binding site ., The structures of the two frag6 correspond to conformations that are less well-approximated in the fragment library: the best conformers display 1 . 8 Å RMSD when fitted to the bound form , versus 0 . 3 Å to 1 . 1 Å for the other fragments ( Table 2 ) ., Additionally in 1CVJ , the nucleotides from fragment 6 establish interactions not only with the protein but also with the RNA of symmetrical units in the crystal ( fragment 6 1st and 3rd nucleotides ) , and with a soluble adenosine-5-monophosphate ( fragment 6 2nd nucleotides ) ., This makes it more difficult to sample the correct pose of this fragment on the protein alone ., All nucleotides in both complexes establish H-bonds via their bases and/or phosphates , except the 3rd nucleotide of 1B7F frag6 ., This nucleotide binds the protein by H-bonds via its O3 and O2 oxygens , which position in the coarse grain representation is more loosely defined than of the other partially charged atoms ., This could also contribute to a worse sampling of that fragment ., Apart from these limitations , our docking results indicate that ATTRACT was able to sample and rank solutions close to the optimal position of each conformer in the 20% top-ranked poses ., To account for a decreased sampling quality of the terminal fragments , we decided to build 5-fragment chains for the docking poses ., We applied the same chain-propensity filter as for bound docking ., Even more so , the filter eliminated virtually all incorrect poses , ending up with 53 and 24 poses out of 19293 and 17345 , respectively ( Table 2 ) ., Again , the procedure proved highly selective: 38–63% of the retained poses were hits for 1B7F and 1CVJ , respectively , compared to 0 . 4–0 . 1% before filtering ( Table 2 ) ., Moreover , for all fragments for which a hit was in the top 20% , one or more hits were kept after filtering , usually the ones with the best RMSD ., Finally , the filtered poses were assembled into all possible 5-fragment chains ( 166–69 chains ) and compared to the bound ssRNA chain ( nucleotide 1–7 ) ., The best chain had an average RMSD of 1 . 6–1 . 0 Å for 1CVJ and 1B7F , respectively ( Fig 3 ) ., More importantly , this RMSD was representative for the whole result ., For 1CVJ , 64% of the chains had an overall RMSD of better than 2 Å , and all chains were within a 5 Å deviation ., For 1B7F , there was a little more diversity: 21% of the chains within a 2 Å RMSD of the native geometry , and 27% within 5 Å ., We clustered the poses at the 5 Å ( 1B7F ) or 0 . 5 Å ( 1CVJ ) level , in order to get similar numbers of clusters despite the higher diversity in poses on 1B7F ., The correct cluster was the 1st-largest for 1CVJ and the 5th-largest for 1B7F , with all chains in this cluster within 1 . 1–1 . 8 Å respectively ., In conclusion , using approximately correct conformations for the fragments , the correct chain was one of the very few possible ways to build a poly-U/A hexanucleotide onto the protein ., A similar procedure was applied considering not only the best conformers but the whole UUU/AAA sub-library ( 1305/1140 conformers ) ., This should in principle not modify the sampling compared to biased docking , as the poses obtained by biased ( subset of the library ) docking will constitute a subset of the poses obtained by unbound ( whole library ) docking ., However , the inclusion of the other library conformers results in a large number of mostly inaccurate decoys with a potential impact on the ranking of the correct solutions ., In addition , compared to biased docking , the very high number of poses generated by unbound docking ( 30 , 000 * 1305 = ~40 million , compared to just 180 , 000 for biased docking ) causes considerable additional numerical demand and we had to adapt our protocol accordingly ., First , to take into account the redundancy induced by close conformers in the library , we selected only the 5% top-ranked non-redundant poses for each conformer , instead of 20% as for our biased docking ., Second , processing such a large pool of poses was not possible in terms of computational memory ., Therefore , we assembled first a small sub-pool of poses , retained the chain-forming fragments , and selected all related poses ( close in RMSD ) from a larger sub-pool , in an iterative procedure that eventually kept about two-thirds of all top 5% poses before the final filtering ., The docking produced poses within 3 Å RMSD toward all bound fragments but frag6 in 1B7F , similarly to what was obtained by biased docking ( Table 3 ) ., Interestingly , despite the reduced percentage of poses kept per fragment compared to our biased docking ( 5% vs 20% ) , the quality of our sampling was not significantly changed ., The best sampling of some fragments was even improved ( see 1B7F frag5 and 1CVJ frag4-5 , Tables 2–3 ) , which means that a conformer close to the best-fitting conformer was docked better than the best-fitted conformer itself ., Thus , the redundancies in the poses induced by the use of the full library made the 5% top-ranked poses sufficient to reach a good sampling ., However , a notable exception is 1B7F fragment 2 , for which hits were no longer among the top-ranked poses ., In addition , for all fragments , the large increase in the number of candidate poses reduced the fraction of hits among the top-ranked poses by an order of magnitude or more ., The large increase in the number of candidate poses in unbound compared to biased docking ( in the previous paragraphs ) made it much more difficult to select hits and near-hits ., Among all fragments , the best docking solution was kept in the filtered solutions for only 3 of the 12 cases ( Table 3 , Fig 4 ) ., For 1CVJ , the chain-propensity filter performed well: it kept almost half of the hits while selecting only 0 . 6% of all the poses , leading to a 78 fold enrichment ., Still , because the chain-propensity filter selected a few thousand structures , rather than a few dozen for biased docking , the hits represented only 0 . 4% of all selected poses ( compared to 64% for biased docking ) ., The procedure performed less well for 1B7F , selecting no hits at all , which might be explained by the reduced sampling for fragment 2 at the docking stage ( no hit , versus 3 hits for biased docking ) ., However , for both 1CVJ and 1B7F , the procedure led to a significant enrichment of near-hits , increasing their percentage from 3–4% to 10–13% respectively , while keeping less than 1% of the docking poses ., For all experiments , the chain-propensity filter was shown to be highly selective in eliminating incorrect solutions ( that cannot form an ssRNA chain ) ., For biased docking , we found it to be rather sensitive to the chain length parameter ., With a chain-propensity filter based on the capacity of each pose to participate in 6-fragments chains instead of 5-fragments chains , no chains were formed at all ( causing all fragments to be eliminated ) ., In contrast , for unbound docking , changing the chain length to 4 or 6 had little effect: the fraction of near-hits among the selected poses remained at ~14% for both test cases ( supplementary material , S1 Table ) ., Despite the still high number of decoys after filtering , the unbound docking permitted to exactly delineate the binding site ( Fig 5 ) without taking this information into account prior to docking: The worst pose after filtering was at only 16 . 7–14 . 9 Å from the closest fragment in 1B7F and 1CVJ respectively; for each complex , more than 65% of the poses were under 10 Å and more than 95% under 15 Å ., So , our procedure for fragments assembly proved an efficient method to discard remote poses ., These results also suggest that the method could be used for binding site prediction ., The novelty compared to existing methods is that is does not use any information from sequence conservation or homology ., But as we tested it only on a very well conserved pattern , where homology-based methods for binding-site prediction should work very well , this direction would need farther investigation on other patterns ., To reduce further the number of solutions to consider for each fragment , we clustered the 7863–3268 filtered poses at 3 Å , and selected the best-ranked pose in each of the 287–440 clusters obtained for 1B7F - 1CVJ respectively ., By assembling these fragments into chains , with weaker overlap-restraints , a total of 242 and 334 chain-forming poses were selected , among which 24% close-hits ( RMSD < 6 Å ) , all fragments in frag1-5 being well sampled ( Table 3 ) ., These poses could be assembled into 10064–4413 chains , with 3–2% close-hits ( geometric mean over frag1-5 RMSD < 6 Å ) ., Measured over the whole chain , the best precision was 5 . 7–3 . 6 Å , and this was sufficient to define both position and orientation of most of the 7 nucleotides in each complex ( Fig 6 ) ., The chain propensity filter selects an ensemble of poses or chains that is still rather large to carry out subsequent refinement steps ., Therefore , to reduce this number , we investigated if it is possible to assign a ranking to the poses within the selected ensembles ., We used combinations of three statistics: chain-propensity , the number of chains a pose participates in; ATTRACT rank , the rank of the pose according to the ATTRACT force field; and , for chains , overlap , the violation of the harmonic distance restrains between two consecutive poses in a chain ., We tried to rank the poses and chains obtained by biased and unbound docking , according to the scoring functions Sposes and Schains , based on these statistics ., Since we have only two test cases , we emphasize that the performance of such scoring functions should be considered as a proof-of-principle , and should be trained on a much larger benchmark before any predictive power can be credited ., Still , given these caveats , we found that the following scoring functions worked well:, Sposes=log ( chain−propensity ) ATTRACTrank, ( 1 ), Schains=∑0<i<NfragmentsOverlap ( posei , posei+1 ) ∑poses ( ATTRACTrank ) 2Nfragments, ( 2 ), For biased docking , the ranking proved efficient in selecting the best solutions , both at the poses and chains levels , for both complexes ( Fig 7 , S1 File ) ., The ranking was less efficient in selecting the best solutions for unbound compared to biased docking , as expected by the use of non-correct conformers in the docking ., Yet , it still achieved a statistically significant enrichment of good solutions in the best-ranked solutions , both at the poses and chains levels , for both complexes ., At the poses level: 178 of the 1038 poses with RMSD < 2 Å were ranked in the top 1000 out of 7862 for 1B7F and 5 of the 7 poses with RMSD < 1 . 5 Å were ranked in the top 1000 out of 3268 for 1CVJ ( p-value 6x10-6 and 0 . 03 ) ., At the chains level , 74 out of 309 chains with RMSD < 6 Å were ranked in the top 2000 out of 13693 for 1B7F , and 53 out of 119 in the top 2000 out of 6190 for 1CVJ ( p-values 4x10-5 and 0 , 001 ) ., Although below usual precision in classical whole-body docking , these results constitute a considerable improvement compared to the poor success that had been achieved so far in docking protein-ssRNA complexes ., To the best of our knowledge , essentially all current methods are limited to structured RNA ., Only the RNA-lim method 25 has attempted to predict protein-ssRNA structures based on fragments ., The authors dock and assemble ultra-coarse-grained nucleotides ( one bead per nucleotide ) in a pre-defined binding site ., The very small size and the simplistic model of the fragments greatly limit the accuracy of the results , partially compensated by the limitation of the search to the known binding site ., They achieved very limited success , sampling center-of-mass ( COM ) positions ( not orientation ) of six RNA mono-nucleotide fragments with ~10% “coarse starting estimates” ( ~ 5 Å on COM ) on a single test-case ., In contrast , our method correctly sampled both position and orientation for most nucleotides in an heptamer RNA on two test-cases ., Our method worked very well when the conformer library was biased towards the closest-to-bound conformers , achieving a best precision of ~1 . 5 Å at both the fragment level ( best fragment among dozens of poses ) and the chain level ( among a hundred of chains ) ., With fully unbound docking , this precision could only be achieved when large numbers ( 105–106 ) of poses where considered ., Filtering the number of poses down to a few thousands worked well for 1CVJ ( ~1 . 5 Å best precision ) but less so for 1B7F ( ~4 Å best precision ) ., At the chain level , our method achieved a best precision of 3 . 6–5 . 7 Å ( among thousands of chains ) ., In real cases , the chains could be further filtered by experimental data on specific contacts ( e . g . from protein mutagenesis or from RNA sequence specificity ) , especially at the extremities of the chains were the diversity in positioning among the chains is the highest ., To be successful , the building of chains needs each of the fragments to be correctly sampled ., But a correct sampling is possible only if the fragment establishes sufficient contacts with the protein ., To get an idea if the contacts of each fragment in a test-case are sufficient to make the correct sampling possible , one can perform a quick docking test with the bound form of each fragment ( without usage of the library nor chain building ) ., To evaluate the applicability of our methods to other protein families and other contact patterns , we performed such bound-bound fragment docking tests on 5 other RNA-protein complexes ( S2 File ) from different RNA-binding protein families and with different RNA binding modes ( S3 Fig , S2 File ) ., The docking was successful ( best docking pose within 4 . 0 Å RMSD ) for all fragments in all complexes , but for frag-1 in 3V6Y and 4KRF ( S2 Table ) ., Still , for those two complexes , 6 consecutive fragments could be well sampled ., Noteworthy , in 4PMW , the presence of a Mg2+ ion participating in the binding of frag-12 , which is not taken into account in our current method , did not affect the quality of the sampling ., In 3V6Y , the only failure is due to symmetries in the system ( S3 Fig , S2 File ) , the poses for frag-1 having RMSD < 4 . 0 Å when compared to the bound form of other fragments ., In 3V6Y , the presence of a bulged out nucleotide at n7 , in the center of the RNA strand , had no impact on the quality of the sampling for that fragment ., Poses were found with RMSD in 0 . 4–0 . 9 Å for the 3 fragments containing n7 ., Therefore , the absence of contacts at least for one nucleotide can be compensated by contacts made by the adjacent nucleotides ., When using a conformational library instead of the bound form , the bulged-out nucleotide is likely to be less well sampled , but our tests show that this would not affect dramatically the adjacent nucleotide , and consequently the building of chains ., However , the current state of our method is not able to distinguish binding from non-binding RNA fragments , but only to sample the correct positioning of a fragment , assuming that it does bind to the protein ., Therefore , the docking should be limited to the binding fragments ., In a real case , this data could be obtained experimentally , e . g . by comparing the in vitro affinity of RNA with different sequences or with modified bases , or by NMR data ( e . g . intermolecular NOE , differences in bound-unbound chemical shifts , H/D exchange rates ) ., In the present study , we focused on homopolymers for the convenience of rather short CPU time needed for the fragment docking as well as the selection of overlapping poses ., The total procedure took around 14h for each case , the docking of one unique conformational ensemble ( for UUU or AAA sequence ) being run on 8 CPU and the chains built on 1 CPU ., The two test-cases provide an important proof-of-principle for RNA-protein modeling , and the method is in principle extendable to arbitrary sequences ., The extension toward heteropolymer will require more CPU time , as each trinucleotidic sub-sequence will require the docking of the corresponding conformational ensemble ., However , the docking of the fragments are independent and can therefore be run in parallel ., The building of chains begins with the identification of pairs of compatible poses , which can be run in parallel for each pair of fragments ., The absolute time required by the method should therefore not be increased when applied on an heteropolymer rather than homopolymer sequence , if running on 8 CPU * Nb ( fragment ) , but would increase linearly with the length of the chain ., Specific binding of proteins to ssRNA participates in key post-transcriptional regulatory processes ., We developed a method to predict the structure of an RNA homopolymer complexed to a RRM-containing protein , based on the str | Introduction, Results and Discussion, Methods | Protein-RNA docking is hampered by the high flexibility of RNA , and particularly single-stranded RNA ( ssRNA ) ., Yet , ssRNA regions typically carry the specificity of protein recognition ., The lack of methodology for modeling such regions limits the accuracy of current protein-RNA docking methods ., We developed a fragment-based approach to model protein-bound ssRNA , based on the structure of the protein and the sequence of the RNA , without any prior knowledge of the RNA binding site or the RNA structure ., The conformational diversity of each fragment is sampled by an exhaustive RNA fragment library that was created from all the existing experimental structures of protein-ssRNA complexes ., A systematic and detailed analysis of fragment-based ssRNA docking was performed which constitutes a proof-of-principle for the fragment-based approach ., The method was tested on two 8-homo-nucleotide ssRNA-protein complexes and was able to identify the binding site on the protein within 10 Å ., Moreover , a structure of each bound ssRNA could be generated in close agreement with the crystal structure with a mean deviation of ~1 . 5 Å except for a terminal nucleotide ., This is the first time a bound ssRNA could be modeled from sequence with high precision . | Protein-RNA interactions fulfill a large variety of fundamental cellular functions , in particular for regulation of genome expression ., A full understanding of these interactions requires an atomistic description of the interface in the complex ., It can aid in silico design of new therapeutics to modulate these functions ., However , structure determination of these complexes can be costly and in many cases difficult due to the transient nature of many protein-RNA interactions ., Computational docking can help to generate structural models of protein-RNA interactions ., Traditional rigid body docking methods largely fail due to the flexibility especially of single-stranded ( ss ) RNA that often forms the binding region in protein-RNA complexes ., We developed an original approach to cope with ssRNA flexibility by assembling them from small structural fragments ., Tested on two known complexes , our method could model the ssRNA at a level of detail never reached so far ., These results constitute a proof-of-principle and major step towards designing a fully flexible RNA-protein docking methodology with a wide range of possible applications . | sequencing techniques, rna-binding proteins, crystal structure, protein interactions, condensed matter physics, protein structure prediction, protein structure, sequence motif analysis, molecular biology techniques, protein structure databases, crystallography, research and analysis methods, sequence analysis, rna structure, solid state physics, proteins, biological databases, molecular biology, physics, biochemistry, rna, nucleic acids, database and informatics methods, biology and life sciences, physical sciences, macromolecular structure analysis | null |
journal.pgen.1007120 | 2,017 | A conserved maternal-specific repressive domain in Zelda revealed by Cas9-mediated mutagenesis in Drosophila melanogaster | During the first hours following fertilization , the zygotic genome is transcriptionally silent , and maternally deposited products control early development ., These maternal products establish regulatory networks that enable the rapid and efficient transition from two specified germ cells to a population of totipotent cells , which give rise to a new organism ., This dramatic change in cell fate is coordinated with the transition from maternal to zygotic control of development , resulting in a complete reorganization of the transcriptome of the embryo ., The maternal-to-zygotic transition ( MZT ) is comprised of two essential and coordinated events , ( I ) transcriptional activation of the zygotic genome , and ( II ) destabilization and degradation of maternally supplied RNAs 1–4 ., The concerted action of two RNA clearance pathways ensures the timely elimination of maternally deposited transcripts 5–11 ., The first is a maternally encoded pathway that initiates the degradation of maternal RNAs in the absence of fertilization and zygotic transcription ., The second pathway is zygotically triggered and contributes to maternal RNA clearance near the end of the MZT ., Thus , transcriptional activation of the zygotic genome is precisely coordinated with degradation of the maternally provided products 5 , 10 , 12 ., Regulation of these events is required for development , as failure to undergo this transition is lethal to the embryo ., Nonetheless , the mechanisms that precisely control the timing and levels of gene expression necessary to successfully navigate this dramatic developmental transition remain to be elucidated ., In Drosophila melanogaster , the MZT occurs over the first few hours of development ., The transcription factor Zelda ( ZLD; Zinc-finger early Drosophila activator ) is a critical regulator of the MZT , and its absence is lethal to the embryo 13–17 ., zld transcripts are maternally deposited and robustly translated following fertilization leading to ubiquitous protein expression in the pre-blastoderm embryo 14 , 17 , 18 ., ZLD binds to thousands of cis-regulatory modules and is required for transcriptional activation of the zygotic genome 13–15 ., ZLD is necessary for gene expression both early and late during the MZT; ZLD drives expression of a small number of genes as early as the eighth mitotic division and is required for the later activation of hundreds of genes during the major wave of zygotic genome activation at mitotic cycle 14 13 ., Among the genes that require ZLD for expression are components of the RNA degradation pathways that destabilize maternal RNAs 14 , 16 ., These ZLD-target genes include several zygotically expressed miRNAs and lncRNAs , including the miR-309 cluster of miRNAs that mediates degradation of over one hundred maternally loaded RNAs 16 , 19 ., Thus , maternally supplied zld is essential for zygotic genome activation and maternal mRNA decay , driving the coordinated transition from maternal to zygotic control ., ZLD is also required zygotically , such that embryos homozygous for a deletion in zld die late in embryogenesis 14 , 17 ., Maternally deposited zld encodes a protein of 1596 amino acids , including six C2H2 ( Cys-Cys-His-His motif ) zinc fingers , but no known catalytic activity ( Fig, 1 ) 14 , 17 , 20 ., In tissue culture , ZLD is a robust transcriptional activator , and this function requires the C-terminal cluster of four zinc fingers that comprise the DNA-binding domain and a low-complexity region proximal to this domain 20 ., Functional data combined with phylogenetic analysis supports a shared role for ZLD in genome activation among insects and crustaceans 20–25 ., Thus , we were surprised to discover that while transcriptional activation is a conserved function of ZLD , in cell culture this activity does not require highly conserved regions in the N-terminus , including two of the C2H2 zinc fingers and an acidic patch 20 , 25 ., A truncated splice isoform of zld is also conserved throughout the Drosophila genus ., This variant is expressed in late embryos and in larvae , but lacks coding sequence for three of the four C-terminal zinc fingers in the DNA-binding domain and is therefore unable to bind DNA ( Fig 1A ) 20 , 26–28 ., Conservation of these additional domains and splice isoforms suggests a potential function that has been retained through evolution , but which may not have been evident in cell culture ., Our recent development of techniques for Cas9-mediated genome engineering in Drosophila enabled us to directly test the roles of these conserved features of ZLD in vivo 29 ., Previous approaches to investigate the in vivo function of specific protein domains relied largely upon the use of transgenes , which do not always adequately reflect the endogenous expression patterns , levels , or alternative splice isoforms ., We therefore developed a rapid and efficient means to screen for Cas9-mediated point mutations ., Generation of specific point mutations allowed us to interrogate the function of conserved features of zld in vivo ., Using a combination of epitope tags and targeted deletions , we demonstrated that a truncated zld isoform was unlikely to be translated and was not required for viability in D . melanogaster , despite being conserved amongst Drosophilidae ., We generated targeted loss-of-function alleles for conserved domains in the N-terminus , including the two zinc fingers and the acidic patch ., Mutations in either the first C2H2 zinc finger ( ZnF1 ) or the acidic patch ( EDD ) did not affect viability ., To our surprise , the second zinc finger ( ZnF2 ) was required for maternal , but not zygotic function of ZLD ., Embryos laid by mothers homozygous for mutations in the second zinc finger died late in embryogenesis ., Contrary to our expectations , mutations in ZnF2 resulted in a hyperactive version of ZLD that caused precocious activation of the zygotic genome and increased degradation of maternal transcripts ., Together these data demonstrate , for the first time , a separable function for maternally and zygotically expressed ZLD and suggest that the early embryo is exquisitely sensitive to ZLD activity such that too little or too much activity results in embryonic lethality ., zld transcripts are present throughout the Drosophila life cycle ., They are strongly expressed during oogenesis , resulting in ubiquitous protein expression in the pre-blastoderm embryo ., Subsequently , zld is zygotically expressed in the developing embryonic germ layers , nervous system , imaginal disc primordia and in larval wing and eye discs 17 , 27 , 28 , 30 ., In the pre-blastoderm embryo zld contains a single , unspliced open reading frame and a single five-prime intron ( Fig 1A ) ., This open reading frame translates into the 1596 amino acid protein product ZLD-PB ., In addition to this maternally deposited isoform , a truncated isoform , zld-RD , which contains a second unique downstream exon and alternative splice junction , is expressed during zygotic development 26–28 ., This alternatively spliced isoform codes for a 1373 amino acid protein lacking three of the four C-terminal zinc finger motifs required for DNA binding ( Fig 1A ) 20 , 31 ., The truncated product resulting from translation of the zld-RD isoform acts as a dominant negative when co-expressed with the 1596 amino acid isoform in cell culture 20 ., Nonetheless , it was unknown whether this shorter isoform was translated to form a protein product in vivo and if so , whether it was expressed in the same cell as the longer 1596 amino acid isoform , which would be required for any dominant negative effect on ZLD activity ., To determine the expression pattern of a protein product from the zld-RD isoform , we used Cas9-mediated genome engineering to tag each of the two protein isoforms with mCherry ., Because we had previously shown that the N-terminal 900 amino acids of ZLD are dispensable for activating transcription in cell culture 20 , we tagged the N-terminus to avoid interfering with protein function ., Flies carrying this mCherry tag are homozygous viable and fertile demonstrating that the tag does not interfere with any of the essential functions of ZLD ., Since all zld splice isoforms encode proteins with a shared N-terminus , expression of the N-terminal mCherry-tagged protein is indicative of the expression pattern of all known ZLD isoforms ., To specifically determine the expression pattern of a protein product of the shorter zld-RD isoform , we engineered an mCherry tag upstream of the stop codon in the downstream exon that is specific to zld-RD ( Fig 1A ) ., This addition did not affect the zld-PB isoform ., Like flies carrying the N-terminal fluorescent tag , these flies were also homozygous viable and vertile ., We imaged stage 5 , 12–13 , and 14–16 embryos homozygous for either the N-terminal mCherry tag or the zld-RD specific mCherry tag to determine the expression patterns of ZLD protein products ( Fig 1B ) ., zld-RB is ubiquitously expressed in the pre-blastoderm embryo , while post-blastoderm expression is limited to the tracheal primordium , central nervous system ( CNS ) , and midline neurons 14 , 27 ., Similar to the expression pattern for the mRNA , the N-terminally tagged protein was expressed throughout embryogenesis ( Fig 1B ) ., We did not detect fluorophore expression from either of the two strains containing the mCherry-tagged zld-RD ( Fig 1B ) ., Thus , despite high levels of zld-RD in the CNS of stage 12–16 embryos 27 , the ZLD-PD isoform does not appear to be expressed ., We detected gut auto fluorescence in all genotypes , including control w1118 embryos ., To investigate additional tissues that might express ZLD-PD at later stages of development , we imaged imaginal wing discs from third instar larvae ( L3 ) ., Previous reports had suggested that ZLD-PD was expressed in larval tissues 28 ., Using our engineered fly lines , we could detect ZLD expression in several L3 tissues , including ubiquitous , nuclear expression in imaginal wing discs ( Fig 1C ) ., By contrast , we could not detect mCherry expression in the lines specifically tagging ZLD-PD ( Fig 1C ) ., In addition to zld-RD , a second splice isoform of zld , zld-RF , has been identified and is predicted to produce a protein product very similar to the predicted product of zld-RD ., The evidence for zld-RF is weaker than for zld-RD , and it has been speculated to be the result of a cloning artifact 26 , 27 , 32 ., Nonetheless , both truncated isoforms have been reported to be expressed in the wing imaginal disc 28 ., To determine if any truncated protein product is translated from either the zld-RD or zld-RF isoforms , we immunoblotted protein extract from wing imaginal discs using our antibody that recognizes all isoforms of ZLD 20 ., We identified only a single isoform , corresponding to the 1596 amino acid protein ( Fig 1D ) ., This evidence suggests that neither zld-RD nor zld-RF isoforms are translated to a stable protein in the wing imaginal disc ., The zld-RD splice isoform is conserved throughout the Drosophila genus 26 , 27 , suggesting a retained function ., Thus , it remained possible that the truncated RNA or the splicing reaction was instrumental to zld function and could explain the conservation , even if the protein isn’t stably expressed ., To test this possibility , we determined the in vivo effects of eliminating the zld-RD isoform by using Cas9-mediated mutagenesis to delete the splice acceptor and downstream coding region of zld-RD ( Fig 1E ) ., We obtained two strains carrying the deleted sequence , both of which were viable and fertile ( Fig 1E ) ., Because the exons encoding zld-RF are within the required longer isoform , we were unable to make a deletion targeting only this isoform as we did for zld-RD ., Therefore , we cannot rule out the possibility that this isoform is important in vivo ., While zld-RD is expressed in multiple post-blastoderm tissues as an RNA 26–28 , our data demonstrated that this truncated splice-isoform is not required for development and is not abundantly translated ., Because the 1596 amino acid ZLD-PB isoform is the predominantly expressed form of ZLD , we investigated the functional requirements of domains within this large transcription factor ., ZLD-PB is comprised of six C2H2 zinc fingers and many low-complexity regions , but no identifiable enzymatic domains ., Alignment of ZLD orthologs showed sequence conservation within insects of all six zinc fingers as well as an N-terminal acidic patch ( Fig 2A–2C ) 20 , 25 ., We previously demonstrated that the cluster of four C-terminal zinc fingers constituted the DNA-binding domain , and the low-complexity domain just N-terminal to the DNA-binding domain mediated transcriptional activation ( Fig 1A ) 20 ., Therefore , the regions required for both DNA binding and transcriptional activation in cell culture were encompassed within the 600 C-terminal amino acids of ZLD 20 , while the functional significance of the conserved N-terminal zinc-fingers and acidic domain was unknown ., Because domains under high evolutionary constraint possess important structural or functional roles , we hypothesized that these highly conserved domains might have an essential developmental function that was missed in our previous cell-culture assays ., The recent development of Cas9-mediated genome editing has allowed us to facilely create point mutants in vivo ., This strategy enabled efficient creation of endogenous mutant alleles to probe the functional importance of individual protein domains of interest ., A single-stranded donor oligonucleotide ( ssODN ) and a single guideRNA ( gRNA ) construct were injected into flies expressing Cas9 to create loss-of-function mutations in the highly conserved N-terminal domains ., We developed a streamlined protocol to molecularly screen for the desired mutations; ssODNs contained both the desired coding mutations and silent mutations that generated a restriction digest site not found in the endogenous locus , allowing for screening by PCR and restriction enzyme digest ( Fig 2D and 2E ) ., Instead of creating deletions , we introduced point mutations in the conserved N-terminal domains with the purpose of maintaining overall protein stability ., To disrupt the zinc-finger domains , we mutated a subset of the zinc-chelating residues in each of the N-terminal zinc fingers ., Within the acidic domain , we mutated the conserved glutamate and aspartate residues to alanine to abrogate the negative charges in the domain ., Using this streamlined strategy , we generated three distinct mutant alleles , individually disrupting each of these conserved protein domains and allowing us to interrogate protein structure and function in vivo ( Fig 2E ) ., We assessed the viability and fertility of each of the mutants we generated by counting the number of homozygous flies carrying the mutations as compared to heterozygous siblings ( Fig 3 ) ., Flies homozygous for mutation of either the first zinc finger ( ZnF1 ) or the acidic domain ( EDD ) were viable to near wild-type levels ( Fig 3A ) ., Both homozygous males and females were fertile ., Similarly , hemizygous males carrying mutations in zinc finger 2 ( ZnF2 ) were viable , albeit to a reduced degree , and were fertile ., Homozygous ZnF2 mutant females were viable at reduced levels , but , in contrast to their male counterparts , were sterile ( Fig 3A ) ., Mutations in ZnF2 resulted in a maternal-effect lethal phenotype in which homozygous zldZnF2 females lay fertilized embryos that arrest late in embryogenesis during stage 17 after tracheal branches have clearly formed ., A subset ( 18% ) of male and female adults homozygous for zldZnF2 had small , malformed eyes , suggesting additional developmental processes were affected by the mutation ., Thus , contrary to our expectations loss-of-function mutations in all three regions were dispensable for development to adulthood even though their conservation suggested they were required for ZLD function ., Our mutational analysis delineated discrete maternal and zygotic functions for ZLD; maternal deposition of zld mutant for ZnF2 was lethal to the embryo , causing arrest late in embryogenesis , whereas zygotic expression of zld with a disruption in ZnF2 supported development to adulthood ., To determine if protein stability was disrupted in any of the mutants , we compared expression of each mutant allele to a GFP-tagged endogenous , wild-type allele in heterozygous embryos ., Embryos were laid by heterozygous females such that they received maternal deposition of RNA encoding both ZLD protein variants ., Equivalent amounts of protein were expressed from alleles carrying mutations in ZnF1 , ZnF2 , or the acidic domain as compared to the GFP-tagged control ( Fig 3B ) ., Furthermore , the observed phenotype for the ZnF2 mutant was retained in trans-heterozygotes carrying a deletion in zld ( zld294 ) ( S1 Table ) ., Thus , the maternal-effect lethality associated with maternal inheritance of zld mutant for ZnF2 was not a result of protein instability or a background mutation , but instead was the result of changes in ZLD function ., To test the functional importance of these conserved domains on ZLD-mediated transcriptional activation , we used our previously established cell-culture system to transiently express ZLD mutants and assay their ability to activate luciferase reporters 20 ., ZLD with mutations in either ZnF1 or the acidic domain was able to activate transcription to a level similar to the wild-type protein ( Fig 4 ) , consistent with these mutations producing viable and fertile flies ( Fig 3A , S1 Table ) ., The single amino acid change ( C554S ) in the ZnF2 mutant significantly hyperactivated the scute reporter , resulting in luciferase activity at least 3-fold greater than wild type ., None of the ZLD proteins activated gene expression from a mutant promoter , confirming the specificity of the assay ., Immunoblots confirmed the expression of all mutated proteins was at approximately equivalent levels ( Fig 4 ) ., These data suggest that ZnF2 may serve as an inhibitory domain that regulates the level of ZLD-mediated transcriptional activation ., This conclusion is further supported by our previous data demonstrating that truncations to the N-terminus of ZLD that remove ZnF2 elevated transcriptional output 20 ., Prior studies demonstrated that either overexpression of maternal zld or the loss of maternally deposited zld resulted in defects in nuclear division in the blastoderm embryo 17 ., The similarity of the loss-of-function and overexpression phenotypes suggests that the early embryo is sensitive to the precise levels of ZLD activity and that both too little and too much activity is detrimental to embryonic development ., To confirm the impact of ZLD overexpression , we used mat-α-GAL4 to drive overexpression of a UASp-zld transgene ., Overexpression of maternally deposited zld caused a late embryonic lethal phenotype similar to that of animals inheriting maternal zldZnF2 , albeit at a lower frequency ( S1 Fig ) ., Based on our tissue culture data and the fact that zld overexpression phenocopies the ZnF2 mutation , we propose that disruption of zinc finger 2 hyperactivated ZLD protein and that this increased activity was lethal to the embryo ., The second zinc finger in ZLD is the most highly conserved domain in the entire 1596 amino acid protein 25 ., To determine whether conserved residues outside of the zinc-chelating amino acids of zinc finger 2 were required for function , we mutated four conserved residues ( F561 , S563 , Y571 and N578 ) to alanine , generating the zldJAZ allele ( Fig 5A–5C ) ., These residues were chosen because they are shared between the second zinc finger in ZLD and the consensus sequence for JAZ ( Just Another Zinc finger ) -domains ( pfam: zf-C2H2_JAZ ) , a domain initially identified in the mammalian double-stranded RNA-binding zinc finger protein JAZ ( Fig 5A ) 33 ., Alanine substitutions in the JAZ domain did not affect protein stability ( Fig 5D ) ., Animals homozygous for zldJAZ were viable at reduced frequencies , and males were fertile ( Fig 5E ) ., Females homozygous for this allele were sterile , laying embryos that later died ( Fig 5E ) , phenocopying the zldZnF2 mutants ( Fig 3A , S1 Table ) ., Cell-culture assays further demonstrated that mutating the JAZ-like domain hyperactivated transcription , similar to the serine substitution in C554 ( Fig 5F ) ., Thus , both zinc-chelation and residues conserved within the JAZ zinc finger domain are critical for negatively regulating the ability of ZLD to activate transcription ., The cell-culture assays demonstrated that the JAZ-like ZnF2 negatively regulated ZLD activity ., This raised the possibility that the lethality in embryos inheriting zld with mutations in this domain might result from hyperactivation of ZLD targets ., To determine the functional consequences of the zldZnF2 allele on early gene expression , we performed mRNA-sequencing on hand-sorted stage 5 embryos with wild-type maternal zld ( w1118 ) or zldZnF2 ( C554S ) ., The high degree of reproducibility amongst the three replicates ( S2 Fig ) allowed us to identify genes misexpressed in embryos inheriting the mutated version of zld ( Fig 6 ) ., We identified 287 genes that were up-regulated in the zldZnF2 mutant and 270 genes that were down-regulated ( Fig 6A and S2 Table ) ., Stage 5 embryos possess both mRNAs that have been deposited by the mother along with newly transcribed zygotic mRNAs ., To distinguish between these two classes of mRNAs , we used previously published data to determine whether the mis-regulated genes were maternally , zygotically or both maternally deposited and zygotically expressed ( mat-zyg ) 34 ., 74% ( n = 212 ) of up-regulated genes were zygotically expressed , including those expressed exclusively in the zygote and those maternally deposited and zygotically expressed ( Fig 6B ) ., By contrast , only 11 . 5% ( n = 31 ) of the genes that were down-regulated were zygotically expressed ., 73% ( n = 197 ) of the down-regulated genes were maternally deposited ( Fig 6B ) ., Thus , the majority of the up-regulated genes were zygotically expressed while the majority of the down-regulated genes were maternally deposited ., ZLD is required for transcriptional activation of hundreds of zygotic genes during early embryogenesis 13–15 , 35 ., Thus , we tested whether the hyperactive zldZnF2 allele up-regulated expression of direct ZLD-target genes ., We used our previous ZLD ChIP-seq data to identify ZLD-bound regions in the stage 5 embryo and associated them with the nearest gene to identify 3836 potential direct ZLD targets 13 ., More than half of the genes up-regulated in embryos inheriting the zldZnF2 allele overlapped with likely direct ZLD targets ( Fisher’s exact test , p < 0 . 0001 ( S2 Table ) ) , suggesting that these genes were directly hyperactivated by the mutant ZLD protein ( Fig 6C ) ., To determine the regulatory networks influenced by ZLDZnF2 hyperactivity , we identified enriched Gene Ontology ( GO ) terms for the 108 likely direct targets ., The most enriched GO terms were related to transcription-factor activity , DNA binding , and RNA Pol II activity ( Fig 6D ) ., Misexpression of these genes may therefore affect multiple downstream processes required for embryonic development , ultimately leading to the late-stage lethality of these embryos ., Zygotic gene activation is coordinated with the degradation of maternally deposited RNAs during the MZT ., Two sets of machinery remove maternally deposited transcripts from the early embryo with one functioning just after fertilization and one functioning later during genome activation 2 , 5 ., The early decay pathway is encoded by maternal factors and is triggered by egg activation ., The late-decay pathway is encoded by zygotic factors expressed at the onset of zygotic genome activation 10 , 12 ., Because we found an enrichment for maternally deposited mRNAs amongst the down-regulated transcripts in the embryos inheriting maternal zldZnF2 , we hypothesized that mRNAs in these mutant embryos might be precociously degraded due to hyperactivation of the zygotic genome ., To test this , we determined whether the down-regulated mRNAs corresponded to genes subject to either the early or late decay pathways 10 , 12 ., 58% ( n = 114 ) of the down-regulated maternal mRNAs overlap with mRNAs subject to degradation late during the MZT , while just 1 . 5% ( n = 3 ) were degraded early in the MZT ( Fig 6E ) ., These data support a model whereby ZLDZnF2 hyperactivates a set of zygotic genes and that this leads to precocious decay of a set of maternally deposited mRNAs ( Fig 6F ) ., D . melanogaster have been a premier organism for studies of gene regulation and development for over a century , but studies have been limited by the inability to precisely engineer mutations in the genome using homologous recombination ., Our establishment of Cas9-mediated genome engineering in D . melanogaster overcame this limitation 29 ., Here we have used this facile method of gene editing to identify the functional domains of the essential transcription factor ZLD ., We developed a molecular screening strategy that enabled us to generate four distinct mutations to directly query the necessity of conserved protein domains ., Editing the endogenous locus provided confidence that any phenotypes we observed were not due to differences in levels or localization of gene expression that might result from the use of a transgene ., This was supported by confirmation that all the mutations we generated were expressed at endogenous levels ., Thus , we are confident that the absence or presence of a clear mutant phenotype represented the endogenous requirement for specific protein sequences ., Our use of genome editing to determine the requirements for specific protein domains within ZLD highlights more generally the power of Cas9-mediated editing to characterize protein structure and function ., The easy PCR-based screening approach described here allows for the generation and identification of novel alleles in as little as one months time , providing an additional powerful tool to study gene function in Drosophila ., Cas9-mediated genome editing also enabled us to specifically determine the protein expression pattern from a conserved splice isoform of zld that is predicted to produce a truncated protein product ., Using a combination of an isoform-specific mCherry tag , a targeted deletion , and immunoblot , we clearly demonstrated that while zld-RD may be expressed as an RNA it is not translated at detectable levels in either the embryo or the larval wing disc and is not required for viability ( Fig 1 ) ., Visualization of an N-terminal mCherry tag that marks all possible ZLD isoforms demonstrated that ZLD is expressed in embryos well past the MZT ( Fig 1B–1D ) ., Thus , the 1596 ZLD-PB isoform that binds DNA and drives transcriptional activation is likely the predominant protein product at all stages of development and in all cell types ., A single zld ortholog with a set of highly conserved domains is found within the genomes of insects and some crustaceans ., These ZLD orthologs are required for embryonic development and transcriptional activation within multiple insect species 20 , 25 , 36 ., We had previously shown that ZLD-mediated transcriptional activation in Drosophila cell culture did not require either of the conserved N-terminal C2H2 zinc fingers or a recently identified conserved acidic patch 20 , 25 ., Here , we used Cas9-mediated genome editing to test the functional significance of these conserved domains in vivo by generating point mutations that were likely to result in loss of function ., We individually mutated both conserved N-terminal zinc fingers as well as the acidic patch ( Fig 2 ) ., Because coordination of zinc ions plays an essential structural role in zinc finger domains 37 , the cysteine-to-serine mutations are likely to lead to structural changes that abrogate function ., Similarly , removing acidic residues from the acidic patch is likely to disrupt any interactions that rely on the negative charge of these residues ., For example , it has been suggested that this negatively charged domain might contact positively charged histones 25 , and the alanine substitutions would be expected to block this interaction ., Mutation of either the first zinc finger or the acidic patch did not disrupt the ability of ZLD to activate transcription in culture ( Fig 4 ) , consistent with our previous cell-culture studies 20 , and mutant flies homozygous for these mutations were viable and fertile ( Fig 3 ) ., These domains are therefore not necessary for ZLD-mediated transcriptional activation in D . melanogaster ., In contrast to the high degree of sequence conservation of the six C2H2 zinc fingers present in D . melanogaster ZLD , sequence alignments have identified an additional zinc finger ( ZF-Novel ) in the N-terminus of ZLD orthologs in multiple insect species that has been eroded in Drosophila 25 ., Thus , this novel zinc finger domain may have functions specific to other species ., It will be interesting to investigate whether there is a phenotypic consequence of introducing such a zinc finger into D . melanogaster ZLD ., Nonetheless , the high-degree of conservation amongst the six zinc fingers maintained in D . melanogaster suggests they have a function that has led to their retention ., Our data demonstrated that specific conserved residues within these domains are not required for viability or fertility ., It remains possible that additional residues within these conserved domains are sufficient for functionality , that these domains have redundant functions , that they are instrumental in other as of yet unidentified functions for ZLD or that they serve as a buffer against environmental perturbations during development ., We demonstrated that the second zinc finger of ZLD is required for female fertility ., This zinc finger is the most highly conserved domain of the entire protein and has similarity to the double-stranded RNA-binding , JAZ-like zinc finger family 17 , 20 , 25 , 33 ., We generated two distinct loss-of-function alleles by mutating either a required zinc-chelating cysteine or four residues that are shared with the JAZ zinc finger domains ., These mutations resulted in maternal-effect lethality due to an increase in ZLD-mediated transcriptional activation ( Figs 4–6 ) ., Thus , we propose that this domain suppresses the ability of ZLD to activate transcription ., While ZnF2 has the canonical architecture of the JAZ-like C2H2 zinc finger , it lacks positively charged lysine residues that are conserved in double-strand RNA-binding zinc fingers and are thought to be required for RNA binding 38 ., Therefore , it is unlikely that this domain functions through interaction with double-stranded RNA ., Co-immunoprecipitations failed to identify homotypic interactions between differently tagged ZLD molecules , suggesting that ZLD does not multimerize 20 ., Nonetheless , it remains possible that this domain could inhibit ZLD activity by preventing multimerization ., It is likely that this domain interacts with a protein partner , a nucleic acid , or intramolecularly within ZLD to effect its suppression of transcriptional activation ., To date , no such interactions within ZLD or between ZLD and a protein or RNA partner have been identified ., zld is required as a maternally deposited mRNA that is translated following fertilization ., Embryos lacking maternally deposited zld die early in embryogenesis due to a failure to undergo the MZT 14 ., Embryos homozygous mutant for zygotic zld die late in embryogenesis 14 , 17 , but the cause of this lethality is currently unknown ., While ZLD is required throughout embryogenesis for viability , it was previously unclear if there were functions distinctly required at either stage of development ., The maternal-effect lethality we demonstrate for | Introduction, Results, Discussion, Materials and methods | In nearly all metazoans , the earliest stages of development are controlled by maternally deposited mRNAs and proteins ., The zygotic genome becomes transcriptionally active hours after fertilization ., Transcriptional activation during this maternal-to-zygotic transition ( MZT ) is tightly coordinated with the degradation of maternally provided mRNAs ., In Drosophila melanogaster , the transcription factor Zelda plays an essential role in widespread activation of the zygotic genome ., While Zelda expression is required both maternally and zygotically , the mechanisms by which it functions to remodel the embryonic genome and prepare the embryo for development remain unclear ., Using Cas9-mediated genome editing to generate targeted mutations in the endogenous zelda locus , we determined the functional relevance of protein domains conserved amongst Zelda orthologs ., We showed that neither a conserved N-terminal zinc finger nor an acidic patch were required for activity ., Similarly , a previously identified splice isoform of zelda is dispensable for viability ., By contrast , we identified a highly conserved zinc-finger domain that is essential for the maternal , but not zygotic functions of Zelda ., Animals homozygous for mutations in this domain survived to adulthood , but embryos inheriting these loss-of-function alleles from their mothers died late in embryogenesis ., These mutations did not interfere with the capacity of Zelda to activate transcription in cell culture ., Unexpectedly , these mutations generated a hyperactive form of the protein and enhanced Zelda-dependent gene expression ., These data have defined a protein domain critical for controlling Zelda activity during the MZT , but dispensable for its roles later in development , for the first time separating the maternal and zygotic requirements for Zelda ., This demonstrates that highly regulated levels of Zelda activity are required for establishing the developmental program during the MZT ., We propose that tightly regulated gene expression is essential to navigate the MZT and that failure to precisely execute this developmental program leads to embryonic lethality . | Following fertilization , the one-celled zygote must be rapidly reprogrammed to enable the development of a new , unique organism ., During these initial stages of development there is little or no transcription of the zygotic genome , and maternally deposited products control this process ., Among the essential maternal products are mRNAs that encode transcription factors required for preparing the zygotic genome for transcriptional activation ., This ensures that there is a precisely coordinated hand-off from maternal to zygotic control ., In Drosophila melanogaster , the transcription factor Zelda is essential for activating the zygotic genome and coupling this activation to the degradation of the maternally deposited products ., Nonetheless , the mechanism by which Zelda functions remains unclear ., Here we used Cas9-mediated genome engineering to determine the functional requirements for highly conserved domains within Zelda ., We identified a domain required specifically for Zelda’s role in reprogramming the early embryonic genome , but not essential for its functions later in development ., Surprisingly , this domain restricts the ability of Zelda to activate transcription ., These data demonstrate that Zelda activity is tightly regulated , and we propose that precise regulation of both the timing and levels of genome activation is required for the embryo to successfully transition from maternal to zygotic control . | invertebrates, messenger rna, invertebrate genomics, animals, dna transcription, animal models, developmental biology, mutation, drosophila melanogaster, model organisms, experimental organism systems, embryos, drosophila, research and analysis methods, embryology, proteins, gene expression, animal genomics, insects, arthropoda, biochemistry, rna, point mutation, eukaryota, nucleic acids, genetics, protein domains, biology and life sciences, genomics, organisms | null |
journal.pntd.0002355 | 2,013 | Culling Dogs in Scenarios of Imperfect Control: Realistic Impact on the Prevalence of Canine Visceral Leishmaniasis | Visceral leishmaniasis or kala-azar is the most severe clinical form of leishmaniasis , a serious public health problem worldwide 1 , 2 ., In Latin America , the agent of the so-called American visceral leishmaniasis ( AVL ) is Leishmania ( Leishmania ) chagasi transmitted , in Brazil , mainly by sandfly Lutzomyia longipalpis Lutz & Neiva , 1912 ., So far , the findings related to the epidemiology of AVL point to a spatial correlation between the occurrence of disease in humans and high rates of infection in dogs , suggesting that , in the presence of the vector , canine visceral leishmaniasis is a key factor for triggering transmission to humans 3 ., Overall , the incidence of AVL remains high despite the large-scale control strategies that have been implemented ., These strategies focus on early diagnosis and treatment of human cases , vector control to reduce sandfly population , as well as the removal of infected dogs and health education 4 ., Despite the lack of solid evidence in literature , culling dogs with canine visceral leishmaniasis ( CVL ) has been the major strategy for controlling this disease in Brazil ., Many authors argue that this strategy has low cost-benefit and many are against it , often encouraging the non-delivery of animals to slaughter 5 , 6 , 7 , 8 , 9 ., Other professionals , however , admit that this strategy can produce positive results 6 , 10 , 11 ., Two possible factors associated with the low effectiveness of culling programs are: ( 1 ) the discontinuity of these programs , which may occur for several reasons , including the lack of a structured surveillance system , budget issues and lack of adequately trained professionals; ( 2 ) Problems related to the logistics in delivering control measures , for example , low infected dog screening rates and lack of a reliable and valid diagnostic test to detect dogs in the early stages of infection , leaving out asymptomatic infectious dogs that are capable of conveying the parasite to the vectors , thus , cooperating with the continuity of the transmission cycle 12 , 13 ., Mathematical modeling has been applied in studies of visceral leishmaniasis in order to understand the transmission dynamics of this infection 8 , 14 , 15 , 16 , 17 , 18 , 19 and the impact of control strategies ., Hasibeder et al . ( 1992 ) and Dye et al . ( 1992 ) proposed and implemented models to estimate the basic reproduction number ( R0 ) of CVL , which was estimated between 1 . 44 and 11 , this large uncertainty being attributed to the poor performance of the available diagnostic tests ., Their model predicts that in areas where R0 is at the upper limit of the R0 range , culling would be successful only if intensively implemented ., In real settings , however , R0 estimation is highly uncertain as it depends on how the seropositivity is measured , and on the many assumptions of the underlying model , such as the homogeneous exposure of dogs to sandflies and homogeneous response to infection ., Later , Dye ( 1996 ) 20 alerted that culled dogs tend to be rapidly substituted by younger and susceptible ones , reducing the effectiveness of this strategy , compared to alternatives such as vector control , drugs and vaccines ., Other studies have explored the effectiveness of imperfect control programs , assessing the effect of imperfect diagnostic tests 24 , and discontinued dog culling programs 15 ., They found that a high sensitivity test , together with the immediate sacrifice , was sufficient to control the disease ., On the other hand , with a low sensitivity test , the effectiveness of the program was lost , whether or not the dogs were sacrificed immediately ., This paper seeks to revisit this problem , focusing on the relevance of asymptomatic infections in a scenario of imperfect control characterized by sub-optimal screening , diagnosis and slaughter rates ., We further investigated an unexplored component that is the impact of the low specificity of the diagnostic test ., We hope to contribute to the understanding of CVL transmission dynamics and the factors that modulate the control effectiveness ., An expression for the basic reproduction number of CVL was derived from the SEI2D model , without control , using the next generation matrix method 24 ., The mathematical derivation is found in the appendix ( Text S1 ) ., The Basic Reproduction Number is: ( 9 ) For modeling purposes , the intervention program was divided into three components: screening , diagnosis and sacrifice ., Screening measures the monthly capture rate and application of the diagnostic test to dogs in the population ., ( 10 ), The parameter d represents the probability of a dog to be positively diagnosed given it has been subjected to a diagnostic test and is infected ( test sensitivity ) ., ( 11 ), Once positively diagnosed , the dog has a chance f of being put down ., The delay between the screening and the sacrifice is 1/u ., The product r x dz measures the rate of misclassification of uninfected dogs ., This rate depends on r ( screening rate ) and dz which corresponds to the tests probability of false positive ( 1- specificity ) ., By varying these parameters , r , d , dz , u e f , one can investigate the impact of a variety of imperfect control programs ., Here , we considered variations of two hypothetical programs , both of which have been continuously implemented for 40 years ., The first one was based on data from the CVL control program implemented in Belo Horizonte , Brazil , considered to be of good quality , within the possibilities of the country ( data provided by the Subcoordenation of Vector Transmitted Zoonosis and Rabies/SVS/MS ) ., In this program , the screening rate is 6% per month followed by the immediate sacrifice of 85% of the screened dogs with positive diagnosis ., We implemented this scenario , assuming a diagnostic test with 90% sensitivity and 100% specificity ., A second scenario was built representing a worse situation , in which the screening rate is 4% per month and time to culling is 4 months as in Courtenay et al . ( 2002 ) ., In this scenario , diagnostic tests were applied with sensitivity and specificity ranging from 80% to 100% ., In both programs , we investigate two protocols: one targeting exclusively symptomatic dogs and screening all dogs , regardless of the presence or absence of symptoms ., To investigate the impact of diagnostic tests with low specificity , we compared the number of erroneously culled dogs by programs using tests with specificity of 80 , 90 and 100% ., By quantifying the number of dogs that were needlessly put down , we have a measure of the negative impact of the control strategy ., The effectiveness of the control programs was assessed by comparing the prevalence before and after 40 years of the establishment of the Control Program ., The control program was considered successful if it were capable of reducing CVL prevalence below 1% ., Considering that prevalence is measured by imperfect diagnostic methods , we further distinguished between real success and perceived success ., Real success is achieved when the true prevalence decreases below 1% , while perceived success is achieved when the measured prevalence decreases below this threshold ., At last , to investigate the impact of uncertainties in the specification of model parameters in the success of the control programs , we performed a multivariate uncertainty and sensitivity analysis ., The procedure was as follows: First , uniform probability density functions were defined for each life-history parameter ( i , qa , qb , p , λ , a , μ ) with intervals equal to 0 . 75 and 1 . 25 times the default parameter value ., Secondly , one thousand values were draw from each of these distributions , producing 1000 sets of parameters ., To maintain the transmission constant , for each new set of parameters , β was calculated from the R0 equation so that the R0 of all the simulations was kept at the same level ., After running the model SEI2D with each set of parameters , we recorded the success of the control program after 40 years as positive if final prevalence was less than 1% and failure otherwise ., Using the expression of R0 derived from the SEI2D model and the parameter values presented in Table 1 , we obtained R0\u200a=\u200a1 . 09 for the low endemicity scenario and R0\u200a=\u200a1 . 29 for the high endemicity scenario ., These values are low compared with those reported by other authors but were based on prevalence observed in the field 21 , 22 ., In the sensitivity analysis section , we discuss scenarios with higher R0 ., According to the SEI2D model , a CVL control program with 6% monthly screening rate , a diagnostic test with 90% sensitivity , and no delay between screening and culling should be effective in controlling CVL if implemented continuously for 40 years , that is , prevalence is reduced below 1% ., In the low endemicity area , the success is reached by only targeting symptomatic dogs ., Under slightly higher transmission , however , successful control requires the sacrifice of symptomatic and asymptomatic dogs ., That is , limiting the intervention to clinically positive dogs was not sufficient to control the disease below the 1% prevalence level , leaving it at 2% instead ., When screening is reduced to 4% per month , a less sensitive test is used ( 80% ) and elimination time increases to an average of four months , the good performance of the control program targeting symptomatic dogs only is still preserved in the low endemicity area , with final prevalence reaching values below 1% ., As transmission increases , targeting just symptomatic dogs becomes no longer effective , resulting in final prevalence of 6% ., To ensure prevalence below 1% , at least 30% of the asymptomatic dogs should be put down ( results not shown ) ., Figure 2 shows that , if control targets both symptomatic and asymptomatic dogs , the impact of improving the sensitivity from 80 to 90% is negligible ., On the other hand , in a program targeting symptomatic dogs only , improving the tests sensitivity to 90% is very advantageous to improve its effectiveness ., To further investigate the relevance of asymptomatic dogs on control , we parameterized the model once again , assuming that all asymptomatic dogs were non-infectious , but still positive for the diagnostic tests ., These individuals are the dogs considered cured according to Lanotte et al . ( 1979 ) 25 ., In this case , their elimination has no effect on the success of the control program ., The low and moderate endemicity scenarios simulated here are in the low range of the estimated values for R0 ., The performance of imperfect culling programs in an area with extremely high transmission rate , corresponding to R0\u200a=\u200a9 , was evaluated and , in this case , none of the culling strategies were effective ( Table 2 ) ., Actually , the R0 threshold under which CVL is controlled is R0\u200a=\u200a1 . 41 for programs targeting any seropositive dog ., A maximum R0\u200a=\u200a1 . 106 is required for the success of programs targeting clinically positive dogs only ., One of the main arguments against culling programs is the unnecessary sacrifice of healthy dogs that are erroneously diagnosed , leading to speeches against culling , which reduces the number of animals delivered to zoonosis centers , increasing the ethical and social costs of this strategy ., Here , we calculated the number of unnecessarily sacrificed dogs in a program using a diagnostic test with 80% sensitivity and either 80% and 90% specificity , during the five years of the control application ( years 35 to 40 after control implementation ) ., In the high endemicity area , a program using a test with 80% specificity , this number was 5821 animals , which corresponds to 38% of all dogs put down ., Increasing specificity to 90% , only a slight reduction was obtained ., However , restricting culling to symptomatic dogs only is not sufficient to control the disease below the 1% prevalence level ., This result poses a dilemma to control programs in high endemicity areas as the success of culling is only achieved if asymptomatic dogs are included and this is done at the expense of putting down non-infected dogs ., In areas with low endemicity , on the other hand , restricting culling to symptomatic dogs can both control the disease and reduce the risk of putting down healthy animals ., Figure 3 shows the proportion of the parameter space – corresponding to a variety of natural history situations – that were controlled by culling programs with screening rates equal to 4 , 6 or 8% , and tests sensitivity equal to 80 or 90% ., It is clear that the success of the culling programs is highly dependent on the transmission rate and that increasing screening effort is required in areas with high transmission ., Moreover , it is clear that increasing screening effort is more effective than increasing the sensitivity of the diagnostic tests from 80 to 90% ., However , one must consider the costs associated with such effort for a routine program ., Figure 4 shows the life-history parameters associated with the success or failure of the control program that targeted asymptomatic and symptomatic dogs , with 0 . 04% screening effort in an area with R0\u200a=\u200a1 . 41 ., The most important parameters refer to the asymptomatic population ., In summary , the higher the proportion of dogs becoming or remaining asymptomatic , the most effective the program is ., The reason is the lower transmissibility of these asymptomatic dogs ., This study aimed at assessing the effectiveness of culling dogs in the control of canine visceral leishmaniasis in scenarios where implementation occurs imperfectly ., This investigation was based on a mathematical model for CVL that introduces a class of infectious asymptomatic dogs which contributes , at a lower rate , to the transmission cycle 15 ., This model differs from previous models , in which asymptomatic dogs are assumed to be uninfectious 7 , which may be true for European dogs that are well nourished 26 but not necessarily for all dog populations ., The infectiousness and proportion of asymptomatic dogs had strong impact on the success of control strategies ., As a matter of comparison , we also simulated a simple SI model as parameterized for CVL ., This is in line with part of CVL modeling literature using SIR-like models 16 , 27 ., Overall , when compared with the SEI2D model , SID generates more optimistic expectations , with successful control being reached at faster rates ., Most researchers agree that the sensitivity , specificity and reproducibility of the available serological tests are substandard 6 , 11 , 15 , 16 , 25 , 28 , 29 , 30 ., Sensitivity depends on the methodology used , and the specificity varies with the choice of the antigen ., Low sensitivity increases the chance of permanence of false-negative animals in the environment 12 ., An aggravating issue in the permanence of asymptomatic dogs is the difficulty of tracking these dogs , turning them into a silent reservoir 12 ., The main result of our simulations is that , in areas with very low transmission ( baseline prevalence of 3% ) , culling of symptomatic dogs by a realistic program with 4% screening and testing per month and a mean delay to culling of four months , is sufficient to maintain prevalence under 1% , which we considered a successful endpoint ., The advantage of this program is the focus on symptomatic dogs only , what reduces the burden of killing apparently healthy dogs , providing a better grip of the program by the population ., However , the control program was successful in interrupting transmission of CVL in areas of low transmission , possibly because the endemic equilibrium in these simulations was fragile , and a simple disturbance in the system lead to R0<1 ., In areas with slightly higher endemicity ( R0\u200a=\u200a1 . 29 , prevalence\u200a=\u200a15% ) , on the other hand , removing clinically diagnosed dogs is not sufficient as a control strategy because the asymptomatic population is large enough to maintain transmission ., This is in accordance with many studies 12 , 13 , 15 , 31 ., In this case , a program would have to be capable of including at least 30% of the asymptomatic but infectious dog population in order to maintain infection prevalence below 1% ., A further complication of targeting asymptomatic dogs is the increased chance of putting down healthy dogs as the diagnostic tests available have low specificity ., This is a serious problem in areas with lower endemicity , where the positive predictive values of the tests tend to be low ., The models studied here suggest that in the simulated area , 79% of dogs would be wrongly eliminated by tests with 80% specificity ., The unnecessary sacrifice of non-infected dogs burdens the program and feeds the discourse against dog culling and increases societys aversion to the control program 12 ., The emotional onus and social cost of euthanizing dogs , whether they are ill or not , must be considered in evaluating of culling dogs as a control strategy against AVL ., To avoid the erroneous sacrifice of false-positive dogs , it must be ensured that the tests have high specificity reducing the social cost of this strategy ., The transmission rate of CVL in real settings can be much higher than the ones simulated here 17 , 20 ., As the transmission rate increases , the effectiveness of the culling program rapidly declines unless investment in screening is enhanced ( Figure 3 ) ., In high transmission areas , the required effort may become too high to be feasible , and combined strategies , such as vector control , may become necessary ., In any scenario , control effectiveness requires continuity , that is , no interruptions in the application of control measures ., In practical terms , the inclusion of asymptomatic dogs in a control program stumbles in several difficulties: the difficulty of screening and testing these dogs , as well as their diagnosis , and convincing the delivery of apparently healthy dogs for culling 12 , 31 ., An excellent program would be the one which is more efficient and less costly ., A control program aimed only at symptomatic dogs has apparently lower cost than one targeting all infected dogs ., However , such program by itself will not control transmission ., These results overall , suggest that strategies should differ in areas with high and low transmission , with more integrated approaches being the choice in the former and culling of symptomatic dogs being a choice in the latter ., We did not investigate the relative effectiveness of other strategies in the same scenario ., Dye ( 1996 ) suggested that insecticide application can be more effective than culling , but this is based on the assumption that the impact of insecticides on the sandfly population is high , and resistance is absent or low ., Palatnik-de-Sousa et al . ( 2004 ) argues that using diagnostic tests with greater sensitivity collaborates for the greater effectiveness of a culling program , by minimizing the percentage of false-negative dogs ., However , Dye et al . ( 1993 ) assures that even if a highly efficient serological test was used , about 20% of the cases would remain undetected , especially in animals which , at the time of the test , were in the incubation or seroconversion phases ., Using tests with greater sensitivity and lower specificity may incur in greater social cost and reduced efficiency due to low social acceptance ., In summary , the analysis of the models suggests that besides investments on the improvement of diagnostic tests , further effort is required to improve the control program itself , considering the logistics and resources required for implementation of control for longer periods ., The results of this study are limited to cases in which the model is valid ., The model assumed a canine population of constant size , but it is possible that , in some contexts , these populations are actually increasing or decreasing ., Another limitation of the model is not explicitly considering the dynamics of the vector ., It is known from the study of other diseases such as dengue and malaria that the vectorial capacity can be affected by climate and environmental conditions , including variation from one year to another ., The model also does not consider other potential hosts such as wild animals ., The impact of control would be lower if these animals were present ., Furthermore , the model assumes that all dogs are homogeneously exposed to the risk of vector contact ., In real situations , the risk is expected to vary spatially and future studies should consider this dimension ., Finally , control is implemented continuously , but in real situations this is rare ., Future studies should investigate the impact of strategies applied in pulses at different times of the year ., With all this , we must evaluate the results of this study with caution and by a realistic point of view , noting that the canine sacrifice was effective in controlling the CVL only in a scenario in which the control was implemented monthly and with the same effort for 40 years . | Introduction, Materials and Methods, Results, Discussion | Visceral leishmaniasis belongs to the list of neglected tropical diseases and is considered a public health problem worldwide ., Spatial correlation between the occurrence of the disease in humans and high rates of canine infection suggests that in the presence of the vector , canine visceral leishmaniasis is the key factor for triggering transmission to humans ., Despite the control strategies implemented , such as the sacrifice of infected dogs being put down , the incidence of American visceral leishmaniasis remains high in many Latin American countries ., Mathematical models were developed to describe the transmission dynamics of canine leishmaniasis and its control by culling ., Using these models , imperfect control scenarios were implemented to verify the possible factors which alter the effectiveness of controlling this disease in practice ., A long-term continuous program targeting both asymptomatic and symptomatic dogs should be effective in controlling canine leishmaniasis in areas of low to moderate transmission ( R0 up to 1 . 4 ) ., However , the indiscriminate sacrifice of asymptomatic dogs with positive diagnosis may jeopardize the effectiveness of the control program , if tests with low specificity are used , increasing the chance of generating outrage in the population , and leading to lower adherence to the program ., Therefore , culling must be planned accurately and implemented responsibly and never as a mechanical measure in large scale ., In areas with higher transmission , culling alone is not an effective control strategy . | Visceral leishmaniasis is listed as a neglected tropical disease and is considered a public health problem worldwide ., The disease has been documented since 1885 , the first case being reported in India ., After over 120 years , the incidence of the disease remains high despite control strategies implemented ., In areas where the disease is zoonotic , such as in Brazil , identification as well as removal of infected dogs is recommended in highly endemic areas for they are considered to be the reservoir of the Leishmania chagasi parasite ., The theoretical basis that supports the culling of infected dogs is the assumption that the incidence of human infection is directly related to the number of infectious dogs ., However , there is no consensus among researchers on the effectiveness of this strategy for controlling either human or canine visceral leishmaniasis ., In this context , mathematical models can provide a basis for determining the strategies with the greatest potential for success ., This paper aims to contribute to this discussion by introducing further complexities into the problem , in particular , the imperfect diagnosis of this infection and the time gap between laboratory diagnosis and culling and the presence of asymptomatic infections . | population modeling, epidemiology, infectious disease epidemiology, disease dynamics, population dynamics, population biology, infectious disease modeling, biology, computational biology | null |
journal.pcbi.1006681 | 2,019 | Human online adaptation to changes in prior probability | Sensory decision-making involves making decisions under uncertainty ., Furthermore , optimal sensory decision-making requires the combination of uncertain sensory signals with prior expectations ., Perceptual models of decision-making often incorporate prior expectations to describe human behavior ., In Bayesian models , priors are combined with likelihoods to compute a posterior 1 ., In signal detection theory , the effect of unequal probabilities ( signal present vs . absent ) is a shift of the decision criterion 2 ., The effects of prior probability on the decision criterion have been observed in detection 2–4 , line tilt 5 , numerosity estimation 6 , 7 , recognition memory 8 , and perceptual categorization 9 tasks , among others ., These studies generally use explicit priors , assume a fixed effect , and treat learning as additional noise ., However , there are many everyday tasks in which the probability of a set of alternatives needs to be assessed based on one’s past experience with the outcomes of the task ., The importance of experience has been demonstrated in studies examining differences between experience-based and description-based decisions 10 , 11 and in perceptual-categorization tasks with unequal probability , in which response feedback leads to performance that is closer to optimal than observational feedback 12 , 13 ., While these studies demonstrate the importance of experience on decision-making behavior , they do not describe how experience influences expectation formation ., The influence of experience on the formation of expectations has been demonstrated for learning stimulus mean 14–17 , variance 14 , 18 , higher-order statistics 19 , likelihood distributions 20 , the rate of change of the environment 15–17 , 21–23 , and prior probability 24 , 25 ., Here , we add to previous work by investigating how one’s previous experience influences probability learning in a changing environment ., In the previous work on probability learning by Behrens et al . 24 , participants tracked the probability of a reward to learn the value of two alternatives ., This is a classic decision-making task that involves combining prior probability and reward ., In contrast , we are interested in perceptual decision-making under uncertainty , in which prior probability is combined with uncertain sensory signals ., We might expect differences in strategy between cognitive and perceptual tasks , as cognitive tasks can be susceptible to additional biases ., For example , participants often exhibit base-rate neglect ( i . e . , they ignore prior probability when evaluating evidence ) in cognitive tasks 26 but not in perceptual tasks 2 ., On the other hand , Behrens et al . 24 found that participants’ behavior was well described by an optimal Bayesian model , in that observed learning rates quantitatively matched those of a Bayesian decision-maker carrying out the same task ., A more recent study by Zylberberg et al . 25 examined probability learning under uncertainty in a motion-discrimination task ., In this study , probability was estimated from a confidence signal rather than explicit feedback ., Additionally , probability was changed across blocks , allowing participants to infer a change had occurred ., Here , we examine probability learning when feedback is explicit and changes are sudden ., To investigate probability learning in uncertain conditions , we designed a perceptual categorization task in which observers need to learn category probability through experience ., Since our goal was to examine low-level perceptual and decision-making processes , we used a highly controlled experimental environment ., To prevent the use of external cues ( e . g . , the start of a new block indicating a change in probability ) probabilities were changed using a sample-and-hold procedure ., This approach has been used in decision-making 21 , 22 , 24 and motor domains 18 to examine behavior in dynamic contexts ., Observers completed both a covert- and overt-criterion task , in which the decision criterion was implicit or explicit , respectively ., The overt task , previously developed by Norton et al . 16 , provided a richer dataset upon which to compare models of human behavior ., We determined how observers tracked category probability in a changing environment by comparing human behavior to both Bayesian and alternative models ., We find that a number of models qualitatively describe human behavior , but that , quantitatively , model comparison favored an exponential averaging model with a bias towards equal priors and a flexible variant of the Bayesian change-point detection model with incorrect beliefs about the generative model ., Although model comparison did not distinguish between these models , we interpret the exponential model with a conservative bias as a simpler , more biologically plausible explanation ., Our results suggest that changes in the decision criterion are the result of both probability estimates computed on-line and a more stable , long-term prior ., During each session , observers completed one of two orientation-categorization tasks ., On each trial in the covert-criterion task , observers categorized an ellipse as belonging to category A or B by key press ( Fig 1A ) ., On each trial in the overt-criterion task , observers used the mouse to rotate a line to indicate their criterion prior to the presentation of an ellipse ( Fig 1B ) ., The observer was correct if the ellipse belonged to category A and was clockwise of the criterion line or if the ellipse belonged to category B and was counter-clockwise of the criterion line ., The overt-criterion task is an explicit version of the covert-criterion task developed by Norton et al . 16 ., The overt-criterion task provides a richer dataset than the covert-criterion task in that it affords a continuous measure and allows us to see trial by trial changes in the reported decision criterion , at the expense of being a more cognitive task ., In both tasks , the categories were defined by normal distributions on orientation with different means ( μB = μA + Δθ ) and equal standard deviation ( σC ) ; the mean of category A was set clockwise of the the mean of category B and a neutral criterion would be located halfway between the category means ( Fig 1C ) ., The difficulty of the task was equated across participants by setting Δθ to a value that predicted a d′ value of 1 . 5 based on the data from the initial measurement session ( see Methods ) ., All data are reported relative to the neutral criterion , zneutral = ( μA + μB ) /2 ., Prior to testing , observers were trained on the categories ., Only the covert-criterion task was used for training ( see Section 6 in S1 Appendix ) ., During training , category probability was equal ( π = 0 . 5 ) and observers received feedback on every trial that indicated both the correctness of the response ( tone ) and the generating category ( visual ) ., As a measure of category learning , we compute the probability of being correct in the training block and averaged across sessions ., All observers learned the categories to the expected level of accuracy ( p ( correct ) = 0 . 74 ± 0 . 01; mean ± SEM across observers ) , given that the expected fraction of correct responses for an ideal observer with d′ = 1 . 5 and equal priors over categories is 0 . 77 ., As an additional test of category learning , immediately following training observers estimated the mean orientation of each category by rotating an ellipse ., Each category mean was estimated once and no feedback was provided ., There was a significant correlation between the true and estimated means for each category ( category A: r = 0 . 82 , p < 0 . 0001; category B: r = 0 . 97 , p < 0 . 0001 ) , suggesting that categories were learned ., However , on average mean estimates were repelled from the category boundary ( average category A error of 11 . 3° ± 6 . 3° and average category B error of −8 . 0° ± 2 . 6°; mean ± SEM across observers ) suggesting a systematic repulsive bias ., To determine how category probability affects decision-making , during testing category A probability πt was determined using a sample-and-hold procedure ( Fig 1D; category B probability was 1 − πt ) ., For t = 1 , category A probability was randomly chosen from a set of five probabilities Sπ = {0 . 2 , 0 . 35 , 0 . 5 , 0 . 65 , 0 . 8} ., On most trials , no change occurred ( Δt = 0 ) , so that πt+1 = πt ., Every 80-120 trials there was a change point ( Δt = 1 ) , with change point sampled uniformly ., At each change point , category probability was randomly selected from the Sπ excluding the current probability ., On each trial t , a category Ct was randomly selected ( with P ( category A ) = πt ) and a stimulus st was drawn from the stimulus distribution corresponding to the selected category ., We assume that the observer’s internal representation of the stimulus is a noisy measurement xt drawn from a Gaussian distribution with mean st and standard deviation σv , which represents visual noise ( v ) ., The generative model of the task is summarized in Fig 1E ., To understand how decision-making behavior is affected by changes in category probability , we compared observer performance to several Bayesian models ., To compute the behavior of a Bayesian observer , we developed a Bayesian change-point detection algorithm , based on Adams and MacKay 27 , but which also accounts for Markov dependencies in the transition distribution after a change ., Specifically , the Bayesian observer estimates the posterior distribution over the current run length ( time since the last change point ) , and the state ( category probability ) before the last change point , given the data so far observed ( category labels until trial t , Ct = ( C1 , … , Ct ) ) ., We denote the current run length at the end of trial t by rt , the current state by πt , and the state before the last change point by ξt , where both πt , ξt ∈ Sπ ., That is , if a change point occurs after trial t ( i . e . , rt = 0 ) , then the new category A probability will be πt and the previous run’s category probability ξt = πt−1 ., If no change point occurs , both π and ξ remain unchanged ., We use the notation C t ( r ) to indicate the set of observations ( category labels ) associated with the run rt , which is Ct−rt+1:t for rt > 0 , and ∅ for rt = 0 ., The range of times with a colon , e . g . , Ct′:t , indicates the sub-vector of C with elements from t′ to t included ., Both of our tasks provide category feedback , so that at the end of trial t the observer has been informed of C1:t ., In S1 Appendix we derive the iterative Bayesian ideal-observer model ., After each trial , the model calculates a posterior distribution over possible run lengths and previous probability states , P ( rt , ξt|C1:t ) ., The generative model makes it easy to calculate the conditional probability of the current state for a given run length and previous state , P ( πt|rt , ξt , C1:t ) ., These two distributions may be combined , marginalizing ( summing ) across the unknown run length and previous states to yield the predictive probability distribution of the current state , P ( πt|C1:t ) ., Given this distribution over states , in both tasks the observer needs to determine the probability of each category ., In particular ,, P ( C t + 1 = A | C 1 : t ) = E π t = ∑ π t ∈ S π π t P ( π t | C 1 : t ) ., ( 1 ) In the overt task , the ideal observer sets the current criterion to the optimal value z t opt based on the known category orientation distributions and the current estimate of category probabilities ., Further , in the ideal and all subsequent models of the overt task , in addition to early sensory noise ( σv ) we assume the actual setting is perturbed by adjustment noise ( z t = z t opt + ε t , where ε t ∼ N ( 0 , σ a 2 ) ) ., In the covert task , the observer views a stimulus and makes a noisy measurement xt of its true orientation st with noise variance σ v 2 ., The prior category probability is combined with the noisy measurement to compute category A’s posterior probability P ( Ct+1 = A|xt+1 , C1:t ) ., The observer responds “A” if that probability is greater than 0 . 5 ., We consider the ideal-observer model ( Bayesideal ) and four ( suboptimal ) variants thereof , which deviate from the ideal observer in terms of their beliefs about specific features of the experiment ( Bayesr , Bayesπ , Bayesβ , and Bayesr , π , β ) ., Two further variants of the Bayesian model ( Bayesbias and Bayesr , β ) are described in S1 Appendix ., Crucially , all these models are “Bayesian” in that they compute a posterior over run length and probability state , but they differ with respect to the observer’s assumptions about the generative model ., Note that these models differ from the model provided by Gallistel and colleagues 28 , which was used to model a task in which participants explicitly indicated perceived probability and change points ., In addition to the Bayesian models described above , we tested the following alternative models that do not require the observer to compute a posterior over run length and probability state ., In each of the following models , assumptions vary about whether and how probability is estimated ., In the Fixed Criterion ( Fixed ) model the observer assumes fixed probabilities ., In the Exponential-Averaging ( Exp ) , Exponential-Averaging with Prior Bias ( Expbias ) , and the Wilson et al . ( 2013 ) models , probability is estimated based on the recent history of categories ., In the Reinforcement-Learning ( RL ) model , the decision criterion is updated following an error-driven learning rule with no assumptions about probability ., Finally , the Behrens et al . ( 2007 ) model is an alternative Bayesian model with fewer assumptions and restrictions than the Bayesian change-point detection model described above ., We also tested three additional models that are described in S1 Appendix ., Each of the following models , except for the RL model , computes an estimate of category probability ( π ^ A , t ) on each trial and the estimated probability of the alternative is π ^ B , t = 1 − π ^ A , t ., On each trial , the optimal criterion zopt is computed based on these estimated probabilities in the identical manner as for the Bayesian models ., To make a categorization decision in the covert-criterion task , the criterion is applied to the noisy observation of the stimulus ., In the overt-criterion task , the observer reports the criterion , which we again assume is corrupted by adjustment noise ., Fig 2 shows raw data for a single observer in the covert ( Fig 2A ) and overt ( Fig 2C ) tasks ., For visualization in the covert task , the ‘excess’ number of A responses is plotted as a function of trial ( gray line in Fig 2A ) ., To compute the ‘excess’ number of A responses , we subtracted t 2 from the cumulative number of A responses ., Thus , ‘excess’ A responses are constant for an observer who reported A and B equally often , increase when A is reported more , and decrease when A is reported less ., To get a sense of how well the observer performed in the covert task , the number of ‘excess’ A trials ( based on the actual category on each trial rather than the observer’s response ) is shown in black ( Fig 2A , top ) ., For reference , πA is shown as a function of trial ( Fig 2A , bottom ) ., From visual inspection , the observer reported A more often when πA , t > 0 . 5 and B more often when πA , t < 0 . 5 ., Results for all observers in the covert task can be found in S1 Appendix ( gray line in Figs S6A-S17A ) ., In the overt task , the orientation of the observer’s criterion setting , relative to the neutral criterion , is plotted as a function of trial ( gray circles ) ., For visualization , a running average was computed over a five-trial moving window ( gray line ) ., Here ( Fig 2C , top ) , the black line represents the criterion on each trial , given perfect knowledge of the categories , sensory uncertainty , and category probability ., While this is impossible for an observer to attain , we can see that the observer’s criterion follows the general trend ., This suggests that observers update their criterion appropriately in response to changes in probability ., That is , the criterion is set counter-clockwise from the neutral criterion when πA , t > 0 . 5 , and clockwise of neutral when πA , t < 0 . 5 ., Fig 2C ( bottom ) shows πA as a function of trial ., Results for all observers in the overt task can be found in S1 Appendix ( gray line in Figs S6B-S17B ) ., To determine the influence of prior probability on decision-making behavior , we examined changes in the decision criterion ., First , we found that no participant was best fit by a fixed-criterion model ., This finding suggests that observers update decision criteria in response to implicit changes in probability ., This result is consistent with previous studies in which prior probability was explicit 2–9 ., Further , this finding complements recent studies suggesting that individuals can learn and adapt to statistical regularities in changing environments 14–17 , 24 , 25 , 35 , 36 ., Although this finding suggests that observers dynamically adjust decision criteria in response to changes in prior probability , it does not tell us how they do this ( e . g . , do observers compute on-line estimates of probability ? ) ., To uncover the mechanisms underlying changes in decision-making behavior , we compared multiple models ranging from the full Bayesian change-point detection model to a model-free reinforcement-learning ( RL ) model ., How is the decision criterion set ?, Qualitatively , most models appear to fit the data reasonably well in the covert task ., However , when we look at data from the overt task , while the Bayesian change-point detection models captured the overall trend , some variants failed to capture local fluctuations in the decision criterion observed during periods of stability ( i . e . , time intervals between change points ) ., In other words , the criterion predicted by these models stabilized whereas the observers’ behavior did not ., This was less true when the Bayesian change-point detection model was allowed to vary more freely ., In contrast , the exponential-averaging models continually update the observer’s estimate of probability based on recently experienced categories ., It can be difficult to discriminate a hierarchical model ( e . g . , the Bayesian change-point detection models ) from flat models ( e . g . , our exponential models ) ., For example , in a similar paradigm , Heilbron and Meyniel 37 found that predictions of change points could be similar for hierarchical and flat models ., The addition of confidence judgments allowed them to discriminate the two model forms more readily ., How quickly observers updated this estimate is determined in the model by the decay-rate parameter ., From our model fits , we found that observers had an average decay rate that was substantially smaller than the true run length distribution ( on average 4 . 5 vs . 100 trials , respectively ) , leading to frequent , systematic fluctuations in decision criteria ., Although we cannot directly observe these fluctuations in the covert task , because the estimated decay rate was not significantly different across tasks we can assume the fluctuations occurred in a similar manner ., Like the exponential models , the RL model was also able to capture local fluctuations in the decision criterion ., However , the amplitude of the changes in criterion predicted by the RL model was generally too low compared to the data ., This discrepancy was especially clear in the overt task; no participant was best fit by the RL model ., As a more fair comparison to the exponential models , we also fit the Bayesian model developed by Behrens and colleagues 24 with and without a bias towards equal probability ., These models are more fair in that they do not require specifying a run-length distribution and allowed us to increase the number of possible probability states ., While this allowed us to capture local fluctuations , model comparison favored the Expbias model ., These results have two important implications ., ( 1 ) It is important to test alternatives to Bayesian models: observers’ behavior might be explained without requiring an internal representation of probability ., ( 2 ) Using multiple tasks together with rigorous model comparison can provide additional insight into behavior ( see also 38 ) ., Here , the fluctuations in decision criteria between change points led to suboptimal behavior ., Overall , our findings suggest that suboptimality arose from an incorrect , possibly heuristic inference process , that goes beyond mere sensory noise 39–41 ., While the Bayesr , π , β and Expbias models fit the data equally well , the Expbias model provides a simpler explanation—considering that our experiment did not necessarily require subjects to build a hierarchical model of the task ( see 37 ) ., This suggests that observers compute on-line estimates of category probability based on recent experience ., Further , the bias component of the model suggests that observers are conservative , as reflected in a long-term prior that categories are equally likely ., The degree to which observers weight this prior varied across individuals and tasks ., Taken together , these results suggest a dual mechanism for learning and incorporating prior probability into decisions ., That is , there are ( at least ) two components to decision making that are acquired and updated at very different timescales ., Multiple-mechanism models have been used to describe behavior in decision-making 30 and motor behavior 42 ., A model that combines delta rules predicts motor behavior better than either delta rule alone 42 ., Using a combination of delta rules 30 , we were able to capture the local fluctuations in criterion that the ideal Bayesian model missed ., However , we found that a constant weight on π = 0 . 5 fit better than the multiple-node model described by Wilson and colleagues 30 ., Temporal differences between their task and ours might explain some of the differences we observed , as changes occurred much more slowly in our experiment ., Additionally , while fitting Wilson et al . ’s model we set the hazard rate to 0 . 01 ( the average rate of change ) , but observers had to learn this value throughout the experiment and may have had incorrect assumptions about the rate of change 21 , 22 ., Conservatism was an important feature in our models , as it improved model fits each time it was incorporated ., While we observed conservatism in both the covert- and overt-criterion tasks , we found that , on average , observers were significantly more conservative in the covert task ., To understand why conservatism differs across tasks , we need to understand the differences between the tasks ., While the generative model was identical across tasks , the observer’s response differed ., In the covert task , observers chose between two alternatives ., In the overt task , observers selected a decision criterion ., This is an important difference because it allows us to potentially rule out previous explanations of conservatism , such as the use of subjective probability 4 , misestimation of the relative frequency of events 43 , 44 , and incorrect assumptions about the sensory distributions 45 , 46; these explanations predict similar levels of conservatism across tasks ., On the other hand , conservatism may be due to the use of suboptimal decision rules ., Probability matching is a strategy in which participants select alternatives proportional to their probability , and has been used to explain suboptimal behavior in forced-choice tasks in which observers choose between two or more alternatives 6 , 47–50 ., Thus , the higher levels of conservatism in the covert task may have been due to the use of a suboptimal decision rule like probability matching , which would effectively smooth the observer’s response probability across trials ., Probability matching is not applicable to responses in the overt task ., Thus , the use of different decision rules may result in different levels of conservatism ., These differences may also arise from an increase in uncertainty in the covert task due to less explicit feedback ., An observer with greater uncertainty will rely more on the prior ., Thus , conservatism may be the result of having a prior over criteria that interacts with task uncertainty ., This can be tested by manipulating uncertainty over the generative model and measuring changes in conservatism ., It is also possible that our training protocol introduced a bias towards equal probability and , due to the greater similarity between the covert and training tasks , the bias was stronger in the covert task ., Finally , it is also possible that conservatism is the result of both the use of suboptimal decision rules and one or more of the previously proposed explanations ., While we tested a number of Bayesian change-point detection models that explored an array of assumptions about the generative model , clearly one could propose even more variants ( e . g . , a model with incorrect assumptions about category means and variance ) ., Here , we analyzed one such assumption at a time ., A simple way to expand the model space is via a factorial comparison 40 , 51 , which we did not consider here due to computational intractability and the combinatorial explosion of models ., We did however , fit the Bayesr , π , β model , which simultaneously accounted for incorrect assumptions about the run length and probability-state distributions and a bias towards equal probability ., We compared the fit to the Expbias model ., Notably these two models explained the data equally well , despite the higher flexibility of the Bayesian model ., We can thus interpret the Expbias model as a simpler explanation for Bayesian change-point detection behavior with largely erroneous beliefs ., For all models except the RL model we assumed knowledge of the category distributions ., However , Norton et al . 16 found that for the same orientation-categorization task , category means were estimated dynamically , even after prolonged training ., Similarly , Gifford et al . 52 observed suboptimality in an auditory-categorization task and found that the data were best explained by a model with non-stationary categories and prior probability that was updated using the recent history of category exemplars ., This occurred despite holding categories and probability constant within a block ., In fact , similar effects of non-stationarity have been observed in several other studies 53–55 ., In addition to non-stationary category means , observers may also have misestimated category variance 56 , especially since learning category variance takes longer than learning category means 14 ., In sum , our results provide a computational model for how decision-making behavior changes in response to implicit changes in prior probability ., Specifically , they suggest a dual mechanism for learning and incorporating prior probability that operate at different timescales ., Importantly , this helps explain behavior in situations in which assessment of probability is learned through experience ., Further , our results demonstrate the need to compare multiple models and the benefit of using tasks that provide a richer , more informative dataset ., The Institutional Review Board at New York University approved the experimental procedure and observers gave written informed consent prior to participation ., Eleven observers participated in the experiment ( mean age 26 . 6 , range 20-31 , 8 females ) ., All observers had normal or corrected-to-normal vision ., One of the observers ( EHN ) was also an author ., Stimuli were presented on a gamma-corrected Dell Trinitron P780 CRT monitor with a 31 . 3 x 23 . 8° display , a resolution of 1024 x 768 pixels , a refresh rate of 85 Hz , and a mean luminance of 40 cd/m2 ., Observers viewed the display from a distance of 54 . 6 cm ., The experiment was programmed in MATLAB 57 using the Psychophysics Toolbox 58 , 59 ., Stimuli were 4 . 0 x 1 . 0° ellipses presented at the center of the display on a mid-gray background ., In both the orientation-discrimination and covert-criterion tasks , trials began with a central white fixation cross ( 1 . 2° ) ., In the overt-criterion task , a yellow line with random orientation was presented at the center of the display ( 5 . 0 x 0 . 5° ) ., During the ‘measurement’ session , sensory uncertainty ( σv ) was estimated using a two-interval , forced-choice , orientation-discrimination task in which two black ellipses were presented sequentially on a mid-gray background ., The observer reported the interval containing the ellipse that was more clockwise by keypress ., Once the response was recorded , auditory feedback was provided and the next trial began ., An example trial sequence is shown in Fig S3A in S1 Appendix ., The orientation of the ellipse in the first interval was chosen randomly on every trial from a uniform distribution ranging from -90 to 90° ., The orientation of the second ellipse was randomly oriented clockwise or counter-clockwise of the first ., The difference in orientation between the two ellipses was selected using an adaptive staircase procedure ., The minimum step-size was 1° and the maximum step-size was 32° ., Each observer ran two blocks ., In each block , four staircases ( 65 trials each ) were interleaved ( two 1-up , 2-down and two 1-up , 3-down staircases ) and randomly selected on each trial ., For analyses and results see S1 Appendix ., Each training trial was identical to a covert-criterion trial ( Fig 1A ) ., During training there was an equal chance that a stimulus was drawn from either category ., To assess learning of category distributions , observers were asked to estimate the mean orientation of each category following training ., The mean of each category was estimated exactly once ., The order in which category means were estimated was randomized ., For estimation , a black ellipse with random orientation was displayed in the center of the display ., Observers slid the mouse to the right and left to rotate the ellipse clockwise and counterclockwise , respectively and clicked the mouse to indicate they were satisfied with the setting ., No feedback was provided ., We computed the proportion correct for each observer to ensure category learning by comparing it to the expected proportion correct ( p ( correct ) = 0 . 77 ) for d′ = 1 . 5 ., Mean estimates are plotted in Fig S4C in S1 Appendix as a function of the true category means ., We computed the average estimation error for each category and observer by subtracting the estimate from the true mean ., From visual inspection , it appears that training was effective with the exception of one outlier , which we assume was a lapse ., For fitting , all models had one free noise parameter ., In the covert-criterion task , this was sensory noise ( σv ) ., In the overt-criterion task , sensory noise was fixed and set to the value obtained in the ‘measurement’ session , but we included a noise parameter for the adjustment of the criterion line ( σa ) ., Fixing one noise parameter in the overt-criterion task ameliorated potential issues of lack of parameter identifiability 60 , and ensured that models had the same complexity across tasks ., The Bayesideal , Fixed , and Behrens et al . ( 2007 ) models had no additional parameters ., The following suboptimal Bayesian models had one additional parameter: Bayesr ( r ) ; Bayesπ ( πmin ) ; Bayesβ ( β ) ., The Bayesr , π , β model had 5 additional parameters ( r , Δr , πmin , Δπ , β ) ., The Exp and RL models also only had one additional parameter ( α ) , as did the Behrens et al . ( 2007 ) model with a bias towards equal probability ( w ) ., The Expbias model had two additional parameters ( αexp and w ) , and the Wils | Introduction, Results, Discussion, Conclusion, Methods | Optimal sensory decision-making requires the combination of uncertain sensory signals with prior expectations ., The effect of prior probability is often described as a shift in the decision criterion ., Can observers track sudden changes in probability ?, To answer this question , we used a change-point detection paradigm that is frequently used to examine behavior in changing environments ., In a pair of orientation-categorization tasks , we investigated the effects of changing probabilities on decision-making ., In both tasks , category probability was updated using a sample-and-hold procedure: probability was held constant for a period of time before jumping to another probability state that was randomly selected from a predetermined set of probability states ., We developed an ideal Bayesian change-point detection model in which the observer marginalizes over both the current run length ( i . e . , time since last change ) and the current category probability ., We compared this model to various alternative models that correspond to different strategies—from approximately Bayesian to simple heuristics—that the observers may have adopted to update their beliefs about probabilities ., While a number of models provided decent fits to the data , model comparison favored a model in which probability is estimated following an exponential averaging model with a bias towards equal priors , consistent with a conservative bias , and a flexible variant of the Bayesian change-point detection model with incorrect beliefs ., We interpret the former as a simpler , more biologically plausible explanation suggesting that the mechanism underlying change of decision criterion is a combination of on-line estimation of prior probability and a stable , long-term equal-probability prior , thus operating at two very different timescales . | We demonstrate how people learn and adapt to changes to the probability of occurrence of one of two categories on decision-making under uncertainty ., The study combined psychophysical behavioral tasks with computational modeling ., We used two behavioral tasks: a typical forced-choice categorization task as well as one in which the observer specified the decision criterion to use on each trial before the stimulus was displayed ., We formulated an ideal Bayesian change-point detection model and compared it to several alternative models ., We found that the data are explained best by a model that estimates category probability based on recently observed exemplars with a bias towards equal probability ., Our results suggest that the brain takes multiple relevant time scales into account when setting category expectations . | learning, ellipses, decision making, statistics, social sciences, geometry, neuroscience, learning and memory, cognitive psychology, mathematics, probability distribution, cognition, research and analysis methods, behavior, human learning, mathematical and statistical techniques, monte carlo method, probability theory, psychology, biology and life sciences, physical sciences, sensory perception, cognitive science, statistical methods | null |
journal.pgen.1001337 | 2,011 | Quantifying the Underestimation of Relative Risks from Genome-Wide Association Studies | Genome-wide association studies ( GWAS ) have been extremely successful across many diseases in identifying loci harbouring genetic variants that affect disease susceptibility ., Virtually all associated variants identified from GWAS to date have relatively small effects: each additional copy of the risk allele typically increases disease risk by 10%–30% ( see for example 1 ) ., It has become clear that the variants discovered thus far account for only a small proportion of the genetic basis of each of the diseases , and there has been considerable speculation about where the “missing” heritability might lie 1 ., One of several important factors in the success of the GWAS design has been the pattern of linkage disequilibrium in human populations ., The strong correlations between nearby SNPs mean that commercially available genotyping chips , which assay 300 , 000–1 , 000 , 000 SNPs , can capture much of the common variation in the human genome , particularly in Caucasian populations 2 ., Because genotypes at the causative loci will often be correlated with those at SNPs that are typed on the genotyping chip , it is typically not necessary to assay the true causative variant directly in order to detect a genetic association with disease ., While linkage disequilibrium is extremely helpful for GWAS discovery , the downside is that in most reported regions of association , the true causal variant or variants remain unknown ., Therefore it is possible that many of the associated SNPs are only surrogates for the true causal variant ( s ) ., When it comes to quantifying the genetic effect , the genotype at the reported SNP acts as a noisy measurement of the genotype at the causal variant ., This noise can dilute the apparent strength of the effect , and obscure the true relationship between genotype and phenotype ., As we progress towards the identification of the causal variants , estimates of effect sizes for associated loci will thus tend to increase ., In turn , the proportion of disease susceptibility explained by GWAS loci will also increase ., Thus in addition to other plausible sources , such as secondary signals in GWAS loci , rare variants ( <1% frequency ) , copy number polymorphisms , and epigenetic effects , some of the missing heritability is actually contained in loci already identified by GWAS , and is driven by common variation ( >1% frequency ) ., In this paper we use an extensive simulation study to investigate , and quantify , this phenomenon ., We show that estimates of the size of the genetic effect based on the best SNP from the GWAS genotyping chip can often closely approximate the effect size at the true causal SNP ., In some cases the causal SNP has a large effect and is poorly tagged , leading to substantial underestimation of the true effect size ., We investigate how much of the “missing” heritability could thus be hidden in reported GWAS loci , under several sets of assumptions about the nature of the effects at true causal SNPs ., Our results also inform the design and value of fine mapping experiments in GWAS loci ., To begin , we compare the estimated effect size at the replicated hit SNP with the true effect size at the causal SNP in the simulation ., Figure 1 illustrates this comparison for three different values of the true effect size ., For each we see a peak of estimates around the true effect size assumed at the causal SNP ., But note also that there is often underestimation of the true effect size ( mean estimated effect size 1 . 24 , 1 . 86 and 3 . 32 for true relative risk of 1 . 25 , 2 and 4 respectively ) , and that this underestimation can be substantial when the true effect is large ., For example , when the true relative risk is 4 , the estimated effect size was less than two in 12% of simulations of successful GWAS discovery of the effect ., In Figure 2 we plot the relative under- ( or over- ) estimation of the effect size ( estimated effect size divided by true effect size ) as a function of the correlation ( as measured by the r2 which is the square of Pearsons correlation coefficient ) between the hit SNP and the true causal variant ., The underestimation is seen to be due to imperfect tagging: when the true causal variant is not well tagged by SNPs on the genotyping chip ( the correlation is weak ) , the estimated effect at the hit SNP is often much lower than the true effect ., Conversely , when the causal SNP is well tagged by a SNP on the chip , the estimated effects cluster around the true effect size ., Note that while underestimation decreases as the correlation between the hit SNP and the causal SNP increases , there remains systematic underestimation even when the hit SNP has r2≈0 . 8 with the causative SNP ., For example in one third of simulations when the true effect is two , the estimated effect will be under 1 . 8 ., Note also that when the true effect size is large , significant and replicable associations can be detected when the best tag SNP only has r2≈0 . 2 with the causal variant ( Figure 2 , relative risk\u200a=\u200a4 ) ., Imperfect tagging and an ascertainment effect also explain the feature of the plots whereby the underestimation is much less for smaller true effect sizes ., If the true effect is small and the true causal variant is not well-tagged on the genotyping chip , there will not be enough power for the GWAS and subsequent replication to reach significance 5 , with the result that the corresponding simulation will not contribute to the plot ., But if the true effect is large there may still be power to see a significant result when the true variant is not well tagged , so the simulation contributes to the plot and shows the underestimation ., Put another way , if the true effect is small , it will only be detected in an association study if the causal SNP is well tagged , and in this case the effect size will be estimated reasonably well ., This second ascertainment effect explains the lack of underestimation at hit SNPs not strongly correlated to the causal SNP in the left panel of the Figure, 2 . Lastly , as low frequency SNPs are less well tagged by other SNPs 6 , the extent of the underestimation also depends on the frequency of the risk allele ( see Figure S1 ) ., Interestingly , the effect sizes at rare alleles are underestimated to a great extent , but only when the true effect size is large enough for the tag SNP of a rare allele to be detected and replicated in the simulated GWAS ., The results above describe the distribution of estimated effect sizes as a function of known true effect sizes and the frequency of the risk allele ., In practice we are actually interested in the reverse question , namely what true effect sizes are plausible in the light of the effect size actually estimated from a GWAS and follow-up study ?, We will see that this requires assumptions about the true distribution of effect sizes ., Indeed , writing RR for relative risk , and RAF ( risk allele frequency ) for the allele frequency at the risk allele , application of Bayes theorem gives ( 1 ) where “true” refers to the value at the causal SNP and “observed” refers to the value at the hit SNP ., Our simulation study allows us to estimate the first factor on the right hand side of ( 1 ) , and we do so by discretising both the observed and true RR and RAF and creating a matrix of counts based on our simulations over the ENCODE regions ., The second factor on the right hand side of ( 1 ) is the assumed joint distribution of true risk allele frequencies and effect sizes , which is of course unknown ., We proceed by making two different sets of assumptions about these unknowns ., In each case we assume that the distribution of risk allele frequencies is given by the empirical distribution of allele frequencies in the ENCODE regions ., In effect this assumes that any SNP variant is , a priori , equally likely to affect disease status ., What differs between the sets of assumptions is the assumed effect size of a particular variant ., Our first set of assumptions posits that the distribution of effect sizes is the same for all putative causal variants , regardless of their allele frequency , and that effect sizes are close to those observed in GWAS studies ., The second set of assumptions explicitly assumes that there might be substantially larger effects at variants with smaller minor allele frequency ., These priors are described in detail in the Methods section ., Different sets of assumptions about true effect sizes and risk allele frequencies necessarily lead to different conclusions , and it is impossible to study all possibilities ., A number of theoretical analyses 7 , 8 , 9 , 10 have argued for a relationship between effect size , disease model , and minor allele frequency ( MAF ) ., As there is no consensus on the exact form and extent of the relationship we do not rely on them explicitly here , and instead our approach aims to capture two different perspectives on unknown effect sizes , with the subsequent analyses indicating a range of possibilities ., The first perspective is that the range of true effect sizes will be close to those estimated from current GWAS ., The second captures the possibility that low-frequency variants may have considerably larger effect sizes ., Under either set of assumptions , we can use our simulation study , and Bayes Theorem ( 1 ) to estimate the conditional distribution of true effect sizes and risk allele frequency ( RAF ) in the light of the observed data at the GWAS hit SNP ., Figure 3 illustrates this , showing estimates of the posterior distribution of the true effect size conditional on observing a risk estimate between 1 . 2 and 1 . 3 , for different observed risk allele frequencies , and under the two different prior assumptions on effect size distributions ., A common feature of the histograms in Figure 3 is that the mode of the posterior distribution on the true effect size is on , or very closes , to the observed estimate ., That is , current estimates from GWAS studies of effect sizes from a common SNP , in the range 1 . 2–1 . 3 are most likely to be very close to truth ., As expected , estimated effects within this range are more likely to be 1 . 3 than 1 . 2 , because larger effects are more likely to generate a signal of association strong enough to pass the p-value thresholds commonly implemented in GWAS ., This explains the left hand tail of the distributions represented in Figure, 3 . Figure 3 also shows that there is some probability that the effect size at the causal variant is greater than estimated from the most associated SNP ., Interestingly , the observed risk allele frequency impacts our posterior belief about the true effect size , under either set of prior assumptions , with underestimation be more marked when the risk allele at the hit SNP is rarer ., Under the conservative prior , when the risk allele at the hit SNP has less than 20% frequency in the control population , the probability that the relative risk is above 1 . 325 is 55% , compared to 35% when the risk allele frequency is between 20–50% ., The corresponding numbers for the MAF-dependent prior are 77% and 49% ., There are several different phenomena at work here ., If the hit SNP is the causal SNP then , assuming that the association is strong enough to be detected and replicated in the GWAS , there is no systematic under estimation ( and very little over estimation as we assume the effect size is estimated from the replication sample ) ., However , conditional on the hit SNP not being causal , the distribution of LD with true causal SNP , and therefore the propensity for under estimation , depends on its allele frequency ., The posterior distribution on the true effect size given the observed frequency and effect of the hit SNP can be viewed as a mixture of these two scenarios , weighted by their conditional probability ., Rarer SNPs are less likely to be tagged well by single markers , and as noted above , poor tagging leads to underestimation of effect sizes ., In contrast , for a common SNP , the associated allele is more likely to be well correlated with the causal allele , so there is relatively less under estimation ., Under the MAF-dependent prior , when the associated allele is low-frequency the causative allele will tend to be low-frequency as well , and so potentially of larger effect ., In the scenario where we believe in larger effects at rare causal alleles and have observed a SNP with low RAF with estimated relative risk between 1 . 2 and 1 . 3 there is a 24% chance that the source of the signal is a variant which actually doubles or more than doubles risk with each copy of the risk allele ., Our observations are similar when the observed risk allele is the most common allele in the population ( RAF>50% ) and therefore the minor allele is protective ( Figure S2 ) ., Qualitatively , the same conclusions also apply when the estimated effect size at the hit SNP is weaker , for example in the range 1 . 05 to 1 . 2 ( Figure S3 ) ., One consequence of the potential underestimation of effect sizes from GWAS findings is that as we move to better identification of the actual causal variants , through fine mapping and/or functional studies of associated regions , our estimates of their effect sizes might well increase ., Assuming a multiplicative model of risk across loci , these small expected changes could combine to increase the relative risk of disease in those individuals with highest genetic risk of disease ., To investigate this , we simulated genotypes at known associated loci in a population of individuals ( assuming Hardy Weinberg equilibrium and no linkage disequilibrium across loci ) for each of breast cancer , type 2 diabetes and Crohns disease , based on reported risk allele frequencies 11 , 12 , 13 ( see Tables S3 , S4 , S5 for a list of loci ) ., First treating the causal loci and relative risks for each disease as given by current GWAS estimates , we measured the average risk of individuals in the top x% , by risk , of the population ( for differing values of x ) and compared this to the mean risk in the population ., We then repeated this simulation , allowing for the uncertainty in the estimation of true effect sizes by averaging over the uncertainty in both the RAF and effect size of the causal variant on the basis of the posterior distributions of these , given the GWAS findings , under the two priors described above ., We assumed that risks combined multiplicatively across loci ., For NOD2 and IL23R in Crohns disease where the causal variant is thought to be known , here and below , we used the effect sizes for the known variant , and did not average over uncertainty in these ., Because all three diseases have been extensively studied , we approximated the GWAS discovery process as corresponding to a GWAS discovery sample of 5000 cases and 5000 controls , and a replication sample of 10 , 000 cases and controls ., The actual discovery process for each of the diseases is complicated , often involving meta-analysis and/or multistage discovery , and not straightforward to model accurately , but the approach we use should capture the fact that GWAS-discovery were ascertained through study of large numbers of samples ., The results of the three simulations are given in Table 1 . The unadjusted simulations give estimates of how much more at risk individuals with the greatest genetic propensity to disease are , based only on GWAS loci , relative to the average person in the population ., As expected , the fold change in risk of individuals carrying a large fraction of risk variants is dependent on the number and magnitude of known loci ., For example , individuals in the top 0 . 1% of risk for Crohns disease are 20 times more likely than the average person to develop the condition , whereas for breast cancer , where the number of common loci and associated relative risks is typically smaller , the equivalent number is just over two-fold ., The second and third simulations attempt to average over the possible outcomes of our future efforts to map causal mutations , to reveal the likely gains in our ability to stratify individuals on the basis of risk ., These use the methodology above , under both prior distributions , to average over the posterior distribution of the allele frequency and effect size at the causal SNPs underlying reported GWAS loci for the three diseases ., These adjusted estimates are also shown in Table 1 ., Across diseases we see that there is a significant increase in the risk associated with carrying multiple risk variants ., In particular we see that the biggest differences in risk are for those individuals in the extreme tail ., It is these individuals who carry the stronger , likely rarer , risk alleles which are currently insufficiently characterised by the most significant signal of association in some regions identified to be important in disease ., For example , the risk of an individual in the top 0 . 1% of the population for genetic risk typed at the causal loci underlying currently known GWAS loci will likely be increased by a factor of 3–6 . 5 , 5–12 , or 25–50 , compared to an average individual , for breast cancer , type 2 diabetes and Crohns disease ., These are notably greater increases in risk than current prediction based in the hit SNPs from GWAS loci which would be 2 . 4 , 3 . 5 and 20 respectively ., We have shown above that as we move to identification of the true causal variants underlying GWAS associations , through fine mapping and functional studies , their effect sizes will tend to increase , in a minority of cases substantially , compared to current estimates from GWAS ., This will , in turn , increase the amount of heritability explained by these diseases ., We can use the approach developed here to try to quantify this effect ., We investigated this question in the context of the three diseases just described , namely breast cancer , type 2 diabetes , and Crohns disease ., For each disease we took the set of hit SNPs from published associated loci 11 , 12 , 13 ( see Tables S3 , S4 , S5 ) , and for our two prior distributions on effect sizes we estimated the posterior distribution of both the effect size and the allele frequency for the causal SNP at each locus , as described in the previous section ., One commonly used measure of heritability is sibling recurrence risk ratio , often denoted by λS: the relative increase in risk to an individual if their sibling has the disease compared to the baseline risk in the population as a whole 14 ., Assuming , as is usual for heritability calculations 15 , that there is no interaction between loci , λS can be calculated as a function of the risk allele frequency and effect size for each causal variant ., In order to allow for the uncertainty in the allele frequency and likely underestimation of the effect size at the causal variants underlying GWAS associations , we averaged this expression over the posterior distribution of these quantities , given the GWAS findings ( see Methods for details ) ., The results are shown in Figure 4 ., For each disease they show that the heritability due to already identified GWAS loci will be higher than current estimates , under either set of assumptions about true effect sizes , but particularly under the MAF-dependent prior ., Whereas at the time of writing the current estimates of the contribution to λS from GWAS loci are 1 . 03 , 1 . 08 , and 1 . 49 for breast cancer , type 2 diabetes , and Crohns disease , these may well be 1 . 06 , 1 . 14 , and 1 . 61 ( mean under the conservative prior ) and they could plausibly be as high as 1 . 21 , 1 . 39 and 2 . 46 ( mean under the MAF-dependent prior ) ., Whilst some of the “missing” heritability is thus disguised rather than missing , we note that this effect is unlikely to account for the extent of the gap between estimates of sibling relative risk ( 2 , 1 . 8 , and 10 , respectively , from family studies 16 , 17 , 18 ) and those explained by currently known loci ., We return below to a discussion of the discrepancy ., The correlation between alleles along the human genome has allowed GWAS to look for regions associated with disease without having to either genotype all known genetic variation or guess a priori which regions of the genome may be important ., Although this approach has been a significant success , there is a predictable downside of using a subset of variation to tag , or predict , untyped diversity: for the vast majority of the SNPs identified as mediating disease risk , we are left uncertain as to whether they are causally involved in the pathway from genotype to phenotype , or , much more plausibly , are just a surrogate for the causal variation ., GWAS associations will thus typically relate to a noisy measurement of the causal variant ., One consequence of this is that the size of the genetic effect associated with GWAS loci may be underestimated ., We quantified this through an extensive simulation study designed to mimic patterns of linkage disequilibrium in European Caucasian populations ., We draw two broad conclusions from these analyses ., Firstly , a significant proportion of estimated relative risks will be biased downwards because the hit SNP is a powerful , but imperfect , tag for the true causal variation ., In most cases this effect will be relatively minor , but in some instances , the best associated SNP may actually be a poor predictor of a , putatively rarer , SNP with a much larger effect , in which case the effect size estimated from the GWAS finding will substantially underestimate the true effect size ., The exact proportion of reported associations which fall into these two categories depends on properties of the design of the study from which the SNP was identified , and on ones belief about how likely low frequency ( >1% ) variants of large effect are to cause common diseases ., The statistical power afforded by any particular association strategy sets a lower limit on the size of effect that can be under-estimated because an imperfect tag of an allele with a small effect size will simply fail to achieve genome-wide significance ., Other properties of GWAS strategy , such as sample ancestry and the number of markers typed , also change our interpretation of observed effect sizes because they influence the distribution of linkage disequilibrium between putative hit SNPs and causal variants ., Our findings show that at any particular locus , especially if the associated SNP has a low MAF , the true effect could be quite large ., But we would not expect this to be widespread ., Were many true effects this large it would be extremely surprising for so few of them to have been observed: although any one such causal SNP may not be well tagged on the genotyping chips used for GWAS , some of them will happen to be at least moderately well tagged , and their detection would lead to much larger estimates than have been seen from current studies ., In the context of this study these early observations suggest that , of the two prior distributions we investigated , it is the conservative prior that may better reflect the true distribution of effect sizes attributed to low and common frequency variants ., One way of viewing the posterior distribution on the true effects shown in Figure 3 is as a probability distribution on the outcome of efforts to fine map current regions of association ., In this light , our results inform questions of the design and value of fine mapping experiments ., First , simulations similar to those described above ( assuming causal variation to be distributed like SNPs in ENCODE regions ) suggest that less than 8% of the time will the hit SNP actually be the causal SNP ., We note that there may be more reward in terms of gains in predictive ability and increases in effect size from fine mapping SNPs with lower minor allele frequency because they are , on average , more likely to be in poor LD with an unobserved causal variant ., On the other hand , our simulations show that although they are unlikely to be causal , most common hit SNPs are likely to be very good surrogates markers for their causal variant ., Indeed , in 25% of cases , the hit SNP will be a near-perfect surrogate ( ie r2>0 . 99 ) for the causal variant ., Should this be the case , further genotyping will not reveal other SNPs with stronger associations , unless sample sizes are extremely large ., Here we have quantified the increased spread of genetic risk with genotypes just at known loci , and only considering a multiplicative disease model ., But even in this restricted setting , there will be substantial differences in risk between high- and low-risk groups based on these genetic factors ., For example the propensity of individuals in the top 0 . 1% of the population distribution of genetic risk of type 2 diabetes will be increased by a factor of 5–10 , compared to the average ., For breast cancer , in the analogous top-risk group this risk will be increased by a factor of 3–5 ( on the basis of common variation ) ., Importantly , with the growth of GWAS findings , both in terms of numbers of diseases and numbers of loci for particular diseases , more and more of the population will be in this most at risk category for at least one disease: assuming 100 independent diseases , nearly 10% of the population will be in the top 0 . 1% of risk of at least one disease ., Knowing which individuals these are and what diseases they are most at risk of is therefore potentially useful information , both to the individual and at the population level ., The issues involved in utilising such information in screening programmes ( discussed for example in 13 ) are complicated , but our results strengthen the arguments for consideration of this possibility ., We have shown that some of the “missing” heritability for common disease actually resides in known GWAS loci and have estimated this deficit for three particular diseases ., While rather more heritability is likely to be explained by known GWAS loci than has been reported , this effect alone falls well short of explaining all the missing heritability ., Note , however , that there are other reasons why existing loci may explain more heritability than currently thought ., Current calculations ( by others , and above ) focus on a single causal variant in each associated region: more variants within regions will explain more heritability ., They also ignore possible non-multiplicative disease effects , and also ignore interactions between variants at different loci ., Power to detect either is low 19 , so it is misleading to put much weight on the failure of existing designs to find such effects ., As others have noted 20 , parts of the missing heritability could be due to multiple rare variants of large effect , associations with other forms of genetic variation such as copy number polymorphisms , and epigenetic effects ., Indeed it would be surprising if each did not play some role ., Another possibility is that estimates of the “genetic” component of disease susceptibility , from epidemiological studies , confound shared environment with shared DNA , and so inflate heritability estimates 21 , 22 ., In order to model the signal of association generated by disease-causing mutations , we chose to simulate data exploiting empirical surveys of human diversity ., For this purpose we used data from the 10 ENCODE regions 23 within the CEU analysis panel of HapMap II 5 , which have undergone SNP ascertainment by resequencing 48 individuals of diverse ancestry ., These regions therefore show a fuller spectrum of SNPs than are represented in the HapMap data at large , and haplotypes are expected to be accurate due to the trio design of the CEU HapMap panel 24 ., The regions over which we simulate data are centred on each of the 10 ENCODE regions ( listed in Table S1 ) and include 500kb of flanking HapMap variation at the boundaries of each region ., As the typical sample size of most GWAS is much larger than the number of CEU HapMap individuals , we simulated 100 , 000 chromosomes using the HAPGEN software package ., These 100 , 000 haplotypes we call the reference panel ., GWAS case and control samples were then subsampled from the reference panel , as described below ., HAPGEN uses a population genetic model that incorporates the processes of mutation and fine-scale recombination to generate individuals from an existing set of known haplotypes ., We ran HAPGEN with an effective population size of 11418 ( as recommended for the CEU population ) , a population scaled mutation rate of 1 per SNP , a population scaled recombination rate from estimates described in 25 , with the known set of haplotypes taken from the CEU analysis panel of HapMap II as described above ( see http://www . stats . ox . ac . uk/~marchini/software/gwas/hapgen . html ) ., For SNPs greater than 1% in frequency in the ENCODE regions we performed two hypothetical GWAS by letting each of the two alleles be causal in turn ., We denote the causal allele by A and the protective allele by a ., To generate the control sample we sampled the required number of haplotypes , without replacement , from the reference panel and combined these in pairs to form diploid individuals ., This mimics the common use of population controls , rather than controls explicitly chosen for not having the disease under study ., For the case sample , we sampled pairs of haplotypes from the reference panel according to the genotype frequencies at the causal SNP dictated by the assumed disease model: If δ is the risk of the AA genotype , and α is the risk of the Aa genotype , both relative to the aa genotype , then we sample case individuals ( without replacement ) on the basis of their genotypes at the SNP assumed to be causal with success probabilities proportional to: ( 2 ) where f is the frequency of the risk allele A in the reference panel ., Throughout , for definiteness , we adopted a multiplicative model for disease risk ( additive on the log scale ) defined by δ\u200a=\u200aα2 ., We refer to α as the relative risk ( RR ) or effect size associated with the causal variant ., To approximate a GWAS , we thinned the generated data set to include only those SNPs present on the Affymetrix 500K array that had a minor allele frequency in sampled controls of greater than 1% ., This set may or may not include the assumed causal SNP ., For analyses involving only simulated data , we sampled 2 , 000 cases and 2 , 000 controls from the reference panel to emulate a typical large GWAS ., For the subsequent analyses of heritability and individual risk profiling for type 2 diabetes , breast cancer and Crohns disease that studied particular reported associations , we simulated 5 , 000 cases and 5 , 000 controls to obtain results more comparable to the size of study from which the associations were ascertained ., We simulated under a range of relative risks at 24 grid points from 1 . 05 to 6 ., In attempting to simulate the signal of disease at rare alleles ( 1% to 5% ) in a GWAS of 5000 cases and controls there were a small number of simulations in which there were insufficient haplotypes in our reference panel to generate the required number of genotypes at the causal SNP for large effect sizes ., These simulations were discarded , but as the numbers were small ( 3% when the RR\u200a=\u200a4 and 11% when RR\u200a=\u200a6 ) we do not believe this greatly affects the results presented below ., Following common practice , for each simulated case control sample , we tested for association between genotype and case control status using the Cochran Armitage trend test 26 at each SNP with frequency greater than 1% in the simulated panel of chromosomes ., We calculated the p-value of this test statistic which is distributed with 1 degree of freedom under the null hypothesis of no association ., If any test across the region obtained a p-value<10−6 the location of the most significant SNP ( termed the hit SNP ) was recorded and we simulated this SNP in an independent replication sample ., We simulated the replication experiment in three stages ., First we simulated | Introduction, Results, Discussion, Methods | Genome-wide association studies ( GWAS ) have identified hundreds of associated loci across many common diseases ., Most risk variants identified by GWAS will merely be tags for as-yet-unknown causal variants ., It is therefore possible that identification of the causal variant , by fine mapping , will identify alleles with larger effects on genetic risk than those currently estimated from GWAS replication studies ., We show that under plausible assumptions , whilst the majority of the per-allele relative risks ( RR ) estimated from GWAS data will be close to the true risk at the causal variant , some could be considerable underestimates ., For example , for an estimated RR in the range 1 . 2–1 . 3 , there is approximately a 38% chance that it exceeds 1 . 4 and a 10% chance that it is over 2 ., We show how these probabilities can vary depending on the true effects associated with low-frequency variants and on the minor allele frequency ( MAF ) of the most associated SNP ., We investigate the consequences of the underestimation of effect sizes for predictions of an individuals disease risk and interpret our results for the design of fine mapping experiments ., Although these effects mean that the amount of heritability explained by known GWAS loci is expected to be larger than current projections , this increase is likely to explain a relatively small amount of the so-called “missing” heritability . | Genome-wide association studies ( GWAS ) exploit the correlation in genetic diversity along chromosomes in order to detect effects on disease risk without having to type causal loci directly ., The inevitable downside of this approach is that , when the correlation between the marker and the causal variant is imperfect , the risk associated with carrying the predisposing allele is diluted and its effect is underestimated ., Using simulations , where we know the true risk at the causal locus , we quantify the extent of this underestimation ., We show that , for loci which have a modest effect on disease risk and are common in the population , the risk estimated from the most associated SNP is very close to the truth approximately two thirds of the time ., Although the extent of the underestimation depends on assumptions about the frequency and strength of the risk allele , we predict that fine mapping of GWAS loci will , in rare cases , identify causal variants with considerably higher risk ., Using three common diseases as examples , we investigate the expected cumulative effects of underestimation at multiple loci on our ability to stratify individuals by disease risk and to explain disease heritability . | genetics and genomics/disease models, genetics and genomics/genetics of disease, genetics and genomics/complex traits | null |
journal.pcbi.1002998 | 2,013 | Systematic Prediction of Pharmacodynamic Drug-Drug Interactions through Protein-Protein-Interaction Network | Drug-drug interaction ( DDI ) is a significant cause of adverse drug reactions ( ADRs ) , especially in patient populations on multiple medications ., A recent study indicated that medications were commonly used together in older adults , with nearly 1 in 25 individuals potentially at risk of a major DDI 1 ., Approximately 70% of interactions are clinically relevant and contribute to the majority of ADRs 2 ., DDIs occur when the pharmacologic effect of a given drug is altered by the action of another drug 3 , leading to unpredictable clinical effects ., DDIs can be categorized into three types: pharmaceutical , pharmacokinetic ( PK ) , and pharmacodynamic ( PD ) 4 , 5 ., Pharmaceutical interactions occur because of a physical or chemical incompatibility ., A PK interaction occurs when one medication alters the absorption , distribution , metabolism , or excretion of another , changing the drug concentrations arriving at the target sites ., PD interactions occur if one drug has an antagonistic , additive , synergistic or indirect pharmacologic effect on another ., Current studies mainly focused on PK ( especially Cytochrome P450 enzymes ) DDIs and established experimental and simulation approaches to test for metabolic or transporter-based drug interactions 6 ., However , a large number of DDIs cannot be explained at the PK or pharmaceutical levels and are supposed to be potential PD DDIs ( Figure S1 , Materials and Methods ) ., Many of these interactions are not easily discernible because the endpoint is often a potentially serious adverse event rather than a measurable change in the concentration of the drug 5 ., Typically , the potential PD DDIs were mainly based on sporadic cases reported during clinical trials ., A number of severe PD DDIs are not identifiable in the early stage and result in great losses to human health ., Thus far , the computational solutions to predict DDIs have used two distinct approaches ., The first approach , termed similarity-based , predicted DDIs by measuring the similarity of drug information ., As an example , Gottlieb et al 7 utilizes multiple drug-drug similarity measures to predict DDI ., In this respect , many previous methods which were originally designed for inferring novel potential targets of drugs based on various types of data , such as structures 8 , targets 9 , indications 10 , side-effects 11 and gene expression profiles 12 , can also be used to infer drug interactions ., The second approach is the knowledge-based approach that predicts DDI from scientific literature 13 , an electronic medical record database 14 and the FDA Adverse Event Reporting System 15 ., However , both approaches suffer from several limitations , such as the necessity to distinguish drug classes and the inability to handle novel drugs for which limited reports exist 7 ., More importantly , they seldom consider drug actions and their clinical effects in the context of complex biological networks ., To ameliorate this situation , we adopted a network pharmacology strategy 16 , which considers drug actions and their clinical effects in the context of molecular network systems , and proposed an algorithm to systematically predict PD DDIs ., Using known PD DDIs as golden standard positives ( GSPs ) , we demonstrated the superiority of our algorithm over previously published methods ., The predictions also agreed with similar clinical side effects between the drugs , which was further incorporated with S-score to increase the prediction performance through a Bayesian probabilistic model ., Importantly , our methods provided not only a comprehensive list of potential PD DDIs , but it also opportunities for further understanding of the molecular mechanism and physiological effect underlying DDIs ., To determine whether the network pharmacology strategy can be used to understand DDIs , specifically PD DDIs , we first investigated whether PD DDIs are reflected at the network level ., We examined the distribution of the targets of drug pairs with known PD DDIs among the 1 , 249 FDA-approved drugs collected in DrugBank 17 in a protein-protein interaction ( PPI ) network from HPRD 18 ., The connection for any possible drug pairs was measured by the minimum shortest path between their targets in the PPI network ( Figure 1A , Materials and Methods ) ., Out of the 21 , 049 drug pairs that have minimum target distances of zero , 924 ( 4 . 4% ) were known PD DDIs ( Figure 1B ) , which represents a ∼6-fold enrichment compared with all possible drug pairs ., This is expected because drug pairs with minimum distances of zero are those sharing at least one overlapping target , and these have been reported to have a high probability to form DDIs 9 ., More importantly , we found that the smaller the minimum distance between two drugs targets the more likely a PD DDI occurs ( Figure 1B ) , suggesting PD DDIs can be discerned at the PPI network level ., In fact , the drug pairs with the minimum distance ≤3 already cover the majority ( >80% ) of the known PD DDIs ( Figure 1C ) ., Overall , the average distance of known PD DDI targets is significantly shorter than the global average of possible drug pairs in the network ( P-value<2 . 2E-16 , Wilcoxon rank sum test ) ., Based on the above observation , we designed a metric for systematically predicting PD DDIs by considering drug actions in the context of the PPI networks ., First , drugs were mapped onto a PPI network based on their drug-target associations ( Figure 2A ) ., Second , many drugs exert their therapeutic or adverse effects by interfering with tissue-specific molecular targets that are usually located in the same tissue where a pathological process occurs 19 ., Therefore , we weighed the PPI in the network by Pearsons correlation coefficient ( PCC ) of their encoding genes expression profile across 79 human tissues 20 ( Figure 2B ) ., Then we defined a target-centered system for each drug , which includes drug targets and their first-step neighboring proteins in the PPI network ( Figure 2C ) ., Finally , we defined a system connection score ( S-score ) to describe the connection between two target-centered systems in the PPI network as the following ( Figure 2D ) :where , and represent the mean , standard deviation and number of the cross-tissue expression PCC of edges connecting two drug-centered systems , respectively; represents the average PCC of all edges in the network as background ., In addition , if two target-centered systems share a gene , an artificial edge with PCC of 1 is added between the two systems ., Thus , S-score reflects the tightness of connection between two target-centered systems in the network , which not only depends on the number of edges connecting the genes in these two target-centered systems but also on the similarity in expression patterns across tissues ., To evaluate our scoring scheme , we calculated S-scores for all possible drug pairs among the list of FDA-approved drugs ., Using known PD DDIs collected in DrugBank as GSPs , we first evaluated the correlation between S-score and the likelihood that a PD DDI occurs ., Indeed , the occurrence of PD DDIs decreased with decreasing S-scores among all possible drug pairs ( Figure 3A ) ., Additionally , there was a highly significant correlation between S-score and the hits enrichment of GSPs ( R2\u200a=\u200a0 . 66 , P-value\u200a=\u200a4 . 3E-52 ) ( Figure 3B ) ., It indicated that the likelihood of a PD DDI to occur is high if the two drugs targets are highly connected in PPI network and co-expressed in the same tissues ., Next , we used receiver operating characteristic ( ROC ) curves to examine the performance of our algorithm ., We compared our prediction with previously published methods ( Materials and Methods , Text S1 ) : ( 1 ) target overlap , connecting two drugs if share at least one target 9; ( 2 ) target distance , connecting two drugs by their minimum distance of shortest path between targets on PPI network ( Materials and Methods ) ; ( 3 ) P-score , connecting two drugs by their side-effect similarities 11; ( 4 ) C-score , connecting two drugs by their gene-expression signatures connectivity 12; ( 5 ) indication overlap , connecting two drugs if they share a similar indication 10; ( 6 ) text mining , connecting two drugs based on a co-occurrence scheme 13; ( 7 ) TWOSIDES , a database of polypharmacy side effects for pairs of drugs mined from FDA Adverse Event Reporting System 15; ( 8 ) INDI , a method predicted DDIs utilizing multiple drug-drug similarity measures 7 ., Targets-based methods ( target overlap , target distance and S-score ) are better than those using indication , gene-expression signatures or side-effect similarities to connect drugs ( Figure 3C ) ., Importantly , S-score , by integrating the information from drug-target associations , PPI network topology and cross-tissue gene co-expression , has the best performance ( Figure 3C ) ., Interestingly , using different types of known DDIs as GSPs ( Materials and Methods ) , we found that S-score mainly predicted PD , but not PK or pharmaceutical DDIs ( Figure 3D ) ., Using the DDIs recorded in DrugBank as GSPs , we also observed that our method outperformed previous methods ( Figure S2A ) ., By using the drug-drug associations with medium text mining confidence score from the STITCH database 13 as another evaluation criterion , we also confirmed the robustness of S-score in predicting potential DDIs ( Figure S2B ) , even for our novel predictions which excluded the known DDIs in DrugBank ( Figure S2C ) ., These results excluded the possibility that the performance of S-score was associated with biases of our semi-automatic text-mining method of classifying known DDIs into three types , and demonstrated the good performance of S-score is independent of the GSPs used ., Expectedly , taking a negative set with a different size had a negligible effect on the result ( Figure S2D ) ., To further validate our predictions , we examined the phenotypic effects of our predictions using published drug clinical side effect data 21 ., Based on the observation that similar disorder phenotypes indicate overlapping molecular mechanisms 22 , we asked whether two drugs have similar clinical outcomes if they are highly connected in their target-centered systems ( Figure 2E and 2F ) ., We measured the phenotypic connections between two drugs by their side-effect similarities ( P-score ) following a published algorithm , which was originally designed to infer novel potential targets of marketed drugs 11 ., The drug pairs with high S-scores indeed had more similar phenotypes ( Figure 3E , P-value\u200a=\u200a2 . 0E-72 , Wilcoxon rank sum test ) ., Thus , S-score calculated using PPI network might partially explain the drug phenotypic overlap ., To further increase the prediction performance , we integrated the evidences from S-score and P-score as a likelihood ratio ( LR ) using a Bayesian probabilistic model ( Figure 2G , Materials and Methods ) ., As a result , we observed a clear improvement of prediction specificity and sensitivity ( Figure 3C ) ., The area under the ROC curve ( AUC ) increased from 0 . 674 to 0 . 731 ., In particular , for drug pairs with both evidences , the AUC of LR ( defined as LR ( S-score and P-score ) ) , approached 0 . 812 ( Figure 3C ) ., We applied the algorithm to the FDA-approved drugs and generated a list of prioritized drug pairs where PD DDIs might likely occur ., Overall , the list of 9 , 626 drug pairs with LR ( S-score and P-score ) >2 were 7 . 5-fold enriched for known PD DDIs against all possible drug pairs ( Table S1 ) , which represents an accuracy of 82% and a recall of 62% ( Materials and Methods ) ., To further assess our novel predictions , we evaluated the potential side effects of our novel predictions against the TWOSIDES database 15 , which collected polypharmacy side effects for pairs of drugs from the FDA Adverse Event Reporting System ( Figure 2J ) ., We observed a significant overlap between our novel predictions and TWOSIDES ( P-value<2 . 2E-16 , Fishers exact test ) , where 27% of the novel predictions overlapped the list of TWOSIDES ., The percentage approached 60% for our top 100 novel predicted drug pairs ( Table S1 ) ., The prioritized list together with the available drug indication information , such as whether two drugs were likely co-used , can provide the rationale for which PD DDIs we should be mindful of during clinical trials or treatment ., The most common drugs at the top of the prioritized list of potential PD DDIs were associated with tricyclic antidepressants ( TCA ) ( Table S1 ) , which are primarily used in the clinical treatment of mood disorders such as major depressive disorder ( MDD ) and dysthymia ., It has been reported that patients taking antidepressants have more opportunities to experience DDIs , because antidepressants are often prescribed for months or years ., In addition , patients with depressive disorders typically have comorbid symptoms that require administration of concomitant medications 23 ., Although many of these drug interaction mechanisms remain unclear , it is recommended that concomitant therapy of TCAs should be used with caution considering the major clinical significance 24 , 25 ., As an example , within the top 10 predicted DDIs , a potential interaction was predicted between two TCAs ( desipramine and trimipramine ) ( Table S1 ) ., Such an interaction has been reported to increase the risk of additive QTc-prolongation and serious ventricular arrhythmias in DrugBank 17 ., In our network model , the target-centered systems of these two drugs highly overlapped and connected with correlated cross-tissue gene expression ( Figure 4A ) , which is indicated by an S-score of 9 . 6 ( Students t-test P-value<2 . 2E-16 , compared to all possible drug pairs ) ., Interestingly , both of two drugs target-centered systems are enriched in genes associated with the Gene Ontology “regulation of heart contraction” ( GO:0008016 ) ( P-value\u200a=\u200a6 . 9E-5 and P-value\u200a=\u200a1 . 6E-3 , respectively ) , which might help in explaining the molecular basis of the potential outcome of the concomitant administration of the two drugs ., Our novel predictions together with the information from TWOSIDES provided opportunities for better understanding the potential molecular mechanisms or physiological effects underlying DDIs ( Figure 4B , Figure S3 and Text S1 ) ., As an example , an interaction was predicted to exist between zonisamide and memantine ( Figure 4B ) ., Zonisamide is a sulfonamide anticonvulsant approved for using as an adjunctive treatment of partial seizures in adults with epilepsy by blocking sodium and calcium channels , which leads to the suppression of neuronal hypersynchronization ( i . e . convulsions ) 17 ., Memantine , an amantadine derivative used in the treatment of Alzheimers disease , exerts its action through uncompetitive NMDA receptor antagonism , which protects against elevated concentrations of synaptically released glutamate in the brain of demented patients 17 ., The two drugs do not have common targets , but do have similar cross-tissue expressions between their drug-centered systems ( S-score\u200a=\u200a6 . 5 , P-value<2 . 2E-16 ) and similar side effects ( P-score\u200a=\u200a73 . 5 , P-value<2 . 2E-16 ) ., Although it has not been reported in DrugBank 17 , TWOSIDES recently reported that this drug pair has an significant association with the adverse event thrombocytopenia ( P-value\u200a=\u200a1 . 36E-177 ) in the FDA Adverse Event Reporting System , which cannot be clearly attributed to the individual drugs alone 15 ., Our analysis reveals that the genes in two drug target-centered systems are highly enriched in genes significantly highly expressed in the “Platelet” ( UP_TISSUE ) ( P-value\u200a=\u200a8 . 8E-3 ) ., Interestingly , such an interaction cannot be predicted based only on the knowledge of their drug targets as neither of the individual drugs target gene set is related to the thrombocytopenia symptom ., Yet , consistent with their intended effects , emantines targets are enriched for “N-methyl-D-aspartate selective glutamate receptor complex” ( GO:0017146 ) ( P-value\u200a=\u200a1 . 4E-2 ) , which is involved in Alzheimers disease 26 , while zonisamides targets are enriched for “voltage-gated sodium channel complex” ( GO:0001518 ) ( P-value\u200a=\u200a3 . 1E-4 ) , which is involved in pathological alterations in epilepsy 27 ., Additional examples of novel predictions of the PD DDIs can be found in Figure S3 and Text S1 ., Despite the many methods previously applied to identify potential drug interactions from different aspects , these approaches have various limitations ., To our knowledge , we for the first time , present an algorithm for systematically predicting PD DDIs by considering drug actions and their clinical effects in the context of complex PPI networks ., The integration of various sources of information such as drug targets , network topology , cross-tissue gene expression correlations and side effect similarity indeed give rise to a better performance in predicting DDIs than those obtained with individual data sources ., Finally , our network model provides opportunities for better understanding the potential molecular mechanisms or physiological effects underlying DDIs ., However , like other computational-based techniques in this field , there still exists a gap between our scientific predictions in theory and clinical application ., First , limited by the current knowledge of the molecular network as well as the robustness of the biological system itself , our prediction only provides the relative likelihood of the occurrence of a PD DDI ., Second , as currently only a few types of data were used for prediction , the prediction power is bound to improve when integrated with more clinical data , if available , and complemented with recently published DDI prediction methods from different aspects ., Last , the predicted potential PD DDIs are not necessarily always harmful but sometimes can also be beneficial 28 ., Even though the current GSPs include only a small number of beneficial interactions , such interactions may occur through the same mechanism - overlapping network , in which case can be predicted by our method ., With these further improvements , our method can be potentially applied in drug discovery and development , serving as an in silico systematic screen to provide a list of prioritized potential PD DDIs in a cost-effective manner or be applied to relabeling drug interaction warnings for marketed drugs ., Our method can also reveal potential mechanisms or effects underlying DDIs and provide the necessary scientific evidence for further investigation of the drugs during clinical trials ., These mechanisms could be valuable for rational poly-medication among existing drugs for new purposes to enhance beneficial drug combinations while avoiding harmful DDIs ., Drug information was downloaded from DrugBank database ( http://www . drugbank . ca/ ) on May 9 , 2011 ., In DrugBank , a drug target is defined as “a protein , to which a given drug binds , resulting in an alteration of the normal function of the bound molecule and a desirable therapeutic effect” ., In our further analysis , we mainly focused on the list of 1 , 249 FDA-approved drugs which include 4 , 776 associations with 1 , 289 targets ., A PPI network , including 34 , 998 edges , was taken from Human Protein Reference Database ( HPRD; http://www . hprd . org/ ) 18 on Dec 7 , 2010 ., To weight the edges in the network , we used PCC based on the pair-wise gene expression profiles in 79 human tissues 20 ., To compare the prediction performance of our algorithm with previously published methods , we selected several representative methods in the field of DDI prediction 7 , 13 , 15 and also covered some approaches that were originally designed for inferring novel potential targets of drugs but can also be used to infer drug interactions 8 , 9 , 10 , 11 ., To integrate the evidences from network system connectivity score ( S-score ) and drug phenotypic similarity score ( P-score ) , we used a Bayesian probabilistic model described in Xia et al 29 , where the Bayesian model has been proven to be particularly competent in predicting PPIs by integrating various evidences ., The method has also been used to combine the different types of clues for predicting PPIs in a paper recently published 30 ., Briefly , in the Bayesian probabilistic model , each score is automatically weighted according to its confidence level ., The scoring schemes ( S-score , P-score ) were integrated as a likelihood ratio ( LR ) for drug pairs to be true positive DDIs versus true negative DDIs by multiplying from all the independent evidences as following:where LR ( i-score ) represents the likelihood ratio of evidence i-score ., It relates prior and posterior odds according to the Bayes rule:where the terms ‘posterior’ and ‘prior’ refer to the condition before and after considering the evidence information i-score; the prior odds ( Oprior ) of finding the positive and negative hits can be can be calculated by considering the total number of GSP/GSN DDIs within all the possible drug pairs; the posterior odds ( Oposterior ) can be calculated by binning all possible drug pairs into discrete intervals according to the evidence i-score ., We defined LR ( S-score and P-score ) for drug pairs with both evidences , while LR ( S-score or P-score ) for those with at least one evidence , respectively ., Functional annotation analysis was performed using the DAVID web-server 31 ., The datasets used in this paper and the core code in calculating the S-score were packaged and provided on our website http://www . picb . ac . cn/hanlab/DDI . | Introduction, Results, Discussion, Materials and Methods | Identifying drug-drug interactions ( DDIs ) is a major challenge in drug development ., Previous attempts have established formal approaches for pharmacokinetic ( PK ) DDIs , but there is not a feasible solution for pharmacodynamic ( PD ) DDIs because the endpoint is often a serious adverse event rather than a measurable change in drug concentration ., Here , we developed a metric “S-score” that measures the strength of network connection between drug targets to predict PD DDIs ., Utilizing known PD DDIs as golden standard positives ( GSPs ) , we observed a significant correlation between S-score and the likelihood a PD DDI occurs ., Our prediction was robust and surpassed existing methods as validated by two independent GSPs ., Analysis of clinical side effect data suggested that the drugs having predicted DDIs have similar side effects ., We further incorporated this clinical side effects evidence with S-score to increase the prediction specificity and sensitivity through a Bayesian probabilistic model ., We have predicted 9 , 626 potential PD DDIs at the accuracy of 82% and the recall of 62% ., Importantly , our algorithm provided opportunities for better understanding the potential molecular mechanisms or physiological effects underlying DDIs , as illustrated by the case studies . | Drug-drug interaction ( DDI ) is an important problem in clinical practice ., In this study , we developed a novel algorithm for systematically predicting pharmacodynamic ( PD ) DDIs through protein-protein-interaction ( PPI ) networks ., We calculated a score to predict potential PD DDIs by integrating the information from drug-target associations , PPI network topology and cross-tissue gene expression correlations ., The scoring system was validated by known PD DDIs and agreed with similarities in drug clinical side effects , which we further integrated to increase the prediction performance ., Our approach not only outperformed previously published methods in predicting DDIs , but also provided opportunities for better understanding the potential molecular mechanisms or physiological consequences underlying DDIs . | medicine, protein interactions, drugs and devices, pharmacodynamics, genetics and genomics, pharmacology, biology, proteomics, drug discovery, biochemistry, drug research and development, genetics, genomics, drug interactions, gene networks, computational biology, pharmacogenomics | null |
journal.pcbi.1002694 | 2,012 | Tissue-Specific Functional Networks for Prioritizing Phenotype and Disease Genes | Phenotypes caused by mutations in genes often show tissue-specific pathology , despite organism-wide presence of the same mutation 1 , 2 , 3 , 4 ., Therefore , a logical genomics approach to infer candidate genes and their functions is to integrate large-scale data in a tissue-specific manner ., However , such efforts are hampered by the lack of adequate tissue-specific training and feature data and by the methodologies to model tissue-specificity systematically in human or other mammalian model organisms ., Functional relationship networks , representing the likelihood that two proteins participate in the same biological process , provide invaluable information for phenotype gene discovery , pathway analysis , and drug discovery 5 , 6 , 7 , 8 , 9 , 10 , 11 ., In human and model mammalian organisms , these networks have been used to predict genes associated with genetic diseases or phenotypes through computational mining of the network structure 5 , 6 , 7 , 8 , 10 , 11 ., For example , we have previously generated a mouse functional relationship network and used it to identify that Timp2 and Abcg8 are bone-mineral density ( BMD ) -related genes 11 , though neither of these were previously detected in quantitative genetics studies ., So far , these analyses have been limited to global functional networks representing the overall relationships between proteins without accounting for tissue specificity ., Analyses based on global functional relationship networks , while effective , ignore a critical aspect of biology that could significantly improve their utility: genetic diseases often target specific tissue ( s ) and thus perturbations of proteins or pathways may have differential effects among diverse tissues ., For example , Timp2 , which we have previously identified to be related to BMD 11 , is also involved in the control and/or development of neurodegenerative disease 12 ., Such multi-functionality is not directly reflected by the global network but would be revealed by different connections in tissue-specific networks ., Therefore , computational modeling and analyses of tissue-specific networks are needed to identify phenotype-associated genes that exhibit tissue-specific behavior ., Current approaches to create functional relationship networks are difficult to apply in a tissue-specific manner ., Typically , networks are constructed by integrating data sources that vary in terms of measurement accuracy as well as biological relevance for predicting protein functions ., Machine learning methods , such as Bayesian networks , learn the relative accuracy and relevance of datasets when given a ‘gold standard’ training set , which consists of gene pairs that are known to work in the same biological process ., Then probabilistic models are constructed to weigh and integrate diverse datasets based on how accurately they recover the ‘gold standard’ set ., The networks generated by this approach lack tissue-specificity information , because systematic collections of large-scale data or ‘gold standard’ pairs with quantitative tissue-specific information are often not available ., Here , we address the tissue-specificity challenge by simulating the natural biological mechanism that defines tissue-specificity: co-functionality in most cases would require the presence of both proteins in the same tissue ., Inspired by our previous efforts to establish biological process-specific networks , such as networks specifically related to the cell cycle or to mitochondrial biogenesis 13 , 14 , 15 , we integrate low-throughput , highly-reliable tissue-specific gene expression information ( e . g . RT-PCR , in situ hybridization , etc . ) from the Mouse Gene Expression Database ( GXD ) into our probabilistic framework when learning the reliability of each data source ., Such an approach is more intuitive for the tissue-specific network setting because it is relatively less likely that a non-expressed gene would collaborate with an expressed gene even though they are ‘functionally related’ in the global sense ( i . e . co-annotated to either a GO term or a KEGG pathway ) ., There are exceptions to this guideline , including signaling and hormonal pathways that traverse multiple organ systems ., However , many cellular processes important for phenotypes are largely restricted to specific tissues ., Therefore , by constraining the ‘gold standard’ to pairs of genes that are both expressed in a tissue , we are able to establish functional networks that are highly specific in capturing the dynamic properties of different tissues ., In addition to generating the first tissue-specific networks for the laboratory mouse , we also explicitly tested the potential of using such networks to predict phenotype-associated genes ., To do so , we mapped diverse phenotypes to their respective tissues in the laboratory mouse , according to the terminology and description of the phenotypes ., We show that the tissue-specific functional relationship networks can improve our prediction accuracy for phenotype-associated genes compared to a single global functional relationship network through computational analyses , and through experimentally confirmed predictions of novel fertility-related genes and visualization of their local networks ., We further identified candidate genes specifically predicted by the cerebellum network to be related to ataxia , which are supported by both literature and experimental evidence ., Our networks are publicly available at http://mouseMAP . princeton . edu , which features the ability to compare networks across tissues for analyzing the dynamics of functional relationships ., Our current framework covers 107 major tissues in the laboratory mouse and focuses on cross-network comparison and phenotype-associated gene discovery ., However , as more data become available , this approach will serve as a prototype for applications to pathway analyses and drug screening ., A common mechanism resulting in tissue-specific protein functionality is the modulation of gene expression levels between tissues 16 , 17 , 18 ., This observation is our theoretical foundation for establishing tissue-specific networks , in which links between proteins represent the probability that they are involved in the same biological processes within a specific tissue ., To simulate such tissue-specificity , we developed a Bayesian approach ( Figure 1 ) that incorporates highly-reliable , low-throughput measures of tissue-specific gene expression into training set , which we utilized to produce networks focused on the real functional relationships occurring within the tissue under consideration ., This Bayesian framework essentially learns how informative each dataset is given a set of ‘gold standard’ training pairs , i . e . pairs of proteins known to be functional in the same biological process and both expressed in the tissue of interest ., In the global ( non-tissue-specific ) sense , following previous definitions 5 , ‘gold standard positives’ are defined by co-annotation to specific Gene Ontology ( GO ) biological process terms 19 , while ‘gold standard negatives’ are defined as pairs that both have specific GO annotations yet do not share any annotations ., For each tissue-specific gold standard set , a positive pair has to meet two requirements: first , the pair must be ‘co-functional’ as defined in the global sense , and second , both genes must be expressed in the tissue under consideration as evident in highly reliable , low-throughput expression datasets , which , in most cases , is necessary for the pair to have a functional relationship in that tissue ., These tissue-specific gold standards are then used to quantify how relevant each genomic dataset is in recovering tissue-specific functional relationships , regardless of the tissue of origin for each genomic dataset ., This allows us to leverage the entire compendium of high-throughput genomic data to generate accurate tissue-specific networks , even for tissues which do not have existing tissue-specific whole-genome experiments , by relying on non-tissue-specific datasets , heterogeneous samples , and potentially related tissues and experiments ., For example , biliary tract , which is not specifically represented in our current collection of high-throughput features used for classification , can still be accurately predicted by utilizing information from related , heterogeneous samples , such as gene expression microarrays of whole liver or the hepatic system , as well as non-tissue-specific information , such as sequence phylogeny and in vitro binding assays ., Thus our approach can leverage the implicit relationships between a dataset and a tissue and therefore enables generation of tissue-specific networks even from feature data that is not resolved for a specific tissue type ., For tissue-specific expression information , our gold standards rely on the Gene Expression Database ( GXD ) of the Mouse Genome Informatics group ( MGI ) ., GXD provides an extensive , hierarchically structured dictionary of anatomical expression results for mouse to allow us to carry out our analysis 20 ., The data in GXD are derived from traditional , “small-scale” expression experiments , such as in situ hybridization , RT-PCR , and immunohistochemistry , which simply reflect presence or absence of a gene within the tissue examined ., No high-throughput expression data were used for our gold standard construction ., In total , there are 107 tissues included in our analysis ., We pursue two main goals in this study: First , we generate tissue-specific networks that synthesize as much data as possible and provide these networks to the public through an online visualization interface at http://mouseMAP . princeton . edu ., For this , we gathered diverse genomic data for mouse as inputs ( Dataset S1 ) to support the functional relationships , including protein-protein physical interactions 21 , 22 , 23 , 24 , homologous functional relationship predictions from simpler organisms 9 , phenotype and disease data 19 , 25 and 960 expression datasets , totaling 13632 experimental conditions 26 , 27 , 28 , 29 ., The reliability of each dataset is learned through Bayesian network classifier training , using the tissue-specific gold standards described above ., Essentially , a dataset deemed more relevant and accurate for the tissue under consideration will be given higher weight , and the final probability of pair-wise functional relationships is determined by updating the initial probability ( prior ) based on the weighted input of all genomic datasets ., This procedure resulted in tissue-specific probabilistic functional relationship networks for the laboratory mouse that effectively summarize these diverse data sources and enable biology researchers to easily explore the resulting functional landscape ., Second , we test the hypothesis that tissue-specific networks could assist us to predict phenotype-related genes more accurately ., In this case , to prevent circularity in our methodology , phenotype and disease data were excluded from network generation , and the results were used to predict novel phenotype-associated candidate genes ., We demonstrate that tissue-specific networks enhance biological clarity and result in more accurate predictions ., Our resulting networks and predictions provide biology researchers with functional interactions specific to each tissue as well as phenotype hypotheses of genes ., One key application of tissue-specific networks is to identify novel genes and relationships between genes that may be specific to a particular tissue ., To computationally evaluate our ability to identify novel relationships , we used cross-validation to test whether our tissue-specific Bayesian scheme is more accurate than the global network ., Cross-validation was used to assess predictions by evaluating the accuracy of recovering subsets of known annotations withheld during the training process ., Specifically , we performed 3-fold cross-validation , by holding out one third of the tissue-specific ‘gold standard’ pairs in each of the three iterations ., We learned the parameters in the Bayesian networks , i . e . , the reliability of each dataset , through the other two thirds of the ‘gold standard’ , and then used these networks to predict the probabilities for the held-out one third of the protein pairs ., Compared to a single global functional relationship network , our approach significantly improved our ability to predict tissue-specific functional linkages ., The mean AUC ( area under the receiver operating characteristic curve , which represents the accuracy in recovering tissue-specific functional relationships ) for the global network estimated through three-fold cross-validation was 0 . 68 ., Tissue-specific networks achieved median AUC of 0 . 72 ., With a random baseline of 0 . 5 in AUC , this represents a ∼20% improvement of the tissue-specific networks over the predictive power of the global network ., This improvement is consistent over all 12 major organ systems defined by GXD 20 ., ( Figure 2A ) ., Immune system-related networks acquired the most median improvement of 22 . 7% and digestive system-related networks achieved least median improvement of 14 . 3% ., For example , for lymphoid system ( MA:0002435 ) , we improved our AUC from 0 . 65 to 0 . 72 and for ventricular zone , brain , we improved from 0 . 65 to 0 . 77 ., Such improvement is consistent across the entire precision-recall spaces ( Figure 2B , Dataset S2 for all precision-recall curves ) ., In all cases , tissue-specific networks performed better than the global network in predicting functional relationships specific to that tissue , which demonstrates the robustness of our integration approach across different systems and tissues in the laboratory mouse ., One important application of our tissue-specific networks is to identify functional relationships between genes that change significantly across tissues ., This provides a platform for analyzing tissue-specific molecular interactions , as well as tissue-specific roles for genes that are ubiquitously expressed but play different roles in different tissues ., For example , Wnt10b ( wingless related MMTV integration site 10b ) is expressed in many tissues throughout development and participates in many biological processes including bone trabecular formation 30 and cell differentiation involved in skeletal muscle development 31 ., The interactors of Wnt10b in our muscle-specific and bone-specific functional networks reflect its differential roles in these two tissues ., The top neighbors in the muscle-specific network consist of genes responsible for skeletal muscle development ( Figure 3A ) ., For example , BIN1 participate in the biological process muscle cell differentiation ( GO:0042692 ) 32 , PLAU is involved in the process skeletal muscle tissue regeneration ( GO:0043403 ) in rat and MYF6 directly function in muscle cell biogenesis 33 ., In fact , 8 out of the 19 top connected nodes of Wnt10b in the muscle-specific network are involved in skeletal muscle cell development , reflecting the functional role of Wnt6b ., On the contrary , in the bone-specific functional network , the top neighbors of Wnt10b consist of genes involved in bone mineralization and bone structure formation ( Figure 3B ) , representing 12 out of the 19 top connected nodes ., This observation suggests that our networks can provide a resource for comparing the dynamic functions of a single gene across different tissues ., To quantify gene connectivity changes across networks , we developed a metric that captures how much the edges involving a gene differ across networks ( see methods ) , and we implemented a web-based visualization interface ( http://mouseMAP . princeton . edu ) allowing users to query genes of interest and compare the local network between tissues ., Essentially , connectivity change of a gene is defined by the sum of absolute values of fold changes ( over prior ) of connections between this gene to all other genes ., Some genes vary greatly in their connectivity between tissues , potentially reflecting their tissue-specific roles ., Of the top 100 altered genes , they were significantly enriched for “anatomical structure development” ( GO:0048856 ) and “organ development” ( GO:0048513 ) ., Additionally , genes with connectivity altered in specific tissues compared to the global network , tend to be enriched for GO terms related to the tissue under consideration ., For example , when comparing the nervous system-specific network ( MA:0000016 ) against the global network , the top changed genes are enriched in “central nervous system development” ( GO:0007417 ) , “diencephalon development” ( GO:0021536 ) , and “brain development” ( GO:0007420 ) ( Table 1 ) ., The full enrichment analysis is provided in Dataset S3 ., A key hypothesis in this study is that analyzing tissue-specific networks may improve our ability to identify phenotype-related genes ., To test this hypothesis , we regenerated tissue-specific networks using the same Bayesian approach as above , but excluded all phenotype and disease data as inputs to avoid circularity in our cross-validations ., Then , we mapped 451 phenotypes to their most related tissue in the laboratory mouse according to the terminology and description of these phenotypes in the Mammalian Phenotype ontology 19 ., For each phenotype , we compared novel predictions made using the appropriate tissue-specific network as compared to using the global network ., This method is based on our previously developed machine learning scheme ( network-based SVM ) 11 that mines information in functional relationship networks to prioritize candidate genes according to their links to known genes related to a disease or phenotype ., To test whether our tissue-specific networks are more capable of identifying phenotype-associated genes than the global network , we used bootstrap bagging 34 to evaluate which network performs better ., Bootstrap bagging is suitable for phenotype predictions , where positive examples ( known phenotype-associated genes ) and negative examples ( random genes ) are highly imbalanced 35 ., Its stability and comparably good performance in estimating error rates has been tested in extensive simulations for positive example set sizes ranging from less than 20 35 to >200 36 , which is the approximate range we are using in our evaluation ., For the 451 mapped phenotypes , the median AUC when utilizing tissue-specific networks is 0 . 794 , representing an improvement of 11 . 8% over utilizing the global functional network ., For many phenotypes , using tissue-specific networks can improve our ability to extract potentially experimentally-verifiable predictions ., For example , at one percent recall ( the low recall end is where most of the follow-up experimental confirmations will focus on ) , we achieved a precision of 1 . 00 compared to 0 . 33 using global network for the phenotype abnormal spleen white pulp morphology ( MP:0002357 ) , and a precision of 0 . 5 compared to 0 . 28 for abnormal malpighian tuft morphology ( MP:0005325 ) ., Additionally , the AUC for “abnormal osteogenesis” ( MP:0000057 ) was 0 . 77 using the global network , but 0 . 81 for tissue-specific networks ., The AUC for “abnormal nervous system electrophysiology” ( MP:0002272 ) using the global network was 0 . 716 , but was 0 . 763 using the nervous system-specific network ( Figure 4C for example precision-recall curves ) ., Such significant improvement demonstrates the potential of mining tissue-specific networks to prioritize phenotype-associated genes ., Performance improvements were consistent across phenotypes of different sizes ( Figure 4A ) ., For phenotypes with 300–1000 annotated genes ( around 1 . 5% to 5% of genome ) , we achieved a median AUC of 0 . 814 ( improvement of 8 . 7% ) ; for phenotypes with 100–300 genes , the median AUC was 0 . 792 ( improvement of 13 . 0% ) ; and for phenotypes with 30–100 genes , the median AUC was 0 . 769 ( improvement of 11 . 0% ) ., At 10 percent recall for the 300–1000 , 100–300 , and 30–100 groups , we achieved precisions of 14 . 8 , 17 , and 20 fold over random , respectively ., This consistency indicates the robustness of tissue-specific networks against the number of known genes in predicting phenotype-associated genes ., Performance improvements were also consistent across different major organ systems ., Phenotypes involved in the endo/exocrine system achieved the most significant improvement in AUC ( +35% , compared to global networks against baseline of 0 . 5 ) and those in cardiovascular system achieved 21 . 8% improvement in AUC ., However , prediction accuracy was improved across all major systems , with the least improvement of 5 . 9% in renal/urinary phenotypes ., Phenotypes related to musculoskeletal systems achieved the highest AUC of 0 . 82 and the group with lowest AUC was digestive system , which still achieved an average of 0 . 78 ., The consistency in improvements across different organ systems demonstrates the robustness of our modeling framework to predict phenotype-related genes in a tissue-specific manner ., We focused on two cases to illustrate how our tissue-specific networks can facilitate disease gene discovery ., These two phenotypes represent two extremes of the phenotype/disease-associated gene prediction problem ., The first , reduced male fertility , is a broadly defined , common phenotype with many causative genes already known ., The second , ataxia , is a rare neurological disorder affecting ∼3–10/100 , 000 of the general population 37 , 38 , 39 ., Roughly 40 genes are known to be associated with this disease , but the majority of both familial and sporadic cases remain unexplained ., Predicting candidate genes related to rare genetic diseases is challenging in that little prior knowledge is available for these diseases ., These phenotypes are related to two different tissue-categories ( reproductive and neurological systems ) , enabling us to highlight the broad applications of our approach across organ systems ., We used these two examples and experimental confirmations to demonstrate the power of tissue-specific networks to discover disease genes ., First , we used male fertility related phenotypes to test the performance of tissue-specific networks to predict phenotype-related genes ., To do so , we utilized a recent , nearly comprehensive literature review of genes involved in mammalian spermatogenesis and male fertility phenotypes 40 , which we organized into a hierarchy of male fertility-related phenotypes ( Dataset S4 ) ., This curation effort is independent of , and more comprehensive than , the current GO or MP annotations related to male fertility , which makes these lists excellent , non-circular test sets ., We tested whether the testis-specific network could predict male fertility genes more robustly than the globally integrated network , and found that the testis-specific network significantly improved our ability to predict spermatogenesis-related phenotypes ., For example , for predicting genes related to ‘spermatid head and nuclear modifications , ’ we achieved 4 . 5-fold improvement in precision at 1 percent recall; for ‘acrosome-related genes , ’ we achieved 3 . 6-fold improvement; and for ‘germ/Sertoli cell interaction genes , ’ we achieved 3 . 3-fold improvement ., On the other hand , for terms that are not specifically related to male-reproductive systems , such as ‘association with methylation and acetylation , ’ and ‘association with Golgi Apparatus , ’ we observe no performance improvements using the testis-specific network ., This illustrates that tissue-specific functional relationship networks are tuned to predict phenotypes closely related to these tissues ., We selected Mybl1 to demonstrate the specific utility of the male-reproductive network to predict fertility related genes ., Mybl1 ( MGI:99925 ) is among our top candidates in multiple phenotypes related to male fertility , including ‘association with chromatoid body and manchette’ , ‘transcription factor involved in spermatogenesis’ and , ‘spermatogenesis’ ., However , in the global network , Mybl1 was not a strong candidate for these phenotypes , as it was predicted with negative values ., Therefore Mybl1 is an ideal candidate to test the accuracy of our tissue-specific network-based phenotype predictions ., In our male-reproductive network , the majority of the top interactors of Mybl1 are indeed well-known male fertility genes ( Figure 5A ) , including Dmc1 ( required for meiosis and male fertility 41 ) , Ddx4 ( a DEAD-box helicase required for male , but not female , germ cell development 42 ) , Cyct ( encoding testis-specific cytochrome c 43 ) and Lhx9 ( a LIM homeobox required for sex differentiation and normal fertility 44 ) ., Moreover , Mybl1 was independently identified recently in an unbiased mutagenesis screen for infertility phenotypes involving meiotic arrest 45 ., We found that the Mybl1 mutants are characterized by low testis weight and depletion of male germ cells , as shown in Figure 5B ., Additionally , analysis of the mutant testis transcriptome suggested that MYBL1 is a “master regulator” of the meiotic cell cycle and transcriptional program 46 , and at least one gene regulated by MYBL1 , Cyct , is among the top interactors of MYBL1 predicted by our network ., Together , these findings on an infertility phenotype and suggestions of a corresponding potential mechanism confirm the accuracy of predictions from our tissue-specific network and show that when taken with expression analyses and other data , they can be used as a basis for functional testing ., In addition to the well-studied phenotype of male infertility , we also examined a less well-understood disease , ataxia , to investigate whether our tissue-specific networks can identify genes related to phenotypes or diseases with limited prior knowledge ., Gene identification through genetic approaches , such as pedigree analyses , has had a major impact on our understanding of ataxia ( over 40 candidate genes identified so far ) ., Genetic testing is now an integral part of assessment ., Routinely , a blood sample of any new ataxia case is mailed in for laboratory evaluation ., However , the majority of the sporadic cases as well as the familial cases are so far unexplained ., We curated the known gene list ( 43 in total ) related to human ataxia , mapped these genes to their mouse orthologs , and used this list as seeds to predict additional candidate genes using our cerebellum-specific network , which is the major tissue affected by ataxia ., Our cerebellum-specific network reveals connections of ataxia-related genes not shown in the global network ., A key , known ataxia gene is Atcay ( ataxia , cerebellar , Cayman type homolog ( human ) ) , and in the cerebellum-specific network , two of its top interactors are Cacna1e ( with connection confidence 0 . 943 , ranked 18 ) and Grm1 ( 0 . 902 , ranked 46 ) ( Figure 6 ) ., These are plausible candidate genes since Grm1 is a known mouse ataxia gene 47 , and Cacna1e encodes a subunit of an R-type calcium channel , while mutations to the related protein family member Cacna1a , encoding a subunit of an L-type calcium channel , causing spinocerebellar ataxia ., However , in the global network these interactions are much weaker ( 0 . 647 for Cacna1e and 0 . 763 for Grm1 respectively ) , and would not be identified in the top 100 connections of Atcay , which supports the utility of tissue-specific networks relevant to ataxia to identify candidate genes ( Figure 6B ) ., In addition to identifying these novel , likely correct edges , we also identified novel candidates using our SVM-based approach described above ., Out of our top 10 novel candidates , we found strong evidence in the literature for 4 of these genes to be associated to ataxia ( Table 2 ) ; suggesting at least a 40% success rate at low levels of recall ., Among these , SORBS1 physically interacts with ATXN7 , an autosomal dominant gene causing cerebellar ataxia 48 ., RBFOX1 physically interacts with the c-terminus of ATXN2 , another autosomal dominant gene causing cerebellar ataxia 48 , 49 ., It is thought that RBFOX1 might contribute to the restricted pathology of spinocerebellar ataxia type 2 ( SCA2 ) 50 ., The homozygous mouse knockout of a third gene , Plcb4 induces ataxia , although no human patients have been identified with mutations in this gene ., A fourth gene , Plp1 , is implicated in Spastic paraplegia-2 and Pelizaeus-Merzbacher diseases 51 , which are disorders closely related to ataxia ., It is also a homologue of Pmp22 , which is involved in Charcot Marie Tooth disease type 1A , a sensory neuropathy common in some forms of ataxia 52 ., Thus , even in the case of less well-studied phenotypes or diseases , our tissue-specific approach is able to identify likely candidates as evidenced by our success rate of at least 40% for ataxia-related predictions based on the cerebellar network , compared to a background detection rate of less than 1/500 ., Genetic diseases often manifest tissue-specific pathologies 1 , 2 , 3 , 4 ., Therefore , acquiring tissue-specific functional information is essential for biomarker identification , diagnosis , and drug discovery ., Current integrative functional genomics approaches to study diseases or phenotypes generally do not analyze them in the context of specific tissues ., Our work represents a conceptual advance to address tissue-specificity in genome-scale functional studies of phenotypes ., We describe a strategy to systematically generate tissue-specific functional networks that are robust and accurate for mining phenotype-related genes , demonstrating the importance of tissue-specific approaches for understanding human diseases ., Our approach addresses the twin challenges of incomplete systematic knowledge of tissue-specific protein functions and of limited availability and coverage of tissue-specific high-throughput functional data ., Due to this lack of systematically defined tissue-specific genomic data , our approach uses highly reliable , low-throughput measures of gene expression to constrain our gold standard examples into tissue-specific sets ., As more tissue-specific protein functions are defined systematically , perhaps with the help of hypotheses generated by approaches such as this , tissue-specific functional interactions will be directly used for experimental testing ., Many genomic datasets , especially physical interaction studies , such as yeast 2-hybrid screens , and large-scale genetic screens , utilize artificial or in vitro contexts that may or may not reflect tissue-specific functional roles ., Other data , however , such as high-throughput gene expression datasets ( e . g . microarrays or RNA-seq ) , is often collected in a specific tissue or cellular context and may thus reflect a more restricted , tissue-specific set of genes or proteins ., In our approach , we use the power of Bayesian machine learning to learn the predictive power of each dataset , whether in vivo or in vitro , by utilizing training sets restricted to gene pairs that are both expressed in the same tissue or context ., In this way , data from empirically relevant contexts are trusted , while irrelevant data are disregarded ., While our current study focuses on predicting genotype-phenotype associations using tissue-specific functional relationship networks , the potential application of tissue-specific networks extends far beyond predicting phenotype-associated genes ., For example , just as perturbations of the same gene may lead to different phenotypic outcomes across different tissues; treatments with bioactive chemicals or drugs may manifest differential effects across different tissues ., Our broad conceptual framework of utilizing tissue-specific expression to refine a global network could be brought into these application domains such as drug target identification ., For each pair of networks , we quantify how much each gene | Introduction, Results, Discussion, Methods | Integrated analyses of functional genomics data have enormous potential for identifying phenotype-associated genes ., Tissue-specificity is an important aspect of many genetic diseases , reflecting the potentially different roles of proteins and pathways in diverse cell lineages ., Accounting for tissue specificity in global integration of functional genomics data is challenging , as “functionality” and “functional relationships” are often not resolved for specific tissue types ., We address this challenge by generating tissue-specific functional networks , which can effectively represent the diversity of protein function for more accurate identification of phenotype-associated genes in the laboratory mouse ., Specifically , we created 107 tissue-specific functional relationship networks through integration of genomic data utilizing knowledge of tissue-specific gene expression patterns ., Cross-network comparison revealed significantly changed genes enriched for functions related to specific tissue development ., We then utilized these tissue-specific networks to predict genes associated with different phenotypes ., Our results demonstrate that prediction performance is significantly improved through using the tissue-specific networks as compared to the global functional network ., We used a testis-specific functional relationship network to predict genes associated with male fertility and spermatogenesis phenotypes , and experimentally confirmed one top prediction , Mbyl1 ., We then focused on a less-common genetic disease , ataxia , and identified candidates uniquely predicted by the cerebellum network , which are supported by both literature and experimental evidence ., Our systems-level , tissue-specific scheme advances over traditional global integration and analyses and establishes a prototype to address the tissue-specific effects of genetic perturbations , diseases and drugs . | Tissue specificity is an important aspect of many genetic diseases , reflecting the potentially different roles of proteins and pathways in diverse cell lineages ., We propose an effective strategy to model tissue-specific functional relationship networks in the laboratory mouse ., We integrated large scale genomics datasets as well as low-throughput tissue-specific expression profiles to estimate the probability that two proteins are co-functioning in the tissue under study ., These networks can accurately reflect the diversity of protein functions across different organs and tissue compartments ., By computationally exploring the tissue-specific networks , we can accurately predict novel phenotype-related gene candidates ., We experimentally confirmed a top candidate gene , Mybl1 , to affect several male fertility phenotypes , predicted based on male-reproductive system-specific networks and we predicted candidates related to a rare genetic disease ataxia , which are supported by experimental and literature evidence ., The above results demonstrate the power of modeling tissue-specific dynamics of co-functionality through computational approaches . | genomics, biology, computational biology, genetics and genomics | null |
journal.pcbi.1003737 | 2,014 | A Scalable and Accurate Targeted Gene Assembly Tool (SAT-Assembler) for Next-Generation Sequencing Data | As gene families of interest are used as input , our algorithm employs homology search against gene families in assembly graph construction ., Using genomes or proteome of related species to boost and optimize genome assembly has been proposed or implemented in a group of assembly programs 23 , 24 , 27–35 ., The contigs belonging to a single gene or a block of genome in the related species are ordered , oriented , and assembled ., Most of these programs are designed to improve genome assembly ., A few of these comparative or gene-boosted assembly programs are specifically designed for RNA-Seq or metagenomic assembly ., For example , Surget-Groba et al . 30 carefully considered the highly heterogeneous sequence coverage of transcripts and employed multi-k and proteome of a related species to optimize transcriptome assembly ., Dutilh et al . 29 used one closely related reference genome to increase assembly performance of microbial genomes in metagenomic data ., Yes group used homologous genes to stitch gene fragments for gene assembly in metagenomic data 24 ., Our work is different from existing comparative assembly approaches in the following aspects ., First , our tool does not require any related species as input ., Most of the existing comparative approaches are limited by the availability of closely related reference genomes ., Low similarity between the to-be-assembled genes or genomes and the related genes or genomes can lead to wrongly assembled contigs ., Our tool uses well-characterized genes of particular interest or ubiquitously represented sequence families such as those from family databases of proteins , domains , or functional sites as input to guide assembly ., Second , as we use sequence families rather than a single sequence as reference , profile-based alignment methods rather than pairwise sequence alignment or exact sequence mapping are applied to conduct homology search ., Profile-based methods tend to be more sensitive for remote homology search ., Third , to our best knowledge , SAT-Assembler is the first tool that uses consistency between sequence overlap and alignment overlap for edge creation in an overlap graph ., Another type of RNA-Seq assembly tool , Cufflinks , assembles gene isoforms due to alternative transcription and splicing and improves transcriptome-based genome annotation 17 ., Cufflinks and SAT-Assembler require different types of input and are targeted at different applications ., Their major differences are summarized below ., First , Cufflinks needs the reference genome as input while SAT-Assembler uses sequence families as input ., The input families do not need to contain any genomic sequence or protein products from the reference genome ., Instead , they may contain a large number of gene or protein sequences from other species with variable evolutionary distances ., For example , the average sequence identity of protein ( domain ) families in Pfam ranges from 20% to over 90% ., The conservation and evolutionary changes of the member sequences are summarized into a profile HMM , which can share high or low similarity with the genes in the reference genome ., Second , Cufflinks and SAT-Assembler have different applications ., Cufflinks is used to annotate transcripts and gene isoforms for species with known reference genomes ., SAT-Assembler is applied to RNA-Seq data of non-model organisms or metagenomic data , which do not have reference genomes ., Third , although both tools conduct sequence alignment between reads and the reference genomes or sequence families in the first step , the alignment algorithms are highly different ., The read alignment in Cufflinks relies on read mapping tools such as TopHat 36 and Bowtie 37 , which allow only minor changes caused by , for example , sequencing errors ., On the other hand , the profile HMM-based alignment in SAT-Assembler can handle a large number of evolutionary changes including substitutions , insertions , and deletions ., Our tool can be divided into two main stages ., First , we align reads against profile hidden Markov models ( HMMs ) , which effectively represent the underlying gene families ., This stage classifies the whole input data set into subsets of reads sequenced from different gene families ., Second , SAT-Assembler constructs a family-specific overlap graph and assembles reads from the same family into contigs using a graph traversal algorithm ., The graph construction is supervised by the alignment information from the first stage and aims to obtain maximum connectivity between reads while avoiding false connections ., In particular , it can accurately capture small overlaps between reads from lowly sequenced regions and improves the assembly of lowly transcribed or encoded genes ., The graph traversal is guided by multiple types of information to avoid generation of chimeric contigs ., Finally , paired-end reads are used to scaffold contigs from the same genes into super contigs , which are sets of contigs that are from the same scaffolds ., Fig . 1 is a schematic representation of the pipeline of SAT-Assembler ., Our method conducts homology search on reads first ., Depending on the algorithms and the target databases , homology search methods can be divided into two types ., The first type compares the sequences against protein sequence databases using pairwise alignment tools such as BLAST 38 ., The second type uses profile-based homology search to classify queries into characterized protein domain or family databases such as Pfam 39 , 40 , TIGRFAM 41 , FIGfams 42 , InterProScan 43 , etc ., Applying profile-based homology search to NGS reads has several advantages ., First , the number of gene families is significantly smaller than the number of sequences , rendering a much smaller search space ., For example , there are only about 13 , 000 manually curated protein families in Pfam ., Together these cover nearly 80% of the UniProt Knowledgebase and the coverage is increasing every year as enough information becomes available to form new families 40 ., As the profile-based homology search tool HMMER is as fast as BLAST 44 , using profile-based search provides a shorter search time ., Second , previous work 45 has shown that using family information can improve the sensitivity of remote protein homology search ., For the transcriptomes of non-model organism and metagenomic data , sensitive remote homology search is especially important for identifying possible new homologs ., Third , although short reads can pose challenges to both types of homology search 46 , 47 , empirical studies on thousands of families 47 showed that the performance of profile-based homology search improved quickly with increasing read size ., For a read length of 85 bases , the sensitivity is close to 1 . 0 for moderately and highly conserved domains ., Thus , for read lengths produced by modern NGS technologies , profile-based homology search methods are capable of classifying many reads into their native families with high specificity ., SAT-Assembler aligns query reads against input families using HMMER with the default E-value threshold 10 ., Reads that generate HMMER hits are classified into the corresponding family and fed into the next stage ., In most cases , a read can only be classified into a single family ., However , because some input families share similarities , some reads may be classified into multiple input families ., In practice , we only classify a read into at most three families with the three smallest E-values ., The first stage not only classifies query reads into their native families but provides important alignment information for de novo assembly ., The alignment positions are used to guide overlap graph construction ., A standard overlap graph is defined as , where each non-duplicate read is a node and an overlap larger than a given cutoff is indicated by a directed edge ., Our overlap graph is different from a standard overlap graph 48–50 in the edge creation criteria and graph construction procedure ., In a standard overlap graph , edge creation only depends on the sequence overlap , which is not ideal for genes of heterogeneous sequence coverage ., We add edges by considering the relationship between two types of overlaps: alignment overlap and sequence overlap ., As HMMER outputs alignments represented by amino acids , all overlaps are converted into the unit of bp for consistency ., For simplicity of explanation , a read corresponds to vertex in G . For two reads and , an edge is created from the corresponding node to node if the following criteria are satisfied:, i ) the alignment position of is smaller than ;, ii ) the alignments of and overlap by at least , a user-defined threshold;, iii ) the sequence overlap of the two reads is consistent with the overlap in their alignment positions ., Suppose aligns to the model between and , and aligns to the model between and , where and are alignment starting positions in the model and and are alignment ending positions in the model ., The alignment overlap is the number of bases converted from the number of amino acids in the overlapping region between and ., For example , the overlapping region between and in Fig . 2 . ( C ) contains 22 amino acids , which are converted into 66 bases of alignment overlap ., Criterion 3 is the key observation to connect reads that are sequenced from the same gene rather than from orthologous or paralogous genes because the latter can have very different sequence and alignment overlaps ., An example is given in Fig . 2 , in which read and read are from two homologous genes ., Their sequence overlap and alignment overlap are 25 and 66 respectively ., Other assembly tools such as Trinity will create an edge between and when the k-mer size is 25 ., However , because their alignment overlap and sequence overlap are inconsistent , SAT-Assembler does not connect them , avoiding a wrong connection between reads from homologous genes ., The consistency-based edge creation also allows us to improve connectivity in regions with low sequence coverage ., For relatively small overlaps , we still allow an edge if the alignment overlap and sequence overlap are similar to each other ., The intuition is that the chance that reads with random overlaps can be aligned to the same model with similar alignment overlap is small ., To quantify the consistency between the two types of overlaps , we introduce the relative overlap difference defined by , where is alignment overlap and is sequence overlap ., Criterion 3 is satisfied only when , where is a predefined cutoff with a default value of 0 . 15 ., We examined relative overlap difference in both real RNA-Seq and metagenomic data sets ., For reads sequenced from the same gene and from different genes , the average relative overlap differences are 0 . 072 and 0 . 89 respectively ., To avoid small random overlaps , we use 20 as the default value for , the alignment overlap threshold ., Our overlap graph construction is different from standard overlap graph construction in that it does not need all-against-all sequence comparison ., We first sort the reads by their alignment positions in a non-decreasing order ., We only check the sequence overlap between two reads if their alignment overlap passes ., Therefore , the alignment information increases the efficiency of graph construction ( running time analysis can be found in the section of Running Time Analysis ) ., To incorporate substitution sequencing errors introduced by some NGS platforms , we allow a certain number of mismatches in the sequence overlap ., That is , the overlap between two hits and is the longest suffix of that has a Hamming distance to a prefix of ., In our current implementation , ., The parameter can be adjusted to fit the error rate of the input data ., Transitive edges correspond to edges whose two end nodes are connected by an alternative path ( usually with higher coverage ) ., They add unnecessary edges without contributing to the connectivity of the graph and are removed before de novo assembly ., Before removing them , SAT-Assembler keeps a record of all the pairs of nodes connected by transitive edges because a transitive edge indicates that a pair of nodes are from the same gene region ., This information will be used to guide the graph traversal ., If a node has only one outgoing edge that points to another node that has only one incoming edge these two nodes can be merged as one node ., Tips are identified and removed using the topology-based pruning methods as in Velvet 11 ., Although our edge creation method excludes most random sequence overlaps , some erroneous edges still exist ., An edge is highly likely to be erroneous if it is inferior to another edge that shares a head node or tail node with it ., An edge is inferior to another edge if the following two criteria are met:, i ) the sequence overlap of is smaller than half of that of ;, ii ) the Hamming distance of sequence overlap of is larger than that of ., A random overlap is more likely to be much smaller and have more mismatches than a true overlap ., Therefore , these two criteria will help us remove most erroneous edges ., Once a family-specific graph is constructed and optimized , the goal is to conduct a graph traversal to output paths corresponding to genes ., The traversal starts with nodes without incoming edges ., The challenge arises when two or more genes contain a common or similar subsequence , leading to chimeric nodes such as and in Fig . 3 ., Chimeric nodes add to the complexity of the graph traversal by leading to chimeric contigs ., For example , the path contains nodes exclusively from both genes and thus is a chimeric path ., SAT-Assembler employs three types of information to guide the graph traversal to recover correct gene paths: transitive edges , paired-end reads , and coverage ., We describe the key steps of our graph traversal algorithm using Fig . 3 ., The goal is to output two correct paths corresponding to the two genes ., A paired-end read represents two reads appearing in the same genome with known order ( by our homology search ) and distance range ( insert size ) ., Although transitive edges are removed at the stage of graph pruning they can act as a paired-end read with a small insert size ., Therefore , both transitive edges and paired-end reads can be used to examine whether two nodes are from the same gene ., Two nodes that are not connected by an edge are said to have supports or be supported if there are transitive edges or paired-end reads between them ., For paired-end read supports , we further require that their distance in the path be consistent with the insert size ., Different from previous traversal algorithms , we divide supports into two types , critical supports and non-critical supports ., Critical supports can be used to resolve branching in graph traversal while non-critical supports are not able to distinguish different gene paths ., For example , a graph traversal generates a path ., The node has two outgoing edges and ., If there is a support between and , such as the transitive edge in Fig . 3 , the traversal will be guided to visit in next step ., This transitive edge provides a critical support for correct traversal ., However , the support between and is not critical for guiding the graph traversal because any path that has visited needs to visit ., In Fig . 3 , the support between and is a non-critical support while all the other supports are critical supports ., When there is no support between two non-chimeric nodes , node coverage will be used to resolve the branches ., The coverage of a node is the total size of reads normalized by the length of the assembled sequence of the node ., For protein-coding genes , although the sequence coverage is usually not uniform along the genes its change is gradual rather than sharp ., Thus , the coverage of two consecutive non-chimeric nodes in the same path should reflect this observation ., Any sharp change indicates a wrong path ., For example , in Fig . 3 , and have similar coverage , as do and ., and , however , have significantly different coverage ., Therefore , a path that has visited and should next visit instead of ., We use a bounded depth-first search ( DFS ) algorithm to generate correct paths ., While a typical DFS takes exponential time to generate all simple paths between two nodes , our graph traversal method makes use of critical supports to bound the search and only visits the correct paths , effectively reducing the time complexity of path generation ., During search , we will proceed to those successors of the current node that provide critical supports ., If none of the successive non-chimeric nodes has supports with any of the previously visited non-chimeric nodes , we will proceed to the one that has a similar coverage to the recently visited non-chimeric node given that their coverage is similar enough ., Otherwise , it is highly likely that the current node is not from the same gene as any of its successors ., Therefore , we will output the current path and start a new path from its successors ., The traversal stops when there is no appropriate successive node available ., All paths with critical supports above a given threshold will be output ., Assembly tools may output multiple contigs from the same gene ., There are two main reasons for the fragmentation:, i ) some regions between contigs are not sequenced due to sequencing bias , PCR bias , low transcription level or abundance;, ii ) reads from lowly conserved regions of the gene may not pass the homology search and thus are not used to construct the graph ., The contigs are oriented and connected using their alignment positions against the underlying profile HMM and paired-end reads ., The scaffolding results in super contigs ., SAT-Assembler can distinguish not only different target genes but also gene isoforms caused by AS events ., Here we classify the seven different types of alternative splicing events 51 into four different groups ., Fig . 4 shows how overlap graphs are constructed for these four groups of AS events ., Each of the group is represented by one AS event in 51 ., All the other types fall into one of these groups ., In Cases ( A ) , ( C ) , and ( D ) of Fig . 4 , different isoforms correspond to different paths of the overlap graphs ., Therefore , isoforms generated by AS events from these groups can be distinguished by generating contigs from these paths ., In Case ( B ) , there are two paths that begin with a root node and end with a sink node: and ., The first path recovers the first isoform and the second path corresponds to a chimeric contig ., However , reads in and are not from the same isoform ., Therefore , there will be no paired end support between them ., Our graph traversal algorithm will stop in node without proceeding to node ., Therefore , SAT-Assembler can still correctly recover both isoforms in this case ., In practice , different types of alternative splicing events can occur together , further compounding the assembly ., In these cases , SAT-Assembler relies on multiple types of information such as paired end reads , transitive edges , and coverage to distinguish different isoforms , as described in the section of Guided Graph Traversal Using Multiple Types of Information ., The performance of assembly tools on distinguishing gene isoforms can be found in the section of Performance of Recovering Gene Isoforms ., Let the number of input reads be and the average read length be ., The time complexity of the homology search stage is for one input family , where is length of the profile HMM and ., Suppose reads have passed the homology search stage ., Usually , ., The time complexity of graph construction is , where is the average number of overlapping alignments longer than a given cutoff ., During graph construction , we use alignment positions to guide the overlap computation , avoiding the all-against-all comparison needed in a standard overlap graph construction ., The time complexity of graph traveral is , where is the number of nodes , is the number of edges , is the number of read pairs that have critical supports , and is the number of correct paths in the graph ., The time complexity of the scaffoding stage is ., Because of various optimization techniques and heuristics , the latest version of HMMER is as fast as BLAST 44 ., Considering , the time complexity of scaffolding is much smaller than graph traversal ., Therefore , the overall running time is determined by the graph traversal stage ., In this experiment , we applied SAT-Assembler to an RNA-Seq data set sequenced from a normalized cDNA library of Arabidopsis generated using paired-end Illumina sequencing 52 ., There were a total of 9 , 559 , 784 paired-end reads of 76 bp ., Pfam was used as our database of input families ., Some Pfam families use sequences from Arabidopsis to train their profile HMMs ., Therefore , we eliminated Arabidopsis sequences from these families and recomputed the profile HMMs for them ., We compared the performance of SAT-Assembler with Velvet , Oases , Trinity , IDBA-Tran , and Trans-ABySS ., Velvet is a widely used short read de novo assembly tool ., Oases , Trinity , IDBA-Tran , and Trans-ABySS are assembly tools specially designed for transcriptomic data ., To determine which genes are transcribed in this data set , we conducted read mapping ( using Bowtie 37 ) on all the coding sequences ( CDS ) of Arabidopsis thaliana of version TAIR10 53 ., 59 . 62% of the input reads were mapped to the CDS with at most 2 mismaches allowed ., There is no commonly accepted criterion to define transcribed genes ., In this work , we defined CDS with at least 10 mapped reads as transcribed CDS ., Assembly results of different tools were compared on these transcribed CDS ., There are 29 , 030 different transcribed gene isoforms corresponding to 21 , 452 genes ., A total of 3 , 163 protein or domain families from Pfam that can be aligned to these CDS using HMMER with gathering thresholds ( GAs ) were used as input to SAT-Assembler ., Among the mapped reads , 65 . 39% generated HMMER hits against these protein or domain families using HMMERs default E-value threshold 10 ., The rest of the mapped reads failed to be aligned by HMMER due to the following main reasons:, i ) some Arabidopsis genes are not covered by Pfam families;, ii ) the average sequence identities of some Pfam families that Arabidopisis genes belong to are low , rendering marginal alignment scores , especially for short reads 47;, iii ) some Arabidopsis genes are too remotely related to the Pfam families ., In this experiment , we conducted targeted gene assembly using a metagenomic data set sequenced from highly diverse bacterial and archaeal synthetic communities with 16 archaea members and 48 bacteria members 57 ., We downloaded all reference genomes from NCBI ftp site ( ftp . ncbi . nih . gov/genomes/ ) ., The metagenomic data set was downloaded from NCBI Sequence Read Archive ( SRA ) ( http://www . ncbi . nlm . nih . gov/sra ) using Accession No ., SRA059004 ., After we trimmed low-quality reads , there were 51 , 933 , 622 paired-end reads with an average read length of 100 bp ., We were interested in assembling the family of butyrate kinase pathway genes , which play important roles in butyrate synthesis ., We downloaded the profile HMM of the family from RDPs functional gene repository 58 ., It was built from 77 seed butyrate kinase pathway genes ., The seed genes are not in the genomes ., We annotated the butyrate kinase gene regions in the genomes by aligning reference genomes against the gene family using HMMER with gathering thresholds ( GAs ) ., We compared the performance of Velvet , IDBA-UD , MetaVelvet , and SAT-Assembler on assembling contigs from these regions ., IDBA-UD and MetaVelvet are both de Bruijn graph based and specially designed for de novo assembly of metagenomic data ., We used VelvetOptimiser to search for the best assembly result by trying k-mer sizes from 53 to 83 bp with “” as the optimization function ., VelvetOptimiser reported 55 as the optimal k-mer size ., For IDBA-UD , which accepts multiple k-mer values , we used the same range of k-mer sizes as Velvet in a single run ., Meta-Velvet used the hash table generated by Velvet and its k-mer size was thus 55 as well ., For SAT-Assembler , we used its default parameters ., A total of 15 , 254 reads were classified into the butyrate kinase family by the homology search stage , which accounted for 0 . 15% of the query reads ., Table 4 shows a performance comparison between these assembly tools ., SAT-Assembler had the best gene coverage , chimera rate , and memory usage ., Its contigs were usually shorter than IDBA-UD ., A closer examination reveals the reason: a number of reads did not pass the profile HMM homology search and thus were not used as input to assembly ., Gene coverage of assembly tools in this experiment was much lower than in the first experiment because of lower sequencing depth and higher data complexity ., MetaVelvet had the best running time performance because it directly used the optimal k-mer size and hash table from VelvetOptimiser while Velvet and IDBA-UD both ran a range of k-mer sizes ., The low memory usage of SAT-Assembler further showed the advantage of using homology search in targeted gene assembly for large-scale NGS data ., Due to the high complexity of the metagenomic data set , SAT-Assembler constructed a much more complex overlap graph compared with the overlap graphs in the first experiment , leading to higher runtime overhead ., Table 4 shows that none of the tested assembly tools is the best in all metrics ., If users prefer high gene coverage and high accuracy , especially on hardware with limited resources , we recommend SAT-Assembler ., If long contigs and high contig coverage are more important , IDBA-UD is the best choice ., In this experiment , we compared the performance of SAT-Assembler with Velvet , IDBA-UD , and MetaVelvet on a human gut metagenomic data set ., There were 47 , 117 , 906 paired-end and 5 , 528 , 102 unpaired reads of various lengths ., The average length of the query reads was 95 . 72 bp and 75% of them were 100 bp ., We were interested in assembling butyrate kinase pathway genes as in the second experiment ., The profile HMM of the gene family was built from 77 seed genes from RDPs functional gene repository 58 ., We also downloaded a set of 2 , 352 annotated genes of butyrate kinase family and eliminated the seed genes from them ., By using read mapping , a total of 58 genes with at least 10 mapped reads were identified and were used to evaluate the performance of all assembly tools ., We used VelvetOptimiser to search for the best assembly result by trying k-mer sizes from 51 to 81 bp with “” as the optimization function ., VelvetOptimiser reported 51 as the optimal k-mer size ., For IDBA-UD , which accepts multiple k-mer values , we used the same range of k-mer sizes as Velvet in a single run ., Meta-Velvet used the hash table generated by Velvet and its k-mer size was thus 51 as well ., For SAT-Assembler , we used its default parameters ., A total of 16 , 136 reads were classified into the butyrate kinase family by the homology search stage , which accounted for 0 . 31% of the query reads ., Table 5 shows a performance comparison between these assembly tools ., In this experiment , the result of performance comparison was similar to the second experiment ., SAT-Assembler still had the best gene coverage , chimera rate , and memory usage ., IDBA-UD had the best contig length and contig coverage ., Compared with the second experiment , the chimera rates of all assembly tools increased ., Without knowing all reference genomes , the computed chimera rates might be an over-estimation for all tools because the assembled contigs may contain novel members of the family ., The experiments on RNA-Seq and metagenomic data sets show that our novel consistency-based edge creation strategy and guided graph traversal can effectively avoid chimeric contigs ., Moreover , by reducing the original search space into a much smaller subset of reads from targeted genes , the memory usage was significantly decreased , making it a more economical tool for the assembly of targeted genes from a single or multiple pathways ., We have also tried to use Velvet and Trinity on the reads that passed the homology search stage on the Arabidopsis RNA-Seq data set ., The gene coverage and chimera rate of Velvet were 64 . 38% and 17 . 25% respectively ., The gene coverage and chimera rate of Trinity were 78 . 72% and 22 . 90% respectively ., Compared with the performance when using Velvet and Trinity directly on the input data set , their gene coverages were decreased ., One reason is that the homology search stage does not have 100% sensitivity ., The missed reads may lead to poorer performance of Velvet and Trinity ., SAT-Assembler also provides an easier way for users to set the parameters ., Our edge creation strategy is based on both the overlap threshold ( ) and the consistency between alignment overlap and sequence overlap ( ) ., The consistency strategy poses a strong constraint on the overlap between two reads ., The alignment overlap threshold is mainly used to avoid random overlaps , which are generally very small ., The default overlap threshold 20 is chosen based on the length of the reads in our experimental data sets ., This value is smaller than the k-mer value chosen by VelvetOptimiser for other assembly tools ., This helps us generate better connectivity between reads from the same genes ., At the same time , the consistency constraint guarantees the accuracy of the edge creation ( Table 1 ) ., We have also tried different values from 15 to 30 and found that the edge creation performance is not sensitive to the choice of unless a very large value of is used ., The values of and control the trade-off between sensitivity and of edge creation ., Users can adjust them based on their specific need ., Based on our observation , an overlap threshold that is 20% of the read length is recommended ., The value of is independent of the read length and we suggest that users use the default value ., There are still some challenges to address to further improve SAT-Assemblers performance ., First , gene segments from some poorly conserved gene regions are fragmented because some reads from these regions fail to pass the homology search ., We have aligned all the reads in the human gut metagenomic data set against protein/domain families in Pfam using HMMER and 38 . 65% of them have HMMER hits ., There are three main reasons for the low coverage of Pfam domains in the metagenomic data set:, i ) Pfam is a collection of protein/domain families ., Therefore , reads sequenced from intergenic regions will not have hits ., In addition , even reads sequenced from protein-coding regions may not be part of any domain ., They will not have hits either ., ii ) Some genes of the microbial species are very remotely homologous to the families in Pfam ., iii ) Some reads in the metagenomic data set are very short , resulting in low sensitivity of HMMER 47 ., This problem can be alleviated by increasing the sensitivity of the homology search ., In the future , we will incorporate our proposed position-specific score threshold ( PSST ) 47 , 56 into SAT-Assembler to classify more reads into their native families ., Second , although the edge creation strategy of SAT-Assembler captured more overlaps between reads from the same genes , some positive overlaps still failed to be captured ., When the conservation between the input family and target genes is not good , the alignment overlap and sequence overlap may not always be consistent ., Therefore , reads from poorly conserved regions of the genes may lose consistency between their alignment overlaps and sequence overla | Introduction, Methods, Results, Discussion | Gene assembly , which recovers gene segments from short reads , is an important step in functional analysis of next-generation sequencing data ., Lacking quality reference genomes , de novo assembly is commonly used for RNA-Seq data of non-model organisms and metagenomic data ., However , heterogeneous sequence coverage caused by heterogeneous expression or species abundance , similarity between isoforms or homologous genes , and large data size all pose challenges to de novo assembly ., As a result , existing assembly tools tend to output fragmented contigs or chimeric contigs , or have high memory footprint ., In this work , we introduce a targeted gene assembly program SAT-Assembler , which aims to recover gene families of particular interest to biologists ., It addresses the above challenges by conducting family-specific homology search , homology-guided overlap graph construction , and careful graph traversal ., It can be applied to both RNA-Seq and metagenomic data ., Our experimental results on an Arabidopsis RNA-Seq data set and two metagenomic data sets show that SAT-Assembler has smaller memory usage , comparable or better gene coverage , and lower chimera rate for assembling a set of genes from one or multiple pathways compared with other assembly tools ., Moreover , the family-specific design and rapid homology search allow SAT-Assembler to be naturally compatible with parallel computing platforms ., The source code of SAT-Assembler is available at https://sourceforge . net/projects/sat-assembler/ ., The data sets and experimental settings can be found in supplementary material . | Next-generation sequencing ( NGS ) provides an efficient and affordable way to sequence the genomes or transcriptomes of a large amount of organisms ., With fast accumulation of the sequencing data from various NGS projects , the bottleneck is to efficiently mine useful knowledge from the data ., As NGS platforms usually generate short and fragmented sequences ( reads ) , one key step to annotate NGS data is to assemble short reads into longer contigs , which are then used to recover functional elements such as protein-coding genes ., Short read assembly remains one of the most difficult computational problems in genomics ., In particular , the performance of existing assembly tools is not satisfactory on complicated NGS data sets ., They cannot reliably separate genes of high similarity , recover under-represented genes , and incur high computational time and memory usage ., Hence , we propose a targeted gene assembly tool , SAT-Assembler , to assemble genes of interest directly from NGS data with low memory usage and high accuracy ., Our experimental results on a transcriptomic data set and two microbial community data sets showed that SAT-Assembler used less memory and recovered more target genes with better accuracy than existing tools . | biology and life sciences, computational biology | null |
journal.ppat.1002708 | 2,012 | Induction of GADD34 Is Necessary for dsRNA-Dependent Interferon-β Production and Participates in the Control of Chikungunya Virus Infection | During their replication in host cells , RNA and DNA viruses generate RNA intermediates , which elicit antiviral responses mostly through type-I interferon ( IFN ) production 1 , 2 ., Several families of proteins are known to sense double-stranded RNA ( dsRNA ) , including endocytic Toll-like receptor 3 ( TLR3 ) 3 , the dsRNA-dependent protein kinase ( PKR ) 4 and the interferon-inducible 2′-5′-oligoadenylates and endoribonuclease L system ( OAS/2-5A/RNase L ) 5 ., Viral dsRNA and the synthetic dsRNA analog polyriboinosinic:polyribocytidylic acid ( poly I:C ) are also detected by different cytosolic DExD/H box RNA helicases such as the melanoma differentiation-associated gene 5 ( MDA5 ) , DDX1 , DDX21 , and DHX36 , which , once activated , trigger indirectly the phosphorylation and the nuclear translocation of transcription factors such as IRF-3 and IRF-7 , resulting predominantly in abundant type-I IFN and pro-inflammatory cytokines production by the infected cells 1 , 6 , 7 ., Alphaviruses such as Chikungunya virus ( CHIKV ) are small enveloped viruses with a message-sense RNA genome , which are known to be strong inducers of type-I IFN in vivo 8 , 9 , a key response for the host to control the infection 10 , 11 , 12 ., In vitro , however , response to RNA viruses is heterogeneous , since Sindbis virus ( SINV ) , do not elicit detectable IFN-α/β production in infected of murine embryonic fibroblasts ( MEFs ) 13 ., The specific points of blockage of type-I IFN production during infection are still not well delineated , but SINV and other alphaviruses could antagonize IFN production by shut-off of host macromolecular synthesis in infected cells 14 , 15 , 16 ., Recently , human fibroblasts infection by CHIKV was shown to trigger abundant IFN-α/β mRNA transcription , while preventing mRNA translation and secretion of these antiviral cytokines 13 , 15 ., Contrasting with these reports , other groups using different CHIKV strains have observed abundant type-I IFNs release in the culture supernatants of CHIKV-infected human monocytes 17 , human lung cells ( MRC-5 ) , human foreskin fibroblasts and MEFs 10 ., Type-I IFN stimulation of non-hematopoietic cells has also been shown to be essential to clear infection upon CHIKV inoculation in mouse , but CHIKV was found to be a poor inducer of IFN secretion by human plasmacytoïd dendritic cells 10 ., Thus , great disparities regarding alphavirus-triggered IFN responses exist between viral strains and the nature of host cells or animal models ., Once bound to their receptor on the cell surface ( IFNAR ) , type-I IFNs activate the Janus tyrosine kinase pathway , which induces the expression of a wide spectrum of cellular genes including Pkr 18 ., These different genes participate in the cellular defense against the viral infection ., PKR is a serine–threonine kinase that binds dsRNA in its N-terminal regulatory region and induces phosphorylation of translation initiation factor 2-alpha ( eIF2α ) on serine 51 19 , 20 , leading to protein synthesis shut-off and apoptosis ., PKR has been also been shown to participate in several important signaling cascades , including the p38 and JNK pathways 21 , as well as type-I IFN production 22 , 23 ., Inhibition of translation , IFN responses and triggering of apoptosis combine to make PKR a powerful antiviral molecule , and many viruses have evolved strategies to antagonize it 24 , 25 ., Interestingly , several positive RNA-strand viruses ( eg . Togaviridae or Picornaviridae ) have been shown to activate PKR , resulting in phosphorylation of eIF2α and host translation arrest 26 , while viral mRNA could initiate translation in an eIF2-independent manner by means of a dedicated RNA structure , that stalls the scanning 40S ribosome on the initiation codon 25 ., Despite the existence of these viral PKR-evading strategies , the importance of PKR for type-I IFN production has been strongly debated over the years and even considered dispensable since the discovery of the innate immunity function of the DExD/H box RNA helicases 27 , 28 ., However , several PKR-deficient cell types have reduced type-I IFN production in response to poly I:C 23 , 29 , 30 , while PKR was demonstrated to be required for IFN-α/β production in response to a subset of RNA viruses including Theilers murine encephalomyelitis , West Nile ( WNV ) and Semliki Forest virus ( SFV ) , but not influenza , Newcastle disease , nor Sendai virus 31 , 32 , 33 , 34 ., These studies raise therefore the possibility that some but not all viruses induce IFN-α/β in a PKR-dependent and cell specific manner ., Infection of PKR or RNAse L deficient mice demonstrated that these enzymes were not absolutely necessary for type I IFN-mediated protection from alphaviruses such as SFV or WNV , but still contributed to levels of serum IFN and clearance of infectious virus from the central nervous system 25 , 35 ., Similarly , deficient mice for both PKR and RNAse L showed no increase in morbidity following SINV infection , although , like during WNV infection , increased viral loads in draining lymph nodes were observed 35 , 36 ., These observations support a non-redundant and cell specific role for these enzymes in the amplification of type-I IFN responses to viral infection more than in their initiation 31 , 32 , 35 ., Nevertheless , the exacerbated phenotypes observed upon alphavirus infection of mice deficient for type-I IFN receptor ( IFNAR ) , underlines the limits of the individual contributions of PKR and RNAse L to the anti-viral resistance of adult animals 10 , 35 , 36 ., In addition to dsRNA detection , different stress signals trigger eIF2α phosphorylation , thus attenuating mRNA translation and activating gene expression programs known globally as the integrated stress response ( ISR ) 37 ., To date , four kinases have been identified to mediate eIF2α phosphorylation: PKR , PERK ( protein kinase RNA ( PKR ) -like ER kinase ) 38 , GCN2 ( general control non-derepressible-2 ) 39 , 40 and HRI ( heme-regulated inhibitor ) 41 , 42 ., ER stress–mediated eIF2α phosphorylation is carried out by PERK , which is activated by an excess of unfolded proteins accumulating in the ER lumen 38 ., Activated PERK phosphorylates eIF2α , attenuating protein synthesis and triggering the translation of specific molecules such as the transcription factor ATF4 , which is necessary to mount part of a particular ISR , known as the unfolded protein response ( UPR ) 43 , 44 ., Interestingly DNA viruses , such as HSV , that use the ER as a part of its replication cycle , have been reported to interfere with the ER stress response through different mechanisms , such as the dephosphorylation of eIF2α by the viral phosphatase 1 activator , ICP34 . 5 45 , 46 ., We show here , using SUnSET , a non-radioactive method to monitor protein synthesis 47 , that independently of any active viral replication , cytosolic poly I:C detection in mouse embryonic fibroblasts ( MEFs ) promotes a PKR-dependent mRNA translation arrest and an ISR-like response ., During the course of this response , ATF4 and its downstream target , the phosphatase-1 ( PP1 ) cofactor , growth arrest and DNA damage-inducible protein 34 ( GADD34 , also known as MyD116 and Ppp1r15a ) 48 , are strongly up-regulated ., Importantly , although the translation of most mRNAs is strongly inhibited by poly I:C , that of IFN-ß and Interleukin-6 ( IL-6 ) is considerably increased under these conditions ., We further demonstrate that PKR-dependent expression of GADD34 is critically required for the normal translation of IFN-ß and IL-6 mRNAs ., We prove the relevance of these observations for antiviral responses using CHIKV as a model: we show that GADD34-deficient MEFs are unable to produce IFN-ß during infection and become permissive to CHIKV ., We further show that CHIKV induces 100% lethality in 12-day-old GADD34-deficient mice , whereas WT controls do not succumb to infection ., Our observations demonstrate that induction of GADD34 is part of the anti-viral response and imply the existence of distinct and segregated groups of mRNA , which require GADD34 for their efficient translation upon dsRNA-induced eIF2α phosphorylation ., We monitored protein synthesis in MEFs and NIH-3T3 cells after poly I:C stimulation , using puromycin labeling followed by immunodetection with the anti-puromycin mAb 12D10 47 ., Poly I:C delivery to MEFs and NIH-3T3 , rapidly and durably inhibited protein synthesis , concomitant with increased eIF2α phosphorylation ( P-eIF2α ) ( Fig . 1A and Fig . S1A ) ., In MEFs , a strong eIF2α phosphorylation was observed after 4 h of poly I:C treatment , followed by a steady dephosphorylation at later times ( Fig . 1A ) ., Protein synthesis arrest was confirmed in individual cells by concomitant imaging of poly I:C delivery , mRNA translation and P-eIF2α ( Fig . 1B and Fig . S1B ) , and with a wide range of dsRNA concentrations ( Fig . S1C ) ., Poly I:C-induced eIF2α phosphorylation and subsequent translation arrest were not observed in PKR-deficient MEFs ( Fig . 1C and 1D ) , while eIF2α phosphorylation induced by the UPR-inducing drug thapsigargin ( th ) ( an inhibitor of SERCA ATPases ) or arsenite ( as ) was unchanged in PKR−/− cells ( Fig . 1C ) ., PKR is therefore necessary to induce protein synthesis inhibition in response to cytosolic poly I:C ., When levels of IFN-ß were quantified in culture supernatants and compared to total protein synthesis intensity , we found that most of the cytokine production occurred after 4 to 8 h of pIC delivery ( Fig . 1E , WT , and S1D ) , a time at which mRNA translation was already considerably decreased ( Fig . 1A and S1E ) ., We measured the amount of cytokine produced in NIH-3T3 cells at a time ( 7 h ) at which translation was already strongly inhibited ( Fig . 1G and 1F ) ., To prove that IFN-β production truly occurred during this poly I:C-induced translation arrest , cells exposed for 7 h to poly I:C were washed and old culture supernatants replaced with fresh media for 1 h ( with or without CHX ) , prior translation monitoring ( Fig . 1F , right ) and IFN-ß dosage ( Fig . 1G , right ) ., We observed that close to 30% of the total IFN-ß produced over 8 h of poly I:C stimulation is achieved during this 1 h period , despite a close to undetectable protein synthesis in the dsRNA-treated cells ( Fig . 1F ) ., The neo synthetic nature of this IFN was further demonstrated by the absence of the cytokine in CHX-treated cell supernatants ., IFN-β production in response to poly I:C is therefore likely to be specifically regulated and occurs to a large extent independently of the globally repressed translational context ., As previously observed in MEFs , IFN-ß production in response to poly I:C was independent of PKR ( Fig . 1E ) 31 ., This suggests that although its production occurs during cap-mediated translation inhibition , it does not directly depend on a specialized open reading frame organization , as described for the translation of the mRNAs coding for the UPR transcription factor ATF4 or the SV 26S mRNA upon eIF2α phosphorylation 26 , 49 ., This hypothesis is also supported by the ability of MEFs expressing the non-phosphorylatable eIF2α Ser51 to Ala mutant ( eIF2α A/A ) , to produce normal levels of IFN-ß in response to poly I:C ( Fig . 1E ) , while global translation was not inhibited by poly I:C in these cells ( Fig . S2 ) ., We went on to investigate the molecular mechanisms promoting this paradoxical IFN-ß synthesis in an otherwise translationally repressed environment ., Induction of eIF2α phosphorylation by PERK during ER stress promotes rapid ATF4 synthesis and nuclear translocation , followed by the transcription of many downstream target genes important for the UPR 50 ., Similarly , in presence of PKR , nuclear ATF4 levels were found to be up-regulated in MEFs responding to cytosolic poly I:C , albeit less importantly than upon a bona fide UPR induced by thapsigargin ( Fig . 2A ) ., One of the key molecules involved in the control of eIF2α phosphorylation is the protein phosphatase 1 co-factor GADD34 , which relieves translation repression during ER stress by promoting eIF2α dephosphorylation 50 , 51 , 52 ., GADD34 is a direct downstream transcription target of ATF4 53 ., Expression of GADD34 was quantified by qPCR and immunoblot in WT and PKR−/− MEFs ( Fig . 2B ) ., In WT cells GADD34 mRNA expression was clearly up-regulated ( 20 fold ) in response to poly I:C , while GADD34 protein induction was equivalent in poly I:C- and thapsigargin-treated cells ., GADD34 mRNA transcription and translation were not observed in PKR−/− cells responding to poly I:C , but occurred normally upon thapsigargin treatment , paragoning eIF2α phosphorylation ( Fig . 2B , right ) ., We next investigated the importance of ATF4 for GADD34 transcription by monitoring the levels of GADD34 mRNA in ATF4-deficient cells ., ATF4−/− MEFs displayed higher basal levels of GADD34 mRNA than WT cells ., However , in absence of ATF4 , MEFs were unable to efficiently induce GADD34 mRNA transcription in response to any of the stimuli tested ( Fig . S3 ) ., GADD34 mRNA expression was induced only 2 fold in ATF4−/− MEFs exposed to poly I:C , suggesting that its transcription is mostly dependent on ATF4 in this context ., We further investigated P-eIF2α requirement for GADD34 expression and found that eIF2α A/A expressing MEFs were incapable of up-regulating GADD34 in response to poly I:C ( Fig . 2C ) ., Phosphorylation of eIF2α by PKR in response to cytosolic poly I:C induces therefore a specific integrated stress response ( ISR ) , that allows ATF4 translation , its nuclear translocation and subsequent GADD34 mRNA transcription ., We next evaluated the relevance of GADD34 induction , by treating WT and GADD34ΔC/ΔC fibroblasts with poly I:C or with drugs known to induce ER stress , such as thapsigargin and the N-glycosylation inhibitor tunicamycin 52 ., As expected , in WT cells eIF2α phosphorylation was rapidly increased in response to all ISR-inducing stimuli and decreased concomitantly with the expression of GADD34 over time ( Fig . 3A and S4 ) 52 ., Consequently eIF2α phosphorylation was greatly increased in GADD34ΔC/ΔC MEFs in all the conditions tested ( Fig . 3A and S4A ) ., In thapsigargin-treated cells , protein synthesis was reduced in the first hour of treatment and rapidly recovered ( Fig . 3B ) 54 ., Poly I:C , however , nearly completely inhibited translation despite active eIF2α dephosphorylation ., This was particularly obvious when poly I:C was co-administrated together with thapsigargin ., Indeed , poly I:C dominated the response by preventing the translation recovery normally observed after few hours of drug treatment ( Fig . 3B ) ., Surprisingly , in absence of functional GADD34 , although eIF2α phosphorylation induction by poly I:C was augmented dramatically , no further decrease in protein synthesis was observed upon treatment of GADD34ΔC/ΔC cells with the dsRNA mimic ( Fig . 3A and 3C ) ., The functionality of GADD34 in translation restoration was , however , fully demonstrated , when the same cells were treated with thapsigargin , and protein synthesis was completely inhibited by this treatment 52 ( Fig . 3C ) ., Thus , cytosolic dsRNA delivery induces a type of protein synthesis inhibition , which requires eIF2α phosphorylation for its initiation , but conversely cannot be reverted by GADD34 induction and subsequent GADD34-dependent eIF2α dephosphorylation ., The potential contribution of the OAS/2-5A/RNAse L system to this P-eIF2α-independent inhibitory process was evaluated by investigating RNA integrity in MEFs exposed to poly I:C ., We used capillary electrophoresis to establish precise RNA integrity numbers ( RIN ) computed from different electrophoretic traces ( pre- , 5S- , fast- , inter- , precursor- , post-region , 18S , 28S , marker ) and quantify the degradation level of mRNA and rRNA potentially resulting from the activation of this well characterized anti-viral pathway ., No major RNA degradation could be observed upon poly I:C delivery ( Fig . S5 ) , suggesting that global RNA degradation does not contribute extensively to the long term translation inhibition observed upon poly I:C delivery in our experimental system ., We have observed that GADD34 expression counterbalances PKR activation by promoting eIF2α dephosphorylation , however it has little impact on reversing the global translation inhibition initiated by poly I:C ., We next monitored the production of specific proteins and cytokines in WT and GADD34ΔC/ΔC MEFs ( Fig . 4 ) ., Cystatin C , a cysteine protease inhibitor was chosen as a model protein , since its secretion ensures a relative short intracellular residency time so that its intracellular levels directly reflect its synthesis rate 55 ., This is confirmed by the N-glycosylated- and total Cystatin C accumulation in cells treated with brefeldin A ( Fig . 4A , left panel ) ., Cystatin C levels were found to follow a similar trend to that observed with total translation , being strongly reduced upon poly I:C exposure and not profoundly influenced by GADD34 inactivation ( Fig . 4A , right panel ) ., Thapsigargin treatment induced a brief drop in cystatin C levels , prior to some levels of GADD34-dependent recovery ., 6 hours of tunicamycin treatment affected more cystatin C accumulation than anticipated ( Fig . 4A , right panel ) , probably due to interference with the N-glycosylation and associated folding of this di-sulfide bridge containing protein 55 , thereby promoting its degradation by endoplasmic reticulum-associated protein degradation ( ERAD ) 56 ., We next turned towards PKR , which displayed a pattern of expression completely different from cystatin C ( Fig . 4B ) ., As expected from its IFN-inducible transcription , levels of PKR were increased in poly I:C-treated MEFs ( Fig . 4B ) , despite the strong global translation inhibition observed in these cells ( Fig . 3 ) ., GADD34 inactivation appeared to influence the accumulation of PKR , since the cytoplasmic dsRNA sensor levels were not up-regulated and even decreased in poly I:C-treated GADD34ΔC/ΔC MEFs ( Fig . 4B ) ., Control treatment with tunicamycin and thapsigargin did not alter significantly PKR levels ( Fig . 4B ) , suggesting that ER stress did not influence the kinase expression ., The absence of PKR up-regulation in the poly I:C-treated GADD34ΔC/ΔC MEFs led us to investigate the capacity of these cells to produce anti-viral and inflammatory cytokines , which normally drive PKR expression through an autocrine loop ., We ruled out any interference from the UPR in triggering IFN-ß production in our experimental system , since , as anticipated from PKR expression , tunicamycin and thapsigargin treatments were not sufficient to promote cytokine production in MEFs ( Fig . S6 ) 43 , 44 ., We therefore investigated IFN-ß and IL-6 production in response to dsRNA in WT , GADD34ΔC/ΔC and CReP−/− MEFs ., CReP−/− MEFs were used as a control , since CReP ( Ppp1r15b ) is a non-inducible co-factor of PP1 and displays some functional redundancy with GADD34 57 ., Although basal levels of eIF2α phosphorylation were higher in CReP−/− , PKR expression and translation inhibition upon poly I:C delivery were equivalent in WT and CReP−/− MEFs ( Fig . S7A and S7B ) ., Quantification of IFN-ß and IL-6 levels in culture supernatants indicated that , although abundant and comparable amounts of these cytokines were secreted by WT and CReP−/− cells , they were both absent in poly I:C-treated GADD34ΔC/ΔC MEFS ( Fig . 4C and S7C ) ., Quantitative PCR analysis revealed that , IFN-ß , IL-6 and PKR transcripts were potently induced in poly I:C treated GADD34ΔC/ΔC MEFs ( Fig . 4D ) , thus excluding any major transcriptional alterations in these cells , as confirmed by the normal levels of cystatin C mRNA , which remained constant in all conditions studied ., Moreover , using confocal immunofluorescence microscopy , we could not detect intracellular IFN-β in poly I:C-stimulated GADD34ΔC/ΔC MEFs , in contrast to WT cells , which abundantly expressed the cytokine , despite the global translation arrest ( Fig . S8 ) ., Thus , we could attribute the deficit in cytokine secretion of the GADD34ΔC/ΔC MEFs to a profound inability of these cells to synthesize cytokines , rather than to a defect in transcription or general protein secretion ., GADD34 induction by poly I:C is therefore absolutely necessary to maintain the synthesis of specific cytokines and probably several other proteins in an otherwise translationally repressed context ., Importantly , GADD34 exerts its rescuing activity only on a selected group of mRNAs including those coding for IFN-ß and IL-6 , but not on all ER-translocated proteins , since cystatin C synthesis was strongly inhibited by poly I:C in all conditions tested ., Interestingly , in GADD34ΔC/ΔC MEFs , PKR mRNA strongly accumulated in response to poly I:C ( Fig . 4D ) , despite the absence of detectable IFN-ß production and PKR protein increase ( Fig . 4B ) ., This continuous accumulation of PKR mRNA in response to poly I:C suggests the existence of alternative molecular mechanisms , capable of promoting PKR mRNA transcription and stabilization independently of autocrine IFN-β detection ., Nevertheless in these conditions PKR expression , like IFN-β , was found to be dependent on the presence of GADD34 for its synthesis ( Fig . 4B ) ., Recent results indicate that PKR participates to the production of IFN-α/ß proteins in response to a subset of RNA viruses including encephalomyocarditis , Theilers murine encephalomyelitis , and Semliki Forest virus 31 ., Even though IFN-α/ß mRNA induction is normal in PKR-deficient cells , a high proportion of mRNA transcripts lack their poly ( A ) tail 31 ., As GADD34 induction by poly I:C was completely PKR-dependent , we wondered whether the phenotypes observed in PKR−/−cells and GADD34ΔC/ΔC MEFs could be related ., Oligo-dT purified mRNA extracted from cells exposed to poly I:C were therefore analyzed by qPCR ., PolyA+ mRNAs coding for IFN-ß and IL-6 were equivalently purified and amplified from WT and GADD34ΔC/ΔC MEFs ( Fig . S9 ) ., This confirms that albeit the phenotypes of PKR−/− and GADD34ΔC/ΔC cells might be linked , mRNA instability is not the primary cause of the cytokine production defect observed in GADD34ΔC/ΔC ., Taken together these observations suggest the existence of a specific mRNAs pool , encompassing cardinal immune effectors such as IFN-ß , IL-6 , and PKR , which are specifically translated in response to dsRNA sensing and increased levels of P-eIF2α ., This mRNAs pool requires GADD34 for their translation during the global protein synthesis shut-down triggered by dsRNA detection ., We verified that GADD34 inactivation , and no other deficiency , was truly responsible for the loss of cytokine production by complementing GADD34ΔC/ΔC MEFs with GADD34 cDNA prior poly I:C delivery ., IFN-ß secretion was partially restored in transfected GADD34ΔC/ΔC cells while eIF2α was efficiently dephosphorylated in both WT and GADD34ΔC/ΔC transfected MEFs ( Fig . 4E ) ., To further demonstrate that the phosphatase activity of GADD34 controls cytokine production upon dsRNA detection , we treated WT MEFs with guanabenz , a small molecule , which selectively impairs GADD34-dependent eIF2α dephosphorylation 58 ., Upon treatment with this compound , a dose dependent inhibition of IFN-ß secretion was observed in poly I:C-treated MEFs , confirming the importance of GADD34 in this process ( Fig . S10 ) ., Fibroblasts of both human and mouse origin constitute a major target cell of Chikungunya virus ( CHIKV ) during the acute phase of infection 59 ., In adult mice with a totally abrogated type-I IFN signaling , CHIKV-associated disease is particularly severe and correlates with higher viral loads ., Importantly , mice with one copy of the IFN-α/ß receptor ( IFNAR ) gene develop a mild disease , strengthening the implication of type-I IFN signaling in the control of CHIKV replication 59 ., Recently , human fibroblasts infection by CHIKV was shown to induce IFN-α/ß mRNA transcription , while preventing mRNA translation and secretion of these antiviral cytokines ., CHIKV was found to trigger eIF2α phosphorylation through PKR activation , however this response is not required for the block of host protein synthesis 15 ., We tested the importance of PKR during CHIKV infection by infecting WT and PKR−/− MEFs with CHIKV-GFP , at a multiplicity of infection ( MOI ) of 10 and 50 ., Productive infection was estimated by GFP expression ( Fig . 5A , left panel ) , while culture supernatants were monitored for the presence of IFN-β ( 5A , right panel ) ., PKR was found to be necessary to control CHIKV infection in vitro , since at least 60% of PKR–inactivated cells were infected after 24 of viral exposure , compared to only 15% in the control fibroblasts population ., WT MEFs produced efficiently IFN-β , while the hypersensitivity to infection of the PKR−/− MEFs was correlated to a reduced type-I IFN production capacity after infection ., Thus , during CHIKV infection , PKR is required for normal IFN production by MEFs ., We also monitored protein synthesis in infected WT and PKR−/− fibroblasts using puromycin labeling followed by immunofluorescence confocal microscopy ( Fig . 5B ) ., CHIKV-GFP positive PKR−/− MEFs were found to incorporate efficiently puromycin , while in their infected WT counterpart protein synthesis was efficiently inhibited ., Thus CHIKV , in this experimental model , induces a PKR-dependent protein synthesis inhibition and is therefore particularly relevant to further confirm our observations on the role of GADD34 in controlling type-I IFN production during response to viral RNAs ., GADD34ΔC/ΔC MEFs were exposed to CHIKV-GFP ( MOI of 10 or 50 ) for 24 and 48 h ., Productive infection was estimated by GFP expression and virus titration ( Fig . 6A ) , and culture supernatants monitored for the presence of type-I IFN ( Fig . 6B , left ) ., Only minimal CHIKV infection ( 15% ) could be observed at maximum MOI in WT MEFs ( Fig . 6A , left ) , while robust IFN- β amounts were already produced at the lowest MOI ( Fig . 6B ) ., Contrasting with WT cells and regardless of the MOI used , a higher level of viral replication was observed in GADD34ΔC/ΔC MEFs ( Fig . 6A ) ., The GADD34-inactivated cells were clearly more sensitive to CHIKV , displaying a 50% infection rate after 24 h of infection ( MOI 50 ) and a log more of virus titer in culture supernatants ( Fig . 6A , right ) ., Correlated with their susceptibility to CHIKV infection , IFN-β production was nearly undetectable in GADD34ΔC/ΔC MEFs ( Fig . 6B ) ., Such observation confirms the incapacity of GADD34-deficient cells to produce cytokines in response to cytosolic dsRNA , a deficiency likely to facilitate viral replication ., This interpretation is further supported by the abrogation of viral replication in both WT and GADD34ΔC/ΔC MEFs briefly treated with IFN-β ( Fig . 6C ) ., Thus , GADD34 inactivation does not favor viral replication per se , but is critical for type-I IFN production ., Interestingly infection levels were found to be higher in PKR−/− than in GADD34 ΔC/ΔC MEFs , although this difference could be attributed to clonal MEFs variation , it more likely suggests that PKR-dependent translation arrest could be key in preventing early viral replication in this system ., In addition , the relatively lower permissivity of GADD34ΔC/ΔC MEFs to infection at high MOI could indicate the existence of GADD34-dependent defense mechanisms , which could be independent from IFN production and eIF2-α dephosphorylation ., To strengthen and generalize these observations , we treated a different strain of WT MEFs with guanabenz and examined the consequences for CHIKV infection ., Biochemically , GADD34 expression was induced upon CHIKV infection , and guanabenz treatment resulted in a clear increase in eIF2α phosphorylation , demonstrating the importance of GADD34 in limiting this process during infection ( Fig . 6D , right ) ., As observed with GADD34ΔC/ΔC cells , pharmacological and RNAi inhibition of GADD34 was found to increase significantly the sensitivity of MEFs to infection , while reducing their IFN-β production ( Fig . 6D and S10 ) ., Thus , induction of GADD34 and its phosphatase activity during CHIKV infection , in vitro , participates to normal type-I IFN production and control of viral dissemination ., Several components of the innate immune response have been shown to impact on the resistance of adult mice and to restrict efficiently CHIKV infection and its consequences in vivo 10 ., We decided to investigate the importance of GADD34 upon intradermal injections of CHIKV to WT ( FVB ) and GADD34ΔC/ΔC mice ., Neither strain of adult mice was affected by intradermal injections of CHIKV , with little statistically significant differences in the virus titers found in the different organs ., Thus , GADD34 deficiency does not annihilate all the sources of type-I IFN in the infected adult animals , a situation exemplified by the capacity of GADD34ΔC/ΔC bone-marrow derived dendritic cells to produce reduced , but measurable IFN-β in response to poly I:C 60 ., This also infers that the light impact of GADD34 inactivation on mouse development 61 does not render these animals more sensitive to CHIKV infection ., As in Humans , CHIKV pathogenicity is strongly age-dependent in mice , and in less than 12 day-old mouse neonates , CHIKV induces a severe disease accompanied with a high mortality rate 59 ., GADD34 function was therefore evaluated in this more sensitive context by injecting intradermally CHIKV to FVB ( WT ) and GADD34ΔC/ΔC neonatal mice ., As previously observed for C57/BL6 mice 59 , when CHIKV was inoculated to FVB neonates , a rate of 50% of mortality was observed 3 days after the infection of 9-day-old mice , while 12-day-old pups were found essentially resistant to the virus lethal effect ( Fig . 7A ) ., Strongly contrasting with these results , all CHIKV infected GADD34ΔC/ΔC neonates died within 3–5 days post inoculation whatever their age ( Fig . 7A ) ., When infection was monitored 5 days post-inoculation of 12-day-old mice at , GADD34ΔC/ΔC pups displayed considerably more elevated CHIKV titers ( 10–100 folds ) in most organs tested , including liver , muscle , spleen and joints , the later being primarily targeted by the virus ( Fig . 7B , left ) ., As expected , and in full agreement with the in vitro data , infected GADD34ΔC/ΔC tissues showed a considerably reduced IFN-ß production ( 40–50% ) compared to control tissues ( Figure 7B , right ) , while serum levels were reduced by 20% ( not shown ) ., Although Infectious virus was poorly detected in the heart of WT animals , elevated titers of virus were observed in the heart of GADD34-deficient pups , matching the limited production of IFN in this organ ., We further investigated the possible pathological consequences of cardiac tissue infection by carrying-out comparative histopathology ., Hearts of infected GADD34-deficient animals displayed severe cardiomyocytes necrosis with inflammatory infiltrates by monocytes/macrophages and very important calcium deposition ( Fig . 8 ) , all being indicative signs of grave necrotic myocarditis ., As a consequence , the left ventricles were strongly dilated , being probably the cause of acute cardiac failures and of the important death rate observed in GADD34ΔC/ΔC infected pups ., Histology of infected FVB mice hearts was , however , normal with only few inflammatory cells ( mainly lymphocytes ) observed in the close vicinity of capillaries ., GADD34 expression is therefore necessary to allow normal type-I interferon production during viral infection and to promote the survival of young infected animals ., We could circumvent the age-related acquisition of viral resistance in GADD34ΔC/ΔC mice to 17 days , since mice inoculated at that age survived CHIKV inoculation ., In these animals , 3 days post-infection , enhanced viral replication was observed in | Introduction, Results, Discussion, Materials and Methods | Nucleic acid sensing by cells is a key feature of antiviral responses , which generally result in type-I Interferon production and tissue protection ., However , detection of double-stranded RNAs in virus-infected cells promotes two concomitant and apparently conflicting events ., The dsRNA-dependent protein kinase ( PKR ) phosphorylates translation initiation factor 2-alpha ( eIF2α ) and inhibits protein synthesis , whereas cytosolic DExD/H box RNA helicases induce expression of type I-IFN and other cytokines ., We demonstrate that the phosphatase-1 cofactor , growth arrest and DNA damage-inducible protein 34 ( GADD34/Ppp1r15a ) , an important component of the unfolded protein response ( UPR ) , is absolutely required for type I-IFN and IL-6 production by mouse embryonic fibroblasts ( MEFs ) in response to dsRNA ., GADD34 expression in MEFs is dependent on PKR activation , linking cytosolic microbial sensing with the ATF4 branch of the UPR ., The importance of this link for anti-viral immunity is underlined by the extreme susceptibility of GADD34-deficient fibroblasts and neonate mice to Chikungunya virus infection . | Nucleic acids detection by multiple molecular sensors results in type-I interferon production , which protects cells and tissues from viral infections ., At the intracellular level , the detection of double-stranded RNA by one of these sensors , the dsRNA-dependent protein kinase also leads to the profound inhibition of protein synthesis ., We describe here that the inducible phosphatase 1 co-factor Ppp1r15a/GADD34 , a well known player in the endoplasmic reticulum unfolded protein response ( UPR ) , is activated during double-stranded RNA detection and is absolutely necessary to allow cytokine production in cells exposed to poly I:C or Chikungunya virus ., Our data shows that the cellular response to nucleic acids can reveal unanticipated connections between innate immunity and fundamental stress pathways , such as the ATF4 branch of the UPR . | biology | null |
journal.pcbi.1006512 | 2,018 | Chromatin remodelers couple inchworm motion with twist-defect formation to slide nucleosomal DNA | Eukaryotic genomes are compacted into the cell nucleus via the formation of nucleosomes , each of them consisting of ~147 base pairs ( bp ) of DNA wrapping around a protein histone octamer 1 ., After having been initially considered as passive building blocks of chromatin organization , nucleosomes became to be recognized as active regulators of DNA transcription and replication 2 ., An origin of this regulation is the steric effect that inhibits other DNA-binding proteins , such as transcription factors , from accessing nucleosomal DNA 3 , suggesting the requirement of fine control of nucleosome positioning along the genomic sequence 2 , 4 ., For instance , repositioning of nucleosomes enables the dynamic regulation of gene expression during cell differentiation 5 and in response to stresses such as heat shock 6 , 7 ., While in vitro the nucleosome locations are solely determined by DNA mechanics 8 , 9 , precise positioning in vivo is largely controlled by chromatin remodelers 2 , 4 , which are ATP-dependent molecular machines 10 , 11 ., High-resolution structures of some of these remodelers bound to nucleosomes were obtained very recently by cryo-EM 12–15 ., While these static structures provide crucial insights , currently missing are the dynamic aspects of how these molecular machines work , on which the current study focus by molecular dynamics simulations ., Remodelers are molecular motors that consume ATP to perform a wide variety of functions related to genome organization 10 , 11: facilitating nucleosome assembly 16 and precise spacing 17–19 , controlling DNA accessibility via nucleosome sliding 20 or histone ejection 21 , and nucleosome editing via exchange between different histone variants 22 ., These activities enable remodelers to maintain chromatin organization after disruptive events such as replication 2 , 23 , and to regulate gene expression via the dynamic control of nucleosome positions 6 , 7 , 21 , 24 , 25 ., Although the changes in chromatin organization induced by remodelers have been widely documented 17 , 18 , the precise molecular mechanisms are still far from being clear 10 , 11 , 26 ., The complexity comes in part from the existence of a wide variety of remodelers with different structures and functions 11 , which has led to several possible classifications into remodeler sub-families 27 ., Each remodeler consists of many distinct domains , which act in concert to confer specificity to the remodeling activity ( e . g . nucleosome sliding vs ejection ) 10 and to fine-tune it via substrate recognition ( e . g . of histone tail modifications ) 28 , 29 ., Despite this complexity , all remodelers share a conserved translocase domain 10: an ATPase motor capable of unidirectional sliding along DNA via binding and hydrolysis of ATP between its two RecA-like lobes , structurally similar to those found in helicases 12 , 27 ., The translocase domain of most remodelers binds nucleosomes at the superhelical location ( SHL ) 2 12 , 13 , 30 , i . e . two DNA turns away from the dyad symmetry axis ( SHL 0 ) 31 ( Fig 1A ) ., Many remodelers induce sliding of nucleosomal DNA towards the dyad from the translocase binding location 30 , 32–34 ., Sliding may represent a shared fundamental mechanism at the basis of most remodeling activities; the interactions with the additional domains would then confer specificity to the remodeler , allowing for substrate recognition and determining whether the final outcome is nucleosome repositioning , histone ejection or histone exchange 10 , 35 ., For instance , recent experiments suggested that the INO80 remodeler causes histone exchange by sliding nucleosomal DNA from its translocase binding site around SHL 6 14 , 15 , 34 ., Therefore , the detailed characterization of active nucleosome sliding by the translocase domain would represent a significant step forward in our understanding of chromatin remodeling ., There is much experimental evidence suggesting that the translocase domain of remodelers , as well as some helicases , slides unidirectionally along DNA via an inchworm mechanism 10 , 36 , which may also be viewed as a molecular ratchet 13 , 37 , processing by 1 bp every ATP cycle ., A minimal inchworm model requires 3 distinct chemical states , apo , ATP-bound , and ADP-bound , which are coupled to conformational changes of the translocase ( Fig 1B ) ., The two lobes are either distant in the open form or in contact in the closed form ., On top , conformational changes modulate interaction strengths with DNA 37 ., To translocate along DNA in the direction from lobe 1 to lobe 2 ( which would correspond to sliding nucleosomal DNA in the direction from SHL 2 towards the dyad , as in most remodelers 12 ) , ATP binding to the translocase first induces the transition from an open to a closed conformation ( first and second cartoons in Fig 1B ) ., During closure , lobe 2-DNA interactions are stronger than those of lobe 1 , so that lobe 1 will detach from the DNA and move towards lobe 2 by 1 bp , which maintains its position ., Then , ATP hydrolysis is accompanied by weakening of the lobe 2 interactions with DNA relative to those of lobe 1 , so that during the conformational change from the closed to the open state lobe 2 now moves away from lobe 1 by 1 bp ( third cartoon in Fig 1B ) ., The cycle is then completed by the release of ADP and the change of the interaction strengths to their initial apo-state values ., This mechanism was firstly suggested for helicases from their crystal structures , which show ATPase closure and changes in lobes-DNA interactions as a function of the chemical state 36 , 37 ., For remodelers , the same mechanism was suggested based on the structural similarity to helicases 11 , 12 , 27 and the recent cryo-EM structures of nucleosome-bound remodelers in the open and closed conformations 12 , 13 ., However , while the inchworm model can explain the motion of remodelers along naked DNA 38 , its application to nucleosome repositioning is far from trivial , since such motion would eventually result into steric clashes between the remodeler and the histone octamer ., Furthermore , complete nucleosome repositioning necessarily involves breakage of the many histone-DNA contacts that stabilize the nucleosome structure 31 , and it is not clear how the remodeler may perturb the contacts far away from the binding location at SHL 2 ., Experimental studies have also highlighted many diverse structural changes of the nucleosome occurring during remodeling , such as DNA twisting 26 , loops 39 or histone deformations 40 , suggesting these may be directly responsible for nucleosome sliding ., Interestingly , similar structural changes are also believed to mediate spontaneous sliding 41 ., Therefore , insights from research on spontaneous nucleosome repositioning may shed light into the more complex active case ., Indeed , DNA sliding on nucleosomes can also be simply driven by thermal fluctuations 42 ., Modes of nucleosome repositioning can be classified in two types depending on whether sliding is accompanied by the rotation of DNA around its axis 43 ., In the rotation-uncoupled mode , sliding proceeds via large steps of about a DNA turn ( ~10 bp ) 43–45 , possibly facilitated by the formation of loops 44 , 45 ., On the other hand , in the rotation-coupled mode , DNA sliding proceeds at small steps of 1 bp via a screw-like motion 43 , facilitated by the formation of twist defects 1 , 46–50 ., Twist defects are the structural deformations of DNA allowing to accommodate different numbers of base pairs between the strong histone-DNA contact points , which correspond to half-integer SHL locations 31 , 48 ., Between two adjacent contact points , while canonical DNA turns contains ~10-bp , 9 bp ( a missing bp ) or 11 bp ( an extra bp ) may also be accommodated; we refer to these deformations as -1bp and +1bp twist defects respectively ( see an example illustrated in Fig 1A , where the anti-clockwise motion of DNA at the contact point at SHL 1 . 5 generates a -1bp defect at SHL 2 , in green , and a +1bp defect at SHL 1 , in brown ) ., The spontaneous formation and propagation of twist defects around the nucleosome causes repositioning by 1 bp at the time 47 , 49 ., Notably , the small characteristic step sizes observed during nucleosome sliding by the ISWI and RSC remodelers 20 , 51 argues in favor of a role for DNA twisting in the molecular mechanism 26 ., Furthermore , having the same step size , active sliding via twist defects is compatible with the inchworm motion of the translocase domain ., While some remodelers have been shown to induce the formation of large DNA loops in nucleosomes 39 , this may be due to interactions with extra domains in the remodeler , and their presence should not rule out the contribution of twist defects to chromatin remodeling ., Due to the ubiquitous importance of remodelers for chromatin organization , gene expression and replication 2 , the detailed understanding of their molecular mechanism of action would be extremely valuable ., Coarse-grained molecular dynamics ( MD ) simulations 52 represent an ideal tool for approaching this problem , since their resolution can be high enough to accurately represent the potential key steps occurring during active nucleosome repositioning 43 , 49 , 53 , while achieving a speed-up of several orders of magnitude relative to all-atom simulations 52 , for which the system size and the relevant time scales would result in an exceedingly large computational cost ., Notably , coarse-graining approaches have been successfully applied to the study of nucleosome dynamics 9 , 54–56 , including spontaneous repositioning 43 , 45 , 49 , and ATP-dependent molecular motors 53 , 57 ., In this work , we investigate the fundamental mechanism of active repositioning in a minimal system consisting of the ATPase-translocase domain of the Snf2 remodeler from yeast in complex with the nucleosome 12 ., Firstly , we test that our model can reproduce the expected inchworm mechanism and unidirectional sliding during ATP consumption ., By comparing to the spontaneous case , we show how the remodeler induces directed repositioning by modifying the nucleosome free energy landscape via steric effects and long-range electrostatic interactions , explaining past experimental data on Snf2 mutants 12 ., Nucleosome repositioning occurs by coupling the ATPase inchworm motion to the formation and propagation of twist defects starting from the remodeler binding location at SHL 2 ., Finally , we reveal how DNA sequence can be exploited to control the kinetics of the system , consistently with its role in determining the repositioning outcome of many remodelers 18 , 58 , 59 ., We performed coarse-grained MD simulations of the ATP-dependent translocase domain of the Snf2 remodeler both on naked DNA and when bound to nucleosomes ( Fig 1C ) ., The nucleosome model is the same as that previously employed to study spontaneous nucleosome repositioning 43 , 49 , whereas the remodeler model and its interactions with the DNA are based on the cryo-EM structure of the Snf2-nucleosome complex with PDB id 5X0Y 12 ., Our computational model coarse-grains proteins at the level of individual residues 60 and DNA at the level of sugar , phosphate , and base groups , capturing the sequence-dependent flexibilities of base steps 61 , 62 ( see the Materials and Methods section and Refs . 43 , 60 , 61 for more details ) ., In the first set of simulations , the nucleosomal DNA consists of a 2-bp periodic sequence formed by repeating ApG base steps , polyApG ., This sequence was chosen because it was shown to display an intermediate flexibility on the nucleosome 49 , and it is used as a reference against which we compare other more experimentally and biologically relevant sequences ., In particular we consider the effects due to strong positioning along DNA ( 601 sequence 63 ) , and the introduction of poly ( dA:dT ) tracts and TpA repeats ., Based on the inchworm model ( Fig 1B ) , each remodeler chemical state ( apo , ATP or ADP ) corresponds to slightly different force-field parameters of the coarse-grained potential , and we simulate an ATP cycle ( apo→ATP→ADP→apo ) via switching the potential during the MD simulation , a common strategy in coarse-grained studies of molecular motors 57 , 64 ., Initially , in the open apo state , the remodeler configuration and the strengths of the interactions between ATPase lobes and DNA are as found in the cryo-EM structure with PDB id 5X0Y 12; then , switching to the ATP-bound potential enhances the attraction between the two lobes , favoring the closed conformation of the remodeler ., ATP hydrolysis is emulated by switching to the ADP-bound potential , which reduces the lobe 2-DNA interactions by a factor of 0 . 8 and weakens the attractive interaction between the two lobes to favor the open conformation ., In all our MD trajectories ( 40 on naked DNA , 100 on nucleosome for each DNA sequence ) , switching from apo to ATP states occurs at time 0 after 2x107 MD equilibration steps in the open conformation , ATP hydrolysis occurs after 107 MD steps , which are sufficient for the full relaxation of the system in the closed conformation , and finally switching back to the apo state occurs after 107 steps ( which are sufficient to observe translocase opening ) ., This simulation protocol is mainly motivated by the mechanism of helicase sliding identified from the analysis of crystal structures 37 ., More details are provided in the Materials and Methods section ., To analyze the repositioning dynamics in our MD trajectories , we track the motion of the DNA base pairs relative to the two individual ATPase lobes and relative to the histone octamer at the 14 histone-DNA contact points , located at the half-integer SHLs where the DNA minor groove faces the octamer ., We refer to these collective variables as the contact indexes: ΔbpL1 and ΔbpL2 for the remodeler lobes and Δbpi for the histone contacts , where i is the half-integer valued SHL of the contact ( these contacts will be indicated by their SHL value , e . g . contact point 1 . 5 ) ., As in our previous work 49 , the nucleosome contact indexes are evaluated relative to the 147-bp conformation found in the crystal structure with PDB id 1KX5 65 , which does not display twist defects ., The contact indexes take fractional values because they are continuous collective variables computed from the system coordinates 49 ( see Supporting Information S1 Text for the full description of their calculation ) ., Using these variables , we can fully characterize the remodeler’s inchworm dynamics , the sliding of DNA in the nucleosome , and the potential role played by twist defects ., These DNA deformations are distributed around integer-valued SHLs lacking direct histone-DNA contacts , and can accommodate either an extra base pair relative to the 1KX5 reference ( +1bp defect ) , or a missing base pair ( -1bp defect ) 1 , 46 , 47 , 66 ., A twist defect at SHL i can be simply evaluated by the difference between the neighboring contact indexes ( i-1/2 and i+1/2 ) : a defect value close to zero corresponds to the standard non-defect case , a value close to 1 to an extra base pair and a value close to -1 to a missing base pair ( see an example in Fig 1A ) ., In this section , we present the simulation results of both Snf2-naked DNA and Snf2-nucleosome systems , but focusing on the motion of the remodeler relative to the DNA ., In Fig 2A , we show that by switching between the remodeler chemical states during the MD simulation , this can slide along both naked and nucleosomal DNA by 1-bp ( as evidenced by the change in the average lobe contact index during the ATP cycle ) ., While sliding on naked DNA is not a key function of remodelers , this process has been documented in experiments 38 ., Interestingly , we note that under the current computational settings sliding by 1 bp by the end of an ATP cycle occurs with higher probability when in complex with nucleosomes ( 98% ) than when on naked DNA ( 45% ) , whereas in the remaining cases the remodeler simply goes back to its original position ., Fig 2B displays two representative trajectories ( one on the naked DNA and one on the nucleosome ) projected onto the contact indexes of the two separate ATPase lobes ., In both cases , this projection clearly highlights the inchworm motions ., Specifically , starting from the apo state in the open conformation ( bottom left in the figure ) , switching the potential to the ATP state induces the closure of the remodeler , with lobe 1 moving by 1 bp towards lobe 2 , which maintains its position due to its stronger grip to the DNA ( bottom right ) ., Then , simulating ATP hydrolysis via switching to the ADP-state potential induces the domain opening , but since the lobe 2-DNA interactions are also decreased , now it is this lobe that usually moves by 1 bp away from lobe 1 ( top right ) ., Switching again to the apo-state potential simply restores the original lobes-DNA interaction strengths , maintaining the same open configuration and completing a full ATP cycle with the remodeler shifted by 1 bp relative to where it started ., On naked DNA , this mechanism is sufficient to explain the translocase’s unidirectional motion ( see S1 Movie for a visualization of the trajectory in Fig 2B ) ., However , what is not clear from this analysis is how the translocase motion may induce sliding of nucleosomal DNA; the next sections are devoted to the characterization of the complete active repositioning process ., Our MD simulations show that the ATP-driven translocase closure is followed by sliding of nucleosomal DNA ., Specifically , the DNA at the remodeler binding location slides unidirectionally towards the dyad , as indicated by the 1-bp increase in the average nucleosome contact index around SHL 2 , ( Δbp1 . 5+Δbp2 . 5 ) /2 ( see two representative trajectories and the cumulative distribution at the end of the ATP-state in Fig 3A , upper panel ) ., This is consistent with the directionality of repositioning observed in experimental studies 32 and with what expected from the comparison with the structure of helicases 12 ., On the other hand , in the absence Snf2 , sliding of the same polyApG nucleosomal DNA occurs in a random direction ( Fig 3A , lower panel ) ., To better characterize the origin of the unidirectional motion , in Fig 3B , we compare the free-energy profiles of nucleosome sliding along the average contact index around SHL 2 for different scenarios ., In the absence of Snf2 , as expected for the uniform polyApG sequence and the random motion reported in Fig 3A , sliding by 1 base pair in either direction does not change the free energy of the system , but it involves climbing significant free-energy barriers ( ~6 kBT ) ., The presence of the remodeler modifies the original nucleosome landscape in a chemical-state-dependent fashion ., In the initial open conformation before ATP binding , there is a single free-energy minimum at ( Δbp1 . 5+Δbp2 . 5 ) /2 = 0 , so that nucleosomal DNA sliding is strongly inhibited ., Instead , after ATP binding and translocase closure , a second deeper free-energy minimum appears around ( Δbp1 . 5+Δbp2 . 5 ) /2 = 1 , favoring DNA sliding towards the dyad ., After the last opening conformational change following ATP hydrolysis , the nucleosome landscape returns to have a single free-energy minimum now at ( Δbp1 . 5+Δbp2 . 5 ) /2 = 1 , so that further sliding is inhibited ., The switching among these different free-energy landscapes reveals a clear ratchet mechanism , as often employed for the theoretical modeling of molecular motors 67 ., The changes in the free energy profiles can be in part explained by the inchworm motion of the translocase domain and steric effects ., In the open conformation , DNA sliding at SHL 2 by 1 bp in either direction would cause steric overlap between lobe 2 and histone octamer on one side or overlap between lobe 1 and the opposite DNA gyre around SHL -6 on the other side ( see inset in Fig 3B ) , blocking nucleosome repositioning and explaining the single free energy minimum when Snf2 is in the apo state ., Since the translocase closure upon ATP binding involves the motion of lobe 1 towards lobe 2 , DNA sliding is now allowed to occur towards the dyad , causing the translocase to swing on the opposite side of its binding location ( see cartoons in Fig 3A ) ., However , this argument does not explain the large extent to which the closed ATPase favors unidirectional repositioning , i . e . the decrease in free energy by ~4 kBT from ( Δbp1 . 5+Δbp2 . 5 ) /2 = 0 to 1 ., From the crystal structure of the nucleosome-bound Snf2 remodeler 12 , it was shown that apart from the main interactions at SHL 2 , the translocase domain also interacts with the opposite DNA gyre around SHL -6 via long-range electrostatics mediated by residues K855 , R880 and K885 , located within lobe 1 ( see bottom view in Fig 1C and inset in Fig 3B ) ., It was also experimentally shown that changing these residues from positively- to negatively-charged markedly reduced the remodeling activity of Snf2 12 ., To investigate this effect , we performed MD simulations where the three key residues have all been mutated to glutamic acid ( K855E-R880E-K885E mutant ) ., While still possible , DNA sliding around SHL 2 is no longer accompanied by a large decrease in free energy ( Fig 3B ) ., This change can be understood in terms of the movement of the ATPase lobe 1 during repositioning ., In the open state , lobe 1 is close to the contact point 1 . 5 and also interacts with the opposite gyre at SHL -6 via the basic patch in wild-type ( WT ) Snf2 ( K855 , R880 and K885 ) ., After ATP binding , lobe 1 moves by 1 bp towards lobe 2 , becoming further apart from both contact point 1 . 5 and the DNA at SHL -6 , weakening the electrostatic interaction ( specifically , the average distance between the center of mass of the lobe 1 patch and the DNA phosphate backbone increases from ~6 . 1 Å to ~8 . 3 Å upon translocase closure ) ., The sliding of nucleosomal DNA causes lobe 1 to swing back towards the dyad , restoring also the original interactions between the basic patch and SHL -6 ., Comparing to the initial open apo structure , the translocase closure and subsequent sliding of DNA at SHL 2 makes it appear that lobe 2 moved by 1 bp towards lobe 1 , and not the opposite ., Notably , this observation is consistent with the recent cryo-EM structure of the nucleosome-Chd1 complex in the presence of an ATP analog 13 , where lobe 1 overlaps with the corresponding lobe in the open conformation of the Snf2 remodeler 12 , whereas lobe 2 appears to have moved by 1 bp 13 ., While so far we only considered a simple uniform polyApG sequence , genomes are rich in positioning motifs that contribute to specify the optimal location of nucleosomes along DNA 4 ., These motifs , such as T/A base steps periodically spaced every 10 bp , cause the intrinsic bending of DNA , which lowers the free energy cost of nucleosome assembly , and favor a specific rotational setting , as they preferentially locate where the DNA minor groove faces the histone octamer 9 ., These signals strongly inhibit nucleosome sliding relative to random DNA sequences , since repositioning would proceed either by DNA screw-like motion via a high-energy intermediate with a non-optimal rotational setting 47 , 49 , or via alternative repositioning mechanisms uncoupled with DNA rotations , which involve the energetically-costly breakage of many histone-DNA contacts 43 , 45 ., Nevertheless , chromatin remodelers are still able to actively reposition nucleosomes made with strong positioning sequences such as 601 12 , 20 , 63 ., To investigate the robustness of the active repositioning mechanism against changes in DNA sequence , we next run MD simulations of Snf2 in complex with nucleosomes made with the 601 sequence 63 ., In the starting configuration , we shifted the DNA by 3 bp relative to the optimal configuration found in the 5X0Y structure , in the direction from the remodeler site towards the dyad ., We refer to this sequence as 601Δ3 ., Because of the non optimal location of the T/A steps relative to the histone octamer ( see cartoon in Fig 3C ) , starting from here in the absence of the remodeler will be most likely followed by sliding backward away from the dyad , i . e . towards the optimal configuration ( in about half of the cases within 107 MD steps , Fig 3C , lower panel ) ., Instead , not only the remodeler prevents backward sliding , but upon ATP binding , in about half of the cases , it can also induce sliding of nucleosomal DNA forward towards the dyad ( Fig 3C , upper panel ) , in the same way as observed with the uniform polyApG sequence ., A comparison of the free-energy landscapes along DNA sliding at SHL 2 with and without remodeler ( Fig 3D ) shows indeed that in the case without remodeler the free energy strongly increases with sliding forward towards the dyad and decreases away from the dyad , whereas the closed translocase is able to lower the free energy cost of forward sliding to ~0 kBT , while preventing sliding backward in the opposite direction via steric effects ., The free energy profile obtained with the K855E-R880E-K885E Snf2 bound to 601Δ3 nucleosomes , shows that this mutant cannot slide these strong positioning sequences , due to an extra free energy penalty of ~3 kBT upon sliding by 1 bp ., This is consistent with the results from experiments on similar Snf2 charge mutants sliding 601 nucleosomes 12 ., While the limitations of our computational model ( e . g . the assumptions on the precise ATP hydrolysis kinetics ) prevent us from making quantitative predictions of remodeling activity , our simulations provide a mechanistic understanding of the important role of electrostatic interactions in directing repositioning 12 ., To show that Snf2 is also able to reposition natural genomic sequences ( both polyApG and 601 are artificial ) , we also performed MD simulations using the weakly positioning 5S rDNA sequence commonly studied experimentally 68 , starting with a nucleosome at the expected equilibrium location 69 ., Even in this case , remodeling by 1 bp at the end of an ATP cycle occurs successfully in most trajectories ( 15 out of 20 ) ., Apart from the interactions between lobe 1 and DNA at SHL -6 , the cryo-EM structure of Snf2 also highlighted electrostatic contacts between the H4 N-terminal tail and an acidic patch located on lobe 2 ( E1069 , D1121 ) 12 ., Analyzing our trajectories , we found that indeed the tip of the H4 tail often localizes in the vicinity of the Snf2 acidic patch ( ~1 nm distance between Cα atoms ) ., However , these interactions do not have any particular correlation with the sliding of nucleosomal DNA or the inchworm motion of the remodeler , suggesting that they do not play a fundamental role ., Although mutations in the H4 tail do have a minor effect on the remodeling activity of the Snf2 translocase 12 , we suspect this should be mainly due to a reduction in the binding affinity to the nucleosome ., So far , we focused on the inchworm motion of the translocase domain and on nucleosomal DNA sliding at the SHL 2 binding site , establishing how these two are tightly coupled ., However , a full characterization of the repositioning mechanism requires the analysis of DNA sliding at the individual histone-DNA contact points on the entire nucleosome ., In Fig 4A , we plot the timelines of the contact index coordinates of both remodeler and nucleosome for two representative trajectories during which repositioning by 1 bp occurs ., These plots show how nucleosomal DNA sliding is initiated near the translocase binding location at the contact point at SHL 1 . 5 , with the creation of opposite-type twist defects at the neighboring SHLs ., The diffusion of these defects then completes repositioning of the entire nucleosome ., To aid the understanding of the dynamics , we label the key metastable conformations of the system according to the following rules: the first letter , o or c , corresponds respectively to open or closed translocase conformation; when it is closed ( c ) , the domain can adopt distinct configurations with DNA and histone octamer within its binding site at SHL 2 , which will be indicated by a capital letter as A , B , C or D ( see below for definition ) ; finally , a last integer number , 0 , 1 or 2 , indicates the number of +1bp defects which may form near the dyad at the three central SHLs ( SHL -1 , 0 , and +1 , these defects are most favorably found at SHLs +/-1 ) ., In the first trajectory ( the left panel in Fig 4A ) , starting from an open translocase bound to a nucleosome in a standard 1KX5-like configuration lacking twist defects ( state o0 ) , switching to the ATP-bound potential at time 0 quickly induces the closure of the remodeler via the motion of lobe 1 towards lobe 2 ( 0 . 03x106 MD steps , state cA0 ) ., In this first closed configuration , the lobe 1-DNA interface is destabilized relative to the one observed in the reference 5X0Y structure and the motion towards lobe 2 is only partial ( ΔbpL1~0 . 6 ) ., Only after some time ( 0 . 09x106 MD steps ) the motion of lobe 1 is complete ( ΔbpL1~1 , state cB0 ) ., From this closed configuration , we observe motion of nucleosomal DNA towards the dyad relative to the histone octamer starting from SHL 1 . 5 ( 0 . 13x106 MD steps , state cC1 ) , causing the accumulation of an extra base pair at SHL 1 , and a missing base pair at SHL 2 ( where the remodeler is bound ) ., Soon afterwards ( 0 . 2x106 MD steps ) , the nucleosomal DNA further slides from the remodeler site up to the closest nucleosome entry/exit , releasing the -1bp defect ( state cD1 ) ., As highlighted in the previous section , while these two steps do not involve remodeler’s motion relative to the DNA , the DNA motion relative to the histone octamer causes the remodeler to swing by 1 bp towards the dyad and enables to re-establish the electrostatic contacts between lobe 1 and SHL -6 , which were lost during the initial ATPase closure ., State cD1 , for the polyApG sequence considered here , is the most stable configuration among the closed ones ., Repositioning is usually completed only after ATP hydrolysis ( 10x106 MD steps ) , which causes translocase opening via lobe 2 motion by 1 bp ( state o1 ) ( the full pathway is o0→ cA0→ cB0→ cC1→ cD1→ o1→ o0 ) ., The very last step consists of the sliding of nucleosomal DNA from the translocase up to the far nucleosome entry/exit , releasing the +1bp defect near the dyad ( state o0 , see S2 and S3 Movies for visualizations of this trajectory ) ., The second trajectory ( Fig 4A , right ) is qualitatively similar to the first , except that all states have an additional defect near the dyad at the starting time ( the full pathway is then o1→ cA1→ cB1→ cC2→ cD2→ cD1→ o1 for trajectory 2 ) ., In particular , motion at the remodeler and nucleosome contact points proceeds in the same order ., These two pathways are representative of the most common ones found in our 100 MD trajectories: the first one being observed in 16 cases , while the second one in 21 cases ., In all trajectories , repositioning involves twist-defect formation and propagation starting from the remodeler binding location , displaying only small deviations from those shown in Fig 4A ., Trajectories can be projected onto a low dimensional space defined by the sum of the contact indexes around the remodelers binding location ( ΔbpL1+ΔbpL2+Δbp1 . 5+Δbp2 . 5 , the horizontal axis in Fig 3B ) and by the size of the twist defect around the dyad ( Δbp1 . 5-Δbp-1 . 5 , the vertical axis in Fig 4B ) ., On this space , all the key metastable states involved in repositioning can be clearly separated ( in Fig 4B we show trajectories 1 and 2 from panel a ) ., From this figure we notice that most key conformational changes occur in the closed conformation ( between two dotted lines ) ., To test the importance of the system relaxation in this portion of the phase space , we in | Introduction, Results, Materials and methods | ATP-dependent chromatin remodelers are molecular machines that control genome organization by repositioning , ejecting , or editing nucleosomes , activities that confer them essential regulatory roles on gene expression and DNA replication ., Here , we investigate the molecular mechanism of active nucleosome sliding by means of molecular dynamics simulations of the Snf2 remodeler translocase in complex with a nucleosome ., During its inchworm motion driven by ATP consumption , the translocase overwrites the original nucleosome energy landscape via steric and electrostatic interactions to induce sliding of nucleosomal DNA unidirectionally ., The sliding is initiated at the remodeler binding location via the generation of a pair of twist defects , which then spontaneously propagate to complete sliding throughout the entire nucleosome ., We also reveal how remodeler mutations and DNA sequence control active nucleosome repositioning , explaining several past experimental observations ., These results offer a detailed mechanistic picture of remodeling important for the complete understanding of these key biological processes . | Nucleosomes are the protein-DNA complexes underlying Eukaryotic genome organization , and serve as regulators of gene expression by occluding DNA to other proteins ., This regulation requires the precise positioning of nucleosomes along DNA ., Chromatin remodelers are the molecular machines that consume ATP to slide nucleosome at their correct locations , but the mechanisms of remodeling are still unclear ., Based on the static structural information of a remodeler bound on nucleosome , we performed molecular dynamics computer simulations revealing the details of how remodelers slide nucleosomal DNA: the inchworm-like motion of remodelers create small DNA deformations called twist defects , which then spontaneously propagate throughout the nucleosome to induce sliding ., These simulations explain several past experimental findings and are important for our understanding of genome organization . | electricity, dna-binding proteins, electrostatics, sequence motif analysis, epigenetics, thermodynamics, chromatin, research and analysis methods, sequence analysis, chromosome biology, proteins, bioinformatics, gene expression, chemistry, histones, nucleosomes, free energy, physics, biochemistry, biochemical simulations, hydrolysis, cell biology, database and informatics methods, genetics, biology and life sciences, chemical reactions, physical sciences, computational biology, atp hydrolysis | null |
journal.pgen.1003888 | 2,013 | tRNA Methyltransferase Homolog Gene TRMT10A Mutation in Young Onset Diabetes and Primary Microcephaly in Humans | Type 2 diabetes ( T2D ) is a heterogeneous polygenic disease with dramatically increasing worldwide incidence as a consequence of the obesity epidemic 1 ., Environmental factors ( energy dense diets rich in saturated fat and sedentary lifestyle 2 , 3 ) and genetic predisposition contribute to its pathogenesis ., T2D develops when β-cells fail to compensate for peripheral insulin resistance by increasing insulin secretion 4 , 5 as a consequence of β-cell dysfunction and reduced β-cell mass ., Genome-wide association studies have identified a number of loci where genetic polymorphisms associate with T2D 6 ., Inherited mutations in genes at some of these loci have been shown to cause monogenic forms of diabetes , indicating that genetic variants of different severity can generate a spectrum of monogenic and polygenic forms of diabetes 7 ., An example of a T2D risk gene is CDK5 regulatory associated protein 1-like 1 ( CDKAL1 ) ., Polymorphisms in this gene have been associated with T2D across ethnic populations 8 ., CDKAL1 encodes a transfer RNA ( tRNA ) methylthiotransferase that catalyzes the methylthiolation of tRNALys ( UUU ) 9 ., Cdkal1-deficient β-cells have impaired glucose-induced insulin secretion , and Cdkal1 knockout mice develop glucose intolerance due to aberrant insulin synthesis 9 ., tRNAs undergo modifications of their bases or sugar moieties that are crucial for proper cellular function ., Mammalian cells have an average of 13–14 modifications per tRNA 10–12 , methylation being the most common one 12 ., Chemical modifications of nucleotides surrounding anticodons of tRNAs are important to preserve translational efficiency and fidelity 13 , modifications in the main body of the tRNA affect its folding and stability , and other modifications at various positions influence tRNA identity 14 , 15 ., Here we identified a nonsense mutation in TRMT10A ( also called RG9MTD2 ) in a new syndrome of young onset diabetes and microcephaly ., The TRMT10A yeast ortholog YOL093w codes for the protein TRM10 that has tRNA methyltransferase activity ., TRM10 specifically methylates tRNA-Arg , -Asn , -Gln , -Thr , -Trp , -Met and -Lys at position 9 ( m1G9 ) , using S-adenosylmethionine ( SAM ) as methyl donor 16 ., TRM10 was shown to be the major if not the only m1G9 methyltransferase in yeast , but its knockout did not alter cell survival or growth 16 ., Mutational analysis in yeast revealed potential interactions between TRM10 , TRM8/TRM82 , and TRM1 17 ., These latter proteins have tRNA methyltransferase activity towards m7G46 and m22G26 , respectively 12 ., The concomitant deletion of TRM10 with TRM8 , TRM82 or TRM1 induced growth arrest in S . cerevisiae exposed to high temperature , suggesting enhanced tRNA instability 17 ., Here we describe the affected siblings and the identification of the TRMT10A mutation ., We followed this up with studies of TRMT10A expression in tissues and subcellular localization , and interrogated the functional consequences of TRMT10A deficiency ., The proband was born to consanguineous parents of Moroccan origin , her paternal and maternal grandmothers being sisters ( Figure 1 ) ., Head circumference , weight and length at birth are unknown ., At age 26 years she had short stature ( 143 cm ) , microcephaly ( adult head circumference 49 cm , -5SD ) and intellectual disability , with a history of petit mal epilepsy in adolescence ., Magnetic resonance imaging of the head showed a small brain with no malformation or other abnormality ( Figure 1 ) ., She had developed diabetes at the age of 22 years ., At diagnosis her body mass index ( BMI ) was 26 . 9 kg/m2; plasma glucose was 176 mg/dl and HbA1c 11 . 3% ., Other features were a short neck , wide nose , low hairline , buffalo hump , retraction of the right 5th toe , scoliosis , and joint laxity ., She also had osteoporosis , with dual-energy X-ray absorptiometry T-scores of -2 . 7 and -3 . 5 at the lumbar spine and femoral neck , respectively ., A skeletal survey revealed no epiphyseal dysplasia or other bone abnormality ( e . g . normal X-ray of the hands , Figure 1 ) ., Her sister had short stature ( 154 cm ) , microcephaly ( adult head circumference 51 cm , -3SD ) and intellectual disability ( IQ 69 ) ., She developed diabetes at the age of 19 years , presenting with a fasting glucose of 365 mg/dl and HbA1c 13 . 2% ., Her BMI was 21 . 7 kg/m2 ., A younger brother had short stature ( 141 cm at age 14 years and final height of 157 cm at 21 years ) , microcephaly ( head circumference 51 cm , -3SD ) and mental retardation ( IQ 52 ) ., His head circumference at birth was reportedly normal ( 36 cm ) ., He was diagnosed with diabetes at 14 years of age , with a plasma glucose of 251 mg/dl and HbA1c 11 . 1% ., His BMI was 20 . 6 kg/m2 ., None of the patients had ketoacidosis and all three were treated with insulin at diagnosis ., They were negative for anti-insulin , anti-GAD65 , anti-IA2 and islet cell autoantibodies and had a HLA genotype that did not confer risk for type 1 diabetes ., Endogenous insulin secretion persisted , shown by C-peptide measurements for up to 20 years of follow-up ., The insulin requirements were moderate with an average insulin dose of 0 . 4–1 . 2 U/kg/day; glycemic control ranged from good to insufficient ( HbA1c 6 . 5–8 . 5% ) ., After 18 years of diabetes , the probands ophthalmologic examination revealed bilateral diabetic retinopathy and cortical cataract ., The parents and non-affected siblings had normal size ( parents 166 and 157 cm , siblings 160 , 175 , 183 and 159 cm ) and head circumference ( both parents 58 cm , P97 ) ., The parents developed diabetes at age 58 years ( BMI 30 . 9 and 31 . 6 kg/m2 , plasma glucose 124 and 169 mg/dl and HbA1c 8 . 3 and 7 . 6% in the mother and father , respectively ) and were treated with metformin and a sulphonylurea ., One grandfather and two aunts had adult onset diabetes ( Figure 1 ) ., One sister had gestational diabetes at the age of 22 years; her fasting plasma glucose was normal ( 90 mg/dl ) at age 30 ( Figure 1 ) ., The GeneChip SNP array analysis identified only one large ( >3 cM ) homozygous genomic region that was common to the three affected siblings ., It was located on chromosome 4q22-23 and spanned 12 . 4 Mb between heterozygous SNPs rs4128340 and rs10516462 ., In this segment , we genotyped 15 microsatellite markers , which confirmed homozygosity and biparental inheritance of a haplotype shared by both parents ( Figure S1 ) ., The multipoint LOD score was 3 . 0 ., Microsatellite analysis in the unaffected sister with a history of gestational diabetes ( Figure, 1 ) showed inheritance of the non-mutated maternal haplotype and of the mutated paternal haplotype ., In an additional unaffected brother with normal fasting plasma glucose ( 84 mg/dl ) and HbA1c ( 5 . 1% ) at age 23 years , we observed a critical meiotic recombination event , resulting in homozygosity for all disease-associated markers except those distal to microsatellite D4S1628 ., This recombinant chromosome reduced the critical linkage region to a 3 . 1 Mb segment at 4q23 ., We initially sequenced the following genes located in the 3 . 1 Mb segment and considered as candidates: H2AFZ ( H2A histone family , member Z ) , LAMTOR3 ( late endosomal/lysosomal adaptor , MAPK and MTOR activator 3 ) , DDIT4L ( DNA-damage-inducible transcript 4-like ) , RAP1GDS1 ( RAP1 , GTP-GDP dissociation stimulator, 1 ) and METAP1 ( methionyl aminopeptidase 1 ) , but no mutation was identified ., Exonic sequences-enriched DNA ( whole exome ) sequencing was performed in one proband and results were analyzed for variants that were not found in: dbSNP135 database , the Thousand Genomes database , the Exome Variant Server , or in-house exome sequencing on 51 individuals ., There was only a single candidate mutation in the 3 . 1 Mb critical linkage segment , a homozygous G to A transition in exon 4 of gene TRMT10A ( tRNA methyltransferase 10 homolog A ( S . cerevisiae ) at position 379 of the coding DNA sequence , predicted to replace an Arginine residue with a premature termination codon at position 127 of the polypeptide ( c . 379, G>A; p . Arg127Stop ) ., Sanger sequencing confirmed the mutation ( Figure 2A ) , which was homozygous in the three affected patients and heterozygous in both parents as well as in the unaffected brother with the critical recombination event ., A comparison across species shows that Arg127 and the surrounding region are highly conserved ( Figure S2 ) ., Outside the linkage region , exome analysis in the proband identified biallelic , potentially damaging mutations in the six following genes: BCLAF1; CES1; EVC2; PTPN22; ST13; ZNF626 ., As none were concordant in the three affected siblings , we rejected them as candidate mutations ., We sequenced the 8 exons and flanking intronic sequences in 20 patients with a similar phenotype of young onset diabetes associated to intellectual disability , microcephaly , epilepsy , developmental delay and/or short stature , five of whom were born to consanguineous parents , but failed to identify another patient with biallelic disease-causing mutations ., We furthermore sequenced TRMT10A in 26 patients with non-autoimmune diabetes with onset before 25 years and a positive family history of diabetes , in whom no mutation was identified in known MODY-associated genes , but did not identify any mutation in TRMT10A ., To examine the outcome of the TRMT10A nonsense mutation on TRMT10A protein and mRNA expression , we performed Western blot and real-time PCR on lymphoblasts from two patients , a heterozygous carrier of the mutation , and three healthy controls ., TRMT10A protein was absent in lymphoblasts from patients homozygous for the Arg127Stop mutation ( Figure 2B ) ., TRMT10A mRNA expression was much reduced in patients , and intermediate in the carrier ( Figure 2C ) ., This finding is consistent with nonsense-mediated mRNA decay induced by the premature translation-termination codon ( PTC ) and/or by PTC-induced transcriptional silencing of the affected gene , a mechanism known to prevent the synthesis of potentially deleterious truncated proteins 18 , 19 ., We next evaluated TRMT10A transcript and protein expression in rat tissues ., TRMT10A was ubiquitously expressed with similar mRNA abundance in liver , kidney , spleen , lung , fat , and brain ., Heart and muscle showed lesser TRMT10A mRNA expression , while pancreatic islets were enriched in TRMT10A transcripts ( Figure 3A ) ., TRMT10A protein was ubiquitously present and 2- to 3-fold more abundant in brain and pancreatic islets compared to other tissues ( Figure 3B–C ) ., In situ hybridization studies were performed in human embryonic brain samples at 8 , 11 , 17 and 19 gestational weeks ( GW ) ., TRMT10A was expressed throughout the whole thickness of the dorsal telencephalon ( presumptive cerebral cortex ) at 8 and 11 GW , with higher expression in the ventricular zone and marginal zone ( Figure 4 ) ., The ventricular zone contains most neural progenitors at early stages of corticogenesis , while the marginal zone is the region where the first post-mitotic neurons migrate ., At later stages TRMT10A expression was not detected in the dorsal telencephalon but was found in the cerebellar cortex and cerebellar nuclei ( Figure S3 and data not shown ) ., To examine TRMT10A subcellular localization we first performed in silico TRMT10A topology prediction using PSORII and WoLF PSORT 20 ., These softwares detected monopartite and bipartite nuclear localization signals in the first 89 amino acids of the protein ., This was confirmed with cNLS Mapper 21 , 22 suggesting predominant nuclear localization ., To experimentally demonstrate the TRMT10A subcellular localization we took two approaches:, 1 ) Expression of a fluorescent recombinant fusion protein , human TRMT10A ( hTRMT10A ) -humanized Renilla green fluorescent protein ( hrGFP ) ;, 2 ) Detection of endogenous TRMT10A by immunofluorescence ., Confocal analysis of clonal rat INS-1E β-cells expressing the TRMT10A-hrGFP fusion protein showed nuclear fluorescence with intense signal accumulation in nuclear regions of low Hoechst 33342 staining ( Figure 5A ) ., Cells expressing hrGFP alone showed homogeneous cytosolic and nuclear fluorescence ., The identity of the recombinant fusion protein expressed in these cells was confirmed by Western blot ( Figure 5B ) using an antibody raised against purified recombinant hTRMT10A ., Similar results were obtained in dispersed rat and human islet cells expressing the recombinant fusion protein ( Figure S4 ) ., To identify the nuclear compartment enriched in TRMT10A , we performed immunofluorescence in rat and human islet cells using antibodies against hTRMT10A and fibrillarin , a nucleolar marker 23 ., Immunostaining of endogenous TRMT10A ( Figure 6 , red ) mimicked the fluorescence profile of recombinant TRMT10A-hrGFP ., Fibrillarin immunolabeling showed a similar punctuate nuclear pattern ( Figure 6 , green ) ., TRMT10A and fibrillarin images were superimposable ( Figure 6 , merge ) indicating that TRMT10A expression is enriched in the nucleolus ., RNA interference technology was used to knock down TRMT10A in β-cells ., TRMT10A mRNA and protein expression was reduced by 50% in INS-1E cells ( Figure S5 ) ., TRMT10A silencing did not modify glucose-induced insulin secretion and insulin content ( Figure S6 ) , but enhanced total protein biosynthesis by 25% in clonal rat β-cells ( Figure 7 ) ., We next examined whether TRMT10A silencing affects β-cell survival ., TRMT10A knockdown induced apoptosis in clonal and primary rat β-cells and dispersed human islets ( Figure 8 ) ., TRMT10A deficiency further sensitized rat β-cells to oleate- , palmitate- and ER stress-induced apoptosis ( Figure 8A–D ) ., These results were confirmed by Western blot for cleaved caspase-3 , showing increased caspase-3 activation basally and after palmitate and cyclopiazonic acid exposure ( Figure 8E ) ., High glucose-induced β-cell apoptosis was also increased by TRMT10A silencing ( Figure 8A ) ., We observed that TRMT10A expression in β-cells is modulated by ER stress ., Exposure of rat or human β-cells to the saturated FFA palmitate , previously shown to induce ER stress 3 , 24 , 25 , or to chemical ER stressors enhanced TRMT10A expression ( Figure S7 ) to an extent that was correlated with the intensity of ER stress ( measured by the expression of ER stress markers , Figure S8 ) ., TRMT10A expression was induced in a PERK- but not IRE1-dependent manner ( Figure S9 ) ., TRMT10A silencing did not induce expression of the ER stress markers BiP , XBP-1s , ATF3 and CHOP ( data not shown ) ., In a large consanguineous family of Moroccan origin , we identified a new syndrome of severe insulinopenic young onset diabetes and microcephaly with intellectual disability ., We used linkage analysis and whole exome sequencing to identify the causal mutation ., We found only one region of homozygosity by descent shared by the three affected patients , and only one potentially damaging rare genetic variant in this region , located in the TRMT10A gene , changing an arginine codon at position 127 of the protein into a stop codon ( Arg127Stop ) ., In the rest of the patients exome , we found no potentially damaging , rare biallelic variants shared by the three patients that might have qualified for a causal mutation ., Among the family members , four were heterozygous carriers of a mutant allele ., Of these , the parents developed diabetes in their fifties , one sister had gestational diabetes , and one brother had normal plasma glucose levels at the age of 23 ( Figure 1 ) ., Other family members were not available for testing ., It is possible that TRMT10A haploinsufficiency increases the risk for adult onset diabetes ., TRMT10A contains 8 exons , the first exon being non-protein coding ., The mutated codon 127 is in exon 4 ., The protein environment of Arg127 is extremely conserved across species ., Little is known about the role of TRMT10A in mammals ., A single study suggested altered TRMT10A mRNA expression in colorectal cancer 26 ., Blast analysis indicated that TRMT10A is the mammalian ortholog of S . cerevisiae TRM10 , previously shown to be involved in guanine 9 tRNA methylation m1G9 16 ., TRMT10A has seven transcripts in the Vega database ., Two of them are non-protein coding due to a retained intron , three contain 8 exons coding for identical proteins of 339 amino acids , and differ only in their untranslated regions ., InterProScan analysis indicates that these three proteins have a tRNA ( guanine 9-N1 ) methyltransferase domain as well as tRNA ( guanine-N1 ) methyltransferase domain , both of them present in TRM10 ., The last two TRMT10A transcripts contain only 6 exons and code for shorter proteins of 200 and 206 amino acids ., These two variants are truncated at the C-terminus and only have the tRNA ( guanine-N1 ) methyltransferase domain ., In rat only one isoform of TRMT10A containing both domains is found ., Based on these analyses , we suggest that TRMT10A functions as a tRNA-modifying enzyme , but this remains to be experimentally confirmed ., The Arg127Stop mutation is predicted to block the expression of the five coding human TRMT10A isoforms ., The nonsense mutation abolished TRMT10A protein expression , and also significantly reduced its mRNA expression ( Figure 2 ) , probably by nonsense-mediated decay and/or transcriptional silencing 18 , 19 ., We show that TRMT10A is ubiquitously expressed but enriched in brain and pancreatic islets ( Figure 3 ) , consistent with the tissues affected in this new syndrome of diabetes and microcephaly ., In silico topology prediction indicates that the five human TRMT10A isoforms , as well as the rat enzyme , have predominant nuclear localization ., This was confirmed by immunofluorescence and confocal microscopy , with TRMT10A mainly localizing in the nucleolus of β- and non-β-cells ( Figure 5–6 and S4 ) ., tRNA transcription and early processing occurs in several subcellular compartments including the nucleus , cytoplasm and cytoplasmic surface of the mitochondria 14 ., tRNA genes are recruited to the nucleolus for transcription 27 , 5′ leader sequence removal and 3′ end modification , removal of the 3′ trailer and addition of the CCA , which is required for efficient tRNA nuclear export 28 ., Mature tRNAs are exported to the cytosol for aminoacylation and function in translation ., This transport is not unidirectional; cytosolic tRNAs can follow a retrograde transport to the nucleus ( e . g . during nutrient deprivation ) , to be re-exported to the cytosol following nutrient availability 14 ., Some tRNA modifications occur on initial tRNA transcripts , while others are introduced in end-matured tRNAs 29 ., Since tRNA transcription and maturation occurs in the nucleus it is expected that the enzymes catalyzing these modifications have a nuclear localization ., Studies in yeast confirmed that a subset of tRNA methyltransferases is located in the nucleus 28 , 30 , 31 , with distinct subnuclear distribution , i . e . nucleolus , nucleoplasm , or inner nuclear membrane; the reason for these different localizations is not known 14 , 31 ., The predominant nucleolar localization of TRMT10A is consistent with its proposed tRNA modifying activity ., Alterations in tRNA modification are expected to affect protein translation ., We showed that TRMT10A knockdown in rat β-cells enhances total protein biosynthesis ( Figure 7 ) ., TRMT10A silencing does not impair glucose-induced insulin secretion or content in β-cells ( Figure S6 ) , suggesting that TRMT10A deficiency has no major impact on β-cell function ., TRMT10A knockdown sensitizes β-cells to apoptosis in control condition and after exposure to FFAs , high glucose or synthetic ER stressors ( Figure 8 ) , conditions related to T2D ., It has been proposed that mammalian cytosolic and mitochondrial tRNAs prevent apoptosis by blocking the binding of cytochrome c to Apaf-1 , thus preventing the formation of the apoptosome 32 , 33 ., It is not known whether tRNA modifications affect this tRNA-cytochrome c interaction ., Primary microcephaly refers to a congenitally small but otherwise normally structured brain , with a head circumference later in life that remains 3 SD below the mean for age and gender ., Primary microcephaly is a very rare disorder affecting approximately 1/100 , 000 live births , mainly inherited as an autosomal recessive trait , and is associated with a high rate of parental consanguinity 34 ., Microcephaly and young onset diabetes co-segregate in the present family , as both features were present in the three affected siblings and absent in the six unaffected siblings , defining a new syndrome ., Our linkage analysis identified a single region where all affected siblings were homozygous over a significant length of genomic DNA ., It is hence likely that the whole phenotype results from pleiotropic effects of a single mutated gene ., Microcephaly in our patients was associated with intellectual disability and no other neurological feature , except for a history of petit mal seizures in the proband ., This clinical presentation fits with the phenotype of primary microcephaly 35 ., Primary microcephaly is vastly heterogeneous , and many genes that cause primary microcephaly play a role in mitotic spindle organization and/or DNA repair , presumably affecting the proliferation of neural progenitors and the generation of an adequate pool of neurons in the developing brain 36 ., The expression pattern of TRMT10A in the ventricular zone of the developing cortex is consistent with its influence on neural progenitor properties , including control of survival that is known to affect brain size ., In addition it may act in subsets of differentiated neurons , as suggested by its expression in cortical marginal zone and cerebellum ., Early onset diabetes has been associated with microcephaly in other genetic disorders ., Homozygous mutations in the IER3IP1 gene encoding the immediate and early response 3 interacting protein 1 result in infantile diabetes and congenital microcephaly with simplified gyration , hypotonia , intractable seizures , and early death 37 , 38 ., Cases of microcephaly with severe neurological expression were also described in Wolcott-Rallison syndrome , which includes permanent neonatal diabetes , multiple epiphyseal dysplasia , osteoporosis and liver dysfunction ., This syndrome is due to biallelic mutations in EIF2AK3 encoding translation initiation factor 2-α kinase-3 39 ., EIF2AK3 is activated upon the accumulation of unfolded proteins in the ER and inhibits protein translation initiation 40 ., Other human diseases are caused by mutations in genes encoding tRNAs and tRNA modifying enzymes ., Pontocerebellar hypoplasia , characterized by hypoplasia and atrophy of ventral pons , cerebellum and the cerebral cortex , is caused by mutations in genes encoding tRNA splicing endonuclease subunits ( TSEN ) or mitochondrial arginyl-tRNA synthetase ( RARS2 ) 41 ., Mutations in mitochondrial tRNA genes and in aminoacyl-tRNA synthetases cause myopathies and neurodegenerative diseases , sometimes in association with diabetes ., Recently , a syndrome of mental retardation , microcephaly and short stature was described , caused by mutations in NSUN2 , encoding a methyltransferase that catalyzes the intron-dependent formation of 5-methylcytosine at C34 of tRNA-leu ( CAA ) 42 , 43 ., NSUN2 is the ortholog of yeast TRM4 ., Wild-type NSUN2 localized to the nucleolus , whereas mutant NSUN2 accumulated in the nucleoplasm and cytoplasm 42; other NSUN2 mutations resulted in nonsense-mediated mRNA decay 43 ., Inactivation of the X-linked gene FTSJ1 , another RNA methyltransferase and ortholog of yeast TRM7 , gives rise to non-syndromic intellectual disability 44 ., In addition to causing microcephaly and short stature , the TRMT10A mutation causes a severe form of diabetes , which was not reported for these other RNA methyltransferase mutations ., This may be related to cell-specific requirements of RNA modifications ., It is of particular interest that CDKAL1 polymorphisms predispose to insulin secretion defects and T2D 8 ., CDKAL1 was recently shown to methylthiolate tRNALys ( UUU ) 45 ., The β-cell-specific Cdkal1 knockout mouse develops impaired glucose tolerance , due to misreading of Lys codons in proinsulin , defective insulin biosynthesis and increased susceptibility to ER stress and high fat diet 9 ., In conclusion , we describe a nonsense mutation in the TRMT10A gene in a new syndrome of young onset diabetes and microcephaly ., Based on its cellular localization and by homology with its yeast counterpart , we propose that TRMT10A has methyltransferase activity ., We show that TRMT10A is expressed in human fetal brain; TRMT10A silencing does not impair β-cell function but induces apoptosis , suggesting that TRMT10A deficiency may negatively affect β-cell mass and the pool of neurons in the developing brain ., Our findings may have broader relevance for the understanding of the pathogenesis of T2D and microcephaly ., The ethics committee of the Erasmus Hospital , Université Libre de Bruxelles approved of the study ., The three patients , their parents , and two unaffected siblings gave informed consent ., Human fetal brain was collected and used according to the guidelines of the local ethics committees on research involving human subjects ( Erasmus Hospital , Université Libre de Bruxelles and Belgian National Fund for Scientific Research ) ., Adult male Wistar rats were housed and used following the rules of the Belgian Regulations for Animal Care , with approval of the ethics committee of the Université Libre de Bruxelles ., A peripheral blood sample was obtained for genetic analysis from the three patients , their parents , and two unaffected siblings ., Leukocyte DNA was extracted using proteinase K digestion followed by phenol-chloroform extraction and ethanol precipitation 46 and samples were stored at 4°C in T10E1 buffer ., We used Affymetrix 11K-GeneChip microarrays representing 10 , 000 autosomal single nucleotide polymorphisms ( Affymetrix , High Wycombe , United Kingdom ) to genotype the three patients DNA ( 500 ng each ) on an Affymetrix platform following the instructions of the manufacturer ., Regions of homozygosity were delineated using the ExcludeAR algorithm 47 ., In chromosomal regions with apparent homozygosity by descent , microsatellites were genotyped in individual subjects ., Marker order was obtained from the University of California at Santa Cruz ( UCSC ) physical map ( http://genome . ucsc . edu/cgi-bin/hgGateway ) ., A multipoint LOD score was computed using the MAPMAKER/HOMOZ software 48 assuming a gene frequency of 0 . 005 and marker allele frequencies as observed in a series of control subjects , with a minimal minor allele frequency of 0 . 10 ., Genomic DNA from the proband ( Figure 1 , arrow ) was sonicated and enriched for exonic sequences by hybridization on an Agilent SureSelect All Exon v1 capture kit ., Exon-enriched DNA was paired-end sequenced over 90 bp by an Illumina HiSeq2000 sequencer ( Beijing Genomics Institute ) ., An average of 55 . 6 million paired-end reads were filtered to eliminate reads with more than 6 undetermined nucleotides or 40 identical bases in tandem ., The filtered reads were then aligned to the human genome GRCh36 assembly using the SOAPaligner 2 . 20 software 49 and the genotypes were called using the SOAPsnp program 50 ., Resulting single nucleotide variants ( SNVs ) were filtered according to the following rules: base quality larger than 20 , read depth equal to or larger than 4 , and a distance between two variants larger than 4 ., Insertions and deletions were identified separately , through alignment to GRCh36 using the Burrows-Wheeler alignment tool 51 , and detection using the Genome Analysis Toolkit 52 ., SNVs and indels were annotated using the Ensembl V54 database ., We considered SNVs and indels that were not found in the dbSNP135 database , nor in the Thousand Genome ( www . 1000genomes . org ) database , nor in the Exome Variant Server ( http://evs . gs . washington . edu/EVS/ ) , and that were not found in our other in-house exome sequencing results ., PCR primers for all exons and flanking intronic sequences were designed using the Exonprimer software ( http://ihg . helmholtz-muenchen . de/ihg/ExonPrimer . html ) ., All exons and flanking intronic regions of the candidate genes were sequenced by the Sanger method using the Big Dye Terminator cycle sequencing kit v2 ( Applied Biosystems , Foster City , California , USA ) , and analyzed on a 3130 Genetic Analyser sequencing machine ( Applied Biosystems ) ., Sequences were analyzed in silico for mutations using the SeqScape software V . 2 . 0 ., ( Applied Biosystems ) ., In situ hybridization was done on human fetal brain ( GW 8 , 11 , 17 , 19 ) as previously described 53 ., Riboprobe template was generated by PCR using TRMT10A specific pairs of primers: F: CCAAGCTAATACGACTCACTATAGGGAGATGTGAACCAATATCTAAACGACAAA – R: GGATCCATTAACCCTCACTAAAGGGAGAGATTTTCCTTATCCTGCTTTTCTTC ., Clonal rat INS-1E cells ( a kind gift from Dr C Wollheim , Centre Médical Universitaire , Geneva , Switzerland ) were cultured in RPMI medium as previously described 54 , 55 ., Tissues were obtained from adult male Wistar rats ( Charles River Laboratories ) ., Rat islets were isolated by collagenase digestion followed by hand picking under a stereomicroscope ., Islets were dispersed and β-cells purified by autofluorescence-activated cell sorting ( FACS , FACSAria , BD Bioscience ) and cultured as described 56 , 57 ., Human islets from non-diabetic organ donors ( n\u200a=\u200a13 , age 68±4 years , BMI 27±1 kg/m2 ) were isolated by collagenase digestion and density gradient purification 58 ., The islets were cultured , dispersed and transfected as previously described 59 ., The mean percentage of β-cells of the human islet preparations was 50±5% , as determined by insulin immunofluorescence 25 , 60 ., Human lymphoblasts from three control individuals , two patients and one heterozygous carrier of the mutation were cultured in RPMI 1640 medium supplemented with 20% FBS , 100 mU/ml penicillin and 100 mU/ml streptomycin ., hTRMT10A was amplified by PCR from lymphoblast cDNA using oligonucleotides spanning the start and stop codons of the TRMT10A open reading frame ( ORF ) , using primers F CGGAATTCATGTCATCTGAAATGTTGCC and R CGCTCGAGGTGTGGCAGAGAGTTCACTG ., The restriction sites EcoRI and XhoI ( underlined ) were added to facilitate the directional cloning into the expression vector pGEX-6P-1 ( GE Healthcare ) ., This vector allows the expression of recombinant proteins fused to glutathione-s-transferase ( GST ) at its N-terminus ., E . coli BL21 cells were transformed with the pGEX-6P-1-TRMT10A plasmid by electroporation ., Positive clones were selected by colony PCR and sequenced ., For recombinant protein expression , a single colony was grown overnight at 37°C in LB medium containing 100 µg/ml ampicillin ., Cells were then diluted 1∶50 in the same medium and grown at 37°C until an optical density of 0 . 6 at 600 nm was reached ., Isopropyl-β-D-thiogalactoside ( 0 . 25 mM ) was then added and cells were grown at 28°C for 3 h to induce recombinant protein expression ., Cells were harvested by centrifugation at 3000×g for 10 min , lysed by sonication in 20 mM Tris buffer pH 8 containing 0 . 5% Triton ×100 , 10 mM dithiothreitol , 0 . 1 mM PMSF and protease inhibitor cocktail ( Roche ) , and centrifuged for 15 min at 15 , 000×g at 4°C ., The supernatant was applied to 1 ml glutathione spin columns ( Pierce ) and washed with ice-cold lysis buffer ., The recombinant hTRMT10A was separated from the GST moiety by in column site-specific proteolysis using PreScission protease ( GE Healthcare ) following the manufacturers instructions ., The purified recombinant hTRMT10A was used for rabbit pol | Introduction, Results, Discussion, Materials and Methods | We describe a new syndrome of young onset diabetes , short stature and microcephaly with intellectual disability in a large consanguineous family with three affected children ., Linkage analysis and whole exome sequencing were used to identify the causal nonsense mutation , which changed an arginine codon into a stop at position 127 of the tRNA methyltransferase homolog gene TRMT10A ( also called RG9MTD2 ) ., TRMT10A mRNA and protein were absent in lymphoblasts from the affected siblings ., TRMT10A is ubiquitously expressed but enriched in brain and pancreatic islets , consistent with the tissues affected in this syndrome ., In situ hybridization studies showed that TRMT10A is expressed in human embryonic and fetal brain ., TRMT10A is the mammalian ortholog of S . cerevisiae TRM10 , previously shown to catalyze the methylation of guanine 9 ( m1G9 ) in several tRNAs ., Consistent with this putative function , in silico topology prediction indicated that TRMT10A has predominant nuclear localization , which we experimentally confirmed by immunofluorescence and confocal microscopy ., TRMT10A localizes to the nucleolus of β- and non-β-cells , where tRNA modifications occur ., TRMT10A silencing induces rat and human β-cell apoptosis ., Taken together , we propose that TRMT10A deficiency negatively affects β-cell mass and the pool of neurons in the developing brain ., This is the first study describing the impact of TRMT10A deficiency in mammals , highlighting a role in the pathogenesis of microcephaly and early onset diabetes ., In light of the recent report that the type 2 diabetes candidate gene CDKAL1 is a tRNA methylthiotransferase , the findings in this family suggest broader relevance of tRNA methyltransferases in the pathogenesis of type 2 diabetes . | The inherited predisposition to type 2 diabetes is attributed to common variants in over 60 loci ., Among these risk variants is CDKAL1 , which has recently been shown to be a tRNA modifying enzyme ( methylthiotransferase ) ., Genetic variants of different severity can generate a spectrum of monogenic and polygenic forms of diabetes ., Here we describe a new syndrome of young onset diabetes , short stature and microcephaly ( small brain size ) with intellectual disability in a large consanguineous family ., By linkage analysis and whole exome sequencing we identified a nonsense mutation in TRMT10A , a gene that has hitherto not been studied in mammals ., The yeast homolog TRM10 has been shown to be a tRNA modifying enzyme with methyltransferase activity ., We demonstrate that TRMT10A mRNA and protein are absent in cells from the affected siblings ., TRMT10A localizes to the nucleolus , where tRNA modifications occur ., TRMT10A silencing induces cell death in insulin-producing pancreatic β-cells , suggesting that TRMT10A deficiency may reduce β-cell mass and the pool of neurons in the brain ., This is the first study describing the impact of TRMT10A deficiency in man ., Our findings may have broader relevance for the understanding of the pathogenesis of type 2 diabetes and microcephaly . | null | null |
journal.pcbi.1006434 | 2,018 | A k-mer-based method for the identification of phenotype-associated genomic biomarkers and predicting phenotypes of sequenced bacteria | The falling cost of sequencing has made genome sequencing affordable to a large number of labs , and therefore , there has been a dramatic increase in the number of genome sequences available for comparison in the public domain 1 ., These developments have facilitated the genomic analysis of bacterial isolates ., An increasing amount of bacterial whole genome sequencing ( WGS ) data has led to more and more genome-wide studies of DNA variation related to different phenotypes ., Among these studies , antibiotic resistance phenotypes are the most concerning and have garnered high public interest , especially since several multidrug-resistant strains have emerged worldwide ., The detection of known resistance-causing mutations as well as the search for new candidate biomarkers leading to resistance phenotypes requires reasonably rapid and easily applicable tools for processing and comparing the sequencing data of hundreds of isolated strains ., However , there is still a lack of user-friendly software tools for the identification of genomic biomarkers from large sequencing datasets of bacterial isolates 2 , 3 ., While microbial genome-wide association studies ( GWAS ) can be successfully used in case of previously known genotype-phenotype associations caused by a single gene or only a set of few and specific mutations , more complex and novel associations would remain undetected ., In addition , many bacterial species have extensive intra-species variation from small sequence-based differences to the absence or presence of whole genes or gene clusters ., Choosing only one genome as a reference for searching for the variable components would be highly limiting ., Alternative approaches use previously detected genomic features , either single nucleotide variations or longer sequences , behind the phenotype to create and train models using those features as the predictors ., Not only antibiotic resistance but wide range of other phenotypes can be predicted , e . g host adaptation in invasive serovars 4 , needed minimum inhibitory concentrations of antibiotics 5 or virulence of the strains 6 ., Using longer sequence regions , such as full genes in those models , requires assembled genomes as an input which adds data preprocessing step ., The solution to avoid this is using k-mers , which are short DNA oligomers with length k , that enable us to simultaneously discover a large set of single nucleotide variations , insertions and deletions associated with the phenotypes under study ., The advantage of using k-mer-based methods in genomic biomarker discovery is that they do not require sequence alignments and can even be applied to raw sequencing data ., In recent years several publications using different machine learning algorithms and k-mers for detecting the biomarkers behind different bacterial phenotypes have been published ., Among the latest , short k-mers and machine learning ( ML ) has been used to create minimum inhibitory concentration prediction models in assembled Klebsiella pneumoniae genomes for several antibiotics 7 ., PATRIC and RAST annotation services include prediction of antimicrobial resistance with the species- and antibiotic-specific classifier k-mers which are selected using publicly available and collected metadata and the adaptive boosting ML algorithms 8 ., Though providing a framework or predictive models for a specific species with a certain phenotype , those studies have not been concentrating on the creation of a software easily applicable by a wider public without an access to extensive computing resources but still having the need for analyzing large scale bacterial genome sequencing data with a reasonable amount of computing time ., Only few papers describe software which we were able to compare with PhenotypeSeeker ., The SEER program takes either a discrete or continuous phenotype as an input , counts variable-length k-mers and corrects for the clonal population structure 6 ., SEER is a complex pipeline requiring several separate steps for the user to execute and currently has many system-level dependencies for successful compilation and installation ., Another similar tool , Kover , handles only discrete phenotypes , counts user-defined size k-mers and does not use any correction for population structure 9 ., The Neptune software targets so-called signatures differentiating two groups of sequences but cannot locate smaller mutations , such as single isolated nucleotide variations , being the reason , it was not used in the comparison in current paper ., The signatures that Neptune detects are relatively large genomic loci , which may include genomic islands , phage regions or operons 10 ., We created PhenotypeSeeker as we observed the need for a tool that could combine all the benefits of the programs available but at the same time would be easily executable and would take a reasonable amount of computing resources without the need for dedicated high-performance computer hardware ., PhenotypeSeeker consist of two subprograms: PhenotypeSeeker modeling and PhenotypeSeeker prediction ., PhenotypeSeeker modeling takes either assembled contigs or raw-read data as an input and builds a statistical model for phenotype prediction ., The method starts with counting all possible k-mers from the input genomes , using the GenomeTester4 software package 11 , followed by k-mer filtering by their frequency in strains ., Subsequently , the k-mer selection for regression analysis is performed ., In this step , to test the k-mers’ association with the phenotype , the method applies Welch’s two-sample t-test if the phenotype is continuous and a chi-squared test if it is binary ., Finally , the logistic regression or linear regression model is built ., The PhenotypeSeeker output gives the regression model in a binary format and three text files , which include the following: ( 1 ) the results of association tests for identifying the k-mers most strongly associated with the given phenotype , ( 2 ) the coefficients of k-mers in the regression model for identifying the k-mers that have the greatest effects on the outcomes of the machine learning model , ( 3 ) a FASTA file with phenotype-specific k-mers , assembled to longer contigs when possible , to facilitate an user to perform annotation process , and ( 4 ) a summary of the regression analysis performed ( Fig 1 ) ., Optionally , it is possible to use weighting for the strains to take into account the clonal population structure ., The weights are based on a distance matrix of strains made with an alignment-free k-mer-based method called Mash 12 ., The weights of each genome are calculated using the Gerstein , Sonnhammer and Cothia method 13 ., PhenotypeSeeker prediction uses the regression model generated by PhenotypeSeeker modeling to conduct fast phenotype predictions on input samples ( Fig 1 ) ., Using gmer_counter from the FastGT package 14 , the tool searches the samples only for the k-mers used as parameters in the regression model ., Predictions are then made based on the presence or absence of these k-mers ., PhenotypeSeeker uses fixed-length k-mers in all analyses ., Thus , the k-mer length is an important factor influencing the overall software performance ., The effects of k-mer length on speed , memory usage and accuracy were tested on a P . aeruginosa ciprofloxacin dataset ., A general observation from that analysis is that the CPU time and the PhenotypeSeeker memory usage increase when the k-mer length increases ( Fig 2 ) ., Previously described mutations in the P . aeruginosa parC and gyrA genes were always detected if the k-mer length was at least 13 nucleotides ., We assume that in most cases , a k-mer length of 13 is sufficient to detect biologically relevant mutations , although in certain cases , longer k-mers might provide additional sensitivity ., The k-mer length in PhenotypeSeeker is a user-selectable parameter ., Although most of the phenotype detection can be performed with the default k-mer value , we suggest experimenting with longer k-mers in the model building phase ., All subsequent analyses in this article are performed with a k-mer length of 13 , unless specified otherwise ., PhenotypeSeeker was applied to the dataset composed of P . aeruginosa genomes and corresponding ciprofloxacin resistance values measured in terms of minimum inhibitory concentration ( MIC ) ( μg/ml ) , which is defined as the lowest concentration of antibiotic that will inhibit the visible growth of the isolate under investigation after an appropriate period of incubation 15 ., We built two separate models using a continuous phenotype for one and binary phenotype for another ., Binary phenotype values were created based on EUCAST ciprofloxacin breakpoints 16 ., Both models detected k-mers associated with mutations in quinolone resistance determining regions ( QRDR ) of the parC ( c . 260C>T , p . Ser87Leu ) and gyrA ( c . 248C>T , p . Thr83Ile ) genes ( Fig 3 , S2 Table ) ., These genes encode DNA topoisomerase IV subunit A and DNA gyrase subunit A , the target proteins of ciprofloxacin 17 ., Mutations in the QRDR regions of these genes are well-known causes of decreased sensitivity to quinolone antibiotics , such as ciprofloxacin 18 ., The classification model built using a binary phenotype had a F1-measure of 0 . 88 , prediction accuracy of 0 . 88 , sensitivity of 0 . 90 and specificity of 0 . 87 on the test subset ( Table A in S3 Table ) ., The MIC prediction model built using the continuous phenotype had the coefficient of determination ( R2 ) of 0 . 42 , the Pearson correlation coefficient of 0 . 68 and the Spearman correlation coefficient of 0 . 84 ( Table M in S3 Table ) ., In addition to the P . aeruginosa dataset , we tested a C . difficile azithromycin resistance dataset ( S2 Table ) studied using Kover in Drouin et al . , 2016 9 ., ermB and Tn6110 transposon were the sequences known and predicted to be important in an azithromycin resistance model by Kover 9 ., ermB was not located on the transposon Tn6110 ., PhenotypeSeeker found k-mers for both sequences while using k-mers of length 13 or 16 ., Tn6110 is a transposon that is over 58 kbp long and contains several protein coding sequences , including 23S rRNA methyltransferase , which is associated with macrolide resistance 19 ., The predictive models with all tested k-mer lengths ( 13 , 16 and 18 ) contained k-mers covering the entire Tn6110 transposon sequence , both in protein coding and non-coding regions ., In addition to the 23S rRNA methyltransferase gene , k-mers in all three models were mapped to the recombinase family protein , sensor histidine kinase , ABC transporter permease , TlpA family protein disulfide reductase , endonuclease , helicase and conjugal transfer protein coding regions ., The model built for the C . difficile azithromycin resistance phenotype had a F1-measure of 0 . 97 , prediction accuracy of 0 . 97 , sensitivity of 0 . 96 and specificity of 0 . 97 on the test subset ( Table A in S3 Table ) ., In addition to antibiotic resistance phenotypes in P . aeruginosa and C . difficile , we used K . pneumoniae human infection-causing strains as a different kind of phenotype example ., K . pneumoniae strains contain several genetic loci that are related to virulence ., These loci include aerobactin , yersiniabactin , colibactin , salmochelin and microcin siderophore system gene clusters 20–24 , the allantoinase gene cluster 25 , rmpA and rmpA2 regulators 26 , 27 , the ferric uptake operon kfuABC 28 and the two-component regulator kvgAS 29 ., The model predicted by PhenotypeSeeker for invasive/infectious phenotypes included 13-mers representing several of these genes ., Genes in colibactin ( clbQ and clbO ) , aerobactin ( iucB and iucC ) and yersiniabactin ( irp1 , irp2 , fyuA , ybtQ , ybtX , and ybtP ) clusters showed the most differentiating pattern between carrier and invasive/infectious strains ( Fig 4; S2 Table ) ., A 13-mer mapping to a gene-coding capsule assembly protein Wzi was also represented in the model ., The model built for K . pneumoniae invasive/infectious phenotypes had a F1-measure of 0 . 88 , prediction accuracy of 0 . 88 , sensitivity of 0 . 91 and specificity of 0 . 78 on the test subset ( Table A in S3 Table ) ., To measure the average classification accuracies of logistic regression models , all three datasets were divided into a training and test set of approximately 75% and 25% of strains respectively ., A K-mer length of 13 was used , and a weighted approach was tested on binary phenotypes ( Table 1 ) ., To reduce the influence of sequencing errors when using sequencing reads instead of assembled contigs as the input , we only counted 13-mers as being present in one of the input lists if they occurred at least 5 times in that input list ., The PhenotypeSeeker prediction accuracy is not lower when using raw sequencing reads instead of assembled genomes , and therefore , assembly building is not required before model building ., Our results with K . pneumoniae show that PhenotypeSeeker can be successfully applied to other kinds of phenotypes in addition to antibiotic resistance ., In our trials , the model building on a given dataset took 3 to 5 hours per phenotype , and prediction of the phenotype took less than a second on assembled genomes ( Table 1 ) ., The CPU time of model building by PhenotypeSeeker depends mainly on the number of different k-mers in genomes of the training set ., The analysis performed on our 200 P . aeruginosa genomes showed that the CPU time of the model building grows linearly with the number of genomes given as input ( S1 Fig ) ., The memory requirement of PhenotypeSeeker did not exceed 2 GB if default parameter settings are used , allowing us to run analyses on laptop computers ( S2 Fig ) if necessary ., The p-value cut-offs during the k-mer filtering step influence the number of k-mers included in the model and have a potentially strong impact on model performance ., Tables A-E in the S1 Table show the effects of different p-value cut-offs on model performances ., We ran SEER and Kover on the same P . aeruginosa ciprofloxacin dataset and C . difficile azithromycin resistance dataset to compare the efficiency and CPU time usage with PhenotypeSeeker ., In the P . aeruginosa dataset , SEER was able to detect gyrA and parC mutations only when resistance was defined as a binary phenotype ., In cases with a continuous phenotype , those k-mers did not pass the p-value filtering step ., Since Kovers aim is to create a resistance predicting model , not an exhaustive list of significant k-mers , it was expected that not all the mutations would be described in the output ., gyrA variation already sufficiently characterized the resistant strains set , and therefore , parC mutations were not included in the model ., The same applies to the PhenotypeSeeker results with 16- and 18-mers ., parC-specific 16- or 18-mers were included among the 1000 k-mers in the prediction model ( based on statistically significant p-values ) but with the regression coefficient equal to zero because they were present in the same strains as gyrA specific predictive k-mers ., In the C . difficile dataset , our model included the known resistance gene ermB and transposon Tn6110 ., We were able to find ermB with both SEER and Kover ., We also detected Tn6110-specific k-mers with SEER while running Kover with 16-mers instead of 31-mers as in the default settings ., Regarding the CPU time , PhenotypeSeeker with 13-mers was faster than other tested software programs ( 3 . 5 hrs vs 14–15 hrs ) without losing the relevant markers in the output ( Table 2 ) ., Using 16- or 18-mers , the PhenotypeSeeker’s running time increases but is still lower than with SEER and Kover ., PhenotypeSeeker works as an easy-to-use application to list the candidate biomarkers behind a studied bacterial phenotype and to create a predictive model ., Based on k-mers , PhenotypeSeeker does not require a reference genome and is therefore also usable for species with very high intraspecific variation where the selection of one genome as a reference can be complicated ., PhenotypeSeeker supports both discrete and continuous phenotypes as inputs ., In addition , this model takes into account the population structure to highlight only the possible causal variations and not the mutations arising from the clonal nature of bacterial populations ., Unlike Kover , the PhenotypeSeeker output is not merely a trained model for predicting resistance in a separate set of isolates , but the complete list of statistically significant candidate variations separating antibiotic resistant and susceptible isolates for further biological interpretation is also provided ., Unlike SEER , PhenotypeSeeker is easier to install and can be run with only a single command for building a model and another single command to use it for prediction ., Our tests using PhenotypeSeeker to detect antibiotic resistance markers in P . aeruginosa and C . difficile showed that it is capable of detecting all previously known mutations in a reasonable amount of time and with a relatively short k-mer length ., Users can choose the k-mer length as well as decide whether to use the population structure correction step ., Due to the clonal nature of bacterial populations , this step is highly advised for detecting genuine causal variations instead of strain-level differences ., In addition to a trained predictive model , the list of k-mers covering possible variations related to the phenotype are produced for further interpretation by the user ., The effectiveness of the model can vary because of the nature of different phenotypes in different bacterial species ., Simple forms of antibiotic resistance that are unambiguously determined by one or two specific mutations or the insertion of a gene are likely to be successfully detected by our method , and effective predictive models for subsequent phenotype predictions can be created ., This is supported by our prediction accuracy over 96% in the C . difficile dataset ., On the other hand , P . aeruginosa antibiotic resistance is one of the most complicated phenotypes among clinically relevant pathogens since it is not often easily described by certain single nucleotide mutations in one gene but rather through a complex system involving several genes and their regulators leading to multi-resistant strains ., In cases such as this , the prediction is less accurate ( 88% in our dataset ) , but nevertheless , a complete list of k-mers covering differentiating markers between resistant and sensitive strains can provide more insight into the actual resistance mechanisms and provide candidates for further experimental testing ., Tests with K . pneumoniae virulence phenotypes showed that PhenotypeSeeker is not limited to antibiotic resistance phenotypes but is potentially applicable to other measurable phenotypes as well and is therefore usable in a wider range of studies ., Since PhenotypeSeeker input is not restricted to assembled genomes , one can skip the assembly step and calculate models based on raw read data ., In this case , it should be taken into account that sequencing errors may randomly generate phenotype-specific k-mers; thus , we suggest using the built-in option to remove low frequency k-mers ., The k-mer frequency cut-off threshold depends on the sequencing coverage of the genomes and is therefore implemented as user-selectable ., One can also build the model based on high-quality assembled genomes and then use the model for corresponding phenotype prediction on raw sequencing data ., PhenotypeSeeker was tested on the following three bacterial species: Pseudomonas aeruginosa , Clostridium difficile and Klebsiella pneumoniae ., The P . aeruginosa dataset was composed of 200 assembled genomes and the minimal inhibitory concentration measurements ( MICs ) for ciprofloxacin ., The P . aeruginosa strains were isolated during the project Transfer routes of antibiotic resistance ( ABRESIST ) performed as part of the Estonian Health Promotion Research Programme ( TerVE ) implemented by the Estonian Research Council , the Ministry of Agriculture ( now the Ministry of Rural Affairs ) , and the National Institute for Health Development ., Isolated strains originated from humans , animals and the environment within the same geographical location in Estonia and belonged to 103 different MLST sequence types ( Laht et al . , Pseudomonas aeruginosa distribution among humans , animals and the environment ( submitted ) ; Telling et al . , Multidrug resistant Pseudomonas aeruginosa in Estonian hospitals ( submitted ) ) ., Full genomes were sequenced by Illumina HiSeq2500 ( Illumina , San Diego , USA ) with paired-end , 150 bp reads ( Nextera XT libraries ) and de novo assembled with the program SPAdes ( ver 3 . 5 . 0 ) 30 ., MICs were determined by using the epsilometer test ( E-test , bioMérieux , Marcy lEtoile , France ) according to the manufacturer instructions ., Binary phenotypes were achieved by converting the MIC values into 0 ( sensitive ) and 1 ( resistant ) phenotypes according to the European Committee on Antimicrobial Susceptibility Testing ( EUCAST ) breakpoints 16 ., The resulted dataset consisted of 124 ciprofloxacin sensitive P . aeruginosa isolates ( 62% ) and 76 ciprofloxacin resistant P . aeruginosa isolates ( 38% ) and is deposited in the NCBI’s BioProject database under the accession number PRJNA244279 ( https://www . ncbi . nlm . nih . gov/bioproject/ ? term=PRJNA244279 ) ., The C . difficile dataset was composed of assembled genomes of 459 isolates and the binary phenotypes of azithromycin resistance ( sensitive = 0 vs resistant = 1 ) , adapted from Drouin et al . , 2016 9 ., The isolates originated from patients from different hospitals in the province of Quebec , Canada and the genomes were received from the European Nucleotide Archive EMBL:PRJEB11776 ( ( http://www . ebi . ac . uk/ena/data/view/PRJEB11776 ) ., The dataset consisted of 246 azithromycin sensitive isolates ( 54% ) and 213 azithromycin resistant isolates ( 46% ) ., The K . pneumoniae dataset was composed of reads of 167 isolates , originating from six countries and sampled to maximize diversity , and the binary clinical phenotype of human carriage status vs human infection ( including invasive infections ) status ( carriage = 0 vs infectious = 1 ) , adapted from Holt et al . , 2015 31 ., The reads were received from the European Nucleotide Archive EMBL:PRJEB2111 ( https://www . ebi . ac . uk/ena/data/view/PRJEB2111 ) and de novo assembled with SPAdes ( ver 3 . 10 . 1 ) 30 ., The dataset consisted of 36 isolates with human carriage status as phenotype ( 22% ) and 131 K . pneumonia isolates with human infection status as phenotype ( 78% ) ., Abstractly , each test dataset was composed of pairs ( x , y ) , where x is the bacterial genome x∈{A , T , G , C}* , and y denotes phenotype values specific to a given dataset y ∈ {0 . 008 , … , 1024} ( continuous phenotype ) or y ∈ {0 , 1} ( binary phenotype ) ., All operations with k-mers are performed using the GenomeTester4 software package containing the glistmaker , glistquery and glistcompare programs 11 ., At first , all k-mers from all samples are counted with glistmaker , which takes either FASTA or FASTQ files as an input and enables us to set the k-mer length up to 32 nucleotides ., Subsequently , the k-mers are filtered based on their frequency in strains of the training set ., By default , the k-mers that are present in or missing from less than two samples are filtered out and not used in building the model ., The remaining k-mers are used in statistical testing for detection of association with the phenotype ., By default , PhenotypeSeeker conducts the clonal population structure correction step by using a sequence weighting approach that reduces the weight of isolates with closely related genomes ., For weighting , pairwise distances between genomes of the training set are calculated using the free alignment software Mash with default parameters ( k-mer size of 21 nucleotides and sketch size of 1000 min-hasehes ) 12 ., Distances estimated by Mash are subsequently used to calculate weights for each genome according to the algorithm proposed by Gerstein , Sonnhammer and Chothia 13 ., The calculation of GSC weights is conducted using the PyCogent python package 32 ., The GSC weights are taken into account while calculating Welch two-sample t-tests or chi-squared tests to test the k-mers’ associations with the phenotype ., Additionally , the GSC weights can be used in the final logistic regression or linear regression ( if Ridge regularization is used ) model generation ., In the case of binary phenotype input , the chi-squared test is applied to every k-mer that passes the frequency filtration to determine the k-mer association with phenotype ., The null hypothesis assumes that there is no association between k-mer presence and phenotype ., The alternative hypothesis assumes that the k-mer is associated with phenotype ., The chi-squared test is conducted on these observed and expected values with degrees of freedom = 1 , using the scipy . stats Python package 33 ., If the user selects to use the population structure correction step , then the weighted chi-squared tests are conducted according to the previously published method 34 ., In the case of continuous phenotype input , the Welch two-sample t-test is applied to every k-mer that passes the frequency filtration to determine if the mean phenotype values of strains having the k-mer are different from the mean phenotype values of strains that do not have the k-mer ., The null hypothesis assumes that the strains with a k-mer have different mean phenotype values from the strains without the k-mer ., The alternative hypothesis assumes that the means of the strains with and without the k-mer are the same ., The t-test is conducted with these values using the scipy . stats Python package 33 , assuming that the samples are independent and have different variance ., If the user selects the population structure correction step , then the weighted t-tests are conducted 34 ., In that case , the p-value is calculated with the function scipy . stats . t ., sf , which takes the absolute value of the t-statistic and the value of degrees of freedom as the input ., To perform the regression analysis , first , the matrix of samples times features is created ., The samples in this matrix are strains given as the input and the features represent the k-mers that are selected for the regression analysis ., The values ( 0 or, 1 ) in this matrix represent the presence or absence of a specific k-mer in the specific strain ., The target variables of this regression analysis are the resistance values of the strains ., Thereupon , input data are divided into training and test sets whose sizes are by default 75% and 25% of the strains , respectively ., The proportion of class labels in the training and test sets are kept the same as in the original undivided dataset ., In the case of a continuous phenotype , a linear regression model is built , and in the case of a binary phenotype , a logistic regression model is built ., The logistic regression was selected for binary classification task as it showed better performance on our datasets than other tested machine learning classifiers like support vector machine ( with no kernel and with Gaussian kernel ) and random forest ., The performance of logistic regression models on our tested datasets in comparison to performance of other machine learning classifiers are shown in S3 Fig and in Tables A-L in S3 Table ., The performance of linear regression model on P . aeruginosa dataset is shown in Table M in S3 Table ., For both the linear and logistic regression , the Lasso , Ridge or Elastic Net regularization can be selected ., The Lasso and Elastic Net regularizations shrink the coefficients of non-relevant features to zero , which simplifies the identification of k-mers that have the strongest association with the phenotype ., To enable the evaluation of the output regression model , PhenotypeSeeker provides model-evaluation metrics ., For the logistic regression model quality , PhenotypeSeeker provides the mean accuracy as the percentage of correctly classified instances across both classes ( 0 and 1 ) ., Additionally , PhenotypeSeeker provides F1-score , precision , recall , sensitivity , specificity , AUC-ROC , average precision ( area under the precision-recall curve ) , Matthews correlation coefficient ( MCC ) , Cohen’s kappa , very major error rate and major error rate as metrics to assess model performance ., For the linear regression model , PhenotypeSeeker provides the mean squared error , the coefficient of determination ( R2 ) , the Pearson and the Spearman correlation coefficients and the within ±1 two-fold dilution factor accuracy ( useful for evaluating the MIC predictions ) as metrics to assess model performance ., To select for the best regularization parameter alpha , a k-fold cross-validation on the training data is performed ., By default , 25 alpha values spaced evenly on a log scale from 1E-6 to 1E6 are tested with 10-fold cross-validation and the model with the best mean accuracy ( logistic regression ) or with the best coefficient of determination ( linear regression ) is saved to the output file ., Regression analysis is conducted using the sklearn . linear_model Python package 35 ., Our models were created using mainly k-mer length 13 ( “-l 13”; default ) ., We counted the k-mers that occurred at least once per sample ( “-c 1”; default ) when the analysis was performed on contigs or at least five times per sample ( “-c 5” ) when the analysis was performed on raw reads ., In the first filtering step , we filtered out the k-mers that were present in or missing from less than two samples ( “—min 2—max 2”; default ) when the analysis was performed on a binary phenotype or fewer than ten samples ( “—min 10—max N-10”; N–total number of samples ) when the analysis was performed on a continuous phenotype ., In the next filtering step , we filtered out the k-mers with a statistical test p-value larger than 0 . 05 ( “—p_value 0 . 05”; default ) ., The regression analysis was performed with a maximum of 1000 lowest p-valued k-mers ( “—n_kmers; 1000”; default ) when the analysis was done with binary phenotype and with a maximum of 10 , 000 lowest p-valued k-mers ( “—n_kmers 10000”; default ) when the analysis was performed with a continuous phenotype ., For regression analyses , we split our datasets into training ( 75% ) and test ( 25% ) sets ( “-s 0 . 25”; default ) ., The regression analyses were conducted using Lasso regularization ( “-r L1”; default ) , and the best regularization parameter was picked from the 25 regularization parameters spaced evenly on a log scale from 1E-6 to 1E6 ( “—n_alphas 25—alpha_min 1E-6—alpha_max 1E6”; default ) ., The model performances with each regularization parameter were evaluated by cross-validation with 10-folds ( “—n_splits 10”; default ) ., The correction for clonal population structure ( “—weights +”; default ) and assembly of k-mers used in the regression model ( “—assembly +”; default ) were conducted in all our analyses ., SEER was installed and run on a local server with 32 CPU cores and 512 GB RAM , except the final step , which we were not able to finish without segmentation fault ., This last SEER step was launched via VirtualBox in ftp://ftp . sanger . ac . uk/pub/pathogens/pathogens-vm/pathogens-vm . latest . ova ., Both binary and continuous phenotypes were tested for P . aeruginosa and the binary phenotype in C . difficile cases ., Default settings were used ., Kover was installed on a local server and used with the settings suggested by the authors in the program tutorial . | Introduction, Results, Discussion, Methods | We have developed an easy-to-use and memory-efficient method called PhenotypeSeeker that, ( a ) identifies phenotype-specific k-mers ,, ( b ) generates a k-mer-based statistical model for predicting a given phenotype and, ( c ) predicts the phenotype from the sequencing data of a given bacterial isolate ., The method was validated on 167 Klebsiella pneumoniae isolates ( virulence ) , 200 Pseudomonas aeruginosa isolates ( ciprofloxacin resistance ) and 459 Clostridium difficile isolates ( azithromycin resistance ) ., The phenotype prediction models trained from these datasets obtained the F1-measure of 0 . 88 on the K . pneumoniae test set , 0 . 88 on the P . aeruginosa test set and 0 . 97 on the C . difficile test set ., The F1-measures were the same for assembled sequences and raw sequencing data; however , building the model from assembled genomes is significantly faster ., On these datasets , the model building on a mid-range Linux server takes approximately 3 to 5 hours per phenotype if assembled genomes are used and 10 hours per phenotype if raw sequencing data are used ., The phenotype prediction from assembled genomes takes less than one second per isolate ., Thus , PhenotypeSeeker should be well-suited for predicting phenotypes from large sequencing datasets ., PhenotypeSeeker is implemented in Python programming language , is open-source software and is available at GitHub ( https://github . com/bioinfo-ut/PhenotypeSeeker/ ) . | Predicting phenotypic properties of bacterial isolates from their genomic sequences has numerous potential applications ., A good example would be prediction of antimicrobial resistance and virulence phenotypes for use in medical diagnostics ., We have developed a method that is able to predict phenotypes of interest from the genomic sequence of the isolate within seconds ., The method uses a statistical model that can be trained automatically on isolates with known phenotype ., The method is implemented in Python programming language and can be run on low-end Linux server and/or on laptop computers . | sequencing techniques, medicine and health sciences, gut bacteria, pathology and laboratory medicine, statistics, pathogens, microbiology, pseudomonas aeruginosa, genome sequencing, antibiotic resistance, mathematics, forecasting, genome analysis, pharmacology, molecular biology techniques, klebsiella, bacteria, bacterial pathogens, research and analysis methods, clostridium difficile, pseudomonas, antimicrobial resistance, genomics, medical microbiology, mathematical and statistical techniques, microbial pathogens, molecular biology, gene identification and analysis, genetics, microbial control, biology and life sciences, klebsiella pneumoniae, physical sciences, computational biology, mutation detection, statistical methods, organisms | null |
journal.pntd.0004132 | 2,015 | Structural and Functional Characterization of the Enantiomers of the Antischistosomal Drug Oxamniquine | For more than 25 years , the mainstay of treatment for Schistosoma mansoni infections in Brazil was the drug oxamniquine ( OXA , ( RS ) -1 , 2 , 3 , 4-tetrahydro-2- isopropylaminomethyl-7-nitro-6-quinolylmethanol ) 1 , 2 ., OXA is species-specific , killing S . mansoni ( 67 million cases worldwide ) but not other schistosome species in Africa ( S . haematobium , 119 million cases ) or in SE Asia ( S . japonicum , 1 million cases ) 3 , 4 ., OXA is no longer manufactured because the drug praziquantel , which is effective against all schistosome species , is now available at a reasonable price due to the expiration of its patent ., The mode of action of OXA was recently elucidated5 ., As predicted by Pica-Mattoccia et al . 6 , OXA is a prodrug that is taken up by the parasite and sulfonated by an endogenous sulfotransferase ( SmSULT , GenBank AHB62207 . 1 , UniProt V9PWX8 ) in the presence of 3’phosphoadenosine 5’phosphosulfate ( PAPS ) ., The resulting sulfate ester of OXA is an unstable species that spontaneously decays to form a reactive electrophilic product capable of alkylating DNA , proteins and other macromolecules ., The ensuing disruption of synthetic and metabolic cellular functions eventually leads to parasite death5 , 7 ., OXA possesses one asymmetric carbon atom and its two enantiomers are both present in the marketed drug ( Fig 1 ) ., The structure of SmSULT with bound OXA and depleted co-factor PAP was determined previously at a resolution of 1 . 75 Å pdb code 4MUB , 5 ., Although the crystals were soaked with racemic OXA , the structure revealed only S-OXA in the central cavity of the L-shaped , predominantly α-helical enzyme with its hydroxyl group ( the target of sulfonation ) centered at the end of a shaft running from the surface of the molecule ., The relative positions of the accepting OXA and donating PAPS groups are entirely consistent with the formation of a sulfonated OXA hydroxyl group ( sulfate ester of OXA ) ., Previous studies have succeeded in separating the enantiomers , but have not provided information about their absolute configurations or relative antischistosomal properties 8–10 ., The question of which or if both enantiomers are active , and which or if both can occupy the binding pocket of SmSULT is addressed in this report ., Racemic OXA was a gift from Dr . D . Buggey ( Pfizer Ltd . ) ., The separation of OXA enantiomers was carried out by HPLC on a modified cellulose chiral stationary phase ( Chiracell OD-H , Daicel Chemical Industry ) 4 . 6 mm i . d . x 150 mm , eluted at 1 mL/min with a mixture of 5% isopropanol in n-hexane ( HPLC grade , Carlo Erba , Italy ) , added with 0 . 1% dimethylamine ( Aldrich , 99 . 5% purified by distillation ) ., The column consists of Cellulose tris ( 3 , 5-dimethylphenylcarbamate ) physically coated on microparticulate silica gel ., This chiral selector is versatile and it shows a particularly good selectivity towards aromatic compounds with substituents containing N or O atoms ., Dimethylamine was added to the eluent to prevent the ionization of OXA amino groups , but it was removed immediately after fraction collection , to avoid degradation of purified OXA isomers ., The chromatogram was obtained at 254 nm ., Fifty μL of a racemate solution ( 1 mg/mL ) in n-hexane:isopropanol ( 1:1 ) were injected into the column and the separated enantiomer peaks were collected ., The racemate solution was stored at –20°C ., In order to obtain sufficient material , the separation was repeated several times and the collected eluates were immediately evaporated to dryness under vacuum , in order to remove the dimethylamine ., The pool of evaporated fractions corresponding to the two enantiomers were re-dissolved in 1 mL of n-hexane:isopropanol and 10 μL analyzed by HPLC in the chromatographic system described above ., This allowed control of purity and quantitative determination by comparison with a calibration curve obtained with a known amount of racemate ., A solution of 0 . 58 mg/mL of OXA racemate was prepared in CH3OH:H2O ( 60:40 ) and then diluted with the running buffer ( 50 mM pH 3 phosphate ) to obtain a 35 μg/mL solution ., The CZE separation was performed as described by Abushoffa & Clark 8 ., The background electrolyte consisted of running buffer with 1 mM heparin as a chiral selector ., Separation was performed in a 62 cm , 75 μm i . d . capillary tube at 30°C , with an applied voltage of 20 kV ., Samples were hydrodynamically injected ( 50 mbar , 5 sec ) ., Fifty μL of enantiomer #1 ( i . e . the first eluting compound in HPLC ) was evaporated to dryness with a N2 flux and subsequently dissolved in 20 μL of 5% methanol in running buffer ., Ten μL of this enantiomer #1 solution was added to 600 μL of racemate ( 35 μg/mL ) and the mixture was analyzed by CZE under the same conditions described above for the racemate ., The sulfotransferase from S . mansoni ( SmSULT ) was crystallized as described previously 5 ., The depleted co-factor PAP ( 3’phosphoadenosine 5’phosphate ) was added to achieve a 4 to 1 stoichiometric ratio over protein and incubated for one hour on ice prior to crystallization ., Freshly grown crystals ( 4–8 days post-setup ) were soaked overnight in saturating conditions of each purified OXA enantiomer and flash-cooled in liquid nitrogen prior to data collection ., All diffraction data were measured at the Advanced Photon Source NE-CAT beamline 24-ID-C and integrated and scaled using the program XDS 11 ., Structures of the OXA enantiomer•SmSULT•PAP complexes were isomorphous with the published SmSULT structure ( Protein Data Bank entry 4MUA ) and the OXA enantiomers were built into difference electron density with coefficients mFo-DFc 12 ., Model coordinates were refined using the PHENIX program suite 13 , including simulated annealing with torsion angle dynamics , alternating with manual model adjustment using the program COOT 14 ., Figures depicting protein and OXA structure were created using the program PyMOL 15 ., Coordinates and structure factors have been deposited in the Protein Data Bank 16 under accession codes 5BYJ and 5BYK ., For the Institute of Cell Biology and Neurobiology , experimental protocols involving the use of animals were reviewed and approved by the Public Veterinary Health Department of the Italian Ministry of Health ( Authorization N . 25/2014-PR ) ., For the University of Texas Health Science Center ( S2 Fig legend ) , the use of animals in his study was approved by the University of Texas Health Science Center IACUC ( Protocol 11087x ) that adheres to the NIH Animal Care and Use guidelines ., A Puerto Rican strain of S . mansoni that has been maintained in the laboratory for several decades was used throughout this study ., An albino strain of Biomphalaria glabrata served as the intermediate host , while CD1 female albino mice ( Harlan , Italy ) were used for the mammalian stages ., Unisexual infections were obtained by exposing snails to a single miracidium and then sexing the emerging cercariae by PCR using female-specific W1 primers 17 ., Mice infected by tail immersion with 160 male cercariae were perfused ≥7 weeks later and the worms obtained were used for drug assays ., Eight to 13 male worms were distributed in tissue culture dishes ( 3 . 5 cm ) in Dulbecco modified Minimum Eagle’s Medium ( bicarbonate buffered ) supplemented with 10% fetal calf serum , 100 U/mL penicillin , 100 μg/mL streptomycin and 0 . 5 μg/mL amphotericin B . Cultures were kept at 37°C in an atmosphere of 5% CO2 in air and were observed daily under a Leica MZ12 . 5 stereomicroscope ., Male worms were used as they are more sensitive to the effects of oxamniquine then are female worms 18 ., Parasites were exposed to racemic OXA or its purified enantiomers for 30 min and subsequently washed three times and transferred to new dishes containing drug-free medium ., At the end of the observation period ( 2 weeks at high doses; 3 weeks at low doses ) , worms were classified on the basis of various vitality indicators , as: normal ( similar to untreated controls ) and assigned score 100; slow ( decreased motility and slight morphological changes ) , score 60; moribund ( only tiny movements , marked morphological changes , opaque appearance ) , score 30; dead ( no movement , severe morphological changes , dark appearance ) , score 0 ., The number of worms in each category was recorded ., The scores of all worms were added , divided by the number of worms present in the dish and reported as average scores ., The enantiomers of OXA were separated by semi-preparative HPLC on a chiral column and the results are illustrated in Fig 2A ., Two major peaks were present , practically with baseline separation , and were provisionally labeled as #1 and #2 ., The area under the curve was essentially the same for the two peaks , consistent with the enantiomers being present in equal amounts ., Since the quantity of compounds obtained from a single separation was limited , fractions #1 and #2 from several runs were pooled , respectively , and re-applied to the same column to check purity and to estimate quantity ., As shown in Fig 2B and 2C , the separated enantiomers were reasonably pure and their total amounts were estimated to be about 200 μg for each enantiomer ., The small amount of material required by CZE prompted us to use this technique in order to assign the optical activity of each enantiomer as either dextro- or levo-rotatory ., In a previous separation by CZE , Abushoffa & Clark 8 showed that the OXA levorotatory ( – ) enantiomer has a higher electrophoretic mobility than the dextrorotatory ( + ) one ., A racemate solution spiked with enantiomer #1 , obtained from the chromatographic separation , was then analyzed by CZE as previously described ., Since compound #1 co-migrated with the faster peak of the racemic mixture , it was identified with the levorotatory ( – ) enantiomer ( Fig 3 ) ., In our previous work , SmSULT crystals were soaked with racemic OXA prepared for medicinal use 5 ., Although the preparation contained an approximate 1:1 mixture of enantiomers , only one ( S-OXA , see below ) was observed in the crystal structure ., In the present study , single SmSULT crystals were soaked with purified preparations of either R- or S-OXA enantiomers ., Data collection and refinement statistics for the two new structures determined in this study are shown in Table 1 ., The OXA enantiomers bind in similar overall orientations ( Fig 4A and 4B ) ., The crystal structures clearly reveal the chirality of each compound and identify the ( – ) -enantiomer ( peak #1 ) as R-OXA and the ( + ) -enantiomer ( peak #2 ) as S-OXA 19 ., The positions and orientations of amino acid residues contacting the two enantiomers of OXA are virtually unchanged in the two structures and it is the relative positions of the two enantiomers of OXA and the water structure that adjust to best accommodate each enantiomer in the SmSULT binding cavity ., Some of the ordered water structure in contact with OXA is preserved between enantiomer complexes including water molecules that form hydrogen bonds to the hydroxy moiety , the amine in the isopropylaminomethyl moiety , or make a van der Waals contact to the isopropyl moiety ( S1 Fig ) ., A water molecule unique to the R-OXA complex is observed in hydrogen bonding distance to the piperidine amine ., The orientation of S-OXA prevents a water molecule from occupying this same position , but two additional water molecules unique to the S-OXA complex are observed nearby within van der Waals contact distances ( S1 Fig ) ., The isopropylaminomethyl and piperidine moieties ( Figs 1 and 5A and 5B ) of the OXA enantiomers are observed in different configurations ., The terminal methyl groups of the isopropyl moiety in each enantiomer orient in the same direction and the adjacent secondary amino groups are both observed in hydrogen bonding distance to Asp144 ( Figs 4 and 5A ) ., However , the chiral carbon atoms of the piperidine moieties joining the isopropylaminomethyl moieties in each enantiomer force the linked methyl groups to orient in opposite directions ( Fig 5B ) ., Additionally , the piperidine rings bend in opposite directions at the carbon position next to the chiral carbon , relative to the plane of the rings thereby adopting opposite ring puckers ( Fig 5B ) ., While the OXA nitro and hydroxymethyl substituent groups are rotated with respect to each other about an axis perpendicular to the aromatic ring , the accepting hydroxyl groups in both enantiomers essentially overlap and are therefore both in position to hydrogen bond to the side chain of Asp91 ( Fig 5A ) ., The nitro moiety of S-OXA maintains a hydrogen bond ( 2 . 7Å ) with Thr157 as previously observed in the SmSULT complex structure determined in the presence of the racemic mixture ., The nitro moiety of R-OXA is significantly rotated relative to that in S-OXA and positioned oriented at a greater distance from Thr157 ( 3 . 5Å ) suggesting a comparatively weak hydrogen bond ( Fig 5A ) ., Overall , the ring structure of R-OXA rotates ~10 degrees about an axis perpendicular to the plane of the ring relative to S-OXA ., The separated OXA enantiomers , together with the racemic mixture , were tested in vitro for their relative schistosomicidal properties ., In these experiments , the concentration of racemate was double that of the enantiomers , under the assumption that only one of the two enantiomers might be active , so that the racemate would contain only 50% of active substance ., In order to take into account as much as possible the variety of effects exerted on schistosomes by the different compounds , we preferred to adopt a vitality scoring system rather than a simple live/dead classification ., Four different experiments were performed at the Institute of Cell Biology and Neurobiology , two of them at low doses ( upper half of Table, 2 ) and two at high doses ( lower half of Table 2 ) ., The results of the two experiments at each dose range were pooled and averaged ., In the set of experiments carried out at low drug concentration , OXA racemate at 8 μg/mL reduced worm vitality to 14 . 4% of controls , while S-OXA ( at half that concentration ) was less effective , decreasing vitality to 32 . 6% of controls ., R-OXA , on the other hand , had only a very modest effect , lowering worm vitality to about 89% of controls ., Thus , under these conditions , S-OXA had about 3X the activity of R-OXA ., When 5X higher drug concentrations were used ( lower half of Table 2 ) , the effect on worms was obviously more pronounced for all compounds , and even R-OXA displayed a sizeable activity ( bringing worm vitality down to 22 . 2% of controls ) ( see S1–S4 Videos ) ., In an independent set of experiments conducted at UTHSCSA , using a different schistosome strain and minor methodological variations , adult schistosome male worms were treated with 40 μg/mL of racemate OXA and 20 μg/mL of each enantiomer ., Under these conditions , S-OXA clearly showed a much higher activity than R-OXA , thus confirming the above results ( S2 Fig ) ., Taken together , these results suggest that the bulk of antischistosomal activity is exerted by S-OXA , but even R-OXA can have antischistosomal effects when present at high concentrations ., This is confirmed by the fact that the racemate at 8 μg/mL is more effective than S-OXA enantiomer at half the concentration , possibly due to a contribution of R-OXA to the overall activity ., In order to rule out the possibility that the separated enantiomers might induce some non-specific toxicity , we had preliminarily ascertained that the OXA-resistant schistosome strain HR 5 was completely unaffected by these substances ., The separation of the two OXA enantiomers had been previously described using either chromatographic or electrophoretic approaches 8–10 , but the relative contributions of each enantiomer to antischistosomal activity and their individual involvement in the molecular mechanism of action had not been addressed ., Data presented here show that the S- ( + ) -enantiomer is responsible for the majority of the antischistosomal activity , while the R- ( – ) -enantiomer is capable of exerting a moderate activity that is best detected when present alone and at relatively high concentration ., Indeed , the crystal structures reveal both enantiomers bind similarly in the central cavity of SmSULT , although when crystals are exposed to a racemic mixture of OXA electron density for only S-OXA is observed ., The structures of both enantiomer complexes with SmSULT reveal that the positions of amino acid residues surrounding OXA in the central cavity do not vary significantly ., Instead , it is the orientation and configuration of the OXA enantiomers that varies in order to occupy the central cavity ., Although R- and S-OXA occupy much of the same space in the cavity , the positions of the piperidine moiety and its isopropylaminomethyl substituent demonstrate the most variation due to influence of the chiral carbon while preserving the positioning of the methylhydroxy group which is the target of modification by the sulfotransferase ., The combined results of these molecular and biological analyses suggest that when schistosomes are exposed to the racemic OXA mixture , the activating enzyme SmSULT preferentially binds and sulfonates the S-OXA , which has an overall better steric fit for the central cavity of the protein ., The piperidine rings in the enantiomers show opposite puckers adjacent to the chiral carbon , but the adjoining isopropylaminomethyl groups of R- and S-OXA occupy similar positions in the binding pocket ( Fig 3 ) ., R- and S-OXA also maintain similar hydrogen bonding distances ( 2 . 8 and 2 . 6 Å , respectively ) between the isopropylamino group and the Asp144 side chain Oδ ., However , a more favorable hydrogen bond is formed by Thr157 to the nitro group of S-OXA ( 2 . 7 Å ) compared to R-OXA ( 3 . 5 Å ) ., Thus , the S-OXA enantiomer may out-compete R-OXA due to a more favorable energy of binding in the racemic mixture ., Kinetic data , which we are attempting to generate , may help address the issue of of why S is better than R in terms of activity ., As with many other drugs , it would be desirable that stereochemically homogeneous compounds be employed as antischistosomal agents ., Our ongoing efforts to generate novel wide spectrum OXA derivatives will definitely take this option into account . | Introduction, Methods, Results, Discussion | For over two decades , a racemic mixture of oxamniquine ( OXA ) was administered to patients infected by Schistosoma mansoni , but whether one or both enantiomers exert antischistosomal activity was unknown ., Recently , a ~30 kDa S . mansoni sulfotransferase ( SmSULT ) was identified as the target of OXA action ., Here , we separate the OXA enantiomers using chromatographic methods and assign their optical activities as dextrorotary ( + ) -OXA or levorotary ( - ) -OXA ., Crystal structures of the parasite enzyme in complex with optically pure ( + ) -OXA and ( - ) -OXA ) reveal their absolute configurations as S- and R- , respectively ., When tested in vitro , S-OXA demonstrated the bulk of schistosomicidal activity , while R-OXA had antischistosomal effects when present at relatively high concentrations ., Crystal structures R-OXA•SmSULT and S-OXA•SmSULT complexes reveal similarities in the modes of OXA binding , but only the S-OXA enantiomer is observed in the structure of the enzyme exposed to racemic OXA ., Together the data suggest the higher schistosomicidal activity of S-OXA is correlated with its ability to outcompete R-OXA binding the sulfotransferase active site ., These findings have important implications for the design , syntheses , and dosing of new OXA-based antischistosomal compounds . | Schistosomes , parasites that cause the disease schistosomiasis in humans , are blood flukes that infect an estimated 200 million people in 76 countries ., Control of schistosomiasis is currently based on repeated doses of the drug praziquantel ( PZQ ) ., Parasites showing reduced susceptibility to PZQ have been recovered from patients that failed PZQ treatment and have been obtained by experimental selection ., New anti-schistosomal drugs are therefore needed that can be used with PZQ to minimize the probability of resistance ., The older anti-schistosomal drug oxamniquine ( OXA ) has an excellent efficacy and safety record but is only active against one of the three species infecting humans ., Recently , a combination of genetic and structural analyses resulted in the determination of the structure of OXA in complex with its target enzyme in the parasite , providing opportunity for structure-guided modifications of OXA to make it more effective against all three schistosome species ., Synthesis of OXA results in a racemic mixture ., Here , we isolate OXA enantiomers and find that one is more effective than the other at killing schistosomes ., Crystal structures of both OXA enantiomers bound to the target enzyme suggest a molecular basis for this observation that should be considered in ongoing and future OXA-based drug design efforts . | null | null |
journal.pcbi.1005534 | 2,017 | Locking of correlated neural activity to ongoing oscillations | To date it is unclear which channels the brain uses to represent and process information ., A rate-based view is argued for by the apparent stochasticity of firing 1 and by the high sensitivity of the network dynamics to single spikes 2 ., In an extreme view , correlated firing is a mere epiphenomenon of neurons being connected ., Indeed , a large body of literature has elucidated how correlations relate to the connectivity structure 3–14 ., But the matter is further complicated by the observation that firing rates and correlations tend to be co-modulated , as demonstrated experimentally and explained theoretically 4 , 5 ., If the brain employs correlated firing as a means to process or represent information , this requires in particular that the appearance of correlated events is modulated in a time-dependent manner ., Indeed , such modulations have been experimentally observed in relation to the expectation of the animal to receive task-relevant information 15 , 16 or in relation to attention 17 ., Oscillations are an extreme case of a time-dependent modulation of the firing rate of cells ., They are ubiquitously observed in diverse brain areas and typically involve the concerted activation of populations of neurons 18 ., They can therefore conveniently be studied in the local field potential ( LFP ) that represents a complementary window to the spiking activity of individual neurons or small groups thereof: It is composed of the superposition of the activity of hundreds of thousands to millions of neurons 19 , 20 and forward modeling studies have confirmed 21 that it is primarily driven by the synaptic inputs to the local network 22–24 ., As the LFP is a quantity that can be measured relatively easily , this mesoscopic signal is experimentally well documented ., Its interpretation is , however , still debated ., For example , changes in the amplitude of one of the components of the spectrum of the LFP have been attributed to changes in behavior ( cf . e . g . 25 ) ., A particular entanglement between rates and correlations is the correlated firing of spikes in pairs of neurons in relation to the phase of an ongoing oscillation ., With the above interpretation of the LFP primarily reflecting the input to the cells , it is not surprising that the mean firing rate of neurons may modulate in relation to this cycle ., The recurrent network model indeed confirms this expectation , as shown in Fig 1A ., It is , however , unclear if and by which mechanisms the covariance of firing follows the oscillatory cycle ., The simulation shown in Fig 1B indeed exhibits a modulation of the covariance between the activities of pairs of cells ., Such modulations have also been observed in experiments: Denker et al . 26 have shown that the synchronous activation of pairs of neurons within milliseconds preferentially appears at a certain phase of the oscillatory component of the LFP in the beta-range—in their words the spike-synchrony is “phase-locked” to the beta-range of the LFP ., They explain their data by a conceptual model , in which an increase in the local input , assumed to dominate the LFP , leads to the activation of cell assemblies ., The current work investigates an alternative hypothesis: We ask if a periodically-driven random network is sufficient to explain the time-dependent modulation of covariances between the activities of pairs of cells or whether additional structural features of the network are required to explain this experimental observation ., To investigate the mechanisms causing time-dependent covariances in an analytically tractable case , we here present the simplest model that we could come up with that captures the most important features: A local network receiving periodically changing external input ., The randomly connected neurons receive sinusoidally modulated input , interpreted as originating from other brain areas and mimicking the major source of the experimentally observed LFP ., While it is obvious that the mean activity in a network follows an imposed periodic stimulation , it is less so for covariances ., In the following we will address the question why they are modulated in time as well ., Extending the analysis of mean activities and covariances in the stationary state 13 , 27 , 28 , we here expose the fundamental mechanisms that shape covariances in periodically driven networks ., Our network model includes five fundamental properties of neuronal dynamics: First , we assume that the state of low and irregular activity in the network 1 is a consequence of its operation in the balanced state 29 , 30 , where negative feedback dynamically stabilizes the activity ., Second , we assume that each neuron receives a large number of synaptic inputs 31 , each of which only has a minor effect on the activation of the receiving cell , so that total synaptic input currents are close to Gaussian ., Third , we assume the neurons are activated in a threshold-like manner depending on their input ., Fourth , we assume a characteristic time scale τ that measures the duration of the influence a presynaptic neuron has on its postsynaptic targets ., Fifth , the output of the neuron is dichotomous or binary , spike or no spike , rather than continuous ., As a consequence , the variance of the single unit activity is a direct function of its mean ., We here show how each of the five above-mentioned fundamental properties of neuronal networks shape and give rise to the mechanisms that cause time-dependent covariances ., The presented analytical expressions for the linear response of covariances expose two different paths by which a time-dependence arises: By the modulation of single-unit variances and by the modulation of the linear gain resulting from the non-linearity of the neurons ., The interplay of negative recurrent feedback and direct external drive can cause resonant behavior of covariances even if mean activities are non-resonant ., Qualitatively , these results explain the modulation of synchrony in relation to oscillatory cycles that are observed in experiments , but a tight locking of synchronous events to a particular phase of the cycle is beyond the mechanisms found in the here-studied models ., To address our central question , whether a periodically-driven random network explains the experimental observations of time-modulated pairwise covariances , we consider a minimal model here ., It consists of one inhibitory ( I ) population and , in the latter part of the paper , additionally one excitatory population ( E ) of binary model neurons 6 , 27 , 29 , 32 ., Neurons within these populations are recurrently and randomly connected ., All neurons are driven by a global sinusoidal input mimicking the incoming oscillatory activity that is visible in the LFP , illustrated in Fig 2 . The local network may in addition receive input from an external excitatory population ( X ) , representing the surrounding of the local network ., The fluctuations imprinted by the external population , providing shared inputs to pairs of cells , in addition drive the pairwise covariances within the network 13 , c . f . especially the discussion ., Therefore we need the external population X to arrive at a realistic setting that includes all sources of covariances ., In the following , we extend the analysis of cumulants in networks of binary neurons presented in 6 , 13 , 27 , 28 , 33 to the time-dependent setting ., This formal analysis allows us to obtain analytical approximations for the experimentally observable quantities , such as pairwise covariances , that expose the mechanisms shaping correlated network activity ., Binary model neurons at each point in time are either inactive ni = 0 or active ni = 1 . The time evolution of the network follows the Glauber dynamics 34; the neurons are updated asynchronously ., At every infinitesimal time step dt , any neuron is chosen with probability d t τ ., After an update , neuron i is in the state 1 with the probability Fi ( n ) and in the 0-state with probability 1 − Fi ( n ) , where the activation function F is chosen to be, F i ( n ) = H h i - θ i h i = ∑ k = 1 N J i k n k + h extsinω t + ξ i H ( x ) = 1 if x ≥ 0 0 if x < 0 ., ( 1 ), We here introduced the connectivity matrix J with the synaptic weights J i j ∈ ℝ describing the influence of neuron j on neuron i ., The weight Jij is negative for an inhibitory neuron j and positive for an excitatory neuron ., Due to the synaptic coupling the outcome of the update of neuron i potentially depends on the state n = ( n1 , … , nN ) of all other neurons in the network ., Compared to the equations in 13 , page 4 , we added an external sinusoidal input to the neurons representing the influence of other cortical or subcortical areas and Gaussian uncorrelated noise with vanishing mean 〈ξi〉 = 0 and covariance 〈 ξ i ξ j 〉 = δ i j σ noise 2 . The threshold θi depends on the neuron type and will be chosen according to the desired mean activity ., We employ the neural simulation package NEST 35 , 36 for simulations ., Analytical results are obtained by mean-field theory 6 , 13 , 27 , 28 , 37 , 38 and are described for completeness and consistency of notation in the section “Methods” ., In the main text we only mention the main steps and assumptions entering the approximations ., The basic idea is to describe the time evolution of the Markov system in terms of its probability distribution p ( n , t ) ., Using the master Eq 14 we obtain ordinary differential equations ( ODEs ) for the moments of p ( n , t ) ., In particular we are interested in the population averaged mean activities mα , variances aα , and covariances cαβ, m α t ≔ 1 N α ∑ i ∈ α n i t ( 2 ), a α t ≔ 1 N α ∑ i ∈ α n i t - n i t 2 ( 3 ), c α β t ≔ 1 N α N β ∑ i ∈ α , j ∈ β , i ≠ j n i t n j t - n i t n j t , ( 4 ), which are defined as expectation values 〈〉 over realizations of the network activity , where the stochastic update of the neurons and the external noisy input presents the source of randomness in the network ., The dynamics couples moments of arbitrarily high order 33 ., To close this set of equations , we neglect cumulants of order higher than two , which also approximates the input by a Gaussian stochastic variable with cumulants that vanish for orders higher than two 39 ., This simplification can be justified by noticing that the number of neurons contributing to the input is large and their activity is weakly correlated , which makes the central limit theorem applicable ., In a homogeneous random network , on expectation there are Kαβ = pαβ Nβ synapses from population β to a neuron in population α ., Here pαβ is the connection probability; the probability that there is a synapse from any neuron in population β to a particular neuron in population α and Nα is the size of the population ., Mean Eq ( 2 ) and covariance Eq ( 4 ) then follow the coupled set of ordinary differential equations ( ODEs , see section II A in S1 Text for derivation ), τ d d t m α t = - m α t + φ ( μ α ( m t , h extsinω t ) , σ α ( m t , c t ) ) ( 5 ), τ d d t c α β t = { - c α β t + ∑ γ S μ α m t , h extsinω t , σ α m t , c t × K α γ J α γ c γ β t + δ γ β a β t N β } + α ↔ β , ( 6 ), where α ↔ β indicates the transposed term ., The Gaussian truncation employed here is parameterized by the mean μα and the variance σ α 2 of the summed input to a neuron in population α ., These , in turn , are functions of the mean activity and the covariance , given by Eqs ( 18 ) and ( 19 ) , respectively ., Here φ is the expectation value of the activation function , which is smooth , even though the activation function itself is a step function , therefore not even continuous ., The function φ fulfills limm → 0 φ = 0 and limm → 1 φ = 1 and monotonically increases ., Its derivative S with respect to μ has a single maximum and is largest for the mean input μ within a region with size σ around the threshold θ ., S measures the strength of the response to a slow input and is therefore termed susceptibility ., The definitions are given in “Methods” in Eqs ( 17 ) and ( 20 ) ., The stationary solution ( indicated by a bar ) of the ODEs Eqs ( 5 ) and ( 6 ) can be found by solving the equations, m ¯ = φ m ¯ ( 7 ), 2 c ¯ = S K J c ¯ + a ¯ N + transposed ( 8 ), numerically and self-consistently , as it was done in 13 , 27 , 33 ., The full time-dependent solution of Eqs ( 5 ) and ( 6 ) can , of course , be determined numerically without any further assumptions ., Besides the comparison with simulation results , this will give us a check for the subsequently applied linear perturbation theory ., The resulting analytical results allow the identification of the major mechanisms shaping the time-dependence of the first two cumulants ., To this end , we linearize the ODEs Eqs ( 5 ) and ( 6 ) around their stationary solutions ., We only keep the linear term of order hext of the deviation , justifying a Fourier ansatz for the solutions ., For the mean activities this results in m α ( t ) = m ¯ α + δ m α ( t ) = m ¯ α + M α 1 e i ω t with, M α 1 = ∑ β U α β M β 1 = ∑ β U α β h ext U - 1 S μ ¯ , σ ¯ β - i τ ω + 1 - λ β τ ω 2 + 1 - λ β 2 ., ( 9 ) The time-dependence of σ was neglected here , which can be justified for large networks ( “Methods” , Eqs ( 22 ) and ( 30 ) ) ., The matrix U represents the basis change that transforms W ¯ α β ≔ S ( μ ¯ α , σ ¯ α ) K α β J α β into a diagonal matrix with λα the corresponding eigenvalues ., We see that , independent of the number of populations or the detailed form of the connectivity matrix , the amplitude of the time-dependent part of the mean activities has the shape of a low-pass-filtered signal to first order in hext ., Therefore the phase of δm lags behind the external drive and its amplitude decreases asymptotically like 1 ω , as can be seen in Fig 3A and 3B ., If we also separate the covariances into their stationary part and a small deviation that is linear in the external drive , c α β ( t ) = c ¯ α β + δ c α β ( t ) , expand S ( μα ( t ) , σα ( t ) ) and a ( t ) around their stationary values , keeping only the terms of order hext and neglect contributions from the time-dependent variation of the variance of the input σ2 ( see “Methods” , especially Eq ( 30 ) for a discussion of this point ) , we get the ODE, τ d d t δ c t + 2 δ c t - W ¯ δ c t - W ¯ δ c t T = { W ¯ diag 1 - 2 m ¯ N diag δ m t ︸ modulated-autocorrelations-drive + diag K ⊛ J δ m t ︸ recurrent drive + h extsinω t ︸ direct drive diag ∂ S ∂ μ t K ⊛ J c ¯ total } + . . . T , ( 10 ), where we introduced the point-wise ( Hadamard ) product ⊛ of two matrices A and B see 40 , for a consistent notation of matrix operations as ( A ⊛ B ) ij ≔ AijBij , defined the matrix with the entries diag ( x ) ij := δij xi for the vector x = ( x1 , ‥ , xn ) and set c ¯ t o t a l ≔ c ¯ + d i a g ( a ¯ N ) to bring our main equation into a compact form ., We can now answer the question posed in the beginning: Why does a global periodic drive influence the cross covariances in the network at all and does not just make the mean activities oscillate ?, First , the variances are modulated with time , simply because they are determined via Eq ( 3 ) by the modulated mean activities ., A neuron i with modulated autocorrelation ai ( t ) projects via Jji to another neuron j and therefore shapes the pairwise correlation cji ( t ) in a time-dependent way ., We call this effect the “modulated-autocovariances-drive” , indicated by the curly brace in the second line of Eq ( 10 ) ., Its form in index notation is W ¯ d i a g ( ( 1 − 2 m ¯ ) / N ) d i a g ( δ m ( t ) ) α β = W ¯ α β ( 1 − 2 m ¯ β ) / N β δ m β ( t ) ., This is the low-pass-filtered input ., The other contributions are a bit more subtle and less obvious , as they are absent in networks with a linear activation function ., The derivative of the expectation value of the activation function , the susceptibility , contributes linearly to the ODE of the covariances ., As the threshold-like activation function gives rise to a nonlinear dependence of φ on the mean input μ , the susceptibility S = φ′ is not constant , but depends on the instantaneous mean input ., The latter changes as a function of time by the direct external drive and by the recurrent feedback of the oscillating mean activity , indicated by the terms denoted by the curly braces in the third line of Eq ( 10 ) ., Together , we call these two term the “susceptibility terms” ., Both terms are of the same form, diag δ μ ( t ) diag ∂ S ∂ μ t K ⊛ J c ¯ total α β= δ μ α ( t ) ∂ S α ∂ μ α ∑ γ K α γ J α γ ( c ¯ γ β + δ γ β a ¯ β N β ) , ( 11 ), but with different δμα ., This form shows how the time-dependent modulation of the mean input δμα , by the second derivative of the gain function ∂ S α ∂ μ α = φ ″ , influences the transmission of covariances ., The sum following ∂ S α ∂ μ α is identical to the one in the static case Eq ( 8 ) ., For the “recurrent drive” , the time-dependent input is given by δμα ( t ) = ∑β Kαβ Jαβ δmβ ( t ) , which is a superposition of the time-dependent activities that project to population α and is therefore low-pass-filtered , too ., The term due to “direct drive” is δμα ( t ) = hext sin ( ωt ) ., We solve Eq ( 10 ) by transforming into the eigensystem of W ¯ and inserting a Fourier ansatz , δ c α β ( t ) = C α β 1 e i ω t ., The solution consists of a low-pass filtered part coming from the direct drive and two parts that are low-pass filtered twice , coming from the recurrent drive and the modulated-autocovariances-drive ., For a detailed derivation , consult the section “Covariances: Stationary part and response to a perturbation in linear order” ., We have calculated higher Fourier modes of the simulated network activity and of the numerical solution of the mean-field equations to check if they are small enough to be neglected , so that the response is dominated by the linear part ., Of course , it would be possible to derive analytical expressions for those as well ., However , we will see that the linear order and the corresponding first harmonic qualitatively and for remarkably large perturbations even quantitatively gives the right predictions ., The limits of this approximation are analyzed in Fig D in S1 Text ., We will therefore constrain our analysis to controlling the higher harmonics through the numerical solution ., In the following we will study three different models of balanced neuronal networks to expose the different mechanisms in their respective simplest setting ., We have left out so far several steps in the derivation of the results that were not necessary for the presentation of the main ideas ., In this section , we will therefore give a self-contained derivation of our results also necessitating paraphrases of some results known from earlier works ., The starting point is the master equation for the probability density of the possible network states emerging from the Glauber dynamics 34 described in “Binary network model and its mean field equations” ( see for the following also 13 , 37 ), ∂ p ∂ t ( n , t ) = 1 τ ︸ update rate ∑ i = 1 N ( 2 n i - 1 ) ︸ ∈ { - 1 , 1 } , direction of flux ϕ i ( n ∖ n i , t ) ︸ net flux due to neuron i ∀ n ∈ { 0 , 1 } N , ( 14 ), where, ϕ i ( n ∖ n i , t ) = p ( n i - , t ) F i ( n i - ) ︸ neuron i transition up - p ( n i + , t ) ( 1 - F i ( n i + ) ) ︸ neuron i transition down = - p ( n i + ) + p ( n i - , t ) F i ( n i - ) + p ( n i + , t ) F i ( n i + ) ., The activation function Fi ( n ) is given by Eq ( 1 ) ., Using the master equation ( for details cf . section II A in S1 Text ) , one can derive a differential equation for the mean activity of the neuron i , 〈ni〉 ( t ) = ∑n p ( n , t ) ni and the raw covariance of the neurons i and j , 〈ni ( t ) nj ( t ) 〉 = ∑n p ( n , t ) ninj 6 , 13 , 27 , 34 , 37 ., This yields, τ d d t n k t = - n k t + F k t d d t n k t n l t = - n k t n l t + n l t F k t + k ↔ l ., ( 15 ), As mentioned in “Binary network model and its mean field equations” , we assume that the input hi coming from the local and the external population is normally distributed , say with mean μi and standard deviation σi given by, μ i ( t ) ≔ h i = J n i + h ext sin ( ω t ) σ i 2 ( t ) ≔ h i 2 - h i 2 = ∑ k , k ′ = 1 N J i , k J i , k ′ n k n k ′ - n k n k ′ + σ i noise 2 = J T c J i i + J ⊛ J n ⊛ 1 - n + σ i noise 2 , ( 16 ), where the average 〈〉 is taken over realizations of the stochastic dynamics and we used the element-wise ( Hadamard ) product ( see main text ) ., The additional noise introduced in Eq ( 1 ) effectively leads to a smoothing of the neurons’ activation threshold and broadens the width of the input distribution ., It can be interpreted as additional variability coming from other brain areas ., Furthermore , it is computationally convenient , because the theory assumes the input to be a ( continuous ) Gaussian distribution , while in the simulation , the input ∑ l = k N J i k n k , being a sum of discrete binary variables , can only assume discrete values ., The smoothing by the additive noise therefore improves the agreement of the continuous theory with the discrete simulation ., Already weak external noise compared to the intrinsic noise is sufficient to obtain a quite smooth probability distribution of the input ( Fig 8 ) ., The description in terms of a coupled set of moment equations instead of the ODE for the full probability distribution here serves to reduce the dimensionality: It is sufficient to describe the time evolution of the moments on the population level , rather than on the level of individual units ., To this end we need to assume that the synaptic weights Jij only depend on the population α , β ∈ {exc ., , inh ., , ext ., } that i and j belong to , respectively , and thus ( re ) name them Jαβ ( homogeneity ) ., Furthermore , we assume that not all neurons are connected to each other , but that Kαβ is the number of incoming connections a neuron in population α receives from a neuron in population β ( fixed in-degree ) ., The incoming connections to each neuron are chosen randomly , uniformly distributed over all possible sending neurons ., This leads to expressions for the population averaged input hα , mean activity mα and covariance cαβ , formally nearly identical to those on the single cell level and analogous to those in 13 , sec . Mean-field solution ., The present work offers an extension of the well-known binary neuronal network model beyond the stationary case 6 , 13 , 27 , 28 , 33 ., We here describe the influence of a sinusoidally modulated input on the mean activities and the covariances to study the statistics of recurrently generated network activity in an oscillatory regime , ubiquitously observed in cortical activity 18 ., Comparing with the results of the simulation of the binary network with NEST 35 , 36 and the numerical solution of the full mean-field ODE , we are able to show that linear perturbation theory is sufficient to explain the most important effects occurring due to sinusoidal drive ., This enables us to understand the mechanisms by the help of analytical expressions and furthermore we can predict the network response to any time-dependent perturbation with existing Fourier representation by decomposing the perturbing input into its Fourier components ., We find that the amplitude of the modulation of the mean activity is of the order h ext / ( ( 1 − λ α ) 2 + ( τ ω ) 2 ) 1 2 , where λα , α ∈ {E , I} are the eigenvalues of the effective connectivity matrix W , i . e . the input is filtered by a first order low-pass filter and the amplitude of the modulation decays like ∝ ω−1 for large frequencies ., This finding is in line with earlier work on the network susceptibility 27 , esp . section V ., The qualitatively new result here is the identification of two distinct mechanisms by which the covariances δc are modulated in time ., First , covariances are driven by the direct modulation of the susceptibility S due to the time-dependent external input and by the recurrent input from the local network ., Second , time-modulated variances , analogous to their role in the stationary setting 13 , drive the pairwise covariances ., Our setup is the minimal network model , in which these effects can be observed—minimal in the sense that we would lose these properties if we further simplified the model: The presence of a nonlinearity in the neuronal dynamics , here assumed to be a threshold-like activation function , is required for the modulation of covariances by the time-dependent change of the effective gain ., In a linear rate model 10 , 46 this effect would be absent , because mean activities and covariances then become independent ., The second mechanism relies on the binary nature of neuronal signal transmission: the variance a ( t ) of the binary neuronal signal is , at each point in time , completely determined by its mean m ( t ) ., This very dependence provides the second mechanism by which the temporally modulated mean activity causes time-dependent covariances , because all fluctuations and therefore all covariances are driven by the variance a ( t ) ., Rate models have successfully been used to explain the smallness of pairwise covariances 6 by negative feedback 10 ., A crucial difference is that their state is continuous , rather than binary ., As a consequence , the above-mentioned fluctuations present due to the discrete nature of the neuronal signal transmission need to be added artificially: The pairwise statistics of spiking or binary networks are equivalent to the statistics of rate models with additive white noise 46 ., To obtain qualitative or even quantitative agreement of time-dependent covariances between spiking or binary networks and rate models , the variance of this additive noise needs to be chosen such that its variance is a function of the mean activity and its time derivative ., The direct modulation of the susceptibility S due to the time-dependent external input leads to a contribution to the covariances with first order low-pass filter characteristics that dominates the modulated covariances at large frequencies ., For small—and probably biologically realistic—frequencies ( typically the LFP shows oscillations in the β-range around 20 Hz ) , however , the modulation of the susceptibility by the local input from the network leads to an equally important additional modulation of the susceptibility ., The intrinsic fluctuations of the network activity are moreover driven by the time-dependent modulation of the variance , which is a function of the mean activity as well ., Because the mean activity follows the external drive in a low-pass filtered manner , the latter two contributions hence exhibit a second order low-pass-filter characteristics ., These contributions are therefore important at the small frequencies we are interested in here ., The two terms modulating the susceptibility , by the direct input and by the feedback of the mean activity through the network , have opposite signs in balanced networks ., In addition they have different frequency dependencies ., In networks in which the linearized connectivity has only real eigenvalues , these two properties together lead to their summed absolute value having a maximum ., Whether or not the total modulation of the covariance shows resonant behavior , however , depends also on the third term that stems from the modulated variances ., We find that in purely inhibitory networks , the resonance peak is typically overshadowed by the latter term ., This is because inhibitory feedback leads to negative average covariances 13 , which we show here reduce the driving force for the two resonant contributions ., In balanced E-I networks , the driving force is not reduced , so the resonant contribution can become dominant ., For the biologically motivated parameters used in the last setting studied here , the effective coupling matrix W has complex eigenvalues which cause resonant mean activities ., If the inhomogeneity was independent of the driving frequency , δc would have resonant modes with frequency fres and 2fres ., Due to the mixing of the different modes and by the frequency dependence of the inhomogeneity driving the modulation of covariances , these modes determine only the ballpark for the location of the resonance in the covariance ., Especially the resonances are not sharp enough so that each of them is visible in any combination of the modes ., Different behavior is expected near the critical point where ℜ ( λ ) ≲\u20091 . For predictions of experimental results , however , a more careful choice of reasonable biological parameters would be necessary ., In particular , the external drive should be gauged such that the modulations of the mean activities are in the experimentally observed range ., Still , our setup shows that the theory presented here works in the biologically plausible parameter range ., The goal of extracting fundamental mechanisms of time-dependent covariances guides the here presented choice of the level of detail of our model ., Earlier works 6 , 28 , 29 showed that our setup without sinusoidal drive is sufficient to qualitatively reproduce and explain phenomena observed in vivo , like high variability of neuronal activity and small covariances ., The latter point can be explained in binary networks by the suppression of fluctuations by inhibitory feedback , which is a general mechanism also applicable to other neuron models 10 and even finds application outside neuroscience , for example in electrical engineering 47 ., The high variability observed in binary networks can be explained by the network being in the balanced state , that robustly emerges in the presence of negative feedback 29 , 30 ., In this state , the mean excitatory and inhibitory synaptic inputs cancel so far that the summed input to a neuron fluctuates around its threshold ., This explanation holds also for other types of model networks and also for biological neural networks 48 ., We have seen here that the operation in the balanced state , at low frequencies , gives rise to a partial cancellation of the modulation of covariances ., Our assumption of a network of homogeneously connected binary neurons implements the general feature of neuronal networks that every neuron receives input from a macroscopic number of other neurons , letting the impact of a single synaptic afferent on the activation of a cell be small and the summed input be distributed close to Gaussian: For uncorrelated incoming activity , the ratio between the fluctuations caused by a single input and the fluctuations of the total input is N − 1 2 , independent of how synapses scale with N . However , the input to a neuron is actually not independent , but weakly correlated , with covariances decaying at least as fast as N−1 6 , 29 ., Therefore this additional contribution to the fluctuations also decays like N − 1 2 . The Gaussian approximation of the synaptic input relies crucially on these properties ., Dahmen et al . 39 investigated third order cumulants , the next order of non-Gaussian corrections to this approximation ., They found that the approximation has a small error even down to small networks of about 500 neurons and 50 synaptic inputs per neuron ., These estimates hold as long as all synaptic weights are of equal size ., For distributed synaptic amplitudes , in particular those following a wide or heavy-tailed distributions ( e . g . 49 , 50 , reviewed in 51 ) , we expect the simple mean-field approximation applied here to require corrections due to the strong effect of single synapses ., The generic feature of neuronal dynamics , the threshold-like nonlinearity that determines the activation of a neuron , is shared by the binary , the leaky integrate-and-fire and , approximately , also the Hodgkin-Huxley model neuron ., An important approximation entering our theory is the linearity of the dynamic response with respect to the perturbation ., We estimate the validity of our theory by comparison to direct simulations ., To estimate the breakdown of this approximation we compare the linear response to the first non-linear correcti | Introduction, Results, Methods, Discussion | Population-wide oscillations are ubiquitously observed in mesoscopic signals of cortical activity ., In these network states a global oscillatory cycle modulates the propensity of neurons to fire ., Synchronous activation of neurons has been hypothesized to be a separate channel of signal processing information in the brain ., A salient question is therefore if and how oscillations interact with spike synchrony and in how far these channels can be considered separate ., Experiments indeed showed that correlated spiking co-modulates with the static firing rate and is also tightly locked to the phase of beta-oscillations ., While the dependence of correlations on the mean rate is well understood in feed-forward networks , it remains unclear why and by which mechanisms correlations tightly lock to an oscillatory cycle ., We here demonstrate that such correlated activation of pairs of neurons is qualitatively explained by periodically-driven random networks ., We identify the mechanisms by which covariances depend on a driving periodic stimulus ., Mean-field theory combined with linear response theory yields closed-form expressions for the cyclostationary mean activities and pairwise zero-time-lag covariances of binary recurrent random networks ., Two distinct mechanisms cause time-dependent covariances: the modulation of the susceptibility of single neurons ( via the external input and network feedback ) and the time-varying variances of single unit activities ., For some parameters , the effectively inhibitory recurrent feedback leads to resonant covariances even if mean activities show non-resonant behavior ., Our analytical results open the question of time-modulated synchronous activity to a quantitative analysis . | In network theory , statistics are often considered to be stationary ., While this assumption can be justified by experimental insights to some extent , it is often also made for reasons of simplicity ., However , the time-dependence of statistical measures do matter in many cases ., For example , time-dependent processes are examined for gene regulatory networks or networks of traders at stock markets ., Periodically changing activity of remote brain areas is visible in the local field potential ( LFP ) and its influence on the spiking activity is currently debated in neuroscience ., In experimental studies , however , it is often difficult to determine time-dependent statistics due to a lack of sufficient data representing the system at a certain time point ., Theoretical studies , in contrast , allow the assessment of the time dependent statistics with arbitrary precision ., We here extend the analysis of the correlation structure of a homogeneously connected EI-network consisting of binary model neurons to the case including a global sinusoidal input to the network ., We show that the time-dependence of the covariances—to first order—can be explained analytically ., We expose the mechanisms that modulate covariances in time and show how they are shaped by inhibitory recurrent network feedback and the low-pass characteristics of neurons ., These generic properties carry over to more realistic neuron models . | resonance frequency, perturbation theory, neural networks, random variables, neuroscience, covariance, probability distribution, mathematics, algebra, network analysis, quantum mechanics, computer and information sciences, animal cells, resonance, probability theory, physics, cellular neuroscience, cell biology, linear algebra, neurons, biology and life sciences, cellular types, physical sciences, eigenvalues | null |
journal.pcbi.1000849 | 2,010 | Computational Analysis of Whole-Genome Differential Allelic Expression Data in Human | In a diploid cell , each gene is present in two copies ., The vast majority of microarray-based or RNA sequencing-based gene expression studies do not distinguish between the two copies and measure the sum of the expression of the two alleles ., This hides the fact that the two alleles are not necessarily expressed at equal levels , a phenomenon called allelic imbalance ( AI ) 1 ., The complete shut down of one allele results in monoallelic expression ( ME ) ., The most drastic example of ME is X-chromosome inactivation , where , in females , one of the two copies of the X chromosome is inactivated and packaged into heterochromatin 2 ., Less drastic is random monoallelic expression , whereby a randomly selected copy of a gene or chromosomal region is silenced by epigenetic mechanisms ( e . g . methylation ) ., In contrast , imprinting results in parent-of-origin specific inactivation of the maternal or paternal allele , depending on the locus ., While monoallelic expression completely silences one of the two alleles , less drastic allelic expression differences can result from a heterozygous regulatory site ., For example , allele of a transcription factor binding site may allow binding and result in normal expression of the target gene on that chromosome , while allele may disrupt the binding site , resulting in lower expression ., While the lower expression of allele may be compensated by an increased transcription rate at allele in heterozygous individuals , this may not be the case for individuals who are homozygous , which may result in phenotypic variation ., Researchers have tried to identify causative regulatory variants by measuring the total expression ( i . e . expression of both copies ) of a particular gene across multiple individuals , treating this as a Quantitative Trait Locus ( eQTL ) , and mapping nearby cis-regulatory regions to the gene expression ( reviewed in 3 ) ., A key problem with this type of approach is that environmental differences across individuals can affect gene expression , making the mapping problem very challenging ., Instead , a focus on the relative expression of two alleles within the same cell has been suggested to factor out environmental sources of variation , allowing for more sensitive and specific detection of epigenetic and genetic phenomena related to local control of gene expression 4 ., Combining AI measurements obtained from a set of individuals with genotyping information about these same individuals , one can map cis-regulatory variants 5–8 or detect epigenetic variation in allelic expression 9 , 10 ., Past studies with the goal of detecting AI have typically relied upon panels of SNPs with relatively low density , located in only a subset of transcribed genes of the genome 10–12 ., A simple threshold for the ratios of expression of the two alleles at a heterozygous locus is usually established ( e . g . 1 . 5 or 2-fold ) and a gene is called as imbalanced based upon whether or not the SNP ( s ) within it exceed this threshold ., Optimal AI profiling in a genome-wide manner would require high-density sampling of expressed heterozygous sites in the genome ., We recently generated the first large-scale , high-resolution assay of allelic expression 13 ., In this study , Illumina genotyping arrays were used to measure differential allelic expression at 755 , 284 polymorphic sites in lymphoblastoid cell lines ( LCL ) derived from 53 CEU samples included in the HapMap project 14 ., Because of the noise in single point AI measurements made at each heterozygous locus , sophisticated analytical methods are required to make the most out of this data ., In this paper , we develop signal processing approaches for the accurate identification and delineation of transcripts with allelic imbalance , either in a single individual at a time , or in a collection of samples ., To our knowledge , no hypothesis-free computational approaches have been proposed for the analysis of this type of data ., Detection of AI in Ge et al . 13 relied heavily upon RefSeq , Vega , and UCSC gene annotations , and SNPs were first partitioned into windows corresponding to these annotated regions as well as intergenic regions and windows with significant AI were reported ., Sophisticated bioinformatics approaches have been developed for a related , but simpler , problem in the past , that of detecting Copy Number Variants ( CNV ) or Loss Of Heterozygosity ( LOH ) in cancer cells using array-based Comparative Genomic Hybridization ( CGH ) 15–18 or genotyping arrays 19–25 ., These include the PennCNV program 26 and the QuantiSNP program 27 , that use a Hidden Markov Model related to one of the approaches considered here ., However , CNV or LOH regions have properties that make them easier to detect than regions of allelic imbalance:, ( i ) the signal , coming from genomic DNA is generally quite strong , whereas gene expression can be very low;, ( ii ) the number of copies of an allele is a small integer , whereas the allelic expression ratio is a real number;, ( iii ) the regions affected are typically quite large , whereas AI can affect a single , short gene , or even only part of a gene ., The approaches listed above are thus not easily applicable to the detection of AI in gene expression ., An alternate family of statistical approaches called changepoint methods has been proposed for segmenting array CGH data into regions exhibiting consistent signals 28 , 29 ., These non-parametric , model-free approaches have the benefit of segmenting real-numbered data without enforcing discretization ., However , they are difficult to generalize to a situation like ours , where signals come from a mixture of discrete ( sites with no expression , sites with expression but no imbalance ) and continuous ( sites with real-valued imbalance ) state space ., In this paper , we introduce a family of signal processing approaches for the analysis of AI data obtained from genotyping arrays ., We consider both statistical approaches ( Z-score computation ) and machine learning approaches ( Hidden Markov Models ) to identify transcripts that show AI and to quantify the latter ., We introduce a new type of left-to-right HMM for the joint prediction of allelic imbalance in the 53 samples considered ., Our algorithms are evaluated using permutation testing and succeed at identifying regions with known AI ., Our approaches reveal that more than 25% of transcripts ( coding or non-coding ) are subject to differential expression between the two alleles and that patterns of AI are varied and complex ., The tools and data sets described here will help biologists and geneticists to identify regions of allelic imbalance , understand the mechanisms at play , identify the genetic or epigenetic causative agents , and associate expression polymorphisms with disease susceptibility ., Allelic imbalance was assayed using Illumina Infinium Human1M/Human1M-Duo SNP bead microarrays ., These arrays , originally designed for genotyping , have probes for approximately 1 . 1 Million polymorphic sites from HapMap , of which 755284 where used for this study ., Each probe estimates the abundance of each of the two possible alleles in the sample ., Normally , genomic DNA is hybridized onto the chip and the genotypes are easily inferred from the probe intensities ., We have previously described how one can take advantage of this technology to measure allelic expression in a high-resolution , genome-wide manner 13 ., Briefly , total RNA is extracted and cDNAs are synthesized based on a protocol on heteronuclear RNA , allowing us to measure unspliced primary transcripts 8 ., The cDNA sample is hybridized onto the array and each probe estimates the abundance of each of the two alleles in the sample ., In parallel , genomic DNA from the same cell line is hybridized , which provides the basis for normalization of the cDNA hybridization while providing us with the genotype of each sample ., Details for the full process of experimentally obtaining the raw imbalance information , as well as the sample information , can be obtained from 13 ., Data obtained from technical replicates show that although the total expression level ( sum of RNA abundance in both alleles ) measured at a given SNP is highly reproducible ( =\u200a0 . 864 ) , single point allelic expression ratios are much more noisy ( =\u200a0 . 632 ) , especially for low expression levels ( see 9 ) ., This suggests that careful data analysis is required to extract as much information as possible ., Let be the set of two alleles present at polymorphic site in the population , for ( the rare cases where three or more alleles exist at the same site are ignored in this study ) ., For notational simplicity , we assume that the genome consists of a single pair of chromosomes ., In reality , the analysis that follows is repeated separately for each autosome ., Genotype phasing consists of the decomposition of the genotype of an individual into its two homologous chromosomes ., For individual , let and , be these two chromosomes , where ., Phasing remains a computationally and statistically challenging problem 30 ., In the case of HapMap individuals , phased genotypes are available , although they are not error free ., Removal of SNPs not phased in CEU HapMap release R22 resulted in 755284 SNPs which were utilized in our study ., Let and be the intensity read outs obtained from the probes interrogating site when hybridizing the genomic DNA of individual ., If individual is heterozygous at site ( i . e . ) , then we expect both and to be large ., When it is homozygous , say for , ( i . e . ) , we expect to be large and to be small ., The genotype of an individual can thus be deduced from the ratio of the two measurements ., Consider now and , the intensity read outs obtained from the probes interrogating site when hybridizing cDNA obtained from whole cell RNA extraction ., When heterozygous site sits in a transcribed region with no allelic imbalance , both and will be relatively large ., Any difference between the two may indicate allelic imbalance ., Regions that are not transcribed will obtain low values for both alleles ., We consider the following pair of observations at each site :measures the total transcript abundance , andwhich measures the fold imbalance between the expression of the two alleles ., Normalization with the DNA sample , which , for heterozygous sites , is known to be balanced , normalizes for probe sensitivity and biases ., Values for and were collected at 755284 sites ., Those sites are not uniformly distributed in the genome , with genic regions ( exonic and intronic ) having roughly 1 . 3 times the SNP density as intergenic regions ( one SNP per 3 . 5 kb in genic regions , one SNP per 4 . 5 kb in intergenic regions ) ., Figure 1, ( a ) shows the distribution of over all genic and intergenic positions ., The distribution of expression levels in gene regions is clearly bimodal: a good fraction of genes are not transcribed in LCL , and most but not all intergenic sites are not transcribed ., Assuming that 50% of genes and 10% of intergenic sites are expressed , we can deconvolve these distributions to obtain the distribution of for expressed and non-expressed regions ( Figure 1, ( b ) ) ., For two individuals , experiments were done in triplicates ., As seen in Figure S1, ( a ) and, ( b ) , the technical noise in the measurement of both and is quite significant ., As expected , values are particularly noisy at low expression levels ., The main problem addressed in this study is the statistically robust identification of genomic regions with significant and consistent allelic imbalance ., We start by noting that the data is too noisy to accurately call imbalance based on each SNP individually ( e . g . by simply using on ) , especially for regions whose expression level is relatively low ., We thus consider approaches that take advantage of the fact that most regions with AI are relatively long and are expected to contain more than one SNP ., Four main approaches were designed , implemented and compared ., Each method aims to robustly assign a score to each SNP , so that SNPs that belong to transcripts with significant allelic imbalance obtain large ( positive or negative ) scores ., In all our AI detection algorithms , AI is detected without reference to any kind of gene annotation , contrasting with the annotation-driven approach used by Ge et al . 13 , which allows us to identify regions of AI whose boundaries does not necessarily correspond to annotated genes ., The first three approaches consider data from each sample individually while the last considers data from all samples jointly in order to improve the detection of AI in individual samples ., The four approaches considered are first summarized below and then described in details ., The code implementing each algorithm is available at http://www . mcb . mcgill . ca/~blanchem/AI/code . zip ., Consider heterozygous site and define window W ( ) to be the set consisting of heterozygous sites to the left of , heterozygous sites to the right of , and itself ., The simple smoothing approach estimates ., Any site with would then be reported as having imbalance , for some appropriate threshold ., Based on False Discovery Rate assessment ( described below ) , a value of was determined to be the optimal window size and was used for all results reported ., At sites with no allelic imbalance , the value of is modeled adequately using a normal distribution centered at 0 ., However , the variance is inversely correlated with the total expression , as AI is difficult to estimate when the total expression is low ( see Figure S1b ) ., The range of possible values of are subdivided into 100 bins of equal size and the mean and variance of values were determined for SNPs belonging to every expression level bin ., A site-specific Z-Score is assigned to heterozygous site as ., Homozygous sites , being uninformative with respect to allelic ratios , are excluded from the analysis ., Consider now a collection of consecutive heterozygous ( ignoring possibly intervening homozygous sites ) SNPs ., We define the regional Z-score as ., Assuming the normality of noise in measurements , follows a Normal ( 0 , 1 ) distribution under the null hypothesis of absence of allelic imbalance ., Regional Z-Scores are first computed for every possible window of heterozygous sites ., The region with the highest regional Z-score ( in absolute value ) , is selected first and we set for all sites heterozygous within the region ., This region is then masked out and the next highest scoring non-overlapping window is selected ., The process is repeated until all heterozygous sites have a Z-Score assigned ., We note that because the is obtained based on the best window that contains site , there is an complex issue of multiple hypothesis testing that makes that this measure will not follow a Normal ( 0 , 1 ) distribution under the null hypothesis ( i . e . absence of AI ) ., In consequence , one cannot easily translate into a p-value ., We also considered a variant of the Z-Score approach where each SNP is assigned the Z-Score of the fixed-size window centered around it ., This approach , which can be seen as an improved version of our simple smoothing approach , indeed improves on the latter ( based on permutation testing and comparison to transcripts with known AI - see below ) , but is far from being as accurate as the proposed Z-Score approach , because it leads to bleeding edges at transcript boundaries ., We also investigated a version of the Z-Score approach where SNPs are not binned by expression level prior to Z-Score computation; this resulted in a small but significant decrease in accuracy , showing that the appropriate modeling of the dependency between the noise in allelic ratio and the total expression level is an important feature of our approach ., The linear nature of the data in question lends itself well to a Hidden Markov Model ( HMM ) in which each data point corresponds to a particular SNP , the hidden states correspond to qualitative descriptions of the allelic imbalance ( e . g . positive imbalance , negative imbalance , no imbalance ) , and emissions correspond to the total expression and the allelic log-ratio observed at site ., We built an HMM consisting of a total of eight hidden states ( see Figure 2a ) ., Seven of these states correspond to SNPs take belong to expressed transcripts in the LCL sample in question , with various levels of imbalance: , corresponding to strongly positive imbalance ( ) , moderately positive imbalance ( ) , slightly positive imbalance ( ) , balance ( ) , slightly negative imbalance ( ) , moderately negative imbalance ( ) and strongly negative imbalance ( ) ., There is also a state ( ) that corresponds to SNPs located in regions that are predicted not to be transcribed , and for which allelic imbalance is meaningless ., The emission probability for each state is modeled with a pair of normal distributions for the and values , with parameters ( , ) , and ( , and ) respectively ., Whereas both total expression and allelic imbalance measurements are observed at heterozygous sites , only the expression is measured at homozygous sites ., In the latter case , the imbalance data is left unobserved ( i . e . all 8 states are equally likely to have generated the observation ) ., Homozygous SNPs can thus be included in the model training and predictions , and can help delineating regions of based on expression levels ., An HMM with a realistic correspondence to the data can in principle be built with states , where represents the number of levels of positive ( and negative ) imbalance that the model represents ., Larger values of should in principle be favorable as they allow a finer discretization of allelic ratios ., Models with were trained and the false discovery rate measured and compared ( see section 0 ) ., It was found that performed better than and , and similarly to ( Figure S2 ) , so this value was used for both the ergodic and left-to-right models ., Certain parameters of the HMM are trained using the Baum-Welch algorithm , while others are fixed ., For , the emission probability distribution for is modeled non-parametrically by the histogram of Figure 1, ( b ) ( black curve ) whereas all expressing states share the same total expression distribution from Figure 1, ( b ) ( red curve ) ., These emission probability distributions are kept constant during the training procedure ., The Baum-Welch algorithm 31 is used to find maximum likelihood estimators for and , for , as well as all transition probabilities and the initial state probability ., The Baum-Welch algorithm is an expectation-maximization ( EM ) 32 approach that alternates between the Expectation step ( or E-step ) , in which the posterior probability over states is computed for each site using the Forward-Backward algorithm , and the Maximization step ( or M-Step ) where the parameters of the emission and transition probability distributions are adjusted to best reflect the observed data given these posterior probabilities ., Formulas for updating the emission probability parameters and transition probabilities are adapted straightforwardly from Mitchell 33 ., We considered training one HMM per individual ( which would allow the flexibility to model inter-experiment variation in noise , for example ) , or to train a single HMM based on the data from all individuals ( which would have the benefit of being based on more data ) ., The latter option produced slightly better results and this is the strategy we used for the rest of the study ., We also considered filtering out sites with low total expression , as their allelic expression ratio may be less reliable ., However , slightly better results were obtained without any filtering ( allowing non-expressed SNPs to naturally be classified as belonging to state ) ., Training on the whole data set took less than Baum-Welch 20 iterations and 3 hours to converge on a standard desktop computer ( convergence is defined as two consecutive iterations where no parameter or transition probability changed by more than or 1% of their value ) ., Restarts from different initial values converged to nearly the same values ., The Viterbi algorithm 34 can then be used to identify , in each individual , predicted regions of different levels of positive or negative imbalance ., The Forward-Backward algorithm 35 yields an estimate of the posterior probability of each state at each site ., In the latter case , a useful summary score for each site is the posterior expected allelic expression log-ratio , which we use as AI predictor: ., Until now we have assumed homogenous transition probabilities , regardless of the distance in base pairs between consecutive SNPs along the chromosome ., However , a more accurate model would factor in the distance between neighboring SNPs , to increase the probability of self-loops ( i . e . staying in the same state ) when the two sites are nearby but increase the probability of state change for two distant sites ., Such an approach has been used previously in HMMs designed to detect CNVs 27 ., We obtained a unit transition probability matrix as the -th root of the transition matrix obtained via Baum-Welch training of the homogeneous model , where is the average distance ( in base pairs ) between two consecutive SNPs in our data ., Then , the transition probability matrix used for a pair of sites separated by base pairs will be , which is efficiently computed using the eigenvalue decomposition of ., To ensure that our training procedure was not subject to overfitting , we used 2-fold cross validation ( dividing the 53 samples into one 26-sample data set and one 27-samples data set ) and trained our 8-state ergodic HMM separately on each half the samples ., The parameters and transition probabilities obtained were nearly identical , and so were the FDR estimates obtained by running each HMM on the complementary data set , indicating that overfitting is not an issue ., The previous HMM is called ergodic because it models an ergodic , homogeneous Markov chain over the state space ( i . e . the set of transition probabilities is independent of the position along the genome ) ., One limitation of this HMM is that it does not take full advantage of the fact that data exists for multiple individuals and that , while not all individuals are expected to have AI in exactly the same regions , one does expect AI hotspots where a significant fraction of the individuals would have imbalance ., That would be the case , for example , for genes where one allele is commonly or always silenced via epigenetic mechanisms , or when AI is due to a common regulatory variant ., The approach proposed in this section aims at predicting AI regions separately in each individual , while taking into consideration the data observed in all individuals ., In doing so , we still want to be able to identify AI regions that are unique to a given individual , but are hoping to improve the detection of regions with common AI ., For example , AI regions containing only a few SNPs , or those where the imbalance is only moderate , may be missed when present in a single individual , but may be detectable if present in a large fraction of the population ., In addition , we may be able to detect boundaries of AI regions with more accuracy when they are shared among individuals ., The approach utilized to address this is termed the left-to-right HMM 35 ( see Figure 2, ( b ) ) , similar to profile HMMs 36 ., Each site has its own copy of the set of states and transitions can only occur between states associated with neighboring sites , from left to right ., Each copy of a given state shares the same emission probability distributions that are modeled the same way as with the ergodic HMM ., However , transition probabilities will vary across positions , making the model non-homogeneous ( in contrast to our ergodic HMM approach ) ., This configuration allows for greater fine tuning at the level of each individual SNP or region , though at the cost of a substantially larger set of transition probabilities to be learned ., The training of our left-to-right HMM is a two stage process ., In the first stage , emission probabilities , transition probabilities , and start probabilities are estimated for the ergodic version of the HMM using the Baum-Welch algorithm described above , using all available individuals ., The parameters of the emission probabilities of the states in the left-to-right HMM will be set to those obtained on the ergodic training and will not be re-estimated ., The obtained ergodic non-homogeneous distance-corrected transition probabilities will be used as prior for those of the left-to-right HMM ., In the second stage , we now switch to learning the transition probabilities of the left-to-right HMM ., We assume that the data set from each individual is the result of an independent run of the HMM: , and we seek to identify the set of transition probabilities of the left-to-right HMM that maximizes this joint likelihood ., Consider a site that is not imbalanced in any individual but where site is positively imbalanced in a large fraction of the individuals ., The maximum likelihood estimator for the transition from state to state will be higher than at other positions where few individual enter an imbalanced region ., Now consider an individual where there is only weak evidence of AI starting at position ., When using an ergodic HMM for our predictions , the weak AI region will probably not be detected ., However , in the left-to-right HMM , with the increased transition probability , the AI path becomes more likely , so provided that there is sufficient imbalance , the most likely path may now to go through one of the imbalanced state ., Estimating transition probabilities between two sites separated by base pairs is done using a simple modification to the standard Baum-Welch algorithm , where the update rule for transitions is: where is the -th power of the unit transition probability obtained previously and indicates the pseudocount weight described in the following paragraph ., The regularization obtained by using the ergodic transition probability as prior reduces the risks of overfitting while improving the convergence of the training procedure ., In practice , based upon permutation tests and resulting FDR scores , a parameter of was determined to be optimal ( data not shown ) ., Once the left-to-right HMM is trained using the data from all 53 individuals ( which took 161 Baum-Welch iterations - less than 4 hours on a standard desktop computer ) , the standard Viterbi or Forward-Backbward algorithms are used to identify AI regions separately for each individual ., As with the case of the ergodic HMM , we use the posterior expected allelic expression log-ratio to summarize AI evidence at SNP ., Overfitting is a possible issue with our left-to-right HMM , as the number of parameters estimated is much larger than for the ergodic HMM ., We performed 5-fold cross-validation , training on 4/5 of the data and predicting on 1/5 ., Thanks to our regularization procedure , the predictions obtained were very similar to those obtained by training and testing on the full data set , with only a marginal decrease in FDR ., Upon study of some of the regions where AI was predicted in most or all individuals but where not known imprinted regions existed , we found that nearly half were a likely artifact of cross-hybridization ., All these suspicious regions were the results of a segmental duplication , where a fragment of a gene was duplicated ., Because the fragments still matches the genic region , sites within them will appear to be expressed ( as they match the transcript of the paralogous region ) , and polymorphisms will cause mismatches between the probe and the true transcript , which will result in apparent AI ., We thus used the human Blastz self-alignment from the UCSC Genome Browser 37 , 38 to filter out regions corresponding to recent duplications ., A possible alternate approach would consist of using the results of the genomic DNA hybridization to identify probes that match more that one location in the genome , with the possible added benefit of detecting DNA possible copy-number variation ., Due to the relatively small number of “gold standard” regions known to exhibit AI , the best available option for comparison of the various models is through permutation tests ., The goal was to preserve some of the structure of the genome such that only SNPs with approximately equal expression levels and heterozygosity would be swapped , i . e . , the only factor that is swapped freely is that of the allelic imbalance ratio ., Permuted data sets were generated as follows ., Sites were partitioned into five levels based on the number of individuals in which they are heterozygous ., Five bins were also assigned based on the average level of expression seen across all individuals ., Each SNP was then finally assigned to one of 25 bins , with one bin for each of the possible combinations of heterozygosity frequency and expression levels ., Sites were randomly permuted within each bin , preserving the correspondence between sites in different individuals ( in the case of the left-to-right HMM , the first stage of training of global HMM parameters was first done on non-permuted data , and then the second stage of model training was done on permuted data ) ., Preserving expression levels and heterozygosity is important to create permuted data sets that are as realistic as possible , in particular with respect to the fact that expressed sites are found in contiguous genomic regions rather than dispersed randomly in the genome ., Each of the prediction methods described produces one AI score per site and per individual ., For each method , the number of regions of consecutive SNPs exceeding a given score threshold , and was determined in the real and permuted data , resulting in a False-Discovery Rate of ., We use two examples to highlight the features of the data and the methods developed ., Figure 3 gives a sample of the raw data and predictions made by each method in the BLK locus ., BLK is a gene that has previously been described as allelically imbalanced in LCL 13 ., Interestingly , in this individual , two other neighboring genes have strong allelic imbalance , with FAM167A showing expression on the opposite allele compared to BLK and GATA4 also obtaining strong an consistent signals ., Although in this example the boundaries of allelic expression domains align nicely with known gene boundaries , this is not the case in general ., As is obvious from the figure , the raw expression and allelic ratio data are quite noisy ., The simple smoothing approach succeeds at identifying the main regions of allelic imbalance but does so much less reliably and precisely than the other three approaches ., Notice that this individual has no heterozygous sites in the 5′ end of FAM167A ., This results in different behaviors for each method ., The ergodic approach assigns gradually decreasing expected allelic log-ratios in that region , while the Z-Score approach only predicts imbalance in the 3′ end of the gene ., However , the left-to-right HMM has the benefit of considering data from other individuals , which have some heterozygous sites in the 5′ region of the gene , which allows it to predict strong and consistent negative allelic log-ratios over the whole gene , and a sharp transition entering the BLK transcript ., A similar phenomenon is observed for GATA4 ., Figure 4 shows the set of predictions made by the Viterbi algorithm using the left-to-right HMM on the extended GATA3 locus , in all 53 samples ., The region exhibits a large diversity of patterns of AI ., In some cases , the region of AI closely matches an annotated gene ( e . g . SFTMBT2 in several individuals ) ., Often , AI regions do not overlap any known gene ( e . g . the region located upstream of SFMBT2 ) ., Such regions , es | Introduction, Methods, Results, Discussion | Allelic imbalance ( AI ) is a phenomenon where the two alleles of a given gene are expressed at different levels in a given cell , either because of epigenetic inactivation of one of the two alleles , or because of genetic variation in regulatory regions ., Recently , Bing et al . have described the use of genotyping arrays to assay AI at a high resolution ( ∼750 , 000 SNPs across the autosomes ) ., In this paper , we investigate computational approaches to analyze this data and identify genomic regions with AI in an unbiased and robust statistical manner ., We propose two families of approaches:, ( i ) a statistical approach based on z-score computations , and, ( ii ) a family of machine learning approaches based on Hidden Markov Models ., Each method is evaluated using previously published experimental data sets as well as with permutation testing ., When applied to whole genome data from 53 HapMap samples , our approaches reveal that allelic imbalance is widespread ( most expressed genes show evidence of AI in at least one of our 53 samples ) and that most AI regions in a given individual are also found in at least a few other individuals ., While many AI regions identified in the genome correspond to known protein-coding transcripts , others overlap with recently discovered long non-coding RNAs ., We also observe that genomic regions with AI not only include complete transcripts with consistent differential expression levels , but also more complex patterns of allelic expression such as alternative promoters and alternative 3′ end ., The approaches developed not only shed light on the incidence and mechanisms of allelic expression , but will also help towards mapping the genetic causes of allelic expression and identify cases where this variation may be linked to diseases . | Measures of gene expression , and the search for regulatory regions in the genome responsible for differences in levels of gene expression , is one of the key paths of research used to identify disease causing genes , as well as explain differences between healthy individuals ., Typically , experiments have measured and compared gene expression in multiple individuals , and used this information to attempt to map regulatory regions responsible ., Differences in environment between individuals can , however , cause differences in gene expression unrelated to the underlying regulatory sequence ., New genotyping technologies enable the measurement of expression of both copies of a particular gene , at loci that are heterozygous within a particular individual ., This will therefore act as an internal control , as environmental factors will continue to affect the expression of both copies of a gene at presumably equal levels , and differences in expression are more likely to be explicable by differences in regulatory regions specific to the two copies of the gene itself ., Differences between regulatory regions are expected to lead to differences in expression of the two copies ( or the two alleles ) of a particular gene , also known as allelic imbalance ., We describe a set of signal processing methods for the reliable detection of allelic expression within the genome . | genetics and genomics/genomics, computational biology/population genetics, genetics and genomics/gene expression, computational biology/molecular genetics, computational biology/genomics, genetics and genomics/epigenetics, genetics and genomics/bioinformatics | null |
journal.pcbi.1006503 | 2,018 | Demonstrating aspects of multiscale modeling by studying the permeation pathway of the human ZnT2 zinc transporter | Multiscale computer simulations provide a general philosophy that allows one to explore complex systems while choosing the proper level of details for different regions of the system 1 ., Thus , such approaches use coarse-grained ( CG ) treatments where molecular systems are simplified by , for example , treating groups of atoms as a single particle , to decrease the required resources and allow longer or larger simulations ., The success of such systems in predicting and explaining biological phenomena has been exemplified by many studies of complex systems ( e . g . 1–5 ) ., Studying a system at CG-level resolution is particularly beneficial when the systems are very large , the process investigated takes place over long time-scales ( ≥ microseconds ) , or when the system is of low resolution ., In the latter case , the advantage of CG modeling is that it does not treat all atoms explicitly , and therefore low-resolution structures ( commonly cryo-EM ) or models ( with the accompanied uncertainties and inaccuracies ) are good examples where CG modeling can perform better than full atomistic simulations ( e . g . 6 ) ., In fact , in studies of complex systems it is recommended to start by charting the system with CG modeling ., Lastly , the advantage of CG is that it results in a smoother landscape and thus results in faster convergence , which is arguably one of the key requirements in computational biology simulations ., In the current study we demonstrate the use of several aspects of multiscale modeling by studying the physiologically important human zinc transporter ZnT2 and investigating its permeation pathway ., Zinc is the second most abundant trace element in the human body and it is estimated that over 10% of the proteins in the human proteome are capable of zinc binding 7 ., Zinc is crucially important for numerous physiological processes including metabolism of nucleic acids , regulation of gene expression , signal transduction , cell division , immune- and nervous-system function , wound healing and apoptosis 8 ., In humans , intracellular zinc homeostasis is tightly regulated via the transport functions of two transporter families containing 24 different transmembrane carriers , ZIP1-14 and ZnT1-10 9 ., Moreover , various metalloproteins bind free zinc , hence buffering cytoplasmic zinc levels 10 ., In the past decade , the human zinc transporter 2 ( ZnT2/SLC30A2 ) was found to be the predominant transporter mediating the translocation of zinc into breast milk in lactating mammary epithelial cells 11 , involved in clinical cases of transient neonatal zinc deficiency ( TNZD ) 11–19 ., Specifically , mothers harboring loss of function mutations in ZnT2 produce breast milk containing very low zinc levels; consequently , their exclusively breastfed infants suffer from severe zinc deficiency ., These TNZD infants present with dermatitis , diarrhea , alopecia , loss of appetite , and consequently display growth and developmental delays 9 ., Importantly , the unaffected healthy ZnT2 allele in TNZD was recently shown to be insufficient to provide the necessary high levels of zinc in breast milk , which are strictly required for proper infant growth and development 9 , 16 ., Currently , no zinc transporter other than ZnT2 is known to play such a vital role in zinc transport into human breast milk ., Taking into consideration the crucial role of ZnT2 in human health , we undertook the current study to understand the molecular mechanism underlying transmembrane zinc transport through ZnT2 ., As a step towards this end , we used multiscale computational analyses and structural modeling to delineate the zinc permeation pathway and study the conformational dynamics of the transporter ., We then functionally validated our proposed permeation pathway using site-directed mutagenesis and experimental zinc transport assays ., To date , there is no high-resolution structure of the clinically significant ZnT2 transporter ., However , E . coli’s YiiP , the closest ZnT2 homologue with a known crystal structure , is assumed to harbor a similar 3D structure 13 ., YiiP and ZnT2 share ~20% sequence identity ( ~28% similarity ) along the region aligned ( ZnT2 residues 70–372 ) , allowing homology-based 3D model reconstruction of ZnT2 ., Interestingly , YiiP was recently suggested to be involved in antibiotic resistance in Pseudomonas Aeruginosa 20 which enhances the importance of identification of the ion permeation pathway of YiiP as well ., Herein , we conducted an array of calculations on the available structures of the bacterial YiiP ( X-ray and cryo-EM ) , and the 3D model of the human ZnT2 both in the inward- and outward-facing conformations ., We investigated the zinc permeation pathway with different multiscale strategies , ranging from PDLD/S-LRA binding free energy calculations , CG evaluation of conformational change process and Monte Carlo simulations ., It should be noted , with respect to zinc binding , that the large experimental free-energy of solvation for a zinc ion ( -467 kcal/mol 21 , 22 ) should be compensated by interactions with the transporter residues , and this is a computationally challenging task which was handled in our case by using explicit ligand particles ( see below as well as the Methods section ) ., We complemented our computational work with site-directed mutagenesis , subcellular localization and functional zinc transport assay to delineate the putative zinc permeation pathway , highlighting key residues predicted to facilitate transmembrane zinc ion translocation through ZnT2 ., The current study demonstrates the strengths of multiscale modeling and highlights the benefits of combining computational and experimental approaches to address medically important questions ., We present the first proposed zinc permeation pathway of a human zinc transporter , bearing important implications for pathophysiological states of zinc deficiency and possible development of proper therapeutic interventions for zinc deficiency-associated disorders ., The ZnT2 models were constructed using the X-ray structure of the outward-facing ( OF; 23 ) and the cryo-EM structure of the inward-facing ( IF; 24 ) conformations of YiiP ( Fig 1 ) , based on their respective sequence alignments by satisfaction of spatial constraints ( S1 Fig ) using the Memoir 25 modeling suite ., Memoir is specifically designed for transmembrane proteins and performs better than HHpred 26 and Swiss-Model 27 ., The quality of the models was verified using Verify3D , providing a global and local assessment of the correctness of fold structures at the residue level ., We found that Memoir models have better Verify3D scores than HHpred models and the scores were further improved after refinement with ModRefiner 28 ., Accordingly , four systems constructed with Memoir and minimized with ModRefiner were chosen for the computations ., We then minimized their energies and equilibrated them using the Molaris simulation package 29 , 30 ., All systems were stable as manifested by low RMSD values to their initial coordinates ( less than 2Å for the backbone atoms ) ., Following the construction and equilibration of these models and considering the expected homodimeric nature of ZnT2 13 , 31 , we initially explored the possible location of the zinc permeation pathway along the dimerization interface , as previously suggested for YiiP 32 ., However , a detailed examination of the structures and models revealed that there are almost no polar residues present along the dimeric interface , and in the OF conformation the monomers are not close enough to form a polar pore to stabilize zinc ions ., In this respect , the bacterial Na+/H+ antiporter also functions as a dimer , whereas its sodium translocation pathway is located within each monomer and not between them 33–35 ., We therefore searched for a highly polar zinc permeation pathway located within each YiiP/ZnT2 monomer and not between them ., Consistent with previous studies , the X-ray structure of the bacterial YiiP monomer reveals a cavity leading from the zinc binding site to the extracellular milieu , lined with many polar and negatively-charged amino acid residues ., However , the cavity grows very wide , and we therefore explored our models for several putative entry and exit routes and zinc permeation pathway ., Prior to computing the binding energies of the zinc ion , we sought to study the protonation state for the binding site residues ( site A , see Fig 1 and 23 ) , harboring Asp and His residues ., Determining the most stable protonation state is critical to attain correct binding energies of the ion because the charge distribution has arguably the biggest contribution to the interaction between the transporter protein and its zinc substrate ., To that end , we calculated the relative energies of the different protonation states for YiiP and ZnT2 using the Molaris package ( see Methods for more details ) ., The results are summarized in S2 Table , and the zinc binding calculations ( see below ) were conducted using the lowest-energy protonation state ., As a control , we performed the binding calculations for several other protonation states for the OF YiiP structure and ZnT2 model , and the curves were very similar qualitatively , but showed a gradual difference quantitatively as the net negative charge increases , mostly around zinc binding site A ( S2 Fig ) ., This hints at the interaction between protons and zinc ions , where protonation of side-chains weakens the binding of zinc ions , as expected for a putative proton-coupled zinc exchanger ., Further investigation of the coupling between the protonation state and the translocation of zinc ions is beyond the scope of this work and will be explored thoroughly in a subsequent study ., The zinc binding energy profile for the most stable protonation states ( see above ) on the equilibrated structures of ZnT2 and YiiP are depicted in Fig 2A ., The binding energies were calculated using the Molaris PDLD/S-LRA method developed and refined over the years by the Warshel group 36 , 37 and proven valuable in numerous studies starting over 35 years ago , exploring various and diverse biological systems 29 , 37 , 38 ., The curves were produced by averaging over several pathways selected along the cavities , where the MD simulations allow extensive sampling for the zinc ion position as well as the local conformation of the transporter protein ( see the Methods section for more details ) ., We computationally explored several alternative permeation pathways based on the volume available in the open-direction cavity ( outward in the OF conformation and inward in the IF conformation , converging at site A ) ., For the cavity on the closed side , we extrapolated the positions based on residues interacting with the zinc ion in the opposite conformation , as well as inspecting the conformational change in the CG and morph trajectories ( see below ) ., The energy curves of the OF conformation revealed a strong zinc-binding at the known binding site ( site A in Fig 3 and 23 ) , as would be expected for a zinc transporter ( corresponding to an affinity in the pM-nM range , see note regarding quantitative evaluation at the end of this subsection ) ., Although there are no published experimental Kd values for zinc binding per se , in the study of Chao and Fu 39 on the E . coli zinc transporter ZitB ( 25% sequence identity to ZnT2 ) , zinc translocation is composed of a two-steps process: a relatively rapid binding of zinc followed by a rate-limiting step in which the transporter undergoes conformational changes ., Our energy profiles suggest a qualitatively similar pattern ., Their kinetic study on ZitB reveals KM values in the high μM range ., However , this discrepancy can be attributed to different functions and properties of ZitB and ZnT2 , as well as possible deviations between the value of KM and Kd ., In this respect , the abundant intracellular zinc binding protein metallothionein ( capable of transferring zinc to the apo-forms of zinc-dependent enzymes and presumably to zinc transporters as well ) displays a KZn of 3 . 2x1013 M-1 ( i . e . a high metal binding constant ) , hence being consistent with the predicted concentration range 40 , 41 ., Moving further along the energy curves , from the binding site to the open side direction , the binding energy steadily increases , albeit with relatively small barriers , as the interactions between the binding site residues and zinc are weakening , until a plateau is observed towards the bulk ., To ensure that the suggested permeation pathway allows for selective zinc transport via alternating access , we additionally computed the energy profile in the closed direction ., Indeed , Fig 2A shows high energy barriers for ions on the closed side of the transporter , compared to the open side , typical for alternating access ., We then repeated this process for the IF conformations ., Here we qualitatively obtained curves similar to the ones for the OF conformations , but as mirror images ., The selection of the permeation pathway was not as obvious as above , because the cavity was not as large and visually apparent ., Thus , we generated several putative trajectories for the conformational transition from OF to IF and examined very carefully the changes between one conformation and the other , i . e . where do these cavities form , and which interactions are disrupted ., We produced the trajectory using two strategies:, ( i ) our newly developed CG normal mode MC simulation ( see the Methods section ) ; and, ( ii ) a Cartesian morph ., Although these trajectories are not guaranteed to represent the precise physiological conformational change , we found them very instructive and were able to suggest several pathways , albeit more tightly packed than in the OF paths ., Our energy calculations support the suggested permeation pathways , as they represent expected curves for zinc ion translocation across a transporter: low energies for the binding site , with barriers on both sides and a higher barrier on the closed side ( Fig 2A ) ., In Fig 2B we show the energy landscape for the zinc ion as a function of its position within the protein and the ZnT2 conformation ( using the OF and IF end points ) ., We estimated the conformation change barrier roughly at ~16 kcal/mol based on the ~1 . 5 sec-1 rates reported for YiiP 42 ., Walking along the energy landscape ( Fig 2B and S1 Movie ) , we divided the path into three sections for clarity; the zinc ion enters the IF conformation from the cytoplasm and binds at the binding site ( site A 23; section 1 ) ; then , ZnT2 undergoes a conformational change to the OF state ( section 2 ) , and finally the zinc ion exits to the vesicular lumen ( for ZnT2 ) or to the extracellular milieu ( for YiiP; section 3 ) ., In this model , the transport of the zinc ion along the membrane necessitates a conformational change of the transporter protein , since the IF or the OF conformations alone harbor a high energy barrier ., As mentioned above , coupling zinc translocation to the proton gradient and the directionality will be explored in a subsequent dedicated study ., Since the energy profiles are a crucial component of the current study , we sought to assess their convergence ., To validate the robustness of our methods and models and to prove proper sampling for the position of the ion and the local conformation of the interacting protein residues at each point , we chose the ZnT2 OF system as a control , and performed the same calculations using 10-fold longer simulations ., The results obtained were essentially the same regardless of simulation length ( see S3 Fig ) , indicating that our calculations most probably converged within the original simulation times used ., We wish to emphasize that the validity of our calculated permeation pathway is reinforced by the experimental mutational analysis and their consequences presented below ., Therefore , the precise qualitative nature of the binding curves found in the current study ( e . g . the Kd we obtained and the conformational change energy ) are prone to uncertainty resulting mainly from the modeling process as well as limitations of the computational methods ( e . g . simulating a single monomer , not considering the probability average of all the different protonation states that are slightly higher in energy ) ., Thus , the computational results and conclusions of this work are likely to be correct , while the calculated values should be still considered as a qualitative trend rather than actual quantitative numbers ., The consistency of the calculated results with the observed mutational experimental analyses below supports the acceptance of this mode of calculation as a proper representation of the functional key residues along the putative zinc permeation pathway ., To provide functional validation to our proposed zinc permeation pathway , we next aimed at predicting the amino acid residues that might play a functional role in zinc permeation ., We searched for residues that contributed significantly to the calculated binding free energy ( see Fig 2 ) along the permeation pathway , in several trajectories , and in several positions ., We note that considering stabilizing interactions at the energy barriers is important as well , because they lower the barrier on the open-side of the transporter ., Notably , one of the binding sites revealed by Lu et al . , ( site B; 23 ) does not appear to directly participate in the zinc permeation pathway ., Although this site appears to be sufficiently close to the permeation pathway to participate , a careful examination of the CG and morph trajectories strongly indicates that this is not the case ., Site B is deformed in the IF conformation , and its residues are too far from the putative permeation pathway that we suggest ., In this context , we propose that site B may act as a ‘waiting-area’ for zinc ions; that is , zinc ions initially bind to site B and are then translocated to the zinc permeation pathway ., This might prove functionally crucial since zinc is found at very low intracellular concentrations , and it is bound and shuttled by zinc-binding proteins such as metallothioneins 9 , 10 ., Thus , based on our computations and models we hypothesize that an auxiliary binding site to capture zinc when the transporter is undergoing conformational changes ( i . e . during the transport cycle ) might render a more continuous and robust ion flux ., This too will be further investigated in dedicated future studies ., Interestingly , according to ZnT2 modeling , the residues that participate in zinc ion binding ( site A ) are very similar in ZnT2 ( two Asp and two His ) and in YiiP ( three Asp and one His , see S1 Fig ) ., In support of this suggestion , Hoch et al . , previously showed 42 that these differences between YiiP and ZnT5 or ZnT8 sequences contribute to the selectivity of the transporters towards zinc as a substrate , compared to YiiP which transports cadmium and zinc 42 ., After careful consideration of the residues interacting with the zinc ion along the putative permeation pathway , we considered their evolutionary conservation ( see S1 Fig ) and assembled a list of candidate residues that are both important for zinc interaction and hydration , and are also conserved ., This was done following the rationale that evolutionary-conserved residues are more likely to be functionally important ., Herein , the residues studied by site-directed mutagenesis , shown in Fig 3 , were carefully selected according to their evolutionary conservation and their calculated energy contribution to zinc binding ., To provide experimental validation for the computational analyses delineating the putative zinc permeation pathway of ZnT2 , we mutated selected residues and assessed their impact on actual zinc transport activity in live human cells ., While employing Ruby-tagged ZnT2 expression plasmids ( emitting red fluorescence ) , the green fluorescence of the selective zinc probe FluoZin 3-AM , was used ., Hence , zinc-containing vesicles were detected as green fluorescent vesicles solely in cells transiently transfected with an active ZnT2 transporter after incubation with zinc as previously described 19 ( Fig 4 ) ., Our choice of specific protein residues to study was made by energetic , structural , and evolutionary conservation basis , as explained above ., Fig 5B shows a ZnT2 monomer color coded by conservation from the extracellular side of the membrane ., Only residues located at the transmembrane domain and with the highest degree of conservation ( 9 ) and the lowest ( 1 and 2 ) are shown for clarity ., This panel shows that the vast majority of highly conserved residues face the internal pore of the transporter , while the fast-evolving residues are facing the lipid core of the membrane ., This fits the model fold and the expected biological function of ZnT2 , where functional conserved residues lining the permeation pathway cannot tolerate evolutionary changes , while more rapidly evolving hydrophobic residues facing the membrane are able to accommodate a range of hydrophobic residues without impacting the membrane-protein interaction ., Therefore , the positions we chose have an important energetic contribution to zinc binding , are highly conserved , face the pore according to our structural model , and in consequence , are expected to impact zinc binding ., The results of the site-directed mutagenesis are summarized in Table 1 ., Additionally , we computed the evolutionary conservation of the ZnT2 family ., Fig 5 shows that most mutated positions are highly conserved ( in purple ) and face the putative zinc permeation pore ( Fig 5 ) , with the exception of H197 and Q198 which are slightly variable ., All mutants displayed similar or even higher red fluorescence levels ( Ruby fluorescence ) compared to the WT-ZnT2 , indicating the proper expression of these ZnT2 mutants ( Table 1 , right column ) ., For our analysis , we grouped mutated positions according to their physicochemical , biological and evolutionary conservation relevance , and hence in this way their biological impact and phenotypic characteristics could be more easily understood ( notice that the groups are not necessarily exclusive ) ., The first group includes residues E88 , D103 and E140 , which are likely involved in direct interaction with zinc; these residues are negatively charged , highly conserved and located along the putative zinc permeation pathway according to the proposed 3D model ., Indeed , site-directed substitution of these residues to Ala proved to be highly deleterious for ZnT2 function with about 90% loss of WT ZnT2 zinc transport activity ( Table 1 ) ., Thus , we considered these positions to be important for zinc transport ., Notably , whereas the E140A mutant showed a low but significant 20% decrease in the overall number of ZnT2 vesicles when compared to the WT ZnT2 , the decrease in its zinc transport function was much more profound ., The second group includes residues M85 , M114 , N189 and N214; these residues are conserved and are spatially located below and above the zinc-binding site with respect to the membrane plane ( Fig 5 ) ., However , in contrast to the first group , this group of residues showed a moderate impact on zinc transport , upon substitution to Ala , with 24–52% decrease in zinc transport capacity ( Table 1 and Fig 4 ) ., This suggests that the impact of polar residues along the permeation pathway on binding and/or transport of zinc is manifested collectively , as the contribution of each single residue to zinc permeation is smaller compared to the charged residues aforementioned ., We also observed an additional impact on ZnT2 function; for example , E88A and N189A displayed loss of vesicular localization ( Fig 4 ) and furthermore , M114A , E140A and N214A exhibited a decreased number of vesicles per cell when compared to the WT-ZnT2 ( Table 1 ) ., Supporting our findings and applying a different experimental setup , a recent study on the closely related ZnT1 revealed that ZnT1’s equivalent of ZnT2’s mutants , E88A and N189A , rendered ZnT1 dysfunctional 43 ., Our findings suggest that these five residues ( E88 , M114 , E140 , N189 , and N214 ) have a significant role in protein structure and stability in addition to their role in zinc binding and zinc permeation ., In contrast , D103A displayed a high number of ZnT2 vesicles while showing very little zinc accumulation , suggesting that this residue has an important role in zinc transport with little impact on transporter stability and subcellular localization ., The third group of mutants including M85A , H197A and Q198A retained the canonical subcellular vesicular localization ( Fig 4 and Table 1 ) ., In this respect , H197 and Q198 are not evolutionary conserved residues , are located near the cytoplasmic region , and point away from the zinc permeation pathway ( Fig 5B ) , according to the proposed structural model ., Hence , they were not expected a priori to impair zinc transport ., Indeed , H197A and Q198A retained 80–90% of WT-ZnT2 zinc transport activity , with no significant differences when compared to zinc accumulation mediated by WT ZnT2 ( Fig 4 and Table 1 ) ., We also focused on a fourth group including residues H201 , H203 , and H205 , which are part of the conserved GHGHSH motif ( His-rich loop ) located between TM helices IV and V in ZnT2 but are not present in the bacterial homologue , YiiP ., This motif was previously suggested to be involved in sensing cytosolic zinc levels 44 , 45 or in mediating the activation of tissue-nonspecific alkaline phosphatase ( TNAP ) by ZnT5 46 ., However , based on our model of the human ZnT2 , all three His residues are pointing away from the central permeation pore and are positioned in close proximity to the exit of the putative permeation tunnel ., In different initial structural models obtained for ZnT2 , these three His residues were in very different conformations , since the GHGHSH motif is a highly charged flexible loop and lacks complete template information ( see S1 Fig ) , thus had to be modeled ab initio by FREAD in Memoir suite ., In most models , these three His residues face the water bulk or other peripheral residues and their high conservation and positive charge suggests a role in the stabilization of ZnT2 with acidic phospholipid head groups or an allosteric zinc regulation role , rather than direct interaction with the transported zinc ion ., Indeed , substitution of all three His residues to Gly did not exert any deleterious effect on the zinc transport capacity of ZnT2 ., This further suggests that the GHGHSH motif in ZnT2 is not directly involved in zinc translocation to site A and across the transporter , i . e . there are no direct interactions between the zinc ion ( or its first hydration shell water molecules ) and the histidine side chains of the GHGHSH motif ., A very recent publication strengthen our findings , showing that deletion or substitution of these His residues to Ala in ZnT2 did not affect zinc transport activity 71 ., However , an allosteric regulatory role of this motif cannot be excluded ., In summary , site-directed mutagenesis of seven key residues along the putative zinc permeation pathway of ZnT2 , markedly impaired its zinc transport function ., In contrast , site-directed mutations at non-conserved residues ( H197A and Q198A ) , and at the conserved GHGHSH motif that is suggested to point away from the zinc permeation pathway , had only a minor deleterious impact on the zinc transport function of ZnT2 ., Taken together , residues predicted by our model and free-energy calculations , supported by the conservation analysis , agree well with our experimental validation of ZnT2 functionality ., In this study we demonstrated the power and potential of multiscale approaches by analyzing the zinc permeation pathway of ZnT2 , highlighting the constructive synergism of computations and functional validation experiments ., The principles of multiscale modeling used in this study , include representing the model structure using a CG model with a reduced number of atoms ., Such approaches are highly advisable to accelerate a simulation ( by reducing the degrees of freedom ) and to perform calculations that would otherwise be extremely challenging ( such as NMA ) ., The binding free-energy calculations were also performed with a semi-macroscopic method , which is a form of reduced-dimensionality modeling , allowing faster convergence and thus more stable results ., The reader is directed to other studies presented in this special issue , regarding other various multiscale approaches and techniques ., More specifically , in the current study we undertook extensive CG simulations of our proposed model of ZnT2 , which provided structural and evolutionary information which delineates , for the first time , the putative zinc permeation pathway of ZnT2 , from the cytoplasm into the lumen of intracellular vesicles ( or the extracellular milieu in the case of YiiP ) ., This proposed permeation pathway harbors one central zinc binding site ( site A 23 and Fig 3 ) and two cavities showing alternating-access in the two principal conformations of YiiP and ZnT2 and possibly in other zinc transporter homologues as well ( Fig 2 ) ., To functionally validate this permeation pathway experimentally , we performed site-directed mutagenesis to target various residues along the zinc permeation pathway ( Fig 4 ) ., Our rationale was to try to properly predict , based on the YiiP crystal structure and homology-based model of ZnT2 , the correct deleterious impact of site-directed mutagenesis of key residues along the permeation pathway on the zinc transport activity of ZnT2 in viable cells ., Indeed , this process lent strong experimental support to our structure modeling and zinc permeation pathway prediction ., Interestingly , a recent cryo-EM structure of the ZnTB zinc transporter was revealed 47 , showing a pentameric architecture ., ZnTB is proton-driven , precisely as ZnT2 is considered to be , however their sequence similarity is very low , and therefore ZnTB was not deemed a suitable template for ZnT2 in our present work ., We selected residues facing towards the putative permeation pathway with high contribution to the calculated zinc binding energy and used conservation analysis as a cross validation prior to the mutagenesis ., Indeed , all the conserved residues that we computationally predicted to be important for zinc permeation , experimentally impaired the zinc transport function of ZnT2 upon substitution to alanine , while five of these mutants exhibited proper vesicular localization ., As a negative control , we showed that residues that were facing away from the zinc permeation pathway including H197A , Q198A , and the GHGHSH motif , or having a tendency towards a less organized secondary structure ( i . e . disordered region ) , showed a minimal deleterious effect on zinc transport activity of ZnT2 , hence further supporting our hypothesis ., Thus , multiscale computational analyses complemented with functional experimental validation lead to the construction of valid models for the 3D functional conformations of human ZnT2 and the zinc translocation pathway ., Furthermore , we were able to successfully predict important residues around the putative binding site involved in zinc binding and translocation , both in the direction of the cytoplasm and the extracellular milieu ., From a translational medicine perspective , such zinc permeation pathway studies may facilitate the screening and identification of small molecules that can correct the proper folding and/or function of mutant transporter proteins ., In this respect , pharmacoperones are recently emerging as a novel class of hydrophobic small molecules that can bind to mutant misfolded and inactive proteins , thereby restoring their proper folding , subcellular sorting , and function 48 , 49; this novel approach is currently known as pharmacoperone drug therapy ., These mutant proteins which are | Introduction, Results, Discussion, Methods | Multiscale modeling provides a very powerful means of studying complex biological systems ., An important component of this strategy involves coarse-grained ( CG ) simplifications of regions of the system , which allow effective exploration of complex systems ., Here we studied aspects of CG modeling of the human zinc transporter ZnT2 ., Zinc is an essential trace element with 10% of the proteins in the human proteome capable of zinc binding ., Thus , zinc deficiency or impairment of zinc homeostasis disrupt key cellular functions ., Mammalian zinc transport proceeds via two transporter families: ZnT and ZIP; however , little is known about the zinc permeation pathway through these transporters ., As a step towards this end , we herein undertook comprehensive computational analyses employing multiscale techniques , focusing on the human zinc transporter ZnT2 and its bacterial homologue , YiiP ., Energy calculations revealed a favorable pathway for zinc translocation via alternating access ., We then identified key residues presumably involved in the passage of zinc ions through ZnT2 and YiiP , and functionally validated their role in zinc transport using site-directed mutagenesis of ZnT2 residues ., Finally , we use a CG Monte Carlo simulation approach to sample the transition between the inward-facing and the outward-facing states ., We present our structural models of the inward- and outward-facing conformations of ZnT2 as a blueprint prototype of the transporter conformations , including the putative permeation pathway and participating residues ., The insights gained from this study may facilitate the delineation of the pathways of other zinc transporters , laying the foundations for the molecular basis underlying ion permeation ., This may possibly facilitate the development of therapeutic interventions in pathological states associated with zinc deficiency and other disorders based on loss-of-function mutations in solute carriers . | Herein we employed multiscale modeling and electrostatic energy calculations to delineate , for the first time , a putative zinc permeation pathway , from the cytoplasm into intracellular vesicles ( for ZnT2 ) or to the extracellular milieu ( for YiiP ) , along the membrane-spanning domain of the human zinc transporter ZnT2 and its E . coli homologue , YiiP ., These computational findings were functionally validated using site-directed mutagenesis of ZnT2 residues predicted to reside along the putative zinc permeation pathway and zinc transport assay ., Our results shed light on the transport mechanisms of ZnT2 and YiiP and pave the way towards the elucidation of the zinc translocation mechanism in other ZnT family members ., Furthermore , these findings could also be harnessed to the possible development of therapeutic interventions in zinc-associated pathologies . | medicine and health sciences, vesicles, zinc transporters, built structures, engineering and technology, site-directed mutagenesis, simulation and modeling, nutrition, biological transport, molecular biology techniques, mutagenesis and gene deletion techniques, cellular structures and organelles, research and analysis methods, sequence analysis, sequence alignment, bioinformatics, zinc, chemistry, molecular biology, nutritional deficiencies, biochemistry, cell biology, database and informatics methods, biology and life sciences, structural engineering, physical sciences, metabolism, chemical elements, micronutrient deficiencies | null |
journal.pcbi.1000010 | 2,008 | A Novel Bayesian DNA Motif Comparison Method for Clustering and Retrieval | Transcription initiation is modulated by transcription factors that recognize sequence-specific binding sites in regulatory regions ., The organization of binding sites around a gene specifies which factors can bind to it and where , and consequently determines to what extent the gene is transcribed under different conditions ., To understand this regulatory mechanism , one must specify the DNA binding preferences of transcription factors ., These preferences are usually characterized by a motif that summarizes the commonalities among the binding sites of a transcription factor 1 ., Multiple tools were developed for finding motifs ( e . g . , 2–5 ) , however there are several problems in interpreting their output ., Typically these algorithms output multiple results which require filtering and scoring ., Moreover , different motif discovery methods have complementary successes , and therefore it is beneficial to apply multiple methods simultaneously and collate their results 6 ., In addition , the motif discovery algorithms frequently produce a redundant output and the transcription factor that binds each motif is usually unknown ., As similar motifs may represent binding sites of the same factor , eliminating this redundancy is essential for elucidating the true transcriptional regulatory program ., The general strategy is thus to cluster similar motifs and merge motifs within each cluster to create a library of non-redundant motifs 6 ( Figure 1B ) ., Next , in order to interpret the meaning of the discovered motifs , they are compared to databases of previously characterized motifs ( Figure 1C ) ., In large-scale experiments , where the motif output set is very large , the tasks of scoring , merging and identifying motifs need to be automated ., To address both the clustering and the retrieval challenges , we need an accurate and sensitive method for comparing DNA motifs ., In the literature there is an ongoing discussion regarding the best representation of DNA motifs 1 , 7–10 ., Here we use a Position Frequency Matrix ( PFM ) , which has the benefits of being relatively simple yet flexible ., A PFM is a matrix of positions in the binding site versus nucleotide preferences , where each row represents one residue and each column represents the nucleotide count at each position in a set of aligned binding sites ., This representation assumes that the choice of nucleotides at different positions is independent of all other positions ., To compare two PFMs , we need to consider all possible alignments between them ., Given two aligned PFMs , we utilize the position-independence assumption to decompose the similarity score into a sum of the similarities of single aligned positions ., Several similarity scores can be used to compare a pair of aligned positions ., One approach uses the Pearson correlation coefficient ( e . g . , 11 , 12 ) ., This measure , however , might inappropriately capture similarities between probabilities ( Figure 2 and Figure S1 ) ., Alternative approaches are based on similarity between two distributions ., This can be a metric distance , such as the Euclidean distance 13 or an information-theoretic measure , such as the Jensen-Shannon divergence 14 ., While these latter distances do not have the artifacts of the Pearson correlation , they equally weight positions with similar nucleotide distributions that are specific ( e . g . , a strong preference for an A ) and similar positions that are non-informative ( e . g . , identical to the background distribution ) ( Figure 2 and Figure S1 ) ., It is important to differentiate between these two situations: Two positions whose similarity is due to a resemblance to the background distribution are less relevant to motif similarity , as they do not contribute to sequence-specific binding of proteins 15 , 16 ., In this work we use this intuition to develop a novel method for comparing and merging DNA motifs , based on Bayesian probabilistic reasoning ., We define a new similarity score that combines the statistical similarity between the motifs with their dissimilarity to the background distribution ., To calculate this score we estimate the probabilities of DNA nucleotides in each position of the DNA motif , by a Bayesian estimator with a Dirichlet mixture prior 17 , 18 to model the multi-modal nucleotide distribution at different binding site positions ., This motif similarity score is used by us to identify similar motifs that represent binding sites of the same factor and for clustering motifs ., For the latter we devised a hierarchal agglomerative clustering procedure that is based on our motif similarity score ., Our results show that the new method outperforms other alternatives in accuracy and sensitivity in both the clustering and retrieval tasks ., Using our new similarity score and the clustering method based upon it , we developed a large-scale analysis pipeline of DNA motif sets ., This pipeline is designed for analysis following concurrent motif search by a combination of methods ( using the TAMO package 19 ) ., The goal is to process the set of DNA motifs into a set of reliable non-redundant motifs ., We use our method to identify sets of DNA motifs from a large-scale ChIP-chip assay in S . cerevisiae 13 ., This allows us to examine how transcription factors alter their DNA binding preferences under various environmental conditions and elucidate mechanisms for condition-specific preferences ., Our goal is to determine whether two DNA motifs represent the same binding preferences ., Since the less informative positions in a motif do not contribute to sequence-specific binding of proteins , we developed a similarity score that measures the similarity between two DNA motifs , while taking into account their dissimilarity from the background distribution ., We now develop the details of the score ., We can view DNA motifs as a list of binding sites from which the nucleotide distribution at each position is estimated ., This view allows us to perform statistical evaluations ., We assume that each binding site was sampled independently from a common distribution over nucleotides , which satisfies the position independence properties ( in correspondence with the motif PFM representation described above ) ., Then , we can evaluate the likelihood ratio of different source distributions of the sampled binding sites ., In practice , we keep only the sufficient statistics allowing us to evaluate the likelihood of the binding sites ., These sufficient statistics are the counts of each nucleotide in each position , represented as a PFM ., Our score is composed of two components: the first measures whether the two motifs were generated from a common distribution , while the second reflects the distance of that common distribution from the background ( see Methods ) ., Our Bayesian Likelihood 2-Component ( BLiC ) score for comparing motifs m1 and m2 is: ( 1 ) Under the position independence assumption , the score decomposes into a sum of local position scores ., More precisely , our likelihood-based score measures the probability of the nucleotide counts in each position of the motif given a source distribution ., For two aligned positions in the compared motifs , let n1 and n2 be the corresponding positions ( count vectors ) in the two motifs , the similarity score is then: ( 2 ) where are the estimators for the source distribution of n1 , n2 and the common source distribution , respectively , Pbg is the background nucleotide distribution , and NT\u200a=\u200a{A , C , G , T} ., Since the source distribution is unknown , we must estimate it from the nucleotide counts at each position ., We used a Bayesian estimator , where a priori knowledge and the number of samples were integrated into the estimation process ., We considered two alternative priors ., The first is a standard Dirichlet prior 20 , which is conjugate to the multinomial distribution , enabling us to compute the estimations efficiently ( see Methods ) ., However with this prior we cannot model our prior knowledge that a position in a DNA motif tends to have specific preference to one or more nucleotides ., Such prior knowledge can be described with a Dirichlet mixture prior 17 , 18 , which represents a prior that consists of several “typical” distributions ., Specifically , we used a five-component mixture prior , with four components representing an informative distribution , giving high probability for a single nucleotide: A , C , G , or T . The fifth component represents the uniform distribution ( see Methods ) ., In the above discussion we assumed that the motifs are aligned , but in practice , we compare unaligned motifs ., Thus , we defined the similarity score for two motifs as the score of the best possible alignment ( without gaps ) between them , including the reverse complement alignment ., In addition , we need statistical calibration of the similarity scores , since a high similarity score might be due to chance events 21 , 22 ., In particular , when comparing a single motif against motifs of different lengths , the probability of similarity by chance depends on the query motif and the length of the target ., To circumvent these problems we use the p-value of the similarity score , which is computed empirically for each query against the distribution of scores of random motifs of a given length ( see Methods and Figure 3 ) ., We set out to compare our similarity score to existing ones in the literature , in the context of both motif comparison and clustering ., We use two different data sets ., The first data set , which we refer to as “Yeast” is a synthetic one where we know the true labeling of motifs and use it to benchmark different procedures by relating their results with the underlying truth ., To generate synthetic motifs in a realistic manner that reflects true binding properties of transcription factors , we use the genome-wide catalogue of transcription factor binding locations in S . cerevisiae 13 ., This catalogue has high-confidence binding sites ( based on combination of experimental assays with evolutionary conservation considerations ) ., From these , we selected nine transcription factors to represent different binding patterns ( in terms of inner arrangements of informative positions and length ) ., From the binding sites of each factor we sampled sets of binding sites , and from each set generated a motif ( see Figure 3A ) ., For each factor we generated noisy motifs that differ in their quality ., To do so , we varied the number of binding sites ( sizes of 5 , 15 or 35 ) and the coverage of the motif ( full site , its beginning , middle , or end ) ., We repeated this for each motif 20 times , creating a set of 240 random motifs for each of the nine transcription factors ., The second data set , which we refer to as “Structural” , was compiled by Mahony et al . 24 ., Their evaluation is based on structural information ., Since structurally related transcription factors often have similar DNA-binding preferences , the best match to a given motif is expected to be a motif associated with a member of the same structural class ., Mahony et al . compiled a data set that contains the motifs of the families with four or more profiles in JASPAR 25 ., Using these two data sets we compared different possible similarity scores for DNA motifs ., Specifically , we compared the Pearson correlation coefficient; the information-theory based Jensen-Shannon divergence; the Euclidean distance; and our BLiC score ., As a real life application of this pipeline we examined genome wide ChIP-chip measurements in S . cerevisiae of 177 transcription factors under several environmental conditions ., In total we analyzed 301 experiments for different factors and conditions 13 ., We used seven motif discovery algorithms to produce a set of motifs for each ChIP-chip experiment ., These motifs were clustered , filtered , ranked and compared to known motifs from the literature ( as described above and in the Methods ) ., This resulted in a concise set of DNA motifs attributed to each transcription factor under each environmental condition ( all the motif sets can be found at the Supplementary Web site http://compbio . cs . huji . ac . il/BLiC ) ., To further analyze the resulting Yeast DNA motif library , we contrast it against the wealth of genomic annotations in the yeast literature ., To do so , we scanned each motif in the library against the promoters of yeast genes ( see Methods ) and created a target gene set for the motif ., We then scored the enrichment of these motif gene sets against different types of gene annotations: the original ChIP-chip data 13 , GO functional annotations 29 , and groups of genes which are up or down regulated according to gene expression data ( assembled by 30–32 ) ., This allowed us to relate each motif to specific genomic annotations ., To visualize these relationships we created a combined clustering of motifs and annotations using EdgeCluster - a clustering algorithm recently developed in our lab 33 ., The novelty of EdgeCluster is in the integration of various sources of information into the clustering process ., These information sources can be attributes of motifs ( e . g . , extent of enrichment in different gene sets ) and pairwise information about motifs ( i . e . , the similarity of motif pairs ) ., Figure S4 demonstrates the clustering of all the motifs ., Clustering of a partial set of motifs is presented in Figure 6 ., In the works of Harbison et al . 13 and MacIsaac et al . 34 , the same ChIP-chip data was used to construct a global transcriptional regulatory map in yeast ., The motif analyses performed in these two works differ from ours in the similarity score used ( the Euclidean distance ) and in the different motif clustering and merging methods ., In addition , the output of these two works was a single motif for each transcription factor ., To be consistent with these previous works in the comparison , we narrowed down our set of motifs for each ChIP experiment to a single motif ., We first looked only at transcription factors with previously characterized motifs ., Our criterion for comparison is measuring the similarity to known motifs from the literature ( TRANSFAC 26 , SCPD 27 , YPD 28 ) , using our BLiC score ., To narrow down our motif set to a single motif for each factor we chose ( as done in these previous works ) the motif most similar to the known motif ., In 65% of the cases our motifs have the highest similarity to the known motifs ( Figure 7 , Table S1 ) ., The motifs learned by the algorithms of MacIsaac et al . and Harbison et al . , had the highest similarity only in 22% and 12% of the motifs , respectively ., For transcription factors with no previously known binding motif in the literature , we compared the enrichment of the motifs within the ChIP-chip groups of sequences ., For the comparison , we narrowed the motif sets by choosing the most significant motif for each factor and environmental condition ( similarly to what was done in these previous studies ) ., We scanned the genomic sequences and computed the enrichment of each motif ( see Methods ) , using the same procedure and parameters for motifs from all three methods ., Our motifs were found to have the highest enrichments in 80% of the cases ( see Figure 7 and Table S1 ) ., To ensure that the improvement we see is not due to differences in motif discovery methods , we repeated the analysis using the original output of the motif discovery of Harbison et al . ( data not shown ) ., This lead to slight changes in the output motifs , as our original analysis used a superset of these motifs ., Comparing these modified results against the results of Harbison et al . and MacIssac et al . we see essentially improvement as the one we reported above ( in 62% of the cases our motifs have the highest similarity to the known motifs , and in 65% of the cases our motifs were found to have the highest enrichments ) ., Using the motif sets we have learned , we next turned to examine the change in the binding specificities of the transcription factors under different conditions ., We distinguish between two types of factors ., A condition-independent factor binds the same targets in multiple conditions , while a condition-dependent factor changes its set of targets between conditions ., An example of a condition-independent transcription factor in yeast is Fhl1 , a master regulator of ribosomal genes , which according to the ChIP data remains bound to 75% of its targets under different conditions ( see Figure S5A ) ., This is consistent with previous work 35 and with the motif analysis , where similar motifs are related to Fhl1 in all three conditions ( see Figure S5B ) ., A condition-dependent regulator can show a range of behaviors in response to a change in condition ., It may expand and bind additional targets , it may alter and bind to a different set of targets , or it may even not bind any targets 13 ., Various mechanisms may be involved in monitoring condition-dependent binding ., A factor may expand its targets , due to dosage change of the active transcription factor in the nucleus 13 ., Alternatively , a factor may alter its targets due to several probable mechanisms ( see Figure S6 ) ., One mechanism is changing the factors specificity to the DNA , which we can trace by identifying variations in the DNA motif ( Figure S6A ) ., Another possible mechanism is a change in the factors binding partner , which may be detected through co-occurrence of motifs of different factors ( Figure S6B ) ., In addition , a change of targets may be caused by a change in the accessibility to the binding site , which we cannot identify by analyzing motifs ( Figure S6C ) ., We focus here on factors that alter their targets under different conditions and try to elucidate the mechanism ., We defined a transcription factor as altering its target genes between two conditions , if the number of target genes in the intersection is less than half of the number in each condition separately ., In addition , we considered only factors with at least 20 target genes in each of the two conditions ( a sufficient number for motif discovery ) ., Out of the 72 transcription factors for which ChIP-chip experiments were carried out in more than one condition , 50 factors alter their target genes between two conditions ( in total , 112 pairs of differential conditions ) ( Table S2 ) ., We searched for differential motifs in the motif set of each factor at every condition ., We say a motif is differential if there is a significant difference ( p<0 . 05 , chi-square test ) in the fraction of ChIP targets containing the motif between the two conditions ( excluding the genes in the intersection ) ., This analysis can potentially elucidate the mechanism through which a factor changes its DNA targets , by finding different variants of motifs , or co-occurrence of motifs of different factors as explained above ., In about half of these pairs we did not find statistically significant motifs in at least one of the compared conditions and thus could not search for differential motifs ., Finding a motif only for one condition could be meaningful on its own , since this may indicate that in the other condition there is no direct binding of the factor to the DNA ., On the other hand it could result from technical reasons , such as noise in the input set of sequences , and thus in this work we do not analyze these cases ., Out of the remaining 52 pairs ( spanned over 27 different transcription factors ) , we found differential motifs for 88% of the factors ( 47 cases spanned over 24 factors , see Table S3 ) with a p-value of less than 0 . 05 ., An example of a transcription factor that shows condition-dependent binding is Ste12 , which activates genes in two alternative pathways—mating and filamentous growth 36 , 37 ( Figure 8A ) ., Under filamentous growth signaling ( Butanol induction ) we found that Ste12 binds promoters enriched with its known motif 38 , as well as the known recognition sequence of Tec1 38 , a co-factor that binds the DNA with Ste12 under filamentous growth 37 , 39 ( Figure 8B ) ., However , under mating conditions ( Alpha factor induction ) we find that Ste12 binds promoters with another variant of the motif more highly enriched than the known one ., This variant is a near-perfect tandem repeat of its known site , suggesting that Ste12 binds the DNA as a homodimer following Alpha factor induction 40 , 41 ( Figure 8B ) ., An additional player found in our analysis is Mcm1 , whose known motif 42 is enriched among promoters bound by Ste12 under both conditions ., This is consistent with the role of Mcm1 inhibiting expression of mating genes in diploid cells 42 ., Mcm1 may play a similar role in the filamentous growth pathway , in which haploid cells undergo invasive growth , and diploid cells undergo pseudohyphal growth ., Interestingly , the exact same motifs were learned for the ChIP targets of the cofactor Dig1 , under all the conditions stated above , which indicates that Dig1 does not bind the DNA directly 37 ., Thus , looking at the discovered motif sets , we can reveal the regulators involved and propose a mechanism through which a transcription factor alters its targets under different conditions ., Here we propose the altered binding is caused by a change in the DNA binding partner: Ste12 binds the DNA with Tec1 under filamentous growth and as a homodimer under mating conditions ., Another interesting example is provided by the iron-regulated transcription factor Aft2 , required for iron homeostasis and resistance to oxidative stress 43 ., This factor exhibits a significant environmental-dependent binding , switching targets between low and high H2O2 conditions ( Figure 9A ) ., The role of Aft2 in iron homeostasis and resistance to oxidative stress is poorly understood ., In low H2O2 , we find that Aft2-bound promoters are highly enriched with a motif similar to the known recognition sequence of Aft2 ( GgGTG ) 43 ., However , in high H2O2 we find abundant occurrences of a low complexity Poly-GT motif ( Figure 9B ) ., This result indicates that a possible explanation for the change in Aft2 DNA targets is a change in its DNA binding specificity over these conditions ., We reach this conclusion due to the lack of the known motif or motifs of other factors in the bound targets under high H2O2 and due to the similarity of the Poly-GT to the known motif ., Furthermore , the poly-GT motif under high H2O2 may suggest that Aft2 binds the DNA as a homodimer ., Interestingly , the known motif of Aft1 ( Rcs1 ) 43 , a paralog of Aft2 , was enriched among the Aft2-bound promoters in low H2O2 condition ., This implies a possible overlap between the targets of Aft2 and Aft1 , supported by ChIP-chip data of the two factors ( Figure 9B ) ., Based on our analysis , we report two similar ( but not identical ) motifs for the two paralogs ( as suggested by 43 , 44 ) ., Since it is known that Aft2 and Aft1 have independent and partially redundant roles in iron regulation 43 , 44 , this strengths our assumption that Aft2 binding to the DNA does not depend on Aft1 , but is due to a change in its specificity to the DNA ., The ChIP-chip data and our motif analysis suggest that under high H2O2 conditions Aft2 has a unique role in gene regulation ., Here again , by looking at the motif sets , we propose a mechanism for condition dependent binding of a transcription factor ., In this case we propose the cause is a change in the factors specificity to the DNA ., We used our BLiC score to develop a hierarchical agglomerative clustering algorithm for merging similar motifs , in which we ensure that the motifs within every sub-tree are properly aligned ., Furthermore , such an approach allows us to trim the cluster tree at various levels , thus allowing us to merge motifs at different resolutions ., In our method a new agglomerative node results from aligning and merging the motifs of its descendent nodes , and then computing the similarly of this new motif to all other nodes ., As a consequence , the hierarchical progression ensures that each sub-tree is coherent ., This is in contrast to many clustering methods , such as k-means and typical hierarchical clustering 45 which find a set of motifs that are all similar to each other , but are not necessarily coherent in the sense that they cannot all be aligned ., Our motif analysis pipeline is designed to process discovered DNA motifs into a set of non-redundant motifs and compare these with known motifs ., As we have shown , our approach improves the sensitivity and specificity in the analysis of the outputs of standard motif discovery methods ., By automating all the steps , we enable the analysis of hundreds of input groups ., In addition , we achieve a wide view on transcription regulation by running several motif discovery algorithms in parallel , and integrating their outputs ., By comparing motifs from different input groups we are able to connect between transcription factors that play a role in different processes ., Our analysis does not focus on finding the “best” single motif for each input group ( e . g . , targets of ChIP-chip assay ) , but rather we find a set of non-redundant motifs and their relations ( enrichment ) to each input group ., This output better captures the complexity of the underlying regulatory program ., For example , in many cases we find motifs of co-factors ( e . g . , Ste12 and Tec1 ) ., In other cases we see that a factor changes its binding specificity under different conditions ( e . g . , Aft2 ) ., For these cases , several DNA motifs better capture the DNA binding preferences of the transcription factor than a single motif ., There are several different approaches attempting to quantify similarities between DNA motifs ., Two previous works 21 , 22 showed that using p-values when comparing motifs is more accurate than the actual similarity scores ., Specifically , Gupta et al . 21 , compared seven motif-motif position similarity functions , including the Pearson Correlation coefficient ( e . g . , 11 , 46 ) , average log-likelihood ratio ( ALLR ) 16 , Kullback-Leibler divergence 47–49 , and the Euclidean distance ( ED ) 13 , 50 ., They found that the Euclidean distance is slightly better than the alternatives they considered ., The data set used by Gupta et al . has a similar design as our data set , but it is based on the TRANSFAC database 26 ., Not surprisingly , our results are consistent with theirs ., Here we also use p-values to calibrate similarity scores , and show that our score is more accurate than the Euclidean distance , which is the second best ., Several resources are available for DNA motif analysis ., There are many open access motif discovery tools available ( e . g . , 2 , 3 , 11 ) and motif comparison tools 11 , 21 , 51 ., In addition there are several available tools that integrate multiple motif discovery tools , and supply additional tools for filtering , comparison and ranking motifs 19 , 49 , 52 ., In our motif analysis pipeline we use the TAMO package 19 , for motif discovery and filtering , with a different genomic scan approach using statistical tools 53 ., The main difference is that for the motif comparison and clustering we use our new BLiC score and a hierarchical agglomerative clustering ( as discussed above ) ., Sequence information is a highly accessible resource , and thus it is interesting to ask whether it can help elucidate mechanisms of transcription regulation ., We examined transcription factors that alter their targets in response to an environmental change , and found a differential motif in 88% of these cases ( 24/27 factors ) ., These differential motifs can suggest the potential mechanism through which the factor changes its targets ., We show that motifs provide an indication for potential mechanisms when the factor changes its binding partner ( Figure S6A ) or its specificity to the DNA ( Figure S6B ) , as we discussed thoroughly for the case of Ste12 and Aft2 ., Nevertheless , motif analysis obviously does not reveal the whole regulatory picture ., For example , chromatin-modeling mediated regulation cannot be inferred from motif analysis ( Figure S6C ) ., Thus , for a complete understanding of the regulatory mechanisms additional information is needed ., A significant limitation of motif analysis in general , is the discrepancy between putative binding sites and actual functional binding events ., This raises the question addressed frequently before 10 , 54 , whether our representation of transcription factor binding preferences is sufficiently accurate ., In this study we overcome a basic obstacle in DNA motif analysis , by developing an accurate motif comparison method ., Our motif analysis pipeline , which includes clustering and retrieval procedures based on our novel score , is fully automated and produces accurate results ., This is highly important in large-scale analysis , such as the one reported here ., We showed the power of motif analyses , which is useful not only for building regulatory maps , but also for understanding more profoundly regulatory mechanisms ., We use a Position Frequency Matrix ( PFM ) representation for a DNA motif ., This is a n×4 matrix , where each i , j cell contains the count of nucleotide j in position i of the motif ., We define the similarity score for two aligned PFMs ., Due to the positional independence assumption in PFMs , the score decomposes into the sum of scores for corresponding positions ., Our score is composed of two components: The first measures whether the two motifs were generated from a common distribution ., The second reflects the distance of that common distribution from the background ., Thus , for positions n1 and n2 , our score is as described in Equation 2 ., Statistically , in the score we sum the log-likelihood-ratio of two pairs of hypotheses ., The first component: H0: The two samples were drawn from a common source distribution ., H1: The two samples were drawn independently from different source distributions ., The second component: H0: The two samples were drawn from a common source distribution that is distinct from the background ., H1: The two samples were drawn from the background distribution ., We estimate the source distributions from the PFM using a Bayesian approach , with a Dirichlet prior ., The Dirichlet prior is specified by a set of hyper-parameters α\u200a= ( α1 , α2 , …αn ) and has the form: Where Γ ( x ) is the Gamma function ., We use two prior variants: The first is a standard Dirichlet prior 20 , with hyper-parameters of ( 1 , 1 , 1 , 1 ) ., When using this prior , the estimated distribution for position n is:where α is the vector of hyper-parameters ., The second prior we use is a five-component mixture of Dirichlet prior 17 ., We merge five Dirichlet priors using uniform weights ., Four of the components give high probability for a single DNA nucleotide: A , C , G , or T . The fifth element represents the uniform distribution ., We use the hyper-parameters ( 5 , 1 , 1 , 1 ) for A , ( 1 , 5 , 1 , 1 ) for C , etc . , For the fifth component we use the hyper-parameters ( 2 , 2 , 2 , 2 ) ., Using this , the estimated distribution for position n is:This is a weighted average , where the weights are the posterior probabilities of each component given the data ., The posterior is: where , To cluster motifs , we implemented a hierarchical agglomerative clustering algorithm , using various motif comparison scores ., In each iteration , the algorithm computes the similarity between all pairs of motifs and then merges the pair with the highest similarity score into a new motif ( see Figure 4A ) ., This merge includes aligning the motifs according to the best scoring alignment between them , and then combining the evidence from both of them , by summing their nucleotide counts at each position ( i . e . , th | Introduction, Results, Discussion, Methods | Characterizing the DNA-binding specificities of transcription factors is a key problem in computational biology that has been addressed by multiple algorithms ., These usually take as input sequences that are putatively bound by the same factor and output one or more DNA motifs ., A common practice is to apply several such algorithms simultaneously to improve coverage at the price of redundancy ., In interpreting such results , two tasks are crucial: clustering of redundant motifs , and attributing the motifs to transcription factors by retrieval of similar motifs from previously characterized motif libraries ., Both tasks inherently involve motif comparison ., Here we present a novel method for comparing and merging motifs , based on Bayesian probabilistic principles ., This method takes into account both the similarity in positional nucleotide distributions of the two motifs and their dissimilarity to the background distribution ., We demonstrate the use of the new comparison method as a basis for motif clustering and retrieval procedures , and compare it to several commonly used alternatives ., Our results show that the new method outperforms other available methods in accuracy and sensitivity ., We incorporated the resulting motif clustering and retrieval procedures in a large-scale automated pipeline for analyzing DNA motifs ., This pipeline integrates the results of various DNA motif discovery algorithms and automatically merges redundant motifs from multiple training sets into a coherent annotated library of motifs ., Application of this pipeline to recent genome-wide transcription factor location data in S . cerevisiae successfully identified DNA motifs in a manner that is as good as semi-automated analysis reported in the literature ., Moreover , we show how this analysis elucidates the mechanisms of condition-specific preferences of transcription factors . | Regulation of gene expression plays a central role in the activity of living cells and in their response to internal ( e . g . , cell division ) or external ( e . g . , stress ) stimuli ., Key players in determining gene-specific regulation are transcription factors that bind sequence-specific sites on the DNA , modulating the expression of nearby genes ., To understand the regulatory program of the cell , we need to identify these transcription factors , when they act , and on which genes ., Transcription regulatory maps can be assembled by computational analysis of experimental data , by discovering the DNA recognition sequences ( motifs ) of transcription factors and their occurrences along the genome ., Such an analysis usually results in a large number of overlapping motifs ., To reconstruct regulatory maps , it is crucial to combine similar motifs and to relate them to transcription factors ., To this end we developed an accurate fully-automated method , termed BLiC , based upon an improved similarity measure for comparing DNA motifs ., By applying it to genome-wide data in yeast , we identified the DNA motifs of transcription factors and their putative target genes ., Finally , we analyze motifs of transcription factor that alter their target genes under different conditions , and show how cells adjust their regulatory program in response to environmental changes . | computational biology/sequence motif analysis, genetics and genomics/gene expression, computational biology/transcriptional regulation, biochemistry/transcription and translation, computational biology/systems biology | null |
journal.pcbi.1004801 | 2,016 | Early Transcriptome Signatures from Immunized Mouse Dendritic Cells Predict Late Vaccine-Induced T-Cell Responses | The development of vaccines against complex chronic diseases such as HIV or cancer has been largely unsuccessful so far ., Novel vaccine technologies are rationally designed to generate appropriate protective immune responses 1 , notably efficient T-cell responses ., Such vaccine vectors include plasmid DNA , viral and bacterial vectors , and virus-like particles ( VLPs ) ., The intrinsic characteristics of these vectors , including their capacity to stimulate innate immunity and to activate and target the antigen to antigen-presenting cells , determine in large part their immunogenicity and thus their potency as vaccine or gene therapy vectors 2–4 ., However the rational design of vectors is limited by various aspects , such as the partial understanding of the factors governing the induction of optimal immunity ( i . e . the activation of the innate immune system by various vector components , the effect upon adaptive immunity… ) or the possible dependence of vector efficacy on the specificity of the target diseases ., Systems biology has been introduced in vaccine development to assist in circumventing these limitations and shorten the vaccine development process ., Systems biology may not only help to better understand , analyze and reconstruct the complex immune interactions between the pathogen/vaccine and host immune system , but may also improve the in silico testing models for vaccine candidates ., Systems biology approaches have proven capable to predict immune responses induced after vaccination 5 , 6 ., For example , expression patterns of genes associated with the efficient processing of peptides for major histocompatibility complex presentation have been identified as useful surrogate markers of vaccine efficacy , obviating the need to perform challenge studies 7 ., Signatures derived from antibody repertoire profiling on peptide microarrays during the natural course of influenza infection were shown to be predictive of the efficacy of influenza vaccines 8 ., Multivariate analysis performed on human peripheral blood mononuclear cell ( PBMC ) microarray data , obtained 3 days after vaccination , identified innate immune response–related signatures that predicted the late adaptive immune response to the YF-17D yellow fever vaccine 9 ., In this manuscript , we describe a methodology that enabled us to successfully predict the adaptive immune responses induced by large sets of vaccine vectors of different classes , ranging from infectious particles to VLPs and DNA ., All these vectors expressed the same antigen , the immune response to which was measured using a validated standardized method ., We developed our model based on the analysis of transcriptomic data , obtained 6 hours after vaccination , that could predict the antigen-specific immune responses induced at the peak of the response , 5–10 days later ., It is noteworthy that this model , developed in mice , successfully predicted vaccine-induced responses from literature-mined human datasets ., Forty-one vectors classified in 13 categories of vaccines and all expressing the same antigen were evaluated and compared for their ability to induce an adaptive T-cell immune response after vaccination ( S1 Table ) ., The forty-one vectors included, ( i ) recombinant viral vectors derived from adenovirus ( rAd ) , vaccinia ( VACC ) , modified vaccinia Ankara ( MVA ) and lentivirus ( LV ) ,, ( ii ) recombinant bacteria vectors derived from Bacille de Calmette et Guérin ( BCG ) ,, ( iii ) recombinant VLPs made of the AP205 10 or Qbeta ( Qb ) 11 proteins from bacteriophage , the VP2 proteins from murine polyoma virus ( MPY ) 12 or murine pneumotropic virus ( MPT ) , the Gag capsid proteins from murine leukaemia virus ( MLV ) 13 , the core from hepatitis B virus ( HBc ) , and, ( iv ) plasmid encoding a recombinant protein ( DNA ) or recombinant MLV-VLPs ( plasmoVLPs ) 13 , 14 ., Each vaccine platform was engineered to display or express the immunodominant LCMV gp33-41 epitope model antigen 15 in order to compare the different vaccine-induced CD8+ T-cell specific responses ., In the framework of CompuVac ( www . compuvac . eu ) , we standardized the method for measuring the gp33-41-specific T-cell response using tetramer staining ( Fig 1A ) ., Mice were immunized with each vector and we evaluated the gp33-41-specific T-cell response in PBMCs at days 5 , 7 and 10 , following the frequency of circulating gp33-41/H-2Db tetramer+ CD8+ T cells ., In each experiment we included control mice that were injected with PBS or rAd ( rAd_1 batch ) to provide negative and positive controls ., Data for each experimental group were normalized as the experimental to rAd vector response ratio allowing cross-laboratory data comparisons ., We observed a wide range of immune responses that were triggered by the different vectors ., The maximal CD8+ T-cell expansion was induced with bacteriophage-derived VLPs , while very low but significant responses were observed with MPT and HBc VLPs ( Fig 1B ) ., Interestingly , different vector designs within the same vector platform led to different responses ., As an example , Qb-derived VLPs induced variable CD8+ T-cell expansion depending on their production processes that were designed to modify their TLR-ligand composition ( i . e . Qb_5 devoid of viral RNA and CpG in contrast to Qb_1; Fig 1A ) ., We took into consideration all the vectors and performed hierarchical clustering on normalized values that defined 3 clusters ( C ) ., The first cluster comprised vectors with low ratio values , characterizing weak inducers of antigen-specific T cells , hereafter referred as “Weak” vectors ., The other 2 clusters included vectors inducing high or intermediate responses , defining the “Strong” vector class ., This class comprised the different recombinant viral vectors ( rAd , MVA , VACC , LV ) expressing rather than displaying the antigen , and which have been extensively developed as CD8+ T-cell vaccines 16–18 ., It also contained bacteriophage-adjuvanted VLPs , in agreement with previous reports 10 , 19 ., As dendritic cell activation is key to the initiation of immune responses , we investigated whether transcriptome data from sorted spleen dendritic cells ( DCs ) sampled 6 hours after immunization could be predictive of the antigen-specific T-cell response measured several days later , at the peak of the response ., To address this question , we devised a stepwise modelling scheme ., DC-sorted transcriptome datasets were initially produced for 19 vectors on the Codelink platform , corresponding to 7 different vaccine platforms , for which the antigen-specific T-cell response was also measured ( S1 Table ) ., The rationale for looking at signatures instead of individual genes was motivated by, ( i ) the need to detect slight gene expression modifications ( captured as the overall expression changes of correlated genes ) ,, ( ii ) the technical constraints of working on different microarray platforms ( CodeLink , Illumina and Affymetrix ) , and, ( iii ) the objective of producing a predictive model working across microarray platforms ., Thus , our modelling scheme was based on our recently described strategy for signature discovery , using independent component analysis ( ICA ) followed by gene set enrichment analysis ( GSEA ) 20 ., This allows circumventing the limitations due to the use of different platforms when analyzing individual gene expressions , by comparing statistical signature’s enrichment across datasets ., ICA is an unsupervised algorithm extracting independent components Y from original datasets X by searching for the demixing matrix W:, Y=X×W, W matrix is calculated by maximizing the non gaussianity of the components measured as the negentropy J:, J ( y ) =H ( yGauss ) −H ( y ) ,, where H ( y ) and H ( yGauss ) are the Shannon entropy for a vector y and a random Gaussian vector with same variance as y 21 ., The use of ICA to analyze microarray data is justified by the hypothesis that X is a mix of signals from underlying cellular pathways ., Therefore , columns of Y contain a summary of gene contributions in the extracted components ., The RNA expression value of a gene is thus the superposition of several signals of this gene in each component which add up ., From each component y , two reduced gene sets can be extracted by selecting genes with critical contribution on both sides of the distribution 22 ., We first performed ICA on the 19 available datasets , yielding 210 molecular signatures characterizing the variability within each dataset , and likely linked to vector properties ., We then analyzed the differential gene expression between the controls and the tested vectors using bootstrapping 23 , 24 , in order to increase the model’s sensitivity ., Bootstrapping consists in sampling series of additional datasets by randomly drawing samples with replacement of equal size from an original dataset , as described in Fig 2 ., We sampled 100 consecutive bootstrapped datasets from each of the 19 original datasets and generated 100 corresponding ranking lists of genes based on modified t-test statistics ., The previously identified signatures were then tested for their behavior vis-à-vis the gene lists using GSEA , generating normalized enrichment scores ( NES ) ., Molecular signatures from GSEA software ( >5000 ) were added at this step in order to increase the efficiency of the normalization procedure ., NES of molecular signatures from ICA were then extracted for the next steps ., This yielded a matrix , containing 1900 columns ( 100 bootstrapped datasets for each of the 19 original datasets ) and 210 lines ( the number of calculated NES ) ., This matrix was then used to create random forest ( RF ) classification models ( Fig 2 ) ., NES values and T-cell response classification were used as predictors and dependent variables , respectively , in the randomForest package , which as output provides classification results and associated probabilities for each T-cell response class ., An initial predictive model was built with 9 vector datasets ( in red in Fig 1B ) for which the antigen-specific T-cell responses were available ( 900 bootstrapped datasets and 100 signatures ) ., Predictions of 10 additional datasets , including independent experiments done with the same or different batches of these vectors , were very consistent ( see Tables 1 and S2 ) ., The model sensitivity for the “Weak” and “Strong” vector classes ( respectively equal to the specificity for the “Strong” and “Weak” classes ) are 0 . 89 and 0 . 98 , respectively ., The positive predictive value ( PPV ) is stable for the two classes ( “Weak”: 0 . 96 , “Strong”: 0 . 93 ) ., This 9-vector model is already efficient to classify the vector platform with 0 . 94 accuracy ., These results led us to construct the final predictive model ( called RFM model ) including all the 19 datasets , based on the analysis of the 210 signatures across the 1900 bootstrapped datasets ., This complete training set contained enough information to discriminate clearly between the 2 vector classes , as demonstrated by the misclassification rate parameter reaching zero after 100 simulated trees ., The RandomForest algorithm provides a ranked list of the signatures based on their importance to the efficacy of the classification in the model ., This score is based on the decrease of the Gini impurity criterion for each child node of a split ., The result of this calculus is the mean of this decrease for each signature present in the trees of the forest ., 27 most important signatures , having a mean decrease score higher than ten , were selected ., Clustering methods were then applied, ( i ) on NES values of these 27 signatures calculated on original datasets ( Fig 3A ) and, ( ii ) on the mean NES values calculated on the bootstrapped datasets ( Fig 3B ) ., The interest of bootstrap is clearly revealed with clusters more explicitly defined after bootstrap ., We then asked whether RFM was biased toward particular vector datasets ., We first used the leave-one-out methodology , where 19 models were iteratively built using only 18 out of 19 datasets , and then assessing how accurately such models predict the 100 bootstraps from the left-out dataset ., All vectors were classified as expected for at least 96 of the 100 bootstrapped datasets , except MPY_3 for which 16 bootstrapped datasets were misclassified ( S3 Table ) ., This result shows overall very high prediction stability and no significant bias of the RFM model ., We verified that RFM was not biased for a given vector platform ., One hundred new models were constructed , each based on one randomly selected representative of the 7 vector platforms ( rAd , AP205 , MVA , MPY , MPT , MLV and BCG ) ., For each vector , the probabilities to be classified as expected were calculated and the prediction distribution across the 100 models is shown in Fig 4 ., Vaccines from the “Strong” vector class ( in red ) showed good consistency in their prediction distribution , with no value under 0 . 6 ( 100% confidence ) ., Vaccines from the “Weak” vector class showed more variability: in particular , 2 MPY vaccines ( MPY_3 & MPY_3bis; same vector batch ( #3 ) used in 2 independent experiments ) were not classified as expected in 16 models out of 100 ( 84% confidence ) ; these 16 misclassifying models all used MPY_2 as the MPY representative ., Note that this specific preparation ( #2 ) of MPY vaccine was produced using baculovirus machinery in insect-derived cells , while the other MPYs were produced in yeast ., RFM was then used to predict the vector class of 4 new vectors belonging to 3 vector platforms: 2 batches of lentivirus ( LV ) vectors -a category of vaccine not represented during the model establishment , one new batch of AP205 ( AP205_3 ) and one of MLV ( MLV_2 ) ., We had independently determined that LV vectors induced strong antigen-specific T-cell responses after immunization and were classified in the “Strong” vector class ( Fig 1 ) ., As shown in Tables 1 and 2A , these 4 bootstrapped datasets were classified as expected with high precision ( >95% ) while sensitivity and PPV of the model increased compared to the 9-vector model , especially the sensitivity for the “Weak” vector class now reaching 0 . 97 ( from 0 . 89 ) with RFM ., These results highlight that RFM, ( i ) is not vaccine platform-dependent ,, ( ii ) correctly predicts a vector platform unknown to the model , and, ( iii ) efficiently predicts both “Weak” and “Strong” vectors ., RFM was built on transcriptome data obtained from sorted spleen DCs ., In our next experiment , we assessed whether RFM would be sensitive enough to classify transcriptome datasets derived from whole spleen samples obtained 6 hours after immunization , where DCs represent 1–2% of total splenocytes ., As summarized in Table 2B , all bootstrapped datasets from whole spleens were well classified , with at least 91% of the expected classification , thus demonstrating our model’s sensitivity in classifying vectors in whole spleen transcriptome datasets ., We then tested microarray datasets for whole spleen samples obtained 6 , 48 and 72 hours after vaccination with one vector , the rAd vector that we used as a standard ., Strikingly , only datasets sampled 6 hours after injection were classified as expected ( as “Strong” ) ( Table 2D ) ., Similarly , we tested the performance of our model in classifying vectors using PBMC-derived microarray datasets ., The rationale for this experiment is that PBMCs , less than 1% of which are DCs , offer a more accessible sample source than spleen , especially in humans ., As shown in Table 2C , all but one vectors were classified as expected with high precision ( ≥ 90% ) ., AP205_1 was classified as expected , though with less confidence ( 73% ) ., Finally , we tested whether our model could classify datasets obtained from the literature ., We found datasets from the Merck Ad5/HIV trial reported by Zak et al . 25 PBMC transcriptome data were generated from samples obtained at 6 , 24 and 72 hours after vaccination ., We bootstrapped the samples of Zak et al . , taking patient-paired samples before and after vaccination ., 100% and 91% of the bootstrapped paired samples were predicted as “Strong” at 24 and 72 hours , respectively ( Table 3 ) , in line with the authors’ original observations ., The same analysis performed with the 6-hour time point gave a “Strong” prediction for 31% of the bootstrapped paired samples ., The latter finding is consistent with the conclusion of Zak et al . that transcriptomic modifications at 6 hours were not significant ., These results demonstrate the capacity of RFM generated from mouse DC transcriptome datasets to classify human PBMC datasets ., Biological annotation of the 27 most important signatures of RFM reveals one signature ( Sig1 ) with statistical functional enrichments related to immune processes ( FDR p-values 10−4–10−8 ) ., This signature is highly focused on STAT-1 with 51 genes having strong biological connections ( Fig 5A ) ., Interestingly , Sig1 is upregulated in all the vectors , but with higher intensity in the “Strong” as compared to the “Weak” vectors ., No specific molecular pathway was clearly identified by QIAGEN’s Ingenuity Pathway Analysis ( IPA ) functional analysis for the other 26 important signatures in our model ., However , visual inspection of these signatures identified the CH25H gene as highly modulated by strong vectors ., Since this gene has been recently described as playing a role in DC maturation 26 , we analyzed its network of connected genes with IPA ( Fig 5B ) ., This network was also globally more modulated by “Strong” rather than “Weak” vectors , and comprised genes implicated in DC function such as MYD88 , DUSP5 and ABCG1 ., Understanding and predicting innate immune response to vector platforms is primordial for fast and effective production of new vaccination or gene therapy protocols ., Systems biology tools efficiently extract information from large datasets in computing predictive models and have already played a major role in recent discoveries in this field 5 , 27 ., In this paper , we initially focused on early transcriptomic changes of DCs since these are first-line players in the innate immune response and directly contribute to the triggering of the adaptive response ., Our aim was to identify transcriptomic signatures predictive of the late CD8+ CTL responses to the LCMV gp33-41 model antigen conveyed by a variety of vaccine vectors ., Based on molecular signatures extracted using the non-supervised ICA method 20 , 22 , we produced and validated a prediction model taking into account 19 available datasets generated with different vector platforms ., We chose the random forest learning algorithm for its reported efficiency among classification methodologies 28–30 ., The originality of our strategy was the use of signatures rather than genes to classify samples ., Our results showed that this model consistently predicts both “Weak” and “Strong” vectors , with greater confidence for the latter ., This suggests that there are shared gene expression modifications induced by “Strong” vectors , while changes induced by “Weak” vectors are more diverse ., Consistent with this , Li et al . recently reported that different types of vaccine lead to different transcriptomic modifications in humans 3 days after vaccination 31 , with vaccines inducing high transcriptomic modifications being those that induce robust antibody responses ., Among the 27 signatures selected for their importance in the RFM model , one ( Sig1; see S4 Table ) is related to immune components , including “viral infection” , “role of RIG1-like receptors in antiviral innate immunity” and “interferon signalling” pathways ., Previous studies have characterized gene expression modifications in the early stages of vaccination consistent with Sig1 annotation ., Querec et al . investigated the transcriptome of patient PBMCs at days 0 , 1 , 3 , 7 and 10 after vaccination with yellow fever vaccine 9 ., Of 65 regulated genes , 26 were related in part to interferon and the antiviral response , including MX1 , IFIT1 , IFIT2 , IFIT3 , OAS1 , OAS2 , OAS3 and OASL , and 7 were related to signal transduction , including STAT1 and IRF7 ., Similarly , Zak et al . 25 applied the modular transcriptome analysis framework described in Chaussabel et al . 32 to study the innate immune response to MRKAd5/HIV in PBMCs 6 , 24 , 72 and 168 hours after patient vaccination ., They identified genes highly regulated at 6 and 24 hours , including STAT1 , STAT2 , IFITs , MXs and OASs ( also identified in Querec et al . ) ., Strikingly , all these genes are also part of Sig1 , emphasizing further their key role in the early response to the vaccine ., Furthermore , DDX60 , a newly described antiviral factor that induces Rig-1-like receptor-mediated signaling 33 , present in Sig1 , was reported by Querec et al . as well 9 ., Interestingly , Sig1 is upregulated in vaccinated samples compared to control group , but to a lesser extent in “Weak” vs . “Strong” vectors ( see Figs 3 and 5 ) ., Our cross-analysis of Zak et al . ’s microarray data on Merck Ad5/HIV-vaccinated human PBMC samples , which yield good predictions for the 24- and 72-hour time points , demonstrates that our prediction model , solely based on mouse DC-sorted transcriptome data , efficiently predicts human transcriptome data ., This can be explained by the high similarity of gene expression in immunological cell lineages between mice and humans 34 , although the kinetics of the immune response to vaccine is different ., No specific molecular pathway was clearly identified by IPA annotations for the other 26 important signatures in our model ., This is somewhat surprising since these signatures have been selected by the model to best distinguish “Strong” and “Weak” vectors and are therefore expected to represent differentially regulated biological pathways ., In this line , none of the 27 signatures corresponds to a peculiar behavior of a vector but they rather reveal similar behavior within “Strong” or “Weak” groups ( Fig 3 ) ., Moreover , the identified signatures were extracted from 13 out of 19 different vector datasets ( 9 “Strong” and 4 “Weak” vectors ) ., We believe that these signatures are unlikely artifactual but related to yet undefined biological processes ., Indeed , the constant improvement of annotation databases can reveal secondary or additional functions of genes ., For example , CH25H , a gene found in one of the 26 signatures and clearly upregulated in “Strong” vectors , is primarily involved in cholesterol metabolism , but has recently been shown to play a role in the early stage of DC maturation 26 ., Fig 5 shows how the expression of this gene is related to dendritic cell through direct or indirect interactions with STAT-1 or IFNγ , both members of Sig1 , and with several genes known to be important in early dendritic cell activation: for example , MYD88 is a gene involved in toll-like receptor signaling 35 , DUSP5 is known to be upregulated during dendritic cell maturation 36 , and ABCG1 is a gene playing a role in adaptive immune responses 37 ., The comparative analysis of gene expression modulation of this interaction network shown in Fig 5 reveals a similar pattern of differential expression for “Strong” vectors ( rAd_1 , AP205_1 ) different than that observed for “Weak” vector datasets ( BCG_2 , MPY_3bis ) ., This again points at a significant difference in early dendritic cell activation-related gene behavior in “Strong” vs . “Weak” vectors ., Altogether , our results underline the relevance of the CompuVac initiative that consisted in producing , in a standardized manner , immunological and transcriptome data related to vaccine candidates in order to predict their capacity to elicit strong antigen-specific responses ., Our model was based on transcriptome data from sorted spleen DCs of mice vaccinated with various “Strong” and “Weak” T-cell inducer vectors ., This prediction model accurately predicted the behavior of these and other candidate vaccines only 6 hours after injection ., The model was powerful enough to produce a relevant vector classification even when using whole mouse spleen and PBMCs , or even human PBMCs ( Fig 6 ) , and across 3 microarray platforms ( CodeLink , Illumina and Affymetrix ) ., The accuracy and sensitivity of the model are likely high because it is built with very different vaccine platforms therefore representative of possible vector behaviors in triggering the early immune response ., This study further supports the potential of systems immunology approaches in facilitating the development and characterization of vaccines , offering robust in silico solutions to study the early events of the immune response to vaccines ., Experimental protocols complied with French law ( Décret: 2001–464 29/05/01 ) and EEC regulations ( 86/609/CEE ) for the care and use of laboratory animals and were carried out under Authorization for Experimentation on Laboratory Animals Number 75-673-R ., Our animal protocol ( Ce5/2009/042 ) was approved by the “Charles Darwin” Ethics Committee for Animal Experimentation ( CNREEA 05 ) and performed in the licensed animal facility A75-13-08 ., Recombinant adenovirus- and MVA-derived viral vectors , BCG-derived bacterial vector , AP205 10 or Qb 11 bacteriophage- , MPT- and MPY- 12 or MLV-derived 13 VLPs used as an antigenic platform and DNA vaccines were included in this study ., According to the CompuVac evaluation scheme , each vaccine platform was engineered to display / express the LCMV gp33-41 model antigen 15 in order to measure the vaccine-induced T-cell specific responses and dendritic cell transcriptome changes ( see following sections ) ., The sequence IITSIKAVYNFATCGILAL corresponding to the GP33-41 epitope flanked upstream and downstream by 5 of its natively neighboring amino acids was used ., The 53 vectors considered in this paper ( S1 Table ) are displayed in 13 vector platforms 7 of which were used for a training set ( rAd , MVA , AP205 , MPT , MPY , MLV and BCG ) and 2 for prediction of new platforms ( LV and Qb ) ., Groups of three to five 7-week-old female C57BL/6 mice ( Charles River , France and Germany ) were immunized with a controlled quantity of vector particles as defined in CompuVac assay protocols ( www . compuvac . eu ) ., For monitoring T-cell responses , each vector was injected with its “best” route of administration: subcutaneously for VLP vectors; intramuscularly for recombinant antigen-expressing vectors and by intra-dermally by gene gun for DNA vaccines ., Control mice were injected with 100 μL of phosphate buffered saline solution ( PBS ) ., For each vector ( n = 41 ) , the T-cell immune response measurement was performed independently one to three times ., T-cell immune responses induced against the LCMV gp33-41 model antigen were measured by MHC-I gp33-41/H-2Db tetramer ( ProImmune , UK ) staining of PBMCs at 5 , 7 and 10 days after injection ., The highest measure was kept for each mouse and the mean value was then calculated for the group ., Values were normalized against measures monitored in parallel in mice immunized with the rAd_1 control vector ., Experimental groups comprised of 3 to 6 mice immunized with vaccine candidates by the intravenous route ., Mice were sacrificed 6 hours after immunization ., Spleen DCs were purified with CD11c+-conjugated MACS magnetic beads ( Miltenyi Biotec ) according to the manufacturers instructions ., After incubation for 20 minutes at 4°C , cells were washed and passed over a MACS column ., Purity was checked routinely by FACS and found to be greater than 96±2% ., 2x106 CD11c+ cells were used for total RNA extraction using Nucleospin RNAII ( Macherey Nagel ) ., For test dataset generation , whole PBMCs and/or whole splenocytes and/or sorted spleen DCs were collected at 6 hours , and at 48- and 72-hour time points for the kinetic follow-up ., RNA was checked for quality using gel electrophoresis and for quantity using a Nanodrop spectrophotometer ( Thermo Scientific ) ., Microarrays were performed using either Applied Microarrays ( CodeLink Mouse Whole Genome Bioarray ) or Illumina ( WG6 Mouse BeadArray ) technologies ( S1 Table ) ., The MessageAmp II aRNA Amplification Kit ( Ambion ) was used for cDNA and cRNA production from 1 μg of total RNA ., 10 μg of amplified cRNA was subsequently fragmented and hybridized for 20 hours using the Applied Microarrays hybridization and washing buffer kit ., Slides were scanned using the GenePix Personal 4100A scanner for CodeLink array or the Illumina BeadArray 500GX Reader for Illumina array ., Hybridization and raw data extraction were performed using either GenePix Pro 6 . 0 ( for CodeLink array ) or BeadStudio ( for Illumina array ) software , respectively ( GEO accession GSE66991 ) ., Each tested vector dataset comprised “vector-immunized” and corresponding PBS control samples ., Quantile normalization was performed with the limma package 38 on R software 39 , and then a log2 transformation was applied ., Probes with a detection p-value above 0 . 05 in all samples in a dataset were discarded ., Following our two-step ICA→GSEA signature discovery strategy 20 , signatures were extracted using the fastICA algorithm R package 40 following modifications in 22 ., Parameters were set as default , except for the unmixing matrix A-1 convergence threshold set to 10-6 ., Ranked gene lists were calculated using the limma modified t-test ., ES were calculated using GSEA 41 with the pre-ranked gene list protocol ., Normalized ES are then calculated based on the permutation performed on gene sets collection , allowing comparison between experiments ., The ICA-extracted signature database was complemented with the MsigDB C2 ( curated gene sets of biological pathways ) and C5 ( Gene Ontology gene sets ) databases ( www . broad . mit . edu/gsea ) in order to increase universe of genes available for permutation of gene sets ., Signatures with fewer than 7 detected genes were ignored ., For each model produced in the Results section , classification was performed on a matrix of fastICA extracted signature NES values ( see above section ) calculated for bootstrapped vector datasets ( 100 bootstraps per vector dataset ) , using the random forest algorithm implemented in the randomForest R package to produce a forest of 2000 trees 42 ., The number of randomly selected signatures used at each of the 2000 runs was set according to the mtry function implemented in the randomForest package ., The class prediction of the new dataset was deduced by the probability to be “Weak” or “Strong” > 0 . 5 ., The overall vector class was then obtained as the majority of “Weak” or “Strong” class assignments over the 100 bootstraps ., For classification model validation , we implemented the leave-one-out methodology consisting in creating models with n-1 datasets , where n is the total number of datasets , and classifying the dataset left out ., In addition , we implemented a “multi-model” methodology based on the classification of bootstrapped datasets over 100 models created as above ., Each model was computed on an NES matrix of a random selection of one representative vector dataset of each of the 7 represented vector platforms ( see Vector platforms section and S1 Table ) ., Vector mean probabilities were calculated as the average probability of being “Weak” or “Strong” over the 100 bootstrapped vector datasets , and their distribution over the 100 models was analyzed ., For biological insight evaluation of the signatures , microarray data were analyzed through the use of QIAGEN’s Ingenuity Pathway Analysis ( IPA , QIAGEN Redwood City , www . qiagen . com/ingenuity ) . | Introduction, Results, Discussion, Material and Methods | Systems biology offers promising approaches for identifying response-specific signatures to vaccination and assessing their predictive value ., Here , we designed a modelling strategy aiming to predict the quality of late T-cell responses after vaccination from early transcriptome analysis of dendritic cells ., Using standardized staining with tetramer , we first quantified antigen-specific T-cell expansion 5 to 10 days after vaccination with one of a set of 41 different vaccine vectors all expressing the same antigen ., Hierarchical clustering of the responses defined sets of high and low T cell response inducers ., We then compared these responses with the transcriptome of splenic dendritic cells obtained 6 hours after vaccination with the same vectors and produced a random forest model capable of predicting the quality of the later antigen-specific T-cell expansion ., The model also successfully predicted vector classification as low or strong T-cell response inducers of a novel set of vaccine vectors , based on the early transcriptome results obtained from spleen dendritic cells , whole spleen and even peripheral blood mononuclear cells ., Finally , our model developed with mouse datasets also accurately predicted vaccine efficacy from literature-mined human datasets . | Vaccines are designed to elicit effective immune responses against antigens ., The various vector platforms used in vaccine development are diverse and complex , rendering the selection of promising vaccines vector challenging ., We have designed a modeling strategy that predicts the propensity of vaccine vectors to elicit strong late T-cell responses using transcriptome material obtained 6 hours after vaccination ., Our model , designed with mouse datasets , also predicted vector efficacy from mined human data ., Thus , molecular signatures obtained 6 hours after vaccination can predict vaccine efficacy at 2 weeks post vaccination , which should help in vaccine development . | blood cells, medicine and health sciences, immune cells, immune physiology, spleen, immunology, vaccines, preventive medicine, mathematics, forecasting, statistics (mathematics), genome analysis, bioassays and physiological analysis, vaccination and immunization, research and analysis methods, public and occupational health, white blood cells, genomics, animal cells, mathematical and statistical techniques, t cells, immune response, microarrays, cell biology, physiology, transcriptome analysis, genetics, biology and life sciences, cellular types, physical sciences, computational biology, statistical methods | null |
journal.pcbi.1002848 | 2,013 | Network Self-Organization Explains the Statistics and Dynamics of Synaptic Connection Strengths in Cortex | The computations performed by cortical circuits depend on their detailed patterns of synaptic connection strengths ., While the gross patterning of connections across different cortical layers has been well described in some cases 1 , 2 , the detailed connectivity structure between groups of cells and its relation to information processing have been notoriously difficult to investigate 3 ., This detailed structure could either be largely random – the product of somewhat arbitrary growth processes , or it could be highly organized ., On the one hand , randomly structured networks have been shown to possess powerful computational properties 4–6 and they are easy to generate ., On the other hand , a precise non-random organization could be the product of network self-organization , where network structure determines neural activity patterns and activity patterns in turn shape network structure through plasticity mechanisms ., At the macroscopic and mesoscopic scales , models based on self-organization have already explained fundamental features of brain networks ., Examples are the formation of topographic mappings 7 or properties of orientation preference maps in primary visual cortex 8 , 9 ., Here we show that fundamental aspects of the microscopic structure of cortical networks can also be understood as the product of self-organization ., Self-organization typically relies on a combination of self-reinforcing ( positive feedback ) processes that are combined with a competition for limited resources ., In the context of Neuroscience , an example of a self-reinforcing process may be that correlated firing of two groups of neurons may strengthen synaptic connections between them according to Hebbs postulate of synaptic plasticity , while the strengthened connections will in turn amplify the correlated firing of the neurons ., An example for competition for a limited resource may be a synaptic scaling mechanism that limits the sum of a neurons synaptic efficacies such that one synapse can only grow at the expense of others ., The combination of self-reinforcing mechanisms with limited resources often gives rise to the formation of structural patterns , which may or may not have specific functional advantages ., Here , we will offer an explanation for fundamental aspects of the fluctuations of synaptic strength and the distribution of synaptic efficacies based on self-organization ., Specifically , recent evidence shows that the distribution of synaptic efficacies is highly skewed 10 , 11 , having an approximately lognormal distribution 12–14 ., Only around 20% of synapses are responsible for 50% of total synaptic weight ., Importantly , synaptic contacts are constantly being created and destroyed and sizes of dendritic spines are fluctuating over time scales of hours and days 14 , 15 ., In the face of this highly dynamic network structure , stable long-term memories are thought to be based on subsets of synapses with long life times 16 , 17 , which may also be comparatively strong 16 ., In line with this , the daily fluctuations of dendritic spine sizes , which are closely related to synaptic efficacies , are such that weak synapses can change their size by as much as a factor of 6 , while strong synapses are much more stable 15 ., To investigate whether and how these properties can arise from self-organization induced by neuronal plasticity mechanisms , we have developed a self-organizing recurrent network ( SORN ) model ., It extends a previous model 18 , and consists of noisy binary threshold spiking neurons ( 80% excitatory and 20% inhibitory ) and uses five different forms of plasticity ( see Materials and Methods for details ) ., Connections between excitatory neurons are subject to an additive spike-timing dependent plasticity ( STDP ) rule that changes synaptic strength in a temporally asymmetric causal fashion as reported experimentally 19 , 20 ., A synaptic normalization mechanism keeps the sum of all excitatory weights to a neuron constant and models classic findings on multiplicative synaptic scaling of synaptic efficacies 21 , 22 ., An intrinsic plasticity mechanism adjusts the firing thresholds of excitatory neurons to maintain a low average firing rate ., This mechanism models homeostatic changes in neuronal excitability through modification of voltage gated ion channels observed experimentally 23 , 24 ., Connections from inhibitory neurons onto excitatory neurons are subject to an inhibitory spike-timing dependent plasticity ( iSTDP ) rule that balances the amount of excitatory and inhibitory drive that the excitatory neurons receive as reported in recent studies 25–27 ., Finally , a structural plasticity rule generates new synaptic connections between excitatory cells at a small rate ., This models the constant generation of new synaptic contacts observed in cortex and hippocampus 15 , 28 ., We simulated networks of 200 excitatory and 40 inhibitory neurons for 10 , 000 time steps and observed the resulting activity patterns ( Fig . 1 ) and distributions of synaptic strength ( Fig . 2 ) ., The network shows irregular activity patterns reminiscent of cortical recordings ( Fig . 1A ) ., Inter-spike interval ( ISI ) distributions are well fitted by an exponential function ( Fig . 1B ) and coefficient of variation ( CV ) values are close to one ( Fig . 1C ) as would be expected from a Poisson process ., Neurons show only very weak correlations of their firing during this phase of network development ( Fig . 1D ) ., To estimate the probability distribution governing excitatory-to-excitatory synaptic strengths we bin connection strengths and divide the number of occurrences in each bin by the bin size ., The bin sizes are uniform on the log scale ., To mimic experimental procedures 15 , very small synapses ( ) are excluded ., Fig . 2A–D shows the distribution of synaptic connection strengths after 10 , 000 time steps and compares it to EPSP data from rat visual cortex 12 ., With distinctly different initial conditions ( Fig . 2E ) , the network faithfully develops a long-tailed distribution of connection strengths that is similar to the biological data ( see Text S1 for details ) ., Experimental data and model results are both well fit by lognormal distributions ., As the network evolves it goes through different phases ( Fig . 3 ) ., The initial phase is characterized by a decay of connectivity , where a substantial fraction of the excitatory-to-excitatory synaptic weights get eliminated ( Fig . 3A ) ., In the subsequent growth phase , the network connectivity recovers through the integration of newly created synapses produced by the structural plasticity ., Eventually , the degree of connectivity stabilizes and the network enters into a stable regime ., Here , connectivity fluctuates very little ( Fig . 3A inset ) ., Newly created synapses tend to quickly disappear and there is a large stable backbone of connections with extremely long life times ( as long as we simulated ) ., The distribution of excitatory-to-excitatory connection strengths is lognormal-like throughout most of the networks evolution ( Fig . 3B–D ) ., ( see Fig . S2 in Text S2 for more results with different parameters ) ., An exception is the transition from the decay to the growth phase , where large deviations from the lognormal shape are observed ( not shown ) ., However , the distribution of synaptic weights maintains a long tail and a positive skewness throughout its development ., The thresholds of the excitatory units in the network develop an approximately Gaussian distribution ., In the stable regime of the network , this distribution is exhibiting only small fluctuations ., As a next step , we assessed the dynamics of synaptic connection strengths in SORN ., Fig . 4A shows traces of 6 synaptic connection weights as a function of time ., The distribution of life times of newly created synapses is well described by a power law with an exponent close to −3/2 during this phase as expected for random walk behavior ( Fig . 4B ) ., We next compared the weight changes occurring in SORN over 3000 time steps with experimental data from time lapse imaging of dendritic spine sizes in rat hippocampus 15 ., In both SORN and the experimental data , strong synapses are found to have comparatively small fluctuations ( Fig . 4C–F ) ., This is not a simple ceiling effect , since synaptic weights could , in principle , grow much larger than the typical values for very strong synapses we observe in the model , which lie between 0 . 2 and 0 . 3 ., There exists a small population of synaptic connections in both model and experimental data which decays completely ( horizontal lines in Fig . 4C , D and oblique lines in Fig . 4E , F ) ., The population of synapses clustered on the Y-axis in Fig . 4E , F represents newly established synaptic connections ., The big fluctuations are mostly seen in decay phase and imply that the network is far from stability in this regime ( see Fig . S6 in Text S2 for additional results with different parameters showing weight fluctuations during different phases of network evolution ) ., To better understand the mechanism through which the network self-organizes its connectivity and dynamics , we examined how the strength of a synaptic connection influences its probability of undergoing further growth or decline ., Among all the plasticity mechanisms , only STDP and synaptic normalization adjust the weights of EE connections ., While synaptic normalization will only scale all incoming excitatory-to-excitatory connections linearly , STDP has the power to change the shape of the distribution of synaptic weights impinging onto a neuron ., When we recorded the isolated effect of STDP , i . e . independently of the synaptic normalization , we found that over a large range of synaptic weight strengths , the expected increase in strength of a connection due to STDP grows approximately linearly with the strength of the synapse ( Fig . 5A ) ., The fraction of connections undergoing depression depends much less on connection weight ( Fig . 5B ) ., Thus , the net effect is that stronger synaptic connections have a higher chance to be potentiated by STDP establishing a rich-get-richer behavior ( Fig . 5C ) ., This mechanism is kept in check by the synaptic normalization mechanism , which scales weights in a multiplicative fashion ., We estimated the mean absolute change of synaptic connection strengths due to STDP and synaptic normalization over 200 time step intervals during the initial 10 , 000 time steps ., The mean absolute sizes of fluctuations grow roughly linearly with weight ( Fig . 5D ) as observed experimentally 14 ., Note that this approximately linear dependence on weight strength occurs despite the additive STDP rule we are using and does not require a multiplicative STDP rule 12 ., With all forms of plasticity present , the network will show irregular firing activity and develop a lognormal-like weight distribution ., These results are stable over a large range of parameter values ( see Text S2 for details ) ., To investigate the extent to which the different forms of plasticity contribute to these results , we performed simulations where we switched off individual plasticity mechanisms ., When synaptic normalization is switched off , the network will show bursts of high activity separated by long periods of inactivity ., As shown in Fig . 4 , the network keeps eliminating synapses as a result of STDP ., The structural plasticity counteracts this process ., If we switch off the structural plasticity , a large number of neurons eventually lose all their postsynaptic targets ., No lognormal-like weight distribution will emerge if one or both forms of plasticity are missing ., Intrinsic plasticity and inhibitory STDP both try to maintain a low average firing rate of excitatory cells and both are important to keep healthy network dynamics ., If both are switched off , some units will exhibit very high firing rates while others remain essentially silent and all the phenomena shown in Fig . 1–5 will disappear ., To study the individual effects of intrinsic plasticity and iSTDP , Fig . 6 shows a scatter plot of the fraction of active excitatory units in subsequent time steps ., With all plasticity mechanisms active , the network activity is confined within a small area ., Activity never dies out or becomes very big ., When either intrinsic plasticity or inhibitory STDP is switched off , the network activity exhibits big fluctuations and can temporarily die out completely ., In certain parameter regimes the network may function without one or the other , but with both mechanisms being present , we obtain robust results over a large range of parameter values ., We conclude that all five plasticity mechanisms are important for proper self-organization ., Understanding the structure and dynamics of neural circuits and reproducing them in neural network models remains a major challenge ., Classic models of STDP have been shown to lead to physiologically unrealistic bimodal weight distributions under certain conditions 29 ., This has lead to the proposal of a number of modifications to STDP rules to remedy the problem ., Specifically , multiplicative STDP rules have received much interest recently 30 , 31 ., Here we have shown that an additive STDP rule when operating together with other plasticity mechanisms in a recurrent network is sufficient to explain both the statistics and fluctuations of synaptic connection strengths observed in cortex ., Associative synaptic plasticity induces a rich-get-richer dynamics of synaptic weights , while homeostatic mechanisms induce competition ., With distinctly different initial conditions , the ensuing self-organization faithfully develops Poisson-like irregular firing patterns , lognormal-like weight distributions and the characteristic pattern of fluctuations of synaptic strengths reminiscent of cortical recordings ., Beyond this , our model predicts a power-law scaling of the lifetimes of newly established synaptic connections during development ., Our results suggest that the statistics and dynamics of neural circuits are the product of network self-organization , and that the combined action of multiple forms of neuronal plasticity plays an essential role in the formation and maintenance of cortical circuits ., It is important , however , to also consider alternative explanations ., One of the simplest ways to obtain lognormal distributions is by virtue of Gibrats law , which was originally developed in Economics ., It describes the growth of companies by random annual growth rates which are independent of the companies sizes ., This process by itself , when applied to the growth of synaptic connections , would predict that the variance of the synaptic weight distribution would grow without bounds , which is clearly at odds with biological reality ., Adding a multiplicative normalization mechanism such as our synaptic normalization rule to Gibrats proportionate growth process retains the development of a lognormal-like distribution while avoiding the problem of unbounded growth ., However , this model does not reproduce the pattern of weight fluctuations observed experimentally ., Furthermore , such a model is purely phenomenological and does not describe the mechanism that causes the synaptic fluctuations in the first place ., Similarly , the models proposed in 15 and 14 describe the fluctuations of synaptic weights as independent random walk processes , but do not explain what causes the synaptic fluctuations ., In contrast , our model offers a mechanistic account that explains the patterns of weight fluctuations and the distribution of synaptic strength in terms of fundamental processes of neuronal plasticity in a recurrent network ., This approach is consistent with the finding in 15 that the fluctuations of dendritic spine sizes seem to strongly depend on activity-driven synaptic plasticity ., Specifically , they found strongly reduced fluctuations of spine sizes and fewer spine eliminations when inhibiting NMDA receptors with APV or MK-801 ., Interestingly , the generation of new spines was unaffected by this manipulations ., This is consistent with our models assumption that the generation of new spines occurs via a process of structural plasticity that is independent of activity-driven synaptic changes ., A further advantage of our model is that it can also be used to derive predictions regarding the emerging network topology in terms of clustering , network motifs , etc ., This topic is left for future work ., If our model is essentially correct , despite its very abstract formulation , then one should be able to replicate the present results in more realistic network models of spiking neurons ., As a first step in this direction , we have constructed a version of the model using leaky-integrate-and-fire neurons with realistic parameter values ., We have also adapted the plasticity mechanisms for this network ., Initial explorations show that major features such as the lognormal-like weight distribution and the pattern of synaptic fluctuations can also be found in this less abstract network model ., Future work will elaborate on these preliminary results ., Since the structure of cortical circuits determines the dynamics of neuronal activity , it also determines how information is encoded and propagated ., The existence of a small number of very strong synaptic connections may greatly facilitate the highly reliable propagation of signals along pools of neurons 32 ., In fact , SORN networks have previously been shown to spontaneously develop encoding strategies based on trajectories through their high-dimensional state space of unit activations 18 ., In this work , the networks were fed with structured time series of input letters and were shown to learn internal representations of these input sequences that allowed large performance increases in prediction tasks ., This was found to be due to the ongoing self-organization in the network driven by the networks plasticity mechanisms ., They were shown to effectively increase the separation of network states belonging to different input conditions ., More recently , we have found evidence that such networks may naturally self-organize to perform computations resembling Bayesian inference processes 33 ., Further work is needed to better understand how the networks self-organization enables it to behave this way ., Many computational models of local cortical circuits assume random network structure 4–6 , sometimes with distance-dependent or layer-dependent connection probabilities 34 ., Such random network structure is at odds with recent evidence that changes to the connectivity structure such as the generation of stable new spines are associated with the formation of new memories 35 ., Hence , we believe that the study of random networks where only connection statistics are matched to those in the brain , may be quite misleading when the goal is to understand processing in cortical circuits ., Instead , self-organizing networks , which can faithfully develop brain-like activity and connectivity patterns , seem a much more promising subject of study ., We use a SORN ( self-organizing recurrent neural network ) model 18 that uses noisy units , incorporates additional plasticity mechanisms , and receives no external input ., The network is composed of excitatory and inhibitory threshold neurons connected through weighted synaptic connections ., is the connection strength from neuron to neuron ., We distinguish connections from excitatory to excitatory neurons ( ) , excitatory to inhibitory connections ( ) and inhibitory to excitatory connections ( ) ., Connections between inhibitory neurons and self-connections of excitatory neurons are forbidden ., The connections onto excitatory cells ( and ) are subject to synaptic plasticity mechanisms described below ., and connections have sparse random initial connectivity with connection probabilities of 0 . 1 and 0 . 2 , respectively ., The remain fixed at their random initial values ., They have all-to-all topology and are drawn from the interval and subsequently normalized such that the incoming connections to an inhibitory neuron sum up to one: ., The networks activity state , at a discrete time , is given by the binary vectors and corresponding to the activity of the excitatory and inhibitory neurons , respectively ., The evolution of the network state is described by: ( 1 ) ( 2 ) The and are threshold values for the excitatory and inhibitory neurons , respectively ., They are initially drawn from a uniform distribution in the interval and ., The Heaviside step function constrains the activation of the network at time to a binary representation: a neuron fires if the total drive it receives is greater then its threshold , otherwise it stays silent ., and represent white Gaussian noise with and ., The time scale of a single iteration step in the model corresponds to typical membrane time constants and widths of spike-timing dependent plasticity ( STDP ) windows — lying roughly in the range of 10 to 20 ms . Note that in order to save computation time the homeostatic plasticity mechanisms described below are simulated to be much faster than in reality ., The network relies on several forms of plasticity: STDP of EE and EI connections , synaptic scaling and structural plasticity of EE connections , and intrinsic plasticity regulating the thresholds of excitatory neurons ., The set of synapses adapts via a causal STDP rule that strengthens the synaptic weight by a fixed amount whenever neuron is active in the time step following activation of neuron ., When neuron is active in the time step preceding activation of unit , is weakened by the same amount ( or set to zero if necessary to prevent it from becoming negative , which triggers synapse elimination ) : ( 3 ) Synaptic normalization proportionally adjusts the values of incoming connections to an excitatory neuron at each time step so that they sum up to one: ( 4 ) This rule does not change the relative strengths of synapses established by STDP but regulates the total incoming drive a neuron receives and limits weight growth ., It leads to a competition among excitatory-to-excitatory connections impinging onto the same neuron such that growth of some connections is compensated by the decay of others ., An intrinsic plasticity rule maintains a constant average firing rate in every neuron ., To this end , a neuron that has just been active increases its threshold while an inactive neuron lowers its threshold by a small amount: ( 5 ) where sets the target firing rate ., For simplicity , one can also set the same target firing rate for all the excitatory neurons ., Note that the synaptic normalization and intrinsic plasticity mechanism operate faster in the model than they would in biological brains ., This choice is warranted because of a separation of time scales and speeds up the simulations ., Compared to the original SORN model , we introduce two additional forms of plasticity ., Structural plasticity adds new synaptic connections between excitatory cells to the network at a small rate , which balances the synapse elimination induced by STDP ., With probability a new connection is added between a random pair of excitatory cells that are unconnected ., The strength of this weight is set to 0 . 001 ., Inhibitory spike-timing dependent plasticity ( iSTDP ) adjusts the weights from inhibitory to excitatory neurons to balance the amount of excitatory and inhibitory drive a neuron is receiving ., If the inhibitory neuron spikes and the excitatory neuron remains silent in the subsequent time step ( the inhibitory spike was “successful” in preventing the excitatory cell from spiking ) , the inhibitory weight is reduced by an amount ( or set to a small positive value of 0 . 001 if necessary to prevent it from being eliminated ) ., If , however , the inhibitory neuron spikes and the excitatory neuron also spikes in the subsequent time step ( the inhibitory spike was “unsuccessful” in preventing the excitatory cell from spiking ) , the inhibitory weight is increased by the larger amount ., In all other cases the weight remains unchanged: ( 6 ) Equivalently , we can write: ( 7 ) Unless otherwise specified , the initial weights of , and are drawn from a uniform distribution as shown in Fig . 2E , and the simulations are conducted using the following parameters ., , , , , , , , , . | Introduction, Results, Discussion, Materials and Methods | The information processing abilities of neural circuits arise from their synaptic connection patterns ., Understanding the laws governing these connectivity patterns is essential for understanding brain function ., The overall distribution of synaptic strengths of local excitatory connections in cortex and hippocampus is long-tailed , exhibiting a small number of synaptic connections of very large efficacy ., At the same time , new synaptic connections are constantly being created and individual synaptic connection strengths show substantial fluctuations across time ., It remains unclear through what mechanisms these properties of neural circuits arise and how they contribute to learning and memory ., In this study we show that fundamental characteristics of excitatory synaptic connections in cortex and hippocampus can be explained as a consequence of self-organization in a recurrent network combining spike-timing-dependent plasticity ( STDP ) , structural plasticity and different forms of homeostatic plasticity ., In the network , associative synaptic plasticity in the form of STDP induces a rich-get-richer dynamics among synapses , while homeostatic mechanisms induce competition ., Under distinctly different initial conditions , the ensuing self-organization produces long-tailed synaptic strength distributions matching experimental findings ., We show that this self-organization can take place with a purely additive STDP mechanism and that multiplicative weight dynamics emerge as a consequence of network interactions ., The observed patterns of fluctuation of synaptic strengths , including elimination and generation of synaptic connections and long-term persistence of strong connections , are consistent with the dynamics of dendritic spines found in rat hippocampus ., Beyond this , the model predicts an approximately power-law scaling of the lifetimes of newly established synaptic connection strengths during development ., Our results suggest that the combined action of multiple forms of neuronal plasticity plays an essential role in the formation and maintenance of cortical circuits . | The computations that brain circuits can perform depend on their wiring ., While a wiring diagram is still out of reach for major brain structures such as the neocortex and hippocampus , data on the overall distribution of synaptic connection strengths and the temporal fluctuations of individual synapses have recently become available ., Specifically , there exists a small population of very strong and stable synaptic connections , which may form the physiological substrate of life-long memories ., This population coexists with a big and ever changing population of much smaller and strongly fluctuating synaptic connections ., So far it has remained unclear how these properties of networks in neocortex and hippocampus arise ., Here we present a computational model that explains these fundamental properties of neural circuits as a consequence of network self-organization resulting from the combined action of different forms of neuronal plasticity ., This self-organization is driven by a rich-get-richer effect induced by an associative synaptic learning mechanism which is kept in check by several homeostatic plasticity mechanisms stabilizing the network ., The model highlights the role of self-organization in the formation of brain circuits and parsimoniously explains a range of recent findings about their fundamental properties . | circuit models, computational neuroscience, biology, neuroscience | null |
journal.pntd.0001026 | 2,011 | First Report of Colonies of Sylvatic Triatoma infestans (Hemiptera: Reduviidae) in the Paraguayan Chaco, Using a Trained Dog | Triatoma infestans ( Hemiptera , Reduviidae ) is the main vector of Chagas disease ( American trypanosomiasis ) in the Southern Cone of Latin America ., Through the Southern Cone Initiative against Chagas disease , vectorial transmission to humans has been interrupted in Chile , Uruguay and Brazil , but Argentina and Paraguay have achieved this only in some regions 1 ., In the Gran Chaco region , comprising parts of Argentina , Bolivia and Paraguay , control of the vectors has been limited due to the persistence of domestic infestations despite the efforts of the Vector Control Services in these countries 2 , 3 ., Studies conducted since the 1970s have shown high levels of indoor infestation of T . infestans in the Paraguayan Chaco , characterizing this region as the highest endemic area in the country 4–8 ., However , sylvatic populations of this vector have only occasionally been reported in Paraguay 9 although nymphs of T . infestans were recently reported amongst vegetation near indigenous dwellings 10 ., By contrast , sylvatic T . infestans have been more frequently reported from the Andean valleys of Cochabamba and La Paz in Bolivia , and also in the Bolivian Chaco 11–13 and the Argentine Chaco 9 , 14 ., The finding of dark morph ( DM ) T . infestans in parrot nests in Argentina 14 , and the finding of extensive new foci of sylvatic triatomine populations in Bolivia 15 encouraged the intense search in the Paraguayan Chaco region , but the search for this species using light traps and manual checking of fallen trees and burrows had been unsuccessful ., We report here a novel approach using a trained dog , which has revealed several sylvatic populations of T . infestans in the Paraguayan Chaco ., Domestic dogs ( Canis familiaris ) are used by humans to locate a range of substances because of their superior olfactory acuity ., Their area of olfactory epithelium ( 18 to 150 cm2 ) 16 is much greater than that of humans ( 3 cm2 ) 17 ., They are widely used to detect non-biological ( explosives , chemical contaminants , illegal drugs ) and biological scents ( human odours , animal scents ) and have an important role in conservation 18 ., Dogs have been trained for search and rescue of missing people 19 , to search for brown tree snakes 20 , insects that damage plants 21 , birds 22 , egg masses of gypsy moths 23 , subterranean termites 24 , screwworm-infested wounds 25 , catfish off-flavour compounds 26 , animal scat detection 27 and microbial organisms such as rot fungi , building moulds , and bacteria 28 ., However , as far as we know , there are no previous attempts to train dogs to detect triatomine bugs ., Triatominae produce volatile compounds , which seem to play a role in their defense and alarm processes , as well as in sexual communication and mating ., The Brindleys glands , present in adult Triatominae , seem mainly to secrete isobutyric acid – believed to be involved in defense against predators 29 , 30 ., The metasternal glands , also present in adults , have been associated with sexual communication , and some highly volatile ketones ( 3-pentanone ) and alcohols that are emitted by adults during mating have been identified 29 , 30 ., Moreover , the nymphs do not have Brindleys glands , metasternal glands , or dorsal abdominal glands 31 ., The bug faeces are also a source of attractants 32 and both adults and nymphs respond to faeces from different species 33–37 ., The compounds most commonly found in fresh faeces are ammonia and uric acid , and other compounds such as o-aminoacetophenone , 4-methylquinazoline , 2 , 4-dimethylquinazoline , and 2-pyrolidinone 37 , 38 ., Based on the possibility of detecting bugs by means of their odours we have implemented a new method in which we use a trained dog to search for triatomines ., This has enabled us to find sylvatic T . infestans in the region of the Paraguayan Chaco through a quick , easy and low-cost procedure ., The study in the indigenous communities was approved by the local Ethical Committee of the Fundación Moisés Bertoni ( IDRC Grant No . 103696-009- Revision 07/27/2007 ) and CEDIQUIFA ( Approved 02/18/2008 ) ( Argentina ) ., Following local indigenous conventions for the approval of research in their communities , the local leaders of the villages of 12 de Junio and 10 Leguas were informed of the study objectives prior to commencing the study and they signed an informed consent form on behalf of the members of the community ., This village-level consent process was approved by both ethics committees ., The use and handling of animals in this study was approved by Fundación Moisés Bertoni ( Grant No . 103696-009-Addendum 05/03/2010 ) and the animal care and facilities supporting this activity was maintained according to the standards of the Council for International Organizations of Medical Sciences ( CIOMS , 1985 ) 39 ., Within the framework of an entomological surveillance study of indigenous communities , sylvatic triatomines were sought within the peridomicile of the indigenous communities of 12 de Junio and 10 Leguas in the Department of Presidente Hayes ( Figure 1 ) ., The surrounding area represents typical xeromorphic Chaco woodland , characterized by species such as Aspidosperma quebracho-blanco , Schinopsis quebrachocolorado , Bulnesia sarmientoi , Prosopis nigra , Schinopsis balansae , Calycophyllum multiflorum y Stetsonia coryne 40 , 41 ., The climate in this part of the Chaco is characterized by extreme summer heat and mild winters ., Temperature extremes range from 45° C in spring and summer to −7°C in winter ., Windspeed averages 3 . 3 meters/second ( 11 . 9 km/h ) that increases up to 3 . 9 m/s ( 14 . 0 km/h ) in winter 42 ., For this study geo-referenced points were identified using a GPS ( GARMIN Etrex Legend ) during field trips ., Triatomines were manually captured in demarcated areas during daylight hours with the help of NERO , a 9 month-old gray German Shepherd male dog ( Figure 2D ) ., NERO had basic obedience training and was further trained to locate triatomines by an experienced dog trainer ., The trainer used live , laboratory-reared , uninfected male and female adult bugs throughout the training process ., The specimens were placed individually in plastic containers closed with gauze , with paper as a substrate ., Training was carried out in the trainers home using the method outlined by the United States Customs Service 43 ., First the living triatomines were presented to the dog to stimulate the dogs olfactory memory before being hidden somewhere in a house , and the dog was told to “search” ., After daily training sessions for 3 weeks , the triatomines were no longer presented to the dog at the beginning of the session , and the dog was asked to “search” for hidden bug samples ., In the third phase , several samples were hidden around the house simultaneously ., The dogs ability to locate different intensities of odor was tested by hiding samples of several bugs at some sites and single bugs at other sites ., Tasks with no positive samples were included as well ., When the dog found the sample , he would sit at attention next to the sample and look at his trainer ., Small pieces of sausage were used as rewards ., The training took a total of 3 months ., In the field , the dog was accompanied by his trainer and a field team made up of three or four biologists ., Every time the dog made the appropriate signal the field team made a thorough revision of the area looking for triatomines ., The collection of triatomines was carried out 5 times during the months of May to August 2010 ., The place and characteristics where triatomines were found were geo-referenced and noted with the climatic characteristics of the days when captures were carried out ., Specimens were placed together in plastic cups with paper as a substrate , coded according to capture sites , and transported live to the laboratory where they were classified by species , sex , and stage following standard taxonomic keys 31 ., Faecal matter expressed from each specimen was also checked microscopically at 400× for possible trypanosome infection ., Specimens were then preserved in 70% ethanol for subsequent DNA extraction from legs ., For DNA extraction , four legs from each specimen were ground to a fine powder in the presence of liquid nitrogen , mixed with 1 mL of lysis buffer , and incubated overnight at 37°C 44 ., DNA was extracted sequentially with phenol , phenol-chloroform-isoamyl alcohol , and chloroform-isoamyl alcohol , and precipitated with ethanol in 0 . 3 M sodium acetate 45 ., The mitochondrial cytochrome B gene was targeted for amplification as described by Lyman et al 46 and a frangment of 415 bp ( primer regions not included ) with no insertions or deletions was considered in the analysis ., PCR products were sequenced directly and in both directions ., Sequences from sylvatic bugs were compared with GenBank Triatoma spp ., sequences by Blast analysis with Genbank default parameters ., To determine if the dog was able to differentiate between nymph and adult triatomines , laboratory-reared 3rd and 5th stage T . infestans were placed in plastic containers and hidden for the dog to search for them ., Similarly , two trials were conducted to assess which triatomine odours the dog could detect ., In the first , a plastic vial containing a filter paper impregnated with 50 uL of commercial isobutyric acid ( MERCK ) was hidden ., The second trial used papers impregnated with fresh or dried faeces from adult and nymph stage T . infestans ., Each of these trials was done on two occasions in the trainers house ., A total of 70 triatomines was collected during 5 field trips with NERO ., All specimens were captured alive from vegetation such as dry branches , hollow or standing trees of different species like quebracho blanco ( Aspidosperma quebracho-blanco ) , verawood better known by its spanish name palo santo ( Bulnesia sarmientoi ) and dried cactus ( Stetsonia coryne ) ., In the case of quebrachos , the bugs were found inside hollow dry branches , while in palo santo they were captured from the cortex ., Triatomines were also found in a Tabara major nest in a fallen quebracho blanco tree ( Figure 2A ) , in rodent burrows inside a fallen palo santo tree ( Figure 2B ) and in piles of quebracho blanco branches cut for firewood ( Figure 2C ) ., The house nearest to capture sites was located 408 meters from the town of 10 Leguas , while the distances from capture sites to the nearest community averaged 2 . 8±1 km ( Figure 1 ) ., Of the 70 bugs collected , 22 specimens ( 16 adults and 6 nymphs ) corresponded to dark morph ( DM ) forms of T . infestans ( Table 1 ) ., Purified DNA from 14 of these successfully amplified the target cyt-b fragment , resulting in two haplotypes that differed in 4 synonymous substitutions ., The Blast analysis with sequences available in GenBank confirmed them as T . infestans ., One of the haplotypes presented 100% similarity to already published sequences ( accession numbers AY062165 . 1 and EF639038 . 1 ) ., The other haplotype is now deposited in Genbank under accession number HQ848648 ., The remaining bugs comprised 18 specimens of Triatoma guasayana , and 30 specimens of Triatoma sordida ( Table 2 ) ., There was a predominance of T . sordida in relation to T . infestans , and although both species were found in the same period of time they were never found sharing the same habitat ., None of the specimens of any species examined under the microscope appeared to be infected with trypanosomes ., Following the discovery of triatomine colonies in the forested areas of the Chaco we attempted further trials to see if the dog was capable of identifying nymphs and adults independently , and what specific scent the dog was detecting ., The dog was exposed to nymphs , fresh and dry bug faeces , and isobutyric acid , in independent experiments ., The dog consistently marked the location of the nymphs , but did not find the fresh or dry faeces on any occassion ., When the dog was exposed to a flask containing isobutyric acid the dog was able to locate the flask , but did not clearly indicate the location as when finding a live triatomine bug ., DNA sequence of 660 bp of the new haplotype including the cytochrome B gene was deposited in GenBank under accession number HQ848648 ., This sequence was obtained using the primers described by Monteiro et al 58 . | Introduction, Materials and Methods, Results, Discussion | In the Gran Chaco region , control of Triatoma infestans has been limited by persistent domestic infestations despite the efforts of the Vector Control Services ., In Paraguay , this region is the highest endemic area in the country , showing high levels of indoor and outdoor infestation ., Although sylvatic T . infestans have been found in the Bolivian and Argentine Chaco , similar searches for sylvatic populations of this species in Paraguay had been unsuccessful over the last 20 years ., Here we present a new approach to detecting sylvatic Triatominae , using a trained dog , which has successfully confirmed sylvatic populations of T . infestans and other triatomine species in Paraguay ., A total of 22 specimens corresponding to dark morph forms of T . infestans were collected , and 14 were confirmed as T . infestans by the mitochondrial cytochrome B gene analysis ., Through this analysis , one of which were previously reported and a second that was a new haplotype ., Triatomines were captured from amongst vegetation such as dry branches and hollows trees of different species such Aspidosperma quebracho-blanco , Bulnesia sarmientoi and Stetsonia coryne ., The colonies found have been small and without apparent infection with Trypanosoma cruzi ., During the study , Triatoma sordida and Triatoma guasayana have also been found in ecotopes close to those of T . infestans . | Confirmation of sylvatic colonies of Triatoma infestans has a significant connotation for Paraguay ., Prior to our findings , we believed this vector —unlike in other regions of the Gran Chaco—was living exclusively in domestic and peridomestic habitats ., We never considered the possibility of sylvatic species re-infesting domiciliary dwellings ., After this discovery , the frame of transmission dynamics of Trypanosoma cruzi in the Paraguayan Chaco proposes new research perspectives ., This also opens the door to promote knowledge regarding potential genetic flows between different T . infestans populations , reservoirs associated with their colonies , as well as their impact over control actions ., Fieldwork for wild species identification is difficult and often unsuccessful , we used several techniques and tools , proven by others such as light traps , and mouse-baited sticky traps however , the triatomine collection in our study area was scarce or null ., Incorporating a trained dog – NERO – to our work team has been a highly successful and productive initiative ., The surprising ability NERO has shown will enable us to provide specific data regarding the still unknown wild ecotopes of T . infestans , as well as the potential use of trained dogs as a community surveillance tool of triatomine species considered particularly important for public health . | community ecology, zoology, ecology, entomology, biology, species interactions, biodiversity, parasitology | null |
journal.pcbi.1006165 | 2,018 | Lipidated apolipoprotein E4 structure and its receptor binding mechanism determined by a combined cross-linking coupled to mass spectrometry and molecular dynamics approach | Apolipoprotein E ( apoE ) is a member of the superfamily of exchangeable apolipoproteins ., It mediates cellular uptake of cholesterol-rich lipoproteins by acting as a high affinity ligand for cell surface receptors belonging to the low-density lipoprotein ( LDL ) receptor family 1 ., An imbalance in cholesterol homeostasis increases the risk for cardiovascular diseases and is also linked to neurodegenerative disorders 2 , 3 ., Therefore , the receptor binding property of apoE stresses its importance in the transport of lipids and metabolism of cholesterol both within the plasma and the central nervous system 4 , 5 ., In blood plasma , the receptor mediated uptake and endocytosis of apoE-containing lipoproteins lowers the overall levels of circulating lipoproteins , explaining the anti-atherogenic effect of apoE 6 ., In the brain , although apoE is involved in lipid redistribution and neuronal growth and repair , the presence of the ε4 allelic form of the apoE gene also represents the most significant genetic risk factor of developing Alzheimer’s disease 7 ., An abnormal trafficking of lipids and cholesterol by apoE4 is among the pathogenic mechanisms that are proposed to contribute to the susceptibility of ε4 carriers for Alzheimer’s disease 8 , 9 ., ApoE is a ~34 kDa protein composed of 299 amino acids ., Single point variations at positions 112 and 158 distinguish the three main isoforms of apoE: apoE2 ( Cys112 , Cys158 ) , apoE3 ( Cys112 , Arg158 ) and apoE4 ( Arg112 , Arg158 ) 10 ., These sole amino acid substitutions result in structural differences between these isoforms 11 and marked effects on their lipid binding abilities 12 , providing grounds to explain their different physiological role ( s ) in cardiovascular and Alzheimer’s diseases 13 ., In the lipid-free state , all three apoE isoforms possess two independently folded structural domains linked by a protease sensitive loop 14 ., The N-terminal ( NT ) domain ( res . 1 to 191 ) comprises an elongated four-helix bundle that contains the binding region to the members of the LDL receptor family on the fourth helix 15 ., The C-terminal ( CT ) domain ( res . 210 to 299 ) presents the major lipid binding region 16 and is particularly challenging to study , as it is involved in the oligomerization of apoE in the absence of lipids 17 ., Several mutations had to be introduced in the CT domain to generate a stable monomeric protein leading to the so far only available full-length high resolution three-dimensional ( 3D ) structure of a lipid-free apoE protein ., In this structure , the CT domain variant contains three α-helices folded upon the NT domain conferring a globular shape to apoE 18 ., Upon binding to lipid particles , apoE undergoes a large conformational conversion to accommodate and stabilize the lipids through its amphipathic α-helices , allowing thereby their trafficking in the circulation 19 ., Additionally , lipid binding induces apoE to adopt a biologically active conformation that is a prerequisite for the binding of lipoproteins to cell surface LDL receptors and their internalization 1 ., Analysis of reconstituted discoidal phospholipid-apoE particles ( rHDL , more recently termed nanodisk ) presented a major step forward towards a structure of lipid-bound apoE 19 ., These particles mimic in vivo nascent high density lipoproteins ( HDL ) in shape , size , density and functional properties 20 ., It was demonstrated that in these systems , the α-helices of apoE are oriented perpendicularly to the acyl chains of the lipids and the apolipoprotein molecules circumscribe the edge of the discoidal particles 21–23 ., Lipid-binding also triggers the elongation of NT domain helix 4 which was proposed to represent a key lipid-induced conformational change allowing for the recognition of apoE by LDL receptors 24 , 25 ., However , the conformation adopted by apoE molecules at the surface of these discoidal particles remains an open question ., While it is accepted that the CT domain adopts an extended α-helical structure 19 , 22 , 23 , the conformation of the NT domain has not converged towards a single model ., Based on calorimetry measurements , it was proposed that the four-helix bundle opens to expose the hydrophobic faces of the amphipathic helices towards the lipids and that further reorganization of helices occurs , triggered by lipid binding 26 ., Although this bundle opening was suggested to ultimately lead to a fully extended conformation of apoE that wraps around the entire circumference of the lipid bilayer of the disc 22 , several studies have indicated that apoE adopts a hairpin structure for which distinct hinge localizations were proposed 27–29 ., Supported by low resolution X-ray density and electron paramagnetic resonance ( EPR ) measurements , an alternative model was developed ., In this case , even though apoE also folds in a hairpin structure , the hydrophobic faces of apoE helices are suggested to interact with each other , while the polar faces contact the phospholipids leading to ellipsoidal lipoparticles 30 , 31 ., Despite two decades of intensive structural studies , a consensus on the conformation of lipidated apoE has not yet been reached ., With the aim of deciphering the molecular structure adopted by apoE at the surface of rHDL in solution , we designed an approach where complementary low resolution structural data were combined with 3D structural modeling ( Fig 1 ) ., We decided to focus our present work on rHDLs containing only apoE4 , considering the prevalent role of this isoform in Alzheimer’s disease 8 , 9 ., Experimental data were primarily generated from chemical cross-linking ( XL ) coupled to mass spectrometry ( -MS ) which produces covalently connected pairs of peptides that provide a set of distance restraints between cross-linked residues on the native protein , enabling low resolution models to be elaborated ., XL-MS has seen significant progress recently 32–37 and has been successfully applied to a large number of protein complexes 38–41 ., The distance restraints from our intramolecular XLs , together with additional experimental data obtained in this work and information from the literature were then used in our hybrid molecular modeling approach ., Two alternative models of lipidated apoE4 were validated by our XL-MS results and assessed by molecular dynamics simulations ., Our resulting models represent the most detailed structures obtained so far on full-length apoE4 associated to rHDL and they provide unprecedented insight into the active structure of apoE4 ., Taken together the data allowed us to propose a novel molecular mechanism that explains how apoE is recognized by the members of the LDL receptor family ., ApoE can bind to lipoproteins of variable sizes and shapes due to its conformational flexibility 42 ., To obtain detailed information on lipidated apoE4 conformation , it was desirable to obtain highly homogeneous lipoproteins , in order to stabilize a uniform apoE4 conformation ., For the preparation of such rHDL , we used the cholate dialysis method 43 and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine ( POPC ) as a source of phospholipids ., The initial lipid:protein molar ratio was optimized to enhance homogeneity of apoE4/POPC particles by following the resulting rHDL migration on native PAGE ( S1 Fig ) ., At an initial apoE4/POPC molar ratio of 1:110 , a single population was apparent , displaying a Stokes diameter of ~105 Å ( Fig 2A ) ., A single population was also detected by gel filtration of apoE4/POPC rHDL ( Fig 2B ) ., Finally , quantification of the protein content indicated that the apoE4/POPC reconstitution allowed recovering up to ~50% of the protein initially engaged in rHDL ., These reconstitution parameters allowed us to prepare highly homogeneous apoE4/POPC rHDL particles in a reproducible manner ., To provide detailed input data for the structural modeling of lipid-bound apoE4 , we extensively characterized their composition and shape ., Quantification of the concentration of lipid by phosphorus assay revealed that each particle contained about 200 POPC molecules ., Chemical XL with large excess of bis ( sulfosuccinimidyl ) suberate ( BS3; molar ratio BS3:apoE4 200:1 ) followed by SDS-PAGE analysis resulted in a single apoE4 dimer band at approximately 70 kDa , indicating that two molecules of apoE4 were present on each apoE4/POPC rHDL ., Infrared measurement revealed a sharp peak for the amide I band ( 1700–1600 cm-1 ) centered at 1652 cm-1 characteristic of α-helical structures 44 ( Fig 2C ) ., An α–helical content of approximately 60% was estimated by curve-fitting of the amide I band ., ApoE4/POPC rHDL were next characterized by both negative staining ( NS ) and cryo-transmission electron microscopy ( TEM ) ., Most representative class-averages of NS-TEM images revealed that apoE4/POPC rHDL mainly appeared to be of circular shape with a diameter of 115 ± 10 Å ( S2A Fig ) ., To determine the overall shape of the rHDL particles in hydrated state and to avoid possible artifacts due to sample drying and heavy metals on the observed shape , we visualized the apoE4/POPC rHDL by cryo-TEM ( S2B Fig ) ., The homogeneity of the apoE4/POPC reconstitution enabled us to perform single particle analysis ., Two-dimensional averages identified both top and side views of the rHDL particles ( Fig 2D ) ., These projections displayed a diameter similar to the one previously measured in NS-TEM and a thickness of 50 ± 10 Å , in good agreement with the expected thickness of a POPC bilayer 45 ., NS- and cryo-TEM images therefore strongly support a discoidal shape for the apoE4/POPC rHDL ., These particles will further be designated as apoE4 nanodiscs in this work ., The apoE4 protein conformation at the surface of nanodiscs was investigated by XL-MS using the homobifunctional disuccinimidyl suberate ( DSS ) cross-linker that reacts with primary amino groups ( Lys residues and protein N-termini ) ., The extended Cα-Cα distance for lysine pairs that can be cross-linked by DSS is usually considered to have an upper limit of about 30 Å 46 , 47 ., An equimolar mixture of light DSS ( DSS-H12 ) and heavy DSS ( DSS-D12 ) was used , providing a unique isotopic signature to cross-linked peptides and facilitating their detection and identification by MS 48 , 49 ., The apoE4 molecules at the surface of the nanodiscs were cross-linked with 8 moles of DSS for 1 mole of apoE4 ., The low DSS/apoE4 molar ratio was chosen to minimize the risk of disturbing the structure adopted by apoE4 in the nanodiscs ., The resulting species were isolated by SDS-PAGE revealing two bands of comparable intensity at approximatively 35 and 70 kDa , which were assigned to cross-linked monomeric apoE4 and cross-linked dimeric apoE4 , respectively ( Fig 3A ) ., To generate exclusively unambiguous intramolecular XL products , the monomeric apoE4 band at 35 kDa was processed by in-gel digestion with trypsin and analyzed by liquid chromatography MS/MS ., The resulting fragment ion spectra were analyzed using the dedicated software pipeline xQuest/xProphet 46 , 48 ., 27 cross-linked peptides were identified for monomeric apoE4 ( S1 Table ) , which corresponded to 22 unique Lys-Lys distance restraints ( Fig 3B and Table 1 ) ., Evaluation of the intramolecular XL data set obtained for the apoE4 nanodiscs revealed that both CT and NT regions of the protein are covered by the ensemble of XLs , with 11 out of the 12 apoE Lys residues involved in at least one XL ., The XLs can be classified into two main categories ., The first , and largest , group comprises XLs that were formed between Lys residues located in the NT domain and the CT domain ( Fig 3B and 3C , dotted lines ) ., The second group contains pairs of Lys residues belonging to the NT domain only , the vast majority connecting different helices forming the four-helix bundle adopted by apoE in its soluble form ( Fig 3B and 3D , dashed lines ) ., Topological information on the conformation of lipidated apoE4 could be deduced from the distance restraints derived from the XL data ., The distribution and number of intramolecular XLs between Lys residues of the NT and CT domains were inconsistent with a completely extended conformation of apoE4 at the surface of the nanodiscs ., They rather suggested a hairpin conformation ( Fig 3C ) ., Besides , the scattering and number of intra-NT domain XLs were indicative of a relatively compact state of the NT helix bundle ( Fig 3D ) ., Once the in-depth experimental characterization of the nanodiscs was achieved , we set out to generate a model of apoE4 bound to rHDL ., To do so a two-step procedure was set up ( Fig 1 ) : first , monomeric conformations of apoE4 were constructed by molecular modeling using experimental data to guide the modeling process ., Then , dimer assemblies of these monomer structures were wrapped around an explicit lipid disc and the evolution over time of these systems was investigated by molecular dynamics simulations ., In a first modeling approach , we directly used all the intramolecular XLs as long and medium-range distance restraints so as to generate a structural model of monomeric lipidated apoE4 ., However , this attempt was unsuccessful as the ensemble of XLs restraints could not generate any concluding structures that would fit the experimental characterization of the nanodiscs ( shape and size ) ., From this first approach , it appeared evident that the ensemble of XL data would not be satisfied by a single ultimate model , hinting at the presence of at least one alternative conformation ., We therefore devised a second approach in which here-acquired structural data were rationalized in the light of current knowledge on lipidated apoE to narrow the range of conformational states apoE4 could adopt at the surface of nanodiscs ( Fig 1 ) ., They were implicitly included in sets of constraints for the structure generation ( S1 Text and S2 Table ) ., First , to fulfill the hairpin conformation suggested by the NT-CT spatial proximity , evidenced by our XL-MS data ( Fig 3C ) , we inserted a hinge , allowing the CT domain to fold back along the NT domain ., To determine the apex of the hairpin , we tested three different hinge positions in the non-structured portions of apoE4 connecting the NT and CT domains ( res . 164 to 168 , 186 to 193 , or 201 to 208 ) ., Although our XL data pointed toward a relatively compact conformation of the NT domain ( Fig 3D ) , we conjectured that an apoE4 NT domain conformation completely folded as in solution would be hardly compatible with a receptor active conformation , as it is commonly accepted that opening of the NT bundle upon lipid interaction is a prerequisite for exposure of NT helix 4 containing the region involved in recognition of LDL receptors 19 ., Therefore , based on literature 18 , 22 , 28 , 29 , we decided to partially open the NT four-helix bundle by unfurling the turn in between NT helices 3 and 4 and aligning these two helices with the CT domain in a hairpin conformation by using a zipping procedure ., Second , we maintained NT helices 1 to 3 bundled together by applying a zipping procedure between NT helices 2 and 3 , thus promoting their spatial proximity to comply with the XL-MS data and to place NT helix 2 outside of the implicit lipid disc ., On the other hand , due to the lack of XL data for NT helix 1 , which does not contain any Lys residue , preventing us to rule on its position , we chose to keep this helix in contact with NT helix 2 by using the distances and angles from the NMR study of full length mutated apoE3 18 ., A partially opened state comprising a NT three-helix bundle with NT helix 4 detached was hence generated ., Finally , we imposed a curvature to adapt the conformation of apoE4 molecules to the experimental discoidal shape of the nanodiscs and applied a distance constraint to move the flexible CT end outside of the nanodisc ., Validation of our models by the XL data revealed that , from the structures generated with the three different hinge positions , the model featuring a hairpin structure containing the hinge formed by res . 186 to 193 best matched the XL pattern , satisfying 12 out of 22 XLs ( Table 1 ) ., This model was named “opened hairpin” model ( Fig 4A ) and the selected hairpin apex placed NT helices 3 and 4 in juxtaposition to the CT domain , in good agreement with 6 XLs ( out of 11 ) formed between these two domains ( Table 1 ) ., Nevertheless , the opened hairpin model left out 10 XLs that failed to comply with its structure ., These non-satisfied XL were either intra NT domain ( helices 2/3 connected to helix 4 ) or NT-CT domains XL ( helices 2/3 connected to a different region of the CT domain ) ( Table 1 ) ., Careful inspection of the opened hairpin model suggested that these XLs were likely to be satisfied if the NT domain adopted a four-helix bundle ., We thus constructed a second monomeric apoE4 model , using the same hinge region ( res . 186 to 193 ) but adjusted the constraint list ( S2 Table ) to retain a compact state of the NT domain bundle ., Remarkably , in this second model , named “compact hairpin” model in the following ( Fig 4B ) , 19 out of the 22 identified XLs were validated ( Table 1 ) ., The three non-satisfied XLs involved a subset of the NT-CT links ( helices 2/3 with res . 262 and 282 of the CT domain ) that otherwise supported the opened hairpin model ., The two conformations proposed here may therefore represent distinct states of lipidated apoE4 that dynamically co-exist in solution ., Both the opened and compact hairpin monomeric models were dimerized in either a head-to-head or head-to-tail orientation ., They were wrapped around a solvated POPC disc producing four different molecular systems ( S1 Text and S3 Fig ) ., In all 4 setups , the final number of lipids contained in the nanodisc is in good agreement with the experimental values we measured , providing a first validation of our models before we further studied their dynamic behavior using molecular dynamics simulation ., In the first nanoseconds of the trajectories , the amphipathic α-helices were observed to rearrange so as to more efficiently protect the hydrophobic acyl chains of the lipids located at the edge of the nanodiscs from the solvent ., By adjusting their α-helical segments contacting the lipids , two apoE molecules are able to accommodate the number of lipids contained in each lipoprotein particle and match the average diameters of the nanodiscs as they were observed in this study by native PAGE ( Fig 2A ) , NS- and cryo-TEM ( Fig 2D and S2 Fig ) ., Further , our 75-ns long trajectories highlighted that the lipid structures kept their disc shape in all cases ( Fig 5A and S4 Fig ) and the majority of the XLs remained satisfied at the end of our simulations ( S3 Table ) ., The α–helical content at the end of the simulations calculated with DSSP 50 ranged between 51% and 66% in good agreement with the 60% estimated from our infrared measurements ., No significant differences could be evidenced between the systems featuring either a head-to-head or head-to-tail apoE dimer and we therefore could not discriminate between both orientations ., However , during the trajectories local changes in the secondary structure were observed in some regions of the protein ., Remarkably , a short stretch ( res . 164 to 168 ) at the end of NT helix 4 switched from a random coil to an α-helical conformation and remained α-helical for the rest of the simulation in one of the monomers in all models ( Fig 5B ) ., This structural change , close to the binding region to LDL receptors , promoted an extension of helix 4 resulting in a long amphipathic helix spanning res . 131 to 180 ( Fig 5B ) ., Furthermore , upon this change Arg172 , known to be involved in the recognition of LDL receptors 51 and other upstream basic residues , also known to interact with the receptor 52 , underwent a reorientation leading to their respective alignment ( Fig 5B ) ., Comparison of the solvent accessibility of these residues in our two models ( S5 Fig ) indicated that , while most residues binding to the LDL receptors featured a low accessibility in the compact hairpin model , they really pointed into the solvent in the opened hairpin model regardless of the dimer arrangement ., Therefore , although both conformations may co-exist in solution , they may exert variable binding activities towards receptor recognition with the opened hairpin model representing the active conformation of lipidated apoE4 ., The XL-MS distance restraints obtained here from the cross-linked monomeric apoE4 molecules argued against a model where apoE could adopt a completely extended structure surrounding the nanodisc , with two molecules of apoE running along each other in a ‘double-belt’ organization as was proposed previously 22 ., A large subset of our intramolecular XL data rather inferred a hairpin fold of lipidated apoE as previously proposed in other studies 29 , 31 ., Alike previous models of full-length apoE , our XLs implied the hinge of the hairpin to be situated in the unstructured region connecting the NT and CT domains but with subtle differences resulting in significant structural and mechanistic implications ., Specifically , in the so far most detailed Xray/EPR model of lipidated apoE4 30 , 31 , the hinge is situated at res . 162 to 169 ( vs res . 186 to 193 here ) and suggested to bring in close proximity regions that are known to be important for the interaction with LDL receptors , the region spanning res . 134 to 150 and Arg172 ., However , due to the hinge location in this model , the α-helical extension of NT helix 4 , suggested to be essential for receptor binding activity 24 , 25 , is no longer possible ., This hinge location was also not supported here , as the model we built with the hinge on res . 164 to 168 only satisfied 9 out of the 22 identified XLs ., In spite of the difference in hinge localization , spatial proximity of significant pairs of residues could be reconciled between our and previous hairpin models ., For instance , for apoE4 , the spatial proximity of two residues , Arg61 and Glu255 , that are proposed to form a salt bridge promotion the interaction between NT and CT domains in the lipid-free form 60 , was confirmed to be maintained in the lipid-bound state in discoidal particles 29 ., The proximity of these residues was also preserved here , thanks to the partially closed conformation of the NT domain ., Further , a significant number of EPR constraints 31 were also validated in our models , including the intramolecular spatial proximity of residues 76/77 with residues 239/241 that were established in our study to be intramolecular by the selection of the monomeric band for in-gel digestion ( Fig 3A ) ., The here-produced hairpin models thus allow at the same time both spatial proximity of recognized pairs of important residues in CT and NT regions and the opportunity for the extension of helix 4 needed for the recognition of LDL receptors ( Fig 5B ) ., However , a limitation of our study is that , in the current setting , we did not specifically discriminate between intra- and intermolecular cross-links within homodimeric apoE proteins and the respective organization of the two apoE molecules on the lipid particle could therefore not be deduced ., The head-to-head and head-to-tail dimerizations , as presented in S3 Fig , therefore remain to a certain degree speculative ., Two alternative models , featuring three or four bundled amphipathic helices from the NT domain , were constructed that together satisfied the ensemble of XL derived spatial restraints ( Fig 4 and Table 1 ) ., The compact hairpin model features a NT four helix bundle laid along the CT domain and interacting with the lipids only via helix 4 ., In this conformation , helix 4 , that contains essential residues for recognition of the members of the LDL receptor family 52 , was shielded from the solvent by helix 3 ( Fig 4B and S5 Fig ) ., In the opened hairpin model , the turn between helices 3 and 4 was unfurled , in agreement with previous studies 18 , 28 , and allowed an opening of the bundle with NT helix 3 now interacting with the lipids ., This partial opening of the bundle was sufficient to expose helix 4 to the solvent ( Fig 4A and S5 Fig ) ., The opening movement from the compact to the opened hairpin model therefore provides us with a possible regulatory mechanism of apoE4 lipoproteins ( Fig 6A and 6B ) ., In contrast to earlier studies that indicated that the interaction of the NT domain with the lipids would engage an open and active conformation of the receptor binding region 18 , 26 , our models strongly suggest that , in both the open and compact state , the NT domain of apoE is associated with lipids at the surface of the nanodisc ., The domains outside the lipid disc in the compact hairpin model ( S3C and S3D Fig ) are not clearly resolved by NS-TEM ( S2 Fig ) ., Heterogeneity in the disc size , dynamic structure of NT domain switching between compact and open conformation , and small size of the folded domain preclude its visualization by single particle technique based on averaging of projections of individual aligned particles ., Moreover NT helix 1 that was considered in our modeling as part of the NT helix bundle despite the absence of structural data could instead adopt a more extended conformation ., Each of these two cases would then contribute to decrease the compactness of the NT portion that may be observed by NS-TEM ., We speculate that both the open and compact hairpin model co-exist in a dynamic equilibrium where the different forms could concurrently be captured by our XL experiments ., Further , we propose that in presence of the receptor this equilibrium is shifted to the opened hairpin model , the model that represents the state accessible to LDL receptors , and therefore allows us to draw a mechanism of accessibility of the LDL receptor binding region ( Fig 6A and 6B ) ., A relatively small structural change , observed in all models during the molecular dynamics trajectories , elongated helix 4 and connected it with a subsequent small helix spanning res . 169 to 180 , leading to the formation of a 50 residue-long amphipathic helix ( res . 131 to 180 ) ( Fig 5B ) ., This helix extension upon lipidation has already been proposed experimentally by NMR and EPR 24 , 25 and it was suggested to act as a molecular switch that stably anchors the receptor binding region on the lipid surface or/and correctly positions residue 172 with other basic residues ( in the region 136 to 150 ) known to be required for an optimal interaction with the LDL receptors 51 , 52 ., These receptors share highly conserved structural domains , including ligand-binding domains containing cysteine-rich ligand binding type-A ( LA ) repeats ., For the LDLr , the most prevalent member of this family of receptors , it is now well established that among the 7 LA repeats , LA5 is essential for binding of apoE lipoproteins 61 and that the pair LA4-LA5 is sufficient to bind apoE in rHDL 62 ., The residues known to interact with the LDLr on apoE 51 , 52 span a too large region to be recognized by a single LA repeat of the LDLr and would thus allow for the binding of the two LA repeats to the same apoE molecule as we proposed earlier based on the lipid-bound structure of an apoE-derived peptide 25 ., In addition , our study showed that the elongation of NT helix 4 upon lipidation led to a reorganization of the LDLr binding residues that could promote their binding to the LDLr LA4-LA5 repeats ( Fig 5B ) ., To support this hypothesis , we performed docking assays in which LA4 and LA5 were docked individually to such an elongated NT helix 4 ( S4 Table ) ., The results confirmed that the distance between the two docked modules was in agreement with the long loop between LA4 and LA5 repeats ( S6 Fig ) ., This distance is unique between this pair of LA repeats 63 , highlighting its importance in lipidated apoE recognition ., Contrary to the soluble form of apoE4 , the elongation of NT helix 4 conferred upon lipidation would therefore represent an additional prerequisite for binding to LDL receptors ( Fig 6D ) ., In summary , our data advocate that several requirements need to be met to provide a fully receptor-active apoE ( Fig 6D ) : lipid binding , exposure of the receptor binding region and elongation of NT helix 4 ., We speculate that the here proposed compact hairpin model is a stable conformation co-existing with the active , receptor-competent open structure , explaining why these two alternative conformations could be trapped in the XL-MS experiments ., Both conformations may therefore be part of a regulation mechanism of apoE function at the surface of lipids ., Our work represents a building stone towards a better understanding of the strong anti-atherogenic effect of apoE and the models we are proposing could prove useful in the study of lipidated apoE in totally different contexts , such as understanding its role in Alzheimer’s disease ., Overall our hybrid approach , compatible with the presence of lipids , results in 3D structures of lipidated apoE4 that represents the most comprehensive model of the active form of apoE4 to date and might be applied to the study of other ( membrane ) proteins where such complementary low resolution structural data are available ., Unless otherwise stated , all chemicals were obtained from Sigma-Aldrich at the highest purity available ., Water was double-distilled and deionized using a Milli-Q system ( Millipore ) ., The human full-length apoE4 gene fused to a self-cleavable intein tag and a chitin binding domain cloned into a pTYB2 vector was a kind gift of Dr . Vasanthy Narayanaswami ( University of Long Beach , California , U . S . A . ) ., ApoE4 was expressed in a T7 expression strain of Escherichia coli ( ER2566 ) in 2xYT medium by the addition of isopropyl β-D-thiogalactopyranoside ., Pelleted cells were resuspended in buffer A ( 20 mM HEPES , 500 mM NaCl , 1 mM EDTA , pH 8 ) , supplemented with 0 . 5% ( v/v ) triton-X-100 ( TX100 ) and anti-proteases ( cOmplete EDTA-free protease inhibitor cocktail , Roche ) , and apoE4 was released by high-pressure homogenizer ., The protein was purified following standardized protocols previously described for intein-labelled proteins 64 ., Briefly , the clarified cell lysate was loaded onto chitin beads ( Impact system , New England Biolabs ) equilibrated with 5 column volumes ( CV ) of buffer A containing 0 . 5% ( v/v ) TX100 and incubated at 4°C during 1 h on a rotating wheel ., The flow through was discarded and the beads were washed with 10 CV of buffer A containing 0 . 3% ( v/v ) TX100 ., ApoE4 was released by incubation of the chitin beads with buffer A containing 30 mM dithiothreitol ( DTT ) at 4°C during 40 h and finally eluted with 3 CV of buffer A containing 5 mM DTT ., ApoE4 was then buffer exchanged against buffer B ( 20 mM ammonium bicarbonate , pH 8 ) with PD-10 desalting column ( GE healthcare ) and lyophilized overnight ., Prior to utilization , lyophilized apoE4 was solubilized in buffer C ( 20 mM HEPES , 150 mM NaCl , pH 7 . 4 ) containing 6 M guanidine-HCl and further purified by size exclusion chromatography on a Superose 6 matrix ( GE Healthcare ) eluted with buffer C containing 4 M guanidine-HCl ., Fractions containing apoE4 were pooled together and dialyzed against buffer B during 48 h at 4°C ., ApoE4 concentration and purity were assessed by absorbance at 280 nm and SDS-PAGE ., ApoE4 rHDL were formed using 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine ( POPC , Avanti polar lipids ) following a modified version of the protocol initially developed for apoA-I by Matz and Jonas 43 ., POPC solubilized in chloroform was dried under nitrogen and resuspended to a concentration of 20 mg/ml in buffer C . Sodium cholate was added at a POPC:sodium cholate molar ratio of 1:2 and the mixture was sonicated for 1 h with vortexing every 15 min ., ApoE4 was added to the mixture at different molar ratio and incubated overnight at 4°C on a rotating whee | Introduction, Results, Discussion, Materials and methods | Apolipoprotein E ( apoE ) is a forefront actor in the transport of lipids and the maintenance of cholesterol homeostasis , and is also strongly implicated in Alzheimer’s disease ., Upon lipid-binding apoE adopts a conformational state that mediates the receptor-induced internalization of lipoproteins ., Due to its inherent structural dynamics and the presence of lipids , the structure of the biologically active apoE remains so far poorly described ., To address this issue , we developed an innovative hybrid method combining experimental data with molecular modeling and dynamics to generate comprehensive models of the lipidated apoE4 isoform ., Chemical cross-linking combined with mass spectrometry provided distance restraints , characterizing the three-dimensional organization of apoE4 molecules at the surface of lipidic nanoparticles ., The ensemble of spatial restraints was then rationalized in an original molecular modeling approach to generate monomeric models of apoE4 that advocated the existence of two alternative conformations ., These two models point towards an activation mechanism of apoE4 relying on a regulation of the accessibility of its receptor binding region ., Further , molecular dynamics simulations of the dimerized and lipidated apoE4 monomeric conformations revealed an elongation of the apoE N-terminal domain , whereby helix 4 is rearranged , together with Arg172 , into a proper orientation essential for lipoprotein receptor association ., Overall , our results show how apoE4 adapts its conformation for the recognition of the low density lipoprotein receptor and we propose a novel mechanism of activation for apoE4 that is based on accessibility and remodeling of the receptor binding region . | Among the proteins involved in the transport of lipids and their distribution to the cells , apolipoprotein E ( apoE ) mediates the internalization of cholesterol rich lipoproteins by acting as a ligand for cell-surface receptors ., In the central nervous system , while apoE is the major cholesterol transport protein , a dysfunction of apoE in the transport and metabolism of lipids is associated with Alzheimer’s disease ., A molecular understanding of the mechanisms underlying the receptor binding abilities of apoE is crucial to address its biological functions , but is so far hindered by the dynamic and complex nature of these assemblies ., We have designed an original hybrid approach combining experimental data and bioinformatics tools to generate high resolution models of lipidated apoE ., Based on these models , we can propose how apoE adapts its conformation at the surface of lipid nanoparticles ., Further , we propose a novel mechanism of regulation of the activation and receptor recognition of apoE that could prove valuable to interpret its role in Alzheimer and apoE-related cardiovascular diseases . | pattern recognition receptors, medicine and health sciences, molecular dynamics, immunology, lipoprotein receptors, lipid structure, immune system proteins, lipids, proteins, lipoproteins, chemistry, physics, biochemistry, biochemical simulations, signal transduction, molecular structure, cell biology, biology and life sciences, immune receptors, physical sciences, computational chemistry, chemical physics, computational biology | null |
journal.pgen.1008340 | 2,019 | An estimator of first coalescent time reveals selection on young variants and large heterogeneity in rare allele ages among human populations | The age of an allele of a given frequency can be reveal the forces acting upon it , with rare alleles being particularly sensitive to recent evolutionary processes 1 ., A functional allele that is younger than expected given its frequency is likely to have been under directional selection ., This is not surprising for favored alleles , but it is also true for harmful alleles 2 , including those with negative impacts on health that are under negative selection 3 ., If researchers are able to estimate allele age , they could combine this with other information ( e . g . allele frequency , geographical distribution , functional annotation ) to improve predictions of an allele’s effect on human health ., Alternatively , an allele that is older than expected given its frequency is also a candidate for having an interesting history , as functional alleles older than expected can be the result of balancing or negative frequency-dependent selection 4 ., As population genomic samples grow in size , the density of variable sites rises approximately in proportion to the log of the sample size 5 , and very large data sets will have large numbers of SNPs and other variants in every gene ., If information could be gleaned on the ages of a large number of variants for a functional region of the genome , this could be used to develop a detailed portrait of the history of natural selection specifically on that region ., Allele ages are also shaped by processes that act in aggregate across the genome ., The overall distribution of ages will be strongly shaped by the demographic history of the population , and for the rarest alleles , that distribution will be acutely sensitive to recent admixture 6 ., The age spectrum will also fluctuate spatially along the genome , both stochastically and as a function of the intensity of background selection 7 ., In developing a way to study allele age , we considered several constraints ., An estimator should not be a function of allele frequency , as we wish to glean information about allele history that is distinct from its frequency ., We also prefer an approach that is not a function of the demographic history of the population , as some estimators are 8 , 9 ., An estimator that can get close to the true value of the unknown , regardless of demographic history , enables analyses in cases when the history is not known and it enables comparisons between populations that are not confounded by errors in our knowledge of the demographic history of the populations ., We also wish to be able to study the age of the very rarest alleles , including those that appear only once in a sample ( singletons ) ., This last criterion leads to an approach that is different from existing methods that are based on the variation observed among copies of an allele , or in flanking regions 8–13 ., An estimator should also be applicable for very large sample sizes for which it becomes increasingly possible to find low frequency alleles that arose by multiple mutations 14 ., For these cases , as with singletons , we need a method that is applicable for each individual gene copy ., Finally , an estimator should not be highly sensitive to the details of the bioinformatics pipeline used to process the data , such as whether the data were statistically phased ., We developed a new estimator that focuses , not directly on allele age , but rather on the time when a base position in a particular chromosome first coalesces in the genealogy ., The mutation causing a derived allele at that base will have occurred since this first coalescent time , and so first coalescent time can be used as a proxy for allele age ., For example , a neutral singleton allele will have a uniform probability of having arisen anywhere along an external branch , and therefore an expected age of half the first coalescent time ., We assessed performance using simulated data , and show that it performs well and substantially overcomes the challenges describe above ., We also applied it to SNP alleles from the 1000 Genomes Project 15 and the UK10K genome panel 16 ., For these data sets , we assessed basic predictions regarding population-specific variation in the ages of rare alleles , the ages of private alleles and alleles shared by populations , and we compared ages of alleles that are expected a priori to have functional impacts with those that are not ., Our estimator shares structural similarities with existing methods developed to estimate demographic histories from shared-haplotype tract-length distributions 17–20 , but is uniquely able to discern the specific histories of individual rare alleles ., Consider one chromosome ( the focal chromosome ) from a sample of chromosomes drawn at random from a population , and an individual base position on that chromosome ( the focal base ) ., As shown in Fig 1 , the focal base can be thought of as the terminal point on a branch A of the genealogy of the sample of chromosomes at that base position ., If the focal base is a singleton variant ( i . e . a derived allele that occurs only once in the sample ) , then the mutation causing that allele must have occurred on this branch ., Also shown in Fig 1 is that branch A connects the focal chromosome to the most recent common ancestor of that chromosome and a sister branch S , which is ancestral to one or more sister chromosomes of the focal chromosome ., Branches A and S share their most recent common ancestor at a time tc generations prior to the sample generation , and branch A is therefore tc generations long ., We denote the length of branch S as φ generations ., If the sample contains a single unique sister chromosome ( that is , the focal chromosome is also the closest relative of the sister at the focal base ) , then φ is equal to tc ., When multiple chromosomes are all equally and most closely related to the focal chromosome at the focal base , they form a clade of sisters , and φ < tc ., Consider first the case when there is no recombination and the focal chromosome has just a single sister ., In this situation , any differences between the focal and sister chromosomes will have been caused by mutations on branches A and S . We model mutation as a Poisson process , where each base has a constant probability , μ , of mutating in each generation ., Treating a chromosome as a continuous line , and considering events just to one side of the focal base ( in either the 5’ or 3’ direction ) , the probability density of distance x from a focal base to the first base that is not identical between the two chromosomes can be approximated with an exponential distribution having a rate of μ2tc:, p ( x ) =μ2tce-xμ2tc ., ( 1 ), If we knew which chromosome was the sister chromosome , we could compare them and identify the distance from the focal base to the nearest difference between the chromosomes ( i . e . x ) , and use this to estimate tc ., This basic idea captures the underlying rationale of our approach ., The final formulation takes into account the remaining issues: that we do not know which chromosome is the sister to the focal chromosome; that the focal chromosome may have multiple sisters; and that Eq ( 1 ) assumes no recombination ( see Materials and methods ) ., Ultimately , an expression that resembles Eq ( 1 ) , but that differs in replacing x with the longest observed tract of identity between the focal chromosome and each of the other chromosomes in the sample , proves applicable ., We call this quantity the maximum shared haplotype ( msh ) , and show that it arises on either branch A or S ( Fig 1 ) with high probability ., We use t^c to denote the estimator , and because tc values range over several orders of magnitude and Poisson processes have variances proportional to their means , we focused primarily on the logarithm , log10 ( t^c ) ., Performance was assessed in terms of the root mean squared error ( RMSE ) , bias ( mean of estimated minus true values ) , and correlation ( Pearson’s r for the true and estimated values of log10 ( tc ) ) for alleles at all frequencies in a series of large simulated data sets ., We varied sample size and recombination rate , and considered three demographic histories that varied in terms of population sizes , exponential growth , historical bottlenecks , and intercontinental migration ., We also considered samples of chromosomes with known and with estimated phase ., All of these results are summarized in S1 and S2 Tables ., For low and intermediate recombination , over a wide range of circumstances , the estimator exhibits an RMSE of about 0 . 4 log-transformed generations , corresponding to estimates that are typically within a factor of 2 . 5 of the true value ., Across the models , correlations of true and estimated values ranged from 0 . 4 to 0 . 95 with a mode of 0 . 9 ., For recombination rates equal to or less than the mutation rate , bias varies from -0 . 4 to 0 . 1 with a mode of -0 . 2 , which corresponds to an average underestimate by a factor of 0 . 63 ., Performance was consistent across the spectrum of allele frequencies , and with respect to particular independent variables , performance was better: when recombination was low; when population size was constant; when phase was known; and when sample sizes were larger ., A key factor that determines how informative estimates are is having chromosomes that are much more closely related to their closest relatives than to unrelated chromosomes ., Thus , strong recent growth can produce more star-like genealogies that reduces these differences and reduce the quality of the estimates , while larger sample sizes improve them ., When recombination rate is appreciably higher than the mutation rate , such that typical msh values are small enough to be within the range of distances between average pairwise SNPs , estimates worsens and the bias shifts from negative to positive ( S2 Table ) ., The amount of recombination that is too high relative to the mutation rate will depend on the length of the region of high recombination ( e . g . if it is associated with hotspots of short length ) , and on sample size and the demographic history of the sample ., We observed that larger sample sizes exert a greater improvement on the quality of estimates in the case of high recombination than with low to intermediate recombination ( S1 and S2 Tables ) ., The change in sign of the bias , with high recombination , and that the observed absolute value of the bias is lower with intermediate levels of recombination , suggests that there are multiple contributions to the bias ., Fig 2 shows plots of estimated versus true values for different ranges of allele frequencies and for two different sample sizes ( box plots are shown in S1 Fig ) ., Also shown in Fig 2 is an idealized estimator of log10 ( tc ) that is based on allele frequency and for which all alleles at a given frequency generate the same estimate–shown as a band of red points ., We show both the msh-based and the idealized frequency-based estimators to highlight the contrast; the former does not make use of information about a variant’s frequency while the latter does not make use of any information about a variant’s msh value ., Unlike the msh-based estimator , where each variant may have a unique msh from which to generate a unique tc estimate , no frequency-based estimator can distinguish among the potentially large number of variants occurring at identical frequencies within a sample ., When considered over the full range of allele frequencies , the idealized estimator can explain nearly as much of the variation in log10 ( tc ) as the msh-based estimator ( Fig 2f ) ., However , for rare alleles , the msh-based stimator retains strong performance , whereas the frequency-based estimator explains little to none of the variation in log10 ( tc ) ., An important application is assessing upon which of the two chromosomes of an individual a singleton allele correctly resides ., As shown in Fig 3 , the phasing accuracy for singleton variants in simulated data rises as the ratio of the genetic lengths of the alternative msh tracts diverge ., For variants with msh tracts of similar length ( and thus high probability of misassigned phase ) , similar t^c values will be found regardless of how phase is assigned ., For non-singleton variants , the estimator is expected to be relatively immune to switch errors introduced by haplotype phasing software ., This is because switch errors typically involve low frequency variants among similar haplotypes 21 , 22 , and these do not typically affect the distribution of msh values which are often terminated by relatively common variants or recombination events between unrelated haplotypes ., S2 Fig shows that results for statistically phased chromosomes are quite similar to those for the correct chromosomes ( r = 0 . 941 ) based on analyses of singleton alleles in male X-chromosome UK10K data ., From S1 and S2 Tables we see that mean error is slightly increased when the data ( including singletons ) are phased statistically ., The bias also becomes more negative by a small amount with phasing ., Singletons are phased by assigning them to the chromosome that reveals the shorter msh and thus the longer t^c ., The proximal effect of this will be to introduce a positive bias that applies in those instances when this phase assignment is not correct ., However , the observation that bias becomes slightly more negative with phasing , including for singletons , suggests a greater effect in the other direction ., It is possible that the phasing of singletons prior to determining msh values causes them to cluster on fewer chromosomes , thereby lengthening the msh values that are observed ., Based on analysis of variation ( ANOVA , Table 1 , S6 Table ) , we learned that different populations show different distributions of log10 ( t^c ) values for singleton variants ( d . f . = 25; F = 19248; p < 1 × 10−128 ) ., As shown in Fig 4 and S3 Table , populations in Africa have higher geometric means ( 6353 to 7178 generations ) than populations from regions that have not had substantial admixture from African populations ( East Asia , South Asia , and Europe ) , which have lower geometric means ( 2018 to 3882 generations ) ., Admixed American populations spanned a wide range ( 5610 to 7194 ) , with values near those of the African populations ., The finding of younger ages for singletons from non-African old-world populations is expected under a general Out-of-Africa model in which those populations have passed through a bottleneck and have had an overall lower effective population size and thus shorter coalescent times ., The 1KGP data allow us to compare the age of rare variants that are found only in a single population to the age of variants that have the same low frequency within that population , but that also are found in one or more other populations ., For all of the 1KGP populations , private singletons had distributions that were shifted to the left ( younger ) relative to singletons that were shared ., Fig 5 shows these distributions for a single population that is representative of those observed from each the five super-populations , each of which showed a characteristic distribution ( S9–S13 Figs ) ., In every population , private singleton alleles are younger than singleton alleles that are also shared with other populations ( d . f . = 1; F = 896298; p = 0 , Table 1 ) , with private singleton variants being 70% younger than comparable shared variants ., However , the size of the reduction varies considerably by population ( d . f . = 25; F = 3437; p = 0 , Table 1 ) , with the largest reduction observed in admixed American populations ( 82% to 92% ) ., With respect to the functional impact of rare alleles , we expect that the rarest variants in a sample will be enriched ( relative to more common variants ) for alleles that have an impact on fitness , and therefore that rare alleles will be targeted more by natural selection ., This is because such alleles pass through frequency space more quickly than neutral alleles and , if not lost from the population , will reach a given frequency more quickly than a neutral allele 2 ., Consistent with this prediction we observed in the 1KGP data that , conditional on population and geographic spread , singleton variants that alter proteins are 3 . 6% younger than those that do not ( d . f . = 1; F = 24; p = 1 × 10−6 , Table 1 ) ., There is no significant difference to the effect of protein-changing status in individual populations ( d . f . = 25; F = 0 . 58; p = 0 . 95 , Table 1 ) , and there is no significant interaction between protein-changing status and a variant being private to its population ( d . f . = 1; F = 2 . 86; p = < 9 × 10−2 , Table 1 ) ., The theory that says that alleles that affect fitness will be younger applies to both harmful and beneficial alleles 2 ., However , if it is the case that harmful rare alleles are much more common than beneficial rare alleles , then the age difference between functional alleles and neutral alleles should be greater the rarer are the alleles under comparison ., This is because most harmful alleles are rapidly removed from a population and are more likely to be observed at the lowest frequencies in larger samples ., We therefore predict that we should observe a greater difference in ages between alleles that change proteins and those that do not in the larger UK10K dataset than in the 1KGP data ., In fact , the singletons in the 7 , 242 samples of UK10k data showed a considerably larger effect of protein-changing status than those identified as singletons in the 100-sample populations of the 1KGP data ., The geometric mean t^c value for the 105 protein-changing singleton variants was 12% lower than it was for the 1 . 5 × 107 non-protein-changing variants ( 305 generations vs 347 generations; Wilcoxon-Mann-Whitney p = 0 ) ., Every mutation can be envisioned as occurring on a branch of the genealogy or gene tree for a sample of genomes at the locus where the mutation occurred ., Previous estimators of allele age have focused on the time point at which different copies of an allele coalesce with each other , i . e . the time of the top of the gene tree edge upon which the mutation occurred 8–12 ., Here we have taken a different approach and focused on the time at the bottom of that same edge , the time of first coalescence for an edge carrying a mutation ., These first coalescent times have direct connections to msh values , which are easily measured from aligned genomes ., The resulting estimator is not without noise , but estimates covary roughly linearly with true values , with moderate error and bias over a wide range of circumstances ., Our estimator also meets a variety of desired criteria established at the outset ., It is not a function of allele frequency , and thus can be used to study how an allele came to reach its observed frequency ., It is only very weakly a function of demographic history , and thus it can be used to compare the ages of alleles of different populations that have unknown or widely varying histories ., It can be applied to alleles that occur only once in a sample , and to genomic data with very large sample sizes ., It can be applied individually to each copy of an allele , and thus can be used in cases when an allele has arisen by multiple mutations ., And it suffers little degradation in performance when run on statistically phased chromosomes ., The estimator is not expected to be highly sensitive to some additional issues that can arise with population genomic data , including sequencing errors ., The UK10K data for example are low coverage ( 7x ) , however , because new singleton variants are only called when the data are strongly supportive ( high false negative rate , low false positive rate ) only a small minority of the singletons in the data set are expected to be errors ., In fact , the proportion of called variants in the case of the UK10K sample that are true variants was estimated to be quite high ( ~94% for singleton variants ) , based on measurements across monozygotic twins ., This is shown in the first figure of the extended data for that paper ( panel k , row AC = 1 , MZ twins section , column non-ref genotypes % , divided by two and subtracted from 100 ) 16 ., However the variants that are missing from low coverage data can affect the distribution of msh values in ways not accounted for by the method ., For example , many of the mutations that terminate tracts of identify and determine msh are themselves singletons , and so a coverage bias against singletons will tend to lengthen msh values and reduce t^c values ., Although the method can be used to study selection , implicit in the method is an assumption of neutrality , such that mutations on branches A and S ( Fig, 1 ) do not affect the distribution of tc values ., But of course a significant fraction of the mutations in evolutionarily constrained regions are not neutral , and the tc values in these regions are reduced , as we see in our ANOVA results ., The question then is , how does this kind of variation in the non-deleterious mutation rate affect msh and t^c values ?, A closely related question applies to factors such as variation in background selection and flucations in rates of gene flow or admixture , that cause polymorphism levels to wax and wane across chromosomes , and that alter the distribution of msh values that terminate due to recombination events ., The basic expectation is that these kinds of variations will have a greater effect on long tc values , when flucations in the detminants of what should be short msh values can have a greater affect ., However short tc values are associated with very long msh values that span large portions of chromosomes and can be expected to be more immune to local flucations in the factors that terminate msh tracts ., Another potential difficulty is that as sample sizes grow very large , some rare alleles will have been caused by multiple mutations 14 ., While this is not a concern for singletons , it is possible that , for example , a doubleton will actually represent two mutations ., In these cases , the likelihood estimator will simply generate a composite of two different singleton coalescent times ., Unlike methods that estimate the time of most recent common ancestry among individuals carrying a rare variant , t^c is the estimated time since each copy of the rare variant last shared a common ancestor with an individual not carrying the variant ., If two copies of an allele are actually the result of independent mutations , the composite t^c estimation will produce an intermediate value and not an aberrantly large one , as would occur with a method based on intra-allelic differences ., S3 Fig illustrates the impact of analyzing two independently derived minor variants as copies of a single mutation ., Some additional and important benefits of the estimator described here emerged during the course of this study ., Estimated tc values , or msh values , can be used to phase singleton alleles ( i . e . estimate the correct chromosome for placement in diploid heterozygous individuals ) ., Another benefit is that the method can be applied relatively quickly to very large data sets with the benefit of the PBWT algorithm 23 ., Computing time increase linearly with both sample size and numbers of variants , and all the singleton variants for the full UK10K panel for chromosome 22 can be analyzed in under 30 minutes , for example ( S4 Table ) ., A third likely benefit , that has not been explored here , is that it should be possible to work with genomic data in which only dispersed portions of the genome have been sequenced ., In particular , msh tracts for rare alleles from large samples are typically much longer than the distances between genes , and consequently it should be possible to apply the method to data sets of aligned exome sequences ., Our analysis of variants of 1% frequency among 26 populations of the 1KGP data , revealed considerable systematic variation in allele ages depending on the variant’s function , location , and geographic distribution ., Different human populations have variants at 1% frequency with very different distributions of ages ., Population-specific demographic histories , including population bottlenecks , expansion , and admixture events , have likely all contributed to these broad differences , and therefore changing , in effect , what it means to be a 1% frequency variant on a population-by-population basis ., We also observed that the ages of variants found at 1% frequency in one population depend greatly on whether or not they are also found in other populations ., To be found in multiple populations , a variant that arose by a single mutation must have either:, 1 ) originated prior to the divergence of those populations and risen to sufficient frequency that it has not been lost in either; or, 2 ) persisted in its population of origin long enough and risen to sufficient frequency that migrants have had an opportunity to bring it to at least 1% frequency in another population ., Both of these phenomena are reflected in substantial differences of ages of variants found in a population at 1% frequency depending on their total geographic range ., The survey of singleton t^c distributions for 1KGP revealed bimodal distributions in each of the non-African populations ( S9–S13 Figs ) ., All of them exhibit a large peak at less than a thousand generations , and another smaller one at more than ten thousand generations , a pattern that is readily interpreted in terms of an older bottleneck associated with the out-of-Africa history of these populations ., S14 Fig shows these same patterns arise for simulated data generated with parameter estimates for an Out-of-Africa model 24 ., The younger larger peak is characteristic of variants that arose after the population’s ancestors had emigrated from Africa ., These are relatively new variants that have not yet had an opportunity to rise to high frequency or spread to other populations ., The older peak consists of variants that arose within an ancestral African population and predate the modern human expansion into Eurasia ., Although they are identified as singletons in individual populations , these variants are often shared across multiple populations , and they typically occur at frequencies greater than 1% in some populations ., In the populations where they are ascertained as singletons , founder events , genetic drift , and natural selection have made these older alleles rare , or even eliminated them entirely before re-introduction by migration from other populations ., Independently of geography and demographics , protein-changing variants in the 1KGP populations that are found at 1% frequency are younger on average than comparable variants that do not change proteins , consistent with previous reports 25 ., These are expected to be mostly deleterious variants that have not yet been removed from the population , but they may also include some beneficial variants that have not yet been pushed to higher frequencies ., At the 1% frequency level , the impact of protein-changing status on the distribution of allele ages is statistically significant , but considerably smaller than the systematic differences introduced by demography and geography ., Among the singleton protein-changing variants ascertained in the UK10k data , however , we find a larger difference with respect to variants that do not change proteins ., The contrast with the 1KGP data is consistent with the younger , rarer , and more numerous variants of the UK10K data are subject to greater selective forces than the variants at 1% frequency in the 1KGP populations ., Our ANOVA did not reveal evidence that local adaptation contributes to the variance in ages of alleles at 1% frequency ., If selectively important alleles were disproportionately prevented from migrating between populations or remaining in multiple populations , we would have observed the pool of shared protein-changing variants to be enriched for protein-changing variants without selective function ., There would have been less differentiation between non-shared and shared , non-protein-changing variants , and there would have been a significant ( negative ) interaction term between private status and protein-changing status , which we did not observe ., We also did not detect an interaction between populations and protein-changing status ., If different populations were broadly experiencing greater or lesser amounts of directional selection on rare variants , or if there were substantial differences between populations in the ability of natural selection to remove harmful variants or raise the frequency of beneficial variants ( such as due to variation in effective population size ) , we would expect to have seen a significant interaction between populations and protein-changing status ., However , we did not; and while both of these phenomena may be taking place , they are not major drivers of the distribution of ages of variants found at 1% frequency , in contrast to the contributions of that demography , geographic spread , and global properties of natural selection ., These population genetic results are all presented as conditioning on alleles at 1% frequency that have been down-sampled to equal sizes ., By considering only alleles of a particular frequency , we are able to draw meaningful comparisons and contrasts among populations ., While the formula for the estimator itself is not a function of the frequency of the allele to which it is being applied , the distributions of t^c estimates will vary greatly as a function of the frequencies of the alleles being studied ( see e . g . S15 Fig ) ., Indeed , it is because of these very different age distributions that alleles of different frequencies can have very different histories , including the action of natural selection and the amount of movement among populations ., Any future work that would make comparisons between populations or estimate demographic models based on t^c estimates will need to explicitly condition on allele frequencies ., With an estimate of an allele’s age , an investigator has an important new piece of information to bring to bear on the possible functional impact of a mutation ., As shown here , for 1KGP data and even more strongly for the larger UK10K data , functional alleles are younger , and therefore , concomitantly , any allele that is discovered to be especially young , given its frequency in the sample , is a good candidate for having an effect of fitness ., In this context , an increase in sample sizes will have multiple important effects ., First , the rarest alleles in large samples will be rarer and younger , on average , than those found in smaller samples , and they will thereby be relatively enriched for more alleles of harmful effect and for alleles of more harmful effect ( these are the alleles that would not reach those higher frequencies observable with smaller sample sizes ) ., Second , the numbers of alleles in the rarest class rises with sample size , and thus so does the number of very rare alleles observed for a given gene ., For example , the average number of autosomal singletons for the 26 1KGP populations ( 100 genomes ) was 3 , 311 , 984 , while the count for the UK10K data ( 7242 genomes ) was approximately 6 times higher at 19 , 078 , 777 ., Based on those values , and assuming the number of SNPs is a function of the log of the sample size 5 , a sample of 2N = 100 , 000 genomes would have 36 million singletons , and a sample of 1 million genomes would have 45 million singletons ., The increasing density of very rare alleles , as sample sizes grow , opens the door to a kind of mapping of functional constraint across a gene that will become increasingly fine-grained ., Third , as data sets get very large so grows the potential for studying the impact of selection on variation that has arisen at different times and it will become increasingly possible to assess whether the action of selection has been changing for different genes or regions of the genome ., The es | Introduction, Results, Discussion, Methods, Web resources | Allele age has long been a focus of population genetic research , primarily because it can be an important clue to the fitness effects of an allele ., By virtue of their effects on fitness , alleles under directional selection are expected to be younger than neutral alleles of the same frequency ., We developed a new coalescent-based estimator of a close proxy for allele age , the time when a copy of an allele first shares common ancestry with other chromosomes in a sample not carrying that allele ., The estimator performs well , including for the very rarest of alleles that occur just once in a sample , with a bias that is typically negative ., The estimator is mostly insensitive to population demography and to factors that can arise in population genomic pipelines , including the statistical phasing of chromosomes ., Applications to 1000 Genomes Data and UK10K genome data confirm predictions that singleton alleles that alter proteins are significantly younger than those that do not , with a greater difference in the larger UK10K dataset , as expected ., The 1000 Genomes populations varied markedly in their distributions for singleton allele ages , suggesting that these distributions can be used to inform models of demographic history , including recent events that are only revealed by their impacts on the ages of very rare alleles . | We developed a way to estimate the time when a copy of a gene most recently shared ancestry with other copies of that gene ., This is also an estimate of the upper bound of when a mutation has arisen , and it can be used to study the ages of alleles that are found in a population ., The method can be applied to the very rarest alleles found only once in a sample , even in studies of many thousands of genomes ., We tested the method extensively , found it performs well , and can be used under a wide variety of conditions ., We applied it to 1000 Genomes project data ( 26 populations ) and the UK10K data ( over 7000 genomes ) and found clear evidence that alleles that change proteins are younger than alleles that do not , as expected ., We also observed wide variation in the ages of alleles at low frequency among the 1000 Genome project populations , indicating that our method could be used to study the demographic history of human populations ., Going forward , the estimator should be useful for many kinds of questions in population genomics , particularly as sample sizes continue to grow . | alleles, genetic mapping, simulation and modeling, mutation, molecular biology techniques, research and analysis methods, sequence analysis, gene mapping, bioinformatics, sequence alignment, chromosome biology, molecular biology, genetic loci, haplotypes, cell biology, natural selection, heredity, database and informatics methods, genetics, biology and life sciences, evolutionary biology, evolutionary processes, chromosomes | null |
journal.pcbi.1005266 | 2,016 | Feedback Loops of the Mammalian Circadian Clock Constitute Repressilator | An autonomous circadian clock controls daily rhythms in physiology and behaviour in a large variety of species ., Such an endogenous timing system has evolved to adapt to the 24h period of the solar day ., Circadian rhythms are generated by intracellular transcriptional feedback loops involving cis-regulatory elements such as E-boxes , D-boxes , and ROR-elements ( RREs ) ., In mammals , more than 20 core clock genes assemble a sophisticated gene regulatory network with multiple negative and positive feedback loops 1 ., Given the complexity of this network , we here investigate , which network motifs are necessary and sufficient for generating self-sustained rhythms ., The classical view of the circadian oscillator considers the E-box mediated negative feedback of Period ( PER ) and Cryptochrome ( CRY ) proteins towards the transcriptional activator complex CLOCK/BMAL1 as the major driver of circadian rhythms 2 ., More recent studies also suggest that another negative feedback loop with the nuclear receptors Rev-erb-α and Rev-erb-β acting through RORE enhancers is not merely an auxiliary loop , but is capable of generating self-sustained oscillations 3 , 4 ., Indeed , double-knockouts of Rev-Erb genes destroy rhythmicity 5 , 6 ., The relative importance of clock genes and their regulatory interactions is consequently debated 7 ., Here , we explore which gene regulatory motifs are most relevant for 24h oscillations ., To this end , we systematically analyzed a recently published circadian oscillator model 8 ., This model includes Bmal1 as a driver of E-box mediated transcription , Per2 and Cry1 as early and late E-box repressors , respectively , as well as the D-box regulator Dbp and the nuclear receptor Rev-erb-α ., The model design is based on experimentally verified regulatory interactions , degradation rates and post-transcriptional delays ., The unknown parameters describing transcriptional regulation have been fitted to four qPCR data sets using an evolutionary optimization algorithm 8 ., The resulting gene network involves 17 regulatory interactions forming multiple negative and positive feedback loops and therefore contains several potential oscillation generating mechanisms ., Such a quantitative model is well suited to study the principles of circadian rhythm generation ., We comprehensively and systematically analyze the capability of sub-networks to generate oscillations ., Interestingly , we identify the “repressilator” motif 9–12 as a central loop of the mammalian circadian oscillator ., The repressilator comprises a series of three inhibitions including the genes whose knockouts lead to arrhythmicity , i . e . Cry , Per and Rev-erb ., To study the complex gene regulatory network of the mammalian circadian oscillator , we constructed a mathematical model with only the key components as explicit variables ., For example , transcriptional profiles reveal clear redundancies in the network of core clock genes 1 , 4 with RORE-binding activators ( Rorα , -β , -γ ) exhibiting opposite phases as the RORE-binding inhibitors ( Rev-erb-α , -β ) ., This allows to describe the regulatory actions by a single term controlled by Rev-erb-α levels , while the systems behaviour remains the same ., The additional effects by Ror-genes and Rev-erb-β can be taken into account by changes of parameters describing the strength of Rev-erb-α regulation ., Analogously , we combine the regulations via D-boxes into one term ., The Dbp-gene represents the combined effects of the activators Dbp , Hlf and Tef and the inhibitor E4bp4 ., Transcriptional regulation via E-boxes is particularly complex 13 ., In our model , Bmal1 quantifies the positive regulation after dimerization with Clock or Npas2 , while the genes Per2 and Cry1 represent early and late E-box driven genes , respectively ., The essential role of a rather late Cry1 phase has been demonstrated in detail elsewhere 14 , 15 ., Overall , we designed a regulatory network consisting of five variables only 8 ., Fig 1A shows that even this core clock network exhibits multiple negative and positive feedback loops ., Importantly , our model successfully describes published phase relations , amplitudes and waveforms of clock gene expression profiles ( Fig 1B ) ., A detailed comparison with experimentally measured profiles is described in 8 ., Our gene regulatory network model contains 34 kinetic parameters ., Since quantitative details of post-transcriptional processes including phosphorylations , complex formations and nuclear translocation are not known , we represent these processes by explicit delays taken from experimentally determined phase-differences between peak expression of mRNA and protein 4 ., Degradation rates were taken from large scale studies of mRNA decay 16 , 17 and protein measurements 18–20 ., Exponents in transcriptional regulation terms are derived from the number of experimentally verified cis-regulatory elements 4 , 21 ., The remaining parameters describe the kinetics of transcriptional regulation , which is not known in quantitative detail ., Thus , we applied global optimization techniques to fit parameters to carefully normalized qPCR data sets from mouse liver and adrenal gland 8 ., For both tissues data from light-dark cycles ( LD ) and constant darkness ( DD ) were available ., Interestingly , all four expression profiles show clear similarities and thus we fitted a consensus model to these four data sets ., The complete set of equations and parameters is provided in ( S1 Appendix ) ., The resulting data-driven gene regulatory network model allows to address the following questions:, ( i ) Which kinetic parameters are most relevant for 24h rhythm generation ?, ( ii ) Are oscillations of all five genes necessary for self-sustained rhythms ?, ( iii ) What are the most essential regulatory interactions required for rhythm generation ?, We will answer these questions in the next sections by systematically varying parameters and clamping gene expression levels to their mean values ., Thereby , we identify design principles in the network necessary and sufficient for generating circadian oscillations ., Our set of default parameters has been fitted to mRNA expression profiles of circadian clock genes from mouse liver and adrenal gland ., It is conceivable that the chosen kinetic parameters are different among tissues and also depend on the specific physiological conditions ., In order to test which parameters are most relevant for 24h oscillations , we varied all parameters by two orders of magnitude around the default values ., Fig 2 represents the results for four particularly interesting parameters ., The periods are plotted for parameter values where self-sustained oscillations occur ., It turns out that oscillations persist for wide ranges of kinetic parameters supporting the robustness of the model ., The period increases with the delay between Per2 transcription and its function as an inhibitor ( Fig 2A ) ., Indeed , the FASPS mutation of PER2 reduces protein life-time , leading to a faster turn-over and hence to shorter delays 22 , thereby implying a shorter period and much earlier phases 23 ., Increasing the degradation rate of Cry1 mRNA leads to period shortening as expected ( Fig 2B ) and consistent with the shorter period of the Cry1-/- knockout mice 24 ., In addition , appropriate interactions of Bmal1 , Rev-erb-α , Per2 and Cry1 are required to generate self-sustained rhythms ( Fig 2C and 2D ) ., Variations of kinetic parameters associated with transcriptional regulations have minor effects on the period near their default values , consistent with the observation that the clock is resilient to changing transcription rates 25 ., The most surprising observation , however , are period jumps for somewhat larger parameter changes ( Fig 2 ) ., In particular , the detection of long and short periods within a very narrow parameter range suggests that multiple mechanisms might co-exist which can generate self-sustained rhythms ., Indeed , the systematic analysis described below allows us to attribute oscillations with different period to specific loops in the model ., For example upon increasing the Per2 delay , the period falls down to 15h after rising up to more than 30h ( Fig 2A ) ., Here , the period jump occurs , when the explicit delay is very large ( above 10 . 5h ) and influences a subsequent cycle of the oscillations rather than the current one ., The rhythm-generating loop , however , remains the same ., Further increase of the delay then leads again to an increase of the period until the phase pattern and period before the jump emerges again ., An animation of this gradual parameter variation is provided as S1 Video ., Interstingly , there is hysteresis near the period jumps , indicating coexisting limit cycles ( also termed “birhythmicity” 26 ) ., A period jump also occurs upon variation of the Cry1 degradation rate ( Fig 2B ) ., Here , the jump to a short period of < 10h is associated with a change of the rhythm-generating loop: Cry1 self-inhibition generates these oscillations ., Since the self-inhibitory loop exhibits a rather short delay of τCry1 = 3 . 13h , the resulting period is consequently quite small ( comprehensive list of feedback loops and delays in the S4 Appendix ) ., We show in S2 Appendix that in the transition region two rhythms persist ( termed “torus” ) ., If we increase Per2 inhibition by Cry1 , oscillations vanish via a supercritical Hopf bifurcation ., At much larger parameter values another Hopf bifurcation leads to a limit cycle governed by a double-negative feedback loop involving Per2 , Dbp and Cry1 ., Taken together , even a relatively small network of just five genes can establish multiple mechanisms generating oscillations , some with periods in the circadian range ., While it is generally believed 2 that the negative feedback loop via Per/Cry is the primary driver of circadian oscillations , these multiple regulatory mechanisms even within a relatively small network raise the possibility that the underlying key mechanism for circadian rhythm generation is more complex ., To investigate , which network nodes ( genes ) are essential for circadian rhythm generation we systematically studied all possible sub-networks under default parametrization ., Our gene network with 7 positive and 10 negative regulations exhibits multiple feedback loops ( Fig 1A ) ., Delayed negative feedback loops constitute the basic elements of self-sustained oscillations 27 , 28 ., Often these negative feedback loops are interlinked with positive feedbacks ensuring robust and tunable rhythms 29–33 ., Thus , we focused on which sub-networks forming feedback loops are able to generate sustained rhythms for physiologically plausible parameters ., To this end , we systematically clamped all possible combinations of gene-subsets to their respective oscillation mean values ., The mean values are obtained from simulations of the complete network ., Clamping the level of Dbp or Bmal1 , for example , retains the corresponding positive regulations , but excludes Dbp and Bmal1 as drivers or transmitters of oscillations , thereby focusing on the remaining genes ., The clamping to mean values ensures that the system remains near the carefully tuned and physiologically reasonable default configuration ., Clamping of genes corresponds to constitutively expressed genes using non-rhythmic promoter constructs 15 , 34–36 ., Compared to knockout studies , our clamping protocol is less invasive and keeps the system close to its physiological ranges ., There are 5 combinations of 4 genes resulting from clamping only one single gene ., For all of the resulting networks there exist certain parameter configurations with oscillatory solutions ( blue bars in Fig 3 ) ., Clamping Rev-erb-α , Per2 or Cry1 has strong effects: Using the default parameters of the complete network , oscillations vanish ., In order to explore the rhythm-generating capabilities of the sub-networks more extensively around the default parameter set , we varied each parameter of the system in a range from 5-fold reduction to 5-fold increase in repeated simulations with 200 points on a log scale ., For every simulation we tested , whether or not the sub-network oscillates ( Fig 3A ) ., It turns out that in principle all sub-systems of 4 genes are capable of generating oscillations with reasonable periods ., Interestingly , clamping Dbp ( first blue bar , Fig 3A ) or Bmal1 ( last blue bar , Fig 3A ) sustains oscillations in about 90% of parameter combinations with a median period close to 24h ., This is in line with experimental findings showing that Bmal1 cycling is not necessary for circadian rhythms 36 , 37 , Dbp-/- knock-out mice are still rhythmic 38 and triple-knockouts of D-box regulators have only minor effects 39 ., Thus , both experimental evidence and our modelling results underline that Bmal1 and Dbp cycling is not essential for sustaining oscillations ., Simultaneously clamping two genes to their mean values results in ( 53 ) =10 sub-networks of 3 genes ., We find that 5 of these networks are capable of generating self-sustained oscillations , when allowing up to 5-fold adjustments of single parameters ., Interestingly , Rev-erb-α is present in most of these oscillatory sub-systems ( as an example , see Fig 4B ) ., Simultaneously clamping 3 genes leads to ( 52 ) =10 sets of only 2 remaining genes ., Surprisingly , 3 of these pairs are still able to oscillate for appropriate parameter adjustments ( Fig 3B ) ., Notably , the negative feedback loop involving Bmal1 and Rev-erb-α oscillates with a period of about 24 hours after only a minor parameter change ( compare Fig 4A ) ., It turns out that some of the previously identified oscillations in larger sets of 3 and 4 genes can be traced back to this simple loop ., This finding confirms earlier observations that the feedback loop via nuclear receptors can serve as a possible mechanism for rhythm generation 3 , 4 , 7 ., In the previous section gene levels ( nodes ) were clamped to their mean values , allowing sub-networks to be identified as possible rhythm generators ., Now we expand our approach to combinatorial clamping of regulatory interactions ( edges in the network graph in Fig 1A ) allowing the identification of sub-networks on a process-level ., Thereby , network motifs most essential for the generation of 24h rhythms can be identified ., In our model , transcriptional regulations are described by products of activating and inhibiting terms corresponding to the influence of regulating genes 4 ., If the expression value of a regulating gene is set constant to its mean value in the term of only one specific target-gene , we call the corresponding interaction “clamped” ., For more details on the method , see ( S3 Appendix ) ., Since the gene network contains altogether 17 regulatory interactions , there are 217 = 131 , 072 combinations , or ON/OFF configurations , if OFF means clamping ., For all these combinations we tested in detail , whether or not oscillations persist , but did not consider additional variation of kinetic parameters ., We found that 14 , 125 ( about 10% ) of all network configurations oscillate ., In order to evaluate the importance of specific regulatory interactions we calculated for each interaction the relative frequency of inclusion in the set of oscillatory network configurations ., Among all possible configurations any given process is ON or OFF in 50% of the cases ., Thus , considering the set of oscillatory ON/OFF configurations , an edge that is not part of the essential loop would still occur in one-half of the cases ., Indeed , the analysis of all oscillatory ON/OFF configurations reveals that most of the processes occur in 50% of the oscillating configurations as expected for a non-essential process ., However , a distinct set of regulatory interactions turned out to be present in almost 100% of the oscillating network configurations ., To our surprise , only 3 of the 17 regulatory interactions are exceptionally important to keep the network rhythmic , occurring in almost every oscillating configuration ( marked in Fig 5 by thick red lines ) ., All other regulations can be clamped to prevent them from transmitting rhythms: The remaining 3 regulatory interactions still retain oscillations ., While the period generated by this 3-gene sub-network in isolation is somewhat longer upon default parametrization , the full network compensates this by fine-tuning through other regulations , including a feedforward loop 40 ., Interestingly , the three regulations are all inhibitory: Per2 inhibits Rev-erb-α , Rev-erb-α inhibits Cry1 and Cry1 inhibits Per2 ., Such a symmetric triangular motif of inhibitory interactions has been introduced as a paradigm of synthetic oscillators termed “repressilator” 9 ., In contrast to most models of the circadian clock , which are essentially based on variations of the Goodwin model 22 , 41 , the repressilator comprises three subsequent inhibitions rather than a single negative feedback ., It is known that classical Goodwin-based models need strong negative cooperativity ( minimal Hill coefficient of 8—probably unrealistic biochemically ) and long balanced degradation times to obtain self-sustained oscillations 28 , 33 , 42 ., Within the repressilator , however , the delay and the required non-linearities can be distributed over the three inhibitions ., To test the compatibility of the repressilator , we performed a robustness analysis of two simple prototypic models with a single feedback loop , one with only one inhibition and one with the repressilator motif ( for details see S5 Appendix ) ., In particular , we generated random parameter sets for both models and compared the frequencies of self-sustained oscillations and the minimal Hill coefficients necessary to generate oscillations ( see Fig . 3 in S5 Appendix ) ., We found that the repressilator model has a higher fraction of oscillations and can oscillate with fairly low Hill coefficients of about 2 ., Note , that modified Goodwin oscillators with additional nonlinearities allow reductions of the Hill coefficient as well 43 , 44 ., Generally , systems with multiple nonlinearities and delayed feedbacks allow robust oscillations with reasonable Hill coefficients 32 , 45 , 46 ., The repressilator motif allows to distribute nonlinearities and delays ., The repressilator motif is represented as a serial inhibition of Cry1 via Rev-erb-α , of Rev-erb-α via Per2 and of Per2 via Cry1 ., The two activators Bmal1 and Dbp can be clamped to their mean values without loosing oscillations ., It has indeed been reported that constant Bmal1 levels can sustain rhythms 36 , 37 and that triple-knockouts of D-box regulators have only minor effects on circadian rhythmicity 39 ., In contrast , double-knockouts of Cry , Per and Rev-erb genes lead to arrhythmicity 5 , 24 , 47 supporting our finding that circadian rhythms are not generated by a single negative feedback loop , but by a gene regulatory network with a repressilator as a core motif ., Double-knockouts induce behavioral arrhythmicity ., Since core clock genes oscillate in surprisingly similar phase relationships in almost all tissues 8 , 59 , 61 , it is very likely that the KO experiments imply also tissue arrythmicity ., Indeed , studies of double-knockouts include data on arrhythmic tissues and cells 5 , 47 , 62 ., In previous studies , models have been adapted to available mutant phenotypes 3 , 32 , 63 ., Since our variables group together genes with similar regulatory effects , a comparison with knockout data is not easy ., Our clamping protocols resemble constitutive expression or overexpression , and thus we discuss related experiments ., It has been shown that constitutive expression or overexpression of Per genes impairs rhythms 34 , 35 , 64 , 65 ., Similarly , constitutive or out-of-phase expression of Cry1 impairs rhythmicity 15 and overexpression of Cry1 leads to arrhythmicity 58 ., Knockouts and knockdowns of Cry1 lead to arrhythmicity in tissues and cells 62 , 66 , even though the coupling within the SCN can rescue rhythmicity 62 corresponding to a short-period phenotype of Cry1 knockouts 24 ., Interestingly , knockouts and knockdowns of Cry2 , an early E-box target not regulated by Rev-erb-α , stay rhythmic with large amplitudes 62 , 66 , 67 ., The essential role of Rev-erb-α inhibition of Cry1 is demonstrated by the removal of the intronic ROR-elements leading to early phases of Cry1 and vanishing amplitudes in single cells 14 ., In summary , there is strong experimental evidence that the cycling of the 3 repressilator genes is of central importance for a cellular clock ., Our 5-gene model is based on carefully normalized qPCR data of liver and adrenal gland 8 ., More recently , expression profiles of 14 different tissues have been published 59 ., In all of these tissues the repressilator genes are oscillating with significant amplitudes and with serially ordered phases consistent with the repressilator mechanism ( see S6 Appendix ) ., Similar observations were reported by Yamamoto et al . 61 ., In addition to mRNA rhythms protein oscillations are relevant to understand regulatory processes ., Unfortunately , liver proteome studies could not quantify core clock protein rhythms due to limited resolution 68 , 69 ., A recent quantification of clock proteins confirms early protein expression of REV-ERBα , followed by peaks of PER2 and CRY1 70 ., Recent ChIP-Seq experiments allow the estimation of binding phases of regulatory clock proteins 5 , 6 , 13 , 71 ., It turns out that REV-ERBα binds early ( Circadian Time CT = 6–10 ) , followed by PER2 binding around CT16 and CRY1 binding at around CT24 ., These subsequent binding peaks are fully consistent with the proposed repressilator mechanism ., Our starting point was a gene-regulatory model based on expression profiles of core clock genes in mouse liver and adrenal gland ., As shown in Fig 5 , the repressilator is the dominant motif of this gene-regulatory network ., However , Figs 3 and 4 illustrate that also other negative feedback loops are capable of generating oscillations ., Furthermore , positive feedback loops are known to support rhythm generation 33 ., A comprehensive list of loops within our gene regulatory network is given in ( S4 Appendix ) , showing the interrelations and coherence of loops ., Our results suggest , that multiple loops support the generation of circadian oscillations , while the repressilator constitutes an essential core mechanism: While the pure repressilator generates oscillations with increased periods , the addition of other regulations including a feedforward loop 40 from Cry1 to Per2 via Dbp tune the period to values of about 24h ., In summary , our comprehensive analysis of a data-driven core-clock model suggests that the synergy of multiple regulatory motifs allows robust and tunable self-sustained oscillations ., We further propose , that a series of subsequent inhibitions known as repressilator constitutes a core motif of the mammalian circadian clock gene-regulatory network . | Introduction, Results, Discussion | Mammals evolved an endogenous timing system to coordinate their physiology and behaviour to the 24h period of the solar day ., While it is well accepted that circadian rhythms are generated by intracellular transcriptional feedback loops , it is still debated which network motifs are necessary and sufficient for generating self-sustained oscillations ., Here , we systematically explore a data-based circadian oscillator model with multiple negative and positive feedback loops and identify a series of three subsequent inhibitions known as “repressilator” as a core element of the mammalian circadian oscillator ., The central role of the repressilator motif is consistent with time-resolved ChIP-seq experiments of circadian clock transcription factors and loss of rhythmicity in core clock gene knockouts . | Circadian clocks are endogenous oscillators that drive daily rhythms in physiology , metabolism and behavior ., The recent years have witnessed enormous progress in our understanding of the mechanistic and genetic basis of these clocks ., While mathematical modelling has made important contributions to our current view of the circadian clock network , it is still debated , which network motifs are necessary and sufficient for generating self-sustained oscillations ., Exploiting a data-driven mathematical model we here identify the “repressilator” as a key design principal ., The central role of the repressilator motif is consistent with time-resolved binding data of circadian clock transcription factors and loss of rhythmicity in core clock gene knockouts . | synthetic genetic systems, engineering and technology, mechanisms of signal transduction, gene regulation, synthetic biology, circadian oscillators, network analysis, chronobiology, synthetic gene oscillators, computer and information sciences, feedback regulation, network motifs, gene expression, biochemistry, genetic oscillators, circadian rhythms, signal transduction, cell biology, gene regulatory networks, genetics, biology and life sciences, computational biology | null |
journal.pgen.1007048 | 2,017 | Role of Neuropilin-1/Semaphorin-3A signaling in the functional and morphological integrity of the cochlea | Age-related hearing loss ( ARHL ) , or presbycusis , is a progressive bilateral symmetrical sensorineural hearing loss 1 characterized by four types of pathology: ( 1 ) sensory deficits resulting from loss of outer hair cells as seen in loss of high frequency auditory brainstem response , ( 2 ) neural deficits from auditory nerve degeneration resulting in poor speech recognition , ( 3 ) degeneration of the stria vascularis leading to flat audiometric losses across frequencies; and ( 4 ) cochlear conductive deficits associated with increased stiffness of the basilar membrane resulting in evenly sloping audiometric losses 2 ., Familial studies of presbycusis have attributed approximately half of audiometric variances to hereditary factors; however , the highly variable age of onset , disease progression , and severity of ARHL demonstrate the current uncertain contribution of individual genetic factors to cochlear integrity 3 ., Our group has recently demonstrated that ARHL in humans is a polygenic trait 4 ., Human genetic studies suggest associations between ARHL and several genes including GRHL2 , ITGA8 , IQGAP2 , GRM7 , PCDH15 , PCDH20 , APOE , EDN1 , ESRRG 2 ., Although very little is known about ARHL in humans , numerous studies have been published on ARHL in mice ., A genetic component to ARHL in inbred mice has been described with approximately 18 Mendelian loci reported to date 5–8 ., It has been our overriding hypothesis that true ARHL in mice , as in humans , is a polygenic trait with the composite phenotype resulting from genomic variation at multiple loci likely different from the Mendelian loci described thus far ., To define the genetic architecture of ARHL in mice , we undertook a genome-wide association study ( GWAS ) using a meta-analysis strategy by combining data sets from five cohorts containing 937 samples in total 9 ., The results of the meta-analysis led us to an approximately 2 Mb interval containing Nrp1 , a gene that is involved in cardiovascular and neuronal development and is closely related to Neuropilin-2 ( Nrp2 ) , a gene involved in cochlear epithelial innervation 10 ., Neuropilin-1 is a transmembrane receptor type I protein that is known to bind both vascular endothelial growth factor beta ( VEGFb ) and semaphorin classes including subtypes 3A , 3B , 3C , and 3D ., Semaphorin-3A is involved in axonal guidance via chemorepulsion 11 ., Semaphorin-3A -induced neuronal growth cone collapse has been shown to require neuropilin-1 in conjunction with Plexin-A co-receptors 12 ., Previous cardiovascular studies have shown that altered endothelial cell migration , abnormal blood flow , and enlarged vessels are major defects caused by targeted inactivation of the Nrp1 gene in mice ., Additionally , homozygous Nrp1 mutant mice are known to have a perinatal lethal phenotype due to impaired heart development 13 ., In this study , using an inner ear-specific knock out , we investigated the role of Nrp1 in the functional and morphological integrity of the cochlea in mice ., The results of this study suggest Nrp1 may be involved in ARHL ., We first used in situ hybridization to characterize the expression of Nrp1 and its ligand Sema3a at different stages of cochlear development ., Between E13 . 5 and P1 , the SGNs migrate along the extending cochlear duct then extend peripheral axons toward the cochlear epithelium and central axons toward the brainstem 14 ., In situ hybridization of WT cochleae at E13 . 5 , E15 . 5 , E18 . 5 , and P1 showed weak expression of Nrp1 at E16 . 5 and E18 . 5 with more robust expression starting at P1 ( Fig 1A and 1C ) ., Our in situ hybridization data also showed that Sema3a expression started around E13 . 5 and continued at E16 . 5 and E18 . 5 on the abneural side of the cochlear epithelium and SGNs ( Fig 1D–1F ) ., These data suggest that both Nrp1 and Sema3a are expressed in the cochlea during time points when SGNs begin to innervate the organ of Corti ., We next performed immunostaining using P5 cochleae to determine the precise distribution of neuropilin-1 and semaphorin-3A in the postnatal cochlea ., As shown in Fig 2 , neuropilin-1 is visible in SGNs and the SV ( Fig 2A , 2B , 2D and 2E ) , but not expressed after conditional deletion of Nrp1 ( Fig 2C and 2F ) ., Semaphorin-3A protein was visible within the organ of Corti ( Fig 2G and 2H ) and SGNs ( Fig 2J and 2K ) , but not after the semaphorin-3A antibody was pre-adsorbed by the blocking peptide ( Fig 2I and 2L ) ., Overall , these data suggest that Nrp1 is expressed at minimal levels by SGNs and cells of the stria vascularis before birth while Nrp1 levels become elevated in these locations after birth ., Sema3a is expressed by cells within the cochlear epithelium and SGNs , which is complementary to the expression patterns of Nrp1 ., The simultaneous expression of both neuropilin-1 and semaphorin-3A shortly after birth suggested these factors may be involved in the process of SGN pruning and refinement , which occurs during this time in cochlear development ., These findings prompted further investigation of Neuropilin-1/Sema-3A signaling in cochlear innervation ., Much of the genetic data from our meta-analysis GWAS came from the original backcrossing data ( C57BL/6J x DBA/2J ) during the mapping of Ahl8 6 ., In their mapping study of ahl8 , a locus on chromosome 8 was also identified ., This led us to determine the possibility of Nrp1 expression variation in the cochlear tissue of C57BL/6J and DBA/2J mice ., Real-time PCR showed 1 . 78-fold higher Nrp1 expression for DBA/2J mice ( 1 . 96 ) when compared to C57BL/6J ( 1 . 09 ) ( S1 Fig ) supporting our gene selection ( p<0 . 01 ) ., To investigate the function of neuropilin-1 in the cochlea , we generated an inner-ear specific conditional knockout mouse using Pax2Cre and loxp-driven Nrp1 removal ( see Methods for details ) ., Using this line , we first wanted to determine whether Nrp1 is required for the formation or maintenance of ribbon bodies , which represent glutamatergic synapses connecting hair cells and SGNs ., To visualize and quantify ribbon bodies , cochlear whole mount preparations from WT , Nrp1fl/+;Pax2Cre and Nrp1fl/fl;Pax2Cre mice at P5 ( n = 4 per group ) and 4 months ( n = 5 per group ) were immunostained with antibodies that bind to ribeye , a splice variant of CtBP2 ., The tissue samples were stained with Hoechst33342 to confirm whether the ribbon bodies were located on either OHCs or IHCs ., The time points described above were chosen so that we could track any possible changes in synaptic connectivity from just after birth to full maturity within the cochlea ., CtBP2 counts of both IHCs and OHCs at P5 showed no differences between WT and Nrp1 CKO samples ( Fig 3 ) ., No significant changes in immunostaining of IHCs was observed between Nrp1fl/fl;Pax2Cre and WT at 4 months; however , at this time point , the synaptic ribbon density ( CtBP2 counts ) in the OHC region in 4-month-old mice was significantly reduced ( p<0 . 01 ) for Nrp1fl/fl;Pax2Cre mutants ( 1 . 9 puncta/cell ) compared to the controls ( 1 . 4 puncta/cell ) ., Given this reduction in ribbon synapses , we next wanted to determine whether Nrp1 loss also conferred a loss of SGNs ( possibly through apoptosis ) ., Thus , we quantified SGN density at the apical , middle , and basal turns of the cochlea at P5 ( n = 4 per group ) and at 4-months ( n = 5 per group ) for WT , Nrp1fl/+;Pax2Cre and Nrp1fl/fl;Pax2Cre mice ( Fig 4 ) ., As expected from our ribbon synapse counts , at P5 no significant changes in SGN density were observed among the different genotypes or regions of the cochlea ( Fig 4E ) ., However , the SGN counts in 4-month-old mice ( Fig 4F ) showed decreased density of the neuronal cell bodies in Nrp1fl/fl;Pax2Cre mice compared to WT controls ( apical turn p = 0 . 03 , middle turn p = 0 . 03 , and basal turn p = 0 . 002 ) ., Interestingly , the number of SGNs lost in the absence of Nrp1 is unexpectedly high compared to the number of IHC ribbon synapses lost ( see Fig 3 and Discussion ) ., Nevertheless , the loss of OHC synaptic ribbons and diminished SGN density in 4-month-old Nrp1fl/fl;Pax2Cre mice , suggests that Nrp1 may play a role in maintaining SGN integrity during postnatal stages ., To examine the mechanism of SGN cell loss , we performed caspase-3 immunostaining ., Caspase-3 , a molecule necessary for the cellular apoptotic cascade , was identified by immunostaining in WT and Nrp1fl/fl;Pax2Cre cochleae to ascertain the fate of the SGNs once mice reached 4 months of age ., Caspase-3 positive neurons were found in Nrp1fl/fl;Pax2Cre mutants but not in WT mice , suggesting that the loss of OHC ribbon synapses resulted from pruning or the apoptosis of mature neurons ( Fig 4C and 4D ) ., Taken together , these data suggest a gradual loss of contacts between OHCs and SGNs in the absence of Nrp1 , which correlates with the death of SGNs around 4 months of age ., We next wanted to see if Nrp1fl/fl;Pax2Cre mutants also showed defects in cochlear innervation patterns to determine the extent to which Nrp1 may function in axon guidance in the cochlea ., Cochleae from WT and Nrp1fl/fl;Pax2Cre mutants at P5 and 4 months were immunostained with TUJ1 antibodies and assessed as whole-mount preparations ., At P5 , disorganized outer spiral bundles ( type II fibers ) were evident in cochleae of the Nrp1fl/fl;Pax2Cre mice at basal , mid , and apical turns ( n = 3 per group ) , but the radial fibers ( mostly type I SGNs ) appeared normal ( Fig 5A and 5B ) ., Compared to controls , we also observed significant disruptions to the normal patterns of innervation in cochleae from 4-month-old Nrp1fl/fl;Pax2Cre mice ( Fig 5C and 5D ) ., In a normal cochleae , 90–95% of the SGNs innervate IHCs; the remaining 5–10% of neurons travel beyond IHCs to innervate OHCs in an en passant fashion 15 ., TUJ1 immunostaining of cochlear nerve fibers extending into the hair cell region in 4-month-old Nrp1fl/fl;Pax2Cre mutants ( basal turn ) revealed aberrant axons with abnormal innervation of OHCs ., Mid-modiolar cross-sections of the cochlea of the Nrp1fl/fl;Pax2Cre mice ( 4-month-old ) also showed disorganized innervation of the outer hair cells ( Fig 6 ) ., Neuropilin-1 can be activated by both secreted semaphorins and Vascular Endothelial Growth Factors ( VEGFs ) 16 ., Given this , we wanted to ask next whether mice with a variant of Nrp1 that fails to bind secreted semaphorins ( Nrp1sema- ) showed cochlear innervation defects similar to the Nrp1 CKO line 17 ., We first used anti-TUJ1 antibodies to examine the overall distribution of nerve fibers in mutant and WT cochleae at E16 . 5 ( Fig 7A–7F ) ., For each sample , we also performed anti-Myo6 and anti-Sox2 immunostaining to identify the hair cells and supporting cells , respectively ., Compared to cochleae from WT littermates ( Fig 7A–7C ) , cochleae from Nrp1sema-/sema- mice showed nerve fibers in great excess with many that extended past the OHC region and even sometimes past the Deiters’ cell region ( Fig 7D–7F ) ., Cochleae from Nrp1sema-/sema- mice showed a normal distribution of hair cells and supporting cells ( Fig 7B , 7C , 7E and 7F ) indicating the innervation defects here were not due to changes in organ of Corti formation ., Using E18 . 5 samples from the Nrp1sema-/sema- mice , we next delineated the distribution of SGN afferents using Syt1 antibodies and cochlear efferents using Gap43 antibodies 10 ., Compared to the apex and middle regions of control cochleae , Nrp1sema-/sema- cochleae showed a significant increase in Syt+ fibers in the OHC region ( Fig 7K and 7O ) , but no changes in the distribution of Gap43+ efferent fibers ., At the base , we found no significant increases in Syt+ fibers in the OHC region of Nrp1sema-/sema- cochleae overall ( Fig 7 ) but did often see unusual nerve bundles that were both Syt1- and Gap43-positive ( see arrowheads in 7M and N ) ., These unusual bundles often took torturous paths toward the organ of Corti and terminated in the hair cell region or just beyond ., In addition , the anatomical origins of these neurons were not clear in that the processes appeared to come from outside of the cochlea and not Rosenthal’s canal where the SGN cell bodies are located ., Nevertheless , cochleae from Nrp1sema-/sema- mice showed innervation defects that , to a large extent , mirrored the phenotypic defects in the Nrp1 CKO mice ., This indicates that neuropilin-1 receptor activation by secreted semaphorins is necessary for normal cochlear innervation ., To further investigate the role of Neuropilin-1/Sema-3A in mediating SGN migration and refinement , we used small interfering RNAs ( siRNAs ) targeting Nrp1 mRNA to reduce neuropilin-1 protein levels in SGNs in cell culture ., A transient transfection with predesigned siRNA oligonucleotides decreased neuropilin-1 protein expression as measured by Western immunoblotting ( Fig 8 ) ., The SGN explant culture showed that semaphorin-3A , at a concentration of 250 ng/mL , mediated axonal repulsion ( Fig 8B ) ., The neurite outgrowth experiment was continued using the concentration of Nrp1 siRNA oligonucleotide ( 50nM ) that produced the greatest knockdown ( approximately 60% decrease from the control ) ., Nrp1 knockdown caused by siRNA transfection decreased neurite outgrowth and abolished the ability of Sema3a to decrease neurite outgrowth , suggesting that Sema3a repulses SGNs in an Nrp1-depended manner ( Fig 8I ) ., In contrast , the negative control scrambled siRNA neither decreased Nrp1 protein nor abolished Sema3a activity ., These experiments further support a key role for Nrp1/Sema3a signaling in cochlear innervation ., To investigate the interaction between semaphorin-3A and SGNs , we established whole cochlear cultures from E17 . 5 mice and placed them in media containing either control IgG-Fc or semaphorin-3A-Fc ( 20nM ) ., To determine whether semaphorin-3A altered hair cell innervation , the tissue samples were labeled with TUJ1 and Myo6 antibodies and imaged by confocal microscopy ., Compared to control samples that showed robust hair cell innervation ( Fig 8G ) , samples treated with semaphorin-3A-Fc showed significantly reduced innervation of the sensory epithelium ( Fig 8H ) ., To quantify this change in innervation , high-resolution confocal z-stacks were taken from the volume of tissue occupied by the HCs ., Compared to controls , semaphorin-3A decreased innervation around the sensory epithelium by nearly 60% ( Fig 8J ) ., These data suggest a possible role for semaphorin-3A in inhibiting SGN outgrowth ., For a detailed analysis of the entire auditory pathway in Nrp1fl/fl;Pax2Cre mutants , we next evaluated OHC activity using DPOAE and neuronal responses by ABR wave I peak-to-peak amplitudes ., DPOAEs , cochlear responses generated after two simultaneous pure tone frequencies , are objective indicators of OHC functional status 18 ., OHC function was determined to be normal in Nrp1 mutants as DPOAE levels for Nrp1fl/fl;Pax2Cre , Nrp1fl/+;Pax2Cre , and WT groups did not differ significantly at 2 and 4 months of age ., ABR test results of the 2-month-old mice showed that the hearing thresholds of the Nrp1fl/fl;Pax2Cre group were significantly higher than WT controls at 4kHz , 8kHz , 16kHz , 24kHz , and 32kHz ., At 4 months of age , Nrp1fl/fl;Pax2Cre mice developed elevated hearing thresholds at all tested frequencies except 12 kHz when compared to WT controls ( Fig 9 ) ., Peak-to-peak analysis of wave I was calculated from the ABR data described above ., Wave I is thought to indicate the summed activity of SGN contact with hair cells , so a normal DPOAE with a diminished wave I peak would suggest dysfunction of the SGNs , IHCs , or the synapses between them 19 ., At 2 months of age , no significant changes in wave I amplitude were found among the three groups of mice; however , the Nrp1fl/fl;Pax2Cre mutants at 4 months of age recorded significantly lower wave I amplitudes than WT mice at 8 kHz , 12kHz , and 32 kHz ., Paired with our immunostaining results of IHC defects in 4-month-old Nrp1fl/fl;Pax2Cre mice , the reduced wave I amplitude suggests the contribution of cochlear neural damage in the hearing loss seen in 4-month-old Nrp1 mutants ( Fig 10 ) ., The composition of endolymph and the maintenance of the endocochlear potential are determined by ion balance regulated by the stria vascularis ( SV ) 20 ., We hypothesized that abnormal vascularization of the SV could lead to electrolyte imbalance , resulting in abnormal hearing thresholds ., To test this hypothesis , we investigated the morphology of the micro-vessels of the SV at the basal turn of the cochlea ( n = 3 per group ) ., The lectin immunostaining of the Nrp1fl/fl;Pax2Cre cochleae at P5 and 4-months-old demonstrated grossly enlarged SV microvessels ( Fig 11 ) ., The minimum and maximum microvessel diameter of the Nrp1fl/fl;Pax2Cre mice were 4 . 17μm and 43 . 67μm at P5 , and 3 . 22μm and 95μm in 4-month-old mice , respectively ., The minimum and maximum microvessel diameters of the WT mice were 5 . 32μm and 23 . 90μm at P5 , and 5 . 2μm and 26 . 15μm in 4-month-old mice , respectively ., Overall , the maximum microvessel diameter in 4-month-old Nrp1fl/fl;Pax2Cre mice was 3 . 6 fold higher than WT controls ., Thus , future endocochlear potential studies are needed to pinpoint the effect of Nrp1 knockout on normal functioning of auditory hair cells ., There exists a growing literature supporting the notion that ARHL may be associated with degenerative changes in the cochlear nerve and its synapses 21 , 22 ., Although this phenomenon and that of the classically defined neural presbycusis are now well studied histologically , little is known about the molecular basis for this pathology ., Using a meta-analysis GWAS approach we have defined several candidate regions for ARHL in mice , one of which included Nrp1 9 ., Nrp1 is a well-known factor in neuronal and cardiovascular development 13 , 17 ., Its homolog , Nrp2 has been shown to be involved in inhibiting type I SGNs from the OHC region in the developing cochlea 10 ., While substantial data exists for the role of Nrp1 in tumorigenesis and embryonic development , to date , the role of Nrp1 in the postnatal development of the cochlear apparatus remains unclear ., According to previously published cochlear nerve microarray data , Nrp1 expression in the spiral ganglion shows a peak at E16 . 5 followed by a dip between E16 and P0 , and a general trend of increased expression up to two weeks into postnatal development 23 ., These findings are consistent with our data as we found upregulated Nrp1 postnatal expression in the organ of Corti and in the SGNs in the first postnatal week , a critical period for maturation of hair cell innervation which suggests a role for Nrp1 in this process ., In the peripheral auditory system , the type I SGN afferent fibers undergo significant reorganization during embryonic development in mice 10 ., In this study , we have identified Nrp1 to be a critical component of this reorganization process ., Previously , mice lacking normal Nrp1 function ( Nrp1Sema- mutants ) showed pathfinding defects in vestibular ganglion neurons 17 , defasciculation of the intercostal nerves , and crossing bundles to neighboring nerves of the Nrp1Sema- mutants 24 ., Many previous studies have also demonstrated a role for semaphorin-3A in axonal chemorepulsion , including repulsion of sensory and cortical axons 11 , 25 ., Our SGN explant and semi-intact cochlear cultures also demonstrated semaphorin-3A repels SGNs , which suggests semaphorin-3A can normally inhibit SGN outgrowth ., Complementary to these findings , Nrp1sema-/sema- cochleae showed dramatically enhanced innervation by Syt-positive fibers during developmental stages , suggesting excessive innervation by SGNs possibly due to the absence of a repulsive signal ., We do not yet know whether this was due to increased numbers of SGNs , increased complexity of individual SGNs , or ectopic innervation of the cochlea by a different population of neurons ( e . g . vestibular neurons ) ., During embryonic development , Type II SGNs pass by the IHCs and reach the OHC area then extend toward the base of the cochlea forming en passant contacts with 3 to 10 OHCs within the same row ., These projections gather beneath the rows of OHCs to form 3 outer spiral bundles 15 ., Here , we demonstrate that Nrp1 conditional mutants show disorganized outer spiral bundles at early neonatal stages ( P5 ) and that these pathfinding defects are apparent in the older ( 4-month-old ) mice ., In addition , Nrp1sema-/sema- cochleae showed excessive numbers of SGNs present in the OHC region in late embryonic stages ., There were several pieces of evidence shown here that implicate Nrp1 in age-related hearing loss ., First , we found that SGN density was lost in the Nrp1fl/fl;Pax2Cre cochleae over time ( Fig 4 ) ., Since most of these neurons are likely type I SGNs , this potentially explains why the Nrp1fl/fl;Pax2Cre mice showed elevated ABR thresholds and wave 1 amplitude reductions ., Larger ABR wave I amplitude shifts at equal sound pressure levels are associated with greater auditory nerve threshold elevation 26 ., Second , Nrp1fl/fl;Pax2Cre cochleae showed conspicuous defects in the stria vascularis , which normally maintains the ionic composition of the endolymph and promotes normal auditory transmission ., Although we did not detect any significant changes to otoacoustic emissions in the Nrp1fl/fl;Pax2Cre mice ( suggesting normal OHC function ) , it is possible that these mice have defects in IHC function that also , like the loss of SGNs , contributes to their altered hearing thresholds ., The Nrp1fl/fl;Pax2Cre mice did show reduced numbers of OHC ribbon synapses , but type II SGNs do not contribute to the canonical auditory pathway 27 , so it is unlikely that this phenotypic defect contributes to the changes in hearing thresholds ., One curious finding here was that the Nrp1fl/fl;Pax2Cre mice showed a profound loss of SGNs and only a mild loss of IHC ribbon synapses ., Although overall the differences between controls were not statistically significant ( Fig 3C ) , we did observe an almost 50% decrease in IHC ribbon synapses in 3 out of 5 of the 4-month-old Nrp1 mutants examined ., Since the majority of SGNs are type I and terminate on IHCs , it would be expected that there would be a commensurate reduction in IHC ribbon bodies ., One obvious cause for this discrepancy is that many of the CtBP2-positive bodies in IHCs from the Nrp1fl/fl;Pax2Cre cochleae may represent orphaned synapses that lack a postsynaptic terminal ., A second less likely possibility is that some remaining type I SGNs extend collateral processes that form synapses with the IHCs ., Overall , these data suggest that reduced SGN density in 4-month-old Nrp1 mutants , in addition to abnormal axonal pathfinding , lead to impairment of the OHCs synaptic integrity ., Consistent with the abnormal neuronal phenotype in 4-month-old Nrp1 deficient mice , we also observed a decline in ABR wave I amplitudes as they matured ., Nrp2 , the other member of the neuropilin family , is responsible for encoding a transmembrane receptor protein with sequence homology to Nrp1 but with different ligand binding affinities ., Neuropilin-2 receptors bind Sema-3 subtypes 3C and 3F , VEGF-A and VEGF-C isoforms 28 ., While Nrp2 plays a role in neuronal pathfinding , it has not been linked to neuron survival 10 ., We have shown , however , that Nrp1 likely plays a role in long-term neuronal survival and maintenance throughout life as SGN cell counts diminish with age in Nrp1 CKO mice ., These results were consistent with previously published data showing that Nrp1 is an essential factor in survival of the GnRH and trigeminal neurons by interacting with vascular endothelial growth factor ( VEGF ) ligands 29 , 30 ., While our results support a model in which Nrp1 signaling is necessary for the establishment of SGN projections in the postnatal period , Nrp1 also appeared to play an essential role in normal vascular development of the cochlea ., Our results show that Nrp1 deletion leads to enlarged vessels in the stria vascularis of early postnatal and adult mice , which may have an impact on the maintenance of the endocochlear potential 31 ., Animal procedures were performed at the Zilkha Neurogenetic Institute in accordance with the guidelines of the Institutional Care and Use Committee ( IACUC ) of the University of Southern California ., Nrp1fl/fl mice of mixed backgrounds ( CBA/CaJ x C57BL/6J ) strains were kindly provided by Dr . Henry Sucov ., Pax2Cre mice of mixed backgrounds strains ( CBA/CaJ x C57BL/6J ) were kindly provided by Dr . Takahiro Ohyama ., In Pax2Cre mice , Cre mRNA is detectable in the otic placode starting at the late presomite stage 32 ., Nrp1 CKO mice of either sex were obtained by crossing Nrp1fl/fl mice to Nrp1fl/+;Pax2Cre mice ., For postnatal collections , P0 was defined as the day of birth ., For genotyping of Nrp1 knockout mice , polymerase chain reaction ( PCR ) was performed using the following primers: Nrp1 forward 5’- AGGTTAGGCTTCAGGCCAAT-3’ , Nrp1 Reverse 5’ GGTACCCTGGGTTTTCGATT-3’; Pax2Cre Forward 5’-GCCTGCATTACCGGTCGATGCAACGA-3’ , Pax2Cre Reverse 5’-GTGGCAGATGGCGCGGCAACACCATT-3’ ., The Nrp1sema- line 17 was kindly provided by Dr . Alex Kolodkin of Johns Hopkins University ., Nrp1sema- mice were bred and maintained at either the Porter Neuroscience Research Facility ( Bethesda , MD ) under the guidelines of the NIDCD IACUC or at the Division of Comparative Medicine ( Washington , DC ) under the guidelines of the Georgetown University IACUC ., Nrp1sema- mice were maintained on a C57BL/6J background ., Male and female heterozygous mice were bred to generate homozygous mutants and littermate controls ., Genotyping was performed using the following WT and Nrp1sema- specific primers: AGGCCAATCAAAGTCCTGAAA ACAGTCCC and AAACCCCCTCAATTGATGTTAACACAGCCC ., Six-week-old C57BL/6 mice ( n = 8 ) and DBA/2J mice ( n = 7 ) were euthanized , and bilateral inner ears were harvested ., Cochlear tissues were collected , and left and right ear samples were combined and immediately processed with RNAqueous Total RNA Isolation Kit ( Life Technologies ) according to manufacturer’s instructions ., Total RNA was then converted to cDNA using the SuperScript III First-Strand Synthesis SuperMix ( Life Technologies ) ., PCR was performed using the primer pairs acquired from applied biosystems: assay ID: Mm00435379_m1 ., Each sample was run in triplicate along with the housekeeping gene , GAPDH ., Relative quantities of the transcripts were determined using the 2−ΔΔCt method using GAPDH as a reference ., Cochlear whole mount sample preparation: Mouse cochleae were dissected after the second hearing measurement and were fixed with 4% PFA overnight ., Fixed samples were decalcified using 10% EDTA , and dissected using the mouse cochlear dissection method from Eaton Peabody Laboratories at the Massachusetts Eye and Ear Institute website ( http://www . masseyeandear . org/research/otolaryngology/investigators/laboratories/eaton-peabody-laboratories ) ., For the Nrp1sema- mice samples , embryonic cochleae were fixed for 30 minutes in 4% PFA and then rinsed extensively in 1X PBS before dissection and immunostaining ., Cochlear frozen section sample preparation: Fixed heads were sequentially dehydrated in 15% and 30% sucrose , embedded in Tissue-Tek O . C . T . compound ( Sakura Finetek ) and snap frozen on dry ice ., Blocks were sectioned ( 12 μm thickness ) on a Leica 3050 S cryostat in a cranial-to-caudal coronal direction ., SGN explant three-dimensional culture: Sensory epithelia of cochleae with attached SGNs were dissected from the E16 . 5 embryos and placed in Leibovitz’s L-15 medium ( Invitrogen ) ., The isolated SGNs were cut into four equal pieces starting from one turn away from the apex of cochlea ., Each SGN explant was transferred in a drop of phenol red-free Matrigel ( Corning ) and placed on poly-D-lysine ( 50 μg/ml ) -coated glass coverslips in a 24-well plate ., After complete solidification of the Matrigel , DMEM/F12 medium supplemented with 10% fetal bovine serum , 1% N2 supplement , and 0 . 3 mg/ml ampicillin were added and maintained in the culture for 3 days ., After fixation , dissected cochleae or tissue sections were permeabilized with 0 . 2% TrionX-100 followed by incubation in 10% blocking serum for 2 hours at room temperature ., The samples were incubated with the primary antibody at 4°C for 24 to 48 hours and exposed to secondary antibodies for 2 hours at room temperature ., Using a Carl Zeiss LSM 780 laser scanning microscope ( AxioObserver . Z1 ) , 3 representative images were taken for each slide , and the total , average , and maximal neurite lengths per SGN explant were measured using Metamorph software ( Molecular Devices ) 33 , 34 ., Antibodies used in this study were as follows: Alexa 488-conjugated mouse anti-neuron specific class III beta tubulin ( anti-TUJ1 ) ( 1:300; Covance ) , rabbit anti-Neuropilin-1 ( 1:50;Abcam ) , rabbit anti-Semaphorin-3A ( 1:100; Abcam ) , mouse anti-CtBP2 ( 1:200; BD Biosciences ) , mouse-anti-TUJ1 ( 1:1 , 000 , Covance ) , rabbit-anti-myosin VI ( 1:1 , 000 , Proteus Biosciences ) , goat-anti-Sox2 ( 1:300 , Santa Cruz Biotechnology ) , chicken-anti-synaptotagmin-1 ( 1:1 , 000 , Aves Labs ) , mouse-anti-GAP43 ( 1:2 , 000 , Chemicon ) , Alexa Fluor-488 anti-mouse ( 1:500; Life technologies ) , Alexa Fluor 594 anti-goat ( 1:500; ThermoFisher ) ., Fluorescent dye Hoechst 33342 ( 0 . 1 μg/mL; Southern Biotech ) was used for DNA labeling ., Blocking peptide for Anti-Semaphorin-3A antibody ( Abcam ) was used for Sema-3A immunostaining ( negative controls ) ., For synaptic ribbon-to-hair cell ratios , tissue sections at the basal turn of the cochlea were selected , and the number of synaptic ribbons was compared separately to the number of inner hair cells and to the number of outer hair cells per section ., The synaptic ribbon-to-hair cell ratios for WT and Nrp1fl/fl;Pax2Cre 4-month-old mouse cochleae ( n = 5 per group ) and P5 mouse cochleae ( n = 4 per group ) were assessed ., Spiral ganglion cells of the P5 and 4-month-old mice were counted at apical , mid , and basal turns of the cochlea ( n = 5 per group at each turn ) ., To quantify numbers of Syt1+ fibers in the OHC region of Nrp1sema-/sema- cochleae and their littermate controls , we determined the number of fiber tracks extending into the OHC region and normalized that value to the longitudinal distance of the cochlea within that region ., Sample sizes: 9 WT cochleae and 13 Nrp1sema-/sema- cochleae ., In situ hybridization was performed as previously described 35 ., Briefly , embryonic day E15 . 5 heads were fixed in 4% paraformaldehyde in PBS overnight at 4°C , sunk in 30% sucrose in PBS at 4°C , incubated in Tissue-Tek O . C . T . compound ( Sakura Finetek ) at room temperature for 10 min and frozen on dry ice ., Sections , 14μm thick , were cut using a Leica 3050 S cryostat ., RNA probes for mouse Nrp1 ( GE Dharmacon , Clone ID 6409596 ) and mouse Sema3a ( GE Dharmacon , Clone ID 30532393 ) were synthesized , labeled with digoxigenin , and hydrolyzed by standard procedures ., In situ hybridization images were obtained under bright-field microscopy ( BZ9000; Keyence , Osaka , Japan ) ., SGN explants in antibiotic-free medium were transfected with 50 nM predesigned Nrp1 siRNA oligonucleotides ( Santa Cruz Biotechnology ) using Lipofectamine 3000 ( Invitrogen ) as per manufacturer’s instructions ., Some explants were treated with 250 ng/ml Sema-3A-Fc ( R&D Systems ) ., We used a scrambled siRNA oligonucleotide that did not exhibit homology to any known encoding region as a negative control ., The siRNA-mediated knockdown efficiency was determined by Western immunoblotting ., After Nrp1 siRNA transfection and semaphorin-3A treatments , the Matrigels covering SGN explants were removed with cell recovery solution ( Corning ) and the explants were homogenized in RIPA lysis buffer supplemented with a cocktail of protease in | Introduction, Results, Discussion, Methods, Conclusions | Neuropilin-1 ( Nrp1 ) encodes the transmembrane cellular receptor neuropilin-1 , which is associated with cardiovascular and neuronal development and was within the peak SNP interval on chromosome 8 in our prior GWAS study on age-related hearing loss ( ARHL ) in mice ., In this study , we generated and characterized an inner ear-specific Nrp1 conditional knockout ( CKO ) mouse line because Nrp1 constitutive knockouts are embryonic lethal ., In situ hybridization demonstrated weak Nrp1 mRNA expression late in embryonic cochlear development , but increased expression in early postnatal stages when cochlear hair cell innervation patterns have been shown to mature ., At postnatal day 5 , Nrp1 CKO mice showed disorganized outer spiral bundles and enlarged microvessels of the stria vascularis ( SV ) but normal spiral ganglion cell ( SGN ) density and presynaptic ribbon body counts; however , we observed enlarged SV microvessels , reduced SGN density , and a reduction of presynaptic ribbons in the outer hair cell region of 4-month-old Nrp1 CKO mice ., In addition , we demonstrated elevated hearing thresholds of the 2-month-old and 4-month-old Nrp1 CKO mice at frequencies ranging from 4 to 32kHz when compared to 2-month-old mice ., These data suggest that conditional loss of Nrp1 in the inner ear leads to progressive hearing loss in mice ., We also demonstrated that mice with a truncated variant of Nrp1 show cochlear axon guidance defects and that exogenous semaphorin-3A , a known neuropilin-1 receptor agonist , repels SGN axons in vitro ., These data suggest that Neuropilin-1/Semaphorin-3A signaling may also serve a role in neuronal pathfinding in the developing cochlea ., In summary , our results here support a model whereby Neuropilin-1/Semaphorin-3A signaling is critical for the functional and morphological integrity of the cochlea and that Nrp1 may play a role in ARHL . | Neuropilin-1 is a member of the neuropilin family acting as an essential cell surface receptor involved in semaphorin-dependent axon guidance and VEGF-dependent angiogenesis and lies within our previously identified ARHL GWAS interval ., In this study , we investigated the role of Neuropilin-1/Semaphorin-3A signaling in the functional and morphological integrity of the cochlea , specifically the innervation and vascularization patterns ., Detailed analyses of the cochleae of 4-month-old Nrp1 CKO mice showed abnormalities in ribbon synapses , innervation of the hair cells , and microvessels of the stria vascularis ., We show also that Neuropilin-1/Semaphorin-3A signaling plays an important role in cochlear innervation . | medicine and health sciences, nervous system, ears, gene regulation, electrophysiology, neuroscience, neurites, animal models, inner ear, model organisms, experimental organism systems, nerve fibers, neuronal dendrites, research and analysis methods, specimen preparation and treatment, staining, small interfering rnas, animal cells, gene expression, mouse models, head, biochemistry, rna, immunostaining, cellular neuroscience, anatomy, nucleic acids, cell biology, synapses, cochlea, physiology, neurons, genetics, biology and life sciences, cellular types, non-coding rna, neurophysiology | null |
journal.pgen.1005874 | 2,016 | Six Novel Loci Associated with Circulating VEGF Levels Identified by a Meta-analysis of Genome-Wide Association Studies | Vascular Endothelial Growth Factor ( VEGF ) is secreted largely by endothelial cells and plays a key role in several physiological and pathological conditions ., During growth , development , and maintenance of the circulatory system , VEGF is the principal pro-angiogenic factor and it has additionally , a neurotrophic role ., High levels of circulating VEGF have been observed in individuals with various vascular diseases ( myocardial infarction 1 , stroke 2 , 3 , heart failure 4 , and atherosclerosis 5 ) , neurodegenerative conditions ( age-related cognitive decline 6 and Alzheimer dementia 7 ) , immune inflammatory disorders ( rheumatoid arthritis 8 , inflammatory bowel disease 9 , and Behçet’s disease 10 ) and cancers ( breast 11 , 12 , uterine 13 , gastrointestinal 14 , 15 , lung 16 and prostate 17 ) ., An increase of VEGF levels has also been found in patients with diabetes 18 and various reproductive disorders 19–21 ., Reduced circulating VEGF levels have been observed in amyotrophic lateral sclerosis 22 and spinal bulbar muscular atrophy 23 ., Moreover , since VEGF levels are pharmacologically modifiable , understanding the determinants of circulating VEGF could support efforts directed at risk prediction , prevention and therapy ., Circulating VEGF levels are highly heritable 24–27 leading to a search for specific genetic determinants within the Vascular Endothelial Growth Factor A ( VEGFA ) gene 27–29 ., Several putative candidate genes were then identified but could not be consistently replicated 10 , 30–41 ., A genome-wide linkage study of VEGF levels identified the 6p21 . 1 VEGFA gene region as the main quantitative trait locus determining variation in VEGF serum levels 27 ., Specific variants at this locus were also identified as the strongest associations in the first genome-wide association study ( GWAS ) of circulating VEGF levels based on data from 3 large cohort studies in this consortium , wherein two addition loci , located at 8q23 . 1 , and 9p24 . 2 were also identified 42 ., We have now conducted a new GWAS meta-analysis using an extended sample , the largest to date , and a deeper genomic coverage based on imputation to the 1000 genomes panel to identify additional genetic variants that explain variation in circulating VEGF concentrations ., A GWAS meta-analysis of VEGF levels was performed in 16 , 112 individuals from 10 cohorts of European ancestry ( see Materials and Methods and Section 1 in S1 Text for details ) : the Age Gene/Environment Susceptibility Reykjavik Study ( AGES ) , the Cilento study ( Cilento ) , the Framingham Heart Study ( FHS ) , the Ogliastra Genetic Park ( OGP ) , the Prospective Investigation of the Vasculature in Uppsala Seniors Study ( PIVUS ) , and the Val Borbera study ( VB ) served as discovery cohorts; the Gioi population , the Sorbs population , the STANISLAS Family Study ( SFS ) and a sample of hypertensive adults ( HT ) served as replication cohorts ., The characteristics of study participants are shown in Table 1 ., The mean age of the participants was 54 . 8 years , ranging from 30 . 4 years in SFS to 76 . 2 years in the AGES ., The percentage of females in the overall sample was 54% , ranging from 37% in OGP to 64% in Sorbs ., To account for differences in age distribution and gender among the studies , both age and sex were subsequently used as covariates in the association analyses ., Across studies , median VEGF levels ranged from 27 . 0 to 393 . 6 pg/ml , with the lowest median levels in HT and SFS studies in which VEGF was measured in plasma rather than serum ( see Section 2 in S1 Text for details ) ., This is expected since VEGF levels are higher in serum than in plasma secondary to VEGF release from platelets during clot formation 43 , 44 ., Differences in VEGF levels also partly reflect demographic and assay differences between the cohorts ., An overview of the study design is presented in Fig 1 ., Due to heterogeneity in the distribution of VEGF levels among the cohorts ( Table 1 ) , a sample size-weighted Z-score ( rather than an inverse-variance ) method was chosen for the meta-analysis ., A discovery GWAS meta-analysis was carried out for 6 , 705 , 861 autosomal variants in 13 , 312 individuals from the six cohorts described in the “Characteristics of study participants” section ( Stage 1 ) ., A Quantile-Quantile plot for the investigated variants revealed many more variants with lower observed p-values ( P ) than expected ( S1 and S2 Figs ) ., There were 920 variants in 5 chromosomal regions ( 6p12 . 1 , 8q23 . 1 , and 9p24 . 2 , which have been previously described and two novel regions at 5q14 . 3 and 10q21 . 3 ) that reached genome-wide significance ( P<5x10-8 ) in the discovery sample ( S2 Table ) ., To identify independently associated variants within these 5 genome-wide significant genomic regions , conditional analyses were carried out in the study with the largest number of samples ( FHS ) ., This approach was selected since our use of a Z-score meta-analysis , which does not yield effect size estimates , precluded the use of aggregate results for conditional analyses ., The conditional analyses revealed 10 independent signals ( 4 previously known and 6 novel variants ) ., These 10 Stage 1 variants were carried forward to in-silico ( Stage 2 ) and subsequent de-novo ( Stage 3 ) replication ., Further , 57 variants in 13 loci were suggestively associated at 5x10-8<p-value<1x10-5 ., At each locus , a single independent signal was identified using a clumping procedure , and the most strongly associated variant at each of these 13 loci was also tested in the in-silico replication ., Among them , 2 variants reached a genome-wide level of significance in the joint meta-analysis of discovery and in-silico replication samples and these two were also carried forward for the de-novo replication ., So a total of 12 variants were carried forward to the de novo replication ., Overall , 10 of these 12 variants , 8 of the 10 independent variants identified in Stage 1 and the 2 variants identified in Stage 2 ( combined discovery and in-silico replication ) , were successfully replicated in the Stage 3 meta-analysis of the combined discovery , in-silico , and de-novo replication samples ( Fig 2 and Table 2 ) ., For these variants , an additional inverse variance-weighted meta-analysis was performed as a secondary analysis on the Stage 3 data , including the discovery and both replication cohorts ., These secondary meta-analysis results , reported in the Table 2 , are concordant with our original analysis results ., Forest plots reporting the effects of the 10 replicated variants in all the cohorts and the cumulative effect in the inverse-variance meta-analysis are shown in the Fig, 3 . Among those 10 signals , 4 were located in novel chromosomal regions ( 5q14 . 3 , 10q21 . 3 , 16q24 . 2 , and 18q22 . 3 ) and 6 ( 2 novel , independent variants and 4 previously known signals ) were located in previously identified chromosomal regions ( 6p21 . 1 , 8q23 . 1 , and 9p24 . 2 ) ., The leading SNP on chromosome 5q14 . 3 was rs114694170 ( P = 6 . 79x10-13 ) ., This new association is located in the intronic region of the myocyte enhancer factor 2C ( MEF2C ) gene ., Conditional analyses did not identify additional independent variants in the region ., In the locus on chromosome 10q21 . 3 , the most significantly associated variant was rs74506613 ( proxy rs10761741 used for in-silico replication has r2 of 0 . 97 , P = 1 . 17x10-19 ) located within the intronic region of the jumonji domain containing 1C ( JMJD1C ) gene ., Conditional analyses did not identify any other independent variants in this region ., Two additional loci reached a genome-wide significance level in the meta-analysis of the combined discovery and replication samples ., At the locus on chromosome 16q24 . 2 , the most significantly associated variant was rs4782371 ( P = 1 . 59x10-09 ) located within the intronic region of the zinc finger protein , FOG family member 1 ( ZFPM1 ) gene ., At chromosome 18q22 . 3 , the leading variant was rs111939830 which along with the second leading variant rs2639990 ( used as proxy for de novo replication for rs111939830 , r2 = 0 . 48 , P = 1 . 72x10-08 ) was located in the intronic region of the zinc binding alcohol dehydrogenase domain containing 2 ( ZADH2 ) gene ., The most significant variant on chromosome 6p21 . 1 was rs6921438 ( P = 7 . 39x10-1467 ) , already identified in the previous GWAS 42 ., Two additional independent variants were also identified at this locus after conditional analyses ., One was rs1740073 ( P = 2 . 34x10-17 ) which was in LD with rs4416670 reported in the previous GWAS ( r2 = 0 . 15 ) 42 ., Although the LD between these two SNPs is relatively low , rs4416670 and rs1740073 are in close physical proximity ( 3055 base-pair distance ) and conditional analysis confirmed that rs1740073 eliminated the signal of rs4416670 ( P = 4 . 16x10-21; before adjusting for rs1740073 , P = 0 . 727; after adjusting for rs1740073 ) , hence we believe the two SNPs , rs1740073 and rs4416670 , both represent a single locus of genetic variation ., This rs1740073 SNP is located about 22Kb downstream from rs6921438 and both are located upstream of the gene C6orf223 , which encodes an uncharacterized protein ., The other independent variant identified , about 221kb distant from the main signal rs6921438 , was rs34528081 ( P = 1 . 52x10-18 ) , a novel variant , located upstream of the VEGFA gene and the mitochondrial ribosomal protein S18A ( MRPS18A ) gene ., The values of r2 between the 3 variants at 6p21 . 1 are extremely low ( rs6921438-rs1740073 = 0 . 01 , rs6921438-rs34528081 = 0 . 007 , rs1740073-rs34528081 = 0 . 01 ) , suggesting that the 6p21 . 1 region has 3 independent variants that modulate circulating VEGF levels ., The leading variant identified on chromosome 8q23 . 1 was rs6993770 ( P = 2 . 44x10-60 ) ., This SNP , located within an intron of the zinc finger protein multitype 2 ( ZFPM2 ) gene , was already known to be associated with circulating VEGF levels 42 ., On chromosome 9p24 . 2 the most significantly associated SNP was rs2375981 ( P = 1 . 48x10-100 , which is in strong LD with rs10738760 ( r2 = 0 . 81 ) reported in the previous GWAS 42 ) ., This variant lies downstream of the very low-density lipoprotein receptor ( VLDLR ) and upstream of the potassium voltage-gated channel subfamily V member 2 ( KCNV2 ) genes ., One novel independent signal also found in this region using conditional analyses was rs7043199 ( P = 5 . 12 x10-14 ) located about 71kb upstream of rs2375981 , in the VLDLR-AS1 gene and upstream of the VLDLR gene ., No LD exists between the two variants ( r2 = 0 . 0008 ) ., Thus , in the 9p24 . 2 region , there are 2 independent variants able to influence VEGF levels ., A genetic score was calculated for each individual using information on the 10 VEGF replicated variants ., This genetic score explained 52% of the observed variability in circulating VEGF levels in FHS ., The proportions of variance in circulating VEGF explained by these 10 replicated variants in PIVUS , Cilento , AGES , VB , HT , and SFS are 48% , 46% , 24% , 24% , 21% and 19% , respectively ., The observed differences in the proportion of variance explained might be due to heterogeneity in effect sizes of some SNPs related to the trait variability in distribution of VEGF levels across the cohorts ( Table 2 ) ., Accordingly , the explained variability is similar in the cohorts where a similar distribution of VEGF levels was observed ( Table 1 ) ., To identify putative functional elements at the associated loci , ENCODE data related to chromatin modifications and hypersensitivity DNAse sites ( DHSs ) included in HaploReg 45 were analyzed ., Among the 10 replicated variants and their 126 proxies ( r2>0 . 8 ) , 16 variants were located in regions reported as DHSs in 5 or more different cell lines ., Among these 16 , 11 variants ( rs114694170 on chromosome 5p14 . 3 , rs6993770 on chromosome 8q23 . 1 , rs7043199 on chromosome 9p24 . 2 , 5 proxies of rs74506613 on chromosome 10q21 . 3 and 3 proxies of rs4782371 on chromosome 16q24 . 2 ) were also located in a promoter and/or enhancer histone mark ., These results suggest a potential functional role of these variants ., A large database assembled by one of the authors ( AJD ) that included eQTL association results from 61 studies ( detailed Section 3 in S1 Text ) was queried for the 10 replicated variants identified in the GWAS and their 126 proxies ( r2>0 . 8 ) ., Eighty-four variants in three loci ( 1 replicated variant and 83 proxies of two additional replicated variants ) were found in the database ., The variant rs6993770 on chromosome 8q23 . 1 was a trans eQTL for the CXCL5 gene; rs609303 ( proxy of rs111939830 ) on chromosome 18q22 . 3 was a cis eQTL for the TSHZ1 gene ., On chromosome 10q21 . 3 82 proxies for rs74506613 were identified: 2 variants were trans eQTL for 6 genes ( AQP10 , CXCL5 , GUCY1A3 , ITGA2B , MYL9 , and NRGN ) and 81 were cis eQTLs for 3 genes ( JMJD1C , NRBF2 and REEP3 ) ; one variant rs10761779 is both a trans and cis eQTL ., All 84 variants identified as eQTL in this search are listed in S3 Table ., In order to identify biological pathways involved in the modulation of VEGF protein levels two pathway analysis approaches were applied ., MAGENTA software 46 was applied to the Stage 1 meta-analysis results , to identify the known biological pathways most strongly represented among all the variants associated with circulating VEGF concentrations ( see Materials and Methods ) ., Overall , 3 , 216 biological pathways ( with at least 10 genes ) and 168 , 932 genes were examined ., This pathway analysis identified 18 biological pathways , 3 molecular functions and 2 cellular components significantly associated with VEGF levels at a nominal Gene Set Enrichment Analysis ( GSEA ) p-value ≤0 . 01 ., Among these , only the ERK5 pathway reached statistical significance after correction for multiple testing ( FDR threshold of 0 . 05 ) ., The Ingenuity Pathway Analysis software ( IPA , www . qiagen . com/ingenuity ) was used to explore functional relationships between genes in the VEGF associated loci ., A total of 26 genes located at and adjacent to the 10 replicated variants were selected as focus genes for IPA analysis ( S4 Table ) ., Among them , 17 genes were found to be biologically linked in a unique network of 70 molecules as shown in Fig, 4 . The associated functions for this network were organism development , especially early embryonic and later cardiovascular system development ., The probability that 17 genes would be linked in a randomly designated set of 26 genes using data from the Global Molecular Network was 1 . 0x10-42 ., Thus , it appears extremely unlikely that this network has been identified purely by chance ., In this GWAS meta-analysis of circulating VEGF levels , we identified 10 independent variants located in 7 chromosomal loci; 4 of those variants had been described in a previous GWAS 42 ., We now describe 6 novel variants , 4 of which were in newly identified chromosomal regions ( 5q14 . 3 , 10q21 . 3 , 16q24 . 2 , and 18q22 . 3 ) whereas 2 were identified through conditional analyses at previously described loci ( 6p21 . 1 and 9p24 . 2 ) ., These 10 variants explain about 52% of VEGF phenotypic variance in the largest cohort in this study , with the 6 novel variants increasing the explained variance by 4% compared to the 48% described by Debette et al . for the 4 previously identified loci 42 ., This increase represents a valuable addition to the proportion of variance explained when compared to the results obtained from GWAS of other complex traits 47–50 ., The newly identified regions include many interesting and plausible candidate genes with angiogenic and neurotrophic roles ., The leading variant on chromosome 5 was located within an intron of the MEF2C gene ., This protein has a demonstrated role in cardiac myogenesis , morphogenesis and in vascular development ., MEFC2 knock out is embryonically lethal due to cardiac and vascular abnormalities ., MEFC2 also supports cortical development and variants in this region have been associated with severe neurodevelopmental problems in humans such as developmental retardation , cerebral malformations 51 , 52 , stereotypic movements and epilepsy ., MEF2C was also reported to be associated with retinal vascular caliber in the Cohorts for Heart and Ageing Research in Genomic Epidemiology ( CHARGE ) consortium 53 , which is particularly interesting given the known role of VEGF in proliferative retinopathy and macular degeneration ., MEF2C may be a transmitter of VEGF signaling and has been shown to be regulated by VEGF in-vitro , as a key mediator 54 ., The leading variant on chromosome 10 was located in an intronic region of JMJD1C , a protein-coding gene with an intriguing role in many biological processes ranging from platelet and endothelial cell function to DNA repair 55 ., Thyroiditis 56 and fatty liver disease 57 have been associated with this gene ., A GWAS of plasma liver enzymes revealed an association of rs7923609 ( P = 6 . 0x10-23 , G = risk allele ) with elevated enzyme levels indicating abnormal liver function ., Interestingly , this SNP also showed an association with VEGF levels in our study ( P = 1 . 15x10-12 ) with the G allele associated with higher levels 58 ., In a mouse model , it was noted that VEGF promotes proliferation of hepatocytes through reestablishment of liver sinusoids by proliferation of sinusoidal endothelial cells; thus VEGF may mediate the genetic association observed 59 between JMJD1C variants and hepatic steatosis ., JMJD1C and MEF2C genes were found associated to platelet count and volume in a European ancestry GWAS 49 ., Further , a variant ( rs7896518 , P = 2 . 93x10-15 ) located in an intron of the JMJD1C gene showed an association with platelet counts ( P = 2 . 3x10-12 ) in an African American GWAS 60 ., In a second European ancestry GWAS of platelet aggregation another SNP in the same gene , rs10761741 , showed an association with epinephrine-induced platelet aggregation with the T allele being associated with greater aggregation 61 ., Interestingly , this T allele of rs10761741 was also associated with higher circulating VEGF levels ( P = 7 . 10x10-15 ) ., Because both platelets and VEGF play important roles in the development of atherosclerosis and arterial thrombosis , investigating the intricate relationships among platelet , VEGF , and JMJD1C might identify novel drug targets and biological pathways implicated in atherosclerosis and arterial thrombosis ., In a GWAS of serum androgen levels in European men a variant ( rs10822184 ) in JMJD1C reached genome-wide significance ( P = 1 . 12x10-8 ) with the C allele being associated with lower levels 62 ., This variant was also associated with higher circulating VEGF levels ( P = 4 . 06x10-11 ) ., Further , in a GWAS of sex hormone-binding globulin , the T allele of a variant in JMJD1C ( rs7910927 ) was associated with a decrement of sex hormone-binding globulin concentrations ( P = 6 . 1x10-35 ) 63 ., This T allele was also associated with a decrement of VEGF levels ( P = 1 . 31x10-12 ) ., Sex hormones influence VEGF levels 64 thus suggesting a hormone-dependent VEGF production mediated by JMJD1C ., The leading variant in chromosome 18 was located in an intergenic region downstream of the ZADH2 gene and upstream of the Teashirt Zinc Finger Homeobox 1 ( TSHZ1 ) gene and a variant in strong LD with the lead SNP regulates expression of the latter gene ., Both genes have been reported as candidate genes for congenital vertical talus 65 ., TSHZ1 has been associated with increased expression in Juvenile Angiofibroma ( JA ) 66 ., Because VEGF is secreted by JA , and VEGF contributes to vascularization in JA 67 , the investigation of relationships among TSHZ1 , JA , and VEGF might lead to a new therapy for JA ., The top variant in chromosome 16 was located in an intron of the ZFPM1 gene ., The ZFPM1 gene is also known as Friend of GATA1 ( FOG1 ) gene and is related to ZFPM2 , which was identified in our previous meta-analysis 68 ., Both proteins are transcription factors that play a role in the development of the heart and coronary vessels ., Further , a mutation in the N-finger of the GATA1 gene , abrogating the interaction between GATA1 and FOG1 , showed associations with X-linked macro-thrombocytopenia , non-X-linked thrombocytopenia and dyserythropoiesis 69 ., It is possible that the observed association between ZFPM1 and serum VEGF levels was partly driven by variations in platelet counts ., Biological pathway exploration using IPA showed that the Ubiquitin C ( UBC ) gene directly interacted with 10 of the focus genes ., The encoded protein is a polyubiquitin precursor 70 ., This gene has been associated with progressive accumulation of ubiquitinated protein inclusions in neurodegenerative disorders that involve dysfunction of the ubiquitin-dependent proteolytic pathway 71 and with verbal memory performance 72 ., The UBC gene might play an important role in the association between variants and circulating VEGF serum as either mediator or confounder ., However , a direct role for the UBC gene in determining circulating VEGF levels was not identified and none of the variants within 60kb of the UBC gene were associated with circulating VEGF level even at a nominally significant level ., Gene set enrichment analysis revealed the ERK5 pathway as significantly enriched for VEGF associations ., ERK5 pathway is involved in multiple processes , such as cell survival , anti-apoptotic signaling , cell motility , differentiation , and cell proliferation 73 , 74 ., ERK5 is also involved in the angiogenic process , where it acts as regulator of VEGF expression 75 , 76 ., More recently it has been reported that this molecule is expressed on the platelet surface , and acts as platelet activator in ischemic conditions , such as after a myocardial infarct 77 ., Based on eQTL analysis , we observed that 3 of the replicated variants were themselves , or in strong LD with , variants acting as cis and/or trans eQTLs on different genes ., In particular , among those identified as trans-regulated genes , there were some very interesting candidates ., The C-X-C motif chemokine 5 ( CXCL5 ) gene was a trans-regulated gene for 3 variants in two VEGF associated regions ( rs6993770 on 8q23 . 1 and 2 proxies of rs74506613 on 10q21 . 3 ) ., It encodes a protein that through the binding of the G-protein coupled receptor chemokine ( C-X-C motif ) receptor 2 , recruits neutrophils 78 , 79 , promotes angiogenesis 80 and is thought to play a role in cell proliferation , migration , and invasion in different types of cancer 81–85 ., CXCL5 acts by activating several angiogenic signaling pathways , some of which , including JAK/STAT 86 and Src family kinases 87 pathways , are also activated by VEGF ., Given the involvement of the two genes in the same pathways , it is conceivable that they could be co-regulated ., The GUCY1A3 gene encodes the alpha-3 subunit of the Soluble Guanylate Cyclase ( sGC ) , an heterodimeric enzyme that , acting as main receptor of the nitric oxide ( NO ) , catalyzes the conversion of guanosine-5-triphosphate ( GTP ) in 3 , 5-guanosine monophosphate ( cGMP ) and pyrophosphate ., This NO-sGC-cGMP pathway controls vascular smooth-muscle relaxation , vascular tone , and vascular remodeling , and is activated by VEGF signaling ., Inhibition of sGC reduces VEGF-induced angiogenesis 88 , 89 ., Moreover , activation of sGC inhibits platelet activation 90 ., The protein encoded by the MYL9 gene is a myosin light chain that regulates muscle contraction by modulating the ATPase activity of myosin heads ., In platelets , MYL9 is associated with MYH9 , the major nonmuscle myosin expressed in megakaryocytes and platelets ., Defects in the MYH9 gene are responsible of different autosomal dominant disorders characterized by thrombocytopenia and platelet macrocytosis 91 , 92 ., Moreover , it has been demonstrated that MYL9 is involved in pro-platelet formation 93 ., In megakaryocytic cells , MYL9 expression is regulated by RUNX1 , a major hematopoietic transcription factor whose haplo-deficiency is associated with familial thrombocytopenia , platelet dysfunction , and predisposition to leukemia 94 ., The ITGA2B gene encodes the integrin alpha chain 2b , a subunit of the glycoprotein IIb/IIIa , and an integrin complex expressed on the platelet surface ., On the activated platelets , it acts as receptor for fibrinogen; this binding induces platelet aggregation , an essential event in thrombus formation , and permits clot retraction ., Defects in the ITGA2B gene cause Glanzmann thrombasthenia , an autosomal recessive bleeding disorder characterized by failure of platelet aggregation and by absent or diminished clot retraction 95 ., Moreover , a GWAS on platelet count revealed a SNP in the ITGA2B gene region associated with platelets count ( rs708382 , P = 1 . 51x10-8 ) 49 As for the ZFPM1 and JMJD1C genes , the observed connection between VEGF levels and GUCY1A3 , MYL9 and ITGA2B genes could be due , therefore , to a regulation of the number and/or the functionality of the circulating platelets ., Overall our data suggest that studies clarifying whether the relationship between these genes and VEGF levels is mediated by platelets may be helpful to better understand the role of these genes in VEGF regulation ., In conclusion , the identification of novel genes and pathways associated with circulating VEGF levels could lead to new preventive and therapeutic strategies for a wide variety of diseases in which a pathophysiological role for VEGF has been implicated ., The major strength of this work is that it is the largest GWAS of circulating VEGF to date ., A limitation is that , due to the heterogeneity in VEGF levels among the cohorts , a sample size-weighted Z-score method was used to perform the GWAS meta-analysis , which has lower power to detect associations compared to inverse-variance weighted meta-analysis , hence we may have failed to detect some real associations ., Further , our analysis focused mostly on common and less frequent variants ., Therefore , we could not comprehensively assess the effect of rare variants on VEGF levels ., Identifying rare variants in future studies , could contribute to further increasing the proportion of variance in circulating VEGF explained ., Also , our study was confined to individuals of European ancestry ., The results need to be replicated in other racial and ethnic groups ., Finally , a functional validation of the identified associations is needed ., Six discovery data sets including 13 , 312 samples were analyzed in the Stage 1 ., The participating discovery studies were the Age Gene/Environment Susceptibility Reykjavik Study ( AGES , n = 1 , 548 ) , the Cilento study ( Cilento , n = 1 , 115 ) , the Framingham Heart Study ( FHS , n = 7 , 048 ) , the Ogliastra Genetic Park ( OGP , n = 897 ) , the Prospective Investigation of the Vasculature in Uppsala Seniors Study ( PIVUS , n = 945 ) , and the Val Borbera study ( VB , n = 1 , 759 ) ., Two additional studies , the Gioi population ( Gioi , n = 470 ) and the Sorbs population ( Sorbs , n = 659 ) provided data for an in-silico replication ( Stage 2 ) ., Further a de-novo replication ( Stage 3 ) was undertaken in the STANISLAS Family Study ( SFS , n = 676 ) and in a sample of hypertensive adults ( HT , n = 995 ) from the Biological Resources Center ( BRC ) IGE-PCV “Interaction Gène-Environment en Physiopathologie Cardio-Vasculaire ., The participating cohorts are described further in Section 1 in S1 Text ., The local institutional ethics boards for each study approved the study design ., Each subject signed an informed consent before participating to the study ., Further details can be found in S5 Table ., In the discovery and in-silico replication cohorts , genotyping was performed using various arrays , and imputation was carried out using the 1000 genome v3 as reference panel in all studies ., Details of pre-imputation quality control parameters , genotyping platforms and imputation parameters for each study are provided in S1 Table ., In all cohorts blood samples were collected after an overnight fast , and serum/plasma samples were prepared and stored as described in Section 2 in S1 Text ., Serum VEGF levels ( plasma VEGF were measured in SFS and HT ) were measured using commercial ELISA assays as detailed in Section 2 in S1 Text ., The de-novo genotyping at SFS and HT was undertaken on a competitive allele specific PCR ( KASP ) chemistry array and variants were called using a FRET-based genotyping system ., In each individual study , a natural log-transformation of VEGF levels was applied ., To do that , in a few studies ( AGES , OGP , VB , and Sorbs ) where some individuals had VEGF levels below the detection threshold of the assay , half the minimum value of VEGF found in that cohort was arbitrarily assigned to each such participant 96 ., The transformed trait , adjusted for age , sex and additional study-specific covariates ( e . g . principal components associated with VEGF levels , study center for multi-site studies ) , was related to the variant dosages using a linear regression ., Studies with familial correlation used linear mixed effect models to account for familial relatedness ., Detailed information about the software used in each cohort is reported in the S1 Table ., An additive genetic model with 1 degree of freedom was applied ., Study specific results of genome-wide per-variant associations underwent additional quality control prior to meta-analysis ., Checking of file formatting , data plausibility , and distributions of test statistics and quality measurements was facilitated by the gwasqc function of the GWAtoolbox package v1 . 0 . 0 in R 97 ., Prior to the meta-analysis , variants with low minor allele frequency ( <1% ) and poor imputation quality ( r2< 0 . 4 ) were removed ., Meta-analysis was performed in METAL using an effective sample size weighted Z-score method 98 ., This method was chosen over an inverse-variance meta-analysis because of different covariate-adjusted mean values and standard deviations in VEGF levels among studies ., The results of meta-analysis were adjusted for genomic control inflation factor ., To define the effective sample size , the product of the sample size and the imputation quality for each variant was calculated in each cohort 99 ., The sum of the product of each cohort divided by overall sample size represents the proportion of the effective sample size for each variant Eq ( 1 ) ., ∑i=1CNi×ri2/13 , 312=Effective\xa0sample\xa0size, ( 1 ), where C is the total number of participating cohorts , i indicates the specific cohort , N is the sample size used for the variant association test , and r2 is imputation quality of the variant ., After completing initial quality control checks , 6 , 705 , 861 variants , each of which was informative at an effective sample size of >70% , were included in the meta-analysis ( Stage 1 ) ., The genomic control inflation factor of the metal analysis was 1 . 003 ., All variants having a p-value less than 5x10-8 were considered to be genome-wide significant ., To identify all independent associations within the loci reaching genome-wide significance , conditional analyses were performed in a forward stepwise fashion , examining the most significant association and including in successive association models the next most significantly associated variant ( P<5x10-8 ) in a specific region at each step ( referred to as the top variant in Eq ( 2 ) ) ., We repeated this process until no more genome-wide significant associations were found ., The conditional analysis model follows the formula ( 2 ) ., ln ( VEGF ) =β0+β1variant+∑i=1nβiCovariatesi+∑j=1kβjTop\u2009variantj, ( 2 ), where n is the number of covariates used in the primary GWAS , k is the number of steps ., The | Introduction, Results, Discussion, Materials and Methods | Vascular endothelial growth factor ( VEGF ) is an angiogenic and neurotrophic factor , secreted by endothelial cells , known to impact various physiological and disease processes from cancer to cardiovascular disease and to be pharmacologically modifiable ., We sought to identify novel loci associated with circulating VEGF levels through a genome-wide association meta-analysis combining data from European-ancestry individuals and using a dense variant map from 1000 genomes imputation panel ., Six discovery cohorts including 13 , 312 samples were analyzed , followed by in-silico and de-novo replication studies including an additional 2 , 800 individuals ., A total of 10 genome-wide significant variants were identified at 7 loci ., Four were novel loci ( 5q14 . 3 , 10q21 . 3 , 16q24 . 2 and 18q22 . 3 ) and the leading variants at these loci were rs114694170 ( MEF2C , P = 6 . 79x10-13 ) , rs74506613 ( JMJD1C , P = 1 . 17x10-19 ) , rs4782371 ( ZFPM1 , P = 1 . 59x10-9 ) and rs2639990 ( ZADH2 , P = 1 . 72x10-8 ) , respectively ., We also identified two new independent variants ( rs34528081 , VEGFA , P = 1 . 52x10-18; rs7043199 , VLDLR-AS1 , P = 5 . 12x10-14 ) at the 3 previously identified loci and strengthened the evidence for the four previously identified SNPs ( rs6921438 , LOC100132354 , P = 7 . 39x10-1467; rs1740073 , C6orf223 , P = 2 . 34x10-17; rs6993770 , ZFPM2 , P = 2 . 44x10-60; rs2375981 , KCNV2 , P = 1 . 48x10-100 ) ., These variants collectively explained up to 52% of the VEGF phenotypic variance ., We explored biological links between genes in the associated loci using Ingenuity Pathway Analysis that emphasized their roles in embryonic development and function ., Gene set enrichment analysis identified the ERK5 pathway as enriched in genes containing VEGF associated variants ., eQTL analysis showed , in three of the identified regions , variants acting as both cis and trans eQTLs for multiple genes ., Most of these genes , as well as some of those in the associated loci , were involved in platelet biogenesis and functionality , suggesting the importance of this process in regulation of VEGF levels ., This work also provided new insights into the involvement of genes implicated in various angiogenesis related pathologies in determining circulating VEGF levels ., The understanding of the molecular mechanisms by which the identified genes affect circulating VEGF levels could be important in the development of novel VEGF-related therapies for such diseases . | Vascular Endothelial Growth Factor ( VEGF ) is a protein with a fundamental role in development of vascular system ., The protein , produced by many types of cells , is released in the blood ., High levels of VEGF have been observed in different pathological conditions especially in cancer , cardiovascular , and inflammatory diseases ., Therefore , identifying the genetic factors influencing VEGF levels is important for predicting and treating such pathologies ., The number of genetic variants associated with VEGF levels has been limited ., To identify new loci , we have performed a Genome Wide Association Study meta-analysis on a sample of more than 16 , 000 individuals from 10 cohorts , using a high-density genetic map ., This analysis revealed 10 variants associated with VEGF circulating levels , 6 of these being novel associations ., The 10 variants cumulatively explain more than 50% of the variability of VEGF serum levels ., Our analyses have identified genes known to be involved in angiogenesis related diseases and genes implicated in platelet metabolism , suggesting the importance of links between this process and VEGF regulation ., Overall , these data have improved our understanding of the genetic variation underlying circulating VEGF levels ., This in turn could guide our response to the challenge posed by various VEGF-related pathologies . | blood cells, genome-wide association studies, medicine and health sciences, body fluids, vegf signaling, genomic databases, mathematics, statistics (mathematics), genome analysis, platelets, research and analysis methods, genome complexity, chromosome biology, animal cells, mathematical and statistical techniques, statistical methods, biological databases, genetic loci, hematology, signal transduction, blood, cell biology, anatomy, meta-analysis, physiology, genetics, database and informatics methods, biology and life sciences, cellular types, physical sciences, genomics, cell signaling, computational biology, introns, chromosomes, human genetics | null |
journal.pgen.1004947 | 2,015 | A Genomic Duplication is Associated with Ectopic Eomesodermin Expression in the Embryonic Chicken Comb and Two Duplex-comb Phenotypes | In the domestic chicken ( Gallus domesticus ) the comb serves as a sexual ornament and the size of the comb is associated with mate choice in both sexes as well as fecundity in females 1 , 2 ., The vast majority of chicken populations used for commercial meat and egg production around the world are fixed for the wild-type single comb phenotype ., However , there are three major comb loci found in non-commercial chicken breeds which are primarily used for exhibition purposes; Rose-comb , Pea-comb and Duplex-comb ., The causal mutations for Rose-comb and Pea-comb have recently been identified , both corresponding to structural genomic variants that drive ectopic expression of transcription factors in the developing comb region of the chicken embryo 3–5 ., The Duplex-comb locus harbors three alleles , Buttercup ( D*C ) , V-shaped ( D*V ) and wild-type or normal ( D*N ) ., Chickens that are wild-type at the Rose-comb , Pea-comb and Duplex-comb loci have the single comb phenotype ., D*C corresponds to a cup shaped comb arising from a single central blade ringed with individual points along the perimeter of the cup ( Fig . 1A ) ., This phenotype is somewhat rare , being found in the Sicilian Buttercup , Caumont and Augsburger chicken breeds ., Chickens that appear to have the Buttercup comb phenotype were described by the naturalist Ulisse Aldrovandi in 1600 6 , 7 and are thought to originate from North Africa , possibly being the progenitors of the Sicilian Buttercup breed ., D*V corresponds to a two-pronged horn or V-shaped comb that is restricted to the posterior portion of the comb developing region ( Fig . 1A ) ., The V-shaped comb is found in many breeds from around the world such as the Crevecoeur , Houdan , La Fleche , Merlerault , Padova , Polish , Spitzhauben and Sultan ., Both comb types can vary slightly in shape and size between breeds and strains due to differences in genetic background , with D*C occasionally resembling two distinct single ( wild-type ) combs split down the midline ., Previous experiments have demonstrated that the V-shaped and Buttercup phenotypes are inherited as determined by alleles at a single locus 8 ., In these experiments D*V was completely dominant over D*C ., Both mutant alleles were incompletely dominant over D*N ., D*V was completely penetrant in both sexes while D*C was incompletely penetrant ( 68% ) in females 8 ., From these experiments it was suggested that D*C represents a comb doubling effect while D*V causes doubling and reduction of comb size ., Here we show that both Duplex-comb phenotypes are associated with a 20 Kb tandem duplication and ectopic expression of EOMES , a transcription factor with a known role in mesoderm specification in the developing embryo 9 ., We previously mapped D*V to the 37 . 3–39 . 8 Mb region of Gallus gallus autosome 2 ( GGA2 ) in a backcross population 10 ., Subsequent fine mapping with additional markers identified a region of maximum association with D*V as between markers rs15086167 and rs14167302 , corresponding to GGA2:38 , 554 , 221–39 , 229 , 442 bp ., Genotyping of a diverse breed panel ( D*V , n = 7; D*N , n = 64 ) on the 60K Chicken iSelect chip 11 ( Illumina ) revealed an identical by descent ( IBD ) haplotype located between markers rs15086146 and rs15086500 , corresponding to GGA2:38 , 528 , 939–38 , 910 , 305 bp and consistent with other reports 12 ., A single SNP within the IBD haplotype was observed to be heterozygous in all D*V individuals , ( GGaluGA142157 at 38 , 806 , 246 bp ) , suggestive of a duplication fixed for alternative SNP variant alleles ( Figs . 1B and S1 ) ., Copy number of the IBD region was explored using SYBR Green qPCR analysis with genomic DNA as template ., Iterative rounds of qPCR analysis of copy number analysis ultimately defined the approximate boundaries of a putative 2-fold duplicated region in D*V individuals ., Genomic copy number analysis using a TaqMan assay was in agreement with the SYBR Green assays and confirmed the presence of a duplication in both D*V and D*C individuals as compared to D*N ( Fig . 1C ) ., Successful amplification and sequencing across the duplication junction point revealed this to be a tandem duplication of a ∼20 Kb segment spanning from 38 , 798 , 537 to 38 , 818 , 314 bp ., Analysis of the duplication junction sequences revealed only a single base pair of overlap and no other sequence micro-homologies ( Fig . 1D ) ., The 38 , 818 , 314 bp duplication junction point is within a LTR element as annotated by the UCSC Genome Browser ., A diagnostic PCR test was designed to amplify across the duplication junction point and used to screen a diverse breed panel representing the three known alleles at the Duplex-comb locus ., All D*N homozygotes ( n = 44 ) were wild-type for the duplication junction point while all V-shaped ( n = 48 ) and Buttercup ( n = 35 ) individuals had at least one copy of the duplication ( Table 1 ) ., This PCR test cannot distinguish between heterozygotes and homozygotes for the duplication as no wild-type sequences are disrupted in this tandem duplication ., Three chickens , each representing one of the three alleles at the Duplex-comb locus , were selected for whole genome sequencing to search for mutations other than the 20 Kb duplication that could be responsible for the difference between the V-shaped and Buttercup comb phenotypes ., The average depth of sequence coverage for each bird was in the range 24x to 34x , which gives a high power for SNP detection at most sites ., The D*V ( White Crested Black Polish , USA ) and D*C ( Sicilian Buttercup , Italy ) individuals were from breeds with standardized V-shaped and Buttercup comb phenotypes and were tested with the TaqMan copy number assay to verify homozygosity before whole genome sequencing ., The D*N ( Single Comb Dark Brown Leghorn , USA ) individual was selected due to sharing an identical haplotype as the D*V individual based on the 60K SNP chip genotype data except for the D*V associated heterozygous SNP and 20 Kb duplication ., The largest region for which D*C and D*V individuals were IBD was 89 Kb in size ( 38 , 738 , 016–38 , 827 , 468 bp ) which includes the entire 20 Kb duplicated region ( Fig . 2A , IBD_reseq track ) ., We identified 6 and 17 paired-end reads that spanned the duplication junctions in D*V and D*C individuals , again confirming the exact duplication breakpoints ., We then used the sequencing data to explore if there were any other sequence variants that showed a perfect concordance with D*V and D*C like the 20 Kb duplication ., Stringent SNP calling revealed only one high-quality SNP , at position 38 , 797 , 948 bp , within the IBD region that showed this pattern and that were not found in other chicken populations with the single comb phenotype 13 ., This SNP did not occur at an evolutionary conserved site ., To identify one or more mutations that distinguish the two mutant alleles we first searched for SNPs within the duplicated region ., There were 182 SNPs detected between all three sequenced individuals , with 181 SNPs having identical genotypes in D*C and D*V individuals ( Fig . 2B ) ., The one remaining SNP at 38 , 808 , 838 bp was heterozygous G/A in the D*V individual and homozygous reference ( G ) in the D*C and D*N individuals chosen for sequencing ., Further screening showed that this SNP was homozygous reference ( G ) in 23 of 32 additional D*V individuals from four different breeds and was never found to be homozygous for the mutant allele ., This indicates that the 38 , 808 , 838 bp SNP is not causally associated with D*V , but instead has evolved in some D*V populations in one of the two copies of the 20 Kb duplicated region ., There were no other high-quality SNPs within the 89 Kb IBD region identified from the sequencing data for which the V-shaped and Buttercup individuals were homozygous for alternative alleles ., There were 66 SNPs within the duplicated region that were called as heterozygous in both the D*C and D*V individuals , indicative of the two copies of the duplicated region composing a single haplotype , each copy carrying different sequence variants at 66 SNP positions ., Thus , the nucleotide divergence between the two copies is about 0 . 3% , i . e . three times higher than the average nucleotide diversity in the human genome and close to the average nucleotide diversity of 0 . 5% in the chicken genome 14 ., This implies a scenario where two different haplotypes contributed to the tandem duplication and the majority of the sequence differences are expected to represent sequence differences between the two ancestral haplotypes ., This interpretation is consistent with the observation that most SNPs showing sequence differences between the two copies , such as GGaluGA142157 at 38 , 806 , 246 bp , also segregated among wild-type chromosomes ( Fig . 1B ) ., There are several regions within the 20 Kb duplication that exhibit elevated conservation scores according to the UCSC genome browser and Genomic Evolutionary Rate Profiling ( GERP ) 15 ( Fig . 2B ) , representing putative regulatory elements ., The chicken comb originates from a region on the upper beak , posterior to the fronto-nasal facial process and is first visible as a narrow midline ridge , at embryonic day 6–7 ( E6–7 ) ., The wild-type single comb has one row of papillae that are formed from local mesenchyme condensations along the initial comb-ridge and they present the beginnings of the comb serrations ( S2A and B Fig . ) ., The V-shaped ( D*V ) and Buttercup ( D*C ) combs are initially formed by a split of this single comb anlage ., The split occurs at a variable posterior position in the V-shaped comb ridge with a reduction of the anterior portion of the comb while in the Buttercup the whole ridge is split ., The posterior part of the split ridge in Buttercup is often fused as seen in S2E Fig ., ., The appearance of the developing nostrils is also affected in the duplex phenotype ., The expression pattern of candidate genes located in the proximity of the 20 Kb duplication was investigated using quantitative reverse transcription-PCR ( qRT-PCR ) in samples of comb tissue from developing chicken embryos ., The duplication is located within an intron of CMC1 encoding COX assembly mitochondrial protein homolog ( S . cerevisiae ) as assessed by aligning the predicted CMC1 sequence ( XM_418758 . 4 ) to the chicken genome via BLAT in the UCSC genome browser ., 5-azacytidine induced 2 ( AZI2 ) is the nearest gene on the 3’ side of the duplication and eomesodermin ( EOMES ) is the nearest gene on the 5’ side ., CMC1 and AZI2 were both expressed in comb tissue but did not show any significant difference in expression between genotypes ., In contrast , EOMES showed a dramatic expression difference between genotypes and was more highly expressed at E8 , E9 and E12 in D*V embryonic comb as compared to D*N , while it was not expressed in any genotype at E18 ( Fig . 3 ) ., The spatial distribution of EOMES expression in the developing comb region of the chicken embryo was investigated using immunohistochemistry ., In D*N embryos no EOMES expression was detected from E7-E18 in the ectoderm or mesenchyme of the comb region ., Both D*V and D*C embryos showed clear expression of EOMES in the ectoderm of the comb region already at E7 and continuing through E12/E15 ( Fig . 4E-G , I-K ) ., The expression of EOMES was limited to the ectoderm of the developing comb region at all stages analyzed in D*V and D*C embryos and could not be detected by E18 ( Fig . 4H ) ., The Duplex-comb locus was originally described as having two mutant alleles 8 and being linked to the polydactyly locus 16 , which was subsequently mapped to GGA2 10 , 17–19 ., Through successive rounds of linkage mapping and IBD haplotype analysis using different chicken breeds we have identified an 89 Kb region of GGA2 as encompassing the Duplex-comb locus ., This region contains a 20 Kb tandem duplication that is only found in chicken breeds that have a Duplex-comb phenotype when screened on a diverse breed panel ., Sequence analysis of the duplicated region identified only a single base pair difference within the 20 Kb duplication between the two mutant alleles , however this variant was subsequently excluded as the causal difference between D*C and D*V alleles after finding many D*V individuals that were homozygous reference ., The 20 Kb duplication contains several putative conserved regulatory elements ( Fig . 2B ) that is likely driving the ectopic expression of the downstream transcription factor EOMES in the developing chicken comb region in Duplex comb individuals ., The phenotypic diversity of the chicken comb is primarily governed by a small number of loci with large effects that determine the overall morphology of the comb during embryonic development; the Rose-comb , Pea-comb and Duplex-comb loci ., The Rose-comb and Pea-comb loci are notable in being the first example of classical genetic epistasis , giving rise to the Walnut comb phenotype when mutant alleles are present at both loci 20 ., The Pea-comb mutant allele has recently been described as corresponding to a copy number expansion in an intron of SOX5 , resulting in ectopic expression of this transcription factor in the mesenchyme of the developing comb region of the chicken embryo 5 ., The Rose-comb mutant allele was also recently characterized , corresponding to a 7 Mb inversion that leads to ectopic expression of the transcription factor MNR2 in the mesenchyme of the developing comb region of the chicken embryo 4 ., This overlapping spatial and temporal domain of ectopic expression of SOX5 and MNR2 is a clear demonstration of how the epistasis between Rose-comb and Pea-comb loci is derived at the cellular level through the combined action of two transcription factors 4 ., Here we show that the last major comb locus in the chicken to be characterized at the molecular level also corresponds to a structural variant in the chicken genome that results in ectopic expression of a transcription factor ., However , while Rose-comb and Pea-comb phenotypes are driven by ectopic expression of SOX5 and MNR2 in the mesenchyme of the developing comb region , we show that the Duplex-comb phenotype is mediated by ectopic expression of EOMES confined to the ectoderm ., Eomesodermin ( EOMES ) is a T-box transcription factor that is involved in mesoderm specification during gastrulation as shown in zebrafish , chicken and mouse ., EOMES is expressed in the extraembryonic tissues of the chicken and the mouse as well as the primitive streak , forebrain region and genital ridge 9 , 21 ., However , expression of EOMES in primordial germ cells is only found in the chicken 21 ., Investigation of four upstream and one downstream putative cis-regulatory element ( CRE ) of mouse EOMES indicated that different regulatory mechanisms between mouse and chicken were likely responsible for EOMES expression in extraembryonic tissues while a single CRE located ∼150 Kb upstream drove expression in the brain in both chicken and mouse 21 ., The 20 Kb duplication overlaps several regions of elevated sequence conservation and lies approximately 200 Kb upstream of EOMES in the chicken , suggesting that the duplicated region contains CREs and that an altered dosage of these elements causes perturbed regulation of EOMES expression ., Using qRT-PCR we show that EOMES is upregulated in the comb region of D*V embryos as early as embryonic day 8 as compared to D*N embryos ., There was no difference in expression of CMC1 and AZI2 ( the two other genes located nearest the 20 Kb duplication ) between D*V and D*N embryos , suggesting that this mutation involves a CRE specific to EOMES , at least in comb tissue ., Using IHC we show that EOMES is ectopically expressed in the ectoderm of the developing comb region of D*C and D*V embryos ., There was no detectable EOMES expression in this region of the D*N embryo at these stages , suggesting that EOMES does not normally play a role in comb development ., The major comb phenotypes are all caused by mutations that direct expression of transcription factors to the ectoderm or mesenchyme of the comb ridge ., The development of the comb as part the chicken naso-facial processes is directly induced and regulated by reciprocal ecto-mesenchymal interactions 22 ., Interactions of ectopically expressed transcription factors either in the ectoderm or mesenchyme then cause the similar but not identical comb phenotypes ., The exact regional and temporal expression of the inductive signals or their receptors is instrumental for the morphogenesis 3 ., The D*C phenotype is characterized as a splitting of the comb mass while the D*V phenotype involves both splitting and reduction of comb mass as well as enlargement of the nostrils 6 ., We propose that the initial duplication event is the primary driver of ectopic EOMES expression in the ectoderm of the comb region and causes the majority of the comb duplication phenotype ., A subsequent and unknown mutation is suspected of further modifying the spatial or temporal expression of EOMES to result in two different Duplex-comb phenotypes , but does so in a manner that escapes the resolution of our IHC experiments ., The mutation that distinguishes the two mutant alleles should be found in a D*C or D*V IBD region ., Our initial genotyping data identified a 381 Kb IBD haplotype in D*V individuals ( Fig . 1B ) , however it is uncertain which mutant allele evolved first and we lack similar data for D*C individuals ., We restricted our search for a causal mutation to the 89 Kb IBD region since this study shows that this region contains regulatory elements that affect EOMES expression during comb development , but we found no high-quality SNP where the sequenced D*V/D*V and D*C/D*C birds were homozygous for different alleles ., A causal mutation could have been overlooked due to a gap or lack of adequate coverage in the sequence data although we had on average high sequence coverage ( in the range 24x to 34x ) ; the current assembly of the 89 Kb region contains one gap annotated as comprising about 750 nucleotides ., Although prior experiments 8 found that D*C and D*V were alleles of the same locus , it remains unknown how close the mutation differentiating these two alleles lies to the 20 Kb duplication ., At present we cannot exclude the possibility that the causal difference between D*V and D*C could be affecting a nearby gene other than EOMES ., A common feature of duplicated sequences is that they show copy number variation because nearly identical tandem copies are prone to unequal crossing-over or slippage during replication 23 ., We did not detect any such copy number variation and all D*C and D*V chromosomes analyzed in this study appeared to contain only two copies of the duplicated sequence ., Furthermore , the two copies of the duplicated sequence showed a 0 . 3% sequence divergence and were identical between V-shaped and Buttercup chromosomes ( except at the SNP distinguishing D*V and D*C sequenced individuals ) ., This implies that the sequence divergence between the two copies is sufficient to suppress unequal crossing-over that may otherwise lead to copy number variation and gene conversion , resulting in homogenization of the tandem copies ., The investigation of genetic mechanisms underlying phenotypic diversity in domestic animals has revealed that structural variation plays a significant role , typically affecting spatio-temporal gene expression patterns through rearrangement of regulatory elements 24 ., Examples of such traits are Pea-comb 5 , Rose-comb 4 , Fibromelanosis 25 and Dark Brown plumage 26 in the chicken; Dominant White in the pig 27; Greying with age in the horse 28 and Color Sidedness 29 and Polled in cattle 30–32 ., Here we add the Duplex-comb locus to this list , highlighting how large-scale genomic mutations appear to often result in very noticeable phenotypic effects that are then easily selected by humans during animal domestication and breeding ., The Duplex-comb trait also illustrates another striking feature of genetic diversity in domestic animals , the evolution of alleles 24 ., The evolutionary history of domestic animals is sufficiently long to allow the accumulation of two or more causative mutations on the same haplotype ., This is the case for instance with Dominant white color in pigs 27 , Black spotting in pigs 33 and Rose-comb in chickens 4 ., The Duplex-comb locus can now be added to this growing list of examples since the Buttercup and V-shaped alleles share an 89 Kb IBD region including the 20 Kb duplication but differ at a yet unknown position ., This illustrates why domestic animals present a valuable model to study the genetic mechanisms and processes that likely underlie phenotypic traits in humans and other species ., A custom GoldenGate BeadXpress panel ( Illumina ) containing 28 SNPs on GGA2 was used to fine map the D*V mutation in the same backcross population we previously reported 10 ., The 60K Chicken iSelect chip 11 ( Illumina ) was used to genotype a diverse panel of chicken breeds for IBD haplotype analysis ., All genome coordinates are relative to the May 2006 WUGSC 2 . 1/galGal3 assembly ., SYBR Green assays for genomic copy number were performed using SYBR Green PCR Master Mix ( ABI ) with 800 nM of each primer and 10 ng of DNA in a total volume of 10 μl ., Reactions were performed in quadruplicate and data was analyzed using the 2-ΔΔCt method 34 , correcting for amplification efficiency as measured by a standard dilution series ., TaqMan assays and data analysis for genomic copy number were performed as previously described 25 ., A primer/probe set in an exon of SOX5 was used as a calibrator for both SYBR Green and TaqMan assays ., All primer sequences can be found in S1 Table ., A three primer PCR diagnostic test was developed that amplified over the duplication junction point as well as amplifying a product over one of the duplicated region wild-type sequences ., The KAPA2G Robust HotStart PCR system ( Kapa Biosystems ) was used with 1X KAPA2G GC Buffer , 0 . 2 mM dNTPs , 1 . 5 mM MgCl2 , 200 nM of primers D_5_F and D_5_R , 150 nM of primer D_3_F , 0 . 4 U of KAPA2G Robust HotStart DNA Polymerase , and 50 ng of DNA in a total volume of 10 μl ., A touchdown thermal cycling protocol was used for the diagnostic test of 95°C for 5 min , 16 cycles of 95°C , 68°C ( -1 . 0°C/cycle ) , and 72°C for 30 s each , followed by 24 cycles of 95°C , 52°C , and 72°C for 30 s each ., This test is not capable of differentiating homozygous mutant individuals from heterozygotes ., DNA was prepared from blood samples of single individuals representing the D*V ( White Crested Black Polish , USA ) , D*C ( Sicilian Buttercup , France ) and D*N ( Single Comb Dark Brown Leghorn , USA ) alleles ., The DNA was used to construct paired-end libraries with average insert size of approximately 220 bp and these libraries were subjected to whole genome sequencing using a HiSeq sequencing instrument ( Illumina ) ., Sequencing reads ( 2 x 100bp ) were aligned to the chicken reference genome ( galgal3 ) using the Burrows Wheeler Aligner ( BWA ) 35 , revealing average depths of coverage of 24 , 25 and 34 for D*V , D*N and D*C , respectively ., The aligned reads were subjected to duplicate flagging using Picard Tools ( http://picard . sourceforge . net ) and to SNP calling using the Genome Analysis Toolkit ( GATK ) Unified Genotyper version 2 . 4 . 9 36 ., Identified raw SNPs were filtered based on GATK best practice variant detection and genotypes with a PHRED genotype quality ≥ 20 were used in subsequent steps ., SNP- and genotype calls were compared to SNPs detected in DNA pools from wild- and domestic chickens in a previously published study 13 ., Total RNA was extracted from comb tissue from E8 , E9 , E12 and E18 D*V ( Merlerault , France ) and D*N ( Geline de Touraine , France ) chicken embryos using TRIzol ( Invitrogen ) ., RNA was treated with DNase ( 1 μg/μl ) and cDNA was made from 1 μg of RNA using High Capacity RNA-to-cDNA Kit ( ABI ) ., The qRT-PCR analysis was performed using CFX96 SyBr Green Supermix ( Bio-Rad ) with primers designed by using Primer Express v2 . 0 ( ABI ) , checked for PCR efficiency , linear dynamic range and specificity ., The mRNA levels were normalized to β-actin mRNA levels ., The use of β-actin for normalization purposes was validated by testing for the most stable mRNA expression of TATA box binding protein , β-actin , ß-2-microglobulin and glyceraldehyde 3-phosphate dehydrogenase over the developmental stages using geNorm 37 ., Expression levels were calculated from cycle threshold ( Ct ) and the 2-ΔΔCt method 34 ., The normalized amplification levels of Duplex-comb and single-comb samples relative to the ß-actin amplification levels are shown , and differences were tested by using one-way analysis of variance ( ANOVA ) followed by Tukey’s range test as indicated in figure legend ., Chicken embryo heads from D*V ( Merlerault , France ) , D*C ( Caumont , France ) , and D*N ( Geline de Touraine , France ) breeds were fixed in 4% paraformaldehyde , pH 7 . 4 in PBS for one hour at 4°C , transferred to 30% sucrose in PBS overnight at 4°C , frozen in OCT freezing medium and sectioned 10 μm with a cryostat ., The sections were washed in PBS and used for immunohistochemistry ., Sections were blocked ( PBS with 1% fetal calf serum , 0 . 1% Triton-X , 0 . 02% Thimerosal ) before addition of primary antibodies in blocking solution and incubated overnight at 4°C ., The slides were washed three times for 5 min in PBS before incubation secondary antibodies in blocking solution in room temperature for two hours ., The slides were washed three times 5 min with PBS before mounting ., Primary antibody: TBR2/EOMES ( Abcam #ab23345 ) , rabbit polyclonal 1:1000 ., Secondary antibody: Alexa Fluor 568 , rabbit IgG ( Invitrogen ) was made in donkey ., Images from immunohistochemistry were captured using a Zeiss Axioplan2 microscope and AxioVision 4 . 8 software ( Carl Zeiss ) . | Introduction, Results, Discussion, Materials and Methods | Duplex-comb ( D ) is one of three major loci affecting comb morphology in the domestic chicken ., Here we show that the two Duplex-comb alleles , V-shaped ( D*V ) and Buttercup ( D*C ) , are both associated with a 20 Kb tandem duplication containing several conserved putative regulatory elements located 200 Kb upstream of the eomesodermin gene ( EOMES ) ., EOMES is a T-box transcription factor that is involved in mesoderm specification during gastrulation ., In D*V and D*C chicken embryos we find that EOMES is ectopically expressed in the ectoderm of the comb-developing region as compared to wild-type embryos ., The confinement of the ectopic expression of EOMES to the ectoderm is in stark contrast to the causal mechanisms underlying the two other major comb loci in the chicken ( Rose-comb and Pea-comb ) in which the transcription factors MNR2 and SOX5 are ectopically expressed strictly in the mesenchyme ., Interestingly , the causal mutations of all three major comb loci in the chicken are now known to be composed of large-scale structural genomic variants that each result in ectopic expression of transcription factors ., The Duplex-comb locus also illustrates the evolution of alleles in domestic animals , which means that alleles evolve by the accumulation of two or more consecutive mutations affecting the phenotype ., We do not yet know whether the V-shaped or Buttercup allele correspond to the second mutation that occurred on the haplotype of the original duplication event . | There are three major variant comb types found in the domestic chicken; Rose-comb , Pea-comb and Duplex-comb ., Within the Duplex-comb there are two distinct types , V-shaped and Buttercup ., Previous experiments have shown that these two Duplex-comb types represent different alleles at a single locus ., We have mapped the location of the Duplex-comb locus and identified a 20 Kb duplication that is present only in chickens that have a Duplex-comb phenotype ., The 20 Kb duplication is located 200 Kb upstream of EOMES , a gene that was found to be abnormally expressed in the comb-developing region of V-shaped and Buttercup comb chicken embryos ., This suggests that the 20 Kb duplication contains regulatory elements affecting EOMES expression ., These findings complete our characterization of the genetic basis of the three major comb loci in the chicken , all of which are caused by large-scale structural genomic variants that drive ectopic expression of transcription factors in the comb region during chicken embryo development . | null | null |
journal.pntd.0002577 | 2,013 | Ecology and Geography of Transmission of Two Bat-Borne Rabies Lineages in Chile | Rabies was known to humans as a disease as of about ∼4000 years ago 1 ., Although important advances have been made in immunization and diagnosis , rabies is still considered a neglected disease 2 ., Rabies is a zoonosis: indeed , although all mammals studied to date are susceptible to infection , major reservoirs that maintain and transmit the virus in the long term are limited to Carnivora and Chiroptera 2 ., Rabies virus ( RABV ) is a neurotropic RNA virus ( family Rhabdoviridae , genus Lyssavirus ) , including at least 14 species 3 ., In the Americas , with generally good control of rabid canines , bats are the main reservoirs of RABV 4 ., Rabies transmission from non-hematophagous bats ( mainly insectivores ) to humans is considered an increasing risk in urban and economically developed areas of Latin America 5 , while dog rabies has decreased dramatically in frequency , now occurring only in specific areas of Latin America 6 , 7 ., Viral “strains” are defined as virus populations maintained by a particular reservoir host in a defined geographic region that can be distinguished from other strains based on molecular and antigenic characteristics 8 ., RABV lineages generally show specificity to particular bat hosts 9–11 ., Antigenic typing depends on use of monoclonal antibodies; their power depends on numbers of monoclonal antibodies that bind consistently to antigenic sites that are conserved in a viral strain 8 , 12 ., Antigenic characterization is used widely in rabies surveillance in Latin America 9 , showing differences among viruses in different host species and geographic locations 13 ., Tadarida brasiliensis , an important reservoir of rabies in urban areas , maintains antigenic variant AgV9 in North America , but AgV4 in South America 14 ., Lasiurus cinereus differs , carrying AgV6 across its entire geographic distribution 15 ., Viral specificity to these two host species has been confirmed with molecular analyses 9 , 10 , 13 ., These bat species presently constitute the principal rabies reservoirs in Chile 16 , 17 , but little is known about roles of different hosts in their ecology and distribution ., T . brasiliensis inhabits sites with other species , roosting in colonies over long periods; owing to anthropogenic perturbation , this species is that which has seen greatest negative population effects in Chile 18 ., In contrast , L . cinereus avoids urban areas , roost solitarily , and shows seasonal migrations 19 ., Both species have broad geographic distributions across the Americas ., Previous such geographic and environmental analyses of rabies lineages have focused on RABV in terrestrial mammal hosts in North America , and documented that rabies in raccoons ( Procyon lotor ) is associated with low wetlands coverage , low elevation , low-intensity residential land use , and absence of major roads , and that rivers act as natural barriers 20 , 21 ., Several studies have explored features of host-virus relationships of bat-borne rabies , based on molecular genetic analyses 22–25 ., However , in these key studies , inferences about geographic pattern were made based on points on an empty map , without reference to environmental drives ., Hence , landscape- and niche-based approaches could offer a valuable complement to conclusions generated in molecular genetic studies , evaluating effects of environment and landscape on rabies host and virus distributions , but such methods must be explored and validated first ., To test these approaches , we address a series of questions regarding rabies transmission ecology in Chile ., ( i ) Do rabies lineages have coarse-grained ecological “signatures” ( i . e . , Grinnellian niches ) that can be characterized robustly ?, ( ii ) Do macro-ecological and macro-geographic linkages exist among viruses and hosts ?, Finally ,, ( iii ) do different bat-borne rabies lineages have distinct ecological signatures ?, Answering these questions will help to illuminate details of virus-host dynamics in bat rabies transmission cycles in South America ., Delimitation of the geographic area of analysis is a crucial issue in generating robust niche models , with significant effects on model results 29 ., The study area must be established a priori based on ( 1 ) the dispersal potential of the species involved , ( 2 ) the sampling available by which to characterize distributions , and ( 3 ) the objectives of the study 29 ., We delimited our study area to the area between −28 . 0° and −43 . 5°s latitude in Chile , corresponding both to the enzootic area in recent decades 16 and to the area sampled by the Chilean Ministerio de Salud ( Ministry of Health; Fig . 1 ) ., Another crucial aspect in niche model development is the set of environmental variables used to characterize the environmental space in which the species is distributed 30 ., We used information on land-surface reflectance from remote sensing , in light of its high information content , fine spatial resolution , and minimal need for interpolation and inference 31 ., Environmental variation can be summarized using multiple seasonal values of the Normalized Difference Vegetation Index ( NDVI ) , which has values correlating strongly with photosynthetic mass and primary productivity 31 , 32 ., Numerous previous studies have shown the importance of such vegetation indices as indicators of ecological and geographic dimensions 31 , including in development of robust ecological niche models 33 , 34 ., We used NDVI images available as monthly maximum raster data layers for 1992 , 1993 , and 1995 , which correspond to the middle years of the study period , at a spatial resolution of 0 . 01°×0 . 01°; to standardize these variables and reduce dimensionality , we generated principal components across all of the monthly data sets using ArcGIS 9 . 3 ( ESRI , Redlands , CA , USA ) ., Principal components analysis used the original NDVI layers to generate 27 new , uncorrelated components: we used the first 10 components in model development ( i . e . , the initial 10 axes that best characterized the major dimensions of the cloud of points ) , as they explained 99 . 99% of overall variance ., To characterize spatial patterns of bat-rabies occurrence across Chile , we only digitized bat surveillance data from the Instituto de Salud Pública de Chile ( ISP ) , for 1985–2011 , corresponding to the major enzootic period for bat rabies in Chile ( Fig . 1 ) ., Host mammal occurrences were obtained from both active and passive surveillance programs , with hosts tested for rabies and identified at ISP ., Coordinates of bat occurrences ( both species , regardless of rabies status ) were derived from geographic centroids of municipalities , as they were submitted by municipal agencies for testing ., Further occurrences were obtained through data mediated by the Global Biodiversity Information Facility ( GBIF; see Acknowledgments for full list of institutions ) , with georeferencing derived from original data records ., Virus occurrences were obtained in the form more precise georeferences derived from postal addresses of sites of origin of rabies-positive bats of both species , although the vast majority ( 78% ) came from Tadarida ., These cases were diagnosed by ISP using direct inmunofluorescence ( IFD ) , to confirm virus presence , and monoclonal antibodies to identify virus variants 35 ., To calibrate niche models , we used a maximum entropy algorithm , considering its predictive power and broad acceptance in the scientific community 36 ., The algorithm uses the information theory concept of maximum entropy to optimize estimates of suitability across complex environmental spaces ., The maximum entropy approach seeks to estimate the probability of suitability through finding the probability distribution closest to uniform , subject to certain restrictions; in our case , the restrictions are environmental conditions associated with known occurrences of the species in question 37 ., In Chile active surveillance is initiated after a positive bat is reported from passive surveillance ., ISP samples originated from passive surveillance 16 , 17 associated with human settlements , without anything close to uniform geographic coverage ., We incorporated sampling bias across the study area in model calibration because spatial and environmental biases in data collection can cause biases in model results 38 ., Maxent can use a sampling bias distribution ( σ in Phillips et al . , 2009 ) to establish areas from which to focus extraction of background data with which to calibrate models 38 ., We thus developed a sampling bias surface for T . brasiliensis based on all of the passive surveillance data , using overall numbers of samples submitted to ISP per municipality ( municipalities with no samples set to no data , and thus excluded from background sampling ) , regardless of rabies-positive status , on the final raster , we added 1 to all pixels to avoid zero values , according to Maxent requirements ., This surface appropriately characterized the sampling that underlies the virus-positive records that drove calibration of the niche models ., We calibrated models with and without this bias file to assess the degree to which sampling effort affects results ., We calibrated models using Maxent version 3 . 3 . 3 . k ., Specific options were a bootstrap subsampling with 1000 replicates , random seed , and the median of replicates as output ., We converted raw Maxent output to binary maps considering an error rate of E\u200a=\u200a10% among occurrence points , and thus used the highest threshold that included 90% of training presence points 26 , a modification of the least training presence threshold idea 39 ., The error rate ( E ) is the proportion of the occurrence data expected to place the species erroneously under inappropriate conditions , as a consequence of incorrect species identifications , errors in georeferencing , and errors in environmental data , among other factors , and is estimated via exploration and error-checking of the occurrence data 40 ., We visualized ecological niche models in environmental spaces based on plots of NDVI values in winter and summer from across the study area , comparing this environmental ‘background’ with corresponding values associated with known occurrences of bat species and rabies variant ., Niche models must be evaluated to validate their predictive power , before any use or interpretation 26 ., We evaluated the predictive ability of models for T . brasiliensis; however , sample sizes for L . cinereus were too small and too clumped spatially to permit detailed evaluations ., Two different spatial subsetting schemes were explored , taking advantage of the roughly linear shape of Chile ., First , we subset data latitudinally by quintiles of frequency , dividing occurrences into five subsets , and using subsets 1 , 3 , and 5 for model calibration and subsets 2 and 4 for evaluation 26 ., Second , we divided the study area into five equal-width latitudinal bands , again using subsets 1 , 3 , and 5 for model calibration and 2 and 4 for evaluation ., In the first scheme , subsets had equal sample sizes , whereas in the second scheme , subsets had similar areal dimensions ( Fig . S1 for supporting information ) ., For evaluating models , we avoided traditional receiver operating characteristic ( ROC ) area under the curve ( AUC ) approaches , considering that AUC tests require presence and absence data for proper implementation 41 , and in light of recent critiques 40 , 41 ., Rather , models were first evaluated using areas and points predicted as suitable and unsuitable after thresholding ( based on E\u200a=\u200a10% ) using a cumulative binomial probability distribution 26 ., Second , models ( without thresholding ) were evaluated using partial ROC approaches 42 , 43 , evaluating the predictive ability of niche models considering only omission errors and proportional areas predicted as suitable , and only over a range of omission errors deemed acceptable in light of error characteristics of the input data ( here again we used E\u200a=\u200a10% , and thus allowed up to 10% omission in our partial ROC calculations ) ., In partial ROC , the area under the observed line of model performance is related to the area under the line of random expectations , and a ratio is calculated ., Bootstrap manipulations ( 1000 total ) , in which 50% of evaluation data are resampled with replacement and AUC ratios recalculated , are used to test the hypothesis that model performance is better than random expectations ., When ≥95% of bootstrap-replicate AUC ratios were >1 , we rejected the null hypothesis of performance no better than random expectations 42 ., Partial ROC software is available for free download in http://kuscholarworks . ku . edu/dspace/handle/1808/10059 Finally , to compare niche models between virus strains and bat species , we used niche identity tests to determine whether two niche models are indistinguishable from one other 44 ., Identity tests have the advantage of restricting comparisons to the same set of points , a feature that is particularly relevant for our occurrence data , which did not come randomly from across the entire landscape ., We calculated observed Hellingers modified ( I ) and Schoeners ( D ) distances between niche models ( thresholded using minimum training presence approaches ) , and compared them to a null distribution of comparable distances derived from 1000 replicate random subdivisions of the overall pool of occurrence data between the two species , maintaining observed sample sizes ., We used ENMTools ( version 1 . 3; http://enmtools . com ) for these comparisons 45 ., We evaluated whether niche characteristics were identical between rabies lineages ( AgV6 versus AgV4 ) , between the host species and associated viruses , and between the two host species ., In all comparisons , our critical value was the 5th percentile of similarity ( i . e . , low end ) , as we were seeking evidence of niche differentiation 45 ., In all , 26 , 323 bat samples from active and passive surveillance were submitted to ISP during 1985–2011 , a data set that was captured digitally as part of this study ., However , many records corresponded to the same county centroids , such that sample sizes were nowhere near the number of samples: in all , to model hosts , we found 70 unique occurrences for L . cinereus ( 9% from GBIF; 91% from ISP ) and 238 for T . brasiliensis ( 3% from GBIF and 97% from ISP ) ., For rabies samples , we obtained 910 unique coordinates for rabies AgV4 ( bat rabies-positive associated with T . brasiliensis ) and 52 for rabies AgV6 ( associated with Lasiurus spp . ; Fig . 1 ) ; sample sizes are larger in this case because georeferencing was to street addresses , rather than just to county centroid ., Sampling intensity for T . brasiliensis varied 0–1178 samples submitted per municipality ( Fig . 2 ) , while that for L . cinereus varied 0–164; with only 64 of the 301 counties in the study area submitting L . cinereus samples ., Niche models , whether considering sampling bias or not , all performed significantly better than random expectations , with partial ROC AUC ratios associated with our niche models were >1 ( Fig . 3 ) ., However , considering that models controlling for sampling bias generated predictions with smaller suitable areas , we prefer to use these models in further steps ., For example , quintile subsetting considering sampling bias had less area predicted ( 35 . 2% of the study area ) than comparable models without considering sampling bias ( 38 . 0% of the study area ) ., Bias control also resulted in lower variance in AUC ratios in the partial ROC analyses ( Fig . 3 ) ., With this general confirmation of predictive power , we proceeded to build ecological niche models for each species ( Fig . 4 ) for interpretation ., None of the six identity tests comparing niches between the two host species , between each host species and its associated virus linage , and between the two virus lineages , was able to reject the null hypothesis of niche “identity” ( Table 1 ) ., Figure 5 shows the latter comparison graphically: observed similarity fell well above the critical value in all comparisons ., In sum , at least across central Chile , the two bat species and their associated viruses share very similar ecological niches , at least in the coarse-grained environmental dimensions explored in this study ., The two bat species had broad distributions in environmental space ( Fig . 6 ) ., Rabies infections were found across the great bulk of the environmental distribution of each of the hosts ., However , both hosts appear to avoid areas presenting extremely low NDVI values in summer and winter , corresponding to the high Andes regions ., In Chile , rabies has been reported as far back as 1879 46 ., All data have been centralized in the Sección de Rabia , Instituto de Salud Pública , since 1929 17 ., Via effective monitoring , mass dog vaccination , elimination of biting stray dogs , improvement of diagnosis quality , and post-exposure vaccination in humans , urban canine rabies was eradicated as of about 1990 47 , 48 ., However , over the same period , the zoonotic cycle , wherein the main reservoirs are bats , has been increasing in importance 16 ., Hence , in Chile , reports suggest rabies in a process of re-emergence in the wildlife cycle 16 , 17 , 49 ., Our large-scale data set , broad latitudinal gradient , and dramatic diversity of landscapes and biomes across the study area allowed a robust test and validation in the use of niche modeling in understanding the spatial epidemiology of bat-related rabies , as required when modeling diseases 50 ., Answering our first question , it was possible to characterize ecological niches of rabies viruses and their hosts consistently and with good predictive power ., In the broadest sense , niche models for the two bat species confirmed the obvious: the high Andes Mountains in the east and the Pacific Ocean in the west are natural barriers 18 , while the Atacama Desert to the north and cold regions in the south delimitated our study region naturally 29 ., With this definition of relevant areas , we derived clear predictions of the geographic distribution of both bat species ( Fig . 4 ) , wherein T . brasiliensis may be somewhat more limited in its use of cold and high zones in the Andes and the northern deserts than L . cinereus ( Fig . 4 ) ., The broad suitable areas for both species corroborate the ecological plasticity known in bats 51 and migratory behavior reported in the northern hemisphere for both T . brasiliensis and L . cinereus ., Niche models provided a first view of rabies distributions in geographic and environmental spaces 27 ., Our ecological niche models for rabies lineages using fine-resolution satellite imagery identified putative potential areas of rabies distribution , albeit under stable characterizations of environments averaged across several years of conditions; clearly , more dynamic characterizations of rabies distributions merit future evaluation ., Although we assembled large data sets that are reasonably comprehensive for Chile , we hasten to point out potential gaps and failings in our data and analysis ., A first such caution is that of the uneven spatial and environmental distribution of rabies in Chile: although samples were submitted from across the county , rabies locations were mainly from passive surveillance , producing three clusters of rabies cases in the main cities of central Chile ( Santiago , Valparaiso , Concepción; Fig . 1 ) , biases that we took into account in our analyses ., Using the bias file helped to reduce variance in model performance , allowing clearer discrimination of performance between models ( Fig . 3 ) ., We used sampling bias summaries for T . brasiliensis to consider the availability , quantity , and quality of data available for this species; for Lasiurus , parallel data were not available in sufficient quantity , reflecting the relative rarity of sample submissions for that species ., Incorporating information on sampling intensity in niche modeling for public health applications is an issue that merits further exploration , particularly considering that the more biased the data are , the more benefit that derives from use of sampling bias surfaces ., Our improvements in model performance with bias surfaces were analogous to previous results in biodiversity studies 38 ., As result , our models provide at least a preliminary assessment of risk in several areas that currently represent gaps in surveillance 52 ., Ecological niche models have seen detailed performance testing in challenges centered on estimating niches and predicting species distributions , showing impressive success even in spite of spatial sampling biases ( e . g . , sampling along roads ) 53 , 54 ., Problems arise when sampling is biased with respect to environments , however , since models based on such sampling will be effectively blinded to potential for occurrence in unsampled environments 53 , 55 ., An additional source of potential problems is the precision of georeferencing that was possible for these data , considering that reports of disease occurrence may simply provide the patients address , but not necessarily the site of infection , which is more relevant in spatial epidemiology 56 ., In this study , such problems introduce a basement level of spatial accuracy in model predictions , such that finest-resolution phenomena may not be “visible” in results ., In relation to our second question , it is important to note that , although viruses and hosts share ecological niche characteristics , the virus does not necessarily occupy the full host distribution ( Fig . 4 ) ; the geographic bias , however , at least within our study area , appears to be without consistent environmental correlates ., Our methodology corroborates the rabies-bat relationship that has heretofore gone untested at landscape scales , and our results suggest that niche modeling offers a useful tool for mapping disease occurrences and potential for occurrence in public health 27 ., With respect to our third question , niche identity tests between hosts and viral variants indicated that niches of all actors in the Chilean bat-rabies system are similar in environmental requirements; that is , we were unable to reject the null hypothesis that niche models of host species are not different from niches of associated virus strains , and indeed that the two host species and the two virus strains do not differ from one another either ., Currently , little is known about the ecology and transmission of rabies virus among bats , but phylogenetic evidence gives strong indications of host specificity 9 , 13 ., In this sense , not only do rabies virus variants appear to track the ecology of their respective hosts , but also the pairs of viruses and hosts do not differ from one another ., A recent report offers some corroboration of this assumption via molecular analysis: a rabies strain specific to Lasiurus spp ., bats was found in T . brasiliensis in Chile 13 , which indicates cross-species spillover transmission of virus lineages in taxonomically distant bat species under natural conditions ., These results support the idea that rabies viruses may infect hosts without environmental bias ( see 44 , for parallel results ) ., Restating , the bat species and rabies lineages evaluated appear to share very similar portions of environmental space , even if this result is not manifested as complete overlap in geographic space ( Fig . 4 ) , perhaps because different geographic distributions do not necessarily reflect niche differences 28 ., This result allows a view into how rabies host ecology influences virus biology , and suggests that taxonomic differences in hosts or viruses do not necessarily translate into ecological differences ., Our results and those of similar studies 51 , 57 may help to clarify the ecology of bat rabies lineages in other hosts and geographic regions ., Potential distribution maps of hosts and their viruses can be an important tool by which to understand potential transmission areas for rabies , although these approaches remain little explored 51 ., Bat-borne rabies has seen some events of cross-species transmission in zoonotic cycles in Chile , with AgV 4 ( related to T . brasiliensis ) found in Lasiurus spp ., and AgV 6 ( related to Lasiurus spp . ) found in T . brasiliensis 10 , 13 ., Accidental hosts have also been reported in recent years: for instance , mortality of dogs , cats , farm animals , and a human caused by rabies related to T . brasiliensis 10 , 13 ., Via this scenario , control of stray dogs and feral cats as well as vaccination campaigns must be implemented with priority in those areas where host and virus distribution match ( Fig . 4 ) ., In conclusion , one should take care to avoid the logical , scale-related error that can be termed the “Beale fallacy . ”, Beale et al . 58 , analyzed distributions of European birds with respect to climate , and concluded that their distributions were not limited by climate ., While this conclusion was , to some degree true , it was completely dependent on the particular context of Western Europe and relatively broadly-distributed bird species; a parallel analysis in a different context found abundant climatic determination of ranges 59 ., In this sense , our conclusion about no niche difference among our bat species and rabies lineages must be considered as context-dependent 59: analyses over broader regions may well detect clear and significant differences ., Our results show two viral lineages as sharing similar environmental signatures with two bat host species , regardless of antigenic characteristics , known associations , and phylogenetic position ., Recent years have seen important advances in molecular dimensions of studies of rabies , but few have explored how regional landscapes affect ( or not ) distributions and dynamics of rabies in zoonotic cycles 20 , 21 ., In light of the results reported herein , the spatial epidemiology and ecology of zoonotic bat rabies should see further exploration . | Introduction, Methods, Results, Discussion | Rabies was known to humans as a disease thousands of years ago ., In America , insectivorous bats are natural reservoirs of rabies virus ., The bat species Tadarida brasiliensis and Lasiurus cinereus , with their respective , host-specific rabies virus variants AgV4 and AgV6 , are the principal rabies reservoirs in Chile ., However , little is known about the roles of bat species in the ecology and geographic distribution of the virus ., This contribution aims to address a series of questions regarding the ecology of rabies transmission in Chile ., Analyzing records from 1985–2011 at the Instituto de Salud Pública de Chile ( ISP ) and using ecological niche modeling , we address these questions to help in understanding rabies-bat ecological dynamics in South America ., We found ecological niche identity between both hosts and both viral variants , indicating that niches of all actors in the system are undifferentiated , although the viruses do not necessarily occupy the full geographic distributions of their hosts ., Bat species and rabies viruses share similar niches , and our models had significant predictive power even across unsampled regions; results thus suggest that outbreaks may occur under consistent , stable , and predictable circumstances . | The situation of rabies in America has been changing: rabies in dogs has decreased considerably , but bats are increasingly documented as natural reservoirs of other rabies variants ., A significant gap exists in understanding of bat-borne rabies in Latin America ., We identified bat species known to be connected with enzootic rabies with different antigenic variants in Chile , and compiled large-scale data sets by which to test for ecological niche differences among virus lineages and bat hosts ., Our results begin to characterize important ecological factors affecting rabies distribution; modeling rabies in Chile allows comparisons across different latitudes and diverse landscapes ., We found that rabies virus strains are found in similar environments , regardless of the bat host involved ., This research improves understanding of bat-borne rabies dynamics , and important step towards preventing and controlling this and other emergent diseases linked to bats . | null | null |
journal.pgen.1006153 | 2,016 | Aversive Behavior in the Nematode C. elegans Is Modulated by cGMP and a Neuronal Gap Junction Network | Chemical stimuli , including odorants and tastants , can provide information about food availability and quality to affect appetitive behaviors across species ., While all sensory circuits use specialized cells in the periphery to detect environmental stimuli , how the sensitivity of sensory neurons is tuned and how chemical information is processed and relayed through downstream interneurons remains largely unknown ., As ultrastructural analyses of simple brains and small brain regions are providing connectome data 1–7 , the challenge that lies ahead is in understanding the dynamic properties of circuitry usage ., While mapped physical connections show the potential for information flow , the breadth of possibilities must also be reconciled with a circuit’s potential for neuromodulation as animals interact with a complex and changing environment ., For example , studies in systems ranging from invertebrates to mammals have revealed that chemosensory responses and feeding behaviors are modulated by an animal’s nutritional state 8–16 ., Furthermore , there is incredible complexity in the mechanisms by which nutritional status ultimately modulates chemosensory and feeding behaviors , reflecting the nervous system’s need to integrate information about what an animal has eaten , how much and when ., For example , neuropeptides , neurotransmitters ( e . g . serotonin and dopamine ) , hormones ( e . g . insulin ) and metabolites have all been shown to modulate chemosensory responses 8 , 12 , 15 , 17–21 ., Chemical synapses allow neurons to communicate with each other through the vesicular release of neurotransmitters into synaptic clefts between the cells ., These molecules then bind to receptors on the postsynaptic neurons ., In contrast , gap junctions allow for direct cytoplasmic communication and electrical coupling between neurons ., As such , gap junctions are often referred to as electrical synapses ., Importantly , the presence of gap junctions in the nervous system allows for the establishment of even more complex circuits than can be generated by synaptic signaling alone ., Vertebrate gap junctions are formed through the association of transmembrane connexin proteins ., Within one cell , six connexin proteins assemble to form one connexon hemichannel ., Two hemichannels on neighboring cells then dock which each other , allowing homotypic ( consisting of a single protein species on both cell membranes ) , heterotypic ( consisting of two different homomeric connexons on the two cell membranes ) or heteromeric ( consisting of a mixture of protein species on both cell membranes ) gap junctions to be made 22 , 23 ., Gap junction communication allows for the transmission of action potentials 24 , 25 , diffusion of metabolites and nutrients 26 and diffusion of second messengers , including Ca2+ , IP3 and the cyclic nucleotides cAMP and cGMP 27–32 ., The ability to pass such a diversity of molecules affords them the potential to repurpose hardwired circuitry to modify responses to diverse environmental stimuli ., The gap junction protein family consists of vertebrate connexins and invertebrate innexins ( invertebrate analogues of the connexins ) 33 , 34 ., The C . elegans genome encodes 25 members of this protein family 35 ., Although the functions of many C . elegans innexins remain unknown , the characterized cases have shown that these gap junction components are involved in diverse processes ranging from embryonic development and cell fate determination to adult neural functions 35–37 ., C . elegans is an ideal model system in which to study the link between hardwired connectivity and the functional circuitry usage that drives animal behavior ., The serial electron micrographs that showed the anatomical positioning of each of the 302 C . elegans hermaphrodite neurons also revealed all of the morphologically identifiable connections between neurons and between neurons and muscles 2 , 38–41 ., The recently updated C . elegans wiring diagram shows a total of 6393 chemical synapses , 890 gap junctions and 1410 neuromuscular junctions 42 ., It is through this interconnected neural network that C . elegans exploits a highly developed chemosensory system to detect olfactory and gustatory cues associated with food , danger and mating 43–45 ., In general , animals move towards chemicals that indicate a favorable environment , such as a potential food source , and away from stimuli that suggest a harmful environment ., The 11 pairs of C . elegans head chemosensory neurons extend their ciliated ends to the tip of the animal’s nose , allowing for direct or indirect exposure to sensory stimuli in their environment 2 , 39 , 40 ., Cellular laser ablation studies have been used to reveal the function of individual neuron pairs 43 ., For example , the ASEs sense water-soluble attractants , while the AWA and AWC neurons detect volatile odorants that C . elegans are attracted to and chemotax towards ., In addition , while the ASJ , ASI , ADF and ASG sensory neurons are primarily involved in regulating dauer formation , they do also play a role in other processes , including a minor role in chemotaxis ., Conversely , the ASH , ADL , AWB and ASK chemosensory neurons detect aversive stimuli that animals avoid by initiating backward locomotion upon stimulus detection ., The two bilaterally symmetric ASH nociceptors are particularly important for the avoidance of noxious stimuli , as they are “polymodal . ”, This neuron pair responds to a broad range of aversive stimuli , including not only soluble chemicals ( e . g . the bitter tastant quinine , heavy metals and SDS ) and odorants ( e . g . octanol ) , but also ions ( e . g . Na+ ) , osmotic stress and mechanosensory stimulation ( nose touch ) 46–54 ., ASH activation elicits reversal and stimulus avoidance because these glutamatergic sensory neurons synapse onto command interneurons that drive backward locomotion via their connections with motor neurons ., Thus , the nociceptive sensory system of C . elegans bears resemblance to its mammalian counterparts , in which noxious stimuli ( including chemicals ) are sensed predominantly by peripheral glutamatergic sensory neurons that synapse onto spinal dorsal horn neurons and , following further sensory processing in multiple regions of the brain , generate the perception of pain and aversive behavior 55–57 ., We previously reported a role for the cGMP-dependent protein kinase EGL-4 in the modulation of ASH-evoked nociceptive behavioral responses 58 ., C . elegans lacking EGL-4 function are hypersensitive in their response to a subset of ASH-detected stimuli; egl-4 ( lof ) animals avoid dilute stimuli that wild-type animals do not respond to ., Our data suggested that EGL-4 likely normally acts to dampen ASH sensitivity by phosphorylating and activating the GTPase activating proteins RGS-2 and RGS-3 , which then downregulate G protein-coupled sensory signaling in the ASH nociceptors ., Surprisingly , although EGL-4 requires cGMP binding to negatively regulate ASH sensitivity , no guanylyl cyclases are known to be expressed in ASH 59 ., Herein we provide evidence that the C . elegans transmembrane guanylyl cyclase ODR-1 functions in a non-cell-autonomous manner to provide cGMP to regulate EGL-4 function in ASH ., Like egl-4 ( lof ) animals , odr-1 ( lof ) animals are hypersensitive in their avoidance of a subset of ASH-detected stimuli ., However , while EGL-4 functions directly in the ASHs , ODR-1 expression in the AWB , AWC and ASI head sensory neurons is sufficient to restore normal behavioral sensitivity to odr-1 ( lof ) animals ., We further provide evidence that the pool of cGMP produced by ODR-1 flows through a gap junction network from its site of production to the ASH nociceptors ., Taken together , our data reveal a new way by which an animal’s nervous system can utilize information about the organism’s well-being to set the threshold of nociceptor sensitivity and coordinate behavioral responses that are appropriate for its internal state ., We previously reported that animals lacking the function of the guanylyl cyclase ODR-1 are hypersensitive in their behavioral avoidance response to dilute concentrations of the bitter tastant quinine 58 ., Significantly more odr-1 loss-of-function ( lof ) animals respond to dilute ( 1 mM ) quinine than wild-type animals ( Fig 1 ) 58 ., The ASH sensory neurons are the main cells used to detect quinine in C . elegans , but the ASK neurons also contribute 52 ., ODR-1 is not expressed in ASH , but is expressed in five other head sensory neurons—AWB , AWC , ASI , ASJ and ASK 60 , 61 ., To determine in which cell ( s ) ODR-1 function is sufficient to dampen quinine sensitivity , we restored ODR-1 function in each neuron pair that it is natively expressed in , using the following cell-specific or -selective promoters: str-1p ( AWB ) , ceh-36p3 ( AWC ) , gpa-4p ( ASI ) , trx-1p ( ASJ ) and srbc-66p ( ASK ) 62–66 ., While expressing ODR-1 in ASJ or ASK had only a minimal effect on quinine sensitivity , individually expressing ODR-1 in either the AWB , AWC or ASI sensory neurons partially rescued the odr-1 ( lof ) quinine hypersensitivity phenotype ( Fig 1A ) ., This suggested that ODR-1 function in more than one neuron pair regulates the quinine response ., Co-injection of str-1p::odr-1 , ceh36p3::odr-1 and gpa-4p::odr-1 to simultaneously express ODR-1 in the AWB , AWC and ASI sensory neurons of odr-1 ( lof ) animals returned quinine response to the level seen when odr-1 was expressed under the control of its own promoter ( odr-1p::odr-1 ) , consistent with ODR-1 functioning in multiple neurons to regulate quinine sensitivity ( Fig 1 ) ., odr-1 ( lof ) animals develop with altered membraneous structures at the distal segments of the AWB , but not ASI , dendritic cilia 67; AWC cilia structure was not examined in this study ., To assess when ODR-1 function is required to downregulate quinine sensitivity , the odr-1 gene was placed under the control of a heat shock inducible promoter 68 and introduced into odr-1 ( lof ) animals ., Induction of odr-1 expression by heat shock in adult animal stages returned the behavioral response to dilute quinine to wild-type levels when assayed four hours later ( Fig 1B ) ., Transgenic animals that were not heat shocked remained hypersensitive , similar to odr-1 ( lof ) animals ( Fig 1B ) ., These results demonstrate that , even though cilia morphogenesis is likely an active process that continues through the late larval stages 67 , odr-1 is only required in adult stages for normal behavioral sensitivity to dilute quinine ., This heat shock-induced expression of odr-1 is also after developmental cell fate specification and neuronal connectivity is complete ., ODR-1 is a receptor-type guanylyl cyclase with an extracellular domain ( ECD ) and an intracellular catalytic domain that processes GTP into cGMP 61 ., To assess the contribution of the ECD to ODR-1 function in the regulation of quinine avoidance behavioral sensitivity , we expressed an ODR-1 construct lacking the extracellular domain ( ΔECD ) 61 under the control of its native promoter in odr-1 ( lof ) animals ( Fig 1C ) ., Expression of the odr-1p::odr-1 ( ΔECD ) construct restored quinine sensitivity to wild-type levels , similar to the expression of the wild-type odr-1 construct ( odr-1p::odr-1 ) ( Fig 1C ) ., This suggests that the extracellular receptor region is not necessary for ODR-1 function in modulating quinine sensitivity ., To determine whether ODR-1’s ability to produce cGMP is necessary for regulation of quinine sensitivity , we expressed ODR-1 harboring a point mutation that abolishes GTP binding in the catalytic domain 61 ., odr-1 ( lof ) animals expressing odr-1p::odr-1 ( E874A ) remained hypersensitive in their response to quinine ( Fig 1C ) ., This indicates an important role for cGMP in modulation of the quinine response and is consistent with our previous demonstration that the C . elegans cGMP-dependent protein kinase EGL-4 also regulates quinine response sensitivity 58 ., The above results suggested a modulatory role for the AWB , AWC and ASI neurons in quinine behavioral sensitivity ., To further examine their contribution to the regulation of the quinine response , we genetically ablated these cells in wild-type animals , both as individual neuron pairs and in combination , using a reconstituted caspase approach 69 , 70 ., Ablation of either the AWBs , AWCs or ASIs did not produce a marked hypersensitive phenotype ( Fig 2A ) ., While animals lacking two of the three neuron pairs displayed greater hypersensitivity than the single ablates , ablation of all three neuron pairs together resulted in the greatest degree of quinine hypersensitivity ( Fig 2A ) ., Taken together , our results reveal a role for the AWB , AWC and ASI sensory neurons in the negative regulation of quinine avoidance , and further suggest that ODR-1 function in these cells contributes to their modulatory role ., As the ASH nociceptors are the primary cells used to detect quinine 52 , we sought to determine how the AWB , AWC and ASI sensory neurons might influence signaling through the ASH sensory circuit ., One possibility could be via synaptic signaling between these neurons ., UNC-13 protein is required for synaptic vesicle fusion and neurotransmitter release at synapses 71 , 72 , and AWB and ASI are known to form direct synaptic connections onto ASH 2 ., Because unc-13 null mutations are lethal , we used cell-specific RNAi 73 to knockdown unc-13 in AWB , AWC and ASI and block synaptic transmission from these neurons ., We confirmed efficient unc-13 RNAi knockdown using chemosensory assays that require synaptic signaling from AWB and AWC for proper behavioral responses ( S1 Fig ) ., Animals in which unc-13 was simultaneously knocked down in the AWB , AWC and ASI sensory neurons did not show increased sensitivity to dilute quinine ( Fig 2B ) , suggesting that vesicular synaptic transduction is not the mechanism by which these neurons influence the ASH-mediated response to quinine ., A second way in which neurons can communicate with each other is via diffusion of ions and small metabolites through gap junctions ., Studies in mammalian systems have suggested that cyclic nucleotides can also pass through gap junctions to affect cellular function ., For example , cAMP movement between cells has been visualized following ectopic expression of gap junction components in human tissue culture 29 , 74 , 75 ., cAMP movement through gap junctions also suppresses CD4+ T-cell function 76 and may alter gene expression in myelinating Schwann cells 77 ., In addition , cGMP can pass through gap junctions in human cell culture , as well as between cultured mouse follicle cells and oocytes 29 , 30 , 32 , 78 ., While 25 innexins are encoded by the C . elegans genome , the physiological roles of most are unknown 35–37 ., To determine whether gap junction signaling can modulate quinine sensitivity , we assayed animals with loss-of-function alleles for 16 of the 25 innexins encoded by the C . elegans genome for response to dilute ( 1 mM ) quinine ( see Supplemental Materials and Methods ) ., Two innexin mutants , inx-4 ( lof ) and inx-20 ( lof ) , responded better than wild-type animals to dilute quinine ( Fig 3A ) ., While inx-20 expression has only been reported in the pharyngeal epithelium and the pm2 pharyngeal muscle cell , inx-4 expression was seen in several head and tail neurons , including the ASHs of early larvae and the ADFs of L1s 35 ., Using an inx-4p::gfp reporter construct , we have also confirmed inx-4 expression in the ASH nociceptors of adult animals ( S2 Fig ) ., To determine whether INX-4 function in the ASHs is sufficient to regulate quinine response , the ASH cell-selective promoters osm-10 48 and srd-10 were used to restore INX-4 function and animals were assayed for response to 1 mM quinine ., Expression in ASH using either promoter dampened the inx-4 ( lof ) hypersensitive response , while simultaneous expression of INX-4 in ASH and ADF , using the osm-10 48 and srh-142 79 promoters , respectively , did not result in additional rescue ( Fig 3B ) ., Furthermore , consistent with the reported lack of inx-4 expression in ODR-1-expressing neurons 35 , an odr-1p::inx-4 construct did not rescue the quinine hypersensitivity of inx-4 ( lof ) animals ( Fig 3B ) ., These results suggest that the ASH nociceptors are the primary site for INX-4 function in regulating quinine sensitivity ., Since inx-4 is expressed from larval through adult stages , the inx-4 gene was placed under the control of a heat shock inducible promoter 68 and introduced into inx-4 ( lof ) animals to determine when inx-4 function is required to modulate quinine sensitivity ., Induction of inx-4 expression by heat shock in adult animal stages significantly dampened the behavioral hypersensitivity of inx-4 ( lof ) animals to dilute quinine when assayed four hours later ( Fig 3C ) ., Transgenic animals that were not heat shocked remained hypersensitive , similar to inx-4 ( lof ) animals ( Fig 3C ) ., These results demonstrate that , like ODR-1 ( Fig 1B ) , INX-4 function in adult animal stages is sufficient to modulate behavioral sensitivity to dilute quinine ., We previously found that egl-4 ( lof ) animals are hypersensitive in their response to distinct ASH-detected stimuli and , although EGL-4 functions in the ASH sensory neurons to regulate these behaviors , this cGMP-dependent protein kinase does not regulate ASH sensitivity in general 58 ., For example , egl-4 ( lof ) animals respond normally to the bitter tastant primaquine , the detergent SDS and the heavy metal copper 58 , all of which are detected by the ASH nociceptors 49 , 51–53 ., To determine whether loss of odr-1 or inx-4 increased overall sensitivity of ASH , or also selectively affected sensory signaling , we assayed odr-1 ( lof ) and inx-4 ( lof ) animals for their response to primaquine , SDS and copper ., For each of these stimuli , odr-1 ( lof ) and inx-4 ( lof ) animals responded similarly to wild-type animals ( S3 Fig ) , indicating that ODR-1 and INX-4 also do not regulate ASH signaling in general ., If INX-4 functions in the same pathway as ODR-1 to dampen quinine sensitivity , then odr-1 ( lof ) ;inx-4 ( lof ) double mutant animals should display a behavioral phenotype similar to odr-1 ( lof ) animals and the hypersensitivity should not be additive ., We found that the double mutant animals’ response to dilute ( 1 mM ) quinine was indistinguishable from animals lacking only ODR-1 function ( p > 0 . 2 , Fig 3D ) , suggesting that they do function in the same regulatory pathway ., Consistent with this observation , in the odr-1 ( lof ) ;inx-4 ( lof ) background , expression of only either odr-1 or inx-4 had no effect on the quinine hypersensitivity ( Fig 3D ) ., However , simultaneous expression of inx-4 in ASH and expression of odr-1 either behind its native promoter or using AWB , AWC and ASI cell-selective promoters , rescued hypersensitivity ( Fig 3D ) ., No guanylyl cyclase has been found to be expressed in ASH 58 , 59 and the results described above suggest that cGMP generated at distant sites can dampen ASH response to sensory stimuli ., To determine if the mere presence of cGMP in ASH is sufficient to dampen quinine sensitivity , the ASH cell-selective promoters osm-10 48 and srb-6 80 were used to express a blue light-inducible guanylyl cyclase ( BlgC ) 81 in the ASHs of animals lacking the blue-violet light receptor LITE-1 82 ., When assayed 10 minutes after a 30-second exposure to blue light , animals expressing BlgC in the ASH sensory neurons displayed a diminished response to both 5 mM ( Fig 4A ) and 10 mM quinine ( S4 Fig ) , while transgenic animals that were not flashed with blue light displayed wild-type sensitivity ., Although cAMP generation by BlgC was undetectable in E . coli , BlgC was shown to posses ~10% residual adenylyl cyclase activity in vitro 81 ., To confirm that the dampened quinine response was due to production of cGMP , and not cAMP , we also assayed animals expressing a blue light-inducible adenylyl cyclase , BlaC 81 ., Unlike BlgC , blue light-induction of BlaC to stimulate cAMP production did not alter animals’ response to quinine ( Fig 4A and S4 Fig ) ., Blue light induction of BlgC in the ASHs was also sufficient to significantly diminish the hypersensitivity of odr-1 ( lof ) animals in response to 1 mM quinine ( Fig 4B ) ., Together , these results demonstrate that elevating cGMP levels in ASH is sufficient to dampen behavioral sensitivity to quinine ., Based on the C . elegans hermaphrodite wiring diagram ( we referred to WormWiring . org for the most current wiring annotations based on the original electron micrograph series reported in 2; Scott Emmons , personal communication ) , the ODR-1-expressing sensory neurons AWB , AWC and ASI do not form gap junction connections directly with ASH ., However , these neurons are connected to ASH indirectly through a gap junction network via ADF ( AWB , AWC ) , AFD ( AWC ) and AIA ( AWB , AWC and ASI ) ( Fig 5A ) ., The ADFs are chemosensory , the AFDs are thermosensory neurons , and the AIAs are interneurons 43 , 83 ., ADF is also directly connected to both AFD and AIA by gap junctions ( Fig 5A ) ., These connections reveal a neuronal circuitry wherein ADF , AFD and AIA lay between the cGMP generating neurons ( AWB , AWC and ASI ) and ASH ., We note that RMG and RIC also connect AWB to ASH , but their role in modulating the quinine response was not examined in this study ., To determine whether ADF , AFD or AIA regulate the quinine response , we genetically ablated these cells in wild-type animals , both as individual neuron pairs and in combination , again using reconstituted caspases 69 , 84 ., As shown in Fig 5B , loss of any one of the three neuron pairs resulted in significant behavioral hypersensitivity , as did simultaneous ablation of all three ., In fact , the degree of quinine hypersensitivity seen in the ADF/AFD/AIA ablated animals was indistinguishable from that of the AWB/AWC/ASI animals ( p > 0 . 5 ) ., We noted that ADF and AIA do also chemically synapse onto ASH 2 ( and WormWiring . org ) ., To confirm that vesicular synaptic signaling from these neurons does not underlie their ability to decrease ASH sensitivity , we again utilized cell-specific RNAi 73 to knockdown unc-13 in ADF and AIA to block synaptic transmission from these neurons ., Animals in which unc-13 was knocked down in either the ADF or AIA neurons did not show increased sensitivity to dilute quinine ( Fig 5C ) ., Even though AFD does not synapse onto ASH , we also confirmed that unc-13 RNAi in this neuron pair did not affect quinine sensitivity ( Fig 5C ) ., Taken together , our data suggest that gap junction-mediated communication between AWB/AWC/ASI and ASH , via ADF/AFD/AIA can regulate ASH sensitivity and an animal’s response to environmental stimuli ., No guanylyl cyclases are known to be expressed in the ADF sensory neurons 59 ., Therefore , we next sought to determine whether ectopic cGMP generation in these cells , which lay between the ODR-1-expressing neurons and the ASHs in the gap junction network ( Fig 5A ) , would be sufficient to dampen quinine sensitivity ., The ADF-specific srh-142 promoter 79 was used to drive expression of BlgC 81 in lite-1 ( lof ) or inx-4 ( lof ) ;lite-1 ( lof ) animals ., After exposure to blue light , lite-1 ( lof ) animals expressing BlgC in the ADFs displayed a diminished response to 5 mM ( Fig 6 ) and 10 mM ( S5 Fig ) quine , while transgenic animals that were not exposed to blue light displayed wild-type sensitivity ., Conversely , inx-4 ( lof ) ;lite-1 ( lof ) transgenic animals expressing BlgC in the ADFs did not display a diminished response following blue light exposure ( Fig 6 and S5 Fig ) ., Consistent with the results described above ( Fig 4A ) , blue light-induction of BlaC to stimulate cAMP production in the ADFs did not alter animals’ response to quinine ( Fig 6 and S5 Fig ) ., Together , these data demonstrate that elevating cGMP levels in the ADFs is sufficient to dampen behavioral sensitivity to quinine , and that function of the INX-4 gap junction component is required for this effect ., The sheer number of possible circuit outcomes revealed by anatomical wiring diagrams means that we cannot possibly predict how information might flow through a circuit based on physical connections alone ., Even in the relatively simple nervous system of C . elegans , the function of most of the connections , or which connections will be preferentially used under different circumstances , is not known 85 ., For example , neuropeptide signaling regulates a sensory context-dependent switch in the composition of a C . elegans salt sensory circuit 86 ., In this case , one of the AWC olfactory sensory neurons is recruited to function as an interneuron during response to high salt concentrations 86 ., It was also recently reported that , in a reciprocal inhibition circuit that fine-tunes copper ( a heavy metal ) avoidance , the ADF sensory neurons can be activated by neuropeptides to act as interneurons downstream of the ASI sensory neurons ., The ADFs then dampen ASH sensitivity via serotonin release 87 ., In addition , the C . elegans nose touch circuit appears to utilize electrical synapses to mediate both excitatory and inhibitory interactions between neurons 88 ., Adding another layer of complexity , the internal state of an animal , such as its nutritional status and degree of satiety , can influence its sensitivity to environmental stimuli 58 , 89–93 , and may even alter which neurons participate in stimulus detection 89 ., Sensory systems in particular may be subject to extensive modulation 8 , 15 ., For example , feeding state and food availability can alter gustatory and olfactory responses in diverse species , where the complexity of the nervous systems ranges from just hundreds of neurons to billions of neurons ., To better understand the neuromodulatory mechanisms that regulate chemosensation , we focused on one component of the regulatory pathway that controls C . elegans response to aversive sensory stimuli , the guanylyl cyclase ODR-1 ., As previously reported , the two AWC olfactory neurons require ODR-1 function to mediate chemotaxis towards attractive odorants that they detect , while the AWBs require ODR-1 for 2-nonanone avoidance 61 ., In addition , ODR-1 plays a role within the AWCs to regulate adaption in response to prolonged odor exposure 61 ., Here , we describe a new role for ODR-1 in modulating animals’ avoidance response to the bitter tastant quinine in a non-cell-autonomous manner ., This is , to our knowledge , the first in vivo demonstration of a guanylyl cyclase functioning to modulate the activity of another cell ., odr-1 ( lof ) animals are hypersensitive in their response to dilute concentrations of the bitter tastant quinine , and ODR-1 function in three distinct pairs of sensory neurons ( the AWBs , AWCs and ASIs ) appears to contribute to the modulation of ASH-mediated avoidance ( Fig 1 ) ., However , the evolutionary advantage of this sort of decentralized regulation of sensory signaling is not immediately clear ., One possibility could be to prevent overstimulation of ASH , which is the main nociceptor in C . elegans ., Nociceptors in general have high thresholds of activation , understandably to prevent organisms from unnecessarily reacting to minute or low-risk stimuli 94 ., An additive modulatory role for AWB , AWC and ASI may help to assure that the threshold of ASH sensitivity to noxious stimuli remains high , consistent with the observed behavioral hypersensitivity we observed upon ablation of these three neurons pairs ( Fig 2 ) ., Another potential benefit of decentralized ASH regulation is that by utilizing multiple sensory neurons to regulate ASH function , the integration of diverse sets of environmental information can maximize the appropriateness of an animal’s response to its surroundings ., For example , in addition to detecting the aversive odorants 100% octanol and 2-nonanone 66 , 89 , which may indicate the presence of fungi or pathogenic bacteria 95–97 , AWB also mediates lawn avoidance upon encountering the pathogenic bacteria Serratia marcescens 97 ., All of these roles for AWB could help C . elegans avoid environments that might be harmful to them ., Conversely , scents detected by AWC ( benzaldehyde , butanone , isoamyl alcohol , 2 , 3-pentanedione , and 2 , 4 , 5-trimethylthiazole ) 98 , 99 are primarily produced in nature by plants ( e . g . fruits , nuts and coffee ) and microorganisms ., Therefore , these natural odorants signal potential sources of nutritive bacteria and it has been shown that C . elegans will populate areas of decaying plant matter 100 ., Finally , if developing larvae encounter harsh environmental conditions , including elevated population density and/or a limited food supply , they can enter dauer arrest at the second molt 101 ., This transition is repressed by the ASI , ADF and ASG neurons , which detect the dauer pheromone that serves as an indicator of population density ., A high density of C . elegans in a given area could signify low or dwindling food availability there , which could in turn result in poor nutritional status ., Thus , the AWBs , AWCs and ASIs detect distinct sets of environmental stimuli that can all provide the animal with information about potential food quality and availability ., Following the initial detection of an environmental stimulus , the AWB , AWC or ASI sensory neurons may then subsequently modulate ASH sensitivity to indirectly optimize nociceptive responses in a context-dependent manner ., For example , when animals are well-fed , they are more sensitive to aversive stimuli , including quinine ( Fig 7 ) , than they are upon food deprivation 58 , 89–93 ., This may reflect the need of animals to reprioritize their behaviors to balance the need to avoid potentially dangerous situations with the need to find food ., If starving , minimized aversive responses may maximize entry into new environments to increase the likelihood of encountering new food sources ., We speculate that food somehow suppresses ODR-1 activity in the upstream AWB , AWC and/or ASI sensory neurons , while removal of food allows for ODR-1 activation and cGMP accumulation ( Fig 8A ) ., ODR-1 is most similar to transmembrane guanylyl cyclases , which are often regulated by extracellular peptides 102 ., However , we found that the extracellular domain is not required for ODR-1 function in modulating ASH sensitivity ( Fig 1 ) ., In addition , while soluble ( cytoplasmic ) guanylyl cyclases are generally activated by nitric oxide 103 , C . elegans lack a nitric oxide synthase ., Thus , the mechanistic link between an animal’s feeding status and ODR-1 activity is unclear ., Interestingly , calcium levels rise in the AWCs upon withdrawal of either odorant or bacterial-conditioned media 104 ., An attractive possibility is that this increase in calcium could directly or indirectly activate ODR-1 activity in the AWCs , which would provide a link between food withdrawal , cGMP generation in this neuron pair and the downstream dampening of ASH-mediated responses ., Alternatively , ODR-1 may be constitutively active and phosphodiesterase activity may be regulated by an animal’s feeding status to adjust the pool of available cGMP ., Collectively , our data support a model wherein , upon food removal , the ODR-1-expressing neurons AWB , AWC and ASI provide a pool of cGMP that flows through a gap junction network from the site of its production in these sensory neurons , through ADF , AFD and AIA , to the ASH nociceptors ( Fig 8A ) ., However , the gap junctions reported in the wiring diagram first published by White et al . 2 do not provide a straightforward explanation for our experimental observations ( Fig 8B ) ., For example , White et al . 2 did not report gap junctions between ASH and ADF , AFD or AIA ( Fig 8 | Introduction, Results, Discussion, Materials and Methods | All animals rely on their ability to sense and respond to their environment to survive ., However , the suitability of a behavioral response is context-dependent , and must reflect both an animal’s life history and its present internal state ., Based on the integration of these variables , an animal’s needs can be prioritized to optimize survival strategies ., Nociceptive sensory systems detect harmful stimuli and allow for the initiation of protective behavioral responses ., The polymodal ASH sensory neurons are the primary nociceptors in C . elegans ., We show here that the guanylyl cyclase ODR-1 functions non-cell-autonomously to downregulate ASH-mediated aversive behaviors and that ectopic cGMP generation in ASH is sufficient to dampen ASH sensitivity ., We define a gap junction neural network that regulates nociception and propose that decentralized regulation of ASH signaling can allow for rapid correlation between an animal’s internal state and its behavioral output , lending modulatory flexibility to this hard-wired nociceptive neural circuit . | Studying the logic of small neural circuits is an essential step toward understanding more complex circuits and , ultimately , the computational and integrative properties of whole nervous systems ., With a compact nervous system ( just 302 neurons ) and a well-characterized behavioral repertoire , the small roundworm C . elegans serves as an excellent animal model to study circuit-level modulation of neuronal function ., By employing a combination of genetic and behavioral approaches , we have identified a non-cell-autonomous role for the guanylyl cyclase ODR-1 in the regulation of nociceptive sensory behaviors , and we provide the first evidence for circuit-level modulation of neuronal activity by the second messenger cGMP ., While revealing a new mechanism for the coordination and optimization of animal behavior , our work also supports a growing body of evidence that hard-wired neuronal circuitry can be dynamically repurposed in a context-dependent manner . | biotechnology, cell physiology, invertebrates, alkaloids, medicine and health sciences, chemical compounds, nervous system, caenorhabditis, immunology, junctional complexes, electrophysiology, neuroscience, animals, gap junctions, animal models, caenorhabditis elegans, model organisms, clinical medicine, animal behavior, hypersensitivity, zoology, genetic engineering, research and analysis methods, genetically modified organisms, quinine, animal cells, behavior, chemistry, genetically modified animals, agriculture, sensory neurons, cellular neuroscience, cell biology, anatomy, synapses, clinical immunology, physiology, neurons, nematoda, biology and life sciences, cellular types, physical sciences, agricultural biotechnology, neurophysiology, organisms | null |
journal.pcbi.1003154 | 2,013 | A Systematic Framework for Molecular Dynamics Simulations of Protein Post-Translational Modifications | Proteins in the cell continually get covalently modified in different post-translational , enzyme-controlled reactions 1–3 ., Additionally , protein modifications frequently arise in a non-controlled fashion as well , mainly as a consequence of oxidative stress 4 ., While enzymatic post-translational modifications ( PTMs ) play important regulatory roles in a large number of different cellular processes , non-enzymatic PTMs are predominantly linked with protein damage and are involved in age-related diseases such as neurodegenerative disorders , diabetes and cancer 2 , 4–7 ., Despite the general importance of PTMs in different biological contexts , their effect on protein structure , dynamics and interaction networks at the atomistic level remains poorly understood ., In particular , molecular dynamics ( MD ) simulations , a widely used high-resolution computational method for studying biomolecular properties and behavior 8–10 , have been limited to unmodified , native proteins due to a surprising deficiency of suitable tools and systematically developed parameters for treating PTMs , with only sporadic exceptions 11–16 ., MD simulations capture atomic and molecular motions based on Newtons equation of motion and an empirical potential energy function that defines interactions between simulated particles ., The latter is defined by a force field , i . e . a self-consistent set of physically realistic equations and semi-empirical parameters describing all interactions in a given system ., Force-field parameters are typically obtained by fitting atomic or molecular properties of small molecules against calculated quantum-mechanical or experimentally measured data ., As the applied parameterization strategies often differ from each other , considerably different parameter values have been derived in many cases 17–20 ., Here , we develop force field parameters for over 250 different types of enzymatic and non-enzymatic modifications of amino-acid side chains as well as protein termini within the context of GROMOS 45a3 19 and 54a7 21 , 22 force fields ( Table S1 ) ., We choose GROMOS force fields because of their widespread usage , high accuracy in reproducing experimental results and general transferability of parameters when it comes to identical chemical groups in different compounds 21 ( e . g . from the hydroxyl group of tyrosine to the hydroxyl group of 7-hydroxytryptophan ) ., The functional form of a typical force field is exemplified in equation 1 for GROMOS class force fields , ( 1 ) with parameters highlighted using boldface letters and RF representing a reaction field contribution to the electrostatic interactions ., The non-bonded interaction terms in the GROMOS force field are primarily parameterized against thermodynamic data of small molecules , either in the pure liquid state , or in aqueous or nonpolar solution ., Therefore , we validate the obtained parameters by reproducing experimental hydration free energies ( HFEs ) , a measure of hydrophobicity and arguably one of the most important amino-acid properties with implications in protein folding , ligand binding or protein-lipid interactions ., Finally , we analyze physico-chemical properties related to hydrophobicity of all parameterized PTMs according to their type and compare them against the 20 canonical amino acids ., One of the principal objectives in our parameterization has been the coverage of experimentally known PTMs , which is as complete as possible ., Following an exhaustive literature search and analysis of an online PTM database PTMdb 23 , we have compiled a diverse list of enzymatic and non-enzymatic PTMs , including phosphorylation , methylation , acetylation , hydroxylation , carboxylation , carbonylation , nitration , deamidation and many others ( Figure 1a , Table S1 ) , covering a total of 259 distinct PTM reactions or 110 non-redundant post-translationally modified amino acids and protein termini ., The lower number in the latter case reflects the fact that different PTM reactions can lead to the same modified product ( e . g . glutamic semialdehyde is a product of both arginine and proline carbonylation ) ., We have generated GROMOS 45a3 ( Dataset S1 ) and 54a7 ( Dataset S2 ) force field parameters for the non-redundant set of compounds by either direct transfer or analogy to already parameterized compounds including amino acids , nitrogenous bases and other small molecules or completely novel parameterization ( see Methods for more details ) ., How well do the obtained parameters cover the space of biologically relevant PTMs ?, To address this question , we have analyzed PTMs that have been experimentally verified ( 72 , 984 ) and annotated as such in the UniProt database 24 ( 21 , 411 protein entries , Dataset S3 ., Phosphorylation is by-far the most abundant modification type in the UniProt database ( 78 . 5% of all UniProt PTMs ) , followed by acetylation , hydroxylation and methylation ( Figure 1b ) ., Note that terminal PTMs account for a sizable fraction of all annotated modification at 8 . 3% ., Strikingly , the parameterized compounds reported herein match every annotated phosphorylation modification , 99 . 9% of acetylation , 99 . 2% of hydroxylation and 99 . 7% of methylation modifications , for a grand-total coverage of 98 . 5% of all PTMs reported in UniProt ( Figure 1c ) ., Concerning PTMs that are not covered by our parameters , they are all extremely rare , each accounting for less than 0 . 5% of all UniProt PTMs ., Finally , we provide parameters for 33 PTMs ( Table S1 ) , mostly non-enzymatic ones , that have to date not been reported in UniProt ., HFE , a free energy difference between a compound solvated in water and the same compound in the gas phase , is an experimentally measurable property related to hydrophobicity , and it has been originally used to re-parameterize the GROMOS force field in 2004 21 ., A proper description of functional groups in the hydrated phase is of crucial importance for virtually all relevant biomolecular processes , so we have used the same thermodynamic quantity to validate the parameters obtained in the present study ., To the best of our knowledge , experimental HFEs are available for the exact side chain analogs of 13 parameterized PTMs only and we have therefore in the validation set also included compounds , which are chemically related to PTM side chains for which no experimental HFEs were available , for a total of 26 different molecules ( only a single representative compound was included for each group of PTMs involving the same chemical moiety , Table 1 ) ., Note that the additional compounds related to PTM side chains have been parameterized in the same way as the relevant PTMs ., We have used MD simulations and the thermodynamic integration ( TI ) approach 25 ( see Methods for more details ) to calculate the HFEs for neutral forms small-molecule analogs of the canonical amino-acid side chains and for the compounds in the validation set using both the 45a3 ( Table S2 ) and 54a7 ( Table, 1 ) parameter sets of the GROMOS force field ., As a consequence of the parameterization strategy behind them , the canonical amino acids exhibit an excellent agreement with experimental HFEs when it comes to the 54a7 parameter set , with a root-mean-square error ( RMSE ) of 3 . 3 kJ/mol ( RT\u200a=\u200a2 . 5 kJ/mol at room temperature ) and an almost perfect correlation with experimental HFEs ( correlation coefficient R2\u200a=\u200a0 . 98 ) ( Figure 2 ) ., Remarkably , the newly generated GROMOS 54a7 force field parameters of PTM-related compounds exhibit a nearly equal level of matching of experimental HFEs with an RMSE of 4 . 2 kJ/mol ( Table, 1 ) and a correlation coefficient R2 of 0 . 94 ( Figure, 2 ) over 25 different compounds , excluding a single outlier , 2-nitrophenol ( Figure 2 , red X symbol ) ., This compound , containing nitro and hydroxyl groups attached to a benzene ring , deviates from the experimental value by 14 . 6 kJ/mol ., Considering the outlier 2-nitrophenol in more detail , additional calculations have shown that p-cresol ( a tyrosine side-chain analog ) , o-cresol , m-cresol and nitrobenzene , compounds containing either a hydroxyl group or a nitro group attached to a benzene ring , agree well with experimental HFEs with an overall RMSE of 2 . 7 kJ/mol only ., This suggests that , although parameters of individual groups do reproduce experimental HFEs , the agreement with experiment may significantly worsen if they appear in combination ., In order to test this , we have calculated HFEs of 3- and 4-nitrophenol and compared them against experimental values ., Interestingly , the calculated HFEs of both compounds match experimental values ( Table, 1 ) suggesting either that these groups exert a specific influence on each other only in 2-nitrophenol or that the experimentally measured HFE may simply not be reliable for this compound ., To account for the former possibility , we have derived a set of parameters de novo for 2-nitrophenol that closely match its experimental HFE with an absolute value of the deviation of 1 . 8 kJ/mol ( Table 1 ) ., Note that we report both versions of nitrotyrosine ( Table S1 ) , a cognate PTM to 2-nitrophenol ., Finally , we have also excluded 4-methylimidazole ( a histidine side-chain analog ) and 1-methylimidazole from the HFE analysis of the canonical amino acids and PTMs , respectively , even though experimental HFEs are available for both compounds ., Since histidine exists in two tautomeric states , described by different parameters , the calculated HFE depends on the choice of the state used for calculations , with one matching the experimental HFE and the other varying by approximately 20 kJ/mol ( Table 1 ) ., Consequently , the same problem exists for 1′- and 3′-methylhistidine , whose parameters are based on those of histidine , where one tautomer matches while the other deviates from the experimental HFE ( Table 1 ) ., In contrast to GROMOS 54a7 , the 45a3 parameter set does not reproduce experimental HFEs well ( Table S2 and Figure S1 ) ., Namely , the slope of 0 . 79 and the offset of 3 . 8 kJ/mol of the regression line suggest that the calculated HFEs are largely overestimated ( RMSE\u200a=\u200a10 . 8 kJ/mol ) for the amino-acid side chain analogs , as observed previously 21 ., The same effect persists for the PTM compounds , with a RMSE from experimental HFEs of 15 kJ/mol ( Figure S1 ) ., As the GROMOS 45a3 parameter set was not parameterized to match experimental HFEs for polar compounds , such level of deviation was to be expected ., Due to a lack of pertinent experimental data , seven parameterized PTMs ( carboxylysine , homocitrulline , citrulline , S-carbamoyl-cysteine , S-nitrosocysteine , 2-oxo-histidine and pyruvic acid ) have remained unrepresented in the validation set , and therefore unverified in terms of reproducing experimental HFEs ., To further assess the quality of the parameters for these compounds , we have compared them to those obtained by the Automated Topology Builder 26 , a widely used online service for automated parameterization of small molecules compatible with the GROMOS 54a7 force field ., While manually curated approaches are arguably superior to automated ones , it is reassuring to see that the two sets of parameters match closely ., For example , we have observed close agreement between the sets of partial charges obtained using the two methods for these seven compounds , with a Pearson correlation coefficient R of 0 . 93 and an overall RMSD of 0 . 2 e− ., As an application of the newly developed PTM parameters , we focus on the changes in several key physico-chemical properties of amino acids introduced by PTMs ., Interestingly , the majority of post-translationally modified amino acids are larger in size than their native counterparts , with more than 85% of PTMs increasing the molecular weight and more than 80% of PTMs increasing the solvent accessible surface area ( SASA ) of the affected residues ( Table S3 ) as calculated on energy-minimized ( using the GROMOS 54a7 parameter set ) configurations of PTMs and canonical amino acids ., What is more , PTMs introduce significant changes in the electrostatic properties of target residues as illustrated in the case of net charge and dipole moment ( Table S3 ) ., For example , 42% of all PTMs studied here undergo a charge change of 1 e− or more in absolute value , with 88% of such changes resulting in a more negatively charged species ., Moreover , the average absolute value of the change in dipole moment upon PTM equals 1 . 7 Debye , which is comparable in magnitude to the average dipole moment of 2 . 7 Debye or its standard deviation of 1 . 9 Debye as calculated in both cases over all unmodified residues using GROMOS 54a7 parameters and energy-minimized configurations ., Finally , given the general importance of hydrophobicity in various biological processes , it is critical to understand in a quantitative manner how PTMs modulate the hydrophobicity of target amino acids ., To address this question , we have used TI and GROMOS 54a7 parameters to calculate HFEs of all parameterized PTMs in neutral protonation states , since the available experimental data is insufficient for such an analysis ., Our results show that methylation and carbonylation modifications increase HFEs on average by 18 . 6 kJ/mol and 20 . 5 kJ/mol , respectively , while hydroxylation modifications exhibit an opposite effect and decrease HFEs by on average 25 . 1 kJ/mol ( Figure 3a ) ., These changes are extremely relevant if one considers the fact that the two central quartiles of the distribution of HFEs for canonical amino acids span the range from approximately −40 kJ/mol to −20 kJ/mol ( Figure 3a ) ., Furthermore , the most extreme cases , i . e . symmetric di-methylation of arginine ( ΔHFE\u200a=\u200a46 . 2 kJ/mol ) and di-hydroxylation of phenylalanine ( ΔHFE\u200a=\u200a−60 . 3 kJ/mol ) are comparable in absolute values to the total span of the canonical amino acid HFEs ( −49 . 4 kJ/mol to −3 . 2 kJ/mol , Figure 3a ) ., In other words , the effect of some PTMs on the HFEs of target amino acids is as large as the difference which would arise by mutating the most hydrophobic to the most hydrophilic canonical amino acid or vice versa ., While some of these effects agree well with what one would qualitatively expect , for a number of PTMs our results are the first to provide a quantitative framework for such an analysis ., As both calculation and experimental measurement of HFEs are limited to neutral compounds only , the above analysis does not take into account charged modifications such as phosphorylation ., To address this , we have used the molecular hydrophobicity potential ( MHP ) 27 approach to estimate hydrophobicity of all parameterized PTMs using their protonation states at physiological pH . MHP values are semi-empirical estimates of logP , a given compounds partition coefficient between water and the non-polar solvent octanol and are widely used in computational drug design 28 , 29 ., Similarly to the HFEs analysis , MHP calculations show that carbonylation and methylation are hydrophobicity-increasing modifications ( Figure 3b ) , in contrast to phosphorylation and hydroxylation , which are hydrophilicity-increasing modifications ., Finally , this analysis shows that PTMs can drastically change hydrophobic/hydrophilic properties of affected residues , e . g . arginine carbonylation shifts a highly hydrophilic to a highly hydrophobic residue , while cysteine oxidation does exactly the opposite ( Figure 3c ) ., By changing the chemical nature of affected residues , PTMs frequently completely alter their physico-chemical properties such as hydrophobicity , a feature with potentially far-reaching biological implications 11 , 12 , 30 ., Despite the importance of understanding PTMs at the molecular level , MD simulations of post-translationally modified proteins lag significantly behind the studies of unmodified proteins , and this seems primarily due to a general lack of suitable computational tools and simulation parameters for treating PTMs ., This study is to the best of our knowledge the first-ever effort to develop force-field parameters for the large majority of known PTMs in a systematic fashion ., We have generated GROMOS force field ( 45a3 and 54a7 ) parameters for over 250 different enzymatic and non-enzymatic PTMs , spanning a wide range of modification types with a close to complete coverage of experimentally verified PTMs ( Figure 1 ) ., Since GROMOS 54a7 force field parameters were fitted to reproduce experimental HFEs , we have tested the quality of the PTM parameters , obtained by manually curating the parameters of different groups mostly in analogy to canonical amino acids , by comparing the calculated HFEs against the experimental values ., The newly generated parameters compatible with the GROMOS 54a7 parameter set reproduce experimental HFEs almost equally well as the original ones ( Table 1 and Figure 2 ) ., Overall , only a few parameterized PTMs have not been directly validated against experimental HFEs due to a lack of experimentally available data ., In those cases , however , good matching with the parameters obtained using an orthogonal , fully automated approach 26 lends support to the general validity of the reported parameters ., However , one should emphasize that the full range of validity of the presented parameters could and should be delineated only by directly comparing MD simulations of different post-translationally modified proteins in biologically relevant contexts with relevant experimental data ., To date , PTMs in MD simulations have been treated in separate studies using different procedures and force fields , typically focusing on a single modification at a time 11 , 13 , 16 ., Additionally , there are some available tools for automated generation of parameters ( e . g . the AMBER 31 feature antechamber and online tools SwissParam 32 , PRODRG 33 , ATB 26 and q4md-forcefieldtools 34 ) , however , envisioned for small molecules rather than protein PTMs ., The parameters reported herein have comparative advantage over these sources along three principal directions ., First , we provide exclusively human curated and validated PTM force-field parameters , which are mutually fully consistent as well as being consistent with canonical amino acids ., Second , we provide PTM parameters in both GROMOS 35 and GROMACS 36 format , widely used MD simulation packages ( supporting GROMOS version 11 and GROMACS versions 3 . × and newer ) , suitable for immediate simulation of modified proteins without any additional work required ., This should be contrasted with the above tools that provide parameters for isolated compounds only ., Finally , in combination with a publicly available online tool for introducing PTMs of choice to a user-supplied protein 3D structure ( Vienna-PTM server , http://vienna-ptm . univie . ac . at ) 37 , we provide a comprehensive , user-friendly toolkit for studying PTMs using MD simulations ., During their lifecycle in the cell , almost all proteins undergo one or more different PTMs affecting their structure , dynamics and interaction networks and , subsequently , their function through direct alteration of chemical and physico-chemical properties of target residues ( Figure 3 ) ., The force field parameters presented here , together with the Vienna-PTM webserver , provide a systematic framework required to study the effects of PTMs using MD simulations ., As a first step in this direction , we have here compared the hydrophobicity-related variables ( HFEs and MFP values ) of native and modified amino acids and quantitatively showed that PTMs can have an extremely strong , biologically significant effect in this context ., It has already been documented that some PTMs exert their biological effect through a general modification of the hydrophobicity of their targets ., For example , lysine trimethylation is known to directly affect the binding of retinoic acid receptors , which regulate genes involved in growth , differentiation and apoptosis , to their partners via an increase in site-specific hydrophobicity 38 ., Moreover , acetylated and methylated lysine residues in histones , i . e . , some of the key components of the histone code , are recognized by the hydrophobic binding pockets of bromo- and chromo-domains based on the difference in hydrophobicity between the modified and unmodified lysines 39 ., Furthermore , we have recently shown that carbonylation , which affects lysine , arginine , proline and threonine residues , drastically increases local propensity for aggregation in proteins by affecting the hydrophobicity of the modified sites 11 ., While other , more specific effects of PTMs on the structure , dynamics and interaction profile of target proteins are certainly important , a major change in hydrophobicity , net charge , isoelectric point or any other general physico-chemical property caused by a PTM at a given site could certainly have major biological repercussions ., We believe that our present study will provide a solid foundation for exploring such timely and important issues in the future ., However , this is only one possible application of the PTM force-field parameters reported herein ., From direct MD simulations to biomolecular structure refinement to computational free energy estimation and drug design , these parameters expand the range of MD methodLology to a large class of biomolecular systems of paramount importance ., It is our hope that this advance will play a catalytic role in bringing together realistic cell biology , dominated by PTMs , and the quantitative , reductionist power of structural biology and chemistry , as embodied in the MD method , and help shed light on a broad spectrum of important biological questions at the microscopic level ., One of the aims of the GROMOS force fields is to allow for the transfer of parameters between chemically similar groups in different compounds ., Accordingly , we have derived GROMOS 45a3 and 54a7 force field parameters describing 110 post-translationally modified amino acids and protein termini ( Table S1 ) by either novel parameterization or direct transfer from or analogy to already parameterized compounds including amino acids , nitrogenous bases and other small molecules according to the following principles and rationales ., General principles: Modification type-specific principles: We include detailed descriptions of parameter choices as comments in Dataset S1 and Dataset S2 ., We have used the thermodynamic integration approach 25 , a widely used computational method based on MD simulations , to calculate hydration free energies ( HFEs ) of neutral forms of small-molecule analogs of 14 amino-acid side chains ( the same set as in Oostenbrink et al . 21 ) , compounds from the validation set and side chain analogs of all parameterized PTMs with a charge neutral protonation state ., Non-bonded ( van der Waals and Coulomb ) interactions , coupled to a parameter λ , were scaled down to zero in a stepwise manner in vacuum and water ., Free energy changes of these processes were calculated as the integral of the ensemble average of the derivative of the total Hamiltonian of the system with respect to λ , over the interval from λ\u200a=\u200a0 to λ\u200a=\u200a1 ., For vacuum calculations , three independent simulations , each 5 ns long , were run at 21 equally spaced λ-points with the temperature kept at 500 K and additional random kicks introduced by Langevin dynamics integration method 44 , in order to avoid convergence problems due to inefficient sampling of the conformational space ., Water simulations were run in five independent copies , each 0 . 5 ns long , at 21 equally spaced λ-points , together with 10 additional λ-points placed in the regions of the least smoothness of the integrated curve , using SPC explicit water 45 , a reaction field electrostatic scheme with a cutoff of rc\u200a=\u200a1 . 4 nm and the dielectric constant of εrf\u200a=\u200a65 and a Berendsen thermostat and barostat keeping the temperature and pressure at 300 K ( τT\u200a=\u200a0 . 05 ps ) and 1 bar ( τp\u200a=\u200a1 ps and compressibility\u200a=\u200a4 . 5×10−5 bar−1 ) 46 ., A soft-core formalism 47 was used to avoid singularities of the potential energy ., The aforementioned integrals were evaluated by the generalized Simpsons rule for non-equidistant nodes using the averages over the independent simulations at each λ-point ., HFEs were calculated as the difference between the change in free energy upon the removal of non-bonded interactions calculated in vacuum and calculated in water . | Introduction, Results, Discussion, Methods | By directly affecting structure , dynamics and interaction networks of their targets , post-translational modifications ( PTMs ) of proteins play a key role in different cellular processes ranging from enzymatic activation to regulation of signal transduction to cell-cycle control ., Despite the great importance of understanding how PTMs affect proteins at the atomistic level , a systematic framework for treating post-translationally modified amino acids by molecular dynamics ( MD ) simulations , a premier high-resolution computational biology tool , has never been developed ., Here , we report and validate force field parameters ( GROMOS 45a3 and 54a7 ) required to run and analyze MD simulations of more than 250 different types of enzymatic and non-enzymatic PTMs ., The newly developed GROMOS 54a7 parameters in particular exhibit near chemical accuracy in matching experimentally measured hydration free energies ( RMSE\u200a=\u200a4 . 2 kJ/mol over the validation set ) ., Using this tool , we quantitatively show that the majority of PTMs greatly alter the hydrophobicity and other physico-chemical properties of target amino acids , with the extent of change in many cases being comparable to the complete range spanned by native amino acids . | Post-translational modifications , i . e . chemical changes of protein amino acids , play a key role in different cellular processes , ranging from enzymatic activation to transcription and translation regulation to disease development and aging ., However , our understanding of their effects on protein structure , dynamics and interaction networks at the atomistic level is still largely incomplete ., In particular , molecular dynamics simulations , despite their power to provide a high-resolution insight into biomolecular function and underlying mechanisms , have been limited to unmodified , native proteins due to a surprising deficiency of suitable tools and systematically developed parameters for treating modified proteins ., To fill this gap , we develop and validate force field parameters , an essential part of the molecular dynamics method , for more than 250 different types of enzymatic and non-enzymatic post-translational modifications ., Additionally , using this tool , we quantitatively show that microscopic properties of target amino acids , such as hydrophobicity , are greatly affected by the majority of modifications ., The parameters presented in this study greatly expand the range of applicability of computational methods , and in particular molecular dynamics simulations , to a large set of new systems with utmost biological and biomedical importance . | biomacromolecule-ligand interactions, molecular dynamics, classical mechanics, statistical mechanics, macromolecular assemblies, molecular mechanics, semi-empirical methods, newtons laws of motion, protein folding, protein structure, biophysics simulations, biophysics theory, chemistry, biology, biophysics, macromolecular complex analysis, physics, biochemical simulations, protein chemistry, computer science, computer modeling, biophysic al simulations, computational chemistry, computational biology, macromolecular structure analysis | null |
journal.pcbi.1000511 | 2,009 | Estimating the Continuous-Time Dynamics of Energy and Fat Metabolism in Mice | Mouse models of obesity have become critically important research tools for discovering molecular mechanisms of body weight regulation ., But understanding these mechanisms in the context of whole-body physiology requires knowledge of food intake , energy output , and fuel selection 1 ., Furthermore , measurements made at an isolated time point cannot explain why body weight has its present value since body weight is determined by the past history of energy and macronutrient imbalance 2 ., While food intake and body weight changes can be measured frequently over several weeks ( the relevant time scale for mice ) , correspondingly frequent measurements of energy output and fuel selection are not currently feasible ., Expensive indirect calorimetry systems can be used to measure energy expenditure and respiratory exchange over periods of a few days and most systems require removing mice from their normal environment which can alter their behavior 3 ., Alternatively , the doubly labeled water method can give an estimate of average energy expenditure , but this method requires specialized equipment for sample analysis as well as prior knowledge of fuel selection as measured by the respiratory quotient ( RQ ) 4 ., Furthermore , significant quantities of blood need to be collected which could impact the behavior of the mouse and makes repeat measurements untenable 4 ., Here , we present a mathematical method that quantitatively relates food intake , body weight and body fat to calculate the dynamic changes of energy output and net fat oxidation rates during the development of obesity and weight loss in male C57BL/6 mice ., The mathematical model is based on the law of energy conservation , makes very few assumptions , and provides the first continuous-time estimates of energy output and fuel selection over periods lasting many weeks ., Our methodology also revealed the relationship between diet , fuel selection , and body composition change in male C57BL/6 mice by identifying a time-invariant curve relating body fat and fat-free masses ., As previously described 5 , male C57BL/6 mice were given ad libitum access to standard chow ( C ) , high fat diet ( HF ) , or high fat diet plus liquid Ensure ( EN ) for 19 weeks , while some mice were fed the high fat or the high fat plus Ensure for 7 weeks before being switched back to chow for the remaining 12 weeks ( HF-C and EN-C , respectively ) ., Figure 1A shows the body weight changes of the various groups during the development of obesity on the HF and EN diets as well as the weight loss and persistent obesity of the HF-C and EN-C groups following a switch back to the chow diet at 7 weeks ( error bars have been omitted for clarity ) ., A single curve was able to describe the adjusted fat-free mass as a function of body fat mass for all groups at all time points ( Figure 1B ) and is analogous to the curve discovered by Forbes describing human body composition change 6 ., Our mathematical model used this fitted curve along with the body weight data to compute the body fat mass changes ( Figure 1C ) ., Without adjusting any parameters , the model also accurately predicted the fat mass changes measured in a separate experiment with high-fat feeding of C57BL/6 mice followed by a switch to chow after 20 weeks ( Figure 1D ) ., Our model calculated the first continuous-time estimates of the energy output dynamics underlying the observed body weight changes ( Figure 2A ) ., The 95% confidence interval surrounding the calculated energy output rates resulted primarily from variability of the measured energy intake rate ( individual data points are depicted along with the average black curve used for each group ) but also included the effect of body composition variability ( Figure 1B ) ., The HF and HF-C groups had a transient decrease of energy output at the onset of high fat feeding at 0 days ., In contrast , the EN and EN-C groups did not show a significant transient reduction of energy output at the onset of the high energy diet ., Energy output gradually increased with weight gain in all of the groups ., Following the return to the chow diet , the HF-C group had a transient increase of energy output which was not seen in the EN-C group ., Note that these transient changes account for significant fractions of the overall energy imbalances and would be difficult to detect using indirect calorimetry or doubly labeled water methods ., Net fat oxidation rates increased sharply at the onset of high fat feeding in the HF and HF-C groups , but did not rise sufficiently to match the increase of fat intake ( Figure 2B ) ., Interestingly , despite similar increases of fat intake in the EN and EN-C groups compared with the HF and HF-C groups , the initial increase of net fat oxidation was significantly attenuated ., Net fat oxidation gradually increased in all the groups as body weight increased ., Following the switch to chow , there was a transient increase of net fat oxidation in both HF-C and EN-C groups before falling to match the low level of fat intake after a few weeks ., A useful measure of fuel selection is the respiratory quotient , RQ , where a value of 0 . 7 reflects a state of pure fat oxidation whereas a value of 1 . 0 reflects a state of pure carbohydrate oxidation and intermediate values represent a fuel selection mixture ( see Methods ) ., The estimated 24 hour RQ ( Figure 3 ) demonstrates the impact of both diet and body composition on fuel selection ., The HF group had an immediate decrease of RQ due to the diet followed by a slow progressive decrease as body fat gradually increased ., The EN group showed little initial change of RQ which then progressively decreased to an intermediate value ., After switching to the chow diet , the HF-C group had a rapid increase of RQ towards that of the C group whereas the EN-C group had a transient decrease of RQ before increasing towards the C group ., The mouse has become the most popular organism for investigating molecular mechanisms of body weight regulation ., But understanding the physiological context by which a molecule exerts its effect on body weight requires knowledge of energy intake , energy expenditure , and fuel selection ., Our simple mathematical method calculates the dynamics of energy output and fuel selection over extended time periods using longitudinal measurements of body weight , food intake , and body composition ., We showed that our method can detect both transient changes of energy expenditure and net fat oxidation rates as well as longer timescale changes found with weight gain and loss ., Similar methodology has been previously developed by our group to relate human body-composition changes with dynamic adaptations of fuel selection in both adults 7 and infants 8 ., The method is especially well-suited for mouse studies because food intake can be accurately measured over the extended time periods required to measure significant changes of body weight and body fat ., While we have applied the model to data averaged within groups of mice , it would be also interesting to examine individual mouse trajectories as a way of investigating inter-individual variability ., Our equations extract information about energy output that is already present in the body weight and food intake data ., Other than the law of energy conservation , the only assumption was that the relationship between changes of body fat and fat-free mass were described by a well-defined function in accordance with the Forbes theory of body composition change 6 ., This assumption was confirmed in the present study for mature male C57BL/6 mice ( Figure 1B ) and we hypothesize that genetic manipulations can alter the shape of this function ., However , once the function has been determined we showed that it provided accurate estimates of body fat changes in an independent feeding experiment using body weight measurements alone ( Figure 1D ) ., Therefore , knowledge of the Forbes curve for a given mouse model eliminates the need for frequent body composition measurements ., To estimate the net fat oxidation rate and RQ , an additional assumption regarding carbohydrate balance was required ( see Methods ) ., We found that the Forbes function ( Figure 1B ) determined the relationship between food intake , body composition change , and net fat oxidation rate 7 ., While both humans and mice have Forbes functions that increase with body fat mass , the concavity of the curves is opposite 6 ., Therefore , great caution should be exercised when extrapolating fuel selection results in mice to predict human responses ., The physiological reason for this difference is presently unclear ., Our research group is actively engaged in developing detailed models of the complex interactions between carbohydrate , fat , and protein metabolism in humans 9 to better understand the relationship between the physiological drivers of fuel selection and the Forbes body composition curve ., We plan to develop similar models in mice to help understand these relationships and the differences between the species ., In contrast to our method , currently available techniques for estimating energy expenditure are expensive , involve a plethora of assumptions , and can impact the behavior of the mice 3 , 4 ., These factors make it common to find reports of energy expenditure rates that are quantitatively inconsistent with the measured energy intake and body weight changes found in mice that were not subjected to these procedures ., As an illustrative example , consider the recent publication by Funato et al . where the energy intake rate of the wild type mice was at least 17 kcal/d and the energy expenditure measured by indirect calorimetry was less than 5 kcal/hr/ ( kg BW ) 0 . 75 ., This translates to an absolute expenditure rate of less than 10 . 7 kcal/d for a mouse that was at most 40 grams at the time of measurement 10 ., Such a large positive energy balance would translate to a rate of weight change of at least 4 . 7 g/week ( if all excess energy was deposited as fat ) versus the measured weight gain which was less than 1 g/week ., The purpose of this example is not to criticize the work of Funato et al . , but rather to highlight how even careful indirect calorimetry and food intake measurements can lead to estimates of energy imbalance that are inconsistent with the weight gain measurements ., Our own attempt to use indirect calorimetry to validate the model predictions of energy expenditure and fuel selection highlighted two important issues ., First , the mice that were consuming the high energy diets lost significant amounts of weight when moved to the indirect calorimetry cages indicating that their behavior was not representative of the mice not subjected to the procedure ., Second , the measured energy expenditure rates were unrealistically high compared to the model predictions for all groups of mice ., In fact , the measured energy expenditure rate was higher than the measured energy intake in the chow-fed mice that did not lose weight ( an impossibility ) and greatly exceeded the expenditure required to explain the weight loss in the mice fed the high energy diets ., These discrepancies led us to diagnose a technical problem with the indirect calorimetry equipment ., Thus , we were unable to validate the model estimates of energy expenditure and fuel selection ., The field of farm animal nutrition has a long and rich history of using mathematical modeling to analyze animal growth and identify nutritional factors that potentially limit growth rate 11–14 ., The simplest models describe the efficiencies of various diets in their ability to deposit body energy , often specified in terms of body fat and protein 11 , 13 , 14 ., Inputs to such models include energy intake , body weight , and the rates of body fat and protein deposition ., The model outputs include the efficiencies of protein and fat deposition as well as the so-called maintenance energy requirement which is roughly defined as the energy intake required when the animal is not growing ., An alternative representation uses energy intake , body weight , total energy expenditure ( by calorimetry methods ) , and protein deposition rate ( via nitrogen balance ) as model inputs and predicts the maintenance energy requirement , fat deposition rate , and body protein and fat deposition efficiencies ., At the next level of complexity , animal growth models prescribe an energy partitioning rule that specifies how body protein will accumulate for a given food intake rate as a function of body weight , age , or body protein ., Energy partitioning rules are often complex 12 , 13 , but can be thought of as similar to the Forbes function that specifies how energy imbalances are partitioned between body fat and fat-free mass ., A significant difference is that our approach is applied to mature mice whose overall growth rate was minimal despite their ability to gain and lose fat-free mass in response to the various diets ., Once the partitioning rule is specified , the outputs of animal growth models include body fat mass , maintenance energy requirement , as well as body fat and protein deposition efficiencies given the food intake and body weight as model inputs ., In contrast , our model outputs are body fat mass , fuel selection , and total energy expenditure which are more relevant for mouse obesity studies and avoids the known problem of arbitrarily distributing total energy expenditure between tissue deposition costs versus maintenance energy requirements 11 , 14–16 ., Animal growth models have often used power-law functions of body weight to model the maintenance energy requirements that were previously calculated using the above methods ., Once specified , the model of maintenance energy requirements can be used along with the calculated efficiencies of protein and fat deposition and the energy partitioning rule to predict body weight and body fat change as a function of the food intake 11 , 14 ., We are presently developing a model of total energy expenditure in mice that will allow prediction of body weight and composition changes as well as fuel selection when food intake is the only input to the model ., A weakness of our methodology is that it does not distinguish the various components of energy output including resting metabolic rate , thermic effect of feeding , adaptive thermogenesis , physical activity , or any changes of energy excreted in urine and feces that are unaccounted for by the estimates of diet metabolizability ., Furthermore , the method does not operate on a within-day time scale and therefore cannot address changes between day versus night or transitions between fed and fasted states ., Indirect calorimetry is required to address these issues and would provide important information for the interpretation of our calculated longer-term estimates of energy output and fuel selection ., We believe that the combination of our continuous-time methodology with indirect calorimetry measurements at judiciously chosen time points can be applied to various mouse models of obesity as a powerful tool for characterizing the metabolic dynamics underlying experimentally observed body weight changes ., We certify that all applicable institutional and governmental regulations concerning the ethical use of animals were followed during this research ., All procedures were approved by the National Institute of Diabetes and Digestive and Kidney Diseases Animal Care and Use Committee ., Full details of the experiment were previously described 5 ., Briefly , forty seven 3 month old male C57BL/6 mice weighing 25 . 9±1 . 2 g ( The Jackson Laboratory , Maine ) were housed individually and randomly assigned to five weight-matched groups: 1 ) C group ( N\u200a=\u200a12 ) continued on the chow diet; 2 ) HF group ( N\u200a=\u200a12 ) on a high fat diet ( F3282; Bio-Serv Inc . , NJ; 5 . 45 kcal/g with 14% energy derived from protein , 59% from fat , and 27% from carbohydrate ) ; 3 ) EN group ( N\u200a=\u200a11 ) on the high fat diet plus liquid Ensure ( Abbott Laboratories , Kent , UK ) , which had an energy density of 1 . 06 kcal/ml with 14% of energy derived from protein , 22% from fat , and 64% from carbohydrate; 4 ) HF-C group ( N\u200a=\u200a6 ) switched from high fat to chow after 7 weeks;, 5 ) EN-C group ( N\u200a=\u200a6 ) switched from high fat plus Ensure to chow after 7 weeks ., All animals received free access to water and food throughout the study ., The high fat diet was provided using Rodent CAFÉTM feeders ( OYC International , Inc . , MA ) , and liquid Ensure was provided in a 30-ml bottle with a rodent sip tube ( Unifab Co . , MI ) and liquid intake was measured every day ., Solid food intake was corrected for any visible spillage and was measured every day for the high fat diet and every other day for the chow diet using a balance with a precision of 0 . 01 g ( Ohaus model SP402 ) ., Body composition was measured once per week using 1H NMR spectroscopy ( EchoMRI 3-in-1 , Echo Medical Systems LTD , Houston , TX ) after body weight was determined ., We begin with the law of energy conservation , also known as the energy balance equation: ( 1 ) where F is the body fat mass , FFM is the fat-free mass defined as the measured body weight , W , minus the fat mass , and and are the energy densities for changes in fat and fat-free masses , respectively 17 ., IT is the total metabolizable energy intake rate corrected for spillage , and E is the energy output rate ., We distinguish the energy output rate from the energy expenditure rate since we did not measure any changes of energy excreted in urine or feces ., In other words , if the metabolizable energy content of each diet is constant then our calculation of the energy output is equivalent to energy expenditure ., Analogous to the Forbes theory of human body composition change 6 , we hypothesized that there is a well-defined , time-invariant function , α , that describes the relationship between changes of FFM and F in male C57BL/6 mice: ( 2 ) Once the function α is specified , equation ( 1 ) can be solved for the energy output rate as a function of the measured energy intake rate and the rate of body weight change as follows: ( 3 ) The fat mass is given by solving the following differential equation: ( 4 ) Alternatively , if the Forbes assumption does not apply for a given mouse model ( for example , during periods of significant growth ) , a curve could be directly fit to the measured fat mass time series data and used in place of equation 4 ., While this procedure would give equivalent results , it necessitates frequent body composition measurements for every experiment ., Note that very few assumptions were made in the development of our equations to estimate energy output ., All of the above equations were derived from the law of energy conservation ( 1 ) and the only assumption was that there exits a well-defined Forbes relationship , α , relating changes of body fat and fat-free masses – an assumption that was directly confirmed by comparison to measured body composition data ., Since we are also interested in fuel selection , we must consider the fates of dietary macronutrients including their oxidation rates , storage in the body , as well as major inter-conversion fluxes where carbohydrate can be converted to fat ( i . e . , de novo lipogenesis ) and amino acids can be converted to the carbohydrate glucose ( i . e . , gluconeogenesis ) ., The following macronutrient balance equations represent these changes: ( 5 ) where P is body protein , G is glycogen , GNG is the gluconeogenic rate , DNL is the de novo lipogenic rate , and IF , IP and IC are the intake rates of dietary fat , protein and carbohydrate , respectively ., The oxidation rates of fat , protein , and carbohydrate ( FatOx , ProtOx , and CarbOx , respectively ) sum to the total energy output , E . To simplify the macronutrient balance equations , we note that glycogen stores are small , especially when compared with daily carbohydrate intake rates ., For example , humans have a glycogen pool size of about 500 g which is equivalent to the typical amount of carbohydrate consumed over ∼2 days and equilibrates on a time scale of ∼1 day 9 , 18 ., The equilibration time is likely even more rapid in mice since they typically consume carbohydrate at a rate of ∼2 g/d and their glycogen stores are probably less than 0 . 6 g ( assuming maximal glycogen pool sizes of 8% of liver weight and 0 . 6% of muscle weight as observed in rats 19 and assuming that mouse liver is less than 5 g and muscle is less than 30 g 5 ) ., Thus , over the time-scale of interest the system is in a state of average carbohydrate balance: ( 6 ) Therefore , ( 7 ) If we define the net fat oxidation rate as follows: ( 8 ) then the equation for body protein change becomes: ( 9 ) Finally , we assume that FFM is proportional to body protein such that ( 10 ) Therefore , we have a simple a two-compartment macronutrient partitioning model which we have previously shown has an invariant manifold as its attractor 20: ( 11 ) From equations 4 and 11 , the net fat oxidation rate can be written as a function of the measured fat intake rate and the rate of body weight change: ( 12 ) Note that the carbohydrate balance assumption was only required to calculate the estimate of net fat oxidation , but was not required to calculate the energy output rate ., The shape of the Forbes curve has direct implications for how fat oxidation rate is related to changes of body fat ., This can be seen by calculating the partial derivative of the net fat oxidation rate with respect to body fat: ( 13 ) Interestingly , this quantity has opposite sign in humans versus mice ., Thus , great care must be taken when fuel selection measurements in mice are extrapolated to humans ., The respiratory quotient , RQ , is the carbon dioxide production rate divided by the oxygen consumption rate and was approximated by: ( 14 ) This approximation assumes a negligible contribution of de novo lipogenesis and gluconeogenesis which is reasonable since these fluxes act to offset each other with respect to CO2 production ., Since the carbohydrate oxidation rate is approximately equal to the carbohydrate intake rate on long time scales , the calculated RQ may have slight inaccuracies during rapid transitions immediately after diet switches , but will be reasonably accurate thereafter ., To apply our mathematical model to data from our mouse experiment , food intake measurements were averaged over each diet period and we assumed stepwise transitions immediately after each diet switch followed by a smooth approach to the average intake of the final diet period ., These curves are depicted as solid black lines in Figure 2 and represent the average of the individual intakes shown by the data points ., Body weight measurements for the C , HF , and EN groups of mice were fit using third order polynomial functions of time , as depicted by the solid curves in Figure 1A ., Following the diet switch in the HF-C and EN-C groups , the body weight curves were fit to exponential functions ., The rates of change of body weight were then calculated by computing derivatives of the fitted curves ., Other than their ability to adequately describe the model input data , the precise mathematical form of these curves is not important ., The Forbes body composition function , α , was fit to an exponential function of the body fat mass as shown in Figure 1B ., Specifically , we assumed that the individual data points for fat-free mass versus body fat for each group of mice were described by the following equation: ( 15 ) The Forbes function , α , is then given by: ( 16 ) Since the intercept parameter , b , does not influence the Forbes function , we adjusted the FFM data for each group by subtracting the difference between the calculated intercept parameter for each group and its average value across groups ., We then simultaneously fit the adjusted FFM data from all groups to arrive at our final Forbes function used for all of the groups ., The parameter values for the Forbes body composition function were determined via a Markov Chain Monte Carlo ( MCMC ) method 21 implemented in MATLAB ( version R2008a; MathWorks Inc , Natick , MA ) ., To approximate the posterior distribution of the parameters in the Forbes function ( equation 16 ) , we drew 100 , 000 MCMC samples of parameter values , of which the first 30000 were discarded as burn-in period; afterwards one fifth of the rounds were retained ., Parameter sets were drawn from a proposal density that were normally distributed and centered on the previous value ., The variance of the proposal density was tuned for an average acceptance rate of ∼0 . 25 during the burn-in period ., The convergence of the chain was assessed both by visual inspection of the trace plots for all the parameters and through the Geweke test 22 ., At each sampling , the probability of accepting the new parameter set given current parameter set was where r is the Metropolis ratio 21 ., The posterior distribution of energy output ( equation 3 ) was calculated from the joint distribution of the parameters in the function and the energy intake in each group of the animals assuming no correlation existed between the two distributions ., The energy intake in each group of animals was normally distributed with a standard error of 0 . 39 , 0 . 39 , 0 . 41 , 0 . 55 , and 0 . 55 Kcal/d for the C , HF , EN , F-C , and EN-C groups , respectively ., The 95% confidence intervals of the predicted energy output were obtained by calculating the 2 . 5th and 97 . 5th percentiles of the posterior distribution of energy output . | Introduction, Results, Discussion, Methods | The mouse has become the most popular organism for investigating molecular mechanisms of body weight regulation ., But understanding the physiological context by which a molecule exerts its effect on body weight requires knowledge of energy intake , energy expenditure , and fuel selection ., Furthermore , measurements of these variables made at an isolated time point cannot explain why body weight has its present value since body weight is determined by the past history of energy and macronutrient imbalance ., While food intake and body weight changes can be frequently measured over several weeks ( the relevant time scale for mice ) , correspondingly frequent measurements of energy expenditure and fuel selection are not currently feasible ., To address this issue , we developed a mathematical method based on the law of energy conservation that uses the measured time course of body weight and food intake to estimate the underlying continuous-time dynamics of energy output and net fat oxidation ., We applied our methodology to male C57BL/6 mice consuming various ad libitum diets during weight gain and loss over several weeks and present the first continuous-time estimates of energy output and net fat oxidation rates underlying the observed body composition changes ., We show that transient energy and fat imbalances in the first several days following a diet switch can account for a significant fraction of the total body weight change ., We also discovered a time-invariant curve relating body fat and fat-free masses in male C57BL/6 mice , and the shape of this curve determines how diet , fuel selection , and body composition are interrelated . | The unrelenting obesity epidemic has resulted in intensive basic scientific investigation into the molecular mechanisms of body weight regulation—with the mouse being the organism of choice for such studies ., We know that any mechanism of body weight regulation must exert its effect by influencing food intake , energy output , fuel selection , or some combination of these factors over extended time scales ( ∼weeks for mice ) ., While food intake and body weight can be frequently measured in mice , current methods prohibit corresponding measurements of energy output or fuel selection on such long time scales ., We address this deficiency by developing a mathematical method that quantitatively relates measurements of food intake , body weight and body fat to calculate the dynamic changes of energy output and net fat oxidation rates during the development of obesity and weight loss in male C57BL/6 mice ., The mathematical model is based on the law of energy conservation , makes very few assumptions , and provides the first continuous-time estimates of energy output and fuel selection over periods lasting many weeks ., Application of our methodology to various mouse models of obesity will improve our understanding of body weight regulation by placing molecular mechanisms in their whole-body physiological context . | mathematics, diabetes and endocrinology/obesity, nutrition/obesity, physiology/integrative physiology, biochemistry/theory and simulation | null |
journal.pcbi.1000948 | 2,010 | Critical Dynamics in the Evolution of Stochastic Strategies for the Iterated Prisoners Dilemma | The evolution of cooperation is difficult to understand within Darwinian theory 1–3 ., Indeed , cooperation is intrinsically vulnerable to exploitation because evolution rewards individual success , while any detrimental long-term effects for the group are secondary 4 , 5 ., The tension between the short-term benefits of defection and the long-term benefits of cooperation has been studied using the Prisoners Dilemma as a paradigm of social conflicts 3 , 6–9 ., Previous work has shown that cooperation can only emerge in the presence of different enabling mechanisms ., The main ones are direct reciprocity 6 , 10 ( which can emerge when players play against each other repeatedly ) , spatial reciprocity 7 , which is ensured if players only play neighbors on a regular grid ( or more generally , on arbitrary graphs , giving rise to “network reciprocity” 11 ) , tag-based selection 12 ( where players can recognize each other using some observable trait ) , kin selection 13 , indirect reciprocity 14 , 15 ( where cooperative or altruistic acts increase a players reputation ) , or group selection 16 ., Social diversity , where either the payoffs or the neighborhoods vary from player to player 17 , 18 can also enhance cooperation , as can “active linking” 19 , 20 , where players differ in the rate at which they maintain interactions with other players ., Generally speaking , the co-evolution of strategies with the different enabling mechanisms can also increase cooperation 21 ., In all the discussed scenarios , a players strategy is such that they either cooperate or defect in a deterministic manner , sometimes conditionally on previous plays ., If a cooperating strategy accidentally defects ( or a defector accidentally cooperates ) the noise that is introduced in this manner can have a dramatic effect on the competition ., For example , among the ( deterministic ) strategies that take one previous move into account in order to decide how to play , the reciprocating strategy “TFT” ( Tit-for-Tat ) dominates 6 , but is outcompeted 22–24 by “Win-Stay-Lose-Shift” ( WSLS ) , which can correct for occasional mistakes 23 ., Experiments with bacteria 25 and social amoeba 26 indeed suggest that the decision to cooperate or defect ( in a general sense ) is stochastic , and moreover that these decisions are controlled by genetically-encoded probabilities that are evolvable 27 ., Rather than assuming that noisy decisions are either due to fuzziness in perception or lack of control over ones action 22 , here we allow these probabilities to be fine-tuned by adaptation in response to the environment ., We find that if a players stochastic decisions are under genetic control , then the level of uncertainty about an opponents next move ( given their previous encounter ) determines whether cooperation or defection evolves ., Because this uncertainty is a direct consequence of environmental conditions , we conclude that when decisions are based on previous interactions , these conditions alone are sufficient to explain the evolution of cooperation in populations ., Note that the stochasticity introduced by probabilistic play controlled by genes is fundamentally different from other random effects that can be introduced into evolutionary game dynamics , such as a probability to inherit a neighbors strategy 28 , or stochastically fluctuating payoffs 29 , 30 , because neither of them can evolve ., In its simplest form , PD players have only two play options: cooperate ( C ) or defect ( D ) ., Both players are awarded a payoff R for mutual cooperation and a payoff P for mutual defection ., Unequal moves award S to the cooperator and T to the defector ., In standard PD 6 , the values of the payoffs are constrained so that T>R>P>S and R> ( S+T ) /2 ., The first equation ensures that for a single round of play , defection is an evolutionary stable strategy 4 , while the second equation ensures that reciprocation of cooperation is favored over the trading of cooperative with defective moves ., In the repeated PD ( iterated PD or IPD ) that we study here , two players meet more than once , and can establish cooperation by means of direct reciprocity 6 ., In particular , we study exclusively the IPD with memory , that is , where players base their decision on previous plays ( except for the first move with a new opponent ) ., The term “memory” is not meant to imply that only higher organisms can engage in such strategies ., Rather , stochastic decisions can be based entirely on the levels of protein on a cells receptor , for example , and where these protein levels are the result of a cellular “decision” at an earlier time ., A simple example for such a stochastic decision in response to the decision of other cells is quorum sensing in bacteria ( see , e . g . , 31 ) ., We contend that the introduction of information exchange between players ( via conditional strategies ) is crucial for the evolution of cooperation ., We evolve strategies in spatially-structured and well-mixed finite populations , as it is known that the evolutionary dynamics depend on population structure as well as size ( small fitness differences are effectively neutral only in finite populations 32 ) ., Evolution experiments are carried out with populations on a regular 32×32 grid with wrapping boundary conditions , where the manner of replacement determines the population structure ., Players engage their eight closest neighbors exactly once every update ( playing one move ) , for 500 , 000 iterations ., At the end of each update , a proportion r ( the replacement rate ) of players is randomly eliminated using a Moran-like process 33 , 34 , establishing a finite probability of future encounters between players beyond the first 35 ., For spatially-structured populations , each player marked for death is replaced by an offspring of one of his neighbors , while for well-mixed populations the entire grid of players is considered for filling the empty position ., In both population types , replicating players are chosen in proportion to their fitness , defined as the accumulated score ., Scores are awarded according to the standard payoff matrix of Axelrod 35 throughout , with T\u200a=\u200a5 , R\u200a=\u200a3 , P\u200a=\u200a1 , S\u200a=\u200a0 . For memory-one strategies , each player is represented by a genotype ( strategy ) composed of five genes , four of which encode the conditional probabilities PXY representing the probability that a player will cooperate , given that his last historical play was X and his opponents response was Y , along with the unconditional probability PC to cooperate on the first move 24 ., Each population is seeded with the “random” genotype where each of the five probabilities is set to 0 . 5 ., At each replication event , genes are subject to a per-gene mutation rate μ , replacing that genes probability to cooperate with a uniformly distributed number between 0 and, 1 . For each evolutionary run , we record the genotype as well as phenotype ( play statistics πCC , πCD , πDC , and πDD , given by the fraction of that type of play among all plays ) for each organism on the line of descent ( LOD ) 36 ., The LOD is generated by randomly selecting a genotype at the end of each run and tracing back its ancestry to the seeding genotype ., Compared to the previously discovered deterministic memory-one strategies 37 , our genetic implementation leads to the evolution of novel and drastically different successful strategies , depending on mutation rate , replacement rate , and population structure ., None of the 32 deterministic strategies ever appear on the LOD , but instead , strategies evolve that are either cooperative or defective , depending on the experimental setting ., Using the LOD averaged over 80 runs ( see Fig . S1 ) , we can obtain a consensus genotype for the particular experiment by averaging all genotypes in the latter half of this average LOD , removing any influence from the starting conditions ( see Methods ) ., The consensus genotype for spatially-structured populations at low mutation and replacement rates is that of a cooperative strategy ( PC , PCC , PCD , PDC , PDD\u200a=\u200a0 . 647 , 0 . 989 , 0 . 234 , 0 . 318 , 0 . 448 ) , as is evident from a commitment to exchange C plays ( i . e . , PCC≈1 ) and a tendency to cooperate on the first move ., By having a low PCD probability this strategy maintains a low tolerance to opponent defection and displays an unwillingness to be exploited ., Maintaining a PDD value close to 0 . 5 with a slight bias towards defection , the consensus genotype expresses indifference in propagating defection but willingness to return to cooperation , a behavior not previously seen among stochastic strategies 24 ., When faced with defective play , the strategy will acquire a deficit in lifetime payoff , which can be offset by exploiting naïve cooperators ( and occasionally similar strategies ) as indicated by a low PDC probability ., Consensus strategies for cooperation in well-mixed populations , as well as defectors in both population structures ( that appear at high mutation and replacement rates ) are listed in Table S1 and described in Text S1 ., In order to monitor the evolution of strategies , we reduce strategy space by performing a principal component analysis ( PCA ) of the probabilities on the average LOD obtained from 80 runs at mutation rate μ\u200a=\u200a0 . 5% and replacement rate r\u200a=\u200a1% , and use these components to display the average trajectory at other mutation rates as well ., For the spatially-structured population the first two principal components explain 83% and 10% of the variance , respectively ( see Methods ) ., Within the two-dimensional window defined by these principal components , we can also mark the location of some well-known strategies ( see Fig . 1 ) ., We find that evolutionary trajectories obtained from the average LOD move towards a fixed point defined by a consensus genotype ( see Methods ) that represents the dominant strategy in the particular regime , while the actual genotypes on the LOD form a cloud in strategy space around the consensus that defines the strategy attractor ., Strategies form clouds around this attractor because in a genetic implementation of IPD , the selective pressures acting on genes depend on the population a player finds himself in ., For example , the DD gene in a cooperating population will begin to drift , only to return to its adaptive value when an invasion of defectors reinstates the selective pressure ., Similarly , the CD and DC genes are under weakened selection in spatial populations because they are only expressed at the boundaries of homogeneous clusters ., The path in strategy space along the average LOD depends strongly on the mutation rate , and shows a qualitative switch—reminiscent of a phase transition—from the cooperative attractor RC ( Fig . 1A ) to the defecting attractor RD ( Fig . 1C ) at a critical value ( Fig . 1B ) , as the mutation rate is increased ., Studying the trajectories that emanate from the 16 ( ignoring the first gene ) deterministic strategies ( Fig . 1D ) suggests that the evolutionary fixed points are unique attractors for a given environment ., We characterize the attractors with an order parameter m generated from the average play frequencies: ( 1 ) which is the normalized difference between frequencies of cooperative and defective play , averaged over the genotypes on the LOD after equilibration ( see Methods ) ., This parameter crosses zero at a critical mutation rate ( Fig . 2A ) , indicating a transition from cooperative to defective strategies ., We find that a transition can also be forced by changing the replacement rate r , as well as other parameters discussed below ., We can study the evolution of cooperation by plotting the order parameter Eq ., 1 as a function of r and μ in a phase diagram that shows that both low replacement rate and low mutation rate lead to cooperation ( Fig . 2B ) , but that the cooperative phase is much smaller for well-mixed populations ( Fig . 2C ) ., As μ approaches 0 . 5 ( a per-genome mutation rate of 2 . 5 mutations per replication event ) , both the spatially-structured and the well-mixed populations begin to drift randomly , signaling that selection has become incapable of maintaining the genetic information ., This transition is likely a quasispecies delocalization 38 , but is smooth rather than abrupt owing to the small genome size 39 ., That all strategies occur with equal frequencies in the population when taking the limit of very high mutation rate has been noted before 40 ., Previous studies have only investigated small slices of this phase diagram by varying the average number of rounds between players 6 ( for deterministic strategies ) or varying the mutation rate in analytic calculations and numeric simulations of an infinitely iterated Markov process 11 , 40 , concluding that cooperation is favored in spatially-structured population but not in well-mixed ones 41 ., The phase diagram suggests instead that both cooperation and defection are possible in either population structure , but that the parameter range that facilitates cooperation in well-mixed populations is more restricted ., As the order parameter Eq ., 1 is obtained from play statistics that represent the phenotype of players , we may ask how this transition is reflected in the genotype instead ., The consensus genotype shows a marked decrease of the PCC probability as mutation rate increases , with clear differences between strategies in spatial ( Fig . 3A ) versus well-mixed ( Fig . 3B ) scenarios , as has been noted before 24 ., At the critical mutation rate ( Fig . 3 , dashed vertical lines ) , the probability to cooperate after CC equals the probability to defect after DD ., Thus , the consensus genotypes mirror the play statistics obtained to define the critical point ., Cooperation is inherently more risky than defection because it forgoes a guaranteed return ( P ) with the expectation of a benefit ( R ) , rather than keeping the guaranteed return hoping for a windfall ( T ) ., This risk is mitigated if the uncertainty about receiving the benefit is reduced ., For example , spatial reciprocity allows kin strategies to preferentially play each other ( because kin place offspring close to themselves ) thus increasing trust ., In our model , an increase in mutation rate decreases the probability that kin play the same strategy ( because mutations change the strategy of kin ) , and thus increases the uncertainty about the identity of the strategy a player will face ., An increased replacement rate has a similar effect , as increasing r shortens the average number of plays that a pair engages in , and this again decreases the probability to face a kin strategy ( mutated or not ) ., Previously , a general theory for the evolution of cooperation has been proposed 42 , 43 that posits that positive assortment between a players genotype and the opponents phenotype is sufficient to promote cooperation , using arguments that ultimately recapitulate Quellers 44 extension of Hamiltons rule ., In our experiments with stochastic conditional strategies , the assortment between a players genotype and an opponents phenotype is generated via the evolution of conditional interactions between the players 45 , i . e . , their ability to base their decisions on information about past behavior ., In a sense , evolutionary adaptation creates this assortment by forging a “model” of the environment ( in terms of the probabilities PXY ) that is adapted to the phenotype given by the play frequencies πXY ., For example , the cooperative fixed point represents a strategy that cooperates with cooperators , retaliates against defectors , but also forgives mistakes ., Thus , it models an environment where cooperators dominate , errors happen , and sometimes defectors try to invade ., More uncertain environments reduce the accuracy of the model , thereby reducing positive assortment , leading to reduced cooperation ., Can changing environmental conditions then drive a population from a cooperating to a defecting phenotype and vice versa ?, In Fig . 4 , we show the order parameter of an adapting population on the line of descent where we changed the mutation rate abruptly from one favoring defection to one favoring cooperation , and back ., We see that the population responds quickly ( in terms of evolutionary time ) and predictably to the changes ., If consistent environments enable cooperative behavior of strategies that rely on “sensing” their environment , we should also be able to influence the critical mutation rate ( where cooperation turns into defection ) by changing other parameters that affect uncertainty ., For example , it is possible to increase player memory so that the last two moves by both players are taken into account to make decisions about cooperation or defection ., In this case , player strategies are encoded in 21 genes , which can be used to predict future moves ., As expected , the critical mutation rate is pushed to higher genomic mutation rates μL ( where L is the number of genes ) for memory 2 ( Fig . 5A ) , and even higher for memory 3 ( data not shown ) ., Another source of unpredictability is the maximal strategy uncertainty given by the Shannon entropy 46 of the genome ., In the present implementation , the probabilities that affect player decisions are coarse-grained to a resolution of 32 , 768 different alleles for each probability , or 15 bits of entropy per gene ., Decreasing this resolution decreases the uncertainty generated by mutations ., Fig . 5B shows the dependence of the order parameter on mutation rate for coarse-grainings of strategy space down to 1 bit ( the deterministic strategies ) ., In this limit , the critical mutation rate ( for 1% replacement ) is pushed towards μ\u200a=\u200a10% , implying that higher mutation rates result in defective play even though cooperation is expected 7 , 47 ., Thus , obtaining more information about the environment , for example by basing decisions on more than one past move , increases the amount of information that a player can use to model the environment , and therefore gives rise to a more close assortment between genotype and opponent phenotype , which increases cooperation ., A framework where evolutionary game theory is implemented via genes that are under mutation and selection could also be used to predict how manipulation of the environment will affect the evolutionary fixed point in other systems ., For example , defection has been observed in a number of biological systems whose dynamics can be described by a PD payoff matrix 48 , 49 ., It is tempting to imagine that these systems can be coaxed into cooperation if mutation rate or turnover rate can be manipulated ( as is shown in Fig . 4 ) ., Evolution can be viewed as a process in which organisms increase their fit to the world by acquiring information about their environment 36 , 50 ., Via this process , genomes become correlated to their environment , that is , genotypes that are adapted to their niche covary with the niches character ., Clearly , such a covariance is greatly enhanced if organisms can sense their environment , and thus base their decisions appropriately on the context ., Therefore , we can expect that the evolution of sensory circuits that inform decisions should ultimately lead to a sufficient amount of covariance so that cooperation is expected according to Quellers rule 44 ( unless environments are so inherently uncertain that they must remain uninformative to any player ) ., If this is indeed true , then it appears that cooperation does not need to be added as a “third fundamental principle of evolution beside mutation and natural selection” as was suggested before 9 , because it is a consequence of evolution ., The payoff for different moves was kept constant at Axelrods values for all simulations:At each update , every player on the 32×32 grid ( with wrapping boundary conditions ) plays each of its neighbors exactly once ., Upon birth , each player begins by consulting its PC gene for each opponent , and one of the four conditional genes thereafter , depending on its own play and the opponents response ., Players are selected for removal randomly with a probability given by the replacement rate r , giving rise to overlapping generations ( asynchronous updating ) 51 , 52 ., As long as the player and its opponent are not replaced , they continue to consult their conditional genes to make decisions , so the replacement rate determines the average length of play history between two players ( if a players partner is replaced , the partner is greeted by consulting the unconditional gene ) ., For most replacement rates , the first gene is consulted so rarely that it drifts neutrally , with a mean around 0 . 5 and a variance of 1/12 , as expected for a uniformly distributed random variable bounded by zero and one ., As a consequence , we often do not show any statistics for this gene ., To implement well-mixed populations using our grid structure , we only changed the identity of the pool used for replacing individuals marked for death , thus keeping the rest of the dynamics consistent ., For structured populations , the eight neighbors of the marked individual are candidates for replication , with a probability proportional to their fitness given by their lifetime accumulated score ., For well-mixed populations , the pool is given by all 1 , 023 remaining strategies in the population ( in a Moran process , it is not usual for the individual to be marked for death to be included in the candidates for replication ) , but each strategy still plays eight neighbors ., The player to be replaced , on the other hand , is chosen randomly among all 1 , 024 players in the population , irrespective of population structure or fitness ., After replication , a genotype is mutated with a probability μ , which is the mean number of mutations per gene per individual , implemented as a Poisson process ., For most of the results in this study , the genes probabilities are coarse-grained to 15 bits , which means that the probabilities are chosen from among 215\u200a=\u200a32 , 768 possible values , representing the number of possible alleles at that locus ., This resolution affects the critical mutation rates as shown in Fig . 5B , but increasing the resolution past 15 bits does not ( data not shown ) ., Because the mutation probabilities are thought to represent the decision of entire pathways of perhaps hundreds of genes , they should not be compared to per-nucleotide mutation rates ., Rather than collecting population averages of plays , we instead study the evolution of strategies by following the line of descent ( LOD ) of player genotypes for each replicate run ., The LOD is obtained by choosing a random player at the end of the run and following its direct ancestors backwards to the first genotype 36 ., Fig . S2A shows a typical sequence of genotypes , while Fig . S2B shows the play statistics for the same LOD ., The population average of play statistics for the same experiment is shown for comparison in Fig . S2C ., Average lines of descent and average play statistics along the line of descent can be created by averaging , for each update , the probabilities of the genotypes as well as the probabilities of play , of the organism on the LOD of each of the 80 replicates at that update ., Fig . S1A shows such an average genetic LOD , while Figure S1B and C shows the average play statistic on the line of descent for two different mutation rates ., The latter two figures show that the average play statistics converge towards evolutionary fixed points that we term the consensus genotype , but that the time to achieve this fixed point depends on the mutation rate ., The consensus genotype for each set of replicates is obtained by averaging the second half of the average genetic LOD minus the last 50 , 000 updates , which removes most or all of the transient and also the variance due to picking random genotypes as the originators of the LOD ., Indeed , because the LOD splits at the most recent common ancestor ( MRCA ) of the population at the end of the run , the LOD past the MRCA is not necessarily representative of the evolutionary dynamics ( as seen for example in Fig . S1B . ) Discarding the last 50 , 000 updates truncates the LOD to genotypes before the MRCA for almost all runs ., Using the MRCA genotype instead of the consensus genotype as representative of the fixed point does not change the results ., We create the evolutionary trajectories in Fig . 1 and Fig . S3 by performing a principal component analysis of the set of probabilities ( PCC , PCD , PDC , PDD ) from all of the 500 , 000 data points on the average genetic LOD of the 80 replicates at mutation rate μ\u200a=\u200a0 . 5% and replacement rate r\u200a=\u200a1% , for both the spatially-structured and the well-mixed population , respectively ., Because the first gene ( PC ) is consulted so rarely it drifts almost neutrally and is for that reason omitted from the PCA ., Including it does not significantly affect the four other principal components ( data not shown ) ., For the spatially structured population we obtain PC1\u200a= ( −0 . 86 , 0 . 192 , −0 . 055 , −0 . 47 ) and PC2\u200a= ( −0 . 348 , 0 . 442 , −0 . 065 , 0 . 824 ) ., These components explain 83% and 10% of the variance respectively ., For the well-mixed population , the principal components are PC1\u200a= ( −0 . 714 , 0 . 132 , −0 . 162 , −0 . 668 ) and PC2\u200a= ( −0 . 393 , 0 . 54 , 0 . 646 , 0 . 37 ) , explaining 86% and 7% of the variance , respectively ., To depict the evolutionary trajectories at higher mutation rate ( panels B and C in Fig . 1 and panels B–D in Fig . S3 ) , we keep the principal components obtained with the low mutation rate strategies so that the landmarks given by the common deterministic strategies such as TFT ( Tit-for-Tat ) , WSLS ( Win-Stay-Lose-Shift ) , ALL-C , and ALL-D remain at the same positions ., These fixed components are also used to plot the location of the consensus genotype at mutation rate 0 . 5% ( RC , the “robust cooperator” ) , and the consensus genotype at mutation rate 5% ( RD , the “robust defector” ) ., The consensus strategies RC and RD for spatially-structured and well-mixed populations are different , and described in the supplementary text below ., Using the principal components implied by the average LOD obtained at 5% mutation rate ( defecting attractor ) instead does not change the nature of the results ( data not shown ) . | Introduction, Results, Discussion, Methods | The observed cooperation on the level of genes , cells , tissues , and individuals has been the object of intense study by evolutionary biologists , mainly because cooperation often flourishes in biological systems in apparent contradiction to the selfish goal of survival inherent in Darwinian evolution ., In order to resolve this paradox , evolutionary game theory has focused on the Prisoners Dilemma ( PD ) , which incorporates the essence of this conflict ., Here , we encode strategies for the iterated Prisoners Dilemma ( IPD ) in terms of conditional probabilities that represent the response of decision pathways given previous plays ., We find that if these stochastic strategies are encoded as genes that undergo Darwinian evolution , the environmental conditions that the strategies are adapting to determine the fixed point of the evolutionary trajectory , which could be either cooperation or defection ., A transition between cooperative and defective attractors occurs as a function of different parameters such as mutation rate , replacement rate , and memory , all of which affect a players ability to predict an opponents behavior ., These results imply that in populations of players that can use previous decisions to plan future ones , cooperation depends critically on whether the players can rely on facing the same strategies that they have adapted to ., Defection , on the other hand , is the optimal adaptive response in environments that change so quickly that the information gathered from previous plays cannot usefully be integrated for a response . | The observed cooperation between genes , cells , tissues , and higher organisms represents a paradox for Darwinian evolution , because the individual success of cheating is rewarded before its long-term detrimental consequences are felt ., The tension between cooperation and defection can be represented by a simple game ( the “Prisoners Dilemma” ) , which has been used to study the conflicts between decisions to cooperate or defect ., Here , we encode these decisions within genes , and allow them to adapt to environments that differ in how well a player can predict how an opponent is going to play ., We find that evolutionary paths end at strategies that cooperate if the environment is sufficiently predictable , while they end in defection in uncertain and inconsistent worlds because inconsistency favors defection over cooperation ., This work shows that cooperation or defection , in populations of players that use the information from previous moves to plan future ones , can be influenced by changing the environmental parameters . | computational biology/evolutionary modeling, computational biology/population genetics | null |
journal.pntd.0005918 | 2,017 | Effectiveness and economic assessment of routine larviciding for prevention of chikungunya and dengue in temperate urban settings in Europe | During the last decade , Europe has faced outbreaks of mosquito-borne diseases ( MBD ) such as dengue and chikungunya , following the continuous importation of human cases in areas with established competent vectors such as the invasive mosquito Aedes ( Stegomyia ) albopictus ( Skuse ) 1 ., Vector control interventions can be implemented by local authorities to keep in check mosquito abundance and consequently reduce the epidemiological risk ., Adulticide spraying rapidly reduces the number of mosquitoes , but its effect is short-lived 2 ., For this reason , it is particularly indicated in situations where the transmission risk needs to be reduced drastically and quickly , such as when an individual is diagnosed with an MBD , to prevent or curtail an outbreak 3 ., Since the effectiveness of reactive measures decreases with the delay between outbreak initiation and implementation of control 4 , a better approach may consist in preventive interventions ., Treatment of potential breeding sites with larvicide products has a delayed impact in reducing adult populations 3 , but experimental studies show that their effect lasts for several weeks 5 , making them better suited for preventive routine control ., The main limit to larviciding as a control option is the proportion of breeding sites that are actually accessible to interventions by public health authorities ., To overcome this limit , education campaigns may be carried forward to encourage citizens to remove and treat potential breeding sites from their private premises during the mosquito season 6 , 7 ., Mathematical modelling of MBD associated with cost-effectiveness analyses can help optimizing routine vector control interventions 8 with respect to constraints in human and financial resources 9 ., With the aim of assisting European municipalities in planning and timing preventive vector control , we assessed the potential epidemiological impact on chikungunya and dengue , and the ensuing economic benefits for the health system , produced by routine larviciding against Ae ., albopictus within urban sites in temperate climates ., Mosquito monitoring via adult trapping was carried out in ten municipalities from the Northern Italian provinces of Belluno and Trento , characterized by a temperate climate 10 ., Mosquitoes were collected using Biogents ( BG ) Sentinel traps ( Biogents AG , Regensburg , Germany ) baited with lures and CO2 from dry ice ., After each trapping session , mosquitoes were killed by freezing at -20°C , identified using taxonomic keys 11 , 12 and confirmed by PCR if found in a location for the first time 12 ., We simulated the transmission dynamics associated with chikungunya and dengue using a standard SEI-SEIR approach 13 in which mosquitoes develop lifelong infection after an ( extrinsic ) incubation period since the bite to an infectious human ( SEI sub-model ) , whereas humans develop temporary infection , followed by the development of immunity , after an ( intrinsic ) incubation period since the bite from an infectious mosquito ( SEIR sub-model ) ., We considered temperature-dependent extrinsic incubation periods and per-bite transmission probabilities for dengue 14 , whereas only temperature-independent estimates were available for Chikungunya 15 , 16 ., The transmission model was initialized with a single infectious human , representing an imported case at a date sampled uniformly between January 1st and December 31st ., The population size of female Ae ., albopictus mosquitoes over time in the transmission model was estimated by fitting a population model to capture data collected in the absence of larvicidal treatments , following the same approach already adopted in 13 , 17 ., The model considers four developmental stages of mosquitoes ( eggs , larvae , pupae and adults ) and reproduces their life cycle by means of temperature-dependent parameters regulating the stage-specific rates of mortality and development ., Free model parameters ( i . e . the site- and year- specific habitat suitability and the capture rate of BG traps ) were estimated via a Monte Carlo Markov Chain approach based on a Poisson likelihood 13 , 17 ., We then included the effect of routine larviciding in the population model ., Experimental studies of several available commercial larvicide products show that 99% of existing larvae and hatching eggs are killed within a given breeding site , with constant efficacy of about 30 days , independently of the specific product used 5 , 18 , 19 ., We considered a standard approach targeting breeding sites in publicly accessible spaces ( e . g . , catch basins placed in public parks and along the road system ) , and an additional strategy where public interventions were integrated by the involvement of citizens in treating and removing breeding sites within private premises ., The latter was parametrized on results from a pilot project conducted in two municipalities within the same area of this study 7 , in which larvicide products were delivered door-to-door and free of charge to house dwellers , who were sensitized and educated to mosquito control interventions ., A key determinant of effectiveness is the fraction of existing breeding sites in a given area that are actually treated ( coverage ) ., We adopted a coverage range between 30% and 50% for larviciding of public catch basins only , and between 60% and 75% for interventions that additionally involve citizens ., These ranges were computed from available data on the density and proportion of reachable breeding sites in public and private premises 7 ., Other strategies aimed at extending the coverage ( e . g . removal of other breeding sites such as water buckets , plant saucers , tarpaulins , etc . ) were not considered ., Since the effect of larvicides is transitory , treatment of catch basins may be repeated multiple times within a given season ., We considered several different starting dates and from 1 to 4 applications of larvicide treatments within a given mosquito season ( hereafter referred to as “effort level” ) , implemented with monthly frequency ., To evaluate the economic acceptability of the two considered strategies , a cost-utility analysis for the prevention of dengue and chikungunya was conducted , taking the number of infections as input from the transmission model ., Disability Adjusted Life Years ( DALYs ) averted and net costs were derived comparing an intervention scenario to the case in which no control programs were put in place ( baseline ) ., The baseline was set to reflect a municipality where only the monitoring of mosquito presence via ovitraps was performed 7 ., The analysis was conducted from a public healthcare system perspective through the maximization of the net health benefit ( NHB ) 20 ., This measure is defined as the difference between the DALY averted and the incremental cost due to the intervention , the latter divided by the willingness to pay ( WTP ) by public authorities for each DALY averted ., Following WHO recommendations 21 , we assumed such value approximately equal to the Gross Domestic Product ( GDP ) , which is about 35 , 000 euro per capita in our study area 22 ., Probabilities of each infected case of being symptomatic , notified , severe , hospitalized and of dying , and the length of stay in hospital , were derived from published studies 23 , 24 and from analyzing data from the Italian Hospital Discharge System ( Schede di Dimissioni Ospedaliere ) , accounting for all hospital admissions for chikungunya and dengue recorded in Italy ., The cost of illness was estimated according to expert opinion ., The costs of intervention were estimated from actual costs during control activities against Ae ., albopictus recently performed in two municipalities from the study area 7 ., For all the considered scenarios , the NHB was computed on a set of 100 , 000 stochastic realizations accounting for the uncertainty in both the transmission and the economic model’s parameters ., Full details on this analysis are provided in S1 Text ., To assess the feasibility and sustainability of public interventions , we used responses from a questionnaire administered in 2013 to municipalities of the province of Trento , aimed at collecting information on the actual public expenditure on vector control activities ., The estimated density of adult female mosquitoes ( averaged between April 10th and September 30th ) was between 4 and 88 per hectare in 2014 and between 9 and 198 in 2015 , depending on the municipality ( see Table 1 ) ., The higher abundance in 2015 is mostly due to the much higher temperatures recorded during summer ., The initial reproduction numbers and the threshold for autochthonous transmission of chikungunya and dengue over time were estimated in a previous study 17 ., Here , for each site and year , we computed the probability of autochthonous transmission of chikungunya and dengue originated by an imported infection in the absence of interventions ., Higher vector densities during 2015 resulted in an increased risk of local transmission for both infections , compared to the previous year ., The probability of observing at least one secondary case was estimated to be up to 30% for chikungunya and 15% for dengue in highly infested towns in 2015 ., Corresponding maximum probabilities in 2014 were around 20% for chikungunya and 5% for dengue ., This means that 7 importations of chikungunya and 15 importations of dengue in towns most at risk would have a >90% probability of causing at least one secondary case in 2015 ., Sporadic transmission ( less than 10 secondary cases ) is by far the most likely scenario , especially for dengue ( Fig 1 ) ., However , we found a low , but non-negligible , probability ( up to 2 . 7% ) that an uncontrolled chikungunya outbreak would produce more than 50 cases in several sites during 2015 ., Routine preventive larvicide treatments can reduce significantly mosquito populations and consequently the probability and size of outbreaks triggered by sporadic importation of infected cases ., To evaluate the overall effectiveness , we considered the expected number of total secondary infections per imported case ., Under the baseline scenario of no control interventions , this index ranged from 0 . 1 to 5 . 2 , depending on the site and year; corresponding numbers for dengue were everywhere below 0 . 5 ., Because of the smaller epidemiological risk of dengue , we discuss only the cost-effectiveness analysis on chikungunya , leaving corresponding results for dengue to the S1 Text ., For each site and year , and for each timing , effort level and assumed coverage , we evaluated the relative reduction in the expected number of secondary infections per imported case as a measure of effectiveness ., Fig 2 and Table 2 show that all interventions with optimal effectiveness covered the month of July , which corresponds to the estimated period of steepest growth of the adult Ae ., albopictus population in both years ., We selected for further analyses only interventions with optimal timing for each effort level ( Table 2; the reduction in mosquito abundance corresponding to the optimally timed interventions is reported in the S1 Text ) ., We found that an increase in the effort level does not proportionally reduce the expected number of cases ( Fig 3 ) ., In particular , an expansion in the coverage of breeding sites from 30% to 50% would be more effective than doubling the effort level while keeping the coverage at 30% ., In general , interventions are most beneficial when the baseline risk is highest ., Towards an optimal allocation of resources , the benefits of reducing the potential number of transmitted cases needs to be compared with the intervention costs ., Taking into account all possible clinical outcomes , including the probability of severe illness and of hospitalization , the estimated average cost per infection is 424 . 9 euros ( 95% CI 342–533 ) for chikungunya and 275 . 88 euros ( 95% CI 151–422 ) for each dengue infection ., The corresponding average DALY loss per case is higher for chikungunya ( 0 . 45 , 95% CI 0 . 10–1 . 12 ) than for dengue ( 0 . 29 , 95% CI 0 . 15–0 . 44 ) ., In Fig 4 , we show the relative probability that each effort level ( including the no-intervention scenario ) will maximize the NHB for each site , year , and coverage ., Three main outcomes can be identified ., The first is represented by larger cities ( Trento , Belluno and Rovereto , all above 35 , 000 inhabitants ) where non-intervention has the highest likelihood of being optimal ., In these sites , the poor economic effectiveness of larviciding depends on the relatively low number of expected secondary cases even in the absence of treatment ( Fig 3 ) , combined with the high intervention costs due to the extent of the area to be covered ., The second group consists of smaller towns where intervention is always beneficial ( Povo , Santa Giustina , Tenno and Tezze , all below 10 , 000 inhabitants ) and where higher effort levels have the highest probabilities of being optimal ., Strigno ( about 3 , 400 inhabitants ) represents an exception to this rule , where the low intervention costs are counterbalanced by a very small transmission risk in the absence of interventions ., Nonetheless , even in Strigno a low-effort intervention ( single treatment ) might be beneficial because of its low cost ., The third situation occurs in towns of intermediate size ( Feltre and Riva del Garda , between 20 , 000 and 35 , 000 inhabitants ) where both the intervention costs and the transmission risks are high ., In these cases , depending on the larviciding coverage , absence of intervention might be the optimal strategy in seasons of lower mosquito abundance ( 2014 ) while a low-to-moderate effort ( 1 to 3 treatments ) might be the best choice in years of high infestation ( 2015 ) ., Overall , the probability that a more intensive intervention will be optimal increases with the coverage and with higher transmission risk ( 2015 vs . 2014 ) ., We also tested the cost-effectiveness of expanding the coverage by involving private citizens 7 ., We found that this type of intervention might achieve significant additional reductions in the expected number of secondary cases and probability of local transmission ( reported in the S1 Text ) ., However , they are rarely optimal from the economic perspective because they require labour-intensive activities ., Fig 5 reports results of the NHB analysis for a single larvicide treatment , but qualitative inferences are similar for more intensive efforts ( see S1 Text ) ., The only two instances where involvement of citizens was found to be economically beneficial were Povo and Tezze and only during the 2015 mosquito season , i . e . only where the urban size is small enough to keep intervention costs low and where the transmission risk at baseline is sufficiently high ., Two municipalities under study , Trento and Riva del Garda , had responded to a previously administered questionnaire on public expenditure on vector control , declaring an overall budget of 0 . 254 euro and 0 . 532 euros per inhabitant , respectively ., In Trento , the most cost-effective activity predicted by our model was monitoring by ovitraps ( Fig 4 ) , which has an estimated average cost of 0 . 016 euro per inhabitant; in Riva del Garda , one or two larvicide applications per year would be likely optimal and would cost between 0 . 256 and 0 . 512 euros per inhabitant ., Therefore , the most cost-effective strategies are sustainable with respect to the current allocated budget ., We provide full details on questionnaires , municipality-specific answers and intervention costs in the S1 Text ., In this work , we evaluated the effect of routine larviciding against dengue and chikungunya , two viruses transmitted by bites of Ae ., albopictus mosquitoes ., We used data from two seasons of entomological surveillance in multiple sites from northern Italy to parametrize a mathematical model of mosquito population dynamics and control ., The population model was coupled with a transmission dynamics model and a cost-effectiveness analysis to identify suitable routine vector control strategies for temperate climate municipalities in Europe ., We found that , in the absence of interventions , the risk of autochthonous dengue transmission was low and limited to sporadic transmission in both years , because of the relatively low competence of European strains of Ae ., albopictus ., On the other hand , the risk of a chikungunya outbreak was estimated to be up to 30% in 2015 , with a non-negligible probability of observing outbreaks larger than 50 cases in most sites ., We found that the most effective interventions in reducing the amount of expected locally transmitted cases were those for which the window of larvicide efficacy covered at least the month of July ( Fig 2 , Table 2 ) ., Larviciding reduced the probability of secondary cases only moderately , but it had an important impact in avoiding larger outbreaks ., Our analysis included two seasons that were representative of a broad range of mosquito abundances , due to the remarkable temperature differences ., The cost-effectiveness of larviciding depends on the actual mosquito abundance in a given year; however , general rules could be identified independently of the considered year: small villages ( <10 , 000 inhabitants ) with moderate-to-high mosquito abundances will maximally benefit of intense larviciding efforts made of season-round monthly treatment of public catch basins ., For medium-sized towns ( 20–35 , 000 inhabitants ) with high infestation rate , the benefits are partially offset by the higher cost of intervention; in these cases , a moderate larviciding effort ( 1 to 3 treatments within the season ) is recommended ., Larger cities in our study ( >35 , 000 inhabitants ) were characterized by a low or intermediate transmission risk , and the high costs of an intervention covering the entire urban area made it economically disadvantageous ., In these situations , treating specific neighbourhoods with highest mosquito abundance ( called ‘hot spot approach 25 ) may be cost-effective ., In order to evaluate such a scenario , however , it would be necessary to model the complex effect of the urban layout on the spatial distribution of breeding sites and on the dynamics of mosquito populations 7 , which is out of the scope of our study ., Treatment of private breeding sites via the direct involvement of citizens by door-to-door visits was recommended only in small towns with high mosquito infestation ., A survey on the allocated budget for mosquito control programs across different municipalities showed that expenses required for the most cost-effective interventions are sustainable for the considered area ., These results need to be contextualized with respect to our simplifying assumptions ., First , all results are given conditionally on a uniform probability of importation of an infectious individual within a given epidemiological year ., For comparison , in the considered provinces of Trento and Belluno , three imported cases of dengue and one imported case of chikungunya were recorded in 2014 ( C . Rizzo , personal communication ) ; however , the actual importation rate may vary significantly by year and time of the year , depending on spatio-temporal patterns of global epidemics and international travel ., We did not consider reactive interventions that are implemented when a case of chikungunya or dengue infection is detected or after an outbreak has started ( e . g . , insecticide air spraying in the neighbourhood of the index case 26 ) ., In addition , our results are relative to the prevention of arboviral transmission; however , there may be other purposes in vector control activities , such as the reduction of nuisance for citizens , which were not included in our analysis ., For what concerns the economic assessment , we did not consider the impact of local transmission detection on the blood supply chain ., Upon clinical confirmation of a locally transmitted arboviral infection , restrictions on the usage of blood bags collected in the region are applied to prevent transmission via transfusions , and screening tests on available blood supplies are implemented 26 ., These additional interventions are quite expensive , and savings associated to the reduction of transmission risk granted by larvicides may dramatically offset the cost-benefit balance in favour of the intervention ., However , these costs are difficult to estimate because of the lack of sufficient data ., We did not include other arboviroses transmitted by Ae ., albopictus because of their lower epidemiological relevance to the considered area ., For example , the risk of Zika virus transmission was found to be close to zero in the study region , even under conservative scenarios 17 ., Nonetheless , we note that larvicides produce simultaneous benefits in preventing multiple diseases transmitted not only by Ae ., albopictus but also by other affected mosquito species ( e . g . West Nile virus associated to Culex pipiens L . ) ., Furthermore , larviciding may assist in limiting the spread of other invasive mosquito species such as Aedes ( Hulecoeteomyia ) japonicus ( Theobald ) and Aedes ( Hulecoeteomyia ) koreicus ( Edwards ) 1 , 27 ., An interesting research question is how the balance of ecological interactions between mosquito species 28 may be offset by such interventions ., Other studies 2 , 6 , 7 have investigated the effectiveness of vector control in Europe using different approaches ., The cost-effectiveness of larvicidal treatment against Ae ., albopictus in temperate climates has been evaluated only in combination with other interventions during an ongoing outbreak 29 , 30; other studies were based on endemic ( extra-European ) settings where transmission is mainly mediated by Aedes ( Stegomyia ) aegypti ( Linneus ) 31 , 32 ., Overall , results from different studies and approaches , including our own , are consistent in highlighting the potential of larviciding towards reducing mosquito populations; however , this reduction will not result in a complete elimination of the risk of local chikungunya or dengue transmission ., Additional strategies may integrate the control of risks from mosquito-borne diseases , including source reduction methods ( e . g . identification and removal of breeding sites ) , mass trapping ( e . g . via lethal ovitraps ) and approaches leveraging ecological interactions ( such as the use of Wolbachia bacteria or the release of genetically sterilized male mosquitoes ) ., A comprehensive review of the potential for these strategies can be found in 9 , but specific cost-effectiveness studies are needed to identify optimal strategies for vector control ., European municipalities with temperate climate where Ae ., albopictus is established may take advantage of results from this study when planning and timing routine larviciding interventions aimed to prevent or reduce epidemiological risks ., Temperate European areas share with our study collection area similar temperature suitability for the transmission of arboviruses 33 and similar abundances of Ae ., albopictus 34 , so that results on the epidemiological effectiveness of larviciding should not differ significantly ., More caution should be paid when extrapolating cost-effectiveness conclusions to different countries , given potential differences in health and intervention costs and in the choice of the WTP ., Finally , we suggest that the proposed methodological approach may also be extended to European areas with different climates , conditional on the availability of local data on mosquito abundances estimated via entomological surveillance activities . | Introduction, Methods, Results, Discussion | In the last decades , several European countries where arboviral infections are not endemic have faced outbreaks of diseases such as chikungunya and dengue , initially introduced by infectious travellers from tropical endemic areas and then spread locally via mosquito bites ., To keep in check the epidemiological risk , interventions targeted to control vector abundance can be implemented by local authorities ., We assessed the epidemiological effectiveness and economic costs and benefits of routine larviciding in European towns with temperate climate , using a mathematical model of Aedes albopictus populations and viral transmission , calibrated on entomological surveillance data collected from ten municipalities in Northern Italy during 2014 and 2015 . We found that routine larviciding of public catch basins can limit both the risk of autochthonous transmission and the size of potential epidemics ., Ideal larvicide interventions should be timed in such a way to cover the month of July ., Optimally timed larviciding can reduce locally transmitted cases of chikungunya by 20% - 33% for a single application ( dengue: 18–22% ) and up to 43% - 65% if treatment is repeated four times throughout the season ( dengue: 31–51% ) ., In larger municipalities ( >35 , 000 inhabitants ) , the cost of comprehensive larviciding over the whole urban area overcomes potential health benefits related to preventing cases of disease , suggesting the adoption of more localized interventions ., Small/medium sized towns with high mosquito abundance will likely have a positive cost-benefit balance ., Involvement of private citizens in routine larviciding activities further reduces transmission risks but with disproportionate costs of intervention ., International travels and the incidence of mosquito-borne diseases are increasing worldwide , exposing a growing number of European citizens to higher risks of potential outbreaks ., Results from this study may support the planning and timing of interventions aimed to reduce the probability of autochthonous transmission as well as the nuisance for local populations living in temperate areas of Europe . | Larvicides are a key tool to prevent the growth of mosquito populations and decrease both the risks of outbreaks of mosquito-borne diseases and the nuisance deriving from bites ., In order to assist municipalities from temperate areas in Europe in effectively planning vector control programs , we modelled the effect of larviciding in public areas on populations of Aedes albopictus using mosquito collection data from 10 municipalities in Northern Italy , over two years with very different temperature conditions ., We then evaluated the resulting probabilities of potential outbreaks of chikungunya and dengue and their expected final sizes , and we compared the intervention costs to the economic and health benefits due to the avoidance of clinical cases ., By assessing several different intervention strategies , we found that the optimal timing should be centred on the month of July , corresponding to the period of maximal growth of the mosquito population ., Municipality-wide interventions are likely to be cost-effective in small-to-medium towns ( below 35 , 000 inhabitants ) even where mosquito infestation is moderate , whereas for larger cities a neighbourhood-based intervention should be considered ., The involvement of citizens to apply larvicides within private premises resulted effective but generally too costly . | larvicides, invertebrates, medicine and health sciences, pathology and laboratory medicine, togaviruses, chikungunya infection, cost-effectiveness analysis, economic analysis, pathogens, tropical diseases, microbiology, geographical locations, social sciences, animals, alphaviruses, viruses, chikungunya virus, rna viruses, pest control, neglected tropical diseases, infectious disease control, insect vectors, infectious diseases, agrochemicals, medical microbiology, epidemiology, microbial pathogens, economics, disease vectors, insects, agriculture, arthropoda, pesticides, people and places, mosquitoes, eukaryota, viral pathogens, biology and life sciences, viral diseases, species interactions, europe, organisms | null |
journal.pcbi.1006143 | 2,018 | Latent environment allocation of microbial community data | Microbial communities are present worldwide in almost all possible environments ., Because the composition ( structure ) of a microbial community and its surrounding environment are closely related to each other , it is important to understand what kinds of structural patterns are possible and how environmental factors affect community formations ., Over the past decade , the structures of tens of thousands of microbial samples derived from various natural environments , including those in symbiosis with humans , have been analyzed ., Using these datasets , global patterns of microbial diversity have been characterized that show that community structures constitute distinct clusters among at least certain environments1–4 ., In addition , the community structure of each examined environment has been evaluated using a clearly defined environmental ontology5 , 6 ., However , the granularities ( i . e . , the levels of detail ) of human-classified environmental categories do not necessarily coincide with structural patterns of microbial communities , and this unavoidable arbitrariness in the granularities of environmental labels may bias the interpretation of results of comparative analysis , such as an enrichment analysis of environments among communities1 ., There are three types of incongruences between environmental labels and community structures ., First , there are different subtypes in the microbial community structures associated with certain environments , e . g . , enterotypes in the human gut and vaginal community types7–9 ., Second , in contrast to the first case , a nearly identical community structure may be observed across different environmental labels ., For example , microbial communities of the surface of the home environment and their inhabitants show highly similar structural patterns10 ., Third , because an environment varies continuously over time and space , it is impossible to define it using a strict segmentation or hierarchical structure ., For example , the brackish water of an estuary can have various mixtures of fresh water and seawater , for which the relative proportion continuously shifts11 ., Although an environment is difficult to definitively define owing to uncharacterized factors , these factors can potentially be defined indirectly using microbial community data because microbes respond quickly to environmental changes12 , 13 , and their community structure reflects the state of the environment14–16 ., Microbial community structures have been analyzed by various data clustering methods ., Most of these approaches are based on the evaluation of data densities on high-dimensional feature space in which microbiome samples are distributed ., Microbial community structures are complex as they are described by a large number of features ( taxa ) , although not all the features necessarily vary independently ., There are groups of taxa that show co-occurrence patterns in samples17–19 ., Herein , we refer to such co-occurrence relationships of microbes as “sub-communities” ., Summarizing community structure data as mixing ratios of sub-communities makes it easier to interpret community dynamics according to environmental changes20 ., To extract such partial structures from mixed data , the machine-learning technique denoted a topic model has been extensively studied in recent years since introduction of the Latent Dirichlet Allocation ( LDA ) model21 ., The LDA model is a probabilistic generative modeling approach , mainly used in natural language processing , for discovering the latent ( unobserved ) structures of the dataset ., The LDA model and its extended models have been used to analyze microbiomes20 , 22 , 23 , however , it is often difficult to interpret extracted sub-communities of microbial taxa ., Sub-communities have been characterized by evaluating their relationships with occurrences of sample metadata ( i . e . , data describing information about the samples , such as body sites and gender ) after modeling24 , or by explicitly modeling associations between metadata and sub-communities20 ., These methods cannot be applied unless all samples have standardized metadata with a uniform granularity , and thus tend not to be practical for comprehensive analyses of microbial samples from various environments ., All metagenomic data registered in public databases have such metadata , that is , natural language data described by the researchers who registered the samples ., The paired dataset of community structures and description documents can be used for modeling the conditional relationship between them ., However , natural language descriptions in databases are not always sufficiently described as their content for many samples is often incomplete and has widely variable resolution ., For example , in a sample , various information on the host such as race and gender , experimental conditions , the purpose of the research project , etc . are described , but in another sample , it is described only as “human gut metagenome” and is needed to be treated as a sample with missing values ., Therefore , for robust modeling , it is necessary to assume stochastic-generating processes not only for community structures but also for documents within the framework of the probabilistic generative model ., For such purpose , the Correspondence-LDA ( Corr-LDA ) 25 , 26 model can be applied ., Corr-LDA is a probabilistic modeling approach that is used to extract correspondence between various types of elements occurring in the same dataset , for example , the correspondence between sub-regions of pictures and their captions25 , between topics of blog documents and their annotation tags26 , or between brain regions and their cognitive functions2728 ., In this research , we attempted to find relationships between patterns of microbial community structures and patterns of “environments” that the human recognizes and describes ., To this end , we applied the Corr-LDA model to pairs of taxonomic compositions and natural language sample descriptions for tens of thousands of sequenced 16S rRNA gene amplicon samples reanalyzed by the unified analysis pipeline ., Using this dataset , “topics” extracted by the Corr-LDA model would represent the core elements of environmental mixtures ., By integrating training results , we developed an interactive web application denoted LEA ( Latent-Environment Allocation ) , which is freely available at http://leamicrobe . jp ., The extracted connections between microbial sub-communities and subsets of English words via topics are applicable to various analyses ., LEA enables researchers to do the following:, 1 ) clarify the relationship between environments and patterns of microbial community structures ., 2 ) predict the “latent environments” of new samples from , for example , the ocean , a diseased gut , or another unexpected environment , and quickly compare new samples with tens of thousands of existing samples based on their environmental similarity , which makes it easy to detect dysbiosis of the microbiome in the human gut or contaminants in natural environments ., 3 ) search for samples in the >30 , 000-sample dataset based on an ecological perspective , without depending on exact word matching of queries and sample descriptions ., In this paper , we show the patterns found in the human gut and vaginal communities as an example of separations and connections of extracted “environments” , and show how the LEA global map and a semantic search method on the map make it easy to explore patterns in microbial community structures ., In addition , as an example of environmental predictions for newly acquired microbiome samples , we show the LEA mapping results for the datasets of the human gut microbiome and the microbiome derived from various natural environments ., We collected sequenced 16S rRNA gene amplicon samples from the MicrobeDB . jp database and performed a phylogenetic analysis using VITCOMIC229 , which is the metagenomic analysis pipeline improved from VITCOMIC30 , on all samples ., This resulted in a dataset containing 30 , 718 samples with genus level information on their taxonomic composition linked to a document containing sample description information ., For this dataset , model inference runs were performed by Corr-LDA with a varying number of topics ., The perplexity , which is the performance evaluation index of the model ( smaller values indicate a better performance ) , was sufficiently small for the model with 80 topics ( S1 Fig ) ., In the following sections , we discuss the results for the model inferred with 80 topics ., The inferred word subsets and microbial sub-communities for each topic are shown in S2 and S3 Figs ., Each topic has a unique subset of words and a microbial sub-community ., The structure of a microbial community sample that has a large proportion of a certain topic is likely to contain microbes in the sub-community of that topic , and the description of the sample is likely to contain words in the word subset of that topic ., For most topics , the word subset associated with the topic represents a single natural environment or a symbiotic environment with humans ., Based on the topic composition of each sample , the similarities among the samples were visualized using parametric t-SNE31 ., In Fig 1 , the dots represent the 30 , 718 samples and the pictures represent 80 topics ., A sample that is mapped near a picture indicates that the sample has a large proportion of that topic ., On the map in Fig 1C , the sample ( SRS425923 ) , which is located approximately midway between the two pictures , has the two topics ( topic #37 and #52 ) mixed in similar amounts ., The 80 topics can be regarded as latent environments that affect the formation of microbial community structures ., Topics form several clusters with dense connections formed by many samples but with sparse or no connections between clusters ., Topics can be roughly divided into gut , skin , vagina , oral cavity , ocean , and soil ., In addition , there are several isolated topics , which include a coral reef , a mosquito , phyllosphere , etc ., The gut microbiome in healthy adult humans are reported to consist of three7 or four24 community types ., However , whether truly discrete clusters exist as individual gut microbial communities remains in doubt32 ., For the gut community types ( enterotypes ) , the key genera characterizing each community type have been identified—Bacteroides , Prevotella , and Ruminococcus7—although the abundance of key genera varies between samples instead of being discretely clustered32 , 33 ., Therefore , unlike discrete clusters , e . g . , blood types , the compositions of microbial communities are continuously shifting , perhaps as a result of environmental factors ., Such continuous variation of the structure of a microbial community can be discerned by our method ., 22 topics are related to the gut according to the word subsets associated with each of the topics ( S8 Fig ) , including those with a large proportion of Bacteroides , Ruminococcus , and Prevotella ( topics #79 , #51 , and #24 . Fig 1B ) ., However , most of the samples do not reside near a single topic but instead occupy an intermediate position between multiple topics , meaning that the samples are a mixture of several topics ., Because many samples with intermediate properties owing to multiple topics exist , there is variability across a limited area of the gut microbiome ., Bacteroides is often found in the guts of people who eat diets rich in protein and fat , and Prevotella is often found in the guts of vegetarians12 , 34 ., In fact , words denoting herbivores , e . g . , “pig” , “swine” , “horse” , “bovine” , and/or “rumen” , are frequently found in the Prevotella-rich topic and in topics that are peripherally connected to the Prevotella-rich topic ( Fig 1B ) ., Regarding the vaginal flora , three related topics were found ( Fig 1C ) : the Lactobacillus-rich topic ( #37 ) ; the topic including Gardnerella , Sneathia , and Atopobium ( #52 ) ; and the Shuttleworthia-rich topic ( #43 ) ., Vaginal community types ( Community State Type; CST ) have previously been examined in detail with five CSTs recognized to date: four ( CSTs I , II , III , and V ) in which Lactobacillus species dominates and one ( CST IV ) with various obligate or facultative anaerobes and very few Lactobacillus9 , 35 ., The two topics detected in our model are consistent with the above results ( Fig 1C ) ., Because we used community structure data found at the genus level , we cannot distinguish between CSTs I , II , III , and V , so these communities were identified as a single topic dominated by Lactobacillus ., For the second topic corresponding to CST IV in which Gardnerella and Atopobium dominate , the associated samples likely were obtained from African-American women ( as estimated by the word subset of topic #52 ) ., Interestingly , samples in which the Lactobacillus-rich and Gardnerella-rich topics are mixed in various proportions are frequently found , as indicated by the many dots that connect these two pictures ( Fig 1C ) ., Vaginal bacterial communities are known to be stable throughout pregnancy and to be relatively stable throughout the menstrual cycle although changes in the Lactobacillus spp ., populations have been observed36 , 37 ., Therefore , an environmental gradient of unidentified factors may exist in the vagina , which would cause a community structure to exist as an intermediate state between two topics ., Another topic related to the vaginal environment is a topic dominated by Shuttleworthia ., The presence of Shuttleworthia may be related to bacterial vaginosis38 or to squamous intraepithelial cervical lesions39 , but its ecology is not well understood ., Interestingly , the continuous transition of samples to the Shuttleworthia-rich topic links only with the Gardnerella-rich topic ( Fig 1C ) ., LEA can predict latent environmental topics of newly acquired samples using the Bayesian prediction method with the identified 80 topics ( see Methods ) ., By examining the word subsets associated with the mixed topics , the environment in which new samples are found can be estimated ., In addition , by training the dimension-reduction function of t-SNE in our system using a neural network procedure , it is possible to arrange the locations of new samples on the global map ( Fig 1A ) according to their topic compositions without changing the coordinates of previously mapped samples ., The topic prediction of a new sample and its placement on the map are implemented by the LEA web application ., By uploading the taxonomic assignment file of VITCOMIC2 , the placement of the sample on the map can be viewed in a few seconds ., As examples of LEA mapping results , we have analyzed the dataset of a time-series human gut microbiome analysis40 , which consists of fecal samples obtained every day from two male subjects from the US ( subjects A and B ) ., The results are shown in S4 Fig . The LEA visualization reproduces the results of David et al . 40 , such as the stability of the gut microbiome of subject A over the course of the experiment with the exception of his time in Southeast Asia , and the change in the gut microbiome of subject B caused by an infection ., Such results can be easily obtained using the LEA web application ., In a meta-analysis of a large-scale dataset , the existence of systematic bias due to the difference in methods across studies often becomes a problem ., We tried to address this problem by processing all samples with a unified information analysis pipeline , but there is a possibility that a further upstream , sample preparation protocol could be a confounding factor ., In particular , a bias due to differences in DNA extraction methods often becomes a problem41 ., To assess the impact of different DNA extraction methods of the human gut microbiome analysis on the locations on LEA global map , we conducted LEA environment predictions for the Microbiome Quality Control ( MBQC ) dataset42 ., This dataset contains 16S amplicon sequencing data from human stool samples , chemostats , and artificial microbial communities ., For the same biological sample , there are multiple sequencing data analyzed with different wet laboratories or different DNA extraction methods ., The results are shown in S10 Fig . First , most of the samples derived from human feces in the MBQC dataset were properly mapped to the “gut” area of the LEA global map ( S10A Fig ) ., As a whole , there was no tendency for samples processed with a specific DNA extraction kit to be mapped only to a specific topic ., Therefore , separation of topics on LEA is not necessarily influenced by differences in experimental protocols ., However , considering samples that have a same biological origin , some samples were mapped to the nearly same position on the map , and the others were mapped on the location biased by the DNA extraction kits ( S10B–S10S Fig ) ., The direction of biases probably differs depending on the position of the true taxonomy composition ., Therefore , topics may partially contain systematic bias due to differences in studies , and caution is necessary for interpretation ., As an example of LEA involving a natural environment , Fig 2 shows the topic predictions for 38 samples of microbial communities obtained over a short period of time and at a high density from various points along the Tamagawa river in Japan ., The upstream region of the river begins in a deep mountainous region; the middle region flows through a densely-populated zone where there is water from sewage treatment plants and from tributaries that joins the river; and the downstream-most region flows into Tokyo Bay ., On May 26 and 27 , 2015 , we sampled the surface water of the river at 38 points ( S5 Fig , S1 Table ) and identified the microbes contained in the samples by VITCOMIC2 after sequencing of their PCR-amplified 16S rRNA genes ., The genus level taxonomic compositions are shown in Fig 2A ., Limnohabitans is the major genus found in the samples from the river ., The microbial community structure of the river continuously shifted as it flowed to the estuary , but sample 200 had a greatly different structure ., Sample 200 was obtained just under the sewage treatment facility , and its composition probably reflects the microbiome of the treated water ., The community structures of samples 10 , 20 , and 30 , which were obtained from brackish water in the estuary , also differed greatly from those collected elsewhere along the river ., The topic predictions for the river samples are shown in Fig 2B ., After performing LEA , two topics related to “river” were found: one was topic #3 , which frequently occurs together with words such as “Baltic sea , ” “lake , ” and “river , ” and the second was topic #53 , which is associated with the words , “river , ” “wastewater , ” and “urban . ”, The aforementioned words belong to the dominant topics in Fig 2B ., Topic #3 is primarily associated with the upstream region of the river and topic #53 with all areas of the river ., The relative proportions of these two topics gradually change as the river flows downstream ., Given the word subsets associated with the two topics , topic #3 represents freshwater ecosystems , such as lakes and rivers , and topic #53 represents river areas adjacent to cities ., Samples 10 and 20 ( from the estuary ) are largely associated with topics #11 and #63 , which represent the ocean , and , along with sample 30 , are associated with topic #56 , which represents activated sludge ., For the Tamagawa river , about half its water that flows into the estuary is treated water43 ., Thus , our results suggest that the mixing of the river water with seawater greatly changes the community structure and that the river’s ecosystem is greatly affected by it interaction with the urban area ., Topic #45 is associated with the upstream region of the Tamagawa river ( sample 360 to 250 ) ; the words associated with this topic include “pet” and an indoor environment ( Fig 2B ) ., Many of the 16S rRNA sequences associated with topic #45 belong to Blastomonas , a genus associated with domestic wastewater , which is found in tap water , faucets , and shower hoses and is resistant to disinfection44–46 ., Advanced sewage treatment facilities are not found in this region of the Tamagawa river , suggesting that untreated household wastewater is being dumped into the river ., Most of the Tamagawa river samples were mapped near “freshwater” topics #53 and #3 on the global map ( Fig 2C ) , although the topics of sample 200 , taken near the sewage treatment plant , and samples 10 , 20 , and 30 , taken from the estuary , diverged , to some extent , from the freshwater topic ., Specifically , sample 10 were mapped within “ocean” topics ., In this way , the LEA web application can place a new sample appropriately on the global map of existing samples and enables visual and intuitive operation to evaluate its characteristics , e . g . , deviations from the expected environments ., For further testing of LEA using external dataset , we conducted environmental predictions of microbiome data derived from a highly diverse environment produced by the Earth Microbiome Project ( EMP ) 47 ., One of the good points about this dataset is that every sample is given an environmental label based on a controlled vocabulary , the EMP Ontology ( EMPO ) ., EMPO is a hierarchical framework that captures the major axis of the microbial community diversity and is used to assign samples of EMP to its habitat47 ., Therefore , by comparing the result of LEA mapping with each EMPO label , we can estimate the accuracy of environmental prediction by LEA ., For each of the lowest layer label ( level 3: most specific habitat name ) of EMPO , we examined the location of the samples given that label on the LEA map ., The results are shown in S11 Fig . Environmental prediction results have well captured the influence of salinity known as the main axis that determines the community structure2 ., For most samples of water , sediments , biofilms , and soils , saline samples were mapped around the ocean area , and non-saline samples were mapped to freshwater or soil area ( S11A–S11H Fig ) ., Regarding the samples derived from the host-associated environments , it was observed that the mapping pattern varied depending on the host species even with the same EMPO label ., Also , the EMPO label “Plant surface” intuitively evokes leaf surface of land plants , but most of the EMP samples labeled with “Plant surface” mapped to “ocean” area on LEA ., This is because most of the EMP samples used in this study with the label “Plant surface” are derived from the kelp as the host ( S11J Fig ) ., In such a case , environmental prediction by LEA gives interpretable results ( microbial communities on the kelp surface reflects the oceanic community structure pattern , etc . ) ., When the environmental ontology and the community structure pattern seem to conflict , LEA can be used to infer the reason from the mapping results ., The topic-model approach can be used to semantically search documents related to a user’s query48 , 49 ., By using the trained model parameters in LEA , existing samples can be searched using natural language such as “forest soil” , or “hot spring” ., Instead of needing to search for samples by exactly matching the queried words and the description information associated with samples , we can find the sample using latent environmental topics , using the probability of each sample to generate the query sentence as the score of the sample ., As an example , Table 1 shows the top five scoring samples obtained by querying “What kind of microorganisms are in an oil sands tailings pond ? ”, Oil sands tailings ponds are slag ponds accompanying oil sand development and are highly toxic environments as they contain heavy metals , naphtha , bitumen , and other toxic chemicals ., Tailing ponds have heterogeneous environments , being aerobic at their surfaces and anaerobic at their bottoms ., Many members of the class Methanomicrobia , including Methanoculleus , Methanolinea , Methanosaeta , Methanobrevibacter , and Methanocorpusculum , which are methanogenic archaea found in the sediment of tailing ponds , contribute to the decomposition of hydrocarbons50 , 51 ., Given the query , “What kind of microorganisms are in oil sands tailings ponds ? ” , LEA returned the samples derived from oil sand tailing ponds and oil-water mixtures ( Table 1 ) ., In addition , LEA returned the sample derived from ocean sediments , although the description of this sample did not contain words such as “oil , ” “sand , ” or “pond . ”, Although the microbial community structures of these samples varied in terms of their taxonomy , almost all were composed of methanogenic archaea ., Within the machine-learning process , these members of Methanomicrobia are considered simply as variables in the microbial community structure data and their shared characteristics are not recognized ( although humans would recognize properties common to all of them given that “Methano-” is at the beginning of each of their names ) ., All high-scoring samples were associated with a large proportion of topic #8 related to methanogenesis ( S2 Fig ) ., Therefore , the fact that samples containing many methanogenic archaea were retrieved after querying for “oil sands tailing ponds” indicates that LEA can automatically extract the following two linkages:, 1 ) the association between words such as “hydrocarbon , ” “oil , ” “tailing , ” and “methane , ” and the latent environmental topic that represent “methanogenesis , ” and, 2 ) the association between methanogenic archaea of various genera and the latent environmental topic that represent “methanogenesis” ., For this study , we applied a correspondence topic model to more than 30 , 000 samples of microbial community structure data and extracted the latent environments of each sample as topics ., By doing so , we obtained microbial sub-communities that can be regarded as “base variables” to describe an entire dataset and associated word subsets that characterize the environments corresponding to the base variables ., By visualizing each sample , which is expressed as a linear combination of these base variables , in two-dimensional space , LEA clarifies continuous variation of the microbial community structures linking two or more environments ., The difference between continuously connected environments and an isolated environment might mean that only a few samples have been characterized that bridge the isolated environments ., Such environments currently include wastewater , the phyllosphere , and environments related to insect symbiosis ., Conversely , human-related environments have been vigorously sampled , and therefore we believe that the visualization reported in this manuscript represents a nearly complete picture of those environments related to healthy human adults ., Using the extracted environmental topics , LEA can infer what mixture of core environments influenced the taxonomic compositions of newly acquired samples ., In the river microbiome analysis , we showed that samples taken from the brackish water area of the Tamagawa river can be expressed as a mixture of a “freshwater” topic , a “seawater” topic and a “wastewater” topic ., Environmental prediction of new samples is performed by a Bayesian approach similar to that used in a microbial source tracking algorithm52 , but using topic sub-communities extracted from a large-scale dataset as source communities , instead of using the samples pre-specified by a user as sources ., This allows to compare new samples virtually with tens of thousands of samples related to diverse natural environments and human body sites ., Environmental prediction of new samples may be done by fixing the granularities of environmental labels to be used and comparing with samples to which those labels are added in advance53 ., In such a method , however , it is difficult to set the level of granularities , especially when there are multiple structural patterns of microbial communities in a single environment ., When analyzing the dynamics of community structures in a single environment , for example , the time series analysis of human gut microbiome or the spatial distribution of river microbiome , it is more useful to use fine-grained environmental labels than to use simple labels such as “river” or “human gut” ., Our method clarifies the structural patterns naturally existing in various environments and provides the way to evaluate how new samples transit among them ., By using a neural network algorithm that maps the data to a two-dimensional space , LEA can position new samples onto the existing global map ., This mapping system can be regarded as a microbial global positioning system54 used to specify the position of a new sample based on the positions of existing sample and allows a user to intuitively evaluate the properties of new samples ., Dysbiosis , a deviation from the ordinary distribution of a microbial community structure that exists in symbiosis with humans , has been discussed in relation to diseases , especially those of the gut55 ., Because a newly acquired sample , such as one from an ill patient , can be located anywhere on the map , identifying the ideal end-point from a clinical perspective and defining its vector may be useful information when choosing a specific treatment that can transition its community structure to another state54 ., To perform comparative metagenomics based on environmental information , a huge amount of environmentally labeled data ordered as a dataset is required ., However , manually labeling such data is nearly impossible , as the amount of available data is increasing too rapidly at present ., In addition , as microbial community structures from new environments are characterized , much work will be needed to design the ontologies of the corresponding environmental labels at the appropriate granularities while incorporating all new environments ., Furthermore , because binary environmental labels ( presence or absence of an environmental property ) are often used to characterize the samples , it is not possible to manually and appropriately evaluate samples that have intermediate properties associated with several environments ., Our method automatically extracts the relationship between microbes and their environments by assuming that the microbial community structure and the natural language description for a given sample are both generated from a state in which several environments are mixed ., The accuracy of the model should increase as more training data are incorporated ., Prior to extending our method for future work , several problems must be solved ., First is how many topics are needed to model microbiomes in highly diverse environments ., The number of topics in this study , 80 , is an arbitrarily determined value in a sense ., In fact , the prediction accuracy of the model for the validation set shows that 80 topics are still inadequate and that a more accurate model can be constructed by setting the more number of topics ( perplexity , S1 Fig ) ., However , increasing the number of topics may lead to overfitting , and too large a number of topics may make the map difficult to visualize and interpret ., Therefore , we aimed to explain the data with as few topics as possible while keeping the overall prediction accuracy ., We are not claiming that microbial communities can be explained by a combination of 80 patterns ., The model used in this study is a practical choice to facilitate the interpretation of the whole picture of the microbial community structures and to provide a tool to explore interesting patterns ., In the future , as the number of samples acquired from various environments increases , it will be necessary to set a larger number of topics ., Nevertheless , from the results of experiments with a large number of topics , the characteristics o | Introduction, Results, Discussion, Materials and methods | As data for microbial community structures found in various environments has increased , studies have examined the relationship between environmental labels given to retrieved microbial samples and their community structures ., However , because environments continuously change over time and space , mixed states of some environments and its effects on community formation should be considered , instead of evaluating effects of discrete environmental categories ., Here we applied a hierarchical Bayesian model to paired datasets containing more than 30 , 000 samples of microbial community structures and sample description documents ., From the training results , we extracted latent environmental topics that associate co-occurring microbes with co-occurring word sets among samples ., Topics are the core elements of environmental mixtures and the visualization of topic-based samples clarifies the connections of various environments ., Based on the model training results , we developed a web application , LEA ( Latent Environment Allocation ) , which provides the way to evaluate typicality and heterogeneity of microbial communities in newly obtained samples without confining environmental categories to be compared ., Because topics link words and microbes , LEA also enables to search samples semantically related to the query out of 30 , 000 microbiome samples . | In the past decade , microbiomes from various natural and human symbiotic environments have been thoroughly studied ., However , our knowledge is limited as to what types of environments affect the structure of a microbial community ., In the first place , how can we define “environments” , in particular , the environmental entities that are often continuously varying and difficult to discretely categorize ?, We assumed that environments could be represented from microbiome data because the structure of microbial communities reflect the state of the environment ., We applied a probabilistic topic model to a dataset containing taxonomic composition data and natural language sample descriptions of >30 , 000 microbiome samples and extracted “latent environments” of the microbial communities , which are core elements of environmental mixtures ., Integrating the training results of the model , we developed a web application to explore the microbiome universe and to place new metagenomic data on this universe like a global positioning system ., Our tool shows what kinds of the environment naturally exist and are similar to each other on the perspective of the structural patterns of microbiome , and provides the way to evaluate the commonality and the heterogeneity of users’ microbiome samples . | taxonomy, ecology and environmental sciences, surface water, microbiome, rivers, community structure, microbiology, data management, metagenomics, aquatic environments, bodies of water, microbial genomics, research and analysis methods, hydrology, sequence analysis, computer and information sciences, bioinformatics, medical microbiology, marine and aquatic sciences, controlled vocabularies, biological databases, microbial taxonomy, community ecology, freshwater environments, sequence databases, ecology, earth sciences, database and informatics methods, genetics, biology and life sciences, genomics | null |
journal.pntd.0000845 | 2,010 | CD8+ T Cells as a Source of IFN-γ Production in Human Cutaneous Leishmaniasis | Leishmaniasis is expanding both by increasing the incidence rate in endemic foci and extending the disease to new regions 1 , 2 ., Control measures against leishmaniasis are not fully effective , chemotherapy is not always successful , and drug resistant is emerging 3–5 ., Although theoretically development of an effective vaccine against leishmaniasis is feasible but yet there is no vaccine available against any form of leishmaniasis 6 , 7 ., CD4+ T cells upon activation differentiate into functional effector Th1 and/or Th2 subsets and the outcome of Leishmania major infection in murine model is dependent upon the type of immune response generated: in most strains of mice L . major infection induces a Th1 type of response associated with a high level of IFN-γ , low level of IL-4 , and similar to human cutaneous leishmaniasis the lesion ( s ) heals spontaneously and the animals are protected against further infection; whereas L . major infection in BALB/c mice induces a Th2 response and a high level of IL-4 and low level of IFN-γ , as a result the disease is fetal 8 , 9 ., The mechanism ( s ) of protection in human leishmaniasis is not well characterized; however , the role of T lymphocytes and Th1/Th2 cytokine profile are extensively studied 10–16 ., In human leishmaniasis , peripheral blood mononuclear cells ( PBMC ) are routinely collected from patients with different clinical pictures of cutaneous leishmaniasis ( CL ) for immunological investigations ., Results from the majority of these studies showed that PBMC of healing or cured cases of CL produce significant amount of IFN-γ in response to Leishmania antigens 17 , 18 ., There is evidence demonstrating CD4+ T cells collected from patients with CL or mucocutaneous leishmaniasis ( ML ) or individuals with history of CL produced a high level of IFN-γ in response to Leishmania antigens which is an indication of a Th1 like response 10 , 11; Conversely , T cells from patients with diffuse CL ( DCL ) failed to express IL-2 receptor and did not produce IFN-γ in response to Leishmania antigens , whereas IL-4 mRNA markedly increased in DCL lesions 17 , 18 ., A clear Th1/Th2 dichotomy similar to murine model is not yet defined in human leishmaniasis 19 ., There are reports which showed that CD8+ T cells play a role in controlling intracellular pathogens including protozoal and viral infections ., CD8+ T cells are shown to confer a significant role in protection against acute and chronic form of Toxoplasma gondii infection 20 ., In early stage of murine toxoplasmosis , CD8+ T cells hamper parasite dissemination by either direct lysis of infected cells or through release of cytokines ., During chronic infection CD8+ T cells limit Toxoplasma cyst formation in tissues 21 , 22 ., Immunity against malarial sporozoites is mediated partially by neutralizing antibodies , but largely depends on antigen specific CD8+ T cells , thus vaccines are designed based on induction of infection-blocking CD8+ T cells 23 , 24 ., CD8+ T cells are also important in the control of HIV infection 25–27 ., During HIV infection , CD8+ T cells recognize infected cells through an MHC-I dependent process and viral infected cells are lysed by secretion of perforin and granzymes 28 ., Most patients chronically infected with HIV show CD8+ T cell response against HIV virus , but the response is not enough to successfully control viral replication 25–27 ., In Listeria monocytogenes infection , both CD4+ and CD8+ T cells contribute in induction of protection , but the major bactericidal role is attributed to CD8+ T cells 29 ., In experimental models of leishmaniasis , CD8+ T cells , in cooperation with CD4+ T cells , appear to be involved in the induction of host immunity against both primary infection and reinfection of Leishmania parasite 30–32 ., In L . major infected CD8+ depleted BALB/c mice , during lesion healing the frequency of IFN-γ producing CD4+ T cells and the amount of IFN-γ are diminished resulted in a higher parasite burden 30 ., In a study performed on C57BL/6 mice , infection with low dose of L . major induces a transient Th2 type response and then shifts to a Th1 response associated with healing ., Induction of this Th1 type of response partly depends on the activation of IFN-γ producing CD8+ T cells and in the absence of CD8+ T cells , the Th2 response is sustained 33 ., In mouse model of both genetically resistant and susceptible ( that were rendered resistant ) backgrounds , CD8+ T cells have been demonstrated to produce IFN-γ and contribute to the rapid healing of secondary lesions which develop after primary challenge with L . major 31 ., There are reports from New World leishmaniasis which showed that CD8+ T cells are involved in healing process of CL due to L . braziliensis 34–37 ., However , to our knowledge there is no data available about the possible role of CD8+ T cells and their cytokines in CL due to L . major ., In leishmaniasis , most of the data generated so far is drawn from PBMCs culture without separation of T cell subtypes 10 , 12 , 13which makes it difficult to judge the role of Th1/Th2 CD4+ cells and CD8+ T cells ., In the current study two major lymphocyte subtypes , CD4+ and CD8+ T cells , were purified from individuals with history of self-healing CL and cytokine pattern were analyzed at transcript and protein levels in response to Leishmania antigens ., The study was approved by Ethical Committee of Tehran University of Medical Sciences ., Potential candidates were invited and those who were willing to participate and sign a written informed consent were recruited ., Fourteen volunteers with history of self-healing CL ( HCL ) caused by L . major and with leishmanin skin test ( LST ) more than zero and as control 18 healthy volunteers from non-endemic area with no response to LST were included ., HCL volunteers were selected among the previous Centers patient who received no treatment for the CL lesion ( s ) and the lesion ( s ) cured spontaneously within one year of onset ., The causative agent of every CL patient was previously identified as L . major using PCR method ., Leishmania major ( MRHO/IR/75/ER ) was cultured on NNN medium and passaged on RPMI 1640 ( Gibco Invitrogen , Carlsbad , CA , USA ) supplemented with 10% fetal calf serum ( FCS ) ., Promastigotes were harvested at day 5 , washed 3 times with PBS ( pH 7 . 2 ) and used for preparation of soluble Leishmania antigen ( SLA ) as previously described 14 ., Briefly , protease inhibitor cocktail enzyme ( Sigma , St . Louis , MO , USA ) was added at 100 µl per 1×109 promastigotes , and then the parasites were freeze-thawed 10 times followed by sonication at 4°C with two 20-sec blasts ., Parasite suspension was centrifuged at 30 , 000×g for 20 min , the supernatant was collected and re-centrifuged at 100 , 000×g for 4 hours ., SLA protein concentration was measured using Bradford method 38 ., Finally the supernatant was sterilized using 0 . 22 µm membrane filter , aliquoted and stored at −20°C until use ., Twenty mL of blood sample was collected from each volunteer and Peripheral Blood Mononuclear Cells ( PBMCs ) were isolated using Ficoll–Hypaque ( Sigma , St . Louis , MO , USA ) density gradient centrifugation ., CD4+ and CD8+ lymphocytes isolation was performed using magnetic beads system ( StemCell Technologies Inc . , Vancouver , BC , Canada ) by positive selection using anti-CD4 or anti-CD8 coated nanoparticles ., Briefly , cell suspension was prepared at a concentration of 1×107 cells/ml in a 5 ml tube in isolation buffer containing PBS plus 2% ( v/v ) FBS and 1 mM EDTA ., EasySep CD4/CD8 cocktail Abs was added at 10 µl/ml cells , mixed well and incubated at room temperature ( RT ) for 15 min ., Magnetic nanoparticles were added at 5 µl/ml cells and incubated for 10 min at RT ., The cell suspension was brought to 2 . 5 ml by adding buffer and the tube was placed into the magnet for 5 min , then the supernatant was discarded ., The desired cells were remained bound inside the tube ., The steps of placing tube into the magnet were repeated three times ., Monocytes ( CD14+ ) were isolated from autologous PBMC by negative selection according to the manufacturers instruction ( StemCell Technologies Inc . , Vancouver , BC , Canada ) ., Briefly , cell suspension was prepared at a concentration of 5×106 cells/ml in isolation buffer ., EasySep monocyte enrichment cocktail Abs was added at 5 µl/ml cells , mixed well and incubated at 4°C for 10 min ., Magnetic microparticles were added at 5 µl/ml cells for 5 min at 4°C ., The cell suspension was brought to 2 . 5 ml by adding buffer and the tube was placed into the magnet , for 2 . 5 min at RT ., The desired unbound fraction was transferred into a new tube ., The purity of the yielded lymphocytes or monocytes was more than 95% by flow cytometry analysis using specific conjugated mAb ( Fig . 1 ) ., The contamination of CD8+ T cells with NK cells was less than 9% using α-CD56 mAb ., Monocytes were co-cultured with sorted lymphocytes as antigen presenting cells ( APCs ) following mitomycin C ( Merk , Darmstadt , Germany ) treatment at a final concentration of 10 µg/ml for 30 min at 37°C with 5% CO2 ., The cells were cultured in RPMI 1640 media supplemented with 10% heat-inactivated human AB Rh+ serum , 10 mM/L Hepes , 2 mM L-glutamine , 100 U/ml penicillin G and 100 µg/ml streptomycin ( Gibco Invitrogen , Carlsbad , CA , USA ) ., CD4+ or CD8+ lymphocytes were adjusted to 0 . 5–1×106 cells/ml mixed with 1∶10 of autologous monocytes and were cultured in U-bottomed 96-well plates ( Nunc , Roskilde , Denmark ) in the presence of either 10 µg/ml PHA or 50 µg/ml of SLA in a final volume of 200 µl ., Plates were incubated at 37°C with 5% CO2 in humidified atmosphere for 72 hrs ., Culture supernatants were collected at 72 hours , the level of IL-5 , IL-10 , IL-13 and IFN-γ were titrated in culture supernatants using ELISA method ( Mabtech , Stockholm , Sweden ) ., Briefly , the plates were coated with anti-IFN-γ/IL-5/IL-10/IL-13 mAb in PBS , pH 7 . 4 , and incubated at 4°C over night ., After blocking the wells using buffer containing PBS plus 0 . 05% ( v/v ) Tween 20 and 0 . 1% ( w/v ) BSA , supernatants were added to each well ., Biotin-labeled mAb in incubation buffer was added to each well and as enzyme streptavidin-HRP was used ., The reaction was developed using 3 , 3′ , 5 , 5′-tetramethylbenzidine ( TMB ) substrate and stopped with 0 . 5M H2SO4 solution ., The plates were washed after each step of incubation using PBS+0 . 05% ( v/v ) Tween20 ., The plates were read at 450 nm using a reader ( BioTek , Winooski , VT , USA ) ., The mean optical densities ( ODs ) of triplicate cultures were compared with the standard curves prepared using recombinant IL-5 , IL-10 , IL-13 and IFN-γ ., The cytokine levels represent the differences between the ODs of test and background wells ., The detection limit of the assays was 4 pg/ml for IL-5 and 0 . 5 pg/ml for IL-10 , 5 pg/ml for IL-13 and 2 pg/ml for IFN-γ ., After SLA stimulation , part of the cells was used for ICS assay ., Cells were adjusted at 5×105 per ml and stimulated with PMA ( Sigma , St . Louis , MO , USA ) 50 ng/ml plus Ionomycin calcium ( Sigma ) 500 ng/ml and incubated at 37°C , 5% CO2 for 5–6 hrs ., Monensin ( Sigma ) was added at 25 µM/ml during the last 4–5 hrs of culture for blocking ., Cells were harvested and washed 2 times with PBS ( pH 7 . 2 ) plus 0 . 1% bovine serum albumin ( BSA ) ., The cells were permeabilized using BD Cytofix/Cytoperm kit according to the manufacturers instruction ( BD Biosciences , San Jose , CA , USA ) ., In the final step , cells were stained with FITC-conjugated mouse anti-human IFN-γ and PE-conjugated rat anti-human IL-2 ( BD Biosciences , San Jose , CA , USA ) ., Cells were washed ×2 with perm/wash buffer and resuspended in PBS ( pH 7 . 2 ) plus 1% BSA ., Cells were analyzed using Partec flow cytometer ( DAKO cytomation , Glostrup , Denmark ) while isotype matched negative controls were used to set the threshold of autofluorescence ., A minimum of 50 , 000 events were acquired for each sample ., FACS data analysis was performed using FloMax ( DAKO cytomation ) software ., At the time of supernatants collection ( at day 3 ) , the SLA stimulated cells were harvested and used for RNA extraction ., The procedure began with reverse transcription of mRNA to cDNA ., The cDNA was then used as template for Real-time PCR using specific primer of each cytokine ., Solutions were treated and glassware was filled with 0 . 1% ( v/v ) diethylpyrocarbonate ( DEPC ) ( Merck , Darmstadt , Germany ) in H2O ., The cell pellet was resuspended in cold PBS ( pH 7 . 2 ) and lysed by addition of 0 . 2 mL of Trizol ( Sigma , St . Louis , MO , USA ) per 1×106 cells ., RNA was solublized through pipetting and incubated at RT for 5 min ., Then 0 . 2 ml chloroform was added per 1 ml of homogenate , the suspension was shook vigorously and kept on ice for 5 min followed by centrifugation for 15 min , 12 , 000×g at 4°C ., The upper phase was collected and added to an equal volume of isopropanol and incubated at 4°C over night ., Then the cell suspension was centrifuged at 12 , 000×g , 4°C for 15 min , the supernatant was discarded and the RNA pellet was washed with 1 ml 75% ethanol at 7 , 500×g for 8 min ., At the end , the pellet was drayed briefly and dissolved in DEPC treated water ., The purity of RNA samples was assessed by the ratio of ODs at 260/280 nm using UV spectrophotometry ., Reverse transcription was carried out using RevertAid M-MuLV enzyme ( Fermentas life sciences , York , UK ) in a 30 µL reaction mixture ., Briefly , 1 µL ( 0 . 5 µg ) of oligo dT18 primer was added to about 3 µg of total RNA , mixed and incubated at 70°C for 5–10 min ., The tube was placed on ice for a few minutes , centrifuged briefly and added with: 4 µL of 5× reaction buffer , 1 µL 10 mM dNTPs , 20 u RNase inhibitor ( RiboLock; Fermentas life sciences ) and DEPC treated water up to 30 µL ., Tube was incubated at 37°C for 5 min , the contents were mixed gently and then added with 100 U of enzyme ., The reaction mixture was incubated at 42°C for 60 min ., The enzyme was inactivated by heating at 70°C for 10 min . and chilled on ice ., In a Real-time PCR MicroAmp optical 96-well reaction plate ( Applied Biosystems , Foster City , CA , USA ) for 25 µl reaction mixture the followings in each optical well were prepared: 12 . 5 µl QuantiTect SYBR Green I ( Qiagen , Hilden , Germany ) , 3 µl cDNA , 2 µl primer pair mix ( 0 . 5 µM each ) , 7 . 5 µl dH2O ( For sequences of primer pairs see Table 1 ) ., For normalizing the difference in the amount of inputting cDNAs , the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase ( GAPDH ) was used as the internal standard ., Samples and internal standard were amplified in separate wells of the plate ., Two-step thermal profile as a PCR program was set up on the SDS software ( version 1 . 3 . 1 ) of Applied Biosystems 7 , 500 machine ( Applied Biosystems , Foster City , CA , USA ) ., The dissociation curve on the instrument software was set as follows:Software generates reports including amplification plots and dissociation curves ., Any bimodal dissociation curve or abnormal amplification plots were checked to see if there is an indication of different Tms and nonspecific products ., The Ct ( threshold cycle ) of each sample was used in gene relative expression calculation ., The 2−ΔΔCt method was used to calculate relative changes in the gene expressions:While To have a valid calculation of 2−ΔΔCt , the amplification efficiencies of the target and reference ( GAPDH ) genes must be approximately equal ., For this purpose , the Ct values variation with cDNA template dilutions was checked ., A pooled cDNA preparation was diluted over a 10-fold range and PCR was performed for each dilution using specific primers ., A plot of the log cDNA dilution versus CT was prepared ., The slope of each line was obtained from regression equation and the efficiencies of the target and reference genes were calculated using the equation: Efficiency ( E ) =\u200a10 ( 1/slop ) −1 Non-parametric tests of Mann-Whitney , Kruskal-Wallis and Dunns post-test for paired comparisons were used for statistical analysis of the data using SPSS version 11 . 5 ( SPSS Inc . , Chicago , IL , USA ) and GraphPad Prism version 5 . 01 ( GraphPad Software Inc . , La Jolla , CA , USA ) softwares ., Nonparametric tests were chosen because the samples did not follow a Gaussian distribution ., P value of <0 . 05 considered to be significant ., Using ELISA method , cytokine profile ( IFN-γ , IL-10 , IL-13 ) were measured on supernatants collected at 72 hrs of SLA stimulated PBMC or CD4+ , CD8+ T cells culture ., The amount of IFN-γ level was significantly higher in PBMC culture of HCL volunteers compared with that of healthy controls ( P<0 . 05 ) ., Results of purified T cell culture showed that stimulated CD4+ T cells from HCL volunteers induced a significantly higher IFN-γ production compared with cells from healthy controls ( P<0 . 05 ) ( Fig . 2A ) ., Similarly , stimulated CD8+ T cells from HCL volunteers induced a significantly higher IFN-γ production compared with the cells from healthy controls ( P<0 . 05 ) ( Fig . 2B ) ., The levels of IL-10 and IL-13 were not significantly different in either CD4+ or CD8+ T cells between HCL volunteers and healthy controls ( Fig . 2A and B ) ., IL-5 level was not detectable in culture supernatants of either CD4+ or CD8+ T cells ., The relative quantities of the target genes were normalized against the relative quantities of the internal standard ( GAPDH ) ., Ct values of amplified templates of antigenic stimulated T cells were used for calculation of different cytokine gene expressions using 2−ΔΔCT method ., The expression amount was compared with unstimulated cells of culture and relative fold-expression was reported ., Result for each donor was calculated and then data was pooled and presented as a mean of HCL volunteers against healthy controls ., Results showed that the upregulation of IFN-γ gene expression in CD4+ cells from HCL volunteers was significantly higher than that of healthy controls ( P<0 . 001 ) ( Fig . 3A ) ., Similarly , fold-expression changes of IFN-γ gene was significantly higher in CD8+ cells from HCL volunteers compared to the cells from healthy controls ( P\u200a=\u200a0 . 006 ) ( Fig . 3B ) ., Comparing CD4+ and CD8+ T cells , the significantly higher fold-expression of IFN-γ gene was seen in CD4+ cells than CD8+ cells of HCL volunteers ., In both CD4+ and CD8+ T cell cultures , the changes in the gene expression of IL-5 , IL-10 and IL-13 were not significantly different between HCL volunteers and healthy controls ., By amplification of serially diluted pooled cDNA , the amplification efficiency of the target ( cytokines ) compared to reference ( GAPDH ) genes was examined using SYBR Green detection ., Using the equation pointed out in methods , efficiency of GAPDH was 96% while that of targets were between 91% and 95% ., Seventy two hrs after SLA stimulation of CD4+/CD8+ T cells , part of the cells were harvested and stimulated with PMA plus Ionomycin for 5–6 hrs , stained for intracellular IFN-γ with conjugated mAbs and the frequency of positive cells was analyzed using flow cytometry ., In CD8+ cells compartment , antibody to CD56 marker allowed to exclude IFN-γ positive populations of natural killer cell sources ( Fig . 4 A and B ) ., Results of analysis of cells from HCL volunteers and healthy controls were pooled separately and presented as median number of intracellular IFN-γ positive CD4+ and CD8+ T cells ( Fig . 4C ) ., Based on this analysis , HCL volunteers showed that a significantly higher number of CD4+ T cells were positive for intracellular IFN-γ production than CD8+ cells ( P\u200a=\u200a0 . 014 ) ., Resistance and susceptibility to L . major infection in murine model depend upon induction of Th1 or Th2 response , respectively 8 , 9 , 39 ., Recovery from CL usually is accompanied with long lasting protection and strong immune response generation indicated by in vivo LST and in vitro lymphocyte response to Leishmania antigens 10 , 12 , 13 , 19 , 40 , yet in human leishmaniasis the surrogate marker ( s ) of protection is not well defined ., Most of the studies performed on human immune response against leishmaniasis is carried out on crude PBMCs without purifying CD4+ T cell and CD8+ population and there is no report to show a clear-cut CD4+ Th1/Th2 response 34 ., In the current study cytokines patterns of CD4+ Th1/Th2 and CD8+ T cells in volunteers recovered from CL is evaluated at the transcript and protein levels ., The results of soluble Leishmania antigens ( SLA ) stimulated cells showed that pure CD4+ T cells induced a significantly higher IFN-γ production in HCL volunteers compared to that of healthy controls , IL-5 as a Th2 type cytokine was not detectable and IL-10 and IL-13 levels were not significantly different in culture of T cells from HCL volunteers compared with that of healthy controls ., At the same time , the level of cytokines mRNA expression was evaluated to detect T cell cytokine responses to Leishmania stimulation at transcript level , after several experiments to explore the optimum stimulation time points for mRNA analysis ., Simultaneous analysis of cytokines gene expression showed a strong up-regulation of Th1 cytokine IFN-γ mRNA in CD4+ T cells ., In line with the results of secreted proteins , level of Th2 cytokines transcripts including IL-5 , IL-10 and IL-13 showed no significant increase in HCL volunteers compared with healthy control ., The up-regulation of IFN-γ transcripts expression in Leishmania stimulated CD4+ T cells consistent with IFN-γ secretion is an indication of Th1 type of response in HCL volunteers parallel with no Th2 response indicating by low level of IL-5 , IL-10 and IL-13 cytokines ., Analysis of cytokine secretion and transcript expression of SLA stimulated CD8+ T cells showed also a significantly higher IFN-γ production in HCL volunteers compared to the healthy control volunteers ., Similar to CD4+ T cells , the levels of IL-10 and IL-13 were not significantly different in CD8+ T cell culture between HCL volunteers and healthy controls ., When CD4+ and CD8+ T cell response was compared , the level of IFN-γ secretion in SLA stimulated cells was not significantly different between CD4+ and CD8+ cells , but real-time PCR analysis revealed that expression level of IFN-γ mRNA was higher in CD4+ T cells than CD8+ T cells ., To confirm the real-time PCR results , part of CD4+/CD8+ sorted T cells were harvested following SLA stimulation and intracellular production of IFN-γ was assessed using flow cytometry ., Results of intracellular cytokine staining ( ICS ) in HCL volunteers confirmed that a significantly higher number of CD4+ T cells produced intracellular IFN-γ compared with CD8+ T cells ( median\u200a=\u200a15% vs . 8% ) ., Based on the results it seems that the source of IFN-γ production is both CD4+ Th1 cells and CD8+ cells in individuals with history of CL ., The role of CD8+ T cells in human Leishmania infection is not well known and existed reports are controversial ., In a study performed on Sudanese individuals it was suggested that IFN-γ production is associated with CD4+ T cells rather than CD8+ T cells in individuals with history of CL due to L . major 40 ., A report from New World leishmaniasis showed that in both asymptomatic and antimonial treated CL individuals caused by L . braziliensis , a higher proportions of CD4+ than CD8+ T cells was present 34 ., In another report the authors showed that after treatment of CL due to L . braziliensis , the frequency of CD4+ and CD8+ T cells was the same with approximately constant production of IFN-γ 41 ., On the other hand , some clinical studies reported high numbers of Leishmania specific CD8+ T cells in the lesions and peripheral blood during acute phase and healing process in L . major or L . braziliensis CL patients 42 , 43 ., In mice , the requirement of CD8+ T cells for the control of L . major infection is shown to be partly dependent on the procedure of challenge: β2-microglobuine or CD8+ deficient C57BL/6 mice when challenged with high dose of L . major have the ability to cure the lesion , which indicates that CD8+ T cells are not necessary for the control of primary 44 infection , while in the intradermal challenge with low dose ( 100 metacyclic promastigotes into the ear dermis ) the outcome of primary L . major infection in anti-CD8 Ab treated or CD8 deficient mice was dependent on the CD8+ T cells 45 ., The role of CD8+ T cells was studied in CBA and anti-CD4 mAb treated BALB/c mice healed from L . major infection ., The cured mice were re-challenged with L . major in the contralateral footpad and lymph nodes cells were depleted of CD4+ T cells and stimulated in vitro ., The remaining CD8+ T cells produced a significant amount of IFN-γ 31 ., It is believed that in the resolution of the primary Leishmania infection and induction of protection in murine model CD8+ T cells play an important role 31 ., In the present study , following the isolation of CD4+/CD8+ T cells , Th1/Th2 cytokines were titrated on culture supernatant of in vitro restimulated T cells to check the type of immune response elicited against Leishmania antigen ., The main cytokine produced was found to be IFN-γ in the volunteers T cells ., IFN-γ eliminates intracellular pathogens primarily through macrophage activation ., Macrophages upon activation produce nitric oxide ( NO ) which inhibits growth of intracellular pathogens ., It is shown that during active lesion of CL due to L . major the proliferative response and IFN-γ production of PBMC was increased 12 , 46 and T cells from healed CL produced a significantly higher level of IFN-γ but a low level of IL-10 than the cells from controls 14 , 15 , 40 ., Similarly , studies in L . braziliensis infection demonstrated a Th1/Th2 mixed response in early stage of active CL 11 , 42 and then a sustained Th1 response with elevated level of IFN-γ and down-regulation of IL-4 and IL-10 production were seen apparently associated with healing 11 ., Likewise , the presence of high level of IFN-γ in the skin lesions of CL patients support the role of IFN-γ in healing process 46 ., Using RT-PCR , the cytokine patterns of skin lesions of CL patients showed predominance IFN-γ , and low levels of IL-5 and IL-10 17 ., The contribution of IFN-γ to the recall of immunological memory against L . major reinfection was assessed in mice ., The neutralization of IFN-γ at the time of reinfection reduced the specific DTH response , showing the involvement of IFN-γ in the recall of memory response to L . major 31 ., Similarly in intracellular infection with T . gondii , it is shown that CD8+ T cells confers resistance against acute infection 20 and IFN-γ producing CD8+ T cells play a significant role in controlling chronic T . gondii infection and inhibits encephalitis in mouse model 21 , 22 ., In the current study , even though using real-time PCR the expression level of IFN-γ transcripts in CD8+ cells was less than CD4+ cells , but interestingly a significant amount of IFN-γ was produced by CD8+ T cells in cell culture and around 5–12% of CD8+ cells was positive for IFN-γ secretion by ICS assay ., It is concluded that CD8+ T cells contribute along with CD4+ Th1 cells in IFN-γ production in individuals with history of CL ., Despite the limited reports of CD4+ Th1 cells as the main source of IFN-γ production in CL patients 47 , 48 in most studies of CD4+ Th1/Th2 paradigm in human CL , PBMCs rather than purified T cells were used , hence the role of IFN-γ producing CD8+ T cells should not be ruled out when reporting a “Th1” type response in PBMC culture ., The strong lymphoproliferative and IFN-γ response in self healing CL caused by L . braziliensis is previously shown 49 , 50 ., In the current study , HCL volunteers with spontaneous healing during 1 . 5–5 months were recruited ., Individuals with history of self healing CL are presumed to be protected against further Leishmania infection ., The blood samples were collected a few months to years after cure of CL lesions ., The strong LST response ( mean LST\u200a=\u200a10 . 7±7 . 5 mm ) and IFN-γ production is an indication of sustaining cell mediated immune response ., This sustaining protective immunity is mediated not only through the expansion of antigen-specific IFN-γ producing CD4+ Th1 cells , but also through IFN-γ producing CD8+ T cells ., The question that which one of these T cell subsets plays a more important role in IFN-γ production at the initiation of exposure to sand fly bite needs to be explored . | Introduction, Materials and Methods, Results, Discussion | In human leishmaniasis Th1/Th2 dichotomy similar to murine model is not clearly defined and surrogate marker ( s ) of protection is not yet known ., In this study , Th1/Th2 cytokines ( IL-5 , IL-10 , IL-13 and IFN-γ ) profile induced by purified CD4+/CD8+ T cells in response to Leishmania antigens were assessed at transcript and protein levels in 14 volunteers with a history of self-healing cutaneous leishmaniasis ( HCL ) and compared with 18 healthy control volunteers ., CD4+/CD8+/CD14+ cells were purified from peripheral blood using magnetic beads; CD4+/CD8+ T cells were co-cultured with autologous CD14+ monocytes in the presence of soluble Leishmania antigens ( SLA ) ., Stimulation of either CD4+ T cells or CD8+ T cells of HCL volunteers with SLA induced a significantly ( P<0 . 05 ) higher IFN-γ production compared with the cells of controls ., Upregulation of IFN-γ gene expression in CD4+ cells ( P<0 . 001 ) and CD8+ cells ( P\u200a=\u200a0 . 006 ) of HCL volunteers was significantly more than that of controls ., Significantly ( P<0 . 05 ) higher fold-expression of IFN-γ gene was seen in CD4+ cells than in CD8+ cells ., In HCL volunteers a significantly ( P\u200a=\u200a0 . 014 ) higher number of CD4+ cells were positive for intracellular IFN-γ production than CD8+ cells ., Collectively , the volunteers have shown maintenance of specific long-term immune responses characterized by a strong reaction to leishmanin skin test and IFN-γ production ., The dominant IFN-γ response was the result of expansion of both CD4+ and CD8+ T cells ., The results suggested that immune response in protected individuals with a history of zoonotic cutaneous leishmaniasis ( ZCL ) due to L . major is mediated not only through the expansion of antigen-specific IFN-γ producing CD4+ Th1 cells , but also through IFN-γ producing CD8+ T cells . | Cutaneous leishmaniasis ( CL ) is usually a self-healing skin lesion caused by different species of Leishmania parasite ., Resistance and susceptibility of mice to Leishmania major infection is associated with two types of CD4+ T lymphocytes development: Th1 type response with production of cytokine IFN-γ is associated with resistance , whereas Th2 type response with production of cytokines IL-4 and IL-5 is associated with susceptibility ., A clear Th1/Th2 dichotomy similar to murine model is not defined in human leishmaniasis and we need as much information as possible to define marker ( s ) of protection ., We purified CD4+/CD8+ T cells , stimulated them with Leishmania antigens and analysed gene and protein expression of Th1/Th2 cytokines in volunteers with a history of self-healing CL who are presumed to be protected against further Leishmania infection ., We have seen significant upregulation of IFN-γ gene expression and high IFN-γ production in the Leishmania stimulated CD4+ T cells and CD8+ T cells ., We concluded that both antigen-specific IFN-γ producing CD4+ Th1 cells and IFN-γ producing CD8+ T cells contribute to the long term protection in individuals with a history of CL ., This proves the importance of CD8+ T cells as a source of IFN-γ in Th1-like immune responses . | immunology/immune response | null |
journal.pcbi.1005966 | 2,018 | A computational model for how cells choose temporal or spatial sensing during chemotaxis | Chemotaxis is the process whereby cells move towards a region of higher chemical stimulus concentration ., Cellular movements towards the favorable direction enables , for example , prokaryotic unicellular organisms such as Escherichia coli ( E . coli ) to move towards food and eukaryotic cells such as neutrophils and macrophages to move towards the site of infection to phagocytize external parasites ., Information about the external chemical gradient is transduced into the cell by binding of chemoattractant and chemorepellant molecules to specific receptors at the cell surface ., These binding events then trigger downstream intracellular signaling to modulate the cell’s motility ., To move up or down the gradient , cells can adopt two distinct strategies: temporal sensing or spatial sensing ( Fig 1 ) ., In temporal or sequential sensing , cells compare the intensity of receptor stimulation at different times ( Fig 1 , left ) and modulate their probability of moving in the same direction or switching directions ., In E . coli , an organism exhibiting temporal sensing , rotation of its flagella in the counter-clockwise direction results in directed motion whereas rotation in the clockwise direction results in tumbling and a random change in direction 1 , 2 ., Binding of chemoattractant decreases the switching probability from counter-clockwise to clockwise rotation , thus reducing tumbling and increasing the run length when the cell is moving in the favorable direction ., In spatial sensing , cells simultaneously measure the intensity of receptor stimulation at its two ends ( Fig 1 , right ) ., The different receptor stimulation leads to cell polarization and motility in the preferred direction ., In neutrophils , G protein-coupled receptors ( GPCRs ) are originally evenly distributed along the plasma membrane ., Binding of chemoattractant results in activation of signaling pathways involving small Rho guanosine triphosphatases ( Rho GTPases ) and phosphoinositide 3 kinases ( PI3Ks ) and asymmetric polymerization of actin at the up-gradient edge of the cell , facilitating motion up the gradient 3 ., The decision whether to employ temporal or spatial sensing has largely been attributed to cell size ., It is thought that large cells have an advantage for spatial sensing as the intensities of receptor stimulation are expected to be very different at its two ends ., In contrast , small cells of around or less than a micron in diameter are unable to exhibit spatial sensing as chemical gradients are rapidly homogenized by fast diffusion ., For example , in an E . coli of 2um , the cytoplasmic CheY chemotaxis signal transduction protein with a diffusion constant of 4 . 6±0 . 8um2s−14 will take only 0 . 9s to transerve the cell ., However , spatial localization of MinC , MinD and MinE proteins to bring about proper cell division 5 and polar localization of the chemoreceptor complex of cytoplasmic CheA and CheW proteins 6 suggest that spatial segregration of proteins can be established at the micron scale in small cells ., Berg and Purcell also showed theoretically that , in principle , an immobile E . coli cell is able to perform spatial sensing 7 ., Dusenbery , based on arguments of signal-to-noise ratio , also found that the cell size limit for spatial sensing ( < 1um ) is close to that for temporal sensing and is actually smaller than the size of many prokaryotes 8 ., These works cast doubts on previous arguments for the inability of small cells of around a micron in diameter to perform spatial sensing and suggests that most cells , whether big and small , are able to perform both spatial and temporal sensing ., Here , we use computational model to show that the decision to perform either temporal or spatial sensing is instead determined by the performance of each type of sensing ., To determine the performance of temporal and spatial sensing , we need to integrate the sensing mechanism with the network circuits use for chemotaxis ., A key goal in systems biology is to identify network motifs capable of achieving certain biological function ., For chemotaxis to be effective , cells need to exhibit adaptation ., Adaptation refers to a cell’s ability to respond to a change in the input stimulus and then return to its original level , even when the input stimulus remains high ., This property allows cells to respond to a high range of chemoattractant concentration ., Extensive efforts to understand the ability of E . coli to remain sensitive to a wide range of chemoattractant has led to the identification of the negative integral feedback ( NFB ) circuit for chemotaxis 9 , 10 ( Fig 2 , step 1 , left ) ., In NFB , following stimulation of the output protein ( protein C ) by the activator ( protein A ) , a buffering component/inactivator ( protein B ) integrates the difference between the response and the baseline level and feeds this difference back into the response , enabling the output protein to return to the basal level after each pulse of chemoattractant ., On the other hand , modeling efforts in eukaryotic gradient sensing have identified the incoherent feedforward ( IFF ) circuit ( Fig 2 , step 1 , right ) for amplification of the signaling response to shallow gradients 11–13 ., In IFF , two nodes , an activator ( protein A ) and a repressor ( protein B ) , are activated proportionally to the stimulus but act with opposite effects on the output protein ( protein C ) ., Like the NFB , the IFF circuit also has the adaptive property needed for sensing a wide range of chemoattractant ., A comprehensive survey of all possible three-node network topologies had been carried out to search for networks that yield biochemical adaptation response 14 ., They found that minimal circuits containing NFB and IFF motifs yield adaptation and that more complicated circuits that yield adaptation contain at least one of these two motifs ., Hence we will use the NFB and IFF circuits to study cells’ chemotaxic response as they are the basic building blocks for three-node circuits that can yield adaptative property , an essential property for chemotaxis ., We compare the performance of temporal and spatial sensing when a cell uses the NFB and IFF circuits by determining the conditions that favor one mode of sensing over the other ., In temporal sensing , the cell compares the level of C with the steady state level of C ( area highlighted in green ) ( Fig 2 , step 4 , temporal ) whereas in spatial sensing , the cell compares the level of C at the front half and back half of the cell ( area highlighted in red ) ( Fig 2 , step 4 , spatial ) ., We identify five dimensionless terms , namely the diffusivities of the activator ( protein A ) , repressor ( protein B ) and output ( protein C ) proteins , all normalised to the deactivation rate of the output protein , the effective chemoattractant gradient experienced by the moving cell , and the ratio of cell speed to the product of diameter and signaling rate that characterize the response of the negative integral feedback and incoherent feedforward circuits ., By varying these five terms and comparing the performance of temporal and spatial sensing on the negative integral feedback and incoherent feedforward circuits , we find that spatial sensing performs better than temporal sensing in the regime where the cell velocity is small relative to the product of cell diameter and the circuit reaction rate , and when the repressor protein ( protein B ) diffuses faster than protein A ) and the diffusibility of the output protein ( protein C ) is low ., In all other cases , temporal sensing performs better ., By incorporating noise into our analysis , we also found that temporal sensing is more robust to noise than spatial sensing ., Here , we want to determine whether cell size is the determining factor or there are other factors contributing to the choice between temporal and spatial sensing ., We assume that the mode of sensing that yields higher signaling output will be adopted by cells ., In general , the signaling output will depend on both the signaling ( i . e . , molecular ) and physical properties of the chemotactic cell as well as the properties of the chemoattractant ., Hence , we need to identify these important variables and determine how they affect the signaling outputs ., However , one obstacle is that , very often , the values of these variables have not been measured experimentally ., Thus , we will adopt a network motifs approach where the exact parameter values are not so critical as long as the parameter values lie within certain regimes , since the same behavior is typically observed over a range of parameter vaues ., Hence , we will identify all possible behaviors of the networks by sweeping through parameter space ., This approach has been widely adopted in modeling papers ( e . g . , Ma et al . , 2014 ) ., In our analysis , we have swept through 4 to 5 orders of magnitude of parameter values and obtained the perfect adaptive behavior expected for the network motifs ., Our analysis consists of four steps ( Fig 2 ) ., We first set up the equations for the negative integral ( NFB ) and incoherent feedforward ( IFF ) circuits ( Fig 2 , step 1 ) and identify 100 sets of parameters that lead to high sensitivity and adaptation precision for these circuits ( Fig 2 , step 2 ) ., High sensitivity is responsible for signal amplication in shallow gradients whereas high adaptation precision is required for signal adaptation ., These are properties that enables a cell to perform chemotaxis effectively ., Next , we determine the protein dynamics as the cell moves through a linear gradient for the sets of parameters that we have identified in step 2 for the NFB and IFF circuits ( Fig 2 , step 3 ) ., The cell is modeled as a one dimensional ring , of diameter d , with an activator protein ( A ) , inactivator protein ( B ) and output protein ( C ) that can diffuse freely on the cell membrane ., At time τ = 0 , the cell moves with velocity , v , into a linear chemoattractant gradient with slope , k ., The cell experiences the gradient for a fixed time , Ts , before moving into a region with constant I = IH ., The cell uses both the NFB and IFF circuits to process the chemoattractant input and interprets the results using temporal or spatial sensing ( Fig 2 , step 4 ) ., More details can be found in S1 Text ., Extending the equations for incoherent feedforward and negative integral feedback circuits to account for spatial differences of protein levels on the cell’s membrane and a changing external chemoattractant gradient ( see S1 Text ) , we find that the equations are fully described by the following variables: cell diameter d , cell velocity v , chemoattractant gradient k ( which has the unit of inverse length ) , signaling rates , of which we choose lBC , the deactivation rate of C , to be representative ( i . e . , other signaling rate can be expressed as ratios of it ) , and the activator ( A ) , inactivator ( B ) , and output protein ( C ) diffusivities , DA , DB , and DC , respectively ., These variables can be grouped into the following five dimensionless variables below ., The five dimensionless variables are as follows: First , we consider the effect of D C ′ on the choice between temporal versus temporal sensing ., For spatial sensing , the cell compares the level of protein C at different parts of the cell ., Therefore , C has to diffuse slowly to allow for spatial sensing ., When D C ′ is big , any spatial information will be rapidly homogenized and temporal sensing will be favored ., We use D C ′ = 0 for our analysis to study the effects of other parameters on the sensing choice ., We hypothesize that α would not affect signaling outcome as a steeper or more gentle external gradient would affect the output from both sensing choice equally ., To test this hypothesis , we vary α = 0 . 00001 , 0 . 0001 , 0 . 001 , 0 . 01 for D A ′ = 1 ., 0 , D B ′ = 100 ., 0 , D C ′ = 0 and β = 0 . 125 , 0 . 5 , 2 . 0 , 8 . 0 ., For each set of parameters , we systematically simulate the dynamics for the selected sets of parameters and determine the output for spatial and temporal sensing ., The strategy yielding the higher output will be selected ., As shown in S1, ( a ) and S1, ( b ) Fig , α does not affect the choice of temporal and spatial sensing ., We also plot the output using temporal sensing ( green ) versus spatial sensing ( red ) at β = 0 . 125 ( S1c Fig ) and β = 8 . 0 ( S1d Fig ) for different values of α ., We observe that both the outputs scale linearly with the increase in α ., Since α affects both outputs equally , it does not affect the sensing choice ., Since α does not affect the sensing choice , we have fixed α = 0 . 001 and focus on the effects of D A ′ , D B ′ and β ., We simulate the protein dynamics for β = 0 . 125 , 0 . 25 , 0 . 5 , 1 . 0 , 2 . 0 , 4 . 0 , 8 . 0 , D A ′ = 0 ., 1 , 1 ., 0 , 100 , 1000 and D B ′ = 0 ., 1 , 1 ., 0 , 100 , 1000 ., We plot the percentage of runs that yield higher signaling output adopting the temporal ( green ) and spatial ( red ) strategy for different values of β in Fig 3, ( a ) and 3, ( b ) ., Although the negative integral feedback ( NFB ) and incoherent feedforward circuits ( IFF ) have been associated with temporal 9 , 10 and spatial sensing 11 , 12 respectively , we find that the two circuits yield similar results ., This shows that NFB can be used for spatial sensing and that the IFF can be used for temporal sensing ., We find that when β is high ( cell velocity is high or cell diameter is small ) , temporal sensing yields higher output than spatial sensing independent of the value of D A ′ and D B ′ ., To examine the effect of β , we plot the protein dynamics of the output protein for one set of parameter for the incoherent feedforward circuit at various values of beta for D A ′ = 1 ., 0 and D B ′ = 1 ., 0 ., When β is small ( cell velocity is small or cell diameter is large ) , the front and back halves of the cell experience a big delay in the time that they observe the chemoattractant and the levels of the output protein , C , at the rear end ( blue curve ) of the cell only increase after the level of C at the front end ( green curve ) starts to decrease ( Fig 3c , left ) ., The average level of C ( red curve ) , which sums over the two halves , shows a net increase at all times when the cell is moving through the gradient ., As β increases , the time difference in which the front and back halves of the cell experiences the chemoattractant decreases and their dynamics began to converge ( Fig 3c , right ) ., The signaling output for temporal sensing ( green ) and spatial sensing ( red ) were plotted in Fig 3d ., When the cell uses the temporal sensing mechanism by comparing the average output C with the baseline level , it observes a net increase in output ( area shaded in green ) as the cell moves through the gradient ., When spatial sensing is used to comparing the ratio of output at the front and back of the cell , the cell observes an increase in output ( area shaded in red above the x-axis ) as the cell entered the gradient ( entering phase ) followed by a decrease in output ( area shaded in red below the x-axis ) as the cell exits the gradient ( exit phase ) ( Fig 3d ) ., However the area above the x-axis is always larger than the area below the x-axis indicating an overall positive response ., As the difference between the level of the output protein at the front and back of the cell decreases with increasing β , so is the signal obtained from spatial sensing ., This explains why at high β , temporal sensing is favored ., Rather than size cell , we show that the relevant parameter for sensing is the ratio of cell velocity to the product of signaling rate and cell diameter ., This suggests that cells moving faster than its cell diameter in the time it takes for signaling to propagate across the cell diameter should adopt temporal sensing , whereas cells moving slower than its cell diameter in that time should adopt spatial sensing ., This can be reasoned as follows: a fast-moving , small cell performs better comparing the chemoattractant at different times in its trajectory; whereas , a slow-moving , big cell that is not travelling much performs better by comparing the chemoattractant concentration at its two ends ., As shown in Fig 3a and 3b , both temporal and spatial sensing can occur when β is small and we will next focus on the effects of D A ′ and D B ′ on this sensing choice ., As shown in Fig 4, ( a ) and 4 ( c ) , temporal sensing is favored when D A ′ > D B ′ and spatial sensing is favored when D A ′ < D B ′ ., The dynamics of protein C are plotted for different values of D A ′ and D B ′ ( Fig 4b and 4d ) ., At low diffusion ( D A ′ = 1 and D B ′ = 1 ) , the front and back halves behave like separate uncommunicating entities as discussed before and temporal signaling yields slightly higher output than spatial sensing ( Fig 4b and 4d , bottom row , left ) ., When diffusion of the activator is slow and diffusion of the inactivator is fast ( D A ′ = 1 and D B ′ = 100 ) , coupling between the front and back of the cell occurred ., Once the cell entered the gradient , inactivator B is produced and diffuses to the back of the cell to suppress the output level of protein C , amplifying the difference in levels of protein C between the front and the back ., This amplification led to a reduction of C from its basal level at the back of the cell ( Fig 4b and 4d , bottom row , right ) ., Hence spatial sensing yields much higher output signal than temporal sensing ., Furthermore , this coupling ensured that the levels of protein C at the back of the cell is lower than that at the front even during the exit phase ., This is consistent with previous models adopting a local acting activator and a globally acting inactivator for spatial sensing 11 ., On the other hand when diffusion of the activator is fast and diffusion of the inactivator is slow ( D A ′ = 100 and D B ′ = 1 ) , the global activation and local inhibition happens with activator diffusing to the back of the cell ., In IFF , this leads to higher level of protein C at the back than the front during the entering phase ( Fig 4b , top row , left ) ., This occurs as protein A produces at the front end of the cell rapidly diffused to the back , homogenizing level of protein A throughout the cell ., The higher level of inactivator , protein B , leads to greater repression and lower level of protein C at the front ., In this case , the level of protein C becomes higher at the back and the signaling output from spatial sensing becomes negative , making spatial sensing an inviable option ., This effect is not observed in NFB circuits as , protein B was activated by protein A rather than the external chemoattractant ( Fig 4d , top row , left ) ., Hence level of protein B is always be proportional to that of protein A . However in this case , spatial sensing is also not favored as the rapid diffusion of protein A led to loss of information about the external chemoattractant gradient ., Finally when both activator and inactivator diffuse fast ( D A ′ = 100 and D B ′ = 100 ) , the amplification effect observed for local excitation and global inhibition is still observed , albeit at a lower value ( Fig 4b and 4d , top row , right ) ., In summary , we find that spatial sensing is favored when the repressor diffuses faster than the activator ., This is because repressor produces at the front end is able to diffuse to the back to lower the signaling level of the output protein ., This magnifies the difference between the signal output at the two ends , leading to higher signaling output for spatial sensing ., When repressor diffuses slower than the activator , this amplification does not occur and temporal sensing is favored ., To check that our findings are independent of the exact gradient profile , we repeat our analysis for an exponential gradient ( S2 Fig ) ., We find that similar to results of the linear gradient , high ratio of cell speed to cell diameter favors temporal sensing and diffusivity of activator has to be smaller than diffusivity of repressor for spatial sensing to be preferred at low values of β ., In our simulations , the cell is moving from a region of constant chemoattractant , into a region with a linear increase in chemoattractant and finally into another region of a higher constant chemoattractant level ., In general , cells may be moving inside a steady state gradient ., To show that the motion from a region of constant chemoattractant into a gradient does not affect the findings , we simulate the response of cells into a step change in chemoattractant ( S3 Fig ) ., This will simulate the case where a cell suddenly encounters a gradient and moves into it , as opposed to moving inside a steady-state gradient ., We find that the main findings are consistent with those for a linear gradient ., We also repeat the simulations using a longer Ts = 20 and obtain similar findings ( S4 Fig ) ., In Fig 5a , we summarize our findings ., We find that sensing outcomes are determined by three dimensionless parameters: 1 ) the ratio of cell speed to the product of cell diameter and rate of signaling , 2 ) the diffusivities of the output protein of the two circuits and 3 ) the ratio of the diffusivities of the activator to inactivator protein ., Temporal sensing is usually preferred whereas spatial sensing is preferred when all three parameters are low ., To compare our theoretical results with experimental observations , we need to determine the diffusion rates , cell sizes and speeds of a wide range of chemotactic cells and organisms ., While cell sizes and speeds are readily available , values of diffusion rates are much harder to find ., Hence , we first compare our findings based on the ratio of cell speed to cell diameter with that of the sensing decisions of chemotactic cells and organisms ., The most well-studied chemotactic organism is E . coli ., E . coli is 2 μm in length 15 and swims at about 20 μm/s ., The dephosphorylation rate of Che-Y has been found to be 2 . 2s−116 , 17 ., This yields β = 4 . 5 , agreeing with our analysis that E . coli will adopt temporal sensing ., Since reactions rates are difficult to characterize and the circuitry controlling chemotaxis is usually much more complicated than our canonical NFB and IFF circuits , we are unable to obtain lBC for many chemotactic cells ., Nonetheless , we estimate reaction rates to be of the order of seconds based on the dephosphorylation rate of Che-Y 16 , 17 and the fast response time observed in chemotactic cells ., Micropipette stimulation experiments showed that neutrophils took between 5–30s to extend their surface towards the chemotactic pipette 18 ., We conduct an extensive literature search to obtain the diameters and velocities of many chemotactic cells and unicellular organisms such as bacteria 19–22 , Paramecium caudatum ( P . caudatum ) 23 , Tetrahymena thermophila24 , alga 25; sperm cells 26 , 27; mammalian cells 28–33; insect cells 34 , 35; and amoeba 36–38 ., We classified these chemotactic cells based on their mechanisms of motion , namely lamellipodia/filopodia , flagellar , pseudpodia and cilia ., In general , the eukaryotic and insect cells are in the lamellipodia/filopodia group; bacteria and sperm cells are in the flagellar group; amoeba are in the pseudpodia group; and Tetrahymena thermophila and alga are in the cilia group ., We find that cells using flagellar and cilia to move have higher ratio of velocity over cell diameter than cells using lamellipodia/filopodia and pseudpodia ( Fig 5b ) ., In our simulations , we find that cells and organisms with high ratio of cell speed to cell diameter adopt temporal sensing ., Assuming lBC = 1s−1 , cells and organisms above the the black horizontal line in Fig 5b will adopt temporal sensing ., In general , lBC may be different in each cell , if lBC lies between 0 . 2s−1 − 5s−1 then the yellow region will be the separating boundary between cells with high and low values of β ., Cells with high β values includes cells in the flagellar ( green ) group and agrees with the broad categorization that these cells adopt temporal sensing ., Sperm cells have been shown to utilize temporal sensing despite being relatively big 26 ., Our results suggests that temporal sensing is utilized as its high cell velocity makes temporal sensing more advantageous ., One exception to the classification is the bipolar flagellated vibrioid bacteria that has been suggested to adopt spatial sensing 39 ., This bacteria has a very fast response time as it was able to correct deviations from its swimming direction within a second ., Further work elucidating the chemotaxis circuitry and reaction rates in this organism is necessary to determine the value of β ., It is currently unclear whether P . caudatum adopts spatial or temporal sensing ., The other ciliated organism ( blue ) , Tetrahymena thermophila , has been proposed to utilize temporal sensing 40 , agreeing with our prediction ., Cells and organisms below the yellow region in Fig 5b have low values of β ., We find that these cells would adopt spatial sensing if the activator diffuses slowly whereas the inactivator diffuses fast ., Unfortunately it is difficult to obtain these diffusion rates as many of the activator and inactivator proteins involved are unknown ., For example , in Dictyostelium discoideum , some literature suggests that the locally acting activator ( Protein A ) , PI3-kinase , and globally acting inactivator ( Protein B ) , PTEN , work together to control G-protein ( Protein C ) activation during chemotaxis 13 whereas other literature suggests that RasGEF and a RasGAP are the activator and inactivator proteins instead 12 ., As diffusion rate is inversely proportional to the square root of the molecular weight , one could estimate the ratio of PTEN to PI3K diffusion rate and the ratio of RasGEF to RasGAP diffusion rate to be 83 , 598 47 , 166 2 = 1 ., 33 and 57 , 010 54 , 556 2 = 1 ., 02 respectively ., The slight differences in these estimated diffusion rates are clearly inconsistent with the local and global activation roles suggested ., This shows that even when there are candidate proteins for the activator and inactivator proteins , molecular weight is not a good approach for estimating diffusion rates in cells and suggests the presence of other active biological processes in controlling the movements of these proteins ., From Fig 4, ( b ) and 4 ( d ) , we observe that the signaling output , OS , is highest at low activator diffusion rate and high inactivator diffusion rate ( D A ′ = 1 . 0 and D B ′ = 100 ) for low value of β ., From an evolutionary point of view , this suggests that organisms would evolve towards having high D B ′ and low D A ′ to achieve better chemotactic response ., Indeed , it has been shown experimentally that lamellipodia/filopodia ( black ) and pseudpodia cells ( red ) utilize spatial sensing ., Hence we find that β is the most important determinant in the choice between spatial and temporal sensing ., Next , we consider the effect of noise on the decision choice ., Noise can exist in both the external chemoattractant and the internal signaling pathway and affects chemotaxis 41 ., We focus our analysis on the regime where D A ′ is low and D B ′ is high as this was the region of parameter space that yields most interesting behavior in the deterministic analysis ., We examine the decision choice for the following cases: ( 1 ) β = 0 . 25 , ( 2 ) β = 1 . 0 and ( 3 ) β = 4 . 0 at D A ′ = 1 and D B ′ = 100 as the amount of external or internal noise increases ., Since each run is stochastic , ten runs are performed on each set of parameters and noise level to determine the average performance from spatial and temporal sensing ., First , we focus on the presence of external noise in the chemoattractant gradient ., The dynamics of protein C is plotted at different noise levels for β = 0 . 25 and β = 4 . 0 ( S5 Fig ) ., η quantifies the amount of fractional noise ., At low level of noise , η = 0 . 0625 , the dynamics of protein C is well behaved with the levels of protein C at the front always higher than that at the back ( S5a and S5d Fig ) ., As η increases , the dynamics becomes noisier with levels of protein C showing more fluctuations ( S5b and S5e Fig ) ., Furthermore , the level of protein C at the front of the cell is sometimes lower than that at the back ., However , the average levels of protein C is still rather well-behaved , rising as the cell enters the chemoattractant gradient and adapting back to basal level as the cell exits the gradient ., At high level of noise , η = 1 . 0 , the noise level dominates over the signal and the levels of protein C fluctuated randomly ( S5c and S5f Fig ) ., Next , we want to determine which sensing strategy is more susceptible to noise ., For each set of parameters , ten stochastic runs are performed ., If all the runs yield positive signaling output for a particular sensing strategy that strategy is considered to be viable for that set of parameters ., We plot the fraction of parameter set that fulfil the above criteria for spatial and temporal sensing ( Fig 6a and 6b ( red ) ., We find that temporal sensing ( green ) was less susceptible to noise than spatial sensing ( red ) ., Intuitively , this can be understand as taking average in temporal sensing is more robust than taking difference in spatial sensing ., We also find that the fraction of parameters fulfilling the criteria increased as β decreased ( Fig 6a , ( red ) ) ., This showed that spatial sensing is less susceptible to noise when the cell diameter is larger than cell velocity ., Lastly , we determine the fraction of parameters that chooses temporal or spatial sensing ., When noise level becomes too high , both sensing mechanisms fail as the signal had been completely dominated by noise ., As shown in ( Fig 6c and 6d ) , spatial sensing performs better than temporal sensing for low values of β and low values of noise ., As noise level increases , temporal sensing yields better results ., Finally at very high noise levels , sensing using both strategies are infeasible ., To introduce noise into the internal signaing pathway , we allow all the kinetic parameters ( kIA , kIB , lFA , lFB , kAC , lBC , kCB ) to be random variable with mean equal to their values in the noiseless case and variance , ν ., We find that the sensing decision is independent of the amount of noise , ν ( Fig 7 ) ., We examine the dynamics of protein C when subjected to external chemoattractant noise and internal signaling noise ( Fig 8 ) ., We find that in the presence of internal noise , protein C fluctuates at high frequency about the expected value of C for the noiseless case ., Integrating over time , the noise would cancel out , leading to an average performance similar to that of the noiseless case ., On the other hand , protein C fluctuates at low frequency in the presence of external noise and its mean averaged over time can be quite different from the expected value of C for the noiseless case ., Hence in this case , temporal sensing performs better ., Here , we determine the conditions favoring temporal and spatial sensing ., We find that the behavior of the negative integral feedback and incoherent feedforward circuits were determined by five dimensionless constants , namely the three diffusion rates , D A ′ = D A d 2 l B C , D B ′ = D B d 2 l B C and D C ′ = D C d 2 l B C , the ratio of cell speed to the product of cell diameter and signaling rate , β = v d l B C , and the effective chemoattractant gradient , α = v k l B C . Both the negative integral feedback and incoherent feedforward circuits yielded similar behaviors when we varied the dimensionless constants ., We summarize our findings in Fig 5a ., In brief , temporal sensing is favored in most situations whereas spatial sensing is only favored when values for D C ′ and β are small and D A ′ , diffusion rate of the activator , is slower than that of D B ′ , diffusion rate | Introduction, Results, Discussion | Cell size is thought to play an important role in choosing between temporal and spatial sensing in chemotaxis ., Large cells are thought to use spatial sensing due to large chemical difference at its ends whereas small cells are incapable of spatial sensing due to rapid homogenization of proteins within the cell ., However , small cells have been found to polarize and large cells like sperm cells undergo temporal sensing ., Thus , it remains an open question what exactly governs spatial versus temporal sensing ., Here , we identify the factors that determines sensing choices through mathematical modeling of chemotactic circuits ., Comprehensive computational search of three-node signaling circuits has identified the negative integral feedback ( NFB ) and incoherent feedforward ( IFF ) circuits as capable of adaptation , an important property for chemotaxis ., Cells are modeled as one-dimensional circular system consisting of diffusible activator , inactivator and output proteins , traveling across a chemical gradient ., From our simulations , we find that sensing outcomes are similar for NFB or IFF circuits ., Rather than cell size , the relevant parameters are the, 1 ) ratio of cell speed to the product of cell diameter and rate of signaling ,, 2 ) diffusivity of the output protein and, 3 ) ratio of the diffusivities of the activator to inactivator protein ., Spatial sensing is favored when all three parameters are low ., This corresponds to a cell moving slower than the time it takes for signaling to propagate across the cell diameter , has an output protein that is polarizable and has a local-excitation global-inhibition system to amplify the chemical gradient ., Temporal sensing is favored otherwise ., We also find that temporal sensing is more robust to noise ., By performing extensive literature search , we find that our prediction agrees with observation in a wide range of species and cell types ranging from E . coli to human Fibroblast cells and propose that our result is universally applicable . | Unicellular organisms and other single cells often have to migrate towards food sources or away from predators by sensing chemicals present in the environment ., There are two ways for a cell to sense these external chemicals: temporal sensing , where the cell senses the external chemical at two different time points after it has moved through a certain distance , or spatial sensing , where the cell senses the external chemical at two different locations on its cellular surface ( e . g . , the front and rear of the cell ) simultaneously ., It has been thought that small unicellular organisms employ temporal sensing as their small size prohibits sensing at two different locations on the cellular surface ., Using computational modeling , we find that the choice between temporal and spatial sensing is determined by the ratio of cell velocity to the product of cell diameter and rate of signaling , as well as the diffusivities of the signaling proteins ., Predictions from our model agree with experimental observations over a wide range of cells , where a fast-moving , small cell performs better comparing the chemoattractant at different times in its trajectory; whereas , a slow-moving , big cell performs better by comparing the chemoattractant concentration at its two ends . | blood cells, cell motility, medicine and health sciences, immune cells, engineering and technology, signal processing, immunology, germ cells, protozoans, network analysis, cellular structures and organelles, neutrophils, sperm, computer and information sciences, white blood cells, network motifs, animal cells, amoebas, chemistry, cell membranes, chemotaxis, physics, mass diffusivity, eukaryota, cell biology, biology and life sciences, cellular types, physical sciences, chemical physics, organisms | null |
journal.ppat.1004657 | 2,015 | Human Adenovirus 52 Uses Sialic Acid-containing Glycoproteins and the Coxsackie and Adenovirus Receptor for Binding to Target Cells | Human adenoviruses ( HAdVs ) are classified into seven species ( A-G ) , with more than 50 different types known to date 1 ., Most HAdVs cause disease in the eyes ( members of species HAdV-B , -D , -E ) , airways ( HAdV-A , -B , -C , -E ) and gastrointestinal tract ( HAdV-F mainly ) 2 ., HAdV-52 was recently identified as a novel , human pathogen associated with gastroenteritis 3 , and was found to be divergent from other HAdVs placing this virus in a new species ( HAdV-G ) ., HAdVs from species HAdV-A and HAdV-C through HAdV-F use the coxsackievirus and adenovirus receptor ( CAR ) as a primary adhesion receptor 4–6 ., Members of species HAdV-B that cause ocular , respiratory and/or urinary tract infections utilize CD46 and/or desmoglein-2 as cellular receptors 7–10 ., Specific members of species HAdV-D cause a more severe ocular infection , epidemic keratoconjunctivitis , and engage glycoproteins that carry glycans mimicking those in the GD1a ganglioside: Neu5Acα ( 2–3 ) Galβ ( 1–3 ) GalNAcβ ( 1–4 ) ( Neu5Acα ( 2–3 ) ) Galβ ( 1–4 ) Glc as receptors 11–13 ., In addition to these AdVs , canine AdV-2 ( species CAdV-A ) is another glycan-binding AdV , engaging Neu5Acα ( 2–3 ) 6SGalβ ( 1–4 ) GlcNAc-containing glycans at a different location on the knob14 ., The locations of the two known glycan binding sites are distinct from the regions that allow some knobs to engage CAR 15 or CD46 16 ., In the case of CAR , the length and flexibility of the fiber shaft also seem to play a role in infection , as short and sturdy fibers cannot bend to bind CAR on a cell surface 17 ., HAdVs can also enter cells through interactions with coagulation factors that mediate indirect binding to heparan sulfate proteoglycans on target cells 18 , 19 ., With a few exceptions , HAdVs are equipped with a single type of capsid fiber protein that interacts with receptors via its knob domain ., HAdV-40 , -41 , and -52 on the other hand are equipped with two different fibers , one long and one short 20 , 21 ., The long fibers of HAdV-40 and -41 bind to CAR 6 , but no function has been described for any of the short fibers ., Phylogenetic analyses have shown that the closest human relatives to the knobs of the HAdV-52 long and short fibers are the knobs of the long and short fibers of species HAdV-F ( HAdV-40 and -41 ) 22 ., Adenoviruses are frequently used as vectors for diverse applications including vaccination 23–25 , treatment of cancer 26 and hereditary disorders 27 , cardiovascular applications 28 , and stem cell research 29 ., Three of the main challenges for the most commonly used HAdV-5 ( species HAdV-C ) based vectors are:, i ) pre-existing , neutralizing antibodies 30 ,, ii ) poor access to CAR 31 , and ,, iii ) coagulation factor-dependent off-target transduction of the liver 18 ., Potential solutions to these obstacles have been to use vector candidates based on less common HAdV types 24 , and HAdV types that use receptors alternative to CAR 32 , 33 ., Inefficient targeting have been addressed by ablating CAR- and/or coagulation factor-interactions , and/or by retargeting to receptors that are overexpressed on target cells 34 ., Ideally , a multi-purpose vector would therefore be based on a rare type that efficiently target host cells by means of specific receptor interactions and low or absent off-target transduction ., The seroprevalence for HAdV-40 and -41 is relatively high ( 40–50% ) in the human population 35–37 ., The seroprevalence for HAdV-52 in humans has not been investigated , but the close relationship with simian AdVs and the low frequency of detection in humans 3 , 38 suggest that the seroprevalence in humans is low ., In combination with its uncommon capsid organization this prompted us to gain more insight into HAdV-52 interactions with host cells and more specifically to identify cellular receptors used by HAdV-52 for attachment to host cells ., To investigate whether CAR , CD46 or sialic acid-containing glycans can function as receptors for HAdV-52 , we first analyzed 35S-labelled HAdV-52 virion binding to CHO cells expressing or lacking these receptors ., HAdV-52 bound with similar efficiency to sialic acid-expressing control CHO ( Pro-5 ) cells , CD46-expressing CHO cells and CHO MOCK ( with respect to CAR ) , but with increased efficiency to CAR-expressing CHO cells and with decreased efficiency to sialic acid-lacking Lec2 cells ( derived from Pro-5 ) as compared to the other cells ( Fig . 1A ) ., Pretreatment of cells with sialic acid-cleaving V . cholerae neuraminidase reduced HAdV-52 binding to background levels for all cells except to CHO-CAR ., To test if this neuraminidase removed sialic acids with equal efficiency from all cells , we treated the cells with V . cholerae neuraminidase and quantified MAL-II lectin binding ., This treatment reduced MAL-II binding to background levels ( S1 Fig . ) and we therefore concluded that HAdV-52 could bind to CHO-CAR independently of sialic acid ., As HAdV-52 bound with equal efficiency to Pro-5 , CHO-MOCK , and CHO-CD46 , and as neuraminidase treatment of CHO-CD46 cells reduced HAdV-52 binding efficiently , these results indicate that CD46 is probably of no or low importance as a receptor for HAdV-52 ., HAdV-52 also infected Pro-5 cells more efficiently than Lec2 cells , and pretreatment of Pro-5 cells with neuraminidase abolished infection ( Fig . 1B ) ., HAdV-52 is associated with gastroenteritis , but the number of human cases described is limited and the cellular tropism of the virus is unclear ., We therefore investigated the relative contributions of sialic acid and CAR using respiratory A549 cells , which support productive infection of most HAdVs and express both sialic acid and CAR at the cell surface ., HAdV-52 binding to these cells was reduced by 20% and 25% , respectively , when preincubating HAdV-52 virions with soluble CAR-D1 ( consisting of the N-terminal , most membrane-distal immunoglobulin-like domain ) , or when preincubating cells with monoclonal anti-CAR antibodies ( clone RmcB ) prior to virion binding ( Fig . 1C ) ., CAR-D1 and anti-CAR antibodies reduced HAdV-5 binding with 50% and 75% , respectively ( S2 Fig . ) , thus demonstrating their function ., On the other hand , HAdV-52 binding was reduced by 75% and 80% after preincubating virions with sialic acid or when pretreating cells with neuraminidase , respectively , prior to virion binding ., Pretreatments with CAR-D1 or anti-CAR antibodies in combination with either sialic acid or neuraminidase reduced binding to background levels ., The involvement of sialic acid-containing glycans as functional human cell receptors for HAdV-52 was confirmed by neuraminidase pretreatment of A549 cells , which reduced HAdV-52 infection by at least 80% ( Fig . 1D ) ., Finally , preincubation of virions with coagulation factor IX and X efficiently enhanced HAdV-5 binding to and infection of A549 cells but had no or limited effect on HAdV-52 ( Fig . 2A , B ) ., These results show that HAdV-52 does not use FIX , FX , or CD46 for attachment to A549 cells ., We conclude that HAdV-52 binds to A549 cells mainly via sialic acid-containing glycans , and that the role of CAR is dwarfed by that of the sialylated receptors ., However , we cannot exclude that the role of CAR as an attachment receptor for HAdV-52 may be more pronounced on other cell types than on A549 cells ., To characterize the nature of the sialic acid-containing glycans as receptors and the mechanism of interaction , we next quantified binding of HAdV-52 virions and HAdV-52 long and short fiber knobs ( 52LFK and 52SFK ) to A549 cells pretreated with enzymes , lectins or metabolic inhibitors that alter the expression levels of cell surface molecules ., Whereas inhibitors of glycolipid biosynthesis ( P4 ) and N- ( via Asp ) linked glycosylation ( tunicamycin ) did not reduce virion binding to A549 cells significantly ( Figs . 3A , S3A , B ) , benzyl N-acetyl-α-D-galactosaminide ( benzyl-α-GalNAc , an inhibitor of O-linked glycosylation , via Ser or Thr ) reduced binding of both the virions and 52SFK , but not of 52LFK ( Fig . 3A , B ) ., Protease ( ficin , proteinase K , and bromelain ) treatments of the same cells reduced binding of both 52SFK and 52LFK by 55–85% ( Fig . 3C , D ) ., These results suggest that on A549 cells , in contrast with the 52LFK which engages proteins directly without involvement of glycans , mucin type O-linked glycans are the dominant receptors for 52SFK and glycolipids and N-linked glycans appear not to play a major role ., To determine the relative contribution of each fiber to cell attachment and infectivity , we first performed western blot analysis to characterize the relative fiber content in virus particles ., Unlike HAdV-41 virions , which contain short and long fibers in a 6:1 ratio 39 , HAdV-52 virions contained equal amounts of long and short fibers according to western blot analysis using a monoclonal antibody , which recognizes an epitope that is conserved in all HAdVs ( Fig . 4A ) ., This suggests that the apparent key role of sialic acid cannot be accounted for by the short fiber being more abundant in the HAdV-52 virion ., We also found by flow cytometry analysis that A549 cells expressed higher levels of CAR compared with another epithelial cell line ( human corneal epithelial cells; HCE ) ( Fig . 4B ) , suggesting that the modest function of CAR during HAdV-52 binding to A549 cells was not due to low expression levels on these cells ., Homology alignment of the long and short fiber knob sequences with corresponding sequences of sialic acid-interacting HAdV-37 ( ocular tropism ) and CAR-interacting HAdV-5 ( respiratory tropism ) and HAdV-12 ( respiratory and intestinal tropism ) revealed that , while the majority of the known CAR-interacting residues 15 , 40 are conserved in 52LFK , only a few of these residues are conserved in 52SFK ( S4 Fig . ) ., Furthermore , when examining the potential for interactions with sialic acid based on the structure of the HAdV-37 knob bound to sialic acid 41 , only two out of the seven sialic acid-contacting residues are conserved in 52SFK and none of these are conserved in 52LFK ., Flow cytometry analysis confirmed that 52LFK can only bind to CAR-expressing cells ( Fig . 5A ) ., 52SFK bound with similar efficiency to all cells ( including CAR-expressing cells ) but not to sialic acid-deficient Lec2 cells ( Fig . 5B ) ., Neuraminidase treatment of A549 cells reduced binding of 52SFK to A549 cells but not of 52LFK ( Fig . 5C ) , confirming that 52SFK binds to sialic acid-containing receptors on human target cells ., ELISA experiments showed that 52SFK ( in solution ) bound efficiently to sialylated fetuin glycoprotein ( immobilized ) but not to two desialylated variants of fetuin ( Fig . 5D ) , this was also confirmed with surface plasmon resonance ( SPR ) where fetuin bound to immobilized 52SFK with an affinity of 37 μM , while for the desialylated fetuin type II a KD could not be determined ( S5 Fig . ) ., 52LFK did not bind to any of these proteins ., SPR analysis demonstrated that the 52LFK:CAR-D1D2 ( full length extracellular domain ) and 52LFK:CAR-D1 interactions were of high affinity ( 5 and 2 . 6 nM , respectively; Figs . 5E and S6 ) , which is in the same range as of other CAR:HAdV-knob interactions 42 ., According to SPR analysis , 52SFK did not interact with CAR at all ( S7 Fig . ) ., As the sialic acid-containing glycan ( s ) used by 52SFK for binding to A549 cells are not known , we can only speculate that such monovalent interactions would probably be of low affinities , as most other protein:glycan interactions , and thereby lower than the affinity of the LFK:CAR interaction ., We conclude from these results that the HAdV-52 long fiber binds to CAR and that the short fiber binds to sialic acid-containing glycans ., It has been shown that cells infected with HAdV-2 ( species HAdV-C ) secrete an excess of fibers that unlocks junctional , intercellular CAR-CAR homodimers , resulting in increased extracellular space and improved intercellular transport of subsequently released virions 43 , and similar effects have been shown for HAdV-3 ( species B ) penton dodecahedra 44 ., It is therefore tempting to speculate that a possible function of the HAdV-52 short fiber is to mediate virion attachment to non-infected cells whereas excess of long fibers are secreted from infected cells and facilitate transmission of subsequently released virions within a tissue , or between tissues ., Further support for this hypothesis is provided in that sialic acid-containing , O-linked glycans are abundant on the apical side of polarized epithelial cells in vivo , whereas CAR is mainly expressed laterally and basolaterally 45 ., Thus it is plausible that virions approaching non-infected cells from the apical side have access to sialylated glycans , but not to CAR ., HAdV-37 has been shown to interact primarily with sialic acids linked via α2 , 3-glycosidic bonds to galactose ( Siaα2 , 3Gal ) ., We found here that Siaα2 , 3Gal-binding M . amurensis type II ( MAL-II ) lectins and/or Siaα2 , 6Gal-binding S . nigra ( SNA ) lectins did not compete with HAdV-52 virion binding to A549 cells ( S8 Fig . ) ., We noted that α2 , 3-specific neuraminidase inhibited HAdV-52 virion binding to A549 cells ( Fig . 5F ) , but only at 100-fold higher concentrations than what has been observed for inhibition of HAdV-37 virion binding 11 ., Pretreatment of A549 cells with neuraminidase from V . cholerae , which cleaves α2 , 3/6/8-linked sialic acids with similar efficiencies , inhibited HAdV-52 binding at much lower concentrations ., By means of glycan microarray screening we identified a number of α2 , 3-sialylated probes that are bound by 52SFK , whereas no binding was detected with probes that contain exclusively α2 , 6-sialyl linkage ( Fig . 6 ) ., The probe most strongly bound was a synthetic glycolipid with type II ( Galβ-4GlcNAc ) backbone sequence ( GSC-273 ) ., In contrast , no binding was detected to the type I ( Galβ-3GlcNAc ) analog ( GSC-272 ) ., Weaker binding was observed to three of the four sulfated sialyl analogs with or without 3-linked fucose ., There was also weak binding to a neoglycolipid derived from GD1a glycan , the previously described ligand for HAdV-37 35 ., No binding was detected to GD1a glycosylceramide ., It should be mentioned that , although five of the ligand-positive sialyl probes in the array were glycolipids , their glycan sequences are common to glycoproteins ., Among the N-glycan probes analysed , there was binding to probes with α2 , 3-linked terminal sialic acids ., Collectively , we cannot exclude α2 , 3-linked sialic acid-containing N-glycans from contributing to HAdV-52 binding to A549 cells , but it is likely that other types of sialic acid-containing glycans also contribute ., Our results suggest that the short fiber is capable of binding to sialic acids on O-glycosylated proteins on A549 cells , but that on other cell types binding to sialyl-N-glycans may also occur ., In order to define the interactions of the HAdV-52 fiber with sialic acid , we solved the crystal structure of 52SFK in complex with 2-O-methyl-sialic acid ( a stereochemically uniform analogue of sialic acid ) at a resolution of 1 . 65 Å ., Similar to all other known AdV fiber knob structures 13 , 46–48 the 52SFK has a nine-stranded antiparallel β-sandwich fold and forms a stable trimer in solution ., The three shallow sialic acid binding sites of 52SFK are formed at the contact site of two neighboring monomers by the EG and GH loops at the side of the short fiber knob domain ( Fig . 7A-C ) ., Two of the three binding sites are partially blocked by crystal contacts , and therefore the structure contains only one fully occupied sialic acid , while a second sialic acid is visible with partial occupancy in one of the two partially blocked binding sites ., The location of this binding site is distinct from those of the other structurally characterized sialic acid-binding fiber knobs , HAdV-37 and canine adenovirus type 2 ( CAdV-2; included here since it is the only known sialic acid-interacting AdV besides those of species HAdV-D ) ( Fig . 8; PDB_IDs 1UXA and 2WBV ) ., The bound 2-O-methyl-sialic acid is well defined by electron density ( Fig . 7C ) and engages the 52SFK mainly through contacts between the sugar’s carboxylate group and the side chains of R316 and N318 ( Fig . 7B ) ., The bidentate salt bridge formed by R316 is a prominent binding motif among glycan binding viruses 49 , 50 ( S9 Fig . ) ., In addition , the backbone carbonyl oxygens of R316 , G317 , and G303 form hydrogen bonds with the sialic acid O4 , N-acetyl and glycerol-like functions , respectively ( Fig . 7B ) ., The tripeptide R316-G317-N318 located on the GH loop forms a hook-shaped motif ( RGN motif ) that is contributing most of the interactions , and that therefore largely defines the specificity of SFK52 for sialic acid ., The pattern of polar contacts formed by this motif is highly similar to HAdV-37 ( S9 Fig . ) 50 ., Mutating either R316 or N318 to alanine , replacing the R316 side chain with a negatively charged glutamate , or introducing a steric clash ( and a polar clash , introducing a charge ) at position 308 ( N308E ) all abolished the attachment of 52SFK to Pro-5 and A549 cells ( Fig . 7D , E ) ., The sugar’s O2 function , to which additional sugars would be attached in a glycan chain , is pointing away from the protein and towards the tip of the knob , suggesting that more complex glycan receptors that bind the knob with their sialic acid caps would have to face towards the capsid in order to be bound by the virus ., The methyl group attached to this oxygen in the compound used for structural analysis ( 2-O-methyl-sialic acid ) does not participate in interactions with the knob ., Alignment of multiple knob sequences suggests that the RGN motif is conserved in the knob domain of the short fibers of other members of species HAdV-G: simian AdV-1 and -7 ( S10 Fig . ) , and we therefore predict that the ability to engage sialylated receptors is shared by these HAdVs ., The RGN motif is not conserved in any other known human and non-human AdV knob sequences , including the short fiber knobs of HAdV-40 and -41 ., In conclusion , we have identified two types of cellular receptors used by HAdV-52 , the only human member of species HAdV-G ., By analogy with HAdV-40 and -41 , we identified CAR as a receptor for the HAdV-52 long fiber ., We also present evidence that O-glycosylated proteins carrying sialic acid-containing glycans serve as receptors on A549 cells for HAdV-52 short fibers ., The 52LFK:CAR interaction is probably of higher affinity than the 52SFK:sialic acid interaction , however the relative importance of cellular receptors is not only determined by the affinity but also to a high extent on the abundance of the receptors ., It has been shown for example by us and others that HAdV-37 binds with lower affinity to CAR ( 20 nM ) than other HAdVs , but with even lower affinity to sialic acid-containing glycans ( 19 μM ) 41 , 51 ., Still , HAdV-37 uses sialic acid-containing glycoproteins as the main receptor 11 , 52 ., Accessibility may also influence receptor usage ., In vivo , CAR localizes to the lateral and basolateral side of polarized epithelial cells , which are not easily accessible for HAdVs , while sialic acid is abundant on the apical surface and may therefore be more available for interaction ., Thus we suggest that sialylated proteins rather than CAR function as primary receptors for HAdV-52 virions on A549 cells ., HAdV-52 might have retained the ability to bind to CAR as secretion of CAR-interacting fibers can disrupt CAR-CAR homodimers in the tight junctions and thereby facilitate virion escape and transmission within a tissue , or between tissues ., The mode of interaction between 52SFK and its sialylated receptors is fundamentally different , both in location on the protein and in contacts formed to the ligand , from the known interactions between HAdV-37 fiber knob and sialic acid ., As the sialic acid-binding RGN motif of 52SFK is not conserved in any other HAdV fiber , it appears that HAdV-52 and other members of species HAdV-G employ a unique strategy for engaging sialic acid ., This is the first information presented about the receptors used by viruses in species HAdV-G , and the first describing a receptor recognized by an AdV short fiber ., These findings shed light on AdV biology and tropism and may be useful for development of vectors based on members of species HAdV-G ., A549 cells ( gift from Dr . Alistair Kidd ) were grown in Dulbecco´s modified Eagle medium ( Sigma-Aldrich ) supplemented with 5% fetal bovine serum ( FBS: Invitrogen ) , 20 mM HEPES ( Sigma-Aldrich ) and 20 U/ml penicillin + 20 μg/ml streptomycin ( Invitrogen ) , human corneal epithelial ( HCE ) cells ( gift from Dr . Araki-Sasaki ) were grown as previously described 53 ., Pro-5 and Lec2 cells 54 , 55 ( both purchased from LGC Promochem ) , Chinese hamster ovary ( CHO ) -CAR , CHO-MOCK ( gift from Dr . Jeffrey Bergelson ) 4 , and CHO-CD46 ( isoform BC1; gift from Dr John P . Atkinson ) 56 were grown as described ., Species G HAdV-52 ( strain TB3-2243 ) 3 and species C HAdV-5 ( Ad75; source ATCC ) virions were produced with or without 35S-labeling in A549 cells as described previously 57 , with the exception that the virions were eluted in sterile phosphate buffered saline ( PBS ) when desalting on a NAP column ( GE Healthcare ) ., Serotype-specific rabbit polyclonal antisera to each HAdV was a gift from Dr Göran Wadell 58 ., Antiserum produced against HAdV-41 virions was used for detection of HAdV-52 antigens in infection experiments ., Cells were detached with PBS containing 0 . 05% EDTA , reactivated in growth medium for one hour at 37°C ( in solution ) , pelleted in 96 well plates ( 2x105 cells/well ) and washed with binding buffer ( BB: Dulbecco´s modified Eagle medium supplemented with 20 mM HEPES , 20 U/ml penicillin + 20°g/ml streptomycin and 1% bovine serum albumin ) . 35S-labeled virions ( 2x109 virions diluted in BB , 100 μl/sample ) were added to the cells and incubated for 1 h on ice ., Unbound virions were washed away with BB and the cell associated radioactivity was measured in a Wallac 1409 liquid scintillation counter ( Perkin-Elmer ) ., This experiment was performed with the following additions/variations: Pro-5 , Lec2 , or A549 cells , grown as monolayers on glass slides in 24-well plates , were washed three times with serum-free medium and treated with or without 10 mU/well of Vibrio cholerae neuraminidase for 1 h at 37°C ., Virions were added to the cells and incubated for 1 h on ice ., After incubation , the wells were washed three times with serum-free medium in order to remove unbound virions ., Cell culture medium containing 1% FBS was added and the plates were incubated for 44 h at 37°C ., Thereafter the glass slides were washed with PBS ( pH 7 . 4 ) once , fixed with methanol and stained with polyclonal rabbit anti-HAdV diluted 1:200 for 1 h at room temperature ., The slides were washed twice with PBS and incubated for an additional hour with a FITC-conjugated swine anti-rabbit IgG antibody ( DakoCytomation ) diluted 1:100 in PBS ., After washing , the slides were mounted and examined in a fluorescence microscope using 20 X magnification ( Axioskop2 , Carl Zeiss ) ., Ten pictures was taken of each well and the number of infected cells was calculated using ImageJ 60 ., In one experiment , virions were preincubated with or without physiological concentrations of FIX ( 5 μg/ml , equal to 1:60000 virion:FIX ratio ) or FX ( 10 μg/ml , equal to 1:110000 virion:FX ratio ) for 1 h on ice before addition to cells ., DNA isolation from HAdV-52 virions was performed by using the Blood & Cell Culture DNA Mini kit ( Qiagen Nordic ) ., DNA fragments encoding HAdV-52 long fiber knob ( 52LFK ) and HAdV-52 short fiber knob ( 52SFK ) were amplified by polymerase chain reaction ( PCR ) using KOD Hot Start ( Novagen , Merck ) and the following primers ( DNA Technology ) : 52LFK forward ( 5´-aaaaggatccggaaacatagctgtttctcct ) , reverse ( 5´-aaaacccgggcggaggaagccttactgtgcgtgt ) , 52SFK forward ( 5´-aaaaggatccaggtttaacagcagt-ggagcc ) , reverse ( 5´-aaaacccgggagggttttattgttcggtaatgtagca ) ., Fragments were then cloned into a pQE30Xa expression vector encoding an N-terminal His-tag ( Qiagen ) using restriction sites for BamHI and XmaI ( Fermentas , ThermoFisher Scientific ) ., All constructs were confirmed by sequencing ( Eurofins MWG Operon ) ., Proteins were expressed in Escherichia coli ( strain M15 ) and purified with Ni-NTA agarose beads according to protocol from the supplier ( Qiagen ) ., Proteins were analyzed by denaturing gel ( NuPAGE Bis-Tris , Invitrogen , Life Technologies ) and western blot with monoclonal antibodies directed against the His-tag ( Qiagen ) ., Six different 52SFK mutants were created using a QuikChange mutagenesis kit ( Agilent ) according to their protocol ., The following mutants were created: 1 ) R316A , 2 ) R316E , 3 ) N318A , 4 ) N308E , 5 ) R316A/N318A , and 6 ) R316E/N318A ., Correct trimerization of all proteins was confirmed with gas-phase electrophoretic mobility molecular analysis ( GEMMA ) 61 ( S11 Fig . , showing 52SFK wt and one representative mutant; R316A ) , which is a protein oligomer measurement technique where the protein solution is converted into gas phase by a charged reduced electrospray process ., The particles are separated according to size in a differential mobility analyzer and quantified by a particle counter ., All mutant and wt fiber knobs were analyzed in the same manner: G-25 columns ( GE Healthcare ) were used for buffer exchange to 20mM ammonium acetate buffer , pH 7 . 8 containing 0 . 005% ( v/v ) Tween 20 ., Buffer exchange was done to remove the non-volatile salts from the protein solution ., The concentrations for 52SFK wt and mutants 1 , 2 and 3 were 0 . 05 mg/ml while for mutants 4 , 5 and 6 it was 0 . 06 mg/ml ., Three to five scans were taken with the GEMMA system ( TSI Corp . ) for each sample with a capillary pressure of either 1 . 7 or 3 . 7 psi ., These parameters depended on the protein sample and the stability of the signal ., Each sample was scanned for 120 seconds per scan at the size range of 2 . 55–255 nm ., For molecular mass calculations , a particle density of 0 . 58 g/cm3 was used ., Cells were detached with PBS-EDTA , reactivated in growth medium for one hour at 37°C , pelleted in 96 well plates ( 2x105 cells/well ) and washed once with BB ., The cells were then incubated with 10 μg/ml of 52SFK or 52LFK in 100 μl BB for one hour on ice ., Unbound fiber knobs were washed away with PFN ( PBS containing 2% FBS and 0 . 01% NaN3 ) and the cells were then incubated with an anti RGS-His mouse monoclonal antibody ( Qiagen; diluted 1:200 in PFN ) for 30 min ., Followed by one wash with PFN , the cells were incubated with polyclonal rabbit-anti-mouse FITC antibodies ( Dako Cytomation; diluted 1:20 in PFN ) for 30 min on ice ., Thereafter the cells were washed with PFN and analyzed with flow cytometry using FACSLSRII instrument ( Becton Dickinson ) ., Results were analyzed using FACSDiva software ( Becton Dickinson ) ., This experiment was performed with the following additions/variations: The cells were, i ) grown in the presence or absence of benzyl-α-GalNAc ( as described above ) ,, ii ) preincubated with or without different concentrations of proteases ( ficin , proteinase K and bromelain; all from Sigma-Aldrich ) for 30 min at 37°C before incubation with fiber knobs , and, iii ) treated with or without Vibrio cholerae neuraminidase for 1 h at 37°C before incubation with fiber knobs ., Purified HAdV-52 virions were resolved on 10% Bis-Tris denaturing gels ( NuPAGE , Invitrogen , Life Technologies ) and transferred to Trans-Blot nitrocellulose membranes ( Bio-Rad Laboratories , Solna , Sweden ) by electroblotting ., The membrane was blocked with 5% milk in PBS-T ( PBS supplemented with 0 . 05% Tween20 ) ., Staining was carried out using 1:5000 dilution of a monoclonal anti-adenovirus fiber antibody ( epitope region suggested by the manufacturer: MKRARPSEDTFNPVYPY , clone 4D2 , ab3233 , Abcam ) in PBS-T with 2 . 5% milk , followed by a 1:1000 dilution of a HRP-conjugated rabbit anti-mouse IgG antibody ( Dako Cytomation ) in PBS-T with 2 . 5% milk ., The fibers were then detected by chemiluminescence using super signal west pico or femto ( Thermo Scientific ) and visualized using the multipurpose CCD camera system FujiFilm LAS-4000 ., Pictures were taken every 10s and the relative abundance of the two fibers were evaluated using ImageJ ., A549 and HCE cells were detached with PBS-EDTA , reactivated in growth medium for one hour at 37°C , pelleted in 96 well plates ( 2x105 cells/well ) and washed once with PFN ., The cells were then incubated with a mouse monoclonal antibody directed against CAR ( E1-1 , Merck Millipore ) for 30 min on ice followed by one wash with PFN ., A polyclonal rabbit-anti-mouse FITC antibody ( Dako Cytomation ) was added to the cells and incubated for 30 min on ice followed by one wash with PFN before flow cytometry analysis ., 96-well plates ( Nunc maxisorp , Thermo Scientific ) were coated with 1 μg/ml of fetuin or asialofetuin type I or II ( Sigma-Aldrich ) for 2 h at room temperature ( RT ) in coating buffer ( bicarbonate/carbonate coating buffer 100 mM , pH 9 . 6 ) ., Meanwhile fiber knobs ( 0 . 4 μg/ml ) were preincubated with monoclonal anti RGS-His antibodies ( Qiagen; dilution 1:1000 ) in PBS-T for 1 h at RT ., The wells were then washed four times with PBS-T and incubated with the fiber knob mixtures for 1 h at RT ., After washing , the plate was incubated with a HRP-conjugated rabbit anti-mouse IgG antibody ( Dako Cytomation; diluted 1:2000 in PBS-T ) for 1 h at RT ., The wells were washed again and incubated with 100 μl enhanced K-Blue TMB substrate ( Neogen Europe ) for 15 min and the reaction was then stopped by addition of 100 μl 1 M H2SO4 ., The absorbance was measured at 450 nm using Tecan infinite F2000 Pro ( Tecan Nordic AB ) ., All SPR experiments were performed at 25°C with a Biacore T100 instrument and a data collection rate of 1 Hz ., For CAR interaction studies: CM5 sensor chips , amine-coupling kit , and HBS-EP+ buffer ( 10 mM HEPES , 150 mM NaCl , 3 mM EDTA , 0 . 005% vol/vol surfactant P20 , pH 7 . 4 ) were all purchased from GE Healthcare ., Recombinant human CAR ( CXADR Fc chimera; R&D Systems; full length extracellular D1D2 domain ) , or CAR-D1 was coupled to the CM5 sensor chip by using the amine coupling reaction according to the manufacturer’s instructions , resulting in an immobilization density of 900–1100 RU ., The surface of the upstream flow cell was subjected to the same coupling reaction in the absence of protein and used as reference ., All binding assays were carried out at 25°C , and HBS-EP+ buffer was used as running buffer ., The analytes ( 52LFK and 52SFK ) were serially diluted in running buffer to prepare a two-fold concentration series ranging from 8 nM to 2 μM , and then injected in series over the reference and experimental biosensor surfaces for 120 s at a flow rate of 30 μl/min ., Blank samples containing only running buffer were also injected under the same conditions to allow for double referencing ., After each cycle , the biosensor surface was regenerated with a 60 s pulse of 10 mM Tris-Glycine pH 1 . 5 at a flow rate of 30 μl/min ., For 52SFK interaction studies: Ni-NTA sensor chips , and HBS-EP+ buffer were purchased from GE Healthcare ., 52SFK was diluted in running buffer ( HBS-EP+ ) to a concentration of 0 . 03μM and captured on the Ni-NTA sensor chip according to the manufacturer’s instructions , resulting in an immobilization density of 700 RU ., In short: an automated program cycle of the following sequence: ( 1 ) activation of the sensor chip with Ni ( II ) , ( 2 ) capture of 52SFK ( 3 ) analyte injection , ( 4 ) regeneration of the surface with 0 . 3 M EDTA , and ( 5 ) rinse with HBS-EP+ without EDTA ., All steps we | Introduction, Results and Discussion, Materials and Methods | Most adenoviruses attach to host cells by means of the protruding fiber protein that binds to host cells via the coxsackievirus and adenovirus receptor ( CAR ) protein ., Human adenovirus type 52 ( HAdV-52 ) is one of only three gastroenteritis-causing HAdVs that are equipped with two different fiber proteins , one long and one short ., Here we show , by means of virion-cell binding and infection experiments , that HAdV-52 can also attach to host cells via CAR , but most of the binding depends on sialylated glycoproteins ., Glycan microarray , flow cytometry , surface plasmon resonance and ELISA analyses reveal that the terminal knob domain of the long fiber ( 52LFK ) binds to CAR , and the knob domain of the short fiber ( 52SFK ) binds to sialylated glycoproteins ., X-ray crystallographic analysis of 52SFK in complex with 2-O-methylated sialic acid combined with functional studies of knob mutants revealed a new sialic acid binding site compared to other , known adenovirus:glycan interactions ., Our findings shed light on adenovirus biology and may help to improve targeting of adenovirus-based vectors for gene therapy . | HAdVs are common pathogens in humans , causing disease mainly in eyes , airways and gastrointestinal tract ., Most HAdVs are equipped with twelve protruding fiber proteins that mediate attachment to host cell receptor molecules ., Recently , a new human gastroenteritis-associated adenovirus ( HAdV-52 ) was identified and classified as the first member of a novel species ( HAdV-G ) ., Unlike most other HAdVs , this virus contains two different fiber proteins , a long and a short one , a feature shared only with the two members of species HAdV-F ( HAdV-40 and -41 ) ., To gain further insights into the mechanisms of HAdV-52 infection of human cells , we set out to identify the host cell receptors used by the long and short fibers ., We find that the long fiber binds to a protein-based receptor known as the coxsackievirus and adenovirus receptor ( CAR ) , and that the short fiber binds to glycoproteins that contain sialic acid-capped glycans ., The crystal structure determination of a complex of the short fiber knob bound to sialic acid demonstrates that this interaction is unique among HAdVs , and bioinformatic analysis indicates that simian AdVs may also engage sialic acids in the manner seen in HAdV-52 ., The results presented here provide insights into the plasticity of adenovirus-host cell interactions . | null | null |
journal.pgen.1003612 | 2,013 | ENU-induced Mutation in the DNA-binding Domain of KLF3 Reveals Important Roles for KLF3 in Cardiovascular Development and Function in Mice | Congenital heart defects are the most common congenital malformations in humans affecting 1–2% of live births 1 and 18% of stillbirths 2 ., Causative mutations have been identified in families with inherited congenital heart defects 3 but in most cases remain unknown 2 ., A strong genetic role is nevertheless likely given high heritability scores , for example >0 . 7 for left-sided congenital heart defects 4 , 5 , 6 ., To discover new genes important in cardiovascular development , we measured aortic blood velocity in an ultrasound screen undertaken to assess left ventricular outflow function , in the offspring of N-ethyl-N-nitrosourea ( ENU ) mutagenized male mice 7 ., One mutant had very high aortic blood velocities due to aortic valvular stenosis and this trait was heritable ., Additional abnormalities in cardiovascular development and function were found in subsequent phenotyping of this mutant mouse line ., A dominant point mutation in the region encoding the DNA binding domain of Klf3 was found by linkage analysis and gene sequencing ., KLF3 is a zinc finger transcription factor that has discrete regions of expression that are widely distributed among embryonic and adult tissues in mice 8 , 9 ., KLF3 functions predominantly as a gene repressor 10 although it also has activator functions 11 ., KLF3 had hitherto no described role in heart or vascular development or function ., In prior work , homozygous deletion of the region encoding the Klf3 zinc finger DNA binding domain caused partially penetrant perinatal lethality in mice and significant abnormalities in adiposity 12 , B cell development 13 , and erythroid maturation 14 whereas cardiovascular defects were not reported ., However , embryonic lethality 12 occurred at a stage of development consistent with death due to cardiovascular dysfunction 15 ., Furthermore , several Klfs are expressed in cardiomyocytes and vascular smooth muscle cells 16 including Klf3 ( current study and 17 ) ., KLF3 is enriched at promoters of several muscle-specific genes including muscle creatine kinase ( MCK ) where it interacts with Serum Response Factor to act as a transcriptional activator 11 ., Thus , despite known molecular mechanisms whereby KLF3 may alter cardiac or vascular development or function at a cellular level , a cardiovascular phenotype remained unidentified ., Herein we report the characterization of a new ENU-induced mouse mutant ., Results reveal important and novel roles for KLF3 in cardiovascular development and function ., Strong similarities in phenotype with homozygous Klf3 gene trap mice , where KLF3 is largely eliminated , suggest a predominantly dominant negative effect of the point mutant protein ., However , intriguingly , the existence of divergent traits suggests the involvement of additional interactions ., At the molecular level , the point mutation illuminates the critical importance of a highly conserved residue in the DNA binding domain of KLF3 ., The discoveries reported here provide impetus for exploring the KLF3 pathway to discover new causative factors contributing to cardiovascular disorders in humans ., In a screen of 1770 adult heterozygous offspring from ENU mutagenized C57BL/6J male mice crossed with wild-type ( WT ) C3H/HeJ females , we identified a mutant mouse with an aortic blood velocity >7 standard deviations ( SD ) above the mean ., Using a cut-off aortic blood velocity of 150 cm/s ( i . e . >3 SD above the mean of all animals ) , we found that the trait ( Figure 1A ) was heritable when mutants were bred to BALB/cJ females although only 10% ( 17 of 165 ) had the trait ., Nevertheless linkage analysis localized the mutation to chromosome 5 between 4 . 9 and 75 . 6 Mb ( Figure S1A ) ., The LOD score exceeded 4 in this interval ( Figure S1B ) whereas it was <2 . 5 elsewhere in the genome ( not shown ) ., The incidence of the trait was higher on a C57BL/6J ( B6 ) background ( 136 of 584; 23% ) so we performed fine mapping by crossing affected animals with B6–Chr 5 A/J consomic mice ( incidence of trait was 40 of 183; 22% ) ., We narrowed the interval to a 12 . 6 Mb region on chromosome 5 ( Figure S1B ) , which contained 35 genes ( Table S1 ) ., Genomic sequencing of 7 candidates ( Table S1 ) revealed only one point mutation predicted to affect the protein product ( Figure 1B ) ., The mutation in exon 5 of Klf3 ( Krüppel-like factor 3 ) ( Figure 1C ) changed a histidine residue ( CAC ) at amino acid 275 to arginine ( CGC ) ( KLF3H275R ) ( Figure 1B ) ., This histidine is conserved across species ( www . ncbi . nlm . nih . gov/homologene ) and across all but one of the 22 Sp/Klf family members 18 ., It is the central of 3 amino acids predicted to make contact with DNA in the DNA binding region of the first of three zinc finger domains in KLF3 18 , 19 ., We predicted that mutation at this site would be highly likely to affect the DNA binding function of KLF3 and thereby its function in transcriptional control ., The Klf3H275R line was subsequently maintained by breeding with B6 mice ., High peak aortic blood velocity in adult heterozygous Klf3 point mutants ( Klf3H275R/+ ) was caused by aortic valvular stenosis as shown by augmented valvular gradients in blood velocity ( Figure 2A ) and blood pressure in Klf3H275R/+ mice ( Figure 2B , C ) , and by abnormal valve morphology detected by gross dissection ( not shown ) , histopathology ( Figure 3A ) , and scanning electron microscopy ( Figure 3B ) ., Aortic valves were tricuspid although bicuspid valves were occasionally observed ., The leaflets were thickened , often partially fused , and sometimes exhibited blebs or small hematomas ( Figure 3A ) ., When genotype was used to identify mutants , most Klf3H275R/+ mice had peak velocities >150 cm/s ( 20 of 31 or 65% ) in contrast with WT littermates ( 0 of 43 or 0% ) ( Figure 3D ) ., Males and females were similarly affected ., Significant aortic valve regurgitation was not observed ., In humans , aortic valvular stenosis is often associated with post-stenotic aortic dilatation 20 ., We therefore measured diastolic diameter of the ascending aorta in vivo and found a significant 27% post-stenotic enlargement in male and female Klf3H275R/+ mice ( 1 . 95±0 . 07 vs . 1 . 53±0 . 08 mm in males ( n\u200a=\u200a5 ) and 1 . 70±0 . 09 vs . 1 . 34±0 . 03 mm in females ( n\u200a=\u200a4 ) ; P<0 . 01 ) ( Figure 3C ) ., We next examined blood velocities through the other heart valves ., Peak blood velocity was ∼30% higher in the main pulmonary artery , and at the atrioventricular valves during early ventricular filling ( E-wave ) in Klf3H275R/+ mice ( Table 1 ) ., E-wave fusion with the atrial filling wave ( A-wave ) occurred significantly more often in the left or right filling waveforms in Klf3H275R/+ ( 9 of 20 ) than WT littermates ( 0 of 20 ) ., There were no abnormalities in peak inflow velocities during atrial contraction ( i . e . A-wave ) or in heart rate ( Table 1 ) , and no evidence of significant valve regurgitation in Doppler waveforms ( not shown ) ., No structural abnormalities in the pulmonary valves ( Figure S2A , B ) or the atrioventricular valves ( not shown ) were detected in adults by gross or histopathology examination ., Thus , the mutation appeared to predominantly impact the aortic semilunar valve ., At weaning , we observed 100% lethality of homozygous offspring and only ∼50% of the anticipated Klf3H275R/+ pups from Klf3H275R/+ intercross breeding ( Table S2 ) ., At E14 . 5–16 . 5 , significant lethality of homozygotes , but not heterozygotes , was observed ( Table S2 ) ., We used histology to examine embryonic heart structure of Klf3H275R embryos at E12 . 5 ( i . e . before the age of lethality ) and at E14 . 5 in heterozygotes and in the few surviving homozygous embryos ., At E12 . 5 ( Figure 4A ) and E14 . 5 ( Figure 4B ) , homozygous embryos exhibited a thinned and disorganized ventricular myocardium and septum suggesting cardiac failure as the cause of their later demise ., At E14 . 5 , ventricular and atrial septation defects were also observed ( Figure 4B ) ., In contrast , heterozygotes at E12 . 5 had apparently normal cardiac anatomy ( Figure 4A ) and at E14 . 5 showed disorganization and thickening of the septal myocardium ( Figure 4B , C ) ., Some had atrial septation defects similar to homozygotes ( Figure 4B ) and enlarged atrioventricular cushion tissue that may have obstructed flow ( Figure 4C ) ., At birth , heterozygous neonates had abnormally thickened myocardial walls by magnetic resonance imaging ( MRI ) ( Figure 5A ) and aortic valve leaflets by histology ( Figure 5B ) and by optical projection tomography ( Figure 5C ) ., No abnormalities in placental weight or histology were detected at E12 . 5 and E14 . 5 ( not shown ) so placental dysfunction was unlikely to play a causative role ( e . g . as in 21 ) ., To better define the age of lethality in heterozygotes , we delivered 3 litters of Klf3H275R/+ crossed with B6 mice by caesarean section at term ( E18 . 5 ) ., Three Klf3H275R/+ embryos had recently died in utero and 2 Klf3H275R/+ died within 30 min with only occasional breathing ( Table S2 ) ., Klf3H275R/+ pups that survived for up to 2 h were significantly smaller ( 1 . 02±0 . 02 g ( n\u200a=\u200a9 ) ) than WT littermates ( 1 . 13±0 . 02 g ( n\u200a=\u200a10 ) ; P\u200a=\u200a0 . 002 ) ., We next allowed 5 litters of Klf3H275R/+ crossed with B6 mice to deliver naturally at term ., One day after birth , cardiac hypertrophy in Klf3H275R/+ pups was significant in surviving pups ( 7 . 5±0 . 2 mg/g body weight ( n\u200a=\u200a9 ) vs . WT 6 . 5±0 . 1 mg/g ( n\u200a=\u200a23 ) ; P <0 . 001 ) whereas it was striking in dead or dying pups on day 1 ( 15±2 mg/g ( n\u200a=\u200a3 ) ; P<0 . 001 ) ., Imaging showed that pups that died within 1 d of delivery had markedly diminished ventricular lumens , markedly thickened ventricular and septal myocardia , and aortic valve leaflets that were short and thick ( Figure 5A , B , C ) ., Thus , heterozygous Klf3H275R/+ pups that had the most pronounced ventricular hypertrophy apparently died in the perinatal period ., Aortic blood velocity was not higher in Klf3H275R/+ pups at day 1 of age ( 49±4 cm/s; n\u200a=\u200a9 ) relative to WT littermates ( 48±4 cm/s; n\u200a=\u200a21 ) ., In Klf3H275R/+ pups assessed on day 1 and again at 8 wk; 6 of 7 developed high velocities ( >2 SD ) by 8 wk ( Figure 5D ) ., Peak aortic blood velocity did not increase further between 9 wk and 1 y ( n\u200a=\u200a15; not shown ) ., Thus , the Klf3 mutation may alter prenatal aortic valve development but the development of sufficient stenosis to elevate aortic blood velocity is a postnatal event occurring by 8 wk in Klf3H275R/+ mice ., Heterozygous mice that survived the perinatal period survived into adulthood ., However the number of adults that died before 60 wk was significantly increased; 21% of Klf3H275R/+ mice with high aortic blood velocities at 8 wk died under 60 wk of age ( 41 of 195 ) vs . 4% of WT cage-mates ( 2 of 54 ) ( P<0 . 0001 ) ., Premature death in adulthood was associated with a rapid deterioration in health and all 4 mice found moribund exhibited marked cardiac enlargement ( e . g . Figure S3; heart weight 0 . 367±0 . 040 g ( n\u200a=\u200a3 ) vs . 0 . 159±0 . 008 g ( n\u200a=\u200a3 ) WT cage mates ) ., In the Klf3H275R/+ group as a whole at ∼20 wk , cardiac enlargement was less pronounced and lung weight was not elevated ( Table S3 ) ., Results are consistent with premature death caused by heart failure ., We anticipated that intraventricular pressures would be elevated due to aortic valvular stenosis in surviving adult Klf3H275R/+ mutants and that this would lead to concentric ventricular hypertrophy ( i . e . increased wall thickness ) secondary to increased afterload ., While the hearts were hypertrophic ( i . e . the heart to body weight ratio was significantly elevated; Table S3 ) , hypertrophy was not concentric because there was no increase in wall thickness in mutant adults ( Table 1 ) ., Furthermore , the heart to body weight ratio did not correlate with aortic blood velocity ( r2\u200a=\u200a0 . 17; P\u200a=\u200a0 . 1; n\u200a=\u200a13 ) or with the transvalvular pressure gradient ( r2\u200a=\u200a0 . 06; P\u200a=\u200a0 . 5; n\u200a=\u200a7 ) in adult Klf3H275R/+ mice ., Indeed , left ventricular systolic blood pressure was not significantly elevated when directly measured in isoflurane-anesthetized mice ( Figure 2C ) ., Instead , we found that surviving adult Klf3H275R/+ mice had eccentric hypertrophy ( i . e . increased chamber dimensions ) ., Thus , other prominent cardiovascular abnormalities were caused by the Klf3H275R allele; abnormalities not due to aortic valve defects ., Echocardiography on Klf3H275R/+ mice was therefore performed to discover other effects of this mutation on adult cardiac function ., We found that the diastolic volume of the left ventricle was significantly increased , and cardiac output was nearly doubled despite their smaller body weight ( Table 1 ) ., The left and right atrial areas measured by ultrasound were also significantly increased in Klf3H275R/+ hearts ( Table 1 ) , as was the heart to body weight ratio ( Table S3 ) ., Left ventricular wall thicknesses and heart rate were unchanged ( Table 1 ) ., Klf3H275R/+ mice had improved systolic function as suggested by a significant increase in ejection fraction ( Table 1 ) and improved diastolic filling as suggested by a 30% increase in the early filling ( E-wave ) velocity for left and right atrioventricular filling ( Table 1 ) ., The high E-wave may explain its more frequent fusion with the A-wave in Klf3H275R/+ mice ( Table 1 ) ., Improved systolic and diastolic function was also supported by a change in the Tei Index ( Table 1 ) , a global indicator of cardiac function 22 ., The change in the Tei Index occurred due to significantly shorter isovolumetric contraction and relaxation times , and a significantly longer ejection time ( Table 1 ) ., Other than eccentric hypertrophy , there was no abnormality in histological structure of the heart , or atrial or ventricular myocardium detected by light microscopy ( Figure S4A ) or electron microscopy although myocardial contraction bands were more frequent in Klf3H275R/+ mice ( Figure S4B ) ., The Klf3H275R/+ mutation caused a doubling of cardiac output ( Table 1 ) ., Physiological increases in cardiac output can be evoked by increased metabolic rate or by reduced oxygen carrying capacity of the blood ., Both would tend to increase tissue requirements for perfusion ., We therefore measured oxygen consumption but found that it was only 15% higher in Klf3H275R/+ mice ( 3112±28 ml h−1 kg−1 vs . 2709±22 ml h−1 kg−1 in WT littermates , P\u200a=\u200a0 . 03; n\u200a=\u200a4 males per group averaged over 24 h ) and thus was insufficient to explain the doubling in cardiac output in Klf3H275R/+ mice ., Similarly , although Klf3H275R/+ mice were slightly anemic with a significant 10% reduction in RBC count ( Figure S5 ) and haemoglobin concentration ( not shown ) , reduced oxygen carrying capacity of the blood was insufficient to explain the large increase in cardiac output in these mutants ., High cardiac output can also be induced physiologically by low total peripheral vascular resistance ., This mechanism was implicated by the significant reduction in arterial blood pressure in Klf3H275R/+ mice both when awake measured by tail cuff ( 98±2 mmHg vs . 108±1 mmHg in WT littermates at 9–12 wk; n\u200a=\u200a10/genotype ) and in the ascending aorta of anesthetised mice using a catheter-tip pressure transducer ( Figure 2C ) ., Heart rate did not significantly differ by genotype whether measured awake ( 677±13 min−1 Klf3H275R/+ vs . 636±21 min−1 in WT; n\u200a=\u200a10/genotype ) or under anesthesia ( Table 1 ) ., We found that low blood pressure was not caused by low plasma volume ., Indeed , plasma volume measured by Evans Blue dilution was 34% greater in Klf3H275R/+ ( 44±3 ml/kg vs . 33±2 ml/kg in WT; P\u200a=\u200a0 . 015; n\u200a=\u200a6 per group ) ., We were unable to find peripheral vascular malformations by color Doppler echocardiography , gross dissection , MRI or histology that could explain the low total peripheral vascular resistance ., These results suggest that the Klf3H275R/+ mutation caused low peripheral vascular resistance by increasing peripheral vascularity and/or vascular calibre ., The resulting low arterial pressure resulted in intraventricular pressures that were not elevated despite aortic valvular stenosis and this likely explains the absence of left ventricular wall thickening in adult Klf3H275R/+ mutants ., Thus the Klf3H275 mutation caused other prominent cardiovascular abnormalities in addition to defects in aortic valve development ., A defect in adipogenesis leading to reduced body weight and fat mass was the primary phenotype reported for homozygous mice with targeted deletion of the KLF3 zinc finger DNA binding domain 12 ., An in vitro role for KLF3 in adipocyte differentiation was also found ., In young adult Klf3H275R/+ mutants in our study , percent body fat was reduced by 14% ( P<0 . 04; n\u200a=\u200a13/genotype ) and body weight by 8% ( P<0 . 001; n\u200a=\u200a13/genotype ) in surviving Klf3H275R/+ mice at 10–11 wk ., At 18–25 wk , the relative weight of the superficial abdominal fat pad was reduced by 43% , which contrasted with increased relative weights of the heart , spleen , kidney , and brain ( Table S3 ) ., Relative lung and liver weights were not affected ( Table S3 ) ., The similarity in body weight and fat mass phenotype between Klf3H275R/+ and Klf3 DNA binding domain deletion mutants 12 suggested that Klf3H275R is a loss of function allele ., However , divergent traits ( below ) suggest that the interaction is more complex ., Klf3 mRNA ( Figure 6A ) and protein ( Figure 6B ) were expressed at wild-type levels in homozygous and heterozygous Klf3H275R embryos at E12 . 5 ., H275R protein exhibited reduced binding to KLF3s canonical CACCC binding region of the β-globin gene promoter both using recombinant bacterial GST-Klf3 zinc finger 1–3 protein ( Figure 6C ) or full length KLF3 protein expressed in COS cells ( Figure 6D ) , but did not interfere with the ability of WT KLF3 to bind to DNA ( Figure S6 ) ., In vivo , the H275R protein significantly opposed the repression of Lgals3 , a gene that is normally repressed by KLF3 14 , in Klf3H275R homozygous embryos at E12 . 5 ( Figure S7 ) ., However , expression of other known KLF3 targets including Klf8 , Crip1 , and Pqlc3 ( Crossley M et al . unpublished ) was unaffected ( not shown ) ., Because a role for KLF3 in cardiovascular development and function was hitherto unknown , we asked whether reduced KLF3 function could cause abnormal cardiovascular development in zebrafish embryos by using anti-klf3 morpholinos to inhibit translation of klf3 transcript ., At 48 hour post-fertilization ( hpf ) , embryos appeared to be developing normally suggesting that initial differentiation and morphogenesis of the heart proceeded normally ., However by 72 hpf , 65% of klf3 knockdown embryos exhibited cardiac edema indicative of cardiovascular dysfunction ( i . e . 49 of 59 , and 27 of 58 in 2 replicate experiments ) and some hearts were visibly dysmorphic and did not properly loop ( Figure 7 ) ., Of those with edema , blood flow was visible in the embryonic vasculature in 60% of embryos at 72 hpf ( i . e . 22 of 49 , and 8 of 27 in the 2 replicates ) with no occlusion of the outflow tract evident ( data not shown ) ., This supports cardiac dysfunction as the primary cause of cardiac edema and heart defects ., Results were consistent over 6 replicate experiments ( 50–100 embryos per experiment ) ., Injection of the mismatch control morpholino had no effect on the developing heart ., To further validate that altered KLF3 function was the cause of the cardiovascular defects observed in Klf3H275R mutants , we generated mutant mice from two embryonic stem cell lines with gene trap vectors inserted near the start of the Klf3 gene ( XS and CH; Figure 8A ) ., These insertions largely eliminated Klf3 mRNA ( Figure 8B ) and KLF3 protein ( Figure 8C ) expression in homozygotes ., If Klf3H275R was a simple loss of function allele , then these gene trap mutants would be anticipated to exhibit perinatal lethality and cardiovascular defects similar to Klf3H275R mutant mice ., At weaning , heterozygosity did not affect survival in Klf3 gene trap mutants ., This contrasted with ∼50% of Klf3H275R heterozygotes dying in the perinatal period ., However , there was significant lethality prior to weaning age in homozygous gene trap mutants compared to WT littermates ( Table S4 ) ., This was reported previously for homozygous mutants lacking Klf3s DNA binding region where ∼half those anticipated were found at weaning whereas the anticipated ratio was observed at E14 . 5 12 ) ., We found that some homozygous gene trap mutants survived to adulthood whereas homozygosity of the point mutation was always embryonic lethal ., To evaluate the effect of Klf3 gene trap mutations on the aortic valve , we measured ascending aortic blood velocity in homozygous adults ., We found that aortic velocity was elevated in a significantly greater proportion of surviving homozygotes of both gene trap lines ( Figure 9A ) ., Heterozygotes were not significantly affected ( not shown ) ., Homozygote gene trap mutants often exhibited thickened aortic valve leaflets by gross morphology ( 5 of 7 mutants vs . 0 of 6 WT littermates ) ( e . g . Figure 9B ) ., Thus , homozygous Klf3 gene trap mutations caused stenotic , malformed aortic semilunar valves that resembled those of the heterozygous point mutant mice ( Klf3H275R/+ ) ., Like adult Klf3H275R heterozygotes , homozygous gene trap mutants also had other diverse cardiac defects ., They had significantly higher cardiac output ( Table 1 ) with no change in left ventricular wall thickness or heart rate ( Table 1 ) and , in homozygous CH mice , significantly higher left ventricular end diastolic volume ( Table 1 ) , consistent with a phenotype of left ventricular eccentric hypertrophy ., Homozygous CH mice also had significantly reduced arterial pressure ( −26 mmHg ) ( Figure 9C ) , which was similar to Klf3H275R heterozygotes ( −10 mmHg ) ( Figure 2C ) ., Also like adult Klf3H275R heterozygotes , homozygous gene trap mutants had significantly reduced body weights , enlarged hearts , and decreased abdominal fat pad weights ( Table S5 ) ., Homozygous Klf3 gene trap mutants exhibited a pronounced right ventricular trait not observed in Klf3H275R/+ mice ., They often had marked pathological enlargement of the right ventricle ( 8 of 15; e . g . Figure 9D ) and an abnormal leftward septal deviation in early diastole ( Movie S1 and S2 ) ., Many also exhibited abnormally thickened pulmonary valve leaflets ( Figure S2C , D ) , and pulmonary valve regurgitation ( 13 of 15 ) that was often associated with tricuspid valve regurgitation ( 7 of 15 ) ( Figure 9E ) ., Lung weight was significantly increased ( Table S5 ) consistent with pulmonary congestion ., No septal defects were detected ., It is noteworthy that valve closure and septal deviation abnormalities could be secondary to a primary right ventricular enlargement defect ., There were also differences in the haematological phenotypes of the point mutant and gene trap mutants ., All 3 mutant lines exhibited greater variation in red cell volume ( %RDW; Figure S5 ) and a larger number of reticulocytes ( immature RBC ) in blood smears suggesting a defect in erythrocyte production , structure , and/or elimination ., However , only Klf3H275R heterozygotes were anaemic and had increased RBC cell size , and only homozygous gene trap mutants had increased white blood cell counts ( WBC ) ( Figure S5 ) ., In Klf3H275R/+ mice , slight anaemia was deemed insufficient to explain the large increase in cardiac output in these mutants ., This is supported by the finding that homozygous gene trap mutants had increased cardiac output ( Table 1 ) but no significant change in RBC count ( Figure S5 ) or haemoglobin concentration ( not shown ) ., Staining for LacZ generated by the XS Klf3 gene trap vector indicated strong Klf3 expression in the E10 . 5–12 . 5 embryonic aorta and cardiac outflow tract ( where heart valve primordia form ) ( Figure 10A–D ) ., LacZ staining in the embryonic myocardium was diffuse and punctate ( not shown ) ., LacZ staining was also observed in the adult atrial and ventricular myocardium , heart valves , and endothelial and vascular smooth muscle of the vasculature ( Figure 10E–L , Figure S8 ) ., Thus local alterations in KLF3 transcriptional activity in cardiovascular cells may directly cause aortic valvular stenosis , hypotension , and abnormal myocardial growth ., Strong LacZ staining was observed at other discrete albeit widespread sites within the embryo ( e . g . Figure 10A , C ) as observed previously by ISH in embryos 8 , 23 and by qRT-PCR in adult tissues 9 so indirect hormonal or neural mechanisms may also play a role in cardiovascular abnormalities in Klf3 mutants ., Gene expression was evaluated in Klf3H275R and CH homozygotes , Klf3H275R heterozygotes , and WT by microarray ( Figure S9 ) ., To increase the likelihood of revealing immediate downstream targets of KLF3 , we used mRNA isolated from whole embryos collected at E12 . 5 d of gestation ., At this stage , Klf3H275R homozygosity was not yet lethal , and Klf3 LacZ expression was widespread ( e . g . Figure 10A ) ., When comparing Klf3H275R homozygotes , heterozygotes , or both groups combined versus WT , no genes were identified as significantly changed using a false discovery rate threshold of 0 . 1 ., For CH homozygotes vs . WT , 18 genes were changed significantly and 2 were changed >2-fold; Klf3 ( 0 . 07× ) and Hsd3b6 ( 0 . 23× ) ( Table S6 ) ., Overall , CH homozygous embryos had 16 genes differentially expressed when compared to Klf3H275R homozygotes ( including 4 genes changed by >2-fold; Klf3 ( 0 . 07× ) , Glmap4 ( 2 . 07× ) , Snora31 ( 2 . 20× ) , and Fah ( 2 . 31× ) ) and 185 genes were differentially expressed when compared to Klf3H275R heterozygotes ( including 5 genes changed by >2-fold; Klf3 ( 0 . 07× ) , Hsd3b6 ( 0 . 38× ) , Mir1948 ( 0 . 48× ) , Snora31 ( 2 . 04× ) , and Slc6a16 ( 2 . 24× ) ) ( Table S6 ) ., These differences in embryonic gene expression profiles may explain phenotype differences between heterozygous point mutant and homozygous gene trap lines that were observed later in development ., We then used qRT-PCR to validate the large down regulation of Hsd3b6 expression found in E12 . 5 embryos by microarray ., Expression was significantly reduced in Klf3H275R heterozygotes ( 0 . 45× ) , and in Klf3H275R homozygotes ( 0 . 27× ) vs . WT ( Figure S10A ) whereas a similar trend in CH homozygotes was not statistically significant possibly due to the smaller sample size ( Figure S10B ) ., We also evaluated Lilra6 by qRT-PCR; among genes with apparent up-regulation by visual inspection of the heat map ( Figure S9 ) , Lilra6 was the most consistently high in Klf3 mutant embryos ( 8 of 11 ) and low in WT ( 4 of 4 ) ., Lilra6 was significantly elevated in Klf3H275R heterozygotes ( 2 . 4× ) , and in Klf3H275R homozygotes ( 13 . 8× ) ( Figure S10C ) whereas the much smaller increasing trend in CH homozygotes ( 2 . 2× ) was not statistically significant ( Figure S10D ) ., These results show that the heterozygous and homozygous presence of the point mutant protein can down or up regulate normal gene expression in E12 . 5 embryos , and that effects may differ from that caused by reduced expression of the native protein ( i . e . in CH homozygotes ) ., Herein we report the discovery of novel and important roles for Klf3 in cardiovascular development and function ., Although KLF3 was identified and cloned nearly 20 years ago 8 and Klf3 knock out mice have been studied 12 , KLF3s role in cardiovascular biology had not been revealed until the current unbiased genome-wide ENU screen , in which prominent abnormalities were discovered in heterozygous Klf3H275R point mutants ., We found abnormalities in embryonic gene expression during organogenesis and in heart morphology at E12 . 5 , increased perinatal lethality associated with marked biventricular myocardial hypertrophy and aortic valve leaflet thickening , and adult survivors exhibited hypotension , aortic valvular stenosis , aortic dilatation , and myocardial hypertrophy with increased chamber size ., KLF3 in vitro is enriched at promoters of muscle-specific genes 11 , and in vivo is expressed by multiple cell types integral to the cardiovascular system including cardiomyocytes , heart valves , and vascular endothelial and smooth muscle cells based on LacZ expression patterns reported here ., Cardiovascular abnormalities may therefore result directly from abnormal KLF3 function in cardiac and/or vascular cells ., Abnormal renal or brain regulatory mechanisms may also contribute to hypotension given that Klf3 is expressed at these sites as well ( e . g . figure 9 A , B; ref #10 ) ., Elucidating the likely multifactorial roles of KLF3 in cardiovascular development and function will require temporal- and cell-type specific control of Klf3 expression in future studies ., A role for KLF3 in cardiac valve development was detected in our ENU mutagenesis screen , in which high aortic blood velocity revealed aortic semilunar valve stenosis in heterozygous adults with Klf3H275R point mutations ., In contrast , other cardiac valves were relatively unaffected although they similarly expressed Klf3 based on LacZ staining ., Valve specificity may arise due to differences in developmental mechanisms in semilunar versus atrioventricular valves 24 ., For example cells derived from the secondary heart field 25 and cardiac neural crest 26 selectively contribute only to semilunar valve formation ., Homozygous gene trap mutants also exhibited a high incidence of aortic valvular stenosis in adults but , in contrast with point mutants , the pulmonary semilunar valve leaflets sometimes appeared thickened histologically and often failed to close sufficiently to prevent regurgitation ., In gene trap mutants , it is possible that pulmonary valve defects were secondary to a primary chamber enlargement defect of the right ventricle , a trait that was also never observed in heterozygous Klf3H275R point mutants ., Although aortic valve cusps were thickened in late gestation in Klf3H275R/+ embryos , aortic valvular stenosis sufficient to elevate aortic blood velocity developed only after birth when the normal separation , elongation , and thinning of the valve leaflets occurs 27 , 28 ., This suggests KLF3 plays a particularly critical role in these later events in aortic valve maturation ., This finding is especially interesting given the paucity of knowledge about the genetic regulation of aortic semilunar valve development and the prevalence of aortic valve defects in humans 24 ., Additional phenotypic characterization revealed broader roles for KLF3 in cardiovascular development and function , roles that were independent of its role in aortic valve development ., In adults , high cardiac output and cardiac hypertrophy due to chamber enlargement ( i . e . eccentric growth ) with no change in left ventricular wall thickness was paradoxical given aortic valvular stenosis in adult Klf3 point mutant and gene trap mutants ., This was explained by the surprising observation that stenosis did not elevate intraventricular systolic blood pressure in Klf3H275R/+ hearts relative to WT ., Instead , normal intraventricular pressure , low arterial blood pressure , and augmented blood volume in adult Klf3H275R/+ were likely caused by low systemic vascular resistance , resulting in a hyperdynamic circulation and the high cardiac output that we observed ., Intriguingly , hypotension in a wide variety of other mouse models does not elicit cardiac hypertrophy and/or an increase in cardiac output ., Examples include transgenic mice with overexpression of eNOS ( −18 mmHg and no change in heart weight t | Introduction, Results, Discussion, Materials and Methods | KLF3 is a Krüppel family zinc finger transcription factor with widespread tissue expression and no previously known role in heart development ., In a screen for dominant mutations affecting cardiovascular function in N-ethyl-N-nitrosourea ( ENU ) mutagenized mice , we identified a missense mutation in the Klf3 gene that caused aortic valvular stenosis and partially penetrant perinatal lethality in heterozygotes ., All homozygotes died as embryos ., In the first of three zinc fingers , a point mutation changed a highly conserved histidine at amino acid 275 to arginine ( Klf3H275R ) ., This change impaired binding of the mutant protein to KLF3s canonical DNA binding sequence ., Heterozygous Klf3H275R mutants that died as neonates had marked biventricular cardiac hypertrophy with diminished cardiac chambers ., Adult survivors exhibited hypotension , cardiac hypertrophy with enlarged cardiac chambers , and aortic valvular stenosis ., A dominant negative effect on protein function was inferred by the similarity in phenotype between heterozygous Klf3H275R mutants and homozygous Klf3 null mice ., However , the existence of divergent traits suggested the involvement of additional interactions ., We conclude that KLF3 plays diverse and important roles in cardiovascular development and function in mice , and that amino acid 275 is critical for normal KLF3 protein function ., Future exploration of the KLF3 pathway provides a new avenue for investigating causative factors contributing to cardiovascular disorders in humans . | Cardiac defects are among the most common malformations in humans ., Most causative genetic mutations remain unknown ., To discover new causative genes important in cardiovascular development and function , we examined 1770 mice with randomly mutated genes and found a mutant with aortic valvular stenosis , and increased risk of fetal and neonatal death ., Using linkage analysis and sequencing , we identified a protein-altering point mutation in the gene regulatory protein KLF3 ., Mice that survived into adulthood with one mutant copy of the Klf3 gene had low arterial blood pressure , enlarged hearts , and increased mortality due to heart failure ., When both copies of the Klf3 gene was mutant , then embryos had heart defects , and all died before birth ., KLF3 had no previously known role in heart development so to confirm these findings , we ( 1 ) knocked down klf3 expression in zebrafish embryos and ( 2 ) examined mice with a mutation that effectively eliminated the KLF3 protein ., In both cases , cardiovascular dysfunction was observed ., In conclusion , we have discovered that KLF3 plays diverse and important roles in cardiovascular development and function in mice ., Future exploration of the KLF3 pathway provides a new avenue for investigating causative factors contributing to cardiovascular disorders in humans . | medicine, developmental biology, genetics, biology, anatomy and physiology, cardiovascular | null |
journal.pcbi.1007121 | 2,019 | Coupling water fluxes with cell wall mechanics in a multicellular model of plant development | Plants grow throughout their lifetime at the level of small regions containing undifferentiated cells , the meristems , located at the extremities of their axes ., Growth is powered by osmosis that tends to attract water inside the cells ., The corresponding increase in volume leads to simultaneous tension in the walls and hydrostatic pressure ( so-called turgor pressure ) in the cells ., Continuous growth occurs thanks to the yielding of the walls to these stretching forces 1–3 ., This interplay between growth , water fluxes , wall stress and turgor was first modelled by Lockhart in 1965 4 , in the context of a single elongating cell ., Recent models focused on how genes regulate growth at more integrated levels 5–9 ., To accompany genetic , molecular , and biophysical analyses of growing tissues , various extensions of Lockhart’s model to multicellular tissues have been developed ., The resulting models are intrinsically complex as they represent collections from tens to thousands of cells in 2- or 3-dimensions interacting with each other ., To cut down the complexity , several approaches abstract organ multicellular structures as polygonal networks of 1D visco-elastic springs either in 2D 7 , 10–12 or in 3D 6 , 13 submitted to a steady turgor pressure ., Other approaches try to represent more realistically the structure of the plant walls by 2D deformable wall elements able to respond locally to turgor pressure by anisotropic growth 8 , 14 , 15 ., Most of these approaches consider turgor as a constant driving force for growth , explicitely or implicitly assuming that fluxes occur much faster than wall synthesis ., Cells then regulate the tissue deformations by locally modulating the material structure of their walls ( stiffness and anisotropy ) 6 , 16–20 ., However , the situation in real plants is more complex: turgor heterogeneity has been observed at cellular level 21 , 22 , which challenges the assumption of very fast fluxes ., As a matter of fact , the relative importance of fluxes or wall mechanics as limiting factors to growth has fuelled a long standing debate 3 , 23 and is still an open question ., Moreover , from a physical point of view , pressure is a dynamic quantity that permanently adjusts to both mechanical and hydraulic constraints , which implies that a consistent representation of turgor requires to model both wall mechanics and hydraulic fluxes ., The aim of this article is to explore the potential effect of coupling mechanical and hydraulic processes on the properties of the “living material” that corresponds to multicellular populations of plant cells ., To this end , we build a model that describes in a simple manner wall mechanics and cell structure , but do not compromise on the inherent complexity of considering a collection of deformable object hydraulically and mechanically connected ., The article is organized as follows ( see Fig 1 ) : we first recall the Lockhart-Ortega model and its main properties ., Then we explore two simple extensions of this model: first we relax the constraint of uniaxial growth in the case of a single polygonal cell; then we study how two cells hydraulically connected interact with each other ., Finally we describe our multicellular and multidimensional model and numerically explore its properties ., A table of notations is provided in Supplementary Information ( S1 Table ) ., In 1965 , Lockhart 4 derived the elongation of a plant cell by coupling osmosis-based fluxes and visco-plastic wall mechanics ., Ortega 24 extended this seminal model to include the elastics properties of the cell walls ., We recall here the main properties of this model , see Fig 1a for the geometrical configuration ., A multicellular extension of the Lockhart-Ortega model adapted to the study of morphogenesis requires first to relax the constraint of uniaxial growth and allow multidimensional geometries , and second is complexified by the possibility of fluxes between cells ., We study separately the effect of each of these extensions before presenting the complete model ., In the Lockhart-Ortega model , the compatibility between wall enlargement and cell volume variation is automatically enforced through the geometrical constraint of uni-directional growth that leads to the identity between the relative growth rate of the cell and the strain rate of the walls ., In contrast , in the multicellular model , this identity is no longer true ., One has to solve the closed set of Eqs ( 7 ) – ( 11 ) and ( 12 ) with respect to the unknowns X , P , and εe ., Despite its apparent simplicity , the problem to be solved is not straightforward as water fluxes induce potentially long range interactions ., In this respect , it differs from most vertex-based models ( e . g 11 , 26 ) where turgor is an input of the model ., The numerical resolution required the development of an original algorithm ( see S5 Text ) implemented in an in-house code ., The properties of this model cannot be as thoroughly studied as those of the simpler models presented above , first because of the numerical cost of the resolution , but above all because it allows an infinite variety of geometries and spatial distribution of its parameters ., We present here a numerical experiment that illustrates on the one hand how the properties of the simple multidimensional and multicellular submodels are combined in the generalized model; in turn the study of these models helps us to anticipate the properties of the generalized model ., And on the other hand , we show that this model is readily applicable to the study of systems of biological interest ., Growth heterogeneities can be triggered by the local modulation of the mechanical properties of the cell walls 27 ., In SAMs , new organs are initiated by a local increase in growth rate that leads to the appearance of small bumps ., Measurements show that physico-chemical properties of walls are modified so that mechanical anisotropy and elastic modulus are decreased ., Our 2D model is used to represent a cross section of a SAM and we explore what effect a local softening of the walls has on growth rate and turgor heterogeneities; based on our previous analysis of the model in simple configurations , we expect that the growth heterogeneities will be maximal for parameters such that the growth is restricted by fluxes rather than wall synthesis ( αa close to 0 ) , cell-cell conductivity is large ( αs close to 1 ) , and the walls deformations are just above the growth threshold , which can be enforced by a low value of the osmotic pressure ( yet large enough to ensure growth ) ., The set of parameters ( REF ) is chosen according to these criteria; then we explore the effect of a higher αa ( ( ALPHA+ ) set ) and lower cell-cell conductivity ( ( CC- ) set ) that should both decrease the growth heterogeneities , and also test the effect of a lower osmotic pressure ( ( PM- ) set ) that should conversely increase the growth heterogeneity ., See S6 Text for detailed explanations on the values of the parameters corresponding to these sets ., We build a mesh made primarily of hexagons ( see Fig 3a ) and first let it grow with homogeneous parameters until the elastic regime ends and plastic growth occurs ., Then we divide by two the elastic modulus of a small group of cells ( marked with a white star in Fig 3a ) that will be referred to as “bump cells” thereafter ., First , Fig 3b shows that the multicellular system grows globally in the same way as the single hexagonal cell studied above; it diverges from the Lockhart predictions because the ratio A/V of the cells is not constant: the ( ALPHA+ ) simulations exhibit a very large initial growth rate that decreases only when the cells are so large that water fluxes become limiting ., The ( PM- ) set leads to a roughly twice lower growth rate than ( REF ) ., The set ( CC- ) leads to the same dynamics at the tissue level as ( REF ) , because the total influx of water is not affected by fluxes between cells in this setup ., Then we turn to the observation of heterogeneities: we focus on the differences between the bump region and the rest of the tissue ., For all the parameters sets , Fig 3c shows that turgor is in general lower in bump cells , but the gap varies depending on the parameters , as it has been predicted by the study of the two-cells model: compared to ( REF ) , the heterogeneity in turgor is increased by a lower cell-cell conductivity ( set CC- ) , and decreased by a larger value of αa ( set ALPHA+ ) ., Decreasing the value of PM ( set PM- ) does not alter much the turgor heterogeneity compared to ( REF ) ., The maps of turgor ( Fig 3e , 3g , 3i and 3k ) confirm visually these observations ., Fig 3d shows the time evolution of γ ˙ / γ ˙ * where γ ˙ * is the relative growth rate predicted by the Lockhart model ( see ( 6 ) ) ; its value is 2% h−1 for ( REF ) , ( CC- ) and ( ALPHA+ ) , and 0 . 5% h−1 for ( PM- ) ., In the considered time frame , the relative growth rate of bump cells is always higher except for ( ALPHA+ ) : after an initial fast increase where bump cells grow faster , the tendency is inversed at t ≈ 20h because the bump cells have grown so much that fluxes become limiting ., In the ( REF ) simulation , while the growth rate of non bump cells is almost constant and close to γ ˙ * , the growth rate of the bump cells is up to 6 times γ ˙ * at the beginning of the simulation and progressively decreases toward γ ˙ * ., As a result of this large discrepancy , the bump region can be clearly distinguished from the rest of the tissue ( Fig 3e and 3f ) ., In ( CC- ) , the growth rate of the non bump cells is close to that of ( REF ) , but the growth rate of the bump cells is much lower ( Fig 3d ) ., As a result , the global shape remains convex and the bump is not clearly detached from the rest of the tissue ( Fig 3i and 3j ) ., Note that ( CC- ) corresponds to a lower value of αs compared to ( REF ) , which corresponded to a lower growth heterogeneity with the two-cells model studied above; this is also confirmed by the lower cell-cell fluxes towards the bump cells for ( CC- ) , see the arrows in Fig 3e and 3i ., The ( ALPHA+ ) simulation exhibits also a convex shape ( Fig 3k and 3l ) ; it corresponds to a larger value of αa than ( REF ) , and similarly to the two-cells model studied above , the growth rate heterogeneity is lower than ( REF ) ., Finally , the set ( PM- ) corresponds to an increase of the dimensionless parameter ρ ( see ( 10 ) ) , and accordingly to an increase in growth rate heterogeneity as can be seen with Fig 3d ., Consequently , the bump region can be clearly distinguished from the rest of the tissue , even better than ( REF ) ( Fig 3g and 3h ) ; moreover , the growth of the cells close to the bump seems to be inhibited by fluxes as explained in the two-cells model described above and further explored below ., The model proposed in this article is a minimal multicellular and multidimensional extension of the Lockhart 1-D single cell model; it can be regarded as a conceptual tool to study the interplay between fluxes and wall mechanics in a multicellular tissue ., Wall expansion is modeled with a visco-elasto-plastic rheological law , while fluxes derive from water potential gradients ., These two contributions are integrated into the mechanical equilibrium and interact through the pressure term ., Contrary to most previous approaches , turgor is not an input of the model but a variable that adjusts simultaneously to mechanical , hydraulic , and geometrical constraints ., First of all , this leads to a physically consistent representation of turgor: for instance , the model predicts that cells with softer walls have a lower turgor ., Moreover , this has deep implications at tissue level: in the previous example , lower turgor is associated with a faster growth which can be itself amplified by fluxes that follow decreasing pressure gradients ., Thanks to the simplicity of the model , the predicted behavior can be analyzed and interpreted with two submodels built from the Lockhart model: first , in a 1-D system , cells are only elongating and their surface-to-volume ratio is constant ., We thus extended the Lockhart model in two dimensions , where cells have more degree of freedom to change their shape ., In particular their allometric surface-to-volume ratio may then vary ., This new possibility induces additional complexity in the tissue development as the rate of growth of cell surfaces may become a limiting factor for growing cells ., Second , a 1-D multicellular submodel was build with two or more side-by-side cells; it was used to study the growth of competing cells with heterogeneous properties ., Key ingredients here are the wall synthesis threshold , the fact that fluxes and growth can relax turgor , and cell to cell fluxes that allow long range interactions ., Depending on mechanical and hydraulic parameters of tissue regions , the model exhibits different growth regimes corresponding to either uniform or differential growth ., One unexpected consequence of such an hydraulic-mechanical coupling at the tissue level is the observation that in certain regions of the parameters space where cell-to-cell hydraulic exchanges are non-limiting , growing tissue may exert an inhibiting influence on the growth of neighboring regions ., This may be interpreted as a lateral inhibition mechanism ., It has for long been recognized that lateral inhibitory mechanisms play a key role in setting some morphogenetic patterns in procaryotes ( e . g . 28 ) , animals ( e . g . 29 , 30 ) or plants ( e . g . 31 , 32 ) ., Lateral inhibition operates in these systems via chemical signals , such as delta-notch in animals or auxin in plants ., Our model predicts the existence of a novel type of lateral inhibition mechanism based on the coupling between mechanics and water fluxes ., Previous observations of tissue growth suggest that such a phenomenon may occur in real tissues ., In the shoot apical meristem for instance , detailed quantification of growth with cellular resolution indicates that the region surrounding primordia growth may have a negative growth rate ( 33 , Figs 2G and 3K ) ., According to our model , this decrease of volume in boundary regions might be due to the primordium growth attracting locally most of the water supply and depriving lateral regions from water , and thus conforts the hypothesis of a new hydraulic-mechanical component of primordium lateral inhibition , beyond already identified auxin and cytokinin signals 34 ., Throughout the development of the model , we made several key choices concerning the abstraction of a multicellular plant tissue ., First , our model was developed in 2-D for reasons of computational efficiency ., In principle , it can be extended in 3-D , though at the expense of more complex formalism and implementation ., Second , the current model considers that water transport is performed in the plant tissue through two conceptually different pathways 1 ., Water can first move within the apoplasmic compartment between the cells and finally enter a cell ., Water can also move locally from cell to cell ., This movement includes itself conceptually both symplasmic movements ( water circulates between cells through plasmodesmata without crossing membranes ) and movements from cell to cell with intermediate steps in the wall ( water is for example exported locally out of the cell by water transporters like aquaporins into the wall and immediately re-imported by water transporters into neighboring cells ) ., For the sake of simplicity in this first analysis , we represented the apoplasm as a single abstract compartment able to exchange water with every cell ., To analyze precisely the effect of water transporters and their genetic regulation or to assess the impact of wall resistance to water movement in the processes , explicit spatial representation of the apoplasm , of plasmodesmata and of membrane water transporters could be integrated into the model in the future ., Finally , we considered a simplified situation here by imposing constant cell osmolarity ., Allowing osmolarity variations ( for instance higher values in faster growing regions ) may impact turgor distribution ( e . g 35 ) ., However , this should not affect the ability of the system to build up growth heterogeneities ., Similarly , we further simplified our model by keeping constant the apoplasmic water potential ., Relaxing this hypothesis would increase cell-cell water fluxes ( via the apoplasm ) and could also shift the model in the direction of the flux-limiting regime ., This would therefore favor regimes where growth heterogeneities are amplified by fluxes ., This model may impact our understanding of various biological questions at the interface between mechanics and hydraulics in plants thanks to its emergent properties that are far more complex and rich than the Lockhart model it is based on ., We showed here the impact of fluxes on turgor and growth rate heterogeneities at tissue level and how they can impact morphogenesis ., In a recent study 22 , we focused on heterogeneities at cell level and compared the model to experimental measurements; in particular , we correctly predicted that the number of cell neighbors is negatively correlated with cell turgor ., Finally , based on its ability to provide quantitative insights in growing multicellular systems , this model could contribute to revisit the long-standing debate initiated by Boyer and Cosgrove regarding the relative importance of fluxes and wall softening in the limitation of growth in plants . | Introduction, Models and results, Discussion | The growth of plant organs is a complex process powered by osmosis that attracts water inside the cells; this influx induces simultaneously an elastic extension of the walls and pressure in the cells , called turgor pressure; above a threshold , the walls yield and the cells grow ., Based on Lockhart’s seminal work , various models of plant morphogenesis have been proposed , either for single cells , or focusing on the wall mechanical properties ., However , the synergistic coupling of fluxes and wall mechanics has not yet been fully addressed in a multicellular model ., This work lays the foundations of such a model , by simplifying as much as possible each process and putting emphasis on the coupling itself ., Its emergent properties are rich and can help to understand plant morphogenesis ., In particular , we show that the model can display a new type of lateral inhibitory mechanism that amplifies growth heterogeneities due e . g to cell wall loosening . | Plant cells are surrounded by a rigid wall that prevents cell displacements and rearrangements as in animal tissues ., Therefore , plant morphogenesis relies only on cell divisions , shape changes , and local modulation of growth rate ., It has long been recognized that cell growth relies on the competition between osmosis that tends to attract water into the cells and wall mechanics that resists to it , but this interplay has never been fully explored in a multicellular model ., The goal of this work is to analyze the theoretical consequences of this coupling ., We show that the emergent behavior is rich and complex: among other findings , pressure and growth rate heterogeneities are predicted without any ad-hoc assumption; furthermore the model can display a new type of lateral inhibition based on fluxes that could complement and strengthen the efficiency of already known mechanisms such as cell wall loosening . | cell walls, plant anatomy, stem anatomy, classical mechanics, plant cell biology, cell cycle and cell division, cell processes, plant physiology, osmotic pressure, shoot apical meristem, plant cell walls, developmental biology, plant science, morphogenesis, cellular structures and organelles, pressure, hydrostatic pressure, relaxation (physics), physics, plant cells, cell biology, biology and life sciences, cellular types, physical sciences, meristems | null |
journal.pcbi.1005826 | 2,017 | Allosteric modulation of cardiac myosin dynamics by omecamtiv mecarbil | Sarcomeric modulators are small molecules that can modify the function of proteins in the sarcomere , the fundamental repeating unit of skeletal and cardiac muscle cells ., New promising avenues for the pharmacological treatment of different muscle and heart diseases rely on direct sarcomeric modulators , which are molecules that can directly bind to sarcomeric proteins and either inhibit or enhance their activity 1 ., Part of the current research is focusing on modulators of myosin II , the motor protein responsible for muscle contraction , with different drugs either in preclinical development or in clinical trials 1–3 ., The possibility to modulate myosin function by either up or down-regulating it is particularly appealing for the treatment of inherited cardiac diseases ., Indeed , myosin mutations are associated with cardiomyopathies with different phenotypes , including hypertrophic ( HCM ) and dilated cardiomyopathy ( DCM ) , and myosin modulators could be potentially used to counteract their damaging effect , with specific drugs tailored for specific mutations 1 , 4 , 5 ., The action of myosin modulators is closely related to the allosteric nature of the protein and in particular of its motor domain ( Fig 1A ) , which is the domain responsible for the hydrolysis of ATP and the conversion of the resulting chemical energy into mechanical work ., The motor domain is composed of four main subdomains , namely the N-terminal ( N-term , green ) , the upper 50-K ( U50K , red ) , the lower 50-K ( L50K , grey ) and the converter ( blue ) subdomains , connected by linkers ( cyan and yellow ) ., The motor domain is then connected to the rest of the myosin molecule through the lever arm helix ( blue ) , which is strongly coupled to the converter ., According to most of the current models of the molecular mechanisms at the basis of myosin function , the relative orientation of the subdomains is controlled by the conformation of the linkers , which is in turn regulated by the biochemical state of myosin ., In particular , Switch 2 ( SW2 ) , the relay helix and the SH1 helix adopt different conformations in the different stages of the acto-myosin cycle , where myosin switches from actin-bound to actin-unbound states according to the nature of the nucleotide bound to it ( Fig 1B ) ., The conformational changes occurring in the linkers are propagated and amplified by the reorientation of the subdomains connected to them , so that small changes in the ATP-binding site in the U50K subdomain can ultimately lead to the powerstroke , a large swinging motion of the lever arm that is caused by the rotation of the converter subdomain and is responsible for the production of mechanical work when myosin is bound to actin ., The powerstroke conformational changes are reversed upon ATP binding and actin unbinding in the recovery stroke , which is the transition from the post-rigor to the pre-powerstroke state that restores the up position of the lever arm 6 , 7 ., In addition to actin and ATP binding sites , the motor domain of myosins has different pockets that can be bound by small molecules 3 , 9 ., Allosteric modulators binding to these sites have been shown to affect myosin function in different ways and using different mechanisms ., In particular , myosin II can be targeted by both inhibitors and activators ., Most of the myosin II inhibitors discovered so far have been found to decrease the release of inorganic phosphate ( Pi ) after ATP hydrolysis , the rate limiting step of the acto-myosin cycle 3 ., While many of these molecules have been used only as research tools because of their toxicity , the recently developed MYK-461 5 and MYK-491 have passed the preclinical development stage and are currently in clinical trials 10 ., In particular , MYK-461 has been tested in mouse models of HCM , where it is has been shown to slow down the progression of the disease 5 ., Structural information on the binding site is available only for some of these inhibitors ., Blebbistatin 11 binds to the 50-K cleft between the U50K and L50K subdomains , while the smooth muscle myosin inhibitor CK-571 has been recently shown to bind to a pocket between the relay and SH1 helices 12 ., Interestingly , CK-571 uses a unique inhibition mechanism where the drug stabilises a previously unknown intermediate step along the recovery stroke and prevents myosin to reach the pre-powerstroke state , thus hindering the ATP hydrolysis 12 ., Myosin activators , namely compounds that enhance myosin activity instead of inhibiting it , have been relatively less studied 3 ., A recent breakthrough in the pharmacological treatment of cardiac disease has been the discovery of the myosin activator omecamtiv mecarbil ( OM ) 13 ., The overall effect of OM on the acto-myosin cycle is to increase the duty ratio , which is the fraction of myosin molecules in the sarcomere that are strongly bound to actin ., This effect is considered to result from an increase in the rate of transition from weakly to strongly bound states that leads to the faster release of Pi measured in the presence of actin13 ., The larger duty ratio causes an increase in the force produced by the sarcomere ( ensemble force ) 4 , so that the overall effect of OM binding is an increased heart contractility 13 ., At the same time , OM has been shown to have an inhibitory effect on the velocity of actin filaments in in vitro motility assays 14 , 15 and on the powerstroke rate in time resolved FRET experiments 16 , while contrasting results have been found when studying its effect on the actin-activated ATP hydrolysis rate 13 , 15 ., The increased force generation induced by OM makes it suitable for the treatment of heart conditions that are characterised by a decreased cardiac contractility ., Indeed , OM is currently due to start phase III clinical trials for the treatment of heart failure 17 , 18 , after previous investigations showed that it was well tolerated and it had beneficial pharmacological effects such as an increased systolic function , a reduced left ventricular wall stress and a beneficial left ventricular remodelling 19–24 ., The trials did not show any evidence of the adverse effects observed with the administration of the drugs currently used to treat heart failure , such as increased heart rate , arrhythmias or hypotension , while the presence of other dose-dependent side effects ( increased troponin levels and reduced diastolic filling times ) will be further investigated in larger patient cohorts 18 ., Understanding the molecular basis of OM action is essential to design strategies for the development of new modulators and their tailoring to specific diseases ., The opposing effects of OM on the kinetics of the acto-myosin cycle suggest that its binding has a complex effect on myosin at the atomistic level ., Recent structural studies 8 showed that OM binds to a deeply buried pocket between the L50K and N-terminal subdomains , close to the SH1 helix and the converter ( Fig 1A ) ., Since these regions are critical for the allosteric communication between the lever arm and the nucleotide binding site , it was suggested that OM can affect the coupling between the lever arm motion and the nucleotide state 8 ., Moreover , small differences were found between Apo and OM-bound structures of the central β-sheet ., Modifications in this part of the protein were suggested to be related to the increase of strongly actin-bound states induced by OM since the central β-sheet is part of the transducer , the element of the motor domain that mediates the communication between the nucleotide and actin binding sites 25 ., The lack of atomistic information on the OM-mediated modulation of myosin dynamics prompted us to perform Molecular Dynamics simulations on the cardiac motor domain bound to OM , where the effects of the drug on the dynamical properties of the protein are investigated for the first time with atomistic resolution ., We find that OM has a double effect on myosin dynamics , inducing, a ) an increased coupling of the motions of the converter and lever arm subdomains to the rest of the protein , which produces a strong reduction in the amplitude of their motions and, b ) a rewiring of the network of dynamic correlations , which produces preferential communication pathways between the OM binding site and functional regions in the U50K subdomain ., The residues and interactions mostly responsible for these effects are also identified and we discuss the possible use of these findings for the future development of improved drugs and the targeting of specific pathogenic mutations ., The overall dynamics of cMotorD during the Apo trajectories was first analysed in terms of the Root Mean Square Fluctuation ( RMSF ) of its Cα atoms ., A high flexibility was found for the long loops in the U50K ( loop1 and cardiomyopathy loop ) and L50K ( loop2 ) subdomains ( Fig 2 ) , as expected on the basis of their length and of the disordered nature of loops 1 and 2 26 ., Interestingly , large RMSF values were observed for the CLD and , to a less extent , the nearby SH3 domain and relay helix ., These large amplitude motions of the CLD domain were found in all the Apo trajectories , albeit with a reduced extent for ApoB2 , indicating that they are independent from either the overall initial conformation of the domain ( chain A or B ) or the loop model used ( 1 or 2 ) ., The cMotorD dynamics was further investigated by identifying the collective motions with a Principal Component Analysis ( PCA ) ., PCA was performed on the Cα atoms only and by excluding the modelled loops to reduce the noise from these disordered regions ( Methods ) ., The first two principal components ( PC1 and PC2 ) of each trajectory , describing the most important collective motions observed during the simulation , accounted together for a significant portion of the overall variance ( S2 Table ) , ranging from 33% ( ApoB2 ) to 65% ( ApoB1 ) ., Two recurring types of motions were found , involving mainly quasi-rigid rotations of the CLD around two different hinge axes , with hinge regions located in the SH1 helix and the relay helix/loop ( Fig 3A and 3B ) ., These CLD motions were the main contributions to the PCs in almost all the cases , even if in different proportions , so that each type of rotation was found either in PC1 or in PC2 according to the specific simulation ( S1 Fig ) ., The exceptions were PC2 from ApoA2 and PC1 from ApoB2 , where rotational motions of the upper 50K domain were found to be of comparable or larger amplitude than the CLD ones ., It is interesting to note that the CLD rotation described by PC1 ( ApoA1 ) has similarities with the conformational changes occurring in the same region during the acto-myosin cycle ., In particular , the transition from the post-rigor to pre-powerstroke state ( ‘recovery stroke’ ) as described by X-ray structures involves a rotation of the CLD around an axis with similar direction as PC1 and similar hinge regions ( Fig 3B ) , but with an amplitude much larger than the one observed here ( ~70° instead of ~ 30° in ApoA1 ) ., The CLD rotation in the recovery stroke is also known to be associated with the formation of a kink in the relay helix 6 ., Remarkably , a significant bending of the relay helix was found during the ApoA1 and B1 simulations ( Fig 4 ) ., In particular , at the end of the ApoB1 simulation the helix formed a kink similar in amplitude and position to that found in the X-ray structures of the pre-powerstroke state ( Fig 4 , left inset ) ., The bent part of the helix was found to adopt different orientations in addition to those observed in the experimental structures ., Overall , these data suggest that the motor domain can partially sample motions similar to those involved in the recovery stroke even in the absence of ATP ., To summarise this section , we found that the intrinsic dynamics of cMotorD in the Apo state is dominated by the CLD , which is relatively free to move as a quasi-rigid domain ., These types of CLD motions , namely rigid rotations around hinges located in the SH1 and relay helices , are consistent with the changes associated with the actomyosin cycle , even if of smaller amplitude ., Moreover , in two of the simulations the CLD motions were associated with a significant bending of the relay helix ., As mentioned above , OM binds to a critical region of cMotorD where it can interact with multiple subdomains at the same time ( Fig 5A ) ., Indeed , the X-ray structures show that in both chain A and chain B conformations OM is in contact with residues of the N-terminal domain ( A91 , M92 , L96 , S118 , G119 , F121 ) , the relay helix ( M493 , E497 ) , the SH1 helix in the L50K subdomain ( V698 , G701 , I702 , C705 ) and the CLD ( P710 , N711 , R712 , L770 ) 8 ., OM was stably bound to the protein during the simulations ( S2A Fig ) ., Its interactions with M90 , A91 , M92 , L96 , S118 , M493 , I702 , C705 , P710 and R712 were found to be particularly stable , since the OM-residue distance was below 4 Å for at least 50% of the frames in all the OM-bound trajectories ( Fig 5B and S3 Table ) ., Chain A and chain B simulations produced similar contact fingerprints ., Residues A91 , S118 , L120 , C705 , N711 , R712 and K762 formed transient hydrogen bonding interactions with OM , which were dependent on the specific conformation adopted by OM ( S4 Table and S3 Fig ) , while strong hydrophobic interactions with M90 , A91 , M92 , L96 , M493 and I702 were observed in all the simulations ( side chains of non-polar residues within 4 Å from OM for at least 50% frames , S3 Table ) ., The molecule was relatively flexible ( S2B Fig ) , as expected from the fact that it adopts different conformations in the X-ray structures of chain A and chain B ( blue and light blue sticks in S2C Fig ) ., A cluster analysis ( S1 Text ) performed on the concatenated trajectories ( S5 Table and S2C Fig ) showed that the most populated cluster ( 82% ) was sampled almost equally in all the simulations ( except for OMA1 , where it contributes with a smaller proportion ) , with the representative structure having an RMSD from the X-ray structures of 2 . 2 ( chain A ) and 2 . 1 ( chain B ) Å ( red sticks in S2C Fig , left panel ) ., A significantly different conformation ( cluster 4 , S5 Table ) was sampled for a short amount of time ( 2% ) at the end of the OMA1 simulation , where the methyl-pyridinyl ring was shifted upwards toward the β1-β2 loop ( S2C Fig , right panel ) ., This conformation might be related to the restructuring of the OM binding pocket recently observed in the pre-powerstroke state , where the pocket is shifted upwards as a result of the lever arm motion during the recovery stroke 27 ., The behaviour of the OM-binding site was further analysed by monitoring the inter-residue contacts in proximity of OM and comparing their stability with that found in the Apo trajectories ( Fig 6 and S4 Fig ) ., OM-bound simulations in general presented a larger number of stable inter-residue contacts ( inter-residue distance < 4 Å for at least 70% of the simulation ) ., OM-stabilised contacts were found between the N-terminal domain and the converter ( T94-N711 for OMA simulations and T94-G771 for OMB ) , and the relay helix and the converter ( Y501-R712 for OMA and E497-R712 and E500-K762 for OMB ) ., Moreover , OMA trajectories showed additional hydrophobic contacts between the central β-sheet and the SH1 helix ( F121-V698 and F121-G697 ) and the β-sheet and the relay helix ( L120-F489 ) , while stronger intra-CLD contacts ( F709-R712 and N711-G768 ) were found in OMB trajectories ., The reduced number of CLD OM-stabilised contacts found in chain A simulations is probably due to the small re-organisation of CLD observed in these trajectories ., Indeed , chain A and B initial structures have small differences in the orientation of the CLD , with the chain B CLD slightly rotated in the direction of the recovery stroke ( S5 Fig ) ., During the OM-bound simulations , chain A relaxed towards the chain B conformation , while the chain B conformation was stable throughout the whole trajectory ( S6 Fig ) ., The overall effect of OM on cMotorD interactions was then to enhance the contacts between the different subdomains that compose its binding site , both by directly interacting with them and by stabilising the inter-residue contacts around it ., This had a dramatic effect on the overall flexibility of the protein , as shown by the RMSF profiles ( lower panels in Fig 2 and S7 Fig ) ., Indeed , a significant reduction in mobility was found for the CLD , together with the neighbouring SH3 domain and the SH1 and relay helices ., Correspondingly , the first two collective motions ( PC1 and 2 ) , while showing a higher diversity across the OM-bound replicas compared to the Apo simulations , were consistently found to have higher contributions from the other subdomains ( the upper domain for chain A simulations and either the N-terminal domain or the lower domain for chain B ) rather than the CLD ( Fig 3A and S8 Fig ) ., Moreover , when the CLD contributed significantly to the PCs , its motion was correlated with the neighbouring subdomains rather than anti-correlated ( insets in Fig 3A ) ., The overall amplitude of the global motions was also significantly reduced compared to the Apo simulations ( S9 Fig ) ., The results obtained comparing the single simulations were confirmed by a PCA performed on the combined Apo and OM-bound trajectories ( S10 Fig ) ., The directions in the conformational space that best discriminate between the two binding states ( 43% of the overall variance ) are represented by the two types of CLD hinged rotations observed in the single Apo simulations ( S10A Fig ) ., A projection onto PC1 and 2 shows a clear separation of the Apo and OM simulations , since the Apo trajectories span a much larger range of values along both components while the OM trajectories are located in the same region of the space ( S10B Fig ) ., Finally , no significant bending was found for the relay helix in the presence of OM ( blue hues and right inset in Fig 4 ) ., A comparison of representative structures from Apo and OM-bound trajectories is presented in S11 Fig , where it is possible to see the different arrangement adopted by the CLD and the relay helix at the end of the simulations ., The previous analysis indicates that the CLD motions found in the Apo simulations are significantly decreased in the presence of OM and that CLD moves in a concerted way with other subdomains rather than freely rotating around the SH1 hinge as was instead observed in the unbound state ., In agreement with this , the comparison of the dynamical cross-correlation matrices ( DCCM ) showed that OM binding induces in all the simulations an increase of the correlations between the CLD and the rest of the protein ( red lines in Fig 7 and red dots in S12 Fig ) , in particular the N-terminal domain , while at the same time decreasing the CLD intra-domain correlations ( green lines in Fig 7 and green dots in S12 Fig ) ., The OM-induced changes in the dynamical correlation within cMotorD were further analysed by reconstructing the network of local correlated motions using the M32K25 Structural Alphabet ( SA ) ( Methods ) ., This type of analysis highlights correlations between changes in the conformational state of 4-residue fragments of the protein backbone during an MD simulation ., While the PCs and the DCCM networks presented in the previous sections are usually dominated by hinge motions of quasi-rigid dynamical domains , local correlations represent more subtle effects involving correlated changes in the local shape of the protein backbone ., Local correlation networks were calculated for each simulation of the Apo and OM-bound state ., A consensus network was then generated for each binding state ( S1 Text ) , resulting in two networks to be compared ( Apo and OM ) ., The networks were first analysed by calculating a preferential connection score ζ between each 4-residue fragment in the OM binding site and the rest of the cMotorD domain ( Methods ) ., Fragments with negative ζ values have a network distance from OM-binding fragments that is smaller than the average , so that they can be considered as preferentially connected to the OM-binding site ., The scores obtained from the OM-bound simulations were then compared with the Apo ones and Δζ differences were calculated by subtracting the Apo from the OM-bound profiles ( Fig 8 ) ., In the following , we will focus on the fragment starting with V698 ( or fragment V698 ) , but consistent results were obtained for the other OM-binding site fragments ( S13 Fig ) ., The ζ values in most of the functional regions show either no change or a decrease upon OM binding ( Δζ < 0 ) , while increased values ( Δζ > 0 ) were usually observed outside these regions ( Fig 8A ) ., This indicates that the OM-binding site tends to have a stronger preferential connection to the functional regions in OM-bound simulations compared to the Apo ones ., Indeed , the Apo simulations showed either a reduced ( smaller |ζ| values compared to the OM ones ) or no preferential connection ( ζ = 0 ) to these regions ( S14 Fig ) ., The most pronounced increases in preferential connection were observed for the G helix , the β5 strand and Switch 2 in the ATP binding site ( Fig 8B ) , which consistently showed negative Δζ values for all the OM-binding site fragments ( S13 Fig ) ., Increased preferential connections were also observed for the β3 and β4 strands and part of Loop 1 in most of the cases ., Interestingly , mapping the mutations known to be associated with dilated cardiomyopathy ( DCM ) 28 onto the Δζ profiles ( yellow points in Fig 8 ) , shows that some of them are in regions preferentially connected to OM ., In particular , I201T ( Loop, 1 ) and A223T ( G helix ) are found in functional regions with consistently negative Δζ values ., This would suggest that their effect could potentially be counteracted by OM binding ., The ζ score changes induced by OM in regions related to myosin function suggest the presence of significant differences in the local correlation network of OM-bound and Apo states ., To investigate this further , the shortest paths were determined between fragment V698 and the preferentially connected functional regions identified above ( Methods ) ., The comparison of the OM-bound and Apo results ( Fig 9A ) shows that the endpoints are more directly connected in the OM-bound network ( bottom panels ) than the Apo one ( top panels ) ., As expected , the paths connecting the OM-binding site ( V698 , magenta ) on one side and the functional regions ( coloured cartoon ) on the other involve a smaller number of edges ( purple lines ) and are thus shorter in terms of distance in the network ( Fig 9B ) in the OM-bound simulations ., The network representation also shows that the paths in the Apo networks involve more nodes that are distant in space from the end points compared to the OM-bound one ( Fig 9A ) ., Moreover , both networks contain a hub node with a large number of connections ( fragment T177 for the OM-bound network and fragment C705 for the Apo one ) , but while T177 is directly connected to most of the functional regions , C705 needs to go through other nodes to reach the endpoints ., In order to identify the residues and interactions mostly involved in this reorganisation of the local correlation network , we analysed the modifications in inter-residue contacts induced by OM binding ( i . e . contacts stabilised or destabilised by OM ) in the region between the OM-binding site and the G-helix ( Fig 10 ) ., Contact matrices were first determined for each Apo and OM-bound simulation by calculating the fraction of the trajectory for which each residue pair was found in contact ., A consensus contact matrix was then derived for each binding state and the OM-Apo difference was used to obtain a matrix representing a network of contact changes ( S1 Text ) ., Calculating the paths connecting V698 ( magenta sphere ) and the G-helix residues ( red cartoon ) in the network ( yellow edges ) shows chains of contact changes going through the two sides of the central β sheet ., The side closer to Switch 2 contains the nodes with the largest number of paths going through them ( larger spheres ) , suggesting that the corresponding residues ( namely F121 , N696 , L693 , T177 , G178 , I462 , K246 , Y266 , L277 and A223 ) are important in mediating the effects on the local correlation network induced by OM binding ., Interestingly , residues T177 and G178 , which are part of the hub fragment T177 in the OM-bound local correlation network ( Fig 9A ) , are involved in most of the paths ( 83% of the paths contain either T177 or G178 , S6 Table ) and they participate in contacts that have among the largest changes in frequency upon OM binding ( S7 Table ) ., In this work , we used MD simulations to investigate the effect of the sarcomeric modulator omecamtiv mecarbil ( OM ) on cardiac myosin dynamics with atomistic resolution ., Simulations were performed to reconstruct the sub-microsecond dynamics of the motor domain of cardiac myosin ( cMotorD ) in the absence and presence of OM , starting from the recently solved structures of the Apo and OM-bound cMotorD in the near-rigor state8 ., The light-chain containing regulatory domain ( RD ) was not considered here since no experimental structure is currently available for the cardiac isoform , however the simulated system included a small fragment of the lever arm helix ., The regulatory role of the RD in cardiac myosin is mediated by the phosphorylation of a disordered portion of the regulatory light chain ( RLC ) , which is thought to be involved in the regulation of the transition from an inactive ( off ) to an active ( on ) form of the two-headed myosin molecule in the thick filament 29 , 30 ., OM has been shown to leave the RLC phosphorylation levels unchanged 31 , suggesting that the molecular mechanisms mediated by OM are independent from those mediated by the RD ., To the best of our knowledge , the present simulations represent the first fully atomistic simulations of cardiac myosin bound to a small molecule modulator ., Previous computational works have focused so far on the effect of nucleotide 32 or actin binding 33 on myosin dynamics , the interactions between actin and myosin 34 , the release of Pi 35 , the modelling of the recovery stroke 36–39 or in general of the coupling between the actin binding site , the nucleotide binding site and the converter 40–43 ., A significant part of these studies used enhanced sampling techniques to accelerate the transitions between the different states in the actomyosin cycle 35 , 36 , 38–41 , 43 , 44 , while unbiased simulations with length > = 50 ns have been performed only recently 32 , 33 , 37 , 45 thanks to the increase of the available computational power ., The simulations presented here show that , in the absence of OM , the dynamics of cMotorD is dominated by the hinge motions of the converter+lever arm helix subdomain ( CLD ) ., These have some resemblance to the CLD rotation observed during the transition from the experimental near-rigor to pre-powerstroke structure ( recovery stroke ) in that, a ) they involve similar hinge regions and, b ) they are associated with the bending of the relay helix ., However , the CLD motions in the simulations seem to be more heterogeneous since , in addition to recovery stroke-like rotations ( Fig 3B ) , the domain can perform rotations in other directions ( e . g . Apo PC2 in Fig 3A ) ., Moreover , the amplitude of the rotations is much smaller than in the recovery stroke , which is expected on the basis of the length of the simulation and the absence of ATP ., The motions observed here , while not representing an actual transition to the pre-powerstroke state and while they might have been enhanced by the absence of the Regulatory Domain , suggest a conformational selection scenario for the recovery stroke , where the type of motions involved in the stroke are already partially sampled by the CLD before the binding of ATP ., This finding is in agreement with the emerging model that CLD rotation is stabilised by the closure of Switch 2 upon ATP binding rather than being induced by it , so that it can occur at least partially before the changes in the nucleotide binding site take place 12 , 43 , 46 ., This would also explain the recent observation that myosin can be trapped by inhibitors in intermediate states of the recovery stroke without closure of Switch 2 12 ., Moreover , a decoupling between Switch 2 and lever arm motions is consistent with the recent observation of a new structural state of Myosin VI where a large conformational change of Switch 2 does not produce a corresponding change in the lever arm position 47 ., OM is shown by the simulations to have a double effect on myosin dynamics ., The first is to dramatically reduce the CLD motions observed in the Apo state and couple them to the rest of cMotorD , as indicated by the large decrease in CLD motions and the increase of positive correlation between the CLD and the other regions , in particular the SH3 subdomain in the N-terminal region ., A possible role of this subdomain in myosin activation is also suggested by the recent observation that one of the few other myosin activators currently known might bind to SH348 ., Our results indicate that OM acts as a “glue” between the different subdomains that compose its binding site , both by directly interacting with them and by stabilising pre-existing inter-domain interactions ., This effect was observed albeit with different extent in all the simulations , indicating that it is independent from the specific starting conformation ., The second OM-induced effect emerging from the simulations is a reorganisation of the network of correlated local motions , which results in a more efficient and direct connection between the OM-binding site and functional regions compared to the Apo state ., In particular , the networks show the formation of preferential pathways between the OM binding site and distant U50K regions close to ATP binding site , namely the G helix and Switch 2 ., The communication between sites is mediated by a chain of OM-induced contact changes involving residues in the central β-sheet ., Preferential connections in the dynamic correlation networks have been previously observed to be involved in allosteric communication 49 ., Our results thus indicate that OM can modulate the cMotorD dynamics through at least two different molecular mechanisms , which would explain its complex and apparently contradictory effects on the kinetic parameters of the actomyosin cycle16 ., In particular , assuming that OM can induce similar effects on the pre-powerstroke state , the reduction of CLD rotational motions upon OM binding might explain the strong reduction in the powerstroke rotation rate measured with FRET 16 and the overall decrease of the actin sliding velocity 14 , 15 ., This inhibitory effect on the lever arm motions is considered to be consistent with the overall increase in muscle contractility produced by OM binding , since it increases the fraction of time spent by myosin in the force generating state where it is strongly bound to actin 14 , 16 ., On the other hand , the enhanced correlation with the G helix and Switch 2 might be related to the change in the number of myosin molecules strongly bound to actin ., Indeed , the G helix has been shown to move concertedly with Switch 1 during the opening of the actin binding site when myosin dissociates from actin 40 ., Changes in the conformation and/or dynamics of the G helix could affect the energetics of the reverse process , where strongly bound myosin states are produced upon closure of the actin binding site ., Moreover , a Pi release mechanism involving mainly Switch 2 rather than Switch 1 motions has been recently suggested on the basis of a newly found structural state of Myosin VI 47 ., During the revision process of this paper , a new OM-bound structure has become available , where OM | Introduction, Results, Discussion, Methods | New promising avenues for the pharmacological treatment of skeletal and heart muscle diseases rely on direct sarcomeric modulators , which are molecules that can directly bind to sarcomeric proteins and either inhibit or enhance their activity ., A recent breakthrough has been the discovery of the myosin activator omecamtiv mecarbil ( OM ) , which has been shown to increase the power output of the cardiac muscle and is currently in clinical trials for the treatment of heart failure ., While the overall effect of OM on the mechano-chemical cycle of myosin is to increase the fraction of myosin molecules in the sarcomere that are strongly bound to actin , the molecular basis of its action is still not completely clear ., We present here a Molecular Dynamics study of the motor domain of human cardiac myosin bound to OM , where the effects of the drug on the dynamical properties of the protein are investigated for the first time with atomistic resolution ., We found that OM has a double effect on myosin dynamics , inducing, a ) an increased coupling of the motions of the converter and lever arm subdomains to the rest of the protein and, b ) a rewiring of the network of dynamic correlations , which produces preferential communication pathways between the OM binding site and distant functional regions ., The location of the residues responsible for these effects suggests possible strategies for the future development of improved drugs and the targeting of specific cardiomyopathy-related mutations . | Cardiac myosin is a motor protein responsible for the contraction of the heart muscle ., New strategies for the cure of heart diseases are currently being developed by using myosin modulators , which are small molecules that can interact with myosin and modify its activity ., The advantage of this approach over traditional drugs is that by directly targeting cardiac myosin it is possible to have drugs with reduced side effects ., Moreover , the availability of a spectrum of molecules to finely tune myosin to a desired level of activity opens the possibility to develop more precise and personalised drug therapies ., In this work , we study a recently discovered activator of cardiac myosin , omecamtiv mecarbil , in order to understand its mechanism of action ., In particular , we use Molecular Dynamics simulations to unravel the effects of the drug on myosin motions , which are closely related to its function ., We find that omecamtiv has a strong effect on myosin dynamics and it changes the way regions of the protein that are critical for its function interact with each other ., We use these data to identify genetic mutations associated with heart diseases that could be targeted by the drug and to suggest a possible route to design drugs with different therapeutic properties . | medicine and health sciences, simulation and modeling, multivariate analysis, molecular motors, actin motors, mathematics, statistics (mathematics), pharmaceutics, motor proteins, research and analysis methods, contractile proteins, proteins, mathematical and statistical techniques, principal component analysis, drug therapy, biochemistry, cytoskeletal proteins, biochemical simulations, cell biology, myosins, biology and life sciences, physical sciences, computational biology, statistical methods | null |
journal.pcbi.1003008 | 2,013 | An Integrated Computational/Experimental Model of Lymphoma Growth | Monoclonal antibodies and small molecule inhibitors of intracellular targets are being developed alongside a host of anti-non-Hodgkins lymphoma therapeutic options 1 ., Yet the tumor tissue-scale effects from these molecular-scale manipulations are not well-understood ., With the ultimate goal to more rationally optimize lymphoma treatment , we integrate pre-clinical in vivo observations of lymphoma growth with computational modeling to create a platform that could lead to optimized therapy ., As a first step towards this goal , we develop the capability for simulation in order to gain insight into the tissue-scale effect of molecular-scale mechanisms that drive lymphoma growth ., We use the modeling to study these mechanisms and their association to cell proliferation , death , and physical transport barriers within the tumor tissue ., Tumor growth and treatment response have been modeled using mathematics and numerical simulation for the past several decades ( see recent reviews 2–9 ) ., Models are usually either discrete or continuum depending on how the tumor tissue is represented ., Discrete models represent individual cells according to a specific set of bio-physical and -chemical rules , which is particularly useful for studying carcinogenesis , natural selection , genetic instability , and cell-cell and cell-microenvironment interaction ( see reviews by 10–20 ) ., Continuum models treat tumors as a collection of tissue , applying principles from continuum mechanics to describe cancer-related variables ( e . g . , cell volume fractions and concentrations of oxygen and nutrients ) as continuous fields by means of partial differential and integro-differential equations 2 ., A third modeling approach employs a hybrid combination of both continuum and discrete representations of tumor cells and microenvironment components , aiming to develop multiscale models where the discrete scale can be directly fitted to molecular and cell-scale data and then upscaled to inform the phenomenological parameters at the continuum scale ( see recent work by 21–23 ) ., There is a paucity of mathematical oncology work applied to the study of non-Hodgkins lymphoma , with some notable exceptions providing insight into the role of the tumor microenvironment heterogeneity in the treatment response 24 , 25 and the disease origin 26 ., Like many other cancers ( solid tumors ) , two critical tissue-scale effects in lymphoma are hypoxia and angiogenesis , as observed in our studies and other work 27 ., Supporting previous qualitative observations of physiological resistance , mathematical modeling and computational simulation have shown that the diffusion barrier alone can result in poor tumor response to chemotherapy due to diminished delivery of drug , oxygen , and cell nutrients 28 , 29 ., Local depletion of oxygen and cell nutrients may further promote survival to cell cycle-specific drugs through cell quiescence ., In order to study these effects in lymphoma , we implement an integrated computational/experimental approach to quantitatively link the processes from the cell scale to the tumor tissue-scale behavior in order to gain insight into their cause and progression in time ., We extend a version of our 3D continuum model 30–32 , building upon extensive mathematical oncology work 2 , 3 , 33–35 , and calibrate both parameters and equations , i . e . , functional relationships that are not conservation laws , from detailed experimental data to produce a virtual lymphoma ., We obtain the experimental data by very fine sectioning of both drug-sensitive and -resistant lymphomas , thus visualizing molecular , cellular , and tissue-scale parameter information across the whole tumor geometry ., We further develop the protocols for calibration of parameters by building on recent work based on patient histopathology 36 , 37 ., We also use the data to derive the relationships between model parameters for apoptosis , proliferation , and vasculature ., We verify the model results at the tumor-scale through tissue-scale observations in vivo of tumor size , morphology , and vasculature using intravital microscopy and macroscopic imaging of the inguinal lymph node ., We note that comparison of model results to experimental data has been done to various extents for different cancers ( see reviews above ) ; here , we perform a tissue-scale comparison after extensive calibration of cell-scale parameters in order to validate the model results ., We undertake simulations to study how the growth of drug-resistant Non-Hodgkins lymphoma may be governed by the cellular phenotype , and use this information to better elucidate the links between physical drug resistance and molecular-scale phenotype by experimental and computational comparison to drug-sensitive tumors ., This process yields a lymphoma simulator as an initial step to study detailed tumor progression and provide further insight into drug resistance , and , ultimately , may provide a tool to design better personalized treatments for Non-Hodgkins lymphoma ., Since the cell-scale measurements used for calibration are different from those at the tissue-scale used for verification , this methodology enables the model to bridge from the cell to the tumor scale to calculate tumor growth and hypothesize associated mechanisms predictively , i . e . , without resorting to fitting to the experimental data ., This process quantitatively links the cellular phenotype to the tumor tissue-scale behavior , and may serve to highlight the importance of physical heterogeneity and interactions in the tumor microenvironment when evaluating chemotherapeutic agents in addition to consideration of chemo-protective effects such as cell-specific phenotypic properties and cell-cell and cell-ECM adhesion 38 ., We choose an Eμ-myc murine orthotopic lymphoma experimental model because of its similarity to human Non-Hodgkins Lymphoma 39 , and select five parameters to measure based on their importance to lymphoma progression: viability , hypoxia , vascularization , proliferation , and apoptosis ., In order to investigate the role of physical heterogeneity in the development of drug resistance , including the impediment of transport barriers , we focus on two types of lymphoma cells: Eμ-myc Arf-/- cells ( Doxorubicin ( DOX ) and Cyclophosphamide ( CTX ) sensitive , with IC50\u200a=\u200a3 . 5 nM and 16 . 0 µM , respectively; the IC50 is the amount of drug needed to kill 50% of a cell population ) , and Eμ-myc p53-/- cells ( DOX and CTX resistant: IC50\u200a=\u200a46 . 2 nM and 75 . 8 µM , respectively ) ., The Eμ-myc transgenic mouse model expresses the Myc oncogene in the B cell compartment , resulting in mice with transplantable B cell lymphomas ., We chose this in vivo model because it captures genetic and pathological features of the human disease and , given the appropriate genetic mutation , drug-resistant and drug-sensitive tumors can be directly compared 39 , 40 ., Eμ-myc/Arf-/- and Eμ-myc/p53-/- lymphoma cells , which harbor loss-of-function regions in the Arf and p53 genes respectively , were previously derived by intercrossing Eμ-myc transgenic mice with Arf-null and p53-null mice , all in the C57BL/6 background as described previously 39 ., Eμ-myc/Arf-/- lymphoma cells and Eμ-myc/p53-/- lymphoma cells were cultured in 45% Dulbeccos modified Eagle medium ( DMEM ) and 45% Iscoves Modified Dulbeccos Medium ( IMDM ) with 10% fetal bovine serum ( FBS ) and 1% penicillin G-streptomycin onto the feeder cells – Mouse Embryonic Fibroblasts ( MEFs ) ., C57BL/6 mice were obtained from Charles River Laboratories ( Wilmington , Massachusetts ) ., All animal studies were approved by The Stanford University Institutional Animal Care and Use Committee ., Lymphoma cells ( 1×106 ) Eμ-myc/Arf-/- and Eμ-myc/p53-/- were diluted with 200 µl of PBS and injected intravenously via the tail vein as described previously 39 ., The intravital microscopy and macroscopic tumor observations were obtained for at least n\u200a=\u200a4 mice per tumor group ., We isolated both Eμ-myc/Arf-/- and Eμ-myc/p53-/- driven tumors at day 21 after tail-vein injection of lymphoma cells ., Typical murine lymphomas were observed to range from about 4 to 6 mm in diameter prior to fixation ., Lymph node tissues were fixed and paraffin-embedded ., The tissues were used for immunohistochemical ( IHC ) identification of cell viability ( H&E staining ) , hypoxia ( HIF-1α ) , vascularization ( CD31 ) , proliferation ( Ki-67 ) , and apoptosis ( Caspase-3 ) ., Five 2-µm thick sections were cut 5 µm apart from each other in order to stain for these markers ( Figure 1 ) ., A total of five sets ( S1 through S5 ) of five stained sections each was collected every 100 µm along the lymphoma , in order to section and stain the entire tumor for sequential microscopic scanning of the stained sections ., Sections S1 and S5 were at the tumor top and bottom , respectively , while the other sections were towards the center with S3 being in the middle ., Note that due to tissue processing and dehydration , the tumors as cut were smaller than measured when removed from the animal ., All the sections were de-paraffinized and rehydrated in PBS ., Then the sections in each set were incubated at 4°C with the primary antibody overnight: rabbit anti-mouse HIF-1 antibody ( Abcam , Santa Cruz , CA ) , rabbit anti-mouse Ki-67 antibody ( Labvision , Fremont , CA ) , rabbit anti-mouse Caspase-3 antibody ( Cell Signaling Technology , Beverly , CA ) , and rat anti-mouse CD31 antibody ( BD Pharmingen , San Diego , CA ) , and incubated for 1 hour at room temperature with a peroxidase-conjugated secondary antibody ., The samples were fully scanned and stitched together using a digital pathology BioImagene instrument ( Ventana Medical Systems , Tucson AZ ) at ×20 magnification ., The model treats tissue as a mixture of various cell species , water , and ECM; each component is subject to physical conservation laws described by diffusion-taxis-reaction equations ( see below ) ., Briefly , the tissue microstructure is modeled through the proper choice of parameter values and through biologically-justified functional relationships between these parameters , e . g . , cellular transitions from quiescence to proliferation depend upon oxygen concentration 41 ., The model simulates non-symmetric tumor evolution in 2D and 3D , and dynamically couples heterogeneous growth , vascularization , and tissue biomechanics ( Figure 2 ) ., In 36 we calibrated models using cell-scale data to predict tissue scale parameters such as size and growth rate ., These models are predictive because they are not calibrated with the same data used for model validation , which avoids data fitting ., While in 36 we focused on the final predicted tumor sizes , here we focus on the growth rate as an essential first step; in follow-up work , we will evaluate the complex problem of drug response ., Our approach to constrain the computational model involves both cell- and tumor-scale approaches as described in Figure 3 ., We approximate the healthy lymph node as a sphere to represent the experiments in the mouse model ( Figure 4 ) ., To simulate node expansion and deformation of surrounding tissue to accommodate the growing tumor , as a first step we delineate the tumor boundary by decreasing the value of the cell mobility parameter beyond the sphere diameter ( see below ) ., For the multigrid algorithm , we pick a computational domain that is a 6 . 4 mm× 6 . 4 mm× 6 . 4 mm box , with finest mesh grid size\u200a=\u200a100 microns; this grid size provides adequate resolution to resolve the tumor boundaries without incurring excessive computational cost ., We assume that the tumor is a mixture of cells , interstitial fluid , and extracellular matrix ( ECM ) ., The temporal rate of change in viable and dead tumor tissue at any location within the tumor equals the amount of mass that is pushed , transported , and pulled due to cell motion , adhesion , and tissue pressure , plus the net result of production and destruction of mass due to cell proliferation and death: ( 1 ) The rate of change in the volume fraction ρi of cell species i ( V: viable tumor; D: dead tumor; H: host ) is specified throughout the computational domain by balancing net creation ( Si: proliferation minus apoptosis and necrosis; see below ) with cell advection ( ∇· ( uiρi ) , where ui is the velocity of the cell species ) and cell-cell and cell-ECM interactions ( adhesion , cell incompressibility , chemotaxis , and haptotaxis , incorporated in a flux Ji ) 31 , 32 ., The reticular network within the lymph node contains a variety of extracellular matrix proteins , many of which are known ligands for integrin cell surface adhesion receptors 42 , 43 ., Cell-cell and cell-ECM mechanical interactions are modeled through J using a generalized Ficks Law 31 ., Tumor angiogenesis is driven by excessive accumulation of cancerous cells , leading to a chronic under-supply of oxygen and cell nutrients ( generically here labeled “nutrients” ) in tumor regions farther removed from pre-existing vessels 44 ., Hypoxic cells in lymphoma release a net balance of pro-angiogenic factors such as VEGF-A , bFGF , PDGF and VEGF-C , which promote neo-vascularization mainly through sprouting angiogenesis of mature resident endothelial cells and , to a lesser extent , through vasculogenesis from recruitment of bone marrow-derived progenitor cells 45 ., Accordingly , the model incorporates angiogenesis into the lymphoma by coupling with a multiscale representation of tumor vessel growth , branching , and anastomosis based on earlier work 46–48 ( further details in Text S1 ) ., The vasculature releases oxygen and nutrients n that diffuse through the tissue and are uptaken by cells during metabolism , while tumor cells secrete VEGF ( nV ) in response to hypoxia 32 ., The oxygen and nutrients are non-dimensionalized by the maximum inside vessels , hence their levels are ≤1 and are assumed to be stationary ., The transport can be described as: ( 2 ) where Dn and are diffusion constants ( 1×10−5 cm2/sec for oxygen 49 and 1×10−7 cm2/sec for VEGF 50 ) , δvessel ( Dirac delta function ) is the indicator function of vasculature ( 1 where it exists and 0 otherwise ) , ν is the delivery rate ( depends upon a , capillary vessel cross-sectional area , and ub , blood velocity ) , , , and are the uptake rates , and are the decay rates ( for simplicity , assumed to be zero ) , and is the secretion rate ., The tumor species viable ( V ) volume fraction ρV is assumed to increase through proliferation and decrease through apoptosis and necrosis ., We assume that normal host cells ( H ) do not proliferate , but may also undergo apoptosis ( A ) and necrosis ( N ) ; the total volume fraction of dead cells ( D ) is ρD ., For simplicity , we assume these primarily affect tumor mass through the transport of water within the tissue and hence neglect their solid fraction ., Under the assumption that a dense viable cell population prevents nutrient saturation , we model the proliferation as directly proportional to ( non-dimensionalized ) nutrient substrate n above a threshold level nN , resulting in the net creation of one cell by removing the equivalent water volume from the interstitium ., Cells experiencing a substrate level below nN are considered quiescent ( e . g . , due to hypoxia ) ., Apoptosis transfers cells from the viable tumor and host cell species to the dead cell species , where cells degrade and release their water content; this models phagocytosis of apoptotic bodies by neighboring viable cells and the subsequent release of the water of lysed cells ., Necrosis occurs when the nutrient substrate concentration falls below the threshold nN and ultimately releases the cellular water content ( i . e . , we assume that the main mode of cell death due to lack of nutrients is mainly represented by necrosis ) ., The resulting model is: ( 3 ) where λM , i , λA , i , and λN , i are mitosis , apoptosis , and necrosis rates , λD is the cell degradation rate ( varies due to the differences between apoptosis and necrosis ) , and H ( x ) is the Heaviside “switch” function ., The movement of a cell species is determined by the balance of proliferation-generated oncotic pressure , cell-cell and cell-ECM adhesion , as well as chemotaxis ( due to substrate gradients ) , and haptotaxis ( due to gradients in the ECM density ) ., We model the motion of cells and interstitial fluid through the ECM as a viscous , inertialess flow through a porous medium ., Therefore , no distinction between interstitial fluid hydrostatic pressure and mechanical pressure due to cell-cell interactions is made ., Cell velocity is a function of cell mobility and tissue oncotic ( solid ) pressure ( Darcys law ) ; cell-cell adhesion is modeled using an energy approach from continuum thermodynamics ( see Text S1 ) ., For simplicity , the interstitial fluid is modeled as moving freely through the ECM ( i . e . , at a faster time scale than the cells ) ., ( 4 ) The variational derivative δE/δρi of the cell-cell interaction potential , combined with the remaining contributions to the flux J ( due to pressure , haptotaxis , and chemotaxis; see Text S1 ) , yields a generalized Darcy-type constitutive law for the cell velocity ui of a cell species i , determined by the balance of proliferation-generated oncotic pressure p , cell-cell and cell-ECM adhesion , as well as chemotaxis ( due to gradients in the cell substrates n ) , and haptotaxis ( due to gradients in the ECM density f ) 32 ., ki is cellular mobility , reflecting the response to pressure gradients and cell-cell interactions , γj is the adhesion force , and χn and χh are the chemotaxis and haptotaxis coefficients , respectively ( see Table S1 ) ., For the host cells , χn\u200a=\u200aχh\u200a=\u200a0 ., The Supplemental Text S1 further describes the ECM density f as well as the effect of the cell velocity on the lymph node geometry ., We used the IHC staining to estimate the number and spatial localization of cells that were viable ( from H&E ) , proliferating ( from Ki-67 ) , apoptotic ( from Caspase-3 ) , hypoxic ( from HIF-1α ) , and with vascular endothelial characteristics ( from CD31 ) ., These estimates were calculated for both Eμ-myc Arf-/- and Eμ-myc p53-/- cells for each set of five sections obtained every 100 µm across the lymphoma ( Figures 5 and 6 ) ., A comparison of viable Eμ-myc p53-/- to Eμ-myc Arf-/- cells along the lymphoma ( Figure 5 ) indicates that the viability is higher for the drug-resistant tumors in the middle of the tumor ( Section S3 ) compared to the drug-sensitive tumors , with a corresponding statistically significant increase in cell density ( p\u200a=\u200a0 . 024; Students t-test with α\u200a=\u200a0 . 05 ) ., In contrast to the Eμ-myc p53-/- tumors , the Eμ-myc Arf-/- seemed to be more dense in the peripheral regions ( p\u200a=\u200a0 . 002 on one end ( Section S1 ) and p\u200a=\u200a0 . 009 on the other end ( Section S5 ) ) , whereas they were about the same for both tumor types in the intermediate sections S2 and S4 ., Tumors with drug-resistant cells have a 4-fold increase in endothelial cells in the core of the tumor ( Section S3 ) compared to drug-sensitive tumors ( Figure 6A ) ., Hypoxia is higher in the peripheral regions for the Eμ-myc p53-/- ( Figure 6B ) even though for both tumor types the peripheral regions seem to be equally vascularized ( based on the endothelial cell density ) ., This could be due to the vasculature on the periphery not being fully functional , with a potential difference in vascular function between the two tumor types leading to a more hypoxic phenotype for the Eμ-myc p53-/- ., Although the core proportionally holds almost twice the number of proliferating cells for the drug-resistant tumors as compared to the drug-sensitive case ( Figure 6C ) , a correlation between proliferation and vascularization/hypoxia is precluded ., Interestingly , the number of apoptotic cells is consistently higher for Eμ-myc p53-/- ( Figure 6D ) , suggesting non-hypoxia driven apoptosis for these tumors ., By analyzing each IHC section longitudinally along the tumor , a range of baseline values can be calculated from the experimental data for key model parameters ( Table S1 ) , inspired by recent methods in mathematical pathology 36: cell viability , necrosis , and spatial distribution pattern ( from H&E ) , cell proliferation ( from Ki-67 ) , cell apoptosis ( from Caspase-3 ) , oxygen diffusion distance ( from HIF-1α ) , and blood vessel density ( from CD31 ) ., These values are obtained for both Eμ-myc Arf-/- and Eμ-myc p53-/- tumors for each of the five sections obtained longitudinally along the tumor , with values sampled from the middle ( core ) and the edge ( periphery ) of each section ., The measured values are not resolved in space but averaged over each section , thus yielding information averaged over space ., The periphery was defined as the region approximately within 200 µm of the tumor boundary ., Figure S1 shows an example of this calibration process for proliferation at the periphery and middle from two histology sections in the center of the tumor ( Section S3 ) ., Taking an average proliferation cycle of 20 hours that we observed for the lymphoma cells in culture , the proliferation calculation in units of day−1 is λM*<n>\u200a=\u200a ( stained/ ( stained+unstained ) ) /20 hours/prolif . * 24 hours/day ., The average nutrient <n> indicates that this proliferation rate depends on the model diffusion of cell substrates such as glucose and oxygen in the 3D space ( Eq . 2 ) ., Similarly , since the apoptosis cycle was detectable up to 5 hours , the apoptosis calculation in units of day−1 is λA\u200a=\u200a ( stained/ ( stained+unstained ) ) /5 hours/apoptosis * 24 hours/day ., We calculate the average nutrient from the blood vessel density by assuming a uniform nutrient delivery rate from the blood to the tissue adjacent to the vessels ( Eq . 2 ) ., Estimating blood vessel area versus surrounding tissue provides a measure of the magnitude of cell substrates transferred into the tumor ., Thus , we calculate the fraction of cells supported per endothelial cell in a unit volume to be ( number unstained/ ( number stained+unstained ) ) 3/2 ., When the viable cell fraction in the simulations matches what is directly observed from microscopy , this implies that the vascular and nutrient distributions have been correctly represented in the model ( Figure 3A , middle ) ., Similarly , we calculate the hypoxic cell fraction per unit volume as ( number stained/ ( number stained+unstained ) ) 3/2 ., The node is represented by the computational model initially as a spherical capsule in 3D with a membrane boundary separating it from the surrounding tissue ( Figure, 4 ) ( see Text S1 ) ., Lymphoma cells are assumed to enter the lymph node through the afferent lymph vessels ., As they accumulate in the node during tumor progression in time , they compete for cell substrates such as oxygen and nutrients with the normal lymphocytes ., These substrates are assumed to diffuse radially outward toward the node periphery from the pre-existing vasculature , situated mainly in the core of the node ( see Figure 4A and Figure 3B , left , at the intersection of three large blood vessels ) ., Once a tumor has begun to form in the core of the node , this diffusion process presents a transport barrier for oxygen and nutrients to the lymphoma cells incoming through the afferent vessels into the node ., We investigate the effect of initially available oxygen and cell substrates needed for cell proliferation , since lymphoma growth is hypothesized to depend on access to these through the vasculature ., Preliminary calculations suggested that the initially available nutrient level has a significant effect on the growth phase of the tumor but not on its terminal size , which according to a theoretical analysis of the model ( Text S1 ) depends mainly on the ratio of apoptosis to proliferation 51 ., A further investigation revealed that the initial guess of parameter values results in a mismatch between the ratio of hypoxic cells and the average apoptosis rate: where the range of hypoxic ratio matches the experiments , the apoptosis rate range in the model is too low ., Accordingly , we calibrated the cell necrosis rate so that the key parameter values remain invariant when the initial nutrient is set to a threshold of 0 . 5 ., With this set of parameters , a necrosis rate from 5 to 7 ( non-dimensional units ) would satisfy the experimentally observed ranges of both the hypoxic fractions and the average apoptosis rate ( Figure S2A–B ) ., We then varied the initial nutrient threshold while maintaining the necrosis rate invariant to confirm that the fraction of hypoxic cells and average apoptosis rate would remain within the experimentally observed range of values ( Figure S2C–D ) ., This calibration suggests that the initially available nutrient still affects the growth phase of lymphoma ., In this model , the lymphoma tumor and the lymph node greatly outgrow the original lymph node size , which we consistently observed in vivo in addition to the distortion of the lymph node geometry ( we are currently implementing the Diffuse Domain Method 52 to better represent this geometry ) ., After using the IHC data to perform a cell-scale calibration of the lymphoma model , we verify the simulated tissue-scale lymphoma size from in vivo macroscopic observations and intravital imaging at the tissue scale ., Recently , it has been discovered with bioluminescence imaging by Gambhir and co-workers that lymphoma cells coming from the spleen and bone marrow seed the inguinal lymph node around Day 9 in vivo 53 ., Using this seeding as the initial condition for the simulations , the model predicts the tumor diameter to be ∼5 . 2±0 . 5 mm by Day 21 ( Figure 7 ) ., This figure also shows the gross tumor size from our caliper measurements in time , indicating that the model-predicted tumor diameter for the maximum possible value of initial nutrient falls within the range of the measurements in vivo ( the experiments show that there is no statistical difference in the tumor growth between the two cell types , Figure 3B , right ) ., The model simulations are based on an oxygen diffusion distance from the vessels estimated to be directly proportional to the distance at which hypoxia is detected away from blood vessels , measured experimentally from the HIF-1α staining to be 80±20 µm ., The variation in this measurement leads to the variation in the simulated diameter ., If the lymphoma is begun at sites within the lymph node other than the center ( Figure 4 ) , similar growth curves are computationally obtained as the whole node volume is eventually taken over by the proliferating tumor cells ( results not shown ) ., We note that since there is a distributed source of vessels in the tumor , the proliferation is relatively weakly sensitive to additional outside sources ., The tumor growth from the model calibrated from the cell-scale can be validated through theoretical analysis of the model based on previous mathematical and computational work 51 , 54–56 ( see Text S1 ) ., Assuming that the lymph node geometry is approximated by a 3D sphere , the model can be used to predict the tumor radius in time based on the ratio A of the rates of apoptosis to proliferation calculated from the experimental IHC data ., The average ratio A\u200a=\u200aλA/λM ∼0 . 4 for both drug-sensitive and drug-resistant cells ., In comparison with the simulations based on the cell-scale calibration , this analysis predicts that the tumor would reach a diameter of ∼6 mm ., Both the theoretical analysis and the tumor growth obtained through the simulations agree with the similar diameters observed experimentally in vivo ( ∼5 to 6 mm ) ( Figure 7 ) ., In the model , simulations of the vasculature were qualitatively compared to independent intravital microscopy observations in vivo of a Eμ-myc p53-/- tumor in the same animal over time ( Figure 8 ) ., The density of simulated viable tumor tissue ( Figure 3A , right ) as a function of the vascularization at day 21 qualitatively matches the density of the tissue observed experimentally ( fraction of simulated viable cells in the 2D plane , >90% per mm2 in inset in Figure 8 , vs . the average fraction of viable cells measured from H&E staining , 87%±6% per mm2 ) , indicating that the overall vasculature function was modeled properly ., The density of simulated endothelial tissue is also highest in the tumor core , as observed from histology ., The increase in the lymphoma cell population disturbs the homogeneous distribution of cell substrates ( such as oxygen and cell nutrients ) , leading to diffusion gradients of these substances that in turn affect the lymphoma cell viability ., If the cell viability is established heterogeneously within the tumor , e . g . , as observed experimentally in IHC with the Eμ-myc Arf-/- cells near the tumor periphery , the model predicts that the diffusion gradients would not be as pronounced ., If the cell viability is higher near the center of the tumor , which is observed in IHC with the Eμ-myc p53-/- cells ( Figure 5 ) , then the gradients are predicted to be steeper and more uniform 28 ., We integrate in vivo lymphoma data with computational modeling to develop a basic model of Non-Hodgkins lymphoma ., Through this work we seek a deeper quantitative understanding of the dynamics of lymphoma growth in the inguinal lymph node and associated physical transport barriers to effective treatment ., We obtain histology data by very fine sectioning across whole lymph node tumors , thus providing detailed three-dimensional lymphoma information ., We develop a computational model that is calibrated from these cell-scale data and show that the model can independently predict the tissue-scale tumor size observed in vivo without fitting to the data ., We further show that this approach can shed insight into the tumor progression within the node , particularly regarding the physical reasons why some tumors might be resistant to drug treatment – a critical consideration when attempting to quantify and predict the treatment response ., We envision that the modeling and functional relationships derived in this study could contribute with further development to patient-specific predictors of lymphoma growth and drug response ., Although the number of mice used for the experimental in vivo validation is limited , the model results are consistent with previous work ., For example , a well-studied mechanism of physiological resistance is the dependence of cancer cell sensitivity to many chemotherapeutic agents on the proliferative state of the cell 28 ., This physical mechanism is likely important in the difference in drug-sensitivity between the tumors formed from the two cell lines and will be explored in further studies ., We found that the Eμ-myc Arf-/- cells tend to congregate at the periphery of the tumor ( Figure 5 ) , even though there are vessels in the interior of the tumor ., This suggests the hypothesis that the more drug-sensitive Eμ-myc Arf-/- cells maintain better oxygenation at the expense of higher drug sensitivity by growing less compactly in the interior of the tumor – where there would be stronger competition for oxygen and cell nutrients – whereas the Eμ-myc p53-/- lymphoma cells may enhance their survival by closer packing in the core of the tumor ., Cell packing density may present a barrier to effective drug penetration 57 , which we have also modeled previously 28 ., Closer packing could further increase the number of cells that would be quiescent due to depletion of oxygen and nutrients , as we specify in the model ( Materials and Methods ) and as we have simulated in previous work 28 ., However , the proportion of chemoresistance inherent with Eμ-myc p53-/- that can be attributed to resistance at the genetic level compared to what can be attributed to suboptimal drug delivery and quiescence is unclear ., In follow-up work we plan to m | Introduction, Materials and Methods, Results, Discussion | Non-Hodgkins lymphoma is a disseminated , highly malignant cancer , with resistance to drug treatment based on molecular- and tissue-scale characteristics that are intricately linked ., A critical element of molecular resistance has been traced to the loss of functionality in proteins such as the tumor suppressor p53 ., We investigate the tissue-scale physiologic effects of this loss by integrating in vivo and immunohistological data with computational modeling to study the spatiotemporal physical dynamics of lymphoma growth ., We compare between drug-sensitive Eμ-myc Arf-/- and drug-resistant Eμ-myc p53-/- lymphoma cell tumors grown in live mice ., Initial values for the model parameters are obtained in part by extracting values from the cellular-scale from whole-tumor histological staining of the tumor-infiltrated inguinal lymph node in vivo ., We compare model-predicted tumor growth with that observed from intravital microscopy and macroscopic imaging in vivo , finding that the model is able to accurately predict lymphoma growth ., A critical physical mechanism underlying drug-resistant phenotypes may be that the Eμ-myc p53-/- cells seem to pack more closely within the tumor than the Eμ-myc Arf-/- cells , thus possibly exacerbating diffusion gradients of oxygen , leading to cell quiescence and hence resistance to cell-cycle specific drugs ., Tighter cell packing could also maintain steeper gradients of drug and lead to insufficient toxicity ., The transport phenomena within the lymphoma may thus contribute in nontrivial , complex ways to the difference in drug sensitivity between Eμ-myc Arf-/- and Eμ-myc p53-/- tumors , beyond what might be solely expected from loss of functionality at the molecular scale ., We conclude that computational modeling tightly integrated with experimental data gives insight into the dynamics of Non-Hodgkins lymphoma and provides a platform to generate confirmable predictions of tumor growth . | Non-Hodgkins lymphoma is a cancer that develops from white blood cells called lymphocytes in the immune system , whose role is to fight disease throughout the body ., This cancer can spread throughout the whole body and be very lethal – in the US , one third of patients will die from this disease within five years of diagnosis ., Chemotherapy is a usual treatment for lymphoma , but the cancer can become highly resistant to it ., One reason is that a critical gene called p53 can become mutated and help the cancer to survive ., In this work we investigate how cells with this mutation affect the cancer growth by performing experiments in mice and using a computer model ., By inputting the model parameters based on data from the experiments , we are able to accurately predict the growth of the tumor as compared to tumor measurements in living mice ., We conclude that computational modeling integrated with experimental data gives insight into the dynamics of Non-Hodgkins lymphoma , and provides a platform to generate confirmable predictions of tumor growth . | biotechnology, medicine, applied mathematics, cancers and neoplasms, hematologic cancers and related disorders, biomedical engineering, oncology, mathematics, lymphomas, bioengineering, non-hodgkin lymphoma, computer science, computer modeling, engineering | null |
journal.pcbi.1004389 | 2,015 | A Diffusive Homeostatic Signal Maintains Neural Heterogeneity and Responsiveness in Cortical Networks | Nitric oxide ( NO ) is a diffusive neurotransmitter which is widely synthesized in the central nervous system , from the retina to the hippocampus 1 , 2 ., Its properties as a small nonpolar gas molecule allows rapid and unconstrained diffusion across cell membranes , a phenomenon often called volume transmission 3 ., An important role of NO signaling is to regulate neural excitability through the modulation of potassium conductances in an activity-dependent manner , effectively mediating a form of homeostatic intrinsic plasticity ( HIP ) ., Experiments characterizing this effect also demonstrated that NO-synthesizing neurons can induce changes in the excitability of neurons located up to 100 μm away 4 , 5 ., These findings are corroborated by a recent study demonstrating neurovascular coupling mediated through activity-dependent NO diffusion 6 ., We build upon these observations , postulating a general form of HIP mediated by a diffusive neurotransmitter such as NO which we will refer to as diffusive homeostasis ., This contrasts with canonical models of HIP , here referred to as non-diffusive homeostasis , which assume that each neuron has access to only its own activity 7 ., Theoretical studies of HIP have generally focused on its role in maintaining stable network dynamics 8 , 9 ., It has also been recently demonstrated that HIP can improve the computational performance of recurrent networks by increasing the complexity of network dynamics 10 ., However , little is known about the effects of HIP on the heterogeneity typically observed in cortical networks; in particular , a growing body of evidence supports the finding that even neurons of the same type have a broad and heavy-tailed distribution of firing rates 11 ., Rather than an epiphenomenon of biological noise , neural heterogeneity has been proposed to improve stimulus encoding by broadening the range of population responses 12 , 13 ., However , this form of heterogeneity is difficult to reconcile with canonical models of HIP , which generally suppress cell-to-cell variability 14 ., While some degree of heterogeneity in populations of the same type of neuron may emerge naturally 15 , we found that such independent sources of variability will generally limit the responsiveness of a network through neuronal saturation ., Using network models and dynamic mean field analysis , here we show that networks with HIP mediated by diffusive neurotransmission exhibit a very different and unexpected behavior ., Firstly , we report that diffusive homeostasis provides a natural substrate for flexibly maintaining substantial heterogeneity across a network ., Secondly , the resulting population heterogeneity enables linear network responses over a wide range of inputs ., This not only improves population coding , but enables a good use of available resources by ensuring that all neurons remain functionally responsive to changes in network dynamics ., Finally , we demonstrate that these effects are preserved in networks whose recurrent synaptic inputs undergo Hebbian plasticity ., Fig 1C illustrates that both forms of homeostasis stabilized network activity following an increase in input ., There was however a crucial difference in how the neurons reacted to this change ., While for non-diffusive homeostasis each neuron simply returned to its target firing rate , diffusive homeostasis caused each neuron to sense a mixture of its own activity level and that of the rest of the network ., This can be seen in the spatial concentration profiles in Fig 1C ., It is important to note that the spatial position of each neuron was random and independent of its connections , meaning that there was no explicitly defined structure in the NO concentrations ., As a result , these networks exhibited a very different steady state behavior ., The firing rate distribution was narrow as expected for non-diffusive homeostasis , but broad and heavy-tailed for diffusive homeostasis ( Fig 1D ) ., The latter is consistent with recent experimental results indicating that firing rate distributions in cortex are generally heavy-tailed , approximating log-normal distributions 11 ., There were no noticeable differences in inter-spike interval statistics between networks with diffusive and non-diffusive homeostasis ( not illustrated ) ., We investigated the difference in firing rate distributions by modeling the relation between activity read-out and homeostatic compensation in these two cases using a dynamic mean-field model ( see Methods ) ., This approach considered an unconnected population of neurons with random inputs , where each of the two scenarios was simulated by using an appropriate activity read-out ., HIP was implemented as in the full spiking model , but the degree of diffusive signaling was now controlled by a single parameter , α ( Eq 11 in Methods ) , which determined the balance between local and global activity read-out ., If small , neurons used primarily their own activity to modulate their firing threshold , while increasing α caused the firing threshold to depend more strongly on the average population activity ., Setting , for instance , α = 0 . 8 led to a broad and heavy-tailed rate distribution similar to the full model , while α = 0 yielded a narrow distribution as in the non-diffusive case ( Fig 1E ) ., This model provides a simple and intuitive explanation for this effect ., For a non-interacting population , non-diffusive homeostasis can be thought of as precisely matching a neuron’s input μi and its threshold θi to maintain the target firing rate ., We can imitate this by introducing a covariance σ ( μ , θ ) between μi and θi , such that a high input rate implies a high firing threshold and a low input rate a low firing threshold ., Since setting α > 0 ( analogous to diffusive homeostasis ) introduces a correlation between a neuron’s threshold θi and the average population threshold θ ¯ , this effectively results in a decorrelation of μi and θi in comparison with setting α = 0 ( analogous to non-diffusive homeostasis ) ., In line with the previous results , populations with for instance σ ( μ , θ ) = 0 . 6 yielded a broader and more heavy-tailed distribution of firing rates than populations with σ ( μ , θ ) = 0 . 99 ( Fig 1F ) ., Since non-diffusive homeostasis directly relates the firing threshold of a neuron to its input , we observed a wider distribution of firing thresholds , which in turn ensured that all neurons assumed similar firing rates ., Diffusive homeostasis , on the other hand , yielded similar firing thresholds across the population ( Fig 3A and 3B ) ., When combined with the nonlinear input-output relation of neurons 17 , this gave rise to the broad firing rate distributions we observed ( see also Discussion ) ., This result was robust to changes in the rate of NO diffusion ., While decreasing the rate of diffusion , D , did result in slightly narrower firing rate distributions , they were broader than in networks with non-diffusive homeostasis across a wide range of values ( Fig 3C ) ., A similar trend was observed when varying the width of the external input rate distribution ., While decreasing this width led to a decrease in the width of the firing rate distribution , they were consistently broader in networks with diffusive homeostasis ( Fig 3D ) ., Since one may argue that diffusive homeostasis is merely adding variability to each neuron’s homeostatic signal due to the influence of neighboring neurons’ activity , we now ask whether it is possible to achieve broad firing rate distributions with non-diffusive homeostasis ., Indeed , by introducing variability of homeostatic targets ( see Methods ) , we could produce a distribution of firing rates similar to that observed with diffusive homeostasis ( Fig 1D , red histogram ) ., However , as we will show next , the effect of diffusive homeostasis is quite distinct from that of activity-independent , ‘quenched’ heterogeneity arising from randomly distributed homeostatic targets ., To investigate the functional consequences of heterogeneity caused by a diffusive homeostatic process , we next simulated specific changes in external input ., First , we stimulated small random groups of neurons at higher input rates of 5 Hz and 10 Hz ( versus a baseline of 2 . 5 Hz ) , as illustrated in Fig 4 ., Such inputs may , for instance , reflect developmental or other plastic changes that lead to a long-lasting change in network input ., In these simulations , the average network firing rate was reliably brought back to the original target firing rate by both forms of homeostasis ( Fig 4A–4C , black traces ) ., As above , in networks with non-diffusive homeostasis this was achieved by returning the rate of each neuron to the target firing rate regardless of their external input ( Fig 4A , colored traces ) ., In contrast , for networks with diffusive homeostasis , we found that the separability of firing rates of individual groups are maintained according to their input , while the firing rates of all groups were simultaneously reduced so that the average network firing rate again reached the target ( Fig 4B , colored traces ) ., Introducing variability in homeostatic targets for the non-diffusive case , as described previously , did not maintain separability of individual groups as in the diffusive case ., Instead , the different groups returned to their mean firing rates that existed before inputs were elevated ( Fig 4C ) ., The distribution of final firing thresholds explains these differences ( Fig 4D–4F ) ., For non-diffusive homeostasis , neurons in the group receiving 10 Hz input had the highest thresholds since they needed to reduce their firing rate the most , followed by the 5 Hz and 2 . 5 Hz groups respectively ., This led to the final threshold of each neuron reflecting its input ., Note that the distribution of firing thresholds is broader in this setup than in ( Fig 3A and 3B ) , as a broader range of inputs is given to the network ., For a diffusive signal , a neuron’s firing threshold is modulated by the activity of nearby neurons ., Since group membership of a neuron is independent of its position , this effect again introduced a correlation between each neuron’s threshold and the mean threshold of the entire network , resulting in a distribution of final thresholds which are less segregated according to their input compared to a network with non-diffusive homeostasis ., Thus , firing thresholds in neurons undergoing diffusive homeostasis were more weakly related to their external input ., This in turn preserves local firing rate differences in input groups while maintaining constant average network activity ., Introducing variable targets for non-diffusive homeostasis caused the thresholds to depend more strongly on their external input , similar to the original non-diffusive case ., We could broadly reproduce the distinctions between diffusive and non-diffusive homeostasis in the dynamic mean-field approach by varying α ., For α = 0 , modeling non-diffusive homeostasis , we obtained identical firing rates in input groups , as in the recurrent network ( Fig 4G ) ., Note that changing the input of groups of neurons in the recurrent network also affects the activity of neurons with fixed input ( Fig 4B , red traces ) due to recurrent connections , an effect that is obviously absent in the dynamic mean-field description ., Increasing α led to local firing rate differences persisting for longer periods of time ., However , these differences eventually decay very slowly , only remaining stable for the case where α = 1 ( Fig 4H and 4I ) ., The reason this occurs is , even after the population activity has quickly reached its homeostatic target , the deviations of the input groups still exert a small homeostatic force when α < 1 ., For example , if α = 0 . 95 , there will be a relatively fast change in thresholds as the population activity reaches its target , followed by much slower changes , at 1 − α = 0 . 05 times the speed ( Fig 4I ) ., This does not happen to the same extent in the spiking network simulations with diffusive homeostasis , as diffusion of NO ensure that deviations from the population activity are directly compensated for by neighboring neurons ., Differences persist for 3245 ± 440 s , compared with 115 ± 6 s and 140 ± 40 s for non-diffusive homeostasis with uniform and variable homeostatic targets , respectively ( ± symbol denotes standard error of the mean of 6 independent network realizations in each case , see Methods ) ., Since we have increased the speed of homeostasis in order to reduce simulation time ( see Methods ) , a more realistic time course of 15 minutes for NO modulation would cause input differences to persist in networks with diffusive homeostasis for many hours to days 5 ., Taken together , this shows that diffusive homeostasis can retain input heterogeneity due to the influence of neighboring neurons’ activity on an individual neuron’s firing threshold ., In the simulations shown so far , each neuron received a static input throughout since we were interested in the final network states ., We now investigate how these networks respond to fast changes in input; specifically how faithfully each neuron represents its change in input ., Since networks with diffusive homeostasis simultaneously maintain constant average network activity and firing rate heterogeneity , we expected that this should allow input modulations to be followed more precisely due to a greater representational capability ., After the network reached steady state under an initial distribution of external inputs , we froze homeostasis so as to simulate fast changes in activity , since we assume that homeostasis is not active over these time scales ., We then regenerated the external inputs to each neuron from the same distribution presented during homeostasis ., This can be thought of as a re-configuration of inputs due to external fluctuations ., To best represent such changes in a simple population coding paradigm , each neuron should respond linearly to a change in input; non-linear transformations may lead to an information loss and hence affect neural computations , although this may indeed be desirable in some brain regions ., We interpreted the range of changes in input over which this response is linear , or non-saturating , as the range over which homeostasis does not interfere with the network response ., Fig 5A–5C show the change in input rate versus change in output rate of each neuron ., A highly nonlinear response was observed in networks with non-diffusive homeostasis , with rectification for large decreases in input and superlinear responses for large increases in input ., This effect was quantified by an R2 value of 0 . 57 from a linear regression ., Conversely , networks with diffusive homeostasis exhibited a linear response across the entire range of input changes , with an R2 value of 0 . 85 ., Population heterogeneity can also be achieved , as discussed before , by introducing target variability during non-diffusive homeostasis ., This yielded a similar non-linear response as in the non-diffusive network with homogeneous targets , with an R2 value of 0 . 38 ., A consequence of the asymmetry in responses to input changes for networks with non-diffusive homeostasis was that the population rate increases upon regenerating inputs , despite the fact that mean input to the network remained unchanged ( Fig 5D ) ., This did not occur for networks with diffusive homeostasis , suggesting that these networks are more adept at maintaining a target level of activity in conditions where external inputs are dynamic and fast-changing ., Crucially , the benefits of a diffusive homeostatic signal can be achieved by a relatively broad range of values for the rate of diffusion , D , indicating that the effects we describe are robust to precise parameter choices ( Fig 5E ) ., Increasing the rate of NO decay , λ ( see Methods ) , has a similar effect to decreasing D ( S1 Fig ) ., This difference in responses to input changes could again be reproduced in the dynamic mean-field approach ., This allowed us to characterize population responses across different effective diffusive ranges , using the R2 value from a linear regression as a measure of response linearity ., Fig 5F shows R2 values across a range of different input distribution widths , δ , as α is varied to model different diffusion coefficients ( see Methods ) ., This revealed a dependency on δ: While values of α ∼ 1 exhibited the best response for smaller δ , hence cases where the inputs are rather narrow , the optimal α decreased as δ increased , as well as the overall response linearity ., This dependence on input width can be explained by considering the manner in which a population of neurons with a distribution of dynamic ranges span a range of inputs ., If this range of inputs is small , then all neurons will span it regardless of their dynamic range ( determined by their firing threshold ) , hence the high values of R2 for δ = 0 . 1 ., For an intermediate range of inputs , neurons whose dynamic range is best adapted to the average input are most responsive ., This is achieved by increasing α ., If the range of inputs is very large ( δ = 1 . 0 ) , R2 values are low since the dynamic ranges of the population cannot span the inputs ., This effect is stronger at high α , as firing thresholds are more correlated , and the dynamics range of most neurons cannot capture the full input variance ., Since connection probability falls off with spatial distance in cortical networks 18 , we additionally simulated recurrent networks featuring such connectivity profiles ., These networks exhibited qualitatively similar behavior under diffusive and non-diffusive homeostasis compared to networks without any spatial dependence in connectivity ( S2 Fig ) ., Up until this point we have presented diffusive and non-diffusive homeostatic mechanisms as dichotomies , which has enabled a clear investigation of their distinct effects on network properties ., However , it is more biologically relevant to investigate networks in which both mechanisms are simultaneously active ., Fig 6 shows that the increased neural heterogeneity and response linearity observed in networks with diffusive homeostasis are also present in networks with both diffusive and non-diffusive homeostasis , and that the degrees of neural heterogeneity ( Fig 6A ) and response linearity ( Fig 6B ) are determined by the relative timescales of these mechanisms ( see Methods ) ., As the ratio of timescales of non-diffusive homeostasis to diffusive homeostasis is increased ( i . e . as non-diffusive homeostasis becomes slower than diffusive homeostasis ) , the network goes from narrow steady state firing rate distributions to broad , and exhibits an increase in response linearity , thus becoming more similar to networks with only diffusive homeostasis ., Overall , these results suggest that networks undergoing diffusive homeostasis are better suited to linearly represent a range of inputs ., We investigated this by presenting the networks with time-varying inputs after freezing homeostasis ., Groups of excitatory neurons received additional inputs which were randomly and independently generated after fixed time intervals ( see Methods ) ., Fig 7A shows the representation of such a time-varying input pattern ( dotted black line ) for each network ( colored lines ) ., Networks which have undergone diffusive homeostasis were capable of tracking this input significantly better than their non-diffusive counterparts , as characterized by the RMS error between the network response and input pattern ( 0 . 12 for diffusive homeostasis; 0 . 23 and 0 . 19 for non-diffusive homeostasis with uniform and variable targets , respectively; Fig 7B ) ., We can explore these differences further by constructing a simplified task in which a population of orientation-selective neurons respond to the orientation of a stimulus ( see Fig 7C–7F , Methods ) ., This is not intended to represent circuits which perform this task in the brain , but to serve purely as a demonstration of the relative merits of linear and non-linear network responses ., Neurons in the network are randomly assigned a preferred stimulus orientation , independent of their spatial position ., A stimulus of a certain orientation can then be presented to the network by varying the external input rates of each neuron , with neurons whose preferred orientation is closest to the stimulus orientation receiving the highest input rate ., The stimulus orientation can be decoded from the network by taking the vector average of the stimulus response across all neurons ., The orientation of this vector average , or population vector , is the decoded stimulus orientation ., Networks with linear responses perform better than those with non-linear responses in decoding stimulus orientation , as measured by the standard deviation of errors in the orientation of the population vector compared to the stimulus orientation ( 41° , 63° and 72° for diffusive homeostasis , non-diffusive , and non-diffusive with variable targets respectively , Fig 7G ) ., In the networks described so far , we have used static and uniform synaptic weights for recurrent connections ., We next considered whether the observed properties of diffusive homeostasis are altered by the presence of plastic synaptic weights , in particular when Hebbian spike-timing-dependent plasticity ( STDP ) is introduced ( see Methods ) ., Using a standard model of STDP with additive depression and potentiation for all recurrent excitatory synapses , we simulated networks with both STDP and homeostasis active until synaptic weight and firing rate distributions reached a steady state 19 ., As before , firing rate distributions were broader in networks with diffusive homeostasis ( Fig 8Bi ) ., Broad distributions could also be achieved by introducing variability in homeostatic targets ., Spiking activity remained asynchronous after STDP , as shown by the distribution of inter-spike intervals and the spike autocorrelograms , although STDP caused weakly synchronous activity in networks without any form of homeostasis ( Fig 8Bii–8Biv ) 20 ., The additive STDP rule led to a bimodal distribution of synaptic weights ( Fig 8A ) , as previously reported 19 ., STDP amplified the differences in response linearity that were observed between homeostatic cases ., Inputs to each neuron were regenerated from the same distribution presented during plasticity , and the corresponding change in output rate was compared to the change in input rate , as in Fig 5A–5C ., While the response linearity , given by the mean R2 value , for networks with diffusive homeostasis was 0 . 16 , networks with non-diffusive homeostasis exhibited much lower mean values of 0 . 01 and 0 . 02 , for uniform and variable homeostatic targets respectively ( Fig 8C ) ., Networks without any homeostasis had a mean value of 0 . 1 ., These R2 values were lower than those from networks without any STDP ( Fig 5E ) , which was likely due to a combination of weaker external input given to these networks ( gext of 40 ns compared to 80 ns , see Methods ) , and stronger recurrent excitation received by neurons in these networks due to potentiation during STDP ., This was tested by assessing response linearity in networks with static weight matrices obtained by shuffling the steady-state weight matrix of a network which had undergone STDP without any homeostasis ( see Methods ) ., These networks were run for each different homeostatic case until a steady state was reached , and inputs to each neuron were regenerated in order to measure response linearity ., While shuffling synaptic weights did increase R2 values across all networks ( Fig 8C , crosshatched bars ) , indicating that STDP plays a role in decreasing response linearity , they remained lower than in networks from Fig 5E , confirming that the reduced influence of external input compared with recurrent input is largely responsible for this difference ., We observed qualitatively similar retention of broad firing rate distributions and response linearity with diffusive homeostasis when a weight-dependent update rule was used ( not illustrated ) , which has been argued to lead to more realistic weight distributions 21 ., Neural homeostasis is commonly thought of as a local process , where neurons individually sense their activity levels and respond with a compensatory change if activity changes ., Here we investigated a complementary mechanism , where homeostasis is mediated by a diffusive molecule such as NO that acts as a non-local signal ., Using a generic recurrent network model , we show that this form of homeostasis can have unexpected consequences ., First , we found that it enables and maintains substantial population heterogeneity in firing rates , similar to that observed experimentally in intact circuits 11 , and that input heterogeneities can be preserved in the population activity ., Second , the specific form of neural heterogeneity brought about by diffusive homeostasis is particularly suited to support linear network responses over a broad range of inputs ., It is important to note that this behavior differs from that of networks where heterogeneity is simply introduced by randomly assigning a different target to each neuron ., These results predict that disrupting neural diffusive NO signaling can affect perceptual and cognitive abilities through changes of neural population responses ., While other non-diffusive homeostatic mechanisms would continue to stabilize neural activity , the lack of a signal related to the average population activity may disrupt the flexible maintenance of firing rate heterogeneity , and as a result the ability to represent network inputs ., Mean-field analysis revealed that these differences are essentially due to the diffusive messenger providing each neuron with a combination of the average network activity and its own activity as the homeostatic signal ., Diffusion of the signal from highly active neurons causes a reduction in the activity of their neighbors , such that firing rates of highly active neurons do not have to be completely reduced in order for the population to achieve a target rate ., As a consequence , diffusive homeostasis furnishes a network with an efficient way of flexibly maintaining heterogeneity of firing rates ., These effects can also be understood by considering the neural transfer functions , as illustrated in Fig 9A and 9B , which provides an intuitive explanation for the differences in firing rate distributions observed under diffusive homeostasis 17 , 22 ., For non-diffusive homeostasis the transfer function of each neuron is brought to center around its input , leading to a narrowing of the firing rate distribution ., Diffusive homeostasis decorrelates the input and threshold of individual neurons , resulting in a population of neurons residing along the entire transfer function ., This preserves the non-linear shape of the transfer function , causing broad and heavy-tailed firing rate distributions ., Narrow firing rate distributions are an obvious consequence of local homeostatic processes , as for instance shown recently with homeostasis implemented as local synaptic metaplasticity 14 ., This is in apparent conflict with the growing body of experiments documenting broad and heavy-tailed distributions of firing rates in cortex 11 ., One could argue that a straightforward explanation is a process , for example genetic or developmental , which randomly assigns neurons heterogeneous homeostatic targets ., While we show here that this can result in broader firing rate distributions , we also found that this generally leads to networks with a mismatch between the neural dynamic ranges and input statistics , which in turn limits the responsiveness of the network ., A striking feature of diffusive homeostasis is the lack of requirement for any such distribution of homeostatic targets , as the diffusive signal can be effectively exploited through providing a context for heterogeneity—neurons which maintain a significantly higher firing rate than the rest of the network also synthesize a higher level of the diffusive signal , thus ensuring that their deviation from the average firing rate is counterbalanced by lowering neighboring neurons’ firing rates ., This mechanism essentially allows neurons to differ in activity from the population as long as the population as a whole provides some compensation for these deviations ., Moreover , this mechanism is compatible with the recent finding that a minority of cells were found to consistently be the most highly active and informative across brain states 23 ., While non-diffusive homeostasis would have a disruptive effect on such a ‘preserved minority’ of neurons by reducing their activity towards those of the less active majority , diffusive homeostasis provides a substrate for maintaining their differentiated activity ., A significant distinction between the effects of diffusive and non-diffusive homeostasis appears when network responses to rapidly changing input are considered ( Fig 5 ) ., We show that networks with diffusive homeostasis represent input changes more faithfully than those with non-diffusive homeostasis ., Saturation of neurons’ responses to large changes are observed in networks without diffusion—this effect is further illustrated in Fig 9C and 9D ., Across a spatially homogeneous network , diffusing signals act to effectively shift the transfer function of each neuron towards the average network input , ensuring that neurons are responsive across the entire range of inputs presented to a network ., This is in contrast to networks with non-diffusive homeostasis , in which individual neurons are only responsive in a range around their current input ., Moreover , the asymmetric response of networks with non-diffusive homeostasis causes the average network activity to increase after fast input changes , while it is constant for a network with diffusive homeostasis ( Fig 5D ) ., The latter case is consistent with observations that mean population firing rates are preserved across novel and familiar environments and across different episodes of slow-wave sleep 24 , 25 Networks with diffusive homeostasis have an improved ability to accurately track time varying inputs ( Fig 7A and 7B ) as a direct consequence of their linear responses ., Beneficial effects of neural heterogeneity for population coding have been suggested before 13 , 26 , but here we find that the broad linear response regime maintained by diffusive homeostasis further improves network performance ., This improvement in network performance is also observed in a simplified stimulus orientation decoding task ( Fig 7C–7G ) ., Networks with diffusive homeostasis perform better than those with non-diffusive homeostasis when a population vector is constructed from the neural responses in order to decode stimulus orientation ( Fig 7G ) ., Although there exist alternative methods for decoding stimuli , the population vector has been shown to exhibit performance close to the optimal maximum likelihood procedure for broad tuning , as was used in our example 27 ., These distinctions between diffusive and non-diffusive homeostasis are conserved in networks with STDP ( Fig 8 ) ., This demonstrates that the limitations of non-diffusive homeostasis in maintaining neural heterogeneity and responsiveness extend beyond the case of static inputs , towards more realistic situations in which neurons receive ongoing and diverse perturbations ., Indeed , networks with non-diffusive homeostasis lost almost all sensitivity to external inputs after STDP , while networks with diffusive homeostasis retained this sensitivity ( Fig 8C ) ., The consequences of diffusive and non-diffusive homeostasis coexisting were also explored , by implementing these mechanisms simultaneously in a single network ( Fig 6 ) ., Stable activity could be robustly maintained , with the resulting network behavior depending on the relative timescales of the non-diffusive and diffusive mechanisms ., | Introduction, Results, Discussion, Methods | Gaseous neurotransmitters such as nitric oxide ( NO ) provide a unique and often overlooked mechanism for neurons to communicate through diffusion within a network , independent of synaptic connectivity ., NO provides homeostatic control of intrinsic excitability ., Here we conduct a theoretical investigation of the distinguishing roles of NO-mediated diffusive homeostasis in comparison with canonical non-diffusive homeostasis in cortical networks ., We find that both forms of homeostasis provide a robust mechanism for maintaining stable activity following perturbations ., However , the resulting networks differ , with diffusive homeostasis maintaining substantial heterogeneity in activity levels of individual neurons , a feature disrupted in networks with non-diffusive homeostasis ., This results in networks capable of representing input heterogeneity , and linearly responding over a broader range of inputs than those undergoing non-diffusive homeostasis ., We further show that these properties are preserved when homeostatic and Hebbian plasticity are combined ., These results suggest a mechanism for dynamically maintaining neural heterogeneity , and expose computational advantages of non-local homeostatic processes . | Neural firing rates must be maintained within a stable range in the face of ongoing fluctuations in synaptic connectivity ., Existing cortical network models achieve this through various homeostatic mechanisms which constrain the excitability of individual neurons according to their recent activity ., Here , we propose a new mechanism , diffusive homeostasis , in which neural excitability is modulated by nitric oxide , a gas which can flow freely across cell membranes ., Information about a neurons’ firing rate can be carried by nitric oxide , meaning that an individual neurons’ excitability is affected by neighboring neurons’ firing rates as well as its own ., We find that this allows a neuron to deviate from the target population activity , as its neighbors will counteract this deviation , thus maintaining stable average activity ., This form of neural heterogeneity is more flexible than assigning different target firing rates to individual neurons ., Consequently , networks endowed with this diffusive mechanism have an improved representational capability compared to canonical , local homeostatic mechanisms , and allow for more efficient use of neural resources . | null | null |
journal.pntd.0005434 | 2,017 | Human Treponema pallidum 11q/j isolate belongs to subsp. endemicum but contains two loci with a sequence in TP0548 and TP0488 similar to subsp. pertenue and subsp. pallidum, respectively | Treponema pallidum subsp ., endemicum ( TEN ) is the causative agent of bejel ( endemic syphilis ) , a chronic human infection usually affecting children under 15 years of age ., The primary stage of endemic syphilis is often localized to the mucosa of the oral cavity or nasopharynx and frequently remains undetected ., Secondary lesions often mimic syphilitic lesions and are found on both mucosal and skin surfaces including the oral cavity , pharynx , and larynx ( for review see 1 ) ., The tertiary stage is characterized by gummatous or destructive lesions of mucosa , skin , and bones ., Recently reported cases of bejel have come from African countries with dry climates including Mauretania , Niger , Chad , Mozambique and from countries in the Middle East including Turkey , Saudi Arabia , and Iran 1 ., Moreover , several imported cases of bejel have been described in France 2 and Canada 3 in children coming from countries where endemic syphilis has been reported ., Compared to the syphilis-causing Treponema pallidum subsp ., pallidum ( TPA ) and the yaws-causing Treponema pallidum subsp ., pertenue ( TPE ) ( reviewed in 4 , 5 ) , TEN is the least well characterized and least studied human pathogenic treponeme ., There are few genetic studies on TEN strains 3 , 6–14 , which is likely due to a limited number of available TEN samples ., In fact , most studies on TEN strains described one of the two reference strains , i . e . , Bosnia A or Iraq B . The Bosnia A strain was isolated in 1950 in southern Europe ( Bosnia ) from a 35-year old male with several mucosal and skin lesions 15 , and the Iraq B strain was isolated in 1951 in Iraq from a 7-year old girl who had oral mucous lesions and an anal condylomata 15 ., Because of the low number of available reference strains , only a single complete genome sequence of TEN Bosnia A has been published to date showing a close relatedness ( higher than 99 . 9% ) to TPE strains and several sequences surprisingly similar to TPA strains 16 ., First reported in 2013 , an unusual 11q/j subtype ( defined by enhanced CDC typing ) 17 , 18 was found among samples taken from a syphilis patient in Paris 19 who had returned from Islamabad , Pakistan , where he admitted having had sex with commercial sex worker ., Based on a partial sequence type reported for TP0548 , Mikalová et al . 20 pointed out that this sequence was more related to the yaws-causing strains rather than to syphilis-causing strains ., Further analyses resulted in the classification of the 11q/j isolate as a TEN treponeme 21 ., In this study , we characterized the 11q/j isolate in a set of 44 independent chromosomal regions ., Sequencing of these loci revealed that the 11q/j isolate belongs to the T . pallidum subsp ., endemicum with two loci having sequences that were related to either TPE ( TP0548 ) or TPA ( TP0488 ) ., The relevance of these findings is discussed here ., The study was approved by the institutional review board of the Comité de Protection des Personnes d’Ile de France 3 ( S . C . 3005 ) ., The sample ( a swab from an indurated genital ulceration ) was collected from a 42-year-old heterosexual man who attended the outpatient STD clinic of Hôpital Saint-Louis ( Paris ) and was analyzed anonymously ., Isolated DNA ( 20 μl ) from this sample ( referred as 11q/j ) was obtained from the National Reference Center for Syphilis in France ( CNR Syphilis , www . cnr-syphilis . fr ) that had performed a routine analysis on DNA from clinical samples 22 ., A nested PCR protocol for detection of the polA gene was performed using the previously described outer primers , polA_outer_F1 ( 5´-TTCTGTGCTCACGTCTGGTC-3´ ) and polA_outer_R1 ( 5´-TGCAACCATCGTATCGAAAA-3´ ) , which resulted in a 637 bp amplicon 23–25 and inner primers for nested polA PCR , polA_F1 ( 5´-TGCGCGTGTGCGAATGGTGTGGTC-3´ ) and polA_R1 ( 5´-CACAGTGCTCAAAAACGCCTGCACG-3´ ) , resulting in a 377 bp amplicon , were used as described in Liu et al . 26 ., This nested PCR protocol was shown to be able to detect 1–10 copies of treponemal DNA in a 1 μl of sample 23–25 and was used for detection of the number of treponemal genome equivalents in 1 μl of DNA ., The original 11q/j DNA sample ( 3 μl ) was randomly amplified using a REPLI-g Single Cell kit ( Qiagen , Hilden , Germany ) according to the manufacturers instructions ., Randomly amplified sample of 11q/j was then used for, ( i ) direct nested PCR amplification with specific primers ( listed in S1 Table ) according to a previously published protocol 25 , 27 ,, ( ii ) whole DNA sequencing using an Illumina MiSeq Next-gen sequencer ( Illumina , San Diego , CA , USA ) , and, ( iii ) the subsequent amplification with T . pallidum specific primers used in the pooled segment genome sequencing ( PSGS ) method 16 , 28–30 , which was followed by Illumina sequencing ., Regions successfully amplified from the 11q/j isolate were also amplified from another available DNA reference sample , i . e . , TEN Iraq B . The TEN Iraq B DNA was provided by Dr . Kristin N . Harper from the Department of Population Biology , Ecology , and Evolution , Emory University , Atlanta , Georgia , USA , in 2005 ., The Iraq B DNA was amplified with PCR or the nested PCR protocol with the same specific primers used for nested PCR amplification of the 11q/j isolate ( S1 Table ) ., The obtained partial sequences from the 11q/j isolate and TEN Iraq B were either assembled from Sanger and/or Illumina sequencing reads using SeqMan or SegMan NGen software ( DNASTAR , Madison , WI , USA ) , respectively , with default assembling parameters ., Genes were annotated according to the whole genome sequence of TEN Bosnia A ( CP007548 . 1; 16 ) and the 11q/j isolate and the TEN Iraq B genes were tagged with TEND11qj_ and TENDIB_ prefixes , respectively ., The resulting sequences of the 11q/j isolate and TEN Iraq B were analyzed and compared to the following genomes: TPA Nichols ( CP004010 . 2; 31 ) , TPA SS14 ( CP004011 . 1; 31 ) , TPE Samoa D ( CP002374 . 1; 29 ) , TPE CDC-2 ( CP002375 . 1; 29 ) , TPE Gauthier ( CP002376 . 1; 29 ) , TPE Fribourg-Blanc ( CP003902 . 1; 30 ) , and TEN Bosnia A 16 ., Alignments of treponemal sequences were performed using SeqMan software and MEGA7 software 32 ., Phylogenetic trees were constructed in MEGA7 software 32 using the Maximum Likelihood method based on the Tamura-Nei model 33 ., The following formula was used to calculate the probability that the observed nucleotide sequences were caused by accumulation of individual mutations instead of a recombination: pmut = ( pmut_gen x pmut_nuc ) n , where pmut = the end probability of mutations resembling recombinant events , pmut_gen = the frequency of mutation per single nucleotide , pmut_nuc = the probability of a nucleotide substitution into the nucleotide sequence in the putative recombinant region , n = the number of mutated nucleotides within the putative recombinant region ., pmut_gen was calculated based on the number of variable sites identified within all available sequences of the 11q/j sample ( total length of 29 , 753 bp , except for loci TP0488 and TP0548 ) and the corresponding sequences of TEN Bosnia A and TEN Iraq B . pmut_nuc had a constant value of 0 . 333 reflecting 3 possible substitutions changing the original sequence at each nucleotide site ., Different probabilities of transitions and transversions were not considered in this analysis ., The “n” was calculated based on the number of different nucleotide positions between the 11q/j isolate and one of the two TEN strains that matched either the TPA or TPE orthologous sequence ., The resulting sequences of the TEN 11q/j isolate and the TEN Iraq B with length ≥ 200 bp were deposited in the GenBank under following accession numbers: KY120774-KY120814 for TEN 11q/j isolate; KY120815-KY120855 for TEN Iraq B . The detailed overview of sequenced loci is shown in S2 Table ., The only available DNA-containing sample ( 20 μl ) was obtained from the CNR Syphilis that performed the isolation of DNA from the original swab sample 22 ., As revealed by the nested polA PCR reaction 23 with detection limit of less than 10 molecules 26 , the sample contained undetectable amounts of treponemal DNA , i . e . less than 10 molecules of treponemal DNA per 1 μl ., Following whole genome amplification with random primers , nested PCR protocol revealed positivity in a 10−2 dilution indicating , at least , 1x102 copies of treponemal genome equivalents per 1 μl in a total of 50 μl of amplified sample ., This randomly amplified sample was used for further analyses ., The randomly amplified sample was used for direct Illumina sequencing and resulted in 1 , 786 , 712 individual reads ., Of those , only 10 reads were mapped to the TEN Bosnia A genome indicating that the ratio of treponemal DNA to DNA from other species ( mostly human ) is less than 1:105 ., Subsequently , the randomly amplified sample was used for specific amplification with the PSGS technique 16 , 28–30 and primer pairs from Pool 1 amplifying the first quarter of the treponemal genome ( Fig 1 ) ., Specific amplification resulted in a total of 353 , 006 individual reads , of which 41 , 308 reads were mapped to the TEN Bosnia A genome ., Consensus sequences from at least 2 individual reads represented sequenced DNA regions of the 11q/j isolate ., All regions determined by Illumina sequencing are shown in Fig 1 and S2 Table ., Altogether , 15 genomic loci were obtained for the 11q/j isolate with lengths ranging from 63–2 , 455 bp , with total length of 9 , 626 bp and with coverage ranging from 2–15 , 832x ., In addition to Illumina sequencing , a nested PCR of 31 chromosomal loci was performed from the randomly amplified 11q/j sample using 1 μl of the starting DNA template ., The resulting amplicons were Sanger sequenced ., Loci for nested PCR were selected based on whole genome comparisons of published TPE strains ( Samoa D , CDC-2 , Gauthier ) and TEN Bosnia A . Preferentially , loci with accumulated single nucleotide variants ( SNVs ) and/or indels between TPE and TEN strains were selected as well as conservative genes suitable for unambiguous distinction between TPE and TEN subspecies ., All regions amplified using nested PCR and sequenced using the Sanger method are presented in Fig 1 and S2 Table ., The 16S and 23S rRNA loci were amplified from both genome positions 12 ., The length of resulting sequences of the 11q/j isolate ranged from 352–2302 bp and represented a total of 23 , 979 bp ., Illumina and Sanger sequencing of the 11q/j isolate resulted in sequences obtained from 44 chromosomal DNA regions covering , altogether , 32 , 635 bp ( 2 . 87% ) of the TEN Bosnia A genome length ( S2 Table ) ., Two genomic regions within TP0121 and TP0136 genes , where both sequencing techniques partially overlapped , revealed identical sequences ., The average length of sequenced regions in the 11q/j isolate was 742 bp ( range 63–2 , 455 bp ) ., The sequenced chromosomal regions were dispersed throughout the entire chromosome with distances ranging from 0 . 1–124 . 7 kb ( Fig 1 ) ., All sequenced genomic regions of the 11q/j isolate were also amplified and Sanger sequenced from the TEN Iraq B DNA and these regions are described in S2 Table ., Interestingly , sequencing of a short gene fragment ( 548 bp ) of TENDBA_0488 between positions 684–1231 revealed that the sequence of the 11q/j isolate was very similar to the sequence in TPA Nichols , but not to TEN strains ( Fig 2A ) ; suggesting the occurrence of a recombination event at this locus ., The minimal size of recombinant DNA sequence was 505 nucleotides ( between coordinates 715–1219; Fig 2A ) ., A set of 21 nucleotide positions of the 11q/j isolate were different from TEN Bosnia A as well at TEN Iraq B , but identical to TPA strains ., A partial sequence of the TEND11qj_0488 from the 11q/j isolate , representing the recombinant part ( 505 bp long fragment ) , was used for construction of a tree ( Fig 3A ) that revealed clustering of the 11q/j isolate within TPA strains , not within TEN strains ., The probability that the observed nucleotide sequence within this locus was caused by an accumulation of individual mutations instead of a recombination was tested using the following formula: pmut = ( pmut_gen x pmut_nuc ) n ( see Materials and Methods ) ., pmut_gen was calculated based on the number of variable sites identified within all available sequences of the 11q/j sample ( 22 variable positions in a total length of 29 , 753 bp from the 3 analyzed TEN genomes; 0 . 00074 nt differences per 1 bp ) ., Loci TP0488 and TP0548 were not included in this calculation ., The “n” was calculated based on the number of variable positions detected in the sequence alignment presented in Fig 2 ., In the TP0488 gene , there was a total of 23 nucleotide positions in the 11q/j sequence that differed from TEN strain Bosnia A but were identical to TPA strain SS14 ., With the assumption that the 11q/j isolate represents a TEN strain , the probability that the accumulated SNVs within the TP0488 of the 11q/j isolate were due to accumulation of individual mutations would be: pmut = ( 0 , 00074 x 0 . 333 ) 23 , i . e . pmut = 1 . 01987 x 10−83 ., To rule out potential co-infection with TPA and TEN in this patient , Illumina sequencing reads of the 11q/j sample , especially in regions with positions that differ between TPA and TEN , were evaluated and revealed 20 informative sites with coverage ≥ 4x ( range 4x–94x ) ., However , there was no heterogeneity in these positions , excluding co-infection with multiple strains ., As shown previously , the 11q/j isolate within its 86 bp-long fragment of TP0548 gene revealed a new sequence type that is , in fact , related to TPE strains 20 ., Analysis of a larger 2 , 302 bp-long region comprising TP0547 , TP0547a , and TP0548 genes ( positions 589 , 926–592 , 227 corresponding to the whole genome sequence of the TEN Bosnia A ) revealed that the sequence of the 11q/j isolate , within the TP0548 gene , was very similar to TPE strains , especially to a sequence from TPE Samoa D ( Figs 2B and 3B ) ., The minimal size of recombinant DNA sequence was 613 nucleotides ( between coordinate 69 of the TENDBA_0547a and coordinate 623 of the TENDBA_0548 ) and comprised 56 variable positions ( Fig 2B ) ., Thirty-seven of the nucleotide positions of the 11q/j isolate were different from TEN Bosnia A and TEN Iraq B , but identical to at least one of the TPE Samoa D or Gauthier strains ( Fig 2B ) ., Both TEN Bosnia A and Iraq B showed 23 nucleotide positions identical to at least one of the TPA strains , i . e . , to Nichols or SS14 , but different from the 11q/j isolate ., A partial sequence of TEND11qj_0548 ( 613 bp ) was used for construction of a tree ( Fig 3B ) and reveled clustering of the 11q/j isolate among TPE strains , but not among TEN strains ., The probability that the observed nucleotide sequence within this locus was caused by an accumulation of individual mutations instead of a recombination was pmut = 1 . 12245 x 10−65 ( pmut = ( 0 . 00074 x 0 . 333 ) 18 ) , since there was a total of 18 SNVs that differed from TEN strain Bosnia A but were identical to TPE strain Gauthier ., Indels were omitted from the calculation ., The sequences of 42 chromosomal regions , excluding TP0488 and TP0548 sequences , were concatenated and used to construct a phylogenetic tree to visualize the relatedness of 11q/j isolate to other treponemal genomes ( Fig 3C ) ., The corresponding genome regions from the published whole genome sequences of two TPA strains ( Nichols , SS14 ) , four TPE strains ( CDC-2 , Gauthier , Samoa D , and Fribourg-Blanc ) , and TEN Bosnia A were used ., Moreover , the dataset was supplemented with sequences of TEN Iraq B . All positions in the alignment containing gaps and missing data were eliminated resulting in a total of 29 , 447 positions in the final dataset having 509 variable sites ., Overall , the 11q/j isolate clustered with both TEN Bosnia A and TEN Iraq B , indicating that most chromosomal loci of the11q/j isolate were consistent with TEN classification ., In this work , we analyzed an interesting human clinical isolate , 11q/j , that was first reported in 2013 as a case of syphilis 19 , but due to an unusual sequence pattern at the TP0548 locus , similar to TPE , it was thought to be an imported case of yaws 20 ., In 2016 , the 11q/j isolate was further characterized in 7 genomic loci and classified as subspecies TEN 23 ., Due to the unusual syphilis-yaws-bejel history of the 11q/j isolate we characterized larger genome regions of this clinical sample using different sequencing approaches ., The small amount of treponemal DNA within the only available sample of the 11q/j isolate ( copy number less than 10 molecules of treponemal DNA per 1 μl ) with an excessive amount of contaminating human DNA , which exceeded the treponemal DNA by at least 100 , 000 times , precluded the use of other techniques that have been recently reported to be effective in sequencing treponemal DNA directly from clinical samples 34 , 35 ., Efficient enrichment of treponemal DNA requires the number of treponemal copies > 1x104 per 1 μl 34 ., In fact , enrichment of the TEN Iraq B DNA sample , containing 104 copies per 1 μl , revealed genome coverage less than 12 . 4% 35 ., For these reasons , we mostly used nested PCR in this study ., In all cases , amplification was done from samples containing at least 102 copies of treponemal DNA to avoid introduction of sequencing errors ., As reported in a previous study based on analysis of 7 chromosomal regions , classification of the clinical isolate 11q/j was consistent with T . pallidum subsp ., endemicum 21 ., In this work , we confirmed this finding based on analyses of 42 chromosomal regions ( excluding the TP0488 and TP0548 loci ) , which were independently amplified and analyzed ., The corresponding phylogenetic tree revealed a clear clustering of the 11q/j isolate with TEN strains ( Fig 3C ) ., In addition , the genetic distance between the 11q/j isolate and TEN Bosnia A and Iraq B was greater than the distance between Bosnia A and Iraq B , indicating that the ancestor of the 11q/j isolate diverged before TEN Bosnia A and TEN Iraq B diversified ., Although the sequenced portion of the 11q/j isolate represented less than 3% of the total genome length , the number of analyzed nucleotide positions informative for differentiation between TPE and TEN was much larger ., Considering the extent of similarity of the genome sequences of available TPE and TEN strains ( i . e . , they are 99 . 91–99 . 94% similar ) , there were relatively few ( 711–970 ) variable sites between TEN Bosnia A and TPE strains Gauthier , CDC-2 and Samoa D 16 ., Within the 11q/j isolate , 196 ( 20–28% ) of these variable sites were sequenced and 98% of them revealed sequence similarity to TEN strains ., Therefore , it is very likely that the classification of 11qj isolate as TEN strain will remain the same even after acquisition of additional genomic sequences ., Sequence analysis of TP0488 of the 11q/j sample revealed a sequence very similar to TPA strains ., A similar situation has been previously reported in the genome of Bosnia A , where several chromosomal regions including TP0326 , TP0488 , TP0577 , TP0858 , TP0968 , and TP1031 showed striking similarity to TPA treponemes 16 ., However , the TPA-like sequences at the TP0488 locus of Bosnia A and Iraq B strains were different from the TP0488 sequence of the 11q/j sample and were located between positions 1175–1195 ( Fig 2A ) , indicating that the 11q/j recombination event was independent of the recombination event at the TP0488 locus in the ancestor of TEN Bosnia A and TEN Iraq B . Interestingly , in the TPA Mexico A , TP0488 was found to contain a sequence very similar to that found in TPE strains , suggesting that the TP0488 locus is prone to gene recombination 36 ., The TP0488 gene encodes a methyl-accepting chemotaxis protein ( Mcp2-1 ) 37 and , as shown by expression profiling of treponemes isolated from rabbit infections , is highly expressed in TPA strains 38 ., Moreover , the Mcp2-1 protein has been shown to elicit a humoral response 37 ., In the TPA Mexico A genome , 8 out of 18 TPE-like changes were located in the Cache domain ( domain binding small molecules ) 39 ., Similarly , 13 out of 21 amino acid replacements resulting from recombination in the 11q/j isolate were also located in the Cache domain , suggesting differences in binding properties of the Mcp2-1 protein ., As discussed in a previous work 36 , the observed sequence patterns are consistent with recombination events that have likely occurred during parallel human infections with both TPA and TEN or TPA and TPE treponemes ., Gene TP0548 , on the other hand , was for the first time found to be recombinant ., TP0548 from the 11q/j isolate appeared to be composed of sequences ( in addition to TEN sequences ) originating from TPE treponemes ., Moreover , TEN Bosnia A and TEN Iraq B showed TPA-like sequences within this locus ., For this reason , as well as the fact that the ancestor of the 11q/j isolate diverged before the ancestor of TEN Bosnia A and Iraq B strains , it is more plausible that the recombination event occurred in the common ancestor of Bosnia A and Iraq B rather than in the 11q/j isolate ., The divergence of the ancestor of the 11q/j isolate before the ancestor of TEN Bosnia A and Iraq B strains is supported by greater genetic distances between the 11q/j isolate and both TEN strains ( Bosnia A , Iraq B ) compared to distances between Bosnia A and Iraq B ( Fig 3 ) ., According to this scenario , the recombination occurred in a TEN strain that was ancestral to both the Bosnia A and Iraq B , which incorporated the TPA sequence into this locus ( Fig 3C ) ., The sequence of the 11q/j isolate thus represents the original TEN sequence that is similar to TPE strains ., The TP0548 was predicted to encode for a rare outer membrane protein 40 and , as shown by molecular typing studies , is highly variable among syphilis isolates 18 , 22 , 25 , 27 ., The tendency of this locus to recombine , although shown only in TEN , should be considered during interpretation of data from both enhanced CDC and sequencing-based typing of syphilis-causing strains ., The calculated probability that the observed SNVs within TP0488 and TP0548 were caused by random mutations was extremely low , i . e . 1 . 01987 x 10−83 and 1 . 12245 x 10−65 , respectively , suggesting that recombination occurred in these loci ., Since mutation frequency per 1 bp within the TEN subspecies was calculated based on the sequences obtained for the 11q/j isolate , there was a potential bias in preferential sequencing of TEN variable regions ., However , inclusion of additional chromosomal loci would likely lower the final probability even more ., Moreover , the calculated SNV density in TEN strains ( 0 . 74 nt per 1000 bp ) differed only slightly compared to densities within other treponemal subspecies ( 0 . 36 nt per 1000 bp in TPA; 0 . 14 nt per 1000 bp in TPE; 29 ) ., Both intra-genomic and inter-genomic recombination events have been identified in uncultivable pathogenic treponemes and are summarized in Table 1 ., While intra-genomic homologous recombination have been found in tpr genes 7 , 10 , 41 , 42 , several inter-genomic recombination events have already been described in the literature 16 , 36 ., The fact that the infection caused by the TEN 11q/j isolate resembled early syphilis with lesions located on the genitals supports previous findings that both TPA and TEN strains form similar , clinically undiscernible primary lesions ., In a similar case , TEN Bosnia A was isolated from genital lesions of a 35-year old male , although in this case , lesions were also found in the oral cavity and pharynx and the patient showed secondary lesions on the face , trunk and extremities ., The patient which was the source of the 11q/j isolate in this study , reported that he returned to France from Islamabad , Pakistan , where he admitted having had sexual contact with commercial sex worker ., Since Pakistan is located close to countries that have recently reported cases of endemic syphilis , including Saudi Arabia and Iran 1 , such an infection should not be surprising ., Given the known number of repetitions in the arp gene , the restriction pattern of the amplified tprEGJ genes , and the TP0548 sequence , the enhanced CDC genotype 18 can be deduced for TEN strains ., For TEN Bosnia A and Iraq B , the 10q/c and 8q/c genotypes can be predicted based on published data , respectively 8 , 16 , 20 ., Interestingly , similar subtypes have already been identified among tested clinical isolates from China including 9h/c , 10h/c , and 9o/c 43 ., In fact , the electrophoretic tpr pattern “h” differs from the pattern “q” by one fragment ( i . e . , 804 bp in pattern “h” vs . 726 bp in “q” ) and “o” differs from pattern “q” by the absence of a 315 bp fragment 44 ., Close similarity of identified subtypes of human syphilis isolates to predicted subtypes of TEN Bosnia A and TEN Iraq B suggests that TEN strains could and should be sporadically detected among human samples from patients suspected of having syphilis ., In such situations , suspicious samples should be further analyzed to obtain an unequivocal classification of either a TPA strain or a TEN strain ., Taken together , analysis of the 11q/j isolate revealed a TEN genome seemingly containing two recombination events and highlights the fact that TEN strains could cause syphilis-like lesions in humans ., A more detailed analysis revealed that the 11q/j isolate had just one recombinant locus , TP0488 ., The recombination of TP0548 took place in a treponeme that was the ancestor of both TEN Bosnia A and TEN Iraq B . | Introduction, Materials and methods, Results, Discussion | Treponema pallidum subsp ., endemicum ( TEN ) is the causative agent of endemic syphilis ( bejel ) ., An unusual human TEN 11q/j isolate was obtained from a syphilis-like primary genital lesion from a patient that returned to France from Pakistan ., The TEN 11q/j isolate was characterized using nested PCR followed by Sanger sequencing and/or direct Illumina sequencing ., Altogether , 44 chromosomal regions were analyzed ., Overall , the 11q/j isolate clustered with TEN strains Bosnia A and Iraq B as expected from previous TEN classification of the 11q/j isolate ., However , the 11q/j sequence in a 505 bp-long region at the TP0488 locus was similar to Treponema pallidum subsp ., pallidum ( TPA ) strains , but not to TEN Bosnia A and Iraq B sequences , suggesting a recombination event at this locus ., Similarly , the 11q/j sequence in a 613 bp-long region at the TP0548 locus was similar to Treponema pallidum subsp ., pertenue ( TPE ) strains , but not to TEN sequences ., A detailed analysis of two recombinant loci found in the 11q/j clinical isolate revealed that the recombination event occurred just once , in the TP0488 , with the donor sequence originating from a TPA strain ., Since TEN Bosnia A and Iraq B were found to contain TPA-like sequences at the TP0548 locus , the recombination at TP0548 took place in a treponeme that was an ancestor to both TEN Bosnia A and Iraq B . The sequence of 11q/j isolate in TP0548 represents an ancestral TEN sequence that is similar to yaws-causing treponemes ., In addition to the importance of the 11q/j isolate for reconstruction of the TEN phylogeny , this case emphasizes the possible role of TEN strains in development of syphilis-like lesions . | Treponema pallidum subsp ., endemicum ( TEN ) is an uncultivable pathogenic treponeme that causes bejel ( endemic syphilis ) , a chronic human infection mostly affecting children under 15 years of age , occurring mainly in several African and Middle East countries ., In this work , we characterized a TEN 11q/j isolate from France that was obtained from an adult male with genital lesions , who was suspected of having syphilis and who received benzathine penicillin G . DNA sequencing of the isolate revealed two loci that were , rather than to TEN , related either to T . pallidum subsp ., pertenue or to T . pallidum subsp ., pallidum and likely resulted from recombination events ., The recombination event in TP0488 as well as the recombination in TP0548 , of the 11q/j , helped clarify the phylogeny of the TEN strains indicating that the recombination in TP0548 took place in a treponeme that was ancestral of Bosnia A and Iraq B , but was not an ancestor of the 11q/j isolate ., In contrast , a recombination event in TP0488 appeared in the ancestor of the 11q/j isolate after separation of the ancestral treponeme of Bosnia A and Iraq B . This case also points to a possible role of TEN strains in development of syphilis-like lesions in countries with endemic syphilis . | medicine and health sciences, pathology and laboratory medicine, pathogens, geographical locations, microbiology, sequence assembly tools, genome analysis, molecular biology techniques, bacterial pathogens, research and analysis methods, sequence analysis, artificial gene amplification and extension, bioinformatics, proteins, medical microbiology, microbial pathogens, recombinant proteins, molecular biology, genetic loci, people and places, biochemistry, biomolecular isolation, treponema pallidum, dna sequence analysis, asia, dna isolation, database and informatics methods, genetics, biology and life sciences, genomics, iraq, computational biology, polymerase chain reaction | null |
journal.ppat.1004744 | 2,015 | Elucidation of Sigma Factor-Associated Networks in Pseudomonas aeruginosa Reveals a Modular Architecture with Limited and Function-Specific Crosstalk | The ability to maintain homeostasis even in changing environments and under extreme conditions is one of the key traits of living organisms ., Pseudomonas aeruginosa is a ubiquitous gram-negative bacterium that can be distinguished by its exceptional high capability to adapt and survive in various and challenging habitats 1 ., The reason for the remarkable ecological success of P . aeruginosa can be attributed to its large metabolic versatility and environment-driven flexible changes in the transcriptional profile ., P . aeruginosa is not only an adaptive environmental bacterium but also an important opportunistic pathogen which exhibits an extremely broad host range 2 , 3 ., It is the causative agent of acute and chronic , often biofilm-associated , infections particularly in the immunocompromized host and cystic fibrosis patients 4–6 ., Genome sequencing of P . aeruginosa reference strains revealed a large genome with highly abundant global regulators and signaling systems that form a complex and dynamic regulatory network responsible for phenotypic adaptation and virulence 7–9 ., Among transcriptional regulators , sigma factors are of exceptional importance as they confer promoter recognition specificity to the RNA polymerase 10 , 11 ., They are essential for transcription initiation 12 which is the key step in gene regulation 13 ., Alternative sigma factors and in particular extracytoplasmic function ( ECF ) sigma factors can provide effective mechanisms for simultaneously regulating expression of large numbers of genes in response to challenging conditions 14 ., P . aeruginosa encodes more than 25 sigma factors most of which , including one strain-specific sigma factor , were reviewed in 2008 14 ., Among them are at least 21 ECF sigma factors 15 16 whose presence has been linked to bacterial virulence and pathogenicity 15 , 17–19 ., The advent of microarray technology has promoted the elucidation of bacterial genetic regulatory networks involved in adaptation to various environmental stresses and physiological processes 20 ., Subsequently , the combination of DNA microarray technology and chromatin immunoprecipitation ( ChIP-chip ) offered the opportunity to distinguish direct binding sites of transcription- and sigma-factors from those bound indirectly 21–23 ., With these valuable tools at hand , sigma factors gained greater attention and their impact on gene expression has become a major research focus 19 , 24–29 ., In this study , we constructed strains expressing his-tagged sigma factors in trans and/or sigma factor deletion mutant strains and performed mRNA profiling as well as chromatin immunoprecipitation coupled to high-throughput sequencing to uncover the direct and indirect regulons of 10 alternative sigma factors in P . aeruginosa ., Our results contribute to a deeper understanding of global gene regulation in bacteria and provide a reliable scaffold for the elucidation of the transcriptional regulatory network of the important pathogen P . aeruginosa ., Sigma factor genes were amplified by PCR using a forward primer harboring a ribosomal binding site and the ATG start codon and a reverse primer with the stop codon TGA ( S1 Table ) ., PCR products were introduced into pJN105 30 under control of PBAD resulting in pJN105-RBS-σ ., For ChIP-seq experiments pJN105-RBS-σ-8xhis was constructed using a reverse primer additionally encoding for an 8xHis-tag and for bioluminescence assays selected sigma factor target promoters were ligated into pBBR1-MCS5-TT-RBS-lux 31 ., Vectors were transferred into respective P . aeruginosa PA14 strains by electroporation as previously described 32 ., The PA14Δσ::Gmr deletion mutants were constructed according to a modified protocol using overlap extension PCR 33 ., The gene replacement vector pEX18Ap 34 was modified by inverse PCR to remove the coding sequence for 5S rRNA ., In addition , the resulting vector pEX18Ap2 encompasses a novel MCS established by primer extension ., Regions up- and downstream of the sigma factor gene were amplified by PCR ( S1 Table ) ., The primer Mut-σ-up-RV and Mut-σ-down-FW harbored complementary sequences coding for three shifted stop codons and a KpnI restriction site ( XmaI for rpoN ) ., The two corresponding PCR products were fused in a second PCR and the obtained fragment was introduced in pEX18Ap2 resulting in pEX18Ap2-up-σ-down-σ ., pEX18Ap2-up-σ-Gm-down-σ vectors were produced by ligation of a FLP-excisable gentamicin cassette amplified from pUC18-mini-Tn7T-Gm-lacZ into pEX18Ap2-up-σ-down-σ ., Single crossovers in PA14 were selected on gentamicin ., Counter-selection in LB low salt supplemented with sucrose resulted in PA14Δσ::Gmr ., Counter-selection for PA14ΔsigX::Gmr was performed in BM2 35 and PA14ΔrpoN::Gmr in LB supplemented with 1 mM glutamine ., The gentamicin cassette was excised from PA14ΔsigX::Gmr and from PA14ΔrpoN::Gmr using the FLP expression vector pFLP3 36 to obtain PA14Δsig and PA14ΔrpoN ., RNA was prepared from PA14 wild-type , PA14Δσ:Gmr , PA14 ( pJN105 ) and PA14 ( pJN105-RBS-σ ) in two independent experiments each containing a pool of three individual main cultures ( in 10 ml medium at 37°C ) ., 0 . 5% L-arabinose was added to PA14:pJN105-RBS-σ and the corresponding control PA14:pJN105 for at least 35 min ., To maximize expression of the sigma factor dependent regulons the strains were cultivated under conditions previously shown to induce the activity of the various sigma factors ., Therefore , PA14 ( pJN105-RBS-fliA ) was harvested in the exponential phase ( OD600 = 1 . 1 ) , PA14 ( pJN105-RBS-rpoS ) and PA14 ( pJN105-RBS-rpoN ) were cultivated to the early stationary phase ( OD600 = 2 . 0 ) ., PA14 ( pJN105-RBS-algU ) was grown to an OD600 of 2 . 3 and exposed to 50°C for 5 min ., PA14 ( pJN105-RBS-rpoH ) was grown at 28°C up to an OD600 of 1 . 4–1 . 5 including 35 min induction of rpoH expression and was exposed to 42°C for 5 min ., PA14 ( pJN105-RBS-sigX ) was cultivated in low osmolarity LB containing 8 mM NaCl ., PA14 ( pJN105-RBS-pvdS ) , PA14 ( pJN105-RBS-fpvI ) , PA14 ( pJN105-RBS-fecI ) and PA14 ( pJN105-RBS-fecI2 ) were exposed to iron starvation ( growth in 50% LB to OD600 of 1 . 5 , incubation with the iron-chelating agent 2 , 2’-bipyridyl ( 200 μM ) for 70 min ) ., PA14Δσ::Gmr deletion mutants were cultivated as the sigma factor in trans expressing strains ., PA14ΔrpoN::Gmr was also cultivated under nitrogen-limitation in BM2 containing 0 . 1% casein amino acids as sole nitrogen source to an OD600 of 1 . 2 and the growth-impaired PA14ΔsigX strain was grown under low osmolarity condition to the same OD as the corresponding PA14 wild-type strain ., RNA extraction , cDNA library preparation and Illumina sequencing were performed as previously described 37 ., In brief , cells were harvested after addition of RNA protect buffer ( Qiagen ) and RNA was isolated from cell pellets using the RNeasy plus kit ( Qiagen ) ., mRNA was enriched ( MICROBExpress kit ( Ambion ) ) fragmented and ligated to specific RNA-adapters containing a hexameric barcode sequence for multiplexing ., The RNA-libraries were reverse transcribed and amplified resulting in cDNA libraries ready for sequencing ., All samples were sequenced on an Illumina Genome Analyzer II-x in the Single End mode with 36 cycles or on a HiSeq 2500 device involving 50 cycles ., Sequences were mapped to the PA14 genome using stampy 38 with default settings and the R package DESeq 39 for differential gene expression analysis ., Differentially expressed genes were identified using the nbinomTest function based on the negative binomial model after pre-filtering by overall variance 40 ., The Benjamini and Hochberg correction was used to control the false-discovery rate at 0 . 05 to determine the list of regulated genes 39 ., The quality control output in PDF format is available for download as part of the supplementary information accompanying GEO dataset ., Genes were identified as differentially expressed if they fulfilled the following criteria:, i ) an at least three-fold down-regulation in the sigma factor mutant as compared to the corresponding wild-type strain or an at least three-fold up-regulation in the strains expressing the sigma factor in trans as compared to the cognate empty vector control strain and, ii ) the Benjamini-Hochberg corrected P value was smaller than 0 . 05 with the exception of PA14ΔfpvI:Gmr and PA14ΔfecI:Gmr whose cut-off values were set to a fold change of at least 2 using the uncorrected P value ., To appraise sigma factor competition , we determined also the negative impact of the expression of the sigma factor in trans on genes and considered genes which were at least three-fold down-regulated with a maximal P value of 0 . 05 ., We computed the pair-wise Pearson correlation between the log2-normalized read counts ( nRPKs ) of all but the rRNA/tRNA genes in PA14 using data from all transcriptome replicates generated for a given condition and sigma factor ., We performed hierarchical clustering of the genes in the resulting expression matrix ( 10 alternative transcription factors * 2 replicates * knockout/in-trans expressing conditions ) by progressively grouping them: at each step of the iterative algorithm the two genes or gene clusters that have the smallest distance were merged to form a new cluster , and two branches of a growing tree were joined ., The lengths of the branches are equal to the half of the distance between two genes or gene clusters ., We used the average linkage rule; this means that the distance between two clusters is computed as the mean of all the distances between the genes in the first cluster and the genes in the second cluster ., All calculations were performed in R using the hclust function ., ChIP-seq was applied to four 20 ml cultures ( with pooling of two individual cultures ) of PA14 ( pJN105-RBS-σ-8xHis ) and PA14 ( pJN105 ) as a control ., ChIP-seq samples were treated under the same condition as described for mRNA profiling with the exception of PA14 ( pJN105-RBS-rpoH-8xHis ) which was exposed to a heat-shift from 37°C to 42°C ., Following treatment with 0 . 5% formaldehyde and glycine cells were harvested , washed and suspended in 0 . 5 ml of lysis buffer ., DNA was fragmented to an average size of 200 to 250 bp and subjected to chromatin immune-precipitation with 15 μl of anti-6xHis tag antibody ( ab9108; abcam ) overnight at 4°C ., Following an incubation step with 1 μl of RNase A ( 100 mg ml−1 ) and proteinase K ( 20 mg ml−1 ) immunoprecipitated DNA was recovered using a QIAquick PCR Purification kit ( Qiagen ) and subjected to a modified linear DNA amplification ( LinDA ) protocol 41 ., For next generation Illumina sequencing , up to 50 ng of DNA was used in a TruSeq DNA sample preparation kit ( Illumina ) according to the low-throughput protocol ., ChIP-seq data was analyzed by removing adapter sequences using the fastq-mcf script that is part of the EA-utils package 42 ., Reads were trimmed allowing for minimal quality of 10 at their ends ., We used the Bowtie aligner 43 to map the reads ., Model-based analysis of ChIP-seq 44 was applied for peak detection using a P value cut-off value of 0 . 05 and shift size 30 for the peak modeling , making use of the relevant control samples ., Details on the promoter hits from the individual replicates are available in S5 Table ., Promoter hits were considered significant when they were detected in both ChIP-seq replicates with an enrichment factor of at least 3 and a P value of less than 0 . 01 ., Statistical analysis of the obtained candidates was performed to assess the number of false positives and the corresponding P value according to the hypergeometric test in R using the phyper command ., ChIP results using an anti-RpoS polyclonal antibody followed by microarray analysis was included in this study 45 ., DNA was purified and amplified and approximately 7 . 5 μg of amplified DNA from the control and the RpoS ChIPs were sheared to a fragment size of 50 to 500 bp and terminally labeled using the GeneChip WT double stranded DNA terminal labeling kit ( Affymetrix ) ., The biotin-labeled DNA was hybridized to an Affymetrix P . aeruginosa genome chip as described previously 46 ., Enrichment of hybridization signals was calculated with the Tiling Analysis Software ( TAS , Affymetrix ) from two independent RpoS ChIPs compared to two independent mock ChIPs ( bandwidth parameter was set to 150 bp ) ., For each gene , the promoter region was defined as the sequence from -300 to -1 bp based on the PAO1 annotation 47 ., Threshold levels for significantly enriched promoter region were log2 enrichment factor of at least 0 . 5 with P value less than 0 . 05 across at least 70 bp DNA segments ., We selectively sequenced the 5’-ends of primary transcript samples of PA14 cultivated under six different growth conditions: transitional phase between exponential and stationary growth phase ( OD600 = 1 . 5–2 . 0 ) , exponential phase ( OD600 = 0 . 5–1 . 5 ) , heat shock at 42°C and 50°C , low osmolarity and iron depletion ., Total RNA was extracted and treated with terminator exonuclease as described previously 48 to remove processed and incomplete transcripts ( including tRNA and rRNA ) ., The remaining primary transcripts were ligated to 5’-RNA adapters using T4 RNA ligase and subsequently reverse transcribed using SuperScript III ( Invitrogen ) reverse transcriptase with a modified RT-N8 primer containing an octameric random sequence at its 3’-end ( 5‘-GCT GAA CCG CTC TTC CGA TCT NNN NNN NN -3‘ ) ., The resulting cDNA contains random 5’-ends while the 3’-ends conserve the 5’-ends of the original RNA species ., A PCR with primers equivalent to the Illumina Paired End Primers was performed on the cDNA yielding double-stranded cDNA libraries that were subsequently sequenced on an Illumina Genome Analyzer IIx ., The sequenced 5´end primary transcript data were clipped to remove low quality sequences and adapters and were subsequently mapped to the reference genome of P . aeruginosa PA14 using bowtie 43 ., In order to detect putative TSS , read counts were normalized by the total number of reads present in each sample obtaining Reads Per Million mappable reads ( RPM ) and TSS were detected by selecting sites that exceeded 10 RPM in any of the 8 samples grown under six different environmental conditions ., TSS separated by less than 3 base pairs were merged , and the position of the TSS having the highest RPM was set as position of the putative TSS ., This resulted in a final list of 5583 TSS that were classified as described previously 37:, i ) promoter TSS ( 3520 sites corresponding to 2309 genes ) , if they were detected up to 500 base pairs upstream of a gene on the same strand ,, ii ) intragenic TSS ( 1709 sites ) , if they were detected within the margins of a gene on the same strand ,, iii ) antisense TSS ( 1027 sites ) , if they were detected within the margins of a gene on the opposite strand or, iv ) orphan TSS ( 1156 ) , if they were not detected within or upstream of a gene on the same strand ( in other words , neither promoter nor intragenic TSS ) ., The 2309 genes having promoter TSS had additional 1211 alternative TSS ( 2309+1211 = 3520 TSS ) ., Thus , many genes ( 739 ) contained more than one alternative TSS ., We defined TUs by applying a combination of three independent criteria on all annotated genes of the PA14 genome:, i ) a TSS was detected in our TSS-Seq approach within 500 bp upstream of the gene on the same strand ,, ii ) the immediate upstream gene on the same strand shows at least a two-fold difference in the median gene expression across the 47 transcriptomes of P . aeruginosa PA14 wild type , and, iii ) a gene is predicted to be the first ( or only ) gene on an operon in the DOOR database 49 ., The DOOR database includes operon predictions based on intergenic distance , neighborhood conservation , phylogenetic distance , information from short DNA motifs , similarity score between GO terms of gene pairs and length ratio between a pair of genes 50 and was found to predict operons in E . coli with 93 . 7% accuracy 51 ., To increase the accuracy of our overall TU prediction we employed a conservative approach and assigned TUs only if both criteria, i ) and, ii ) or criterion, iii ) were fulfilled , resulting in 3687 TUs ( S4 Table ) ., Most of these TUs were already included in the DOOR database except 159 TUs that were only positive in criteria, i ) and, ii ) ., For 2025 of those 3687 TUs we were able to detect the TSS positions experimentally ., A large fraction , 2499 ( 67 . 7% ) of those 3687 TUs were singleton operons , further 657 TUs contained 2 genes on the operon ., Wurtzel et al . 52 previously defined 3794 TUs ( 2117 of which were detected experimentally ) ., We furthermore analyzed the overlap of 1381 TSS which shared a respective TU as determined previously 52 and in this study ., Whereas 69% of them ( 958 TSS ) were separated by < 2bp , only 15% of them ( 205 TSS ) were positioned 50bp or more apart from each other on the genomic coordinate ., In general , sigma factor binding motif was identified by applying the MEME suite 53 on promoter regions ( 300 bp upstream of start codons ) whose respective genes, ( i ) showed at least a three-fold sigma factor-dependent down-regulation in PA14Δσ::Gmr and at least a three-fold sigma factor-dependent up-regulation in PA14 ( pJN105-RBS-σ ) or alternatively a more than ten-fold down-regulation in PA14Δσ::Gmr only or a more than ten-fold up-regulation in PA14 ( pJN105-RBS-σ ) only and, ( ii ) were defined to be the first gene of a transcriptional unit ., The SigX motif is based on genes which are the first gene of a TU , show a differential expression of at least 3 and whose promoters were enriched at least three-fold ( P value less than 0 . 01 ) in both ChIP-seq replicates ., For RpoD the TOP 3 motifs were elucidated selecting promoters which are enriched at least 5-fold in both ChIP-seq replicates ( P value less than 0 . 001 ) and whose genes are the first gene of a TU and whose gene expression was not affected by alternative sigma factors ., General parameters were selected as followed: occurrence ( 0 or 1 per sequence ) , number of sites ( minimum , 7 ) and activated DNA option ‘search given strand only’ ., The motif width was adapted to each sigma factor ., Furthermore , a background Markov model was supplied ., The obtained motif was forwarded to FIMO 54 to identify putative sigma factor binding sites in all promoter regions across the PA14 genome ., Promoter hits with a P value less than 0 . 0005 were regarded as significant and were listed in S5 Table ., A gene was defined to be a member of the primary sigma factor regulon if it fulfilled at least two of the following three criteria:, i ) it exhibited sigma factor-dependent regulation of expression ,, ii ) its promoter was enriched in ChIP-seq experiments and, iii ) its promoter contained a sigma factor binding motif ., Since RpoD is an essential gene no deletion mutant could be constructed and in trans expression of RpoD only led to up-regulation of three genes—probably due to high abundance of RpoD in the cell , Thus , no RNA-seq data were available to describe the impact of RpoD of the global transcriptional profile ., We therefore considered those genes that were not differentially regulated by any of the tested alternative sigma factors as belonging to the primary RpoD regulon , however , only if they were either found in the RpoD ChIP-seq approach or harbored an RpoD motif ., Statistical significance of these primary regulon members was checked by performing a hypergeometric test using the phyper command in R on the intersections ChIP-seq/RNA-seq , RNA-seq/motif search , and ChIP-seq/motif search ( S7 Table and S3 Fig . ) ., To include not only first genes but all genes of identified transcriptional units from S4 Table , downstream genes were added , if the first gene met the criteria indicated above ., These final sets of genes ( S6 Table ) were functionally characterized using the PseudoCAP annotation 55 ., To further improve this profiling , the PseudoCAP PA14 annotation was updated by adding the PseudoCAP classes of PAO1 homologs to corresponding PA14 genes ., Over- or underrepresentation of each PseudoCAP category was calculated by comparing normalized PseudoCAP category experimentally detected and normalized PseudoCAP category annotated as previously described 19 ., The enriched categories and their P values obtained using a hypergeometric test are listed in S8 Table ., For each bioluminescence assay , three independent experiments were performed and each experiment included pooling of three biological replicates ., Reporter strains ( S1 Table ) harboring selected sigma factor target promoter fused to the luxCDABE cassette of Photorhabdus luminescens were grown under same conditions as described under mRNA profiling ., Bioluminescence of 200 μl bacterial suspension was measured in a black 96-well microtiter plate with a transparent and flat bottom ., In parallel , cell density was determined using a standard photometer ., Relative light units ( RLU ) were normalized to the optical density at a wavelength of 600 nm and the arithmetic average was calculated ., Next , the average bioluminescence over the three independent assays was calculated and compared to the bioluminescence of the respective control reporter strain , e ., g ., the reporter construct in the corresponding sigma factor mutant strain or the PA14 wild-type strain harboring the empty vector ., The standard error was determined and the statistical significance was examined using the two-tailed Student’s t-test assuming unequal variances ., The raw and processed data have been deposited in the Gene Expression Omnibus ( GEO ) database ( accession numbers GSE54997 and GSE54998 united under SuperSeries GSE54999 ) ., The short read data is available through the GEO interface under projects SRP037770 and SRP037771 ., In this study , we aimed at deciphering the regulons of alternative sigma factors and to quantify their relative contribution to the overall transcriptome plasticity in the opportunistic pathogen P . aeruginosa ., We therefore amended our previously published data on the impact of the alternative sigma factor SigX 19 on the global transcriptional profile in the type strain PA14 and further expressed the his-tagged alternative sigma factors AlgU , FliA , RpoH , RpoN , RpoS , PvdS , FpvI , FecI and FecI2 in trans ( S1 Table ) ., We also inactivated these alternative sigma factors ( with the exception of RpoH and FecI2 ) and recorded transcriptional profiles ( S2 and S3 Table ) under growth conditions that are expected to support sigma factor dependent gene expression ( see Materials and Methods for details ) ., We observed activity of all sigma factor target promoters ( not done for the FpvI , FecI and FecI2 sigma factor targets which are known to respond to low iron medium conditions ) under the selected experimental conditions ., This activity was strictly dependent on the presence of the respective alternative sigma factor in the reporter strain ( AlgU , FliA , RpoN , RpoS , SigX ) or on the presence of inducing conditions ( RpoH , PvdS ) ( S1 Fig . ) ., Overall , 491 genes were up-regulated in response to ( hyper- ) presence and down-regulated in the absence of at least one alternative sigma factor ( Fig . 1A ) ., Additional 1195 genes were up-regulated in response to ( hyper- ) presence and 532 were down-regulated in the absence of at least one alternative sigma factor ., Interestingly , the majority of 1504 out of the overall 2218 genes ( 67 . 8% ) were differentially regulated due to in trans expression and/or inactivation of only one single sigma factor ( colored bars in Fig . 1A ) ., The finding that many genes are exclusively affected by only one alternative sigma factor indicates that the alternative sigma factor regulons are distinct functional modules and that they have only a limited overlap at the level of transcription ., Nevertheless , there was also shared regulation: the expression level of 471/2218 genes ( 21 . 2% ) was influenced by two sigma factors ( white bars in Fig . 1A , colored bars in Fig . 1B ) and 243/2218 genes ( 11% ) were influenced by more than two sigma factors ( S2 and S3 Table ) ., Thus , in addition to the detection of largely isolated regulons ( Fig . 1A ) , transcriptional profiling also uncovered a co-ordinated gene expression pattern in which many genes were affected by distinct sets of alternative sigma factors ( Fig . 1B and C and Table 1 ) ., In this study , transcriptional profiling was performed either in the absence and or in trans expression of the various alternative sigma factors to improve the elucidation of the primary and complete sigma factor regulons ., This strategy was proven valid for numerous sigma factors in P . aeruginosa 15 , 56–58 ., However , since there is a limited amount of RNA polymerase in the cell , we analyzed whether sigma factor expression might negatively impact the global gene expression profile under our experimental conditions ., Overall we found 644 genes ( 10 . 9% of the whole genome ) that were three-fold down-regulated upon in trans expression of any one of the ten alternative sigma factors , 169 genes that were negatively affected by two of the ten sigma factors and only 85 genes affected by more than two sigma factors ., These results indicate that although the expression of a distinct set of genes might be affected by sigma factor competition for the RNA polymerase , there is no notable alternative sigma factor competition on a global scale under our experimental conditions ., It seems that in P . aeruginosa competition of alternative sigma factors for a limiting amount of RNA polymerase does not play a general role and indicates robustness of overall gene expression to shifts of alternative sigma factor levels ., To define the primary regulons of the P . aeruginosa sigma factors we complemented our transcriptome data with chromatin immunoprecipitation followed by high-throughput sequencing ( ChIP-seq ) and in case of RpoS with ChIP-chip experiments ., This allows the differentiation of direct from indirect sigma factor-dependent regulation of genes ., We constructed variants of the housekeeping sigma factor RpoD and ( in addition to SigX 19 ) the nine alternative sigma factors fused to an octahistidine-tag and sequenced sigma factor bound genomic DNA ., In order to define transcriptional start sites ( TSS ) and to predict the transcriptional units ( S4 Table ) , we selectively sequenced the 5’-ends of primary transcript samples of PA14 cultivated under six different growth conditions ( as outlined in material and methods ) ., This served as the basis for elucidating the de novo binding motif of each sigma factor 53 ( Fig . 2 ) ., We used the MEME suite ( 27 ) on those promoter regions whose respective genes exhibited an alternative sigma factor dependent regulation of expression and which were upstream of the first gene within a transcriptional unit ( see Material and Methods section for details ) ., We were able to generate a de novo sequence logo for each of the ten alternative sigma factors ., Furthermore , in 76 . 4% of the genes that harbored a sigma factor binding motif and for which a TSS could be detected experimentally ( 814 genes ) the motif was demonstrated to be located at the expected distance ( max . 60 nucleotides ) from the TSS ., As exemplified in the previously published primary regulon of SigX 19 , we then defined a gene to be a member of the primary regulon of the P . aeruginosa sigma factors if it fulfilled at least two of the following three criteria:, i ) it exhibited sigma factor-dependent regulation of expression ,, ii ) its promoter was enriched in ChIP-seq experiments and, iii ) its promoter contained a sigma factor binding motif ., Detailed view on the intersections between the different approaches for the individual regulons is provided in S3 Fig ., , P values from the hypergeometric tests are listed in S7 Table ., Based on these data , the primary P . aeruginosa sigma factor regulome is depicted in Fig . 3 ( an interactive image is available at https://bactome . helmholtz-hzi . de/ ) ., S5 Table shows the promoter enrichment by means of ChIP-seq or motif search before applying our criteria for definition of primary regulons ., S6 Table lists the discrete sets of genes belonging to the individual primary sigma factor regulons as defined above ., The primary sigma factor regulome covers 2553 genes ( 43% of the genome ) including 598 genes of unknown function ., This number represents the most significant candidates which were obtained by stringent threshold settings ., Due to the low conservation of the RpoD binding motif and no RpoD RNA-seq data , the RpoD regulon ( encompassing 686 genes ) is probably significantly underestimated ., Of note , meeting two out of the three of the criteria to define the sigma factor regulons decreased the regulon size of e . g . RpoH from 268 when meeting just one of the criteria ( RNA–seq ) ( Table 1 ) to 96 ( Table 2 ) ., However , on the other hand the regulon sizes of RpoN , AlgU and RpoS even increased , indicating that ChIP-seq in combination with a motif scan uncovered additional sigma factor binding sites ., The validity of the selection criteria was further verified by functionally categorizing the members of each primary sigma factor regulon by the use of the PseudoCAP annotation 55 ., The results are summarized in Fig . 4 ( the enrichment values and their P values are listed in S8 Table ) ., As expected , the AlgU regulon comprises genes of alginate biosynthesis and cellular homeostasis 59–63 , the motility sigma factor FliA influences genes involved in chemotaxis and motility 64 , 65 and PvdS directs the pyoverdine biosynthesis genes 66 which are assigned to the category adaptation/protection ., The heat-shock sigma factor RpoH governs the gene expression of chaperones and heat-shock proteins 67 , while RpoN controls genes of nitrogen metabolism , chemotaxis , motility and attachment 68–70 ., The stationary phase sigma factor RpoS regulates quorum sensing genes as well as genes involved in general adaptation processes 71 , 72 ., A more detailed description of the individual regulons is provided in the supplementary material ( S1 Text ) ., Beyond the assignment of genes to specific sigma factor regulons , our experimental design allowed us to define sets of genes that are affected by more than one sigma factor ., We were able to assign as many as 1149 genes ( 61 . 6% of the primary alternative sigma factor regulome ) to one distinct sigma factor regulon ., Whereas those genes were exclusively affected by one sigma factor and did not participate in sigma factor crosstalk , 401 genes belonged to the primary regulon of more than one sigma factor ( direct crosstalk ) and 317 genes belonged to the primary regulon of one sigma factor , but were additionally affected on the transcriptional level by the activity of a second alternative sigma factor ( indirect crosstalk ) ( Table 2 ) ., Both , the primary alternative sigma factor regulon and the RNA-seq data , revealed that all alternative sigma factors showed auto-regulation which is well-known for ECF sigma factors 16 ., However , cross-talk among the sigma factors was very limited ., We found only a direct impact of AlgU on rpoH expression , while indirect cross talk was identified between RpoH and algU as well between FpvI and fecI2 ., These results corroborate the finding of insulated sigma factor networks ., Direct crosstalk was mainly found to involve genes of the more complex functional categories adaptation/protection , chaperones/heat shock proteins , chemotaxis , motility/attachment , protein secretion and secreted factors ( Fig . 5 ) ., There was also a preference of sigma factor combinations within the direct crosstalk ., Direct crosstalk with RpoN clearly played the most dominant role ( S2 Fig . ) ., In total , 183 out of the 401 genes affected by | Introduction, Materials and Methods, Results, Discussion | Sigma factors are essential global regulators of transcription initiation in bacteria which confer promoter recognition specificity to the RNA polymerase core enzyme ., They provide effective mechanisms for simultaneously regulating expression of large numbers of genes in response to challenging conditions , and their presence has been linked to bacterial virulence and pathogenicity ., In this study , we constructed nine his-tagged sigma factor expressing and/or deletion mutant strains in the opportunistic pathogen Pseudomonas aeruginosa ., To uncover the direct and indirect sigma factor regulons , we performed mRNA profiling , as well as chromatin immunoprecipitation coupled to high-throughput sequencing ., We furthermore elucidated the de novo binding motif of each sigma factor , and validated the RNA- and ChIP-seq results by global motif searches in the proximity of transcriptional start sites ( TSS ) ., Our integrated approach revealed a highly modular network architecture which is composed of insulated functional sigma factor modules ., Analysis of the interconnectivity of the various sigma factor networks uncovered a limited , but highly function-specific , crosstalk which orchestrates complex cellular processes ., Our data indicate that the modular structure of sigma factor networks enables P . aeruginosa to function adequately in its environment and at the same time is exploited to build up higher-level functions by specific interconnections that are dominated by a participation of RpoN . | Pseudomonas aeruginosa is well known for its high adaptability to a large range of environmental conditions , including those encountered within the human host ., Transcription initiation represents a major regulatory target which drives versatility , and enables bacterial adaptation to challenging conditions and expression of virulence and pathogenicity ., In bacteria , this process is largely orchestrated by sigma factors ., Here , we performed an integrative approach , and by the combined use of three global profiling technologies uncovered the networks of 10 alternative sigma factors in the opportunistic pathogen P . aeruginosa ., We demonstrate that these networks largely represent self-contained functional modules which exhibit a limited but highly specific crosstalk to build up higher-level functions ., Our results do not only give extensive information on sigma factor binding sites throughout the P . aeruginosa genome , but also advance the understanding of sigma factor network architecture which provides bacteria with a framework to function adequately in their environment . | null | null |
journal.pcbi.1004910 | 2,016 | Tuning Curves for Arm Posture Control in Motor Cortex Are Consistent with Random Connectivity | The dependence of neuronal responses on stimulus- or movement-related parameters is often characterized by tuning curves ., A natural assumption is that a smooth , regular tuning curve reflects structured , orderly input to a neuron ., However , neuronal responses inevitably involve a degree of irregularity , even when responses are averaged across trials ., Do such irregularities simply reflect noise in the inputs , or might they suggest something more complex such as unstructured input ?, Here we address this question using data recorded from primary motor cortex ( M1 ) during an arm posture task , augmented by a neural network model of M1 neurons and their inputs ., Interpreting neural activity during arm movements is difficult because motor and sensory activity as well as limb biomechanical variables all change simultaneously ., In this study , we focus on the often ignored task of actively maintaining arm posture ., Arm posture control is a natural behavior without the dynamic complexity of time-varying movements ., It is therefore well suited to address the nature of neuronal tuning curves and their relationship to input ., To reveal fine-scale tuning curve structure , we employed a task with 54 different arm postures , consisting of 27 target positions with two forearm rotation angles ., Previous studies of arm posture control 1–3 concluded that neuronal responses in M1 vary as linear functions of hand position , which , by a change of coordinates , is equivalent to cosine tuning ., The finding of broad , singly peaked tuning curves in both motor and visual areas led to proposals suggesting that both are generated by similarly structured cortical microcircuit 4 , 5 ., The M1 tuning curves we extract have a strong linear component , in agreement with previous studies , but we also find sizable and significant nonlinear elements ., We examine the quality of the fit of linear tuning curve models across the entire population of tuned neurons , we analyze the nonlinear components in detail , and we determine how both linear and nonlinear elements contribute to the accuracy of inferring EMGs from M1 neuronal activity ., Finally , we construct a neural network model of M1 activity driven by target-position related inputs that accurately replicates the data ., Surprisingly , the synaptic connectivity in this model is completely unstructured , in fact , random ., This model shows that the presence of a strong regular and , in this case , linear component in the tuning of a population of neurons does not necessarily imply structured connectivity ., The responses of each neuron , for a particular forearm angle , are summarized by a set of 27 mean firing rates , one for each hand position ( Fig 2 ) ., To study the relationship between firing rate and hand position , we fit spatial linear tuning curves 2 to the responses of all the tuned neurons ( N = 411; Methods ) , separately for pronation and supination ., Neurons with perfectly spatial linear tuning would fire at rates that increase linearly when the hand is held at different positions along a Preferred Position ( PP ) vector and are constant across locations in planes orthogonal to this vector ., During pronation , the response of the neuron shown in Fig 2A is well fit by a linear tuning curve ( Fig 2C ) ., However , this is one of only 9 neurons with R2 > 0 . 9 for the fit to linear tuning ., Across the population of tuned neurons , the distribution of R2 values for the spatial linear tuning fit is broad ( Fig 3B ) ., We pooled the R2 distributions from both forearm angles because they are not significantly different ( Wilcoxon rank-sum test , p = 0 . 56 ) and , in fact , there is only a small reduction in goodness of fit if the PP vectors for each neuron are required to be the same for both forearm angles ( Methods , Eq 3 ) ., The R2 distribution has a median of 0 . 52 , and its breadth indicates that neurons range from being well fit to poorly fit by a linear tuning curve ., We tested whether the deviations of the data from the linear fit are due to noise by generating artificial data with perfectly spatial linear tuning , using the model parameters obtained by fitting the real data , and including noise extracted from the real data ( resampled 1 , 000 times; Methods ) ., We calculated mean responses for these artificial data and re-fit them to the linear tuning curves ., The resulting distribution of R2 values is skewed towards 1 ( Fig 3C ) , with a median of 0 . 75 , indicating significantly better fits than those for the real data ( Wilcoxon rank-sum test , p < 10−46 ) ., Generating Poisson spike trains with rates given by the generated linear tuning curves over the same number of trials as the real neurons yielded similar results ., We also studied whether the observed neuronal responses had linear structure above chance level by generating another artificial dataset constructed from randomly generated responses drawn across the same range of firing rates as the real neurons ., Noise traces extracted from the data were added to these responses ., When fit to the linear tuning curve , this resulted in significantly worse fits than for the real data , with an R2 distribution ( Fig 3A ) skewed towards zero with a median of 0 . 12 ., These analyses indicate the linear model is not a complete description of the tuning curves extracted from the data , but that the linear component is significantly larger than what would be expected from a random mapping between hand position and firing rate ., The neuronal responses deviate from spatial linearity in myriad ways ., Some are bimodal , with high firing rates in two regions of space and low firing rates for the targets in between ( Fig 2E ) ., Other neurons have spatially localized tuning , responding to only one or two adjacent target locations ( Fig 2I ) ., Still others display more complex nonlinear spatial patterns , either on top of a linear component ( Fig 2M ) or without one ( Fig 2N ) ., Strikingly , most of the responses exhibit high spatial frequencies , with large differences in firing rates between neighboring targets in a seemingly unsystematic manner ., This causes a zigzag pattern in the projected responses ( Fig 2D , 2G , 2O and 2P ) ., Such “salt and pepper” tuning does not support the explicit encoding of motor parameters ( hand position , joint angles , etc . ) by single neurons , because these should vary systematically with changes in hand position ., We quantified the spatial irregularity of each neuron’s responses using a complexity measure , based on the distribution of firing rate differences between all pairs of neighboring targets ( Methods ) ., The resulting distribution of spatial complexities across the real neurons ( Fig 3D ) covers values that are significantly larger than for linear and regular nonlinear ( threshold-linear , exponential , sigmoidal; with added noise as described above ) artificial datasets , yet significantly smaller than for an artificial dataset constructed from random responses ., Stated another way , the real responses contain more power at high spatial frequencies than conventional parametric linear and nonlinear tuning curves , but not as much as expected for completely random tuning ( Fig 3D ) ., Because the neuronal responses are not well described by standard parametric tuning curves ( e . g . threshold-linear , exponential , sigmoidal , etc . ) , we used principal component analysis ( PCA ) to quantify their shapes non-parametrically ., We also performed PCA on noise traces resampled from our data to implement a procedure developed by Machens et al . 6 for separating signal from noise ( Methods ) ., The first 12 PCs of the full data each account for more variance than the 1st PC of the noise ( Fig 3E; 99% confidence interval , bootstrap ) ., Using this strict threshold , we find that the responses occupy 12 of the possible 54 dimensions , with the rest assumed to be noise ., If , instead , we take the mean of the noise variance across all of its PCs as the noise threshold , the responses are 36 dimensional ., These results do not depend on whether or not the data is normalized ( Methods ) ., We also performed PCA on the separate pronation and supination data sets , obtaining in each case 12 or 20 dimensions above the maximal or mean noise thresholds , respectively ., Regardless of where in the range from 12–36 the exact dimensionality lies , it is much higher than the 3 dimensions expected for linear tuning , and it is presumably sufficiently high to allow these neurons to control the arm muscles required for this task ., To study the signal variance in the data , we subtracted away the noise variance and compared the resulting cumulative explainable variance to that of the artificial linear dataset , which is described by only 3 PCs ( Fig 3F ) ., The difference between these curves reveals the contribution of the heterogeneous nonlinear components ., The first 3 PCs of the real responses roughly describe the linear component described above ( F-Test , p < 0 . 01 , Fig 3G , Fig 3H ) ., The 4th and higher PCs ( Fig 3I ) provide a basis for the heterogeneous nonlinear structure found in the responses ( S3 Fig ) ., How much better can a general model describe the neuronal responses than the linear tuning curves ?, To answer this question we used the non-noise PCs to construct nonlinear “models” of the neuronal responses and compared these with the data ., Using only 12 PCs , results in a dramatically improved description relative to the linear fit , with 93% of neuronal responses having an R2 > 0 . 52 ( the median value for the linear tuning curve fit ) and a median R2 of 0 . 84 ( Fig 3J ) ., Using 20 PCs produces an excellent fit for every neuron ( median = 0 . 95; Fig 3K ) ., To investigate the relationship between position-dependent tuning of M1 neurons and arm muscle activity , we studied how the EMG of the primary muscles used in the task could be decoded from M1 activity ., To obtain sufficiently accurate measurements of EMG activity , we used the mean EMG ( averaged over a time window , trials , and 6 days; Methods ) recorded at 18 arm postures ( 9 targets times 2 forearm rotation angles; Fig 4A ) rather than the full 27 ., Like the neuronal activities , the EMG of all 5 muscles had a significant nonlinear component , with 3 of the muscles significantly fit by a linear tuning curve as well ( F-test , p < 0 . 01 ) ., To reconstruct these EMGs from M1 activity , we used single-trial neuronal firing rates combined from neurons recorded across different days ( pseudo-simultaneous population activity; Methods ) ., This is a reasonable strategy because the EMGs were averaged over trials and days as well ., However we expect this to limit our ability to decode the EMG because using the mean EMG and pseudo-population activity breaks single-trial correlations ., We performed decoding using 1 , 000 different random cross-validation splits of the data into training and test sets ( C . V . repetitions ) each consisting of 18 test trials ( one per arm posture ) ., We obtained the best performance in cross-validation tests by using optimal subset selection 7 in which the linear decoder is constructed with a LASSO algorithm 8 ., This selects the optimal subpopulation of neurons for decoding each muscle , rather than using all of the neurons to decode all of the EMGs ( Methods ) ., This has the additional benefit of providing information about how the signals that affect different muscles are distributed across the M1 population ( S2 Text ) ., We first determined how well the EMGs could be decoded using the full neuronal responses ( see Methods ) ., The decoding predictions using the neural activity capture the nonlinear structure in the EMG even on a single-trial basis , outperforming the fit of the EMG to a linear tuning curve , which is based on the mean EMG values ., The latter is the best that could be accomplished from the neurons if they had purely spatially linear responses and no noise ., We assessed decoding performance using the full responses and also different response components separately , by dividing the data into distinct spatially linear and nonlinear components ., To make this split , each single-trial firing rate was expressed as the sum of its underlying mean linear component ( the value predicted by the linear tuning curve ) , the mean nonlinear component ( the total mean firing rate minus the linear component ) , and a noise fluctuation for that trial ( Methods; S4D Fig ) ., We then constructed spatially linear and nonlinear datasets by keeping either the linear component and the noise or the nonlinear component and the noise , and we repeated our decoding procedure twice , using one or the other of these datasets ., Predicted EMG for the test trials of one C . V . repetition are compared with data for each muscle in Fig 4A ., Using the full data , the median correlation coefficient between the decoded and actual EMG signals across the 1 , 000 C . V . repetitions is 0 . 81 ( Fig 4B ) , and decoding errors are distributed around zero with a standard deviation of 13% of the min/max range of the EMG ( Fig 4C ) ., The decoding accuracy using only the linear component ( median CC = 0 . 73 , S . D . decoding error = 14%; Fig 4B and 4C ) is somewhat worse than decoding using the full data , and decoding using only the nonlinear components is slightly worse still ( median CC = 0 . 62 , S . D . decoding error = 16%; Fig 4B and 4C ) ., Chance level , as quantified by a shuffle control , resulted in a median correlation coefficient of 0 . 01 , and decoding errors with a standard deviation of 29% ( Fig 4B and 4C ) ., Thus , the EMG could be decoded far above chance level using either the linear or nonlinear datasets , and the nonlinear component contributed to the accuracy of the full decoding ., In a second approach , we decoded the EMG using increasingly complex nonlinear datasets by projecting the data onto increasing numbers of PCs of the neuronal responses ( as described previously ) ., As PCs representing nonlinear structure are added , the decoding accuracy improves , achieving the maximal ( median ) correlation coefficient with 7 PCs ( Fig 4D ) and the minimal ( standard deviation ) decoding error with 14 PCs ( Fig 4E ) ., These results confirm that nonlinear tuning contributed significantly to the information about muscle activity contained in the recorded M1 responses ., We now address the impact of forearm angle on hand-position tuning , starting by looking for two simple effects , an additive shift and a multiplicative gain change between the two forearm angles ., To examine a possible linear shift , we computed the baseline firing rate for each neuron , defined as the lowest firing rate across targets ., The distribution of baseline firing rate differences ( S . D . = 4 . 3 spike/s ) was significantly greater than what would be expected purely from noise ( bootstrap control; S . D . = 1 . 4 spikes/s; F-test for variances , p < 10−48 ) , indicating a significant additive shift between the firing rates for the two forearm angles ., To look for gain changes , we examined the range of the firing rates across all targets for the two forearm rotation angles ., Again , the distribution of gain differences ( S . D . = 8 . 7 spikes/s ) was significantly broader than for the bootstrap control ( S . D . = 3 . 3 spikes/s , F-test for variances , p < 10−31 ) , indicating a gain change ., If baseline shifts and gain changes were the only impact of forearm rotation , the correlation coefficient between the firing rates of any neuron across the two forearm angles would differ from 1 only due to noise effects ( Methods ) ., Instead , we found that the distribution of correlation coefficients between responses for pronation , and supination is broad ( Fig 5B , median = 0 . 54 ) and very different from the distribution expected solely from noise ( Fig 5C , median = 0 . 93 , bootstrap of the medians , p < 10−64 ) ., On the other hand , most of these correlation coefficients are positive , and their distribution is not at all like that for correlation coefficients between random pairs of responses ( Fig 5A ) ., Thus , the shapes of the hand-position tuning curves change as a function of forearm angle beyond a baseline shift and gain change , but not as much as if the tunings in these two cases were unrelated ., This suggests that the independent task variables ( the joint angles constrained by target position , and the forearm rotation angle ) are non-separable at the level of M1 , and thus they appear to have undergone nonlinear mixing ., In order to further understand the data , we searched for a minimalistic neural network model with biologically realistic constraints that could give rise to the empirical results ., The nonlinear component of the responses we have studied appears quite random , but the presence of a strong linear component would seem , at first sight , to rule out a model based purely on random target-related inputs to M1 neurons ., There is , however , a complication because physiological constraints impose a degree of smoothness on the tuning curves , and a random set of responses might exhibit a degree of regularity , including linearity , purely because of smoothness ., To address the degree to which smoothness constraints impose a linear component on responses to random inputs , we constructed a model in which M1 neurons are driven by inputs conveying information about target location ., In the model , target position is represented by a population of neurons that have Gaussian receptive fields ( one dimension of which is illustrated in Fig 6A ) centered on particular 3-dimensional preferred target locations ., This is consistent with the properties of parietal reach area neurons 9–11 , but this population could also correspond to neurons in premotor cortex ., The model input neurons have receptive fields with different preferred target locations and , at first , they have the same widths ., Later , we extend the model to include more realistic heterogeneity in input receptive field widths ., Each target generates a Gaussian “bump” of activity across the population of input neurons ., Each M1 neuron is connected to a random selection of 10% of the input neurons ., As a result , each M1 neuron received about 330 inputs that are active somewhere in the workspace and about 50 active inputs at each target location ., These inputs are multiplied by random synaptic weights ( Methods ) and the result is summed to produce the total input for each neuron ., The response of each model M1 neuron is then computed by passing its total input through a threshold-linear firing-rate function ., The threshold for this response function depends on a threshold parameter that , in this initial model , is set to the same value for all M1 neurons to match the mean coding level of the real neurons ( 0 . 85 ) ., The coding level is the fraction of conditions that cause a neuron to respond either at a level significantly different from 0 ( p < 0 . 01 , t-test , Bonferonni corrected ) or ≥ 5 spikes/s ., Once the coding level is set , this first-round model has only a single free parameter , the input tuning curve width ( our model fits are not sensitive to other features of the model , such as the total number of input-layer neurons , and the number and strength of synapses per M1 neuron ) ., This width determines the smoothness of the M1 responses and this , in turn , controls the tradeoff between spatial-linearity and complexity ( Fig 6B ) ., Wider input receptive fields lead to increased smoothness and spatial linearity , whereas narrower fields increase spatial complexity ., For a receptive field width corresponding to a visual angle of 12° , the model M1 responses have the same mean complexity measure as the real neurons ( Fig 3D ) ., Interestingly , this width is consistent with the values reported for parietal cortex neurons 9 , 12 ., Surprisingly , this simple random model does a good job of matching both the linear and nonlinear components of the real neurons , ( S5 Fig , compare to Fig 2 ) ., Even though we only used the mean complexity and mean coding level of the data to set the 2 parameters of the model ( the input tuning-curve width and the M1 threshold ) , the entire complexity measure distribution for the model responses is very similar to that of the real neurons ( Fig 6C ) ., In addition , the distribution of R2 values for the fit to the linear tuning curve is similarly broad ( Fig 6D ) ., Thus a feedforward model receiving input with structured tuning through random synapses can account for both the spatially linear and nonlinear components without any explicit tuning of its synaptic connections ( i . e . with purely random connectivity ) ., The 15% undershoot in the standard deviation of the linear tuning curve R2 distribution ( Fig 6D ) for the two-parameter model can be improved by extending the model slightly ., In the extended model , each M1 neuron is assigned a threshold to match the coding level of one of the real neurons ( randomly chosen ) , and the input receptive field widths are drawn independently from a uniform distribution parameterized by a mean and range ( Methods ) ., We chose the mean and range of the tuning curve width distribution to obtain the best match between the model and data for the mean and standard deviation of the linear tuning curve R2 distribution ( Fig 6D ) ., The neuronal responses of the enhanced model match the real data extremely well ( Fig 6C–6J ) ., The fraction of variance explained by the PCs of the model data matches the general shape seen for the real data , although the nonlinear components of the model are represented slightly more uniformly across its PCs ( Fig 6E ) ., This shows that the model responses capture approximately the same dimensionality as the real data ., In addition , the first 3 PCs of the model neurons capture the spatial linear component as they do for the real data ( Fig 6F ) ., We also checked whether the M1 model could reproduce the recorded EMG activity ., To do this , we added a set of linear readouts to the model ( Fig 6A , bottom ) and adjusted the weights of these readouts to fit the EMGs , using the decoding scheme used for the real neurons ( described above ) ., Relative to the small number of muscles , the neurons form an overcomplete basis ., As a result , using the noise-free model neurons decodes the EMGs almost perfectly ., For a fair comparison to the performance of the real data , we added noise samples to the model neurons , resampled from the real data ( Methods ) ., Decoding the EMGs using the model neurons with noise produced decoding correlation coefficients ( median CC = 0 . 82 for the model vs . 0 . 81 for the data; Fig 6I ) and decoding errors ( S . D . = 12% of EMG range vs . 13% , respectively; Fig 6J ) comparable to that of the real neurons ., These decoding results provide additional confirmation that the model M1 neurons capture the dimensionality and frequency content of the real neurons ., To model the effects of forearm rotation angle , we let the input neurons that drive the model M1 neurons depend on forearm angle as well as on target location ( Fig 6A , top right ) ., The forearm angle dependence involves multiplying the Gaussian function of target location for each neuron by a factor , chosen independently for each input neuron , that takes a different value for pronation or supination , with the values chosen randomly ( tuning of the forearm factor could have any shape for intermediate values , since these were not sampled in our task ) ., Thus , the forearm angle inputs act as a gain modulation of the visual inputs consistent with responses found in premotor cortex 13 and eye-position gain field modulation of visual responses in parietal cortex 12 , 14 ., With forearm angle included in this way , the correlations between the model neuronal responses for pronation and supination are extremely similar to those for the real neurons ( 98% similar in the mean and S . D . of the two distributions; Fig 6G ) , without any additional parameters or fitting ., The distributions of differences in gain across forearm angles are also very similar ( S . D . for the data = 8 . 7 versus for the model = 8 . 4 spikes/sec , Fig 6H ) ., These results are a consequence of the random nonlinear interaction of visual target and forearm angle signals at the input stage of the model , suggesting a mechanism that may underlie the non-separable mixing in M1 discussed above ., Cosine tuning has played a dominant role in discussion of the tuning of M1 responses during reaching movements 15–18 but , due to many conflicting findings , this approach has fallen out of vogue 19–23 ., During arm posture control , we found that M1 neuronal responses show a heterogeneity of irregular nonlinear responses that includes high spatial frequencies ( Fig 2 ) , in addition to the linear ( cosine-tuned ) component described previously 1–3 ., Due to this complex nonlinear component , it is unlikely that any reasonable coordinate choice could lead to a simple parametric description of these responses ., This is in contrast to the long-standing tradition of using parametric tuning curves and encoding models to describe neuronal responses ., Despite its irregularity , the nonlinear response component contributes to decoding the EMG of the major muscles used during arm posture control ( Fig 4 ) ., It is natural to assume that regular neuronal responses imply structured ( and in network models , learned ) input synaptic connectivity 24–26 ., However , we found that random connectivity could reproduce the data ( Fig 6C-6J ) , with regularity , in particular linearity , arising from smoothness ., This smoothness stems from assumptions about the nature of the inputs to M1 , assumptions that are consistent with the known physiology of the target-position coding regions 9 , 12 ., Importantly , this smoothness was not imposed by limits on the resolution provided by the task ( as an extreme example , a task with only two targets would unavoidably produce linear tuning ) or by the analyses we performed 27 , 28 ., Despite its simplicity and small number of adjustable parameters , the model accounts for not only the linear ( Fig 2A ) and nonlinear ( Fig 3D ) components , but also their distributions across the neuronal population ( Fig 6 ) ., This is a rare example in which both the regularity and the diversity of a population of neural responses have been replicated by a neural network model ., More generally , the model provides support for the idea that some neural circuits use random connectivity to generate rich and high-dimensional representations , and produce their outputs by tuning only their output synaptic weights 29–32 ., Firing rates were constant during posture control and feedback did not have a measurable effect on the neuronal responses ( S1 Text ) ., This implies that the population has either relaxed into fixed points or is dominated by the ( target position and intended forearm angle ) inputs external to M1 ., Either way , any internal recurrent inputs from other M1 neurons cannot generate the tuning to arm posture , which is why we focused on a feedforward network architecture ., Recurrent dynamics are likely to play an important role in how M1 generates time-varying movements , such as reaching 33 , 34 ., Such recurrent connections are unlikely to be random and may require synaptic modification 35 ., A number of influential theoretical models of M1 function have taken a normative engineering-inspired approach 36–40 ., By combining the equations of motion of a model arm with putative population M1 responses with features such as cosine tuning , these models can account for many psychophysical features of arm movements ., The normative approach has helped clarify the relationships between the arm’s biomechanics and resulting movement variables , and has highlighted the caution necessary in interpreting the correlations between movement variables and neuronal responses ., However , as has long been acknowledged 36 , due to the considerable dimensionality reduction from M1 to muscles , biomechanical properties do not generally suffice to constrain models of cortical computation ., The model we have presented here does not attempt to be a general model of M1 function , but it does incorporate representations in the inputs to M1 that are consistent with known physiology and neuronal properties that are biologically realistic ., We believe that understanding M1 computation will require models that integrate neural constraints , in addition to constraints imposed by the periphery ., The idea that sensory inputs are mapped to motor commands through a series of sequential coordinate transformations has received a lot of attention 41–45 ., In this view , sensory representations lead to kinematic representations and only finally to joint torques and muscle activity ., However , many daily actions require combinations of task demands , involving hand position , arm posture , endpoint and segmental forces ., Therefore , intended endpoint forces , for example , must sometimes be represented by the inputs to M1 and not only by its outputs ., In general , the sensorimotor system must be able to process parallel sensory and intentional signals , each ultimately shaping the activity of the same muscles ., Our experiment involved demands on both hand position and forearm rotation angle , a degree of freedom that does not affect hand position ., The neural responses show an interaction between these two independent inputs that is non-separable ( i . e . cannot be decomposed into a sum or product of functions; Fig 5 ) ., This suggests that these signals are already mixed nonlinearly before the level of M1 ., Most encoding models introduced in the M1 literature , whether linear 46–48 or nonlinear 49 , 50 , produce separable functions ., In this view , M1 responses are an example of nonlinear mixed-selectivity 32 , representations that have been shown to have considerable computational power in tasks requiring nonlinear combinations of multiple parameters 51 ., We incorporated this finding into our model by basing the inputs to the model M1 neurons on the multimodal visual and arm-posture-dependent receptive fields found in premotor cortex 13 , 52–56 and posterior parietal cortex 57–62 ., The model places the idea that such multimodal activity reflects the multiplexing of several sensorimotor parameters 63 , 64 within the context of a class of basis-function network models 65–67 ., In this view , the multimodal responses found in parietal and premotor cortex may be evidence for the nonlinear mixing of parallel task variables , in addition to sequential sensorimotor transformations ., Because of this basic feature of its construction , our model predicts that during experiments with additional concurrent task demands , M1 responses should show non-separable interactions across all task dimensions ., This can be tested experimentally , by , for example , repeating our experiment with two different load conditions at the hand , in addition to varying hand position , and forearm rotation ., Posture control could , in principal , be implemented by a simple “on-off” cortical signal that causes ongoing spinal feedback to maintain current muscle activation levels ., Even a transiently tuned cortical motor command could activate ongoing spinal control 68 ., In contrast to these possibilities , we found that most M1 neurons are tuned even 1–1 . 8 seconds after a target is reached and posture is held constant ., By using a powerful decoding algorithm , some of the hand jitter during the target hold epoch could be predicted 69 ., However , having found no correlations between firing rates and hand jitter suggests that such signals are weak and that the ongoing M1 responses primarily represents a feedforward ( intentional ) motor command , with ongoing corrections executed by sub-cortical feedback loops ., This is in contrast to the strong and fast feedback corrections seen in M1 in response to large unexpected per | Introduction, Results, Discussion, Methods | Neuronal responses characterized by regular tuning curves are typically assumed to arise from structured synaptic connectivity ., However , many responses exhibit both regular and irregular components ., To address the relationship between tuning curve properties and underlying circuitry , we analyzed neuronal activity recorded from primary motor cortex ( M1 ) of monkeys performing a 3D arm posture control task and compared the results with a neural network model ., Posture control is well suited for examining M1 neuronal tuning because it avoids the dynamic complexity of time-varying movements ., As a function of hand position , the neuronal responses have a linear component , as has previously been described , as well as heterogeneous and highly irregular nonlinearities ., These nonlinear components involve high spatial frequencies and therefore do not support explicit encoding of movement parameters ., Yet both the linear and nonlinear components contribute to the decoding of EMG of major muscles used in the task ., Remarkably , despite the presence of a strong linear component , a feedforward neural network model with entirely random connectivity can replicate the data , including both the mean and distributions of the linear and nonlinear components as well as several other features of the neuronal responses ., This result shows that smoothness provided by the regularity in the inputs to M1 can impose apparent structure on neural responses , in this case a strong linear ( also known as cosine ) tuning component , even in the absence of ordered synaptic connectivity . | Relationships between the activity of single neurons and experimental parameters are often characterized by functions called tuning curves ., Regular tuning-curve shapes are typically assumed to arise from structure in the synaptic inputs to each neuron ., We found that the activities of neurons in primary motor cortex during an arm posture task exhibit both a regular component that fits a well-known tuning curve description , and heterogeneous irregular components that do not ., Such complex components are often assumed to reflect residual noise ., However , both the regular and irregular components are needed to optimally decode the commands that guide the muscles used in the task ., We then asked what type of input structure was needed to generate neuronal responses with both regular and irregular elements ., We constructed and analyzed a mathematical model , based on known physiology of the relevant brain regions that replicates the full spectrum of recorded neuronal responses ., Surprisingly , the synaptic connectivity in this model is completely random . | medicine and health sciences, limbs (anatomy), vertebrates, social sciences, neuroscience, animals, mammals, primates, bioassays and physiological analysis, muscle electrophysiology, vision, forearms, neuronal tuning, research and analysis methods, curve fitting, musculoskeletal system, monkeys, mathematical functions, animal cells, hands, mathematical and statistical techniques, electrophysiological techniques, arms, cellular neuroscience, psychology, cell biology, anatomy, electromyography, neurons, biology and life sciences, cellular types, sensory perception, amniotes, organisms | null |
journal.pcbi.1000110 | 2,008 | The Ascent of the Abundant: How Mutational Networks Constrain Evolution | Despite its familiar slogan—“survival of the fittest”— evolution by natural selection may not always yield optimal organisms ., In particular , it will be fundamentally constrained by the variation introduced into populations by mutation or migration ., If better traits never arise , then natural selection will never have the opportunity to favor them ., Whereas adaptive constraints are central to evolutionary theory 1–3 , there have been relatively few empirical characterizations of them 4–8 ., Several of these studies suggest that selection can overcome putative constraints 6–7 ., Yet , one study of the enzyme beta-isopropylmalate dehydrogenase ( IMDH ) concludes that adaptation is constrained by its spectrum of mutations 8 ., With the introduction of the fitness landscape metaphor , Sewell Wright was one of the first to argue for the importance of adaptive constraints 9 ., In contrast to Fishers panselectionist views 10 , Wright suggested that fitness valleys—low-fitness genotypes separating high-fitness genotypes—may preclude simple incremental evolution 9 ., He argued that adaptation depends on both the structure of the fitness landscape ( that is , the spectrum of possible mutations ) and demographic conditions ., Since the 1930s , the theory of evolutionary constraints has matured , but is largely premised on hypothetical fitness landscapes or very local estimates of mutational effects 11 , 12 ., For most phenotypes of interest , we cannot yet model complete fitness landscapes ., It requires knowing the fitnesses across large sets of genotypes , typically too vast to exhaustively study either empirically or computationally ., There are , however , a few biologically important phenotypes for which this is tractable ., In particular , Eigen and Schuster pioneered the study of RNA molecules , using RNA secondary-structure folding algorithms as tractable genotype-to-phenotype maps 12 , 13 ., In their model , the genotype of a molecule is its primary sequence and the phenotype is its predicted minimum free energy secondary structure; fitness is based entirely on the similarity of a phenotype to an ideal target structure ., Through extensive sampling ( that is , folding many diverse sequences ) and evolutionary simulations , this system has motivated and clarified several important ideas in modern evolutionary theory , including error catastrophes , quasispecies , neutral networks , and punctuated equilibria 14–23 ., The most influential concept to emerge from these RNA studies is that of “neutral networks” , which are sets of genotypes with identical fitness that are interconnected by neutral mutations 15 ., In the RNA model , the genotypes in a neutral network are sequences that fold into the same shape and are connected to each other by paths of neutral point mutations ., The neutral networks of RNA and protein molecules appear to share three basic characteristics:, ( i ) most neutral networks are small ( contain few genotypes ) , whereas relatively few are large ( contain many genotypes ) ;, ( ii ) large neutral networks are mutationally adjacent to a greater diversity of phenotypes than small neutral networks; and, ( iii ) large neutral networks span the entire sequence space 15 , 24–26 ., Based on these characteristics , researchers have proposed that large neutral networks should facilitate evolution by allowing populations to explore vast regions of regions of fitness landscapes through neutral drift 15 , 18 , 24 , 26 , 27 ., There is some evidence to support this assertion , though it is largely based on sampling studies 15 , 24 , 26 or simulation studies with strong assumptions about the fitness landscape 26 ., Most recently , Wagner ( 2008 ) showed that populations evolving on large neutral networks sample more alternative phenotypes than those evolving on small neutral networks , yet these populations were constrained to explore a single neutral network ., Whether large neutral networks actually facilitate the evolution of optimal phenotypes fundamentally depends on the global structure of mutational connections between different neutral networks ., If large neutral networks are almost exclusively connected to other large neutral networks , then populations will easily move among common phenotypes , but be unable to evolve rare phenotypes ., Theoretical and computational characterizations of RNA fitness landscapes suggest that this may , in fact , be the case ., Yet , these predictions are largely based on relatively small samples of sequences which may include only the most common phenotypes in the fitness landscape 15 , 24 ., Here , we use the RNA folding model to determine the complete structure of fitness landscapes and how neutral network size and adjacencies constrain evolutionary dynamics ( for better or for worse ) ., Specifically , we fold all RNA molecules of lengths 12 to 18 nucleotides , and then develop a network model describing the patterns of mutational connectivity among the phenotypes produced by molecules of the same length ., We build on previous characterizations of RNA neutral network structure 15 , 25 , 28 , 29 , and argue that the mutational connectivity among phenotypes follows simple predictable patterns that fundamentally constrain evolution ., RNA molecules fold into secondary structures that are the essential scaffolds for functional tertiary structures and are evolutionarily conserved for most functional RNA molecules 30 ., The formation of secondary structures is relatively well understood and can be rapidly predicted using thermodynamic minimization 31–34 ., We used the Vienna RNA folding software version 1 . 6 . 1 with the default parameter set; 33 to predict the lowest free energy shapes of all RNA molecules of lengths 12–18 nucleotides ., We assume that the shape of a molecule is a reasonable proxy for its fitness 19 , 21 , 23 and refer to each map from sequences of length n to their predicted shapes as an n-mer fitness landscape ., We studied evolutionary dynamics on the 12-mer fitness landscape by computationally simulating a population of evolving RNA molecules ., The molecules stochastically replicate at each discrete generation in proportion to their fitnesses , and evolve by point mutations ., We and others have used similar models to study many aspects of RNA evolutionary dynamics 18–23 , 35 ., An important feature of the RNA system is that the fitness effect of a point mutation stems from a biologically explicit model of molecular structure and is not simply selected from a probability distribution of mutational effects , as in simpler evolutionary models ., To compute the fitness of a molecule , we first predict its minimum free energy secondary structure ( that is , its groundstate ) , and then compare this predicted structure with a pre-specified target structure ., Specifically , if σ is the groundstate of a molecule m and t is the target structure , then the fitness of the molecule W is given by ( 1 ) where α\u200a=\u200a0 . 01 and β\u200a=\u200a1 are scaling constants , d ( σ , t ) is the Hamming distance between the parenthetical representations of σ and t , ( parenthetical notation represents paired bases with pairs of parentheses and unpaired bases with dots ( e . g . , ( ( ( . . . . ) ) ) is a simple stem-loop structure ) and L\u200a=\u200a12 is the length of the sequence ., The range of fitness values possible given our choice of parameters is 0 . 99 - 100 . 0; except the open-chain shape , which was assigned a fitness of zero ., Several other studies using this computational model have shown that the qualitative results are largely insensitive to the choice of parameters and even the shape of the fitness function 18–21 , 23 ., For every starting structure-target structure combination , we adapted 20 replicate populations for τ\u200a=\u200a1 , 000 , 000 generations ., The population size was held fixed at N\u200a=\u200a1000 , which was chosen both for computational tractability and to limit the effects of genetic drift ., The genomic mutation rate was maintained at U\u200a=\u200a0 . 0003 ( NU\u200a=\u200a0 . 3 ) for all bases in the RNA alphabet ., We used soft-selection ( constant N ) to maintain the population size when genotypes that fold into the open-chain shape occasionally appear ., The expansive and intertwining neutral networks smooth the fitness landscape so that virtually every phenotype can mutate to at least one fitter phenotype , except , of course , the optimal ( target ) phenotype ., Yet the likelihood of finding a more fit mutation while drifting on a large neutral network may be exceedingly small ., Specifically , 96 . 7% of all neutral networks have at least one beneficial mutation ( across all fitness functions considered in this study ) , and there always exists a path of beneficial and neutral mutations leading to the target phenotype ., In our simulations , the average time to target was 339111 . 7 generations; and there is no significant correlation between time to target and the abundance of the target ., The simulations were allowed to run for approximately three times longer than the typical time to acquire the target , and 100 times longer than the evolutionary simulations reported in other studies using this system 18–21 , 23 ., Two sets of simulations with different parameter sets ( N\u200a=\u200a500 , U\u200a=\u200a0 . 05 , τ\u200a=\u200a5 , 000; N\u200a=\u200a1000 , U\u200a=\u200a0 . 005 , τ\u200a=\u200a250 , 000 ) produced similar results to those reported here ( not shown ) ., The parameters were selected to be biologically reasonable and do not appear to strongly affect the outcome ., Although even the most unlikely phenotype can evolve given infinite time , we believe that our results reflect the likely course of evolution ., Rfam is a curated database of functional RNA genes , which are those genes in which the RNA molecule itself takes parts in a biological reaction 36 ., Here , we used version 7 ( 2006 ) of the database ., We restricted our analysis to families in which the predicted shape of each sequence in the family was at least 60% identical to the consensus structure , thereby minimizing the effects of folding inaccuracies ., This included 239 Rfam families ( about 50% of the entire database ) with representatives of every functional class in the database ., Abundance estimates were obtained by calculating contiguity statistics for the secondary structures predicted by thermodynamic minimization of each sequence in a family ., We then determined the rank percentiles of these abundance estimates in a null distribution of abundance estimates from random sequences ., To generate the null distributions , we randomized each sequence in a family 500 times ( preserving nucleotide composition ) , and then calculated the contiguity statistics of the ground-state shapes of these random molecules ., We finally determined the fraction of contiguity statistics in the null distributions that were less than the contiguity statistic from the naturally occurring molecule ( Figure 1 ) ., Receiver operating curve ( ROC ) analysis is a technique for assessing the performance of classifier models 37 ., The area under an ROC gives the probability that a model correctly assigns a binary variable ( in this case , natural or random molecule ) to its proper group ., We used ROC analysis to assess relative accuracies of thermostability and contiguity for classifying sequences as natural ( taken from the Rfam database ) or random , under the assumption that natural molecules will have higher contiguity and thermostability than random permutations of those molecules ., Specifically , we performed logistic regressions of molecule class ( natural or random permutation ) on contiguity statistic and thermostability , and compute the area ( A ) under the ROC as:where P and N are the numbers of positive and negative instances in the data set , TP and FP are the counts of true positive and false positive classifications between indices i and j ., We used the ROCR package to perform all such calculations in R 2 . 5 . 0 38 ., We have predicted the groundstate structures of all RNA molecules of lengths 12 through 18 nucleotides; we refer to length n RNA molecules as n-mers ., The map from sequences to shapes is extremely degenerate with large numbers of sequences ( genotypes ) giving rise to identical shapes ( phenotypes ) , as previously observed 15–17 ., We found that the number of unique phenotypes approximately doubles with each single-base addition , from 59 unique 12-mer shapes to 3211 unique 18-mer shapes ., Some of these shapes are quite common , with many unique genotypes folding into them , while others are quite rare , formed by few unique genotypes ., We define abundance as the number of genotypes that produce a particular phenotype ., The distributions of phenotype abundances appear similar across all lengths of molecules ( roughly exponential without the 10% of extreme values in each tail ) , with relatively few highly abundant phenotypes and many rare ones Figure 2 ., This is qualitatively similar to the distributions reported previously for both protein and larger RNA molecules 15–17 , 29 ., Figure 2 shows a portion of the abundance distribution and a sample of shapes present in the 12-mer fitness landscape ., For the 12-mer to 16-mer sequence lengths , the landscapes are composed entirely of variations on stem-loop-structures ., In the 17- and 18-mer landscapes , we observe the emergence of sequences folding into multi-loop shapes , albeit at very low frequencies ( on the order of 0 . 001% of all sequences ) ., The relatively low structural diversity is consistent with known constraints on RNA structural motifs , for example , loops must contain at least three nucleotides 31 , 33 ., A set of genotypes that shares a common phenotype is called the neutral network of that phenotype ( Figure 3 ) 15 ., Neutral networks may be composed of one or more components ., Within any component , all genotypes are connected to each other by a sequence of point mutations that remain within the component; these mutations are , by definition , neutral ., For example , in the bottom network of Figure 3B , the red phenotype has a neutral network with two components , each of which consists of a set of red nodes interconnected by red edges ., The abundance of a phenotype is precisely the size of its neutral network ., Counterintuitively , there is only a weak positive relationship between the abundance of a phenotype and the number of distinct components in its neutral network ( r2\u200a=\u200a0 . 11 , P≈0 . 01 ) ., The majority of the 12-mer RNA neutral networks are dominated by relatively few large components , which each contain approximately 8–10% of the sequences in the neutral network; together these large components account for at least 80% of the neutral network ., Importantly , the large components share many of the same characteristics as the entire neutral network ., In particular , they are each mutationally connected to the majority of the shapes that are adjacent to the entire neutral network ( typically >75% ) ., Figure 2 also reports the number of components ( Nc ) , the maximum Hamming distance between a pair of sequences in a single component ( Dmax ) , and the maximum shortest path length between a pair of sequences in a single component ( Dspl ) for the neutral networks in the 12-mer landscape ., The neutral networks for the most abundant phenotypes percolate through the entire space of genotypes ., The various phenotypes within a fitness landscape are connected to each other by mutations ., If we aggregate all genotypes into their respective neutral networks , we create a mutational network in which each vertex represents a distinct phenotype and edges connect pairs of vertices when there is at least one point mutation that converts one phenotype to the other ( Figure 3 ) ., For example , consider a two-locus , two-allele , haploid model with genotypes AB , Ab , aB , and ab ( Figure 3A ) ., There are three unique phenotypes–the two ( A- ) genotypes produce one phenotype ( blue ) , aB produces another phenotype ( green ) , and ab produces a third phenotype ( purple ) ., Mutational networks , in turn , form the underpinnings for fitness landscapes , which depend on the map from phenotype to fitness ., Figure 3B caricatures a higher dimensional genotype network and its projections to phenotype and fitness networks ., For RNA molecules , the vertices in a mutational network represent unique shapes and the edges represent point mutations that cause a molecule to fold into a new shape ., Roughly speaking , evolution by natural selection moves populations along the edges in a mutational network from one phenotype vertex to another ., We are therefore interested in how the structure of mutational networks influences evolutionary dynamics ., Intuitively , the structure of a mutational network may influence, ( i ) the likelihood that a given phenotype will arise and , ( ii ) if it arises , the likelihood that the population can further evolve other , better phenotypes ., Hereafter , we use accessibility to refer to the likelihood that a phenotype will arise , and evolvability as the likelihood that a phenotype can further evolve other , better phenotypes ., The most straightforward measure of a phenotypes mutational connectivity is its degree in the mutational network , that is , the number of other phenotype that can be reached by a single mutation ., For the 12-mer through 18-mer RNA molecules , there are significant positive correlations between phenotype abundance and degree R\u200a=\u200a0 . 88 ( 12-mer ) to R\u200a=\u200a0 . 91 ( 18-mer ) ; P<2×10−16 ., This has been observed previously and suggests that abundant phenotypes should be both more evolvable and more accessible than rare phenotypes 24 , 26 , 27 ., The degree of a phenotype is , however , a crude indicator of its mutational connectivity to other phenotypes ., It does not reflect the probability that a mutation will actually yield a new phenotype; this probability typically declines as the size of the neutral network increases ., Furthermore , the degree does not quantify whether the non-neutral mutations off a neutral network are evenly divided among the set alternative phenotypes , or are biased towards a select few of these phenotypes ., We therefore developed two novel statistics , which provide a more nuanced perspective on mutational connectivity ., Both of these statistics use the quantity , where νij is the number of point mutations to genotypes in the neutral network for phenotype i that create a genotype in the neutral network for phenotype j , and is the total number of non-neutral point mutations to genotypes in the neutral network for phenotype, i . Thus , fij is the fraction of non-neutral point mutations to genotypes in the neutral network for phenotype i that create genotypes in the neutral network for phenotype, j . Large values of this fraction indicate that phenotype j is relatively easy to find ( via random mutations ) from phenotype, i . Mutational proximity is often not symmetric ( that is , fij≠fji ) , because the denominators differ ., The first statistic estimates the overall accessibility of phenotype i from other phenotypes in the landscape using ., Large values of Ai indicate that phenotype i is relatively accessible from throughout the landscape ., The second statistic quantifies the potential for evolution away from phenotype i using a variation on Simpsons diversity index: ., This index indicates the diversity of other phenotypes that can be easily produced by mutations from a given phenotype , and thus may indicate the potential for further adaptation away from that phenotype ., Specifically , it gives the probability that two randomly chosen non-neutral mutations to genotypes within a given neutral network will result in the same phenotype ., The index is large for phenotypes that are adjacent to many other phenotypes , and its non-neutral mutations are fairly evenly divided among the adjacent phenotypes; it is small for phenotypes that primarily mutate to one or very few alternate phenotypes ., In the 12-mer landscape , A increases significantly with the abundance of a phenotype ( Figure 4 , top pane ) ., In other words , random mutations are more likely to move genotypes to a large neutral network than to a small neutral network ., In contrast , E decays significantly with phenotype abundance ( Figure 4 , middle pane ) , suggesting that it may be more difficult to evolve away from large neutral networks than small neutral networks ., To provide more insight into the mutational networks , we also calculated the average abundance of phenotypes reached by mutation from phenotype i using ., We find that the average abundance of neighboring phenotypes significantly increases with the abundance of a phenotype ( Figure 4 , bottom pane ) , meaning that the majority of non-neutral mutations to abundant phenotypes produce other abundant phenotypes ., Thus far we have characterized the mutational networks formed by single point mutations ., If we instead considered the mutational networks formed by all combinations of one , two or three mutations , then the phenotype network becomes highly interconnected ., The number of adjacent phenotypes significantly increases with multiplicity of mutations considered ( mean node degrees are 42 . 7 , 53 . 6 , and 57 . 2 for the one , two , and three mutant adjacencies , respectively; P<5×10−3 ) , and the network is nearly completely connected for triple mutations ., Thus , under elevated mutation rates , populations may be able to attain rare phenotypes easier than expected based on point mutation adjacencies ., In summary , these observations suggest that abundant phenotypes may be easy to find but difficult to escape , and thus the structure of a fitness landscape may significantly constrain evolutionary dynamics ., Whereas the accessibility of abundant shapes is rather intuitive , the prediction that their vast neutral networks can hinder further evolution contradicts a large body of theory , which suggests that large neutral networks should enhance evolvability 18 , 26 , 27 ., We note that this evolutionary constraint was previously proposed for a simple fitness landscape model 39 ., To test the hypothesis that highly abundant phenotypes are readily accessible , yet poorly poised for further evolution , we ran stochastic simulations of an adapting population of 12-mer RNA molecules using an established model ( see Materials and Methods for details ) 18–21 , 23 ., Since we are interested in the effect of phenotype abundance on the capacity of selection to acquire the optimal phenotype , we selected the phenotypes of the founding populations ( henceforth , founding phenotypes ) and target shapes to span the range of abundances found among the 12-mer phenotypes ., We chose ten founding phenotypes ranks ( abundance ) : 3 ( 183 , 791 ) , 8 ( 117 , 213 ) , 13 ( 76 , 478 ) , 18 ( 61 , 699 ) , 23 ( 39 , 740 ) , 28 ( 27 , 312 ) , 33 ( 11 , 354 ) , 38 ( 2 , 260 ) , 43 ( 1 , 299 ) , 48 ( 713 ) and randomly selected 20 genotypes from the neutral network of each founding phenotype to form 200 isogenic founding populations ., Each founding population was composed of a single genotype and , therefore , a single phenotype ., In essence , we simulated adaptation starting from 20 random points in the neutral network of each founding phenotype ., We separately adapted each founding population to twelve target phenotypes ranks ( abundance ) : 2 ( 218 , 576 ) , 7 ( 122 , 332 ) , 12 ( 93 , 866 ) , 17 ( 61 , 895 ) , 22 ( 41 , 092 ) , 27 ( 27 , 522 ) , 32 ( 15 , 348 ) , 37 ( 2 , 963 ) , 42 ( 1 , 368 ) , 47 ( 800 ) , 52 ( 240 ) , 57 ( 109 ) ., We considered adaptation successful if the population ever acquired the target phenotype , regardless of its frequency in the population ., In the successful runs , however , the target phenotype quickly dominates the populations and rises to frequencies of nearly N ( the population size ) ., The mutational connectivity statistics described above ( Ai and Ei ) will only be good indicators of evolutionary dynamics if the probability of mutating from phenotype i to phenotype j correlates with the fraction of mutations to i that produce j ( fij ) ., To test this basic assumption , we compared the phenotype mutation rates observed in the simulations ( fraction of mutations to i that produce, j ) to fij ( the fraction of non-neutral point mutations to genotypes in the neutral network for phenotype i that create genotypes in the neutral network for phenotype, j ) ., In fact , we find an almost perfect relationship between the two quantities ( Figure 5A ) , suggesting that mutational network structure fundamentally constrains evolution and that Ai and Ei are good indicators of these constraints ., Across the 2400 simulations , we observed a significant positive correlation between the abundance of the target phenotype and the likelihood that a population successfully evolved to the target ( Figure 6A ) ., This is consistent with the positive relationship between phenotype abundance and mutational accessibility , as indicated by the A statistic ( Figure 4A ) ., Phenotype abundance also positively correlates with the number of times a phenotype arises in the evolving populations ( Figure 7A ) ., Taken together , these results support our hypothesis that abundant shapes are more likely to appear via mutation in evolving populations than are rare shapes ., We did not , however , observe a relationship between the founding phenotype abundance and the ultimate evolutionary outcome ( Figure 6B ) ., When a simulation failed to acquire the target , the population was primarily composed of phenotypes of greater abundance than both the target phenotype and the average abundance of a random phenotype , demonstrating that the structure of mutational networks can steer populations towards abundant , but non-optimal , phenotypes ., As suggested by the negative relationship between abundance and the E statistic , evolution away from abundant phenotypes appears to be limited by the improbability of beneficial mutations ., In support of this explanation , we also find a significant positive correlation between the abundance of a phenotype and the duration of the phenotype in the evolving populations ( Figure 7B ) ., These observations appear to be inconsistent with the widely-held belief that neutral networks facilitate evolution by allowing populations to traverse large regions of fitness landscapes without reducing fitness 15 , 18–20 , 26 , 27 , 40 ., In our simulations , populations readily evolve from one abundant shape to another ( that is , from one large neutral network to another ) , but are often unable to evolve rare phenotypes ., Thus , while the hypothesis that neutrality ( the fraction of mutations that are neutral ) allows populations to explore phenotype space is true , the evolutionary outcome of such exploration is generally confined to other abundant phenotypes ., Most of the prior studies addressing this hypothesis are based on relatively small random samples of sequences from large genotype spaces , which may consist of exclusively abundant phenotypes ., The conclusion that neutrality facilitates evolution is reasonable when considering only abundant subsets of fitness landscapes , but is somewhat misleading when one considers the fitness landscapes in their entirety ., These results suggest the following hypothesis: the evolution of phenotypes , whether complex whole-organism phenotypes or RNA shapes , may be biased toward abundant phenotypes , even if those phenotypes are not optimal ., We cannot , however , test this hypothesis by directly measuring the abundances of complex organism-level phenotypes since we cannot yet completely characterize their fitness landscapes ., As a first step in this direction , we have developed a simple structural statistic that allows us to indirectly estimate the abundances of naturally occurring RNA shapes , which are much larger and more complex than those considered thus far ., Across the n-mer phenotypes , we observed that longer contiguous helical stacks ( stems ) form more frequently than shorter contiguous stacks and stacks that contain bulges ( which break up helices ) ., We quantify this with a new statistic ( Figure 8 ) given by This contiguity statistic significantly correlates with log phenotype abundance in the 12- through 18-mer landscapes r ranges from r\u200a=\u200a0 . 71 ( P\u200a=\u200a3 . 6×10−10 ) in the 12-mer landscape to r\u200a=\u200a0 . 69 ( P<2 . 2×10−16 ) in the 18-mer landscape ., The utility of the contiguity statistic is that one genotype is sufficient to estimate the abundance of its phenotype ., We conjecture , therefore , that we can use the contiguity statistic to ask whether naturally occurring RNA molecules are biased towards abundant shapes ., We used the contiguity statistic to estimate the abundances of the RNA molecules in Rfam , a curated database of functional RNA genes 36 ., The Rfam molecules are grouped into families , and every sequence in a family is thought to code for the same functional RNA ., We compared the contiguity statistics calculated for the Rfam sequences to null distributions generated by calculating contiguity statistics for thousands of random permutations of those sequences ., Specifically , for each naturally evolved molecule , we determined whether the contiguity statistics of their predicted shapes were significantly larger than the contiguity statistics of random molecules from the same fitness landscape ( see Methods for details ) ., The structures of the natural RNA molecules indeed have larger contiguity statistics than randomly chosen structures from the same fitness landscapes ( Figure 1 ) ., This observation supports an “ascent of the abundant” hypothesis in which the mutational networks connecting diverse phenotypes may steer populations toward abundant , though not necessarily optimal , phenotypes ., Yet , Figure 1 ( red squares ) shows that natural molecules are also significantly more thermostable than random molecules ., Thus one must ask whether the high contiguity values of natural molecules are simply byproducts of the evolution of thermostability ( or some other advantageous structural property ) or , in fact , exist because of mutational biases towards abundant shapes , or both ., The abundances of the natural molecules ( as estimated by their contiguity statistics ) are even more statistically pronounced than their thermostabilities ., We used logistic regression analysis to ask which of contiguity or thermostability better distinguishes naturally occurring molecules from their random permutations ., We regressed molecule class ( natural or random permutation ) on contiguity statistic and ( separately ) on thermostability ., The area under a receiver operating curve ( ROC ) gives the probability that a model correctly assigns a binary variable ( natural or random molecule ) to its proper group ., The logistic model for contiguity yielded an area under the ROC of 0 . 82 , which is good; the model for thermodynamic stability yielded an area under the ROC of 0 . 62 , which is poor ., Our results are therefore consistent with an apparent biases towards abundant phenotypes in both the small RNA landscapes and natural RNAs are not simply byproducts of natural selection for thermostability ., Evolutionary biologists have long appreciated that the evolutionary potential of a phenotype depends on the breadth of its neutral network ., Eigens error catastrophe theory , an extension of classic mutation-selection balance theory , argues that the evolutionary potential of a phenotype depends on both its fitness relative to alternative phenotypes and its robustness to mutations 41 ., Under high mutation rates , only phenotypes with sufficiently large and connected neutral networks can persist ., The phrase “survival of flattest” has been used to refer to the evolutionary s | Introduction, Materials and Methods, Results, Discussion | Evolution by natural selection is fundamentally shaped by the fitness landscapes in which it occurs ., Yet fitness landscapes are vast and complex , and thus we know relatively little about the long-range constraints they impose on evolutionary dynamics ., Here , we exhaustively survey the structural landscapes of RNA molecules of lengths 12 to 18 nucleotides , and develop a network model to describe the relationship between sequence and structure ., We find that phenotype abundance—the number of genotypes producing a particular phenotype—varies in a predictable manner and critically influences evolutionary dynamics ., A study of naturally occurring functional RNA molecules using a new structural statistic suggests that these molecules are biased toward abundant phenotypes ., This supports an “ascent of the abundant” hypothesis , in which evolution yields abundant phenotypes even when they are not the most fit . | Evolutionary biology tells us much about the immediate fate of a mutation once it appears , but relatively little about its long-term evolutionary implications ., Major evolutionary transitions from one trait to another may depend on a long sequence of interacting mutations , each arising by chance and surviving natural selection ., In this study , we characterize the network of mutations that connect diverse molecular structures , and find that this network biases evolution toward traits that are readily produced by one or a short sequence of mutations ., This bias may prevent the evolution of optimal traits , a phenomenon they call the “ascent of the abundant . ” | computational biology/evolutionary modeling, evolutionary biology, evolutionary biology/bioinformatics | null |
journal.ppat.1003864 | 2,014 | IFNγ/IL-10 Co-producing Cells Dominate the CD4 Response to Malaria in Highly Exposed Children | Clinical immunity to malaria eventually develops in endemic populations , but only after repeated infections with significant morbidity to both individuals and their communities 1 ., Studies in regions of high malaria transmission intensity have consistently shown that the incidence of severe disease decreases considerably after the first years of life , but sterile immunity ( i . e . protection against parasitemia ) develops rarely if ever 2 , 3 ., Moreover , previously immune individuals may lose protection against symptomatic infection in the absence of continuous exposure 4 , 5 ., The reasons underlying the slow acquisition of clinical immunity and the failure to develop sterilizing immunity are unclear , but may include parasite diversity and evasion 6 , age-related differences in immune responses 7–12 , and/or host immunoregulatory mechanisms induced by the parasite 13–19 ., As the incidence of malaria continues to be high in many parts of Africa despite insecticide-treated bednets and artemisinin-based combination therapy 20–22 , there is a tremendous need to better understand mechanisms of immunity to malaria in naturally exposed populations ., The identification of immunologic correlates of exposure and protection in naturally exposed children would significantly help with the rational design of vaccines and other malaria control interventions ., Both CD4+ and CD8+ T cells have been demonstrated to play an important role in protective antimalarial immunity in mouse models 23–30 , and experimental challenge models in humans and mice strongly suggest that malaria-specific T cells contribute to protective immunity 31–36 ., However , the identification of T cell correlates of immunity in field-based studies of naturally exposed humans has proven to be quite challenging ., Prior studies employing cross-sectional or prospective cohort designs have found associations between cellular immune responses and protection from future malaria , including IFNγ responses to liver stage 37–40 and/or merozoite stage malaria antigens 41–44 ., However , such studies may be confounded by the level of exposure to malaria-infected mosquitoes , which varies greatly within populations , leading subjects with lower exposure to be miscategorized as “protected” 45 , 46 ., Because naturally acquired immunity confers relative rather than absolute protection – manifested by a gradual decline in the incidence of clinical disease - careful quantitative outcome measures are essential , but few population-based studies of natural immunity have included careful measurement of malaria incidence over time ., Pathogen-specific T cells exhibit notable functional heterogeneity , largely dependent on the antigen and cytokine microenvironment encountered during activation , and measurement of a single parameter of T cell function ( i . e . IFNγ production ) may overlook others that are more critical for protection 47 ., In other parasitic infections such as leishmania 48 , 49 and toxoplasma 50 , the functional phenotype of the CD4+ T cell response correlates with the success or failure to clear the pathogen ., Recent observations in individuals naturally exposed to malaria suggest an important role for CD4+ T cell production of TNFα , with or without IFNγ , as a potential immunologic correlate of protection 51 ., Conversely , CD4+ T cell production of the regulatory cytokine IL-10 has been implicated in modulating the severity of disease 18 , 52 and may interfere with the development of protective immunity 14 , 42 , 53 ., The role of these inflammatory and regulatory cytokines in mediating protective immunity in naturally exposed children , and in determining the balance between immunopathology and chronic repeated infection , remains unknown ., In this study we performed a detailed functional characterization of malaria-specific T cell responses among four-year-old children residing in a highly malaria-endemic region to determine whether naturally acquired T cell responses correlate with exposure to and/or protection from malaria ., We hypothesized that CD4+ T cells producing the pro-inflammatory cytokines IFNγ and/or TNFα are associated with protection from malaria , and that T cell production of the regulatory cytokine IL-10 may interfere with the acquisition of protection ., Our results suggest that the functional phenotype of the malaria-specific T cell response was heavily influenced by prior malaria exposure intensity , with CD4+ T cells co-producing IFNγ and IL10 dominating this response among highly exposed children ., However , these IFNγ/IL-10 co-producing cells were not independently associated with protection from future malaria , and may be associated with increased risk ., The study cohort consisted of 78 HIV-uninfected children followed from infancy through 5 years of age ( Table 1 ) ., Blood for this study was drawn at four years of age ( range 49–51 months ) , and 92% of children continued to be followed through 5 years of age ., A total of 1855 incident cases of malaria were observed in this cohort through 5 years of age ., All children were treated promptly with artemisinin-based combination therapy , and despite the strikingly high numbers of malaria episodes , only 4 cases of malaria were deemed “complicated” ( all based on a single convulsion ) ., No cases of severe malaria ( including severe anemia ) were observed ., Among children with a lower prior incidence of malaria ( <2 episodes per person year ( ppy ) between 1 and 4 years of age , n\u200a=\u200a10 ) , 90% lived in town; whereas among children with higher prior malaria incidence ( >\u200a=\u200a2 episodes ppy , n\u200a=\u200a68 ) , only 7% of children lived in town ., This suggests that children with the lowest prior incidence had less exposure to malaria-infected mosquitoes ., Episodes of asymptomatic parasitemia were rare in this cohort ( median 1 episode per subject over the entire study period , IQR 0–4 , Table 1 ) and the incidence of malaria declined only slightly in the year following the blood draw ( from 5 . 7 to 5 . 1 episodes ppy ) , suggesting that effective clinical immunity had not yet emerged in most children ., One child had symptomatic malaria ( parasitemia with a fever requiring treatment ) at the time of the blood draw , and 17 ( 22% ) had blood smears demonstrating parasitemia ., To define the frequency and function of malaria-specific T cell responses , PBMC were stimulated with malaria-infected red blood cells ( iRBC ) and analyzed by flow cytometry for production of IFNγ , IL-10 , and TNFα ( Fig . 1a ) ., The median frequency of malaria-specific CD4+ T cell responses producing any of these cytokines , alone or in combination , was 0 . 20% ( IQR 0 . 12%–0 . 35% ) ., Among all children , frequencies of CD4+ T cells producing IFNγ ( median 0 . 16% ) and IL-10 ( median 0 . 14% ) were significantly higher than those producing TNFα ( median 0 . 04% , P<0 . 001 , Fig . 1b ) ., Production of these two cytokines largely overlapped , with a median of 83% of IL-10-producing cells also making IFNγ , and a median of 71% of IFNγ-producing cells also making IL-10 ., Malaria-specific production of IL-2 was tested in a subset of children ( n\u200a=\u200a44 ) , but responses were consistently of low magnitude ( median frequency 0 . 02% , data not shown ) ., At the time of the assay 17 of the 78 children had positive blood smears; however there was no significant difference in the overall frequency of malaria-specific IFNγ+ ( P\u200a=\u200a0 . 20 ) , TNFα+ ( P\u200a=\u200a0 . 29 ) , or IL-10+ ( P\u200a=\u200a0 . 21 ) CD4+ T cells between children with or without parasitemia ., Malaria-specific CD8 T cell responses were not observed in the peripheral blood of any of the 78 children , although this does not exclude their presence in the liver and other tissues as demonstrated by non-human primate studies 54 ., The pattern of cytokine production by malaria-specific CD4+ T cells was noted to differ markedly based on childrens prior incidence of malaria ( Fig . 2a–c ) ., Both IL-10-producing CD4+ T cells and IFNγ-producing CD4+ T cells were present at higher frequencies among children with a higher prior incidence of malaria ( ≥2 episodes ppy ) than among those with a lower prior incidence ( <2 episodes ppy , P<0 . 001 and P\u200a=\u200a0 . 02 , respectively , Fig . 2a ) ., Most strikingly , CD4+ T cells co-producing IFNγ and IL-10 dominated the response among children with higher prior incidence , but were virtually absent among lower incidence children ( P<0 . 001 , Fig . 2b ) ., Production of TNFα followed the opposite pattern , with higher frequencies of TNFα+/IL10− CD4+ T cells observed among children with lower prior incidence than among those with a higher prior incidence ( P\u200a=\u200a0 . 003 , Fig . 2b ) ., Interestingly , despite these differences in cytokine production profiles , the overall frequency of malaria-specific CD4+ T cells ( i . e . those producing any cytokine ) did not statistically differ between the higher and lower incidence groups ( P\u200a=\u200a0 . 13 ) ., We also analyzed the relationship of prior malaria incidence with the “composition” of the malaria-specific response ( i . e . the proportion of each cytokine combination amongst the total malaria-specific CD4+ T cell population ) , and found similar results ., Among children with <2 episodes ppy , TNFα-producing CD4+ T cells ( including TNFα single-producers and IFNγ/TNFα double producers ) comprised a greater proportion of the malaria-specific response than among children with ≥2 prior episodes ppy , whereas in children with a higher prior malaria incidence , IL-10-producing CD4+ T cells ( including IL-10 single-producers and IFNγ/IL-10 double producers ) comprised a far greater fraction of the malaria-specific response ( P<0 . 001 , Fig . 2c ) ., There was no significant difference in the proportion of IFNγ-producing CD4+ T cells between children with higher and lower incidence ., These findings suggest that the functional phenotype of the malaria-specific CD4+ T cell response differs according to prior exposure , and that with more prior episodes , the overall response is more regulatory ( IL-10 producing ) and less inflammatory ( TNFα producing ) ., While the data above demonstrate that there is a strong relationship between the functional phenotype of malaria-specific CD4+ T cells and prior malaria history , we wished to determine whether this phenotype was influenced by the time elapsed since the most recent malaria episode , the cumulative number of prior malaria episodes , or both , as these parameters are both logically and statistically related ( Spearmans Rho\u200a=\u200a−0 . 46 , P<0 . 001 ) ., We observed a strong inverse correlation between the frequency of IFNγ+/IL-10+/TNFα− CD4+ T cells and the duration since the last episode of malaria ( Spearmans Rho\u200a=\u200a−0 . 39 , P<0 . 001 , Fig . 3d ) , with more recent malaria associated with a higher frequency of these co-producing cells , as well as a positive correlation with the total cumulative number of prior episodes ( Spearmans Rho\u200a=\u200a0 . 23 , P\u200a=\u200a0 . 04 , Fig . 3e ) ., However , when assessed in a multivariate model , the frequency of malaria-specific IFNγ/IL-10 co-producing CD4+ T cells remained strongly associated with the duration since malaria , whereas the total prior incidence was no longer significant ., Similar results were observed for total IL-10 ( Fig . 3a ) and total IFNγ-producing ( Fig . 3b ) populations , and when assessing the duration since last episode of parasitemia ( data not shown ) ., Interestingly , the opposite relationship was observed between total TNFα+ producing cells and the duration since last episode of malaria , with more recent malaria associated with a lower frequency of TNFα -producing cells ( Spearmans Rho\u200a=\u200a0 . 23 , P\u200a=\u200a0 . 041 , Fig . 3c ) ., Further , there was no significant correlation between the number of cumulative prior malaria episodes and TNFα+ producing cells ., Together these data suggest that recency of malaria infection , rather than the total number of past episodes , exerts a dominant influence on the functional phenotype of malaria-specific CD4+ T cells ., Similar findings were obtained when analyzing the “composition” ( i . e . the proportion of responding cells producing IFNγ , TNFα , and/or IL10 ) of the malaria-specific response and duration since last malaria infection ., Protection from clinical malaria in naturally exposed individuals can be defined using a number of outcomes , including a delayed time to reinfection 37 , 38 , 41–43 , 51 , a decreased incidence of malaria over time 53 , and/or a decreased probability of clinical disease once parasitemic 46 ., In all cases , identification of immune correlates of protection is challenging due to the difficulty of distinguishing protection from a lack of exposure to malaria-infected mosquitos 45 , 46 ., To address this , we assessed the relationship between malaria-specific T cell functional subsets and protection from malaria , while adjusting for prior malaria ( duration since last episode and/or cumulative number of prior episodes ) as a surrogate measure of exposure intensity ., We also evaluated potential associations with the overall prevalence of asymptomatic parasitemia , as clinical immunity to malaria is normally characterized by a transition from symptomatic to asymptomatic disease 3 ., In univariate Cox proportional hazards analysis evaluating time to next episode of malaria , a higher frequency of CD4+ T cells producing any IFNγ or IL10 , or the combinations IFNγ+/IL-10+/TNFα− and IFNγ−/IL-10+/TNFα− was associated with a significantly increased hazard of malaria ( Table 2 , left columns ) ., However following adjustment for surrogates of exposure intensity ( duration since last episode of malaria and/or cumulative prior malaria episodes ) in a multivariate model , none of these associations remained significant ., Similar relationships were observed when we analyzed the total malaria incidence in the year following the assay in a multivariate regression model ( Table 2 , middle columns ) ., However , in this analysis both IFNγ+/IL-10+/TNFα− ( IRR 1 . 40 per 10 fold increase , P\u200a=\u200a0 . 038 ) and any IL-10-producing CD4+ T cells ( IRR 1 . 41 per 10 fold increase , P\u200a=\u200a0 . 039 ) remained independently associated with an increased risk of malaria after controlling for duration since last malaria infection ., Nearly identical results were obtained when analyzing the total composition of cytokine producing cells: both the fraction of IFNγ+/IL-10+/TNFα− and any IL10+ cells among all cytokine-producing cells were associated with increased malaria risk ( IRR 1 . 47 , P\u200a=\u200a0 . 038 and 1 . 40 , P\u200a=\u200a0 . 039 per 50% increase in fraction of responding cells , respectively ) ., Together , these data suggest that the dominant population of malaria-specific CD4+ cells , which co-produce IFNγ and IL-10 , are not associated with protection from future malaria , and may in fact be associated with an increased risk of malaria ., We next assessed the relationship of TNFα− producing CD4+ T cells with protection ., In RTS/S vaccine recipients , malaria-specific CD4+ T cells producing TNFα in the absence of IFNγ or IL-2 have recently been shown to correlate with protection from malaria infection 55 ., In our cohort , a greater frequency of malaria-specific CD4+ T cells producing TNFα alone ( IFNγ−/IL-10−/TNFα+ ) was associated with a significantly reduced hazard of developing malaria ( HR 0 . 31 , P\u200a=\u200a0 . 015 per 10 fold increase ) and lower prospective incidence ( IRR 0 . 44 , P\u200a=\u200a0 . 004 per 10 fold increase ) in univariate analysis , but in multivariate models controlling for duration since malaria and/or cumulative prior malaria episodes , these associations were no longer significant ( Table 2 ) ., Interestingly , however , the frequency of malaria-specific CD4+ T cells producing any TNFα was inversely associated with the monthly prevalence of asymptomatic parasitemia , even after controlling for duration since last episode of malaria and/or cumulative prior malaria episodes ( PRR 0 . 41 per 10 fold increase , P\u200a=\u200a0 . 011 ) ., Thus , the absence of malaria-specific CD4+ T cells producing TNFα may be associated with the phenotype of asymptomatic infection ., Although IL10 production by T cells was initially believed to occur predominantly within Th2 and FoxP3+ Treg CD4+ T cell subsets , it is now known that additional subsets , including cells expressing the Th1 master regulator T-bet , produce IL-10 under conditions of continuous antigen exposure 56 , 57 ., We assessed transcription factor expression within the dominant population of malaria-specific IFNγ/IL-10 co-producing cells ( Fig . 4a ) and found that these cells uniformly were TBet+ and FoxP3− ( Fig . 4b–c ) ., These IFNγ/IL-10 co-producing CD4+ T cells were predominantly of an early effector memory phenotype ( CD45RA− , CCR7− CD27+; Fig . 4d–e ) ., CD4+ T cell IFNγ/IL-10 responses to the polyclonal mitogen PMA/Io have previously been shown to correlate with relative protection against severe malaria 52 ., We therefore compared the response to iRBC and PMA/Io stimulation , and found a strong correlation between the frequency of IFNγ/IL-10 double producing CD4+ T cells following iRBC or PMA stimulation ( Spearmans Rho\u200a=\u200a0 . 88 , P<0 . 001 , Supplemental Fig . S1 ) ., As PMA/Io stimulation is thought to induce cytokine production by recently activated cells , these data suggest that this mitogen stimulates cytokine production by malaria-specific T cells that have recently seen their cognate antigen ., IL-10 levels measured concurrently in plasma were significantly higher among children with parasitemia at the time of the blood draw compared with children with no parasitemia ( median 30 . 4 pg/ml vs 11 . 4 pg/ml , P\u200a=\u200a0 . 0035 ) , consistent with prior reports 58–61 ., Similar to IL-10 producing CD4+ T cells , plasma IL-10 strongly correlated with recent malaria ( Spearmans Rho\u200a=\u200a0 . 30 , P\u200a=\u200a0 . 009 , Supplemental Fig . S2a ) ., However plasma IL-10 levels did not correlate with the frequency of total IL-10 producing CD4+ T cells ( Spearmans Rho =\u200a0 . 11 , P\u200a=\u200a0 . 35 , Supplemental Fig . S2b ) , suggesting that additional cell types , including cells of the myeloid lineage , may contribute to plasma IL-10 levels during malaria infection 19 ., Immunomodulation through downregulation of antigen-specific CD4+ T cell proliferative responses has been described in the context of several chronic parasitic infections 62–65 , as well as chronic viral infections that result in persistent antigenemia 66 , 67 ., We assessed proliferation of malaria-specific CD4+ T cells by measuring CFSE dilution following stimulation with schizont extract ( PfSE ) in a subset of children ( n\u200a=\u200a42 ) ., A significant inverse correlation was observed between malaria-specific CD4+ T cell proliferation and cumulative prior incidence ( Spearmans Rho\u200a=\u200a−0 . 39 , P\u200a=\u200a0 . 011; Fig . 5a ) , suggesting that heavy antigen exposure may result in a proliferative defect in malaria-specific CD4+ T cells ., We also observed an inverse correlation between CD4+ T cell proliferation following PfSE stimulation and the frequency of IFNγ/IL-10 co-producing CD4+ T cells ( Spearmans Rho\u200a=\u200a−0 . 31 , P\u200a=\u200a0 . 049 ) ., It has previously been suggested that IFNγ/IL-10 co-producing CD4+ T cells may play an autoregulatory role through suppression of proliferative responses in an IL-10 mediated manner 68 ., We therefore assessed whether in vitro IL10 blockade would reverse the observed proliferative defect ., The ability of CD4+ T cells to proliferate in response to PfSE was partially restored in 8 of 9 subjects upon blockade of IL-10 receptor alpha ( fold change 1 . 7 , P\u200a=\u200a0 . 01 , Fig . 5b–c ) , suggesting that the CD4+ T cell proliferative defect observed in heavily exposed children may be in part due to IL-10 mediated suppression ., In this cohort of young children living in an area of very high transmission intensity in Uganda , very little evidence of clinical immunity had emerged by five years of age ., In this setting , the functional phenotype of the malaria-specific CD4+ T cell response was significantly influenced by prior malaria exposure; with less prior malaria , the overall malaria-specific CD4+ T cell response was more inflammatory ( TNFα-producing ) , but with heavier exposure , the overall malaria-specific response was more regulatory ( IL-10 producing ) ., To our knowledge , this is the first study to show that Th1 IFNγ/IL-10 co-producing cells constitute the dominant population of CD4+ T cells responding to malaria in heavily exposed children ., Moreover , we found no evidence that these IFNγ/IL-10 co-producing cells were associated with protection from future malaria ., Interest in IFNγ/IL-10 co-producing Th1 cells has increased in recent years as these cells have been found to be important regulators of the immune response to several infectious , allergic , and autoimmune diseases 18 , 49 , 50 , 52 , 56 , 69 , 70 ., In a murine model of Toxoplasma gondii , IFNγ produced by these cells was shown to be required for pathogen eradication , and concomitant production of IL-10 was vital for the resolution of the inflammatory response and to prevent tissue pathology 50 ., However , in a murine model of Leishmania major , co-production of IL-10 by Th1 cells prevented pathogen eradication , contributing to chronic infection 49 ., These data suggest that IL-10 co-production by Th1 T cells may help prevent immunopathology , but this may come at the cost of chronic pathogen persistence 71 ., IL-10 levels are increased during malaria infection 58 , 59 , 61 and this regulatory cytokine is thought to play a key role in dampening proinflammatory responses and preventing the development of severe malarial anemia and cerebral malaria 72 ., In mice , Th1 cells were elegantly shown to be the major producer of IL-10 and were critical for limiting the pathology associated with malaria infection 18 ., T cell production of IL-10 has also been described in reports of human malaria infection 14 , 52 , 73–77 ., Plebanski and colleagues described a switch in production from IFNγ to IL-10 in CD4+ T cells from Gambian adults stimulated with altered peptide ligands of the circumsporozoite protein , with an associated suppression of proliferative responses in vitro 14 ., T cells co-producing IFNγ/IL-10 following nonspecific PMA/ionomycin stimulation were described in the context of acute malaria infection 73 , and were also more abundant among children with uncomplicated rather than severe malaria 52 , consistent with a role in modulating inflammation ., More recently , Gitau and colleagues described malaria-specific co-production of IFNγ and IL-10 following stimulation of CD4+ T cells with a variety of expressed PfEMP variants , although these co-producing cells represented a minor fraction of the total antigen-specific CD4+ T cell response 75 ., The potential role that malaria-specific IFNγ/IL-10 co-producing CD4+ T cell cells play in mediating or inhibiting protective immunity in humans has not thus far been investigated 77 ., We observed that CD4+ T cells co-producing IFNγ/IL-10 dominate the T cell response to malaria in heavily exposed children , and that the overall frequency and proportion of these cells among malaria-specific T cells was strongly correlated with recent exposure to malaria , more so than cumulative prior exposure ., These IFNγ/IL-10 co-producing cells express T-bet , indicating that they have differentiated along the Th1 pathway ., The dominance of this functional phenotype among malaria-specific T cells has not previously been reported , and may be related to the unusually high malaria exposure intensity of our cohort , as this cell population was of much lower frequency among children with <2 malaria episodes per year ., Further , frequencies of IL-10–producing and IFNγ/IL10 co-producing cells were not associated with protection from future malaria after controlling for recent and/or cumulative prior malaria , but were instead associated with an increased risk of cumulative malaria in the year following the assay , although this may be due to the inability to fully adjust for the level of environmental exposure to malaria using clinical surrogates such as prior malaria incidence ., We further observed that heavy malaria exposure was associated with a decreased ability of CD4+ T cells to proliferate in response to malaria antigens , and that this impaired proliferation is partially reversed by IL-10 blockade ., These data are consistent with in vitro studies of recently activated IL7R− , CD25− , CD4+ T cells which co-produce IFNγ and IL-10 and limit CD4+ T cell proliferation through IL-10 dependent mechanisms 68 ., In addition , prior studies have shown that IL-10 blockade increases malaria-specific IFNγ cytokine production in filaria-coinfected individuals 78 and in cord blood mononuclear cells from neonates born to mothers exposed to malaria 79 ., A similar IL10-dependent functional impairment of CD4+ T cells has been described in other infections such as HIV that are characterized by chronic high-level antigen stimulation 80 , 81 ., Together , these data are consistent with the hypothesis that IFNγ/IL-10 co-producing CD4+ T cells primarily function to limit the immunopathology associated with malaria infection – including cerebral malaria , anemia , and death - through autoregulation of CD4+ T cell proliferation and cytokine production ., A similar role has been attributed to IL-10-producing Th1 cells in other parasitic diseases characterized by heavy continuous antigen exposure 49 , 50 , with evidence that IL-10 produced by Th1 effector cells acts through a negative feedback loop to regulate CD4+ T cell responsiveness , limiting inflammation and tissue pathology at the cost of impaired pathogen clearance 56 , 71 ., It is possible that unmeasured confounders , such as helminthic co-infections , may have been unequally represented in the high and low-incidence groups , particularly as the lower incidence children were more likely to reside in town ., However routine deworming was performed in all study subjects every 3–6 months , lessening the likelihood that co-infection with helminths explains our findings ., Further studies are needed to determine if IL-10-producing Th1 cells contribute to pathogen persistence , and to the failure of humans to develop sterile protective immunity to malaria ., In addition , we found that children with the fewest prior episodes of malaria were significantly more likely to have malaria-specific production of TNFα without IL-10 , and that the absence of this inflammatory cytokine was associated with the phenotype of asymptomatic infection ., Studies in murine models have shown that TNFα plays an important role in inhibiting the development of hepatic stages of malaria 82 , 83 ., Importantly , a recent study of RTS/S vaccine recipients identified antigen-specific CD4+ T cell production of TNFα as a correlate of protection in vaccinees 55 ., In contrast to that study , we found no evidence of protection after controlling for prior malaria , though we did observe that asymptomatic infection was inversely associated with the frequency of TNFα producing CD4+ T cells , independent of prior malaria ., Together our data suggest that production of this inflammatory cytokine may decrease with increasing cumulative malaria exposure , enabling a transition to asymptomatic infections ., A notable strength of this study was the availability of comprehensive malaria clinical histories spanning from early infancy to the time of the immunologic assessment , plus one additional year thereafter , which enabled us to assess for T cell correlates of both exposure to and protection from malaria ., Several prior studies have reported correlations between T cell responses or IL-10 production and protection from malaria in naturally exposed children 37 , 42 , 53 , but such studies have generally been unable to adequately account for prior malaria exposure ., While we did observe associations , both positive and negative , between malaria-specific CD4+ T cells of varying functional phenotypes and the risk of future malaria , most of these associations were not significant after adjusting for recent or cumulative prior episodes of malaria , surrogates for the level of ongoing exposure to malaria-infected mosquitos ., Hence the failure to account for malaria exposure intensity may lead to spurious associations with protection ., Although we did not identify T cell phenotypes that were associated with protection from future malaria , this may be related to the young age of children in this cohort , as there was little evidence that clinical immunity had developed prior to 5 years of age ., Future longitudinal studies examining responses in older children and adults , incorporating more precise entomological measurements of malaria exposure , are underway ., In conclusion , among naturally exposed children living in a high endemicity setting , malaria-specific CD4+ T cells were present in the vast majority of children , and their functional phenotype differed greatly based on the level of prior exposure to malaria , in particular the duration of time since last infection ., IFNγ/IL-10 co-producing Th1 cells dominated the CD4+ T cell response to malaria in these heavily exposed children , but were not associated with protection from future infection ., These CD4+ T cells may play important immunomodulatory roles in the pathogenesis of malaria in childhood ., Samples for this study were obtained from children enrolled in the Tororo Child Cohort ( TCC ) in Tororo , Uganda , a rural district in south-eastern Uganda with an entomological inoculation rate ( EIR ) estimated at 379 infective bites per person year ( PPY ) in 2012 20 ., Details of this cohort have been described elsewhere , and the sub-study described in this report includes only HIV-uninfected children born to HIV-uninfected mothers 20 , 84–87 ., Briefly , children in the TCC were enrolled at infancy ( median 5 . 5 months of age ) and followed for all medical problems at a dedicated study clinic open seven days a week ., Monthly assessments were done to ensure compliance with study protocols and perform routine blood smears ., All children were prophylactically dewormed with mebendazole every 3–6 months per Ugandan Ministry of Health guidelines 88 ., Children who presented with a documented fever ( tympanic temperature ≥38 . 0°C ) or history of fever in the previous 24 hours had blood obtained by finger prick for a thick smear ., If the thick smear was positive for malaria parasites , the patient was diagnosed with malaria regardless of parasite density , and given artemisinin-based combination therapy for treatment of uncomplicated malaria ., Children were followed until 5 years of age unless prematurely withdrawn ., Incident episodes of malaria were defined as all febrile episodes accompanied by any parasitemia requiring treatment , but not preceded by another treatment in the prior 14 days 20 ., The incidence of malaria was calculated as the number of episodes per person years ( ppy ) at risk ., Asymptomatic parasitemia was defined as a positive routine blood smear in the absence of fever that was not followed by the diagnosis of malaria in the subsequent seven days , and was reported as a count outcome as it was measured via monthly surveillance ., The period prevalence of asymptomatic parasitemia was calculated as the number of episodes/total months observed ., Written informed consent was obtained from the parent or guardian of all study participants ., The study protocol was approved by the Uganda National Council of Science and Technology and the institutional review boards of the University of California , San Francisco , Makerere University | Introduction, Results, Discussion, Methods | Although evidence suggests that T cells are critical for immunity to malaria , reliable T cell correlates of exposure to and protection from malaria among children living in endemic areas are lacking ., We used multiparameter flow cytometry to perform a detailed functional characterization of malaria-specific T cells in 78 four-year-old children enrolled in a longitudinal cohort study in Tororo , Uganda , a highly malaria-endemic region ., More than 1800 episodes of malaria were observed in this cohort , with no cases of severe malaria ., We quantified production of IFNγ , TNFα , and IL-10 ( alone or in combination ) by malaria-specific T cells , and analyzed the relationship of this response to past and future malaria incidence ., CD4+ T cell responses were measurable in nearly all children , with the majority of children having CD4+ T cells producing both IFNγ and IL-10 in response to malaria-infected red blood cells ., Frequencies of IFNγ/IL10 co-producing CD4+ T cells , which express the Th1 transcription factor T-bet , were significantly higher in children with ≥2 prior episodes/year compared to children with <2 episodes/year ( P<0 . 001 ) and inversely correlated with duration since malaria ( Rho\u200a=\u200a−0 . 39 , P<0 . 001 ) ., Notably , frequencies of IFNγ/IL10 co-producing cells were not associated with protection from future malaria after controlling for prior malaria incidence ., In contrast , children with <2 prior episodes/year were significantly more likely to exhibit antigen-specific production of TNFα without IL-10 ( P\u200a=\u200a0 . 003 ) ., While TNFα-producing CD4+ T cells were not independently associated with future protection , the absence of cells producing this inflammatory cytokine was associated with the phenotype of asymptomatic infection ., Together these data indicate that the functional phenotype of the malaria-specific T cell response is heavily influenced by malaria exposure intensity , with IFNγ/IL10 co-producing CD4+ T cells dominating this response among highly exposed children ., These CD4+ T cells may play important modulatory roles in the development of antimalarial immunity . | Despite reports of decreasing malaria morbidity across many parts of Africa , the incidence of malaria among children continues to be very high in Uganda , even in the setting of insecticide-treated bednets and artemisinin-based combination therapy ., Additional control measures , including a vaccine , are sorely needed in these settings , but progress has been limited by our lack of understanding of immunologic correlates of exposure and protection ., T cell responses to malaria are thought to be important for protection in experimental models , but their role in protecting against naturally acquired infection is not clear ., In this study , we performed detailed assessments of the malaria-specific T cell response among 4-year-old children living in Tororo , Uganda , an area of high malaria transmission ., We found that recent malaria infection induces a malaria-specific immune response dominated by Th1 T cells co-producing IFNγ and IL-10 , and that these cells are not associated with protection from future infection ., IFNγ/IL-10 co-producing cells have been described in several parasitic infections and are hypothesized to be important in limiting CD4-mediated pathology , but they may also prevent the development of sterilizing immunity ., These observations have important implications for understanding the pathophysiology of malaria in humans and for malaria vaccine development . | medicine, infectious diseases, adaptive immunity, immune cells, clinical immunology, immunity, t cells, immunity to infections, immune tolerance, malaria, plasmodium falciparum, parasitic diseases, immune response | null |
journal.ppat.1002956 | 2,012 | Post-Transcriptional Regulation of the Sef1 Transcription Factor Controls the Virulence of Candida albicans in Its Mammalian Host | Candida albicans is a ubiquitous component of the mammalian microbiome 1 as well as the most common fungal pathogen of humans 2 , 3 , 4 , 5 ., As this organism transits between its commensal niches ( mammalian skin and gastrointestinal tract ) and those of virulence ( bloodstream and internal organs ) , it experiences profound shifts in the levels of nutrients , the physical environment , and immune surveillance ., We previously demonstrated that a novel C . albicans transcriptional regulatory circuit is required for survival in at least two distinct habitats , the host bloodstream and gastrointestinal tract 6 , where levels of bioavailable iron differ by more than 20 orders of magnitude 7 , 8 ., In the bloodstream , where iron is tightly sequestered by host transferrin 7 , C . albicans defends against iron deficiency through expression of Sef1 , a Cys6Zn2 transcriptional activator of iron uptake genes and an indirect suppressor of the gene for Sfu1 6 ., In the gastrointestinal tract , where iron is abundant thanks to diet and sloughed cells 8 , 9 , C . albicans defends against iron toxicity through the expression of Sfu1 6 , a GATA family transcriptional repressor that inhibits both SEF1 and genes for iron uptake 6 , 10 ., Remarkably , the opposing roles of Sef1 and Sfu1 in iron homeostasis extend to differing relationships with the host , with Sef1 promoting virulence and Sfu1 promoting commensalism in animal models 6 ., However , the details of how these transcriptional regulators are themselves regulated by iron remain to be elucidated ., Sfu1 is broadly conserved among ascomycetes , and orthologs from multiple species have been shown to play a negative role in iron homeostasis through repression of iron uptake genes 10 , 11 , 12 , 13 , 14 , 15 ., The best-characterized ortholog is Fep1 from Schizosaccharomyces pombe that , like Sfu1 , is subject to repression at the transcriptional level when environmental iron is limiting 16 , 17 ., In this species , protein activity is also regulated by iron , since only iron-bound Fep1 can associate with DNA 18 ., By contrast , orthologs of Sef1 have not been extensively characterized , in part because the genomes of only a handful of species in the Saccharomyces and Candida lineages encode this protein 6 ., Moreover , C . albicans Sef1 appears to function differently from its S . cerevisiae ortholog , since iron homeostasis in the latter species is controlled by Aft family proteins 19 , 20 , 21 and is not dependent on Sef1 6 ., Here we describe studies that reveal an unexpected , transcription-independent role of C . albicans Sfu1 in inhibiting Sef1 function , as well as a role for a predicted protein kinase , Ssn3 , in Sef1 activation ., Specifically , we find that , under iron-replete conditions , Sfu1 physically associates with Sef1 and sequesters it in the cytoplasm , where it is destabilized ., In contrast , under iron-depleted conditions , Sef1 forms an alternative complex with Ssn3 , resulting in Sef1 phosphorylation , nuclear localization , and the transcriptional activation of iron uptake genes ., These post-transcriptional regulatory events are of direct consequence to C . albicans virulence , since either overexpression of SFU1 or deletion of SSN3 results in attenuated virulence in a mammalian model ., We hypothesize that these multiple , opposing mechanisms for Sef1 regulation , including a surprising protein-protein interaction with its own transcriptional inhibitor , enable this obligate mammalian parasite to fine-tune its interactions with the host on a spectrum from commensalism to virulence ., Given the important role of Sef1 in promoting C . albicans virulence 6 , we speculated that it would be a prime target for regulation ., We and others had previously shown that , under iron-replete conditions , transcription of SEF1 is repressed by Sfu1 6 , 10 , the C . albicans structural and functional ortholog of S . pombe Fep1 22 ., To determine whether additional regulators contribute to SEF1 gene expression , we used RT-qPCR to compare SEF1 transcript levels in a wild-type strain vs . an isogenic strain lacking the SFU1 gene ., The result was that deletion of SFU1 was sufficient to fully derepress SEF1 , independent of the iron content of the growth medium ( Figure 1a , compare the level of SEF1 in wild-type cells grown in iron-depleted medium bar 2 , derepressing condition to that in the sfu1ΔΔ strain , grown in either iron-replete bar 3 or iron-depleted medium bar 4 ) ; numerical values and statistical analysis are provided in Table S1 ., These results suggested that iron-dependent transcriptional repression by Sfu1 is sufficient to account for SEF1 transcript levels in wild-type cells ., To determine whether forced overexpression of SFU1 could further suppress SEF1 gene expression , we created a strain in which the endogenous promoter of SFU1 was replaced with the strong , constitutively active TDH3 promoter ( SFU1OE ) ; increased levels of SFU1 RNA and protein were confirmed by RT-qPCR and immunoblot analysis , respectively ( Figure S1a and S1b ) ., Overexpressed Sfu1 did not substantially diminish the level of SEF1 mRNA under iron-replete or iron-depleted conditions ( Figure 1a , compare bar 1 with bar 5 and bar 2 with bar 6 ) ., The failure of overexpressed SFU1 to inhibit the transcription of SEF1 under iron-depleted conditions suggested that C . albicans Sfu1 might , like its S . pombe ortholog 18 , require iron as a cofactor for binding to DNA and transcriptional repression ., Although SEF1 mRNA levels were normal in the SFU1-overexpression strain ( Figure 1a ) , this strain demonstrated hypersensitivity to treatment with the iron chelator , bathophenanthroline disulfonic acid ( BPS; Figure S2 ) , suggestive of a potential defect in iron acquisition ., Addition of FeCl3 to the BPS-treated medium was sufficient to reverse the growth defect ( Figure S2 ) , confirming the specificity of the iron-chelation phenotype ., To determine whether Sef1 protein levels were affected in the SFU1OE strain , we utilized an epitope-tagged version of Sef1 in which 13 copies of the Myc epitope were fused in-frame at the C-terminus; this fusion protein is fully functional 6 ., Surprisingly , the steady state level of Sef1-Myc was substantially reduced , particularly under iron-depleted conditions ( Figure S3 ) ., The observations that overexpression of SFU1 does not affect SEF1 mRNA levels but strongly decreases Sef1 protein levels raised the possibility that Sfu1 may have a second function in the post-transcriptional regulation of Sef1 ., To determine whether Sef1 localization is regulated , we used indirect immunofluorescence to visualize Sef1-Myc in wild-type cells exposed to varying concentrations of iron ., Under iron-replete conditions , Sef1-Myc was localized primarily in the cytoplasm ( Figure 1b , WT strain , H; note the absence of green Sef1-Myc signal in the FITC channel in areas that correspond to red DNA signal in the DAPI channel; a negative control showing minimal staining of an isogenic strain that lacks the Myc epitope is shown in Figure S4a ) ., Under iron-depleted conditions , however , Sef1-Myc was primarily nuclear , with prominent areas of yellow overlap when the FITC and DAPI channels were merged ., Notably , examination of Sef1-Myc in an sfu1ΔΔ mutant revealed constitutive nuclear localization , even under iron-replete conditions ( sfu1ΔΔ strain , Figure 1b ) ., Conversely , overexpression of SFU1 resulted in substantial cytoplasmic localization of Sef1-Myc even under iron-depleted conditions in which it is usually nuclear ( SFU1OE strain , Figure 1b ) ., By comparison , an Sfu1-Myc fusion protein was found to be distributed between the nucleus and cytoplasm in wild-type cells propagated under iron-replete conditions and primarily cytoplasmic under iron-limiting conditions ( Figure S4b ) ., These results established that Sef1 localization varies as a function of iron , that Sfu1 promotes Sef1 localization in the cytoplasm , and that the protein localizing activity of Sfu1—unlike its transcriptional repression activity ( Figure 1a ) —does not inherently require iron ., Immunoblot analysis of Sef1-Myc recovered from wild-type cells grown under iron-replete vs . iron-depleted conditions demonstrated an inverse relationship between Sef1 protein abundance and iron levels ( Figure 2a , lanes 1 and 2 ) , which was expected based on the known , iron-dependent inhibitory activity of Sfu1 on SEF1 gene expression ., An unexpected finding was that the electrophoretic mobility of Sef1 also varied in an iron-dependent fashion ., This subtle but reproducible decrease in Sef1 mobility under iron-depleted conditions was observed not only in wild-type cells , but also in an sfu1ΔΔ deletion mutant ( lanes 3 and 4 ) , arguing against a role for Sfu1 in this process ., We hypothesized that the lower mobility form of Sef1 might result from covalent phosphorylation ., To test this hypothesis , we used a tandem affinity purification strategy to recover TAP-tagged Sef1 from C . albicans grown under iron-replete or iron-depleted conditions ., Purified TAP-tagged Sef1 exhibited an iron-dependent mobility shift similar to that of Sef1-Myc , with protein from the iron-depleted cells running with slightly lower mobility ( Figure 2b , compare lanes 1 and 3 ) ., Treatment of the purified proteins with lambda phosphatase , a broad specificity enzyme with activity on phospho-serine , phospho-threonine , and phospho-tyrosine residues , resulted in conversion of the lower mobility form of Sef1-TAP to the higher mobility form ( Figure 2b , compare lane 4 to lanes 1 and 2 ) , in support of our hypothesis ., To identify the kinase responsible for low-iron-dependent phosphorylation of Sef1 , we tested the 31 available homozygous knockout mutants affecting predicted kinases for sensitivity to BPS ., Our reasoning was that , if phosphorylation of Sef1 is required for full induction of iron uptake genes , then a mutant lacking the relevant kinase might be hypersensitive to iron depletion , that is , phenotypically similar to sef1ΔΔ itself 6 , 23 ., Our screen identified the ssn3ΔΔ mutant as being hypersensitive to iron depletion ( Figure S2 ) ., Further , an immunoblot of Sef1-Myc recovered from the ssn3ΔΔ strain revealed persistence of the higher mobility form under iron-depleted conditions ( Figure 2c ) , consistent with a role for Ssn3 in phosphorylation of Sef1 ., The identical result was obtained when Sef1-Myc was examined in a strain encoding a predicted kinase-dead allele of Ssn3 ( Ssn3D325A , Figure S5a ) ., Although C . albicans Ssn3 has not yet been characterized , its S . cerevisiae ortholog is a cyclin-dependent kinase with two known functions: first , it is a component of the Mediator complex with inhibitory activity on RNA polymerase II 24; second , it phosphorylates a number of specific transcription factors to regulate their activity , nuclear-cytoplasmic localization , and/or stability 25 , 26 , 27 ., To determine whether C . albicans Ssn3 influences the localization of Sef1 , we performed indirect immunofluorescence on Myc-tagged Sef1 in the ssn3ΔΔ mutant ., As shown in Figure 2d , deletion of SSN3 resulted in constitutive cytoplasmic localization of Sef1-Myc under both iron-replete and iron-depleted conditions; similar mislocalization was observed in a strain containing Ssn3D325A ( Figure S5b ) ., Unlike the case with SFU1 , however , overexpression of SSN3 via the TDH3 promoter ( SSN3OE ) had no obvious effect on Sef1-Myc localization ( Figure S5b ) , perhaps indicating that the nuclear localization activity of Ssn3 is restricted to low iron conditions ., The preceding results were suggestive of a model in which Sfu1 and Ssn3 have opposite and competing roles in Sef1 localization , with Sfu1 promoting cytoplasmic localization and Ssn3 promoting nuclear localization ., To test this model , we utilized the SFU1-overexpression strain that mislocalizes Sef1-Myc to the cytoplasm under iron-depleted conditions ( Figure 1b ) ., We predicted that , if Ssn3 competes with Sfu1 for localization of Sef1 , then overexpression of SSN3 might rescue this Sef1 mislocalization phenotype ., Indeed , a strain in which both genes are driven by the strong TDH3 promoter exhibits substantial restoration of nuclear Sef1-Myc under iron-depleted conditions , with normal cytoplasmic localization under iron-replete conditions ( Figure 3a ) ., These results indicate that Sfu1 and Ssn3 exert opposing roles on Sef1 localization , but only under iron-depleted conditions ( when Sef1 is phosphorylated ) ., To determine whether Sef1 physically associates with Sfu1 and/or Ssn3 , we created a series of double epitope-tagged strains , each containing a Myc-tagged version of one of the three potentially interacting proteins and a TAP-tagged version of another; the TAP epitope consists of a calmodulin binding domain fused to a TEV cleavage site and a Protein A domain ( Figure S6a; 28 , 29 ) ., Co-immunoprecipitation experiments were performed using whole cell extracts prepared from cells grown under iron-replete or iron-depleted conditions ., Extracts were incubated with IgG-sepharose , which binds to the Protein A component of the TAP epitope , followed by extensive washing of the immunoprecipitated complexes and protein electrophoresis under denaturing conditions ( SDS-PAGE; see Figure S6b for a schematic of the protocol ) ., Finally , immunoblots were probed with anti-Myc antibodies to determine the presence or absence of a Myc-tagged putative binding partner ., Specificity of IgG-sepharose for the TAP tag was confirmed by performing experiments with strains containing Myc-tagged fusion proteins and an unfused TAP tag ( Figure S6c ) , and specificity of the anti-Myc antibodies for the Myc epitope was confirmed using cells containing only the TAP-tagged fusion proteins ( Figure 3b ) ., Shown in Figure 3b are the results with Sfu1-Myc and Sef1-TAP ., Sfu1-Myc was efficiently co-immunoprecipitated with Sef1-TAP when cells were propagated in iron-replete medium ( lane 3 , IP ) , but not when iron-starved cells were used ( lane 4 , IP ) ., On the other hand , when the epitope tags were reversed , co-immunoprecipitated Sef1-Myc was poorly visualized using extracts of iron-replete cells ( Figure 3c , lane, 3 ) but was easily seen using iron-starved cells ( Figure 3c , lane 4 ) ., Together , these results suggest that Sef1 and Sfu1 interact physically in a manner that is independent of iron levels , whereas the sensitivity of our biochemical assay is a function of the relative abundance of the Myc-tagged protein in the extract ., Co-immunoprecipitation experiments combining either Sef1-Myc or Sfu1-Myc with Ssn3-TAP revealed a robust interaction between Ssn3 and Sef1 , but no detectable interaction between Ssn3 and Sfu1 ( Figure 3d ) ., That is , Sef1-Myc was efficiently co-immunoprecipitated with Ssn3-TAP from an extract of iron-depleted cells ( Figure 3d , lane 4 , IP ) , which express relatively high amounts of Sef1-Myc protein ( Figure 3d , lane 4 , input ) , whereas Sfu1-Myc was not co-immunoprecipitated under any condition ( Figure 3d , lanes 5 and 6 , respectively ) ., When the epitope tags were reversed , Ssn3-Myc was efficiently co-immunoprecipitated with Sef1-TAP using either iron-replete ( Figure 3e , lane, 3 ) or iron-depleted ( Figure 3e , lane, 4 ) cells; note that Ssn3-Myc is relatively abundant under both conditions ., These results suggest that Sef1 physically associates with Ssn3 as well as Sfu1 , but these appear to represent alternative complexes since Ssn3 and Sfu1 do not associate with each other ., To learn whether the stability of Sef1 varies with its intracellular localization , we determined the half-life of Myc-tagged Sef1 in wild-type C . albicans and in mutants in which Sef1 is stably localized in either the nucleus or the cytoplasm ., Under iron-replete conditions , Sef1 is predominantly cytoplasmic in wild-type C . albicans but is mislocalized to the nucleus in sfu1ΔΔ ( Figure 1b ) ., To obtain sufficient Sef1 protein for the analysis and to uncouple the role of Sfu1 in Sef1 localization from its effects on SEF1 transcription , we replaced the endogenous SEF1 promoter with a constitutively active TDH3 promoter in both wild-type and sfu1ΔΔ strains; overexpressed Sef1-Myc exhibited the same pattern of iron-dependent nuclear vs . cytoplasmic localization as Sef1-Myc expressed from its endogenous promoter ( Figure S7 ) ., The strains were propagated to mid log phase in iron-replete medium , followed by addition of cycloheximide to block further translation , and serial sampling for measurements Sef1-Myc abundance ., Shown in Figure 4a is a quantitative immunoblot of Sef1-Myc and alpha tubulin , which was used as an internal control for protein loading ., Under these iron-replete conditions , the calculated half-life of cytoplasmic Sef1-Myc was ∼80 minutes ( wild type , R2\u200a=\u200a0 . 94 ) and that of nuclear Sef1-Myc was ∼160 minutes ( sfu1ΔΔ , R2\u200a=\u200a0 . 92 ) ., Next , we examined Sef1-Myc stability under iron-depleted conditions , in which the protein is predominantly nuclear in wild-type cells ( Figure 1b ) but mislocalized to the cytoplasm in the ssn3ΔΔ mutant ( Figure 2d ) ., Wild-type and ssn3ΔΔ strains expressing SEF1-MYC from the endogenous SEF1 promoter were propagated in iron-depleted medium to mid-log phase , then treated with cycloheximide and visualized as above ( Figure 4b ) ., Under these iron-depleted conditions , the calculated half-life of nuclear Sef1-Myc ( ∼150 minutes in wild type; R2\u200a=\u200a0 . 98 ) was once again more stable than that of cytoplasmic Sef1-Myc ( ∼40 minutes in ssn3ΔΔ; R2\u200a=\u200a0 . 96 ) ., The most parsimonious explanation for these results is that Sef1 is degraded more rapidly in the cytoplasm than in the nucleus; however , we cannot exclude the possibility that Ssn3 and Sfu1 exert independent effects on Sef1 degradation that are unrelated to its intracellular localization ., Our current model of Sef1 regulation , which integrates these results with previously published findings 6 , 10 , 22 , is depicted in Figure 4c ., According to the model , Sef1 is subject to two distinct forms of Sfu1-mediated repression when environmental iron is replete:, 1 ) transcriptional repression of the SEF1 gene , through direct binding and repression of transcriptional initiation; and, 2 ) post-translational inhibition of Sef1 protein , through direct binding and retention in the cytoplasm , where Sef1 is more rapidly degraded ., Alternatively , under iron-limiting conditions , when Sfu1 protein is depleted , Sef1 is bound by Ssn3 , phosphorylated , and localized in the nucleus , where it activates expression of iron uptake genes ., Our recent observation that Sef1-Myc is constitutively cytoplasmic in an sfu1ΔΔ/ssn3ΔΔ double mutant strain ( Figure S8 ) suggests that Ssn3 may play actively promote the nuclear localization of Sef1 , beyond merely extricating Sef1 from Sfu1 ., We previously demonstrated that SEF1 gene expression is induced in the iron-limiting environment of the host bloodstream and that SEF1 is required for virulence in a murine model of bloodstream candidiasis 6 ., Conversely , we showed that SFU1 is not required for virulence but rather that the sfu1ΔΔ mutant exhibits increased competitive fitness relative to wild-type C . albicans , presumably because of an enhanced ability to take up extracellular iron 6 ., Our current results suggest that the negative effect of Sfu1 on C . albicans virulence likely results from mislocalization of Sef1 to the cytoplasm rather than from repression of SEF1 gene expression , since only the former activity is observed under conditions of iron depletion ( compare Figure 1a and Figure 1b ) ., We tested this hypothesis by examining the virulence of mutants with moderate ( SFU1OE , Figure 1b , low iron condition ) to severe ( ssn3ΔΔ , Figure 2d , low iron condition ) defects in Sef1 nuclear localization ., As shown in Figure 4d and 4e , both mutants were significantly attenuated in the murine bloodstream infection model , such that mice infected with either mutant survived longer than mice infected with wild type ., Note also that the strength of the virulence defects paralleled the strength of the Sef1 mislocalization defects of the two mutants , with those of ssn3ΔΔ being worse , although contributions from additional misregulated targets of Ssn3 cannot be excluded ., Sef1 plays a central role in C . albicans pathogenesis through promoting the expression of virulence factors as well as iron uptake genes , whereas Sfu1 is essential for commensalism 6 ., Given its role in virulence and , perhaps , in the choice between commensal and virulent lifestyles , we hypothesized that Sef1 would be a prime target for regulation beyond transcriptional repression by Sfu1 ., Indeed , our analysis of Myc-tagged Sef1 in wild-type C . albicans has revealed multiple levels of iron-dependent regulation , including nuclear vs . cytoplasmic localization , phosphorylation , and differential protein stability ., In wild-type cells , Sef1 protein is nuclear , phosphorylated , stable , and competent for transcriptional activation only under iron-depleted conditions such as those encountered in the bloodstream ., Our analysis of Sef1 in C . albicans mutants has shed further light on the mechanisms of Sef1 regulation ., Surprisingly , in the sfu1ΔΔ mutant , Sef1-Myc is constitutively nuclear , whereas in an SFU1-overexpression strain it is predominantly cytoplasmic ., These results clearly suggested a role for Sfu1 in the cytoplasmic localization of Sef1 ., Our screen of C . albicans mutants affecting predicted kinases exposed a role for Ssn3 in promoting cellular resistance to iron depletion as well as phosphorylation of Sef1 ., Co-immunoprecipitation experiments indicating that Ssn3 forms a physical complex with Sef1 supported a direct role for Ssn3 in Sef1 phosphorylation ., Our finding that Sef1-Myc is constitutively cytoplasmic in the ssn3ΔΔ mutant suggested that Ssn3 might oppose Sfu1 by promoting the nuclear localization of Sef1 ., This hypothesis was validated by the ability of overexpressed SSN3 to overcome the cytoplasmic Sef1-mislocalization phenotype ( under low iron conditions ) of an SFU1-overexpression strain ., Finally , our observations that Sfu1 and Ssn3 were both detectable in complexes with Sef1 , but that neither could be found associated with the other , suggested that the functional antagonism between Sfu1 and Ssn3 occurs in part through competitive binding to Sef1 protein ., Meanwhile , the observation that Sef1 is constitutively cytoplasmic in an sfu1ΔΔ/ssn3ΔΔ double mutant argues that Ssn3 plays at least one additional role in Sef1 nuclear localization ., These studies led to a revised model of Sef1 regulation ( Figure 4c ) ., According to the model , under iron-replete conditions , Sfu1 utilizes two distinct mechanisms to inhibit the function of Sef1:, 1 ) Transcriptional repression , via direct binding to the SEF1 promoter , and, 2 ) Post-transcriptional repression , via binding to Sef1 protein and forced localization in the cytoplasm , where Sef1 is unstable and unable to participate in transcription ., To our knowledge , this would be the first example of a regulatory factor that regulates it target by both transcriptional and post-transcriptional mechanisms ., Under iron-limiting conditions , Sfu1 protein is depleted , and Sef1 associates with the predicted protein kinase , Ssn3 ., Ssn3 most likely phosphorylates Sef1 directly , and either the complex or free Sef1 is transported to the nucleus , where Sef1 functions as a transcriptional activator ., A key goal of future studies will be to understand how iron regulates these newly described activities of Sfu1 and Ssn3 ., The findings that Ssn3 and Sfu1 post-transcriptionally regulate Sef1 , an important virulence factor , raised the question of whether these regulatory events impact C . albicans virulence ., Previously , we observed that deletion of SFU1 leads to hypervirulence in the murine bloodstream infection model , with the sfu1ΔΔ mutant significantly better at colonizing host kidneys than wild-type C . albicans 6 ., We attributed this enhanced fitness to derepression of SEF1 and iron uptake genes in the mutant , resulting in an increased capacity for iron acquisition ., In light of our current results showing that Sfu1 requires iron for transcriptional repression activity , a more likely explanation for the fitness advantage of sfu1ΔΔ is that Sef1 is constitutively nuclear ( and therefore transcriptionally active ) in this strain , whereas in wild type some fraction of Sef1 is retained in the cytoplasm and degraded ., Our current observations with SFU1OE and ssn3ΔΔ mutants dovetail with these findings by showing the converse , i . e . that mutants with incremental defects in the nuclear localization of Sef1 have proportional defects in virulence ., Together , these results strongly support the hypothesis that C . albicans iron acquisition ( and therefore virulence ) can be modulated up or down , respectively , through the activities of Ssn3 or Sfu1 on Sef1 localization and stability ., We hypothesize that the evolution of such fine-tuned regulation of a potent transcription factor is particularly advantageous to an obligate commensal-pathogen , such as C . albicans , which must continuously adapt to differing iron concentrations among the various microenvironments of its mammalian host , while avoiding excessive expression of pathogenicity genes during its usual role as a commensal ., All procedures involving animals were approved by the Institutional Animal Care and Use Committee at the University of California San Francisco and were carried out according to the National Institutes of Health ( NIH ) guidelines for the ethical treatment of animals ., C . albicans strains were routinely propagated in YPD , also referred to as “iron-replete” medium ., “Iron-depleted” medium is YPD supplemented with one of the specific iron chelators , bathophenanthroline disulfonic acid ( BPS ) or 2 , 2′-dipyridyl ( DIP ) , as previously described 6 ., All C . albicans strains used in this study are described in Table S2 , primers are listed in Table S3 , and plasmids are listed in Table S4 ., Construction of C . albicans knockout mutants , complemented ( gene addback ) strains , and strains containing Myc-tagged fusion proteins was performed as previously described 6 , 42 , 43 , 44 ., For introduction of TAP epitopes at the C-terminus of Sef1 , Sfu1 , and Ssn3 , a series of plasmids was constructed using PCR and homologous recombination in S . cerevisiae 45 ., The vector was pRS316 46 , and the insert consisted of ( 5′ to 3′ ) : a PmeI restriction site; 350–450 bp of target ORF sequence up to , but not including , the stop codon; the TAP tag coding sequence 29; a SAT1 ( dominant selectable marker ) -flipper cassette 47; 350–450 bp of sequence downstream of the target ORF; and a second PmeI restriction site ., Plasmids were called pSN150 ( Sef1-TAP ) , pSN228 ( Sfu1-TAP ) , and pSN219 ( Ssn3-TAP ) ., PmeI-digested plasmids were transformed into wild-type C . albicans reference strain SN250 42 , and nourseothricin-resistant C . albicans transformants were screened by colony PCR to verify the correct orientation of the C-terminal TAP tag and SAT1-flipper cassette ., Strains expressing both Myc- and TAP-tagged fusion proteins were constructed by transforming strains already expressing the Myc-tagged protein with the appropriate PmeI-digested TAP-tag integration fragment , as described above ., Overexpression strains for SEF1- and SSN3 were created by replacing portions of the endogenous promoters with the highly active TDH3 promoter ., PCR and homologous recombination in S . cerevisiae 45 were used to create plasmids containing ( 5′ to 3′ ) : a PmeI restriction site; 350–450 bp of sequence homology ending ∼500 bp upstream of the target ORF; the SAT1 gene ( dominant selectable marker ) ; the TDH3 promoter; 350–450 bp of sequence homology beginning with the start codon of the target ORF; and a second PmeI site ., The vector was pRS316 46 , the source of NAT1-TDH3 promoter was pCJN542 48 , and the resulting plasmids were named pSN147 ( SEF1OE ) and pSN229 ( SSN3OE ) ., Correct integration of the inserts in nourseothricin-resistant transformants was verified by colony PCR , and overexpression of SEF1 and SSN3 was confirmed by RT-qPCR ., The SFU1 overexpression strain ( SN742 ) was created using an analogous method ., pSN141 was engineered to contain ( 5′ to 3′ ) : a PmeI site; 350–450 bp of sequence upstream of the C . albicans LEU2 ORF; the C . dubliniensis ARG4 gene ( selectable marker ) ; the TDH3 promoter; the SFU1 ORF; 350–450 bp sequence downstream of the LEU2 ORF; and a second PmeI restriction site ., After digestion with PmeI , the plasmid was transformed into SN515 ( sfu1ΔΔ ) ., Correct integration of the insert in Arg+ transformants was verified by colony PCR , and overexpression of SFU1 was confirmed by RT-qPCR ., The Ssn3D325A kinase-dead mutant ( SN977 ) was created in a similar manner to that of the SFU1OE strain ., First , PCR and primers SNO1394 through SNO1397 ( Table S3 ) were used to create a D325A-encoding variant of the SSN3 ORF ., Next , plasmid pSN239 was engineered to contain ( 5′ to 3′ ) : a PmeI site; 350–450 bp of sequence upstream of the C . albicans LEU2 ORF; C . dubliniensis ARG4 ( selectable marker ) ; the TDH3 promoter; the mutant SSN3 ORF; 350–450 bp sequence downstream of the LEU2 ORF; and a second PmeI restriction site ., A Myc-tagged version of Ssn3D325A ( SN987 ) was created using a plasmid ( pSN273 ) that contains ( 5′ to 3′ ) : a PmeI site; 350–450 bp of SSN3 ORF sequence up to , but not including , the stop codon; sequence encoding 13×Myc; a SAT1-flipper cassette 47; 272 bp of sequence downstream of the SSN3 ORF; 350–450 bp of sequence downstream of the LEU2 ORF; and a second PmeI restriction site ., PmeI-digested plasmid was transformed into SN977 , and correct integration in nourseothricin-resistant transformants was verified by colony PCR ., Sequences of all PCR products were verified by DNA sequencing ., C . albicans was grown at 30°C for 5–6 hours in “iron-replete” ( YPD ) or “iron-depleted” medium ( YPD supplemented with 500 µM BPS ) to OD600\u200a=\u200a0 . 8–1 . 0 ., Cell fixation , cell wall digestion , and antibody hybridization were performed as previously described 49 except that the 9E10 anti-c-Myc antibody ( Covance Research ) was used at a 1∶300 dilution and detected with a 1∶400 dilution of Cy2-conjugated secondary antibody ( Jackson ImmunoResearch , 715-225-151 ) ., Images were acquired under 100× oil objective using a cooled CCD camera ( Cooke Sensicam ) mounted on an inverted microscope ( Zeiss Axioplan 200 M; Carl Zeiss MicroImaging ) or a Nikon Eclipse TE2000-E fluorescence microscope ., All images were processed with ImageJ software ( National Institutes of Health ) ., C . albicans protein extracts were prepared under denaturing condition using a protocol adapted from a previously described method 50 ., Lysates corresponding to 1 OD600 of cells were analyzed by SDS-PAGE and immunoblotted with either anti-c-Myc ( 9E10 , Covance Research ) for Myc-tagged proteins or anti-peroxidase soluble complex antibody ( Sigma , P2416 ) for TAP-tagged proteins ., Immunoblots were also probed with anti-alpha tubulin antibody ( Novus Biologicals , NB100-1639 ) as a loading control ., C . albicans strains were grown on YPD medium ( “iron-replete” ) or YPD medium supplemented with the specific iron chelator 2 , 2′-dipyridyl ( DIP ) at a final concentration of 0 . 5 mM ( “iron-depleted” ) ., A sample of 1–1 . 5 OD600 cells was taken immediately ( zero time point ) before addition of cycloheximide to a final concentration of 2 mg/ml ., At the indicated times , 1 OD value of cells was collected and harvested for protein preparations and immunoblotting ., Semiquantitative detection of protein levels was performed using the LiCor Odyssey Infrared Imager ( Lincoln , NE ) ., Integrated fluore | Introduction, Results, Discussion, Materials and Methods | The yeast Candida albicans transitions between distinct lifestyles as a normal component of the human gastrointestinal microbiome and the most common agent of disseminated fungal disease ., We previously identified Sef1 as a novel Cys6Zn2 DNA binding protein that plays an essential role in C . albicans virulence by activating the transcription of iron uptake genes in iron-poor environments such as the host bloodstream and internal organs ., Conversely , in the iron-replete gastrointestinal tract , persistence as a commensal requires the transcriptional repressor Sfu1 , which represses SEF1 and genes for iron uptake ., Here , we describe an unexpected , transcription-independent role for Sfu1 in the direct inhibition of Sef1 function through protein complex formation and localization in the cytoplasm , where Sef1 is destabilized ., Under iron-limiting conditions , Sef1 forms an alternative complex with the putative kinase , Ssn3 , resulting in its phosphorylation , nuclear localization , and transcriptional activity ., Analysis of sfu1 and ssn3 mutants in a mammalian model of disseminated candidiasis indicates that these post-transcriptional regulatory mechanisms serve as a means for precise titration of C . albicans virulence . | Candida albicans is a fungus that resides on the skin and in the gastrointestinal tract of humans and other mammals ., However , this commensal organism is also capable of proliferating and causing disease in people who have received antibiotics , who are immunocompromised , or who have suffered injury to epithelial layers ., We previously identified a novel transcription factor called Sef1 that promotes C . albicans virulence by activating the expression of iron uptake genes in iron-poor environments , such as the host bloodstream ., However , in iron-replete environments such as the gastrointestinal niche , the SEF1 gene is repressed by a second transcription factor called Sfu1 ., Here , we report our discovery of a series of post-transcriptional regulatory events that determine the intracellular localization , stability , and activity of Sef1 protein ., Mutants that disrupt these post-transcriptional events alter C . albicans virulence in a mammalian model of disseminated infection ., The existence of multiple levels of regulation speaks to the importance of Sef1 in C . albicans virulence and suggests that close titration of Sef1 activity is important for adaptation to distinct microenvironments within the mammalian host . | fungal biochemistry, medicine, infectious diseases, mycology, microbial evolution, microbial pathogens, biology, microbiology, fungal diseases | null |
journal.pbio.1001951 | 2,014 | Disentangling Human Tolerance and Resistance Against HIV | In response to pressure by pathogens , host populations can evolve in two ways: They can develop either resistance or tolerance to the disease 1–8 ., Resistance mechanisms reduce the pathogen burden ., Tolerance mechanisms , in contrast , reduce the damage that accompanies infection without affecting the pathogen directly ., One of the best examples for tolerance are sooty mangabeys infected with Simian Immunodeficiency Virus ( SIV ) , which—despite harboring high virus loads—do not develop disease 9 ., Whether hosts evolve resistance or tolerance affects the evolutionary trajectory of host-pathogen systems 2 , 3 , 10–12 ., The evolution of resistance genes in the host provokes counteradaptations of the pathogen that overcome host resistance , resulting in an endless arms race ., In contrast , tolerance genes benefit both the host and the pathogen and are therefore predicted to fix ., It is increasingly recognized that disentangling resistance and tolerance not only advances our understanding of the coevolution between hosts and pathogens but also is relevant clinically 13 ., Like resistance factors , mechanisms of tolerance , once identified , can be exploited for therapy ., In contrast to resistance-based therapy , tolerance-based treatment does not aim at reducing the pathogen load but rather at ensuring the well-being of the host ., For that reason , tolerance-based therapy is also hypothesized to be evolution-proof—that is , not to select for drug-resistant pathogens 4 , 5 , 14 ., It has been argued , however , that the pathogen population might evolve higher virulence in response to tolerance-based treatment 3 , 15 , 16 ., Although numerous review papers have been written on the potential benefits of tolerance research 1–8 , the formal framework for disentangling tolerance and resistance has not been applied to many animal disease systems ., There is a paradigmatic study on mouse malaria 17 and a few on insects 18–20 ., But a quantitative tolerance analysis has , to our knowledge , not yet been conducted for any clinically relevant human disease ., In this study , we apply such an analysis to HIV infection in humans ., Formally , tolerance can be quantified as the change in disease progression across different levels of pathogen burden ( see Figure 1A ) 2 , 4 ., In the context of HIV , excellent measures of disease progression and pathogen burden are available ( see Figures 1B and 2A ) ., A few weeks after infection , HIV attains a level in the plasma of infected individuals that is approximately stable over several years ., This level , called the set-point viral load , is very well suited as a proxy for the “parasite burden” necessary for a formal tolerance analysis ., The rate of disease progression—the second essential parameter for an analysis of tolerance—can be measured quantitatively by the decline of CD4+ T lymphocytes ., Before infection , individuals have on average 1 , 000 CD4+ T cells per µl of blood ., A decline of CD4+ T cells below 200 per µl of blood defines AIDS ., Thus , the decline of CD4+ T cells reflects what we know about the mechanistic basis of the disease ., CD4+ T-cell declines have also been found to be independent predictors of disease progression in the Swiss HIV Cohort 21 that we analyzed here and other cohorts 22 ., Importantly , the rate of decline can be calculated in a much shorter time scale than the direct observation of disease progression requires ., The faster the CD4+ T cells decline , the higher the rate of progression toward disease and death—that is , the higher the virulence of the infection in the sense of evolutionary ecology ., For these reasons , also previous studies on virulence relied on the CD4+ T-cell decline 23 ., To our knowledge , such a well-established , quantitative measure of virulence is not available for any other human infection ., To establish the baseline relationship between CD4+ T-cell decline and viral load , we performed a regression analysis ., We found that this relationship is significantly nonlinear ( see Figure 2 ) ., Although nonlinear tolerance curves are a departure from what has been reported in other systems , this finding is not surprising ., Linearity is an assumption generally adopted in regression analyses mostly for the sake of simplicity and convenience ., Commonly , low sample sizes precluded the assessment of a potential nonlinearity ., The establishment of such a nonlinearity in the context of tolerance , however , is particularly crucial to reliably establish tolerance differences between groups 24 ., The relationship is best described by a quadratic relationship ( see Figure 2B and Text S1 ) ., The intercept of the relationship is not significantly different from 0 ., This is in line with the expectation that uninfected individuals should have relatively stable CD4+ T-cell counts ., Also the linear term is not significantly different from 0 ., Mathematically , we can write the relationship as: ( 1 ) In this equation , denotes the rate of change of CD4+ T cells per µl of blood per day , and the logarithm to the base 10 of the viral load per ml of plasma ., The quadratic model explains 5% of the variation in CD4+ T-cell decline , consistent with previous studies investigating this relationship with linear models 25 ., The parameter α is the quantitative measure of the average tolerance across the entire study population , which we used in the present study ., It describes how the relationship curves downwards; that is , it measures how the decline in CD4+ T cells , —a surrogate measure of disease progression—changes with the set-point viral load ., For a value α\u200a=\u200a0 , CD4+ T cells would not decline irrespective of the set-point viral load ., This case would correspond to complete tolerance ., If α<0 , an increase in the set-point viral load accelerates the progression towards disease ., The lower α , the lower the tolerance ., For the entire study population , we estimated α\u200a=\u200a−0 . 0111±0 . 0003 ., Four individuals with an infection characterized by very high viral load and minimal disease progression are also depicted in Figure 2B ., They lie above the average tolerance curve ., These individuals , referred to as viremic nonprogressors 26 , share the transcriptomic , interferon response , and gut microbial translocation profile of nonpathogenic SIV infection in their natural host species 26–28 ., Thus , the tolerance analysis correctly identified individuals whose tolerance had been previously established ., First we tested if the tolerance parameter differs with sex and the age at which individuals were infected ., Information on these demographic characteristics was available for all 3 , 036 individuals in our study population ( see Materials and Methods ) ., Although females had an almost 2-fold lower viral load set-point than males , we did not find significant differences in tolerance between sexes , either in a univariate analysis ( F test: p\u200a=\u200a0 . 69; Figure 3A ) or in an analysis adjusting for age difference between sexes ( F test: p\u200a=\u200a0 . 45 ) ., This result challenges previous reports , according to which females are less tolerant ( see Discussion ) 29 ., The age at which individuals become infected with HIV , however , was very strongly associated with tolerance ( Figure 3B ) , both in univariate ( F test: p\u200a=\u200a10−9 ) and multivariate analyses controlling for sex ( F test: p<3×10−8 ) ., According to this analysis , at equivalent viral load , the disease progression rate of an individual who contracts HIV at the age of 60 is 1 . 7-fold faster than that of an individual becoming infected at the age of 20 ., Next , we investigated if the tolerance parameter α differs across well-established human genetic polymorphisms associated with HIV control and disease progression—that is , resistance to HIV in the sense of evolutionary ecology ., For more than 850 individuals in our study population , information on HLA class I alleles and the CC chemokine receptor 5 ( CCR5 ) genotype was available ( see Materials and Methods ) ., In a first step , we focused on HLA-B alleles that have been found to associate with lower viral load—that is , with resistance 30 ., We wondered if these alleles are also associated with tolerance ., We found that protective HLA-B alleles are not associated with higher or lower tolerance in a univariate analysis ( F test: p\u200a=\u200a0 . 40; Figure 3C ) ., This is independent of how stringently we define protective HLA-B alleles ( see Materials and Methods and Figure S2 ) ., Thus , the protection these alleles confer can be fully attributed to the effect they have on viral load ., Higher HLA-C expression has been associated with better control of HIV viremia and slower disease progression 31–33 ., The expression level of HLA-C is reasonably predicted by classical HLA-C alleles , which are in strong linkage disequilibrium with a causal polymorphism in the 3 untranslated region of HLA-C 33 ., We could thus predict the HLA-C expression level for 850 individuals in our study population , of which 243 , 434 , and 173 had low , medium , and high expression , respectively ., We found that the tolerance parameter α does not vary significantly with HLA-C expression in a univariate analysis ., We also did not find any association of tolerance with protective HLA-B alleles and predicted HLA-C expression in a multivariate analysis including both factors together with sex and age at infection as covariates ., Another important polymorphism related to HIV acquisition and disease progression is located in the gene coding for the chemokine receptor CCR5 ., About 10% of Europeans carry a CCR5 allele with a 32 base pair deletion ( CCR5Δ32 ) ., Homozygous individuals are almost completely resistant to infection , while carriage of a single allele has been reported to be associated with slightly lower set-point viral load and slower disease progression 34 ., We divided the fraction of our study population , for which we had information on the CCR5 genotype , into individuals with ( n\u200a=\u200a163 , all heterozygous ) and without ( n\u200a=\u200a699 ) CCR5Δ32 ., There was no significant difference in tolerance between these two groups in a univariate analysis ., Again , we obtained the same result in a multivariate analysis including sex and age at infection as covariates ., The analyses above aimed at determining if known resistance genes also induce tolerance ., We found that they do not ., But what if there are yet unknown genes , unrelated to resistance , that confer tolerance ?, As first candidates for such tolerance genes , we considered HLA-B alleles irrespective of their protectiveness ., To assess if there are differences in tolerance associated with HLA-B , we adopted a mixed-effects modeling approach ., We combined the two HLA-B alleles of an individual into a genotype ( see Materials and Methods ) obtaining 375 unique genotypes in our study population ., The frequency distribution of the combined HLA-B genotypes is shown in Figure 4A ., In the mixed-effects models , we used HLA-B genotype as a random effect ., Specifically , we assumed the following relationship between CD4+ T-cell decline , ΔCD4 , and set-point viral load , V , in a univariate analysis: ( 2 ) The parameter characterizes the average tolerance in our study population , and denotes how the tolerance of genotype h deviates from this average ., We treated this parameter as a random effect—which means that we did not estimate it for each genotype but estimated the variance of its distribution ( see Text S1 ) ., We found significant variation in the random effect αh of HLA-B genotypes ., Compared to a model without this random effect with a likelihood ratio test , we obtained a significance level of p\u200a=\u200a0 . 0002 ., This variance is illustrated in Figure 4B: across HLA-B genotypes , tolerance differs approximately 2-fold and the relative standard deviation ( the standard deviation divided by the absolute value of the mean ) is 0 . 34 ., This variance in tolerances translates into an approximately 1 . 7-fold difference in the rate of disease progression for two randomly selected HLA-B genotype groups ., Restricting our analysis to genotypes represented by more than one individual yields an even larger and more significant random effect , and a multivariate analysis that includes sex and age at infection as covariates shows that these two variables do not confound our analysis ( see Text S1 ) ., Table 1 lists 5% ( n\u200a=\u200a18 ) of the HLA-B genotypes with the most extreme tolerance as predicted by the mixed-effects model ., The values in Table 1 are best linear unbiased predictions 35 , rather than estimates of tolerance parameters for each combined HLA-B genotype group , and should therefore be interpreted with care ., Figure S3 shows a histogram of the best linear unbiased predictions of tolerance for the HLA-B genotypes ., As outlined in Text S1 , we could not identify any association of tolerance with particular HLA-B alleles , suggesting that the effects of the two HLA-B alleles on tolerance depend on the specific combination of HLA-B alleles , rather than just on the sum of their effect ( see Figure S4 ) ., A case in point is the least tolerant genotype group “3501/3501” ., Carriage of this allele ( considering homo- and heterozygotes together ) is not associated with higher set-point virus load , faster CD4+ T-cell decline , or lower tolerance ., But HLA-B*3501 homozygotes display the most extreme departure from the average tolerance curve ., This is due to a very fast CD4+ T-cell decline in two individuals in this genotype group ., To further explore the importance of HLA-B allele combination on tolerance , we compared homozygous to heterozygous individuals ., Of the 923 individuals in our study population , for which we have information on the HLA-B alleles they carry , 39 were homozygous , displaying 14 unique genotypes ., A regression analysis of the CD4+ T-cell decline against set-point viral load with HLA-B homozygosity as a covariate confirmed a significant association of homozygosity with tolerance in univariate ( F test: p\u200a=\u200a0 . 00016 ) and multivariate analysis including sex and age at infection ( F test: p\u200a=\u200a0 . 00005 ) ., Figure 3D depicts the difference in tolerance between hetero- and homozygotes according to a univariate analysis ., Homozygotes have higher set-point viral loads than heterozygotes and are therefore expected to display faster CD4+ T-cell declines ., Figure 3D , however , shows that the CD4+ T-cell decline is in fact much faster in homozygotes than their set-point viral load predicts ., Quantitatively , the tolerance paramete α of homozygotes is −0 . 019 ( versus α\u200a=\u200a−0 . 012 in heterozygotes ) ., This difference in the tolerance parameter translates into a 1 . 6-fold faster rate of disease progression of homozygotes compared to heterozygotes with the same set-point viral load ., The tolerance difference between homo- and heterozygotes further supports the view that the effect of HLA-B alleles is not additive and refines our understanding of the well-established HLA-heterozygote advantage with respect to set-point virus load and disease progression 36 , 37 ., In contrast to previous studies on tolerance and resistance 17 , we did not find a trade-off—that is , a negative correlation—between resistance and tolerance across HLA-B genotype groups ( see Text S1 ) ., The lack of a correlation between tolerance and resistance suggests that there are no mechanistic or genetic constraints to display both traits ., If both tolerance and resistance mechanisms are costly , a trade-off could eventually evolve , but the co-evolutionary history between humans and HIV may have been too short for distinct resistant and tolerant lineages to separate ., However , we found a positive relation between tolerance and resistance across age ., As individuals get older they become less tolerant and less resistant ., We also looked for genome-wide associations with tolerance ., To this end , we defined a tolerance phenotype for each individual by calculating the residual in a quadratic regression between an individuals CD4+ T-cell decline and viral load , controlling for the age at infection ( see Figure S5A ) ., This analysis failed to identify any SNPs associated with tolerance ( Figure S5B ) ., It is important to note that this analysis , in addition to setting very stringent requirements for significance by correcting for multiple testing , also assumes additivity of allelic effects—that is , ignores a potential heterozygote advantage ., In summary , we presented the first formal tolerance analysis of a clinically relevant human infection ., HIV infection features well-established measures of pathogen burden and disease progression that are required for such an analysis ., The analysis consistently identified a subset of individuals that tolerate high viral load with minimal disease progression—the so-called viremic nonprogressors 26 , whose biological profile ( transcriptome , interferon response , gut microbial translocation ) is reminiscent of SIV infection in sooty mangabeys 26–28 ., But beyond this consistency with the tolerant profile of these four individuals , adopting the evolutionary ecology framework for tolerance allowed us to assign quantitative tolerance measures to well-defined groups of individuals and to statistically compare them ., In addition to investigating age- and sex-related differences in tolerance to HIV , we could , due to the wealth of information available for individuals in the Swiss HIV Cohort Study , test for potential associations with genes implicated in disease susceptibility and progression , such as HLA class I and CCR5 ., The finding that there is no difference in tolerance between the sexes challenges a previous report by Farzadegan et al . 29 , according to which females are less tolerant than males ., Just like Farzadegan et al . , we found that females have significantly lower viral loads , but do not differ in their disease progression ., In contrast to Farzadegan et al . , however , this pattern did not result in a significant difference in the relationship between disease progression and set-point viral load ., One reason for this discrepancy may be that Farzadegan et al . used data on AIDS diagnosis during a time window of observation , whereas we used CD4+ T-cell decline to measure disease progression ., Furthermore , Farzadegan et al . performed a survival analysis , whereas we performed a regression analysis ., Lastly , in contrast to our analysis , Farzadegan et al . did not adjust for the age at which individuals became infected ., For all these reasons , the previous and present analyses are difficult to compare and the discrepancy remains unresolved ., In all of the figures that show our data , it is apparent that the relationship between the set-point viral load and CD4+ T-cell decline is weak ., The noise in this relation is entirely consistent with previous studies 25 in which 5%–9% of the variation in the CD4+ T-cell decline could be explained by the set-point viral load ., The analysis we performed to identify variation in tolerance aimed at detecting differences in this relationship between different subgroups in our study population ., Given how noisy this relation is , it is remarkable that we could identify significant associations of host factors with tolerance at all ., In our study , we considered the most important host genes but disregarded the potential impact of virus genetics on tolerance ., The viruses harbored by the individuals in our study population differ by subtype ., Although viral subtypes are hypothesized to vary in virulence , this effect is difficult to ascertain due to usually unaccounted differences in the study populations 38 ., However , a large fraction of individuals in the Swiss HIV Cohort carry subtype B virus 39 , 40 ., We therefore do not expect the genetic variation of the virus to confound our analysis ., The framework for investigating tolerance we adopted for this study , despite its internal consistency , has its limits ., The parasite burden—central as the x-axis in our tolerance curve plots—is not simply an external factor affecting virulence but will itself be influenced by the host genotype and phenotype ., If we had virus dynamics models that described the entire course of HIV infection , the relationship between virulence and virus load could be mechanistically derived , and we would not have to rely on the statistical approach adopted here ., Such a comprehensive model has , however , been elusive to date 41 , mostly because the slow depletion of CD4+ T cells cannot be accounted for by HIV targeting and killing these cells ., Rather , a generalized immune activation in infected individuals is currently conceived to be at the heart of the mechanisms of pathogenesis 42 , and a straight-forward relationship between set-point virus load and CD4+ T-cell decline is unlikely to emerge from the probably complex dynamics ., Until a better dynamical understanding of HIV pathogenesis emerges , the low power of the set-point virus load to predict the CD4+ T decline 25 provides some justification of treating these two entities as independent ., Our analysis implicates HLA-B in modulating tolerance ., In particular , we established a tolerance advantage of HLA-B heterozygotes , providing an additional example of a benefit that host diversity affords against pathogens 36 , 43–46 ., Mechanistically , it is conceivable that certain HLA-B alleles cause faster disease progression without increasing viral load by modulating immunopathology , rather than leading to the killing of infected cells by cytotoxicity ., The higher tolerance of individuals , who contracted HIV at a young age , is likely to be explained by the higher thymic output of young individuals that can compensate infection-related CD4+ T-cell loss 47 ., Confirming or refuting these hypothetical mechanisms will be an important direction of future research on tolerance against HIV ., The Swiss HIV Cohort Study was approved by the local Ethics Committees of all participating centers , and written informed consent was obtained from the participants ., This project was approved by the Scientific Board of the SHCS as project 697 ., We used data from the Swiss HIV Cohort Study ( www . shcs . ch ) 48 ., Briefly , the study has enrolled more than 18 , 000 HIV-infected individuals to date ., Sociodemographic and behavioral data are recorded at entry to the study , in particular year of birth , gender , and the date of the last negative HIV test ., Laboratory and clinical data , including viral load and CD4+ T-cell count , are obtained at each semiannual follow-up visit ., Approximately 2 , 000 individuals have been genotyped in the context of previous genome-wide association studies 31 , 49 and/or at loci relevant for HIV acquisition and disease progression , such as those encoding the Human Leukocyte Antigen ( HLA ) class I genes and CCR5 ., We included individuals into our study , for whom viral load measurements and CD4+ T-cell counts were available , to reliably estimate the set-point viral load and CD4+ T-cell decline , as defined below ., We restricted our analysis to data obtained before antiretroviral treatment because the relationship between CD4+ T-cell count and viral load is dramatically altered during treatment ., To exclude the primary infection period , during which viral load and CD4+ T-cell count exhibit strong fluctuations , we discarded results obtained during the first 90 days after the estimated date of infection ., To exclude the late phase of the infection , during which viral load increases and fluctuates due to severe immunosuppression , we discarded measurements obtained when the CD4+ T-cell count was below 100 per µl ., Individuals were included if they had at least two eligible viral load results and three eligible CD4+ T-cell measurements at least 180 days apart ., After applying these inclusion criteria , our study population comprised 3 , 036 individuals ., For 837 , 923 , and 862 individuals , we had information on the HLA-A , -B , and -C alleles , respectively ., The CCR5Δ32 genotype was available for 862 individuals , whereas 852 individuals had genome-wide genotyping results ., Of the 923 individuals , for whom we had information on the HLA-B alleles , a large majority of 850 were of European ancestry ., Set-point viral load was determined as the geometric mean of the eligible viral load measurements in each individual ., Nondetectable viral loads were set to half the detection limit ., The change of CD4+ T-cell count over time was estimated as the slope in a linear regression of CD4+ T-cell count against the date at which they were determined ., Data S1 provides estimates of the set-point viral load and CD4+ T-cell declines for the 3 , 036 individuals included in our study ., We defined an HLA-B allele as “protective” if it has been found to associate with better HIV control and slower disease progression , according to table 1 of 30 ., In addition , we adopted alternative , more restrictive definitions , considering either only HLA-B27 or 57 , or only HLA-B*27:05 and *57:01 as protective ( see Figure S2 ) ., The HLA-C expression levels of the individuals in our study were predicted from the classical HLA-C alleles using data from table S1 in Kulkarni et al . 33 ., For each individual , a combined HLA-B genotype was defined by concatenating and sorting the four-digit alleles they carry ., An example for a genotype thus defined is “0702/3501” ., The statistical analysis is comprehensively described in Text S1 ., Here we just give a brief overview of the logic of our statistical procedures ., We regressed the change in CD4+ T cells over time , ΔCD4 , against the set-point viral load , V , using a least-square fitting algorithm assuming linear and nonlinear relationships ., Sex , age at infection , protectiveness of HLA-B alleles , carriage of CCR5Δ32 , predicted HLA-C expression levels , and HLA-B homozygosity were included into the regression analysis as covariates either individually or in combination ., Formally , we investigated the association of tolerance with a binary factor , such as sex or the carriage of protective HLA-B alleles , by decomposing the parameter α in the baseline model ( equation 1 ) : ( 3 ) Hereby , denotes the tolerance parameter for the subpopulation without the factor , and an offset associated with the factor ., Multiple factors were included into the statistical model by further decomposing the tolerance parameter: ., If a factor had more than two levels , one level was defined as the baseline and an offset parameter was added for each alternative level ., This was the case for HLA-C expression , which can be expressed at low , medium , and high levels ., Consequently , the models including HLA-C expression as a covariate feature two offset parameters ( and —see Text S1 ) ., Age at infection , a , being a continuous variable , was assumed to affect the tolerance parameter linearly: ( 4 ) In this expression , denotes the tolerance when contracting HIV at age 0 , and c describes the increase or decrease of tolerance per life year ., We assessed if a covariate significantly affected tolerance in two ways ., First , we checked if the offset associated with the covariate was significantly different from zero ., Second , we compared the models with and without the covariate with an F test or a likelihood ratio test ., In all cases , these two tests agreed ., Each factor was considered on its own in a univariate analysis and in combination with the other factors in multivariate analyses ( see Text S1 ) ., The coefficient of determination of a model , R2 , was calculated as one minus the ratio between the variance of residuals in the respective model fit and the variance in ΔCD4 50 ., Note that , because our models set the intercept to zero , the variance in ΔCD4 does not represent the residual sum of squares of any special cases of our models—that is , of any model nested in our models ., The inclusion criteria , calculation of set-point viral load and CD4+ T-cell decline , as well as the model fitting and comparisons were implemented and performed in the R language of statistical computing 51 ., Regression analysis was performed using the R-functions lm ( ) and , for the mixed effects models , lme ( ) in the R-package nlme ( ) ., The F tests and likelihood ratio tests were performed using the R-function anova ( ) ., For the genome-wide association study , we assigned a tolerance phenotype to 852 individuals in our study population , for whom we had genomic information and who were of European ancestry ., This phenotype was calculated as the deviation of the individuals set-point viral load and CD4+ T-cell decline from the average tolerance relationship of the population ., Because the age at infection was associated very strongly with tolerance , we calculated the deviation from an age-controlled tolerance relationship ( see Figure S5A ) ., Study participants had been genotyped in the context of previous studies 31 , 49 using Illumina 550 or 1 M chips , and genome-wide SNPs were imputed using the 1000 Genomes Project CEU panel as a reference ., After quality control and exclusion of nonvariable SNPs , seven million variants were available for association testing ., We used linear regression to test for association between each SNP and the tolerance phenotype , including sex and the coordinates of the first five principle components of an EIGENSTRAT analysis 52 as covariates ., We used Bonferroni correction to control for multiple testing ( p threshold =\u200a5×10−8 ) . | Introduction, Results, Discussion, Materials and Methods | In ecology , “disease tolerance” is defined as an evolutionary strategy of hosts against pathogens , characterized by reduced or absent pathogenesis despite high pathogen load ., To our knowledge , tolerance has to date not been quantified and disentangled from host resistance to disease in any clinically relevant human infection ., Using data from the Swiss HIV Cohort Study , we investigated if there is variation in tolerance to HIV in humans and if this variation is associated with polymorphisms in the human genome ., In particular , we tested for associations between tolerance and alleles of the Human Leukocyte Antigen ( HLA ) genes , the CC chemokine receptor 5 ( CCR5 ) , the age at which individuals were infected , and their sex ., We found that HLA-B alleles associated with better HIV control do not confer tolerance ., The slower disease progression associated with these alleles can be fully attributed to the extent of viral load reduction in carriers ., However , we observed that tolerance significantly varies across HLA-B genotypes with a relative standard deviation of 34% ., Furthermore , we found that HLA-B homozygotes are less tolerant than heterozygotes ., Lastly , tolerance was observed to decrease with age , resulting in a 1 . 7-fold difference in disease progression between 20 and 60-y-old individuals with the same viral load ., Thus , disease tolerance is a feature of infection with HIV , and the identification of the mechanisms involved may pave the way to a better understanding of pathogenesis . | When confronted with pathogens , hosts can either evolve to fight them or learn to live with them ., The first of these two strategies is called “resistance” and the second “tolerance” ., In the context of HIV , many genes conferring resistance have been identified , but no tolerance genes are known ., Using statistical techniques originating from plant ecology , we analyzed data from an HIV cohort to look for differences in tolerance between HIV-infected individuals and tested whether they go hand in hand with genetic differences ., We found that younger people are more tolerant to HIV infection ., We also observed that individuals who carry two different alleles of HLA-B , an important immunity gene , are more tolerant ., These findings add to our understanding of how hosts tolerate infections and could open new avenues for treating infections . | infectious diseases, hiv infections, evolutionary ecology, medicine and health sciences, major histocompatibility complex, ecology, clinical immunology, population modeling, aids, biology and life sciences, immunology, infectious disease modeling, computational biology, evolutionary biology, evolutionary immunology, population biology, viral diseases | Title: Human tolerance against HIV An evolutionary ecology perspective on clinical data reveals that human traits can affect how well an individual tolerates HIV infection, and identifies host immunity factors associated with disease tolerance. |
journal.pntd.0000227 | 2,008 | Lower Expression of TLR2 and SOCS-3 Is Associated with Schistosoma haematobium Infection and with Lower Risk for Allergic Reactivity in Children Living in a Rural Area in Ghana | In the last few decades , allergic diseases have become a major health burden in the western world ., Although these disorders clearly have a genetic component , their rapid change in prevalence points to environmental conditions that have changed during this time frame ., In the same time frame , there has been a decrease in exposure to microbial products as a result of changing lifestyle with , among others , improved sanitation and access to clean water ., Interestingly , in the developing world , the prevalence of allergies is relatively low , particularly in rural areas , where exposure to infectious agents is high ., There is increasing evidence that exposure to pathogen-derived compounds influences the maturation of the immune system and therefore the balance reached between pro- and anti-inflammatory responses , such that Th2 responses are kept under control when allergens are encountered ., In rural areas in the developing world , chronic helminth infections are highly prevalent ., These infections not only result in skewing of the immune responses towards Th2 , but also induce the higher production of anti-inflammatory molecules such as IL-10 to prevent the elimination of helminths , which at the same time protect the host against the pathological consequences of excessive inflammation 1 ., Such an anti-inflammatory environment induced by chronic helminth infections might modulate immune responses to other antigens ., For example , chronic infection with schistosomes or Onchocerca was shown to modulate the immune response to tetanus toxoid following vaccination 2 , 3 ., Epidemiological studies have revealed both positive and negative associations between helminth infections and allergies ( reviewed in 4 ) ., It is thought that severe , chronic infections are often associated with suppression of allergic reactivity ., For example chronic infections with intestinal helminth , such as with hookworm , have been shown to suppress allergic diseases 5 , 6 ., These observations have been confirmed for schistosomiasis , demonstrating lower skin reactivity to allergen in infected individuals 7 , 8 ., Additionally , removal of helminths by long-term anti-helminth treatment in Venezuelan or Gabonese children resulted in increased atopic reactivity to house dust mite 9 , 10 , even though a shorter anti-helminth treatment did not show an effect on atopy in one study 11 ., In a population of rural Ghanaian school children a negative association was found between infection with Schistosoma haematobium and skin reactivity to mite allergen ( Obeng et al , submitted for publication ) ., Within this study we aimed to identify the molecular mechanisms by which schistosome infections may modify immune responses and modulate inflammatory reactions such as atopy ., To address this we selected two groups of genes that have been described previously to play a role in allergic diseases ., Toll-like receptors ( TLRs ) have been shown in several studies outside Africa to change in expression levels following exposure to microorganisms 12–15 ., In a European study these molecules were linked to allergy: children of farmers in Alpine regions , exposed to high microbial burden and with a low prevalence of atopy , had altered levels of TLR2 16 , suggesting that exposure to microorganisms might modulate the innate immune system and thereby suppress the development of allergic disorders ., The molecules suppressor of cytokine signalling ( SOCS ) -1 and SOCS-3 have recently been described and reported to be involved in TLR signalling and inflammatory diseases 17–22 ., Elegant studies by Kubo and co workers in animal models have shown SOCS-3 to be involved in regulation of immune responses in allergic disease 23 , 24 ., Given that helminth products have been shown to modulate cells of the innate immune system and to interfere with pathways that are activated via TLR stimulation 15 , 25 , we asked whether in an area in Africa where helminth infections are highly prevalent and allergic disorders are low , we can find molecular pathways that may explain the relationship ., To this end , gene expression not only of TLRs but also of molecules such as SOCS-1 and SOCS-3 , involved in downstream signalling , were studied in whole blood samples of rural Ghanaian school children ., The results of this study showed that high expression of TLR2 and SOCS-3 was associated with allergic skin reactivity , whereas helminth infection was associated with lower expression levels of TLR2 and SOCS-3 , providing a potential regulatory link between helminth infection and allergies at the molecular level ., The study population consisted of schoolchildren between 5 and 14 years of age ., Children whose parents consented by signing or thumb printing an informed consent form were registered to participate in a large study on allergy and parasitic infections ( B . B . Obeng et al , manuscript submitted ) ., The Institutional Review Board of the Noguchi Memorial Institute for Medical Research , Accra , Ghana approved the study ., Skin reactivity to mite was negatively associated with S . haematobium infection ( OR 0 . 5 , 95% CI 0 . 2–1 . 0 , p\u200a=\u200a0 . 05 ) , particularly in areas where prevalence of schistosomiasis is high ( OR 0 . 3 , 5% CI 0 . 1–0 . 9 , p\u200a=\u200a0 . 04 ) ., Blood samples from children from two rural schools with high prevalence of S . haematobium infection were used for RNA isolation to study gene expression ., In these schools , the reactivity to house dust mite was low ( 9% ) compared to school children from Accra ( 15% ) , free of any helminth infections , and with a relatively high socioeconomic status ., The study subjects were fist selected randomly , one out of three children from whom blood samples were available were selected ( 107 children ) ., In order to increase power , all skin prick test ( SPT ) positive children from these schools were added to our randomly selected subjects , along with randomly selected SPT negative children ( 16 children in total ) , resulting in a group of 123 children ( Table 1 ) ., The participants were given specimen bottles and were asked to collect a fresh stool and urine sample for the detection of helminth infections ., Stool examination was performed by the Kato-Katz method for the detection of hookworm and trichuris , and the total number of eggs was calculated per gram of faeces ., Urine samples were used for the detection of S . haematobium by passing 10 ml of urine through a filter with 10-micron pore size ., A subject was considered positive for helminth infection if eggs of any of the helminth species were detected ., Blood samples were collected from all participants for the detection of the malaria infection by Giemsa-stained thick smear ( GTS ) examination ., The immediate hypersensitivity skin prick test with inhalant allergen extracts was performed by using the standard prick method on the volar surface of the right forearm with the standardized extracts of 6 allergens; Dermatophagoides pteronyssinus ( Der P ) , Dermatophagoides farinae ( Der F ) , cat , dog , peanut and grass mix ( HAL Allergen Laboratories , The Netherlands ) ., Histamine dihydrochloride ( 1/1000 ) and glycerinated saline solution were used as positive and negative controls , respectively ., The wheal diameter was measured after 15 minutes and the result considered positive when the wheal size was at least 3 mm in diameter , in the absence of significant reactivity of the diluent negative control ., None of the children were taking anti-allergic medicine that might interfere with SPT or had ever been treated with specific immunotherapy ., Since most of the positive reactions seen were against mite allergen , we have focused on children having a positive skin test for mite allergen ., Serum levels of total IgE were measured by the enzyme linked immunosorbent assay ( ELISA ) as described before 26 ., Results were expressed as international units per ml ( IU/ml ) ., Serum levels of house dust mite ( HDM ) antibodies were determined by radio allergosorbent test ( RAST ) as described previously 27 ( CLB , Amsterdam , The Netherlands ) ., Results were expressed as international units per ml ( IU/ml ) ., One IU is 2 . 4 ng IgE ., Subjects were considered sensitised when concentrations of specific IgE of more than 0 . 7 IU/ml were measured ., Immediately after venapuncture into heparinised tubes , 0 . 8 ml of whole blood was added to 3 . 6 ml of Nuclisens lysis buffer ( Biomérieux , Boxtel , The Netherlands ) to stabilise the RNA ., Samples were stored for a maximum of two weeks at 4°C , after which they were put at −80°C for long-term storage ., The Nuclisens Isolation kit ( Biomérieux ) was used for the isolation of total nucleic acid ( approximately 1 ml of blood mixed with lysis buffer per isolation ) according to the manufacturers instructions ., Genomic DNA was removed by treating the samples with RNAse-free DNAse ( Invitrogen , Breda , The Netherlands ) for 30 minutes at 37°C , followed by the Nuclisens isolation procedure to isolate the purified RNA ., RNA was isolated from the same samples before and after one year of storage at −80°C , and mRNA levels of several genes of interest were compared ., There was no difference in gene expression indicating that mRNA was stable in this buffer for at least one year at −80°C , and that the procedure was consistent ., From a subset of the donors , PBMC were isolated and monocytes and T cells were separated by subsequent labelling and magnetic cell separation of cells with CD3 and CD14 Microbeads ( Miltenyi Biotech , Germany ) ., Fractionated monocytes and T cells were mixed with Nuclisens lysis buffer ., Fluorescence activated cell sorting ( FACS ) of the isolated cells indicated that these fractions were at least 90% pure ., The unlabeled cell fraction depleted for monocytes and T cells was also collected and mixed with lysisbuffer ( remaining cell fraction ) ., Samples were stored and RNA was isolated as described for whole blood samples ., The percentages of monocytes and T cells in all donors were determined by flow cytometry of the PBMC prior to the isolation procedure ., Reverse transcription of RNA was performed using moloney murine leukaemia virus reverse transcriptase ( M-MLV RT ) ( Invitrogen ) ., Samples without RT were regularly taken along to control for genomic DNA contamination ., Gene expression was assessed with real-time quantitative PCR ( Prism 7700 , Applied Biosytems ) ., PCR reactions were performed in duplicate in accordance with the TaqmanTM assay instructions using Taqman probes and qPCR Core kit reagents ( both Eurogentec , Seraing , Belgium ) ., Gene expression was normalized to the housekeeping gene 18S rRNA and calculations were performed as described 28 ., Analysis of the expression of 8 different housekeeping genes in a subset of the samples indicated that 18S rRNA was a stable housekeeping gene in our samples ., Sequences of primers and probes have been obtained from Dr . Roger Lauener ( TLR2 , TLR4 , 18S rRNA , 16 ) and from Dr . Masato Kubo ( SOCS-1 , SOCS-3 , 23 ) ., IgE mRNA levels were determined by a primer and probe set specific for the CH1 region of IgE ( forward primer 5′-CAA TGCCACCTCCGTGACTC-3′ , reverse primer 5′-CGTCGCAGGACGACTGTAAG-3′ and probe 5′-ATCGTCCACAGACTGGGTCGACAACAAA-3′ ) ., For each gene , after normalisation for the housekeeping gene , the donor with the lowest expression was set to 1 ., For the expression levels of TLR2 and SOCS-3 in isolated cell subsets , expression of SOCS-3 or TLR2 in T cells was set to 1 for each donor ., Whole blood from 5 donors was diluted 1∶1 with RMPI 1640 medium ( Gibco ) and stimulated for 16 hours and 24 hours in 96-well round bottom plates with medium , 10 µg/ml SEA ( schistosomal egg antigens ) , 100 ng/ml LPS ( Sigma ) , 100 µg/ml poly I:C or 5 µg/ml TNF-α ( Sanquin , The Netherlands ) ., After 16 hours the blood was mixed with ABI Lysis buffer ( Applied Biosystems ) and RNA was extracted using the ABI6100 according to their protocol ( Applied Biosystems ) ., cDNA synthesis and quantitative PCR for TLR2 was performed as described above ., 24 hours after stimulation , cells were mixed with FACS lysing solution ( BD Biosciences ) to lyse the erythrocytes , washed with PBS and cells stained with anti-TLR2 PE ( clone T2 . 5 , eBioscience ) ., Flow cytometric analyses were performed using a Becton Dickinson FACSCalibur flow cytometer ( BD Biosciences ) and analysed using FlowJo analysis software ( Tree Star Inc . ) ., Association of total and mite-specific IgE with skin reactivity to mite was analysed by logistic regression of log-transformed values ., The association of gene expression with skin reactivity or helminth infection was analysed by logistic regression using a value of each gene to separate high and low gene expression , since the association between gene expression and allergen reactivity might not be a linear one ., This value was based on the geomean of the relative expression data ( low expression: below geomean; high expression: above geomean ) ., The regression analysis was performed adjusting for age , sex and helminth infection ( Table 2 ) or for age , sex and skin reactivity to mite ( Table 3 ) ., The comparison of the non-adjusted means of gene expression between skin prick positive and negative children and between helminth-infected and non-infected children was determined by the non-parametric Mann-Whitney test ., Correlation between mRNA and surface TLR2 levels and between mRNA and serum IgE levels was compared using the non-parametric Spearmans correlation test ., The study population here originates from two schools selected from a large study on allergy and parasitic infections ., The schools were in a rural area highly endemic for helminth infections ( see Material and Methods section ) ., Fifty-four percent of the children were infected with at least one helminth species ( Table 1 ) ., As indicated in Table 1 , the most prevalent helminth species was Schistosoma haematobium , followed by hookworm , Ascaris lumbricoides and Trichuris trichiura ., In the population where mRNA expression was analyzed ( see Materials and Methods ) , 14 out of 74 schistosome negative children had a positive skin reaction to mite ( 19% ) , whereas 5 out of 46 schistosome-infected children were SPT positive for mite ( 11% ) resulting in a significant negative association between infection with S . haematobium and atopy ( OR 0 . 26 0 . 07–1 . 00 , p\u200a=\u200a0 . 05 , adjusted for age , sex , school and levels of mite IgE ) ., Additionally , the odds ratio of the association between the log of mite IgE and atopy is clearly lower in children infected with helminths ( OR 5 . 8 1 . 0–33 . 2; p\u200a=\u200a0 . 05 ) compared to the odds ratio in non-infected children ( OR 18 . 7 2 . 4–145 . 8; p\u200a=\u200a0 . 005 ) ., The level of mite-specific IgE was a strong determinant for the risk of positive skin reactivity to mite ( Table 2 ) ., In contrast , the level of total IgE was not significantly associated with atopy ., However , total IgE was associated with atopy in the children that were not infected with helminths ( OR\u200a=\u200a5 . 6 , CI 95%: 1 . 0–30 . 9; p<0 . 05 ) ., In order to evaluate whether the in vivo status of the immune system could be evaluated by analysing the mRNA expression in whole blood , RNA was isolated from peripheral blood samples collected in our study population ., In agreement with IgE serum levels , the levels of IgE mRNA were significantly higher in helminth-infected children compared to non-infected children ( Figure 1A ) , and a high correlation was seen between the mRNA expression and serum IgE levels ( r\u200a=\u200a0 . 58 , p<0 . 001; Figure 1B ) ., In addition the mRNA levels of TLR2 were compared to surface TLR2 protein expression using flow cytometry in whole blood samples ., There was a strong correlation between TLR2 mRNA expression and TLR2 surface expression ( r\u200a=\u200a0 . 58; p<0 . 001 ) , validating the use of mRNA levels measured in whole blood samples as a reflection of the measured immunological events in vivo ., In Ghanaian school children living in an area highly endemic for parasitic infections , there was a significantly higher expression of TLR2 in subjects with positive skin reactivity to house dust mite ( Figure 2A ) ; high expression of TLR2 doubled the risk of atopy ( OR 2 . 6 , Table 2 ) , whereas there was no such association for TLR4 and skin reactivity ( OR 0 . 9; Table 2 and Figure 2B ) ., Children with high expression of SOCS-1 or SOCS-3 had a significantly increased risk for skin reactivity ( OR 5 . 8 and 4 . 4 , respectively , Table 2 ) ., Both SOCS-1 and SOCS-3 expression were significantly elevated in those who were skin prick positive to house dust mite compared to non-atopic children ( Figures 2C and 2D ) ., To determine the source of TLR2 and SOCS-3 expression in whole blood samples , monocytes and T cells were isolated from five donors ., Although the mRNA expression of TLR2 was high in monocytes ( 63 to 275-fold higher than in T cells , Figure 3A ) , correction for the percentages of monocytes and T cells in the peripheral blood mononuclear cells , indicated that monocytes , T cells and other cells contributed similarly to the TLR2 expression measured in whole blood ( Figure 3C ) ., In contrast , the mRNA expression of SOCS-3 could clearly be attributed to the T cell fraction with little contribution from monocytes or other cells ( Figure 3B–C ) ., TLR expression can be altered following exposure to ligands expressed by microorganisms and parasites ., Helminth parasites carry signature molecules that can interact with TLRs and therefore could affect their expression ., The expression of both TLR2 and TLR4 genes was lower in children with a helminth infection; the effect being more prominent for TLR2 ( Figures 4A and 4B ) ., Indeed , infection with helminths predicted low expression of TLR2 and , to a lesser extent , of TLR4 ( Table 3 ) ., Two major helminth species prevalent in the study area were Schistosoma haematobium and hookworm ., The mRNA expression of TLR2 in helminth-infected children was significantly lower only in children infected with S . haematobium , which was associated with low expression of TLR2 ( Table 3 ) , whereas no such association was found for TLR4 ., Malaria infection was associated with higher levels of TLR2 expression ( not shown ) , and adjustment for malaria infection did not change the association between helminth infection and TLR2 expression ., Thus , induction of lower expression of TLR2 by infection with schistosomes seems to be specific ., The analysis of the expression levels of SOCS-1 and SOCS-3 revealed that children infected with helminths had significantly lower expression levels of SOCS-3 , but not of SOCS-1 ( Figures 4C and 4D ) ., Helminth positivity in a child was associated with low gene expression of SOCS-3 , but not of SOCS-1 ( Table 3 ) ) ., As for TLR2 , low expression of SOCS-3 was associated with S . haematobium rather than hookworm infection ., Using gene expression profiles in whole blood from children living in a rural area in Ghana , we found that high expression of TLR2 , SOCS-1 and SOCS-3 mRNA was associated with positive skin reactivity to house dust mite ., Presence of Schistosoma haematobium infection , reported to decrease the risk of atopy 7 and observed in the current study , affected the expression levels of TLR2 and SOCS-3 , which were significantly lower in infected children ., There are few studies that have looked at the association of TLR expression and allergy , and those that have , are all in European populations 29 , 30 ., Of these , only one study has investigated the levels of TLR expression in an age group comparable to our study ., European children living on farms and reported to have a lower risk of developing atopic disorders were shown to have higher expression levels of TLR2 and CD14 , compared to non-farmer children 16 ., As farmer children would be expected to be exposed to a high burden of environmental microorganisms , the results are in contrast to the lower expression of TLR2 in our helminth infected subjects compared to uninfected Ghanaian school children ., Children living in a rural area in Ghana are expected to have exposures that are higher in intensity , and different in nature in terms of the sort of microorganisms and parasites , compared to European farmer children ., Moreover , the European study did not examine the TLR2 levels in atopic and non-atopic individuals as we do here , showing that atopy was associated with high expression of TLR2 ., Our data support a role for suppression of atopy by current infection with a systemic helminth , Schistosoma haematobium ., The finding that S . haematobium infected children show a lower expression of TLR2 gene , is supported by the results that baseline expression levels of TLR2 protein were also shown to be lower in individuals infected with another systemic helminth infection , the filarial nematode , Wuchereria bancrofti 31 , 32 ., Importantly , lower expression of TLR correlated with a lower expression of co-stimulatory molecules such as CD80 and CD86 and lower production of the inflammatory cytokines IL-6 and TNF-α 32 ., Our results also indicated an association of skin reactivity to house dust mite with higher gene expression for both SOCS-1 and SOCS-3 ., SOCS genes are involved in the pathogenesis of several inflammatory diseases ., They are induced upon cytokine signalling or by stimulation of TLR and limit the production of inflammatory cytokines 33 ., In murine models , transgenic over expression of SOCS-3 has been shown to mediate and maintain allergic responses 23 ., Furthermore , T cell expression of SOCS-3 , but not of SOCS-1 , was associated with atopic disease in humans and increased with disease severity 23 ., These results suggest that SOCS-3 is involved in Th2 skewed responses ., However , although helminth infections are clearly associated with Th2 responses , we have found that SOCS-3 expression is decreased in helminth-infected children ., Cell subset analysis indicated that T cells were the main source of SOCS-3 mRNA leading us to conclude that in helminth infected children , with strong Th2 responses , SOCS-3 expression is low in T cells ., So although both allergic disorders and helminth infections are characterized by Th2 responses , SOCS-3 is associated with allergic disorders , but not with helminth infection ., This would suggest that in allergic subjects , with high expression of SOCS-3 , Th2 responses are associated with inflammation; whereas in allergic subjects with a helminth infection and consequent low expression of SOCS-3 , Th2 cells are not associated with inflammation ., Interestingly , a recent report using T cell-specific SOCS-3 conditional knockout mice indicated that in the absence of SOCS-3 expression , the levels of typical Th2 cytokines in peripheral CD4+ T cells were either unaffected ( IL-5 ) or only slightly lower ( IL-4 ) , whereas following T cell stimulation , the production of the anti-inflammatory cytokines , IL-10 and TGF-β1 , were significantly higher ., The abolition of SOCS-3 expression in T cells ameliorated ovalbumin-induced airway hyperresponsiveness in vivo 34 ., Furthermore , CD4+CD25+foxp3 positive regulatory T cells were shown to have low SOCS-3 expression as compared to Th2 cells , indicating that low SOCS-3 expression in T cells is associated with suppressive function 35 ., These data raise the possibility that a decrease in SOCS-3 T cell expression by helminth infection might shift the balance towards a modified Th2 response 36 with a more anti-inflammatory function , and thereby might suppress allergic inflammation 1 ., As malaria infection was prevalent in our study area , we looked at this infection and found that it had no effect on atopy ( multivariate analysis , data not shown ) and interestingly found that malaria infection was associated with a higher expression of TLR2 ( data not shown ) , indicating that different pathogens might induce different regulation of TLR expression ., Indeed , other protozoa such as Entamoeba histolytica and Trypanosoma cruzi inhibit immune responses by down-regulating TLR2 expression 37 or signalling via TLR2 38 , 39 ., There are numerous studies supporting either up- or downregulation of TLR2 and TLR4 , depending on the stimulus and cell type studied 40 ., Both an increase and a decrease in TLR2 expression might reflect repeated TLR stimulation , the direction as well as the downstream signaling pathways being dependent on the type of pathogen ., Alternatively , the cytokine environment might influence the expression of TLR ., Th1 cytokines such as IFN-γ seems to increase expression levels of TLR 41 , whereas Th2 cytokines as IL-4 and IL-13 , abundantly present in helminth infected individuals , downregulate TLR expression and function 42 , 43 ., Thus , the exposure of European farmers to Th1 inducing agents might be reflected in higher TLR2 expression , whereas in our subjects the exposure to Th2 inducing agents might lead to low TLR2 expression ., The relationship between TLR2 and SOCS-3 expression might not be a direct one ., In rural Ghana , helminth infection is associated with low TLR2 as well as low SOCS-3 expression , whereas the expression of TLR2 is high in a European rural area ., If TLR2 expression is merely the result of exposure to pathogens , and low SOCS-3 expression is associated with protection from allergy , it would be of interest to know whether SOCS-3 is also lower in the European farmers environment despite a higher TLR2 expression ., In summary , ex vivo whole blood analysis of mRNA profiles in children infected with helminths compared to non-infected children living in a rural area in Ghana have shown that chronic helminth infections are associated with a lower expression of TLR2 and SOCS-3 ., The difference in the expression of SOCS-3 in helminth infected and uninfected children might result from the interaction of helminth derived molecules with the immune system leading to modulation of downstream signalling and induction of “modified” Th2 cells ., Larger epidemiological studies will be needed to be able to test this hypothesis directly and to confirm that helminths modify the development of allergy by modulating the expression levels of TLR2 and SOCS-3 . | Introduction, Methods, Results, Discussion | Helminth infections are prevalent in rural areas of developing countries and have in some studies been negatively associated with allergic disorders and atopy ., In this context little is known of the molecular mechanisms of modulation involved ., We have characterized the innate immune responses , at the molecular level , in children according to their helminth infection status and their atopic reactivity to allergens ., The mRNA expression of several genes of the innate immune system that have been associated with microbial exposure and allergy was examined in 120 school children in a rural area in Ghana ., Helminth infections were common and atopy rare in the study area ., The analysis of gene expression in ex vivo whole blood samples reflected the levels of corresponding proteins ., Using this approach in a population of school children in whom the presence of Schistosoma haematobium infection was associated with protection from atopic reactivity , we found that the level of TLR2 and SOCS-3 , genes associated with atopy in the children , were significantly downregulated by presence of S . haematobium infection ., S . haematobium infections modulate the expression of genes of the innate immune system ( TLR2 and SOCS-3 ) ; these are genes that are associated with increased allergic inflammatory processes , providing a molecular link between the negative association of this infection and atopy in rural children in Ghana . | Inflammatory diseases such as atopic disorders are a major health problem in the Western world , but their prevalence is also increasing in developing countries , especially in urban centres ., There is increasing evidence that exposure to a rural environment with high burden of compounds derived from parasites and microorganisms is associated with protection from atopic disorders ., Since urbanisation is progressing at a rapid pace , particularly in less-developed nations , there is a need to understand the molecular processes that control the progress towards the development of allergic diseases in developing countries ., In this study we have examined a population of school children living in a rural area of Ghana , where helminth ( worm ) infections are prevalent and associated with protection from skin reactivity to house dust mite ., Blood samples were collected from these children and analysed for the expression levels of several genes involved in the development of a pro allergic immune system ., The results point at a potential molecular link that might explain the negative association between schistosome infections and allergies . | immunology/immunomodulation, immunology/leukocyte signaling and gene expression, immunology/allergy and hypersensitivity, infectious diseases/helminth infections | null |
journal.pcbi.1003450 | 2,014 | Epigenetics Decouples Mutational from Environmental Robustness. Did It Also Facilitate Multicellularity? | Understanding the evolution of major transitions in the complexity of organisms remains one of the key challenges in modern biology 1 , 2 ., In particular , the transition to multicellularity required the evolution of several innovations at the molecular level in order to satisfy three key requirements: cell-to-cell adhesion , cell-to-cell signaling , and cellular differentiation 3 , 4 ., Such molecular innovations can often be facilitated by genomic duplication and subsequent specialization 5 as well as other evolutionary processes such as exaptation 6 , 7 and coevolution 8 ., In the case of cellular differentiation , the evolution of epigenetic gene regulation is arguably the most important; enabling molecular innovation during the expansion of the Metazoa 9 , 10 ., Of course , molecular innovations are also subject to multiple constraints which may be imposed externally through the environment 11 or internally , for example as a consequence of the developmental process 12 ., Here we will be concerned with robustness as an evolved internal constraint ., Robustness in biological systems is the property of persistent behavior despite genetic and environmental insults ., Previous studies , using gene regulatory network models , have shown that networks will evolve robustness to genetic mutations under conditions of stabilizing selection 13 , 14 ., This result has been experimentally verified in RNA viruses 15 , yeast 16 , 17 , and in the process of RNA folding 18 ., In addition to genetic mutations , organisms are exposed to environmental changes ., Previous studies using gene regulatory network models have shown that environmental and mutational robustness are positively correlated and are therefore expected to increase together under stabilizing selection 16 , 17 , 18 , 19 , 20 , 21 ., Furthermore , studies exploring robustness of miRNA sequence have shown that mutational robustness develops directly in response to evolving environmental robustness 22 ., Indeed computational models of cell differentiation also show the presence of robustness 23 ., However , invariance to the environment poses an obstruction to cell differentiation in multicellular organisms where internal environmental factors dictate cell fate decisions ., Highlighting the Metazoan cell differentiation dependence on the environment is recent work showing that changes in a small number of key growth factors is capable of altering cell fate decisions 24 , 25 ., For example , changes in expression of ct4 , Sox2 , Klf4 and c-Myc can drive conversion of fibroblasts to cardiomyocytes 26 ) ., Furthermore , the developmental impact of environmental sensitivity can be observed in the developing human fetus which is most vulnerable to environmental chemicals such as alcohol within the first few weeks of pregnancy 27 , 28 , 29 ., Therefore , how did multicellular organisms develop sensitivity to the internal environment , promoting cell differentiation , while retaining mutational robustness ?, The available evidence suggests that the transition to multicellularity was accompanied by major innovations in epigenetic regulation 30 , 31 , 32 ., Indeed chromatin states are in large part responsible for the gene expression differences across cell types 33 , 34 , 35 , 36 ., Post-translational modification of histones alters chromatin structure to encourage or repress transcription ., A key group of proteins responsible for marking regions for transcriptional repression during development are the Polycomb Group Proteins ( PcGs ) ., Early studies elucidated the general functionality of this protein group in developing Drosophila embryos ., In particular it was found that the chromosomal regions targeted by PcGs were transcriptionally repressed only if genes in the region were exhibiting low levels of expression when the PcGs became active 37 ., In this manner the PcGs were found to be responsible for turning off discrete sets of genes in different cell types depending on expression levels during early development ., For example , MyoD , a transcription factor required for myogenic commitment , is unable to access its binding sites in non-myoblast cells due to PcG dependent methylation 38 ., In addition , it has been shown that activation of muscle-specific genes in the vicinity of the PcG binding site prevent the PcGs from hypermethylating the site , thus allowing MyoD to exert transcriptional activation effects ., This functionality has motivated speculation that PcGs may have aided in the transition from a unicellular to a multicellular world by promoting differential expression in cell differentiation 39 , 40 ., Supporting this hypothesis , evolutionary analysis of the PcG Polycomb Repressive Complex 2 ( PCR2 ) has revealed that homologs of the core components ( E ( Z ) , ESC , Su ( z ) 12 , and Nurf55 ) existed prior to multicellular lineages but were rarely found present as a functional complex in single cell organisms ( although it is likely the last common unicellular ancestor of Metazoa did have all the components in place ) 39 , 40 , 41 ., In addition , Saccharomyces cerevisiae and other unicellular fungi with multicellular ancestors do not have the full set of functional homologs , correlating the loss of PcGs with reversal of multicellularity 39 ., To explore how a dynamic epigenetic process such as chromatin modification affects robustness and cell differentiation we have extended a well-established gene regulatory network model 13 , 42 with an epigenetic mechanism modeled on the Polycomb system ., In accordance with previous results we find that in the absence of an epigenetic mechanism both mutational and environmental robustness co-evolve by increasing together ., However , with the introduction of the Polycomb mechanism we see a decoupling of environmental and mutational robustness ., Mutational robustness still increases under stabilizing selection in concordance with experimental results but environmental robustness decreases , thus increasing responsiveness to the environmental cues ., In order to evaluate the capacity for cell differentiation in the model , we quantified the ability for producing alternative steady states ( outputs ) in response to novel environmental conditions ( inputs ) ., Consistent with the increase in environmental sensitivity we found that the Polycomb mechanism greatly facilitated the ability to create new input/output mappings , suggesting a strongly increased capacity for generating alternative cell fates ., Our results suggest a clear link between epigenetic regulation and cell differentiation in that the epigenetic mechanism allows a gene regulatory network to be altered dynamically , effectively creating multiple networks out of a single regulatory architecture ., In order to study the evolution of a Polycomb-like epigenetic mechanism we extended an established model of evolution in gene regulatory networks 13 , 42 ., Briefly ( see Methods for details ) , the model functions on two levels: population dynamics and gene regulatory network ( the genotype-phenotype mapping ) ., At the lower level of the genotype-phenotype mapping , the genotype of each individual is represented as a gene regulatory network of genes ., Gene expression dynamics are initiated by an input vector , leading to a steady state of length this defines the phenotype ( individuals not reaching steady state have zero fitness ) ., At the population dynamics level individuals undergo iterations of mutation , reproduction and selection ., We measure mutational robustness as described previously 13 , 14 , 43 by randomly mutating an entry in the interaction matrix ( of size ) and comparing the effect on the phenotype to that for the unmutated matrix ., Following Ciliberti et al 19 , we measure environmental robustness by introducing random changes into the input vector and similarly considering the effect on the phenotype ., Epigenetic regulation through chromatin remodeling is postulated to be a key mechanism through which a single genome can dynamically change gene expression to produce distinct stable cell types 30 , 31 , 32 ., To determine the effect of epigenetic mechanisms on the two distinct forms of robustness we incorporated Polycomb group ( PcG ) -like activity into the gene regulatory model ., Here , we assume that Polycomb is expressed beginning at time during development ., Susceptibility to Polycomb for each gene ( representing the presence of cis-acting Polycomb Response Elements ) is determined by such that from time onwards , the expression of each gene is repressed by the Polycomb protein if and its expression level falls below a threshold level ., This behavior is modeled upon the known function of the Polycomb Repression Complex 1 ( PRC1 ) in the Drosophila embryo where the Hox genes ( whose initial expression is determined by transiently expressed Gap genes ) are permanently repressed by PRC1 , thus maintaining anterior/posterior expression patterning 44 ., More formally the expression dynamics are defined by: ( 1 ) Where is the sigmoid function defined as and is a Heavy-side function that equals 0 if x<0 and 1 if x≥0 ., Susceptibility to Polycomb for each gene is set to for all genes at the beginning of each simulation ( generation 0 ) but is subject to change at a mutation rate such that genes can gain or lose susceptibility ( i . e . the variable transitions between 0 and 1 with probability in each offspring ) ., Here we are modeling the evolution of the Polycomb Response Element ( PRE ) , a small canonical base sequence that is targeted by PcGs in higher metazoans 45 , 46 ., In order to assess the impact of the Polycomb mechanism on the evolution of robustness , we measured both environmental and mutational robustness in simulations over 1000 generations ., First we set the mutation rate for susceptibility to thus eliminating the possibility of evolving any epigenetic function ., In keeping with previous results 19 we found that under these conditions both mutational and environmental robustness are positively correlated and increase in tandem ( Figure 1 , blue lines ) ., However , this relationship was inverted when we allowed the Polycomb mechanism to evolve by setting ( the same mutation rate per individual used for the matrix of regulatory interactions ) ., Here mutational robustness increased while environmental robustness decreased ( Figure 1 , red lines ) ., These results were consistent across a wide variety of parameter values ( see Figure S1 ) ., In addition , we modeled the results while allowing for a changing network topology ( links could be created and destroyed ) and found that mutational and environmental robustness remained decoupled see Figure S2 ) ., In summary , we have shown that introducing a Polycomb-like epigenetic mechanism into a transcriptional regulation network model is capable of decoupling environmental and mutational robustness ., Cell differentiation in multicellular biological organisms usually begins with expression changes in a small number of key differentiation genes in response to environmental cues , often upstream genes in the pathway ., Expression of a upstream gene will in turn trigger larger sets of downstream genes that distinctly define each cell type ., One of the best understood examples of this is during muscle differentiation where the key gene MyoD regulates hundreds of downstream targets 47 including important differentiation factors such as muscle specific creatine kinase ( MCK ) 48 and muscle acetylcholine receptor ( AChR ) alpha subunit 49 ., In multi-celled organisms that use epigenetic regulation , cell types are further determined by chromatin changes that lock the cell fate ., In terms of our model , the early differentially expressed genes can be considered as alternative inputs for our system and the transcription of genes in the differentiated cell can be considered the output ., We therefore assume that each input/output mapping ( → ) is the equivalent of the cell type and evaluate whether an evolving network is capable of handling multiple input/output mappings ( → , → and so on ) and in particular whether the capacity to create new mapping is altered by epigenetic functionality in the model ., We therefore allowed a population to evolve under stabilizing selection for generations ( =\u200a100 in main text results; longer values were tested as well . See Figures S1 and S2 ) and then evaluated whether a randomly selected individual from the population could accommodate a new input state and produce a novel output state ( see Methods ) ., The input for the new state was chosen by flipping ( 0↔1 ) each binary input with probability ( in main text results , though values up to give similar results – see Table S1 ) ., The corresponding stable output , , was compared to the initial output , , and to the founders initial output , , using a normalized distance measures and respectively ( see Methods ) which had to be greater than 0 . 05 in both cases for to be considered a new unique output state ., If no such significantly different output was found , we repeated the attempt to create a new input/output mapping ( random individual , random input state ) up to total of 100 times before considering the network unable to create a new input/output state ., Without epigenetic functionality we found that the system was unable to create a new input/output in 47% of 200 cases ., However , with Polycomb it was able to find a new input/output 100% of the time ( Fig . 2 inset ) , a highly significant difference ( p\u200a=\u200a8 . 62×10−22 , Fishers exact test ) suggesting that introducing the epigenetic mechanism enabled networks to evolve a strongly increased capacity for adding new input/output states ., Multi-stability was found after testing an average of just 7 . 55 individuals compared to the case without Polycomb where we were unable to detect multistability even after testing 100 individuals ., Furthermore , the difference is highly robust to different values in the Polycomb threshold ( ) as shown in Figure 2 , since starting with values of =\u200a0 . 05 we already have a capacity above 99% of accepting a new state across many parameter values ., These results are in accordance with the result described above showing that environmental robustness without Polycomb increases through evolutionary time , making the system less likely to produce a unique output even when inputs are altered ., However , with Polycomb the network becomes more sensitive to changes in the environment ( represented here by changes in the input vector ) and consequently acquires the capacity for producing a new output when the inputs are perturbed ., ( In addition , we tested adding multiple new input/output mappings , see SI Table S1 ) ., The role of Polycomb Group Proteins ( PcG ) , discovered in Drosophila , include transcriptional repression of genes showing low expression during early development , a key process in cell differentiation 37 ., Homologs of the core functional proteins comprising the PRC-2 complex ( a component of PcGs ) are present in some eukaryotic unicellar ancestors but are nearly ubiquitous in the multicellular world 39 , 40 , 41 ., The phylogenetic distribution of PcG components and their role in development suggests that Polycomb has played a key role in enabling cell differentiation 40 ., In order to study the evolutionary consequences of Polycomb functionality we incorporated Polycomb functionality into a modeling framework 13 , 42 which captures key features of gene regulatory networks in an evolutionary context ., The evolution of novel mechanisms for controlling gene expression has evolved in tandem with more complex life forms ., Prokaryotes possess cis-regulatory elements , operons and some species show evidence of histone style chromatin structure 9 ., As the Eukarya evolved from simpler unicellular organisms to complex Metazoa , controlling specialized cell functionality became essential ., At the same time , the repertoire of gene expression control expanded to include mechanisms such as methylation , acetylation , ubiquination , and small RNA mediated transcriptional regulation ( i . e . RNAi ) , all of which sculpt gene expression for specialized function 9 ., As each of these mechanisms arose , they often functioned “orthogonally” of the others in a mechanistic sense ., For example , repression of gene expression can be achieved independently either by cis-regulation ( recruitment of repressing TFs to regulatory region ) or by histone modifications at the relevant locus ., These methods result in the same outcome , transcriptional repression , but work through wholly independent mechanisms ., By utilizing chromatin states , Polycomb effectively modifies the architecture of the gene regulatory network in real time ( Figure 3 ) ., As such Polycomb simplifies the architecture by carving out segments of the network to respond to different environmental cues ., Polycomb-targeted genes that exhibit low expression during early development ( expression of PcGs begins as early as 3 hours post-fertilization in the Drosophila embryo ) are continuously repressed through heterochromatin formation , nullifying their associated cis-regulatory effects ., However , under a different set of environmental conditions ( i . e . , in another developmental context ) the same genes might not be enveloped in heterochromatin , allowing the cis-regulatory elements to control expression ., This method allows cells to use a single set of transcriptional regulators ( PcGs ) and yet create very different patterns of expression in distinct cell types ., For example , undifferentiated mesodermal cells require the expression of MyoD to become myoblast cells ., However , MyoD is repressed through the activity of Polycomb ( in particularPRC-2 ) unless the necessary genes ( controlled via adjacency to the PREs ) are expressed early in cell division 38 ., In this manner Polycomb inhibits MyoD in all cells except those destined to become myoblast cells ., This design pattern effectively stratifies a single network into many networks , suggesting a functional role for Polycomb in the evolution of cell differentiation , a key requirement for the evolution of multicellularity ., To explore the development of differential expression we evaluated the capacity of the model to accommodate multiple input-output mappings , as in previous studies 50 ., We found the ability to adopt multiple input/outputs is greatly facilitated with the functionality of Polycomb ( Figure 3 ) ., This finding is consistent with the evolutionary data showing that the essential components of Polycomb function are almost ubiquitous in the multicellular world but are rarely all present simultaneously in unicellular organisms 39 , 40 , 41 again strengthening the hypothesis that Polycomb played a key role during the evolution of multicellularity 3 , 4 ., Further evidence arises from our finding that evolution under Polycomb decoupled mutational and environmental robustness , suggesting that Polycomb can increase sensitivity to environmental conditions for the purposes of cell differentiation ., Previous work has shown that mutational robustness develops in gene-regulatory networks under conditions of stabilizing selection , and that mutational robustness and robustness to environmental changes are correlated 16 , 17 , 18 , 21 , 43 ., This correlated robustness feature is clearly incongruent with multicellular development where minimal ( though particular ) environmental cues are capable of drastically changing cellular phenotypes ., For example , regulation of only four key transcription factors is needed to change a fibroblast to a cardiomyocyte 26 ., When Polycomb functionality is added to the developmental program in the model , this facilitates the effective real-time changes to network connectivity that in turn promotes environmental sensitivity ., However , each effective network is still under stabilizing selection so mutational robustness develops ., With Polycomb the switch between these effectively distinct network architectures is initiated by changing the initial environmental conditions , making the system more responsive to environmental changes ., This real-time remodeling makes use of sub networks for multiple input/output rather than the creation of separate modules within the network ., Indeed previous work on the same base model as we used by Borenstein and Krakauer 51 showed that only a limited number of phenotypes of the total phenotype space are possible ., It appears that the epigenetic addition to the model makes many of the obtainable phenotypes possible ., Biological evidence for decoupling these types of robustness exists in developing multicellular organisms , such as the human fetus , where slight changes in the environmental conditions ( for example , exposure to alcohol during the first weeks ) can cause severe phenotypic changes 52 , 53 , indicative of high environmental sensitivity ., At the same time , the approximately 70 point mutations acquired on average in each human generation 54 rarely produce catastrophic changes , thus demonstrating high mutational robustness ., These findings are consistent with our modeling predictions for a system developing under Polycomb control ., Epigenetic mechanisms have been suggested to evolve in numerous ways ., As with the evolution of sexual reproduction , no single explanation has become the definite single explanation for their evolution ., Similarly , multicellularity has been suggested to evolve by different means and different mechanisms ., Here we put forward an explanation that ties the evolution of multicellularity to that of epigenetic mechanisms ., Additionally , we hypothesize that the capacity to respond differently to different environmental signals , as is required during the developmental program of multicellular organisms , is only one evolutionary advantage of epigenetic processes ., Other advantages include the contribution of epigenetic mechanisms to the emergence of modularity ., It has been argued previously that network modularity contributes to robustness 55 ., As we have shown , Polycomb , in response to environmental queues , carves the network into sub-networks such that beyond the critical time only a subset of the interacting elements play a role is shaping the final gene expression pattern ., Polycomb , thus , amplifies the effect of environmental perturbation beyond genetic perturbation , and introduces modification at the architectural level ., Such change in network architecture introduces higher sensitivity to environmental changes while maintaining robustness to genetic perturbation that have no effect on network architecture ., It has been shown that under stabilizing selection , our model tends to decrease the mean number of steps to reach a stable output state 13 ., Thus , further analysis of the dependency of time to stable output on the time at which Polycomb is activated ( - in our model ) , would further elucidate the evolutionary role of epigenetic mechanisms ., Metazoan evolution is characterized by specialization of cell and tissue functionality ., During multicellular development cells become specialized in function within the organism ., This differentiation requires real-time analysis of the local environment to direct cellular development ., Our findings , although based on the functionality of Polycomb , suggest a general design principal for evolution in the creation of multicellularity , namely the real-time stratification of the gene network ., The effect of the PcG mechanism is to elegantly limit the useable genetic information for a cell based on the events during development ., By effectively removing genes from the accessible gene network the complexity of millions of potential interactions among thousands of genes is reduced ., Following Siegal and Bergman 13 , the model consists of a gene regulatory network of genes each of which has the ability to regulate the expression of any of the genes ., The topology is held in the form of a matrix , with non-zero entries , wij , representing connections within the regulatory network ( a negative value denotes an inhibitory effect ) ., The non-zero entries in the matrix are randomly assigned at the beginning of each simulation with probability ( connectivity of the network ) ., To initiate the development process a random binary ( i . e . containing either 0 or, 1 ) initial condition vector of length is selected ., Gene expression dynamics are then computed according to Equation 1 ., Once a stable founder individual is found , a population of a given size ( kept constant through the simulation ) is founded by that individual ., Evolution of the gene network is done through a standard population-genetic process ., Mutations occur via changes to the non-zero entries of the matrix with 10% chance of a single mutation per genome ., Mating is carried out by selecting two random individuals from the population and then selecting random rows from each parents matrix to create an offspring genotype ( sexual reproduction ) ., At this point selection is carried out as developmental instability ( if no equilibrium gene expression can be generated , as calculated by all real components of the eigenvalues of the Jacobian matrix being less than or equal to 0 ., The Jacobian matrix is defined as: where is the Kronecker delta ( only when , and 0 otherwise ) and through distance from an optimal phenotype ( is defined as the of the initial founder in stabilizing selection ) using the formula: ( 2 ) with Measuring the mutational robustness of our networks was done in the same manner as multiple previous studies 13 , 43 , 56 ., For each individual in the population we mutated exactly one random connection in the matrix ., We simulate gene expression dynamics until a new steady state is reached , or until , and calculate the phenotypic distance ( ) between the new resulting output vector and the original using Equation ( 2 ) above ., Identical steady-state and vectors would be considered as having absolute mutational robustness ., For sake of clarity we report mutational robustness as ., To measure our networks robustness to environmental changes we used a measure outlined in previous studies 43 ., In this measure we vary the input vector by randomly flipping two members and ( a 0→1 or 1→0 ) , reflecting the small environmental differences needed to alter cell fate in Metazoa ., Using the manipulated input vector we re-compute gene expression dynamics ., After altering the input conditions we calculate the divergence from the original in the same manner as for mutational robustness and report it in the same manner . | Introduction, Results, Discussion, Methods | The evolution of ever increasing complex life forms has required innovations at the molecular level in order to overcome existing barriers ., For example , evolving processes for cell differentiation , such as epigenetic mechanisms , facilitated the transition to multicellularity ., At the same time , studies using gene regulatory network models , and corroborated in single-celled model organisms , have shown that mutational robustness and environmental robustness are correlated ., Such correlation may constitute a barrier to the evolution of multicellularity since cell differentiation requires sensitivity to cues in the internal environment during development ., To investigate how this barrier might be overcome , we used a gene regulatory network model which includes epigenetic control based on the mechanism of histone modification via Polycomb Group Proteins , which evolved in tandem with the transition to multicellularity ., Incorporating the Polycomb mechanism allowed decoupling of mutational and environmental robustness , thus allowing the system to be simultaneously robust to mutations while increasing sensitivity to the environment ., In turn , this decoupling facilitated cell differentiation which we tested by evaluating the capacity of the system for producing novel output states in response to altered initial conditions ., In the absence of the Polycomb mechanism , the system was frequently incapable of adding new states , whereas with the Polycomb mechanism successful addition of new states was nearly certain ., The Polycomb mechanism , which dynamically reshapes the network structure during development as a function of expression dynamics , decouples mutational and environmental robustness , thus providing a necessary step in the evolution of multicellularity . | Understanding the transition to multicellularity remains a key unanswered question in evolutionary biology ., The transition required three essential cellular features to evolve: adhesion , signaling and differentiation ., In particular , cell differentiation requires sensitivity to environmental cues to create distinct cell-specific transcription profiles ., Previous work with model organisms and gene network models showed that biological systems evolve robustness to both mutational and environmental perturbations under stabilizing selection and that furthermore , mutational and environmental robustness are correlated ., Increased robustness to environmental cues will therefore pose a barrier to the development of cell differentiation , and thus multicellularity ., Because several important epigenetic developmental mechanisms , particularly Polycomb-mediated histone modification , appear to have evolved with multicellularity , we hypothesized that such a mechanism facilitated sensitivity to the environment and therefore cell differentiation ., Using a computational model , we integrated Polycomb function with a regulatory model , revealing a clear decoupling between environmental and mutational robustness , allowing increased environmental sensitivity while mutational robustness remained intact ., We also found that Polycomb greatly facilitated the ability for a single gene network to create several distinct transcription profiles - each representing a distinct differentiated cell type ., Our work highlights the simple elegance through which the evolution of a key epigenetic mechanism can facilitate the transition to functional cell differentiation . | systems biology, evolutionary modeling, regulatory networks, biology, computational biology, evolutionary biology, evolutionary theory | null |
journal.pgen.1008004 | 2,019 | The meiotic phosphatase GSP-2/PP1 promotes germline immortality and small RNA-mediated genome silencing | Animals , including humans , are comprised of two broad cell types: somatic cells and germ cells ., Somatic cells consist of many diverse differentiated cell types , while germ cells are specialized to produce the next generation of offspring ., An important difference between these two cell types is that somatic cells undergo aging phenomena while the germ line is effectively immortal and capable of creating new “young” offspring 1 ., Understanding the basis of immortality in germ cells may provide insight into why organisms age ., In C . elegans , disruption of pathways that promote germ cell immortality results in initially fertile animals that become sterile after reproduction for a number of generations ., Many such mortal germline ( mrt ) mutant strains are temperature-sensitive , becoming sterile at 25°C but remaining fertile indefinitely at 20°C 2 ., Mutations that cause a Mrt phenotype have been reported in two distinct pathways: telomerase-mediated telomere maintenance 3 , 4 and small RNA-mediated nuclear silencing 5–9 ., Mutations in the PIWI Argonaute protein cause immediate sterility in many species ., However , disruption of the C . elegans Piwi orthologue PRG-1 , which interacts with thousands of piRNAs to promote silencing of some genes and many transposons in germ cells , results in temperature-sensitive reductions in fertility and a Mrt phenotype 6–12 ., Multiple members of a nuclear RNA interference ( RNAi ) pathway that promote the inheritance of transgene silencing also promote germ cell immortality and likely function downstream of PRG-1/Piwi and piRNAs 10 , 13 ., One nuclear RNAi defective mutant , nrde-2 , a number of heritable RNAi mutants , including hrde-1 , and two RNAi defective mutants , rsd-2 and rsd-6 , only become sterile after growth for multiple generations at the restrictive temperature of 25°C 10 , 12–16 ., The reason for this temperature-sensitivity is not clear ., These ‘small RNA-mediated genome silencing’ mutants fail to repress deleterious genomic loci as a consequence of deficiency for small RNA-mediated memory of ‘self’ vs ‘non-self’ segments of the genome 13 , 17 , 18 ., The transgenerational fertility defects of such mutants could reflect a progressive desilencing of heterochromatin , which is modulated by histone modifications that occur in response to small RNAs , such as H3K4 demethylation and H3K9me2/3 15 , 19 ., The SPR-5 histone 3 lysine 4 demethylase promotes genomic silencing in the context of H3K9 methylation and represses transgenerational increases in sterility 20 ., Deficiency for spr-5 also compromises germ cell immortality in a temperature-sensitive manner 21 , similar to genome silencing mutants that are deficient for RNAi or RNAi inheritance 10 , 12–16 ., However , thorough genetic screens for defects in RNAi inheritance failed to recover mutations in spr-5 16 , and a direct test confirmed that deficiency for spr-5 does not compromise RNAi inheritance 13 ., It is therefore not clear if the role of SPR-5 and small RNA-mediated genome silencing proteins in maintenance of germ cell immortality is a consequence of deficiency for the same genomic silencing pathway ., If this is the case , it is possible that deficiency for spr-5 leads to the upregulation of a compensatory RNAi inheritance mechanism that masks an overt role for SPR-5 in RNAi inheritance ., Pioneering studies in Neurospora demonstrated that unsuccessful pairing of whole chromosomes during meiotic prophase , as well as discrete ‘unpaired’ chromosomal regions within paired meiotic homologs , can trigger small RNA-mediated genome silencing 22 ., Multigenerational transmission of hemizygous transgenes in C . elegans , which results in an ‘unpaired’ ~10 kb genomic segment within paired homologous chromosomes during meiosis , leads to transgene silencing in a manner that depends on small RNAs and the PRG-1/Piwi Argonaute protein 23 ., Therefore , a conserved small RNA mechanism operates during meiosis to promote genomic silencing when either large ( chromosome scale ) or small ( transgene scale ) segments of the genome are not properly paired ., A central function of Piwi/piRNA-mediated genomic silencing is to protect the genome from foreign genetic elements like transposons and viruses 11 ., Horizontal transfer of a transposon into the genome of a naïve species will result in a burst of transposition events that ends when the host mounts a small RNA-mediated genomic silencing response against the transposon ., In this context , de novo transposon insertions that represent a threat to genomic integrity would create small ‘unpaired’ hemizygous discontinuities within paired homologous chromosomes during meiosis ., The discrete ‘unpaired’ meiotic chromosome aberrations created by de novo transposon insertions are structurally analogous to hemizygous transgenes , which are the targets of a multigenerational small RNA-induced genome silencing process 23 ., Small ‘unpaired’ meiotic discontinuities created by de novo transposon insertions are therefore likely to be important for shaping genomic and epigenomic evolution ., C . elegans chromosomes do not have a discrete centromere to maintain cohesion between chromosomes during meiosis ., Therefore they utilize two domains , separated by a crossover , called the long and the short arms ., These arms separate at distinct stages of meiosis to prevent premature separation , with the short arms separating in Meiosis I and the long arms separating in Meiosis II ., The regulation of cohesion occurs through localization of GSP-2 to the long arms of meiotic chromosomes through binding to LAB-1 , where it antagonizes AIR-2 ( Aurora-B kinase ) activity 24–26 ., In addition , LAB-1 is also present on mitotic chromosomes where it likely antagonizes AIR-2 activity 27 ., In C . elegans , LAB-1 and GSP-2 fulfills the roles played by Shugoshin and Protein Phosphatase 2A in many other organisms , by protecting meiotic chromosome cohesion on the long arms in Meiosis I 27–29 ., Once recruited by LAB-1 , GSP-2 keeps REC-8 , a meiosis-specific cohesin subunit , dephosphorylated to protect it from premature degradation and chromatid separation 26 , 27 ., Additionally , recent work has shown that HTP-1/2 , HORMA-domain proteins are responsible for LAB-1 chromosomal recruitment and therefore GSP-2 phosphatase activity 30 ., Here we report the identification of a hypomorphic allele of gsp-2 , a PP1/Glc7 phosphatase , which fails to maintain germline immortality at 25°C ., GSP-2 is one of four PP1 catalytic subunits in C . elegans 31 , 32 ., PP1 phosphatase has roles in many cellular processes including mitosis , meiosis , apoptosis and protein synthesis 33 ., Previously , GSP-2 has been shown to promote meiotic chromosome cohesion by restricting the activity of the Aurora B kinase ortholog AIR-2 to the short arms of C . elegans chromosomes during Meiosis I 26 , 27 ., Here , we demonstrate that GSP-2 is likely to act during meiosis to promote germline immortality via a small RNA-mediated genome silencing pathway ., In a screen for mrt mutants 2 , one mutation that displayed a Temperature-sensitive defect in germ cell immortality , yp14 , was tightly linked to an X chromosome segregation defect manifesting as a High Incidence of Males ( Him ) phenotype , such that 3 . 9% of yp14 self-progeny were XO males , which was significantly greater than the 0 . 05% male self-progeny phenotype observed in wildtype animals at 20°C ( Fig 1A , p < . 0001 ) ., The yp14 mutation was mapped to Chromosome III , and whole genome sequencing revealed missense mutations in 6 genes within the yp14 interval ( S1A and S1B Fig ) ., Three-factor mapping of the yp14 Him and Mrt phenotypes suggested that yp14 might correspond to the missense mutation in gsp-2 ( Fig 1C and 1D ) or to a mutation in the G-protein coupled receptor gene srb-11 ( S1A and S1B Fig ) ., To test whether the chromosome segregation defect of yp14 was due to a mutation in gsp-2 , we performed a non-complementation test with a deletion mutation in gsp-2 , tm301 ., yp14 / tm301 F1 heterozygous hermaphrodites gave rise to F2 male progeny at a frequency of 5 . 7% at 20°C , similar to the 3 . 8% male phenotype observed for yp14 homozygotes ( S1C Fig ) ., Thus , tm301 failed to complement gsp-2 ( yp14 ) for its Him phenotype ., In contrast , neither gsp-2 ( tm301 ) / + nor gsp-2 ( yp14 ) / + control animals displayed a Him phenotype ( S1C and S1F Fig ) ., Additionally , gsp-2 ( tm301 ) null mutants immediately exhibited high levels of embryonic lethality at 20°C with a few F3 embryos that survive until adulthood ( Fig 1B ) , consistent with roles for PP1 in chromosome condensation and segregation during mitosis in several species 24 , 25 , 34 ., High levels of embryonic lethality for F3 gsp-2 ( tm301 ) mutant embryos ( 97% ) , led to uniformly sterile uniformly sterile F3 adults that produced no F4 progeny 25 ( Fig 1B ) ., These very high levels of embryonic lethality contrast with the embryonic lethality observed for gsp-2 ( yp14 ) mutants , which was 6% at 20°C and 41 . 6% for F8 animals grown at 25°C ( Fig 1B ) ., Both the Emb and Him phenotypes were exacerbated at 25°C ( Fig 1A and 1B ) , suggesting that gsp-2 ( yp14 ) has a chromosome segregation defect that may be mechanistically linked to its Mortal Germline phenotype ( Fig 1A and 1E ) ., In gsp-2 ( yp14 ) mutants , the X chromosome non-disjunction defect was more pronounced at both temperatures than the embryonic lethality associated with non-disjunction of the five C . elegans autosomes ( S1 Table ) ., Mutations that cause chromosome non-disjunction during mitosis occasionally lead to loss of an X chromosome during germ cell development , which could result in the stochastic appearance of XX hermaphrodites with high numbers of XO male progeny 35 ., However , jackpots of XO males did not occur when yp14 mutant hermaphrodites were isolated as single L4 larvae at 20°C or as L1 or L4 larvae at 25°C ( Fig 1G , S1D and S1E Fig ) , implying that yp14 is a separation-of-function mutation that specifically compromises the meiotic chromosome segregation function of GSP-2 , with little or no effect on mitotic chromosome segregation ., It is formally possible that gsp-2 ( yp14 ) is deficient for a mitotic function of GSP-2 that is relevant to germ cell immortality that is either distinct from its role in mitotic chromosome segregation or so subtle that we could not detect it in our assays ., At 20°C , gsp-2 ( yp14 ) mutants remained fertile indefinitely , but at 25°C they exhibited sterility between generations F5 and F17 ( Fig 1E and 1F ) ., Given that LAB-1 promotes cohesion of the long arms of meiotic chromosomes via the GSP-2 phosphatase , we asked if LAB-1 is relevant to germ cell immortality by first outcrossing a lab-1 deletion with wildtype and re-isolating lab-1 homozygotes in an effort to eliminate epigenetic defects that could have accumulated in the parental lab-1 strain ., Outcrossed lab-1 mutants displayed a Mortal Germline phenotype at 25°C ( Fig 1E and 1F ) ., We created lab-1; gsp-2 double mutants , which remained fertile indefinitely when grown at 20°C but displayed a slightly accelerated number of generations to sterility at 25°C in comparison with lab-1 mutants ( Fig 1E and 1F ) ., Together , these results suggest that a meiotic function of GSP-2 that is directed by LAB-1 promotes germ cell immortality ., The small acceleration in the time to sterility in the double mutant animals suggests slight additivity between the mutations ., Both the gsp-2 and lab-1 alleles are partial loss-of-function alleles that when combined could conceivably result in a stronger phenotype ., Moreover , the weak Mortal Germline phenotype of lab-1 single mutants at 20°C was suppressed by gsp-2 ( yp14 ) ( Log Rank Test , p = . 001 ) ., One possible explanation for this very slight rescue at the permissive temperature is the loss of lab-1 alone results in GSP-2 being mis-localized and performing an ectopic function that is ablated when GSP-2 function is reduced ., It is likely that this does not occur at 25°C because GSP-2 function is more severely compromised at the higher temperature ., Multiple genes that regulate small RNA-mediated epigenomic silencing promote germ cell immortality at high temperatures , like gsp-2 ( yp14 ) and lab-1 10 , 12 , 16 ., Three small RNA-mediated epigenomic silencing genes that are required for germ cell immortality promote a specific form of transcriptional gene silencing termed nuclear RNA interference , nrde-1 , nrde-2 and nrde-4 10 , 12 , 36 ., The response to a dsRNA trigger that targets lin-26 is dependent on nuclear RNA interference 37 ., Control wildtype and gsp-2 ( yp14 ) mutant animals displayed a completely penetrant Embryonic Lethality phenotype in response to lin-26 dsRNA , whereas nuclear RNAi defective mutant nrde-2 and the RNAi defective mutant rsd-6 did not ( Fig 2A ) , indicating that nuclear RNAi within a single generation is normal in the gsp-2 ( yp14 ) mutant ., Small RNAs can trigger RNAi inheritance 10 , 13 , where silencing of a gene in response to siRNAs can be transmitted for multiple generations ., Transgenerational RNAi inheritance can occur when endogenous genes are targeted by dsRNA triggers 38 , but this can also happen when GFP reporter transgenes are targeted by small RNAs derived from GFP 13 , 17 , 18 ., We tested the transgene cpIs12 Pmex-5::GFP and found that it was silenced in response to GFP siRNAs and that silencing of this transgene was inherited for up to 4 generations after removal from the dsRNA trigger ( Fig 2B , Results summarized S6 Table ) ., In contrast , GFP expression in gsp-2 ( yp14 ) ; cpIs12 was initially silenced but silencing was not inherited over multiple generations ( Fig 2B ) , indicating that gsp-2 ( yp14 ) promotes RNAi inheritance ., Propagation of GFP or mCherry transgenes in the hemizygous state for multiple generations elicits a strong transgene silencing response , which is thought to be due to persistent yet small ‘unpaired’ discontinuities in the structure of paired meiotic homologous chromosomes at the site of the transgene 23 ., We found that hemizygosity for the transgene cpIs12 resulted in progressive transgene silencing in populations of animals over the course of several generations until cpIs12 became fully silenced by generation 5 ( Fig 2C and 2D ) ., In contrast , when cpIs12 was placed in a gsp-2 ( yp14 ) genetic background and propagated in a hemizygous state , we found that cpIs12 was initially weakly silenced but that genomic silencing never became fully penetrant ( Fig 2C and 2D ) ., Together , the above data indicate that gsp-2 promotes the silencing of unpaired hemizygous transgenes , which depends on small RNA-mediated genome silencing 23 ., A central function of small RNA-mediated genomic silencing is to maintain silencing of repetitive elements and transposons in the germline , thereby protecting genomic integrity 15 , 19 , 39 ., We previously reported that RNA expression of tandem repeat loci was upregulated in late-generation rsd-2 and rsd-6 mutants grown at 25°C 12 ., Therefore , we asked if desilencing of tandem repeats occurred in gsp-2 ( yp14 ) mutants using RNA fluorescence in situ hybridization ( FISH ) to examine the expression of multiple repetitive elements ., In wild-type controls grown at 25°C , we detected RNA from tandem repeat sequences using CeRep59 sense and anti-sense probes in embryos but not in the adult germline or somatic cells , consistent with previous observations ( S2 Fig ) 12 ., However , in late-generation gsp-2 ( yp14 ) and rsd-6 mutants , robust expression of tandem repeats was observed throughout the soma and germline of adult animals , indicating that tandem repeats become desilenced in these strains ( S2 Fig ) ., Given that small RNA-mediated genome silencing is dysfunctional in gsp-2 ( yp14 ) mutants , we asked if small RNA populations were perturbed by preparing RNA from early- and late-generation wildtype , gsp-2 ( yp14 ) , rsd-6 and spr-5 mutants grown at either 20°C or 25°C ., We examined rsd-6 and spr-5 mutants as they have known temperature sensitive germ cell immortality defects associated with loss genomic silencing as a consequence of small RNA or histone demethylation defects , respectively 12 , 21 ., Small RNA libraries were prepared and subjected to high throughput sequencing , and we then examined levels of 22G RNAs that are 22 nucleotides in length beginning with a 5’ guanine , as 22G RNAs are the major effectors of genomic silencing in C . elegans 5 , 40 ., 22G-RNAs in all late generation lines , normalized to total small RNA content showed a decrease relative to early generation N2 lines ., The decrease was more pronounced in gsp-2 and rsd-6 mutants ( p = 1 . 2e-7 and 4e-19 , Wilcox paired test; S2 Table , S3 Fig ) but not in spr-5 where the decrease was not significantly different from the difference in N2 ( p = 0 . 13 ) ., Analysis of the 22G-small RNA data revealed that spr-5 and rsd-6 share some genes with reduced levels of 22G RNAs with increasing generations , but there are other genes that show dissimilar behavior for each individual mutant ., This suggests that spr-5 may act both in conjunction with rsd-6 and in a separate pathway to promote germline immortality ., In contrast , 22G RNAs from gsp-2 ( yp14 ) showed strong similarities to those of spr-5 mutants but showed little similarity to 22G RNA changes observed for rsd-6 mutants , suggesting that gsp-2 ( yp14 ) and spr-5 have similar effects on genome maintenance ( S3 Fig ) ., As a control , there is little coherent change in late-generation versus early generation N2 wildtype that overlaps with gsp-2 ( yp14 ) meaning that changes we see in gsp-2 ( yp14 ) are not due simply to passaging animals at 25°C ( Fig 2E and 2F ) ., As germ cell immortality is promoted in part by primary siRNAs termed piRNAs that interact with the Piwi Argonaute protein PRG-1 8 , we also examined piRNA populations , which are enriched for 21 nucleotide RNAs that begin with a 5’ uracil ( 21U RNAs ) 6 , 7 , 9 and found that these were normal ( Fig 2E and 2F ) ., We also examined miRNAs , which have not previously been implicated in the Mortal Germline phenotype ., Interestingly , miRNAs were significantly reduced in late generation spr-5 and gsp-2 ( yp14 ) mutants ( p = 1 . 2e-20 and p = 2 . 05e-25 respectively; S3 Table , S3 Fig ) , but not in rsd-6 mutants ., Since spr-5 does not show global decrease in 22G-RNAs this is unlikely to be a secondary consequence of disturbance of the total small RNA pool ., The relevance of this finding to the Mortal Germline phenotype awaits further investigation ., Together these results indicate that gsp-2 ( yp14 ) and spr-5 display common statistically significant changes in two classes of small RNAs , which implies that their genomic silencing defects may be more similar to one another than to those of rsd-6 mutants ., To study the relationship between gsp-2 ( yp14 ) and the small RNA genome silencing pathway , we created double mutants between gsp-2 ( yp14 ) and small RNA silencing mutants that display temperature-sensitive defects in germ cell immortality , hrde-1 , nrde-2 and rsd-6 ., Because gsp-2 ( yp14 ) is a hypomorphic allele , we predicted that single and double mutants would display a similar number of generations to sterility if it were functioning in the small RNA silencing pathway ., For gsp-2 ( yp14 ) ; hrde-1 and rsd-6; gsp-2 ( yp14 ) , we saw a modest decrease in the number of generations to sterility suggesting a slight additive effect ( Fig 3A and 3C , Log Rank test: p < . 0001 ) ., In contrast , nrde-2; gsp-2 ( yp14 ) double mutants did not differ from the single mutants ( Fig 3B , Log Rank test: p = . 06 ) ., Together , these results indicate that there is a weak additive effect on transgenerational lifespan when gsp-2 is combined with hrde-1 or rsd-6 , but not when it is combined with nrde-2 ., The modest acceleration observed for some small RNA genomic silencing pathway and gsp-2 ( yp14 ) double mutants may be consistent with a single genome silencing pathway , as many single mutants in this pathway that display similar germline phenotypes at sterility also display a consistent , slightly accelerated sterility as double mutants ., There are a number of explanations for this , including transmission of epigenetic defects from germ cells of the grandparents that created these double mutants , or shared but non-equivalent functions in terms of which segments of the genome each protein silences 15 ., We previously reported that sterile , late-generation small RNA genome silencing mutants display a wide range of germline sizes , including many with few or no germ cells 12 , 41 ., Therefore to investigate the cellular cause of transgenerational sterility in gsp-2 ( yp14 ) and lab-1 mutants , we examined germline development in animals that became sterile after multiple generations ., Most sterile generation L4 gsp-2 ( yp14 ) and lab-1 mutant germlines were normal in size , though a small minority had a reduction in total germline length , resulting in a weak but significant difference in germline profile compared to wild-type ( Fig 4A–4E and 4H , S5 Table , Results summarized S6 Table ) ., Differentiating germ cell nuclei in spermatogenesis were observed for sterile generation L4 larvae for all strains ( Fig 4A and 4H ) ., However , the germlines of two-day-old sterile gsp-2 ( yp14 ) and lab-1 mutant adults ranged in size from normal to a complete loss of germ cells ( Fig 4B–4E and 4I ) , resulting in a significant difference when compared to wild-type controls ( S5 Table p<1E-80 ) ., We studied small RNA genome silencing mutants and found that rsd-6 , hrde-1 or nrde-2 mutant L4 larvae that were poised to become sterile displayed predominantly normal-sized germlines ( Fig 4H ) ., In contrast , sterile-generation rsd-6 , hrde-1 and nrde-2 mutant adults had germline profiles that were similar to those of sterile gsp-2 ( yp14 ) mutant adults and markedly smaller than those of sterile generation L4 larvae ( Fig 4I , S4 Table ) ., lab-1 ( tm1791 ) displayed an increased frequency of germline tumors in comparison to other mutants , possibly due to a genetic modifier present in the tm1791 mutant background ., Lastly , we tested if sterile spr-5 mutants displayed similar germline phenotypes as those observed in small RNA mutants and gsp-2 ( yp14 ) ., We found that sterile spr-5 mutant adults displayed similar germline atrophy phenotypes , suggesting the resemblance to gsp-2 ( yp14 ) or lab-1 mutants ( Fig 4H and 4I ) ., Our previous work showed that mutations in the cell death genes ced-3 and ced-4 partially rescued the empty and atrophy phenotypes observed for germlines of rsd-2 , rsd-6 , and prg-1 small RNA genome silencing mutant adults 12 , 41 , suggesting that apoptosis promotes germ cell atrophy as these animals develop from L4 larvae into adults ., To determine if acute loss of GSP-2 causes germline atrophy , we examined gsp-2 ( tm301 ) null mutants grown at 20°C and 25°C ., gsp-2 ( tm301 ) homozygous F2 animals and their few surviving F3 progeny showed normal germlines , with no morphological defects in germline size or development for either L4 larvae or young adults , which significantly differed from the germline profiles of gsp-2 ( yp-14 ) animals ( S4 and S5 Tables ) ., Therefore , the late-generation sterility phenotype of yp14 mutants is distinct from the fertility defects that occur in response to acute loss of GSP-2 in maternally depleted F3 deletion homozygotes ., Mature C . elegans oocytes typically contain 6 bivalents ( pairs of homologous chromosomes held together by crossovers ) , which can be scored as 6 DAPI-stained bodies ., Defects in meiotic pairing , cohesion , synapsis , and crossing over can lead to the presence of univalents , which are observed as greater than 6 DAPI bodies per oocyte ., We previously observed that small RNA nuclear silencing mrt mutants rsd-2 and rsd-6 displayed increased levels of univalents at sterility , which were not observed in either wildtype or in fertile rsd-2 or rsd-6 mutant late-generation animals grown at 25°C 12 ., We measured the presence of oocyte univalents in N2 wildtype control worms grown at 20°C and 25°C , which almost always contained 6 DAPI bodies representing the 6 paired chromosomes ( 5 bodies are occasionally scored when bivalents that overlap spatially cannot be distinguished ) ., However , when gsp-2 ( yp14 ) worms were passaged at 25°C until sterility , only 60% of oocytes contained 6 paired chromosomes with the other 40% contained 7 to 12 DAPI bodies ( Fig 4J , Results summarized S6 Table ) ., This increase in oocyte univalents was not present in fertile gsp-2 ( yp14 ) worms , at 20°C or even for fertile late-generation 25°C gsp-2 ( yp14 ) adults that were close to sterility ( Fig 4J ) ., In contrast , we found no univalents in the null gsp-2 allele tm301 , either for F2 animals or for rare F3 escapers , consistent with previous observations 24 , 25 ., LAB-1 has been previously implicated in the pairing of homologous chromosomes during meiosis 27 ., To determine if homolog pairing is perturbed in gsp-2 ( yp14 ) mutants grown for two generations at 25°C , we examined the X chromosome pairing center protein HIM-8 in fertile 2 day old adults ., When scored at pachytene only one spot was present in the majority of the nuclei suggesting pairing is occurring normally ( Fig 4K , S4 Fig ) ., In addition to gsp-2 ( yp14 ) , we examined HIM-8 foci in fertile lab-1 , rsd-6 and spr-5 mutants grown at 25°C for two generations and found that lab-1 mutants displayed decreased meiotic chromosome pairing consistent with previously reported data 27 but that pairing was relatively normal in the other mutants ( Fig 4K , S4 Fig ) ., Given that LAB-1 and GSP-2 are known to promote meiotic chromosome cohesion , we tested the hypothesis that dysfunction of other factors that promote meiotic chromosome cohesion might be sufficient to elicit germline atrophy ., Mutant strains with defects in cohesion , smc-3 ( t2553 ) and coh-3 ( gk112 ) ; coh-4 ( tm1857 ) double mutants 42–44 became sterile immediately and did not exhibit germline atrophy phenotypes observed in gsp-2 ( yp14 ) ( S5 Fig , S4 Table ) ., Therefore , the late-generation sterility phenotypes of gsp-2 ( yp14 ) and small RNA mutants are not due to acute loss of meiotic chromosome cohesion ., To further characterize the nature of the gsp-2 ( yp14 ) mutation , we examined the localization of LAB-1 and GSP-2 in pachytene nuclei of gsp-2 ( yp14 ) , lab-1 , rsd-6 and spr-5 animals ., Decreased GSP-2 localization was observed in both gsp-2 ( yp14 ) and spr-5 mutants but not in lab-1 or rsd-6 mutants ( Fig 5A ) ., Similar defects in small RNA profiles of gsp-2 and spr-5 mutants are consistent with the localization of GSP-2 being normal in rsd-6 mutants but absent in gsp-2 ( yp14 ) and spr-5 mutants ( Fig 5A ) , which supports the possibility that GSP-2 may promote genomic silencing in response to small RNAs ., The presence of GSP-2 staining in the lab-1 deletion was surprising as animals treated with RNAi against lab-1 show decreased GSP-2 staining ., However , as the tm1791 deletion is a non-null allele , it is possible that GSP-2 can still interact with LAB-1 to some degree ., Additionally , we saw no change in LAB-1 localization in any strain except for the lab-1 deletion , which still exhibited some staining consistent with the tm1791 deletion being a non-null allele ( Fig 5B ) ., Lastly , we assessed LAB-1 localization at diakinesis to determine if LAB-1 localization on the long arms was altered in any of these mutants and we found that localization was relatively normal in gsp-2 ( yp14 ) , rsd-6 and spr-5 mutants ( S6 Fig ) ., The localization of LAB-1 in gsp-2 ( yp14 ) along the long arms was abnormal looking but clearly did not localize to both the long and short arms of the chromosomes ., A previously identified phenotype of gsp-2 null mutants is an increase in Histone 3 Serine 10 ( H3S10 ) phosphorylation due to expansion of the AIR-2-localizing domain 24 , 30 ., In wildtype worms grown at 20°C and 25°C , H3S10 phosphorylation was visible on the condensed chromosomes in the -1 to -3 oocytes , which are defined relative to the spermatheca with the closest being called -1 ( Fig 6A , Results summarized S6 Table ) ., In both early- and late-generation gsp-2 ( yp14 ) mutant oocytes , H3S10 phosphorylation increased when compared with wildtype controls , with increased levels on chromosomes ( Fig 6B and 6M ) ., Late-generation gsp-2 ( yp14 ) mutant animals grown at 25°C had a small but significant increase in H3S10 phosphorylation levels compared to gsp-2 ( yp14 ) mutant controls grown at 20°C ( Fig 6M ) ., Furthermore , we observed increased levels of H3S10 phosphorylation in lab-1 mutants ( Fig 6C and 6M ) , consistent with previous results 27 ., By quantification of fluorescence intensity we measured significant increased levels of H3S10p in lab-1 , rsd-6 , and hrde-1 but not in nrde-2 mutants ( Fig 6C–6F and 6M ) ., The distinct phosphorylation levels in nrde-2 mutants could reflect its small RNA genome silencing function , where NRDE-2 works downstream of RSD-6 and HRDE-1 to promote accumulation of stalled RNA polymerase II at loci that are targeted by small RNAs 45 ., This would suggest that the maintenance of histone marks occurs at the point in the pathway where RSD-6 and HRDE-1 function but not downstream at level of NRDE-2 ., PP1 has been previously shown to dephosphorylate a number of histone amino acids , including Histone 3 Threonine 3 ( H3T3 ) 46 ., When we examined H3T3 phosphorylation in wildtype controls grown at 25°C , staining was visible in the -1 to -3 oocytes ( Fig 4G and 4G’ , Results summarized S6 Table ) ., However , in sterile generation gsp-2 ( yp14 ) mutants , H3T3p staining was significantly brighter than controls when images were taken under the same conditions ( Fig 6H and 6H’ ) ., Sterile generation lab-1 and the small RNA mutants hrde-1 , rsd-6 and nrde-2 all exhibited increased H3T3 phosphorylation signal intensity in the -1 to -3 oocytes ( Fig 6I , 6L and 6N ) ., Furthermore , there was a significant increase in H3T3 phosphorylation in sterile generation gsp-2 ( yp14 ) mutant adults compared to the earlier , fertile generation animals suggesting transgenerational accumulation of H3T3 phosphorylation ( Fig 6N ) ., Together , our results suggest that an increase in phosphorylation of H3T3 consistently occurs in oocytes of gsp-2 and small RNA silencing mutants however , increased H3S10 phosphorylation occurs only in gsp-2 ( yp14 ) , lab-1 , rsd-6 , and hrde-1 but not in nrde-2 mutants ., This defect is sensitive to temperature , as observed for the meiotic chromosome segregation and germ cell immortality defects of gsp-2 ( yp14 ) ( Fig 1E and 1F ) ., Finally , we examined histone marks that promote gene silencing or activation ., H3K9 methylation can be deposited at silenced genomic loci , and H3K9me and H3S10p marks can function as a phospho-methyl switch where H3S10 phosphorylation can block some epigenetic regulators , such as HP1 , from accessing the adjacent H3K9me mark 47–49 ., In late-generation fertile gsp-2 ( yp14 ) , lab-1 , rsd-5 and spr-5 mutant animals grown at 25°C , we observed a significant decrease in H3K9me2 and H3K9me3 intensity in diakinesis oocytes ( Fig 7A and 7B , Results summarized S6 Table ) ., We also assessed the H3K4me3 transcriptional activation mark and found that it was significantly decreased in all mutant genotypes at diakinesis ( Fig 7A and 7B ) ., It is possible that excess H3T3 phosphorylation present in these mutant strains ( Fig 6G–6N ) could affect the activities of enzymes that modify histone H3 , especially H3K4 ., Additionally , the presence of excess phosphorylation on adjacent amino acids could perturb the binding of the histone methylation antibodies , possibly disrupting our ability to assess methylation levels ., We demonstrate for the first time that gsp-2 and lab-1 are required for germ cell immortality at 25°C as strains deficient for these proteins bec | Introduction, Results, Discussion, Materials and methods | Germ cell immortality , or transgenerational maintenance of the germ line , could be promoted by mechanisms that could occur in either mitotic or meiotic germ cells ., Here we report for the first time that the GSP-2 PP1/Glc7 phosphatase promotes germ cell immortality ., Small RNA-induced genome silencing is known to promote germ cell immortality , and we identified a separation-of-function allele of C . elegans gsp-2 that is compromised for germ cell immortality and is also defective for small RNA-induced genome silencing and meiotic but not mitotic chromosome segregation ., Previous work has shown that GSP-2 is recruited to meiotic chromosomes by LAB-1 , which also promoted germ cell immortality ., At the generation of sterility , gsp-2 and lab-1 mutant adults displayed germline degeneration , univalents , histone methylation and histone phosphorylation defects in oocytes , phenotypes that mirror those observed in sterile small RNA-mediated genome silencing mutants ., Our data suggest that a meiosis-specific function of GSP-2 ties small RNA-mediated silencing of the epigenome to germ cell immortality ., We also show that transgenerational epigenomic silencing at hemizygous genetic elements requires the GSP-2 phosphatase , suggesting a functional link to small RNAs ., Given that LAB-1 localizes to the interface between homologous chromosomes during pachytene , we hypothesize that small localized discontinuities at this interface could promote genomic silencing in a manner that depends on small RNAs and the GSP-2 phosphatase . | The germ line of an organism is considered immortal in its capacity to give rise to an unlimited number of future generations ., To protect the integrity of the germ line , mechanisms act to suppress the accumulation of transgenerational damage to the genome or epigenome ., Loss of germ cell immortality can result from mutations that disrupt small RNA-mediated genome silencing , which protects the germ line from foreign genetic elements such as transposons ., Here we report for the first time that the C . elegans protein phosphatase GSP-2 that promotes core chromosome biology functions during meiosis is also required for germ cell immortality ., Specifically , we identified a partial loss-of-function allele of gsp-2 that exhibits defects in meiotic chromosome segregation and that is also dysfunctional for transgenerational small RNA-mediated genome silencing ., Our results are consistent with a known role of Drosophila Protein Phosphatase 1 in heterochromatin silencing , and point to a meiotic phosphatase function that ensures germ cell immortality by promoting genomic silencing in response to small RNAs . | phosphorylation, cell physiology, meiosis, rna interference, gene regulation, cell cycle and cell division, cell processes, dna-binding proteins, invertebrate genomics, germ cells, epigenetics, small interfering rnas, animal cells, genetic interference, chromosome biology, proteins, gene expression, cell immortalization, histones, animal genomics, biochemistry, rna, cell biology, nucleic acids, post-translational modification, genetics, biology and life sciences, cellular types, genomics, non-coding rna | null |
journal.pcbi.1006185 | 2,018 | miRAW: A deep learning-based approach to predict microRNA targets by analyzing whole microRNA transcripts | MicroRNAs ( miRNAs ) are a family of ∼22-nucleotide ( nt ) small RNAs that regulate gene expression at the post-transcriptional level ., They act by binding to partially complementary sites on target genes to induce cleavage or repression of productive translation , preventing the target gene from producing functional peptides and proteins ., Despite advances in understanding miRNA:mRNA interactions , the rules that govern their targeting process are not fully understood 1–4 ., While many miRNA targets have been computationally predicted only a limited number have been experimentally validated ., Moreover , although a variety of miRNA target prediction algorithms are implemented , results amongst them are generally inconsistent and correctly identifying functional miRNA targets remains a challenging task ., The majority of prediction tools are based on the assumption that it is the miRNA seed region—generally defined as a 6 to 8 nucleotide sequence starting at the first or second nucleotide—that contains almost all the important interactions between a miRNA and its target and their focus is on these canonical sites ., This seed-centric view has been supported by structural studies 5 and a widely cited report 6 that investigated the importance of other ( non-canonical ) regions within a miRNA and concluded their contributions had relatively low relevance compared to the ( canonical ) seed region ., However , more recent studies have revealed that many relevant targets are implemented via non-canonical binding and involve nucleotides outside the seed region , indicating that the entire miRNA should be considered in target prediction algorithms 3 , 7 , 8 ., This is also supported by the performance of target prediction tools which typically identify approximately 80% of known miRNA targets , indicating the mechanisms associated with the remaining 20% of non-canonical targets remain poorly understood ., Thus , there is an opportunity for novel approaches to improve knowledge of miRNA-regulated processes ., In turn , this can lead to better understanding the effects of mutations in the non-coding region of the genome in terms of function and disease ., To this end , in this work , we apply deep learning techniques to investigate the role of non-canonical sites and pairing beyond the canonical seed region in microRNA targets ., Almost all target prediction methods are rule-based or adopt machine learning ( ML ) methodology with varying success ., Rule-based systems incorporate various human-crafted descriptors to represent miRNA:gene target binding ( e . g . type of pairs in the site , binding stability , or conservation of the target site among species ) ., Machine learning techniques also use human crafted descriptors , but as input features to machine learning models ., The limitation of both these approaches is the process of feature selection and representation , which is constrained by the use of handcrafted descriptors to model a process that is not fully understood ., Recent increases in computational power have permitted the rise of methods that can dispense with human-crafted features; making it possible to deal directly with raw data and autonomously learn and identify patterns to appropriately represent data ., In particular , deep learning ( DL ) 9 has been shown to be an effective method for classification tasks in domains with complex feature representation ., Generally , DL methods represent raw data by incorporating multiple hierarchical levels of abstraction ., While this approach is typically applied to standard ML problems such as image classification 10 , natural language processing 11 or speech recognition 12 , it is now finding use in the life sciences for applications such as RNA splicing prediction 13 and gene expression inference 14 , 15 ., DL has also been applied to the miRNA target prediction problem ., Cheng et al . 16 used convolutional neural networks to analyze matrices of miRNA:site features , but the selected features were still human-crafted descriptors and thus the method faces similar problems as rule-based and ML approaches ., A more recent work , DeepTarget 17 , relied on recurrent neural networks to identify potential binding sites and assess their functionality ., However this work is still oriented to the identification of canonical sites and relies on a limited small data set for the training phase ., In this paper we present miRAW , a novel miRNA target prediction tool that works with raw input data and which makes no assumptions about suitable input descriptors ., miRAW scans the 3’UTR of the gene to identify potential target sites ., It then uses DL to identify relevant patterns by directly analyzing the whole mature miRNA transcript , rather than focusing on the seed region and analyzing precomputed descriptors ., It is trained and tested against experimentally verified positive and negative datasets ., The resulting predictions can then be refined by incorporating exogenous information ., When compared to other state-of-the-art miRNA target prediction tools , miRAW demonstrates a significant improvement in performance , highlighting the importance of considering pairing beyond the seed region ., In order to gain a deeper understanding of the characteristics of non-canonical targets , we also investigated the prediction results in terms of model design ( i . e . , how different configurations affect the type of predictions obtained ) and from a biological perspective ( i . e . , how different classes of predicted target sites varied in terms thermodynamic stability and binding structures ) ., In particular , results reveal, ( i ) many potential functional non-canonical binding structures that are supported by experimentally verified miRNA:mRNA target data and, ( ii ) commonly prioritized features such as site accessibility energy and seed region structure are relevant but not sufficient for discerning between functional and non-functional target sites ., A key factor for successful application of any ML classification technique is access to a sufficiently variable and representative dataset that will generalize a trained model to new and unseen data ., For the miRNA target prediction problem , this requires a comprehensive dataset of verified positive and negative targets that encompass both canonical and non-canonical examples ., While there are multiple data repositories providing information regarding experimentally validated positive miRNA targets 20–22 , there are significantly fewer experimentally verified negative targets ., This is not an issue for methods that use rule-based approaches to describe positive matches 6 , but it represents a major concern for ML-based approaches that require similar numbers of labeled examples for both classes ., Here , we focused on human data and used, ( i ) Diana TarBase 21—the most comprehensive publicly available dataset , which contains information for both positive ( 121 , 090 ) and negative ( 2 , 940 ) experimentally verified human miRNA:mRNA interactions—and, ( ii ) MirTarBase 20—containing 410 , 000 experimentally verified positive targets—as the knowledge core for our study ., Annotation related to transcripts and miRNA binding site locations were obtained by cross-referencing Diana TarBase identifiers with miRBase release 21 23 and Ensembl release 87 24 entries ., As a preliminary step , the Diana and MirTarBase data were parsed to, ( i ) remove inconsistent entries that were marked both as positive and negative targets—due to contradictory results in different experimental validations—and, ( ii ) combine entries that were validated more than once by different verification methods ., This produced a final dataset of 303 , 912 positive ( + ) and 1 , 096 negative ( - ) miRNA:mRNA interactions ., The data was then split into two parts ( each consisting of 151 , 956+ and 548- interactions ) for the training and testing stages ( see Fig A in S1 and S2 Files ) ., Selection of candidate MBSs in a mRNA is another key step for a miRNA target prediction algorithm as it identifies which regions within a mRNA have the potential to be a target binding site ., Most target prediction methods follow a similar approach for candidate selection: they scan the 3’UTR of the gene looking for sites that are partially complementary to the miRNA transcript; if a site fulfills certain criteria , it is considered to be a candidate site and is subjected to further analysis ., Candidate site selection methods ( CSSMs ) that focus on the retrieval of canonical targets only consider those sites that have perfect complementary within the miRNA seed region ( nucleotides 2 to 8 , see Fig 2a ) and will return the smallest number of predicted targets ., Methods willing to accept non-canonical sites have looser restrictions: some accept a limited number of bulges , mismatches or wobble pairs in the seed region whilst others accept such mismatches only if there are compensatory nucleotide pairs outside the seed region ( Fig 2b and 2c ) ., In an ideal scenario where the training dataset contained sufficient examples of all the possible forms of positive and negative targets , the CSSM would not be required as , theoretically , an ANN would be able to estimate the function acting as CSSM ., In reality , there are limited numbers of reliable experimentally verified miRNA:targets ( especially for negatively validated sites ) and the CSSM step effectively narrows the search space to simplify the ANN classification task ., The CSSM used by miRAW ( CSS miRAW ) for searching the 3’UTR follows a similar approach to other prediction tools –investigating successive 30-mer segments– but employs a more relaxed set of restrictions that reflect recent experimental studies that relax the requirement of perfect pairing in the seed region and acknowledge a possible role for the other nucleotides ., For example , Kim et al 8 report the role of nucleotide 9 in several miRNA binding sites and Grosswendt et al 2 found that a significant number of miRNAs do not require perfect complementarity within the seed region and compensate for this in non-seed nucleotides ., Finally , a recent structural study by Klum et al 31 clarify a role for the 3’ end of the miRNA in the targeting process ., Based on the findings from these and other related studies , we investigated three different approaches that expand the analysis beyond the typical 7mer seed region and relax the broadly adopted requirement for perfect pairing within the seed region ., In particular , we consider a site to be a candidate MBS if there is a minimum number of base pairs—considering both Watson-Crick ( WC ) pairing and wobbles—within an extended seed region and investigated three different configurations: In each case , base pairs do not need to be consecutive in order to accommodate the presence of gaps and bulges ., Thus , these models can accommodate both standard canonical MBSs as well as a broader range of non-canonical target site structures ( see Fig 2 ) , including the vast majority ( up to 97 . 63% ) of experimentally validated sites from Diana TarBase and CLIP/CLASH binding site datasets ., Moreover , while these relaxed conditions for the seed region generate a much larger number of candidate sites , the decision of whether a site represents a functional target is delegated to the ANN ( which considers the entire miRNA & mRNA sequence ) ., In this way , we ensure that minimal assumptions , and hence bias , are incorporated into the analysis ., To further evaluate the impact of choice of CSSM , we also implemented the CSSMs used in two of the most commonly used miRNA target prediction tools: Both these CSSMs are subsets of CSSM-miRAW-6-1:10 and CSSM-miRAW-7-1:10 ( Fig 2 ) ., Implementation of different CSSMs served the primary purpose of fine-tuning miRAW but also allowed us to investigate the targeting process from a biological perspective ., The 5 proposed methods encapsulate different target ranges ., At one extreme , CSS-miRAW-TS and CSS-miRAW-P adopt conservative approaches oriented towards canonical sites but they also consider a limited number of non-canonical sites with small irregularities in the seed region; at the other extreme , the other non-canonical CSSMs follow a greedier approach that allows the consideration of several non-canonical sites with broader irregularities in the seed region ., These differences produce variations in both the canonical and non-canonical predicted targets ., As an ANN requires numerical data for input , we transformed the miRNA and candidate mRNA site transcripts to binary values using one hot encoding ., Each of the mRNA and miRNA nucleotides was translated to a binary vector of dimension 4 , corresponding to the four possible nucleotide values ( see Table 1 ) ., Thus , each miRNA target is represented by two concatenated binary vectors: one composed of dimension 120 ( 4x30nt , where 30nt accommodates the longest known miRNA ) corresponding to the mature miRNA transcript , and a second composed of dimension 160 ( 4x40nt ) corresponding to the mRNA site ( 30 nt ) and 5 additional upstream and downstream nucleotides ., These additional nucleotides seek to capture any influence that the flanking sequence may exert on the target 32 , 33 ., The optimal number of additional upstream/downstream nucleotides was determined by evaluating how it affected the predictive power of the neural network ( see Fig B in S1 File ) ., The number of additional nucleotides also conditions the window step size used when scanning the 3’UTR—a smaller window would result in a redundant analysis of potential sites by the neural network whilst a larger step would result in unscanned regions within the 3’UTR ., Classification of candidate miRNA:MBSs was performed using a feed forward deep ANN ., As we rely on the network to identify the relevant relationships between a sequence and the features that describe the miRNA:mRNA interaction , the input of the network consisted of the binarized transcripts of the miRNA and the MBS ., The network was configured so that the number of inputs in the input layer was equal to the dimensionality of the binarized representation of the miRNA:mRNA transcripts , and the output layer consisted of two outputs ( positive and negative class classification ) ., In addition , transcripts were aligned so the start of the seed region corresponded always to the same input node ., The deep ANN was composed of eight dense hidden layers ( comprising rectifier activation function –RelU– nodes ) whilst the output layer comprised two softmax output nodes ., The shape of the network was consistent with its intended functionality:, ( i ) the first hidden sparse layer increases the dimensionality of the problem allowing the representation of data in a more complex dimension ( over-completion ) ; this layer does not necessarily improve the efficiency of the network autoencoder but it gives it “room” to explore the search space ., ( ii ) Hidden layers one to five aim to identify the relevant features representing the data; they correspond to the first half of a stacked autoencoder ., These layers were pre-trained as an isolated autoencoder in order to learn the features that are most representative of miRNA:MBS duplexes ., ( iii ) the last three layers are responsible for classifying the features learned by the autoencoder; and follow the typical shape of a feedforward classification network ., The number of nodes per layer was chosen experimentally using the guidelines in 34 as a starting point and resulted in the structure shown in Fig C and Fig D in S1 File ., The size of the autoencoder was determined by minmizing the number of layers required to compress the data without losing important information ( relative error < 0 . 05 ) ; i . e . , a smaller network may struggle to capture important information whereas a larger one may require additional training time and would have higher overfitting risk ., To ensure the network’s capacity to deal with newly observed data and to avoid overfitting , training was performed with a dropout rate of 0 . 2 ., The maximum number of epochs was set to 500 in order to prevent excessive training time and overfitting ., We tested two different loss functions for the network: negative log likelihood ( NLL ) and cross entropy ( XENT ) ., After performing cross-validation , and determining the best model configuration ( see “Results:Neural Network Evaluation” ) we generated miRAW’s ANN model by retraining the network using the complete training dataset ( with the same proportion of positive and negative class instances ) ., The two neurons of the output layer correspond to the negative ( output 0 , o0 ) and positive ( output 1 , o1 ) classes ., Therefore , the class of the site is determined by the values of the two output neurons:, c l a s s = { 1 if o 1 - o 0 > 0 - 1 if o 1 - o 0 ≤ 0 ( 1 ) This method will assign a positive or negative classification even if there is only a small difference between the positive and negative output neurons ., This scenario corresponds to situations where the network is not confident about the classification of the input data ., To deal with such uncertainty a constant parameter K was used to define a ‘grey area’ in which the network is not able to provide a reliable prediction:, c l a s s = { 1 if o 1 - o 0 ≥ K - 1 if o 1 - o 0 ≤ - K u n k n o w n if - K < o 1 - o 0 < K ( 2 ) According to 35 , we consider that a miRNA targets an mRNA if any of the potential MBSs of the mRNA are functional ., In the representation of the targeting process implemented within miRAW , we require the neural network classify at least one candidate site as positive to consider a miRNA:mRNA pair as a positive targeting event ., In our model , given a miRNA m and a gene g , a candidate site selection method sm ( m , g ) determines a set of potential MBSs for that pair , i . e ., s m ( m , g ) = C S ( 3 ), C S = { c s 0 … c s i } ( 4 ), To determine if the miRNA is targeting the gene , each candidate site within the miRNA:mRNA segment is binarized and input to miRAW’s deep ANN ., The result of the targeting prediction T ( m , g ) corresponds to the disjunction of the neural network outputs ( ann ( m , csi ) ) for all the candidate sites csi ∈ CS in the gene g ., T ( m , g ) = ⋁ i = 0 | C S | a n n ( m , c s i ) ( 5 ) The fact that it only requires a single candidate site to be classified as positive for the miRNA:mRNA prediction to be positive implies that miRAW is particularly sensitive to false positives ., A false negative for a single candidate site can be abrogated by a positive classification for any of the remaining candidate sites but a single false positive cannot be corrected by any number of negative candidate sites ., Hence , the more potential sites a CSSM identifies , the higher the probability of obtaining a false-positive prediction , reducing the performance of the classification due to a lower precision ., P ( F P C S S M ) = 1 - P ( ! F P C S S M ) = 1 - ( 1 - F D R ) | C S | P ( F P C S S M ) = 1 - p r e c i s i o n | s i t e s | ( 6 ), where FDR corresponds to the false discovery rate of the neural network ( that can also be definied as 1—precision ) ., This also implies that CSSMs that adopt a greedier approach will end up obtaining more false positives by chance ., The presence of false positives in miRAW’s ANN can be partially attributed to the fact that not all the information concerning miRNA targets can be obtained from the miRNA:MBS duplex and , therefore , cannot be inferred by the neural network ., For instance , aspects such as site accessibility 36 require accessing additional external data sources ., This external information can be used to refine ANN outcomes by removing sites unlikely to be functional ., In an attempt to reduce the likelihood of false positives , we included an a posteriori filtering step based on accessibility energy ., It is known that miRNA binding sites that are more easily accessible tend to have higher chances of being functional targets 36; for this reason , several tools usch as PITA , miRMAP 37 or PACMIT 38 combine this information with the binding site minimum free energy ( ΔGduplex ) to produce a refined target prediction ., The site accessibility energy ( ΔGopen ) of a MBS can be defined as the energy required to unfold the secondary structure of the mRNA in order to accommodate the miRNA 36 , 38 ., As the calculation of ΔGopen requires information that extends beyond the MBS and which involves the whole mRNA sequence , it is particularly well suited for use as a posteriori filter in miRAW ., Following the site accessibility energy definition of 38 , we implemented an ΔGopen filter that removed all predicted sites presenting a ΔGopen higher than a threshold thsa ., Based on results from previous studies 36 , 38 , we set thsa = −10kcal/mol ., For accuracy and robustness , we computed local site accessibility following the guidelines defined in 39 and 38 ., Specifically , we used the ViennaRNA package 26 and considered the 200nt surrounding the target rather than folding the whole mRNA sequence as this may result in less accurate and more complex secondary structures 38 ., miRAW was implemented using Java ., RNACoFold from the ViennaPackage 26 was used for computing the candidate sites ., Implementation of the deep neural network was done using the DeepLearning4Java ( DL4J ) library 40 ., DL4J allows the use of both CPU and GPUs for neural network training and classification ., All the analyses presented in this paper were performed using GPUs due to its improved performance; however , a CPU based version of miRAW is also available ., To assess miRAW’s performance , we compared it against the following commonly used target prediction tools: TargetScan release 7 . 1 6 , Diana microT-CDS v . 4 41 , PITA v . 6 36 , miRanda ( built upon the mirSVR predictor ) 42 , mirzaG 43 , Paccmit 44 and mirDB 45 ., These represent the current gold standards ( i . e . , most commonly referenced ) for microRNA target prediction software ., These software periodically release datasets containing all available predicted target databases ( as of January 2017 ) with the test datasets defined in “Methods:Dataset Preparation” ., We performed 10-fold validation for the TarBase test , and repeated the analysis for 100-fold ( Fig F in S1 File ) and the full TarBase test ( Fig G , Fig H and Table D in S1 File ) ., For the 10-fold analysis we also characterized the predictions in terms of the structure and site accessibility energy distributions to try and to gain a better understanding of the identified targets in these terms ., If a miRNA:mRNA was present in the test dataset , but missing in a release dataset , then that interaction was assigned the lowest possible score for that algorithm ., For genes with multiple annotated 3’ UTRs we selected the isoform used to build the test dataset ., In prediction datasets where such an isoform was not available or not specified , we tested the longest isoform available ., The comparison was performed using the optimal reported configuration for each tool ., For the release datasets that provided a prediction score , we used a variable threshold and the pRoc R package 46 to build receiver operating characteristic ( ROC ) curves ., To assess the significance of the results , we performed a Wilcoxon signed rank test for each of the evaluated metrics; Results were considered significant for p < 0 . 05 unless otherwise stated ( see S3 File for specific p-values ) ., To evaluate the correlation between the target predictions for a miRNA and the corresponding gene expression profiles in the Transfection Test Dataset we computed the coefficient of determination ( r2 ) between the reported predictions for each different algorithm and the changes of gene expression levels present within the corresponding transfection dataset ., As some methods such as miRAW or microT adopt binary approaches that only evaluate whether or not a miRNA is targeting a gene , rather than quantifying the repression efficacy of the miRNA target , we also examined the expression changes for the top 1% of predicted targets for each of the algorithms , counting how many of the predicted targets presented significant changes in their expression levels ., In contrast to the r2 test , this approach determines if higher confidence predictions are reliable and does not penalize methods that follow less restrictive approaches in order to favour sensitivity ., As before , we considered the interactions not present in the release datasets as negatives and we assigned them the lowest score for that algorithm ., mirSVR , mirDB and mirza-G have been excluded from this test due to absence of predictions for the transfected miRNAs we evaluated here ., All genes within the testing datasets correspond to the GRCh38 release of the reference human genome whilst all the miRNAs appear in miRBase version 19 and later ., All tested target prediction datasets with the exception of PITA and miRanda were built using these , or more recent , releases of miRBase and the reference human genome ., This might impair the performance of PITA and miRanda as some of their negative predictions may not have been tested ., TargetScan , Paccmit and MIRZA-G offer two different databases in each release , one providing target sites highly conserved among species ( TS_Conserved , Paccmit_Cons MG_Conserved ) and one providing sites not-necessarily conserved among species ( TS_NonConserved , Paccmit_NonCons , MG_NonConserved ) ., In both cases , the two versions of the databases were considered ., Cross validation of miRAW’s ANN presented good results in terms of predicting both positive and negative sites ., This was independent of the loss function used during training , with all evaluated metrics resulting in scores higher than 0 . 90 ( Fig 3 and Tables A and B in S1 File ) ., Nonetheless , accuracy and area under the curve ( AUC ) metrics show that the XENT ( accuracy = 0 . 92 , AUC = 0 . 96 ) loss function resulted in a statistically significantly ( Wilcoxon signed sank test ) better network compared to the NLL function ( accuracy = 0 . 91 , AUC = 0 . 93 ) ., This was reflected in both prediction of positive targets , where the XENT network achieved higher precision , sensitivity and F1-score compared to the NLL network ., For negative target prediction , the XENT network returned a larger number of predictions than the NLL but nevertheless achieved a similar negative precision ., It is worth noting that , across the different folds , the XENT network was less consistent in terms of negative precision than in positive precision and that for most of the folders it presented more FN than FP ., This , combined with the difference in sensitivity and specificity values ( 0 . 92 vs 0 . 94 ) , suggests that the XENT network is slightly biased towards negative predictions as it predicted more negative than positive sites for each fold ., Despite this fact , the statistically significant higher accuracy , AUC and F1-scores ( both positive and negative ) indicate that the XENT network is more appropriate than the NLL network for miRNA target prediction ., Fig 4 shows the receiver operating characteristic ( ROC ) curves for the NLL and XENT networks ., The XENT network has a larger AUC , indicating superior performance ., Moreover , there is a clear difference in shape of the curves and distribution of data points ., The XENT network exhibits a smooth curve with relativity evenly spaced points , the NLL curve is more discontinuous and the data points are concentrated within a smaller region ., This indicates a stronger polarization of the NLL network , where all the predictions are strongly classified as a positive or a negative target ( i . e . class value is very close to 1 or -1 ) ., Conversely , the smoothness of the XENT network represents a more progressive classification , allowing the presence of less polarized predictions , resulting in a more generalized predictive ability ., The shape of the NLL curve also suggests that the NLL network might be overfitted and that it might struggle to classify new observations that significantly differ from the training data—this is also supported by the average epoch numbers used by each network to reach its optimal set of weights , 7 . 32 for the XENT network versus 11 . 21 for the NLL network ., The general consistency of calculated parameters and ROC curves across the different folds in the two networks ( Tables A and B in S1 File ) indicates that the model performance is not dependent on the training and test datasets used ., Fold 7 of the XENT network achieved the highest performance in terms of all the considered evaluation metrics and so this ANN model was selected for testing in the gene prediction stage ., To investigate the impact of the site selection method , we compared the performance of five different CSSMs ( miRAW-6-1:10 , miRAW-7-1:10 , miRAW-7-2:10 , miRAW-TS and miRAW-Pita ) in the presence and absence of a site-accessibility energy ( AE ) filter of -10kcal/mol , summarized in Fig 5 and Table C in S1 File ., All the methods achieve accuracies between 0 . 64 and 0 . 74 with significant differences depending on if the site-accessibility filtering is present ( AE ) or absent ( NF ) ., This effect can be seen when the different CSSMs are ordered by accuracy ., miRAW-Pita-NF , miRAW-TS-NF , miRAW-7-2:10-AE and miRAW-7-1:10-AE obtain similar accuracies ( ≈ 0 . 72 ) with no statistically significant differences in the metric , although miRAW-6-1:10-AE has a slightly poorer performance ( 0 . 71 ) ., However , miRAW-7-2:10-NF , miRAW-7-1:10-NF , miRAW-Pita-AE and miRAW-TS-NF ranked in the bottom of the table in terms of accuracy ., Thus , while the non-canonical CSSMs specifically derived for miRAW obtain better results in the presence of filtering , the canonical derived CSSMs ( miRAW-Pita and miRAW-TS ) exhibit improved performance in the absence of filtering ., The F1-scores summarize how well a particular class is classified by a particular CSSM , an optimal CSSM will perform well for both positive and negative targets ., Fig 5 and Table C in S1 File show that CSSMs with reported low accuracy underperform in at least one of the F1-scores: miRAW-7-2:10-NF , miRAW-7-1:10-NF and miRAW-6-1:10 have high positive F1-scores but a low negative F1-score caused by an excess of false positives ( causing a low specificity ) whilst miRAW-Pita-AE and miRAW-TS-AE have a low positive F1-score cause by the excess of false negatives ( causing a low sensitivity ) ., However , miRAW-Pita-NF , miRAW-TS-NF , miRAW-7-2:10-AE and miRAW-7-1:10-AE all obtain balanced F1-scores ranging between 0 . 71 and 0 . 74 , indicating an ability to effectively predict both positive and negative targets ., Fig 6 summarizes the composition of site types by each CSSM ., Fig 6a shows the average number of canonical ( blue ) and non-canonical ( green ) sites identified for each miRNA:gene pair in the test dataset whilst Fig 6b shows the relative proportions of each identified type ., As expected , CSSMs methods following conservative approaches ( miRAW-Pita and miRAW-TS ) identified more canonical than non-canonical potential sites , whereas the miRAW CSSMs identified larger total numbers of potential sites , of which many more were non-canonical ., The figure also shows that the number of predicted canonical sites varies according to the selected CSSM , with the conservative approaches obtaining more canonical sites than the greedy approaches ., While this seems to contradict the expectation that all the CSSMs should identify similar canonical sites , the difference can be understood when the higher number of accepted binding structures recognized by the non-canonical oriented CSSMs are taken into account ., Several of the sites identified by miRAW-Pita or miRAW-TS overlap with non-canonical binding sites predicted by the miRAW specific CSSMs that present greater stability and are therefore preferentially selected ( Fig 7 ) ., Fig 6b also shows that the application of the site accessibility filter does not signific | Introduction, Materials and methods, Results, Discussion | MicroRNAs ( miRNAs ) are small non-coding RNAs that regulate gene expression by binding to partially complementary regions within the 3’UTR of their target genes ., Computational methods play an important role in target prediction and assume that the miRNA “seed region” ( nt 2 to 8 ) is required for functional targeting , but typically only identify ∼80% of known bindings ., Recent studies have highlighted a role for the entire miRNA , suggesting that a more flexible methodology is needed ., We present a novel approach for miRNA target prediction based on Deep Learning ( DL ) which , rather than incorporating any knowledge ( such as seed regions ) , investigates the entire miRNA and 3’TR mRNA nucleotides to learn a uninhibited set of feature descriptors related to the targeting process ., We collected more than 150 , 000 experimentally validated homo sapiens miRNA:gene targets and cross referenced them with different CLIP-Seq , CLASH and iPAR-CLIP datasets to obtain ∼20 , 000 validated miRNA:gene exact target sites ., Using this data , we implemented and trained a deep neural network—composed of autoencoders and a feed-forward network—able to automatically learn features describing miRNA-mRNA interactions and assess functionality ., Predictions were then refined using information such as site location or site accessibility energy ., In a comparison using independent datasets , our DL approach consistently outperformed existing prediction methods , recognizing the seed region as a common feature in the targeting process , but also identifying the role of pairings outside this region ., Thermodynamic analysis also suggests that site accessibility plays a role in targeting but that it cannot be used as a sole indicator for functionality ., Data and source code available at: https://bitbucket . org/account/user/bipous/projects/MIRAW . | microRNAs are small RNA molecules that regulate biological processes by binding to the 3’UTR of a gene and their dysregulation is associated with several diseases ., Computationally predicting these targets remains a challenge as they only partially match their target and so there can be hundreds of targets for a single microRNA ., Current tools assume that most of the knowledge defining a microRNA-gene interaction can be captured by analysing the binding produced in the seed region ( ∼ the first 8nt in the miRNA ) ., However , recent studies show that the whole microRNA can be important and form non-canonical targets ., Here , we use a target prediction methodology that relies on deep neural networks to automatically learn the relevant features describing microRNA-gene interactions for predicting microRNA targets ., This means we make no assumptions about what is important , leaving the task to the deep neural network ., A key part of the work is obtaining a suitable dataset ., Thus , we collected and curated more than 150 , 000 experimentally verified microRNA targets and used them to train the network ., Using this approach , we are able to gain a better understanding of non-canonical targets and to improve the accuracy of state-of-the-art prediction tools . | transfection, learning, neural networks, engineering and technology, gene regulation, social sciences, neuroscience, learning and memory, artificial neural networks, micrornas, cognitive psychology, mathematics, nuclear engineering, forecasting, statistics (mathematics), artificial intelligence, computational neuroscience, molecular biology techniques, research and analysis methods, computer and information sciences, mathematical and statistical techniques, gene expression, molecular biology, biochemistry, rna, site selection, psychology, nucleic acids, genetics, biology and life sciences, structural engineering, physical sciences, computational biology, non-coding rna, cognitive science, statistical methods | null |
journal.pbio.1002217 | 2,015 | Sustained Pax6 Expression Generates Primate-like Basal Radial Glia in Developing Mouse Neocortex | The evolutionary expansion of the mammalian neocortex is thought to be primarily the consequence of the increasing proliferative capacity of cortical stem and progenitor cells during development 1–9 ., Recent studies have progressively focused on differences between species regarding the type , abundance , and modes of division of cortical stem and progenitor cells , which are thought to contribute to the variety of shapes and sizes of the neocortex present across mammals 1–8 ., A hallmark of the developing cortical wall is its apical–basal polarity , with the apical side corresponding to the ventricular surface and the basal side contacting the basal lamina 4 , 10 ., At the onset of neurogenesis , neuroepithelial cells , the primary cortical stem cells , transform into apical radial glia ( aRG ) 11 , 12 ., aRG , together with apical intermediate progenitors , constitute apical progenitors ( APs ) , as they repeatedly undergo mitosis at the apical surface of the cortical wall 8 , 10 ., Apical intermediate progenitors ( previously called short neural precursors ) undergo self-consuming division generating two neurons 13–15 ., In contrast , aRG undergo self-renewing divisions , generating neurons and , more frequently , basal progenitors ( BPs ) that delaminate from the apical surface , leave the ventricular zone ( VZ ) and move basally to the subventricular zone ( SVZ ) 16–24 ., BPs comprise basal radial glia ( bRG , also called outer radial glia ) and basal intermediate progenitors ( bIPs ) 8 , 10 ., BPs typically undergo mitosis in the SVZ and can undergo , in principle , neurogenic ( i . e . , neuron-producing ) or proliferative ( i . e . , self-amplifying ) divisions , albeit with profound differences in occurrence between species 8 , 16–18 , 20–22 , 25–31 ., bRG can be distinguished from the process-lacking bIPs by their apically and/or basally directed processes at mitosis 8 , 17 , 18 , 21–28 , 31 ., Comparison of BPs in various mammalian brains has revealed key differences in their abundance and mode of cell division 1–6 , 8 , 32–34 ., Thus , such differences have been reported for bIPs , which can be classified into two principal types , neurogenic and proliferative , depending on the mode of cell division ( generating two neurons and two bIPs , respectively ) 8 ., In the mouse and rat SVZ , neurogenic bIPs constitute the vast majority of BPs ( >80% ) 16–18 , 21 , 22 , whereas proliferative bIPs and bRG exist in only small proportions 17 , 28–30 , 35 ., Moreover , mouse bRG typically undergo asymmetric self-renewing neurogenic divisions but not symmetric proliferative divisions 28 ., By contrast , in mammals exhibiting an increased abundance of BPs and an enlarged SVZ , as characterized in detail in species such as ferret , macaque , and human 1 , 4–6 , 8 , 23 , 32 , bIPs are mostly of the proliferative type , and bRG constitute at least half of all BPs 23–27 ., Moreover , in these species , both bRG and proliferative bIPs undergo mostly symmetric proliferative rather than neurogenic divisions 23 , 24 , 31 ., These self-amplifying divisions significantly increase the number of BPs residing in the SVZ , consequently leading to the expansion of the SVZ ., Moreover , the SVZ of these animals comprises not only a rodent SVZ-related layer called the inner SVZ ( iSVZ ) but in addition a novel layer called the outer SVZ ( oSVZ ) 32 ., Importantly , these alterations in the mode of cell division and the resulting increase in BP abundance and formation of an oSVZ have been hypothesized to be major causes underlying the expansion of the neocortex 2–6 , 8 , 32 ., A key question then is how these differences in BP abundance and mode of cell division between rodents and primates are brought about at the molecular level ., A candidate regulatory mechanism is the differential expression of transcription factors ., Of particular interest in this regard is Pax6 ( accession number: AAH36957 ) , a paired-box transcription factor 36–39 ., Several mouse and rat mutant models have demonstrated that Pax6 is required for normal aRG abundance and mode of cell division 37 , 40–49 ., Moreover , although Pax6 mRNA levels are generally lower in BPs than APs , this down-regulation is much greater for mouse than human 50 ., Consistent with this , only a minority of mouse and rat BPs ( <30% ) show Pax6 immunoreactivity ( which is of lower level than in APs ) 3 , 51 , 52 , whereas the opposite is the case for primate , notably human , BPs ( >80% Pax6-positive ) , with essentially all bRG and the majority of bIPs containing this transcription factor 3 , 23–27 , 53 , 54 ., Together , these findings raise the possibility that the differences in Pax6 expression between rodent and primate BPs may be responsible , at least in part , for the greater abundance and proliferative or self-renewal capacity of the latter ., We therefore sought to maintain Pax6 expression specifically in newly generated BPs in order to investigate if such expression would increase the abundance of BPs , notably of bRG , and their proliferative or self-renewal capacity ., Using a novel approach of conditional Pax6 expression 16 , 21 , 55 , we find that sustaining elevated Pax6 levels in BP-genic mouse aRG and the BP progeny derived therefrom increases both the proportion of bRG among the newly generated BPs and the self-renewing capacity of BPs ., In mouse , the aRG subpopulation that gives rise to BPs , in contrast to self-amplifying aRG , specifically expresses Tis21 , a pan-neurogenic progenitor marker 16 , 21 , 55 ., Thus , as a tool towards maintaining Pax6 expression in mouse BPs , we generated a Tis21–CreERT2 knock-in mouse line ., In this mouse line , exon 1 of Tis21 is replaced by CreERT2 containing a herpes simplex virus ( HSV ) tag at its C-terminus via homologous recombination ( Fig 1A; for details , see S1 Fig ) , in order to limit Cre expression to Tis21-positive cells ., To assess the cellular specificity of Cre expression , Tis21–CreERT2 knock-in mice were crossed with Tis21–GFP knock-in mice 16 ., Immunofluorescence of the dorsolateral telencephalon of double-transgenic mice at embryonic day ( E ) 10 . 5 , corresponding to the onset of Tis21 expression , and at E13 . 5 , corresponding to the time point at which the in utero electroporations described below were conducted , showed that Cre was expressed in essentially the same cells as GFP ( Fig 1B and 1C ) , indicating its expression selectively in the neurogenic subpopulations of cortical progenitors ., Specifically , quantitation at E10 . 5 revealed that 97% of the cells containing nuclear Tis21–GFP were also positive for cytoplasmic Cre ( Fig 1D ) , and no Cre was detected in Tis21–GFP-negative cells ., We next ascertained that the Tis21–CreERT2 mouse exhibits tamoxifen-dependent recombination by crossing this mouse line with a conditionally activateable GFP reporter mouse line , RCE:loxP 56 ( Fig 1E ) ., In these double-transgenic mice , GFP should be expressed only when CreERT2 has been translocated from the cytoplasm into the nucleus and excised a stop cassette that prevents the transcription of the GFP mRNA; the estrogen analog tamoxifen induces such CreERT2 translocation 57 ., Indeed , no GFP-positive cells were observed in the absence of tamoxifen ( Fig 1G ) ., In contrast , when treated with tamoxifen ( Fig 1F ) , GFP fluorescence was observed throughout the double-transgenic mouse brain ( Fig 1I ) , and GFP-positive cells were found in all layers of the embryonic neocortex ( Fig 1I’ ) ., This reflected Cre recombinase activity , because no GFP expression was observed when tamoxifen was administered to RCE:loxP offspring lacking the Tis21–CreERT2 allele ( Fig 1H ) ., We conclude that Tis21–CreERT2 mouse embryos can be used to obtain tamoxifen-dependent recombination specifically in the neurogenic subpopulations of cortical progenitors ., To conditionally express Pax6 in BP-genic aRG of developing neocortex , we introduced a floxed Pax6 plasmid at midneurogenesis into APs of tamoxifen-treated Tis21–CreERT2 mouse embryos ., Specifically , we generated a plasmid ( referred to as Pax6-expressing plasmid ) containing a constitutive promoter ( CAG ) followed by a membrane ( GAP43 ) –GFP cassette flanked by two loxP sites , mouse Pax6 , an internal ribosome entry site ( IRES ) sequence , and nuclear RFP ( nRFP ) ( Fig 2A ) ., Upon Cre-mediated recombination , the membrane–GFP cassette would be excised , leading to the simultaneous expression of Pax6 and nRFP ., Introduction of this plasmid into APs of tamoxifen-treated Tis21–CreERT2 mouse embryos should ensure maintenance of Pax6 expression as mouse BPs arise from aRG divisions , as well as during their subsequent migration to , and function in , the SVZ ., An identical plasmid but lacking the Pax6 and IRES sequences served as control ( Fig 2A ) ., We first validated the Pax6-expressing plasmid by transfection of HEK 293T cells , a cell line in which the endogenous PAX6 gene is not expressed ., Transfection with the Pax6-expressing plasmid alone resulted in GFP , but not nRFP , expression ., Cotransfection of the Pax6-expressing plasmid and a Cre-expressing plasmid yielded both Pax6 and nRFP expression , whereas only nRFP expression was observed upon cotransfection of the control plasmid and the Cre-expressing plasmid ( S2 Fig ) ., We then explored whether the Pax6-expressing plasmid could be used in Tis21–CreERT2 mouse embryos to obtain conditional Pax6 expression specifically in the neurogenic subpopulation of APs and their progeny ., To this end , we used the in utero electroporation technique where an electric field is generated across the cortical wall in order to allow for the unidirectional delivery of the negatively charged plasmid DNA , injected into the ventricular lumen , into APs ., Dorsolateral telencephalon of tamoxifen-pretreated ( E12 . 5 ) Tis21–CreERT2 mice was electroporated with Pax6-expressing plasmid at E13 . 5 and analyzed at E14 . 5 , the peak of BP generation from neurogenic aRG 22 ( Fig 2B ) ., For the ease of presentation , we shall refer to this approach from here onwards simply as conditional Pax6 expression ., Analysis of the Pax6 expression pattern yielded the following observations ., First , analysis of the level of Pax6 immunoreactivity revealed that a subpopulation of cells had higher Pax6 immunoreactivity upon conditional Pax6 expression than in the control ( Fig 2C and 2D ) ., Upon closer inspection , all these highly Pax6-immunoreactive cells were RFP-positive , indicating that these cells constituted Pax6-expressing-plasmid–electroporated neurogenic APs and their progeny ( Fig 2C , 2D and 2F ) ., The level of Pax6 immunoreactivity in these cells in the VZ was approximately 3-fold higher than that of the nonelectroporated APs or control-plasmid–electroporated neurogenic APs and their VZ progeny ( Fig 2G ) , essentially all of which are known to express endogenous Pax6 37 , 51 , 52 ., In the SVZ , where mouse BPs normally down-regulate Pax6 expression 3 , 50–52 , this difference was even greater ( ≈6-fold higher ) ( Fig 2H ) ., Second , the appearance of these highly Pax6-immunoreactive and RFP-positive cells upon Pax6-expressing plasmid electroporation was strictly dependent on tamoxifen pretreatment ( S3 Fig ) ., Together , these observations allow us to equate the RFP-positive cells with the cells containing Pax6 due to the electroporation ., To distinguish these conditionally Pax6-expressing cells from the cells expressing Pax6 endogenously , we shall refer to them from here onwards as exogenous Pax6- ( exoPax6- ) expressing cells ., In addition , considering the results shown in Fig 1 , we conclude that these cells constitute specifically the neurogenic subpopulation of APs and their progeny , notably the aRG-derived BPs ., Third , we found that electroporation with Pax6-expressing plasmid did not affect , after 24 h , the distribution of the progeny ( RFP+ cells ) of the electroporated neurogenic APs between ( Fig 2C–2E ) and within ( S4 Fig ) the germinal layers ( i . e . , VZ and SVZ ) ., This implies that conditional Pax6 expression in neurogenic APs and their progeny , even if this expression exceeds the normal endogenous level , does not cause any overt effects on cell migration within the first 24 h after electroporation ., The finding that RFP-positive cells are similarly distributed in control and upon conditional Pax6 expression allows for a valid comparison between germinal layers of the effect of conditional Pax6 expression in subsequent experiments ., Conditional Pax6 expression in aRG has previously been found to induce apoptosis when pan-aRG Cre drivers based on Emx1 and hGFAP promoter and regulatory sequences were used ., However , this phenomenon was not observed with a Cre driver based on Ngn2 expression 58 , which , similar ( but not identical ) to Tis21 expression , is characteristic of neurogenic progenitors 59 ., It was therefore important to ascertain that conditional expression of Pax6 in Tis21–CreERT2 mice would not induce apoptosis ., Indeed , immunofluorescence for the apoptosis marker activated caspase-3 did not reveal any significant difference in the number of caspase-3–positive cells between the progeny of control-plasmid–and Pax6-expressing-plasmid–electroporated neurogenic APs ( S5 Fig ) ., We therefore conclude that the present approach of conditional Pax6 expression is suitable to maintain high levels of Pax6 expression specifically in neurogenic APs and their progeny , notably the aRG-derived BPs , thus recapitulating the Pax6 expression pattern observed in BPs of developing primate neocortex ., In assessing the functional consequences of sustained Pax6 expression in BPs , we sought to obtain initial clues as to the identity of the progeny of the Pax6-electroporated neurogenic APs ., Using the cycling cell marker Ki67 , we first investigated whether the exoPax6-expressing cells exhibited the same proportion of progenitors versus neurons as control cells ( Fig 2I–2K ) ., Whereas conditional Pax6 expression did not alter the percentage of Ki67-positive cells in the VZ , it did result in a significant increase in Ki67-positive cells in the SVZ ( Fig 2K ) ., This suggested that the conditional Pax6 expression increased the population of cycling BPs derived from electroporated aRG ., We noticed in some experiments that in both control and conditional Pax6 expression , more Ki67-positive cells were observed in the basal region of the SVZ , and in particular in the intermediate zone of the electroporated area , but not in the contralateral area nor in nonelectroporated dorsolateral telencephalon ., This reflected a previously described side effect of in utero electroporation , that is , the displacement of some Pax6-positive cells towards the cortical plate 60 ., Importantly , this side effect does not affect the findings described in the present study for three reasons ., First , all our data are comparisons between control and conditional Pax6 expression , both of which involve identical conditions of in utero electroporation ., Second , all our quantifications are confined to electroporated , RFP-positive cells , and the electroporation side effect has been reported to affect mainly nonelectroporated cells 60 ., Third , our quantifications of cells in the SVZ exclude cells in the intermediate zone ., To gain further insight into a possible regulation of the cell cycle of cortical progenitors by conditional Pax6 expression , we examined specific cell cycle parameters ., We first examined the effect of conditional Pax6 expression on the total cell cycle length ( Tc ) of neurogenic aRG by performing live imaging on E14 . 5 organotypic slices prepared from control or Pax6-expressing plasmid–electroporated brains ., The time period between two successive aRG mitoses was taken to indicate the length of the cell cycle , Tc ., In both control and conditional Pax6 expression , we observed no major difference in Tc , although there was a trend for a shorter Tc upon conditional Pax6 expression ( control , 21 . 0 ± 3 . 3 h , n = 8 cells versus Pax6 , 18 . 5 ± 1 . 2 h , n = 9 cells , S1 Table top ) ., To estimate the proportion of the progeny of control-plasmid–and Pax6-expressing-plasmid–electroporated neurogenic APs that were in S-phase , we performed pulse-labeling with the thymidine analog EdU one hour before analyzing the embryos at E14 . 5 ., This revealed that a significantly greater proportion of the exoPax6-expressing progeny than of the control progeny was in S-phase , in both the VZ and SVZ ( Fig 3A–3C ) ., Given that conditional Pax6 expression did not increase the population size of cycling APs ( Fig 2K ) , nor alter much their Tc ( S1 Table top ) , the increase in the proportion of cells in S-phase in the VZ ( Fig 3C ) likely reflected a greater share of S-phase in the AP cell cycle , rather than an increase in cycling APs as such ., To address this directly , we performed a dual pulse chase experiment as previously described 61 ( see S6A Fig and Materials and Methods ) in order to determine the length of S-phase ., We observed a significant increase in the length of S-phase for the sum of the electroporated aRG and their progeny upon conditional Pax6 expression ( S6 Fig ) ., We further corroborated this by analyzing the pattern of immunofluorescence of the cycling cell marker proliferating cell nuclear antigen ( PCNA ) ., Like other cycling cells , cortical progenitors in S-phase show a punctate nuclear PCNA pattern , whereas progenitors in G1 and G2 show diffuse nuclear PCNA immunoreactivity 23 , 52 , 62 ., Based on punctate PCNA staining , we observed a proportion of neurogenic APs in S-phase upon control electroporation that was similar to previously published data on E14 . 5 Tis21-positive APs 52 ( S1 Table middle ) ., Conditional Pax6 expression , however , was found to significantly increase the percentage of PCNA-positive nuclei in the VZ that showed a punctate pattern ( Fig 3D–3F ) , i . e . , increased the proportion of neurogenic APs that were in S-phase ., These findings , together with the Ki67 ( Fig 2K ) and EdU ( Fig 3C ) data , imply that conditional Pax6 expression increases the relative proportion of S-phase within the AP cell cycle ., As there was no significant difference in Tc but an increase in the proportion of cells in S-phase upon conditional Pax6 expression in Tis21-positive APs , we hypothesized that the G1-phase must have been shortened to compensate for the longer S-phase ., Consistent with this hypothesis , a significantly smaller proportion of the exoPax6-expressing progeny in the VZ than of the control progeny of electroporated neurogenic APs was positive for cyclin D1 , a cyclin that is expressed from mid- to late-G1 ( Fig 3G–3I ) ., To estimate the length of the G1-phase , we combined the data obtained from live imaging with the punctate PCNA staining data ( S1 Table bottom ) ., As none of the apical mitoses observed lasted for >1 h and no difference in G2 length was reported between neural progenitors 52 , we assumed that the proportion of neurogenic aRG in G2- and M-phase remained unchanged upon conditional Pax6 expression ., Similar to the data obtained for cyclin D1 ( Fig 3I ) , we estimated a shorter G1-phase upon conditional Pax6 expression ( control 15 . 6 h versus Pax6 12 . 8 h , S1 Table bottom ) ., As to BPs , the increase in the proportion of EdU-positive cells in the SVZ upon conditional Pax6 expression ( Fig 3C ) was consistent with that of Ki67-positive cells ( Fig 2K ) , corroborating our conclusion that the population size of cycling BPs derived from electroporated aRG was increased under this condition ., Further support for this population size increase was provided by immunofluorescence for phosphohistone H3 , a marker of cells in late G2- and M-phase , which revealed a significant increase in mitotic BPs derived from electroporated aRG ( Fig 3J–3L ) ., Also in the case of BPs , conditional Pax6 expression significantly increased the relative proportion of S-phase within the cell cycle as revealed by the pattern of nuclear PCNA immunoreactivity ( Fig 3D–3F ) , albeit not at the expense of decreasing the relative proportion of G1 ( Fig 3I ) ., Our group previously reported a difference in S-phase length between Tis21- positive and Tis21-negative APs 52 ., As Tis21-negative and Tis21-positive APs differ in the type of division ( symmetric versus asymmetric ) and progeny produced ( APs versus BPs ) 16 , 55 , we wondered whether the increase in the relative proportion of S-phase within the cell cycle of the exoPax6-expressing APs ( Fig 3F ) may be indicative of an alteration in their mode of division ., To explore this possibility , we investigated the nature of the cycling BPs that were increasingly observed upon conditional Pax6 expression ( Fig 2I–2K ) by examining the expression of two characteristic transcription factors , Tbr2 ( Fig 4A–4C ) and Sox2 ( Fig 4D–4F ) ., Tbr2 is typically expressed by the differentiating progeny of Tis21-expressing aRG fated to become bIPs 22 , 51 , 52 , 63 , whereas Sox2 expression is characteristic of aRG and bRG 23 , 24 , 26 , 28 , 29 , 31 , 64 ., Upon conditional Pax6 expression , analysis for the abundance of Tbr2-positive cells revealed a significant reduction in the exoPax6-expressing progeny as compared to control ( Fig 4A–4C ) ., This reduction was largely accounted for by the decrease in Tbr2-positive cells in the SVZ , most of which presumably were bIPs ( Fig 4C ) ., Conversely , the abundance of Sox2-positive cells was higher in the exoPax6-expressing progeny as compared to the control ( Fig 4D–4F ) ., Remarkably , this increase occurred in the SVZ rather than the VZ ( Fig 4F ) ., This suggested that conditional Pax6 expression , which increased the population of BPs ( Fig 2K ) , induced Tis21-expressing aRG to increasingly generate BPs with a radial glia-characteristic transcription factor expression ( i . e . , bRG ) , at the expense ( at least relatively ) of bIP production ., To directly investigate a possible effect of conditional Pax6 expression on the mode of cell division of neurogenic APs , we performed a daughter cell pair assay 65 by analyzing areas of dorsolateral telencephalon that contained only a few RFP-positive cells in the VZ 24 h after electroporation ., Tbr2 immunofluorescence allowed us to distinguish three types of RFP+ daughter cell pairs: ( 1 ) Tbr2–/Tbr2– ( no bIP daughter cells ) , ( 2 ) Tbr2+/Tbr2– ( 1 bIP daughter cell ) and ( 3 ) Tbr2+/Tbr2+ ( 2 bIP daughter cells ) ( Fig 4G ) ., Importantly , virtually all Tbr2– daughter cells in the VZ are likely to be radial glia , either aRG or newborn bRG , based on the following considerations ., Essentially all daughter cell nuclei in the VZ were PCNA-positive ( S7 Fig ) ., This was in line with the findings that >80% and almost 90% of the progeny in the VZ that was derived from electroporated neurogenic APs were Ki67+ ( Fig 2K ) and Sox2+ ( Fig 4F ) , respectively ., Hence , the Tbr2– daughter cells were radial glial progenitors rather than neurons ., Consistent with this , almost all cells in the mouse E14 . 5 VZ are cycling 52 , and very few of them are newborn neurons 52 ., Quantification of daughter cell pairs in the VZ showed that in the control , the majority ( 77% ) of these pairs derived from AP divisions that had generated bIPs ., Specifically , 56% of divisions were asymmetric ( and presumably self-renewing ) ( Tbr2+/Tbr2– , Fig 4H , red ) , and 21% symmetric self-consuming ( Tbr2+/Tbr2+ , Fig 4H , green ) ., These findings were in line with the fact that the progeny specifically of neurogenic APs was analyzed ., Of note , only 23% of divisions did not generate any bIPs and hence were presumably symmetric proliferative with regard to the radial glia nature of the daughter cells ( Tbr2–/Tbr2– , Fig 4H , blue ) ., In contrast , upon conditional Pax6 expression , the majority ( 59% ) of the daughter cell pairs were derived from neurogenic AP divisions that did not generate bIPs but radial glia ( Tbr2–/Tbr2– , Fig 4H , blue ) ., This occurred at the expense of bIP-generating divisions , that is , asymmetric self-renewing divisions ( Tbr2+/Tbr2– , reduced to 32% , Fig 4H , red ) , and symmetric self-consuming divisions ( Tbr2+/Tbr2+ , reduced to 8% , Fig 4H , green ) ., The observations that conditional Pax6 expression increased, ( i ) the non-bIP generating divisions ( Tbr2–/Tbr2– , Fig 4H , blue ) and, ( ii ) the Sox2-positive progeny in the SVZ ( Fig 4F ) suggested that the former progeny increasingly consisted of newborn bRG ., As bRG are known to delaminate from the ventricular surface 24–27 , 31 , 35 , we explored whether the radial glia progeny in the VZ observed upon conditional Pax6 expression increasingly showed signs of delamination ., To this end , we measured the distance of the ventricular-most nucleus of each Tbr2–/Tbr2– daughter cell pair from the ventricular surface ( Fig 4I ) ., In light of the observation that the mean distance of the ventricular-most nucleus of the control and exoPax6-expressing Tbr2+/Tbr2– and Tbr2+/Tbr2+ daughter cell pairs was always >40 μm ( S8 Fig ) , whereas that of the Tbr2–/Tbr2– pairs was <26 . 5 μm ( Fig 4I , S8 Fig ) , we focused our attention on the abundance of the ventricular-most nuclei of Tbr2–/Tbr2– daughter cell pairs with a distance from the ventricular surface of ≥27 μm ( corresponding to >3 nuclear diameters and referred to as abventricular location 16 ) ., Whereas only 1 of the 7 ventricular-most nuclei ( 14% ) of the Tbr2–/Tbr2– daughter cell pairs in the control was found in an abventricular location , 7 of the 15 nuclei ( 47% ) analyzed upon conditional Pax6 expression were abventricular ( Fig 4I ) ., This suggested that conditional Pax6 expression promoted a substantial proportion of the radial glia progeny derived from neurogenic AP divisions to delaminate from the ventricular surface , as would be expected for newborn bRG ., In species with a high abundance of bRG in the SVZ , the radial thickness of the VZ decreases concomitant with bRG generation 8 , 23 , 25–27 , 64 ., In light of the findings described above , we investigated a possible reduction in VZ thickness upon conditional Pax6 expression by quantifying the total number of nuclei ( both RFP–and RFP+ ) in the VZ within a 200-μm wide , electroporated region of the dorsolateral telencephalon ., Indeed , we observed a significant , approximately 10% , reduction in the number of nuclei in the VZ upon conditional Pax6 expression ( Fig 4J ) ., The magnitude of this reduction was consistent with the efficiency of electroporation and the estimated increase in the proportion of the progeny of electroporated neurogenic APs that delaminated upon conditional Pax6 expression as compared to control ( Fig 4H ) ., Taken together , the findings presented so far strongly suggest that mouse neurogenic APs and their progeny that constitutively express Pax6 increasingly generate bRG at the expense of generating bIPs ., To corroborate and complement these findings , we next investigated the effect of conditional Pax6 expression on the proportion of bRG in the BP progeny of electroporated aRG ., To this end , we analyzed the morphology of mitotic BPs using phosphovimentin immunofluorescence ( Fig 5A–5C ) , which stains both the cell bodies and processes of mitotic cortical progenitors 66 ., bRG characteristically extend basally and/or apically directed processes 23–29 , 31 , 35 , whereas bIPs do not 17 , 18 , 21–26 , 28 , 35 ., As the apically directed processes have been reported to be thinner than basal processes and may not be easily detected via phosphovimentin staining 23 , we focused our analysis on basal process-bearing mitotic BPs ., In the control , the vast majority ( 91% ) of mitotic BPs were nonpolar and only a small minority ( 9% ) extended a basal process ( Fig 5C ) , consistent with the high abundance of bIPs and low abundance of bRG in the embryonic mouse SVZ 28 , 29 , 35 ., In contrast , upon conditional Pax6 expression , we observed a more than 2-fold increase in the proportion of mitotic BPs with a basal process , i . e . of bRG within the BP population ( 23% , Fig 5C ) ., These data show that , concomitant with the increase in the proportion of BPs among the aRG progeny ( Fig 2K ) , conditional Pax6 expression more than doubled the proportion of bRG within these BPs ., As the apically-directed process of bRGs may be harder to detect via phosphovimentin immunofluorescence at mitosis 23 , we next investigated the diversity of bRG morphology during interphase ., To do this , we made use of the residual membrane-GFP ( Fig 2A ) expressed presumably due to incomplete Cre recombination ( see Materials and Methods , live imaging ) ( Fig 5D and 5E ) ., To distinguish bRG from migrating neurons , we stained for Sox2 , which is expressed in radial glia but not in neurons ., In the control , all of the bRG progeny of the electroporated neurogenic APs exhibited a basal process , and 40% of them an apically-directed process as well ( Fig 5F ) ., Upon conditional Pax6 expression , we found an increase in the proportion of bRG exhibiting both basally and apically directed processes ( Fig 5E and 5F , 53% ) and also observed bRG with an apically directed process only ( Fig 5D and 5F , 7% ) ., Interestingly , in the macaque , bRG with both processes and bRG with an apically directed process only have been reported to have a higher self-renewing capacity as compared to bRG with a basal process only 23 ., Of note , the basal process of the bRG generated upon conditional Pax6 expression sometimes extended all the way to the pia ( S9A Fig ) ., The bRG generated upon conditional Pax6 expression were nestin-positive ( S9B Fig ) , could be Tbr2-negative ( S9C Fig ) , and typically exhibited a perinuclear centrosome ( S9D Fig ) ., Furthermore , these cells underwent mitotic somal translocation , in which the cell soma moves rapidly in the basal or apical direction prior to mitosis 23 , 26 , 28 , 31 , as revealed by live time-lapse imaging ( S9E Fig ) ., The data presented so far show increased bRG generation upon elevating Pax6 levels in neurogenic aRG and sustaining it in the BPs derived therefrom ., We sought to complement these findings by a converse , loss-of-function , approach ., To this end , we investigated the proportion of mitotic ( phosphovimentin-positive ) bRG among BPs in the dorsolateral telencephalon of E14 . 5 homozygous small eye ( Sey ) mutant mice , which lack functional Pax6 because of a mutation that generates a premature translational stop codon ( Fig 5G–5I ) ., We found a significant reduction in the percentage of bRG as compared to littermates that have at least one copy of the Pax6 gene ( Fig 5I ) ., These data indicate that although Pax6 function is not absolutely required for bRG generation , its level of expression is crucial for determining the abundance of these cells in the developing mouse neocortex ., Ferret and primate bRG are known to undergo multiple rounds of self-renewing division 23–26 , 31 , whereas bIPs in mouse and rat embryonic neocortex typically undergo one round of self-consuming division 16–18 , 20–22 ., In light of the increase in cycling BPs ( Fig 2K ) and bRG ( Fig 5C ) upon conditional Pax6 expression , it was therefore of interest to investigate whether conditional Pax6 expression would subsequently lead to increased cell cycle re-entry of the BP progeny derived from electroporated aRG ., To this end , a single pulse of EdU was administered at 24 h after electroporation and analyzed after an additional 24 h for the proportion of cycling , Ki67-positive cells among the EdU-labeled progeny of electroporated APs , in order to identify cells that had re-entered the cell cycle ( Fig 6A ) ., In the control , 75% of such daughter cells present in the VZ , but only 23% of such daughter cells in the SVZ , had re-entered the cell cycle ( Fig 6B and 6D ) ., In contrast , upon conditional Pax6 expression , whereas daughter cell cycle re-entry was the same in the VZ , it nearly doubled in the SVZ ( 41% , Fig 6C and 6D ) ., Again , we used the residual | Introduction, Results, Discussion, Materials and Methods | The evolutionary expansion of the neocortex in mammals has been linked to enlargement of the subventricular zone ( SVZ ) and increased proliferative capacity of basal progenitors ( BPs ) , notably basal radial glia ( bRG ) ., The transcription factor Pax6 is known to be highly expressed in primate , but not mouse , BPs ., Here , we demonstrate that sustaining Pax6 expression selectively in BP-genic apical radial glia ( aRG ) and their BP progeny of embryonic mouse neocortex suffices to induce primate-like progenitor behaviour ., Specifically , we conditionally expressed Pax6 by in utero electroporation using a novel , Tis21–CreERT2 mouse line ., This expression altered aRG cleavage plane orientation to promote bRG generation , increased cell-cycle re-entry of BPs , and ultimately increased upper-layer neuron production ., Upper-layer neuron production was also increased in double-transgenic mouse embryos with sustained Pax6 expression in the neurogenic lineage ., Strikingly , increased BPs existed not only in the SVZ but also in the intermediate zone of the neocortex of these double-transgenic mouse embryos ., In mutant mouse embryos lacking functional Pax6 , the proportion of bRG among BPs was reduced ., Our data identify specific Pax6 effects in BPs and imply that sustaining this Pax6 function in BPs could be a key aspect of SVZ enlargement and , consequently , the evolutionary expansion of the neocortex . | During development , neural progenitors generate all cells that make up the mammalian brain ., Differences in brain size among the various mammalian species are attributed to differences in the abundance and proliferative capacity of a specific class of neural progenitors called basal progenitors ., Among these , a specific progenitor type called basal radial glia is thought to have played an important role during evolution in the expansion of the neocortex , the part of the brain associated with higher cognitive functions like conscious thought and language ., In the neocortex , the expression of the transcription factor Pax6 in basal progenitors is low in rodents , but high in primates , including humans ., In this study , we aimed to mimic the elevated expression pattern of Pax6 seen in humans in basal progenitors of the embryonic mouse neocortex ., To this end , we generated a novel , transgenic mouse line that allows sustained expression of the Pax6 gene in basal progenitors ., This elevated expression resulted in an increase in the generation of basal radial glia , in the proliferative capacity of basal progenitors , and , ultimately , in the number of neurons produced ., Our findings demonstrate that altering the expression of a single transcription factor from a mouse to a human-like pattern suffices to induce a primate-like proliferative behaviour in neural progenitors , which is thought to underlie the evolutionary expansion of the neocortex . | null | Humanizing the expression of the transcription factor Pax6 in cortical progenitors in the developing mouse brain is sufficient to endow these progenitors with a primate-like proliferative capacity. |
journal.pgen.1003763 | 2,013 | Dynamics of DNA Methylation in Recent Human and Great Ape Evolution | The genomic era is characterized by different comparative approaches to understand the effect of genomic changes upon phenotypes ., In the context of human evolution , the genomes of all species of great apes have now been sequenced 1–4 allowing nucleotide resolution comparisons to understand the evolution of our genome ., However , in contrast to these advances in comparative genomic analyses , there has been relatively little progress in the understanding of the evolution of genome regulation 5–9 ., DNA methylation is an important epigenetic modification found in many taxa ., In mammals , it is involved in numerous biological processes such as cell differentiation , X-chromosome inactivation , genomic imprinting and susceptibility to complex diseases 10–13 ., Promoter hypermethylation is generally thought to act as a durable silencing mechanism 14 ., However , the exact relationship between DNA methylation and gene expression is not clear since recent studies have also linked gene body methylation with transcriptional activity and alternative splicing 15–17 ., At some loci DNA methylation patterns are influenced by the underlying genotype 18–20 ., However , due to the fact that patterns of DNA methylation can change during development 16 , 21 , 22 or as a result of environmental factors 23 , 24 , the exact mechanisms governing DNA methylation states remain unclear ., Most efforts to understand DNA methylation changes in primates have focused on the comparison of human with chimpanzee or macaque 6 , 7 , 9 , 25 ., This is largely attributable to the difficulty of obtaining samples from endangered species and the lack of genome sequence for the great apes ., The publication last year of draft sequences of the gorilla 2 and bonobo 3 genomes facilitates a more accurate characterization of the species-specific events in all the great ape phylogeny , and interrogation of this epigenetic modification from an evolutionary point of view ., Studies to date have found that DNA methylation profiles are , in general , more similar between homologous tissues than between different tissues of the same species 9 ., However , differentially expressed genes between human and chimpanzee are often associated with promoter methylation differences , regardless of tissue type , establishing that some differences in the expression rates of genes between the species are associated with differences in DNA methylation ., It is estimated that around 12–18% ( depending on the tissue ) of interspecies differences in gene expression levels could be explained by changes in promoter methylation 9 ., Here we present the first comparative analysis of DNA methylation patterns between humans and all great ape species , allowing us to recapitulate the evolution of CpG methylation over the last 15 million years in these species ., We used Illumina Methylation450 BeadChips to profile DNA methylation genome-wide in blood-derived DNA from a total of 9 humans and 23 wild-born individuals of different species and sub-species of chimpanzee , bonobo , gorilla and orangutan ., We observed that the methylation values recapitulate the known phylogenetic relationships of the species , and we were able to characterize methylation differences that have occurred exclusively in the human lineage and among different great apes species ., We also identified a significant positive relationship between the rate of coding variation and alterations of methylation at the promoter level , indicative of co-occurrence between evolution of protein sequence and gene regulation, We obtained cytosine methylation profiles of peripheral blood DNA isolated from a set of males and females of nine humans , five chimpanzees , six bonobos , six gorillas and six orangutans ( Table S1 ) using the Illumina HumanMethylation450 DNA Analysis BeadChip assay ., Because the probes on the array are designed using the human reference genome , we performed a set of strict filters to remove divergent probes that could bias our methylation measurements ., The filtering was based on the number and location of mismatches with their target site in each species genome assembly tested 1–4 ( Figure S1 and Figure S2 , see Methods ) ., This resulted in the retention of 326 , 535 probes ( 72% ) in chimpanzee , 328 , 501 probes ( 73% ) in bonobo , 274 , 084 probes ( 61% ) in gorilla and 197 , 489 probes ( 44% ) in orangutan , consistent with their evolutionary distance to human ., We also applied a second filtering step to remove probes that overlapped with intra-species common variation ( see Methods ) 26 ., Cell heterogeneity may also act as a confounder when measuring DNA methylation , particularly from whole blood 27 ., Due to the difficulty of obtaining fresh blood samples from wild-born great apes , we were unable to either isolate a specific blood cell type or measure the cellular composition of the blood samples from which our DNA was extracted ., To minimize false positives resulting from different cellular compositions or other confounders , we performed two filtering steps ., First , we removed CpG sites that showed differential methylation in human between whole blood and each of the two most abundant subtypes of blood cell ( CD4+ T-cells and CD16+ neutrophils , see Methods ) ., Second , we required a minimum threshold of at least 10% change in mean methylation ( mean β-value difference ≥0 . 1 ) at each CpG in order to define differential methylation between species ., As a result of this threshold , differences in other cell types that account for <10% of the cellular composition of blood , are unlikely to affect our results ( see Methods ) ., In this work we used two different datasets:, i ) we confined our analysis to 114 , 739 autosomal probes and 3 , 680 probes on the X chromosome that were directly comparable across all the species to facilitate an unbiased comparison of human and all great apes ( 32 individuals ) , and, ii ) we used 291 , 553 shared autosomal probes between humans and chimpanzees to compare these two species ., We performed separate analyses of autosomal and sex-linked probes to prevent confounding effects of X chromosome inactivation on DNA methylation between males and females 13 ., Unless specifically mentioned , all results presented below refer to analysis of autosomal probes only ., To investigate the global correspondence of DNA sequence differences between species and the degree of methylation changes , we examined the Enredo-Pecan-Orthus ( EPO ) whole-genome multiple alignments of human , chimpanzee , gorilla , and orangutan Ensemble Compara . 6_primates_EPO 28 , 29 and we calculated pairwise distances between these four species ., Upon comparison of these sequence distances and methylation data ( see Methods ) , we observed a high global correlation between sequence substitution and methylation divergence ( R2\u200a=\u200a0 . 98 , p\u200a=\u200a0 . 0003 ) ( Figure 1A ) ., We then constructed a neighbor-joining phylogenetic tree based on the methylation levels of the 114 , 739 autosomal CpGs measured in all individuals and species ( Figure S3 ) ., This tree accurately recapitulates the known evolutionary relationships of great apes , including the separation at sub-species level of the Pan , Gorilla and Pongo genera ., These results are also maintained when using only the subset of probes that have a perfect match ( n\u200a=\u200a31 , 853 ) to each of the primate reference genome and contain no common polymorphisms suggesting that that methylation levels are associated with the evolutionary history of these species ( Figure 1B ) ., Due to the relatively recent origin of all partitions within genera of great apes 2–4 and our sample size , we focused our analysis on changes at the genus taxonomic level ., To identify only those methylation differences that represent fixed changes between these groups and to avoid possible artifacts due to intraspecific polymorphism , we retained only those CpGs with low methylation variance within each genus ( intragenus standard deviation <0 . 1 ) ., This filtering step resulted in the removal of 1 , 377 CpGs in human , 5 , 224 in the Pan sp ., , 5 , 289 in Gorilla sp ., and 5 , 740 in Pongo sp ., , with the resulting final set being 99 , 919 CpGs shared across all five species , covering 12 , 593 genes ( ≥2 probes within a 1 kb interval and overlapped with RefSeq genes , −1500 bp transcription start site ( TSS ) to 3′UTR ) ., The proportion of sites removed in this step are consistent with the relative population diversity within each of these species 2–4 , 30 ., Approximately 22% of the sites tested ( n\u200a=\u200a21 , 884 CpGs ) showed no significant changes among any of the species ( conserved sites: Wilcoxon rank-sum test , FDR-adjusted p>0 . 05 and mean β-value difference all cases <0 . 1 ) ., Comparison of genes linked with these sites showed an enrichment of Gene Ontology ( GO ) categories for fundamental cellular processes ., In contrast , we identified 2 , 284 human-specific ( 2 . 3% ) differentially methylated CpGs , 1 , 245 specific to Pan species ( 1 . 2% ) , 1 , 374 specific to Gorilla species ( 1 . 4% ) ., and 5 , 501 changes specific to Pongo species ( 5 . 5% ) ( Wilcoxon rank-sum test , FDR-adjusted p<0 . 05 and mean β-value difference ≥0 . 1 , see Methods ) ( Figure 2 and Table S2 ) ., We clustered these sites into regions with at least two nearby differentially methylated CpGs ( <1 kb interval ) and overlapped with RefSeq genes ( −1500 bp from TSS to 3′UTR ) ., Doing this , we identified 171 genes that show human specific methylation patterns , 101 genes in Pan species , 101 genes in Gorilla species and 445 genes in Pongo species ( Table S3 ) ., We observed that this spatial aggregation of differentially methylated sites is significantly non-random ( random permutation compared to all 99 , 919 CpGs used in our analysis , p<0 . 0001 , see Methods ) and a simple Likelihood Ratio Test also suggested a non-homogenous rate of methylation changes in the human and great ape evolution ( LRT , p<10−5 , see Text S1 ) ., Using the Genomic Regions Enrichment of Annotation Tool ( GREAT ) 31 ( see Methods ) we identified significant enrichments ( FDR-corrected p<0 . 05 ) for several biological processes associated with lineage-specific differentially methylated genes ., Within the human-specific differentially methylated regions most of the categories found were related with the circulatory system , as expected from testing blood-derived DNA ., However , we also found enrichment for terms related to development and neurological functions , including semicircular canal formation and facial nucleus development ( Table S4 ) ., The use of disease ontology terms showed that mutations in several of these genes are known to be associated with diseases including Möbius syndrome , Aspergers syndrome and malignant hyperthermia ., In the Pan genus ( chimpanzee and bonobo ) we observed significant enrichments among genes involved in epithelial development and the respiratory system , while in Pongo species ( orangutan ) enriched categories included a variety of basic metabolic and reproductive processes ( Table S4 ) ., We found a particular set of genes with methylation changes specifically in the human lineage including examples such as ARTN , COL2A1 and PGAM2 ( Figure 3 ) ., ARTN is a neurotrophic factor which supports the survival of sympathetic peripheral neurons and dopaminergic neurons ., COL2A1 encodes the alpha-1 chain of type II collagen , which is found primarily in the cartilage , the inner ear and the vitreous humor of the eye ., Mutations in this gene are associated with several developmental syndromes 32 ., PGAM2 is an enzyme involved in the glycolytic pathway , mutations in which are associated with glycogen storage disease MIM: 261670 , a defect that causes muscle cramping , myoglobinuria and intolerance for strenuous exercise ., In addition to the identification of regions showing changes in a single species , we also detected loci with more complex changes in methylation profiles among great apes ., One example is the promoter region associated with different isoforms of the GABBR1 gene ( Figure 3D ) ., This gene encodes the GABAB receptor 1 , a G protein-coupled receptor involved in synaptic inhibition , hippocampal long-term potentiation , slow wave sleep , muscle relaxation and sensitivity to pain ., While human and gorilla have GABBR1 promoter methylation patterns that are broadly similar to each other , orangutan shows relative hypomethylation across this region ., In contrast chimpanzee and bonobo show increased methylation specifically at the TSS of long GABBR1 isoforms , and intermediate methylation levels associated with the short isoform ., These data suggest some epigenetic differences among primates are associated with isoform regulation ., We observed a highly non-random distribution of the differentially methylated CpGs ( Figure 4A and 4B ) in relationship to gene annotations and CpG density ., From the functional distribution standpoint , there was a significant excess of changes ( p<0 . 0001 , permutation test , see Methods ) for sites located within 1 , 500 bp upstream of gene TSSs , gene bodies and intergenic regions , and from the CpG content standpoint , differential methylation occurred preferentially in CpG shores ( ±2 kb CpG island ) and non-CpG island regions ., These results highlight CpG shores as epigenetically variable regions , as it has been observed in human development and disease 12 , 33 ., In contrast , the regions immediately surrounding gene TSSs ( −200 bp of the TSS and 1st exon ) and CpG islands showed relative conservation of methylation ., We also observed a significant difference in the distribution of methylation levels at differentially methylated sites compared to the rest of the genome ( Figure 4C ) ., While the overall genome-wide pattern of methylation levels shows a strongly bi-modal distribution , with most sites having either very high or very low methylation levels , sites of evolutionary change have a significantly different distribution to genome wide distribution ( p\u200a=\u200a2 . 2×10−16 Kolmogorov-Smirnov test ) , showing intermediate methylation levels , which has been shown to be a hallmark of distal regulatory elements ., 34 ., In female mammals , X chromosome inactivation ( XCI ) is maintained via a number of epigenetic marks , including altered DNA methylation 35 , 36 ., Consistent with a role in XCI , the majority of sites we identified on the X chromosome in great apes showed relatively higher methylation levels in females versus males due to the contribution from the inactive X chromosome ( 63% , p\u200a=\u200a0 . 005 , Figure S4A ) ., We searched for CpG sites on the X chromosome presenting no significant changes between males and females in a specific lineage ( mean β-value difference <0 . 1 ) but showing significant gender differences in all the other species ( see Methods ) ., This analysis identified 22 CpGs in human , 59 in chimpanzee and bonobo , 72 in gorilla and 41 in orangutan ( Table S5 ) ., Some regions are particularly interesting such as the MID1 gene which has been previously reported as a gene subject to X-inactivation in humans but not in mouse 37 ., Our results suggest that this gene may escape XCI in the Pan lineage , but not in all other great apes ., Similarly the HTR2C gene shows multiple probes upstream of the TSS with similar patterns of methylation in both male and female humans , potentially suggesting that this gene escapes XCI in the human lineage ., In contrast , the same sites show significantly higher methylation levels in females compared to males in all other primate species , suggesting that in these species HTR2C may be subject to XCI ( Figure S4B ) ., Using published RNAseq data 38 , we did not observe a female-specific increase in HTR2C gene expression for in humans , although we note that many genes escaping XCI show no clear sex differences in expression levels 39 ., To maximize the identification of altered methylation patterns between human and our closest living relative , the chimpanzee , we performed a pairwise comparison of these two species using a larger dataset of 289 , 007 filtered probes common to human and chimpanzee ., We used the chimpanzee species and not the whole genera to make use of the better annotation in the genome reference assembly for this species compared to the rest of non-human primate genome reference assemblies 1 ., We identified 16 , 365 sites that showed significant hypermethylation in human , and 9 , 693 sites showing significant hypomethylation ( FDR-adjusted p<0 . 05 , β-value difference ≥0 . 1 ) ., This represents ∼9% of the total number of sites tested , and includes ∼2 , 500 genes ( ≥2 differentially methylated CpGs within a 1 kb interval and overlapped with RefSeq genes , −1500 bp TSS to 3′UTR ) ., Using this larger dataset , we then investigated the relationship between the evolution of protein coding sequences and epigenetic change at promoter level ., Using a curated set of 7 , 252 human∶chimpanzee 1∶1 orthologs 1 we identified 745 genes ( ∼10% of those tested ) that showed at least two differentially methylated sites at the promoter ( −1500 bp from the TSS to 1st exon , see Methods ) ., We then compared both the number of amino acid changes and the KA/KI ratio ( the number of coding base substitutions that result in amino acid changes as a fraction of the local intergenic/intronic substitution rate ) of these differentially methylated genes against the remainder 1 ( Figure 5 ) ., We observed a significant difference in both the number and rate of non-synonymous amino acid changes between genes with altered promoter methylation compared to those without significant methylation differences ( p<0 . 0001 , permutation test ) suggesting that rapid evolution at the protein coding level is frequently coupled with epigenetic changes in the promoter ., We also observed similar results when using only those probes with a perfect match to the chimpanzee reference genome ( Figure S5 ) ., An interesting example is the BRCA1 gene , which contains 32 amino acid changes between human and chimpanzee and has a KA/KI ratio of 0 . 69 ( three times the average of all orthologous genes ) ., This gene shows large differences in methylation ∼1–1 . 5 kb upstream of the TSS ( Figure 6 ) ., Previous studies have shown that methylation changes of this same region are associated with altered BRCA1 expression 40 ., In contrast , we also observed 184 genes that show perfect human:chimpanzee conservation at the amino acid level , yet they show significant epigenetic differences at their promoter ( Table S6 ) ., Within this set of genes , we observed significant enrichments for categories related with gene expression ( table S7 ) 41 , 42 ., As our survey of evolutionary changes in primate DNA methylation patterns utilized DNA derived from whole blood , we tested whether these changes are also present in other somatic tissues by comparing against an independent dataset ., A previous study 9 utilized a similar array platform , although with a much reduced probe density , to compare DNA methylation levels in humans and chimpanzees using DNA isolated from heart , liver and kidney ., Comparing the 457 sites common to both datasets that we defined as differentially methylated in blood samples versus these three other tissues , we observed a highly significant trend for methylation differences identified between human and chimpanzee to be conserved across all four tissue types ( Figure 7 ) ., The primary focus to date for understanding human evolution from a comparative genomic perspective has been the study of changes in DNA sequence and gene expression levels 43–45 ., Our study of DNA methylation profiles among human and great apes adds to this wealth of information , reinforcing the view that epigenetic changes contribute significantly to species divergence , and therefore they should be considered in studies of human evolution ., In this study , one of the main challenges was the technical limitation stemming from the use of arrays designed against the human genome to profile methylation patterns in great ape species with divergent genomes ., We utilized a set of filters to account for these differences , and were also able to replicate the results even after limiting our analysis to those probes with 100% identity in each of the non-human reference genome assemblies ., Supporting a biological role for our findings , we observed that the clustering of differential methylation within each species was highly non-random , and showed significant enrichments within functional genomic elements ., From a biological perspective , it is conceivable that differences in the constitutive fractions of whole blood between species might introduce a bias due to the fact that different cell types possess distinct epigenomes 27 ., This limitation is shared by nearly all comparative molecular studies of primary tissues from endangered species ( i . e . great apes ) due to the difficulty of obtaining relevant samples , especially in the case of wild-born individuals as the ones used in this study ., However , in order to minimize this problem we removed all CpG sites that vary significantly between whole blood and the most abundant cell populations in blood ., We further required a minimum threshold of 10% change in global methylation between sites in these species in order to identify differentially methylated sites , meaning that changes in the prevalence of minor cell fractions would not influence the results ., Finally , while all samples were obtained from adult individuals , we could not match the ages perfectly among all samples , so in order to compensate for this effect , and to minimize the effects of intraspecific polymorphism , we focused our study on sites with low intragenus variance ., Our results show that ∼9% of the CpGs we assayed showed significant methylation differences between human and chimpanzee , including the promoter regions of 745 genes ( 10% of those tested ) ., We estimate that over 2 , 500 genes present at least some methylation changes between human and chimpanzees ( ≥2 differentially methylated sites separated by ≤1 kb ) , a higher number than that known to be affected by copy number variation or under positive selection in the same species 46–49 ., Although the arrays we used do not provide a complete and unbiased coverage of the genome , these data suggest that epigenetic changes have been frequent during recent primate evolution and represent an important substrate for adaptive modification of genome function ., Underlining this idea , the changes we observed among primates are highly enriched for sites showing intermediate DNA methylation levels ., Previous studies have shown that such methylation values are often a hallmark of distal regulatory elements 34 , suggesting that many epigenetic changes occurring among human and great ape species impact transcriptional regulation ., Consistent with these findings , we detected global enrichments for epigenetic change within known regulatory regions , including distal regions upstream of gene transcription start sites and regions flanking CpG islands ( termed ‘CpG shores’ ) ., We observed that the great ape phylogeny can be recapitulated from methylation data alone ., Potential explanations for this are that methylation values could be driven by proximal DNA changes that were not controlled in this study , or that epigenetic changes have occurred independently of DNA sequence but are subject to similar rates of change either through selective pressures or neutral drift ., Interestingly , we also identified a significant positive relationship between the rate of coding variation within genes and alterations of promoter methylation , suggesting a co-occurrence between changes in protein sequence and gene regulation that may be related to expression changes in fast evolving genes 50 ., In contrast , and consistent with previous analysis indicating the importance of regulatory changes in evolution 51 , our study also identified scores of genes that are perfectly conserved at the amino acid level between human and chimpanzee , yet showing significant epigenetic change between these two species ., Furthermore , gene ontology analysis of this set showed that they are significantly enriched for the functional category of gene expression ., These observations highlight the evolutionary importance of epigenetic changes that affect gene regulation , and also demonstrate that sequence-based studies are insufficient to capture the full spectrum of evolutionary change ., Overall our analysis identified >800 genes with significantly altered methylation patterns specifically within each species of human and great apes , including 171 with a methylation pattern unique to humans ., Analysis of these 171 genes identified interesting enrichments for a number of functional categories that could suggest a relationship to human-specific traits ., For example , we observed that genes involved in the regulation of blood pressure and development of the semicircular canal of the inner ear among others , were all highly enriched for DNA methylation changes specifically in the human lineage ., While major changes in circulatory physiology are required for upright locomotion , the inner ear provides sensory input crucial for maintaining balance ., Furthermore , a previous study of primates and other mammals has shown that the size of the semicircular canals is correlated with locomotion and with relatively larger canals found in species that utilize fast or agile movement 52 ., While these trends hint at the potential importance of epigenetic changes in the evolution of several human-specific features , we caution that at this stage they should be considered as preliminary , as our studies were performed using DNA derived from whole blood , and it is well known that epigenetic patterns often vary widely between different tissues of an organism 27 ., Therefore further studies in physiologically relevant tissues will be required to confirm the significance of these findings ., However , we note that comparison with previously published data 9 suggests that many of the changes in DNA methylation that we detected between blood of human and chimpanzee appear to be conserved across several other tissues , suggesting that inter-specific differences observed in blood can in some cases be informative for other tissues ., Although sequencing studies have undoubtedly provided major advances in our understanding of primate evolution , our analysis of primate epigenomes unveils many novel differences among the great apes that are not apparent from purely sequence-based approaches ., Of particular note is the fact that we identify enrichments in multiple independent functional gene categories which suggests that regulatory changes may have played a key role in the acquisition of human-specific trait ., Therefore , epigenetic alterations likely represent an important facet of evolutionary change in primate genomes ., Future studies that integrate epigenetic data with recent detailed maps of functional elements , selective constraint and chromatin interactions in the human genome 53–55 will likely provide many novel insights into genomic and phenotypic evolution ., The non-human research has been approved by the ethical committee of the European Research Union ., No living animal has been used and DNA has been obtained during standard veterinary checks ., Methylation profiling of human subjects was approved by the Institutional Review Board of the Icahn School of Medicine at Mount Sinai ( HS#: 12-00567 HG ) ., We obtained methylation data from peripheral blood DNA extracted from 9 adult humans , 5 chimpanzees , 6 bonobos , 6 gorillas and 6 orangutans ., All individuals were unrelated adults and the non-human primates were all wild born ., DNA samples were bisulfite converted , whole-genome amplified , enzymatically fragmented , and hybridized to the Infinium HumanMethylation450 BeadChip which provides quantitative estimates of methylation levels at 482 , 421 CpG sites distributed genome-wide ., The assay was performed according to the manufacturers instructions ., The BeadChip array data discussed in this publication have been deposited in NCBIs Gene Expression Omnibus and are accessible through GEO Series accession number GSE41782 ., Due to the low density of probes targeting non-CpG dinucleotides ( <0 . 7% ) on the array , we focused our study on CpG sites ., Since the 50 bp probes on the array were designed against the human reference genome but we performed hybridizations utilizing DNA from different great ape species , we first mapped the probe sequences to the chimpanzee ( panTro3 ) , bonobo ( panPan1 ) , gorilla ( gorGor3 ) and orangutan ( ponAbe2 ) reference genomes using BWA 56 , allowing a maximum edit distance of 3 ., We then assessed probe performance as a function of the number and relative location of sequence differences at the probe binding site in each primate genome ( Figure S1 and Figure S2 ) ., Based on this analysis , in each species we only retained those probes that had either a perfect match , or had 1 or 2 mismatches in the first 45 bp but no mismatches in the 3′ 5 bp closest to the CpG site being assayed ., We also removed all probes that contained human SNPs with minor allele frequency ≥0 . 05 within the last 5 bp of their binding site closest to the CpG being assayed 57 ., Using published SNP data 26 for each species we removed probes containing SNPs with minor allele frequency ≥0 . 15 within the last 5 bp of their binding site closest to the CpG being assayed ., We also removed all probes that contained more than two SNPs with minor allele frequency ≥0 . 15 in the first 45 bp ., Methylation values for CpG sites in each sample were obtained as β-values , calculated as the ratio of the methylated signal intensity to the sum of both methylated and unmethylated signals after background subtraction ( β-values range from 0 to 1 , corresponding to completely unmethylated and fully methylated sites , respectively ) ., Within each individual , probes with a detection p>0 . 01 were excluded ., We performed a two color channel signal adjustment and quantile normalization on the pooled signals from both channels and recalculation of average β-values as implemented in “lumi” package of R 58 ., The Illumina Infinium HumanMethylation450 BeadChip contains two assay types ( Infinium type I and type II probes ) which utilize different probe designs ., As the data produced by these two assay types shows distinct profiles ( Figure S6 ) , to correct this problem we performed a BMIQ ( beta mixture quantile method ) 59 on the quantile normalized data sets ., Using a published human data set 27 we identified differentially methylated sites between whole blood and CD4+ T-cells , and between whole blood and CD16+ neutrophils , representing the two most abundant cell fractions of blood ( comprising ∼13% and ∼65% , respectively ) ( Wilcoxon rank-sum test , FDR-adjusted p<0 . 05 and mean β-value difference in each case ≥0 . 1 ) ., These sites ( n\u200a=\u200a10 , 151 ) were removed to mitigate potential confounders due to differing proportions of blood cell types among primates , leaving for comparison only those sites that do not significantly vary among the most abundant cell types of blood ., β-values can be interpreted as the percentage of methylation at a given site ., A β-value of 0 . 1 indicates that there has been a change in methylation in 10% of the molecules tested ., Because our analyses required a mean β-value difference >0 . 1 to achieve significance , this threshold means that changes in blood cell fractions representing <10% of whole blood will be unlikely to affect our results ., The final dataset after all filtering steps comprised 114 , 739 probes shared across all great ape species , and 291 , 553 probes shared between human and chimpanzee ., To investigate the global correspondence of DNA sequence differences betwe | Introduction, Results, Discussion, Methods | DNA methylation is an epigenetic modification involved in regulatory processes such as cell differentiation during development , X-chromosome inactivation , genomic imprinting and susceptibility to complex disease ., However , the dynamics of DNA methylation changes between humans and their closest relatives are still poorly understood ., We performed a comparative analysis of CpG methylation patterns between 9 humans and 23 primate samples including all species of great apes ( chimpanzee , bonobo , gorilla and orangutan ) using Illumina Methylation450 bead arrays ., Our analysis identified ∼800 genes with significantly altered methylation patterns among the great apes , including ∼170 genes with a methylation pattern unique to human ., Some of these are known to be involved in developmental and neurological features , suggesting that epigenetic changes have been frequent during recent human and primate evolution ., We identified a significant positive relationship between the rate of coding variation and alterations of methylation at the promoter level , indicative of co-occurrence between evolution of protein sequence and gene regulation ., In contrast , and supporting the idea that many phenotypic differences between humans and great apes are not due to amino acid differences , our analysis also identified 184 genes that are perfectly conserved at protein level between human and chimpanzee , yet show significant epigenetic differences between these two species ., We conclude that epigenetic alterations are an important force during primate evolution and have been under-explored in evolutionary comparative genomics . | Differences in protein coding sequences between humans and their closest relatives are too small to account for their phenotypic differences ., It has been hypothesized that these differences may be explained by alterations of gene regulation rather than primary genome sequence ., DNA methylation is an important epigenetic modification that is involved in many biological processes , but from an evolutionary point of view this modification is still poorly understood ., To this end , we performed a comparative analysis of CpG methylation patterns between humans and great apes ., Using this approach , we were able to study the dynamics of DNA methylation in recent primate evolution and to identify regions showing species-specific methylation pattern among humans and great apes ., We find that genes with alterations of promoter methylation tend to show increased rates of divergence in their protein sequence , and in contrast we also identify many genes with regulatory changes between human and chimpanzee that have perfectly conserved protein sequence ., Our study provides the first global view of evolutionary epigenetic changes that have occurred in the genomes of all species of great apes . | null | null |
journal.pcbi.1005277 | 2,017 | Morphological Transformation and Force Generation of Active Cytoskeletal Networks | The actin cytoskeleton plays an important role in various cellular processes , such as changes in cell shape , cytokinesis , and cell migration 1 ., Much of the mechanical forces required for these processes are generated by interactions between actin filaments ( F-actin ) and myosin II motors 2 ., Actomyosin contractility regulates structural organization of the actin cytoskeleton and its rheological properties by interacting and competing with the dynamics of actin cross-linking proteins ( ACPs ) and actin filaments ., For example , during Dictyostelium furrow ingression , interactions between myosin and ACP dynamics control cytokinesis contractility dynamics and mechanics 3 ., In addition , during fission yeast cytokinetic ring assembly , an increase in ACP density prevents clump formation 4 , 5 ., Representative cytoskeletal structures that are regulated by actomyosin contractility are various types of bundles , such as stress fibers , random polarity bundles , cytokinetic rings , and transverse arcs 6 ., Despite similarity in their structural organization , these bundles are formed via very distinct mechanisms ., Dorsal stress fibers are assembled via formin-driven polymerization of actin filaments occurring outside adhesion sites ., Transverse arcs , that are located at the interface between lamellipodia and lamella , form via actomyosin-driven condensation of actin filaments within the lamellipodia 7 ., During the condensation , actin filaments whose barbed ends are initially biased toward the cell margin are reoriented and thus become parallel to the margin ., Transverse arcs move away from the cell margin and eventually coalesce with dorsal stress fibers , to transmit contractile forces to surrounding environments , without direct attachment to focal adhesions 8 ., Several aspects regarding structural reorganization of a network into a bundle have been investigated in previous numerical studies ., It was shown that an increase in myosin density induces a structural transition from networks into bundles through a series of hierarchical steps 9 with enhancement of forces generated by the actomyosin structures 10 ., In addition , a recent study demonstrated that an increase in ACP density above a threshold value leads to a switch-like transition from random networks to ordered , bundled structures 11 ., However , owing to the highly simplified models and limited scopes of the previous studies , it still remains inconclusive how a network is transformed into a bundle , how force is generated , and what happens on actin filaments during the structural reorganization ., Several biophysical factors are likely to impact network transformation into a bundle ., For example , an extent to which actin filaments are cross-linked will play an important role ., If filaments are loosely cross-linked , they may be reoriented relatively easily to form a bundle , but low network connectivity could be antagonistic to the stability of formed bundles and generated forces ., By contrast , if actin filaments are heavily cross-linked , they may not easily rotate without significant deformation ., Because of the low bending rigidity of actin filaments , myosin motor activity could result in buckling during reorientation and compaction of cross-linked actin filaments ., As suggested by a previous theoretical study 12 , filament buckling may play a critical role in either force generation or bundle formation or in both ., In addition , fast turnover of actin filaments occurring via diverse actin binding proteins within cells has potential to modulate the morphological transformation and force generation ., Using only experiments , it is challenging to accurately evaluate relative importance of each of these factors and isolate their effects ., In this work , using an agent-based computational model , we systematically investigated morphological transformation of an actomyosin network into a bundle and force generation during the transformation ., We investigated effects of diverse biophysical parameters on network compaction into a bundle , which were not systematically studied in previous computational works ., Specifically , we focused on the impacts of the densities of ACPs and motors and of the rigidity , initial orientation , and turnover of actin ., Results from the study were discussed in the context of the assembly of transverse arcs observed in migrating cells 7 ., This study provides new insights into mechanistic understanding of a role of the interplay between various biophysical factors in bundle formation and force generation ., We employed our previous coarse-grained Brownian dynamics model for actomyosin structures 13 ., In the model , actin filaments , actin cross-linking proteins ( ACPs ) , and motors are simplified into interconnected cylindrical segments ( Fig 1A ) ., Actin filaments consist of serially-connected cylindrical segments with polarity ( barbed and pointed ends ) ., ACPs are composed of a pair of cylindrical segments ., Each motor has a backbone structure with 8 arms , each of which represents 8 myosin heads ., Displacement of the segments is governed by the Langevin equation ., Harmonic potentials with bending ( κb ) and extensional stiffnesses ( κs ) maintain equilibrium angles and lengths , respectively , formed by the segments ., Repulsive forces account for volume-exclusion effects between actin filaments ., Stochastic forces satisfying the fluctuation-dissipation theorem are applied to induce thermal fluctuation 14 ., Positions of the segments are updated at each time step using the Euler integration scheme ., ACPs bind to actin filaments at a constant rate and also unbind from actin filaments in a force-dependent manner following Bell’s equation 15 ., A motor arm binds to an actin filament and walks toward the barbed end of the actin filament , generating tensile forces ., Actin undergoes nucleation , polymerization , and depolymerization , staying in either monomeric or filamentous state ., We simulate treadmilling of actin filaments by imposing equal polymerization and depolymerization rates at barbed and pointed ends , respectively ., To alter the treadmilling rate without a large change in average length of actin filaments , a nucleation rate is dynamically adjusted ., Monomeric actin and free ACP and motor that are not bound to any actin filament are considered implicitly by their local concentrations ., Self-assembly of actins , ACPs , and motors in a 3D rectangular computational domain ( 4×8×0 . 5 μm ) results in a homogenous actomyosin network ( Fig 1B ) ., A periodic boundary condition is imposed in the y-direction , whereas boundaries in the x- and z-directions exert repulsive forces on the segments to keep them within the domain ., After network assembly , walking of motors on actin filaments is initiated , facilitating transformation of the network to a bundle ., We measured a macroscopic force generated by a bundle and also microscopic forces acting on ACPs and motors ., Definitions of terms are listed in S1 Table , and detailed values of parameters are listed in S2 Table ., Consistent with previous theoretical and experimental studies 16–18 , densities of ACPs ( RACP ) and motors ( RM ) critically affect bundle formation and tension generation ., With RM = 0 . 08 and RACP = 0 . 01 , a homogeneous network compacted into a bundle spanning the computational domain in the y-direction within ~10 s ( Fig 2A ) ., However , the bundle was heterogeneous at 10 s in terms of actin concentration , showing a few regions with higher actin density ., In addition , the bundle was highly unstable , resulting in a few separate aggregates over time ., Tension measured in the bundle increased up to ~0 . 8 nN and then decreased to nearly zero ( Fig 2C ) ., By contrast , with RM = 0 . 08 and RACP = 0 . 1 , a more compact , uniform bundle was formed within 15 s , and the bundle remained intact for the duration of the simulation ( Fig 2B ) ., Tension increased up to ~4 nN , and then decreased slowly ., Microscopic forces exerted on each motor ( fMmax ) and ACP ( fACPmax ) measured at maximum tension can explain the magnitude and sustainability of the generated tension ( Fig 2D ) ., Note that fMmax and fACPmax are positive when they are exerted toward barbed ends of actin filaments ., With a large number of ACPs ( RM = 0 . 08 and RACP = 0 . 1 ) , fMmax was higher , and fACPmax was smaller ., If there are many ACPs , they share loads exerted by motors , leading to smaller force on each ACP ., Since ACPs are assumed to exhibit slip-bond behavior , the smaller force on ACPs leads to less frequent unbinding events of ACPs ., Thus , stable ACPs can help motors to generate higher force close to their stall force and support the force for a longer time ., By contrast , with fewer ACPs ( RM = 0 . 08 and RACP = 0 . 01 ) , most motors failed to attain their stall force , and each ACP supported a larger force , leading to instability of the bundle and reduction in generated tension ( Fig 2D ) ., We systematically varied RACP and RM to probe their effects on bundle formation and tension generation ., Maximum tension was positively correlated with both densities ( Fig 2E ) , whereas sustainability was proportional to RACP but inversely proportional to RM ( Fig 2F ) ., We measured time evolution of standard deviation of x positions of actins ( σx ) to quantify compaction of networks ( S1 Fig ) ., σx tends to initially decrease , indicating compaction of networks ., After reaching its minimum value , σx remained constant in most cases ., However , in some cases , σx increased over time , which may indicate disintegration of a bundle into aggregates ., Indeed , the increase in σx occurred in cases with higher RM and lower RACP where tension is not sustained well , and bundles are likely to form aggregates ., In cases with very low RM , σx continuously decreased , indicating very slow compaction of networks ., To quantify how fast networks compact , we defined compaction time as time at which the rate of change in σx over time becomes larger than 0 . 01 × ( the average rate of change in σx during first 5s ) ., The compaction time was shorter at higher RM and lower RACP ( Fig 2G ) ., We used the standard deviation at compaction time ( σxc ) as an indicator of how tightly a network is compacted in the x-direciton ( Fig 2H ) ., A tighter bundle was formed with higher RM and RACP ., A sufficient amount of ACPs can tighten bundles by helping force generation of motors and increasing connectivity of bundles ., However , ACPs slow down formation of bundles because a network becomes much more stiffer with more ACPs ., In sum , a network with more motors compacted faster into a tighter bundle exerting larger tension because there are more force generators ., However , the bundle and the tension are likely to be unstable , leading to bundle disintegration into aggregates and significant tension relaxation ., A network with more ACPs compacted more slowly into a tighter bundle generating larger and more sustained tension ., In our previous studies , it was shown that buckling of actin filaments is necessary for contraction of a network and for force generation in a preformed bundle 16 , 19 ., We quantified buckling events occurring in the simulations shown in Fig 2E–2H , by tracking the ratio of end-to-end distance to contour length of actin filaments ., Since most actin filaments have multiple , transiently bound motors and ACPs , buckling takes place in various ways; some of the actin filaments experienced subsequent buckling events at multiple locations over time , and buckled filaments , at times , became straight again ( S2 Fig ) ., We determined the number of actin filaments that underwent buckling at least once in each simulation by assuming that actin filaments with a ratio of end-to-end distance to contour length smaller than 0 . 6 are buckled ., We found that buckling occurred less frequently with higher RACP because the critical force above which buckling occurs becomes larger with higher RACP ( Fig 3A ) ; this is associated with a decrease in distance between adjacent cross-linking points on an actin filament ., Although motors generate larger forces with higher RACP ( Fig 2D ) , the increase in the critical force required for buckling is greater , leading to less frequent buckling events ., With higher RM , buckling took place more frequently since more motors generate larger contractile forces that can induce buckling ., These buckling events mostly occurred during the transformation to a bundle before tension reached its maximum , rather than after the peak tension ( Fig 3D ) ., We tested whether buckling is required for the transformation of a network into a bundle by suppressing the filament buckling via a 100-fold increase in the bending stiffness of actin filaments ( κb , A = 100×κb , A* ) , where κb , A* is the reference bending stiffness ., At both high and low levels of RACP , a bundle rarely formed although some of the actin filaments formed a pseudo bundle at the center ( Fig 3B and 3C ) ., At RM = 0 . 08 and RACP = 0 . 1 , the developed tension in a network with 100×κb , A* was much smaller than that in a network with κb , A* , and buckling rarely occurred ( Fig 3D ) ., Smaller tension for the case with 100×κb , A* can be attributed to low values of fMmax; although some values reached stall force , there was a general tendency for the forces to be smaller overall than those in the case with κb , A* ( Fig 3E ) ., Negative values of fACPmax were also slightly smaller in magnitude for the case with 100×κb , A* since ACPs sustain lower positive fMmax in this case ., Note that negative or positive fACPmax sustain positive or negative fMmax , respectively ., Positive fACPmax showed higher value for the case with 100×κb , A* , since this case exhibits a significant amount of negative fMmax while the case with κb , A* does not ., Due to the catch-bond nature of motors , the lower positive fMmax makes motors stay for a shorter time on actin filaments , which corresponds to a lower duty ratio of motors ., Then , motors are less able to stably generate a large amount of forces ., Suppression of bundle formation and generation of lower tension observed in Fig 3B–3D might originate largely from a decrease in the duty ratio rather than an increase in κb , A ., To confirm the importance of κb , A , we ran a simulation using motors with a much higher unbinding rate ( i . e . lower duty ratio ) than the motors used in the case shown in Fig 2B where a stable bundle was formed ., We varied one of the mechanochemical rates in the parallel cluster model 20 , 21 , which leads to a decrease in the stall force from 5 . 7 pN to 5 . 3 pN and an increase in the unbinding rate from 0 . 049 s-1 to 0 . 49 s-1 ., As shown in S3 Fig , a bundle still formed well , and tension inside the bundle and sustainability were similar to those of the reference case shown in Fig 2B and 2C ., Thus , the inhibition of bundle formation and the decrease in tension result mostly from the change in the κb , A , not the change in the duty ratio of motors ., Maximum tension measured under various values of RM and RACP with 100×κb , A* ( Fig 3F ) was much lower than that measured with κb , A* ( Fig 2E ) ., Dependences of sustainability and compaction time on RM and RACP ( Fig 3G and 3H ) were similar to those in the cases with κb , A* ( Fig 2F and 2G ) ., We also measured time evolution of σx for quantification of network compaction ( S4 Fig ) ., Interestingly , in cases with lower RACP and higher RM , σx increased beyond its initial value after reaching the minimum ., σxc was overall higher in the cases with 100×κb , A* ( Fig 3I ) than that in the cases with κb , A* ( Fig 2H ) , quantitatively showing suppression of bundle formation with stiffer actin filaments ., Interestingly , with more ACPs , σxc was larger , which is opposite to the observation in Fig 2H ., As shown in Fig 3A , buckling occurred less frequently at higher RACP even with κb , A* ., However , since a fraction of actin filaments were still buckled , the number of buckled actin filaments is not a critical factor determining the extent of network compaction ., By contrast , with 100×κb , A* , most of actin filaments cannot be buckled due to a significant increase in the critical buckling force ., Then , network compaction becomes very sensitive to the number of buckled actin filaments because buckling is necessary for network compaction , resulting in less network compaction with higher RACP ., In sum , these results demonstrate that even with a sufficient number of ACPs that sustain tension and help motors reach their stall force , buckling of actin filaments is required for formation of tight bundles and generation of large tension ., Myosin II motors compact actin filaments in lamellipodia into transverse arcs that generate contractile forces 22 ., Since the barbed ends of all actin filaments in lamellipodia are directed toward the cell margin , the lamellipodia is not an isotropic actin network ., We probed the effects of anisotropic initial orientations of actin filaments on bundle formation and tension generation with RM = 0 . 08 and RACP = 0 . 01 by creating three networks consisting of actin filaments with biased initial orientations ( Fig 4A–4C ) ., Note that the case shown in Fig 4B where actin filaments are initially oriented toward the +x direction mimics filament orientation in lamellipodia ., Compared to the reference case with isotropic orientation of filaments ( Fig 2A and 2C ) , the networks with biased orientations showed lower maximum tension and slower bundle formation ( Fig 4A–4D ) because there were a smaller number of antiparallel pairs of actin filaments that are in configuration suitable for motors to produce force ( Fig 4F ) ., Interestingly , a network with barbed ends directed toward +y was effectively transformed to a bundle with significant tension despite the fact that it initially had no antiparallel pairs of actin filaments in the y-direction ., We found that some of the actin filaments changed their orientations ( S5 Fig and Fig 4C , right column ) during network contraction ( Fig 4F ) ., Even in the network with barbed ends oriented toward +x/+y , a bundle could form slowly and generate tension due to changes in filament orientation ( Fig 4A , 4D and 4F ) ., In all cases , bundles eventually collapsed into a few aggregates; this occurred at a rate proportional to the maximum tension because larger tension accelerates destabilization of ACPs , leading to faster disintegration of bundles ., We also tested the influences of initial orientation of actin filaments ( diagonal or horizontal/vertical ) on bundle formation and tension generated in networks , and the results overall showed similar tendencies ( S6 and S7 Figs ) ., At higher ACP density ( RM = 0 . 08 and RACP = 0 . 1 ) , actin filaments tend to rotate less than those at lower RACP because the filaments are confined more by a larger number of ACPs ( S8A , S8B and S8C Fig ) ., However , some of the actin filaments were still able to change their orientations , contributing to tension generation ( S8D and S8E Fig ) ., Note that unlike the case with lower ACP density , the bundles were not disintegrated into aggregates , regardless of initial filament orientation ., This can explain a discrepancy between the unstable bundle shown in Fig 4B formed from a network mimicking the geometry of lamellipodia and a stable bundle observed at the interface between lamellipodia and lamella ., It is expected that actin filaments with numerous branching points in lamellipodia have very high connectivity between actin filaments , preventing a bundle from being disintegrated ., Taken together , these results demonstrate that networks with biased filament orientations can still be transformed to bundles owing to changes in filament orientation occurring during contraction ., However , if orientations are biased , bundles are loose , and generated tension tends to be lower but is sustained for a longer time ., We have observed that buckling is necessary for bundle formation in networks with isotropic filament orientation since contraction of antiparallel pairs of actin filaments requires buckling ., We tested whether buckling is still necessary for bundle formation in networks with a much smaller number of antiparallel pairs by increasing the bending stiffness of actin filaments 100-fold as before ( κb , A = 100×κb , A* ) ., We found that networks with barbed ends directed toward +x/+y or +y were still transformed to bundles because contraction in the y-direction does not need to occur in such configurations ( Fig 5A and 5C ) ., Filaments in the network with barbed ends directed toward +x/+y initially form only parallel pairs of actin filaments , so they can be aligned in the y-direction ( S9A Fig ) ., Filaments forming antiparallel pairs in the x-direction in the network with barbed ends directed toward +y can be aligned in the y-direction via polarity sorting due to the absence of a periodic boundary condition in the x-direction ( S9C Fig ) ., Some of the filaments changed their orientation during bundle formation , resulting in antiparallel pairs in the y-direction that were also connected to other actin filaments in a bundle ( Fig 5E ) ., Due to suppression of buckling , these pairs cannot contract , so the bundles remained curved rather than straight ., Accordingly , forces generated on bundles remained close to zero and even became compressive ( i . e . negative ) ( Fig 5D ) ., By contrast , a network with barbed ends directed toward +x/±y could not form a bundle since the antiparallel pairs of filaments that existed from the beginning were not able to contract ( Fig 5B and S9B Fig ) ., Tension generated in these networks was similar to that in networks with isotropic orientations ( Fig 5D ) ., Therefore , buckling is not always necessary for the transformation of a network to a bundle ., If orientation of actin filaments is highly anisotropic , the transformation can still take place via polarity sorting of filaments by motors ., However , tensile forces are not developed on the formed bundles ., In our previous study , we demonstrated that actin turnover modulates the buildup and sustainability of tension generated by actomyosin networks 13 ., We tested effects of actin turnover on bundle formation and tension generation by imposing actin treadmilling at various rates ( kt , A ) under a condition where bundles generate unsustainable tension and eventually form aggregates in the absence of any turnover ( RM = 0 . 08 and RACP = 0 . 01 ) ., We additionally assumed that depolymerization of actin filaments can be inhibited by bound ACPs or motors to a different extent 2 ., We defined the inhibition factor ( ξd , A ) to represent this effect; with ξd , A = 0 , depolymerization is not inhibited at all , whereas inhibition is complete with ξd , A = 1 ., In a control case without turnover ( kt , A = 0 ) and a case with kt , A = 60 s-1 and ξd , A = 1 , bundles became aggregates within 100 s ( Fig 6A and 6D ) , and generated tension fell to nearly zero ( Fig 6E ) ., With kt , A = 60 s-1 and ξd , A = 0 , some of the actin filaments in the network formed a thin bundle that was converted into aggregates over time ( Fig 6B ) , and tension ultimately relaxed to zero ( Fig 6E ) ., By contrast , with kt , A = 60 s-1 and ξd , A = 0 . 6 , the bundle was maintained much longer , showing highly sustainable tension ( Fig 6C and 6E ) ., We systematically probed the effects of kt , A and ξd , A on the maximum and sustainability of tension ( Fig 6F and 6G ) ., While maximum tension showed no correlation with kt , A and ξd , A , sustainability tended to be higher at intermediate levels of ξd , A because too large ξd , A completely inhibits actin turnover , whereas too small ξd , A precludes bundle formation and destabilizes the bundle by ACP unbinding induced by actin turnover ., The region with higher sustainability is wider with lower kt , A , since less turnover occurs at lower kt , A at the same level of ξd , A ., Networks compacted faster with more turnover ( i . e . higher kt , A and lower ξd , A ) , but formed bundles were loose ( Fig 6H and 6I ) ., This agrees with the observation that compaction occurred faster , and more loose bundles formed at lower RACP ( Fig 2G and 2H ) , because more frequent turnover facilitates unbinding of ACPs , leading to a decrease in the number of active ACPs bound on two actin filaments at dynamic equilibrium ., Also , with low ξd , A , σx increased after reaching its minimum ( S10 Fig ) , which corresponds to disintegration of a bundle into a network ., However , the increase in σx significantly slowed down after some time in several cases , which implies a steady state with coexistence of bundle and network structures as shown in Fig 6C ., At high RACP shown in S11 Fig ( RM = 0 . 08 and RACP = 0 . 1 ) , bundle formation and the maximum tension were both enhanced with slower actin turnover ( i . e . lower kt , A and higher ξd , A ) ., Compaction time , σxc , and σx showed similar trends with those in Fig 6 and S10 Fig ( S11 and S12 Figs ) ., In this case , the bundle and generated tension are already stable without turnover owing to numerous ACPs ., Actin turnover decreases the number of actin filaments involved with bundle formation as can be seen in a change in the diameter of bundles ( S11B , S11C and S11D Fig ) ., Thus , the connectivity of filaments in the bundle is deteriorated , resulting in less sustainable tension ., In addition , since turnover induces unbinding of ACPs which leads to instability , more motors failed to reach their stall force , leading to smaller maximum tension ( S11E Fig ) ., Indeed , fMmax was lower with increasing turnover ( S11F Fig ) ., fACPmax also decreased with increasing turnover , owing to lower tension and facilitated ACP unbinding by actin turnover ., Note that the case with ξd , A = 1 showed more sustained tension than the case without actin turnover ., With ξd , A = 1 , depolymerization occurs in regions of an actin filament which are not bound to ACPs or motors , thus unnecessary for tension generation ., Depolymerized actin can be polymerized at barbed ends of actin filaments , helping sustain tension by increasing a walking distance of motors toward a barbed end ., In sum , with an insufficient number of ACPs , actin turnover with intermediate values of ξd , A enhances the stability of bundles and generated tension , whereas with more ACPs , actin turnover plays only a negative role for the stability of bundles and tension ., Structural reorganization of a cross-linked actin network into a bundle occurs in several cellular phenomena , such as formation of transverse arcs at the interface between lamellipodia and lamella ., Recent experiments have shown that in the absence of stress fibers , cells can still exert large tensions on surrounding environments due to contractile lamella that contain transverse arcs , implying the significance of transverse arcs in cells as a force generator 23 ., To illuminate mechanisms of formation and force generation of transverse arcs , we here presented a computational study regarding transformation of actomyosin networks into bundles under diverse conditions ., Results from this study demonstrate that formation of contractile bundles and force generation in the bundles are tightly regulated by the interplay between concentrations of cytoskeletal elements and the deformability , dynamics , and initial orientation of actin filaments that have not been tested systematically in previous studies ., This study is significantly different from our previous study that employed actomyosin bundles preassembled by stacking straight actin filaments in parallel 16 since actin filaments are not stacked merely without any deformation during the morphological transformation ., We found that during the transition from a network into a bundle , actin filaments undergo buckling and reorientation in various ways , and a large portion of tension is built during the structural reorganization rather than after bundle formation ., In addition , we incorporated systematic variations of initial filament orientation that have not been included in our previous studies 13 , 16 , 24–27 , motivated by observation that transverse arcs located at the interface between lamellipodia and lamella are formed by compaction and realignment of actin filaments with biased orientations within the lamellipodia 28 ., We investigated how the density of ACPs and motors and the buckling of actin filaments govern the bundle formation and tension generation ., It was found that maximum bundle tension is proportional to motor and ACP densities , whereas sustainability of tension is proportional to ACP density but inversely proportional to motor density ., A key factor for determining tension sustainability is how much force is exerted on each ACP because large force can make ACPs unstable by increasing their force-dependent unbinding rate ., This is consistent with our previous studies where forces are generated by cortex-like actomyosin networks 19 and preformed bundles 16 ., We observed that time required for bundle formation is inversely proportional to motor density but proportional to ACP density ., Previous experimental studies showed that condensation of networks into transverse arcs occurs within 20 s 29 , which is comparable with the compaction time measured in this study ., We also observed that buckling of actin filaments plays an important role in bundle formation , and most of the tension is generated during a transition from a network to a bundle ., This is different from our previous study where we found the importance of filament buckling and force generation during contraction of the preformed bundles 16 ., In addition , using networks consisting of filaments with biased orientations , we found that buckling should take place in antiparallel pairs of actin filaments initially aligned in the y-direction in order to induce transformation of networks into bundles ., If there is not such an antiparallel pair in the y-direction , the transformation is possible without filament buckling ., However , development of large tension on a formed bundle is possible only when filament buckling is allowed ., In addition , we showed that networks with isotropic filament orientations result in the best bundle formation and the largest tension ., Interestingly , even if orientations of actin filaments are too biased to initially have antiparallel pairs of actin filaments , some of the actin filaments change their orientations during network contraction , resulting in antiparallel pairs and formation of bundles ., However , compared to the network with isotropic orientations of actin filaments , bundles are loosely formed , and tension is smaller ., Since the smaller tension leads to lower force on each ACP , tension is sustained for a longer time ., Also , we probed influences of actin turnover via treadmilling on bundle formation and tension generation as in our previous study ., However , we made a new assumption that actin depolymerization rate can be varied by cross-linking points based on previous experimental observations 30 ., We observed that actin turnover with moderate inhibition of actin depolymerization by motors and ACPs increases the sustainability of tension and confers structural stability to the bundles at low ACP density ., If there is a selective inhibition of depolymerization , the region of a filament that contributes least to the connectivity of bundles ( from a pointed end to the first cross-linking point ) is depolymerized faster ., Depolymerized actin can be polymerized at a barbed end of the same filament or other actin filaments ., Since motors walk toward barbed ends , the newly polymerized actin can enable motors to walk further ., By contrast , at high ACP density , actin turnover decreases tension sustainability and the stability of formed bundles because the connectivity of the bundles is already maximized by nu | Introduction, Results, Discussion, Methods | Cells assemble numerous types of actomyosin bundles that generate contractile forces for biological processes , such as cytokinesis and cell migration ., One example of contractile bundles is a transverse arc that forms via actomyosin-driven condensation of actin filaments in the lamellipodia of migrating cells and exerts significant forces on the surrounding environments ., Structural reorganization of a network into a bundle facilitated by actomyosin contractility is a physiologically relevant and biophysically interesting process ., Nevertheless , it remains elusive how actin filaments are reoriented , buckled , and bundled as well as undergo tension buildup during the structural reorganization ., In this study , using an agent-based computational model , we demonstrated how the interplay between the density of myosin motors and cross-linking proteins and the rigidity , initial orientation , and turnover of actin filaments regulates the morphological transformation of a cross-linked actomyosin network into a bundle and the buildup of tension occurring during the transformation . | Contractile networks and bundles generate mechanical forces required for various cellular processes , particularly cell division and migration ., In many of these processes , networks are structurally reorganized into bundles by the activity of molecular motors ., During this morphological transformation , filaments constituting networks are reoriented and undergo deformation and turnover , and large tensile forces are generated and sustained in bundles ., However , it remains inconclusive how the morphological transformation and force generation are regulated ., Here , using a rigorous computational model , we quantitatively demonstrated that the interplay between several factors determines the characteristics of generated tensile force and regulates the transformation from networks to bundles ., Thus , results in this study provide insights into the physical and mechanistic basis of the complex transition from networks to bundles observed in cells . | stiffness, cell motility, mechanical properties, actin filaments, classical mechanics, molecular motors, actin motors, materials science, damage mechanics, motor proteins, dynamic actin filaments, buckling, polymer chemistry, proteins, deformation, chemistry, physics, biochemistry, cell biology, biology and life sciences, physical sciences, material properties, depolymerization | null |
journal.pgen.1004851 | 2,014 | A Cbx8-Containing Polycomb Complex Facilitates the Transition to Gene Activation during ES Cell Differentiation | First identified in Drosophila , the polycomb group of proteins share conserved domains and play an important role in coordinated gene repression during vertebrate and invertebrate development 1 ., The prevailing view is that PRC2 and PRC1 act in a sequential manner ., The association of the three proteins Eed , Suz12 and Ezh2 or Ezh1 leads to the formation of the core PRC2 complex ., Ezh1 and Ezh2 are histone methyl transferases that mediate the addition of up to three methyl groups to lysine 27 of histone H3 ( H3K27me1–3 ) ., The trimethylated mark is recognized by PRC1 complexes that further mediate ubiquitination of H2A and gene repression 2–4 ., More recently , it has been shown that PRC1 can also be recruited to chromatin in the absence of a functional PRC2 complex 5–8 ., In contrast to PRC2 , the composition of PRC1 is highly modular and much more variable ., The ubiquitin ligase that provides the catalytic activity to the PRC1 complex can be either Ring1a or Ring1b ., The complex additionally includes one of six Pcgf proteins; one of three orthologs of polyhomeiotic and five mutually exclusive Cbx proteins can occupy the position of the Drosophila Polycomb protein ., Cbx proteins differ in some of their domains suggesting that they could convey different functional and regulatory properties to PRC1 9 ., In addition a variant complex in which RYBP replaces Cbx proteins has been shown to mediate repression independent of the methylation status of H3K27 7 ., Mouse embryonic stem ( ES ) cells are characterized by their ability to self-renew and their potential to differentiate into any of the three germ layers ., PRC maintain the pluripotency of the cells by maintaining the developmental regulators repressed 10–12 ., On differentiation ES cells acquire cell-type specific gene expression patterns that strongly depend on the genome-wide redistribution of the Polycomb proteins 12 ., Activation of tissue specific genes correlates with the displacement of Polycomb proteins and a decrease of the H3K27me3 mark during retinoic acid induced neuronal differentiation 13 ., However , it has been recently shown that Polycomb proteins can also be recruited to activated genes to attenuate the retinoic acid associated transcriptional activation of specific genes 14 ., The important function of Polycomb complexes in the epigenetic changes induced by retinoic acid in mouse embryonic stem cells has been recently reviewed by Gudas 15 ., The composition of the PRC1 complex changes during the differentiation of ES cells ., Cbx7 is the primarily expressed Polycomb ortholog in ES cells but it is quickly downregulated during differentiation while Cbx2 , Cbx4 and Cbx8 are induced 16 , 17 ., These studies showed that the integrity of Cbx7 was required for stable ES cell maintenance , while Cbx2 and Cbx4 were required for balanced lineage specification ., It is worth noting that similar results have been obtained for hematopoietic stem cells 18 ., However , some important questions about Polycomb proteins remain unanswered ., Despite their overt relevance for ES cell differentiation , it is poorly understood how Polycomb repressed states are established and resolved ., How PRCs are initially recruited to target genes is a matter of continuous debate ( discussed in 1 ) ., Similarly it is unclear how the transition from a PRC repressed state to an active state is achieved ., How changes in PRC composition relate to these transitions has not been investigated ., Here , we analyzed the genome wide recruitment of Cbx8 in ES cells induced to differentiate ., We provide compelling evidence suggesting that Cbx8 is part of a transitory PRC1 complex facilitating the activation of Cbx7-PRC1-repressed genes during the commitment to differentiation ., We used retinoic acid ( RA ) to induce mouse E14 ES cells to start differentiating towards the neuronal lineage ., We confirmed previous results 16 , 17 showing that Cbx8 was virtually absent in self-renewing ES cells but potently induced on protein and RNA levels after three days of RA-induced differentiation ( Fig . 1A and S1A Figure ) ., To assess the genome wide distribution of Cbx8 in differentiating ES cells , we enriched Cbx8-bound chromatin by chromatin immunoprecipitation ( ChIP ) using an antibody generated against the unique part of the protein ( S1B-G Figure ) and analyzed the co-precipitated DNA by direct massive parallel sequencing ( ChIP-seq ) ., Taking advantage of the fact that Cbx8 is virtually absent in untreated , self-renewing ES cells ( Fig . 1A ) , we decided to use both IgG as well as Cbx8 ChIPs from untreated cells as negative controls ., We were able to uniquely map 9–16 million reads per sample ( S2A Figure ) ., For our further analysis we used a set of high confidence binding sites that were identified by the overlap of peaks that were called by MACS comparing Cbx8 ChIP from RA-treated ES cells to IgG and those called comparing Cbx8 ChIP from RA-treated ES cells to the antibody-specific background ChIPed from untreated cells ( S2B Figure ) ., By this method we were able to identify a subset of 171 peaks corresponding to Cbx8 binding sites of high confidence ( S1 Table ) ., Peaks were annotated to the nearest gene if the center of the peak was inside a window flanking the transcribed region by 3 kb ., Plotting the average read coverage on genes with annotated Cbx8 indicated that Cbx8 tends to accumulate on gene bodies but also to spread into upstream and downstream regions ( Fig . 1C ) ., Using this annotation we found that the large majority ( 141/171 ) of identified peaks are associated to annotated genes ( Fig . 1D ) ., We performed a microarray analysis comparing self renewing untreated ES cells and differentiating cells after 3 days of RA treatment and crossed the data with our ChIP-seq to identify possible transcriptional changes on Cbx8 target genes ., Taking into consideration the established function of Polycomb proteins in gene repression , we expected target genes to either not change because they are maintained in a repressed state or to be downregulated ., We could extract data for 121 of the 141 gene-associated peaks ., We found that about one third of these Cbx8 binding sites ( 53 peaks ) annotated to genes that displayed a more than 1 . 5-fold change in gene expression ., To our surprise the large majority of these genes ( 44/53 ) was not down- but upregulated ( Fig . 1E ) ., Most of these upregulated genes were repressed in untreated cells as indicated by very low average probe intensities ( Fig . 1F ) ., Though a previous report has already shown that Cbx8 can be found on a handful of activated genes in differentiating cells 19 , our data suggested that this could actually be true for a substantial fraction of Cbx8 target genes ., In order to confirm this , we selected a panel of target and control genes and simultaneously analyzed Cbx8 recruitment and the corresponding mRNA levels ., We were able to confirm differentiation-induced recruitment of Cbx8 on all target genes tested while control genes were negative ( Fig . 2A ) ., Target genes included many important key differentiation genes such as Sox9 , Gata6 and Nkx6-1 whose expression was potently induced ( Fig . 2B ) ., In order to exclude the possibility that Cbx8 binding and active transcription might occur on different and exclusive alleles within the cell population , we analyzed the co-occurrence of Cbx8 and H3K36me3 , which is a mark of active transcription 20 , by coupled ChIP in differentiating ES cells treated for three days with RA ., First , we have analyzed five Cbx8 target genes and four non-target genes that included Oct4 , Nanog , Gapdh and Rpo ., Despite the fact that Nanog is downregulated after three days of RA treatment ( Fig . 2B ) it still retained some H3K36me3 that was in a similar range as on the activated gene Sox9 or the constitutively active gene Gapdh ( Fig . 2C , left panel ) suggesting that removal of the active mark H3K36me3 follows a slower dynamic than the actual gene repression ., Taking advantage of the fact that with Gapdh , Nanog , and Sox9 we had identified target and non-target genes of Cbx8 with comparable H3K36me3 levels , we have used anti-Cbx8 antibody-enriched chromatin as input material for a secondary ChIP with IgG and anti-H3K36me3 antibody ., As shown in the right panel of Fig . 2C , we found a clear enrichment of H3K36me3 over IgG on Sox9 and on the other Cbx8 target genes but not on non-target genes such as Gapdh or Nanog ., As chromatin ranging in size between 300–500 bps has been used for these experiments H3K36me3 and binding of Cbx8 could be occurring on different H3 tails on the same or even neighboring nucleosomes ., The observation that Cbx8 can be simultaneously detected on the same locus provides further support to the idea that Cbx8 is recruited to genes that become actively transcribed ., To study the functional importance of Cbx8 for the activation of differentiation genes , we used lentiviral vectors expressing two different short hairpin RNAs ( shRNA ) directed specifically against the Cbx8 transcript ., After selection stably transduced cells were used for the study ., Both shRNAs efficiently repressed the expression of Cbx8 on both the mRNA and protein levels in ES cells treated with RA ( Fig . 3A , B ) ., The activation of upregulated Cbx8 target genes was significantly decreased in cells depleted for Cbx8 ( Fig . 3C ) ., Among the Cbx8-sensitive genes were several pivotal regulators of differentiation processes such as Sox9 and Nkx6-1 , that have been shown to be important transcription factors required for normal brain development 21 , 22 ., The reduction of Cbx8 occupancy was similarly efficient on up- as well as downregulated genes ( Fig . 3D ) , however , the reduction in Cbx8 levels didnt affect the repression of its target genes Prdm14 and Otx2 ( Fig . 3C ) ., Importantly , the non-target genes Oct4 and Nanog , which encode the regulators of pluripotency , were similarly repressed in Cbx8 deficient and control cells ( Fig . 3C , D ) ., Plotting enriched gene ontologies according to their similarity in a semantic space illustrates a clear overrepresentation of both transcriptional and differentiation regulators ( Fig . 4A ) , which are the classical categories of Polycomb target genes in self-renewing ES cells 12 ., Gene ontologies related to neuronal development were preferentially enriched in the subgroup of activated Cbx8 genes but not those target genes that did not show any change in gene expression ( S3A Figure ) ., Downregulated genes were not sufficient in number to yield a result in gene ontology analysis ., We compared our genome wide Cbx8 binding profile in differentiating ES cells with published ChIP-seq data obtained from self-renewing ES cells 17 ., The binding of Cbx8 in RA-treated differentiating ES cells mirrors the binding of PRC1 proteins Ring1b and Cbx7 within H3K27me3 domains in untreated self-renewing ES cells ( Fig . 4B ) ., As shown in Fig . 4C , this holds true for the vast majority of Cbx8 binding sites in RA-treated ES cells as 133/171 overlapped with sites bound by Cbx7 in self-renewing ES cells ., Similar overlaps were observed with Ring1b and H3K27me3 ( S3B Figure ) ., The unexpected link between Cbx8 recruitment and gene activation prompted us to analyze whether Cbx8 acted alone , as part of a PRC1 or part of a different complex ., First we compared the dynamics in occupancy of Cbx8 and Ring1b , which is the least variant component of PRC1 ., As already suggested by the ChIP-seq data ( Fig . 4B ) , Ring1b was strongly enriched on target genes in self-renewing ES cells ., While the recruitment of Cbx8 initially increased and peaked after three days of retinoic acid induction , the binding of Ring1b decreased progressively reaching background levels at day five of differentiation ( Fig . 5A ) ., At day five of retinoic acid induced differentiation , the Cbx8 occupancy dropped to very low levels similar to Ring1b ., In contrast , H2A ubiquitination , mark set by Ring1b 23 , persisted over the entire time course ( Fig . 5A ) ., In order to understand whether Cbx8 and Ring1b are acting together , we decided to identify the proteins that bind Cbx8 during ES cell differentiation ., Therefore , we generated ES cells stably expressing epitope-tagged Cbx8 ( Fig . 5B ) ., Exogenous Cbx8 expressed in untreated ES cells bound to the same target genes as endogenous Cbx8 in RA-treated ES cells suggesting that the epitope does not affect its function ( S4A Figure ) ., We harvested cells at day three of differentiation which corresponds to the time point with maximal recruitment of endogenous Cbx8 to target genes ( Fig . 5A ) ., Cbx8 and interacting proteins were enriched by affinity purification , analyzed by mass spectrometry and significantly enriched proteins were ranked according to their abundance ( S2 Table ) ., After the bait protein Cbx8 , the top five ranked proteins were the PRC1 subunits Ring1a/b , Phc2 and the Pcgf proteins Mel18 and Bmi1 ., Three additional PRC1 subunits were found in lower abundance ( Fig . 5B ) ., Next , we tested whether Ring1b and H2A ubiquitination co-occur with Cbx8 on genes ., Therefore , we have used anti-Cbx8 antibody-enriched chromatin as input material for a secondary ChIP with IgG , anti-Ring1b and anti-ubiquitinated H2A antibody ., As shown in Fig . 5C we could detect co-enrichment on the target genes Sox9 , Nkx6-1 , Lhx2 and Gata2 but not on the non-target genes Oct4 , Rpo or Gapdh ., Taken together , these results suggested that Cbx8 acts as part of a PRC1 complex ., To further support this , we have analyzed the occupancy of Cbx8 target genes by Ring1b , Cbx7 and H2A ubiquitination in RA-treated cells after Cbx8 knockdown ., The resulting reduction of Cbx8 occupancy ( Fig . 3D ) correlated with a small but consistent reduction of Ring1b on several Cbx8 target genes without affecting non-target genes ( Fig . 6 ) ., Notably , global Ring1b protein levels were not affected ( S4B Figure ) ., However , we observed an increased incorporation of Cbx7 that could partially compensate for Cbx8 loss ( Fig . 6 ) ., This was not the consequence of an upregulation of Cbx7 expression as knockdown of Cbx8 did not affect the mRNA levels of Cbx7 or other Cbx proteins ( S4C Figure ) ., Moreover we found that Cbx8 loss did not affect the levels of H2A ubiquitination on its target genes ( Fig . 6 ) ., In order to address the question whether Cbx8-containing PRC1 on activated genes is repressive or activating , we stably interfered with the expression of Ring1b , the least variant component of PRC1 ., When analyzing Ring1b knockdown cell after 3 days of RA treatment we did not observe any compensation by Ring1a but a slight increase in Cbx8 mRNA ( S4D Figure ) ., Under these conditions the Cbx8 target gene Gata2 was further upregulated while other target genes such as Nkx6-1 and Sox17 were less activated ( S4D Figure ) ., These results are difficult to interpret as knockdown of Ring1b interferes with all PRC1 complexes present prior and after RA treatment ., We then focused our attention on the mechanism that recruits Cbx8 to its target genes ., An obvious possibility is that it binds directly to H3K27me3 ., Although activated Cbx8 target genes progressively reduced their H3K27me3 levels , in contrast to Ring1b , this reduction occurred only very slowly and even after five days of differentiation genes still retained half-maximal or even higher levels of H3K27me3 ( Fig . 7A ) ., Depletion of Cbx8 did not affect the amount of H3K27me3 detectable 3 days after RA treatment ( Fig . 7B ) ., We then analyzed by ChIP-ReChIP if Cbx8 and H3K27me3 co-occur on target genes in ES cells treated for 3 days with RA ., Indeed , we found that Cbx8 and H3K27me3 co-existed on Cbx8 target genes ( Fig . 7C ) ., This suggests that also in the context of gene activation the interaction of Cbx8 with H3K27me3 is the most likely mechanism of recruitment to chromatin ., In addition to H3K27me3 , at day 3 of RA treatment we also detected some acetylation of the same residue on Cbx8 target genes ( S5B Figure ) ., This H3K27ac was of low level when compared to an enhancer that was previously described to be marked by H3K27ac in ES cells ( S5A Figure ) 24 ., Curiously , knockdown of Cbx8 lead to a modest but significant decrease of H3K27ac ( S5B Figure ) ., The presence of these two mutually exclusive H3K27 marks provided us with a valuable tool to interrogate the preferential binding of Cbx8 ., First we titrated both antibodies to reach similar enrichment in ChIP assays ( S5C Figure ) ., Then we used the enriched material to perform a sequential ChIP for Cbx8 ., As shown in S5C Figure , Cbx8 preferentially bound K27me3-marked chromatin ., Cbx7 and Cbx8 are expressed in an almost mutually exclusive manner in self-renewing and differentiating ES cells , respectively 16 , 17 ., To further gain additional insight into the functional relevance of the switch from Cbx7 to Cbx8 on target genes , we generated mouse embryonic stem cells that stably express exogenous Cbx8 and analyzed them in self-renewing conditions while cells expressing exogenous Cbx7 were analyzed in differentiating cells after 3 days of retinoic acid induction ( Fig . 8A ) ., When expressed in self-renewing cells , exogenous Cbx8 was able to efficiently outcompete Cbx7 for its target genes ( Fig . 8B ) ., The enforced recruitment of exogenous Cbx8 achieved under these conditions was two-to-three-fold higher than that observed in differentiating cells for the endogenous protein ( Fig . 8B ) , although , this did not affect the low expression level of these genes ( Fig . 8C ) ., In the converse experiment during differentiation , Cbx7 overexpression significantly reduced the activation of Cbx8 target genes ( Fig . 8C ) ., Although enrichment of overexpressed Cbx7 on target genes did not reach the levels of the endogenous protein in self-renewing cells , it resulted in an efficient displacement of Cbx8 ( Fig . 8D ) ., Taken together our results support a model in which PRC1 containing Cbx8 replace PRC1 containing Cbx7 on developmental genes , which facilitates the transition from a repressed chromatin state to gene activation during early ES cell differentiation ( Fig . 8E ) ., Prolonged gene activation results in eviction of PRC1 complexes despite some persisting H3K27me3 and H2A ubiquitination ., This mechanism affects several key regulatory genes and thus probably contributes to a robust execution of differentiation programs ., Our data supports the idea that Cbx8 acts in the context of an intact PRC1 complex ., Initially , we considered also two other possibilities: Cbx8 could act as monomer competing with repressive PRC1 complexes for H3K27me3 binding sites , or Cbx8 acts in complex with non-polycomb proteins that are activating ., Indeed , in leukemia cells Cbx8 has already been described as a component of activating complexes containing Tip60 and MLL-AF9 25 ., However , performing mass spectrometric analysis of affinity purified Cbx8-complexes from differentiating ES cells we could not detect any of these proteins , but rather found other PRC1-components acting as main binding proteins ( Fig . 5B ) ., Moreover , we have shown that Ring1b was enriched on Cbx8 immunoprecipitated chromatin compared to IgG ( Fig . 5C ) and that the knockdown of Cbx8 resulted in a small but detectable reduction of Ring1b occupancy on genes ( Fig . 6 ) ., Since we would have expected the opposite if Cbx8 acted as a monomer on target genes , this observation taken together with our co-immunoprecipitation data strongly argued for Cbx8 to be functioning in the context of a PRC1 complex ., One fundamental question is how a Cbx8-containing PRC1 is able to contribute to gene activation ., A plausible explanation could be that Cbx8-containing PRC1 is simply less repressive than the PRC1 containing Cbx7 that it replaces ., This prompted us to test whether Cbx8 has an influence on PRC1 activity ., When monitoring the ubiquitination of H2A , we found that its levels on Cbx8 target genes neither decreased during the first five days of differentiation nor changed on depletion of Cbx8 ( Figs . 5A , 6 ) ., It has been shown that Ring1b mediates chromatin compaction and gene repression independently of its catalytic activity 26 ., In that regard , it would be interesting to test whether changes in PRC1 composition affects its capacity to compact chromatin ., If Cbx7-containing PRC1 complexes induced a higher degree of chromatin compaction , this could possibly be relaxed on replacement by Cbx8 ., Knockdown of Ring1b itself let to the further activation of a Cbx8 target gene but to a reduction in the activation of several other Cbx8 target ( S4D Figure ) ., It is intriguing to speculate that the outcome could depend on the ratio of Cbx7 and Cbx8-containing PRC1 complexes present at the time point of analysis and the gene-specific kinetics of activation ., However , the interpretation of such data is complicated by the large number of binding sites of Ring1b as its other target genes could exert indirect effects and the fact that Ring1b occurs in both PRC1 and non-PRC1 complexes 27 ., Finally , of course we cannot exclude the possibility that Cbx8-containing PRC1 complexes are able to recruit activating proteins in a very transient way , which would not be captured by our purification and mass spectrometric analysis ., Both the replacement of Cbx7-containing PRC1 for Cbx8 loaded complexes and the complete eviction of all PRC1 after prolonged activation occurs in the context of persisting H3K27me3 and H2A ubiquitination ., This reiterates that these marks have a slow turnover and are not necessary a reflection of gene activity or repression , in particular during dynamic cell fate transitions ., On Cbx8 target genes we could detect low levels of H3K27ac that were further reduced mildly but significantly in cells depleted for Cbx8 ( S5B Figure ) ., Whether Cbx8-dependent H3K27ac is a cause or consequence of enhanced transcription is unclear ., H3K27ac could positively affect chromatin accessibility for transcription-supporting factors ., However it is more likely that the observed low level of H3K27ac is a collateral consequence of an increased concentration of histone acetylases travelling with Polymerase II ., The modularity of PRC1 is illustrated by the fact that there are 180 theoretical combinations for assembling the different PRC subunits ., The real number of different PRC1 complexes existing under different physiological conditions is likely to be much lower due to mutually exclusive expression patterns and preferential binding between subunits ., The systematic proteomic and epigenomic analysis of all six PCGF proteins shed some initial light on the different compositions and genomic distributions of PRC1 complexes 28 ., On the one hand proteomic studies allowed the identification of new PRC1 components such as RYBP and YAF2 , while on the other hand they increased doubts about the genuineness of subunits that had previously been considered to be canonical; such as Cbx6 ., A major challenge for the field is to understand how the modularity of PRC1 is regulated and how it contributes to cellular functions ., Self-renewing ES cells primarily express two PRC1 complexes containing either Cbx7 or RYBP 7 ., Whereas Cbx7-PRC1 mediates early repression of differentiation genes by binding to H3K27me3 , RYBP-PRC1 binds independently of the methylation status of H3K27 and is associated with lower levels of Ring1b and H2A ubiquitination and occupies genes that are less repressed 29 ., Cbx7 and Cbx8 are expressed in an almost mutually exclusive manner in self-renewing and differentiating ES cells , respectively ., Although in vitro the chromodomain of Cbx7 has higher affinity to H3K27me3 than the one of Cbx8 30 , in cells we found that both proteins were able to efficiently replace each other on genes ., Enforced expression of Cbx7 in differentiating cells was able to compete with Cbx8 for its target genes and to reduce gene activation ., In the converse experiment similarly efficient replacement of Cbx7 by exogenous Cbx8 in self-renewing ES cells was not sufficient to induce derepression ( Fig . 8A–D ) ., These results place Cbx7 over Cbx8 in the functional hierarchy of Polycomb proteins ., Future studies will have to assess in greater detail how different PRC1 complexes regulate cell fate decisions ., Our preliminary data shows that ES cells that maintain 30–50% reduced Cbx8 expression are qualitatively able to differentiate into Tuj1-positive neurons with neurite outgrowth following 11 days of a long-term differentiation protocol ( adapted from 31 ) , but seem to do so in a less efficient way ( S6 Figure ) ., The number of observations linking Polycomb proteins and active gene transcription are increasing ., Here we report that the transient recruitment of PRC1 containing Cbx8 facilitates the transition from a Polycomb-repressed to a fully active state of key regulatory genes during early differentiation ., Others have suggested that binding of Cbx8 to active genes could mark these for later repression 19 ., Our data does not support this as we find Cbx8 recruitment to be only transient and all PRC1 to be entirely evicted after prolonged gene activation ( Fig . 5 ) ., It is worth to point out that Cbx8 target genes that got repressed after treatment with retinoic acid were not affected by the knockdown of Cbx8 ., This can be explained by a possible compensation by other canonical repressing PRC1 complexes or a recent finding showing that Polycomb protein recruitment is rather a consequence than a cause of initial gene silencing 32 ., Studying differentiating myocytes , others have reported that Ezh1 associates with actively transcribed genes and further argued for a positive function in which Ezh1 could be required for the recruitment of Polymerase II 33 ., In contrast , Pombo and colleagues suggested that their observed Polycomb-binding to transcribed metabolic genes was the consequence of a continuous switching between an active and a Polycomb-repressed state restraining transcriptional elongation 34 ., In Drosophila cells , PRC1 was found to indirectly associate with some active as well as inactive genes by binding to structural cohesin proteins 35 ., The authors argued again for a more active role on active genes by suggesting that PRC1 could be required for allowing the phosphorylation that makes Polymerase II elongation competent ., A large body of additional work is needed to sort out the relation between Polycomb-bound and transcribed chromatin states in greater detail ., In this endeavor it will be important to carefully distinguish between passive contributions from reductions in repressive potential and genuine contributions to transcription initiation and elongation ., We produced a specific polyclonal antibody against mouse Cbx8 by immunizing rabbits with a His-tagged fragment of Cbx8 protein encompassing amino acids 201–360 ., Serum was precleared with sepharose and passed over a column containing a fusion protein of glutathione-S-transferase ( GST ) and amino acids 201–360 of Cbx8 covalently cross-linked to Glutathione sepharose ., Anti-Cbx8 antibody was eluated with low pH , dialyzed and stored in PBS with 20% glycerol ., The antibody performed in a similar way to antibodies previously described and kindly provided by Kristian Helin 13; ( S1 Figure ) ., In addition we made use of the following antibodies: anti-IgG ( Abcam ) , anti-H3 C-terminal ( Abcam ) and anti-H3K27me3 ( Millipore ) , anti-Flag M2 ( Sigma-Aldrich ) , anti-Ring1b provided by Luciano Di Croce 36 , anti-H3K27Ac and anti-Cbx7 ( Abcam ) , anti-H3K36me3 ( Abcam ) and rabbit monoclonal anti-ubiquityl-H2A ( Lys119 , D27C4 , Cell Signaling Technology ) ., The amounts and concentrations used for ChIP and western blotting is given in S3 Table ., Expression plasmids and pLKO-1 constructs for shRNA-mediated knockdown were generated with standard PCR and cloning techniques ., For stable expression in ES cells , cDNAs were cloned in frame with a multiple-epitope tag into a vector containing CAG promoter and an IRES-puromycin resistance gene cassette 37 ., A modified and gateway cloning-adapted ( Life Technologies ) version was kindly provided by Diego Pasini ., E14Tg2A . 4 mouse ES cells were cultured as previously described 38 ., Cells were transduced with lentiviral shRNA cassettes essentially as described before 39 ., Transduced cells were selected with 2 µg/ml puromycin ., For shRNA sequences see S3 Table ., For the generation of stably expressing ES cell clones , the above described CAG promoter driven vectors were transfected into mouse ES cells E14 using Lipofectamine 2000 ( Life Technologies ) ., Stable transfectants were selected with 2 µg/ml puromycin and analyzed by RT-PCR and western blot for transgene expression ., For neuronal differentiation , 1 µM all trans retinoic acid ( RA ) was added directly to cells in medium without leukemia inhibitory factor ., Unless indicated otherwise in Figure legends , cells were collected after 3 days of RA treatment ., Lysis and western blot analyses were performed as previously described 40 ., Following the suppliers instructions , RNA was purified from 2×106 cells using the RNeasy minikit ( Qiagen ) , with a DNase I digestion step to avoid any potential DNA contamination ., Total RNA ( 1 µg ) was reverse transcribed using a cDNA synthesis kit ( Roche Diagnostics ) and oligo ( dT ) primers ., Relative cDNA levels were quantified by quantitative PCR ( qRT-PCR ) ., Values were normalized to the expression of two housekeeping genes ( Rpo and Gapdh ) ., For gene expression analysis , four biological replicates of RA treated ( 3 days ) and untreated ES cells were used for each condition and samples were prepared and hybridized to SurePrint G3 Mouse GE 8×60K Microarrays ( Agilent technologies ) following the suppliers instructions ., Analyses were essentially performed as described 41 selecting differentially expressed probes with a FDR of 0 . 05 and fold change of >1 . 5 ., Chromatin fragmented to a size ranging from 300–500 bps and immunoprecipitation ( ChIP ) experiments were performed essentially as previously described 42 ., ChIP-reChIP experiment was performed as described elsewhere 34 ., The sequences of all oligonucleotides used here are provided in the S3 Table ., Unless indicated otherwise , ChIP results are given as the percentage of the amount of ChIP-enriched DNA relative to the amount of DNA isolated from one tenth of input material measured by quantitative PCR ., For ChIP-sequencing ( ChIP-seq ) , 10 ng of DNA was enriched by ChIP and fluorimetrically quantified with PicoGreen ., Library generation and direct massive parallel sequencing on an Illumina genome analyzer were performed according to the suppliers instructions ., Reads obtained were cleaned based on quality , trimmed using the ShortRead package in R 43 and aligned with the mouse genome ( NCBIM37/mm9 ) using Bowtie version 0 . 12 . 7 44 , two mismatches were allowed for the alignment within the seed , only reads mapping to a single position in the genome were used ., To detect genomic regions with significant enrichment we used MACS software version 1 . 4 . 1 45 ., For peak calling of Cbx8 in RA-treated ES cells we used a p-value cut-off of 1×10−4 and a FDR of 5% ., Both IgG and Cbx8 from self-renewing cells ( that do not express Cbx8 ) were independently used as control libraries ., Only peaks called in both cases ( minimal overlap of 50 bps ) were accepted as high confidence target peaks and further analyzed ., A subset was validated by direct ChIP ., Peaks were annotated using ChIPpeakAnno package 46 ., Genes were considered to be target genes if the center of a peak was found in the transcribed region ±3 kb using the transcript set of Mouse Ensembl Gene ( based on assembly NCBIM37/mm9 ) ., In cases where a peak annotated to two genes , the nearest gene was selected and identified by the minimal distanc | Introduction, Results, Discussion, Materials and Methods | Polycomb proteins play an essential role in maintaining the repression of developmental genes in self-renewing embryonic stem cells ., The exact mechanism allowing the derepression of polycomb target genes during cell differentiation remains unclear ., Our project aimed to identify Cbx8 binding sites in differentiating mouse embryonic stem cells ., Therefore , we used a genome-wide chromatin immunoprecipitation of endogenous Cbx8 coupled to direct massive parallel sequencing ( ChIP-Seq ) ., Our analysis identified 171 high confidence peaks ., By crossing our data with previously published microarray analysis , we show that several differentiation genes transiently recruit Cbx8 during their early activation ., Depletion of Cbx8 partially impairs the transcriptional activation of these genes ., Both interaction analysis , as well as chromatin immunoprecipitation experiments support the idea that activating Cbx8 acts in the context of an intact PRC1 complex ., Prolonged gene activation results in eviction of PRC1 despite persisting H3K27me3 and H2A ubiquitination ., The composition of PRC1 is highly modular and changes when embryonic stem cells commit to differentiation ., We further demonstrate that the exchange of Cbx7 for Cbx8 is required for the effective activation of differentiation genes ., Taken together , our results establish a function for a Cbx8-containing complex in facilitating the transition from a Polycomb-repressed chromatin state to an active state ., As this affects several key regulatory differentiation genes this mechanism is likely to contribute to the robust execution of differentiation programs . | Cell fate transitions have long been known to be accompanied by alterations in chromatin structure ., But only during the last few years has it become clear that chromatin modifications form the molecular basis of an epigenetic memory that defines cell identity ., The Polycomb Group Proteins ( PcGs ) form two major protein complexes known as polycomb repressive complexes 1 and 2 ( PRC1 and PRC2 ) ., Their function is essential for the maintenance of transcriptional repression during embryogenesis through the methylation of the lysine 27 on histone H3 and the subsequent ubiquitination of histone H2A ., The chromobox homolog 8 , Cbx8 , which is part of the PRC1 complex , is therefore generally defined as a repressor of gene transcription ., The genome wide profiling of Cbx8 during the early steps of mouse embryonic stem ( mES ) cells differentiation provided us with surprising results involving Cbx8 in gene activation ., Our results point out that Cbx8 is part of a PRC1 complex involved in the transition from a Polycomb repressed state to an active state . | cell biology, chromosome biology, genetics, biology and life sciences, epigenetics, molecular cell biology, chromatin, histone modification | null |
journal.pgen.1001193 | 2,010 | A Coastal Cline in Sodium Accumulation in Arabidopsis thaliana Is Driven by Natural Variation of the Sodium Transporter AtHKT1;1 | Uncovering the genetic polymorphisms that underlie adaptation to environmental gradients is a critical goal in evolutionary biology , and will lead to a better understanding of both the types of genetic changes and the gene functions involved ., Such understanding will not only provide insight into how organisms may respond to future global climate change , but will also provide tools for the development of agricultural systems and ecological services that are more resilient to such changes ., Patterns of phenotypic diversity across environmental gradients can be indicative of adaptive responses to selection , and evaluation of these patterns has the potential to lead to the identification of the genetic polymorphisms underlying these adaptive responses ., Numerous studies in animals and plants have identified phenotypic clines in various life history traits , but only a few have determined the genetic changes driving such traits ., In Arabidopsis thaliana , plasticity in seasonally regulated flowering appears to be modulated by a network of gene interactions responsive to both vernalization and photoperiod signals 1 ., Adaptive clines in resistance to oxidative stress and chilling 2 , and wing size 3 in Drosophila melanogaster are modulated by the Insulin-like Receptor ( InR ) and Drosophila cold acclimation ( Dca ) genes , respectively ., While adaptation to high altitude in Peromyscus maniculatus ( Deer mice ) is associated with enhanced pulmonary O2 loading driven by alterations in α-globin and β-globin genes 4 ., These genetic changes are all associated with adaptation to variation in environmental factors that vary with latitude or altitude ., Such systematic variation has greatly facilitated the discovery of these loci and their adaptive significance ., Clines in various life history traits have also been identified in plants growing on serpentine 5 , saline 6 , 7 , and mine impacted soils 8 ., Progress has been made in outlining the genetic architecture of these adaptive traits 5 , 8–10 , though a molecular genetic understanding is still needed ., A . thaliana is broadly distributed in its native Europe and central Asia , where it experiences a wide range of altitudinal , climatic , and edaphic conditions , leading to a range of selective pressures 11 ., Whether the wide variety of natural phenotypic and genetic variation observed in A . thaliana 12 contributes to its local adaptation is an important unresolved question that is currently attracting a significant amount of attention 13 ., Because of its relevance to crop production , salinity tolerance in plants has been studied intensively 14 , and natural plant populations adapted to such conditions have provided an excellent system for studying the evolutionary mechanisms of adaptation and speciation in coastal 6 , 10 and salt marsh 7 , 9 , 15–18 environments ., The primary effects of excess Na+ on plants are water deficit resulting from a water potential gradient between the soil solution and plant cells , and cytotoxicity due of intracellular Na+ accumulation 14 ., To overcome these effects plants must both accumulate solutes for osmotic regulation , and detoxify intracellular Na+ either by limiting its accumulation , or by compartmentalizing Na+ into the vacuole ., In addition , Na+ compartmentalization facilitates vacuolar osmotic adjustment that is necessary to compensate for the osmotic effects of salinity by maintaining turgor pressure for cell expansion and growth ., Plants therefore need to strike a balance between the accumulation of Na+ to maintain turgor , and the need to avoid Na+ chemical toxicity , and this balance will depend in part on soil salinity levels ., Given the critical role Na+ accumulation plays in salinity tolerance , we used this life history trait to probe the global A . thaliana population for signals of adaptive selection for growth in saline impacted environments ., We grew 349 accessions of A . thaliana in a controlled common garden in non-saline soil , and analyzed leaf Na+ accumulation ., We observed a wide range of leaf Na+ accumulation across the accessions ( 330–4 , 848 mg kg−1 dry weight ) ., If this natural variation in leaf Na+ accumulation capacity is related to adaptation to growth in saline soils we would expect to find evidence of an adaptive cline , or a gradient of leaf Na+ accumulation that correlates with the geographical distribution of variation in soil salinity ., Salinity impacted soils are expected to occur in coastal regions due to air born deposition of sea spray which can occur many tens of km inland 19–22 , but can also occur in areas distant from the coast through high Na+ in the soil or ground water ., Elevated soil salinity can also be caused by inappropriate irrigation practices such as irrigation with saline water or poor drainage ., To test for the existence of an adaptive cline in leaf Na+ accumulation capacity and soil salinity we related leaf Na+ accumulation capacity to the distance of the collection site for each accession to the coast , or to the nearest known saline soil , whichever is the shortest ., We focused on European accessions since a good soil salinity map exists for this region 23 , which left 300 accessions ., Regressing the distance to the coast , or nearest known saline soil , on leaf Na+ for all 300 accessions revealed a significant relationship ( p-value<2e-12 ) , establishing that accessions with elevated leaf Na+ are more likely to grow in potentially saline impacted soils ( Figure 1A and 1B ) ., To investigate the genetic architecture underlying this cline in leaf Na+ accumulation capacity we performed a genome-wide association ( GWA ) study ( previously described for a smaller data set 24 ) to identify regions of the genome at which genetic variation is associated with leaf Na+ accumulation capacity ., The 337 A . thaliana accessions used in our GWA study , which are a subset of the 349 accessions phenotyped for leaf Na+ , were genotyped using the Affymetrix SNP-tilling array Atsnptile1a which can interrogate 248 , 584 SNPs ., To assess evidence of association between SNPs and leaf Na+ accumulation we used a mixed-model approach 25 to correct for population structure , as previously described 24 ., In the current analysis we identified a single strong peak of SNPs associated with leaf Na+ , with the peak centered on AtHKT1;1 ( Figure 2 ) , a gene known to encode a Na+-transporter 26 ., Accessions with a thymine ( T ) at the SNP most significantly associated with leaf Na+ at position 6392276 bp on chromosome 4 ( Chr4:6392276 ) have significantly higher leaf Na+ than accessions with a cytosine ( C ) at this same position ( 2 , 325 vs . 955 mg Na+ kg−1 dry weight , p-value<2e-16 ) ., This SNP explains 32% ( without accounting for population structure ) of the total variation in leaf Na+ accumulation observed ., Previously , in independent test crosses between the high leaf Na+ accessions Ts-1 and Tsu-1 ( both containing a T at Chr4:6392276 ) and the low leaf Na+ accession Col-0 ( containing a C at Chr4:6392276 ) QTLs for leaf Na+ centered on AtHKT1;1 were identified in both F2 populations 27 ., Such genetic evidence provides independent support that the peak of SNPs associated with leaf Na+ observed in our GWA analysis , centered at AtHKT1;1 ( Figure 2 ) , represents a true positive association and not a false positive driven by the high degree of population structure known to exist in A . thaliana 24 ., Reduced expression of AtHKT1;1 in Ts-1 and Tsu-1 was concluded to drive the elevated leaf Na+ observed in these two accessions 27 ., Here , we expand on this observation by establishing the strength of the AtHKT1;1 alleles in four further high Na+ accumulating accessions ( Bur-1 , Duk , PHW-20 and UKNW06-386 ) that all contain a T at Chr4:6392276 , along with a low leaf Na+ accession ( Nd-1 ) with a C at Chr4:6392276 ., By examining the leaf Na+ accumulation in F1 plants from crosses of each of these accessions to Col-0hkt1-1 and Col-0HKT1 , we were able to establish a significant correlation between leaf Na+ accumulation and the strength of the AtHKT1;1 alleles ( Figure 3A ) ., These crosses confirmed that all accessions tested with elevated leaf Na+ , and that contain a T at Chr4:6392276 , have hypofunctional alleles of AtHKT1;1 relative to the Col-0 allele ., Furthermore , analysis of the expression of AtHKT1;1 in the same set of accessions revealed that allelic variation in AtHKT1;1 strength is modulated at the level of gene expression ( Figure 3B ) , consistent with what was previously observed for Ts-1 and Tsu-1 27 ., Though the SNP most significantly associated with leaf Na+ ( Chr4:6392276 ) is unlikely to be causal for these AtHKT1;1 expression level polymorphisms , this SNP can be used as a linked genetic marker to determine the type of AtHKT1;1 allele present , with a T at this SNP being associated with weak AtHKT1;1 alleles ., Using the SNP at Chr4:6392276 as a genetic marker for the type of AtHKT1;1 allele ( strong or weak ) allowed us to test the hypothesis that the leaf Na+ soil salinity cline we observe in European populations of A . thaliana ( Figure 1A and 1B ) is associated with weak alleles of AtHKT1;1 ., By comparing the means of distances to the coast , or known saline soil , for the collection site of all 300 accessions with and without a T at Chr4:6392276 , we determined that a significant association ( parametric test p-value\u200a=\u200a0 . 0001; non-parametric Wilcoxon rank-sum test p-value\u200a=\u200a0 . 0062 ) exists between A . thaliana growing on potentially saline impacted soils and the presence of a weak allele of AtHKT1;1 ( Figure 1A and 1B ) ., Such a strong correlation between the presence of allelic variation at AtHKT1;1 known to drive elevated leaf Na+ , and the observed cline in leaf Na+ and saline soils , is evidence for the involvement of AtHKT1;1 in determining this geographical distribution ., Furthermore , using 13 SNPs within a 20kb region centered on HKT1;1 to define the HKT1;1 haplotype , we identify 7 haplotypes ( 6 if you combine haplotypes with only 1 SNP different ) in accessions with high leaf Na+ ( >2 , 500 ppm ) , suggesting that weak alleles of HKT1;1 have arisen independently multiple times ., However , to be credible it is also important to provide evidence that selection for growth on saline soils could be acting on the phenotype driven by allelic variation at AtHKT1;1; in this case elevated leaf Na+ ., Such evidence is provided by the previous observation that the weak allele of AtHKT1;1 in the coastal Tsu-1 A . thaliana accession not only causes elevated leaf Na+ but is also genetically linked to the elevated salinity tolerance of this accession 27 ., In A . thaliana AtHKT1;1 functions to unload Na+ from xylem vessels in the root , controlling translocation and accumulation of Na+ in the shoots 26 , 28 ., Therefore , modulation of its function would allow the balancing of Na+ accumulation in the shoot with soil salinity ., We note here that the hkt1-1 null mutation in the Col-0 background causes plants to exhibit dramatic leaf Na+ hyperaccumulation and increased NaCl sensitivity 29 , 30 ., We interpret this to mean that expression of AtHKT1;1 in the hkt1-1 null mutant is reduced to such an extent that leaf Na+ accumulation saturates the capacity for cellular detoxification of Na+ by vacuolar compartmentalization ., We propose that the naturally occurring weak alleles of AtHKT1;1 , that we show are associated with populations growing in potentially saline impacted environments , allow sufficient Na+ to accumulate in leaves for osmotic adjustment , conferring elevated Na+ tolerance ., However , these weak , but not complete loss-of-function AtHKT1;1 alleles , do not saturate the mechanism whereby the accessions avoid Na+ cytotoxicity ., The basis of this Na+ detoxification mechanism remains to be determined , though an active leaf vacuolar Na+ compartmentalization mechanism driven by AtNHX1 is one likely candidate ., In conclusion , here we provide evidence supporting the involvement of specific cis-regulatory polymorphisms at AtHKT1;1 in the potentially adaptive cline in leaf Na+ accumulation capacity we observe in A . thaliana populations to saline impacted environments ., We have identified a strong association between the AtHKT1;1 allele frequency in A . thaliana populations and their growth on potentially saline impacted soils ( Figure 1A and 1B ) ., Further , we have confirmed by GWA mapping , experimental complementation crosses , and gene expression studies , that this allelic variation directly causes changes in the clinally varying leaf Na+ accumulation phenotype via cis-regulatory polymorphisms ( Figure 2 and Figure 3 ) ., And , finally , we have previously established that the weak AtHKT1;1 alleles we show to be associated with potentially saline soils , are also linked to elevated salinity tolerance 27 , providing a plausible mechanistic link between selection for growth on saline soils and variation in AtHKT1;1 allele frequency ., Such discoveries provide tantalizing evidence that points to selection acting at AtHKT1;1 in natural populations of A . thaliana in adaptation to growth in saline environments ., Plants were grown in a controlled environment with 10 h light/14 h dark ( 90 µmol m−2s−1 photosynthetically active light ) and 19 to 22°C , as previously described 31 ., Briefly , seeds were sown onto moist soil ( Promix; Premier Horticulture ) in 10 . 5″×21″ 20 row trays with various elements added to the soil at subtoxic concentrations ( As , Cd , Co , Li , Ni , Rb , and Se 31 ) and the tray placed at 4°C for 3 days to stratify the seeds and help synchronize germination ., Each tray contained 108 plants , six plants each from 18 accessions , with three plants of each accession planted in two different parts of the tray ., Each tray contained four common accessions ( Col-0 , Cvi-0 , Fab-2 and Ts-1 ) used as controls , and 14 test accessions ., Trays were bottom-watered twice per week with 0 . 25-strength Hoagland solution in which Fe was replaced with 10 µM Fe-HBEDN , N′-di ( 2-hydroxybenzyl ) ethylenediamine-N , N′-diacetic acid monohydrochloride hydrate; Strem Chemicals , Inc . ) . After 5 weeks plants were non-destructively sampled by removing one or two leaves and the elemental composition of the tissue analyzed by Inductively Couple Plasma Mass Spectroscopy ( ICP-MS ) . The plant material was rinsed with 18 MΩ water and placed into Pyrex digestion tubes . For complementation experiments plants were crossed to Col-0 or Col-0hkt1-1 and approximately 12 F1 plants were grown in the conditions described above . A set of 360 A . thaliana accessions were selected from 5 , 810 worldwide accessions to minimizing redundancy and close family relatedness , based on the genotypes at 149 SNPs developed in a previous study 32 ., Figure S1 and Table S1 show the genetic variation in the core set of 360 accessions vs . a random set of 360 accessions chosen from the genotyped 5 , 810 accessions ., From the selected core set of 360 accessions a subset of 349 were phenotyped using ICP-MS , and of these 337 were genotyped using the Affymetrix SNP-tilling array Atsnptile1 which contains probe sets for 248 , 584 SNPs ., Details of the SNP-tilling array and methods for array hybridization and SNP-calling are the same as previously described 24 ., In brief , approximately 250 ng of genomic DNA was labeled using the BioPrime DNA labeling system ( Invitrogen ) and 16 µg of the labelled product hybridized to each array ., SNPs were called using the Oligo package after slight modifications ., Quality control ( QC ) of the genotypes , and imputation of the missing SNPs were performed following the procedure previously described 24 , except that a 15% mismatch rate was used to filter out low quality arrays ., After QC and imputation , the 337 accessions had genotypes for at least 213 , 497 SNPs ., The core set of 360 accessions selected are all available from the Arabidopsis Biological Resource Center ( http://abrc . osu . edu/ ) , and the SNP genotypes for the 337 accessions used for the GWA study are available from http://borevitzlab . uchicago . edu/resources/genetic/hapmap/BaxterCore/ ., Samples were analyzed as described by Lahner et al . 31 ., Tissue samples were dried at 92°C for 20 h in Pyrex tubes ( 16×100 mm ) to yield approximately 2–4 mg of tissue for elemental analysis ., After cooling , seven of the 108 samples from each sample set were weighed ., All samples were digested with 0 . 7 ml of concentrated nitric acid ( OmniTrace; VWR Scientific Products ) , and diluted to 6 . 0 ml with 18 MΩ water ., Elemental analysis was performed with an ICP-MS ( Elan DRCe; PerkinElmer ) for Li , B , Na , Mg , P , S , K , Ca , Mn , Fe , Co , Ni , Cu , Zn , As , Se , Rb , Mo , and Cd ., A liquid reference material composed of pooled samples of A . thaliana leaves was run every 9th sample to correct for ICP-MS run to run variation and within-run drift ., All samples were normalized to the calculated weights , as determined with an iterative algorithm using the best-measured elements , the weights of the seven weighed samples , and the solution concentrations , implemented in the Purdue Ionomics Information Management System ( PiiMS ) 33 ( for a full description see www . ionomicshub . org ) ., Data for all elements is available for viewing and download at www . ionomicshub . org in trays 1478–1504 ., To quantify the levels of AtHKT1;1 mRNA in roots of the various accessions studied , we used a protocol similar to that of Rus et al . 27 ., Roots from plants grown under identical conditions to those used for ICP-MS analysis were separated from the shoots and rinsed thoroughly with deionized water to remove any soil contamination ., The samples were frozen in liquid nitrogen and stored at −80°C until extraction ., Total RNA was extracted , and DNase digestion was performed during the extraction , using the Invitrogen PureLink RNA Mini Kit ., Two micrograms of total RNA were used as a template to synthesize first-strand cDNA with random hexamers , using SuperScript II Reverse Transcriptase ( Invitrogen Life Technologies ) ., Quantitative real-time PCR ( qRT-PCR ) was performed with first strand cDNA as a template on four technical replicates from three independent biological samples for each accession , using a sequence detector system ( StepOne Plus , Applied Biosystems ) ., For normalization across samples within a qRT-PCR run the expression of the Actin 1 gene ( At2g37620 ) was used with the following primers: CPRD66 , 5′-TGG AAC TGG AAT GGT TAA GGC TG-3′ and CPRD67 , 5′-TCT CCA GAG TCG AGC ACA ATA C-3′ ., For quantification of AtHKT1;1 the following primers were used: HKT-RTF , 5′-TGG GAT CTT ATA ATT CGG ACA GTT C-3′ and HKT-RTR , 5′-GAT AAG ACC CTC GCG ATA ATC AGT-3′ ., The fold induction relative to AtHKT1;1 expression in Col-0 roots was calculated following the method of Livak and Schmittgen 34 ., CT values were determined based on efficiency of amplification ., The mean CT values were normalized against the corresponding Actin 1 gene and ΔCT values calculated as CTAtHKT1;1–CTActin 1 ., The expression of AtHKT1;1 was calculated using the 2∧ ( ΔCT ) method 34 ., To normalize between samples analyzed in separate qRT-PCR runs , we divided the ΔCT for each line by the ΔCT of Col-0 roots in that run ., ICP-MS measurements below zero and extreme outliers ( those values that were greater than the 90th percentile + percentile ) within each tray were removed ., To account for variation in the growth environment , the four control accessions included in each tray were used to create a tray specific normalization factor ., Briefly , for each element , each control accession in a given tray was compared to the overall average for that accession across all trays to obtain an element×line×tray specific normalization factor ., The four element×line×tray factors in a give tray were then averaged to create a tray×element normalization factor for the tray ., Every value for the element in the tray was then multiplied by the normalization factor ., See Figure S2 for data of control accessions before and after the normalization ., The mean of each accession was then used for all subsequent analysis ., Normalized Na+ values and their frequency distribution can be found in Dataset S1 and Figure S3 ., Genotype calls for all 349 accessions were obtained using the methods previously described 24 ., GWA analysis was done with correction for confounding using a mixed-model that uses a genetic random effect with a fixed covariance structure to account for population structure 25 implemented in the program EMMA 24 ., The contribution of the best performing SNP ( C or T at Chr4:6392276\u200a=\u200aisT ) was checked using un-normalized Na+ data and the linear model: ( 1 ) using the lm and anova functions from R v2 . 9 . 1 ., The control accessions were excluded from this analysis ., The output of the statistical model can be found in Text S1 ., Although the samples were nested in trays , Figure S4 indicates that the best performing SNP is essentially evenly distributed across all trays ., The geographical location of each accession was obtained from TAIR ( www . arabidopsis . org ) ., When processing the original data , we found an inconsistency for one of the high-Na accessions , CS28373 ( also known as Jm-1 ) ., The listed latitude and longitude ( 49 , 15 ) of the accession do not match the location name “Jamolice” from where this accession was collected ., The town Jamolice is located at 49 . 0721283 latitude and 16 . 2532139 longitude ( http://www . gpsvisualizer . com/geocode ) ., In the interests of consistency , we used the original coordinates , although altering the location did not materially change the analysis ., The distance to the coast or saline/sodic areas was calculated by obtaining the longitudes and latitudes of the shoreline/coast from the National Oceanic and Atmospheric Administrations National Geophysical Data Center ( NOAAs NGDC http://www . ngdc . noaa . gov/ngdc . html ) and the saline and sodic soils data from the European Soils database 23 ., The pointDistance function in R 2 . 10 . 0 and the package raster were used to calculate the Great-circle distance to the shoreline or saline/sodic areas ., We created a variable ( toSeaSal ) representing the shortest distance from the target accession to the shoreline/coast or saline/sodic area ., The accession coordinates , distance to sea , distance to saline environment and SNP genotype at Chr4:6392276 can all be found in Dataset S1 ., The method used to collect accessions and assemble the population might introduce unintended confounding effects that violate the assumption of independent locations used by our models ., To determine whether the locations of the accessions were spatially dependent we performed a Mantel test 35 on the distances from the 300 accessions to the coast or known saline/sodic areas ., The simulated p-values of 50 permutations tests with 999 repeatedly simulated samples are 0 . 996 , indicating that an assumption of independency for the response variable toSeaSal is acceptable ., To test for associations between leaf Na+ ( Na ) , genotype at the highest scoring SNP ( C or T at Chr4:6392276\u200a=\u200aisT ) , and the distance to the nearest coast or saline/sodic area ( toSeaSal ) , we used the package lm in R 2 . 10 . 0 to fit linear models , with the weights determined by the following approach ., First , to quantify the strength of the relationship between toSeaSal and the leaf sodium Na , we fit the data into a linear model and regressed toSeaSal on Na ., ( 2 ) Second , we applied a regression approach to single-factor analysis 36 between toSeaSal and isT and tested if the average distance to coast or saline/sodic areas of samples having the high Na T allele is significantly different from the average of samples having the C allele ., ( 3 ) Finally , we regressed toSeaSal on the interaction between Na and isT to inspect how the two predictors jointly affect the distance to sea or saline/sodic ., ( 4 ) To perform the significance tests on the linear coefficients , Na should be centered at the mean 36 ., The extent of variation of distances to saline environments changes with both leaf Na+ concentrations and genotypes ( Figure S5 ) ., Therefore , all three models account for this heterogeneity of variation , and parameters of the models are fitted using weighted least squares ., The variances of the error terms in equation 2 , 3 , and 4 are not constant , and are related to the predictors according to the diagnosis on the model residuals ., The models were fit using iterative weighted least squares 36 ., In addition to the parametric test ( model 3 ) , we performed a non-parametric test ( Wilcoxon rank-sum test or Wilcoxon-Mann-Whitney test 37 ) using the wilcox . test function in R package stats , to assess whether toSeaSal is higher in the lines with the T allele than those with the C allele at Chr4:6392276 ., The p-value of the Wilcoxon rank-sum test is 0 . 006224 indicating that both the parametric and non-parametric approaches reach the same conclusion ., The statistical output of all models can be found in Text S1 . | Introduction, Results/Discussion, Materials and Methods | The genetic model plant Arabidopsis thaliana , like many plant species , experiences a range of edaphic conditions across its natural habitat ., Such heterogeneity may drive local adaptation , though the molecular genetic basis remains elusive ., Here , we describe a study in which we used genome-wide association mapping , genetic complementation , and gene expression studies to identify cis-regulatory expression level polymorphisms at the AtHKT1;1 locus , encoding a known sodium ( Na+ ) transporter , as being a major factor controlling natural variation in leaf Na+ accumulation capacity across the global A . thaliana population ., A weak allele of AtHKT1;1 that drives elevated leaf Na+ in this population has been previously linked to elevated salinity tolerance ., Inspection of the geographical distribution of this allele revealed its significant enrichment in populations associated with the coast and saline soils in Europe ., The fixation of this weak AtHKT1;1 allele in these populations is genetic evidence supporting local adaptation to these potentially saline impacted environments . | The unusual geographical distribution of certain animal and plant species has provided puzzling questions to the scientific community regarding the interrelationship of evolutionary and geographic histories for generations ., With DNA sequencing , such puzzles have now extended to the geographical distribution of genetic variation within a species ., Here , we explain one such puzzle in the European population of Arabidopsis thaliana , where we find that a version of a gene encoding for a sodium-transporter with reduced function is almost uniquely found in populations of this plant growing close to the coast or on known saline soils ., This version of the gene has previously been linked with elevated salinity tolerance , and its unusual distribution in populations of plants growing in coastal regions and on saline soils suggests that it is playing a role in adapting these plants to the elevated salinity of their local environment . | genetics and genomics/functional genomics, plant biology/plant-environment interactions, evolutionary biology/plant genetics and gene expression, ecology/spatial and landscape ecology, plant biology/plant genetics and gene expression, genetics and genomics/population genetics | null |
journal.pcbi.1004951 | 2,016 | Modelling Virus and Antibody Dynamics during Dengue Virus Infection Suggests a Role for Antibody in Virus Clearance | In contrast to malaria , dengue is a vector-borne infection with a growing geographical range , which is therefore responsible for an increasing burden of disease 1 ., Much remains to be understood about the epidemiology and pathogenesis of infection , notably how infection with one serotype modifies viral replication and disease in a later infection with a different serotype ., Multiple studies have examined the role of antibody in enhancing infection 2 , antigenic sin in T or B cells 2 , 3 and protection afforded against infection or disease 4 ., However , only a limited amount of past work has examined how the kinetics of the antibody response interact with the dynamics of viral replication within the infected patient , and investigated the causes of viral clearance during infection ., Previous viral dynamic modelling work for dengue has fit mechanistic models of various immune responses to viral titres 5 , 6 ., Here we extend this work to fit to both viral and antibody titres during infection ., One previous study 7 analysed a small number of serial antibody measurements from primary dengue infections to examine whether antibody titres , along with NS1 measurements , could be used as an alternative diagnostic method for detecting infection ., The study showed that IgM antibodies were detectable in 43% of cases on day 3 of symptoms , though in some individuals they were detectable from day 1 and were detectable in 100% of individuals by day 8 ., Some individuals also had detectable IgG antibodies by day 8 ., Though generally only two measurements were available per patient , the study highlighted high levels of heterogeneity between patients in antibody responses ., These results echo what was seen in an older study 8 which showed that in primary infection IgM antibody developed more quickly and to higher levels than IgG , but that the reverse was true in secondary infection ., This work led to the use of the ratio of IgG vs . IgM titres to classify primary and secondary infection ., IgM was also noted to become detectable at around the same time point as virus became undetectable , but since the main focus of the work was the use of antibody titre measurements as a diagnostic tool , mechanistic explanations of antibody and virus dynamics were not considered ., Zompi and colleagues 9 considered the kinetics of antibody and B cell populations during acute secondary DENV3 infection in Nicaragua ., Early in infection they found that the majority of antibody was cross- reactive with more antibody directed towards DENV2 than DENV3 ., Most recently , a study of Mexican patients compared ( at a single time-point ) viral titres in patients with or without detectable IgM 10 ., Lower virus titres were observed in individuals with detectable IgM ., There are two mechanisms by which dengue infection can be controlled: limiting the rate of production of new virus particles ( by blocking virus entering the cell or preventing the cell from releasing virus ) or increasing the clearance of infected cells or virus ( neutralisation or opsonisation and clearance ) ., Antibody can play a role in the clearance of virus through neutralisation 11 and in the clearance of infected cells through antibody dependent cell cytotoxicity ( ADCC ) 12 ., In this paper , we explore whether sequential antibody and virus measurements from a closely observed set of Vietnamese dengue patients are temporally and mechanistically consistent with either or both of these mechanisms for antibody action ., Virus and antibody titres were measured throughout DENV1 and 2 infections ( Figs 1 and 2 ) ., A summary of characteristics of the dataset is given in Table 1 ., The levels of IgG titres in patients with primary infection were too low during infection for IgG to play a role in viral clearance ., We therefore fitted the IgG data only to data from patients with secondary infections and the IgM titres to both primary and secondary infections ., Since measurements only started after patients sought healthcare ( and therefore after symptoms had started ) , data are only typically available from around the time of peak RNA titres ., A peak in RNA titre ( defined as an observed increase in titre relative to the first measurement , followed by a decline ) was observed in 12 out of the 32 DENV1 patients and 7 of 21 DENV2 patients ., Subject 15 ( marked in black in Fig 1 ) was an outlier in having very low peak RNA titres and therefore we excluded this patient from the model fitting ., We explored models of virus replication and immune control with two extreme cases for the action of antibody: direct neutralisation of free virus , and killing of infected cells ( e . g . via ADCC ) ., We found that either assumption was able to fit the data well , pointing towards a dominant role for antibody in shaping DENV RNAemia dynamics , in particular IgM ., Though both models fit qualitatively well , the fit of the virus neutralisation model was statistically significantly better ( judged by the log likelihood difference ) than the ADCC model ( Tables 2 and 3 ) ., This model fit better for 24 out of 31 individuals ., We also see in comparing the fit of the antibody neutralization model to the virus and IgM antibody titres ( Figs 3 and 4 ) with the model fit of the ADCC model ( Figs 5 and 6 ) , that the first model captures the magnitude and timing of the early viral titres better than the second ., Parameter estimates for both model variants are given in Tables 2 and 3 . The model fits for both models to the virus and IgG titres are shown in the Figs A-D in S1 Text with the parameters in Tables A-B in S1 Text ., The scaling factor , SF , relating ELISA measured antibody levels to actual effective antibody titres , was fitted independently for each patient; this is equivalent to assuming that a specific density of antibody has differing effectiveness in clearing virus or infected cell clearance across subjects ., However , the estimated differences between individuals in the value of SF were not large , and it is possible to fit the data reasonably ( though less well , with more predicted target cell limitation ) assuming this parameter takes the same value for all individuals ( see Figs E and F in S1 Text ) ., In this paper , we used dynamical modelling to show that the measured titres of antibody and virus throughout dengue infection are consistent with antibody playing a dominant role in shaping virus dynamics ., Antibody kinetics as measured by IgG and IgM ELISA were able to explain infection dynamics and clearance in secondary dengue cases , while only IgM kinetics were able to for primary cases ., That only IgM can explain the clearance in primary cases points towards a clear role for IgM in RNA clearance ., The strength of this modelling approach is that we can take into account the feedback processes between viral kinetics and the immune response; the immune response is stimulated by the virus and then acts to control viral replication ., We found that the viral and antibody data are consistent with models in which antibody acts on either the virus or infected cells ., This is consistent with previous modelling work that suggested that models of target cell limitation was not able to explain viral dynamics 6 ., The fit was better for the model which assumed antibody directly neutralises free virus ., However the estimated infected cell lifespan was a third of a day for this model variant , a low value compared with other viral infections ( e . g . HIV 14 ) ., Such a short lifespan might suggest an additional important role for dengue infection lysing cells , ADCC or T cells in clearing infected cells later in infection ( i . e . from the peak of RNA titre on ) or for other immune actions still to be understood; unfortunately we do not have data on RNA or antibody titres prior to the onset of symptoms , or on measures of T cell activation ., The small but significant differences in estimates of the SF parameter ( effectively antibody efficacy ) between individuals seen in the best fit model could reflect limitations of the ELISA assay , which captures all anti-dengue antibody ., This crude measure of all anti-dengue antibody will most likely include multiple different levels of responses ( with different epitope-specific affinities ) to each serotype ., The efficacy of this response will depend on the previous infecting serotype , and how long ago this infection occurred , so will likely vary between individuals ., Further work with serotype specific neutralizing titres or epitope specific measures will be of interest here ., We find that even relatively low-levels of antibody ( measured by ELISA ) were able to begin to control infection , possibly suggesting that the immune response substantially overshoots ( in terms of antibody levels attained ) compared with the minimum response required for control ., Our ability to quantify the relationship between antibody and RNA titres is limited by the fact that ELISA assays give results on a linear scale , while RNA titre measurements ( quantified via PCR ) have a dynamic range of 5 or 6 orders of magnitude , given their measurement errors are on a logarithmic scale ., Our analysis suggests antibodies start to control dengue replication at concentrations below the lower limit of quantification of the ELISA assay ., Use of antibody dilution assays would therefore provide better resolution of the detailed relationship between virus and antibody kinetics and would therefore allow us to explore more rigorously whether antibody trends are consistent with antibody playing the dominant role in viral control , or whether another aspect of the immune response ( e . g . the innate response ) must also be playing an important role ., In addition , this ‘RNAemia’ as measured by RT-PCR is an imperfect proxy of infectious virus titre , and the relationship between titre and infectiousness may well break down in the latter stages of infection- close to defervescence ., Non-infectious ( e . g . because it is bound to neutralising Ab ) virus will nonetheless continue to give a signal in the PCR assay ., Measures of infectivity of individuals throughout and particularly in the latter stages of infection will be of use to clarify the magnitude of this effect ., Previous early work by Innis et al . 8 also considered IgG and IgM dynamics during infection ., We adopted the criterion proposed by that work in classifying patients without quantifiable ( <10 ) IgG antibodies by the end of infection as primary infections and the remainder as secondary ., For secondary infections , Innis et al noted IgG developing more quickly than IgM ., In our data , however , we observe a range of IgG and IgM kinetics for secondary cases ( Figs 1 and 2 ) ., Though IgG reaches high levels ultimately in all secondary cases , in some individuals IgG and IgM growth is concurrent , or IgM actually develops more quickly than IgG ., This individual heterogeneity is consistent with observations in a recent work by Hu et al . 7 ., It implies that primary/secondary classification using the IgG to IgM ratio might be highly sensitive to the timings of the measurements used ., In incorporating a single monolithic immunity variable and the one to one relationship of clearance , our model makes highly simplifying assumptions about the development and binding of the immune response to dengue virus ., In reality there are probably multiple arms of the immune response contributing to the control of viral replication and a more complex binding process occurring 15 ., For example , in addition to B-cell mediated responses considered here , there is evidence from mice that the innate immune response may assist in viral clearance and that T-cells may be important 16 ., Our model currently predicts some role for target cell depletion in infection dynamics , which may be the result of our model fits adjusting to cope with the absence of data on other parts of the immune response ., Measurements of anti-viral innate immune responses ( such as nonspecific Type I interferon activation ) and T cell response dynamics throughout infection , paired with virus titres , will therefore be informative in disentangling which arms of the immune response play the dominant role at which stage of pathogenesis ., To understand the antibody response further , multiple antigen-specific antibody measurements ( possibly coupled with measurements of the capacity of sera to neutralise/enhance ) would be highly informative ., It would be particularly valuable to obtain such data ( from human challenge models or otherwise ) from early in infection , as currently we have little data on the early growth kinetics of virus or the immune response ., An understanding of how the dynamics of virus replication and the immune response interact during infection gives insight into pathogenesis and how disease course might be modulated ., In this paper , the fit of mathematical models of immune system and viral dynamics to dengue patient data , sheds light on this key relationship ., We have presented the first study which quantitatively and mechanistically links measured dengue virus and antibody dynamics throughout infection ., We found a mathematical model of dengue antibody playing a role in controlling infection was consistent with the RNA and antibody titres throughout dengue infection ., The trial protocol was approved by Oxford University Tropical Research Ethical Committee and the Scientific and Ethical Committee of the Ministry of Health , Vietnam ., The trial was registered at http://www . clinicaltrials . gov ( NCT01096576 ) ., We use RNA titre data presented in a prior publication of a clinical trial of the drug balapavir to treat dengue infection 13 ., Informed consent was obtained from the study participants as described in 13 ., That study saw no differences between treatment arms , so both were combined and used here ., Patients were enrolled within 48 hours of fever onset ., The trial had 32 subjects with DENV1 infection and 21 with DENV2 infection ., All patients had twice daily viral load measurements 13 ., Antibody titre measurements were also measured throughout infection in this study , and these data are presented for the first time here ., IgG and IgM antibodies were measured using an ELISA assay 17–19 with quantitation via measurement of optical density ., The ELISA assay does not measure antibody to a specific epitope , but overall binding of the antibody to virus ., Measurements are thought to be linearly proportional to total binding levels below 25 optical density units , but above 25 the relationship becomes non-linear as optical density measurements saturate at high levels of binding ., We excluded one patient ( patient 15 in Fig 1 ) from the analysis due to the outlier virus and immune titres seen ., Using the antibody titre measurements , individuals could be classed as primary or secondary infections 8 ., We categorised patients as primary infections if they had IgG titres less than 10 in the specimen collected at the time of patient discharge from hospital and as secondary ( or later ) infections otherwise ., Using this algorithm , 5 of the DENV-1 cases were classified as primary and the remainder as secondary infections ., All DENV-2 cases were classed as secondary infections ., We extended an existing mathematical model of dengue virus and immune dynamics within a host 5 to explore the extent to which antibody kinetics are consistent with a key role for antibody in limiting dengue infection ., In this model ( similar to those used for influenza 20 , 21 ) , the target cells ( x ) and free virus, ( v ) interact to infected cells, ( y ) , which can then go onto produce more virus ., Whilst this is occurring , antibody levels, ( z ) are increasing with the aim of halting infection ( and in future providing protection against a subsequent infection ) ., The following equations define the model ., Parameters of the model and their meaning are given in Table 4 ., We fit two model variants representing different mechanisms of antibody action: Model 1: antibody acting to kill infected cells e . g . via antibody dependent cell cytotoxicity ( ADCC ) , and Model 2: antibody neutralising and clearing the virus ., We model antibody acting to kill infected cells by assigning ε = 0 and α > 0 and virus neutralization and clearance by assigning α = 0 and ε > 0 . Key to modelling the interaction between viral and immune system dynamics is how the different parts of the immune response proliferate in response to infection , represented in the model by the function f ( y , v ,, z ) ., In preliminary fitting we saw that the mass action formulation 5 , f ( y , v ,, z ) = ηyz was unable to fit the IgG and IgM data ., Hence we used a more realistic saturating function of infected cell or virus density: f ( y , v ,, z ) = ( η1 y z / ( η2 +, y ) ) ( infected cell killing model ) or f ( y , v ,, z ) = ( η1 v z / ( η2 +, v ) ) ( virus neutralisation model ) , respectively ., This functional form implicitly incorporates the processes of B cell maturation and antibody production , and similar forms have been used to model immune cell proliferation to viral infection in previous work 22 ., We simultaneously fitted both viral titres and antibody levels ., In order to assess which antibody measures best explained viral dynamics , we fitted the models separately to IgG and IgM data , fitting the same viral titre data in each case ., Since model ( 1 ) allows for both target cell-limited and immune control of the virus , how it fits the data will shed light on the mechanism driving infection control ., If the antibody and the virus dynamics are consistent with a model in which antibody is controlling virus , we would expect to find a minimal role for target cell limitation 23 ., The models were fitted using Markov Chain Monte Carlo ( MCMC ) methods , implemented in R 24 and C , as described in previous work 5 ., For each subject we estimate the length of the incubation period ( i . e . the time from infection to symptom onset ) using the reported day of symptom onset and an informative prior distribution for the incubation period ., Informed by early human challenge studies 25 , the prior distribution used was a left-truncated normal with mean of 5 . 7 days and standard deviation of 3 days ., Antibody measurements were also included in the model likelihood , with the upper limit of reliable quantification modelled using the cumulative distribution function ., The likelihood for a single subject is:, ∏i=1nϕ ( logDiv|log, ( vi ) , σv2 ) 1−civΦ ( logLDv|log, ( vi ) , σv2 ) civϕ ( Dia|SFzi , σa2 ) 1−cai ( 1−Φ ( LDa|SFzi , σa2 ) cai ), Here ϕ and Φ are the probability density function and cumulative density function of the normal distribution , respectively ., The number of observations for a single individual is denoted by n , Div is the ith viral titre measurement and vi is the model prediction of viral titre at the ith measurement ., LDv is the limit of detection of viral titre and σv is the error of viral titre measurements on a log10 scale , assumed to be 1 . The indicator function , civ is 0 if Div > LDv and 1 if not ., In addition , Dia is the ith antibody level measurement , zi is the model prediction of antibody level at the ith measurement and SF is a scaling factor for antibody measurements ( discussed below ) ., LDa is the upper limit of reliably quantification of antibody levels and σa is the error in antibody measurements assumed to be 1 . The indicator function , cai = 1 if the ith antibody level measurement is above the limit of reliable quantification ( LDa ) , and 0 if not ., Since the optical density based measurements of antibody levels obtained via the ELISA assay do not provide a direct measurement of antibody density per ml of plasma , we introduce a fitted multiplicative scaling factor , SF , to transform the state variable z , which represents antibody densities in the model , into the same scale as the antibody level measurements ., The full likelihood is given by the product of the likelihood over all patients and we use the natural logarithm of the likelihood ( log-likelihood ) as a measure of goodness of fit ., We fit some model parameters as patient-specific and others as group-specific , with the groups here being defined by the infecting serotype ( DENV1 or DENV2 ) ., Parameters relating to the host immune response ( z0 , η1 , η2 and SF ) were treated as patient-specific , whilst virus parameters ( β and κ and δ ) were assumed to be the same for all subjects ., The assignment α = 1 or ε = 1 and does not affect model results; α and z0 are unable to be estimated simultaneously since the antibody level measurements available are relative , not absolute measures of antibody density ., We estimated κ and fixed β and x0 to values from previous work; higher for secondary than primary cases , however the results are not sensitive to these values , and with different values for each of the model formulations ( required for each model to reach the peak titers ) 5 ., Table 4 gives a complete list of model parameters and definitions . | Introduction, Results, Discussion, Materials and Methods | Dengue is an infection of increasing global importance , yet uncertainty remains regarding critical aspects of its virology , immunology and epidemiology ., One unanswered question is how infection is controlled and cleared during a dengue infection ., Antibody is thought to play a role , but little past work has examined the kinetics of both virus and antibody during natural infections ., We present data on multiple virus and antibody titres measurements recorded sequentially during infection from 53 Vietnamese dengue patients ., We fit mechanistic mathematical models of the dynamics of viral replication and the host immune response to these data ., These models fit the data well ., The model with antibody removing virus fits the data best , but with a role suggested for ADCC or other infected cell clearance mechanisms ., Our analysis therefore shows that the observed viral and antibody kinetics are consistent with antibody playing a key role in controlling viral replication ., This work gives quantitative insight into the relationship between antibody levels and the efficiency of viral clearance ., It will inform the future development of mechanistic models of how vaccines and antivirals might modify the course of natural dengue infection . | Dengue is a globally important viral disease ., Despite this , there is still much unknown about the immunology , virology and epidemiology of dengue ., As for all viral infections , the interaction between virus and immune response is a complex one ., Using data collected from patients , we model how the virus replicates in an infected individual and how the human antibody response acts to control that replication ., We show that the timing and magnitude of the growth and decline of virus and antibody levels in dengue-infected patients are consistent with antibody playing a key role in controlling infection ., Our results are of use in the evaluation of potential antiviral drugs and vaccines . | cell death, dengue virus, medicine and health sciences, immune physiology, enzyme-linked immunoassays, pathology and laboratory medicine, pathogens, immunology, cell processes, microbiology, mathematical models, viruses, rna viruses, antibodies, immunologic techniques, antibody response, research and analysis methods, immune system proteins, proteins, medical microbiology, mathematical and statistical techniques, immunoassays, microbial pathogens, viral replication, immune response, biochemistry, cell biology, flaviviruses, virology, physiology, viral pathogens, biology and life sciences, organisms | null |
journal.pcbi.1005647 | 2,017 | MrTADFinder: A network modularity based approach to identify topologically associating domains in multiple resolutions | The packing of a linear eukaryotic genome within a cell nucleus is dense and highly organized ., Understanding the role of 3D genome in gene regulation is a major area of research 1234 ., Recently , genome-wide proximity ligation based assays such as Hi-C have provided insights into the complex structure by revealing various structural features regarding how a genome is organized 567 ., Perhaps , one of the most important discoveries is the domain of self-interacting chromatin called topologically associating domain ( TAD ) 89 ., Inside a TAD , genomic loci interact often; but between TADs , interactions are less frequent ., Thus the TAD emerges as a fundamental structural unit of chromatin organization; it plays a significant role in mediating enhancer-promoter contacts and thus gene expression , and breaking or disruption of TADs can lead to diseases like cancers 101112 ., Therefore , a deeper understanding of TADs from Hi-C data presents an important computational problem ., Results of a typical Hi-C experiment are usually summarized by a so-called chromosomal contact map 5 ., By binning the genome into equally sized bins , the contact map is essentially a matrix whose element ( i , j ) reflects the population-averaged co-location frequencies of genomic loci originated from bins i and j ., In this representation , TADs are displayed as blocks along the diagonal of a contact map 89 ., Despite the fact that TADs are rather eye-catching in a contact map , computational identification is still challenging because of experimental factors such as noise and inadequate coverage ., Moreover , it is apparent from a visual examination of the contact map that TADs exhibit various length scales: there are TADs that appear to be overlapping , and within many TADs , there are rich sub-structures ., Mathematically speaking , it is very natural to transform a contact matrix to a weighted network in which nodes are the genomic loci ( or bins ) whereas the interaction between two loci is quantified by a weighted edge ., In network science , a widely studied problem is the identification of network modules , also known as community detection problem 13 ., A module refers to a set of nodes that are densely connected ., In its simplest form , the community detection problem concerns with whether nodes of a given network can be divided into groups such that connections within groups are relatively dense while those between groups are sparse ., Therefore , by viewing the chromatin interactions as a network , the highly spatially localized TADs immediately resemble densely connected modules ., Motivated by the resemblance , we formulate the identification of TADs as a global optimization problem based on the observational contact map and a background model ., As a network-based approach , our method goes beyond a direct adaptation of standard community detection algorithms ., We introduce a novel background model that takes into account the effect of genomic distance , which is specific to the context of genome organization ., The objective function is optimized using a heuristic algorithm that is efficient even if the size of the input contact map is large ., Furthermore , by introducing a tuning parameter , our network approach can identify TADs at different resolutions ., At a low resolution , larger TADs are found whereas , at a high resolution , smaller TADs are identified as the nucleome is viewed on a finer scale ., In other words , the method can identify TADs at different length scales ., We name our method MrTADFinder where the acronym Mr stands for multiple resolutions ., The identification of modules in a network is formulated as a global optimization problem on the so-called modularity function over possible divisions of the network ., Consider an unweighted network represented by an adjacency matrix A . For a particular division ( i . e . a mapping from the set of all nodes to a set of modules ) , the modularity is defined as the fraction of edges within modules minus the expected fraction of such edges in a randomized null model of the network ., Mathematically , the modularity is equal to, 12m∑i , j ( Aij−kikj2m ) δσiσj ., ( 1 ), Here , the summation goes over all possible pairs of nodes , the value of the Kronecker data δσiσj equals one if nodes i and j have the same label σ and zero otherwise , meaning only pairs of nodes within the same module are summed ., In particular , m is the number of edges in the network whereas the expression kikj/2m represents the expected number of edges between i and j in a so-called configuration model ., The configuration model is a randomized null model in which the degrees of nodes ki are fixed to match those of the observed network , but edges are in other respects placed at random ., High values of the modularity correspond to good partitions of a network into modules and similarly low values to bad partitions ., Optimizing the modularity function leads us to the best partition over all possible partitions ., More recently , a so-called resolution parameter γ has been incorporated in Eq ( 1 ) to adjust the size of the resultant modules 14 ., Following the network formalism , given a Hi-C contact map represented by a weighted matrix W , we define a similar objective function Q as, Q=12N∑i , j ( Wij−γEij ) δσiσj ., ( 2 ), Here , i , j index the equally binned genomic loci ., N is the total number of pair-end reads ., Eij is the expected number of contacts between locus i and locus, j . γ is the resolution parameter that could be used to tune the size of resultant TADs ., Very much similar to the network setting , the identification of TADs aims to partition the loci into domains such that Q is optimized ., Nevertheless , it is important to emphasize two points ., First , unlike the case in a network , the bins in a chromosome form a continuous chain and therefore genomic loci belonging to a TAD have to form a continuous segment ., Second , simply because of the physical nature of chromosome , the expected number of contacts between locus i and locus j depends on their genomic distance ., Two loci that are close together in a 1-dimensional sense are expected to have a higher contact frequency as compared to two loci that are far apart ., This point suggests that the null model Eij in Eq ( 2 ) has to be modified ., Thus , given an intra-chromosomal contact map W , the expected null model E is defined as, Eij=κi*κj*f ( |i−j| ) ., ( 3 ), Here , f is the average number of contacts as a function of distance d = |i − j| ., By considering all possible pairs of bins in W in terms of their distance apart and the contact frequency , we estimate f by local smoothing ( see Methods ) ., For intermediate values of d , f follows pretty well with a power-law function d−1 ( see S1 Fig ) , which is a well-known observation first reported in 5 ., As a null model , the resultant E matrix satisfies a set of constraints , namely, ∑jEij=∑jWij=ci∀i ,, ∑ijEij=∑ijWij=2N ., ( 4 ), The first equation means that the coverage ci ,, i . e ., the total number of reads ( one end of pair-end reads ) mapped to bin i , defined in the observed map is the same as the coverage defined in the null model ., The second equation is a direct consequence of the first equation , where N is the total number of pair-end reads mapped to the chromosome ., As f has been estimated from the observed W , we can numerically solve all the unknowns κi* in the system of matrix equations ( see Methods ) ., Mathematically , κi* can be regarded as an effective coverage because of the correlation between κi* and the coverage ci is extremely high ( r = 0 . 95 , S2 Fig ) ., In comparison with Eq ( 1 ) , κi* is conceptually analogous to the degree ki ., As shown in Fig 1 , given a particular matrix W , the contact frequency of the resultant null model E are the highest in the diagonal and decrease gradually away from the diagonal ., With W and E , for any given resolution parameter γ , we employ a modified Louvain algorithm to optimize Q ( see Methods and Fig 1 for details ) ., To ensure robustness , multiple runs of the modified Louvain algorithm are performed , and a boundary score is defined as the fraction of times a bin is called as a boundary ., The final set of TADs is defined based on the set of consensus boundaries ( Fig 1 and Methods ) ., It is important to emphasize that the conventional Louvain algorithm used in network analysis 15 cannot be directly used because chromatin domains are continuous segments ., As a demonstration , we applied MrTADFinder to analyze Hi-C data of hES cell from 8 ., Fig 2A shows a particular snapshot of the contact map ( for chromosome 10 ) and its alignment with the identified TADs ., In general , the TADs displayed agree well with the apparent block structures in the contact map ., Of particular interest is the choice of γ that capture various length scales in domain organization ., As shown in Fig 2A , when γ increases , a large TAD breaks into a few small TADs ., On the other hand , a few large TADs merge together to form an even larger TAD as the value of γ is lowered ., Statistically speaking , γ quantifies to what extent do we accept the enrichment of empirical contact frequency over the expectation ., As γ increases , only matrix elements close to the diagonal contribute positively to the objective function ., Therefore , in general , the size of TADs decreases ( see Fig 2B ) and the number of TADs increases ( see Fig 2C ) ., For example , when γ = 1 . 0 , there are about 1000 TADs in hES cells with a median size of 3Mb ., When γ = 2 . 25 , the number of TADs increases to 2600 and the median size is roughly 1Mb ., We then further compared the TADs identified at different resolutions by MrTADFinder with TADs identified by a previous method ., As quantified by the normalized mutual information ( see Methods for details ) , TADs identified by MrTADFinder best match with TADs identified in 8 when the resolution parameter is 2 . 9 ., In general , unless the resolution is sufficiently small ( γ < 1 . 5 ) , the two methods are quite consistent ( see Fig 2C ) ., Nevertheless , the introduction of the resolution parameter γ opens an extra dimension in domain identification in a sense the algorithm used in 8 focuses on a particular resolution instead ., The interplay between 3D genome organization and various chromatin features has widely been investigated since some of the first Hi-C experiments were reported 589 ., Nevertheless , there is no clear-cut pattern emerges by aligning a variety of chromatin features with TADs ( S3 Fig ) , even though the occurrence of sharp peaks at the boundaries is quite apparent ., By identifying TADs and their boundaries using MrTADFinder , we found the boundary signatures that are consistent with the observations previously reported 8 , for instance , the enrichment of active promoter mark H3K4me3 or active enhancer mark H3K27ac , as well as the depletion of transcriptional repression mark like H3K9me3 ( Fig 3A and S4 Fig ) ., To better understand the relationship between domains organization and different chromatin features , we further examined the chromatin features near different sets of boundaries that were identified in different resolutions ., We found that in general , the enrichment of peak density at boundary decreases as resolution increases ., This is because the number of TADs increases as the resolution increases , various chromatin features appear in the boundaries of low-resolution TADs do not appear in high-resolution TADs ( Fig 3A ) ., More specifically , the enrichment of histone marks like H3K36me3 and H3K4me3 exhibits a monotonic drop whereas certain marks exhibit characteristic resolutions ., For instance , the enrichment of mark H3K27me3 remains high up to a resolution of γ = 2 . 5 ( Fig 3B ) ., The observation suggests that the mark H3K27me3 in general marks the boundary of TADs up to a particular resolution ( length scale ) ., Beside epigenetic signatures , we examined the distribution of protein-coding genes along chromosomes in relation to TAD boundaries formation ., Though the starting positions of genes tend to be enriched near TAD boundaries , the enrichment is much stronger for housekeeping genes as compared to tissue-specific genes ( Fig 4A ) ., As housekeeping genes are essentially active , the pattern resembles the active promoter mark H3K4me3 shown in Fig 3B ., The discrepancy between housekeeping genes and tissue-specific genes was firstly reported in Ref ., 8 ., Nevertheless , by extending the idea to multiple resolutions , we found that the distribution of housekeeping genes follows a different length scale compared to tissue-specific genes ., As shown in Fig 4B , housekeeping genes in general marks the boundary of TADs up to the resolution γ = 1 . 5 ., Apart from histone modifications , it is well known that certain transcription factor binding sites are enriched near the boundary regions of TADs 8 ., Instead of looking at individual factors , we further explored the location of the so-called HOT regions and XOT regions on TADs ., High-occupancy target ( HOT ) regions and extreme-occupancy target ( XOT ) regions are genomic regions that are bound by an extensive amount of transcription factors 16 ., As expected , we found a strong enrichment of HOT regions and an even stronger enrichment of XOT regions near TAD boundaries in hES cells ( Fig 5A ) ., The observation is , in general , true for all tested resolutions ., The observation agrees with the idea that HOT regions are very accessible regions in open chromatin ., Nevertheless , it is still widely unknown if transcription factors bind to HOT regions simply because of thermodynamics , or the binding will result in important biological consequences ., Motivated by the observation that many factors tend to bind to the boundary regions , we further examine which factors are responsible for establishing the domain border , and more interestingly for borders in different resolutions ., There are a few proteins which are widely known to be important in border establishment 17; nevertheless , it is worthwhile to perform a systematic analysis ., To do so , we formulated a classification problem which aims to distinguish , for each resolution , a set of boundaries identified by MrTADFinder ( positive set ) from a set of random boundaries obtained by swapping the TADs along the chromosomes ( negative set ) ., Using a logistic regression model recently proposed by 18 , we integrated the binding signals of 60 transcription factors at a genomic locus to predict if it is TAD boundary ( see Fig 5B and Methods for details ) ., Generally speaking , with 10-fold cross validation , the model is quite successful in low resolutions ( AUC = 0 . 81 , S5 Fig ) ., The result is consistent with an early work based on histone modifications 19 ., Being consistent with the trend that chromatin features are less enriched at the boundaries of high resolution TADs , the predicting power of the model decreases as the resolution increases ., The regression model further quantifies explicitly the influence of each of the transcription factors ., In general , factors that are responsible for border formation are quite consistent across different resolutions ( Fig 5B ) ., For instance , we found that the well-known insulator CTCF , and Rad21 that is a part of cohesin , are direct key components of border establishment ., In addition , the chromatin remodeler Chd7 , which is often found at enhancers 20 , is predicted to be a key component ., On the other hand , factors like MYC have a consistently negative effect ., Nevertheless , the relative importance of factors does change with resolutions ., For instance , Rad21 has a higher predictive power in classifying high-resolution domains in compared with classifying low-resolution domains ., The contact maps of more deeply sequenced Hi-C experiments have exhibited a pattern that a large fraction of TADs has “peaks” in their corner 21 , meaning the contact frequency between the endpoints of such domains is higher than those of their surrounding neighborhood ., The configuration suggests that the boundaries of such domains form a chromatin loop ., We investigated if a similar conclusion could be drawn from the TADs called by MrTADFinder using a set of significant long-range promoter contacts identified by capture Hi-C 22 ., Based on the Hi-C data of GM12878 in 21 , we found that there are indeed potential promoter-enhancer linkages connecting the endpoints of domains ., Moreover , by increasing the resolution parameters , the boundaries of the smaller TADs further capture the potential promoter-enhancer linkages in shorter length scales ( Fig 6 ) ., It is worthwhile to point out that the linkages connecting the endpoints of domains form a small fraction as compared to the total number of significant interactions identified by capture Hi-C ., Therefore , identifying the domain borders is not a direct method to predict potential enhancer-gene linkages ., On the other hand , though the increase in the number of boundaries can capture a higher number of potential interactions , the same analysis for an ensemble of randomly reshuffled TADs shows the observation in TADs called by MrTADFinder is significant ( Fig 6 ) ., In other words , TADs in a higher resolution are potential subTADs that mediate long-range interactions in a finer length scale 23 ., We have examined the interplay between domains organization and chromatin features ., Recently , it has been reported that epigenomic features shape the mutational landscape of cancer 24 ., Motivated by this linkage , we further investigated the occurrence of somatic mutations near the boundaries ., More specifically , we mapped the somatic mutations obtained from breast cancer samples to the TAD boundaries we identified in MCF7 cells ( see Methods ) ., In a given resolution , there are 85 boundary regions identified on chromosome 10 ., The regions can be clustered into 3 groups based on the positional distribution of somatic mutations ., As in shown in Fig 7 , two of the clusters exhibit a step-function behavior ( blue and red ) in which the abrupt transition essentially happens at the boundary ., For boundary regions in the remaining cluster , the mutational burden exhibits no difference across the TAD boundaries ., Because of the close relationship between TADs and replication-timing domains 25 , the observation resonates with a well-known observation that genomic regions with a high mutational burden are replicated at a later stage during DNA-replication 26 ., As shown in the inset , using Repli-seq data in S1 phase , the upstream regions of the boundaries found in the blue cluster have a high mutation rate but a low Repli-seq signal , meaning they are indeed replicated at a later stage during replication ., On the contrary , the upstream regions of the boundaries found in the red cluster are replicated at an early stage and therefore exhibit a low mutation rate ., Motivated by the relationship between TADs and DNA replication , we overlaid TADs in different resolutions with data from Repli-seq experiment ( S6 Fig ) ., We observed that TADs identified in different resolutions match with the Repli-seq data in different stages of a cell cycle ., For instance , while a TAD identified in a low resolution does not replicate at an early phase , say S1 , its sub-structures identified in a higher resolution correspond to two separate peaks at later stages , say S2 and S3 ( S7 Fig ) ., Nevertheless , it is worthwhile to point out that mapping Hi-C reads from cancer cell lines like MCF7 to the reference genome is not perfect because quite some reads may come from copy number variations ., Computational approaches have recently been developed to perform specific normalization 27 as well as to infer those large scale genomic alterations from Hi-C data 28 ., There are quite a few existing methods on identifying TADs using Hi-C data ., Dixon et al . identified TADs based on the so-called directionality index using Hi-C data in hES cell and found an enrichment of CTCF binding sites at the boundary regions 8 ., Since then the enrichment of chromatin features has been used as a benchmark for various TAD calling algorithms 293031 ., As a comparison , we performed the same analysis using TADs based on MrTADFinder ., As shown in Fig 8 , both methods exhibit a similar pattern ., In fact , as reported in 293031 , the enrichment pattern of CTCF binding peaks is qualitatively the same for all the proposed methods ., By repeating the analysis in different resolutions , we observed that the level of enrichment depends on the resolution ( Fig 8 , S7 Fig ) ., At a low resolution ,, i . e . for larger TADs , the enrichment signal is stronger , and the signal tends to extend over a longer distance from the boundary ., At a higher resolution , the signal is weaker and confined to near the boundary ., In general , Fig 8 suggests that boundaries identified in lower resolutions are more likely to be bound by CTCFs ., From a biological standpoint , as a boundary identified in a lower resolution separates two large domains , the results may bring insights on how to mediate chromatin loops at different length scales via an important architectural protein 3233 ., As the level of CTCF enrichment might be the consequence of different chromatin length scales , it might not be fair to use it directly for benchmarking the performance of different algorithms ., Because of the stochastic nature of the modified Louvain algorithm , we explored the robustness of MrTADFinder ., In the current setting based on multiple runs of the modified Louvain procedure , we found the results of two independent callings highly robust ., In fact , the normalized mutual information is 0 . 99 ( see S8 Fig ) ., We further investigated the reproducibility of MrTADFinder in two aspects ., First , we studied the agreement of TADs called in biological replicates ., Using Hi-C data released by the ENCODE consortium , we found that TADs called in a pair of biological replicates agree reasonably well , with normalized mutual information about 0 . 85 ( see S9 Fig and Methods ) ., Secondly , we explored the effects of sequencing depth to our algorithm ., Specifically , we applied MrTADFinder to identify TADs from a deeply sequenced Hi-C data of GM12878 21 ., We then reduced the number of reads included and called TADs again ., We found that the TADs identified using a subset of reads are slightly different from the original , and in general , the discrepancy increases as fewer reads were used ( S10 Fig and Methods ) ., Despite a certain level of discrepancy , nevertheless , the resultant TADs agree well ., For instance , in the extreme case , by comparing using contact maps constructed from 2 . 4 billion reads and 480 million reads respectively , the mean normalized mutual information of various pairs of chromosomes is about 0 . 88 ., If we compare the TADs called from 2 . 4 billion reads to the TADs called from 1 billion reads , the normalized mutual information is higher than 0 . 95 ., MrTADFinder is implemented in Julia ., Julia programmers can import MrTADFinder as a library for calling various functions ., It can also be run in command line if Julia and the required packages are installed ., The performance of MrTADFinder , in general , depends on the size of the input contact map ., We have tested the performance using the contact maps of GM12878 cell generated by the Aiden lab 21 ., The performance is reasonable ., For instance , for chromosome 10 , in a bin-size of 25kb ( i . e . a contact map 5400 by 5400 ) , the time required to arrive at all TADs with 10 runs of Louvain algorithm is about 20 minutes on a laptop with 2 . 8GHz Intel Core i7 and 16Gb of RAM ., The time required is only 6 minutes if the bin size is 50kb ., We have made the source code available on GitHub ( see software availability ) ., Despite the similarity between Eqs ( 1 ) and ( 2 ) , network modules are rather arbitrary collections of nodes , but domains are continuous segments along the chromosome ., In fact , the total number of possible partitions for a chromosome is much smaller than the total number of ways to divide a network into modules ., As a result , while the optimization of Eq ( 1 ) is an NP-hard problem , the optimization of ( 2 ) can be quite efficiently solved using a dynamic programming inspired method ( see Methods and S11 Fig ) ., It is instructive to explore this avenue because quite some algorithms for identifying TADs are based on a similar approach but with different objective functions 293031 ., Moreover , by enumerating all possible ways to decompose a chromosome into TADs , one could write down the partition function and define a probability of occurrence for each of the possible partition in a statistical mechanics’ manner ., The time complexity of this algorithm is in order of O ( n3 ) , where n is the size of the contact map ., Given the time complexity , finding the optimal partition using a bin size of 40kb is quite impractical ., For instance , the calculation takes about an hour for chromosome 21 , as compared to seconds by using the heuristic ., Therefore , though the connection between identifying TADs and problems like finding RNA secondary structure is of theoretical interest , MrTADFinder is developed based on the modified Louvain algorithm ., Nevertheless , we have implemented the approach based on recurrence relation and performed a comparison with the heuristic ., Using a contact map of hES cell ( chromosome 1 ) with a bin size of 500kb , we found the sub-optimal partitions based on our modified Louvain algorithm are very close to the optimal partition ., The normalized mutual information between optimal and sub-optimal values is 0 . 977±0 . 007 ., In this paper , we have introduced an algorithm to identify TADs from Hi-C data and performed several analyses to show the biological significance of the TADs identified ., In particular , by introducing a single continuous parameter γ , we can further examine domains organization and its interplay with a variety of chromatin features in multiple resolutions ., It is important to emphasize that the idea of resolution we introduced in MrTADFinder is different from some other usages of the same term in Hi-C analysis ., From an experimental standpoint , the resolution of a Hi-C experiment refers to the average fragment size as digested by restriction enzymes ( ~4kb to ~1kb ) 521 or more recently by micrococcal nuclease ( ~150bp ) 34 ., Regarding the construction of contact maps , the term resolution has been used to refer to the bin size , where the proper choice usually depends on the number of reads in the stage of data processing ., Both usages are primarily technical ., What we mean by resolution , however , refers to the multiple length scales built inside the organization of the genome ., It is well known that there are structures in different length scales such as compartment , domains , and sub-domains 35 , and chromatin features like histone marks exhibit multiple length scales 36 ., The concept of resolution introduced here points to the integration of these structures and enables one to explore the rich structures hidden in contact maps ., From a practical point of view , γ = 1 seems to be the natural starting point ., One could increase or decrease the value of γ in order to explore the intrinsic structure ., Nevertheless , because of the different contact maps might have various differences like the read coverage , one should be cautious to directly compare the resolution parameters between different contact maps ., A novel contribution of this work is the derivation of an expected model for any intra-chromosomal contact map by solving a system of matrix equations ., The null model preserves the coverage of each genomic bin as well as the distance dependence of contact frequencies in the observed map ., As such features of contact maps are involved in most computational analysis of Hi-C data , apart from the identification of TADs , the expected model can be used for applications like finding compartments 5 and identifying potential enhancer-target linkages 37 ., Mathematically , the expected matrix is solved by an iterative procedure ., The procedure can be regarded as a generalization of a class of matrix balancing methods used for normalizing Hi-C matrices 38 , as the later is merely a different set of matrix equations ., However , it is important to emphasize that the so-called ICE algorithm aims to remove bias in the contact map , whereas our method aims to generate a background model ., While MrTADFinder focuses on intra-chromosomal interactions , recent studies employ various clustering methods to identify inter-chromosomal clusters using Hi-C contact frequency 3940 ., It is worthwhile to point out that similar expected models used in this study can also be derived for inter-chromosomal interactions to better separate signal and noise ., Several methods have been developed for identifying TADs from Hi-C data 41 ., One of the earliest methods is based on the so-called directionality index , a 1D statistic measuring whether the contacts have an upstream or downstream bias 8 , and later the bias is exploited by the so-called arrowhead algorithm 21 ., Later algorithms exploit the block diagonal nature of TADs in a contact map 29 3042 ., Though some of these algorithms do take the distance dependence into the background , but they do not take into account both the genomic distance and the effects of coverage in a compact mathematical formalism ., The algorithm TADtree 30 , and more recent efforts , namely Matryoshka 31 and metaTAD 43 aim to investigate the hierarchical organization of TADs based on a tree structure ., Indeed , merging smaller TADs at the lower level of the hierarchy results at larger TADs similar to the TADs obtained by MrTADFinder at a low resolution ., Nevertheless , MrTADFinder does not impose a hierarchical organization ., The probabilistic nature of Louvain algorithm enables the definition of TAD boundaries in a probabilistic fashion , and therefore a possibility to define overlapping TADs ., To a certain extent , the idea of continuous resolution used in MrTADFinder is distinct in comparison with algorithms based on a bottom-up approach , but similar in spirit to Ref ., 29 ., MrTADFinder is motivated by the community detection problem in network studies ., Although a network perspective of chromosomal interactions has previously been proposed 4445 , a lot of widely studied concepts in networks have rarely been explored in the context of chromosomal organization ., A network representation is arguably more flexible than a simple matrix representation , for instance , transcription factors binding and histone modifications can be easily incorporated into the network , forming a decorated network ., Moreover , one could extend the framework by concatenating multiple Hi-C contact maps to form a multi-layer network ., The same idea has been used for cross-species transcriptomic analysis 46 ., By facilitating the application of a variety of graph-theoretical tools , we believe that network algorithms will be useful for future studies on the spatial organization of the genome ., The Hi-C data of human ES cells and IMR90 cells were reported in Ref ., 8 ., Raw reads were processed using Hi-C Pro 47 , arriving at contact matrices in various bin sizes ., In all analysis , the whole-genome contact map was iteratively corrected for uniform coverage 38 ., Intra-chromosomal contact maps were then extracted from the whole-genome contact map of bin size 40kb for downstream analysis ., Hi-C data and contact maps in MCF7 cells were reported in Ref ., 48 ., The whole-genome contact map provided was binned with 40kb bin size and was normalized by the ICE algorithm ., Data in GM12878 were reported in 21 ., | Introduction, Results, Discussion, Materials and methods | Genome-wide proximity ligation based assays such as Hi-C have revealed that eukaryotic genomes are organized into structural units called topologically associating domains ( TADs ) ., From a visual examination of the chromosomal contact map , however , it is clear that the organization of the domains is not simple or obvious ., Instead , TADs exhibit various length scales and , in many cases , a nested arrangement ., Here , by exploiting the resemblance between TADs in a chromosomal contact map and densely connected modules in a network , we formulate TAD identification as a network optimization problem and propose an algorithm , MrTADFinder , to identify TADs from intra-chromosomal contact maps ., MrTADFinder is based on the network-science concept of modularity ., A key component of it is deriving an appropriate background model for contacts in a random chain , by numerically solving a set of matrix equations ., The background model preserves the observed coverage of each genomic bin as well as the distance dependence of the contact frequency for any pair of bins exhibited by the empirical map ., Also , by introducing a tunable resolution parameter , MrTADFinder provides a self-consistent approach for identifying TADs at different length scales , hence the acronym Mr standing for Multiple Resolutions ., We then apply MrTADFinder to various Hi-C datasets ., The identified domain boundaries are marked by characteristic signatures in chromatin marks and transcription factors ( TF ) that are consistent with earlier work ., Moreover , by calling TADs at different length scales , we observe that boundary signatures change with resolution , with different chromatin features having different characteristic length scales ., Furthermore , we report an enrichment of HOT ( high-occupancy target ) regions near TAD boundaries and investigate the role of different TFs in determining boundaries at various resolutions ., To further explore the interplay between TADs and epigenetic marks , as tumor mutational burden is known to be coupled to chromatin structure , we examine how somatic mutations are distributed across boundaries and find a clear stepwise pattern ., Overall , MrTADFinder provides a novel computational framework to explore the multi-scale structures in Hi-C contact maps . | The accommodation of the roughly 2m of DNA in the nuclei of mammalian cells results in an intricate structure , in which the topologically associating domains ( TADs ) formed by densely interacting genomic regions emerge as a fundamental structural unit ., Identification of TADs is essential for understanding the role of 3D genome organization in gene regulation ., By viewing the chromosomal contact map as a network , TADs correspond to the densely connected regions in the network ., Motivated by this mapping , we propose a novel method , MrTADFinder , to identify TADs based on the concept of modularity in network science ., Using MrTADFinder , we identify domains at various resolutions , and further explore the interplay between domains and other chromatin features like transcription factors binding and histone modifications at different resolutions ., Overall , MrTADFinder provides a new computational framework to investigate the multiple length scales that are built inside the organization of the genome . | gene regulation, applied mathematics, regulatory proteins, dna-binding proteins, simulation and modeling, algorithms, histone modification, optimization, mutation, chromosome mapping, mathematics, transcription factors, epigenetics, molecular biology techniques, chromatin, research and analysis methods, gene mapping, chromosome biology, proteins, gene expression, chromatin modification, molecular biology, genetic loci, somatic mutation, biochemistry, cell biology, genetics, biology and life sciences, physical sciences | null |
journal.pgen.0030102 | 2,007 | COUP-TFII Mediates Progesterone Regulation of Uterine Implantation by Controlling ER Activity | Establishment of uterine receptivity is mandatory for successful embryo apposition , attachment , and implantation; failure to manifest this uterine state is an underlying cause of most pregnancy failures in women ., A multitude of signaling molecules have been shown to play key roles in the elaboration of this uterine response through mesenchymal–epithelial interaction ., Among numerous factors involved in these primary events of pregnancy , two steroid hormone receptors , progesterone receptor ( PR ) and estrogen receptor ( ER ) , and their cognate ligands , undoubtedly play central roles in this biological process 1–3 ., Although estrogen activity is essential for an integrated uterine response , it has been shown that excessive estrogen activity can prematurely close the implantation window 4 , suggesting that estrogen activity is tightly controlled during the peri-implantation period to allow normal development of the receptive uterus ., Importantly , progesterone is known to attenuate estrogen-induced gene expression in uterine epithelial cells 5 ., Intriguingly , this suppression is mediated by stromal progesterone receptors 6 , 7 , suggesting that the coordinated action of estrogen and progesterone depends on crosstalk between the epithelial and stromal compartments of the uterus ., Although the inhibitory effect of progesterone on epithelial estrogen activity has been described 6 , 7 , the mechanism by which progesterone suppresses estrogen action remains poorly defined ., Lydon et al have shown that female PR-null mice are infertile 8 ., The expression of Indian hedgehog ( Ihh ) , a gene highly expressed in the uterine epithelium , is greatly reduced in these null mutants , indicating that Ihh is a downstream target of the progesterone receptor 9 ., To understand the role of Ihh in reproduction , conditional null mutant mice of Ihh were generated 10 ., These mutants exhibit defects in both implantation and decidualization , indicating that epithelial Ihh regulates the decidual response through Patched/Smoothened ( Ptch/Smo ) signaling in the stroma ., Recently , it has been shown that COUP-TFII ( chicken ovalbumin upstream promoter transcription factor II; also known as NR2F2 ) is a downstream target of Ihh in the uterine tissue 9 , 10 ., COUP-TFII is highly expressed in the uterine stromal compartment , and its expression is significantly reduced in the Ihh mutant mice , suggesting that COUP-TFII might mediate the effects of Ihh signaling in the uterine stroma 10 ., This notion is consistent with our previously findings that COUP-TFII is a downstream target of sonic hedgehog ( Shh ) signaling 9–13 and conditional ablation of COUP-TFII in the foregut mesenchyme resembles Shh-null mutant phenotypes 14 ., COUP-TFII belongs to the orphan nuclear receptor superfamily , and has been well characterized 15 , 16 ., Genetic ablation of COUP-TFII results in early embryonic lethality due to cardiovascular defects 17 ., COUP-TFII heterozygous female mice have shown significant reduced fecundity , which is attributed to ovarian and uterine defects 18 , but the uterine-specific function of COUP-TFII remains largely undefined ., We have recently established a COUP-TFIIflox/flox mouse model in order to generate tissue- or cell-lineage–specific knockouts of COUP-TFII ., Phenotypes exhibited by conditional mutants lacking COUP-TFII in endothelial cells 19 , limbs 20 , stomach 14 , or diaphragm 21 are all consistent with the notion that COUP-TFII might play an important role in reciprocal epithelial–mesenchymal cellular cross-talk ., To define COUP-TFII uterine function , we have ablated COUP-TFII specifically in cell lineages of the uterus that express PR by crossing COUP-TFIIflox/flox with PR-Cre knockin mice 22 ., Ablation of COUP-TFII in the uterine stroma results in decidualization failure , resembling the conditional ablation of Ihh ., In addition , we showed that the expression of bone morphogenetic protein 2 ( BMP2 ) is greatly reduced in the COUP-TFII mutants , and that reintroduction of BMP2 into uterine horn rescues the decidualization defects ., Thus , we established a genetic pathway in which progesterone receptor regulates Ihh , which in turn regulates COUP-TFII through Ptch/Smo signaling , and finally , COUP-TFII regulates BMP2 to confer decidualization in the uterus ., Surprisingly , we also found that ER activity and target gene expression in the uterine epithelial cells are markedly elevated in conditional COUP-TFII knockout mice , which alters the window of receptivity and affects embryo attachment and implantation ., Since stromal PR has been implicated to suppress epithelial ER activity , we further asked whether stromal PR expression is downregulated in the COUP-TFII conditional mutants ., Indeed , PR expression is reduced significantly in the uterine stroma , while no obvious change is seen in the epithelium ., This finding indicates an indispensable role for stromal COUP-TFII in the maintenance of progesterone suppression of uterine epithelial ER activity , a prerequisite for the establishment of normal uterine receptivity ., This study also substantiates the importance of epithelial–stromal cross-communication and sheds new light on a complex signaling circuit that spans the uterine epithelial–stromal divide that is indispensable for the development of the receptive uterus and subsequent decidualization ., In addition , not only does this finding further our understanding of steroid hormonal control of uterine receptivity , but it provides a novel signaling paradigm for steroid hormonal dysregulation shown to underlie such female reproductive pathologies as endometriosis and endometrial , ovarian , and breast cancers ., Conditional knockout mice of COUP-TFII , PRCre/+ COUP-TFIIflox/flox were generated by crossing PR-Cre knockin mice with COUP-TFIIflox/flox mice ., PR is highly expressed in the uterine stroma and epithelium , while COUP-TFII is highly expressed in the stroma , but rarely expressed , if ever , in the uterine epithelium 18 , 22 ., Immunohistochemistry of the reproductive tract indicated that COUP-TFII is efficiently ablated in the stroma of the mutant uterus by the PR-Cre ( Figure 1A and 1B ) ., It is also evident from this figure that COUP-TFII is highly expressed in the stroma compartment but is hardly detectable in the luminal and the glandular epithelia of the uterus ., In contrast , PR is expressed in the granulosa cells , while COUP-TFII is expressed in the theca cells of the ovary 18 , 22 ., Since PR and COUP-TFII are not expressed in the same cell , COUP-TFII is not ablated in the theca cells ., As expected , the expression of COUP-TFII in the theca cells of the ovary is not altered in the conditional COUP-TFII mutants comparison with controls as shown by immunostaining ( Figure 1C–1E ) ., To ensure there is no disruption of ovarian function in COUP-TFII mutants , we transferred ovaries from control and mutant mice to wild-type recipients and observed reproduction for a 6-mo period ., Healthy newborns were yielded from PRCre/+ COUP-TFIIflox/flox ovaries in a similar manner as the controlled PRCre/+ and PRCre/+ COUP-TFIIflox/+ ovaries ( Table 1 ) ., In addition , the litter size from the mutant ovaries was not significantly reduced compared with that of controls , indicating that the PRCre/+ COUP-TFIIflox/flox females have no ovarian defects ( Table 1 ) ., Thus , the implantation failure observed in our conditional mutants is likely due to impaired uterine , but not ovarian , function ., PRCre/+ COUP-TFIIflox/flox mutant mice and COUP-TFIIflox/flox control mice were mated with wild-type males ( B6SJLF1; Taconic ) and observed for 6 mo to compare breeding capacity ., PR-Cre mice were also used as a control to distinguish the contribution of the PR-Cre allele ., Pups were not born from mutant females , while both types of controls gave birth regularly ( Figure 1F ) , indicating that ablation of COUP-TFII in the uterus leads to infertility ., The hormone profile during pregnancy showed no significant difference in estradiol ( control , 45 . 3 ± 3 . 8 pg/ml; mutant , 47 . 9 ± 3 . 2 pg/ml; n = 14 , 3 . 5 d postcoitus 3 . 5 dpc ) and progesterone levels ( control , 15 . 8 ± 4 . 4 ng/ml; mutant , 18 . 6 ± 2 . 9 pg/ml; n = 12 , 3 . 5 dpc ) between mutants and controls , further supporting the fact that PRCre/+ COUP-TFIIflox/flox mice have no obvious ovarian defect as stated above ( Figure 1C–1E; Table 1 ) ., To dissect the cause of infertility , we examined whether embryos properly attach to the uterine lumen , an early event of pregnancy that is initiated at midnight of pregnancy day 4 ( 4 dpc ) ., We dissected mice on the morning of pregnancy day 5 ( 4 . 5 dpc ) and counted the number of implantation sites by injecting Chicago Blue dye ., Implantation sites were not detected in the mutant uterine horns , while normal implantation sites were scored in the controls ( Figure 2A–2C ) ., Histological examination also showed embryos failed to attach to the uterine lumen of mutant mice , while normal attachment and induction of the decidual response was observed in all controls ( Figure 2D–2E ) ., Embryo-attachment failure is most likely caused by an altered uterine receptivity response in the mutant model , since the blastocyst still contains an unaltered COUP-TFII allele , and even mutant embryos are able to implant in wild-type mothers as previously described 17 ., Decidualization is the subsequent step in the implantation process 23 ., Although it is not possible to compare the decidual response in natural pregnancies of these mice , decidualization was assayed after hormonal induction 24 ., Induction of decidualization was normal in COUP-TFIIflox/flox control mice using two types of stimuli ( oil injection into the uterine lumen or needle scratching on the antimesometrial side of luminal epithelia ) ., Uterine horns of PRCre/+ COUP-TFIIflox/flox mutant mice failed to decidualize under treatment of either stimuli ( Figure 3A , 3B , and 3E ) , and alkaline phosphatase activity , an indicator of stromal cell differentiation in response to decidualization , was absent ( Figure 3C and 3E ) ., In addition to the failure of decidualization , stromal cell proliferation is also affected since the size of mutant uterine horns appear small at 4 . 5 dpc ( Figure 2A and 2B ) ., Immunostaining of phosphorylated histone H3 ( phospho-H3 ) demonstrated that stromal cell proliferation was significantly decreased as indicated by the phospho-H3–positive cells ( Figure 3F and 3G ) ., The numbers of phospho-H3–positive cells in stroma are quantified and shown in Figure 3H ., In contrast to the stroma , the numbers of phospho-H3–positive cells in the epithelia are increased in the mutant ( Figure 3F and 3G ) ., The increase in the numbers of proliferating cells in the mutant epithelium are quantified and shown in Figure 3I ., In addition to the decreased proliferation in the stroma , vessel density visualized by lectin staining was also lower in the mutant uterus ( Figure 3J and 3K ) ., Reduced angiogenesis could partly contribute to the decrease in size of the uterine horn ., BMP2 is a known specific marker for decidualization in the uterus , and its expression is greatly induced upon decidualization 25 , 26 ., To explore the molecular mechanism of decidualization failure in the COUP-TFII mutant mice , we asked whether expression of BMP2 is altered ., Basal Bmp2 expression levels were unaffected in the mutants in comparison with the controls ., However , the induced expression of Bmp2 upon decidualization was greatly diminished in the mutant uterus ( Figure 4A ) ., Immunohistochemistry confirmed no stromal expression of BMP2 in the mutant uterus ( Figure 4B and 4C ) ., The above results suggest that BMP2 is a downstream target of COUP-TFII that regulates the decidual response ., To address this , we asked whether BMP2 could rescue the decidualization defect exhibited by the COUP-TFII conditional mutant ., Along with artificially stimulating the uterus , recombinant human BMP2 was administered into the uterine lumen ., Mice were dissected 48 h later , and the decidual response was measured ., BMP2 treatment restored the decidual response in the mutant uterine horns ( Figure 4D and 4G ) as measured by the enhancement of alkaline phosphatase activity in the stimulated horns , while no activity was detected in the vehicle ( BSA ) –treated mutant horns ( Figure 4E , 4F , 4H , and 4I ) ., These results strongly support that BMP2 is a major COUP-TFII effector that lies downstream of COUP-TFII to mediate uterine decidualization ., BMP2 has also been shown as a downstream target of hedgehog signaling in other tissues 27 , 28 , and conditional ablation of Bmp2 results in decidualization defects , but embryo attachment is unaffected ( Lee et al . , unpublished data ) ., Therein , our finding provides new evidence in support of the existence of a uterine Ihh–COUP-TFII–BMP2 axis that is required for decidualization ., The lack of embryo attachment indicates that ablation of COUP-TFII not only affects the physiology of uterine stromal cells but also affects the endometrial epithelial compartment ., One of the major roles of progesterone is to down-regulate ER activity in the uterine luminal epithelium , which consequently opens the uterine receptivity window ., Since COUP-TFII mutants have a receptivity defect ( Figure 2A–2E ) , we wondered whether COUP-TFII is a mediator of progesterones suppression of ER activity in the epithelia ., If so , ER activity in this compartment should increase in COUP-TFII mutants ., To address this , the expression level of estrogen-responsive genes was examined by quantitative real-time RT-PCR analysis ( qRT-PCR ) ., The expression of lactoferrin ( Ltf ) , a known estrogen-responsive target in the uterine epithelia 29 , is significantly elevated in the mutant uterus at 3 . 5d pc ( Figure 5A ) ., To exclude the possible involvement of other factors , we also examined the expression of Ltf in mice exogenously treated with hormones , mimicking 3 . 5 dpc of pregnancy ( 30 h after progesterone and estrogen Pe treatment; see Materials and Methods ) ., Although the fold changes vary , Ltf expression level is consistently significantly higher in mutant mice ( Figure 5B ) ., Immunohistological staining detected high lactoferrin expression in mutant epithelia ( Figure 5C–5F ) , demonstrating that estrogen activity is indeed enhanced in the uterine epithelial compartment ., Other well-documented estrogen-responsive genes in the uterine epithelia , including complement component 3 ( C3 ) and chloride channel calcium activated 3 ( Clca3 ) 30 , 31 , were also elevated in the mutant mice ( Figure 5G and 5H ) , indicating that estrogen activity is upregulated in the uterine luminal epithelium of mutant mice ., Mucin 1 ( MUC1 ) is known to be one of the important markers determining uterine receptivity 32 ., MUC1 is an estrogen-responsive target , and its expression is attenuated at the time of implantation to facilitate epithelial remodeling 33 , 34 ., Persistent expression of MUC1 during the peri-implantation period prevents uterine receptivity and embryo attachment 32 ., qRT-PCR showed high expression levels of Muc1 in the mutant uterus ( Figure 6A ) ., In addition , immunohistochemistry detected high expression levels of MUC1 in the apical surface of mutant luminal epithelia ( Figure 6B and 6C ) ., These results suggest that high estrogen activity might be the underlying cause of the uterine receptivity defect displayed by the mutant model ., Consistent with this notion , Clca3 , a gene important for the overproduction of mucus protein 35 , is also shown to be highly upregulated in the mutant uterus ( Figure 5H ) ., Therefore , upregulation of many ER target genes suggests that stromal COUP-TFII is essential for the PR-mediated downregulation of ER activity in the epithelium to open up the receptivity window ., The membrane transformation of uterine epithelia is well documented as a marker of uterine receptivity 36 ., Long microvilli of the epithelial surface are characteristically present under estrogen influence , while progesterone shortens these structures ., Microvilli flattening occurs before implantation and is an important process to facilitate embryo attachment 36 ., Electron microscope ( EM ) studies revealed that mutant epithelia fail to undergo appropriate remodeling to flatten the microvilli ( Figure 6D and 6E ) ., In addition , mutant microvilli exhibit increased glycocalyx expression ( Figure 6F and 6G ) , which is consistent with high expression of MUC1 36 ., Both MUC1 expression and glycocalyx formation prevent embryo attachment 34 , 37 , 38 ., It has been reported that a series of glycosylation enzymes are involved in the glycosylation of mucins , and among them , UDP-N-acetyl-alpha-D-galactosamine: polypeptide N-acetylgalactosaminyltransferase 1 ( GALNT1 ) is a key enzyme 39 ., We examined Galnt1 expression in qRT-PCR and observed that its expression level is increased by 20% ( Figure 6H ) , and , most remarkably , another glycosylation enzyme Galnt7 expression is upregulated by almost 70% in the mutant uterus ( Figure 6H ) ., Although it has not been well established in the mouse uterus , activation of GALNT7 catalytic activity requires prior glycosylation by other enzymes 40 , and GALNT7 cooperatively functions with GALNT1 41 ., The high expression of these enzymes might account for hyperglycosylation of the apical surface of the mutant uterine luminal epithelium ., Another important parameter for uterine epithelial maturation is the presence of desmosomes 36 , 42 , 43 , adherent junctions of the lateral plasma membrane ., Desmosomes are normally lost before implantation to facilitate embryo invasion into the uterine stroma ., However , desmosomes are persistently present in the mutant epithelia ( Figure 6I and 6J ) ., As expected , the expression level of desmocollin-2 ( Dsc2 ) , one of the ubiquitous desmosomal components 44 , is high in the mutant uterus ( Figure 6K ) ., This inappropriate regulation of Dsc2 might contribute to desmosome dysregulation ., Taken together , the high estrogen activity observed in the mutant epithelium alters the uterine receptivity in the mutant mice , which is reflected by striking structural abnormalities in the apical–lateral regions of mutant luminal epithelial cell ., PR in the stroma has been implicated to play a critical role in modulating ER activity in the epithelium 6 , 7 ., Since the activity of ER is enhanced in COUP-TFII mutants , an important question is whether ablation of COUP-TFII in the uterine stroma alters the expression level of stromal PR ., To address this possibility , we used PR-specific immunostaining to assess the expression of PR in the uterus of controls and mutants ., The result clearly shows that the expression level of PR is significantly reduced in the stroma of COUP-TFII mutants ( Figure 7A and 7B ) ., In contrast , there is no significant change in the PR expression levels in the luminal epithelium or the glandular epithelium ., This result indicates that downregulation of PR in the stroma in the absence of COUP-TFII could disrupt stromal–epithelial interactions and contribute to the enhanced ER activity ., In an attempt to further dissect the molecular mechanism of enhanced estrogen activity in the mutant uterine epithelium , we first asked whether uterine ERα levels are altered in mutants ., qRT-PCR showed a 40% increase in levels of ERα ( Esr1 ) mRNA in the whole-uterine tissues of COUP-TFII mutants ( Figure 7C ) ., Immunohistological staining using ERα-specific antibody further confirmed an increased expression of ERα in the epithelial compartment of COUP-TFII mutants ( Figure 7D and 7E ) ., To quantify the difference in expression levels , we isolated uterine epithelia from whole uterus and examined ERα expression by western blot analysis ., The result showed that ERα expression is increased 2- to 3-fold in the mutant uterine epithelia ( Figure 7F ) ., To further ask whether these receptors are activated or not , we examined the phosphorylation status of ERα using antiphosphorylated ERα antibody and observed increased phosphorylation of ER in the uterine epithelium of COUP-TFII mutants ( Figure 7F–7H ) ., Increased phosphorylation levels of ER seem proportional to increased expression levels of ER , but this modification has been shown to couple with growth factor signaling , which might be controlled under paracrine mechanism 45 and is less likely to be an autophosphorylation; therefore , this finding really supports the notion that stromal–epithelial communication is dysregulated in the COUP-TFII mutant uterus ., In addition to ER , members of the steroid receptor coactivator ( SRC ) /p160 family , SRC-1 and SRC-2 , have been shown to play a major role in regulating ER activity and uterine function 46–48 ., Thus , we examined the expression of coactivators by both immunohistochemistry and western blot analysis ., We showed that SRC-1 is upregulated in mutant uterine epithelia ( Figure 7F , 7I , and 7J ) , while SRC-2 and SRC-3 are unchanged ( unpublished data ) ., Taken together , the increase in ER , phosphorylated ER , and SRC-1 levels in the mutant uterine epithelium can together contribute to enhanced uterine ER activity in the COUP-TFII mutant ., Uterine receptivity has been intensely studied in recent years because of its clinical importance 49 , 50 ., Mouse models generated by gene-knockout technology revealed that multiple factors are involved in this process 1–3 ., Although individual factors have proven to be essential for uterine receptivity , most of them are directly or indirectly controlled by estrogen and/or progesterone ., Therefore , we assume that the balance in activities between these two hormones is a major determinant of successful uterine receptivity ., Indeed the levels of estrogen used in in vitro fertilization procedures have recently been suggested as a likely contributor to lower pregnancy successes when using artificial reproductive techniques 51 ., This reappraisal prompts the question of how to control estrogen activity during the peri-implantation period so that higher success rates with in vitro fertilization can be achieved ., An important step toward addressing this question is to define the mechanism by which progesterone modulates estrogen activity in the uterine epithelium ., Understanding this pivotal control mechanism would enable the formulation of better clinical protocols to induce and preserve the receptive uterus ., Based on the findings described herein , we propose a new model to explain estrogen and progesterone control of uterine implantation ., In this model , progesterone activates the Ihh–COUP-TFII–BMP signaling axis to elicit stromal cell differentiation that is required for decidualization ., Importantly , COUP-TFII also mediates progesterone-induced suppression of epithelial estrogen action through decreasing epithelial ER and SRC-1 levels and inhibition of ER activation ( phosphorylation ) during the peri-implantation period ( Figure 8 ) ., All these effects are likely due to its regulation of stromal PR level , which was shown to be responsible for the downregulation of ER activity 6 , 7 ., Because COUP-TFII is expressed in the stroma , a paracrine mechanism of action is proposed by which stromal-derived COUP-TFII controls epithelial ER activity through as-yet-unknown mediator ( s ) that transmits the inhibitory signal from the stromal to the epithelial compartment ., Although beyond the scope of this study , identification of this paracrine signal represents the next most important step to fully understand the complete circuitry of progesterone/estrogen action in reproduction ., The model of progesterone signaling described here is most likely oversimplified ., Many other players in both epithelial and stroma compartments may also participate in the overall regulation ., For example , ER in the stromal compartment and the very low expression of COUP-TFII , if any , in the epithelial compartment , may also participate in some way in this scheme ., Only by compartmental-specific deletion of these genes can we validate and dissect the contributions of these proteins to embryo implantation in the future ., As reported previously , COUP-TFII is regulated by progesterone through Ihh signaling , which emanates from the epithelial compartment of the uterus , but we should not underestimate the roles of stromal PR because tissue-recombinant experimental models have demonstrated that epithelial estrogen activities are suppressed by stromal PR 6 , 7 ., Indeed decreased expression of stromal PR is expectedly observed in COUP-TFII mutant mice , which was examined under proper comparison ( PRCre/+ versus PRCre/+ COUP-TFIIflox/flox; Figure 7A and 7B ) , although detailed regulatory mechanism of this interdependence has yet to be defined ., In addition , leukemia-inhibiting factor ( LIF ) , which has been well documented , may also participate in this scheme ., Lif-null mice exhibit defects in embryo attachment , in which the specific ultrastructural and immunohistological features associated with a receptive uterus are lost 52 ., Since observed phenotypes in Lif-null mice are similar to COUP-TFII mutant mice , LIF also could be placed in our scheme ., LIF is known to be estrogen responsive; when examined , we did not find significant changes in LIF by qRT-PCR and by immunocytochemistry in mutants in comparison to the controls ., It is possible that more complex mechanism underlies our model , but it is still unequivocal that COUP-TFII has access to the principal part of steroid receptor regulation in the uterine biology ., The finding that COUP-TFII antagonizes ER action is intriguing ., ER has been shown to regulate the expression of many glycoproteins during the peri-implantation period 34 , 53 ., Downregulation of the expression of such glycoproteins ( including MUC1 ) is known to pave the way for remodeling of the epithelial surface to facilitate embryo attachment ., Although COUP-TFII has been shown to compete with ER binding in vitro in the regulation of Ltf 54 , 55 , COUP-TFII is not expressed in the same compartment as lactoferrin and MUC1 , and thus it is unlikely that it regulates their expression directly in vivo ., Using tissue-recombinant studies , Buchanan et al . showed that epithelial lactoferrin expression is not only regulated by epithelial ER but also regulated by stromal ER 56 ., This raises the possibility that COUP-TFII might compete with stromal ER and alter the epithelial ER function ., Another possible mechanism is that COUP-TFII regulates local estrogen levels , since COUP-TFII has been shown to compete with SF-1 to regulate aromatase expression 57 ., However , aromatase expression was not altered in the COUP-TFII conditional mutant mice ( unpublished data ) ., We also showed that the expression of ER , phosphorylated ER , and SRC-1 are all increased in the COUP-TFII mutants ., Enhanced expression of these molecules will no doubt contribute to the observed increased ER activity and the subsequent activation of the downstream ER targets ., Since COUP-TFII is highly expressed in the stroma but is barely detectable in the epithelia , the up-regulation of ER activity in the epithelium is unlikely a consequence of direct regulation of the above molecules by COUP-TFII ., It is more likely that the stromal COUP-TFII regulates PR to control a paracrine signal , which acts through its epithelial receptor to suppress epithelial ER activity as well as ER and its coregulator expression ., Unlikely as it might be , we can not exclude the possibility that the low levels of epithelial COUP-TFII expression is sufficient to synergize with other epithelial factors to suppress epithelial ER activity directly ., In conclusion , COUP-TFII controls early molecular and cellular changes in the uterus that are required for embryo implantation and subsequent decidualization ., Based on our previous observation that COUP-TFII is a mediator of the Shh pathway in motor neurons and the stomach 11 , 14 , it is not surprising that COUP-TFII mediates progesterone–Ihh signaling to regulate decidualization ., We also show that BMP2 can rescue the decidual defect elicited by the loss of COUP-TFII , which places BMP2 downstream of the COUP-TFII pathway ., Unexpectedly , stromal COUP-TFII also promotes PR expression to mediate progesterone-induced suppression of estrogen activity in the uterine epithelium; local suppression of estrogen activity is required to establish a receptive uterus ., Therefore , progesterone control of epithelial estrogen activity is projected from the stromal compartment via COUP-TFII through a complex epithelial–stromal cross-communication pathway ., The abnormal increase in estrogen activity following the removal of COUP-TFII may help our understanding of the molecular events that control uterine receptivity as well as female reproductive health ., Generation of COUP-TFIIflox/flox mice and PR-Cre knockin mice has been previously described 14 , 22 ., To obtain uterine tissues of pregnant mice , we started mating with wild-type males ( B6SJLF1; Taconic , http://www . taconic . com ) at 7 wk of age and designated the day of vaginal plug as pregnant day 1 ., Ovariectomy was performed at 6 wk of age and followed by the hormone regimen as described below ., For priming with 1 μg of 17β-estradiol ( E2; Sigma-Aldrich , http://www . sigmaaldrich . com ) was dissolved in 1 ml sesame oil ( Sigma-Aldrich ) , and 0 . 1 ml was subcutaneously administered in a single dose for each mouse ., For daily treatment of Pe , 10 mg progesterone ( Sigma-Aldrich ) and 67 ng 17β-estradiol ( nidatory estrogen ( e ) ) were dissolved in 1 ml sesame oil , and 0 . 1 ml was subcutaneously administered in a single dose for each mouse ., In the implantation study , 1% Chicago Sky Blue 6B ( Sigma-Aldrich ) was prepared in 0 . 9% saline , and 0 . 1 ml was intravenously injected for each mouse before dissection ., For the rescue of decidualization , 25 μg recombinant human BMP2 ( Fitzgerald Industries International , http://fitzgerald-fii . com ) was reconstituted by 10% BSA , and 10 μl was administered for each uterine horn ., All procedures for animal study were approved by the institutional animal care guidelines at Baylor College of Medicine ., All assays were repeated at least three times ., We followed the ovary transfer procedure described previously 58 ., Ovaries from 6-wk-old controls , PRCre/+ or PRCre/+ COUP-TFIIflox/+ mice , or mutant PRCre/+ COUP-TFIIflox/flox mice were isolated and then transferred to a B6129-F1 female mouse ., At 2 wk after transfer , the mice were mated with B6SJL-F1 male mice for a period of 2 to 6 mo ., Each litter was genotyped in order to characterize the origin of the pups ., When two litters came from the transferred ovary , the mating was stopped and the experiment was considered a success ., The details of this method have been previously described 24 ., Briefly , after 2 wk of ovariectomy , we first primed mice with 100 ng of estradiol ( E2 ) for 3 d and then started the daily treatment of 1 mg progesterone and 6 . 7 ng E2 ( Pe ) 2 d later ., Mechanical stimulation was added 54 h after the first Pe treatment ( 54 hPe ) , and mice were dissected 48 h later for decidual response measurement ., The same hormone regimen was used for exogenous hormone treatment mimicking 3 . 5 dpc ., Tissues were isolated at 30 hPe ., Isolated uterine tissues were fixed in 4% paraformaldehyde ( PFA ) /PBS , dehydrated through graded ethanol , and processed for paraffin embedding ., Primary antibodies used in this study are as follows: mouse monoclonal anti-COUP-TFII ( 1:1 , 000; Perseus Proteomics , http://ppmx . com ) , rabbit polyclonal anti–phospho-H3 ( 1:200; Upstate Biotechnologies , http://www . upstate . com ) , goat polyclonal anti-BMP2 ( 1:100; Santa Cruz Biotechnology , http://www . scbt . com ) , rabbit polyclonal anti-lactoferrin ( 1:5 , 000; Abcam , http://www . abcam . com | Introduction, Results, Discussion, Materials and Methods | Progesterone and estrogen are critical regulators of uterine receptivity ., To facilitate uterine remodeling for embryo attachment , estrogen activity in the uterine epithelia is attenuated by progesterone; however , the molecular mechanism by which this occurs is poorly defined ., COUP-TFII ( chicken ovalbumin upstream promoter transcription factor II; also known as NR2F2 ) , a member of the nuclear receptor superfamily , is highly expressed in the uterine stroma and its expression is regulated by the progesterone–Indian hedgehog–Patched signaling axis that emanates from the epithelium ., To further assess COUP-TFII uterine function , a conditional COUP-TFII knockout mouse was generated ., This mutant mouse is infertile due to implantation failure , in which both embryo attachment and uterine decidualization are impaired ., Using this animal model , we have identified a novel genetic pathway in which BMP2 lies downstream of COUP-TFII ., Epithelial progesterone-induced Indian hedgehog regulates stromal COUP-TFII , which in turn controls BMP2 to allow decidualization to manifest in vivo ., Interestingly , enhanced epithelial estrogen activity , which impedes maturation of the receptive uterus , was clearly observed in the absence of stromal-derived COUP-TFII ., This finding is consistent with the notion that progesterone exerts its control of implantation through uterine epithelial-stromal cross-talk and reveals that stromal-derived COUP-TFII is an essential mediator of this complex cross-communication pathway ., This finding also provides a new signaling paradigm for steroid hormone regulation in female reproductive biology , with attendant implications for furthering our understanding of the molecular mechanisms that underlie dysregulation of hormonal signaling in such human reproductive disorders as endometriosis and endometrial cancer . | Pregnancy is established and maintained through a series of precisely choreographed cellular and molecular events that are controlled by two sex hormones , estrogen and progesterone ., Both hormones exert their actions through their distinct nuclear receptors ., During the peri-implantation period , estrogen activity is attenuated by progesterone to facilitate epithelial remodeling and embryo attachment , but the detailed molecular mechanism of how this process is achieved remains largely undefined ., COUP-TFII ( chicken ovalbumin upstream promoter transcription factor II; also known as NR2F2 ) , a member of the nuclear receptor superfamily , is highly expressed in the uterine stroma , and its expression is controlled by progesterone–Indian hedgehog–Patched signaling from the epithelium to the stroma ., To assess the uterine function of COUP-TFII , uterine-specific COUP-TFII knockout mice were generated ., These mutant mice are infertile due to failure of implantation ., We identified a novel genetic pathway in which the epithelial Ihh regulates the stroma COUP-TFII to control BMP2 and regulates decidualization ., Interestingly , enhanced epithelial estrogen activity , which impedes the maturation of receptive uterus , was clearly noted in the absence of COUP-TFII ., This finding reveals that COUP-TFII plays a critical role in maintaining the balance between estrogen and progesterone activities to establish proper implantation ., This finding also provides new insights into womens health care associated with uncontrolled estrogen activity , such as breast cancer and endometriosis . | mammals, physiology, eukaryotes, vertebrates, mus (mouse), animals, genetics and genomics | null |
journal.pbio.0050177 | 2,007 | Development of the Human Infant Intestinal Microbiota | The adult human body typically comprises ten times more microbial cells than human cells , due largely to the extremely high density of microbes found in the human intestinal tract ( typically 1011–1012 microbes/ml of luminal content ) ., This microbial ecosystem serves numerous important functions for its human host , including protection against pathogens , nutrient processing , stimulation of angiogenesis , and regulation of host fat storage 1–7 ., It is clear that this list is not yet complete; as this field of study expands , we are continually discovering new roles and relationships ., Studies of gnotobiotic mice have been particularly enlightening , illustrating the essential role of the gastrointestinal ( GI ) microbiota in normal gut development 2 , 5 ., In addition , numerous diseases in both adults and infants have known or suspected links to the GI microbiota , including stomach cancer 8 , mucosa-associated lymphoid tissue lymphoma 9 , inflammatory bowel disease 10 , 11 , and necrotizing enterocolitis 12 , 13 ., The composition of the adult GI microbiota has been intensely studied , using both cultivation and , more recently , culture-independent , small subunit ( SSU ) ribosomal DNA ( rDNA ) sequence-based methods 14 ., The human colon ecosystem alone has been estimated to contain more than 400 bacterial species , belonging to a limited number of broad taxonomic divisions 15 ., Members of the anaerobic genera Bacteroides , Eubacterium , Clostridium , Ruminococcus , and Faecalibacterium have typically been found to comprise a large majority of the human adult gut microbial community ., Still , each adults gut appears to have a unique microbial community , with a structure that remains stable on the time scale of months 3 , 15 , 16 ., In contrast , the infant GI microbiota is more variable in its composition and less stable over time ., In the first year of life , the infant intestinal tract progresses from sterility to extremely dense colonization , ending with a mixture of microbes that is broadly very similar to that found in the adult intestine 17 ., Although the beginning and end points of this time course are well defined , the path between these points is poorly understood ., There are conflicting reports in the literature regarding the composition of the neonatal GI microbiota and the factors that shape it ., Several studies have reported that Bifidobacteria almost always dominate the GI microbiota of breast-fed infants by several weeks of age 17–20 , while others find that they occur in only a small fraction of infants , or are not numerically dominant 21 , 22 ., The effect of diet on the composition of the infant GI microbiota is also controversial—numerous studies have found a lower abundance of Bifidobacteria and a higher abundance of aerobic bacteria in the GI microbiota of formula-fed infants relative to breast-fed infants 20 , 21 , 23–25 , yet other reports have found no such difference 26 , 27 ., Mode of delivery has frequently been cited as one of the key factors that shape the infant microbiota 18 , 28 , 29 ., The GI microbiota of infants delivered by caesarean section has been reported to differ from that of infants delivered vaginally , both in the timing of colonization and in composition 18 , 30–32 , and in some cases , there are clearly traces of the maternal vaginal microbiota in the neonatal GI microbiota 33 , yet the relative importance of mode of delivery on GI microbiota is unclear ., Because of the increased incidence of GI problems in premature infants , the effect of gestational age has also been extensively studied ., These studies have consistently shown that the microbiota of hospitalized , preterm infants differs from that of healthy , full-term babies 32 , 34–36 ., Attempts to associate specific microbes with the occurrence of necrotizing enterocolitis , a condition with suspected bacterial etiology that is an important cause of morbidity and mortality in premature babies , have yielded mixed results 32 , 36 ., Clearly , there is still much to be learned about the origins and development of the infant GI microbiota and its influence on health and disease ., We focused our study on describing the range of profiles that constitute a healthy infant GI microbiota in the hopes of discovering themes that govern its development , and in order to provide a detailed reference and a solid foundation for later studies examining the factors that influence the GI microbiota ., Our study participants included 14 healthy , full-term babies , born to 13 healthy mothers ( thus including one set of fraternal twins ) ( Table 1 ) ., Stool samples were collected according to a prescribed schedule , beginning with the first stool produced after birth: samples were collected daily at first and then with decreasing frequency over the course of 1 y , with additional sampling around key events ( e . g . , introduction of solid food and administration of antibiotics ) , yeilding an average of 26 stool samples per baby ( Table 2 ) ., In addition , stool samples were collected from parents and siblings , and vaginal swabs and breast milk were collected from the mothers ., We analyzed the microbiota of each of these specimens using a newly developed SSU rDNA microarray designed to give nearly comprehensive coverage of known SSU rDNA species ., A subset of these samples was also analyzed by SSU rDNA clone library sequencing , for the purposes of calibrating and validating our microarray results ., To survey the composition of our sample set and to provide a basis for quantitative calibration of the microarray results , we created a reference pool by combining equal amounts of amplified SSU rDNA from each PCR-amplifiable sample ( except for samples collected when the infants were ≥1 y old ) ., We obtained 3 , 458 high-quality clone sequences from a library constructed from this pool , and taxonomically assigned each sequence using Ribosomal Database Projects Classifier 37 ., The taxonomic distribution of these sequences is summarized in Table 3 ., To assess the performance of our new microarray design relative to SSU rDNA sequencing , we sequenced SSU rDNAs amplified from each of 12 individual biological samples obtained in this study , selected for their diverse profiles by 16S rDNA microarray analysis ., This study set included DNA extracted from eight baby stools , two maternal stools , one vaginal swab , and one breast milk sample ., For each of these samples , we amplified SSU rDNA sequences using the same PCR primers that were used in the microarray analysis , then cloned and sequenced several hundred ( mean = 342 ) of the amplified products for a total of 4 , 100 sequences ., We focused our comparison at levels 2 , 3 , and 4 of the prokaryotic multiple sequence alignment ( prokMSA ) hierarchy , which very roughly correspond to the phylum , class , and order levels in the classical taxonomic hierarchy ., At these broader levels , most sequences are expected to have homology to at least one probe in our current microarray design , and rDNA sequences can generally be unambiguously classified ., Microarray-based relative abundance estimates were obtained for 2 , 149 species and taxonomic groups by integrating data from all probes that represented any subset of the class in question , as fully described in Materials and Methods ., Sequence-based estimates were obtained by taxonomically classifying each sequence by assigning the prokMSA operational taxonomic unit ( OTU ) code of the best BLAST match in the 2004 prokMSA database of 86 , 453 SSU ribosomal RNA ( rRNA ) gene sequences 38 ( Datasets S1 and S2 ) ., Although the relative abundance of a bacterial species cannot be precisely determined from its proportional representation in a pool of amplified rDNA sequences , we expect that such estimates should be accurate within an order of magnitude and usually within a few-fold 39–41 , based on previous studies that compared abundance levels estimated from sequencing SSU rDNA amplicons with counts based on in situ hybridization ., Overall , the microarray results were very similar to those obtained by sequencing , both qualitatively and quantitatively ., Figure 1A shows the comparison of the community profiles of each of the 12 samples derived from our microarray analysis and by sequencing , for each taxonomic group at level 2 of the prokMSA taxonomic tree ., Note that the levels ( e . g . , level 2 ) in the prokMSA taxonomy do not have a consistent correspondence with the levels ( e . g . , phylum ) in the classical taxonomic hierarchy , and thus some of the conventional names associated with prokMSA level 2 groups can appear somewhat incongruous ., Both the sequence analysis and the microarray analysis showed that the samples were dominated by a limited number of taxonomic groups—99% of the 4 , 100 sequences were encompassed by just three of the 22 level 2 prokMSA divisions: 2 . 15 ( Flexibacter-Cytophaga-Bacteroides ) , 2 . 28 ( Proteobacteria ) , and 2 . 30 ( Gram-positive bacteria including Firmicutes and Actinobacteria ) , and the remaining 1% belonged to groups 2 . 10 ( Prosthecobacter ) , 2 . 29 ( Fusobacteria ) , or 2 . 21 ( Cyanobacteria and Chloroplasts ) ., As shown in Figure 1B and 1C , the population profiles obtained by microarray and sequencing analysis were also quantitatively similar—the Pearson correlation of the microarray- and sequencing-based estimates of relative abundance for the 12 samples was 0 . 97 at prokMSA taxonomic level 2 ( Figure 1B ) , 0 . 89 at level 3 ( Figure 1C ) , and 0 . 80 at level 4 ( unpublished data ) ., We estimated the overall density of bacteria in each sample by a real-time quantitative PCR ( qPCR ) assay , using a broad-range bacterial primer and probe set ( see Materials and Methods ) ., We used the total number of rRNA gene copies ( typically about five per genome 42 ) per gram of stool , as estimated by this assay , to approximate the total density of bacteria ., As shown in Figure 2 , the total number of rRNA gene copies was relatively unstable throughout the first week of life , then persisted in most babies in the range of 109 to 1010/g of stool ( wet weight ) ., Although there was no clear effect of method of delivery on the timing of the colonization , it is noteworthy that babies 13 and 14 ( the dizygotic twins ) , who were the only babies delivered by a planned caesarean section , and thus without rupture of the amniotic membrane and exposure to maternal birth canal microbiota during labor or delivery , had low bacterial counts ( <108 rRNA gene copies/g ) until the seventh day of life ., We analyzed the bacterial composition of 430 samples—363 infant stool samples , 43 adult stool samples , two sibling stool samples , 12 breast milk samples , and ten maternal vaginal swabs—by hybridization to the DNA microarray developed in this study ., By combining information across multiple probes ( see Materials and Methods ) , we obtained relative abundance estimates for 2 , 149 nested taxonomic groups and species in each of these samples ( All probes are listed in Dataset S3; All taxa are listed in Dataset S4 ) ., As shown in Figure 3 , the phylum-level diversity in the stool samples analyzed in this study was extremely limited ., The vast majority of samples were dominated by just three of the 22 level 2 bacterial groups represented by our microarray: 2 . 15 ( Flexibacter-Cytophaga-Bacteroides ) , 2 . 28 ( Proteobacteria ) , and 2 . 30 ( Gram-Positive Bacteria Firmicutes and Actinobacteria ) ., A second major finding was the remarkable degree of interindividual variation in the colonization process ., Although the taxa that populate the infant GI tract were limited at the broadest levels , each baby was distinct in the combination of microbial species that it acquired and maintained , and in the precise temporal pattern in which those species appeared and disappeared ., Bacteroides , for example , dominated the early microbiota of some babies but were virtually absent at this stage in other babies ., A third striking feature of this dataset was the relative stability of the microbial populations over time—even early in the course of the colonization of the infant GI tract , most taxonomic groups persisted over intervals of weeks to months ., The main dimensions of variation among the colonization profiles of different taxonomic groups were timing of colonization and temporal stability ., Consistent with previous studies 28 , 35 , 43 , 44 , the earliest colonizers were often organisms predicted to be aerobes ( e . g . , Staphylococcus , Streptococcus , and Enterobacteria ) , whereas the later colonizers tended to be strict anaerobes ( Eubacteria , and Clostridia ) ., The Bacteroides varied greatly from baby to baby in the timing of their first appearance , but were consistently present to some degree in nearly all babies by 1 y ., Several other taxa , including Prevotella , Acinetobacter , Desulfovibrio , Veillonella , and Clostridium perfringens , tended to appear only transiently , sometimes appearing and disappearing repeatedly within a babys first year of life ., We explored the similarities and differences in the composition of all of our samples by hierarchically clustering the 430 samples based on their similarity with respect to their abundance profiles for the set of 53 prokMSA level 4 taxonomic groups that had at least two samples with a relative abundance estimate greater than 1% ., The clustering pattern , as reflected in the dendrogram at the top of Figure 4 , highlights several critical features of the colonization program , and shows that the stool microbiota of babies 1 y of age and older is distinctly different from that at earlier ages and much more similar to that of adults ., Prior to 6 mo of age , stool samples tended to cluster by baby , indicating that the differences from baby to baby are much greater than the changes over periods of weeks or months in the composition of any individual babys microbiota ., There were two notable exceptions to this baby-specific clustering ., First , samples from the first few days of life often clustered away from the rest of a given babys samples , sometimes clustering with other very early samples and sometimes with samples from other sites ( e . g . , baby 8 day 1 with vaginal samples ) ., Second , samples from babies 13 and 14 , who are fraternal twins , tended to intermingle ., Figure 4B shows examples of several of the clustering patterns described above ., Most of the breast milk and maternal vaginal samples clustered perfectly by anatomic site of origin ., As expected , all but one of the vaginal samples were overwhelmingly dominated by lactobacilli , with Staphylococci , Bacteroides , Clostridia , and Veillonella among the groups variably present as minority constituents ., The vaginal sample from one of the mothers ( mother of baby 11 ) had a distinctly different population profile , dominated instead by members of the Gamma Proteobacteria group ., The microbial populations found in the milk samples were diverse , often including mixtures of enterics and species of Bacteroides , Pseudomonas , Haemophilus , Veillonella , and Streptococcus ., In order to compare the infants more systematically , we determined the nearest-neighbor sample for each sample as measured by the Pearson correlation of level 4 relative abundance estimates ., Using this metric , the nearest-neighbor sample of any given baby sample was usually another sample from the same baby—the average percentage of samples from a given baby for which the most similar sample was from the same baby was 82% ., Figure 5 summarizes this analysis and illustrates the interesting finding that by this measure , the most similar pair of babies by far was babies 13 and 14—fraternal twins raised in the same environment—8 of 23 ( 35% ) of baby 13′s nearest-neighbor samples were from baby 14 ( the next most similar pair was babies 11 and 14 , at 17% ) ., The similarity of the microbial community profiles of stool samples from babies 1 y and older to each other and to those of the adult stool samples suggested that the infant GI communities converged over time toward a generalized “adult-like” microbiota ., We explored this phenomenon by calculating , for each age interval , the average pairwise Pearson correlation of the population profiles of all infant samples collected at that age ., As shown in Figure 6A , this analysis revealed that as time progressed , the babies microbiota consistently converged toward a common profile ., We also calculated , for each time point , the average correlation of infant samples at that time point to a generalized adult profile ( centroid of 18 adult samples—nine fathers and nine mothers from this study ) ., This analysis , shown in Figure 6B , confirmed that the profile toward which the infants microbiota converges is similar to that of adults , and highlighted an apparent tendency for a population rearrangement to occur around 5 d after birth ., Notably , the infants GI microbiota was not significantly more similar to that of their parents than to that of other adults , as measured by the Pearson correlations of their level 4 taxonomic profiles ( mean baby–parent correlation of 0 . 55 for within family , versus 0 . 62 between families for nine “triads” of contemporaneously obtained samples from baby , mother , and father obtained at 1–1 . 5 y of age ) ., To visualize the temporal patterns in the particular phylogenetic groups that populate the infant gut , we charted the relative abundance of the nine level 4 taxonomic groups that had a mean relative abundance of 1% or greater over time in each infant ( Figure 7 ) ., This analysis enabled us to identify common themes and interesting differences among the colonization profiles of these babies ., First , we observed that “uneven” populations ( populations heavily dominated by a single taxonomic group ) were common in the first several weeks but rare later in the time courses ., Another notable feature in the temporal program of many of the babies was the occurrence of one or more dramatic shifts in the population structure—such shifts were frequently stabilized within one sampling interval ., We were unable to identify any specific age or signal event consistently associated with such transitions , although the transition to an “adult-like” profile often followed the introduction of solid foods ., Several of the babies were treated with antibiotics either in the neonatal period ( day 0–28 ) or in the later months ( see Table 1 and Figure 2 for more details ) ., In some cases , the treatment was associated with a striking alteration in the density or composition of the GI microbiota ., For example , baby 8 received two courses of amoxicillin , one at 4 mo and one at 6 mo ., In both cases , both the total density of bacteria ( Figure 2 ) and the community composition were dramatically altered ( Figures 3 and 6 ) ., Indeed , in this baby , the bacterial density in fecal samples decreased so much during the antibiotic courses that we were unable to amplify sufficient SSU rDNA for microarray analysis , so we could only compare the populations before and after the antibiotic course ., However , we did not identify any consistent consequences of antibiotic treatment ., The results of both the sequence analysis of the reference pool and the microarray data analyses indicated that Bifidobacteria were only minor components of the population—a result at odds with the conventional wisdom 20 , 21 , 26 ., The primers we used for broad-range PCR amplification of the reference pool ( the source of the sequences ) and samples for microarray analysis were potentially suboptimal for amplification of Bifidobacteria 21 , 26 due to three mismatches in the rDNA sequence of Bifidobacterium longum to the forward primer 8F used in this study ., A survey of the 5′ sequences of full-length SSU rDNA genes showed that Bifidobacteria are outliers in their divergence from the generally conserved 8F primer sequence ., We therefore carried out two independent analyses to determine whether and how the quantitative estimates of Bifidobacteria from the microarray hybridization results would need to be adjusted ., First , we quantitatively evaluated the relative efficiency with which the 8F/1391R primer pair amplified SSU rDNA from two Bifidobacteria species—Bifidobacterium longum and Bifidobacterium infantis—compared to a set of three diverse common fecal bacteria—Escherichia coli , Clostridium perfringens , Bacteroides fragilis—all of which have SSU rDNA sequences with one or more mismatches to the 8F/1391R PCR primer sequences ., Using a range of stoichiometric mixtures of chromosomal DNA extracted from these species , we found that after 20 cycles ( the number of cycles used for our microarray analyses and for amplification of the reference pool prior to sequencing ) , efficiency of amplification of the two Bifidobacterial species DNA was consistently 8-fold lower than that of the three other species , all of which amplified with nearly identical efficiencies ( unpublished data ) ., This result suggests that both the reference pool sequencing results and the microarray-based quantitation underestimated the abundance of the Bifidobacteria group by a factor of eight ., Second , we used a real-time qPCR assay with a primer pair and probe optimized for detection of Bifidobacteria to obtain an independent estimate of the abundance of Bifidobacteria in each sample ., The results confirmed the finding from the microarray analysis that Bifidobacteria were almost always only minor constituents of the fecal microbiota of both the infants and adults in our study population ( Dataset S5 and Figure S1 ) ., The majority of bacterial species identified in our sample set were previously reported constituents of the human GI microbiota ., There were , however , a number of cases in which the microarray results indicated the presence of a bacterial species or group that was both unexpected and not represented in our sequenced reference pool ., We investigated several of these cases using independent assays ., For 12 of the suspect species/taxa , we used the cognate group-specific primers in a PCR assay applied to most or all of the samples in which the suspect species/taxa appeared to be present based on the microarray results , as well as a small set of samples in which the suspect species was not detected by the microarray ., In one case , that of Sutterella wadsworthia , sequencing of the species-specific PCR product confirmed its presence ., In seven of 12 cases , none of the array positive ( or negative ) samples yielded an amplified product in the PCR analysis ., For four remaining cases , the ostensibly species-specific PCR assay yielded an amplified product of the expected size , but the clones sequenced from this product did not correspond to the expected species ., We further investigated these four cases by sequencing a clone library obtained by amplification with the same broad-range primers that were used in preparation for microarray analysis ., Although the sequencing did not confirm the presence of any of the four questionable species/taxa , it provided strong evidence for a major source of false-positive hybridization signals ., Specifically , in three of the four cases , we identified a relatively abundant species whose rDNA sequence was sufficiently similar to the probe sequence that it was likely to account for the observed signal ., In one case ( Legionella pneumophila ) , which was predicted to be present at approximately 1% , we were unable to identify any candidate species that could account for the hybridization signal ( i . e . , none with best BLAST matches scores ≥30 ) , among our set of 192 sequences ., Since our power to detect a species present at a partial abundance of 1% was only 85% , it remains possible that this species , or another species with a similar SSU rDNA sequence , could have been present at a low abundance in the suspect samples ., Both our DNA extraction and rDNA amplification methods were optimized for bacteria and suboptimal for eukaryotes and archaea , thus we separately tested for the presence and abundance of fungi or archaea by means of qPCR assays with broad specificity for the respective taxonomic groups ., Based on our qPCR analysis , fungi were intermittently detectable in stool samples at relatively low abundance ( 104–106 rRNA genes/g fecal wet weight ) , persisting for varying durations in individual babies , through the first year of life ., One of the babies in this study ( baby 10 ) was noted to have a diaper rash , as well as oral thrush , both of which are commonly caused by a fungus ( Candida ) , and which were treated with an antifungal agent ( nystatin ) ., The qPCR analysis detected especially high levels of fungal rDNA in stool samples from this baby , particularly during the period in which these findings were described ., This babys mother also had notably high levels of fungal SSU rDNA sequences in her prenatal vaginal swab sample , but not in her “day 0” stool sample ., The prevalence of archaea was considerably lower and more variable than that of fungi or bacteria; qPCR analysis detected archaeal rRNA genes ( in the range of 103–106 rRNA genes/g ) in only seven babies during their first year of life , and in four of these babies , they were detected in only a single sample ., In these babies , archaea appeared only transiently , and almost exclusively in the first few weeks of life; they were detected in only one infant after the fifth week of life ., Limited analysis of archaeal sequences amplified from the three maternal stool samples that tested positive for archaea ( mothers 4 , 9 , and 12 ) revealed a predominance of Methanobrevibacter smithii ( 7/8 archaeal clones identified , including at least one clone from each mother ) , with one additional ( uncultured ) archaeal phylotype ., Results of qPCR analysis of fungi and archaea are included in Dataset S5 and shown graphically with bacterial qPCR results in Figure S2 ., The microbial colonization of the infant GI tract is a remarkable episode in the human lifecycle ., Every time a human baby is born , a rich and dynamic ecosystem develops from a sterile environment ., Within days , the microbial immigrants establish a thriving community whose population soon outnumbers that of the babys own cells ., The evolutionarily ancient symbiosis between the human GI tract and its resident microbiota undoubtedly involves diverse reciprocal interactions between the microbiota and the host , with important consequences for human health and physiology ., These interactions can have beneficial nutritional , immunological , and developmental effects , or pathogenic effects for the host 2 , 5 , 7 , 18 , 45 ., This study began with the development of a DNA microarray with nearly comprehensive coverage of the bacterial taxa represented in the available database of SSU rRNA gene sequences ., Our microarray design and experimental methods were based on lessons learned in the validation of a less comprehensive SSU rDNA microarray 46 ., These previous experiments enabled us to optimize our methods for computational prediction of SSU rDNA hybridization behaviors , and to develop an experimental protocol that maximized hybridization specificity ., The excellent concordance in the measurements of individual taxa determined using the new microarray design in comparison with sequencing results from corresponding SSU rDNA clone libraries ( Figure 1 ) suggests that these design principles hold true for this platform across a diversity of taxa and give us confidence in both the comprehensiveness and accuracy of the results obtained with our new microarray probe set ., It is important to note , however , that our methods of array design and analysis are imperfect and still evolving ., Several of the unexpected species predicted by the microarray to be present in one or more samples could not be corroborated by sequencing ., In most of these cases , sequence analysis of the sample ( s ) in question revealed that low-level cross hybridization of a highly abundant species was responsible for the false-positive prediction , a result that will be taken into consideration in future rounds of array design and analysis ., We used this microarray in a detailed , systematic , and quantitative study of bacterial colonization of the newborn human GI tract ., We used freshly collected stool samples as surrogates for samples taken from the lumen and mucosa of the colon ., Although there are undoubtedly differences in the population profiles of stool samples and corresponding mucosa , we found in a previous study that the profiles are nonetheless remarkably consistent—sufficiently so that individual stool samples can readily be matched to colonic biopsy samples from the same individual , based on the similarity in their bacterial profiles 15 , 46 ., Thus , we believe that the results of our temporal analysis of the bacterial populations in infant stool samples provide a useful window on the resident colonic microbiota ., In view of the importance of the symbiosis between human host and gut commensals for both human host and microbial colonist , it would be easy to imagine that the program of microbial colonization of the neonatal GI tract would have evolved under strong selective pressure , acting on both the intestinal niche and its microbial colonists , to be highly deterministic and stereotyped ., We might have expected that a highly restricted group of co-evolved commensals would be exceptionally well adapted to this environment and consistently dominate the colonization process in a stereotyped fashion ., Indeed , the bacteria that we found in infant and adult feces , presumably reflecting the colonic microbiota , were largely restricted to only a small subset of the bacterial world—Proteobacteria , Bacteroides , Firmicutes , Actinobacteria , and Verrucomicrobia ., Yet , surprisingly , we found that in the first days to months of life , the microbiota of the infant gut , and the temporal pattern in which it evolves , is remarkably variable from individual to individual ., The seemingly chaotic progression of the early events in colonization , and the similarity in bacterial composition of some early infant samples to breast milk or vaginal swabs , suggests that the bacterial population that develops in the initial stages is to a significant extent determined by the specific bacteria to which a baby happens to be exposed ., Notably , these maternal “signatures” did not persist indefinitely , as evidenced by our failure to find a significantly higher correlation of the overall taxonomic profiles of baby/parent pairs from the same household versus different households ., An important exception to the tale of individuality and uniqueness in the early profiles was the remarkable similarity of the temporal profiles of the fraternal twins ( babies 13 and 14 ) ( Figures 4 and 5 ) ., These twins shared both a common environment and approximately 50% genetic identity , making it impossible to determine from this study to what degree each of these commonalities is responsible for their similar colonization patterns ., However , evidence from this and other studies suggests that the shared environment is a major factor ., One argument in favor of this view is the lack of comparable similarity in the microbial communities of other pairs that also share 50% genetic identity , including mother:baby , father:baby , and sibling:baby ( unpublished data ) , although this dissimilarity may be due in part to their differing stages in development ., Another argument in favor of a strong environmental influence is the coincidental transient appearance of specific organisms in both twins—it is hard to imagine that the appearance of a particular microbe on a particular day could be genetically programmed ., Our final argument rests on evidence from a previous study that the microbiota of genetically equivalent families from a cross of inbred mice was more similar among members of the same “household” ( mother and offspring who share a cage ) than between households 1 ., The definition of a “h | Introduction, Results, Discussion, Materials and Methods | Almost immediately after a human being is born , so too is a new microbial ecosystem , one that resides in that persons gastrointestinal tract ., Although it is a universal and integral part of human biology , the temporal progression of this process , the sources of the microbes that make up the ecosystem , how and why it varies from one infant to another , and how the composition of this ecosystem influences human physiology , development , and disease are still poorly understood ., As a step toward systematically investigating these questions , we designed a microarray to detect and quantitate the small subunit ribosomal RNA ( SSU rRNA ) gene sequences of most currently recognized species and taxonomic groups of bacteria ., We used this microarray , along with sequencing of cloned libraries of PCR-amplified SSU rDNA , to profile the microbial communities in an average of 26 stool samples each from 14 healthy , full-term human infants , including a pair of dizygotic twins , beginning with the first stool after birth and continuing at defined intervals throughout the first year of life ., To investigate possible origins of the infant microbiota , we also profiled vaginal and milk samples from most of the mothers , and stool samples from all of the mothers , most of the fathers , and two siblings ., The composition and temporal patterns of the microbial communities varied widely from baby to baby ., Despite considerable temporal variation , the distinct features of each babys microbial community were recognizable for intervals of weeks to months ., The strikingly parallel temporal patterns of the twins suggested that incidental environmental exposures play a major role in determining the distinctive characteristics of the microbial community in each baby ., By the end of the first year of life , the idiosyncratic microbial ecosystems in each baby , although still distinct , had converged toward a profile characteristic of the adult gastrointestinal tract . | It has been recognized for nearly a century that human beings are inhabited by a remarkably dense and diverse microbial ecosystem , yet we are only just beginning to understand and appreciate the many roles that these microbes play in human health and development ., Knowing the composition of this ecosystem is a crucial step toward understanding its roles ., In this study , we designed and applied a ribosomal DNA microarray-based approach to trace the development of the intestinal flora in 14 healthy , full-term infants over the first year of life ., We found that the composition and temporal patterns of the microbial communities varied widely from baby to baby , supporting a broader definition of healthy colonization than previously recognized ., By one year of age , the babies retained their uniqueness but had converged toward a profile characteristic of the adult gastrointestinal tract ., The composition and temporal patterns of development of the intestinal microbiota in a pair of fraternal twins were strikingly similar , suggesting that genetic and environmental factors shape our gut microbiota in a reproducible way . | developmental biology, ecology, obstetrics, pediatrics and child health, immunology, microbiology, homo (human), eubacteria | Microarray profiling of the microbial communities of infant guts throughout the first year shows initial variation then convergence on the adult flora, providing new insight into this human ecosystem. |
journal.pcbi.0030204 | 2,007 | From Inverse Problems in Mathematical Physiology to Quantitative Differential Diagnoses | The potential of computer-based , algorithmic support for medical decision making in data-rich environments , and in particular in the context of evidence-based practice , was recognized early on and has been pursued extensively 3 , 7–11 ., Few of these efforts , which mostly have consisted of rule-based expert systems , statistical models , or approaches driven by machine learning ideas such as dynamic Bayesian or artificial neural networks , have reached a sufficient level of practicality and usefulness to be accepted into the day-to-day practice of acute care medicine 8 , 12 , 13 ., These tools either attempt to formalize empirical knowledge already available to a physician ( expert systems ) or to capitalize on statistical associations of phenomena and inherent structures of the available dataset ., All largely fail to make direct and quantitative use of known causalities and dynamics in the physiologic systems underlying the observed pathophysiology , which are typically characterized by basic science investigations ., A promising approach to incorporating this knowledge into the medical decision making process would be to use mathematical models of physiologic mechanisms to map clinical observations to quantitative hypotheses about physiologic conditions , leading to improved insight into current patient status and , eventually , predictions about responses to therapeutic interventions ., While complex mathematical models of physiology in general , and the cardiovascular system and its control in particular , have a long history and are still actively being developed 14–23 , their translation to clinically useful tools has proved challenging ., Early examples of using mathematical models to quantify “hidden” parameters based on clinical measurements include the pioneering work of Bergman et al . in the late 1970s on glucose control and insulin sensitivity 24 ., More recent work in the same field has focused on accurately quantifying the uncertainty arising in the resulting parameter estimation problems using current methodology , such as Markov chain Monte Carlo approaches 25 ., In the critical care environment , extremely simple models of the cardiovascular system have been in use for decades and are implemented in commercially available products , examples being the electric circuit analog of the systemic circulation used to calculate total peripheral resistance , which can then become a therapeutic target , or pulse contour analysis , which attempts a model-based assessment of systemic flow from arterial pressure waveforms 26–28 ., The clinical application of mathematical models of physiology to date has failed to extend to models of sufficient complexity to significantly help alleviate the previously discussed problem of information overload in the diagnostic process ., We contend that a key obstacle preventing the successful clinical use of available mathematical models has been the lack of a robust solution to the inverse problem ., That is , any physiologically reasonable mathematical model of components of the human body will typically be nonlinear and have a large number of parameters ., Despite the complexity of such models , if the user fixes the parameter values and initial values of the physiological states in a model , then the model can be simulated to obtain time courses of the physiological states ( solving the forward problem; Figure 1 ) ., However , the corresponding inverse problem , i . e . , inferring parameters and starting conditions of state variables from measured physiological data ( http://www . ipgp . jussieu . fr/~tarantola/Files/Professional/Books/InverseProblemTheory . pdf ) 29 will usually be ill-posed in the sense of Hadamard , meaning that it does not admit a unique solution that depends continuously on the data 30–32 ( Figure 1 ) ., This ill-posedness is directly related to the concept of system identifiability in both the statistical and engineering senses of the term ., The most popular approaches to the inverse problem in physiology , such as nonlinear least squares , which seeks a maximum likelihood estimate by minimizing the sum of squared residuals , inherently assume the existence of a unique “best” solution ., The ill-posedness of the inverse problem corresponds to a violation of this assumption , which often causes solution approaches such as least squares to fail completely , in spite of regularization of the underlying nonlinear programming problem , or to give meaningless or even misleading results ., More recent work attempts to quantify the uncertainty of resulting parameter estimates 25 ., However , given the uncertainty stemming from the fundamental ill-posedness of the inverse problem , together with additional uncertainty from measurement error and model stochasticity , which affect both forward and inverse problems ( Figure 1 ) , a fully probabilistic approach to the inverse problem in quantitative physiology seems appropriate ., We hypothesize that the ill-posedness of the inverse problem is not merely a technical obstacle but reflects clinical reality in the sense that an experienced physician is rarely certain about a patients status , despite a large number of available observations ., More typically , the physician entertains an evolving differential diagnosis , consisting of a list of hypotheses of varying likelihoods about the physiological mechanisms underlying available observations , updated and ranked according to current observations ., We therefore propose to approach the inverse problem in such a way that uncertainty from all sources is quantitatively reflected by the solution , which will consequently take the form of a ( typically multimodal ) probability distribution on parameter and state space ., This distribution will represent the relative likelihoods of the possible values of the physiological elements that these parameters and states represent , in the patient for whom the clinical observations are made ., To explore the feasibility of such an approach , we combine a mechanistic model of cardiovascular physiology with a stochastic model of the observation process and Bayesian inference techniques to infer a posterior probability distribution on parameter and state space from prior ( population-level and individual ) knowledge and quantitative observations ., We illustrate these ideas in a simplified simulation of a clinically relevant differential diagnostic procedure and examine the relationship between the obtained posterior probability density functions and pertinent qualitative differential diagnostic concepts ., When only one blood pressure measurement was made , the probable parameter/state range represented a continuum of various combinations of contractility and hydration status ., As expected , high-precision measurements ( σ = 10 mm Hg; Figures 5A and 6A ) led to more concentrated probability density functions than low-precision measurements ( σ = 30 mm Hg; Figures 5B and 6B ) , independent of the type of prior used ., Two peaks , corresponding to the differential diagnoses of “heart failure” ( low contractility , normal-to-high total intravascular volume ) and “hypovolemia” ( normal contractility , low intravascular volume ) , can be discerned in the case when blood pressure was measured with high precision for both Gaussian and uniform priors ( Figures 5A2 and 6A2 ) ., When the measurement was less precise , the peak corresponding to “heart failure” was nearly absent with a Gaussian prior , but not with a uniform prior ( Figures 5B2 and 6B2 ) ., To illustrate the additional diagnostic knowledge gained from perturbing the system , we simulated a fluid challenge 33 ., Depending on the systems response to the intravenous administration of 1 , 500 ml of fluid , the updated posterior densities on parameter space were altered significantly ( Figures 5A3 , 5A4 , 5B3 , 5B4 , 6A3 , 6A4 , 6B3 , and 6B4 ) ., Specifically , for a high-precision measurement , a fluid challenge differentiated between cardiac causes of hypotension ( “heart failure”; low contractility , low responsiveness to volume resuscitation; Figures 5A3 and 6A3 ) and lack of intravascular volume as cause ( “hypovolemia”; normal or high contractility , high responsiveness to volume resuscitation; Figures 5A4 and 6A4 ) ., With low-precision measurements , the failure to restore blood pressure following the fluid challenge did not eliminate hypovolemia as the cause of hypotension ( Figures 5B3 and 6B3 ) ., While the clinician often wonders whether there is a preferred sequence of diagnostic challenges for ascertaining an accurate diagnosis , the order of fluid challenges of different sizes , with consecutive assimilation of intermediate observations , had little effect on the final posteriors in this highly simplified setting ., When we allowed three parameters to vary , the posterior distributions became truly multimodal ., We depict two different visualizations of the grid points accounting for 95% of the total probability mass of posterior densities for the scenarios described earlier ( Figure 7A and 7C ) , as well as for a more ambiguous post-resuscitation observation of 50 mm Hg ( Figure 7B ) ., As can be seen , the assimilated observations are still sufficient to meaningfully constrain the probable region in parameter/state space ., In the poor ( 30 mm Hg post-resuscitation ) response to volume scenario , an additional probability concentration appears ., This additional probability mass corresponds to the possibility of shock induced by severely decreased peripheral resistance , which corresponds to the differential diagnostic possibility of a failure of vasomotor tone , as observed in septic , anaphylactic , or neurogenic shock states ., For intermediate values of the post-resuscitation observation , the structure becomes even richer ( Figure 7B ) , while the post-resuscitation observation of 70 mm Hg ( good response ) concentrates probability mass in a region of low intravascular volume ( Figure 7C ) ., While a greatly simplified physiological representation , our mathematical model of the cardiovascular system fulfills its design objectives: to be qualitatively correct in its response to variations in hydration status and myocardial contractility while incorporating enough homeostatic mechanisms to create realistic ambiguity in the identification of parameter values underlying observed states ., The conversion of the discrete dynamical system representing the sequential filling and emptying of the heart ( and the resulting “history awareness” of the system ) into a compact system of ODEs that preserves the physiologic meaning of parameters of the discrete system is , to our knowledge , novel ., Physiologically constrained cardiovascular simulations done by previous authors have typically involved either simulating intra-beat dynamics , which rapidly becomes computationally prohibitive , using a more ad hoc approximation of the Starling mechanism at the expense of physiological interpretability of parameters , or resorting to a beat-to-beat discrete time representation ( e . g . , 20 ) ., Our model is therefore particularly suited for simulation scenarios where an accurate description of intra-beat details is not required , yet a continuous form of inter-beat dynamics that preserves parameter meanings is desired ., Our model derivation aims to achieve a reasonable compromise between representing all known mechanisms in full physical detail , which leads to challenges of simulation expense and intractability of the inverse problem , and model reduction , which may result in loss of physiological accuracy and interpretability ., The need for such a trade-off is typical when modeling complex biological systems ., From our perspective , making use of domain-specific knowledge to arrive at meaningfully interpretable model reductions whenever possible , and resorting to multiscale models with a hierarchical arrangement of submodels of different granularity and timescales when the assimilation of data on very different spatial and temporal scales is desired , may be the most promising way to address this issue ., The ideal level of model complexity will generally depend both on the amount of data available for assimilation and the specific application intended ., To what extent the growing theoretical understanding of model selection based on information theoretical measures 34 can be leveraged to facilitate or partially automate this process for physiological applications is an interesting topic for further investigation ., As illustrated by this proof-of-concept implementation , the proposed methodology holds promise as a tool for integrating existing mechanistic knowledge and data generated by measurements in a clinical setting into a quantitative assessment of patient status ., Our approach offers a means to achieve this integration in a way that not only incorporates all available data , but also quantifies the remaining uncertainty , thus avoiding unjustified claims of high certainty that could prove disastrous in a clinical setting ., In particular , the clinical construct of differential diagnoses of different likelihoods is reflected in the observed multimodality of posterior probability distributions ( Figures 5–7 ) ., More generally , representations of probability densities of states and parameters provide a natural setting for linking mathematical models of different scales and levels of detail since the distribution of states of some detailed small-scale models ( “microstates” ) may naturally determine a value or distribution of values for parameters of larger scale/lumped parameter models ., The proposed approach aims to map clinical syndromes described by a set of observations to configurations of physiologically meaningful pre-observation states and parameters appearing within a mathematical model ., Based on the physiological knowledge embodied in the model , certain regions in parameter and state space may in turn be associated with differential diagnoses , similar to the conditions of “hypovolemia , ” “heart failure , ” and “sepsis” in our simplified example ., When this linkage is possible , the quantitative nature of the method presented here allows for the estimation and refinement of probability values associated with certain diagnoses ., This , to our knowledge , is the first time that such a high-level concept central to clinical decision making is shown to emerge naturally from the combination of sequential observations , diagnostic challenges , and physiological principles ., Moreover , we believe that the methods presented herein open novel avenues for exploring theoretical aspects of clinical epistemology , independent of practical applications ., Since measurement characteristics are described stochastically , the method we demonstrate is not fundamentally limited to assimilating data from device-based quantitative measurements , but can also make use of rather qualitative clinical observations such as quality of peripheral perfusion , presence of lung rales , or altered mental status , provided reasonably informative densities on system states or parameters conditional on such observations can be defined ., Similarly , genomic information can be naturally incorporated , since it can provide probability distributions of physiological parameters conditional on individuals genomes ., To what extent a combination of several subjective ( or inaccurate ) observations may exploit physiological coupling of observables and yield informative posterior distributions corresponding , for example , to a carefully performed clinical examination is a matter of current investigation ., While modifying the order of physiological challenges did not have a tangible impact on diagnostic discrimination in our limited exploration , we anticipate that order generally matters , as a systems response to a perturbation can be highly dependent on the system state at the time that the perturbation is delivered ., That is , an initial diagnostic challenge will alter the state of the underlying system , which may impact its response to a subsequent challenge ., Our approach could allow for a theoretical exploration of how to optimize the selection , and order , of diagnostic challenges for maximal information gain in the context of specific clinical scenarios ., The simulations presented here illustrate the importance of congruence between the accuracy of observations and the level of information included in prior distributions ., In particular , the inappropriate use of informative priors can be misleading in this context ., In our results , for example , the combination of informative Gaussian priors with inaccurate observations effectively eliminates the physiologically reasonable “heart failure” peak in both the posteriors after a single observation of low blood pressure and the post-resuscitation posteriors for the low-response case , while the peak is still clearly evident in the case of uniform priors ( Figures 5B2 , 5B3 , 6B2 , and 6B3 ) ., This example demonstrates that a conscious choice needs to be made as to whether an interpretation based on population-level probabilities ( corresponding to the use of informative priors ) or an unbiased assessment of physiological possibilities ( corresponding to the use of uniform priors ) is more appropriate , when only few , low-quality measurements are available ., Whether an optimal degree of incorporation of population-based information exists and how such an optimum could be defined are highly relevant issues that remain to be explored ., Our observations suggest that when addressing inverse problems in quantitative physiology , the traditional approach of requiring a unique optimal solution may be misleading and introduce unnecessary information loss , at least in situations where the additional computational burden of characterizing the posterior distributions more fully is not prohibitive ., Model reduction to eliminate perceived “overparametrization” may weaken the correspondence of components of a physiologically faithful mathematical model with components of the actual physiological system it describes ., Furthermore , we hypothesize that the ambiguity of the mapping from observation to parameter space is at least partially due to a characteristic particular to physiological systems , namely their ability to tightly control certain system states via highly tuned , and often nested , internal regulatory feedback control mechanisms , such as the baroreflex in our simple example ., In situations where such a controlled variable is observed , the ambiguity in the mapping from observation to parameter space is naturally exacerbated since perturbations of the observable will be compensated by alterations of other , possibly unobserved system states , as has already been proposed in a neurophysiological context 35 ., Since a living organism is a system that maintains a state of dynamical equilibrium at energy expenditure , this phenomenon is likely to be the rule rather than the exception ., A more precise formulation of this qualitative observation is the subject of current investigation ., We believe that our approach provides a conceptually new quantitative framework for a theoretical description of the development of differential diagnoses , which may potentially be harnessed to improve this process ., Eventually , this methodology could be extended to an outcome prediction tool and could help to optimize diagnostic and therapeutic interventions in individual patients ., Its practical implementation will require broad interdisciplinary collaborations , because of the significant challenges involved ., We nevertheless believe that the potential gains in diagnostic effectiveness and efficiency that can be made by taking a quantitative approach to uncertainty , based on our ever-growing mechanistic understanding of physiology , will make the effort worthwhile ., The model we developed was designed to be computationally and conceptually simple while achieving a good qualitative reproduction of system responses to alterations in contractility and hydration status ., It consists of a continuous representation of the monoventricular heart as a pump , connected to a representation of the systemic circulation with the large blood vessels treated as linear capacitors ( Windkessel model ) and with arterial pressure controlled by a physiological feedback loop ( baroreflex 39 , Figure 2 ) ., The pulmonary circulation is excluded for simplicity , since the perturbations to be studied in our example are not directly related to it ., The physiological variables and parameters used in the following exposition are summarized in Table 1 ., The heart as a pump ., At a basic level , the heart acts analogously to a piston pump ., During each heartbeat , blood from the venous side of the circulation fills the ventricle ( “piston” ) during the filling phase ( diastole ) ., When the cardiac rhythm generator ( anatomically , the sinoatrial node ) triggers myocardial contraction , the heart starts to contract , increasing the pressure in the ventricle ., This process leads to the ejection of blood toward the arterial side of the circulation , and thus emptying of the ventricle , as soon as intraventricular pressure exceeds the pressure on the arterial side of the circulation ., Simplifying the underlying physiology somewhat , one key factor for the amount of volume entering the ventricle in diastole is the so-called preload , which is related to the pressure in the large veins immediately upstream of the heart ., How much blood is ejected during systole depends on the so-called afterload , corresponding to the pressure in the large arteries downstream of the heart , the strength of the contraction of the heart muscle ( myocardial contractility ) , and the extent to which the ventricle was filled during diastole ., The amount of force the myocardium can develop depends on its current level of stretch , which gives rise to a relationship between the amounts of filling during diastole and ejection during systole , termed the Starling mechanism ., We developed an ODE model of the monoventricular heart , omitting pulmonary circulation , by considering a single-cycle representation of the emptying ( ejection ) and filling of the ventricle ., Systole ., The model of ejection was based on the experimentally observed linearity of the relationship between stroke work WS , which refers to the work performed by the heart during ejection , and end-diastolic volume VED over a wide range of volumes 40 ., This linear relation takes the form, where the slope factor cPRSW is termed the preload recruitable stroke work ., The volume axis intercept of this relationship has been found to be equivalent to the volume at which the passive intraventricular pressure is 0 mm Hg ,, 40 ., Approximating stroke work as pure volume work performed from VED to the end-systolic ventricular volume VES against the arterial pressure Pa yields, where PED is the intraventricular pressure at the end of diastole , and, is the stroke volume ., Based on the finding that the ventricular volume will not usually decrease below, , we define the end-systolic volume as a function of the end-diastolic volume as follows:, The expression for, is obtained by equating the expressions for WS in Equations 1 and 2 , solving for VS , and substituting the result into Equation 3 to obtain, Note that, is a continuous function of VED since , if, , the limit of, as PED approaches Pa from below is smaller than, ., Diastole ., To complete one stroke cycle , we derive an expression for the end-diastolic volume as a function of the end-systolic volume ., Ventricular filling is modeled as a simple passive filling through the linear inflow resistance Rvalve , driven by the difference between pressure in the central veins PCVP and ventricular pressure PLV ( VLV ) , through the ODE, In Equation 6 , the dependence of ventricular pressure on ventricular volume is governed by the experimentally characterized 40 exponential relationship:, Under the assumption of constant PCVP , Equation 6 is of the general form, with constants, which resolves to, by quadrature ., By letting t = 0 at the beginning of diastole and eliminating the unknown constant C using end-systolic volume VES as the initial condition , we obtain, where the final expression is numerically advantageous since it avoids floating point overflow in the exponential terms ., At a given heart rate fHR , and assuming an approximately constant duration of systole TSys ( physiologically , the duration of diastole is much more strongly affected by alterations in heart rate than the duration of systole 41 ) , the end-diastolic volume will therefore be, with V ( t ) given by Equation 11 ., If PCVP exceeds the intraventricular pressure at the beginning of diastole , then passive filling can occur and Equation 12 provides the desired expression for VED as a function of VES , through the VES-dependency of Equation 11 ., Otherwise , no filling will occur ., The overall expression for VED as a function of VES is thus, note that, is a continuous function of VES since the limit of, as PLV ( VES ) approaches PCVP from below is VES ., Joining systole and diastole ., We can now define a discrete dynamical system describing the beat-to-beat evolution of VED ( or , similarly , VES ) ., Specifically , given the current end-diastolic volume, , we can use Equation 4 to compute, and use Equation 13 to obtain, ., Together , these steps yield, To obtain a continuous dynamical system amenable to coupling with continuous representations of the physiologic control loops and simulation with available ODE software over long time intervals , we converted VES and VED to state variables of a continuous time system ., This was done by setting their rates of change to the average rates of change over an entire cardiac cycle that would occur during one iteration of the discrete time system for the current VES and VED values , to obtain, The discrete system ( Equation 14 ) and the continuous system ( Equation 15 ) share identical sets of fixed points ., Indeed , fixed points of the discrete system ( Equation 14 ) are given by, and by applying, to Equation 16 , we have, thus, at fixed points of the discrete system ., Conversely , inspection shows that fixed points of Equation 15 satisfy Equation 16 and hence are fixed points of Equation 14 as well ., The relationship between the stability of fixed points of Equation 14 and the stability of fixed points of Equation 15 is not obvious , since the discrete system ( Equation 14 ) treats systole and diastole sequentially , while VES and VED co-evolve under Equation 15 ., However , linearization shows that any fixed point, is stable with respect to both systems if, , whereas the fixed point is unstable with respect to both systems if, ., ( The stability condition can be evaluated at all points in ( VES , VED ) space , except for on the finite collection of lines where either derivative fails to exist ( see Equations 4 and 13 ) ., In the case that, , the fixed point is stable with respect to the continuous time system and unstable with respect to the discrete , suggesting that the continuous approximation in theory has the potential to eliminate instabilities inherent in the beat-to-beat dynamics ., However , preliminary numerical explorations suggest that, is significantly smaller than one for the relevant parameter range ., When the continuous system ( Equation 15 ) is embedded in the complete circulation model ( described below ) , the full system very quickly settles to a stable fixed point ( Figure 8 ) ., The systemic circulation ., The circulation is represented by a simple Windkessel model ., It consists of linear compliances representing the large arterial vessels of volume Va and venous vessels of volume Vv with respective pressures, where α is “a” or “v” and where, is the respective unstressed volumes , i . e . , the in general non-zero volume at which the pressure in the respective compartment will be 0 mm Hg ., These pressures appear in the equation that links the arterial and venous compartments through a linear resistor , representing the total peripheral resistance RTPR regulating the arterio–venous capillary blood flow IC , namely, The veno–arterial flow ( cardiac output ) ICO generated by the heart is given by the product of the heart rate fHR and the volume VS ejected per beat , which by Equation 3 takes the form, Assuming conservation of volume at the nodes , the evolution of arterial and venous volumes is described by the following differential equations:, where Iexternal represents a possible external blood withdrawal or fluid infusion to or from the venous compartment ., Baroreflex control of blood pressure ., Baroreflex control of blood pressure , which is one of the key regulatory mechanisms in cardiovascular homeostatis , is implemented based on the established representation of the central processing component of the baroreceptor sensor input as a combination of a sigmoidal nonlinearity ( logistic function , in our case ) with a linear system 15 , 18 ., For simplicity , we reduced baroreflex activity to a single activating ( sympathetic ) output instead of the more physiologically accurate balance of stimulating ( sympathetic ) and inhibiting ( parasympathetic ) outputs ., Since our model of the heart is designed to represent timescales significantly larger than a single beat , the linear part of the baroreflex feedback loop is simplified to display first-order low-pass characteristics with a time constant on the order of the slowest actuator response ( unstressed venous volume control ) ., Pure delays associated with the neural transmission of baroreflex signals are neglected ., Under these assumptions , the temporal evolution of the stimulating output from baroreflex central processing is governed by the differential equation, The stimulating output S ( t ) of the feedback loop acts on heart rate fHR , total peripheral resistance RTPR , myocardial contractility cPRSW , and unstressed venous volume, effectors/actuators to adjust blood pressure according to its current deviation from the set point , based on the linear transformations, where α = fHR , RTPR , or cPRSW , and, The form of Equation 23 , in particular , arises since the venous capacitance vessels contract , reducing their unstressed volume , in response to drops in blood pressure ., Combining Equations 15 and 17– 23 , and writing out dependencies relevant to the coupling of the system explicitly , we obtain a system of five ODEs:, It should be noted that for Iexternal = 0 , conservation of total intravascular volume would allow for elimination of one state variable ( either Va or Vv ) to obtain a four-dimensional system ., We chose to leave the system in the above form , however , to preserve direct correspondence between anatomical entities and mathematical representation , at the cost of some loss in computational efficiency ., With regard to the coupling between equations , it should be noted that the sympathetic nervous system activity S , which serves as a central control mechanism critical for functional cardiovascular system homeostasis , links together all components , while the coupling between the equations describing heart and circulation reflects the cyclical structure of the cardiac action and the circulation ., Numerical solution of system Equation 24 , as well as all other algorithms used for this work , was implemented in the MATLAB 7 ( The MathWorks ) programming environment , using the ode15s solver for numerical integration ., The source code used in generating results is available in Text S1 ., Parameter selection ., The parameters, = 2 . 03 mm Hg ,, = 7 . 14 ml , and, = 0 . 066 ml−1 describing the ventricular pressure–volume relationship were estimated from experimental data for the left ventricle from 40 using the Levenberg-Marquard nonlinear least squares algorithm ( Figure 9 ) ., The remaining parameter values and ranges for variables , as well as the respective sources , are given in Table 2 ( for meaning of parameters , see Table 1 ) | Introduction, Results, Discussion, Methods | The improved capacity to acquire quantitative data in a clinical setting has generally failed to improve outcomes in acutely ill patients , suggesting a need for advances in computer-supported data interpretation and decision making ., In particular , the application of mathematical models of experimentally elucidated physiological mechanisms could augment the interpretation of quantitative , patient-specific information and help to better target therapy ., Yet , such models are typically complex and nonlinear , a reality that often precludes the identification of unique parameters and states of the model that best represent available data ., Hypothesizing that this non-uniqueness can convey useful information , we implemented a simplified simulation of a common differential diagnostic process ( hypotension in an acute care setting ) , using a combination of a mathematical model of the cardiovascular system , a stochastic measurement model , and Bayesian inference techniques to quantify parameter and state uncertainty ., The output of this procedure is a probability density function on the space of model parameters and initial conditions for a particular patient , based on prior population information together with patient-specific clinical observations ., We show that multimodal posterior probability density functions arise naturally , even when unimodal and uninformative priors are used ., The peaks of these densities correspond to clinically relevant differential diagnoses and can , in the simplified simulation setting , be constrained to a single diagnosis by assimilating additional observations from dynamical interventions ( e . g . , fluid challenge ) ., We conclude that the ill-posedness of the inverse problem in quantitative physiology is not merely a technical obstacle , but rather reflects clinical reality and , when addressed adequately in the solution process , provides a novel link between mathematically described physiological knowledge and the clinical concept of differential diagnoses ., We outline possible steps toward translating this computational approach to the bedside , to supplement todays evidence-based medicine with a quantitatively founded model-based medicine that integrates mechanistic knowledge with patient-specific information . | Although quantitative physiology has developed numerous mathematical descriptions of components of the human body , their application in clinical medicine has been limited to a few mostly primitive and physiologically inaccurate models ., One reason for this is that the inverse problem of identifying unknown model parameters and states from prior knowledge and clinical observations does not usually have a unique solution ., Hypothesizing that this non-uniqueness might actually convey clinically useful information , we used a simplified mathematical model of the cardiovascular system and its control , in combination with Bayesian inference techniques , to simulate the diagnosis of low blood pressure in an acute care setting ., The inference procedure yielded a distribution of physiologically interpretable model parameters and states that exhibited multiple peaks ., The key observation was that these peaks corresponded directly to clinically relevant differential diagnoses , enabling a quantitative , probabilistic assessment of the simulated patients condition ., We conclude that the proposed probabilistic approach to the inverse problem in quantitative physiology may not only be useful for quantitative interpretation of clinical data , and eventually allow model-based prediction and therapy optimization , but also provides a novel link between mathematically described physiological mechanisms and the clinical concept of differential diagnoses based on patient-specific information . | mathematics, physiology, dog, computational biology, homo (human), cardiovascular disorders | null |
journal.pgen.1001160 | 2,010 | Role for the Mammalian Swi5-Sfr1 Complex in DNA Strand Break Repair through Homologous Recombination | Homologous recombination ( HR ) is a key pathway in mammalian cells for the repair of several types of lesions , including DNA strand breaks ., Its importance is emphasized by the sensitivity of HR mutants to a variety of DNA damaging agents , as well as the loss of genomic integrity seen in these mutants arising from DNA damage ., As a result , HR is a critical DNA repair pathway during development and for tumor suppression 1 , 2 ., Double-strand breaks ( DSBs ) arise in DNA as a result of both endogenous cellular processes and from exogenous sources 3 , 4 ., HR is a precise pathway for the repair of DSBs , during which homologous sequence information is copied from an intact donor template 1 , 2 , most frequently the sister chromatid during late S/G2 in mitotic cells 5 ., A second key pathway for the repair of DSBs is nonhomologous end-joining ( NHEJ ) , where two ends are joined with little or no sequence identity 6 ., In addition to canonical two-ended DSBs , one-ended DSBs also arise in DNA 7 ., These lesions form when a replication fork encounters a DNA single-strand break that is not repaired by base excision repair , for example , from a covalent topoisomerase I-DNA intermediate as a result of exposure to camptothecin 8 , 9 ., HR is the primary mechanism for the repair of one-ended DSBs , given that the joining of two unrelated one-ended DSBs by NHEJ would give rise to genomic rearrangements 7 ., Many of the known HR factors in mammalian cells , including the central Rad51 protein , have been identified by their homology to yeast HR factors 2 , 10 , 11 ., Rad51 , the eukaryotic homologue of Eschericia coli RecA , binds to single-stranded DNA to form a nucleoprotein filament which catalyzes base pairing and strand exchange between homologous DNAs 12–14 ., Single-stranded DNA is formed at DNA ends by resection 15; although a substrate for Rad51 filament formation , single-stranded DNA is also bound by replication factor A ( RPA ) , which binds at high affinity and removes secondary structure 16 , 17 ., While critical for the initiation of HR 18 , RPA interferes with Rad51 loading onto single-stranded DNA ., Several factors , referred to as “mediators” , are required to overcome the inhibition by RPA to facilitate Rad51 nucleoprotein filament formation 19 ., Proposed mediators in yeast include the Rad51 paralogues , Rad55-Rad57 , and Rad52 20 , 21 ., Vertebrates have five Rad51 paralogues , of which a complex of two have been shown to have mediator activity in vitro 22 ., Additionally , the breast cancer suppressor BRCA2 , for which there is no homologue in budding or fission yeast , has been proposed to have mediator activity 23 ., BRCA2 may also function to stabilize Rad51 filaments on single-stranded DNA , by inhibiting ATP hydrolysis while preventing the formation of non-productive filaments on double-stranded DNA 24 ., A distinct complex that functions in fission yeast HR is Swi5-Sfr1 ., Mutation of either Swi5 or Sfr1 results in reduced HR in both mitotic and meiotic cells 25–28 ., Like other HR mutants , Swi5 and Sfr1 mutants have elevated sensitivity to a number of DNA damaging agents , including ionizing radiation , UV , and methyl-methanesulfonate 29 ., In vitro , the Swi5-Sfr1 complex binds to Rhp51 ( the fission yeast Rad51 homologue ) in an Sfr1-dependent manner 29 , 30 , and has been shown to possess mediator activity but importantly also to enhance strand exchange by Rhp51 30 ., While loss of either Swi5-Sfr1 or Rhp55-Rhp57 ( fission yeast Rad55-Rad57 homologues ) reduces HR , loss of both complexes complete abrogates Rhp51-dependent HR 26 ., Both complexes are also required during meiotic recombination 31 ., Budding yeast has a homologous complex to Swi5-Sfr1 termed Sae3-Mei5 , although this complex is only expressed during meiosis where it plays a critical role in meiotic recombination 32–34 ., Swi5 forms a second complex with an Sfr1-related protein , Swi2 , which localizes to heterochromatin at the donor mating-type loci and promotes HR during switching 26 , 29 , 35 ., In budding yeast , the function of Sae3-Mei5 appears to be limited to supporting the function of Dmc1 , the meiosis-specific RecA homologue 33 , 34 ., Previous reports suggest that both Swi5/Sae3 and Sfr1/Mei5 are evolutionarily conserved , while Swi2 is only found in fission yeast 29 , 34 ., In this study , we isolated Swi5 and Sfr1 homologues from mice ., Swi5 and Sfr1 form a complex in vivo and in vitro , and Rad51 binding to Swi5 is detected in vitro in GST-pull down assays , suggesting that the Swi5-Sfr1 complex has a conserved function in mouse ., To investigate their function in vivo , we generated Swi5−/− and Sfr1−/− mouse embryonic stem ( ES ) cell lines ., Although loss of either Swi5 or Sfr1 did not decrease HR frequency by itself , HR was perturbed to a greater extent in these cells by expression of a BRC peptide from BRCA2 ., Interestingly , Swi5−/− and Sfr1−/− cells were sensitive to ionizing radiation , camptothecin , and an inhibitor of poly ( ADP-ribose ) polymerase ( Parp ) , all of which cause strand breaks ., The induction of sister chromatid exchanges ( SCE ) by Parp inhibition was attenuated in the Swi5 and Sfr1-deficient cell lines; moreover , Parp inhibition resulted in increased chromatid breaks and radial chromosomes in Swi5−/− and Sfr1−/− cells ., Thus , Swi5 and Sfr1 have an important role in the maintenance of genomic integrity in mammalian cells , in particular in the repair of DNA strand breaks ., Based on amino acid conservation , putative mammalian homologues of Swi5/Sae3 and Sfr1/Mei5 have previously been reported 29 , 34 ., We cloned the mouse homologues , 2900010J23Rik for Swi5 and 6330577E15Rik for Sfr1 , based on existing database information ( http://www . informatics . jax . org/ and http://uswest . ensembl . org/index . html ) ., The Sfr1 cDNA was successfully amplified by PCR following reverse transcription ( RT-PCR ) of RNA obtained from mouse ES cells ., Sequence analysis of the Sfr1 cDNA revealed that the Sfr1 protein is 303 amino acids and is encoded by four exons ( Figure 1A , Figure S1A and S1B ) ., Ectopic expression of the cloned cDNAs complemented the phenotypes of Sfr1−/−cell lines ( see below ) ., Swi5 cDNAs were also obtained by RT-PCR of RNA from mouse ES cells ., Two differentially spliced forms were detected containing alternative first exons which encoded proteins of 89 and 121 amino acids ( Figure S2A and S2C ) ., When expressed in ES cells , we found that both forms migrated at a lower molecular weight than the endogenous protein ( Figure S2B ) ; attempts to clone a cDNA expressing a larger protein were unsuccessful , possibly because the 5′ end of the mRNA contains a structure which impedes amplification or a non-AUG initiation codon ., Nevertheless , both forms complemented the phenotypes of Swi5−/− cells ( see below and data not shown ) ., In subsequent experiments , we used the Swi5 cDNA encoding the 89 amino acid protein ( Figure 1A ) ., Overall , the sequence identities between mouse and fission yeast proteins were 28 . 6% ( Swi5 ) and 20 . 9% ( Sfr1 ) ., Significant variation was noted between the N-terminus of the various Sfr1 orthologues , even among mouse strains ., Mouse Sfr1 has a proline-rich repeat of 16 amino acids at its N-terminus , which we named the RSfp motif ( rodent Sfr1 proline rich motif ) ( Figure 1A and Figure S1B ) ., In the mouse ES cells used in this study ( E14 ) and in DBA/2J mice ( Q3TI03 ) , there are five repeats of the RSfp motif , whereas in C57BL/6J mice there are six repeats ., The rat Sfr1 homologue ( rCG57555 ) has two repeats ., Repetition of the RSfp motif appears to be unique to rodents , as only a single RSfp motif is present in other mammals , including human , rabbit , dog and pig ( Figure S1C ) ., The RSfp motif is not present outside of mammals , although the downstream region is conserved ( Figure S1D ) ., In fission yeast and in budding yeast , Swi5/Sae3 and Sfr1/Mei5 form a stable complex in vivo and in vitro ., To determine whether mouse Swi5 and Sfr1 interact , we performed a yeast two-hybrid assay ( Figure 1B ) ., Swi5 fused to the Gal4 activation domain ( AD ) and Sfr1 fused to the Gal4 DNA binding domain ( DBD ) gave a positive interaction , suggesting a physical association between the two ., The reverse test was uninformative as Swi5 fused to the Gal4DBD itself allowed growth on the test medium ., In addition , Sfr1 showed self-association , which has also been observed in fission yeast 29 ., We also tested complex formation with a GST pull-down assay using recombinant proteins expressed in E . coli ., Unlike expression of yeast Sfr1/Mei5 , which yields insoluble protein without co-expression of Swi5/Sae3 30 , , mouse His6-Sfr1 was soluble by itself ( Figure 1C ) ., The tagged Sfr1 migrated at a higher molecular weight ( ∼50 kDa ) than the molecular weight calculated from the amino acid sequence ( 36 kDa ) ., An unexpected lower mobility was seen with the endogenous Sfr1 protein ( see below , Figure 2C ) ., The E . coli extract expressing His6-Sfr1 was incubated with GST-Swi5 or GST alone immobilized on magnetic beads ., Pull-down of GST-Swi5 , but not GST , brought down His6-Sfr1 ( Figure 1C ) , again indicating a physical association between Swi5 and Sfr1 ., To determine their interacting domains , two-hybrid and GST pull-down assays were performed with N and C-terminal fragments from both proteins ( Figure 1A ) ., The N-terminal half of Swi5 , but not the C-terminal half , interacted with Sfr1 in both assays ( Figure 1D and 1F ) ., Conversely , the C-terminal fragment of Sfr1 , but not the N-terminal fragment , interacted with Swi5 ( Figure 1E and 1G ) ., The interacting fragments from both proteins contain coiled-coil motifs ( Figure 1A ) , which may be responsible for the interaction ., Consistent with their variability in different species , the RSfp motifs of Sfr1 did not appear to play a role in the interaction ., Co-immunoprecipitations were performed with mouse ES cell extracts to investigate the interaction in vivo ., Using antibodies directed against the endogenous proteins , Swi5 co-precipitated Sfr1 and Sfr1 co-precipitated Swi5 ( Figure 1H ) ., Most of the Swi5 in the cell seems to be in a complex with Sfr1 ., Thus , despite the poor sequence conservation overall , Sfr1 is a major interacting partner for the Swi5 in the cell , consistent with the better conservation of the Sfr1 C-terminal portion , which interacts with Swi5 ., To investigate their cellular functions , we generated Swi5 and Sfr1-deficient mouse ES cell lines ., The Swi5 targeting vector was designed to replace the exons 3 and 4 with a neomycin resistance gene ( neo ) , resulting in deletion of most of the Swi5 coding sequence , including the sequence for Sfr1 interaction ( Figure 2A and Figure S3A ) ., In the Sfr1 targeting vector , the neo gene replaced exons 2 and 3 , which encode amino acids 5 to 240 , removing 78% of the coding region ( Figure 2B and Figure S3B ) ., Two rounds of gene targeting were performed , with an intervening step to delete the neo gene from the first targeted allele using Cre recombinase ( Figure S3A and S3B ) ., Successful gene targeting of both Swi5 and Sfr1 alleles in the respective cell lines was confirmed by Southern blotting ( Figure S3A and S3B ) , and loss of protein was confirmed by Western blotting ( Figure 2C ) and immunofluorescence ( Figure 2F ) ., The Swi5−/− and Sfr1−/− cell lines ( formally Swi5Δ/neo and Sfr1Δ/neo , respectively ) exhibited similar proliferation kinetics and cell cycle distribution as wild-type cells ( Figure 2D and 2E ) , indicating that Swi5 and Sfr1 are not essential for cell viability ., As Swi5 and Sfr1 form a complex , we determined whether loss of one affects the stability of the other by Western blotting ( Figure 2C ) ., Swi5 protein was not detectable in Sfr1−/− cells , indicating that the stability of Swi5 requires association with Sfr1 ., The level of Sfr1 in Swi5−/− cells was also diminished , although the protein was still detectable ., These results provide further evidence for a physical association between Swi5 and Sfr1 in vivo ., To determine the sub-cellular localization of Swi5 and Sfr1 , mouse ES cells and embryonic fibroblasts ( MEFs ) were examined by immunofluorescence ., Swi5 and Sfr1 localized to the nucleus in both cell types ( Figure 2F and 2G ) ., Importantly , Swi5 was not detected in Sfr1−/− ES cells , providing further support that Swi5 is unstable without Sfr1; Sfr1 was detectable in Swi5−/− ES cells , albeit weakly ( Figure 2F ) ., In fission yeast , Swi5 localizes to heterochromatin as well as to euchromatin 26 ., However , neither mouse protein specifically localized to heterochromatin , as marked by trimethyl-lysine 9 of histone H3 and intense DAPI staining ( Figure 2G ) ., Rather , both proteins had a more widespread nuclear distribution that was , nonetheless , somewhat granular ., The fission yeast Swi5-Sfr1 complex interacts with the Rad51 recombinase through the Sfr1 subunit 29 , 30 ., We tested whether Rad51 interaction would be conserved with the mouse proteins by co-immunoprecipitation from ES cell extracts ., Neither Swi5 nor Sfr1 precipitated detectable amounts of Rad51 from either untreated ( Figure 1H ) or γ-irradiated cells ( data not shown ) ., To investigate this further , GST pull-down assays were performed with recombinant Rad51 expressed in E . coli ( Figure 3A ) ., Pull-down of Rad51 was detected with GST-Swi5 , but not with GST-Sfr1 or GST alone ( Figure 3A ) ., Treatment of the extracts with ethidium bromide or DNase I did not affect the association between GST-Swi5 and Rad51 ( Figure S4A ) ., These results indicate a physical association between Swi5 and Rad51 ., We tested whether Swi5 and Sfr1 co-localize with Rad51 in nuclear foci after X-irradiation ., Unlike Rad51 , Swi5 and Sfr1 were distributed throughout the nucleus , as in untreated cells , indicating that there was no specific recruitment of these proteins to DSB sites ( Figure S4B ) ., Further , Rad51 focus formation after X-irradiation was not noticeably affected in either Swi5−/− and Sfr1−/− cells ( Figure S4C ) ., The conservation of the protein complex and the interaction with Rad51 suggested that Swi5-Sfr1 could play a role in HR in mammalian cells ., We examined HR levels in the Swi5−/− and Sfr1−/− ES cells using the DR-GFP assay 37 ( Figure 3B ) ., In this assay , a single DSB is introduced into the chromosomally integrated DR-GFP substrate by the I-SceI endonuclease; repair of the DSB by HR gives rise to cells expressing functional GFP ., After I-SceI expression , Swi5−/− and Sfr1−/− cells gave similar levels of GFP positive cells ( 4 . 9% and 4 . 4% , respectively ) as wild-type cells ( 5 . 2%; Figure 3C ) , indicating that Swi5 and Sfr1 are not essential for HR in mouse cells ., In fission yeast , the Swi5-Sfr1 complex stabilizes Rad51 filament formation on single-stranded DNA 38 ., We hypothesized that if Rad51 nucleoprotein filaments were perturbed in mouse cells , a role for the Swi5-Sfr1 complex in HR might be uncovered ., BRCA2 is a central HR protein in mammalian cells , binding Rad51 at a series of repeats ∼35 amino acids ( BRC repeats ) ; as an isolated peptide , the BRC repeat has been demonstrated to bind Rad51 , to inhibit Rad51 focus formation 39–41 and , importantly , to decrease HR in mammalian cells 42 ., Compared to cells transfected with an empty expression vector ( 5 . 8% ) ( Figure 3D ) , expression of BRC3 in wild-type cells resulted in a significantly reduced frequency of GFP positive cells ( 0 . 55% ) and hence HR ( Figure 3E ) , consistent with previous results ., This inhibitory effect on HR was not observed with expression of the BRC3Δ peptide which is unable to bind Rad51 43 ( 6 . 6%; Figure 3F ) and , further , was rescued by Rad51 overexpression ( data not shown ) ., With BRC3 expression , Swi5−/− and Sfr1−/− cells exhibited a 2 . 1-fold and 1 . 9-fold reduction of GFP positive cells ( 0 . 26% and 0 . 29% , respectively ) compared to wild-type cells ( Figure 3E ) , indicating that Swi5-Sfr1 plays a role in HR when it is compromised ., Consistent with this interpretation , expression of the cognate cDNAs complemented the HR defect ( Figure 3E ) ., The defect in HR was dependent on the ability of the BRC3 repeat to bind Rad51 , as a similar number of GFP positive cells were obtained with BRC3Δ expression ( Figure 3F ) ., These results indicate that the Swi5 and Sfr1 function in HR , but are not required for the process unless it is already compromised ., Because BRC3 perturbs Rad51 focus formation , mouse Swi5-Sfr1 may play a role in stabilizing Rad51 filaments , as in fission yeast ., Given the HR phenotype associated with these cells , we next examined the sensitivity of Swi5−/− and Sfr1−/− cells to DNA damaging agents ., In these assays , Brca2lex1/lex2 cells were included for comparison , as they are known to be defective in HR 44 ., Swi5−/− and Sfr1−/− cells were found to be more sensitive to X-rays than wild-type cells , although their sensitivity was less pronounced than that of Brca2lex1/lex2 cells ( Figure 4A ) ., Expression of the Swi5 or Sfr1 cDNA in the respective mutant cells restored survival to the level observed in wild-type cells , demonstrating that the sensitivity was specifically due to the deletion of Swi5 or Sfr1 ., Cells were also exposed to topoisomerase poisons , which like X-rays lead to strand breaks ., Both Swi5−/− and Sfr1−/− cells exhibited sensitivity to the type I topoisomerase poison camptothecin , although not as severely as Brca2lex1/lex2 cells ( Figure 4B ) ., Interestingly , Sfr1−/− cells were somewhat more sensitive to camptothecin than Swi5−/− cells ., Sfr1−/− cells were also sensitive to the type II topoisomerase poison etoposide ., The two mutants again showed differential sensitivity , with Sfr1−/− cells showing a more severe phenotype ., In this case , Sfr1−/− cells were even more sensitive than Brca2lex1/lex2 cells , whereas Swi5−/− cells were no more sensitive than wild-type cells ( Figure 4C ) ., These results suggest that Swi5 and Sfr1 have a function in repairing DNA strand breaks , the primary lesions from X-irradiation and topoisomerase poisons ., Given the greater sensitivity observed in Sfr1−/− cells , they also indicate that the roles of Swi5 and Sfr1 are not equivalent in the cell ., Cells with defective DNA damage checkpoints often exhibit sensitivity to DNA damaging agents ., Chk1 and Chk2 are two proteins that are phosphorylated upon X-irradiation 45 , 46 ., After X-irradiation , Swi5−/− and Sfr1−/− cells were proficient at phosphorylation of both proteins and showed similar kinetics ( Figure 4D ) ., Checkpoint-proficient cells also arrest after DNA damage rather than proceed into mitosis ., Mitotic populations were reduced to a similar extent in Swi5−/− and Sfr1−/− cells as in wild-type cells ( Figure 4E ) ., These results point to intact DNA damage checkpoints in both mutants ., We also tested the sensitivity of Swi5−/− and Sfr1−/− cells to a variety of other DNA damaging agents ., HR mutants are typically sensitive to interstrand crosslinking agents 47–49 , yet we observed that Swi5−/− and Sfr1−/− cells were not any more sensitive to either mitomycin C or cisplatin than wild-type or complemented cells ( Figure S5A and S5B ) ., In addition , cells were not sensitive to the replication inhibitor hydroxyurea ( Figure S5C ) , implying that the camptothecin sensitivity is specifically related to strand breaks generated by this agent rather than indirectly to problems with replication per se ., Finally , neither mutant was sensitive to ultraviolet light ( Figure S5D ) , indicating that the proteins do not play a role in nucleotide excision repair ., Interestingly , Brca2lex1/lex2 cells were found to be sensitive , suggesting a role for HR repair of UVC lesions ., Poly ( ADP-ribose ) polymerase ( Parp ) plays an important role in the repair of DNA single-strand breaks , such that inhibition of Parp activity leads to the accumulation of the unrepaired single-strand breaks that turn into DSBs when encountered by replication forks ., Since the repair of DSBs arising during replication largely depends on the HR pathway , cells deficient in HR are extremely sensitive to Parp inhibitors 50 , 51 ., To further investigate the effects of Swi5 and Sfr1 deficiency on the repair of DNA strand breaks , Swi5−/− and Sfr1−/− cells were exposed to the Parp inhibitor olaparib ., Consistent with previous reports , Brca2 mutant cells were exquisitely sensitive to olaparib; by contrast , the NHEJ mutant Ku70−/− was not ( Figure 5A ) ., Swi5−/− and Sfr1−/− cells were also significantly more sensitive to olaparib than wild-type cells , although not as sensitive as Brca2lex1/lex2 cells ( Figure 5A ) ., This sensitivity was suppressed by introducing the cognate cDNAs into the Swi5−/− and Sfr1−/− cells ( Figure 5A ) ., Sensitivity of the cell lines to Parp inhibition further implicates Swi5 and Sfr1 in the repair of DNA strand breaks ., To further examine the effect of Parp inhibition on Swi5 and Sfr1-deficient cells , chromosomes were examined for aberrations in metaphase spreads ., In Swi5−/− and Sfr1−/− cells , chromatid breaks were elevated 30 and 20-fold , respectively , after exposure to olaparib compared with untreated cells , significantly more than that observed in wild-type cells ( 9-fold; Figure 5B ) ., Radial chromosomes , which were not observed in untreated cells , were also induced in Swi5−/− and Sfr1−/− cells ., Both of these types of aberrations typically arise from problems encountered during DNA replication ., Brca2lex1/lex2 cells showed a substantial number of chromatid breaks even without olaparib , but chromatid breaks increased and radial chromosomes were observed upon olaparib treatment ., The level of aberrations in olaparib-treated Brca2lex1/lex2 cells was similar to that found in the treated Swi5−/− and Sfr1−/− cells , but aberrations may be underestimated if the G2/M checkpoint was activated ., The observation of increased chromatid breaks and radial chromosomes in Swi5−/− and Sfr1−/− cells suggest that unrepaired DSBs accumulate , which may be responsible for the toxicity observed with Parp inhibition in these cells ., The accumulation of chromatid breaks induced by Parp inhibition may be the result of HR deficiency ., To test this , we examined sister-chromatid exchange ( SCE ) , which is one of outcome of HR ( Figure 5C ) ., The spontaneous SCE frequency was similar among wild-type , Swi5−/− and Sfr1−/− cells ( 9 . 3 , 9 . 3 and 9 . 7 SCEs per metaphase , respectively ) , while Brca2lex1/lex2 cells showed a lower frequency of SCE ( 7 . 1 SCEs per metaphase ) ., With Parp inhibition , SCEs were significantly induced in wild-type cells ( 41 . 1 SCEs per metaphase ) as well as in Brca2lex1/lex2 cells , although the overall level was lower ( 30 . 1 SCEs per metaphase ) ., In Swi5−/− and Sfr1−/− cells , the overall level of SCEs was reduced compared with wild type ( 35 . 0 and 35 . 4 SCEs per metaphase , respectively ) ., These results indicate that SCE induction by Parp inhibition is partially dependent on Swi5 and Sfr1 ., In this study , we identified Swi5 and Sfr1 orthologues in mammalian cells and determined that they have critical roles in the repair of DNA strand breaks ., Despite their low conservation with the respective yeast proteins , we found that mouse Swi5 and Sfr1 form a complex in vivo and in vitro , as do fission yeast Swi5 and Sfr1 and budding yeast Sae3 and Mei5 29 , 30 , 34 ., The integral nature of the protein-protein interactions is emphasized by the mutual interdependence of the Swi5 and Sfr1 for stability , and by the finding that Sfr1 co-immunoprecipitates Swi5 to a similar extent as immunoprecipitation of Swi5 itself ., Although the budding yeast complex is only expressed during meiosis 32 , 34 , mouse Swi5-Sfr1 is expressed in mitotically dividing cells , making it more akin to the fission yeast complex ., We found that Swi5 or Sfr1-deficient mammalian cells are sensitive to agents that cause DNA strand breaks , including X-rays , camptothecin , and the Parp inhibitor olaparib ., Consistent with a DNA damage repair defect in Swi5−/− and Sfr1−/− cells , chromosome aberrations are increased compared to wild-type when cells are challenged with olaparib ., For the most part , the sensitivities of Swi5−/− and Sfr1−/− cells are similar to each other , although unlike Swi5−/− cells , Sfr1−/− cells are also sensitive to etoposide ., In contrast to Swi5 , the stability of Sfr1 is not fully compromised when its partner protein is absent , consistent with Sfr1 functions that are independent of Swi5 in some contexts , as is the case with fission yeast 52 , 53 ., While fission yeast Swi5 acts independent of Sfr1 during mating-type switching , mouse Swi5 is unlikely to have Sfr1-independent functions , given its instability in the absence of Sfr1 ., Sensitivity to camptothecin and olaparib is consistent with a defect in the ability to repair DNA damage by HR ., That Swi5 interacts with Rad51 in vitro , the critical strand exchange protein for HR reactions , supports a role for the mammalian Swi5-Sfr1 complex in HR , like the cognate complexes in fission and budding yeast 26 , 33 , 34 ., Further , DNA damage-induced SCEs are reduced in Swi5−/− and Sfr1−/− cells compared with wild-type cells ., Moreover , although direct assay of DSB-induced HR in these cells did not reveal an intrinsic HR defect , a more severe defect in HR is observed in both the Swi5−/− and Sfr1−/− cells when HR is compromised by interfering with Rad51 function ., Unlike typical mammalian HR mutants , however , Swi5−/− and Sfr1−/− cells are not sensitive to interstrand crosslinking agents or the replication inhibitor hydroxyurea ., Although both agents lead to DSBs during S phase and induce HR , DSBs are detected by pulse field gel electrophoresis only after prolonged incubation with these agents and require the structure-specific nuclease Mus81 for their formation 54 , 55 ., By contrast , when a replication fork encounters a single-strand break , a one-ended DSB is generated with fast kinetics , as DSBs appear within 30 min after camptothecin exposure during S phase 56 ., In fission yeast , evidence points to a role for Swi5-Sfr1 ( or Swi5-Swi2 ) acting specifically at one end of a DSB or at the one-ended DSBs at the mat locus during either mating-type switching or sister chromatid recombination in donorless strains 26 , 57 , 58 ., Taken together , we propose that Swi5-Sfr1 is an evolutionarily conserved complex that acts at specific types of lesions , specifically at one-ended DSBs ., These experiments reveal a role for the mammalian Swi5-Sfr1 complex in HR ., Although Swi5-Sfr1 are required for repair when the DNA damage load is high , the role of the complex appears to be more restricted than that of BRCA2 and the Rad51 paralogues , given the more severe phenotype seen when these other proteins are deficient 37 , 44 , 59 ., In fission yeast , which does not have a BRCA2 orthologue , both Swi5-Sfr1 and the Rad51 paralogue complex Rhp55-Rhp57 are required for high levels of HR 26 ., Thus , a shift in dependence on the Swi5-Sfr1 complex may have occurred during evolution ., How might Swi5-Sfr1 function in HR ?, In vitro , the fission yeast Swi5-Sfr1 complex has mediator activity 30 ., Moreover , the fission yeast Swi5-Sfr1 complex stabilizes the Rad51 filament on single-stranded DNA 38 ., We hypothesize that the mammalian complex plays a similar role , given the reduced recombination in Swi5−/− and Sfr1−/− cells in the presence of the BRC3 repeat , which is known to perturb Rad51 focus formation 40 , 41 ., It is noteworthy , however , that the interaction of the Swi5-Sfr1 complex with Rad51 is through Swi5 , in contrast to fission yeast where the interaction with Rhp51 is through Sfr1 29 , 30 ., In both mouse cells and fission yeast , the interaction between Swi5-Sfr1 and Rad51 is detected in vitro , but not in vivo , as co-precipitation of the endogenous proteins has been unsuccessful , even under DNA damaging conditions 29 ., Thus , Swi5-Sfr1 and Rad51 may interact weakly or transiently in cells ., In fission and budding yeast , Swi5-Sfr1 and Sae3-Mei5 , respectively , bind and promote the activity of Dmc1 30 , 33 , 34 , the meiosis-specific strand exchange protein , which is also critical for mouse meiosis 60 , 61 ., Whether Swi5-Sfr1 plays a similar role in mammalian cells awaits mouse knockout studies of the complex , although notably we have detected high level of expression of the complex in the testis , including a testis-specific isoform of Swi5 ( Y . A . and M . J . , unpublished results ) ., In summary , we have characterized a novel complex critical for DNA strand break repair in mammalian cells ., The importance of strand break repair is well recognized , as defective repair is associated with various neurodegenerative diseases 62 ., Moreover , therapeutic approaches to some cancers are being developed which increase the cellular load of DNA strand breaks through Parp inhibition 63 ., The identification of Swi5-Sfr1 as being important for cellular resistance to agents like olaparib therefore has potential clinical as well as biological relevance ., Primers for Swi5 and Sfr1 cDNA cloning were designed based on annotated transcripts from the Ensembl database ., For Swi5 , forward-reverse primer pairs , YA110 ( 5′ATACCCACCCCTCCCAATAC ) -YA113 ( 5′AGTTTAAGCCCACCCCACTC ) and YA532 ( 5′ATTATTGTCGACATGGGAAGCAGGGGCGGAAC ) -YA127 ( 5′GCCGGCGGCCGCTTACTATCAGTCATTCAGGTTTAGATC ) , were designed based on annotated transcripts ENSMUSG00000044627 and ENSMUST00000113400 , respectively ., For Sfr1 , the forward-reverse primer pair , YA114 ( 5′GGCTGTGTGTACGGTGTGTC ) -YA115 ( 5′CCTCCCTCTAAGCCACAACA ) , was designed based on annotated transcript ENSMUST00000099353 ., The genomic structures presented in Figures S1A and S2A were derived by comparing the amplified cDNA sequences to the genomic structures in Ensembl ., The full length and truncated Swi5 cDNAs were cloned into the GST expression vector pGEX6P-1 ( GE healthcare ) ., GST and GST-Swi5 proteins were expressed in E . coli ( UT481 ) ., The cell lysates were obtained by sonication of cells in R-buffer ( 20 mM Tris-HCl , pH 7 . 6/1 mM EDTA/100 mM NaCl/0 . 1% Triton X-100/1 mM DTT/10% Glycerol ) followed by centrifugation at 15000 ×g for 20 min ., The expressed GST and GST-Swi5 protein in the lysates were immobilized to the MagneGST ( Promega ) ., The His6-Sfr1 and Rad51 proteins were expressed in E . coli BL21-CodonPlus ( DE3 ) from plasmids pET15b or pET21d ( Novagen ) respectively ., The lysates ( 40 µg of proteins ) obtained in R-buffer with sonication followed by centrifugation were mixed with 10 µl of the GST or GST-Swi5 protein immobilized to MagneGST , and incubated three hours at 4°C ., The precipitates were then washed three times with R- buffer and eluted by boiling in SDS-PAGE sample buffer ., The co-precipitations were subjected to SDS-PAGE with Coomassie Brilliant Blue ( CBB ) staining and to Western blotting ., The full length and truncated Swi5 and Sfr1 cDNAs were cloned to pGADT7 or pGBKT7 expression vectors to fuse to the Gal4 activation domain ( AD ) or the Gal4 DBA-binding domain ( DBD ) ., The experiments were performed according to the manufacturers instructions ( Matchmaker Two-Hybrid system 3 from Clontech ) ., The full length of Swi5 and Sfr1 cDNAs were cloned in pET15b vector ( Novagen ) ., The His6-tagged Swi5 and Sfr1 proteins , as immunogens , were expressed in E . coli ( BL21-Codonplus DE3 ) and purified using affinity to TALON ( Clontech ) ., Polyclonal antisera against Swi5 and Sfr1 were generated by Covance ., Each antiserum was affinity purified against the respective protein ., Protein extracts from mouse ES cells were obtained by lysing cells in L-buffer ( 50 mM Tris-HCl , pH 8 . 0/2 mM EDTA/125 mM NaCl/1% NP-4 | Introduction, Results, Discussion, Materials and Methods | In fission yeast , the Swi5-Sfr1 complex plays an important role in homologous recombination ( HR ) , a pathway crucial for the maintenance of genomic integrity ., Here we identify and characterize mammalian Swi5 and Sfr1 homologues ., Mouse Swi5 and Sfr1 are nuclear proteins that form a complex in vivo and in vitro ., Swi5 interacts in vitro with Rad51 , the DNA strand-exchange protein which functions during HR ., By generating Swi5−/− and Sfr1−/− embryonic stem cell lines , we found that both proteins are mutually interdependent for their stability ., Importantly , the Swi5-Sfr1 complex plays a role in HR when Rad51 function is perturbed in vivo by expression of a BRC peptide from BRCA2 ., Swi5−/− and Sfr1−/− cells are selectively sensitive to agents that cause DNA strand breaks , in particular ionizing radiation , camptothecin , and the Parp inhibitor olaparib ., Consistent with a role in HR , sister chromatid exchange induced by Parp inhibition is attenuated in Swi5−/− and Sfr1−/− cells , and chromosome aberrations are increased ., Thus , Swi5-Sfr1 is a newly identified complex required for genomic integrity in mammalian cells with a specific role in the repair of DNA strand breaks . | Our genome constantly undergoes DNA damage as a result of agents in the environment , as well as from metabolic processes ., One method of repairing DNA damage is homologous recombination ( HR ) , in which genetic information from a duplicate sequence ( the sister chromatid ) is copied into the damaged site in DNA ., In model organisms ( the yeasts ) , a protein complex termed Swi5-Sfr1 functions in DNA damage repair by HR ., In this study , we characterize mouse homologues of this complex ., We find that mouse cells lacking this complex are sensitive to DNA damaging agents , in particular , those that cause breaks in DNA strands and that serve as cancer chemotherapeutics ., These cells also have increased numbers of chromosome aberrations when exposed to DNA damaging agents ., Moreover , HR is decreased in Swi5 and Sfr1 mutant cells under conditions where the cell is challenged ., Together , these results demonstrate a requirement for the Swi5-Sfr1 protein complex in maintaining genomic integrity in mammalian cells . | biochemistry/replication and repair, genetics and genomics/cancer genetics, genetics and genomics/chromosome biology, genetics and genomics/gene function | null |
journal.pcbi.1000420 | 2,009 | Investigating the Conformational Stability of Prion Strains through a Kinetic Replication Model | Prions are infectious agents composed solely of proteins , whose replication does not rely upon the presence of nucleic acids 1 ., Although the molecular mechanisms of prion replication are poorly understood , the current working hypothesis is based on the assumption that prions replicate by means of an autocatalytic process which converts cellular prion protein ( ) to the disease-associated misfolded PrP isoform ( ) ., This process of replication of a prion depends upon the capacity of the pathogenic protein form to bind to and to catalyze the conversion of existing intermediate molecules ., Recent studies 2 have observed that the prion protein can misfold into a range of different aggregated forms derived from a continuum of structural conformation templates 3 from which different phenotypic and pathological states derive ., The ability of the same encoded protein to encipher a multitude of phenotypic states is known as the “prion strain phenomenon” 4 ., Prion strains are defined as infectious isolates that , when transmitted to identical hosts , exhibit the following distinct prion disease phenotypes: A reason for the strain phenomenon can be the association of to several disease conformations , characterizable by means of a different stability against denaturation , different post-translational modifications ( e . g . glycosylation ) and distinct cleavage sites ., These observations are reinforced by 5 , where it is reported that the amyloid fibrils ( formed by the 40-residues -amyloid peptide ) with different morphologies have significantly different molecular structures ., These differences are shown to be self-propagating and to be associated with different toxicities , suggesting the possibility for a structural origin of prion strains ., Moreover , recent studies on prion disease have confirmed that the incubation time is related not only to the inoculum dosage and the prion protein expression , but also to the resistance of prion strains against denaturation 3 in terms of the concentration of guanidine hydrochloride ( Gdn-HCl ) required to denaturate 50% of the disease-causing protein ( see Text S1 for further discussions ) ., Other studies have highlighted a strong relationship between the stability of the prion protein against denaturation and neuropathological lesion profiles 6 , 7 ., Lesions due to stable prions tend to show large vacuolations localized in specific small brain regions , whilst lesions due to unstable prion strains show a less intense vacuolation and are more widely distributed in the brain ., Apart from these properties , crucial details of the molecular mechanisms enabling the characterization of different prion strains are still missing ., For example , neither structural characterizations of , nor maps of protein-protein interactions have so far been provided , and even the biological function is unclear ., Hence , in order to use the existing data to gain some insight into the properties of the different prion strains , we decided to follow a model-based approach ., In this paper , using a well established model for the kinetics of the in vivo prion replication 8 , we relate the evidence about conformational stability to the parameters of the model describing the evolution in time of the fibril length ., The main points we deduce from our analysis are: Multiple experimental observations in vitro 9 and in yeast 10 , 11 support our model-based considerations , reinforcing our predictions for in vivo mammalian systems ., Protein polymerization seems to have a central role in the progression of the prion pathology , an aspect shared with several other neurodegenerative diseases associated with different aggregating proteins , such as Alzheimers ( A ) , Parkinsons ( -synuclein ) and Huntingtons ( huntingtin ) diseases ., The aggregation kinetics of amyloid peptides has been studied extensively ( see 12 , 13 ) , and has shed light on the wide range of amyloid aggregation mechanisms observed ., Many modeling approaches have been introduced for this purpose in recent years , e . g . theoretical models consisting of nonlinear ordinary differential equations ( ODEs ) , two-dimensional lattice-based statistical models and molecular dynamics simulations 8 , 13–18 ., In this paper we explore a mathematical description of the prion replication dynamics through nonlinear ODEs ., This class of models explain the appearance of the disease by means of a bistability induced by a quadratic term , as in classical epidemic models 19 ., The model we used is drawn from 8 , 14 and is based on a nucleated polymerization mechanism 20 ( see Materials and Methods ) ., This approach has been shown to overcome the limitations of the “heterodimer model” 1 and to be a reasonable simplification of the “cooperative autocatalysis” approach 18 ., Furthermore , it is able to explain the kinetics of spontaneous generation 18 , the association between infectivity and aggregated PrP , the linear appearance of the fibrils and to take into account fundamental processes of an in vivo replication ( i . e . fibrils splitting ) , all while remaining relatively mathematically tractable ., Moreover , its dynamical behavior has been extensively studied 21 , 22 , and experimental measurements were used in 14 to provide an estimation of the full set of parameters for a particular prion strain ., The model has three state variables ( Eq . 10 ) describing the amount of monomer ( ) , polymer ( ) and the mass of polymer ( ) , and it involves 6 parameters ( see Table 1 ) ., We reproduce here only the features essential to discuss the strain dependence of its parameters; the details are covered in Materials and Methods ., In 8 it has been shown that for any prion strain two parameters , the rate of growth ( ) and the reproductive ratio ( ) , can be estimated from experimental data ., The former ( Eq ., 2 ) represents the exponential growth of the number of infectious particles ., The latter ( Eq ., 3 ) is defined as the average number of prion fibrils that a single infectious particle can give rise to , before splitting into fibrils smaller than a critical size or being degraded ., In other words , represents the ability of the fibrils to survive to critical breakage and degradation events ., The equations for the and parameters obtained from the kinetic model of 8 can be reparametrized in terms of the mean length of the fibrils ( 1 ) obtaining: ( 2 ) ( 3 ) In order to estimate from experimental measures both parameters ( and ) certain assumptions are necessary ( see Materials and Methods for full details ) ., An estimation of and from in vivo experiments and for different prion strains characterized by different values of stability against denaturation ( ) is listed in Table 2 ., The dataset currently available is limited ( as not many prion strains can be fully characterized ) and many error sources are potentially affecting the estimation of the parameters ., Nevertheless , Figure 1 shows the existence of a negative trend between these two empirical parameters ( Pearson correlation\u200a=\u200a−0 . 91 , p-value\u200a=\u200a0 . 01 ) ., If we now turn to the kinetic model and look at the corresponding expressions ( Eq ., 2 ,, 3 ) the interesting question is whether such a behavior is predicted by the model itself , and is explainable in terms of some of its parameters , in a way that is both mathematically and biologically plausible ., Otherwise stated , we investigate which , if any , among the model parameters best describe the strain variability ., The critical size of the nucleus ( parameter in the model ) plays a marginal role in our analysis and is likely to be a fixed integer , in between 2 and 4 , across different strains 23 ., Even though it has been argued that a hexamer is the minimum infectious unit 24 , it can be shown that the model-based conclusions are not conditioned by the value of ., In addition is clearly independent of the prion strains , so we remain with three possible choices: , and ., From Eq ., 1 , increasing means incrementing and this affects and in a similar manner , so that this parameter alone cannot explain the inverse relationship derived in Figure 1 ., The same can be said for and which , if increased/decreased , would induce changes of equal sign in and ., Different conclusions can be drawn when considering as the only strain-varying parameter ., This dependence becomes clearer assuming that fibrils cannot be degraded in the exponential phase ( , identical results can be obtained supposing that the degradation of the fibrils scales as the fibrils breakage rate , , see Text S2 ) ., Such assumption leads to the following expressions: ( 4 ) ( 5 ) ( 6 ) If we keep into account only the dependence from , then Eq ., 4 and Eq ., 5 can be simplified to ( 7 ) ( 8 ) From these simplified formulas it is clear that an increase in the frangibility of the fibers ( i . e . , in ) produces an increment of ( Eq ., 7 ) and a decrement of ( Eq ., 8 ) in agreement with the trend in Figure 1 ., Therefore , from the model we expect to give the best fitting result ., As a matter of fact , this relationship ( black dash-dotted line in Figure 1 ) does not provide the optimal fit , although it reproduces the qualitative observed behavior ( ) ., The fittings of Figure 1 ( see Table, 3 ) suggest that , approximately , ( red line ) implying that we are observing proportional to and to ( see Materials and Methods , Eq . 12 ) ., This means that the estimated exponents for are somewhat different from the expected values of ( ) predicted in Eq ., 7 and 8 ., In order to improve the model prediction , we introduce a strain-dependence on a second parameter ., The simplest solution suggested by the model for this scope ( deducible from Eq . 4 and 5 ) points to the aggregation rate ., By linking to , we are still left with a one-parameter family of models describing the strain-dependence ., In doing so , we obtain the estimate ( see again Materials and Methods , Eq . 13 ) ., This correction yields and , this time respecting the predictions of Eq ., 4 and 5 ., Therefore , on the one hand we can show that at a qualitative level is the only parameter that alone can explain the inverse relationship between and ., On the other hand , the variation of by itself is not able to quantitatively describe the experimental data in a precise way ., An additional correction , obtained relating to , leads to a substantially improved fitting ., Apart from Eq ., 4 and 5 , our choice of alongside as strain-dependent parameter is suggested by the structure of the model of Eq ., 10 , in which , of all parameters , those describing the kinetics of fibril aggregation/breakage are the most likely to vary across strains ., Both the fitting and the model structure suggest an interplay between and , with partially balancing the effect of ., In the following , we will describe how the previous results can be extended to the stability to denaturation of the prion strains , providing experimental observations in support of our claims ., From Figure 2A a direct linear proportionality between and is inferred ., Therefore , combining the fitting between and and and , a similar inverse relationship ( see Figure 2B ) relates and ( see Table 3 ) ., A point of note is that a linear model ( i . e . , ) is not only associated to a low coefficient of determination but is also implausible , as it predicts negative values of in correspondence of very stable prion strains ( such as MK4985 , see Table 4 ) ., Owing to the linear proportionality ( ) of Figure 2A , the inferred functional dependencies from extend to ( i . e . , ) ., This result , in light of the experimental observations in 10 , contributes to validate the results of the kinetic model and provides us with a simple practical tool to interpret prion strain stability ., As a matter of fact , the experimental data in 10 report a relationship between the chemical stability of yeast prion strains and their structural properties , hence reinforcing our conclusions ., In particular , the frangibility of different Sup35NM amyloid conformations was measured and shown to be consistent with an increase in sensitivity to denaturants and proteases ., Thus , confirming the main role of the breakage rate , as predicted here by the model ., Furthermore , the authors observed also a variation in the aggregation rate ( parameter in the model ) , which was however overcome by the stronger effect of the division rate; an additional observation in agreement with our results , where the best match with the experimental data is obtained for a variation of that only partially compensates for that of ., The importance of breakage events for the in vivo prion propagation is also underlined in 25 , where the authors observed that membrane-anchored PrP is necessary for the exponential growth of prion aggregates ., In transgenic mice , expressing anchorless prion protein inoculated with different prion strains , the aggregates seem to grow quadratically in time 26 ., This feature is explainable by a linear aggregation model ( i . e setting equal to 0 ) ., Moreover , in 26 , different prion strains show a common inability to induce the disease ., The absence of fibrils disruption can prevent the formation of oligomeric species , thus hiding the difference between prion strains ., Our model-based analysis suggests that an experiment monitoring the propagation of prion strains lacking the GPI anchor would be useful to characterize in more depth the strain phenomenon ., In the last part of the Section , we investigate how the prion stability ( ) is reflected in the mean length of the fibrils ( ) ., Combining the fitting of Figure 2 with Eq ., 6 , ( and consequently , from Eq . 1 ) can be inferred directly from and : ( 9 ) In Table 4 , we compare the approach of Eq ., 9 with the results obtained in 14 , where the authors give a complete estimation ( including a range of uncertainty ) of all the parameters for the RML prion strain ( , highlighted in bold in Table 4 ) ., The comparison between these two approaches shows that the predictions obtained through Eq ., 9 are similar to the values reported in 14 for the RML strain ., In addition , we can compare the values of for the strains inferred from Eq ., 9 , with the ones computed using Eq ., 11 and then imposing equal to the values of 14 for the RML strain ( see Figure S1 ) ., Our predictions are approximately within the range of values computed considering constant among strains ., This result reinforces the major role of in explaining strain variability ., Owing to the fact that is now strain-dependent ( ) , we can also predict the mean length of the fibrils ( Eq . 1 ) for each considered strain ( see Table 4 , ) ., For instance the mean length of the fibrils population for two prion strains with different stabilities ( e . g . RecMoPrP ( 89–230 ) and Sc237 ) can be compared ., For the unstable prion strain ( Sc237 ) this is approximately 7 monomer units , while for the stable prion strain ( RecMoPrP ( 89–230 ) ) it is approximately 14 monomer units ., This theoretical approach provides a valuable method to simplify the model characterization ., Furthermore , it contributes to understanding the properties associated to prion strains with different stability against guanidine denaturation ., While it is reasonable that the parameters of the kinetic model might all be affected by strain specificities ( i . e . stability against denaturation ) , the dominant contribution seems to be due to the susceptibility to frangibility ( i . e . ) , with only a minor correction due to ., The inverse relationship between and shown in Figure 1 is the main argument in the identification of as the key physical aspect differentiating prion strains ., In addition , is suggested as the most plausible and parsimonious correcting factor , in order to improve the data fitting ., Several aspects can influence the estimation of the parameters and the model predictions ., For example , the uncertainty affecting the estimation of and ( respectively inferred from an exponential curve and from a ratio of exponentials ) ; or the possibility that the breakage rate is not equal across all the different polymer lengths ( e . g . mechanical stress can differently affects longer fibers ) ; or even the impact of the mouse age on the model parameters ( affecting e . g . the production rate ) ., In spite of these ( and potentially many other ) disregarded aspects characterizing an in vivo system , this simple model is able to capture and explain the observed data dependencies through arguments supported by multiple independent experimental observations ., Our analysis reveals that stable prion strains can be characterized by a “stronger” aggregated structure which is less prone to breakage events ., This will further imply a longer mean length of the fibrils ., Instead , unstable prion strains are subject to a higher fragmentation rate ., The role of is essentially to partially balance the increased breakage and is coherent with the experimental observations in yeast ., Furthermore , the increased number of catalytic sites may be also responsible for the shorter incubation time ., As already mentioned , such phenomenon was observed in yeast prions 27 ., The yeast prion proteins , although fundamentally different from the mammalian prion proteins , show the same ability to convert into aggregated forms , propagate and be infectious ., This simpler unicellular system is a valuable model as it enables a deeper analysis of the fibril formation process 28 , not possible to the same extent in higher organisms ., The framework proposed allows for a model-based analysis of these properties in mammalian prions in vivo ., In the context of mammals , our results are consistent with 9 , where fibrils with different conformational stability are generated in vitro from full length mammalian PrP ., In that paper , the authors relate the stability to the size of the smallest possible fibrillar fragment without taking into account the kinetics of the replication ( reproducing the in vivo behavior ) ., We draw similar conclusions from a different point of view ., As a matter of fact , we investigate the dynamic evolution of prion propagation in a multicellular in vivo system , in which molecular and cellular mechanisms are present as well ., Our model-based conclusions provide further evidence that in vitro systems and yeast prion propagation mechanisms can be transposed in mammals ., Moreover , linking the strain phenomena to dynamical features leads to a characterization of the evolution of the length of the fibrils in vivo ., We can , in fact , speculate ( in agreement with 5 ) that stable prion strains exhibit a proliferation of longer fibrils that , upon splitting , still manifest the same stability properties ( Figure 3 ) , giving rise to a preferential proliferation of relatively long fibrils with a low toxic effect ., On the other hand , less stable prion strains tend to form shorter fibrils , to proliferate faster and to be more neurotoxic ., It is worth noting the connection with 13 , where the kinetics of aggregation of amyloid peptides is studied by means of coarse-grained molecular dynamics ., The authors showed how the relative stability of -prone states of a polypeptide can influence the pathway of aggregation ., Their results suggest that the -stable amyloids follow an aggregation pathway without intermediates , while -unstable amyloids seem to involve on-pathway oligomers ., The characterization of prion strains in terms of polymer mean size is per se a significant observation ., It provides a new possible explanation of the observation that stability is correlated with lesion profiles and vacuolation areas ., Several hypotheses have been made , such as the existence of a co-factor that supports the conversion of distinct prion strains in precise brain regions ., Here , another possibility emerges: the increased size associated to stable prions can decrease their ability to diffuse , and can circumscribe them to small brain regions ., On the contrary , oligomers can spread around the brain more easily , causing a more homogeneous damage ., In conclusion , we show that linking the conformational stability property of prions , acquired during in vivo propagation in mammals , to their replication kinetic properties is achievable through a rather simple model ., For a wide range of parameters , the model predicts that a higher breakage rate implies shorter and shorter incubation time ( in Figure S2 two simulations are compared ) ., Our model-based approach suggests that the amount of information that can be extrapolated from the knowledge of goes beyond the expected incubation time ., In vitro prion propagation is characterized phenomenologically by the following properties:, ( i ) a critical concentration threshold below which fibrils cannot form;, ( ii ) a delay before their propagation ( which can be eliminated by the addition of seeds of preformed fibrils ) ;, ( iii ) a direct proportionality between the initial rate of fiber growth and the monomer concentration 29 ., The overall behavior resembles a sigmoidal growth curve 30: an exponential growth of infectious particles followed by a plateau ., The simplest description of the underlying observed mechanism of protein aggregation consists of a slow continuous nucleation followed by a fast autocatalytic growth ., Therefore a simple two-step model is able to reproduce the dynamics of the in vitro prion propagation 12 ., An in vivo prion propagation model should explain the fact that the spontaneous prion-induced disease is rare but progresses inevitably after infection , that the incubation period is long and followed by a brief fatal clinical disease and that prions undergo several molecular processes within the cell ( e . g . fibrils breakage , degradation , endogenous production ) ., The model derived in 8 is obtained as a closed form of an infinite set of differential equations describing the variation in time of the monomer and fibrils of each possible length ( from to ) ., The biological mechanisms taken into account are the lengthening at the fiber end by the addition of monomers , the degradation of polymers , and their splitting into smaller polymers ., Only if several monomeric molecules are mounted into a highly ordered seed , can further monomeric be recruited and form amyloid aggregates ., If , after the breakage , the fibril has a length under the critical size , it degrades instantaneously into normal monomers ., The model in Eq ., 10 has three state variables , describing the amount of monomer ( ) , polymer ( ) and the mass of polymer ( ) , and it comprises of 6 independent parameters: nucleus size ( ) , rates of production ( ) , degradation ( ) , aggregation ( ) , clearance ( ) and breakage ( ) : ( 10 ) The assumption that is negligible , made in in order to simplify the parameters equations , changes the qualitative behavior of the model , that no longer has two stable steady states but only one , which is unstable ., This means that the exponential growth will never reach a plateau ., As mentioned in the text , this does not affect our previous considerations , especially in light of the fact that in vivo death occurs during the exponential growth phase ( see also the Text S2 for similar conclusions on the full model ) ., In this section we summarize the procedures mentioned in 8 and adopted here to derive a measure for and ., The assumptions deemed , in order to measure and from the observed effect of different levels of PrP expression and inoculum dosage , are as follows: Of all assumptions , the last one is the most important ., It is considered valid also for transgenic mice expressing different quantities of cellular prion protein ., Currently there is wide debate about the cause of cell death in prion neurodegeneration ., From knockout mutants , it seems that loss of function is not sufficient to cause cell death ., What has been observed is that the conversion of to the isoform has a key role in the disease ., In spite of their apparent low neurotoxic effect 26 , fibrils have been proven to be the main ingredient in catalyzing variations of protein conformation 31 ., Therefore , it is reasonable to assume that even if toxicity is not directly associated to fibrils aggregates , it has to be closely related to their amount , implying that a critical concentration of is required to provoke cell death ., The current working hypothesis is that oligomeric species are the most infectious 32 and a substantial body of evidence suggests that they are also highly cytotoxic 33 ., According to the previous observations , a possible explanation is that an equal mass of prion fibrils with smaller mean size provides a larger number of active sites for catalysis , hence inducing a higher lethality ., In order to extrapolate a measure for we follow the method described in 8 based on relating the incubation time to the inoculum dose and implying an exponential growth in the number of infectious particles ., Taking advantage of these data ( e . g . incubation time inoculum dosage ) , we can infer the parameter just by fitting an exponential growth curve ., More precisely , before inoculation of prions , PrP ( ) can be considered at steady state ( ) ., After inoculation , it is reasonable to assume that it remains almost constant for a while ., According to the model equations , the steady state of the mean polymers distribution length ( in Eq . 1 ) , is typically reached before the exponential phase ., Immediately after reaching , the polymer amount ( ) and the polymer mass ( ) start to grow exponentially ., Thus , is defined as the dominant mode of this exponential growth ( i . e . , ) ( Eq . 2 ) ., To have an indirect measurement of , the inverse relationship between incubation time and the PrP expression is exploited ., We take into account the previous assumptions reporting that the number of infectious units in two inoculated mice expressing different level of PrP ( , ) at the times of death ( , ) can be considered almost equal ., Thus imposing we can derive : ( 11 ) where and ., For a more detailed description see Appendix of 8 ., It is worth noticing that the incubation times listed in 3 are not the same as those used to estimate ( see Text S1 for more details ) ., Rather that using Eq ., 7 and 8 , the exponents and such that , can be computed from the fitted curve in Figure 1 ., We approximate the numerical value 0 . 38 of Table 3 with 0 . 4 ( i . e . ) ., From the above expressions , , which yields , i . e . , or ., Examples of values on this line are:From Eq ., 6 it is clear that the only admissible pair of values is ( 12 ) If , following Eq ., 4 and 5 , we add the extra functional dependence of from as , we can look for a value of that satisfies simultaneouslyyielding: ( 13 ) | Introduction, Results, Discussion, Materials and Methods | Prion proteins are known to misfold into a range of different aggregated forms , showing different phenotypic and pathological states ., Understanding strain specificities is an important problem in the field of prion disease ., Little is known about which PrPSc structural properties and molecular mechanisms determine prion replication , disease progression and strain phenotype ., The aim of this work is to investigate , through a mathematical model , how the structural stability of different aggregated forms can influence the kinetics of prion replication ., The model-based results suggest that prion strains with different conformational stability undergoing in vivo replication are characterizable in primis by means of different rates of breakage ., A further role seems to be played by the aggregation rate ( i . e . the rate at which a prion fibril grows ) ., The kinetic variability introduced in the model by these two parameters allows us to reproduce the different characteristic features of the various strains ( e . g . , fibrils mean length ) and is coherent with all experimental observations concerning strain-specific behavior . | Prion diseases are caused by the accumulation of a cellular prion protein with an altered conformation , which acts as a catalyst for the further recruitment and the modification of the normal form of the protein ., Protein polymerization appears to have a central role in the progression of the disease , an aspect shared with several other neurodegenerative diseases ., The aim of this work is to investigate at the kinetic level the “prion strain phenomenon” , i . e . , the ability of prion proteins to misfold into a range of different aggregated forms exhibiting different replication and propagation properties ., The dynamics of prion replication is investigated with the help of a mathematical model ., We relate a measurement accessible in vitro ( prion structural stability ) to a mathematical description of the fibrils kinetics in vivo ., The analysis of the model suggests that the replication kinetics of the different prion strains is characterizable by means of two parameters , representing the rates of breakage and aggregation ., This result is coherent with various experimental findings concerning strain-specific behavior , such as , for example , the observation of the fibril mean length of the various strains . | infectious diseases/prion diseases, computational biology/systems biology | null |
journal.pcbi.1000495 | 2,009 | A Structured Model of Video Reproduces Primary Visual Cortical Organisation | It is well established that the receptive fields ( RFs ) of neurons in the early visual cortex depend on the statistics of sensory input and can be modified by perturbations of those statistics during development 1–6 ., This relationship has been studied theoretically in many ways ., Phenomenological models have focused on the mechanisms of synaptic plasticity and axon-guidance , giving mathematical or computational accounts of how Hebbian-like learning rules may combine with sensory stimulation to drive the formation of cortical response properties 7–12 ., Constrained optimality approaches look beyond the details of the synaptic learning rule , and ask whether the observed pattern of cortical responses has been selected to optimise a functional objective ., Many early studies of this type were founded on the information-theoretic ideas of efficient coding and redundancy reduction 13 , 14 , and proposed that RFs had adapted to maximise the transmission of information from the periphery 15–18 ., More recent work has generalised this approach to consider other possible objective functions with different representational or metabolic benefits ., Two established alternatives are the sparseness and temporal stability objective functions ., In the sparse-coding view neuronal properties are optimised so that neurons remain silent most of the time , responding vigorously to only a limited subset of all stimuli 19–21 ., Thus every image is represented by relatively few active neurons ., Such a representation makes it easy to detect “suspicious coincidences” 22 and reduces energy consumption 23 ., It can also be related to the older objective of information efficiency 19 ., Under the temporal stability objective , neuronal RFs are adapted so that their output firing rates vary slowly in time 24–26 ., To achieve stability , neurons must learn to be insensitive to typical rapid transformations of their input , leading to invariant representations that simplify recognition tasks 27 ., The generative modelling approach takes a complementary functional view ., It is based on the Helmholtzian account of perception as inverse inference ( sometimes called analysis-by-synthesis ) ., That is , that the goal of the perceptual system is to infer from sensation the environmental causes most likely to be responsible for producing the sensory experience 28 , 29 ., In this view , sensory cortex implicitly embodies a model of how external causes interact to form the sensory input ( a causal generative model ) ; given a particular sensory experience , cortical processing inverts the model to infer the most likely causes of the sensory activity ., Mathematically , this corresponds to an application of Bayes rule ., This general view that the brain carries out or approximates some form of probabilistic inference is supported by a number of psychophysical , anatomical , and physiological results ( see 30 , 31 for reviews ) ., Many models that have been formulated in terms of the optimisation of an objective function could also be viewed as implementing inference within an appropriate generative model: the assumptions and structure of the model are implicit in the objective function ., Thus , recoding based on the sparseness objective corresponds to inference within a generative model in which a number of independent , sparsely active causes combine linearly to form the image 20 ., Similarly , the goal of redundancy reduction has led to models in which divisive normalisation reduces second-order dependence between linear recodings of an image 32; in the generative view , this corresponds to joint modulation of the variances of otherwise independent sparse causes 33 , 34 ., Finally , the temporal stability objective corresponds to a model with causes that are independent of one another , but stable or predictable in time 35 ., A remarkable success of these functional models , whether formulated generatively or in terms of a representational objective function , is that , when used to learn an appropriate representation from a set of natural images , they yield elements that mirror a number of response properties of primary visual cortical neurons ( though some notable discrepancies do remain 16 ) ., However , despite this success , the generative models involved match only the lowest-level statistics of natural images ., Images generated from the learnt models have naturalistic textural properties , but none of the higher-level structure of the natural world ., If this approach is to provide insight into higher processing within the visual cortex then appropriate structure must be introduced to the models ., In the present study we focused on one basic structural aspect of the environment: The visual world is largely composed of discrete objects , which each contributes a set of discrete visual features to the image ., Moreover , the objects , and therefore their associated features , usually remain in view for some time , although their precise appearances might change gradually due to changes in viewpoint , lighting or in the objects position ., We thus formulated a model in which the identity of the visual elements present was signalled by a set of binary-valued variables , while their appearances each evolved separately under the control of continuous attribute variables ., This independent control of appearance stands in contrast to a related idea of “content” and “style” 36 , 37 where the transformation of appearance is usually shared across the image or image patch ., This comparison is taken up in greater detail in the Discussion ., We fitted this model to natural video images , without using any additional information about which elements were present or what their transformations might be ., We found that the model naturally learned biologically plausible features , with low dimensional manifolds of attributes ., Many aspects of the learnt representation corresponded closely to both anatomical and functional observations regarding simple and complex cells in the primary visual cortex ( V1 ) ., Thus , the model offers a functional interpretation for the presence of two main classes of cells in V1 ., Complex cells represent the probability of presence of an oriented feature , while simple cells parametrise the precise appearance of the feature in the visual input ., We speculate that a similar representation in the form of feature identities and attributes may continue up the visual hierarchy , ultimately contributing to view-independent object recognition ., Figure 1A illustrates the intuitions that underlie the general structure of the model ., The image at each point in time—represented by a vector shown at the bottom of the figure—is composed from a set of visual elements illustrated by the objects in the top row ., Only a small subset of all the possible elements contributes to any one image ., The identity of these active elements is represented by a set of binary-valued variables , where means that the th element appears in the image at time ., If active , the form of the element in the image may vary; for instance the object may appear at any position or orientation ., Each element is thus associated with a set of possible contributions to the image , which form a manifold embedded within the space of all possible images ., The configuration of element at time is then specified by a vector , with dimensionality equal to that of the manifold ., We call the elements of this vector , , the attributes of the visual element ., The shape of the manifold is described by a function , which maps this attribute vector to the partial image it describes ., For concreteness , consider the rightmost panel of Figure 1A , which represents a model for a beverage can ., The fact that the variable takes the value indicates that the object is present in the image at time ., The arrow indicates the point ( encoded by ) on the manifold where the can has a particular position and viewpoint in the input visual space ., If one of the attribute variables were to correspond to the orientation of the can , changing its value would trace a trajectory on the manifold , which would result in a rotation of the object in the image space ., The set of partial images associated with all of the active elements then combine through a function , which could in principle implement occlusion , illuminant reflection , or other complex interactions , to yield the image: ( 1 ) where we have included an additive , independent noise term ., In this abstract form the model is very powerful , and provides an intuitively satisfying generative structure for images ., Unfortunately , for manifolds and combination functions modelling the appearance of entire complex objects and the interactions between them as illustrated in Figure 1A , the task of inferring the elements and their appearances from natural data is intractable ., To explore the potential of the framework we adopted a simplified form of the model , taking the mappings to be linear ( equivalently , we defined the attribute manifolds to be hyperplanes ) and to sum its arguments ., This allowed us to implement the selection of the active elements by multiplication: ( 2 ) where the basis vectors parametrise the linear manifold , and and are the number of identity variables and the ( maximum ) dimensionality of each attribute manifold respectively ., In this simpler form , we expect the visual elements to correspond to more elementary visual features , rather than to entire objects ( Fig . 1B ) ., The complete probabilistic generative model for image sequences includes probability distributions over the identity and attribute variables ., We chose distributions in which objects or features appeared independently of one another , and where the probability of appearance at time t depended on whether the same feature appeared at time ., The attributes of the feature evolved smoothly , again with a Markovian dependence on the preceding state ., The formal definition of the probabilistic model is given in Methods ., The parameters of the model specify the partial images generated by each feature ( represented by the basis vectors ) , the probability of each feature being active , and the degree of smoothness with which the appearance of the feature evolves ., All of these parameters were learnt by fitting the model to natural image sequences ., In previous work on sparse coding the number of basis vectors or components needed has been explored outside of the model fitting procedure ( for example 38; but see 39 ) ., Crucially , here we were able to learn the dimensionalities of the model—the numbers of visual elements and associated attribute variables—from the data directly , using Bayesian techniques described below and in Methods ., Probabilistic models are often fit by adjusting the parameters to maximise the probability given to the observed data—called the likelihood of the model ., In practice , image models have often been fit by maximising the data probability under settings of both the parameters and the unobserved variables ( in our case these would be the identity and attribute variables ) , a procedure which may be severely suboptimal 40 ., Here , we adopted an iterative procedure called Variational Bayes Expectation Maximisation ( VBEM ) 41 , 42 to learn an approximation to the full probability distribution over the parameters and unobserved variables implied by the data—known as the VB posterior distribution ., This posterior provides a more robust estimate of the parameters , with concomitant estimates of uncertainty , and can be used to determine the appropriate dimensionality of the model directly ., More complex models can always be adjusted to give higher probability to any data set , and so the maximum likelihood approach would always favour a model with greater dimensionality ., This effect can lead to overfitting , where an overly complex model is selected ., However , because there are very many more possible parameter settings in a complex model , any one such parameter setting may actually be very improbable even though it might fit the data well ., Thus , when considering the probabilities of parameter settings and models as in the Bayesian approach , a form of “Occams Razor” comes into effect favouring descriptions complicated enough to capture the data well but no more so 43 ., For models similar to the one developed here , one consequence of this “Occams Razor” is that the posterior probability distributions on the values of any superfluous basis vectors concentrate tightly about 0 , effectively pruning the basis dimension away , and leaving a simpler model ., In this context , the effect has been called Automatic Relevance Determination or ARD 42 , 44 ., Bayesian estimation is well-defined only if a prior distribution—that is , an initial probability distribution determined before seeing the data—is specified ., The prior on the basis vectors was of a form often used with ARD , with a so-called hyperparameter determining the concentration about a mean value of 0 ., The prior distributions on the parameters that determine the temporal dependence of identity and attribute variables were broad enough not to influence the posterior distribution strongly ., The exact definitions of the distributions over parameters , along with details of the fitting algorithm , are given in Methods ., We used this model to investigate the visual elements that compose natural images , comparing features of the representation learnt by the model when fit to natural image sequences to the representation found in V1 ., The input data were a subset of the CatCam recordings 45 , which consist of several-minute-long video sequences recorded by a camera mounted on the head of a cat freely exploring a novel natural environment ., Temporal changes in the CatCam videos are caused partly by moving objects , but mostly by the animals own movement through the environment ., Cats make few saccades and use only small eye movements to stabilise the image during locomotion 45 , so that the amplitude and frequency of spatial transformations in the videos ( translation , rotation , and scaling ) is similar to that experienced by the animals ., Computational constraints prevented us from modelling the entire video sequence ., Instead , we fit the model to the time-series defined by the pixel intensities within fixed windows of size pixels over 50 frames ., We initialised the model with 30 identity variables each associated with attribute manifolds of 6 dimensions and let the algorithm learn an appropriate model size by reducing the number of active attribute dimensions and identity variables by ARD ., We performed a total of 500 VBEM iterations , at each iteration taking a new batch of 60 sequences of 50 frames , randomly selected from the entire dataset ., Further computational details are given in Methods ., Given an observed image sequence , the model could be used to infer a posterior probability distribution over the values of the identity and attribute variables at each point in time ., We compared the means of these distributions to the firing rates of neurons in the visual cortex ., The use of the mean was necessarily arbitrary , since there is no generally agreed theory linking probabilistic models to neural activity ., The brain may well represent more than a single point from this distribution ., For example , information about the uncertainty in that value would be necessary to weight alternative interpretations of the data ., Once the model had been fit to the data , however , we found that the attribute variable distributions estimated from high-contrast stimuli were strongly concentrated around their means ., Thus , many different choices of neural correlates would have given essentially identical results ., It is also worth mentioning here that although the identity variables describe the presence or absence of a feature in the generative model and are thus binary-valued , the posterior probability of the feature being present ( which is the same as the posterior mean of the binary identity variable ) is continuous ., Thus , neurons presumed to encode these posterior means would respond to stimuli with graded responses , which would take uncertainty about feature identity ( e . g . , under conditions of low contrast ) into account ., Figure 2A shows the VB posterior mean basis vectors learnt from the CatCam data ., Each row displays the basis vectors of the attribute manifold corresponding to a single identity variable ., Since the manifold was a hyperplane , the set of possible feature appearances was given by all linear combinations of the basis vectors ( Fig . 3D ) ., For every manifold , the mean basis vectors resembled Gabor wavelets with similar positions , orientations , and frequencies , but different phases ( Fig . 4A–C ) ., Thus every point on the manifold associated with a single feature corresponded to a Gabor-like image element with similar shape , orientation , and frequency , but variable phase and contrast ., When presented with a drifting sine grating of orientation and frequency similar to that of the basis vectors , the probability of the feature being present was found to approach 1 rapidly , and then to remain constant , while the means of attribute variable distributions oscillated to track the position of the sine grating on the manifold , as illustrated in Figure 3 ., Attribute variables thus behaved much like simple cells in V1 , in that they responded optimally to a grating-like stimulus and oscillated as its phase changed , while identity variables responded like complex cells , being insensitive to the phase of their optimal stimulus ., In electrophysiological studies , the classification of neurons into simple and complex cells is done using a relative modulation index 46 , 47 , which is defined as the ratio of the response modulations ( F1 ) to the mean firing rate ( F0 ) in response to a grating with optimal orientation and frequency , but varying phase ., Cells that respond to phase changes with large oscillations have relative modulation larger than 1 and are classified as simple cells , while cells that are invariant to a phase change are classified as complex cells ., We computed the relative modulation for the posterior mean values of the variables in our model ., All identity units were classified as complex ( maximum F1/F0 ratio 0 . 28 ) and all attribute units that had not been pruned during learning were classified as simple ( minimum F1/F0 ratio 1 . 45 ) ., The magnitude of relative modulations for attribute and identity units is comparable to that found in simple and complex cells in the primary visual cortex of macaque and cat , although the population distribution is narrower 47 ( Fig . S2 ) ., By contrast to the standard energy model of complex cells 48 , here complex and simple cells did not form a hierarchy , but rather two parallel populations of cells with two different functional roles: the former coding for the presence of oriented features in its receptive fields , the latter parametrising local attributes of the features ( primarily their phase ) ., To explore this connection further we compared the properties of simple cell RFs in V1 as reported in the physiological literature with the ‘RFs’ of the attribute variables ., The RF of an attribute variable was defined by analogy to the conventional physiological definition ., We fixed the posterior distribution over the parameters of the model to that estimated by VBEM from the natural data , and then examined the values of the attribute variables that were inferred given coloured Gaussian noise input ., The RF was defined to be the best linear approximation to the mapping from this input to the inferred mean attribute value , a procedure equivalent to finding the “corrected spike-triggered average” or Wiener filter 49 ( see Methods ) ., Although nonlinearities in the model and inference meant that these RFs differed slightly from the basis vectors associated with the attribute variables , we found them to be visually indistinguishable ( Fig . S1 ) ., We then computed the orientation , spatial frequency and phase for the resulting RFs by fitting a Gabor function to each of the filters ( Methods; Fig . S1 ) ., Figure 4 ( A–C ) shows the orientation , frequency , and phase for each pair of RFs associated with the same identity variable ( thus , a feature with a 4-dimensional attribute manifold contributed 6 points to each graph ) ., In the visual cortex , neurons performing related computations appear to be co-located 50 , 51 ., Since the responses of related neurons are highly dependent given a visual stimulus , this may reflect a computationally efficient solution that minimises wiring length 11 , 52 ., We compared our data to the results reported in 53 for pairs of simple cells recorded from the same electrode in area 17 of the cat visual cortex ( Fig . 4D–F ) ., In both the model and physiological reports , the two orientations in each pair of RFs agreed very closely; the frequencies slightly less so; while no relation was apparent in phase ., The distribution of preferred frequencies and orientations in the RFs of attribute variables are shown in Figure 2 B , D ., The distribution of frequencies is quite broad compared to that found in models based on sparse coding or independent component analysis ( ICA ) 16 , 54 , where RF frequencies tend to cluster around the highest representable value , and compares well with the width of the distribution in simple cells ( Fig . 2C ) 55 ., The joint distribution of orientation and frequency ( Fig . 2E ) covers the parameter space relatively homogeneously ., Note that the CatCam image sequences have less high-frequency power at horizontal orientations , and this bias is reflected in the results ., Figure 5 shows the joint distribution of RF width and length in normalised units ( number of cycles ) in our model and for simple cell RFs as reported by Ringach 56 , 57 for area V1 in the macaque ., The aspect ratios are similar in both cases ( again , contrasting with typical sparse coding results 58 ) , although the model results tend to have larger RFs , possibly again due to the particular content of the video ., The model was initialised using a representation that contained 6-dimensional attribute manifolds for each feature ., However , in the posterior distribution identified by VBEM , the probability of the basis vectors corresponding to many of these dimensions being non-zero vanished—that is , a model in which the image data were described using fewer dimensions was found to be more probable ., In fact , the VB posterior representation was only slightly overcomplete , with 96 basis vectors representing an 81-dimensional input space , and with the dimensionality of most feature manifolds lying between 2 and 4 ( Fig . 6A ) ., Given the proposed identification of identity variables with complex cells , this gives a prediction for the dimensionality of the image-subspace to which a V1 complex cell should be sensitive ., The subspace-dimensionality of a complex cell may be estimated by finding the number of eigenvalues of the spike-triggered covariance ( STC ) matrix 59 that differ from the overall stimulus distribution ., One study 60 has reported , for complex cells in the anaesthetised cat , a distribution of dimensionalities that peaked sharply at 2 , with only a few complex cells being influenced by 1 , 3 , or 4 dimensions ., A more recent paper published by the same group has found a broader distribution in the awake macaque 61 ., An analysis of the RFs of the identity variables made using an equivalent procedure revealed a comparable distribution for our results ( Fig . 6B ) ., ( The number of significant eigenvectors returned by the STC analysis can be slightly different from the dimensionality of the attribute manifold because of the non-linear interactions with other variables in the model . ), The model distribution is skewed slightly towards a larger number of stimulus dimensions; although this may be because the sample in 61 included both simple and complex cells ., A second study 62 performed a similar analysis using spatio-temporal stimuli and found 2 to 8 significant dimensions for complex cells ., This broad range of dimensionalities agrees qualitatively with our results ., Unfortunately , quantitative comparison with this study is unreliable as the physiological RFs were identified in effectively one dimension of space , and one of time , while the basis vectors of the attribute manifolds span two spatial dimensions , without a temporal aspect ., A key aspect of our model is the temporal dependence of the identity and attribute variables ., To ask what role this temporal structure had on the feature basis vectors found , we shuffled the order of frames in the CatCam database , and then refit the model using exactly the same procedure as before ., When using unshuffled data , the learning process adapted the feature manifolds so that the inferred values of identity variables persisted in time , while the inferred attribute variables changed smoothly ., In the shuffled data such a persistent and smooth representation cannot be found ., Instead , learning adjusts the attribute manifolds so as to maximise the independence of the associated identity variables , grouping together attribute dimensions that tend to co-occur in single frames ., This is similar in spirit to Independent Subspace Analysis 63 , or to a Gaussian Scale Mixture model 33 with shared binary-valued scale parameters 64 ., Figure 7 shows the basis vectors and pairwise distributions of their properties found for the shuffled data ., The VB posterior distribution concentrated on a more overcomplete representation ( 122 basis vectors representing 81 input dimensions ) than for the unshuffled data ., Some manifolds were pruned away entirely , while the majority of those that remained preserved the maximum dimensionality of 6 ., The basis vectors still resembled oriented features , although the fit of the linear RFs with Gabor wavelets was worse on average than that obtained for the unshuffled video , or seen in physiological data ., The fractional error of fit ( sum of squares of the residuals divided by the sum of squares of the RFs ) was for simple cells 53 , for the model learnt from unshuffled data , and in this case ( Fig . 8 ) ( see Fig . S1 and S3 , for the reverse-correlation filters and Gabor fits ) ., As shown in Figure 7 ( b–d ) , attribute variables associated with a single identity still agreed in orientation , but not in phase ., However , in contrast to the model learnt from unshuffled sequences and to the physiological results , there was much poorer correspondence in spatial frequency ( compare Fig . 7C to Fig . 4B , E ) ., According to their relative modulation index , identity variables would still be classified as complex cells ( maximum F1/F0 ratio 0 . 63 ) , and attribute variables as simple cells ( minimum F1/F0 ratio 1 . 34 ) ., Despite finding a larger number of basis vectors , the model described a larger proportion of the shuffled data as noise , thereby fitting them more poorly ., We evaluated the probability given to 50 new batches of 3000 frames each by the parameter distributions learnt from the shuffled and unshuffled data ., As estimated by the VB approach , the probability assigned by the unshuffled model was more than times greater ( more precisely , the free-energy—a lower bound on the log probability that is maximised by the VBEM algorithm—was larger by NATS , i . e . between 1 . 7% and 4 . 5% greater; Methods ) ., Overall , when deprived of temporal structure in the observations , the algorithm converged to a worse model of the video , and one which was less similar to the physiological data ., It is interesting to note that despite these deficiencies in the representation learnt from shuffled sequences , the basis vectors of the attribute variables still resembled simple cell RFs ., This observation stands in contrast to results from previous models of complex cells based on temporal stability , which had assumed a hierarchical organisation similar to the classical energy model 25 , 26 ., In those models the only signal available to shape the simple cell RFs derived from the temporal stability imposed on the corresponding complex cells ., If this signal were removed by shuffling the input frames , the simple cells would be unable to develop any sort of organised response ., In our model , however , the independence effect discussed above was still able to provide a learning signal for the attribute manifold in the absence of temporal stability ., Thus , we predict that even if stimulus temporal correlations were disrupted during learning , for example by rearing animals in a strobe-lit environment , simple-cell responses would still emerge; although the receptive fields ( defined by reverse correlation ) would fit Gabor wavelets less accurately , and anatomical subunits would be less well-grouped in spatial frequency ., In fact , experimental evidence from Area 17 in strobe-reared cat seems to support our results ., After strobe rearing at an 8 Hz frequency , the spatial RF structure of simple cells in area 17 remained intact except for their width , which was found to increase; and for direction selectivity , which was mostly lost 65 ., Studies performed with lower strobe frequencies ( 0 . 67–2 Hz ) found other changes in the RF properties , including an increase in the number of cells classified as non-oriented , a slight decrease in orientation selectivity , and a reduction of the frequency of binocular cells 66 ., In addition , given the increase in the dimensionality of the attribute manifold , we predict that an STC analysis of complex cells in strobe-reared animals would show a larger number of relevant dimensions ., We have investigated a new generative model for images which makes explicit the separation between the identity of a visual element and the attributes that determine its appearance ., This structure within the model makes it possible to extract and bind together attributes that belong to the same visual element , and at the same time to construct an invariant representation of the element itself ., We modelled identity with a set of binary-valued variables , each coding for the presence or absence of a different feature ., Their appearances were described by manifolds , parametrised by a set of attribute variables ., Both identity and attribute variables were assumed to exhibit temporal dependence within image sequences ., We were also interested in determining the size of the model , i . e . , the number of attribute and identity variables required to optimally describe the input data ., This was achieved by performing a Bayesian analysis of the model , which avoids over-fitting and involves defining an appropriate prior distribution over the generating basis vectors ., As a result , after convergence of an iterative algorithm , only the basis elements needed to effectively match the data remained active and all redundant attribute directions were pruned away , avoiding overfitting the image data ., The algorithm was applied to natural image sequences in order to learn a low-level representation of visual scenes ., The filters associated with the individual attribute variables were shown to have characteristics similar to those of simple cells in V1 ., The RFs of attributes associated with the same identity variable had similar positions , orientations , and frequencies , but different phases ., As a consequence , the corresponding identity variable became invariant to phase change and behaved like a complex cell ., In the standard energy model of complex cells and in several previous functional models , complex and simple cells form a hierarchy ., Simple cells have the role of subunits and are regarded as an intermediate step on the way to the complex cell ., Their phase-dependent information is then discarded as a first step towards the construction of an invari | Introduction, Results, Discussion, Methods | The visual system must learn to infer the presence of objects and features in the world from the images it encounters , and as such it must , either implicitly or explicitly , model the way these elements interact to create the image ., Do the response properties of cells in the mammalian visual system reflect this constraint ?, To address this question , we constructed a probabilistic model in which the identity and attributes of simple visual elements were represented explicitly and learnt the parameters of this model from unparsed , natural video sequences ., After learning , the behaviour and grouping of variables in the probabilistic model corresponded closely to functional and anatomical properties of simple and complex cells in the primary visual cortex ( V1 ) ., In particular , feature identity variables were activated in a way that resembled the activity of complex cells , while feature attribute variables responded much like simple cells ., Furthermore , the grouping of the attributes within the model closely parallelled the reported anatomical grouping of simple cells in cat V1 ., Thus , this generative model makes explicit an interpretation of complex and simple cells as elements in the segmentation of a visual scene into basic independent features , along with a parametrisation of their moment-by-moment appearances ., We speculate that such a segmentation may form the initial stage of a hierarchical system that progressively separates the identity and appearance of more articulated visual elements , culminating in view-invariant object recognition . | When we look at a visual scene , neurons in our eyes “fire” short , electrical pulses in a pattern that encodes information about the visual world ., This pattern passes through a series of processing stages within the brain , eventually leading to cells whose firing encodes high-level aspects of the scene , such as the identity of a visible object regardless of its position , apparent size or angle ., Remarkably , features of these firing patterns , at least at the earlier stages of the pathway , can be predicted by building “efficient” codes for natural images: that is , codes based on models of the statistical properties of the environment ., In this study , we have taken a first step towards extending this theoretical success to describe later stages of processing , building a model that extracts a structured representation in much the same way as does the visual system ., The model describes discrete , persistent visual elements , whose appearance varies over time—a simplified version of a world built of objects that move and rotate ., We show that when fit to natural image sequences , features of the “code” implied by this model match many aspects of processing in the first cortical stage of the visual system , including: the individual firing patterns of types of cells known as “simple” and “complex”; the distribution of coding properties over these cells; and even how these properties depend on the cells physical proximity ., The model thus brings us closer to understanding the functional principles behind the organisation of the visual system . | neuroscience/natural and synthetic vision, neuroscience/theoretical neuroscience, neuroscience/sensory systems | null |
journal.pgen.1008026 | 2,019 | Quantitative mapping of DNA phosphorothioatome reveals phosphorothioate heterogeneity of low modification frequency | Phosphorothioate ( PT ) modification of DNA , in which the non-bridging oxygen in the phosphate moiety of the sugar-phosphate backbone is replaced by sulfur , is widespread in prokaryotes in and R-configuration and a sequence-selective manner ., PT modification was originally developed as an artificial means to stabilize oligodeoxynucleotides against nuclease degradation1 ., A physiological PT modification from bacterial enteropathogens Escherichia coli B7A was first identified by liquid chromatography-coupled tandem quadrupole mass spectrometry ( LC-MS/MS ) analysis in 20072 ., The physiological PT modifications are governed by a large family of five-gene clusters termed as dndA-E3 , 4 ., The biochemical study of PT modification revealed that PT-modifying enzymes DndACDE function as a large protein complex , with DndB actings as a negative transcriptional regulator5–7 ., Many bacterial PT-modifying enzymes act in concert with and are encoded in close proximity to cognate restriction endonucleases DndFGH ., PT modifications catalysed by PT-modifying enzyme protects DNA from digestion by DndFGH with which it forms a restriction-modification ( R-M ) system8 ., The novel R-M systems protect cells from invading foreign DNA similar to methylation-based R-M systems9 ., However , more than half of all PT bacterial strains lack the dndF-H restriction system in spite of containing dndA-E and PT ., The fact that many strains of bacteria lack the restriction enzyme components of a typical R-M system is consistent with the idea that PT modifications and dndA-E genes provide functions other than R-M , such as control of gene expression10 ., PT modifications occur in DNA in a sequence-specific manner , the LC-MS/MS method was first applied to define the PT sequence contexts and quantify PT modifications in bacteria , and all 16 possible PT-linked dinucleotides in RP configuration have been detected ., In E . coli B7A , PT modifications occur in GpsA and GpsT ( phosphorothioate-containing dinucleotides ) at 370 ± 11 and 398 ± 17 PTs per 106 nt nucleotides , respectively11 ., LC-MS/MS method utilizes the PT-induced nuclease inhibition that produces a limit digest of PT-containing dinucleotides , and thus is limited to quantitative analyses of dinucleotides such as d ( GpsA ) and d ( GpsT ) but not genomic mapping of PT modifications , termed phosphorothioatome ., Recently , two sequencing technologies were applied to map bacterial phosphorothioatome: single-molecule , real-time ( SMRT ) sequencing and deep sequencing of iodine-induced cleavage at PT ( ICDS ) 10 ., The double-stranded PT modification in E . coli B7A was determined by both SMRT and ICDS to occur in the GpsAAC and GpsTTC sequence context , which is consistent with the previous observation of equimolar quantities of d ( GpsA ) and d ( GpsT ) detected in the genome by LC-MS/MS ., Only 7045 ± 470 out of 40701 possible GAAC/GTTC sites in B7A genome were detected as PT-modified in spite of the presence of the PT-dependent R-M system in this strain , indicating that the genome-wide PT distribution is partial in the B7A genome10 ., Similar partial modification has been detected in DNA methylome , which is also been characterized as epigenetic heterogeneity12 , 13 ., DNA methylation heterogeneity is widespread in organisms , which play critical roles in cell physiology , including as a source of phase variation that increase population-level phenotypic plasticity and provides opportunities to modulate transcription in response to changing environmental conditions , the colonization of animals by bacterial pathogens14–16 , and cellular development and tumorigenesis17 , 18 ., Here , we developed two techniques that provide improved resolution to observe heterogeneity of PT modifications: an iodine induced cleavage-based PT sequencing ( PT-IC-seq ) and an iodine induced cleavage-based PT droplet digital PCR ( PT-IC-ddPCR ) ., Our data reveal PT heterogeneity in bacteria ., These characterizations of PT modification will open the door to understanding biology of this novel DNA modification ., DNA sequencing is an essential tool for biological and medical studies , and it can be used not only to determine the precise order of the four nucleotide bases in genetics , but also to analyze specific nucleotide base in epigenetics ., Recently , we developed ICDS approach 10 to map PT locations in bacterial genomes based on adaptation of high-throughput next generation sequencing ( NGS ) technology ., Using this approach , only 7045 ± 470 out of 40 , 701 total possible PT modified sites across the E . coli B7A genome were detected to be PT modified sites10 ., Mapping the detected modification sites in the E . coli B7A genome showed that GpsAAC/GpsTTC distributs sporadically throughout the whole genome ., Deep resequencing has been recently used for the correction of genome sequence19 , rare genetic variation20 , 21 , and discovery of large numbers of single nucleotide polymorphisms ( SNPs ) 22 ., For constructing more accurate high-resolution genetic map of PT sites in the E . coli B7A , we conducted resequencing of E . coli B7A genome DNA by using ICDS approach with increasing sequencing depth ., The genomic DNA from E . coli B7A was isolated and treated with iodine ( S1 Fig ) ., When sequencing depth was increased to approximately 1000 × , 23 , 708 out of a total of 40 , 701 GAAC/GTTC sites were detected as PT modification sites ( Fig 1A and S1 Table ) ., The newly detected PT sites were also distributed sporadically throughout the E . coli B7A genome ., The number of PT sites were ~3 times more than the 7 , 045 sites previously detected by low coverage sequencing ( 200× ) ., If all of these 23 , 708 GAAC/GTTC sites are fully modified homogenously in every cell , it would corresponds to ~4 , 800 PTs per 106 nt in GpsA or GpsT motif , which was ~12 times more than the ~384 PTs per 106 nt quantified by LC-MS/MS previously11 ., This suggests that there are significant heterogeneity in the PT modifications , where the individual GAAC/GTTC sites were phosphotrionated in some cells but not all ., The detection of a larger number of PT modification sites is likely due to the increased sequencing depth , which led to higher sensitivity in the detection of GAAC/GTTC sites with low frequency PT modification ., In order to confirm the above finding , we firstly re-quantified PTs in E . coli B7A by LC-MS/MS to rule out the effects of biological samples ., Consistent with the previous measurements , PT modifications at the same level occurred in GpsA and GpsT contexts at 370 ± 11 and 398 ± 17 PTs per 106 nt , respectively ( Fig 1B ) ., We then randomly sampled the sequencing data to analyze the involvement of the detected PT sites in different sequencing depths ( 200 , 400 , 600 , 800 and 1000 × coverage ) ., The number of GpsAAC/GpsTTC sites achieved were plotted in Fig 1A ., 9 , 098 , 14 , 867 , 18 , 770 , 21 , 596 , 23 , 708 GAAC/GTTC sites across the E . coli B7A genome were detected as PT modified sites at 200 , 400 , 600 , 800 and 1000 × coverage , respectively ., As expected , with the increase of sequencing depth , the number of PTs detected increased rapidly ., To further verify the sequencing depth required for the detection of PT sites at low-frequency , we subsampled the sequencing data ., The analysis of reads correlation showed most relatively low-reads ( < 50 ) sites ( low-frequency sites ) could be identified as modification sites , since their reads increased with the increase of sequencing depth ( Fig 1C and S2 Table ) ., In contrast , the reads from some non-phosphorothioated GAAC/GTTC sites and randomly selected ten other sequence motifs such as CCGA , TTGA , ACGG , were not increased with the increase of sequencing depth ( Fig 1C and S3 Table ) ., These results suggested that most GAAC/GTTC sites on E . coli B7A genome could be modified at a lower frequency ( Fig 1D ) ., To investigate PT modification frequency in genome landscape at single-base resolution , we sought to develop a high-throughput assay to quantitatively determine the PT modification percentage at each site ., In the ICDS approach , iodine reagent was introduced to cleave DNA at GpsAAC/GpsTTC sites , and then ligated to an adaptor with specific index sequence for enriching the DNA fragments with PT modifications by PCR amplification ., Since the PCR amplification step will result in enriched amplicons with PT modifications , ICDS sequencing approach can only be used for identifying PT sites but not for quantification ., Based on these considerations , we developed one approach named iodine induced cleavage-based PT sequencing ( PT-IC-seq ) to quantitatively determine the PT modification percentage without enrichment process ., As shown in Fig 2A , DNA sample was treated with iodine and then sonicated to 150–350 base pair fragments , end-repaired , adenylated , and ligated to Illumina adapters ., After PCR amplification , a standard Illumina DNA library was constructed for high-throughput sequencing ., The modified GpsAAC/GpsTTC motifs would be cleaved and should be presented as the reads ends in the sequencing output ., The unmodified GAAC/GTTC motifs would not be cleaved and should be present in the internal locations of DNA fragments ., The ratio of a specific GAAC or GTTC sites with sequence reads at the end versus internal represents the relative PT modified to unmodified ratio ., Through mapping sequencing reads to the reference genome , PT modification frequency for every GAAC or GTTC site can be calculated ., To test the feasibility of PT-IC-seq for determining PT modification frequency , the DNA of plasmid BlueScript SK ( + ) from Salmonella enterica serovar Cerro 87 we isolated and tested as an example ., By aligning the output reads to the plasmid sequence , as expected , we found that the most of intact GAAC or GTTC motifs appeared at the internal of sequencing reads , which should be from unmodified GAAC/GTTC sites , and some of sequencing reads were initiated with AAC or TTC , which resulted from the double-strand cleavage of PT modified GpsAAC/GpsTTC by iodine ( Fig 2B ) ., For evaluating the PT level of individual GAAC/GTTC sites , we counted and calculated the ratio of reads obtained from fragment terminals to total reads of each site ., The PT modification level of all 25 GAAC/GTTC sites throughout the entire plasmid were calculated and shown in S4 Table ., Meanwhile , to exclude the false-ended reads generated by random shearing , eight non GAAC/GTTC sites were randomly selected as control sites ., The false PT detection possibility was estimated to be below 1% ( ~0 . 54% , S4 Table ) ., Statistics analysis of PT modification frequency showed that most of the GAAC/GTTC sites ( 72% ) were modified with a low PT modification percentage of < 5% ( Fig 2C and S4 Table ) ., These results demonstrated that the PT-IC-seq approach could quantitatively determine the PT ratio of each GAAC/GTTC site and was suitable for investigating PT modification frequency of whole genome simultaneously ., Given the successful detection of PT modification percentage of each modified sites in plasmid DNA , the PT-IC-seq approach was then applied to genomic DNA isolated from E . coli B7A ., The obtained sequencing reads were mapped to a reference genome of E . coli B7A ( GenBank accession No . CP005998 . 1 ) ., About 85 percent of all GAAC/GTTC sites ( 34 , 796/40 , 701 ) across the E . coli B7A genome were detected with PT modification , and the PT modification frequency of each site was shown in S5 Table ., The false PT detection possibility was estimated to be below 1% ( ~0 . 62% , S5 Table ) ., Among the 34 , 796 PT modification GAAC/GTTC sites , 24 , 477 sites ( 60 . 1% ) were found to have a modification percentage below 5% ( Fig 3B ) , which was consistent with the results of plasmid that 72% of the total detected PT sites were modified with a low PT ratio ., The number of PT sites calculated by PT-IC-seq is 392 . 4 per 106 bp , which was very consistent with the MS measurements of 370 ± 11 GpsA and 398 ± 17 GpsT per 106 nt ., In addition , 10 , 326 5 , 379 , 2 , 718 and 360 GAAC/GTTC sites were identified with higher PT frequencies than 5% , 10% , 20% and 30% , respectively ., Only two sites’ PT ratio exceed 35% were 35 . 36% and 35 . 17% respectively ., The low modifcation frequency suggests heterogeneity of PT modification in the bacterial population ., That is , only a small proportion of individual cells contain PT at specific sites , whereas other cells do not ., The low frequency PT modification may indicate distinctive roles yet to be identified ., The majority of GpsAAC/GpsTTC sites were modified with a low percentage was also consistent with the phenomenon that a large part of GAAC/GTTC motif could be PT modified while the actually detected GpsA and GpsT are only at low level ( < 400 PTs per 106 nt ) ., By comparing the 360 high-frequency sites ( >30% ) and the 9098 PT sites determined by the ICDS approach with 200 × coverage , we found that 353 out of the 360 ( 98 . 06% ) sites with a modification frequency more than 30% overlap with the 9098 PT modified sites detected by ICDS with 200 × coverage ( S6 Table ) ., Meanwhile , 8391 out of the 9098 ( 92 . 2% ) PT modified sites detected by ICDS with 200 × overlap with the 10319 sites with a modification frequency more than 5% by PT-IC-seq ., The result from these comparisons shows that the sites with higher modification frequency detected by PT-IC-seq can also be detected as PT modified sites by ICDS with lower sequencing depth ., These results indicate that the two sequencing methods are very consistent for detecting PT modification sites ., The percentage of PT-modified GAAC/GTTC sites with PT modification frequency > 5% in different regions of the genome varied from 17 . 1 to 39 . 8% , with 9 , 537 of 36 , 607 sites ( 26 . 0% ) modified with PT in ORFs , 156 of 392 ( 39 . 8% ) in ncRNAs and 633 of 3702 ( 17 . 1% ) in the noncoding regions ., Among the non-coding region , there are 224 of 725 sites ( 30 . 9% ) in the promoter regions modified with PT modification frequency > 5% ., The relatively high-frequency ( > 20% ) PT sites were distributed in ORF and non-coding regions other than ncRNA; while those PT frequency < 20% distributed relatively evenly across the B7A chromosome ( Fig 4 ) ., We performed an extended motif search based on the 360 high-frequency sites to examine whether there is any additional preference of nucleotides flanking the GAAC or GTTC sequence ., However , no additional consensus nucleotides were observed ( S7 Table ) ., In order to evaluate the creditability of the PT-IC-seq quantitative assays , we then developed another method named PT-IC-ddPCR for validation of the PT modification frequency ., The PT-IC-ddPCR approach was developed based on iodine-induced selectively cleavage at PT sites and inherent property of droplet digital PCR ( ddPCR ) technique for absolute DNA quantification ., In the PT-IC-ddPCR ( Fig 5A ) , a portion of the PT modified genome DNA sample ( system A ) was treated first by iodine ., The double strands of those PT modified molecules broke at specific site while the non-modified DNA remained intact ., When the designed primers and probes ( S8 Table ) were used for quantitative ddPCR , only the non-modified DNA was amplified and quantified , and the results were recorded as X copies ( Fig 5A , system A ) ., Similary , another equal portion of the DNA sample ( system B ) was treated by ethanol instead of iodine , the treated DNAs were then quantified by ddPCR , and the obtained results were determined as Y copies ( Fig 5A , system B ) ., The Y copies in the system B was the sum of modified and unmodified molecular number at target sites , which theoretically was equivalent to the total DNA number in system A . The number of PT modified molecules at this site was calculated as Y-X ., The PT modification frequency of target site can be calculated by the percentage of PT modified molecular to the total sum of the PT modified and non-modified molecular number , which was ( Y-X ) / Y . Using the PT-IC-ddPCR method , we first evaluated its dynamic range and correlation employing 1818096 loci in E . coli B7A genome as an example ., The observed positive droplets matched well with the predicted values , and the R2 value of 0 . 9932 indicated that this standard curve had good linearity and the quantified results were accurate with low RSD ( all < 25% ) ( Fig 5B and 5C ) ., The PT modification frequency at this position obtained from PT-IC-ddPCR method is 8 . 78% , which was consistent with 5 . 19% obtained from PT-IC-Seq method ( S9 Table ) ., PT-IC-ddPCR method were then applied to another 7 sites in E . coli B7A genome , which were detected as PT modification sites by PT-IC-seq ., As can be seen from S9 Table , the PT modification frequency obtained from PT-IC-ddPCR at all 8 sites were consistent with the value obtained from PT-IC-seq method ., Therefore , this assay provided locus-specific validation of the PT-IC-seq results ., Although the PT ratio value from PT-IC-Seq method is a slightly different from those of PT-IC-ddPCR analysis , the overall trend is highly consistent ( S9 Table ) ., As is shown in Fig 5D , all the 8 sites’ PT modification frequency in a population of DNA molecules were no more than 50% ., The PT-IC-ddPCR assay identified 8 GAAC/GTTC sites were partially modified and demonstrated PT modification has heterogeneity in E . coli B7A genome ., After confirming the heterogeneity feature of PT modification in E . coli B7A , we then attempt to explore whether the heterogeneity of PT modification is ubiquitous phenomena among other PT modified strains ., The PT-IC-Seq method was then applied to the genomes of S . enterica serovar Cerro 87 ., By aligning the sequencing output reads to the S . enterica serovar Cerro 87 reference genome ( GenBank accession No . NZ_CP008925 . 1 ) , we found that in addition to these reads ended at each PT site , there were also a greater number of reads crossing over the same sites just as in E . coli B7A genome ( Fig 6A ) ., The existence of both the reads ended and crossing over at the same sites , suggests that the genomic PT modification in the S . enterica serovar Cerro 87 shows heterogeneity as well ., After calculating the ratio of the iodine induced cleaved reads to all of the reads across or ended at this site , we found that all of the detected PT modified sites in S . enterica serovar Cerro 87 genome were also partially modified , and the percentage were also below 40% ( S5 Table and Fig 6B ) ., And about 35% of PT modified sites were modified with a PT modification percentage below 5% ., The fact that PT modification in S . enterica serovar Cerro 87 also is characteristic of heterogeneity suggests that heterogeneity might be ubiquitous among PT modified strains ., The quantitative mapping of PT sites in the Salmonella genome further supports that the PT-IC-seq method is a valuable tool to study PT-modified strains ., DNA and RNA modifications play critical roles in various biological processes , including R-M systems serving to protect bacteria against bacteriophages in prokaryotes , and as epigenetic marks for DNA replication , repair , recombination , chromatin organization and hypermutation in all organisms23 ., The best known DNA modifications in organisms are methylation of cytosine to produce 5-methylcytosine ( m5C ) and adenine to N6-methyladenine ( m6A ) ., The recently discovered PT modification of the DNA backbone in bacteria is an unusual DNA physiological modification , in that the nonbridging oxygen in the sugarphosphate backbone of DNA is replaced by sulfur ., Benefiting from the latest developments of SMRT and ICDS sequencing methods , PT modified sites across the E . coli B7A genome have been mapped in recent studies 10 ., The well characterization of the genomic landscape of PT modifications in bacteria have led to a greatly enhanced interest in the biology of PT ., However , a quantitative map at single-base resolution of the PT modification percentage information is key to understand its biological function ., Here , we developed PT-IC-seq and PT-IC-ddPCR to quantify PT modification sites in the E . coli B7A genome ., As illustrated in Fig 2 , PT-IC-seq combined iodine-induced cleavage at PT sites and high-throughput NGS was first used to quantitatively characterize the PT modification at the scale of whole genome ., The feasibility of the PT-IC-seq technique was demonstrated in not only plasmid but also genomic DNA ( Figs 2 and 3 ) ., Similar to m6A-RE-seq , an approach that uses DpnI to cleave methylated adenine sites in duplex DNA13 , 24 , PT-IC-seq uses iodine reagent to cleave double-strand PT modified DNA ., Iodine reagent sensitively and specifically cleaving PT-containing DNA have been demonstrated10 , and it only cleaves fully PT modified sites , thus PT-IC-seq strategy was limited to genomic mapping of bistranded PT modifications but not the single-stranded modifications ., Although SMRT sequencing can distinguish modified PT from unmodified DNA and has been applied in genomic mapping of PT , the relatively high expense , low detection signals ( the kinetic interpulse duration signals for PT modifications smaller than m6A signals ) and the use of subjective thresholds10 have prevented its wide applications in genomic mapping of PT ., Comparing to ICDS sequencing , PT-IC-seq can quantitatively map PT modifications because the ratio of a specific GAAC or GTTC sites with sequence reads at the end versus internal represents the relative PT modified to unmodified ratio , which ICDS sequencing could not do ., Furthermore , the PT modification frequency of specific target sites were absolutely quantified with the PT-IC-ddPCR to confirm the results of PT-IC-seq approach ., Consistent with the PT-IC-seq results , both of them identified all eight sites have similar PT frequency ., The low PT frequency sites based on PT-IC-seq results were determined to be low by PT-IC-ddPCR as well ( S9 Table ) ., PT-IC-ddPCR also showed good ability in quantifying the PT sites with low modification frequency ., However , the relatively high expense and low throughput also have hindered their applications in genomic mapping of PT ., Furthermore , the LC-MS/MS method developed previously , although sensitive enough to detect and accurately quantify the amount of the very low abundance of PT modifications , can only provide the overall ratio of the modification in total DNA ., Therefore , PT-IC-seq method combining high throughput sequencing and bioinformatics analysis is a powerful strategy to quantitatively map PT modification sites in bacteria genome and to explore the PT biology ., We then applied this PT-IC-seq strategy to E . coli B7A and S . enterica serovar Cerro 87 genomic DNA and obtained genome-wide PT maps at single-nucleotide resolution that show PT heterogeneity in bacterial populations ( Fig 3 and S5 Table ) ., This PT heterogeneity phenomenon in bacterial populations is similar to recent observation of DNA methylation abundance in several distinct prokaryotes and even eukaryotes 12 ., In E . coli , Chromohalobacter salexigens , Geobacter metallireducens , Campylobacter jejuni and Helicobacter pylori , the detected m6A or m4C methylations reveal distinct types of epigenetic heterogeneity12 ., In Plasmodium falciparum and Penicillium chrysogenum , most of the G ( m6A ) TC sites were partially methylated with a low percentage of methylation ratio ( below 10% ) 13 ., Given the fact that solitary PT-modifying enzyme occur in more than half of all PT bacterial genome without an associated restriction enzyme component , PT heterogeneity in E . coli B7A genome suggests that PT modification may work as an marker other than R-M system ., Heterogeneity is beneficial for organisms , because it enables some cells to survive in harsh environments ., When sudden changes in chemical composition 25 , local temperature 26 or other environment conditions occur , the heterogeneous population may contain some individuals that could cope with it , and thereby maintain the survival of the whole population ., Hence , heterogeneity increases the overall population fitness when environmental shifts are unpredictable ., As other organisms heterogeneity , PT heterogeneity may increase population-level phenotypic plasticity in response to changing environmental conditions or better preservation of certain functions of PT modification in the population to ultimately maintain the survival advantage of the population ., These issues are yet to be revealed and more research is needed to answer these questions ., Other than the roles of the PT heterogeneity in bacteria , the mechanism of heterogeneity for modification is also unclear ., We propose two possible mechanisms for this PT heterogeneity phenomenon ., One possibility is that , intracellular PT-modifying enzymes , which function as a large protein complex5 , could keep at a low activity level ., The other possibility is that endogenous oxidants , metals or alkylating agents result in nonspecific desulfuration of PT modified DNA27 ., In summary , we developed PT-IC-seq and PT-IC-ddPCR approaches to quantify PT modification sites in bacteria genome ., With these two methods , we clearly demonstrated that the widespread PT modification has heterogeneity in bacterial population which has important implications for the future study of PT modification ., The following materials were obtained from Sangon Biotech Co ., Ltd . ( Shanghai , China ) : Enantiomerically pure d ( GpsA ) and d ( GpsT ) in Rp and Sp configuration , the custom oligodeoxynucleotides duplex tag which marked the iodine nick in the ICDS , the primers and probes for PT-IC-ddPCR were listed in S6 Table ., The TaqMan probes which were labeled with 6-carboxyfluorescein ( FAM ) at the 5’ end and black hole quencher ( BHQ I ) at the 3’ end ., The plasmid pBluescript SK ( + ) was obtained from Life Technologies ( Grand Island , NY , USA ) ., Salmonella enterica serovar Cerro 87 was supplied by Professor Toshiyuki Murase ( Tottori University , Japan ) , E . coli B7A was obtained from Dr Jaquelyn Fleckenstein ( Departments of Medicine and Molecular Sciences , University of Tennessee Health Science Center ) 2 , 11 ., Bacterial DNA Kit was purchased from TIANGEN ( Cat . no . DP302-02 TIANGEN , Beijing , China ) ., Plasmid Mini Kit and Cycle pure Kit were purchased from OMEGA ( Omega Bio-tek , USA ) ., The following kits and reagents were purchased from New England BioLabs ( Ipswich , MA , USA ) : Antarctic Phosphatase , Quick Blunting Kit , Quick Ligation Kit , Klenow Fragment ( 3′-5′ exo_ ) , dATP solution ., Centrifugal filters ( 10 KD ) were from Millipore ( EMD Millipore , Billerica , MA , USA ) and MicroSpin G-25 columns were from GE Healthcare ( Buckinghamshire , UK ) ., Iodine and nuclease P1 were from Sigma-Aldrich ( St Louis , MO , USA ) ., PCR tubes were from Molecular BioProducts ( San Diego , CA , USA ) , Alkaline phosphatase were from Fermentas ., 2×ddPCR Supermix and droplet generation oil were from Bio-Rad ( Bio-Rad ) ., All water was deionized and filtered using a MilliQ water purification system ( EMD Millipore , Billerica , MA , USA ) ., Phosphorothioate modifications in E . coli B7A were quantified by HPLC and LC-MS/MS ( ACQUITY UPLC & SCIEX SelexION Triple Quad 5500 System ) ., 20 μl of DNA was digested by nuclease P1 ( Sigma-Aldrich , Cat . no . N8630 ) in 30 mM NH4OAc pH 5 . 3 , 0 . 5 mM ZnCl2 in a 100 μl of total volume at 50 °C for 2 h ., After the P1 hydrolyzation , 10 μl of 1 M Tris-HCl , pH 8 . 0 and 5 U of alkaline phosphatase ( Fermentas , FASTAP ) were added for the fully dephosphorylation at 37 °C for another 2 h ., The enzymes were subsequently removed by ultrafiltration ., Then the digested samples were dried and resuspended in 40 μl of deionized water ., The hydrolyzed mono nucleosides subsequently quantified by HPLC according to standard curves of pure A , T , C , and G nucleosides ., Meanwhile , the PT-modified dinucleotides were quantified by LC-MS/MS according to standard curves of pure Rp configuration PT linked dinucleotides GpsA/GpsT ( Rp ) and Sp configuration PT linked dinucleotides GpsA/GpsT ( Sp ) used as internal standards ., Thus , the number of PT modified dinucleotides in a unit length of DNA can be calculated ., Approximately 20 μg of purified genomic DNA was cleaved by I2 in a 100 μl reaction system which is consist of genomic DNA ( 20 μg ) , 50 mM Na2HPO4 , 3 mM I2 at pH 9 . 0 ., Reaction system were incubated in 65 °C for 15 min and then slow cooled ( 0 . 1 °C/s ) to 4 °C using a thermal cycler ., Following iodine cleavage , the DNA segments were subjected to end repairment , 3’-deoxyadenylation and unique tag ligation at double-strand break sites according to instructions provided with the NEBNext DNA Library Prep Reagent Set for Illumina ( New England BioLabs , Beverly , MA ) which essentially described as follows: Terminal phosphates of the iodine cleaved DNA samples were removed with alkaline phosphatase ( Fermentas , FASTAP ( 10 units ) ) at 37 °C for 60 min ., Then the system was heated to 65 °C for 10 min for the inactivation of phosphatase , t and subsequently cooled to 4 °C slowly ( 0 . 1 °C s-1 ) to assure proper complementary re-annealing ., Samples were cleaned up using cycle pure kit ( OMEGA ) and eluted with 30 μl MilliQ water ., Break sites were blunt-ended using the Quick Blunting Kit at room temperature for 30 min ., The enzyme was inactivated by heating the samples to 75 °C for 10 min and then slowly cooling them ( 0 . 1 °C s-1 ) to 4 °C ., Samples were cleaned up using cycle pure kit and eluted with 30 μl MilliQ water ., Next , blunt-ends were 3’-deoxyadenylated ( that is , A-tailing ) in reactions ( 100 μl ) containing 1x NEB Buffer #2 , 0 . 1 mM dATP , and Klenow fragment ( 3’-5’exo- ) ( 15 units ) at 37 °C for 30 min ., The resulted system was treated as before and eluted with 30 μl MilliQ water ., The enzyme was inactivated by heating the samples to 75 °C for 20 min and then slowly cooling them ( 0 . 1 °C s-1 ) to 4 °C ., Finally , a custom 20-mer duplex tag-sequence ( 5’-FWD tag ( 5’-/Phos/TTTAACCGCGAATTCCAG /dideoxyC/-3’ ) / 3’-REV tag ( 5’-GCTGGAATTCGC- GGTTAAAT-3’ ) ) ( 3 μM ) was ligated to 3’-deoxyadenylated ends using T4 ligase at 16 °C for 16 h ., The enzyme was inactivated by heating and cooled to 4 °C as before , and the recovered sample with cycle pure kit was also resolved in 30 μl MilliQ water ., Then the quality of the samples was tested: the purity of the samples was measured using Nano Drop 2000; The concentration was measured using picogreen ., Next , the DNA samples were sheared to fragments of optimal range ( 150–350 bp ) for Illumina sequencing and followed by adaptor ligation ., Unique tag marked and adaptor-ligated DNA segments were PCR amplified for 15 cycles and the segments which only have I2-cleaved ends for which the linked unique tag have been enriched ., The prepared PE library was qualified and quantified by Qubit 3 . 0 Fluorometer ( Life Technologies ) , agarose gel electrophoresis and Agilent 2100 bioanalyzer ., The Library was sequenced on the Illumina HiSeq X Ten platform ., After Illumina sequencing completed , the reads which containing tag were selected and done adaptor and tag trimming and quality control as follows:, 1 ) Clipping the adapter sequences;, 2 ) Removing non-A , G , C , T bases of the 5 end;, 3 ) Trimming low quality sequencing reads ( base quality is less than Q20 ) ;, 4 ) Removing reads with > 10% of“N”base calls;, 5 ) Filtering small fragments less than 25 bp after clipping the adapter sequences and quality trimming ., Then aligned to reference genomes by Burrows-Wheeler Aligner ( BWA ) and the position-wise coverage values were calculated using a custom python script ., The GAAC/GTTC sites that only measured above 50 reads which ended at this site were regarded as PT modified sites ., And 10 non GAAC/GTTC sites were randomly selected as control ., Approximately 20 μg of genome DNA was cleaved by I2 as des | Introduction, Results, Discussion, Materials and methods | Phosphorothioate ( PT ) modifications of the DNA backbone , widespread in prokaryotes , are first identified in bacterial enteropathogens Escherichia coli B7A more than a decade ago ., However , methods for high resolution mapping of PT modification level are still lacking ., Here , we developed the PT-IC-seq technique , based on iodine-induced selective cleavage at PT sites and high-throughput next generation sequencing , as a mean to quantitatively characterizing the genomic landscape of PT modifications ., Using PT-IC-seq we foud that most PT sites are partially modified at a lower PT frequency ( < 5% ) in E . coli B7A and Salmonella enterica serovar Cerro 87 , and both show a heterogeneity pattern of PT modification similar to those of the typical methylation modification ., Combining the iodine-induced cleavage and absolute quantification by droplet digital PCR , we developed the PT-IC-ddPCR technique to further measure the PT modification level ., Consistent with the PT-IC-seq measurements , PT-IC-ddPCR analysis confirmed the lower PT frequency in E . coli B7A ., Our study has demonstrated the heterogeneity of PT modification in the bacterial population and we also established general tools for rigorous mapping and characterization of PT modification events at whole genome level ., We describe to our knowledge the first genome-wide quantitative characterization of PT landscape and provides appropriate strategies for further functional studies of PT modification . | Phosphorothioate ( PT ) modification is a novel DNA modification , previous studies showed that PT modifications in E . coli occure at GpsAAC/GpsTTC motifs , but the modification frequency at each site are not known ., In this study , we introduced two methods: PT-IC-seq , which could quantitatively characterize the genomic landscape of PT modifications; and PT-IC-ddPCR , which could measure PT modification frequency precisely ., Through these two new methods , heterogeneity , an important feature of PT modification , is revealed intuitively and accurately ., The modifications revealed in this study provide a basis for the final revelation of the physiological functions of PT and give valuable reference for other DNA modification studies . | bacteriology, medicine and health sciences, pathology and laboratory medicine, pathogens, microbiology, organisms, bacterial diseases, enterobacteriaceae, genome analysis, bacterial genetics, sequence motif analysis, dna, epigenetics, molecular biology techniques, microbial genetics, bacteria, dna cleavage, bacterial pathogens, microbial genomics, research and analysis methods, bacterial genomics, sequence analysis, infectious diseases, artificial gene amplification and extension, bioinformatics, medical microbiology, gene expression, microbial pathogens, chemistry, salmonella enterica, dna modification, molecular biology, salmonella, biochemistry, iodine, nucleic acids, polymerase chain reaction, database and informatics methods, genetics, biology and life sciences, physical sciences, genomics, computational biology, chemical elements | null |
journal.pcbi.1002863 | 2,013 | A Guide to Enterotypes across the Human Body: Meta-Analysis of Microbial Community Structures in Human Microbiome Datasets | Together with the MetaHIT consortium 1 , the Human Microbiome Project ( HMP ) represents one of the first major attempts to define the microbial diversity comprising the “normal healthy” human microbiome 2 ., The HMP dataset includes 16S rRNA gene sequence data of roughly twice the size of all similarly derived data in previously published studies , effectively tripling the size of combined data available for comparative studies ( Table S1 ) ., In addition , the HMP generated whole-genome shotgun ( WGS ) metagenomic data for a subset of individuals ., These data allowed for the characterization of patterns of microbial diversity across body sites and between individuals 2 ., The HMP data also provides an opportunity to test the generality of the concept of enterotypes in the human microbiome ., Arumugam et al . first articulated the concept of enterotypes as robust clustering of human gut samples based on microbial community composition , and largely driven by the abundances of key bacterial genera 3 ., Although the term ‘enterotype’ refers to microbiota types within the gut , the concept can be applied generally , and here , for convenience , we use the term ‘enterotype’ to refer to microbiota types across different body sites ., The HMP data are ideally suited to test the robustness of the enterotype concept in multiple body sites , and together with recently published community-generated datasets , across multiple populations ., In this report , we combined 16S rRNA gene sequence data generated using next-generation sequencing by the scientific community ( hereafter , ‘community data’ ) together with the MetaHIT WGS data 3 and the recently released HMP 16S rRNA gene sequence data and WGS data 2 ., Because there is currently no community standard for testing for enterotypes , we explore how the detection of enterotypes is affected by the following: clustering methodology , distance metrics , OTU-picking approaches , sequencing depth ( i . e . , rarefaction ) , data type ( 16S rRNA vs . WGS ) , and the specific region of the 16S rRNA gene sequenced ., We find that the emergence of enterotypes is sensitive to the community structure of communities within each body site , and importantly also to the analysis methods employed ., Our comparative analysis of various approaches across datasets informs the discussion on the technical basis for enterotyping and on how to interpret enterotype results ., We constructed a database containing the recently released HMP 16S rRNA gene sequence data 4 and publically available ( published ) human microbiome datasets ( community data ) ., For inclusion , community datasets were required to contain a minimum of 25 samples per study and sequences generated using the Roche 454 platform ( Table S1 ) ., The majority of samples were from healthy controls; however , a small subset of samples was derived from subjects that differed from adult healthy subjects due to age ( i . e . , infants and the elderly ) , use of antibiotics , or possible presence of disease ( Fig . S1 ) ., We acquired raw SFF files and metadata files containing the unique identifiers for each sample within a study ( barcodes ) from the authors and re-processed the data using the default settings in the Quantitative Insights Into Microbial Ecology ( QIIME ) analysis pipeline 5 ., For the majority of samples , quality filtering consisted of rejecting reads <200 nt and >1000 nt , excluding homopolymer runs >6 nt , accepting 0 barcode corrections and 0 primer mismatches; two datasets were processed with slightly different screening parameters , as described in their respective publications 6 , 7 ., When picking operational taxonomic units ( OTUs ) we used the OTU tables generated by the HMP , which were created de novo ., Because the regions of the 16S rRNA gene differed between studies ( and within: the HMP sequenced both V1–V3 and V3–V5 regions ) , we used a reference-based approach ( hence , we did not denoise the data ) to pick OTUs at 97% pairwise identity using as a reference the latest release of the GreenGenes ( GG ) taxonomy 8 ., We also used the phylogenetic tree from GG to calculate weighted ( abundance based ) and unweighted ( presence/absence based ) UniFrac distances between communities 9 , after applying two rarefactions ( 1 , 000 and 2 , 000 sequences/sample ) to standardize sequence counts ., Principal coordinates analysis ( PCoA ) was applied to the distance matrices for visualization ., The HMP and MetaHIT shotgun metagenomic datasets were taxonomically profiled using MetaPhlAn 10 ( version 1 . 1 , default parameter settings ) , which infers relative abundances for all taxonomic levels ( from phyla to species ) for Bacteria and Archaea ., We performed standard quality control on the HMP and MetaHIT samples as reported in the original studies 2 , 3 - other metagenomic pre-processing steps ( e . g . , error detection , assembly , or gene annotation ) are not required by MetaPhlAn ., The taxonomic profiles of HMP metagenomes are available at http://www . hmpdacc . org/HMSMCP/ , and the MetaHIT profiles can be downloaded from http://huttenhower . sph . harvard . edu/metaphlan/ ., The 690 HMP metagenomic samples from 7 different body sites can be accessed at http://hmpdacc . org/HMASM/ , from which we used the ‘WGS’ reads ( i . e . , we did not use the ‘PGA’ assemblies ) , collapsing multiple visits from the same individual into one sample ., The 124 fecal samples from MetaHIT were downloaded from the European Nucleotide Archive ( http://www . ebi . ac . uk/ena/ , study accession number ERP000108 ) ., To evaluate the clustering results in the context of previously published results reporting enterotypes , we merged publicly available data ( genus relative abundance tables ) from MetaHIT 3 with data for the 16S rRNA-based HMP and non-HMP samples , based on genus-level taxonomy assignments ., We performed enterotype testing using the relative abundances of OTUs ( rarified at 1 , 000 sequences/sample for the majority of analyses , except where effect of rarefaction was tested specifically ) , to which we applied five distance metrics: Jensen-Shannon divergence ( JSD ) , Root Jensen-Shannon divergence ( rJSD ) , Bray-Curtis ( BC ) , and weighted/unweighted UniFrac distances ., For the calculation of JSD and BC distances , we first binned the counts of OTUs at the desired level ( 95% and 97% ID for genus and species level OTUs , respectively ) ., We used the R “vegan” package 11 for calculating the Bray-Curtis distance according to this formula for the distance between samples j and k , with taxa/OTUs indexed by i:Clustering was performed via partitioning around medoids in the R package “cluster” 12 ., We chose the number of clusters and quality of the resulting clusters by maximizing the prediction strength ( PS ) 13 and silhouette index ( SI ) 14 ., We applied a criterion of ≥0 . 90 for PS to signify strong clustering ( this implies that 90% of the data points fall within the cluster and 10% are outliers ) ., For SI , we used a score of 0 . 5 for moderate clustering as described by Wu et al . 15 , and ≥0 . 75 for strong clustering ( note this is close to the value of 0 . 71 originally reported for strong clustering 16 ) ., We performed kernel density estimation of the global distribution of gut microbial communities using the R package “ks” 17 ., This included automatic inference of unconstrained ( non-diagonal ) bandwidth parameters using the function “Hscv” ., We also calculated the Caliński-Harabasz ( CH ) statistic for comparison to PS and SI , using the R ‘fpc’ package 18 ., This package uses the following formula for the CH statistic 19:whereandIn this formula , n is the number of data points w , k is the number of clusters , and Ch represents the set of data points in cluster h ., We chose to use PS and SI to assess the support for clustering ( and to choose the number of clusters if supported ) , as they are both absolute measures of the clustering quality , while CH is only a relative assessment of the quality of the clustering ., We generated a synthetic dataset of 100 communities each containing 3 , 000 “sequences” belonging to 500 mock OTUs ., For each synthetic community , 90% ( 2 , 700 sequences , or OTU observations ) was drawn from the same randomly generated lognormal abundance distribution ( shared across all communities ) and the remaining 10% ( 300 sequences ) drawn from one of four unique lognormal distributions , forcing the data into four clusters ., We then applied the enterotyping methods as described above ., Beta-diversity measures provide a view of how diversity differs between sets of samples and quantifies those differences ., We used the unweighted UniFrac measure of β-diversity to contrast the range of bacterial phylogenetic diversity captured by the HMP data to existing community data ( Fig . 1 ) ., This analysis showed that the overall pattern of diversity is similar for HMP and community data , with clear separation between body sites ( Figs . 1A , B , S1 ) as has been described previously 2 , 20 ., Similarly , Fig . S2 shows the locations of the MetaHIT samples relative to the HMP and other community fecal samples ., The HMP and MetaHIT data map onto the community data well ( Fig . 1C; Fig . S2 ) , lending support for the approach of combining these sets in a meta-analysis of body habitats ., The gut microbiota are the most extensively studied of the human-associated microbiota ., Combining community and HMP fecal microbial 16S rRNA gene sequence data effectively extended the subject age range from early infancy to old age ( 3 days to 85 years old ) , with the HMP supplying the majority of the middle years of the human life span ., Interestingly , infant samples ( younger than 2 . 5 years ) were outliers in the range of diversity represented by the healthy adult and elder ( older than 70 years ) gut ( Fig . 1D ) and were more similar to vaginal and skin communities ., Adult HMP samples cluster together with those from the community studies , excluding samples from infants ( <2 . 5 yrs ) and elders ( Table S1 ) ., This combined analysis corroborates the previously described vast difference between bacterial diversity of infants and adults 21 , 22 ., We first tested the effects of different cluster scoring methods using a lognormally distributed synthetic community data containing 4 clusters that served as a positive control for enterotypes ., We applied the JSD , rJSD and BC distance measures to the synthetic dataset and compared cluster scores using prediction strength ( PS , Fig . 2A ) , silhouette index ( SI , Fig . 2B ) and Calińksi-Harabasz ( CH , Fig . 2C ) scores ., This analysis revealed strong support for 4 clusters using PS for the BC , JSD and rJSD distance metrics , but SI provided no support for clustering using BC and rJSD , and only weak support for 3–5 clusters using the JSD distance metric ., The CH index supported 4 clusters using only the JSD distance metric ., Wu et al . also reported a discrepancy in cluster scoring strengths between clustering methods 15: CH indicated that 3 enterotypes were present , but SI provided weak support using rJSD ., Wu et al . also compared clustering with CH and SI together with weighted/normalized unweighted UniFrac , BC and Euclidean distances , and reported concordant numbers of clusters with weighted UniFrac only ., Together these results indicate that these different clustering methodologies can yield inconsistent results , although SI and CH have been reported to be stable and comparable 23 , 24 ., Arumugan et al . used CH as the basis for choosing the number of enterotypes , even when SI values were very low ( all published values were less than or equal to 0 . 25 ) , indicating weak or no support for clustering 3 ., It is important to note that the CH score is a relative measure that alone cannot be used to determine statistical significance of clustering in the data , and that furthermore , CH is intended to indicate the optimal number of clusters based on the assumption that clusters exist ., PS and SI , on the other hand , are absolute measures of how likely cluster structure is to emerge from a dataset ., Based on our results , we recommend using at least one absolute measure ( specifically , we recommend PS ) , and if possible confirming those results with an additional absolute measure ( such as SI ) , when searching for enterotypes ., Depending on the signal-to-noise distribution within individual datasets and data types , PS may have difficulty identifying clusters represented by few samples , as we discuss below ( e . g . , posterior fornix WGS data ) ., In such cases SI may be relied on , but we recommend using a high threshold ( e . g . , ≥0 . 75 ) in identifying potentially reproducible clusters ., We prefer PS over SI for large sample sizes because ( 1 ) it has a clear quantitative and intuitive interpretation , ( 2 ) it allows estimation of the clustering stability of individual samples , and ( 3 ) it performs better than SI in recovering known enterotypes in synthetic datasets ., Note however that there is currently no consensus in the field on the specific thresholds that should be used with these methods for assessing clustering strength , making it all the more important for researchers to clearly state the criteria they apply when reporting enterotypes ., We searched for fecal enterotypes in the HMP and community 16S rRNA gene sequence data using the relative abundances of OTUs across samples , and applying five different distance metrics: JSD , rJSD , BC , and weighted/unweighted UniFrac distances , and three cluster evaluation methods ( PS , CH and SI; Fig . 3 ) ., Using PS , we observed at best moderate support for 2 fecal enterotypes in the HMP data using weighted UniFrac , but little or no support using other distance metrics ( Fig . 3A ) ., We obtained similar results using community data alone and when combined together ( Fig . 3A ) ., Weighted UniFrac scoring for enterotypes was weak with SI ( Fig . 3 ) ., Figs ., S3 , S4 , S5 , S6 , S7 show similar analyses for 3 different vaginal sites , and 9 oral and 3 skin sites ., Moderate to strong clustering is evident in only 3 out of these 15 body sites ., In the mid vagina there is strong support for 2 clusters ( discussed below ) using BC , JSD and rJSD ( Fig . S3; weighted UniFrac provided moderate support; unweighted UniFrac provided no support ) ., In the posterior fornix ( Fig . S3 ) and the attached keratinized gingiva ( Fig . S4 ) , we observed moderate support for 2 clusters using 5 and 4 distance metrics , respectively ( unweighted UniFrac resulted in little or no support in the gingiva ) ., These results indicate that the detection of enterotypes is sensitive to the distance metric used , a result also recently reported by Claesson et al . 25 ., Note that this sensitivity is not dependent on body site ., However , because enterotyping is driven by the relative abundances of specific genera within samples , unweighted UniFrac , which takes into account presence/absence of tree branches but not abundances of sequences mapping to those branches , may not be an ideal distance measure to use for enterotyping ., We include it here because it is widely used in microbiome studies ., In contrast , the weighted UniFrac , BC , JSD and rJSD distance metrics are based on OTU abundances and should in principle be more appropriate ., The lack of concordance between results based on different abundance-based distance metrics raises the following questions: if enterotypes are to be considered robust , must they be observed using more than one distance metric ?, Or does the lack of concordance between results using different distance metrics indicate that ( here at least ) weighted UniFrac is the best choice for enterotyping ?, Because the interpretation of the findings is currently subjective , and in the absence of any community-wide best practices , we recommend using at least 2 or 3 distance metrics and clearly stating the criteria used for calling enterotypes within the context of any particular study ., Particularly , if different metrics yield different results , authors should attempt to understand the discrepancies and justify their choice of distance metric ., The effects of OTU taxonomic levels ( for instance , clustering sequences at genus or species level ) on the recovery of enterotypes are best illustrated with 16S rRNA gene sequence data from vaginal sites ( Figs . 4 and S3 ) ., Ravel et al . 26 reported enterotypes in the vagina based on the abundances of bacterial species ( as opposed to genera used in gut studies ) ., We used the abundances of both species and genus-level OTUs from the Ravel et al . dataset to test for enterotypes ., Our analysis shows strong support for two genus-level enterotypes using 4 of 5 distance metrics ( i . e . , unweighted UniFrac had moderate support ) for the Ravel et al . dataset when using the PS to test the strength of the clustering ( Fig . 4 ) ., We also observed strong support for genus-level mid-vaginal enterotypes using 3 of 5 distance metrics ( BC , JSD and rJSD ) for the HMP dataset ( Fig . 4 ) ., Additionally , using a species-level analysis , we obtained moderate support for five enterotypes using BC and JSD in the Ravel et al . data ( we also scored strong support for 2 enterotypes with weighted UniFrac ) , and moderate to strong support for 2 clusters ( i . e . , little or no support for five clusters ) in the HMP data ( Fig . 4 ) ., We also tested for clustering of vaginal samples using SI and CH ., When using SI ( Fig . S8 ) at the genus level we found strong support for 2 clusters in the HMP and Ravel et al . datasets using 3 and 1 distance metrics respectively ., But when using CH ( Fig . S9 ) on the HMP data at the genus level , the highest scores were obtained for 2–3 and 9–10 clusters , and in the Ravel data the strongest support was for 2 clusters ., At species level , we observed strong support with SI for 2 enterotypes in the HMP data using weighted UniFrac , and for 5 enterotypes using JSD ( Fig . S8 ) ., No strong support was observed for the Ravel data for any number of clusters at the species level using SI ., With CH , at the species level , the highest score was for 10 clusters in the HMP data , while for the Ravel data the highest score was for 2 clusters ., The differences in number of enterotypes found at the genus and species levels underscore the sensitivity of enterotyping to the taxonomic depths used in constructing OTUs ., To test for the influence of the specific variable region of the 16S rRNA gene on the detection of fecal enterotypes , we compared fecal samples from the HMP for which sequence data for both the V1–V3 and V3–V5 regions were available ., Data from the V3–V5 region yielded moderate support for two fecal enterotypes , but no enterotypes were detected using data from the V1–V3 region ., When using SI , we observed moderate support for 2 clusters when using JSD on the V1–V3 data and weak support for the V3–V5 data ., The highest scores using CH were three clusters using BC for V1–V3 data and two clusters using weighted UniFrac for V3–V5 data ( Fig . S10 ) ., Different primers amplifying different regions of the 16S rRNA gene sequence are known to impact the diversity described for a microbial community ., For example , primers for the V1–V3 region ( e . g . , 27F-338R ) are not efficient for amplifying 16S rRNA gene sequences from members of the Bifidobacteria genus , which can dominate the infant microbiota 22 , 27 ., Our analysis demonstrates that the specific region of the 16S rRNA gene that is amplified during PCR is another factor that can affect the outcome when searching for enterotypes ., We compared enterotype clustering using two methods for OTU picking: ( 1 ) de novo sequence clustering into OTUs , in which sequences are clustered based on similarity to one another , and ( 2 ) a reference based approach , in which sequences are clustered based on similarity to sequences in a reference database 27 ., We found that for the HMP dataset , the two OTU picking approaches yielded consistent results for the majority of body sites ( Fig . 5 ) ., However , for the attached keratinized gingiva , posterior fornix and tongue dorsum , the reference-based approach provided moderate support for enterotypes , whereas the de novo approach did not support clustering ., One important difference between the two OTU-picking approaches is that the reference-based method can yield fewer OTUs , particularly at fine taxonomic resolution , because any sequence that fails to find a match in the database is discarded ., In contrast , the de novo approach retains all sequences and has the potential to yield higher OTU counts ., Fewer OTUs would have the effect of increasing the relative abundances of the dominant genera , and may therefore strengthen the gradient effect frequently observed ( see below ) ., Thus , the reference-based OTU picking approach may result in over-confidence in enterotype discovery ., Rarefaction is the commonly used normalization practice of randomly subsampling the data so that an equal number of sequences are drawn for each sample ., We rarefied the sequences from the HMP fecal samples at 2 , 000 sequences per sample and compared the results to those obtained after rarefying at 1 , 000 sequences per sample ( Fig . S11 ) ., Rarefaction depth did not seem to strongly affect the results of the clustering ., We also implemented our methodology in the smaller set of HMP samples for which WGS data were available , in addition to the MetaHIT WGS data 3 ., While the HMP WGS data included fewer samples and body sites than the 16S rRNA gene sequence data ( approximately 700 spanning the gut , nares , three oral habitats , and posterior fornix ) , they provided consistent species-level resolution ., We found a strong gradient effect in the fecal samples ( see discussion on gradients below ) for almost all genera and species , and between species within the genus Bacteroides and members of the Firmicutes ., We also found that the presence of Prevotella ( specifically , P . copri ) was clearly associated with the first principal coordinate in the PCoA using three distance measures ., This feature in turn drove moderate support for two clusters in the HMP data ( using JSD and rJSD ) and strong support for 2 clusters in the MetaHIT data ( Figs . 6 and S12 ) , that appeared to separate roughly according to presence/absence of Prevotella ., Prevotella is similarly influential in driving variation along the first principal coordinate axis when using Jensen-Shannon divergence for the 16S-based samples , albeit not with weighted UniFrac ( Figs . S13-S14 ) ., Although the importance of Prevotella in clustering analysis clearly depends on the choice of distance metric , the genus does exhibit enterotype-like behavior in that it follows a bimodal distribution: high relative abundance in a small fraction of samples , but low or zero relative abundance in many other samples ., Note that the HMP 16S rRNA gene sequence surveys include a smaller fraction of samples containing high relative abundance of Prevotella compared to WGS data ( 12 . 3% and 10 . 9% of samples contained >10% Prevotella in the HMP V1–V3 and HMP V3–V5 data sets , respectively , compared to 13 . 9% and 24 . 2% in the HMP and MetaHIT shotgun metagenomics ) ., Although a bias of certain primers against Prevotella in 16S surveys has been reported previously 28 , this is not likely to have affected the HMP data ., The difference in Prevotella abundance between MetaHIT and HMP samples remains to be explained ., In all other body sites , we again found general agreement between metagenomics clustering results and the 16S rRNA gene sequence-based clustering results regarding cluster quality ( Figs . S15-S16 ) , with the exception of the moderate support for two enterotypes in the buccal mucosa WGS data ( Fig . S15 ) , and lack of consistent support for enterotypes in the posterior fornix ( Fig . S16 ) ., As we described above , we found moderate support for enterotypes in the posterior fornix in 16S data ( Fig . S3 ) ., The discrepancy might be due to the fact that the WGS data included too few samples from the smaller “clusters” to permit detection by the prediction strength approach; SI was highest for two enterotypes at the genus level corresponding to Lactobacillus ( either dominant or absent; JSD jackknifed SI: 0 . 89±0 . 012 ) , and statistically tied as highest for five ( four Lactobacillus species or Lactobacillus absent , JSD jackknifed SI: 0 . 79±0 . 005 ) and eight ( JSD jackknifed SI: 0 . 79±0 . 008 ) enterotypes at the species level ( Fig . S17 ) ., Our results underscore the importance of methodology in assessing whether populations can be categorized by enterotypes ., Table 1 summarizes the factors that might have an effect on enterotyping: the two with the largest effect are the distance metric and the clustering score method ., We recommend using at least one absolute scoring method ( i . e . , PS or SI ) combined with at least 2–3 distance metrics to verify the presence of enterotypes ., When using the different scoring methods , authors should indicate and justify the choice of thresholds for indicating levels of support for enterotypes ., Other factors that should be kept in mind are the data type ( and if using 16S rRNA gene sequence data , the variable region sequenced ) and the OTU picking method ., At the present time , there is no community consensus on how to define an enterotype , and two researchers with the same data can easily come to opposite conclusions regarding the presence of enterotypes if they apply different criteria ., Microbial ecologists and clinicians interested in the enterotype concept need to standardize enterotyping methods for the concept to gain utility ., The large size of the HMP dataset , augmented with the community and MetaHIT data , brought to light the extent of bacterial abundance gradients within body habitats ., The presence of these gradients underscores that discrete enterotypes ( i . e . , enterotypes with distinct boundaries ) are lacking ., Instead , for continuous OTU and genus gradients are the norm for most body sites , although a few body sites had multimodal distributions of samples with modes near the extremes of the gradients , and very few cases ( e . g . , the vagina ) had consistent discrete community types ., The biological drivers of these patterns , and their robustness over time , may be manifestations of host-microbial interactions , especially if they correlate with host factors such as diet , lifestyle , or genetics . | Introduction, Materials and Methods, Results and Discussion | Recent analyses of human-associated bacterial diversity have categorized individuals into ‘enterotypes’ or clusters based on the abundances of key bacterial genera in the gut microbiota ., There is a lack of consensus , however , on the analytical basis for enterotypes and on the interpretation of these results ., We tested how the following factors influenced the detection of enterotypes: clustering methodology , distance metrics , OTU-picking approaches , sequencing depth , data type ( whole genome shotgun ( WGS ) vs . 16S rRNA gene sequence data ) , and 16S rRNA region ., We included 16S rRNA gene sequences from the Human Microbiome Project ( HMP ) and from 16 additional studies and WGS sequences from the HMP and MetaHIT ., In most body sites , we observed smooth abundance gradients of key genera without discrete clustering of samples ., Some body habitats displayed bimodal ( e . g . , gut ) or multimodal ( e . g . , vagina ) distributions of sample abundances , but not all clustering methods and workflows accurately highlight such clusters ., Because identifying enterotypes in datasets depends not only on the structure of the data but is also sensitive to the methods applied to identifying clustering strength , we recommend that multiple approaches be used and compared when testing for enterotypes . | Recent work has suggested that individuals can be classified into ‘enterotypes’ based on the abundance of key bacterial taxa in gut microbial communities ., However , the generality of enterotypes across populations , and the existence of similar cluster types for other body sites , remains to be evaluated ., We combined the Human Microbiome Project 16S rRNA gene sequence data and metagenomes with similar published data to assess the existence of enterotypes across body sites ., We found that rather than forming enterotypes ( note we use this term for clusters in all body sites ) , most samples fell into gradients based on taxonomic abundances of bacteria such as Bacteroides , although in some body sites there is a bi/multi modal distribution of samples across gradients ., Furthermore , many of the methods used in the analysis ( e . g . , distance metrics and clustering approaches ) affected the likelihood of identifying enterotypes in particular body habitats ., We recommend that multiple approaches be used and compared when testing for enterotypes . | biology, computational biology, microbiology | null |
journal.ppat.1002902 | 2,012 | BZLF1 Governs CpG-Methylated Chromatin of Epstein-Barr Virus Reversing Epigenetic Repression | Activity and repression of eukaryotic genes correlate with the level of DNA methylation of promoter regions ., Prominent models are ß-globin genes ., Their sequential developmental activation and silencing in embryonic , fetal , and adult erythroid cells depends on the methylation status of DNA sequences near promoters of globin genes 1 , 2 and references therein ., It appeared that CpG methylation is a stable epigenetic mark transmitting the repressed state of chromatin through mitosis to daughter cells ., Little was known about dynamic demethylation ( and methylation ) events at promoters although demethylation is considered to be a prerequisite for gene activation at highly CpG-methylated promoter elements ., It is now clear that gene activation can involve the rapid gain or loss of 5′-methylcytosine ( 5mC ) residues in estrogen-responsive promoters 3 , 4 ., The methylation status of CpGs close to the transcription start site of the pS2 promoter gene changes upon estrogen induction within minutes indicating that methylation of DNA is dynamic but also involves processes of reactive demethylation 5 ., Erasure of DNA methylation and derepression of silenced chromatin has been observed in zygotes and primordial germ cells during fertilization and embryonic development , respectively ., Recently , the responsible enzyme ( s ) were identified as members of the Tet ( ten eleven translocation ) family of proteins capable of catalyzing the conversion of 5mC to 5′-formylcytosine followed by the excision by thymine-DNA glycosylase and base excision repair 6–10 ., Therefore , Tet proteins may drive the process of active CpG-demethylation , which is thought to be crucial to overcome transcriptionally repressed chromatin ., Epigenetic information like positioned nucleosomes or posttranslational modifications of N-terminal histone tails provides more flexibility to react to environmental cues ., In inducible promoters nucleosome positions change depending on the activation state of the gene 11 for a recent review ., N-terminal modifications of histone tails can be highly dynamic as a distinct epigenetic state can be enzymatically reverted by particular histone-modifying enzymes erasing the previous modifications 12 for a recent review ., Certain histone modifications are flexible in principle but can be stable and heritable through many cell generations ., For example , Polycomb group ( PcG ) proteins are regulators that repress genes by keeping a transcriptionally inactive state , which is mediated by H3K27 trimethylation ., The common view is that the Polycomb repressive complex 2 ( PRC2 ) acts as the “writer” of the repressed state ., It establishes H3K27 trimethylation with its histone methyltransferases EZH1 or EZH2 ., A second Polycomb repressive complex , PRC1 , is regarded as the “reader” of the epigenetic state ., It recognizes histone H3K27me3 and acts as a silencing complex by ubiquitinating histone H2A 13 or by chromatin compaction of defined nucleosome arrays 14 leading to stably repressed chromatin loci ., During specific stages of embryogenesis and stem-cell differentiation certain members of the trithorax group of proteins can remove the methyl groups at lysine 27 of histone 3 to re-install transcriptionally active genes 15 , 16 ., The balance between epigenetic stability and flexibility underlies the adaption of EBV to its human host ., In infected cells this herpesvirus can adopt two different states , which depend on the epigenetic regulation of its genes ., Upon infection of primary human B cells , the virus does not promote de novo virus synthesis but establishes a latent phase , characterized by the expression of a small set of viral genes , which promote cellular proliferation and contribute to viral oncogenesis 17 ., Extensive DNA methylation of viral DNA 18 , 19 is thought to contribute to overall gene silencing 20 whereas histone modifications such as H3K4me3 mark the few active promoters of latent genes together with the chromatin insulator protein CTCF and the viral factor EBNA1 21 , 22 ., EBVs EBNA genes can be alternatively expressed from the latent viral Cp- , Wp- or Qp-promoters ., Their states of DNA methylation were studied in vivo by several groups 23 , 24 and in vitro in cell lines of different origins 25 , 26 ., They found that the methylation status of these promoters determines the expression profile of latent viral genes ., In contrast to latent viral promoters , little is known about epigenetic marks at lytic promoters during latency when epigenetic silencing of most viral genes might guarantee the co-existence of EBV with its cellular host in the absence of viral de novo synthesis ., The eventual switch to EBVs lytic phase in latently infected cells is initiated by the expression of the viral BZLF1 gene encoding the transcription factor BZLF1 ( also called EB1 , ZEBRA , Z , or Zta ) 27 , 28 ., BZLF1 binds sequence-specifically to one class of DNA motifs , termed ZREs , but prefers a second class that contains methylated 5′-cytosine residues ( 5mC ) , termed meZREs 18 , 29–31 ., meZREs predominate in so-called early viral promoters 32 ., Paradoxically , CpG methylation of meZREs is instrumental for the expression of essential lytic genes 33 and indispensable for virus synthesis 18 , 32 , changing the conventional view of DNA methylation solely as a repressive epigenetic feature ., EBVs closest relative , Kaposis sarcoma-associated herpes virus ( KSHV ) , relies on bivalent chromatin in its latent phase to reactivate the lytic phase of its life cycle 34 , 35 ., In contrast to this option , we report here on a novel mechanism that deliberately relies on methylated EBV DNA to promote induced transcription of a distinct class of viral genes ., Our data suggests that , in combination with a very high concentration of highly methylated CpG dinucleotides , Polycomb-group ( PcG ) proteins introduce repressive modification of histones , which form densely arranged nucleosomes in order to shield the binding sites of BZLF1 in certain promoters of Epstein-Barr virus ( EBV ) to maintain latency ., Upon lytic reactivation BZLF1 is expressed and gains access to compacted , highly repressive chromatin , where it binds to its CpG-methylated DNA motifs ., BZLF1 binding induces nucleosomal eviction at BZLF1s cognate sites , erases repressive histone marks , recruits RNA polymerase II , and activates transcription of viral genes to trigger escape from latency ., Surprisingly , CpG methylation is invariably maintained during this early lytic phase of EBVs life cycle and no barrier to active transcription ., These attributes provide a novel paradigm for gene regulation in metazoan cells , which depend on stable chromatin repression to maintain their differentiation state and cellular identity but require the plasticity needed to respond immediately to developmental and environmental cues ., EBV exploits these mechanisms in order to sustain its lifestyle ., In the EBV genome a number of meZREs were identified , which presumably contain methylated CpGs 32 ., We examined selected regions of the latent EBV genome with deep bisulfite sequencing to assess the state of cytosine methylation in the cell line Raji , which is our prototype of a B cell latently infected with EBV ., After bisulfite modification genomic DNA from Raji cells was amplified by PCR with EBV-specific primer pairs ., The analysis encompassed 26 regions of EBV with latent , early lytic , and late lytic gene promoters covering 27 , 869 bp of EBVs genome ( Fig . S1A ) ., The PCR products were pooled , sequenced , and the percentage of 5mCs was assessed 36 ., The coverage of each CpG dinucleotide was 840 reads on average ., A list with all CpGs analyzed is available upon request , a graphical view of selected promoters is shown in Fig . 1D and Fig . S1B ., Cytosines within CpG dinucleotides were methylated to 66% on average in Raji EBV DNA ., We analyzed the distribution of EBVs DNA methylation in this latently infected B cell line ., The analysis revealed that the methylation state of EBV DNA is bimodal with 23% hypomethylated ( <20% methylation ) and 59% hypermethylated CpGs ( >80% methylation; Fig . 1A ) ., Interestingly , the distribution peaks at the very left ( <2 . 5% methylation ) and the very right ( >97 . 5% methylation ) as shown in Fig . 1A , indicating that many viral CpG dinucleotides are either completely methylated or unmethylated ., The CpG-rich BBLF4 promoter is a model of a BZLF1-regulated early lytic promoter 32 and its methylation state in Raji cells is shown in Fig . 1D ., As predicted 32 , the majority of CpG dinucleotides in this promoter carries 5mCs , which include the previously identified meZRE motifs ., Our laboratory strain of Raji cells contains 16 copies of EBV genomes per cellular genome equivalent as determined by quantitative realtime PCR ( Fig . 1B ) suggesting that cytosine methylation in the BBLF4 promoter is prevalent in all copies of EBV genomes present in Raji cells ., Our data so far did not exclude that unmethylated copies of EBV genomes might exist that could bind BZLF1 and e . g . exclusively support transcriptional reactivation of the BBLF4 promoter upon induction of EBVs lytic phase ( see below ) ., To address this point we arranged single reads of the deep sequencing analysis of the BBLF4 promoter according to their average degree of CpG methylation ( Fig . 1C ) ., We observed that, ( i ) the majority of all 31 CpG dinucleotides present in the BBLF4 promoter carries 5mC residues and not a single DNA molecule is entirely unmethylated ,, ( ii ) CpG methylation is variable at two hotspots while, ( iii ) unmethylated meZREs are infrequent and rarely cooccur on the same DNA molecule ., Our analysis excludes the existence of EBV genomes that escape CpG methylation in this model ( Fig . 1C ) but highlights the need to override CpG methylation and reverse epigenetic repression upon induction of EBVs lytic phase ., The BBLF4 promoter is no exception as other early lytic promoters showed an equally high and homogeneous methylation at meZREs with a median methylation rate of 95 . 3% ( Fig . S1B and Table S1 ) ., In contrast , active latent viral genes and their promoters like the Qp promoter ( Fig . 1D , lower panel ) were virtually free of methylated CpG dinucleotides in Raji cells resembling cellular CpG islands with a very high density of CpGs spared from DNA methylation ., Next , we wanted to assess the methylation profile of EBV DNA in cells , which are the latent reservoir of EBV in its human host ., We sorted memory B cells from PBMCs of healthy donors by multiparameter FACS enriching CD19 positive , CD38 negative , IgD negative , and CD27 positive cells ., Cellular DNA was extracted and parts of EBVs genome were analyzed by conventional bisulfite sequencing ., The frequency of EBV-infected memory B cells was reported to be very low with only one EBV-positive cell out of 104–106 cells 37 , but we succeeded in obtaining information of the BBLF4 and BZLF1 promoters from DNA samples of two healthy individuals ., The BBLF4 promoter was highly CpG-methylated in vivo ( Fig . 1E ) but the pattern was not entirely identical to that of Raji DNA ., The four CpGs , which constitute the meZREs of BBLF4 , were fully methylated in memory B cells in vivo ., In contrast , CpG dinucleotides in the BZLF1 promoter are rare and were unmethylated in Raji DNA and weakly or variably methylated in memory B cells ( Fig . 1E ) indicating that this locus is probably not controlled by CpG methylation in vivo or in vitro ., In addition to cytosine methylation , the density and the position of nucleosomes might also contribute to the transcriptionally repressed state of EBV DNA during latent infection ., We therefore investigated the nucleosomal occupancy of EBVs genomic DNA with a particular focus on BZLF1-regulated promoters in Raji cells ., This cell line also provides a model for viral reactivation because ectopic expression of BZLF1 efficiently induces EBVs lytic phase in Raji cells ( see below ) ., Lytic induction does not lead to de novo synthesis of virus progeny because the EBV genome in Raji cells has a deletion of the BALF2 gene abrogating viral DNA amplification in the lytic phase ., As a consequence , the late lytic phase is blocked , which is advantageous for the exact , unequivocal analysis of viral DNA , viral chromatin , and transcripts of early lytic genes shortly after induction ., We introduced a conditional allele of BZLF1 into Raji cells to analyze possible changes of epigenetic modifications of viral chromatin at the onset of EBVs lytic phase ., The coding sequences of BZLF1 and green fluorescent protein ( GFP ) were placed under the control of the bidirectional conditional promoter ( Fig . S2A ) ., Addition of doxycycline led to about 90% GFP-positive cells ( Fig . S2B ) , a rapid induction of BZLF1 ( Fig . S2C ) , and the expression of the BZLF1 target gene BRLF1 in more than 50% of GFP positive cells ( Fig . S2D ) ., In a first approach , chromatin of different Raji cell derivates was studied with an MND-on-Chip analysis ( mono-nucleosomal DNA-on-Chip ) ., MND-on-Chip experiments rely on the different accessibility of DNA to Micrococcal Nuclease ( MNase ) cleavage in the context of chromatin ., Digestion of chromatin with MNase leads to the degradation of free DNA , whereas nucleosomal DNA is protected and can be isolated and identified in e . g . microarray hybridization ., Mononucleosomal DNA and sonicated input DNA were labeled with fluorochromes and hybridized to a high-resolution custom-made EBV microarray ., The experiments were performed with three different cell lines ., Parental Raji cells were analyzed as a model of the latent state ., Chromatin of lytically induced and sorted Raji-BZLF1 cells addressed nucleosomal occupancy during EBVs lytic phase ., Raji-BZLF1ΔTAD cells constitutively expressed a truncated BZLF1 protein lacking its N-terminal transactivation domain 32 to study the effect of BZLF1s DNA binding domain ( DBD ) on nucleosomal occupancy at BZLF1-regulated promoters ., 32 ZREs were selected , which fulfilled certain criteria to ensure their being informative ( Table S2 ) ., The “enriched versus input” log2-ratios indicated the occupancy of DNA with nucleosomes ., The log2-ratios of the selected ZREs were averaged and the average nucleosome occupancy profiles of the three Raji derivates were overlaid in a window ±2000 bp , centered at the start of the ZRE ( Fig . 2A ) ., The average nucleosome occupancy profile displayed increased nucleosome occupancy at ZREs during latency ( parental Raji cells , black line ) , indicated by a high log2-ratio ., The elevated log2-ratio dropped after lytic induction ( Raji-BZLF1 cells , red line ) and after binding of truncated BZLF1 ( Raji-BZLF1ΔTAD cells , green line ) ., This observation indicated that binding of full length BZLF1 but also the C-terminal half of BZLF1 with its DBD caused a general loss of nucleosomes at ZREs ., The log2-ratios of the 32 individual ZREs in lytically induced Raji-BZLF1 cells and parental Raji cells were subtracted and ranked in a heat map shown in the left panel of Fig . 2B ., A cluster analysis was performed to probe for possible functional groups among all ZREs ., The resulting dendrogram shown on the left side of the heat map identified three functional groups of ZREs ., For the subsequent analysis the two most divergent groups , group 1 and group 2 were selected ( Fig . 2B ) ., The average nucleosome occupancy profile of group 1 ZREs in parental Raji cells during latency ( Fig . 2B , right panel , black line ) did not display elevated log2-ratios at the ZRE sites in contrast to group 2 ZREs ., Their average log2-ratios dropped after induction of the lytic phase ( red line ) , and constitutive binding of BZLF1 caused a similar reduction of average nucleosome occupancies at group 2 ZREs ( green line ) ., Next , we analyzed the differences between Raji-BZLF1ΔTAD cells that express the truncated BZLF1 protein constitutively and lytically induced Raji-BZLF1 cells ., This analysis addressed the role of BZLF1s N-terminal transactivation domain on nucleosomal displacement ., Subtraction of the log2-ratios of the eleven ZREs of group 2 in lytically induced Raji-BZLF1 cells and Raji-BZLF1ΔTAD cells are visualized in a second heat map ( Fig . 2C ) ., A cluster analysis and the resulting dendrogram distinguished two subgroups ., Subgroup 2a showed similar average nucleosome occupancies between the two datasets , indicating a similar displacement of nucleosomes , as compared to parental Raji cells ( Fig . 2C , upper right panel ) ., Average nucleosome occupancy profiles of subgroup 2b showed high nucleosome occupancy in latent parental Raji cells ( black line ) as expected ., Binding of truncated BZLF1 ( green line ) caused a small drop in the log2-ratios , suggesting lower nucleosome occupancy on average ., The induction of the lytic phase in Raji-BZLF1 cells resulted in a collapse of the log2-ratios ( red line ) , indicating the formation of hypersensitive sites ., Control experiments with 32 randomly chosen positions of the EBV genome lacking any ZRE did not reveal differences between the datasets ( Fig . S3 ) ., Our findings implied a functional classification of BZLF1-regulated genes ., Seven genes are essential for EBVs DNA replication in the lytic phase: BMLF1 , BRLF1 , BMRF1 , BBLF2/3 , BBLF4 , BALF2 , and BALF5 38 , 39 ( BALF2 is deleted in Raji EBV DNA and excluded from this analysis ) ., Five of the remaining six promoters of these genes make up subgroup 2b or belong to subgroup 2a , indicating that theses promoters are repressed very tightly during latency ., The promoters of BMLF1 , BRLF1 , and BMRF1 ( subgroup 2b ) probably require the local binding of wild-type BZLF1 for their chromatin remodeling upon lytic induction ., The promoters of BBLF2/3 and BBLF4 ( subgroup 2a ) also appeared tightly repressed during latency , but binding of truncated BZLF1 was sufficient to rearrange the promoter nucleosomes ., The seventh gene important for lytic replication , DNA polymerase BALF5 , contains a single ZRE in its promoter 32 , which did not fall into any analyzed group ., Promoters of group 1 are not regulated by BZLF1 at the level of nucleosomal rearrangements ., None of them belongs to the seven genes , which are required for EBVs lytic DNA replication ., Table 1 lists the ZREs hierarchically and according to their functional groups , Fig . S4 provides the graphical representation of examples of two members of each group ., MND-on-Chip experiments indicated a loss of nucleosomes at ZREs , which was validated in two additional , independent experiments with parental Raji cells ( latent ) and doxycycline induced Raji-BZLF1 cells ( lytic ) ., Chromatin immunoprecipitations with a histone H3-specific antibody resulted in the enrichment of nucleosomal DNA containing histone H3 at two different promoters that contained ZREs ( BRLF1 and BMRF1 ) in comparison to the ZRE-free W promoter ( Wp ) during latency in parental Raji cells ( Fig . 3A; 2–3% of input DNA ) ., The ZRE-containing promoters of BRLF1 and BMRF1 showed a clear reduction of the histone H3 signal ( 1% of input DNA ) in lytically induced Raji-BZLF1 cells but the ZRE-free Wp was not affected by lytic induction ( Fig . 3A ) ., These three promoters were re-analyzed in indirect endlabeling experiments ., Chromatin was treated with limited amounts of MNase and cleaved with appropriate restriction endonucleases at sites close to the regions of interest ., DNA was purified , separated on agarose gels , and transferred to nylon membranes by Southern blotting ., The membranes were hybridized to radioactively labeled probes that were complementary to regions downstream and close to the cleavage sites of the restriction endonucleases ., The bands that are visible in the autoradiograms show the boundaries of nucleosomes and hypersensitive sites ., The migration of the bands indicates the distance of the boundaries from the restriction endonuclease cleavage sites in the individual loci as shown graphically on the right side of the autoradiograms in Fig . 3B ., Top panels in Figure 3B show the results in parental Raji cells ., Raji chromatin was partially digested with increasing amounts of MNase as indicated and analyzed for nucleosome occupancies in the promoter regions of BRLF1 , BMRF1 , and Wp ., The restriction endonuclease cleavage sites , the probe location , and selected features of EBVs genome are depicted on the right side of the autoradiograms ., The lower panels in Figure 3B show the same promoter sites in Raji-BZLF1 cells , without or after addition of doxycycline and overnight induction of the lytic phase of EBV ., The uninduced cells displayed a pattern similar to the latent parental Raji cells ( Fig . 3B , upper panels ) as expected ., The induced expression of BZLF1 caused major rearrangements in the promoters of BRLF1 and BMRF1 , but not in the ZRE-free W promoter ., The BRLF1 promoter acquired a pattern that was typical for a hypersensitive site , which is situated in the same location as the BRLF1 ZRE sites , confirming the microarray and ChIP experiments ., The BMRF1 promoter showed overall reduced signal intensities in bands that were more than 1000 bp apart from the restriction endonuclease cleavage site after lytic induction , indicating a loss of nucleosomes in this BZLF1-regulated promoter as well ., We considered that the induced expression of BZLF1 and subsequent eviction of nucleosomes lead to an active demethylation of EBV DNA fostering chromatin reactivation and transcriptional activation ., The methylation state of four selected regions of the viral genome covering the latent Fp/Qp promoter and three lytic promoters representing an immediate early , an early , and a late gene were determined after induction of the lytic cycle ( Fig . 4 ) ., Raji cells with the conditional BZLF1 allele ( ‘Raji-BZLF1’ ) were induced with doxycycline for 15 hours and sorted for GFP-positive cells to obtain a pure population of lytically induced cells ., Cellular DNAs of lytically induced Raji-BZLF1 and parental Raji cells were isolated , bisulfite-treated , amplified with suitable primers spanning the Fp/Qp promoter , the BZLF1 promoter , the BBLF4 promoter , and the BDRF1 promoter , and directly sequenced ., There was no discernable difference between the DNA methylation pattern of parental Raji cells and lytically induced Raji-BZLF1 cells at any locus ( Fig . 4 ) indicating that active demethylation of the viral genome is not part of EBVs lytic phase ., We wanted to challenge the possibility that BZLF1 might induce transcription from partially unmethylated templates of EBV DNA ., BZLF1 could bind to unmethylated meZRE-sites , which , nevertheless , are infrequent in Raji DNA during latency ( see above , Fig . 1 ) and only weakly bound by BZLF1 32 ., Towards this end we employed a BZLF1-specific antibody and performed chromatin-immunoprecipitation experiments followed by direct bisulfite sequencing termed ChIP-BS-seq ( Fig . S5 ) ., If BZLF1 bound to non-methylated meZREs and/or preferentially supported transcription from partially methylated templates , we would expect an enrichment of DNA with a reduced frequency of methylated CpGs as compared to viral DNA present in latently infected Raji cells ., Our results demonstrated that BZLF1-bound DNA was indistinguishable from EBV DNA prevalent in Raji cells during latency ( Fig . S5 ) ., This finding further supported our working hypothesis that, ( i ) lytic reactivation relies on methylated meZREs and, ( ii ) BZLF1-controlled transcription originates from methylated DNA templates ., We analyzed promoters representing all different classes of EBV genes and included cellular control loci with chromatin immunoprecipitations ( ChIPs ) using antibodies against various histone modifications or certain chromatin-associated proteins ( Fig . 5 ) ., panH3-specific ChIP experiments confirmed the results of the microarray analysis: early lytic promoters and late lytic promoters were enriched in histone H3 during latency but induction of the lytic phase caused a selective loss of histone H3 at early lytic promoters , only ( Fig . 5A ) ., The BZLF1 promoter is an exception; lytic induction does not lead to a displacement of histone H3 , which is in accordance with our microarray analysis ( Table 1 ) ., We also determined whether lytic promoters carry specific chromatin marks during latency ., ChIP experiments with an anti-H3K9me3 antibody indicated that this repressive mark is present at some EBV-loci but that it is not important for the regulation of lytic promoters , as induction of the lytic phase did not alter their occupancy with this modification ( Fig . 5D ) ., In sharp contrast , ChIP experiments with an anti-H3K27me3 antibody revealed that all lytic promoters are characterized by this histone modification , which presumably is involved in their efficient repression during latency ( Fig . 5B ) ., Induction of the lytic phase selectively erased or reduced this modification at early lytic promoters , consistent with their reactivation ., High levels of H3K27me3 are a hallmark of Polycomb repression; therefore we assessed the occupancy of EBV promoters with the H3K27me3 methyltransferase EZH2 , which is a protein component of the Polycomb repressive complex 2 , PRC2 ., ChIPs with an EZH2-specific antibody perfectly matched the results of H3K27me3 suggesting that this methyltransferase is responsible for trimethylation of H3K27 in repressed EBV promoters ( Fig . 5E ) ., Interestingly , late lytic promoters , which do not support transcription in this model also showed a reduction in EZH2 levels , but to a lower extent ., The loss of EZH2 at EBV promoters did not result from a reduction of cellular EZH2 steady-state levels after the induction of EBVs lytic phase ( Fig . 5G ) but was a locus-specific phenomenon ., These results demonstrated that, ( i ) Polycomb repression is important to maintain EBVs lytic genes in a silent state during latency , and, ( ii ) induction of EBVs lytic cycle eliminates this mark relieving the tight repression ., We also wanted to assess the mode of activation of early lytic promoters on the chromatin level ., In Fig . 5C , the activation mark H3K4me3 appeared enriched at the transcriptionally active Q promoter ( Qp ) during latency , but H3K4me3 was undetectable at early lytic and late lytic promoters ., Induction of the lytic phase increased the levels of H3K4me3 at Qp as well as early lytic promoters ., Late lytic promoters also showed a minor enrichment of H3K4me3 marks upon lytic induction ., ChIP experiments with an anti-PolII antibody showed no significant levels in the latent state ( Fig . 5F ) but in lytically induced cells , PolII was recruited to the latent Q promoter as well as early lytic promoters ., In contrast , PolII was not detectable at late lytic promoters after lytic induction ., To address the kinetics of lytic gene activation quantitative RT-PCR analyses of Raji-BZLF1 cells documented the induction of PolII-mediated transcription of selected viral genes in a time course experiment ( Fig . 6A ) ., Raji-BZLF1 cells were induced with 100 ng/ml doxycycline for 46 h ., RNA was prepared every four hours ., Expression of a set of early and late lytic genes was tested together with the latent gene EBNA1 ., Absolute transcript levels were calculated on the basis of a single cell ., The kinetics of induction differed among the early lytic genes indicating that some genes are direct targets of BZLF1 , while other genes are probably induced by a combination of transcription factors or are secondary targets of BZLF1 ., The expression levels of the transgene BZLF1 peaked after four hours of doxycycline induction ., Expression of BMRF1 and BMLF1 was equally fast ., The BBLF4 , BBLF2 , and BALF5 genes were maximally expressed eight hours post induction ., BSLF1 and EBNA1 levels continuously increased for a time period of 28 h of induction ., Late lytic genes were not expressed or at very low levels , only ., We also followed the changes on EBVs chromatin over time to analyze the order of events contributing to the activation of early lytic promoters ., ChIP experiments were conducted similar to the experiments described above with two additional , early time points after doxycycline induction of BZLF1 ., Polycomb repression of EBV promoters was substantially reduced as early as three hours post induction with signals decreasing further ( Fig . 6B and D ) ., Surprisingly , H3K27me3 initially also decreased at late lytic promoters three hours after BZLF1 induction but quickly recovered thereafter ., Compared to the rapid loss of repressive chromatin marks , activation marks and recruitment of PolII were slow processes ., A significant increase in H3K4me3 signals became apparent 15 hours post induction similar to PolII ( Fig . 6C and E ) ., Only the BMRF1 and the Q promoter showed a significant enrichment for PolII 7 hours after adding doxycycline followed by H3K4me3 modifications indicating that these promoters respond very early to lytic induction ., As a consequence , BMRF1 together with BZLF1 and BMLF1 was one of the first genes reaching high steady-state levels of transcripts ( Fig . 6A ) ., Our time course studies indicated a sequential order of events at viral chromatin: Polycomb repression is relieved before PolII is recruited to the promoter sites inducing viral gene expression ., Figure 7 summarizes all epigenetic and functional data obtained in this study with the BBLF4 promoter as a representative example ., The promoter comprises several highly methylated meZREs 32 ., During latency its nucleosomal occupancy was high and the complete region displayed high levels of H3K27me3 and EZH2 ., Activation marks and PolII were not detectable during latency ., The situation changed dramatically upon lytic induction as the promoter was completely remodeled: nucleosomes and repressive modifications were evicted and erased , respectively ., Instead , the H3K4me3 levels and PolII occupancy rose , inducing the transcription of BBLF4 ., It is worth noting that all these dramatic changes occur at chromatin encompassing densely arranged CpGs with highly methylated cytosine residues , which remain unaltered throughout the early lytic phase of EBVs life cycle ., Herpes viruses establish a life-long infection in their hosts ., Success of infection relies on two principles:, ( i ) in latently infected cells the promoters of lytic genes must be tightly repressed to evade immune recognition of their products by the infected host and, ( ii ) induction of the lytic phase has to proceed rapidly and synchronously to escape from latency and virus-specific effector T cells ., We found that repression and dynamic reactivation of viral genes are both governed by epigenetic mechanisms ., Repressed lytic promoters of EBV are associated with extensive DNA methylation , high nucleosome occupancy , and H3K27me3 histone marks , which are repressive modifications transmitted by Polycomb group ( PcG ) proteins leading to compaction of chromatin ., Repressive H3K9me3 histone marks were detectable at low levels as reported recently 40 , but very much in contrast to H3K27me3 , H3K9me3 modifications are not removed upon viral reactivation suggesting that the low level of these modifications are not central to maintaining a repressed chromatin in EBV ., The activation mark H3K4me3 and PolII were undetectable at lytic promoters during latency , indicating that EBVs latent DNA is not associated with bivalent chromatin in contrast to latently KSHV-infected cells ., DNA methylation is a prerequisite for the activation of EBVs early lytic promoters that rely on meZREs 18 , 32 ., Upon lytic reactivation , CpG-methylation of viral DNA was unaltered but repressive histone marks were erased and nucleosomes were evicted in a subset of ZREs , which led to a local opening of promoters and loading of the transcription machinery ., | Introduction, Results, Discussion, Materials and Methods | Epigenetic mechanisms are essential for the regulation of all genes in mammalian cells but transcriptional repression including DNA methylation are also major epigenetic mechanisms of defense inactivating potentially harmful pathogens ., Epstein-Barr Virus ( EBV ) , however , has evolved to take advantage of CpG methylated DNA to regulate its own biphasic life cycle ., We show here that latent EBV DNA has an extreme composition of methylated CpG dinucleotides with a bimodal distribution of unmethylated or fully methylated DNA at active latent genes or completely repressed lytic promoters , respectively ., We find this scenario confirmed in primary EBV-infected memory B cells in vivo ., Extensive CpG methylation of EBVs DNA argues for a very restricted gene expression during latency ., Above-average nucleosomal occupancy , repressive histone marks , and Polycomb-mediated epigenetic silencing further shield early lytic promoters from activation during latency ., The very tight repression of viral lytic genes must be overcome when latent EBV enters its lytic phase and supports de novo virus synthesis in infected cells ., The EBV-encoded and AP-1 related transcription factor BZLF1 overturns latency and initiates virus synthesis in latently infected cells ., Paradoxically , BZLF1 preferentially binds to CpG-methylated motifs in key viral promoters for their activation ., Upon BZLF1 binding , we find nucleosomes removed , Polycomb repression lost , and RNA polymerase II recruited to the activated early promoters promoting efficient lytic viral gene expression ., Surprisingly , DNA methylation is maintained throughout this phase of viral reactivation and is no hindrance to active transcription of extensively CpG methylated viral genes as thought previously ., Thus , we identify BZLF1 as a pioneer factor that reverses epigenetic silencing of viral DNA to allow escape from latency and report on a new paradigm of gene regulation . | Latency is a fundamental molecular mechanism that is observed in many viruses ., We reveal that the human herpes virus Epstein-Barr virus ( EBV ) uses cellular functions of epigenetic repression to establish latency in infected B cells and a previously unknown mechanism to escape from it ., We show that the herpesviral DNA genome is transcriptionally silenced by cellular mechanisms during viral latency , which includes excessive methylation of EBV DNA in vitro and in its human host in vivo ., Epigenetic modifications like high nucleosome density and repressive histone marks shield and inactivate lytic viral genes during latency ., EBVs genuinely repressed chromatin poses the problem of efficient reactivation to support virus synthesis ., BZLF1 is the viral switch gene that induces the lytic phase of EBVs life cycle ., We show here that this viral transcription factor erases static , repressive chromatin marks reversing epigenetic silencing ., DNA methylation is preserved but no hindrance to lytic gene activation because BZLF1 directly binds to methylated viral DNA and overcomes heavily repressed chromatin without the need for active DNA demethylation ., DNA demethylation has been thought to be a prerequisite for gene transcription but this virus falsifies this hypothesis and provides a new model for epigenetic gene regulation . | cellular stress responses, dna transcription, histone modification, epigenetics, dna, dna structure, chromatin, gene expression, biology, dna modification, molecular biology, cell biology, nucleic acids, genetics, molecular cell biology, genetics and genomics | null |
journal.pbio.1000494 | 2,010 | Chromosomal Redistribution of Male-Biased Genes in Mammalian Evolution with Two Bursts of Gene Gain on the X Chromosome | In mammals and Drosophila , the X chromosome usually differs dramatically from autosomes since it is hemizygous in males 1 ., Sexual antagonism ( beneficial for one sex , but deleterious for the other ) enriches male-biased genes on the X chromosome , if alleles are generally recessive , and on the autosome if they are generally dominant 2–3 ., On the other hand , inactivation of the X chromosome during spermatogenesis 4–5 drives the accumulation of male-biased genes on the autosomes where they can be expressed in the meiotic or post-meiotic phase 6–7 ., These two processes can explain the gene traffic between the X and autosomes in Drosophila 8 and mammals 9–10 as well as the excess of male-biased genes on the autosomes 11–12 ., However , recent analyses of male-biased genes identified several X-linked genes that originated in the last 1–3 million years ( myr ) in Drosophila 13–15 ., Whether or not these data implicate an effect of evolutionary time on the chromosomal location of male-biased genes remains unknown ., In our investigation of how the various evolutionary forces impact the chromosomal distribution of sex-biased genes , we focused particularly on how the age of genes affects their chromosomal locations ., By dating when genes arose in humans and mouse , we found male-biased genes were distributed at different locations in different phases of mammalian evolution: young male-biased genes are enriched in the X chromosome , but older male-biased genes favor autosomal locations ., Interestingly , this redistribution seems to be associated with feminization of the X chromosome with more X-linked old genes expressed in ovaries ., Besides the recent gene gain contributed by emergence of male-biased genes on the X chromosome , we found another burst of gene gain on X chromosome immediately after the divergence of opossum and eutherian mammals ., Accelerated protein evolution and transcriptional evolution of X-linked genes reveal positive selection occurring in this period ., These data support the recent notion 10 , 16 that our X chromosome originated in the therian ancestor instead of the common ancestor of all mammals ., These two lines of findings significantly extend our knowledge of the origination and evolution of X chromosomes in mammals ., We tracked the relative gene abundance of individual chromosomes across 450 myr and identified two bursts of genes occurring on the X chromosome ( Figure 2 ) ., One burst ( branches 5–7 ) postdated the divergence of eutherian mammals ( human or mouse ) and marsupials ( opossum ) and the other occurred recently after the split of human and chimp and after the split of mouse and rat , respectively ., For both peaks , the X chromosome contributes to 8%∼14% of genes , while it only accounts for 3% of genes in the first 300 myr of vertebrate evolution ., In contrast , autosomes tend to vary less in their relative contribution to the whole genome ( Figure S1 ) ., As the major contributor generating new genes , DNA-level duplication accounts for 73∼95% of genes of these two peaks ., If we only use DNA-level duplicates , the pattern remains the same ., Considering that many more genes arose in branch 5 compared to branch 6 or 7 ( 1 , 200∼1 , 400 versus 400∼500 , Figure 1 ) , the old peak seems to be best explained by the hypothesis that the X chromosome emerged in the therian ancestor and subsequently recruited many genes in an accelerated evolution of sex-related functions , as found with retrogene-based chromosomal movement studies 26 ., In contrast , the recent burst reveals a rapid addition of new genes into the mammalian X chromosome , which may be independent of major chromosomal changes ., Based on human body index data ( GSE7307 , Materials and Methods ) and mouse tissue profiling data 27 at the NCBI GEO database 28 , we identified genes with sex-biased expression ( Materials and Methods ) ., As shown in Figure 3 , both human and mouse demonstrate a similar pattern regarding the proportion of male-biased genes and the age of the branch in which they arose ., For younger branches ( less than 50 myr ) , male-biased genes are enriched in the X chromosome compared to autosomes ( ∼50% versus ∼30% , Chi-square test p<0 . 05 ) , which might be driven by fixation of recessive male-beneficial alleles under sexual antagonism ., This pattern decreases for genes originating in earlier branches ., Male-biased genes older than 300 myr are overrepresented on the autosomes ( ∼30% versus ∼15% , p\u200a=\u200a1×10−9 ) ., This pattern was independently supported by an Affymetrix exon array panel with larger coverage of new genes ( Figure S2 ) ., Thus , the recent peak observed in Figure 2 could be attributed to a burst of male-biased genes on X chromosome younger than 50 myr ., Figure 3 also demonstrates that the X chromosome consists of a similar or even higher proportion of male-biased genes compared to autosomes from 90 myr ago ( branch 7 ) to 130 myr ago ( branch 5 ) ., Thus , many of the genes gained in the first , older peak may also have male-biased expression ., Notably , the proportion of female-biased genes on branch 5 was greater on the X chromosome compared to autosome ( 39% versus 20% in Table 1 ) ., In contrast , for branches 6 and 7 , the proportion of female-biased genes is around 15% for both the X chromosome and autosomes ( Table S4 ) ., Again , this suggests that the newly originated X chromosome was subjected to enhanced positive selection and recruited an excess of both male- and female-biased genes ., The earlier peak in Figure 2 indicates the mammalian X chromosome emerged before the divergence of eutherian and marsupial 10 ., Thus , the nascent X chromosome changed remarkably , gaining an excessive number of genes ., If this scenario is true , those preexisting genes on the ancestral X chromosome might have accumulated many evolutionary changes during this period ( branch 5 ) , as did genes linked to the neo-X chromosome in Drosophila 29 ., That means we would expect these ancient genes on the X chromosome to show signatures of positive selection ., To test this scenario , we investigated the evolutionary path of ancient genes shared by vertebrates by comparing the ratio between non-synonymous substitution rate and synonymous substitution rate ( Ka/Ks ) ( Materials and Methods ) ., In other words , we compared the Ka/Ks of X- and autosomal-linked old genes in separate evolutionary periods ., Across evolution of 450 myr , the X chromosome did not show significantly higher Ka/Ks except in branch 5 ( Table 2 ) , which strongly corroborates the hypothesis that the X chromosome did not acquire sex-chromosome status until this period ., We extended this analysis to genes gained since branch 5 ., We directly estimated the proportion of replacement substitutions ( α ) based on polymorphism and divergence data in 30 and a maximum-likelihood method implemented in the DoEF package 31 ., As shown in Table S5 , young genes generally show higher α compared to old genes , and X-linked male-biased genes show the highest α , 0 . 501 ., This pattern shows that positive selection instead of neutrality drives the evolution of X-linked genes arising since branch 5 , especially those with male-biased expression ., However , positive selection of nucleotide substitutions can only suggest that initial fixation may also be driven by positive selection ., More direct evidence comes from copy number polymorphism ( CNP ) data in Drosophila , which showed that the X chromosome is subject to stronger purifying selection than autosomes 32 ., In human , it was also noted that the X chromosome shows a paucity of CNPs 33 ., Together with bursts of adaptive fixations occurred on the neo-X of Drosophila 29 , it is likely that positive selection instead of drift accounts for two bursts of genes on the X chromosome ., As we noted before , enrichment of young male-biased genes on the X declines for those originating in earlier evolutionary branches ., Using expression data from mouse spermatogenesis , we compared different age groups to investigate which force underlies such a demasculinization process ( Table 3 ) ., As previous studies such as 7 found , old genes are expressed more in the pre-meiosis stage ( spermatogonia ) but are silent from meiosis ( pachytene spermatocyte ) to post-meiosis ( round spermatid ) ., In terms of whole testes , however , old X-linked genes are underrepresented ( Table 3 ) ., New genes show a distinct pattern: while often expressed in spermatogonia , they are not silent in meiosis ., Moreover , a much greater proportion of new genes on the X are expressed in the post-meiosis stage compared to genes on the autosome ( 70% versus 27% , Chi-square test p\u200a=\u200a5×10−10 ) ., This is consistent with a previous observation of X-linked postmeiotic multicopy genes 34 , the vast majority of which we found were very young ( Materials and Methods ) ., Such a pattern suggests that the young X-linked genes are not affected by MSCI ., An independent microarray dataset of mouse spermatogenesis 35 confirms high expression of X-linked young genes in spermatid ( Figure S3 ) ., In addition , we note that the customized array by Khil et al . was comprised mainly of old and conserved genes , with only 1 . 7% of the set being young genes ., In contrast , the Affymetrix array data 36 we used covered 14 , 923 Ensembl genes , 3 . 9% of which are young genes ., This striking contrast between young and old genes suggests that MSCI plays an important role in determining the age-dependent chromosomal distribution of male-biased genes ., In order to investigate how this contrast occurred in such a short time , we analyzed four major cell types including sertoli cells , spermatogonium , spermatocyte , and spermatid between mouse 35 and rat 37 ., We used the Euclidean distance of relative abundance ( RA ) to measure how orthologous genes have diverged in their expression ( Materials and Methods ) ., Consistent with a previous comparison of human and chimpanzee 38 , the testis expression of genes on the X chromosome diverge more between rat and mouse than genes on autosomes ( Wilcoxon rank sum test p\u200a=\u200a4×10−6 , Figure 4 ) ., Furthermore , X-linked young genes show significantly higher divergences , compared to all other three groups ( p<0 . 05 ) ., While we found that expression in various spermatogenesis stages is generally conserved 35 with only about 3% divergence ( Figure S4 ) , X-linked young genes show the largest expression divergence in spermatid ., Specifically , after the split of mouse and rat 37 myr ago 39 , young X-linked genes show 6 . 9% divergence in spermatid , which is much higher than the genomic average for spermatid , 3 . 3% ( Wilcoxon rank sum test p\u200a=\u200a0 . 002 ) ., This increased divergence suggests that , although these genes seem to escape MSCI and preferentially transcribe in post-meiosis , the expression profile is not conserved ., It remains unknown whether these genes get up-regulated or down-regulated in one species ., But if the latter case were true , it indicates that the high post-meiotic expression would be silenced by MSCI in later evolution ., This could also explain how the different pattern between young and old genes in Table 3 is achieved ., We investigated the distribution of female-biased genes on chromosomes and its correlation with gene ages ., Interestingly , female-biased genes are distributed in a pattern symmetrical to male-biased genes ( Figure S5 versus Figure 3 ) : the old X-linked genes are more often female-biased , while young genes are not ., We characterized ovary expression of genes using the Affymetrix mouse exon array panel data ., Consistent with Figure S5 , ovary expression also depends on the age of the genes origination ., Specifically , young autosomal genes show significantly higher expression in ovaries than young X-linked genes ( Wilcoxon rank sum test p\u200a=\u200a5×10−12 , Figure 5 ) ., However , old X-linked genes generally show higher expression in ovaries ( p\u200a=\u200a5×10−7 ) ., Thus , as gene age increases , this expressional excess of autosomal genes reverses and older X-linked genes show significantly higher expression in ovaries ., It can be argued that such an age-dependent pattern of expression is not a specific property of ovary evolution and other organs might also show a similar pattern ., To test this possibility , we investigated gene expression in the major organs: brain , heart , kidney , liver , lung , muscle , spleen , and thymus ., All these tissues , except for brain , showed a significant excess of expression for new genes ( branch≥5 ) on autosomes compared to that of X-linked genes ( Wilcoxon rank sum test p<0 . 01 , Figure S5 ) ., However , for old genes ( branch≤4 ) , they are evenly distributed ( p>0 . 05 ) ., The brain shows a unique pattern ., Young genes ( branch>7 ) are relatively abundant on autosomes ( p\u200a=\u200a0 . 001 , Figure S5 ) , but old genes ( branch≤7 ) are overrepresented on the X chromosome ( p≤0 . 01 ) ., This is consistent with previous findings that X chromosome is enriched with genes expressed in brain 1 , 40 ., Notably , different from ovaries , enrichment in the brain did not show clear age dependence , since genes originating from branches 5 to 7 presented the most significant excess ( Figure S6 ) ., The coincidence that the X chromosome is enriched with both ovary-expressed and brain-expressed genes occurring in branch 5 ( Table 1; Figure S5 ) motivated us to perform more thorough transcriptional profiling to get a more complete picture of how genes from this evolutionary period are transcribed ., We investigated mouse exon atlas data ( GSE15998 ) to ask whether X-linked genes are more frequently expressed in the tissue of interest across different age groups ., We clustered tissues by the proportion of X-linked genes expressed versus the proportion of autosomal genes expressed and identified three major groups: nervous system , testes , and all other tissues ( Figure 6 ) ., Remarkably , the X-linked genes originating in branch 5 are transcriptionally permissive with a larger proportion of them expressed in many tissues compared to autosomal genes ., This excess is most pronounced for brain samples ., Consistently , human data revealed that a greater proportion of X-linked genes emerging on branch 5 are expressed more widely than autosomal genes originating in this period , which is strongest for the brain ( Figure S7 ) ., Since human and mouse share a similar pattern , parsimony suggests this striking transcriptional pattern of branch 5 derived genes is ancestral ., Notably , none of these genes show sex bias in human brain profiling data 41 , which suggests they might be important for both sexes ., We have described evolutionary patterns of protein-coding genes , which could be driven by natural selection in various forms like sexual antagonism or MSCI ., If , however , such a pattern is a product of some mutational bias of gene origination , we would not detect similar evolutionary patterns in non-coding RNA genes , such as X-linked miRNAs ., Therefore , we investigated the chromosomal distribution of miRNA genes annotated in miRBase 42 and found that miRNA duplicates are distributed in a pattern similar to that observed for protein-coding genes ( Table S6 ) ., Specifically , both human and mouse show significant miRNA gene gain in branches 5 to 7 compared to the proportion of all miRNA genes ( 18∼22% versus 10∼13% , Fishers Exact Test p<0 . 05 ) ., Moreover , they also show an excess for the youngest branch ., Although it is not significant for the human data due to small sample size , it is for mouse ( p\u200a=\u200a0 . 02 ) ., Like protein-coding genes , a larger proportion of X-linked miRNAs originating in branch 5 are transcribed in nine tissues ( statistically significant for six of them ) surveyed on Agilent chip 43 compared to autosomal genes ( Table S7; Materials and Methods ) ., Moreover , semi-quantitative PCR data of X-linked miRNAs in 12 tissues 44 show 9 out of 13 ( 69% ) young genes are expressed higher in testes than at least six non-testis tissues ., However , this percentage drops to 23% for old X-linked genes ( 9 out of 39 , Fishers Exact Test p\u200a=\u200a0 . 005 ) ., Consistent with protein-coding genes , these data also show that old genes have moderate or high expression in ovaries and the young genes show only trace levels of expression ( Wilcoxon rank sum test p\u200a=\u200a0 . 01 ) ., The age-dependent locations and expression profiles of miRNAs support that it is evolutionary forces , rather than some mutation bias intrinsic to a certain type of gene , which account for the dynamics of X-linked gene evolution ., It is known that the X chromosome can be divided into five evolutionary strata because of step-wise repression of recombination 45–47 ., The X-conserved region ( XCR ) consists of the oldest strata 1 and 2 , while the X-added region ( XAR ) includes younger strata 3 that is shared by primates and rodents , and much younger 4 and 5 that were derived within primates 46 , 48 ., Since sexual antagonism or other sex related forces like the faster-X process ( see Discussion ) depends on hemizygosity of the X chromosome in male , we expect the accordance between bursts of gene gain with the formation of corresponding strata ., If these forces shape the evolution of gene content on the X chromosome , we should find that X-linked genes originating at a given time period should accumulate only in the strata already formed at that time ., In other words , we should find a correlation between the ages of genes and the strata in which they are located ., Consistent with these predictions , Figure 7 shows that the older strata 1 to 3 are associated with relatively older genes , while strata 4 or 5 are enriched with younger genes ( one sided Fishers Exact Test p\u200a=\u200a0 . 03 ) ., This finding parallels the temporal correspondence between the occurrence of strata and the out-of-X retrogene traffic 49 ., Our analyses demonstrated that the X chromosome evolved dramatically on both the sequence and expression levels after the split of eutherian mammal and marsupials ., Specifically , the X chromosome showed a burst of gene gain during this time , and many of these genes quickly invaded the transcriptional network of various tissues , especially the brain ., Furthermore , genes predating the birth of the X showed rapid protein-level evolution ., A straightforward interpretation is that the newborn mammalian X was subjected to strong positive selection similar to the neo-X chromosome in Drosophila 29 ., Moreover , the X-linked genes arising in branch 5 seem to have played important roles , as shown by their broad expression ., Their transcription pattern suggests that the early evolution of placental mammals was associated with rapid changes in the brain ., Furthermore , analysis of gene ontology showed that many of these genes mainly played regulatory roles in transcription and metabolism ( Table S8 ) ., Thus , regulatory change contributed by gene gain on the X chromosome was extensively involved in the initial evolution of eutherian mammals ., The fact that this peak ranges between branches 5 and 7 suggests remodeling of incipient X chromosome might take about 90 myr ( −160∼−70 myr , Figure 2 ) , which is consistent with one report based on retrogene movement 26 ., However , the selective pressures driving this dramatic change in branch 5 appear to be smaller in subsequent branches ( Table 2 ) ., Our analyses reveal chromosomal redistribution of X-linked male-biased genes ., Sexual antagonism may contribute to the initial fixation of X-linked recessive alleles as described previously 2 , 7 ., The faster-X hypothesis was initially proposed to fix more mutations on the X chromosome only if they are recessive and beneficial 1 ., Recently , it was observed that this force was most pronounced for male-biased genes 50 ., This suggests that the faster-X process could also be involved in the emergence of young X-linked male-biased genes , as the hypothesized sexual antagonism might ., These young X-linked male-biased genes could be later silenced by MSCI as suggested by Table 3 , Figure 4 , and Figure S4 ., At least two processes could be involved in this switch ., First , we found a statistically significant excess of male-biased retrogenes generated in the X→A movement process and X-enrichment of the female-biased parental genes for both human and mouse ( Table S9 ) ., Thus , the demasculinization and feminization of the X chromosome could be coupled in retrogene traffic ., Moreover , our RA analysis ( Figure S6 ) extends the out-of-the-testes hypothesis 51 to non-retroposed new genes ., We found that new genes generally acquire transcription in more tissues during evolution although they are initially enriched in testes ., With increasing MSCI and expanding expression breadth , X-linked male-biased genes might become unbiased or even female-biased as Figure S6 shows ., If new strata on the X chromosome represent regions that did not develop recombination repression until recently , the genes encoded in these regions will often escape MSCI 45 ., Thus , it is expected that the X-linked male-biased genes more likely escape MSCI when located on young strata or pseudoautosomal regions ( PARs ) ., However , out of 13 young male-biased genes in humans , the relatively young strata 4 and 5 encode only one ( Table S10 ) , which does not significantly differ much from the expected number based on its genomic size ., How then did the remaining 12 genes , those situated on older strata , escape from MSCI ?, It was proposed that the excess of inverted repeats ( IRs ) encoded by human and mouse X chromosome could protect genes contained by these IRs from MSCI 52 ., IRs suppress MSCI through formation of cruciforms or other unusual chromatin structures ., Moreover , cancer/testis ( CT ) genes that are often expressed in normal testes and in cancerous tissues frequently overlap with IRs 52 ., Given that X-linked CT genes underwent recent expansion 53 , it is not surprising that some of them could form highly homologous IRs ., In fact , 8 out of 13 young X-linked male-biased genes are CTs ( Table S10 ) ., Thus , the high IR abundance on the mammalian X chromosome might be one reason that these genes can be transcribed in meiosis or postmeiosis ., Furthermore , out of 12 genes encoded by PARs and covered by unique probes ( Table S11 ) , there is only one ( 8% ) male-biased gene , PPP2R3B , which is shared by human and mouse ., Thus , different from our intuition , PARs do not harbor an excess of male-biased genes compared to the remaining strata ( 18% ) and to autosomes ( 24% ) ., Albeit of small sample size , this observation suggests that sex-related forces like sexual antagonism or faster-X process account for the observed excess of young X-linked male-biased genes ., There are only limited number of genes with unique probes on strata 4 ( five ) and 5 ( eight ) ., For the remaining strata , stratum 3 is enriched with male-biased genes , which is much higher than stratum 1 ( 27% versus 17% , one-sided Fishers Exact Test p\u200a=\u200a0 . 02 ) and stratum 2 ( 27% versus 15% , p\u200a=\u200a0 . 03 ) ., This pattern suggests that stratum 3 recruits more young male-biased genes and there was not enough evolutionary time to be feminized as occurred in the oldest strata 1 and 2 ., As shown in Figure 2 , the emergence of young male-biased genes peaks in recent evolution of human and mouse ., However , this peak started 30 myr ago ( before the divergence of mouse and rat ) in the rodent lineage , while the peak appeared in the last 5 myr in human lineage ., This difference is consistent with the fact that the mouse X encodes more young male-biased genes than the human X . Specifically , male-biased genes account for 52% and 74% of X-linked young genes in human and mouse , respectively ( Figure 3; one sided Fishers Exact Test , p\u200a=\u200a0 . 07 ) ., Exon array data are similar ( Figure S2; 45% versus 76% , one sided Fishers Exact Test , p\u200a=\u200a2×10−8 ) ., Origination of significantly more male-biased young genes suggests that stronger positive selection acts on rodents and could explain why the recent peak of gene gain ( Figure 2 ) began earlier in the mouse lineage than in the human ., We developed a genome-alignment based pipeline to infer the origination time of a given genomic region by modifying a previous gene-alignment based method 58 ., We analyzed UCSC 17 netted chained file for human ( hg18 ) and mouse ( mm9 ) to verify whether a given human/mouse locus has a reciprocal syntenic alignment in the outgroup genome such as chimpanzee , rat , chicken , and so on ., In other words , we investigated whether a best-to-best match could be found between human/mouse loci and outgroup loci regardless of chromosomal linkage ., In this way , we can identify orthologous genes; even those with different chromosomal location due to fusions or translocations such as those found in XAR region will be identified as well ., Then , in order to handle occasional sequencing gaps , we scanned multiple outgroups and assigned this locus to a specific branch by following a parsimony rule ., Compared to the previous method 58 , our strategy is independent of gene annotation of outgroups and robust with gene translocation ., Thus , we generated a more stringent young gene dataset ( as described in the Result section ) ., And , as Figure S8 shows , we have not assigned most genes encoded by XAR as young genes simply because this region changed the linkage by fusing to X chromosome ., Conversely , several genes originated in branch 5 are located in strata 1 and 2 that are not XAR ( Figure S8 ) , also supporting that our pipeline is robust with gene translocations ., Notably , for regions without reliable synteny , our method might not work ., This situation would be most pronounced for telomeres , which tend to be repetitive and prone to recombine 59 and thus have very limited synteny ., For example , we dated 17 genes situated on PARs of the X chromosome ( Table S11 ) ., For three genes encoded by PAR2 , repeats contribute less than 16% of the gene loci based on UCSC annotation 17 ., Accordingly , our age assignments for these three genes are always consistent with those inferred by tree reconstruction provided by Ensembl 54 ., In contrast , for 14 genes linked with PAR1 , repeats are prevalent with a median contribution of 55% to the gene loci ., In this case , our results are consistent for only three out of nine cases with Ensembl age information ., We slightly modified the previous pipeline 58 , 60–61 and classified young genes as DNA-level duplicates , RNA-level duplicates ( retrogenes ) , and de novo genes ., Briefly , we performed all-against-all BLASTP search for human and mouse proteins ., It was reported previously that retrogenes can recruit other neighboring genome regions with introns after being retroposed 51 ., Thus , in order to define a new gene as retrogene , we requested that in the aligned region between the most similar paralog ( candidate parental gene ) and child genes , the former contain at least one intron and the latter to be intronless ., Otherwise , it will be classified as DNA-level duplicates ., Notably , if there is no hit with BLAST evalue cutoff 10−6 found 58 and no annotated paralog by Ensembl 54 , the gene will be defined as de novo ., In order to avoid non-specific probes and to cover more recently annotated genes , we used the customized array annotation files ( released on November , 2008 ) downloaded from University of Michigan 62 , HGU133Plus2_Hs_ENSG ( Affymetrix Human 133 plus 2 ) and Mouse4302_Mm_ENSG ( Affymetrix Mouse Genome 430 2 . 0 Array ) for human and mouse , respectively ., For exon array analysis , we used HuEx-1_0-st-v2 , U-Ensembl49 , G-Affy . cdf and MoEx-1_0-st-v1 , U-Ensembl50 , G-Affy , EP . cdf generated by Aroma . affymetrix team 63 ., Thus , we excluded some candidate young genes that were too similar to their paralogs and did not have specific probes ., Based on R 57 and Bioconductor platform 64 , we used RMA 65 to normalize and generate gene-level intensity for 3′ gene array and Aroma . affymetrix to normalize and summarize gene-level signal for exon arrays ., We used MAS5 to call expressional presence and absence for 3′ gene array ., In case of exon array , we used Affymetrix dabg ( detection above background ) algorithm to generate chip specific background signal and then compared gene-level signal to this background with Wilcoxon rank sum one-tail test ., Considering multiple-testing issues , we converted all p values to q values using the qvalue package 66 ., The q value of 0 . 01 was used as the cutoff ., For Agilent miRNA array , we used “gIsGeneDetected” column generated by Agilent Feature Extraction software to define presence or absence calls 67 ., We required a gene to be present in all replicates to be considered a presence and a gene to be absent in all replicates to be considered an absence ., We removed all ambiguous cases from the final statistics ., We used the LIMMA package 68 to call expressional difference , with a false discovery rate corrected p of 0 . 05 used as the cutoff ., Although we compared testis and ovary , we used the term “male-bias” or “female-bias” rather than “testis-bias” or “ovary-bias . ”, The reason is that these two datasets are nearly equivalent ., A previous study showed that the proportion of germline male-biased genes is much higher than that of somatic male-biased genes ( 20% versus 2% ) 12 ., For meta-analyses of mouse and rat spermatogenic data , we followed the concept of RA and euclidean distance ( d ) to measure the between-species expression divergence 69 ., Specifically , we defined RA as the proportion of expression intensity of one tissue out of all tissues and d as the sum of the square of RA difference for all tissues between mouse and rat , i . e . , ., We mapped 20 out of 33 representative genes in 34 to our gene age data using unique NCBI gene names ., Remarkably , 16 ( 80% ) are rodent-specific , with 11 of them originating after the mouse and rat split ., We note here that this dataset does not overlap with what we described in Table 3 , since Table 3 only presents genes with unique probes , which 19 of these 20 genes do not have ., We downloaded the vertebrate-wide 44-way coding sequence alignment from UCSC ., UCSC known genes mapping to multiple Ensembl genes were discarded ., For Ensembl genes mapping to multiple UCSC known genes , we retained only one UCSC gene with the longest coding region ., Then , considering that low quality assembly often causes unreliable estimation of Ka/Ks 70 , we extracted 17 species with relatively better quality ( Figure 1 ) and then removed all in-frame stop codons or gaps in the alignment ., According to our age dating information , taxa conflicting with the age were removed ., Based on the species tree ( Figure 1 ) , we estimated Ka/Ks for each branch using free ratio model in PAML 71 ., We downloaded Gene Ontology ( GO ) annotations for Ensembl V51 ., We used the program analyze . pl V1 . 9 of TermFinder package 72 to identify those significant terms for new genes , with multiple test corrected p of 0 . 05 as the cutoff and the whole genome as the background ., Herein , TermFinder was updated to V0 . 83 , which corrected a mistake in calculating false discovery rate 73 . | Introduction, Results, Discussion, Materials and Methods | Mammalian X chromosomes evolved under various mechanisms including sexual antagonism , the faster-X process , and meiotic sex chromosome inactivation ( MSCI ) ., These forces may contribute to nonrandom chromosomal distribution of sex-biased genes ., In order to understand the evolution of gene content on the X chromosome and autosome under these forces , we dated human and mouse protein-coding genes and miRNA genes on the vertebrate phylogenetic tree ., We found that the X chromosome recently acquired a burst of young male-biased genes , which is consistent with fixation of recessive male-beneficial alleles by sexual antagonism ., For genes originating earlier , however , this pattern diminishes and finally reverses with an overrepresentation of the oldest male-biased genes on autosomes ., MSCI contributes to this dynamic since it silences X-linked old genes but not X-linked young genes ., This demasculinization process seems to be associated with feminization of the X chromosome with more X-linked old genes expressed in ovaries ., Moreover , we detected another burst of gene originations after the split of eutherian mammals and opossum , and these genes were quickly incorporated into transcriptional networks of multiple tissues ., Preexisting X-linked genes also show significantly higher protein-level evolution during this period compared to autosomal genes , suggesting positive selection accompanied the early evolution of mammalian X chromosomes ., These two findings cast new light on the evolutionary history of the mammalian X chromosome in terms of gene gain , sequence , and expressional evolution . | Some evolutionary theories predict that the X chromosome will be enriched for genes with male functions ., However , recent studies showed there had been gene traffic in which autosomal male-biased genes were retroposed from X-linked parental genes ., A question remains about whether this pattern also holds for all types of new genes ., Herein , using comparative genomic analysis , we dated all human and mouse genes to the vertebrate phylogenetic tree ., We found that the X chromosome evolved with two bursts of gene origination events ., The recent burst includes mainly male-biased genes in contrast to older X-linked genes that are often female-biased in expression ., Meiotic sex chromosome inactivation contributes to this dynamic since it silences the older but not the younger X-linked genes ., The older burst was after the split of eutherian mammals and the marsupial opossum , and the genes from this burst were quickly incorporated into transcriptional networks of multiple tissues , especially in the brain ., The transcriptional expansion , together with the rapid protein evolution of the preexisting old X-linked genes , suggests that positive selection was acting in the early evolution of the mammalian X chromosome ., These two lines of findings revealed extensive gene evolution in the mammalian X chromosome . | evolutionary biology/human evolution, evolutionary biology/bioinformatics, evolutionary biology/genomics | Two bursts of gene gains occurred on the mammalian X chromosome contribute to an age-dependent chromosomal distribution of male-biased genes. |
journal.pcbi.1005464 | 2,017 | Functional asymmetry and plasticity of electrical synapses interconnecting neurons through a 36-state model of gap junction channel gating | In most models of neuronal networks , it is assumed that electrical synapses exhibit constant conductance , and that electric synaptic transmission is bidirectional and symmetric ., However , experimental studies show that these assumptions are not always satisfied ., For instance , some synapses formed of gap junction channels exhibit an instantaneous conductance–voltage rectification , which promotes a direction-dependent asymmetry of electrical signaling 1–4 ., In addition , all members of the connexin ( Cx ) family forming gap junction channels exhibit a sensitivity of junctional conductance to the transjunctional voltage 5 ., Moreover , voltage sensitivity of gap junctions can be strongly affected by chemical factors , e . g . by intracellular concentrations of H+ , Ca2+ or Mg2+ 6–8 ., Thus , electrical synapses are not just passive pores , but can exhibit dynamic changes of junctional conductance ., Presumably , these changes in electrical synaptic strength could affect the transfer of an electrical signal ., The purpose of our study was to develop a computational model for evaluation of such an interaction between electrical synapses and signal transmission between coupled neurons ., The first quantitative models describing equilibrium 9 , 10 and kinetic 11 properties of junctional conductance dependence on transjunctional voltage were based on the assumption that the channel can be in two states , open and closed ., Later , single channel studies have shown that transjunctional voltage causes channels to close to a subconductance ( residual ) state 12 , 13 with fast gating transitions , and to a fully closed state with slow gating transitions 14 , 15 ., Thereafter , it was proposed that gap junction channels comprise two types of gating mechanisms , fast and slow , each exhibiting rectification of their unitary conductances depending on the voltage across them ., These properties were described in a stochastic 16-state model ( 16SM ) of gap junction channel gating 16 in which fast and slow gates operate between open ( o ) and closed ( c ) states ., However , experimental data from our and other groups 17 , 18 allowed us to suggest that the slow gate operates between open ( o ) , initial-closed ( c1 ) and deep-closed ( c2 ) states ., Such a suggestion was implemented in a 36-state model ( 36SM ) of voltage gating 19 ., The 36SM allowed us to reproduce experimentally observed gating behavior of gap junction channels more adequately than 16SM , especially regarding the kinetics of conductance recovery , or a low fraction of functional channels clustered in junctional plaques ., Earlier 20 , we combined a 16SM of gap junction channel gating and rectification with the Hodgkin–Huxley ( HH ) equations 21 ., The developed model ( HH-16SM ) allowed us to evaluate the kinetics of junctional conductance during the spread of excitation in neuronal networks ., In this study , we replaced 16SM with 36SM for a better evaluation of junctional conductance kinetics ., We applied the combined model ( HH-36SM ) to investigate the signal transfer between electrically coupled neurons in response to different types of presynaptic inputs , such as electrotonic signals or action potentials ( APs ) ., In this study , we analyse three main aspects of the 36SM with respect to the functional behavior of electrical synapses ., Firstly , transjunctional voltage distribution across each channel gate can result to almost instantaneous asymmetric conductance-voltage rectification of gap junction channel ., We showed that such rectification of gap junctions can affect the asymmetry of the electrical cell-to-cell signaling , especially the spread of a single AP ., Secondly , in the 36SM , the gap junction channel can transit between open and closed states , and probabilities of these transitions depend on voltage across each channel gate ., We demonstrated that closing of gap junction channels could be induced by transjunctional voltage spikes , which develop during the spread of excitation ., More precisely , our modeling results showed that voltage spikes induced by the trains of APs can cause an accumulation of gap junction conductance decay ., As a result , the junctional conductance can significantly decrease in just a few seconds , and substantially modulate electrical signaling between neurons ., This short-term plasticity of electrical synapses was supported by our electrophysiological experiments in HeLa cells expressing connexin45 ., Thirdly , we suggested that some types of chemical modulation of electrical synapses could be explained by an assumption that the values of 36SM parameters depend on chemical factors ., Under such a hypothesis , the chemical modulator would influence the junctional conductance by modifying voltage sensitivity properties of a gap junction channel ., In this case , the chemically-induced variation of junctional conductance would be explained by the changed equilibrium of open and closed voltage sensitive channels , and not by a separate chemical gate ., To illustrate the feasibility of this idea , we fitted the 36SM to explain the kinetics of connexin36 gap junctional conductance under different concentrations of free magnesium ions ( Mg2+i ) ., We demonstrated that a long-term ( a few minutes ) plasticity , which is induced by variation in Mg2+i , can be adequately reproduced through the changes of 36SM parameters ., Thus , the presented model accounts for the complex behavior of electrical synapses under a wide variety of voltage and temporal conditions ., Moreover , all these phenomena can be explained by the underlying mechanisms of gap junction channel voltage gating ., Such a modeling approach allows one to evaluate the response of neuronal networks , which would be very difficult to measure experimentally ., Junctional conductance of electrical synapses was evaluated using a Markov chain 36-state model of voltage gating , which is detailed in 19 ., The model describes the probabilistic behavior of gap junction channels in response to the transjunctional voltage ., In the 36SM , the gap junction channel consists of two hemichannels , each enclosing one fast and one slow gates ( Fig 1 ) ., Thus , the channel comprises four gates ( fast left , slow left , slow right and fast right ) , all arranged in series ( Fig 1B ) ., The fast and the slow gates operate according to a linear kinetic schemes , o↔c and o↔c1↔c2 , respectively ( Fig 1A ) ., Thus , gap junction channel can be in 36 ( 2∙3∙3∙2 ) different states , and overall junctional conductance is estimated as an averaged value of each state conductance weighted to their probabilities ., Transition probabilities between system states depend on transjunctional voltage distribution across gates , which must be evaluated first ., In general , the voltage distribution can be nonlinear due to rectification of unitary conductances of channel gates ., The developed HH-36SM combines Hodgkin–Huxley equations that describe excitability of neurons and a 36-state model ( 36SM ) of gap junction channel gating that evaluates conductance of the electrical synapse ., More precisely , membrane voltages of the neurons are estimated using the Hodgkin-Huxley model ., The resulting transjunctional voltage can affect the junctional conductance , which is evaluated using the 36SM ., Thus , the HH-36SM allowed us to simulate electrical signal transfer between neurons connected through modulatable gap junctions ., Asymmetry of electrical synaptic transmission has been observed in numerous studies 41–44 ., Such asymmetry might arise due to differences in input resistances ( Rins ) of coupled neurons 29 , even when gap junctions themselves are symmetric ., Rin depends on the conductivity of the plasma membrane and its surface area , as well as on the number of neighboring neurons connected through electrical synapses ., Another source of electrical synaptic transmission asymmetry is related to instantaneous conductance–voltage rectification of gap junction channel , which results from the inhomogeneous distribution of charged amino acids lining the pore 45 ., Such rectification of gap junction channels typically arises in heterotypic junctions under normal conditions 4 , 17 , 46 , but it can also develop in homotypic gap junctions under an asymmetry of intracellular milieu , e . g . gradients of Mg2+i 47 ., In addition , electrical signaling asymmetry across heterotypic channels can arise with repeated stimulation due to voltage gating ( see Fig 7 ) ., This type of asymmetry in electric synaptic transmission is not instantaneous and depends on past history ., The other factors that contribute to asymmetry of signaling do not have this property ., Our data show that asymmetry in electrotonic cell-to-cell communication is more affected by the difference in Rins of coupled cells ( see Figs 3D and 5D ) , while gap junctional rectification primarily influences an asymmetry of AP transfer between neurons ( Fig 5A–5C ) ., This can be explained by the conductance–voltage curves in Fig 2 , which show that conductance changes are small at low voltages ( ±10 mV ) , which typically arise during measurements of coupling coefficients ., Significant changes of junctional conductance can only be expressed at high voltages ( ±100 mV ) , which develop during the spread of excitation ., We believe that these observations might have practical applications in electrophysiological experiments when studying the strength and rectification properties of electrical synapses ., The aforementioned sources of functional asymmetry are independent by nature , e . g . Rin of a neuron directly depends on plasma membrane area , while synaptic rectification is determined by properties of gap junction channels 48 ., Thus , they can act antagonistically promoting bidirectionality of electrical synapses , as was demonstrated in the teleost auditory system 4 ., Alternatively , if rectification of gap junctions and differences in Rins acted synergistically , it could facilitate unidirectional AP transfer ., Thus , unidirectionality , which is a genuine property of chemical synapses , could be executed through electrical synapses alone ., Because electrical synaptic transmission is faster than chemical , unidirectional spread of AP through gap junctions might be useful in rapid response warranting behavior such as escape reflex 49 , 50 ., Asymmetry of electrical synaptic transmission plays an important role in spike-timing regulation , as was demonstrated in neurons of the thalamic reticular nucleus 44 ., In larger networks , even a small asymmetry would add up during the spread of excitation and could significantly affect the latency of AP transfer along neural pathways ., This process could be crucial in temporal coding activities , such as coincidence detection , in which gap junctions are reported to play an important role 50 ., Presumably , the effect of asymmetry of electrical signaling would be difficult to measure and observe experimentally in highly complex neuronal networks , and a simulation-based approach could provide valuable insights on the role of rectification in network dynamics 51 ., It is well established that gap junctional conductance depends on voltage 10 ., Our previous 20 and current modeling studies show that decay of junctional conductance can be induced by voltage gating of gap junction channels during bursting activity of neurons ., To our knowledge , at least one study reported such spiking activity-dependent reduction of electrical synaptic strength in brain slices 52 ., Our data showed that even in gap junctions formed of low-voltage-sensitive Cx36 , this decay exceeds 10% while in more voltage-sensitive Cx isoforms it could reach ~50% over several seconds ( Figs 6 and 7 ) ., The magnitude of junctional conductance decrease and duration of its recovery depends not only on Cx properties but also on the firing rates of neurons ( Fig 6 ) ., Because the transfer of electrical signal and its asymmetry depends on junctional conductance 53 , an activity-induced inhibition of electrical synapses can significantly diminish ( Fig 6A-b ) or even abolish AP transfer between neurons ( Fig 8C-c ) ., Such a role of electrical synaptic plasticity was acknowledged in 54 and was demonstrated by an activity-dependent decrease of junctional conductance together with enhanced asymmetry of electrical synaptic transmission in TRN slices 52 ., Heterotypic gap junctions exhibit structure-determined voltage-gating asymmetry , which could result in even more diverse functional behavior with respect to plasticity and directionality than homotypic gap junction ., As we showed in Fig 8 , changes in junctional conductance and the response rate of neurons depends on the direction of AP spread with respect to the orientation of heterotypic gap junctions ., Thus , heterotypic synapses could promote direction-dependent asymmetry of electrical signal transfer not only by its rectification properties but by asymmetric voltage gating as well ., We presume that such processes might have an important functional role in sensory systems where heterotypic electrical synapses are detected 55 , 56 ., Regulation of the strength of electrical synapses by a variety of chemical reagents is well established ., Others and our data showed that junctional conductance decay caused by chemical uncouplers can be reversed by voltage , while some chemical factors can change voltage sensitivity of Cxs 8 , 38 , 39 , 57 ., These observations , as well as the fact that all known chemical uncouplers close gap junction channels fully but not to residual conductance , suggest that some chemical factors act through the slow gate ., We implemented this idea by simulating Mg2+-mediated changes in junctional conductance of Cx36 gap junctions using the 36SM ., The obtained data revealed that an effect of Mg2+i can be relatively well reproduced ( Fig, 9 ) assuming variation in values of 36SM parameters , mainly V0 and probabilities of c1↔c2 transitions of slow gates ., Moreover , because the voltage sensitivity of the gap junction channels is defined by the same parameters ( see Fig 4 in 20 and Fig 6 in this paper ) , chemically modulated gating would also affect spiking activity-dependent short-term plasticity of the electrical synapse ., Our modeling results showed that even a moderate change ( ±20% ) in Mg2+i could result in very significant differences in the spread of APs between two neurons ( see Fig 10 ) ., Thus , chemically modulated gating of Cx36 can expand the time window of electrical synaptic plasticity for as long as chemical factors are present , which could last for minutes or even hours ., Therefore , even Cx36 , which exhibits relatively low voltage sensitivity , could act as a highly modulatable constituent of neuronal networks due to chemically mediated gating ., Our modeling results show that a persistent spiking activity or chemical factors could keep a significant proportion of gap junction channels in a closed state ., We assume that this process could offer at least a partial explanation to a well-documented ‘low functionality’ of gap junctions , especially those expressed in excitable cells , such as neurons or cardiomyocytes ., Low functionality refers to a small fraction of channels residing in the open ( or high conductance ) state ., This applies to all connexins , such as Cx36 58 , Cx43 26 , Cx45 57 and Cx57 59 , examined on this issue , and likely applies to other Cx isoforms ., The strength of electrical synapses directly affects the level of synchronization in neuronal networks , which can underlie various physiological processes and pathological brain conditions ., For example , increased cortical synchronization correlates with reduced information processing capability in the primary auditory cortex 60 ., The rise in junctional conductance can lead to over-synchronization , which is associated with episodes of epileptic seizures ., Interestingly , an activity-induced decrease in the coupling of electrical synapses through an intracellular Ca2+ mechanism was observed in the thalamic reticular nucleus of epileptic rats and was proposed to act as a compensatory mechanism to reduce excessive synchronization 61 ., Thus , both voltage- and chemically induced gating of gap junction channels can play an important role in shaping activity of neuronal networks through modulation of neuronal synchrony ., In addition , short-term plasticity induced through voltage gating of electrical synapses could contribute to lateral inhibition and resulting center-surround effect , which is important in sensory systems of the CNS ., This hypothesis is supported by studies showing that more voltage-sensitive Cx isoforms are expressed in the structures associated with sensory functions ., For example , one of the most voltage-sensitive Cxs , mouse Cx57 and its human homolog Cx62 are expressed in horizontal cells of the retina 62 , while Cx45 , which is significantly more voltage-sensitive than Cx36 , predominates in the olfactory bulb 34 ., The chemically mediated gating could play an important role in regulating longer term changes , especially in less-voltage-sensitive Cx36 ., For example , it was reported that Cx36 plays an important role in shifting between sleep and wake states 63 ., We believe that the unique sensitivity of Cx36 to Mg2+ could contribute to this process ., This view is supported by accompanying changes in ATP levels , which effectively influence Mg2+i ., It was reported that ATP levels increase during the initial hours of sleep in wake-active regions of rat brain 64 ., This should decrease Mg2+i and , consequently , increase conductance of Cx36 gap junctions ., As a result , an increased synchronization could suppress activities in brain regions associated with the waking state , thus maintaining sleep ., In this study , we used the Hodgkin–Huxley equations to describe excitability of neurons ., The developed model can be adapted to various brain regions and circuits by choosing an appropriate set of ionic currents ., For example , the inclusion of Ca2+ currents , which underlie bursting trains of APs in thalamic relay neurons 65 , might be relevant for short-term plasticity as well as for chemical modulation of electrical synapses ., Furthermore , major principles used to develop an HH-36SM can be applied in cardiac tissue modeling , provided that the Hodgkin–Huxley equations are replaced by those specific for cardiomyocytes 66 , 67 ., Cardiomyocytes are predominantly connected through Cx43 , Cx40 and Cx45 , which are more voltage sensitive than Cx36; therefore , it might exhibit more expressed activity-dependent conductance decrease , especially during tachyarrhythmias ., Furthermore , chemically mediated gating of cardiac gap junction channels , e . g . by acidification 68 , could be important in describing enhanced arrhythmogenicity of the ischemic myocardium 69 ., Obviously , the 36SM of gap junction channel voltage gating is a simplification of complex processes underlying changes of electrical synaptic strength ., However , we believe that rectification and voltage gating properties of gap junction channel can be reasonably well reproduced using the 36SM ., On the other hand , an inclusion of chemical modulation into 36SM is far less explored ., So far we made only the first steps in this direction to explain Cx36 mediation by Mg2+i , and presented modeling results ( Fig, 10 ) are obtained from just a few data points ., Moreover , cytosolic conditions are rarely defined by a single chemical factor , and various different reagents might affect electrical synapses synergistically or antagonistically ., For example , our preliminary data suggest that Mg2+i effect on Cx36 gap junctions might depend on the pH level ., In addition , modulation of electrical synapses by other chemical reagents , such as Ca2+ ions , might be more relevant for the spread of excitation than that of Mg2+i ., In this study , we simulated electrical synaptic transmission between two cells connected through a soma-somatic gap junction ., For a more realistic neuronal network simulation , it would be beneficial to include dendro-dendritic connections , which are far more prevalent in mammalian brain ., Another important extension of our model would be an inclusion of chemical synapses ., Presumably , this would allow one to study an interaction between chemical and electrical synapses , which was observed in numerous experimental studies 70 ., However , all physiologically relevant extensions , and especially an increased number of cells and synapses , might require a large amount of computational recourses ., To our knowledge , Hodgkin-Huxley type models are rarely applied for large neuronal network simulation due to computation time constraints ., This problem would be enhanced by our modeling approach , because evaluation of junctional conductance using the 36SM consumes ~95 percent of overall computation time ., We presume that simulation time could be decreased by two different approaches:, 1 ) Creation of a more simplistic model of gap junction voltage gating , which would roughly describe relative changes of junctional conductance in response to a single AP ., Somewhat similar approach is applied in mathematical models of chemical synapses 71 ., This would allow one to combine a model of electrical synapse with integrate-and-fire type models , which are often used for simulation of large neuronal networks ., 2 ) Application of advanced computation techniques , such as an extensive parallelization together with graphic processing unit computation . | Introduction, Materials and methods, Discussion | We combined the Hodgkin–Huxley equations and a 36-state model of gap junction channel gating to simulate electrical signal transfer through electrical synapses ., Differently from most previous studies , our model can account for dynamic modulation of junctional conductance during the spread of electrical signal between coupled neurons ., The model of electrical synapse is based on electrical properties of the gap junction channel encompassing two fast and two slow gates triggered by the transjunctional voltage ., We quantified the influence of a difference in input resistances of electrically coupled neurons and instantaneous conductance–voltage rectification of gap junctions on an asymmetry of cell-to-cell signaling ., We demonstrated that such asymmetry strongly depends on junctional conductance and can lead to the unidirectional transfer of action potentials ., The simulation results also revealed that voltage spikes , which develop between neighboring cells during the spread of action potentials , can induce a rapid decay of junctional conductance , thus demonstrating spiking activity-dependent short-term plasticity of electrical synapses ., This conclusion was supported by experimental data obtained in HeLa cells transfected with connexin45 , which is among connexin isoforms expressed in neurons ., Moreover , the model allowed us to replicate the kinetics of junctional conductance under different levels of intracellular concentration of free magnesium ( Mg2+i ) , which was experimentally recorded in cells expressing connexin36 , a major neuronal connexin ., We demonstrated that such Mg2+i-dependent long-term plasticity of the electrical synapse can be adequately reproduced through the changes of slow gate parameters of the 36-state model ., This suggests that some types of chemical modulation of gap junctions can be executed through the underlying mechanisms of voltage gating ., Overall , the developed model accounts for direction-dependent asymmetry , as well as for short- and long-term plasticity of electrical synapses ., Our modeling results demonstrate that such complex behavior of the electrical synapse is important in shaping the response of coupled neurons . | In most computational models of neuronal networks , it is assumed that electrical synapses have a constant and ohmic conductance ., However , numerous experimental studies demonstrate that connexin-based channels expressed in neuronal gap junctions can change their conductance in response to a transjunctional voltage or various chemical reagents ., In addition , electrical synapses may exhibit direction-dependent asymmetry of signal transfer ., To account for all these phenomena , we combined a 36-state model of gap junction channel gating with Hodgkin–Huxley equations , which describes neuronal excitability ., The combined model ( HH-36SM ) allowed us to evaluate the kinetics of junctional conductance during the spread of electrical signal or in response to chemical factors ., Our modeling results , which were based on experimental data , demonstrated that electrical synapses exhibit a complex behavior that can strongly affect the response of coupled neurons ., We suggest that the proposed modeling approach is also applicable to describe the behavior of cardiac or other excitable cell networks interconnected through gap junction channels . | cell physiology, ion channel gating, medicine and health sciences, action potentials, neural networks, nervous system, membrane potential, junctional complexes, electrophysiology, neuroscience, gap junctions, ion channels, synaptic plasticity, electrical synapses, neuronal plasticity, developmental neuroscience, computer and information sciences, animal cells, proteins, biophysics, physics, biochemistry, cellular neuroscience, cell biology, anatomy, synapses, physiology, neurons, biology and life sciences, cellular types, physical sciences, neurophysiology | null |
journal.pcbi.1003424 | 2,014 | Fast Reconstruction of Compact Context-Specific Metabolic Network Models | Cell metabolism is known to play a key role in the pathogenesis of various diseases 1 such as Parkinsons disease 2 and cancer 3 ., The study of human metabolism has been greatly advanced by the development of computational models of metabolism , such as Recon 1 4 , the Edinburgh human metabolic network 5 , and Recon 2 6 ., These are genome-scale metabolic network models that have been reconstructed by combining various sources of ‘omics’ and literature data , and they involve a large set of biochemical reactions that can be active in different contexts , e . g . , different cell types or tissues 7 ., To maximize the predictive power of a metabolic model when conditioning on a specific context , for instance the energy metabolism of a neuron or the metabolism of liver , recent efforts go into the development of context-specific metabolic models 8–13 ., These are network models that are derived from global models like Recon 1 , but they only contain a subset of reactions , namely , those reactions that are active in the given context ., Such context-specific metabolic models are known to exhibit superior explanatory and predictive power than their global counterparts 10 , 14 , 15 ., Most algorithms for context-specific metabolic network reconstruction ( see ‘Related work’ section for a short overview ) first identify a relevant subset of reactions according to some ‘omics’ information ( typically expression data and bibliomics ) , and then search for a subnetwork of the global network that satisfies some mathematical requirements and contains all ( or most of ) these reactions 8 , 10 , 13 , 16–18 ., The mathematical requirements are typically imposed via flux balance analysis , which characterizes the steady-state distribution of fluxes in a metabolic network via linear constraints that are derived from the stoichiometry of the network and physical conservation laws 19–23 ., The search problem may target the optimization of a specific functionality of the model ( e . g . , biomass production ) or some other objective 24 , and it may involve repeated tests under different conditions and parameter tuning 8 , 14 , 25 , 26 ., The latter calls for fast algorithms ., We present fastcore , a generic algorithm for context-specific metabolic network reconstruction ., fastcore takes as input a core set of reactions that are supported by strong evidence to be active in the context of interest ., Then it searches for a flux consistent subnetwork of the global network that contains all reactions from the core set and a minimal set of additional reactions ., Flux consistency implies that each reaction of the network is active ( i . e . , has nonzero flux ) in at least one feasible flux distribution 19 , 27 ., An attractive feature of fastcore is its generality: As it only relies on a preselected set of reactions and a simple mathematical objective ( flux consistency ) , it can be applied in different contexts and it allows the integration of different pieces of evidence ( ‘multi-omics’ ) into a single model ., Computing a minimal consistent reconstruction from a subset of reactions of a global network is , however , an NP-hard problem 27 , and hence some approximation is in order ., Our key observation is that a minimal consistent reconstruction can be defined via a set of sparse modes of the global network , and fastcore is designed to compute a minimal such set ., Every iteration of the algorithm computes a new sparse mode via two linear programs that aim at maximizing the support of the mode inside the core set while minimizing that quantity outside the core set ., fastcores search strategy is in marked contrast to related approaches , in which the search for a minimal consistent reconstruction involves , for instance , incremental network pruning 10 ., fastcore is simple , devoid of free parameters , and its performance is excellent in practice: As we demonstrate on experiments with liver data , fastcore is several orders of magnitude faster , and produces much more compact reconstructions , than the main competing algorithm MBA 10 ., A metabolic network of m metabolites and n reactions is represented by an m×n stoichiometric matrix S , where each entry Sij contains the stoichiometric coefficient of metabolite i in reaction, j . A flux vector is a tuple of reaction rates , , where is the rate of reaction i in the network ., Reactions are grouped into reversible ones ( ) and irreversible ones ( ) ., For a reaction it holds that this and other imposed flux bounds , e . g . , lower and upper bounds per reaction , are collectively denoted by ( which defines a convex set ) ., A flux vector is called feasible or a mode if it satisfies a set of steady-state mass-balance constraints that can be compactly expressed as: ( 1 ) An elementary mode is a feasible flux vector with minimal support , that is , there is no other feasible flux vector with , where is the support ( i . e . , the set of nonzero entries ) of 19 , 22 ., A reaction i is called blocked if it cannot be active under any mode , that is , there exists no mode such that ( in practice , for some small positive threshold ε ) ., A metabolic network model that contains no blocked reactions is called ( flux ) consistent 19 , 27 ., Given a metabolic network model with stoichiometric matrix S , a problem of interest is to test whether the network is consistent or not ., Additionally , if the network is inconsistent , it would be desirable to have a method that detects all blocked reactions ., It has been suggested that network consistency can be detected by a single linear program ( LP ) 27 ., The idea is to first convert each reversible reaction into two irreversible reactions ( and define a reversible flux as the difference of two irreversible fluxes ) , and then test if the minimum feasible flux on the new set of irreversible-only reactions is strictly positive ( in practice , at least ε ) ., This is equivalent to testing if the following LP is feasible: ( LP - 2 ) This test of consistency , however , can produce spurious solutions ., In Figure 1 we show a toy metabolic network comprising four metabolites ( A , B , C , D ) and six reactions annotated with corresponding fluxes ., Fluxes are bounded as for , and ., All stoichiometric coefficients are equal to one , except for the reaction →2A ., The only reversible reaction is A↔B , which is a dead-end reaction and therefore blocked , whereas all other reactions are irreversible and unblocked ., After converting A↔B to a pair of irreversible reactions , LP-2 achieves optimal value , which implies ( wrongly ) that the network is consistent ., The test here fails because the two irreversible copies of A↔B have equal flux at the solution , thereby nullifying the actual net flux of A↔B ., A straightforward solution to the problem would involve iterating through all reactions , computing the maximum and minimum feasible flux of each reaction via an LP that satisfies the constraints in ( 1 ) ., Reactions with minimum and maximum flux zero would then be blocked ., This is the idea behind the FVA ( Flux Variability Analysis ) algorithm and the reduceModel function of the COBRA toolbox 28 , 29 ., However , iterating through all reactions can be inefficient ., A faster variant is fastFVA 30 , which achieves acceleration over FVA via LP warm-starts ., Another fast algorithm is CMC ( CheckModelConsistency ) 10 , which involves a series of LPs , where each LP maximizes the sum of fluxes over a subset of reactions: ( LP - 3 ) The set is initialized by ( all reactions in the network ) , and it is updated after each run of LP-3 so that it contains the reactions whose consistency has not been established yet ., When cannot be reduced any further , we can reverse the signs of the columns of S corresponding to the reversible reactions in and resume the iterations ., Eventually , all remaining reactions may have to be tested one by one for consistency , as in FVA ., Such an iterative scheme is complete , in the sense that it will always report consistency if the network is consistent , and if not , it will reveal the set of blocked reactions ., However , as we will clarify in the next section , LP-3 is not optimizing the ‘correct’ function , which may result in unnecessarily many iterations ., For example , when applied to the network of Figure 1 , LP-3 will pick up the elementary mode that corresponds to the pathway A→C→D ( because this pathway achieves maximum sum of fluxes ) , and it will set ., To establish the consistency of the reaction A→D , an additional run of LP-3 would be needed , where the set would only involve the reactions A↔B and A→D ., Hence , an iterative algorithm like CMC that relies on LP-3 would need two iterations to detect the consistent part of this network ., However , one LP suffices to detect the consistent subnetwork in this example , as we explain in the next section ., In more general problems involving larger and more realistic networks , CMC may involve unnecessarily many iterations , as we demonstrate in the experiments ., In most problems of interest there will be no single mode that renders the whole network consistent , and an iterative algorithm like the one described in the previous section must be used ., For performance reasons it would therefore be desirable to be able to establish the consistency of as many reactions as possible in each iteration of the algorithm ., Since consistency implies nonzero fluxes , it is sufficient to optimize a function that just ‘pushes’ all fluxes away from zero ., Formally , this amounts to searching for modes whose cardinality—denoted by card ( v ) and defined as card ( v ) =\u200a#supp ( v ) , i . e . , the number of nonzero entries of —is as large as possible ., Directly maximizing card ( v ) is , however , not straightforward , for the following reasons: First , the card function is quasiconcave only for ( the nonnegative orthant ) , and it is nonconvex for general 31 ., Second , even if we restrict attention to nonnegative fluxes in each iteration ( which we can do without loss of generality by flipping the signs of the corresponding columns of S ) , it is not obvious how to efficiently maximize the quasiconcave card ( v ) ., Third , in practice consistency implies fluxes that are ε-distant from zero , in which case some adaptation of the card function is in order ., Here we propose an approach to approximately maximize card ( v ) over a nonnegative flux subspace indexed by a set of reactions ., First note that the cardinality function can be expressed as ( 4 ) where is a step function: ( 5 ) The key idea is to approximate the function θ by a concave function that is the minimum of a linear function and a constant function: ( 6 ) where ε is the flux threshold ., The problem of approximately maximizing card ( v ) can then be cast as an LP: We introduce an auxiliary variable for each flux variable , for , and take epigraphs 31 , in which case maximizing card ( v ) can be expressed asUsing ( 6 ) and assuming constant ε , this simplifies to ( LP-7 ) Note that LP-7 tries to maximize the number of feasible fluxes in whose value is at least ε ( contrast this with LP-2 ) ., Returning to the network of Figure 1 , if comprises all network reactions , then note that the flux vector is an optimal solution of LP-7 ., Hence , a single run of the latter can detect all unblocked reactions of that network ., More generally , a single run of LP-7 on an arbitrary subset of a given network will typically detect all unblocked irreversible reactions of ., The intuition is that LP-7 prefers flux ‘splitting’ over flux ‘concentrating’ in order to maximize the number of participating reactions in the solution , which , in the case of irreversible reactions , corresponds to flux cardinality maximization ., By construction , the above approximation of the cardinality function applies only to nonnegative fluxes ., In order to deal with reversible reactions that can also take negative fluxes , we can embed LP-7 in an iterative algorithm ( as in the previous section ) , in which reversible reactions are first considered for positive flux via LP-7 , and then they are considered for negative flux ., The latter is possible by flipping the signs of the columns of the stoichiometric matrix that correspond to the reversible reactions under testing , in which case the fluxes of the transformed model are again all nonnegative , and the above approximation of the cardinality function can be used ., This gives rise to an algorithm for detecting the consistent part of a network that we call fastcc ( for fast consistency check ) ., Since fastcc is just a variant of fastcore , we defer its detailed description until the next section ., Independently to this work , a similar approach to network consistency testing was recently proposed , called OnePrune 32 ., OnePrune first converts each reversible reaction into two irreversible reactions , forming an augmented set of irreversible-only reactions ( as in LP-2 above ) , and then it employs an LP that coincides with LP-7 for the above choice of and ε\u200a=\u200a1 ., However , such an approach is prone to the same drawback as LP-2 , namely , that the two irreversible copies of a blocked reaction can carry equal positive flux at the solution of LP-7 due to the presence of cycles introduced by the transformation ., The authors acknowledge this problem but they do not fully resolve it ., In our case , we avoid this problem by working with the original reactions and a series of LPs with appropriate sign flips of the stoichiometric matrix , thereby guaranteeing the completeness of the algorithm ., The reconstruction problem involves computing a minimal consistent network from a global network and a ‘core’ set of reactions that are known to be active in a given context ., Formally , given, ( i ) a consistent global network with reaction set and stoichiometric matrix , and, ( ii ) a set , the problem is to find the smallest set such that and the subnetwork induced by the reaction set is consistent ., ( By we denote the submatrix of that contains only the columns indexed by . ), This problem is known to be NP-complete 27 , suggesting that a practical solution should entail some approximation ., ( We note that Acuña et al . 27 prove NP-completeness of this problem by noting that a special case involves being the empty set , in which case the problem comes down to finding the smallest elementary mode of the global network , which , as the authors show , is NP-complete . However , this leaves open the case of a nonempty core set , since a solution to the minimal reconstruction problem need not constitute an elementary mode . We conjecture that the problem remains NP-hard when is nonempty , but we are not pursuing this question here . ), Our approach hinges on the observation that a consistent induced subnetwork of the global network can be defined via a set of modes of the latter: Theorem 1 ., Let be a set of modes of the global network , and let supp ( v ) be the union of the supports of these modes ., The induced subnetwork is consistent ., Proof ., For each , let be the ‘truncated’ after dropping all dimensions not indexed by ., Clearly , , therefore each is a mode in the reduced model ., By construction of , each reaction in is in the support of some , and hence also in the support of some mode of the reduced model ., This simple result allows one to cast the reconstruction problem as a search problem over sets of modes of the global network: ( NLP-8 ) Note that this optimization problem involves searching for a set of modes of , such that the union of the support of these modes ( the set ) is a minimal-cardinality set that contains the core set ., In order to practically make use of this theorem , one has to define a search strategy over modes ., Next we discuss two possibilities ., The first gives rise to an exact algorithm , but this algorithm does not scale to large networks ., The second is a scalable greedy approach that gives rise to fastcore ., Several algorithms have been published in the last years for extracting condition-specific models from generic genome-wide models like Recon 1 ., Among them , mCADRE 26 , INIT 13 , iMAT 35 , MBA 10 and GIMME 8 are the most commonly used ( see Table 1 for an overview ) ., Here we provide a short outline of the different algorithms , and refer to 24 for a more extensive overview ., For GIMME , iMAT , and MBA , we briefly discuss some notable differences to fastcore ., GIMME 8 takes as input microarray data and a biological function to optimize for , such as biomass production ., GIMME starts by removing reactions with associated expression levels below a user-defined threshold , and then it optimizes for the specified biological function using linear programming ., In case the pruning steps compromise the input biological function , GIMME reintroduces some previously removed reactions that are in minimal disagreement with the expression data ., Since GIMME has not been designed to include all core reactions in the solution ( as fastcore does ) , the reconstructions obtained by GIMME and fastcore can differ significantly: Running the createTissueSpecific function of the COBRA toolbox on a set of liver core reactions ( see ‘Results’ section ) treating them as expressed reactions ( and adding a biomass reaction 26 and a sink reaction for glycogen to be used as optimization function ) , only about 50% of the core reactions of the GIMME model were consistent at the solution ., A fairer comparison would require adapting fastcore to explicitly deal with omics data , which is outside the scope of the current work ., iMAT 35 was originally designed for the integration of transcriptomic data ., iMAT optimizes for the consistency between the experimental data and the activity state of the model reactions ., iMAT tries to include modes composed of reactions associated to genes with high expression value , and therefore a threshold needs to be chosen to segregate between low , medium , and highly expressed genes ., The computational demands of iMAT are high due to the repeated use of mixed integer linear programming ., As with GIMME , direct comparison of iMAT to fastcore is problematic ., Nevertheless , we applied iMAT ( own implementation ) on the liver problem ( see ‘Results’ section ) , by setting the liver core reactions to RH ( reaction high ) and all non-core reactions to RL ( reaction low ) ., iMAT determined 549 core reactions as active , while 182 and 338 reactions were classified as undetermined and inactive , respectively ., This means that about 50% of the core reactions were lost during iMAT model building ., As with GIMME , this demonstrates the difficulty of directly comparing fastcore to algorithms that optimize different objectives ., mCADRE 26 is similar to MBA , except that the pruning order is not random , but it depends on the tissue-specific expression evidence and weighted connectivity to other reactions of the network ., Reactions that are associated to genes that are never tagged as expressed and which are not connected to reactions associated to highly expressed genes are first evaluated in the pruning step ., Reactions are effectively removed if the removal does not impair core reactions and metabolic functions to carry a flux ( mCADRE removes core reactions if the core/non-core reaction ratio is below a user-given threshold ) ., mCADRE uses mixed integer linear programming and therefore it does not scale up to large networks ( but it is in general faster than MBA ) ., INIT 13 uses data retrieved from public databases in order to assess the presence of a certain reaction-respective metabolites in the cell type of interest ., INIT uses mixed integer linear programming to build a model in which all reactions can carry a flux ., Contrary to other algorithms , INIT does not rely on the assumption of a steady state , but it allows small net accumulation of all metabolites of the model ., The closest algorithm to fastcore is the MBA algorithm of Jerby et al . 10 ., MBA takes as input two core sets of reactions , and it searches for a consistent network that contains all reactions from the first set , a maximum number of reactions from the second set ( for a given tradeoff ) , and a minimal number of reactions from the global network ., ( fastcore can be easily adapted to work with multiple core sets , by introducing a set of weights that reflect the confidence of each reaction to be active in the given context , and adding appropriate regularization terms in the objective functions of LP-7 and LP-10 that capture the given tradeoff . We will address this variant in future work . ), Both fastcore and MBA involve a search for a minimal consistent subnetwork , however the search strategy of fastcore is very different to MBA: Whereas fastcore iteratively expands the active set starting with , MBA starts with and iteratively prunes the set by checking whether the removal of each individual reaction ( selected in random order ) compromises network consistency ., As the pruning order affects the output model , this step of MBA is repeated multiple times ., MBA builds a final model by adding one by one non-core reactions with the highest presence rate over all pruning runs , and it stops when a consistent final model is obtained ., Due to the multiple pruning runs , MBA has very high computational demands ., Consistency testing in MBA is carried out with the CMC algorithm that is based on LP-3 , as explained earlier ., Hence , fastcores search strategy differs to MBA in two key aspects: First , consistency testing in fastcore involves the maximization of flux cardinality ( LP-7 ) instead of sum of fluxes ( LP-3 ) , which results in fewer LP iterations ., Second , the search for compact solutions in fastcore involves L1-norm minimization instead of pruning ., The advantage of the former is that it can be encoded by a single LP , resulting in significant overall speedups ( see ‘Results’ section ) ., In the first set of experiments we applied fastcc , the consistency testing variant of fastcore , for consistency verification of four input models , and compared it against the FastFVA algorithm of Gudmundsson and Thiele 30 , and an own implementation ( based on fastcc but with LP-3 replacing LP-7 ) of the CMC algorithm of Jerby et al . 10 ., We also tested the FVA algorithm of the reduceModel function of the COBRA toolbox 29 , and the MIRAGE algorithm of Vitkin and Shlomi 36 , but we do not include them in the results as they performed worse than the reported ones ., The input models were the following: The results are shown in Table, 2 . fastcc is faster and it uses much fewer LPs than the other two algorithms ., We note that fastFVA is based on an optimized Matlab/C++ implementation with LP warm-starts , while fastcc is based on standard Matlab ., These results confirm the appropriateness of flux cardinality ( LP-7 ) as a metric for network consistency testing , in agreement with the theoretical analysis and the discussions above ., In the second set of experiments , we used the fastcore algorithm to reconstruct a liver specific metabolic network model from the consistent part of Recon 1 ( c-Recon1 , ) , and we compared against an own implementation of the MBA algorithm of Jerby et al . 10 ., We applied the two algorithms in two settings ., The first setting involves the liver specific input reaction set of Jerby et al . 10 , which is based on 779 ‘high’ core and 290 ‘medium’ core reactions ( the latter set is supported by weaker biological evidence than the former ) ., To allow a comparison with fastcore , we defined a single core set as the union of the high and medium core reaction sets , and we applied the two algorithms on this core set ., The second setting uses the ‘strict’ liver model of Jerby et al . 10 , which contains 1083 high core reactions and no medium core reactions , and therefore allows a direct comparison with fastcore ., The results for the two settings are shown in Table, 3 . We note that for MBA , the reported number of LPs and the runtime refer to a single pruning iteration of the algorithm , whereas the size of each reconstruction refers to the final model after 1000 pruning iterations ., In both settings , fastcore is several orders of magnitude faster than MBA , achieving a full reconstruction of a liver specific model in about one second , using a much smaller number of LPs ., As MBA employs a greedy pruning strategy for optimization , the number of LPs that it uses and its total runtime can be very high , as also indicated by Wang et al . 26 who reported runtime of a single pruning pass of MBA in the order of 10 hours on a 2 . 34 GHz CPU computer ., The reconstructed models by fastcore are also more compact than those obtained by MBA , with a difference of 70–80 non-core reactions ., For the standard liver model , 1687 out of the 1746 reactions ( 96% ) of the fastcore reconstruction appear also in the MBA reconstruction , whereas for the strict liver model the common reactions are 1739 out of 1818 ( 95% ) ., The two algorithms turned out to use alternative transporters to connect the core reactions: In the standard liver model , 46 out of 59 reactions that are present exclusively in the fastcore reconstruction are transporter reactions or other reactions which are not associated to a specific gene and thus are not sufficiently supported in the core set , whereas in MBA the corresponding numbers are 116 out of 139 reactions ., ( In Text S1 we provide more details on the reconstructions obtained by the two methods . ), Note that both MBA and fastcore try to minimize the number of added non-core reactions in order to obtain a compact consistent model ., The above difference in the number of added non-core reactions between MBA and fastcore is the result of the different optimization approaches taken by the two algorithms , and no biological relevance should be attributed to each reconstruction other than the one implied by the makeup of the core set ., From this point of view , fastcore performs in general better than MBA , as it tends to add fewer unnecessary reactions ., We also compared fastcores reconstructions to the exact solutions obtained from MILP-9 , using core sets that are randomly generated from a consistent subset of E . coli core 38 ., This is a small model with and 414 elementary modes ( unfortunately , the dependence of the MILP-9 model on the number of elementary modes did not allow testing larger models ) ., In Figure 3 we show the size of the reconstructed models ( mean values ) obtained with the MILP formulation vs . fastcore , as a function of the size of the core set ., fastcore is capable of obtaining very good approximations to the optimal solutions , which improve with the size of the core set ., To evaluate fastcores performance in correctly identifying liver reactions , we performed repeated random sub-sampling validation in which fastcore was used to reconstruct the liver metabolism based on a reduced , randomly selected ‘subcore’ set of 80% of the original core reactions ., As in 10 , we wanted to test whether fastcore is able to recover a significant number of the 20% left-out core reactions ., To test for the enrichment of the left-out core reactions in the reconstructed model , we used a hypergeometric test , in which the total population is defined by all non-subcore reactions in the global network , the number of draws is defined as the number of non-subcore reactions included in the reconstruction , and the left-out core reactions are the ‘successes’ ., Under the null-hypothesis that there is no enrichment for the left-out core reactions when reconstructing the liver model based on the subcore set , we can compute a p-value for including at least the number of observed left-out core reactions in the reconstruction ., We repeated this random sub-sampling procedure 500 times and computed the corresponding p-values ., The median of these p-values was 0 . 0025 , indicating the ability of fastcore to capture liver-specific reactions that were included in the original core set ., As argued above , the reconstructions obtained by fastcore need not optimize for cellular functions other than the ones implied by the composition of the input core set , and it is an interesting research question how to modify fastcore so that it can explicitly capture functional requirements in its reconstructions ., Nevertheless , it is of interest to test whether the current version of fastcore can produce reconstructions that are functionally relevant , perhaps for slight variations of the core set ., To this end , as in 10 , we checked whether the ( standard ) liver model reconstructed by fastcore can perform gluconeogenesis from glucogenic amino acids , glycerol , and lactate ( altogether 21 metabolites ) ., If not yet included , transporters from the extracellular medium to the cytosol were added to the model ( glycerol , glutamate , glycine , glutamine , and serine ) ., This was necessary as the transport reactions were not sufficiently supported in the core set ., This ‘extended’ liver model was able to convert 17/21 metabolites ( vs 12/21 metabolites of the non-extended model ) ., The extended liver model was then used to simulate the liver disorders hyperammonemia and hyperglutamenia , which affect the capacity to metabolize dietary amino acids into urea 10 ., Loss of function mutations of three enzyme-coding genes , argininosuccinate synthetase ( ASS ) , argininosuccinate lyase ( ASL ) , and ornithine transcarbamylase ( OTC ) were identified in patients suffering from these disorders ., The rates of the reactions controlled by the three genes were fixed to 500 , 250 , or zero , to mimic the healthy homozygote ( no mutation ) , heterozygote ( loss of one allele ) , and the complete loss of function , respectively ., To allow for a comparison with the experimental study of Lee et al . 39 where labeled 15N-glutamine was administrated to patients suffering from inborn errors affecting the three genes , we explicitly shut down the influx of other potential nitrogen sources in the liver model , thereby simulating only the uptake and metabolism of glutamine ., By allowing the influx of only one nitrogen source , the fate of the latter could be determined exactly in the model ., The ratio of urea secretion level over glutamine absorption was computed by sampling over the feasible space 21 ., In accordance with the wet lab observations 39 , the severity of the disorders , characterized by the mean urea over glutamine ratio , increased with the level of loss of function of the three genes ASS , ASL , and OTC ( see Figure 4 ) ., Null patients showed no native production of urea ., Overall , the ratios predicted by the fastcore model faithfully match the experimentally observed ones 39 ., ( The corresponding ratios reported by Jerby et al . when using the MBA algorithm 10 matched less well the experimental observations , probably because of the cross-feeding of nitrogen to urea from multiple nitrogen sources . By running the above procedure on the MBA model , we noticed that both models attained comparable urea/glutamine flux ratios . ), To summarize , the above experiments demonstrate that , by an informed choice of the core set and influx bounds , fastcore can indeed give rise to functionally relevant models ., We also used the fastcore algorithm to build a cell-type specific murine macrophage model from the consistent part of Recon1bio ( comprising reactions ) ., Recon1bio ( ) is a modified Recon 1 model that contains three extra reactions ( biomass , NADPOX , and a sink reaction to balance the glycogenin self-glucosylation reaction ) 15 ., We used a core set comprising 300 ( out of 382 ) proteomics derived Raw264 . 7 macrophage reactions , as described by Bordbar et al . 15 ., ( The remaining 82 reactions could not be added to the core set as they are situated in an inconsistent region of Recon 1 and therefore carry a permanent zero net flux . ), For their macrophage reconstruction , Bordbar et | Introduction, Methods, Results, Discussion | Systemic approaches to the study of a biological cell or tissue rely increasingly on the use of context-specific metabolic network models ., The reconstruction of such a model from high-throughput data can routinely involve large numbers of tests under different conditions and extensive parameter tuning , which calls for fast algorithms ., We present fastcore , a generic algorithm for reconstructing context-specific metabolic network models from global genome-wide metabolic network models such as Recon X . fastcore takes as input a core set of reactions that are known to be active in the context of interest ( e . g . , cell or tissue ) , and it searches for a flux consistent subnetwork of the global network that contains all reactions from the core set and a minimal set of additional reactions ., Our key observation is that a minimal consistent reconstruction can be defined via a set of sparse modes of the global network , and fastcore iteratively computes such a set via a series of linear programs ., Experiments on liver data demonstrate speedups of several orders of magnitude , and significantly more compact reconstructions , over a rival method ., Given its simplicity and its excellent performance , fastcore can form the backbone of many future metabolic network reconstruction algorithms . | Metabolism comprises all life-sustaining biochemical processes ., It plays an essential role in various aspects of biology , including the development and progression of many diseases ., As the metabolism of a living cell involves several thousands of small molecules and their conversion , a full analysis of such a metabolic network is only feasible using computational approaches ., In addition , metabolism differs significantly from cell to cell and over different contexts ., Therefore , the efficient generation of context-specific mathematical models is of high interest ., We present fastcore , a fast algorithm for the reconstruction of compact context-specific metabolic network models ., The algorithm takes as input a global metabolic model and a set of reactions that are known to be active in a given context , and it produces a context-specific model ., fastcore is significantly faster than other algorithms , typically obtaining a genome-wide reconstruction in a few seconds ., High-throughput model building will soon become a common procedure for the integration and analysis of omics data , and we foresee many future applications of fastcore in disease and patient specific metabolic modeling . | algorithms, systems biology, computer science, metabolic networks, biology, computational biology | null |
journal.pcbi.1004756 | 2,016 | Oligomers of Heat-Shock Proteins: Structures That Don’t Imply Function | Passive molecular chaperones inhibit the aggregation of cytosolic proteins and are thus a nearly ubiquitous component of living cells 1–3 ., This class of chaperones comprises clusterin , α-crystallins and many other small heat-shock proteins ( sHSPs ) , which promote tolerance to a wide range of cellular stressors such as elevated temperatures and hazardous nonspecific interactions 4 , 5 ., These chaperones cannot by themselves fold or refold misassembled proteins and do not require ATP to function ., Instead , passive chaperones associate reversibly with aggregation-prone proteins ., Even when present in sub-stoichiometric ratios with their client proteins , sHSPs and similar chaperones are effective at suppressing aggregation and coping with environmental stress 6–8 ., Yet the mechanism by which this class of chaperones stabilizes the cytosol is not well understood despite significant efforts at determining the structural properties of these molecules ., Here we propose that passive chaperones function by increasing the overall solubility of the proteome ., Through this mechanism , passive chaperones reduce the fraction of toxic oligomers in solution and suppress the nucleation of protein aggregates ., It has recently become apparent that some sHSPs can also interact with protein aggregates in order to curtail further protein deposition 9–11 ., These aggregates are often detrimental to cellular survival , in part because they can sequester other crucial proteins 12 ., We provide simulation evidence that this effect on the proteome solubility is a generic feature of passive chaperones that associate promiscuously and reversibly with their clients ., There is substantial experimental evidence that passive chaperones interact promiscuously with client proteins in chemical equilibrium ., Both the rate of client aggregation and the fraction of chaperones associated with insoluble proteins are concentration-dependent 1 , 3 ., Furthermore , chaperone binding responds directly to increases in the available client binding surfaces , including hydrophobic regions of destabilized clients that are only transiently exposed 13 ., The binding of passive chaperones often modifies the size and structure of amorphous aggregates , leading to smaller soluble clusters in which the putative chaperone binding sites are associated with the hydrophobic interfaces of the client proteins 14–16 ., On the basis of these dynamic chaperone–client aggregates , previous studies have suggested that such aggregates might serve as a relatively inert depot of misfolded proteins during cellular stress 2 , 17–20 ., However , client proteins are not the only substrates to which passive chaperones bind: these chaperones are commonly found in chaperone-only oligomers both in vitro and in vivo7 , 14–16 , 21–24 ., Recent experiments indicate that these dynamic oligomers are also under thermodynamic control 15 , 16 , 25 , 26 and vary with the experimental conditions , such as the temperature and the ionic strength of the solution 25 , 27 , 28 ., Because this tendency to form oligomers is highly conserved across the family of sHSPs and similar molecular chaperones , it has long been recognized that dynamic fluctuations in the oligomeric state play an important role in the organization of many passive chaperones 7 , 25 , 29 , 30 ., At present , however , it is unclear whether the formation of chaperone oligomers is a key functional event ., In fact , there is considerable evidence to the contrary: experiments have shown that mutations and post-translational modifications that alter the tendency of chaperones to form oligomers do not necessarily affect their function 27 , 31–34 ., These observations raise the question of how , if at all , the presence of chaperone oligomers contributes to their ability to solubilize aggregation-prone proteins in vivo ., Here we show that both the function and oligomerization of passive molecular chaperones can be explained by identifying the optimal conditions for a thermodynamically stable solution of chaperones and aggregation-prone proteins ., Our results suggest that low concentrations of promiscuous chaperones are a generic means of stabilizing a biological mixture with respect to a variety of nonfunctional interactions ., In protein solutions under physiological conditions , the interactions between proteins are short-ranged in comparison to the size of the monomers , since the high ionic strength characteristic of physiological media leads to an effective screening of electrostatic interactions 37 ., We therefore choose to model protein interactions through nearest-neighbor contacts on a three-dimensional lattice , where unoccupied lattice sites represent an implicit solvent ., Monomers interact if they reside on adjacent lattice sites , and they are free to rotate and to move among lattice sites in accordance with the equilibrium Boltzmann distribution ., We assume that each protein exists in a single coarse-grained conformation and that the interactions between proteins are determined by effective binding free energies ( Fig 1 ) ., This coarse-graining of the internal degrees of freedom allows us to capture the effects of the intermolecular forces in a reduced set of parameters and is particularly suitable for both globular proteins in near-native states and misassembled proteins with exposed hydrophobic regions ., All monomers on adjacent lattice sites experience an orientationally averaged nonspecific interaction , which is assigned a dimensionless free energy of −βϵ ., ( Interaction energies are expressed in thermal units: β−1 ≡ kB T , where kB is the Boltzmann constant and T is the absolute temperature . ), Because aggregation-prone proteins are likely to participate in directional protein–protein interactions via multiple binding sites 1 , 38–40 , which also promote interactions with sHSPs 41 , 42 , we choose a client model with three patches that is susceptible to aggregation by means of directional interactions alone ( Fig 1 ) ., The directional interactions between client monomers are assigned an attractive free energy of −βϵs-s ., These interactions are chosen to be strong enough to form insoluble client aggregates in the absence of both chaperones and additional nonspecific interactions 35 ., A minimal model of a passive chaperone must be capable of binding exposed patches on the client monomers ., Here we assume that the chaperone monomers have a single binding site and that the interaction free energy between chaperone and client patches is −βϵc-s ( Fig 1 ) ., While this assumption is clearly a simplification of the structure of passive chaperones , which may interact with diverse clients via different binding sites , this representation captures the passivation of interactive client binding sites through the burial of hydrophobic surfaces ., Most importantly , this representation has the physical features that are necessary to capture the qualitative effects of passive chaperones on the thermodynamics of a complex fluid ., Because passive chaperones are known to function at low concentrations , we assume that there are always fewer chaperones than client binding sites ., In what follows , the relative amounts of the chaperone and client monomers in solution are indicated by xc and xs , respectively , such that xc + xs = 1 ., We have used Monte Carlo simulations and finite-size-scaling techniques to calculate the miscibility limit of this model , i . e . , the point at which the chaperone–client mixture becomes unstable with respect to aggregation and/or demixing ( see Methods ) ., This miscibility limit coincides with a thermodynamic instability , where small , spontaneous fluctuations are sufficient to establish long-ranged spatial heterogeneity in an initially well-mixed solution ., In a protein solution , a thermodynamic instability may have contributions from directional interactions , which cause the polymerization and demixing of the strongly interacting species , as well as orientationally averaged interactions , which drive the formation of thermodynamic phases with differing protein densities 35 ., Strong directional interactions between the client proteins can thus lead to the formation of disordered aggregates and an accompanying loss of solubility ., As expected , the presence of chaperones inhibits the formation of client oligomers by competing for binding to patches on the client monomers ., However , this passivation of directional interactions is not the only effect of chaperone binding: the interactions between chaperones and client proteins simultaneously increase the strength of the orientationally averaged nonspecific interactions that the solution can tolerate while remaining thermodynamically stable ., This effect can be seen in Fig 2 , which shows the miscibility limit , βϵ* , at which insoluble aggregates first appear in the solution ., When the strength of the orientationally averaged nonspecific interactions increases beyond this limit , i . e . , βϵ > βϵ* , the solution becomes unstable with respect to small fluctuations in the protein concentrations ., Increasing the strength of these nonspecific interactions can thus cause the solution to become unstable without altering the strength of the directional interactions that drive the polymerization of the client monomers ., Our calculations show that passive chaperones dramatically affect the miscibility limit by inhibiting polymerization and solubilizing transient clusters of client proteins , despite the fact that there are far fewer chaperones than there are client binding sites ., In the absence of chaperones , i . e . , xc → 0 , the solution is unstable due to the strong directional interactions between the aggregating client monomers ., In this case , βϵ* is negative , indicating that a solution of sufficiently concentrated client proteins in a well-screened solvent will form insoluble aggregates ., It is important to note that even when βϵc-s = 0 , the chaperones still interact nonspecifically with the client monomers through the orientationally averaged interaction βϵ ., Here we find that the addition of such ‘inert’ chaperones has a negligible effect on the miscibility limit relative to a client-only solution ., This observation also implies that the majority of cytosolic proteins that are not aggregation-prone do not significantly affect the miscibility limit when the dominant instability is driven by strong directional interactions ., Our calculations further indicate that the thermodynamic forces driving these instabilities are qualitatively different in solutions with weakly and strongly binding chaperones ., In the case of weakly binding chaperones ( βϵc-s ≪ βϵs-s ) , the solution demixes into client-enriched and client-depleted phases primarily as a result of directional interactions ., Insoluble client aggregates recruit monomers via the formation of directional contacts and exchange small oligomers with the coexisting solution ., With strongly binding chaperones ( βϵc-s ≫ βϵs-s ) , the solution forms a high-density condensate consisting of both chaperones and client proteins bound by nonspecific interactions ., Under these conditions , the proteins in both the soluble and insoluble phases exist as amorphous clusters that decrease in size as the stoichiometric ratio xc/xs is increased ., The introduction of strongly binding chaperones , even in low concentrations , significantly increases the solution miscibility limit towards the theoretical maximum for this model , β ϵ max * ≃ 0 ., 87 ., Because a chaperone that interacts with a variety of clients is very likely to engage in promiscuous interactions , it is reasonable to assume that chaperones do not distinguish among the various hydrophobic surfaces in solution ., The strength of interactions between chaperone binding sites is likely to be similar to the strength of interactions between chaperones and clients , and thus it is natural to assume that βϵc-c = βϵc-s ., However , if we instead prevent chaperone–chaperone binding by setting βϵc-c = 0 , we find that the effect on the solution miscibility limit is negligible ( Fig 2 ) ., Since the parameter βϵc-c directly controls the probability of chaperone dimerization , our calculations suggest that the formation of chaperone oligomers has a very minor effect on chaperone function ., Experimentally , the relationship between oligomerization and chaperone function has been probed by modifying or truncating sHSPs 27 , 32–34 ., The available experimental evidence indicates that alterations to the putative client binding sites on sHSPs affect the oligomer equilibria and the functionality of the chaperones independently , in qualitative agreement with the predictions of this model ., Putting these results into context , we now ask , “Is there an optimal chaperone–client binding strength for a biological mixture ? ”, Fig 2 shows that strongly binding chaperones are best suited for increasing the miscibility limit ., In this case , producing more chaperones ( or reducing the total concentration of aggregation-prone clients ) increases βϵ* in an approximately linear relationship , allowing an organism to respond effectively to an increase in nonspecific interactions ., Nevertheless , strong promiscuous interactions come at a cost: nearly irreversible binding between a chaperone and any available association site , including other proteins that are not explicitly modeled in our simulations , sequesters both interfaces , thereby preventing their participation in further functional interactions ., The optimal chaperone binding strength must balance these competing requirements for solution stability and reversible binding ., Despite the complexity of naturally occurring protein solutions , we can predict the optimal chaperone binding strength by considering a generic fitness function , which quantifies the trade-offs in biological costs and benefits ., This fitness should be maximized for optimal biological function ., For the present model , the fitness F is a function of the miscibility limit , βϵ* , as well as two biological costs that depend on the variables βϵc-s and xc ., In the absence of any deleterious effects of chaperone action , increasing the solubility of the proteome must be beneficial , and thus F should be an increasing function of βϵ* ., However , one potential cost of chaperone action arises from the sequestration of functional proteins ( which are not explicitly modeled in our simulations but must be present in a naturally occurring protein solution ) due to the promiscuous binding of chaperones ., Another potential cost is associated with the production of chaperone molecules ., These costs imply that the fitness function F β ϵ * ( β ϵ c-s , x c ) , β ϵ c-s , x c should satisfy both ∂ ℱ / ∂ β ϵ c-s | β ϵ * < 0 and ∂ ℱ / ∂ x c | β ϵ * < 0 , respectively , when the miscibility limit βϵ* is held constant ., Taking the total derivative of F with respect to both βϵc-s and xc , we find that this fitness function is maximized where, ∂ β ϵ * ∂ β ϵ c-s = - ∂ F ∂ β ϵ c-s β ϵ * ∂ F ∂ β ϵ * and ∂ β ϵ * ∂ x c = - ∂ F ∂ x c β ϵ * ∂ F ∂ β ϵ * ., All partial derivatives of F depend on the precise nature of the biological system and thus cannot be determined precisely ., Nevertheless , it is reasonable to assume that the cost derivatives of F are approximately constant: at low chaperone concentrations , the law of mass action implies that the cost due to reversible , promiscuous binding is approximately linear in both xc and βϵc-s , while the total cost associated with the production of chaperone molecules is also proportional to their concentration ., We can therefore interpret the ratios of the derivatives in each of the above equations as the importance of each cost relative to the benefit of stabilizing the protein solution ., Assuming that promiscuous chaperone binding and chaperone production are indeed significant biological costs , then these equations imply that we should seek to optimize the fitness by maximizing the response functions ∂βϵ*/∂βϵc-s and ∂βϵ*/∂xc ., More intuitively , maximizing these response functions directs the optimal chaperone design towards the region of parameter space in which the solution miscibility limit is most sensitive to small increases in either the chaperone–client binding strength or the number of chaperone molecules in solution ., The first condition , ∂βϵ*/∂βϵc-s , biases the optimal chaperone design away from values of βϵc-s for which the miscibility limit increases asymptotically , thus discriminating against excessively strong binding between chaperones and clients ., The second condition , ∂βϵ*/∂xc , requires that the miscibility limit be sensitive to changes in the chaperone stoichiometric fraction ., Our calculations show that it is indeed possible to satisfy both conditions simultaneously ., In Fig 3a and 3b , we plot the calculated response functions ∂βϵ*/∂βϵc-s and ∂βϵ*/∂xc , in dimensionless units , as functions of the chaperone stoichiometric fraction and the chaperone binding strength ., We identify a ‘design window’ for optimal chaperone operation by finding the approximate range of chaperone binding strengths over which both response functions are maximized given a fixed chaperone stoichiometric fraction ., The region of parameter space in which both response functions can be maximized is relatively narrow , suggesting that optimized passive chaperones should have tightly constrained binding strengths ., We further find that the optimal range of chaperone binding strengths is only weakly dependent on the chaperone stoichiometric fraction ., This observation implies that chaperones with fixed binding strengths can operate close to optimality over a wide range of sub-stoichiometric concentrations ., We also note that these protein–protein interaction free energies are in the physical range of a few kB T . The optimal chaperone binding strength is generally weaker than the client–client interactions , indicating that the chaperones need not out-compete the aggregation-prone clients for association with exposed binding interfaces ., Remarkably , our simulations reveal that the probability of finding chaperone oligomers is also highest in the region of parameter space where the optimal design conditions for chaperone activity are satisfied ., In Fig 3c , we plot the probability of chaperone–chaperone binding at the miscibility limit , assuming that βϵc-c = βϵc-s ., We find that this probability is maximal in the window of optimal chaperone binding strength over the complete range of simulated chaperone stoichiometric fractions ., Under these conditions , a significant fraction of the chaperone binding sites are not associated with the aggregation-prone interfaces on the client proteins , but are rather buried in chaperone-only oligomers ., This fraction may be even higher in the miscible fluid or in the presence of client proteins that are less prone to aggregation due to weaker directional interactions ., These calculations provide further evidence that the assembly of chaperone oligomers does not play a functional role ., Although the choice of βϵc-c affects the magnitude of the effect shown in Fig 3c , we emphasize that simply allowing chaperone–chaperone binding does not imply that chaperone-only oligomers will be observed at the miscibility limit: there is a large region of parameter space over which this probability is very small ., Furthermore , the results presented in Fig 3 are qualitatively unchanged for all reasonable choices of βϵc-c , i . e . , 0 < β ϵc-c ≲ βϵc-s ., Our simulations thus indicate that the ability to assemble chaperone oligomers affects neither the anti-aggregation function of the chaperones nor their adherence to the proposed design constraints ., We have presented a minimal model of a mixture of passive molecular chaperones and aggregation-prone proteins ., By calculating the limit of thermodynamic stability in this model protein solution , we have shown how passive chaperones that are expressed in sub-stoichiometric ratios with their clients can substantially suppress aggregation ., We have further argued that the biological costs associated with chaperone production and promiscuous , irreversible binding significantly constrain the optimal design of an effective passive chaperone ., We find that if passive chaperones interact promiscuously with exposed hydrophobic surfaces , then the assembly of chaperone oligomers emerges as a nonfunctional side-effect of this thermodynamically optimal design ., Because of the generality of the model , these conclusions are relevant to a broad class of molecular chaperones ., Fully atomistic simulations could provide further information on the parameters governing the interaction strengths between chaperones and their aggregation-prone targets as well as between the passive chaperones themselves ., Such simulations could therefore provide a means of transferring the general thermodynamic principles uncovered by the coarse-grained simulations presented here to detailed models of specific chaperone–client mixtures ., In the lattice model considered here , the limit of thermodynamic stability of a well-mixed solution is encountered at the critical surface for phase separation ., In what follows , we describe the Monte Carlo simulations and finite-size-scaling theory used to calculate points on this critical surface ., Our approach is a generalization of the computational strategy described in detail in Ref . 35 ., In general , the critical surface of a multicomponent mixture has dimension d − 2 , where d is the total number of independent thermodynamic fields 54 ., The independent thermodynamic fields in the present model are the dimensionless chemical potentials of both the chaperones and the clients , βμc and βμs , respectively , as well as the dimensionless interaction energies: βϵ , βϵs-s , βϵc-s and βϵc-c ., The relevant critical surface in this model is thus a 4-dimensional surface ., We perform biased grand-canonical Monte Carlo simulations , as described in Ref . 35 , to collect statistically independent lattice configurations near the critical surface ., We use a L × L × L cubic lattice with periodic boundary conditions and set L = 12 so that all simulations are carried out in the scaling regime ., We then apply the finite-size-scaling theory of Wilding and Bruce 55 , 56 to solve self-consistently for the critical order parameter , M ^ , and the critical orientationally averaged nonspecific energy , βϵ* , at fixed values of βϵs-s , βϵc-s , βϵc-c and xc ., In order to determine each critical point plotted in Fig 2 , we approximate the marginal probability distribution p ( M ) from the grand-canonical samples and then tune this distribution in order to match the known distribution of the critical ordering operator in the three-dimensional Ising universality class , p M . This computational procedure is described below ., In a two-solute solution , with two independent dimensionless chemical potentials βμc and βμs , the critical order parameter must account for fluctuations in the number densities of both the client and chaperone monomers , ρs and ρc , respectively , as well as fluctuations in the internal energy density , u ., The critical fluctuations in the number densities can be described by the vector ν ^ , which indicates the difference in compositions of the two incipient phases 36 ., We therefore define M ^ to be the linear combination, M ^ ≡ ν s ρ s ^ + ν c ρ c ^ - s u ^ , ( 1 ), where both ν ^ and the field-mixing parameter s must be determined self-consistently ., The grand-canonical distribution of M is constructed from the simulation data according to, p gc , k ( M ) ≡ Λ ∑ v w v 1 δ M k ≤ ( ρ s , ρ c , u ) v · M ^ < δ M k + 1 , ( 2 ), where the index v runs over all independent samples and 1{⋅} is the indicator function ., Each sample has a statistical weight wv in the grand-canonical ensemble that depends on the values of the thermodynamic fields 35 ., The system-dependent scaling constant Λ must be determined self-consistently ., The bin size is chosen such that ( δ M k + 1 - δ M k ) = L - 3 , where δ M ≡ Λ ( M - M * ) and M * is the ensemble-averaged mean value of M . We then construct a χ2-function that seeks to minimize the difference between the observed distribution of M and the universal distribution , p M , while obeying the imposed composition constraint:, χ 2 ≡ ∑ k p gc , k ( M ) ( β f → ) - p M ( δ M k / Λ ) 2 σ k 2 + ∑ i ∈ { s , c } ρ i ( β f → ) / ∑ j ∈ { s , c } ρ j ( β f → ) - x i 2 σ i 2 , ( 3 ), where β f → ≡ ( β ϵ , β ϵ s-s , β ϵ c-s , β ϵ c-c , β μ s , β μ c ) and the index k runs over all bins ., In the second term , 〈ρi〉 indicates the ensemble-averaged number density of component i ., We estimate the error in the sampled distribution of M to be, σ k 2 = ∑ v w v 2 1 k , v - ∑ v w v 1 k , v 2 / n samples ∑ v w v , ( 4 ), where 1k , v is the indicator function written out explicitly in Eq ( 2 ) , and we estimate the error in the observed composition at the critical point to be, σ i 2 = 1 ϕ 2 ∑ j , k ∈ { s , c } δ i j - ρ i ϕ 〈 δ ρ j δ ρ k 〉 δ i k - ρ i ϕ , ( 5 ), where 〈δρj δρk〉 ≡ 〈ρj ρk〉 − 〈ρj〉〈ρk〉 , ϕ ≡ ∑j∈{s , c} ρj and δij is the Kronecker delta ., Finally , we calculate the probability of chaperone dimerization , 〈pc-c〉 , directly from the simulation data according to the definition, 〈 p c-c 〉 * ≡ 2 n cc N c * , ( 6 ), where ncc is the number of chaperone–chaperone patch contacts and Nc is the total number of chaperone monomers on the lattice ., In this definition , 〈⋅〉* indicates a grand-canonical average obtained at the critical point with the specified chemical potentials and directional interaction energies . | Introduction, Results, Discussion, Methods | Most proteins must remain soluble in the cytosol in order to perform their biological functions ., To protect against undesired protein aggregation , living cells maintain a population of molecular chaperones that ensure the solubility of the proteome ., Here we report simulations of a lattice model of interacting proteins to understand how low concentrations of passive molecular chaperones , such as small heat-shock proteins , suppress thermodynamic instabilities in protein solutions ., Given fixed concentrations of chaperones and client proteins , the solubility of the proteome can be increased by tuning the chaperone–client binding strength ., Surprisingly , we find that the binding strength that optimizes solubility while preventing irreversible chaperone binding also promotes the formation of weakly bound chaperone oligomers , although the presence of these oligomers does not significantly affect the thermodynamic stability of the solution ., Such oligomers are commonly observed in experiments on small heat-shock proteins , but their connection to the biological function of these chaperones has remained unclear ., Our simulations suggest that this clustering may not have any essential biological function , but rather emerges as a natural side-effect of optimizing the thermodynamic stability of the proteome . | The vast majority of living cells express molecular chaperones that suppress protein aggregation by inhibiting illicit protein–protein interactions ., We refer to this class of chaperones as ‘passive molecular chaperones , ’ since they do not require an external energy source in order to function ., We use simulations of a minimal model of passive chaperones and aggregation-prone client proteins to show how these chaperones increase the solubility of the proteome as a whole ., This anti-aggregation mechanism is surprisingly effective , even when the chaperones are expressed in very low concentrations ., Most importantly , we predict that passive chaperones that are optimized to stabilize the proteome while avoiding irreversible aggregation are likely to cluster in chaperone-only oligomers ., This behavior is not functional per se—that is , it is not required for these chaperones to perform their anti-aggregation function—but nevertheless emerges as a side-effect of this optimization ., Our analysis thus provides an explanation for an unusual behavior that is commonly observed in experiments on passive molecular chaperones . | protein interactions, mathematics, statistics (mathematics), materials science, chaperone proteins, oligomers, thermodynamics, materials by structure, research and analysis methods, proteins, mathematical and statistical techniques, statistical methods, monte carlo method, physics, biochemistry, biochemical simulations, solubility, proteomes, biology and life sciences, physical sciences, material properties, computational biology | null |
journal.pntd.0003004 | 2,014 | Increased Interleukin-17 in Peripheral Blood and Cerebrospinal Fluid of Neurosyphilis Patients | Syphilis , a sexually transmitted multi-stage disease caused by the spirochete Treponema pallidum , remains to be a global public health problem with an estimated 12 million new cases annually 1 ., In recent years , China has experienced a resurgence of syphilis cases , with the national incidence rate of 32 . 04 per 100 , 000 population and with 429 , 677 new cases reported in 2011 2 ., T . pallidum invades the human host through genital or oral mucosa , abraded skin , enters lymphatic system and bloodstream , and then disseminates to different organs ., Without treatment , this spirochetal pathogen is able to survive in the human host for several decades , causing damage in multiple organs including nervous system ( neurosyphilis ) 3 , 4 ., Neurosyphilis is a frequent and protean clinical manifestations ranging from headache and oculopathy to more serious conditions such as cerebrovascular events , paretic and tabes dorsalis 5 ., The mechanisms underlying the development of neurosyphilis remain poorly understood ., T . pallidum can invade the CNS at any stage of infection and provokes robust cellular immune response 6 ., In the non-human primate models , strong T helper ( Th ) 1-type immune response can contribute to the clearance of T . pallidum in CNS 6 ., The immune response elicited during infection , although aimed to eliminate organisms , may also contribute to the pathogenesis ., Cytokines produced by T lymphocytes are critical for regulation of both protective and pathogenic immune responses 7 ., Th17 cells , with the hallmark of producing cytokine IL-17 , were identified as a subset of CD4+ T helper cells ., Emerging evidence has demonstrated that Th17 cells contribute to clearance of diverse organisms ( Mycobacterium tuberculosis , Pneumocystis carinii , Candida albicans and Klebsiella pneumonia et al . ) 8 , 9 , 10 , 11 ., On the other hand , Th17 also mediates strong immunopathology in chronic infection ., Anti-IL-17 and anti-IL-17R treatments could prevent severe Borrelia-induced destructive arthritis 12 ., Hence , Th17 response in infection may be involved in both protection and progression/chronic infection ., Previous studies reported an increase of IL-17 in secondary syphilitic lesion and peripheral blood 13 , 14 ., Recently , Pastuszczak et al . also showed elevated CSF IL-17 levels in early asymptomatic neurosyphilis 15 , suggesting that IL-17 may be involved in local immune response to T . pallidum infection ., In this study , we performed a comparative analysis of Th17/IL-17 in peripheral blood and CSF in asymptomatic and symptomatic neurosyphilis patients , and evaluated the relationship between CSF IL-17 level and the clinical outcomes ., Our results suggested that Th17/IL-17 is a contributing factor to the immunopathology of neurosyphilis , and may be used to monitor the prognosis of treatments of syphilis infected patients ., This study was performed at the Shanghai Skin Disease Hospital between Aug . 2010 and Dec . 2012 ., The study was approved by the Ethics Committee of the Shanghai Skin Disease Hospital ., Written informed consents were obtained from all participants ., Patients were identified and referred for enrollment by dermatologists , neurologists , psychiatrists and ophthalmologists after careful examination and evaluation ., Syphilis was diagnosed at each stage of infection by a combination of compatible history , clinical features and the results of nontreponemal and treponemal tests of serum and CSF samples ., The exclusion criteria include positive HIV infection; prior history of syphilis infection , or history of syphilis treatment ( except for 7 serofast patients ) ; history of systemic inflammation , autoimmune disease , other underlying acute or chronic disease , receiving anti-inflammatory medications , immunocompromised conditions , or use of antibiotics or immunosuppressive medications in the last four weeks ., 70 healthy donors , who visited Shanghai Skin Disease Hospital voluntarily for a medical check-up for the purpose of STD prevention , were recruited to the study ., All healthy subjects were negative for HIV and serological tests for syphilis ( i . e . , both serum RPR and TPPA negative ) , and did not have any clinical symptoms consistent with T . pallidum infection ., In this study , three groups of patients were included: 1 ) neurosyphilis group ( including 40 subjects with asymptomatic neurosyphilis , 4 with meningovasculitis , 39 with general paresis , 8 with tabes dorsalis , and 12 with ocular neurosyphilis ) ; 2 ) non-neurosyphilis group with normal CSF WBC count , CSF protein concentration and CSF-VDRL negative ( including 13 subjects with primary syphilis , 30 with secondary syphilis , 19 with latent syphilis , and 7 with serofast syphilis ) ; 3 ) 70 healthy donors ., In this study , the serofast state must be met the following three criteria:, i ) syphilitic patients , despite receiving recommended standard treatment ( according to Chinese National STI Treatment Guidelines ) , whose nontreponemal test titers ( RPR ) persists positive for at least two years of follow-up evaluation ., ii ) patients who denied high risk sexual behaviors ( re-infection ) following treatments; and, iii ) patients with RPR titers declined fourfold within 6 months after therapy ., Peripheral blood from healthy donors was used for peripheral blood mononuclear cells ( PBMC ) isolation and for measurement of the baseline of the levels of IL-17+ cells and the frequency of Th17 cells ., Since it is difficult to collect CSF from healthy donors , we used a separate control group of 29 patients who underwent orthopaedic or stone surgery ( gall stone , kidney stone ) but were serum RPR and TPPA negative , whose CSF samples were collected prior to spinal anaesthesia ., The baseline level of IL-17 in CSF was determined using samples from the control group ., All groups were well matched in the categories of gender and age ., Additional information on the patient groups were presented in Table 1 . CSF samples were stored at −70°C and thawed once before analyses ., First , all neurosyphilis patients have positive serum RPR and TPPA tests ., The diagnosis of confirmed neurosyphilis includes reactive CSF-VDRL ( Venereal Disease Research Laboratory ) and CSF-TPPA tests in the absence of substantial contamination of CSF with blood ., Presumptive neurosyphilis was defined as a nonreactive CSF-VDRL test but reactive CSF-TPPA with either or both of the following:, ( i ) CSF protein concentration >45 mg/dl or CSF white blood cell ( WBC ) counts≥8/µl in the absence of other known causes for the abnormalities;, ( ii ) clinical neurological or psychiatric manifestations consistent with neurosyphilis without other known causes for such abnormalities 16 , 17 ., Neurosyphilis is categorized as: asymptomatic , meningovascular , paretic , ocular and tabetic neurosyphilis ., Asymptomatic neurosyphilis is defined by the presence of CSF abnormalities consistent with neurosyphilis and the absence of neurological/psychiatrical symptoms or signs ., Meningovasculitis is defined by clinical features of meningitis and magnetic resonance image ( MRI ) evidence of brain lesions and/or a stroke syndrome ., General paresis is characterized by personality changes , dementia and psychiatric symptoms including mania or psychosis ., Patients with sensory loss , ataxia , lancinating pains , bowel and bladder dysfunction were considered as Tabes dorsalis ., Ocular neurosyphilis ( those who had ocular signs or symptoms but with normal CSF index were not included in this study ) is defined by the presence of CSF abnormalities consistent with neurosyphilis and ocular signs or symptoms ( worsening visual acuity and visual fields , floaters , papillitis , uveitis ) ., All these forms of neurosyphilis should have no other known causes for these clinical abnormalities ., A complete list of information of neurosyphilis patients were listed in Table 2 . According to Chinese National STI treatment Guidelines , syphilis patients without CNS involvement were treated with benzathine penicillin 2 . 4MU/qw intramuscular for 1 or 2 weeks for early syphilis , and 3 weeks for late or unknown duration syphilis ., If allergic to penicillin , ceftriaxone 0 . 25 g/day intramuscular for 10 days were given ., Neurosyphilis patients were given aqueous crystalline pencillin G , 4MU intravenously every 4 h for 14 days , or ceftriaxone intravenously with 2 g daily for 10 days if allergic to penicillin ., In the 103 neurosyphilis patients , 80 patients were treated with aqueous crystalline pencillin G , 4MU intravenously every 4 h for 14 days , 23 patients were treated with ceftriaxone intravenously with 2 g daily for 10 days ., All patients were asked for follow-up after treatment ., Patients were selected if the patients written informed consent was obtained ., The exclusion criteria include positive HIV infection; history of syphilis or syphilis treatment; history of systemic inflammatory , autoimmune disease , other underlying acute or chronic disease , patients receiving anti-inflammatory medications , immunocompromised , or using antibiotics or immunosuppressive medications in the last four weeks ., In this study , 44 neurosyphilis patients were enrolled and followed up at Shanghai Skin Disease Hospital ., Patients returned for follow-up visits at 3 , 6 , 9 and 12 months after treatment ., All patients underwent lumbar puncture at the 3-month visit , and lumbar punctures were repeated at 6 , 9 and 12 months if the previous CSF profile was abnormal ., Blood samples were collected at each follow-up visits ., All patients completed their 12 months follow-up visit ., For the CSF-VDRL and the serum RPR test , a 4-fold decrease or more in titer or reversion to a nonreactive result was defined as a normal response ., Stepwise Cox regression models were used to determine the influence of the following factors on the likelihood of normalization of each measure and the improvement of clinical symptoms: ( 1 ) neurosyphilis treatment regimen ( intravenous ceftriaxone , vs . intravenous aqueous penicillin G ) ; ( 2 ) syphilis stage ( secondary and early latent vs . late latent and syphilis of unknown duration ) ; ( 3 ) baseline laboratory values ( greater or less than the median value for those subjects with each abnormality ) ; ( 4 ) CSF IL-17 levels: CSF IL-17 negative ( <0 . 5 pg/ml ) and CSF IL-17 positive ( ≥0 . 5 pg/ml ) ; and ( 6 ) clinical symptoms ., Whole blood samples ( 5 ml ) from syphilis patients and healthy donors were used for peripheral blood mononuclear cells ( PBMC ) isolation ., PBMC were purified from peripheral blood by centrifugation using a Ficoll-Hypaque gradient ( Axis-Shield ) ., Because resting cells do not normally produce cytokines , cells were stimulated in vitro in order for the respective cytokine genes to be activated for intracellular cytokine staining ., Phorbol myristate acetate ( PMA ) and ionomycin are unspecific stimulator that trigger a strong production of cytokines in vitro and are widely used to evaluate intracellular cytokine production from various T lymphocyte subpopulations 18 ., Monensin is used to prevent the intracellular transport of cytokines from Golgi apparatus for enhancing the sensitivity of the detection ., Accordingly , PBMC were seeded into 24-well culture plates ( Corning ) at 2×106 cells/well and stimulated ex vivo with PMA ( 50 ng/ml ) ( Sigma ) and ionomycin ( 1 µg/ml ) ( Sigma ) for 4 hours ., Monensin ( 2 uM ) ( eBioscience ) was then added at the start of stimulation ., CSF ( 10 ml ) was centrifuged at 4°C immediately after spinal tap , and cells were stimulated as described above ., For intracellular staining , cells were first stained with ECD-labeled anti-human CD3 ( Clone UCHT1 , Beckman ) , FITC-labeled anti-human CD4 ( Clone RPA-T4 , Biolegend ) and then fixed and permeabilized using Perm/Fix solution ( Biolegend ) at room temperature for 20 minutes ., Cells were washed with Perm/Wash buffer ( Biolegend ) and stained with PE-labeled anti-human IL-17A ( Clone BL168 , Biolegend ) ., Mouse IgG1 and IgG2 ( BD Biosciences ) were used as isotype controls ., After staining , cells were analyzed with Epics XL ( Beckman Coulter ) and FlowJo software ( Tree Star ) ., Lymphocytes were gated according to forward and side scatter characteristics and CD4+T cells were gated based on CD3 and CD4 expression ., IL-17 positive lymphocytes , CD3+ , CD4+ T cells were defined by setting regions with the lower limits for cytokine positivity determined from isotype antibody ., IL-17 levels in CSF were determined using human IL-17 Quantikine ELISA kits ( eBioscience ) according to manufacturers instruction ., The sensitivity for detecting IL-17 is 0 . 5 pg/ml ., Data were presented as median and range ( min , max ) ., Differences between the groups were analyzed using the nonparametric Mann-Whitney U test ., The detection rates between the groups were assessed using χ2 test or Fishers exact test ., Spearman correlation analysis was performed between the levels of IL-17 and other parameters ., Stepwise Cox regression models were used to determine the influence of the factors on the likelihood of normalization of each measure ., All statistical analyses were performed using SPSS 17 . 0 software ., A value of P<0 . 05 was considered significant ., To investigate the potential role of Th17 in neurosyphilis , we first examined the frequency of IL-17+ among lymphocytes , CD3+ , CD4+ T populations in PBMC ., The baseline frequency of total IL-17+ cells , and IL-17+ CD3+cells , and IL-17+ CD4+ T cells ( Th17 ) of PBMC in healthy individuals were 0 . 86% ( 0 . 19%–1 . 58% ) , 1 . 33% ( 0 . 48%–3 . 2% ) and 1 . 7% ( 0 . 56%–2 . 76% ) , respectively ( Fig . 1A , 1B & 1C ) ., We observed a significant increase in the frequencies of IL17+ , CD3+IL-17+ and Th17 cells in syphilis patients with either non-neurosyphilis or neurosyphilis compared to those in healthy individuals ( Fig . 1A , 1B & 1C ) ., However , there was no significant difference in the frequencies of IL-17+ cells , CD3+IL-17+ and Th17 cells in PBMC between syphilis patients without neurological involvement ( including primary , secondary , latent and serofast syphilis patients ) and neurosyphilis patients ( Fig . 1A , 1B & 1C ) ., To further investigate whether Th17 cells in peripheral blood are different between diverse clinical presentations of neurosyphilis , we divided neurosyphilis patients into two groups , asymptomatic ( n\u200a=\u200a40 ) and symptomatic neurosyphilis patients ( n\u200a=\u200a63 ) ., We then compared the Th17 cell frequency in PBMC between these two groups ., As shown in Fig . 1D , 1E & 1F , patients with symptomatic neurosyphilis had significant higher percentage of total IL-17+cells , CD3+IL-17+ and Th17 in PBMC than the patient group with asymptomatic neurosyphilis ., Since neurosyphilis patients had increased levels of Th17 cells in peripheral blood , we further investigated the IL-17 levels in CSF of these patients ., We first compared the detection rate of IL-17 in CSF between neurosyphilis patients and non-neurosyphilis patients ., We found that there was five-fold higher detection rate of IL-17 in CSF in neurosyphilis patients than that in non-neurosyphilis patients ( Fig . 2A ) ., The average levels of CSF IL-17 was also significantly higher in neurosyphilis patients ( 2 . 29 pg/ml ) ( range of 0–59 . 83 pg/ml ) than that in non-neurosyphilis patients ( 0 pg/ml ) ( range of 0–2 . 60 pg/ml ) ( Fig . 2B ) ., IL-17 was not detected in CSF of the control group ( Fig . 2A & B ) ., We further compared the levels of CSF IL-17 between patients with asymptomatic and symptomatic neurosyphilis ., The detection rates of CSF IL-17 were 47 . 5% and 61 . 9% in asymptomatic and symptomatic neurosyphilis , respectively ., The level of CSF IL-17 in symptomatic neurosyphilis patients ( 4 . 91 pg/ml , range from 0 to 59 . 83 pg/ml ) was significantly higher than that in asymptomatic neurosyphilis patients ( 0 . 715 pg/ml , range from 0 to 44 . 27 pg/ml ) ., Noted that the symptomatic neurosyphilis patient group included meningovascular , paretic , ocular and tabetic neurosyphilis ., Further examination showed that the level of CSF IL-17 was the highest among paretic patients ( 7 . 6 pg/ml , range from 0 to 38 . 07 pg/ml ) ( Table 3 ) ., T . pallidum is capable of invading central nervous system and damaging local tissues ., There are detectable CSF abnormalities in neurosyphilis patients , including positive CSF VDRL , pleocytosis , and/or elevated protein concentration 19 ., These measurements correlate well with the disease activity 19 ., Since the above data showed that neurosyphilis patients had increased CSF IL-17 levels , we further investigated a possible relationship between CSF IL-17 levels and other measurements ., As shown in Fig . 3 , there was a significant positive correlation between CSF IL-17 levels and CSF protein concentrations or CSF VDRL titer , but not with the CSF WBC counts ., In some neurosyphilis patients , CSF IL-17 was not detected ., We thus further investigated whether there are certain factors which may contribute to this phenomenon ., We found that there were no differences in age , sex , the baseline serum RPR titer , or duration of symptoms prior to diagnosis between the IL-17 positive and IL-17 negative neurosyphilis groups ., However , the IL-17 positive group had higher CSF protein concentration and CSF VDRL titer and higher frequency of symptomatic neurosyphilis than that of the IL-17 negative group ( Table 4 ) ., We further analyzed IL-17-producing cells in CSF of neurosyphilis patients ., Because of the limited sample sizes and lymphocyte cell numbers in CSF collected from neurosyphilis patients , CSF cells from 14 neurosyphilis patients who had high levels of CSF pleocytosis ( >50 cells/ul ) were collected and stimulated for intracellular staining for the purpose of this study ., 14 subjects included 6 ( 42 . 9% ) subjects with asymptomatic neurosyphilis , 5 ( 35 . 7% ) subjects with paretic , 2 ( 14 . 3% ) subjects with ocular neurosyphilis , 1 ( 7 . 14% ) subjects with meningovascular neurosyphilis ., The average percentage of CD4+ T cells was 58 . 28% ( 51 . 65%–80 . 1% ) of total lymphocytes , and Th17 ( CD3+CD4+IL-17+ cells ) was 1 . 8% ( 0 . 25%–4 . 6% ) ( Fig . 4 ) ., However , Th17 cells accounted for 88 . 8% ( 45 . 1%–100% ) of total IL-17+ cells ( Fig . 4 ) , indicating that Th17 is the dominant IL-17-producing cells and may play an important role in neurosyphilis ., Among 44 subjects with confirmed neurosyphilis , 22 ( 50% ) subjects were asymptomatic neurosyphilis , 15 ( 34 . 1% ) were paretic neurosyphilis , 5 ( 11 . 4% ) were ocular neurosyphilis , 2 ( 4 . 5% ) were meningovascular neurosyphilis ., All enrolled patients were routinely under followed-up examination and treated with standard therapy according to the Chinese treatment Guidelines ., Factors that were included in the final regression models of normalization of each laboratory measure , the improvement of clinical symptoms and their HRs are shown in Table 5 ., Factors that may improve laboratory markers or clinical symptoms were analyzed using the final regression model ( Table 5 ) ., The neurosyphilis treatment regimen did not influence normalization of any of the 4 laboratory markers and the improvement of clinical symptoms ., Normalization of the CSF protein concentration was more likely to occur in subjects with early syphilis ( p\u200a=\u200a0 . 018 ) ., CSF-VDRL reactivity was less likely to become normal in patients with positive CSF IL-17 ( p\u200a=\u200a0 . 04 ) and with higher baseline CSF-VDRL titer ( p\u200a=\u200a0 . 019 ) ., Serum RPR reactivity was more likely to return to normal in subjects with higher baseline serum RPR titers ( p\u200a=\u200a0 . 008 ) ., T . pallidum remains one of the human pathogens that cannot be cultivated in vitro to-date ., A suitable animal model for studying the pathogenesis of syphilis is also lacking ., These obstacles have greatly hindered the effort of elucidating the basic immunobiological traits of syphilis ., As a consequence , little is known about how T . pallidum causes damage to the central nervous system ., IL-17 , a potent proinflammatory cytokine , plays a key role in the induction and development of tissue injury ., IL-17 results in an increased production of ICAM-1 , IL-6 and IL-8 , and an increased synergy of many effects of IL-1β and TNF-α , which enhances the local inflammation and leads to inflammatory destruction 20 , 21 ., In this study , we observed an elevated CSF IL-17 in neurosyphilis patients ., A similar scenario has been observed in infectious and auto-immune CNS disorders 22 , 23 ., Furthermore , we found that the level of CSF IL-17 is positively associated with CSF VDRL titer and total CSF protein in neurosyphilis patients ., These findings suggest that IL-17 may involve in the CNS damage in neurosyphilis patients ., Syphilis is known as a “great imitator” because it is protean , especially neurosyphilis ., Based on the patients clinical and laboratory features , neurosyphilis is divided into five diagnostic categories , including asymptomatic , meningitis , meningovascular , paretic , and tabetic neurosyphilis 5 ., Meningitis involves diffuse inflammation of the meninges with signs and symptoms of meningitis including headache , photophobia , nausea , vomiting , cranial nerve palsies , and occasionally seizures ., It is diagnosed within 12 months after T . pallidum infection but it is relatively rare 5 ., Unfortunately , no meningitis neurosyphilis patients enrolled in this study , and thus , the involvement of IL-17 in this stage of neurosyphilis remains unknown ., The pathogenic mechanisms underlying different clinical presentations of neurosyphilis are largely unknown ., In this study , we found that higher levels of IL-17 were observed in CSF of symptomatic neurosyphilis patients , especially in paretic patients , compared with asymptomatic neurosyphilis patients ., Moreover , the higher CSF protein and VDRL titer were observed in symptomatic neurosyphilis patients ., These results suggest that IL-17 may be associated with clinical symptoms in neurosyphilis patients ., Asymptomatic neurosyphilis does occur in both early and latent stages of syphilis ., It is reported that there was an elevated CSF IL-17 level in early asymptomatic neurosyphilis patients , which correlated with the extent of CSF abnormality 15 ., In our study , CSF IL-17 could be detected in 66 . 7% ( 12/18 ) of early asymptomatic neurosyphilis patients ., It is believed that asymptomatic neurosyphilis patients may develop to late neurological complications 24 ., Moreover , the extent of abnormalities of CSF positively correlated with the probability of developing late neurological complications 25 ., Based on these notions , we suggested that some early asymptomatic neurosyphilis patients might have persistent IL-17 inflammation response , which could damage the CNS , resulting in neurological symptoms ., Regrettably , there has been no study to compare long-term outcomes between CSF IL-17 negative and positive asymptomatic neurosyphilis patients ., Pastuszczak et al . identified that there was a strong correlation between CSF IL-17 and CSF pleocytosis in early asymptomatic neurosyphilis patients 15 ., But our results indicated that there was no correlated between the level of CSF IL-17 and CSF pleocytosis ., The CSF pleocytosis in neurosyphilis was related to the syphilis stage besides to the CNS damage ., There was a marked pleocytosis in patients with acute meningeal syphilis ., In late stage , CSF WBC counts in some neurosyphilis patients were less or even normal and were inconsistent with clinical symptoms ., In up to 10% of patients with tabes referred to as the “burned out” stage , the CSF cell count may be normal 5 ., In our study , there were early and late stage neurosyphilis patients ., There were some paresis patients the CSF WBC counts were normal , though the clinical manifestations were severe ., Therefore , according to the data , the CSF WBC counts were not always correlated with the degree of CNS damage ., The different stage patients were enrolled in our study , leading to be inconsistent with the previous results ., Besides CD4+ T cells , other cells are capable of secreting IL-17 26 ., It was previously shown that IL-17 is mainly secreted by CD4+ T cells ( Th17 ) in CSF in patients with chronic inflammatory demyelinating polyradiculoneuropathy ( CIDP ) 27 ., In this study , we observed the CD4+ T cells were accumulated in CSF in neurosyphilis patients , and they were the dominant IL-17-producing cells ., This finding suggests that Th17 response is a part of the local CNS response in a sub-population of neurosyphilis patients ., Our results showed that Th17 cells were increased in CSF of neurosyphilis patients ., The mechanisms underlying the increase of Th17 in CNS remain unclear ., IL-17 can disrupt the tight junction molecules and activates the endothelial contractile machinery , leading to disruption of blood brain barrier ( BBB ) 28 ., Thus , Th17 in CSF may be a consequence of passive diffusion from blood ., On the other hand , microbial lipopeptides such as Helicobacter pylori HP-NAP and B . burgdorferi NapA , can induce Th17 differentiation and production of IL-17 29 , 30 ., In this regard , T . pallidum TpF1 is a protein homolog of HP-NAP and NapA 31 , which may be capable of promoting Th17 differentiation and expansion in CNS ., Interestingly , recent data showed that TpF1 can stimulate Treg cell differentiation 32 ., The mechanisms underlying the increase of Th17 in CNS in neurosyphilis need to be further elucidated ., Because Th17 response may induce the immune-mediated CNS injury , we further evaluated the relationship between IL-17 and the clinical outcome of neurosyphilis ., The baseline CSF-VDRL titer , and serum RPR titer influenced the likelihood of normalization of each parameter , consistent with previous studies 33 ., However , we observed that CSF IL-17 positive neurosyphilis patients were 2 . 43 times less likely to normalize CSF-VDRL reactivity , even after taking into account the baseline CSF-VDRL titer and the stage of syphilis ., CSF VDRL titer may reflect the T . pallidum burden and the extent of CNS damage ., T . pallidum invaded CNS can induce Th17 immune response and CSF IL-17 were positively correlated with CSF VDRL titer ., So positive CSF IL-17 in patients may reflect higher number of T . pallidum spirochetes in CSF ., Since longer time would be required to clear a higher number of T . pallidum burden , the likelihood of normalization of CSF VDRL reactivity at the end of the observation would be lower ., Because the sample size is limited ( the number of subjects included in the regression analyses was only ranged from 21 to 44 patients in this study ) , a large sample study is needed to further understanding the true immune response at different stages of neurosyphilis and its clinical significance ., In conclusion , our findings demonstrate that neurological damage in syphilis patients is associated with increased CSF Th17/IL-17 response ., CSF IL-17 may be used to evaluate the clinical outcome of treatment of neurosyphilis . | Introduction, Methods, Results, Discussion | Treponema pallidum infection evokes vigorous immune responses , resulting in tissue damage ., Several studies have demonstrated that IL-17 may be involved in the pathogenesis of syphilis ., However , the role of Th17 response in neurosyphilis remains unclear ., In this study , Th17 in peripheral blood from 103 neurosyphilis patients , 69 syphilis patients without neurological involvement , and 70 healthy donors were analyzed by flow cytometry ., The level of IL-17 in cerebrospinal fluid ( CSF ) was quantified by ELISA ., One-year follow up for 44 neurosyphilis patients was further monitored to investigate the role of Th17/IL-17 in neurosyphilis ., We found that the frequency of Th17 cells was significantly increased in peripheral blood of patients with neurosyphilis , in comparison to healthy donors ., IL-17 in CSF were detected from 55 . 3% neurosyphilis patients ( in average of 2 . 29 ( 0–59 . 83 ) pg/ml ) , especially in those with symptomatic neurosyphilis ( 61 . 9% ) ., CSF IL-17 was predominantly derived from Th17 cells in neurosyphilis patients ., Levels of IL-17 in CSF of neurosyphilis patients were positively associated with total CSF protein levels and CSF VDRL ( Venereal Disease Research Laboratory ) titers ., Notably , neurosyphilis patients with undetectable CSF IL-17 were more likely to confer to CSF VDRL negative after treatment ., These findings indicate that Th17 response may be involved in central nervous system damage and associated with clinical symptoms in neurosyphilis patients ., Th17/IL-17 may be used as an alternative surrogate marker for assessing the efficacy of clinical treatment of neurosyphilis patients . | Syphilis , caused by the bacterium Treponema pallidum , can progress to affect the central nervous system ( CNS ) and cause damage in the brain and spinal cord , which is called neurosyphilis ., While many neurosyphilis patients may not have any symptoms , some patients develop severe symptoms which can be life-threatening ., Th17 cells are a subset of CD4+ T helper cells producing the hallmark cytokine IL-17 , which are essential for effective antimicrobial host defense and are also involved in tissue damage ., In this study we conduct a comparative analysis of Th17/IL-17 in the peripheral blood and cerebrospinal fluid ( CSF ) of syphilis patients without neurological abnormalities , and neurosyphilis patients with or without symptoms ., Th17 frequency in peripheral blood was significantly increased in neurosyphilis ., CSF IL-17 was increased in neurosyphilis patients , especially in those with symptomatic neurosyphilis ., Levels of CSF IL-17 in neurosyphilis patients were positively associated with CNS damage ., Notably , neurosyphilis patients with undetectable CSF IL-17 had better outcome upon treatment ., These findings indicate that the Th17 response may be involved in central nervous system damage , clinical symptoms and prognosis of treatment of neurosyphilis patients . | bacterial diseases, infectious diseases, infectious diseases of the nervous system, medicine and health sciences, neurology | null |
journal.pntd.0004229 | 2,015 | Quality of Sterile Male Tsetse after Long Distance Transport as Chilled, Irradiated Pupae | Tsetse flies are hematophageous insects found in sub-Saharan Africa and are the main vectors of trypanosomes , the causative agents of African Animal Trypanosomosis ( AAT ) and Human African Trypanosomosis ( HAT ) 1 ., The debilitating disease AAT limits the exploitation of fertile land for agricultural activities in an area of 10 million km2 2 and is considered the main constraint to more productive livestock systems in sub-Saharan Africa 3 , 4 ., The direct annual production losses of cattle in terms of decreased meat and milk production , abortions , etc . are estimated at USD 600–1200 million 5 and the overall annual losses in livestock and crop production have been estimated as high as USD 4750 million 6 ., To suppress or eradicate these disease vectors , four methods that are environmentally and economically acceptable can be used in a context of area-wide integrated pest management ( AW-IPM ) approaches 4 , 7 , 8 i . e . the sequential aerosol technique ( SAT ) 9 , 10 , the deployment of insecticide-impregnated traps/targets 11 , the live-bait technology 12 and the sterile insect technique ( SIT ) 13 , 14 ., The SIT is used throughout the world to suppress , eradicate , contain or prevent the introduction of several insect pests such as fruit flies 15 , moths 16 , screwworm flies 17–19 , mosquitoes 20 and tsetse flies 14 ., The effectiveness of the SIT to eradicate tsetse fly populations was demonstrated in Nigeria with Glossina palpalis palpalis 21 , in Burkina Faso with G . palpalis gambiensis , G . tachinoides and G . morsitans submorsitans 13 , 22 and on Unguja Island , Zanzibar with G . austeni 14 ., In Senegal , a programme is underway to eradicate a G . p ., gambiensis population from the Niayes area 23–27 ., This campaign is part of the Pan-African Tsetse and Trypanosomosis Eradication Campaign ( PATTEC ) , an initiative of the African Heads of State and Government to ensure increased food security through better management of the tsetse fly and trypanosomosis problem 28 ., The data of the feasibility study ( 2007–2010 ) indicated the potential to create a sustainable zone free of G . p ., gambiensis in the Niayes 24 , 29 , and therefore , the Government of Senegal opted for an AW-IPM approach that included an SIT component ., An agreement was made between the Government of Senegal and the Centre International de Recherche-Développement sur l’Elevage en zone Subhumide ( CIRDES ) in Bobo-Dioulasso , Burkina Faso and the Slovak Academy of Sciences ( SAS ) in Bratislava , Slovakia to produce the sterile flies for the eradication campaign in Senegal ., The male flies were transported as chilled pupae to Dakar where they could emerged under standard conditions 27 ., In AW-IPM programmes that have an SIT component , the quality of the released sterile males remains one of the most crucial prerequisites for success , as flies of low quality ( i . e . low survival rate and/or deformed wings ) can’t compete with wild males in the field to mate with females and induce sterility in the native population 8 , 30 , 31 ., Therefore , routine quality control procedures are crucial for the SIT component to identify weaknesses in production or handling procedures that result in low quality of the sterile males which may lead to potential failure of these programmes 32 ., In past tsetse eradication campaigns 13 , 14 , 21 , 22 the rearing facility and target area were not far apart , so there was no need for pupal shipments ., As an example , in the eradication programme on Unguja Island , Zanzibar , the sterile male flies were produced in Tanga , mainland Tanzania and the sterile adult flies were collected twice a week with light aircraft and released from the air in biodegradable cartons ., A quality control system was implemented that consisted of taking one release carton before loading the aircraft in Tanga and one carton during the release flights ., Both in Tanga and Unguja , the flies were released in a specially designed release arena and the following quality parameters assessed: number of flies in the box , mortality , number of non-flyers , sexing error , and feeding status 33 ., In this study , we developed and validated a quality control protocol to assess the quality of male G . p ., gambiensis that were irradiated and shipped as pupae under chilled conditions ., Four biological parameters were measured:, i ) adult emergence ,, ii ) percentage of flies with deformed wings ,, iii ) flight ability of the sterile flies in the insectary and in the field and, iv ) survival of the flyers ( those that were capable of flying out of the cylinder in the insectary ) under stress conditions ., These parameters were used to assess the reliability of this quality protocol to, 1 ) predict field performance of the flies ,, 2 ) monitor and compare the performance of flies from two locations with different treatment protocols , and, 3 ) develop quality criteria for use in feedback mechanisms to improve rearing systems ., The study was carried out at the Institut Sénégalais de Recherches Agricoles / Laboratoire National de l’Elevage et de Recherches Vétérinaires , Service Bio-Ecologie et de Pathologies Parasitaires ( ISRA/LNERV/BEPP ) in Dakar ., Insectary conditions were 24–25°C , 75 ± 5% RH and 12:12 light:dark photoperiod for emergence and the monitoring of the flies ., Male G . p ., gambiensis pupae from colonies kept at Burkina Faso and Slovakia were irradiated under chilled ( 4–6°C ) conditions to lower their metabolic rate to prevent emergence 27 , 34 ., The SAS pupae were irradiated in a Gammacell 220 ( MDS Nordion , Ottawa , Canada ) ( dose rate of 3 . 11 Gy . sec-1 on 1 May 2012 and 2 . 19 Gy . sec-1 on 1 January 2015 ) or in an X-ray irradiator ( Radsource 2400 ) ( dose rate of 14 . 30 Gy . min-1 ) located at the FAO/IAEA Insect Pest Control Laboratory , Seibersdorf , Austria ., The CIRDES pupae were irradiated in a 137Cs source for 24 minutes 30 seconds to give a dose of 110 Gy ., The male pupae were packaged in cartons ( for SAS ) and in petri dishes ( CIRDES ) that were placed in insulated transport boxes containing phase change material packs ( S8 ) ( PCM Phase Change Material Products Limited , Cambridgeshire , United Kingdom ) to maintain the temperature at 8–10°C and shipped to Dakar by commercial aircraft 27 ., The study was implemented from May 2012 to January 2015 ., A shipment of CIRDES and SAS pupae was received every week at the ISRA in Dakar ., Each consignment contained two batches ( 1 and 2 ) of pupae that had a different larviposition date and consequently had been exposed to a different chilling period before shipping , i . e . batch 1 was chilled at 8°C for one day longer than batch 2 in the source insectary before transport ., Each batch contained an average of 2500 pupae ., A total of 50 pupae were sampled from each batch for the quality control test ( QC ) and the remaining pupae were emerged to be released in the operational eradication programme , i . e . the flies destined for release in the programme ( RF ) ., The pupae of the QC and RF groups were kept under the same environmental conditions ( 24–25°C , 75 ± 5% RH and a photoperiod of L:D 12:12 h ) ., The 50 pupae for the QC group were selected to assess whether a small sample of each received pupae consignment was adequate to predict the quality of the shipped and released flies ( RF ) ., Pupae from the RF group were placed in Petri dishes under ~1cm of sand mixed with a fluorescent dye ( DayGlo ) ( 0 . 5g dye/200g of sand ) , to mimic the natural emergence conditions in the soil ( Fig 1A ) and to allow discrimination from wild flies in the monitoring traps as these sterile male flies were released in the operational programme ., Emerged flies were sorted and classed as “normal” ( flies with no apparent morphological deficiencies ) and “abnormal flies” ( i . e . with deformed wings ) ., Normal flies were offered at least three bovine blood meals ( originating from slaughterhouse of Dakar , with the consent from the slaughterhouse to obtain the blood samples from livestock ) containing 10 mg of the trypanocidal drug isometamidium per litre of blood using the in vitro silicon membrane feeding system before being transported to the field for release ., The trypanocidal drug prevents the cyclical development of trypanosomes in the released sterile males 35–37 ., Irradiated and marked males were transported by car to the release sites ( ~ 1 hour for Diacksao Peulh and Kayar and 10 minutes for the Parc de Hann 24 ) in Roubaud-type cages ( 4 . 5 x 13 x 8 cm ) that were covered with netting with a mesh size of 1 mm x 1 mm , each containing on average 120 sterile males ., Cages were kept in climate controlled containers ( temperature and humidity of 24–26°C , 75 ± 5% respectively ) during the transport and temperature and humidity were recorded every 5 minutes with a Hobo data logger ., Flies were released every Friday afternoon between 16:00 and 18:00 h over a white cloth ( 2 x 1 . 5 m ) ., Males remaining on the cloth after 5 minutes were counted and considered as non-flyers ., Ground releases of these flies took place from May 2012 to March 2013 ., Thereafter , all sterile male flies were released by air ., The pupae of the QC group were kept under the same conditions as the RF group but the Petri dishes with the pupae were put in a flight cylinder , i . e . a PVC tube 10 cm high and 8 . 4 cm in diameter ( Fig 1B ) ., The inner wall of the cylinder was coated with unscented talcum powder to prevent the flies from crawling out ., This method was initially developed for routine quality assessments of sterile fruit flies 38 , 39 and moths 40 , and adapted here to tsetse flies ., This protocol gave an indication of the propensity of the sterile male flies to fly out of the cylinder and only those flies that managed to escape the flight cylinder after emergence were considered as “available for the SIT” ., Flies with deformed wings and those with normal wings but unable to escape the flight tube were counted , as well as the number of pupae that did not emerge ., The survival of the sterile males of the QC group that escaped the flight cylinder was assessed under stress conditions ( no food ) ., Every morning ( except Sundays ) , the emerged flies were collected and transferred to standard fly holding ( 10 . 3 cm diameter and 4 . 5 cm high ) cages ( Fig 1C ) ., The flies emerged on a given day were pooled in one cage ., Dead flies were counted daily and removed from the cages ., The data sets ( both QC and RF groups ) on percentage emergence , flies with deformed wings and flight ability were each divided into training and test sets ., The training set was used to build the model and the test set to measure its performance 41 ., For the data on emergence and percentage of flies with deformed wings , 60% of the entire data set ( n = 364 rows ) , selected at random , was used as a training set and the remaining as the test set ., For the flight ability , 75% of the entire data set ( n = 80 rows ) was used for the training set and 25% for the test set ., The difference in the proportion of data used for the training set in the first and second cases was related to sample size ., A binomial linear mixed effect model was used to analyze emergence rates ., The emergence rate measured within the QC group , the origin of the pupae ( CIRDES and SAS ) , the batches ( 1 and 2 ) and their second and third order interactions were used as explanatory variables and the emergence rate of the RF group as the response variable ., The shipment date was considered as a random effect ., The best model was selected on the basis of the lowest corrected Akaike information criterion ( AICc ) , and the significance of fixed effects was tested using the likelyhood ratio test 42 , 43 ., The R2 ( coefficient of determination ) was used to describe the proportion of variance explained by the model for the training and test data sets 44 , 45 ., The same analysis was used for the percentage of flies with deformed wings and the percentage of flyers ., Flight ability was analyzed between QC and RF groups using only the CIRDES data sets because field data were not available for the SAS shipments ., Flight ability was compared among years ( 2012 , 2013 and 2014 ) using the same binomial model ., The survival of the sterile males of the QC group that had escaped from the flight cylinder and kept under starvation was analyzed using Kaplan-Meier survival curves ., Survival curves were compared between origins ( CIRDES and SAS ) , batches ( 1 and 2 ) and years using the coxph model 46 ., The median survival was considered to be the average probable survival of the studied flies ., The R Software ( version 3 . 1 . 0 ) was used for all statistical analyzes 47 ., The complete data sets are available in S1 and S2 and S3 Tables ., The study was conducted in the framework of the tsetse eradication campaign in Senegal , led by the Directorate of Veterinary Services , Ministry of Livestock and the ISRA/LNERV , Ministry of Agriculture and Rural Equipment ., This project received official approval from the Ministry of Environment of Senegal , under the permit N°0874/MEPN/DE/DEIE/mbf ., A total of 1 , 581 , 366 irradiated pupae were used for this study of which 1 , 271 , 121 ( 80 . 4% ) originated from the CIRDES insectary ( 123 shipments ) and 310 , 245 ( 19 . 6% ) pupae originated from the SAS insectary ( 53 shipments ) ., The emergence rate of pupae of the RF group was significantly greater for shipments originating from CIRDES than those from SAS ( P < 10−3; Table 1 ) , as well as for batch 2 pupae than batch 1 pupae regardless of the origin of pupae ( P < 10−3; Table 1 ) ., The percentage of flies with deformed wings was significantly lower for the flies that originated from the CIRDES than the SAS flies and for batch 2 pupae than batch 1 pupae regardless of the origin ( P < 10−3; Table 1 ) ., The flight ability of the CIRDES flies in the field was significantly better for flies derived from batch 2 pupae than batch 1 pupae ( P < 10−3; Table 1 ) ., Adult emergence of pupae of the QC group was similar between origins ( P = 0 . 8 ) but differed between batches regardless of the origin ( P < 10−3; Table 1 ) ., The percentage of flies with deformed wings that emerged from the SAS pupae was significantly lower than that for the CIRDES pupae ( P < 10−3; Table 1 ) ., It was similar between batches for CIRDES and different for SAS ( P < 10−3; Table 1 ) ., The flight ability was similar between batches ( P > 0 . 05; Table 1 ) ., The comparison of the different parameters between the QC and the RF groups showed that the emergence rates were similar for the CIRDES flies while they were significantly greater in the QC group of the SAS flies ( P < 10−3; Table 1 ) ., The percentage of flies with deformed wings was lower in the RF group as compared with the QC group for the CIRDES pupae , whereas it was the opposite for the SAS pupae ( P < 10−3; Table 1 ) ., The percentage of the CIRDES flies escaping the flight cylinder in the insectary was significantly lower than the percentage of flies taking off in the field after the release ( P < 10−3 ) i . e . 34 . 9 ± 17 . 8% and 55 . 1 ± 13 . 5% , respectively ., The flight ability of batch 2 flies of the RF group was significantly greater than for batch 1 flies ( P < 10−3 ) whereas no difference was observed between batches of the QC group ( P = 0 . 6; Table 1 ) ., The predicted probabilities of occurrence using the QC data allowed us to predict the results observed in the RF group with good accuracy: the emergence rates , percentage of flies with deformed wings and flight ability were strongly correlated to predictions of the training data set ( P < 10−3; R2 of 0 . 90 , 0 . 94 and 0 . 95 respectively; Fig 2 ) ., For the test data set , the model predicted 55% , 53% and 45% of the variances respectively ( P < 10−3 , Fig 2 ) ., Survival curves of QC flies kept under starvation are presented by batch and origin in Fig 3 ., Flies from batch 2 pupae survived significantly longer than those from batch 1 pupae for CIRDES ( P = 0 . 01 ) whereas for SAS , batch 2 flies survived marginally longer than from batch 1 ( P = 0 . 09 ) shipments ., The CIRDES flies survived marginally longer than the SAS flies ( P = 0 . 06 ) ., The median survival was 6 days regardless of the batch and origin of pupae ( Fig 3 ) ., The maximum survival observed was 12 days after emergence for the CIRDES ( batch 2 ) and 10 days for the SAS ( batch 2 ) flies ., From 2012 to 2014 , the percentage of QC flies escaping the cylinder gradually increased regardless of the origin of pupae ( P < 10−3; Table 2 ) ., Flies lived significantly longer in the survival tests in 2013 and 2014 as compared with 2012 for both CIRDES and SAS flies ( P < 10−3 ) ., Thus , the quality of sterile male flies ( flight ability and survival ) was significantly improved among years and these improvements were more prominent for flies from SAS ( Table 2 ) ., The quality protocol implemented in this study was developed for a programme that required long distance transport of chilled male tsetse pupae and was shown to be a good proxy for the insectary rearing output ., Indeed , the emergence rates , percentage of flies with deformed wings and flyers from the QC and RF groups were highly correlated ., Overall , these results highlight that the quality protocol procedures had no negative impact on adult emergence and predicted well the amount of sterile males available for the SIT component ., In eradication programmes such as the one implemented in the Niayes of Senegal , thousands of sterile male flies need to be processed weekly for release requiring many preliminary activities in the insectary ( to separate and to count normal and abnormal flies after emergence , assess mortality rate and percentage of non-flyers after release in the field ) ., With the results obtained from the QC group , it was shown that all these parameters predicted well the biological quality of the sterile male flies , which will reduce considerably the work load ., More importantly , multiple handling of flies ( generally at 2–4°C for the sorting ) generates stress which reduces their quality which can be avoided using a sample for the quality control test 27 ., Quality control protocols for SIT programs were initially developed for fruit flies , especially the Mediterranean fruit fly Ceratitis capitata and has more recently been extended to Anastrepha and Bactrocera fruit fly species 38 , 39 ., For these insects , the average flight ability after irradiation and transport of pupae was 65% for C . capitata , 75% for Anastrepha suspensa and 55% for Bactrocera oleae 38 , 39 , 48 ., The flight ability obtained in the present study with G . p ., gambiensis , BKF strain was on average 35 . 8 ± 18 . 4% ., Although caution is required when comparing data from different species and when pupae were shipped under different conditions , it provides an indication that our results with G . p ., gambiensis were rather low ., This low propensity to fly could be due to mechanical shocks and vibrations that were absorbed by the pupae during transport or possibly different handling procedures in the different insectaries ., In addition , the length of the cooling period of the pupae seems to be an important quality reducing factor , especially in terms of emergence rate ., The impact of these different variables on emergence of adults was shown before 27 ., Adults emergence may also be affected by excessive temperatures or inappropriate relative humidity during the rearing process 30 ., In addition , it is well established that irradiation could potentially lower the quality of the produced flies especially when the irradiation dose that is required to obtain 95–100% sterility is high and therefore results in severe somatic damage 30 , 49–51 ., The released sterile males must be active to find a blood meal , shelter and to compete with wild males for mating with wild females and successfully transfer the sterile sperm , and they must survive long enough to be able to find the virgin females 30 ., Data of the QC group indicated that about 20% of the flies that emerged were “normal-looking” flies that had their wings deployed but did not escape the flight cylinder ., What is measured here is the propensity of the flies to fly i . e . some of those flies that stayed in the cylinder probably can fly , but for one or the other reason they don’t ., This was confirmed by the data from the field in that most of these flies were able to take off from the release cloth; however , they still might be poor flyers ( but this was not assessed in this study ) ., Indeed , after the preliminary sorting at the insectary , all normal-looking flies ( i . e those that were mobile and had deployed wings ) were transported to the field and released using the ground release protocol where the flies were released on a cloth ( 2 x 1 . 5 m ) and checked after 5 minutes ., These observations confirm the necessity to implement a quality control protocol for sterile males to make eradication campaigns more effective ., Weekly data on the percentage of released sterile males as compared to the number of shipped pupae allows for crucial feedback information to the rearing facility and to better plan the operational phase of the SIT component of AW-IPM programmes 4 ., For example , by improving the packaging and transport protocols ( such as the use of cotton for the CIRDES pupae and sawdust/vermiculite for the SAS pupae to cushion the mechanical shocks ) flight ability was increased significantly reaching 55% in 2014 ., There was no difference between the survival of the sterile males that emerged from the CIRDES and the SAS pupae ., This indicates that the quality of the blood diet and the performance of the females in the colonies of the two rearing facilities were equivalent ., The sterile males did not receive any blood meals during the survival experiment , and hence , their survival depended only on the fat reserves acquired during larval development ., As tsetse reproduce by adenotrophic viviparity 4 , these fat reserves are closely linked to the quality of the blood meals that are taken by the female parents ., Under these conditions , the median survival of sterile males was 6 days regardless of the origin of the pupae with more than 80% and 10% surviving until 4 and 8 days after emergence , respectively ., These results were similar to thoses observed with G . pallidipes , i . e . 90% of G . pallidipes males that emerged from pupae that had been exposed to a low temperature of 15°C survived unfed until 4 days but less than 10% survived after 8 days 52 ., In order to simulate the proposed use of the chilled adult release system for area-wide tsetse SIT , the tenerale male flies of the same tsetse species exposed to a temperature of 7°C for 48 and 72 hours followed by 6 hours at 4°C and monitored without being offered a blood meal showed a median survival of 4 days 52 ., The flies that emerged from batch 2 pupae survived on average longer than those emerging from batch 1 pupae indicating that the duration of the chilling at 8°C had a negative impact on fly quality ., The median survival of the sterile males of 6 days without food as observed under laboratory conditions is encouraging , as the sterile males that are destined for release are being offered a blood meal at least three times before being released ., This will undoubtedly increase their fat reserves thus enhancing their survival until they have found a host and hence , their competitiveness ., In conclusion , although the quality protocol data indicated that the percentage of flyers was less than 40% , the quality of the transported sterile pupae improved with time ., More importantly , the data from the field indicate that the competitiveness of those male flies that were released was very good 53 resulting in excellent progress in the eradication campaign 54–56 ., Research continues to improve the transport conditions of the pupae to potentially further increase the proportion of flyers . | Introduction, Methods, Results, Discussion | Tsetse flies transmit trypanosomes that cause human and African animal trypanosomosis , a debilitating disease of humans ( sleeping sickness ) and livestock ( nagana ) ., An area-wide integrated pest management campaign against Glossina palpalis gambiensis has been implemented in Senegal since 2010 that includes a sterile insect technique ( SIT ) component ., The SIT can only be successful when the sterile males that are destined for release have a flight ability , survival and competitiveness that are as close as possible to that of their wild male counterparts ., Tests were developed to assess the quality of G . p ., gambiensis males that emerged from pupae that were produced and irradiated in Burkina Faso and Slovakia ( irradiation done in Seibersdorf , Austria ) and transported weekly under chilled conditions to Dakar , Senegal ., For each consignment a sample of 50 pupae was used for a quality control test ( QC group ) ., To assess flight ability , the pupae were put in a cylinder filtering emerged flies that were able to escape the cylinder ., The survival of these flyers was thereafter monitored under stress conditions ( without feeding ) ., Remaining pupae were emerged and released in the target area of the eradication programme ( RF group ) ., The following parameter values were obtained for the QC flies: average emergence rate more than 69% , median survival of 6 days , and average flight ability of more than 35% ., The quality protocol was a good proxy of fly quality , explaining a large part of the variances of the examined parameters ., The quality protocol described here will allow the accurate monitoring of the quality of shipped sterile male tsetse used in operational eradication programmes in the framework of the Pan-African Tsetse and Trypanosomosis Eradication Campaign . | An area-wide integrated pest management campaign against Glossina palpalis gambiensis has been implemented in Senegal since 2010 that includes a sterile insect technique component ., The sterile males used for the releases emerged from pupae that were produced and irradiated in Burkina Faso and Slovakia ( irradiation done in Seibersdorf , Austria ) and transported weekly under chilled conditions to Dakar , Senegal ., Tests were developed to assess the quality ( flight ability and survival ) of sterile males ., To assess flight ability , for each consignment a sample of 50 pupae ( QC flies ) was put in a cylinder filtering emerged flies that were able to escape the cylinder ., The survival of these flyers was monitored under stress conditions ., Remaining pupae ( RF flies ) were emerged and released in the target area of the eradication programme ., The quality assessment of the QC flies was a good proxy of the quality of the RF flies ., The quality protocol described here will allow the accurate monitoring of the quality of shipped sterile tsetse males used in operational eradication programmes . | null | null |
journal.pgen.1004547 | 2,014 | Comprehensive Identification of Single Nucleotide Polymorphisms Associated with Beta-lactam Resistance within Pneumococcal Mosaic Genes | Recent research aimed at finding the genetic causes of beta-lactam resistance in Streptococcus pneumoniae has been focused on laboratory mutagenesis 1–4 , sequence comparison 1 , 5 , 6 , and identification of interspecies sequence transfer that promotes penicillin non-susceptibility 7–10 ., Though these studies have increased our understanding , their resolution is limited , and a whole-genome systematic search in real population settings is still lacking ., Indicative of this limited resolution is the frequent use of the term “mosaic genes” to describe pneumococcal resistance alleles 7 ., Although recombinational mosaics are clearly identifiable as regions of several hundred nucleotides in resistance genes , it is likely that only a subset of the observed alterations are important in causing resistance ., Genome wide association studies ( GWAS ) have been used to identify genetic loci associated with complex diseases ranging from cancer to mental illness in human populations 11–13 ., While the method should , in theory , be applicable to bacterial populations , its use has been inhibited by significant difficulties ., These are primarily due to the clonal population structure , and generally limited recombination within bacteria , which make the causal SNPs indistinguishable from other linked SNPs , effectively creating very large haplotype blocks 14 , 15 ., Attempts have been made to take this clonal structure into account in association analyses 16 , 17 , but strong linkage disequilibrium will always restrict the resolution of the approach ., To overcome this , studies will require either populations with elevated recombination , a large diverse sample , or both , to make the statistical analysis robust ., The confounding effect from clonal population structure may be less problematic in highly recombinogenic bacteria ., Homologous recombination brings genetic admixture into bacterial populations in a manner akin to sexual reproduction in humans , although it does not occur every generation , and only affects a small part of the genome in each occurrence ., In S . pneumoniae , homologous recombination involves , on average , 2 . 3 kb of chromosomal DNA 18 , about twice the size of an average pneumococcal gene , suggesting that large numbers of recombinational events must accumulate in order to break up linkage blocks smaller than this size ., However , the recombination signals left after the action of natural selection are not uniformly distributed across the genome but are concentrated at particular loci , which are commonly known as recombination hotspots ., These hotspots are coincident with genes involving the bacterial response to selection pressure , which includes host immune responses and antibiotic utilization , particularly beta-lactams 19 , 20 ., We hypothesized that the frequency of recombination at these sites would therefore be sufficient to allow the identification of causal SNPs associated with resistance to beta-lactams , given a large enough sample size ., Continuing reduction in sequencing costs has allowed the scale of whole-genome bacterial population studies to increase 20 , 21 , and this should provide more robust statistical power for association analyses ., The availability of multiple large bacterial population studies allows a replication of such association studies , providing stronger evidence for common causal SNPs as well as the potential to identify rarer causal SNPs , some of which might only be detected in unique population settings ., Here we conducted an association study using the pneumococcal populations from carriage cohorts in Maela refugee camp , Thailand 20 , and Massachusetts , USA 21 , two recent species-wide pneumococcal studies from which large numbers of whole genome sequences and phenotypes for beta-lactam susceptibility are available ., Given the high recombination frequency in S . pneumoniae generally , the observed recombination hotspots covering antibiotic resistance genes , and the relatively large sample sizes of both studies , we hoped to overcome the challenges in performing bacterial association studies discussed above ., Therefore this study aimed to more precisely identify the sets of variants associated with resistance , where they were located in the genome , and how they were distributed across the population ., We conducted an association study on whole genome SNPs and insertions or deletions ( indels ) to identify variants associated with beta-lactam non-susceptibility ., Two rounds of analyses were performed separately using 3 , 085 genomes from pneumococcal strains collected from a carriage cohort in Maela 20 , and 616 genomes from a carriage cohort from Massachusetts 21 ., Based on the Clinical and Laboratory Standard Institute guidelines ( CLSI , 2008 ) , strains with penicillin minimum inhibitory concentration ( MIC ) ≤0 . 06 µg/ml were classified as susceptible; applying these cutoffs to our data gave 1 , 729 case ( non-susceptible ) and 1 , 951 control ( susceptible ) samples for our study ( with 21 unknown ) ., The Maela and Massachusetts populations comprise strains from multi-lineage backgrounds ., Therefore , taking the population stratification into account is essential to separate the clonal population signals from true phenotypic associations ., The population substructures utilized were those defined previously 20 , 21 , which in both cases were determined using a Bayesian clustering approach ( see Methods ) ., Based on this clustering information , the Cochran-Mantel-Haenszel ( CMH ) association statistic was employed to test for associations between beta-lactam non-susceptibility and specific variants , conditional on the population cluster ., For each population , we screened for common alterations with minor allele frequency >0 . 01 and reported sites with a p value<0 . 01 , incorporating a Bonferroni adjustment for multiple comparisons ( Figure 1 , Tables S1–S3 ) ., We found 858 and 1 , 721 SNPs associated with beta-lactam non-susceptibility in the Maela and Massachusetts populations respectively ( Figure 2 , Tables S2–S3 ) ., Among these , 301 SNPs were found to be associated with non-susceptibility in both populations ., Considering that the two settings have different population structures that have evolved independently , these co-detected SNPs represent a set of candidates in which we can have more confidence ., Rather counter-intuitively , more candidate SNPs were identified in the smaller dataset from Massachusetts than in the larger dataset from Maela ., There are several potential explanations for this observation; one being that it is due to different linkage structure within the two datasets ., Not all of the candidate SNPs will necessarily play a causative role; some may be tightly linked to causative SNPs , with insufficient recombination in the dataset to separate them ( here called “hitchhiking” SNPs ) , and hence form part of the same haplotypes ., To test this , we estimated the size of the haplotype blocks that harbor candidate SNPs in both the Maela and Massachusetts populations , using the criteria described in Gabriel et al . 22 , 23 ( see Methods ) ., Haplotype block sizes detected in the Maela data are significantly smaller than the Massachusetts data ( Mann Whitney test p value 6 . 53×10−9 , Figure S1 ) , indicating that many of the candidate SNPs detected in the Massachusetts data are potential hitchhikers , thereby generating some false positive results ., The second potential explanation is due to the population stratification defined previously 20 , 21 ., As the clustering analyses were performed separately on each dataset , it is possible that the clustering information from the two datasets are not equivalent in their stringency , leading to a more strict control over population stratification in one population than the other ., Nevertheless , 51 candidate loci , comprising a total of 301 discrete and linked SNPs , were co-detected in both the Maela and Massachusetts data; using these should provide a high-stringency data set that overcomes these population-specific effects ., The co-detected loci include three intergenic SNPs , and 298 SNPs detected in coding sequences ., The latter can be divided into 71 non-synonymous and 227 synonymous SNPs ., Of these 51 loci , nine were single SNPs , and 42 were in linkage blocks of between two and 19 SNPs , of which 12 contained only a single non-synonymous SNP ., Based on assembled sequences , we also investigated whether or not indels found in associated genes could contribute to the resistant phenotype ., None of the identified indels showed significant association with resistance , after Bonferroni correction , at a p-value of <0 . 01 ., To estimate how much of the phenotypic resistance in our samples could be explained by the identified SNPs , we performed cross-prediction tests using only the SNPs co-detected in both the Maela and Massachusetts association studies , tested back against each population separately ., We found that close to 100% of the resistance in each population could be explained by all of the co-detected SNPs ( Figure S2 ) ., Unlike human polygenic traits where each locus contributes only a small effect on the phenotype , each of these bacterial loci appears to have a much stronger effect , and indeed some have been shown experimentally to change the phenotype with only a single variant 1 ., This can be demonstrated using odds ratios , which indicate the size of the effect of each associated SNP ., While human GWAS studies report a median odds ratio of 1 . 33 per SNP 24 , 25 , our analysis gives a median odds ratio of 11 . 09 per SNP , indicating a stronger effect size ., For both the Maela and Massachusetts populations , the percentage of resistance explained plateaued after the addition of approximately 10 loci in any order ., This suggests that , at most , about 10 loci are required to make a susceptible strain non-susceptible and that multiple different combinations can achieve this ., However , in each resistant isolate , combinations of more than ten loci are commonly detected , perhaps indicating that not all loci are involved in conferring resistance , but that some may play a compensatory role in reducing the fitness cost of resistance variants ., In total , the co-detected variants are present in 100% and 98% of the Maela and Massachusetts resistant strains respectively , highlighting that a large proportion of possible resistance variants are captured in our study ., For both population settings , loci found to be associated with beta-lactam non-susceptibility show higher enrichment in genes compared to intergenic regions than would be expected by chance ( Fishers Exact Test p-value<0 . 0001 ) ., Candidate loci are not randomly distributed across the whole genome , but clustered within certain genes ( Figure 1 ) ., Co-detected loci in both datasets are localized in genes participating in the peptidoglycan biosynthesis pathway , including penicillin binding proteins ( pbp2x , pbp1a , pbp2b ) and two transferases required for cell wall biogenesis ( mraW , mraY ) , the cell division pathway ( ftsL , gpsB ) , heat shock protein and chaperones ( clpL , clpX ) , the recombination pathway ( recU ) and a metabolic gene known to confer resistance to co-trimoxazole ( dhfR ) ., Some of these sites , particularly in the pbp genes , matched those previously reported to play an important role in beta-lactam resistance in the literature ( Table S1 ) ., To our knowledge , out of 71 non-synonymous SNPs reported here , 43 SNPs are novel and potentially contribute to beta-lactam non-susceptibility in addition to those identified in previous studies ., Since most beta-lactam antibiotics work by inhibiting cell wall biosynthesis , it is not surprising to see significant associations between non-susceptible phenotypes and variants in genes participating in the peptidoglycan biosynthesis pathway , including pbp2x , pbp1a , pbp2b , mraW and mraY ., Many single amino acid alterations in pbp2x , pbp1a and pbp2b have been previously demonstrated experimentally to increase pneumococcal resistance to beta-lactams ., Mutations within or close to the active sites of the transpeptidase domain in penicillin binding proteins have been reported to be associated with penicillin resistance 26–28 ., By interfering with the formation of a covalent complex between the active site serine and antibiotic molecules , these mutations help reduce the binding affinity of beta-lactam rings to the transpeptidase enzyme ., This allows the pneumococci to form a functional cell wall , and thereby become non-susceptible ., We observed many predicted loci co-localizing with or surrounding the transpeptidase active sites ., These are recognized as three conserved amino acid motifs , SXXK , SXN and KT ( S ) G 1 , 29 and are highlighted as vertical dotted lines in the bottom panel of Figure 1 ., Many known structurally characterized alterations in pbp genes have been rediscovered in our analysis , providing independent validation of some of our results ., In pbp2x , we observed an association at T338A , which is located next to the active site 337 ., The side chain of T338 is required for hydrogen-bonding , and the T338A substitution results in a distortion of the active site 1 ., In pbp1a , an alteration from TSQF to NTGY at position 574 , which has been shown to have a lower acylation efficiency in vitro 1 , 30 , was also observed in this analysis ., In addition to candidates known to confer structural changes that lead to resistance , we also observed association with E285Q in pbp1a which might contribute to a fitness compensation mechanism caused by resistance in pbp2b 31 ., Other functional conformational changes , as well as variants that matched previous observations from sequence comparison , are tabulated in Table S1 1 , 2 , 5 , 6 , 30–38 ., Moreover , we observed substitutions outside pbp genes that could potentially affect antibiotic susceptibility , or represent compensatory mutations that interact epistatically with changes associated with resistance ., These include the genes mraY and mraW , which encode transferases ., Both function upstream of the pbp genes in the peptidoglycan biosynthesis pathway ., The genome-wide screen provided us with an opportunity to identify associations outside the pbp genes and the peptidoglycan biosynthesis pathway , which are the direct target of beta-lactams ., In both the Maela and Massachusetts datasets , nine independent loci comprising 31 SNPs were detected outside of these pathways ., These include amino acid alterations in a major heatshock protein , clpL ., Mutants lacking clpL have been previously reported to be more susceptible to penicillin 1 ., The effect was attributed to the ability of clpL to interact with and stabilize the Pbp2x protein ., In the Massachusetts data alone , we observed associations between resistance and ciaH , a histidine kinase sensor , and ciaR , its response regulator , consistent with previous studies reporting a large increase in resistance due to ciaH mutations ., The mutations in the ciaH kinase sensor resulted in hyperactivation of the ciaR regulator , which in turn leads to increased beta-lactam resistance 1 , 39 ., Association signals from ftsL and gpsB genes were detected in both datasets ., These two genes function in cell division and are essential for complete cell wall formation ., Depletion of GpsB leads to cell deformation with a similar phenotype to that observed when the Pbp2x protein is inhibited by methicillin 40 ., These identified candidates potentially interact with pbp genes , either directly or indirectly through regulation or participation in cell wall formation; however , experimental characterization will be required to explore the mechanisms of how these alterations influence beta-lactam susceptibility ., Interestingly , strong discrete associations were also found in genes where specific variants are known to confer resistance to co-trimoxazole , an antibiotic targeting the bacterial DNA synthesis pathway 41 ., We detected associations in dhfR ( encoding dihydrofolate reductase ) and folP ( dihydropteroate synthase ) , which are required for folate synthesis and are essential for nucleotide biosynthesis ., Given that beta-lactam and trimethoprim resistance arise from different mechanisms and that the dhfR and folP loci are not genetically linked to any other detected loci , it is curious as to why we observed these signals ., A possible explanation could be the contemporaneous use of both beta-lactam and trimethoprim antibiotics in the host populations studied , which would drive co-selection for resistance to the two unrelated classes of antibiotics ., In both the Maela and Massachusetts datasets , strains that are phenotypically resistant to beta-lactams are more likely to be phenotypically resistant to co-trimoxazole , suggesting that the two phenotypes did not occur independently ( Fishers exact test p-value<2 . 2×10−16 , Table 1 ) ., Clinical records from Thailand listed beta-lactams and co-trimoxazole as the first and second most frequently prescribed antibiotics for upper respiratory infection treatments 42 , indicating that co-selection pressures may have been possible if the two antibiotics were frequently used together ., As some of the variants detected by our study are known to affect the binding affinity for beta-lactams , we looked to see whether the effect would be equivalent across all classes of beta-lactam antibiotics , or if resistance due to specific variants would be greater for certain classes of antibiotic ., To test this , we replicated the analysis on the candidate loci using the continuous phenotype of the minimum inhibitory concentration ( MIC ) value for two classes of beta-lactam antibiotics; penicillins and cephalosporins ( here represented by ceftriaxone ) ., Penicillins and cephalosporins both possess a characteristic beta-lactam ring , but while the beta-lactam ring is fused to a 5-membered thiazolidine ring in penicillins , it is fused to a 6-membered dihydrothiazine ring in cephalosporins ., The drugs also differ in side chains that differentiate their kinetic properties 43 ., Figure 3 plots the differential association of each locus to the two beta-lactam antibiotics ., Loci with stronger association towards penicillin are distributed along the positive y-axis , while those showing a stronger preference towards cephalosporins are distributed along negative y-axis ., We found that loci do not contribute equally to different classes of beta-lactam antibiotics ( Kruskal-Wallis rank sum test , p value<2 . 2×10−16 ) , with a strong association of some loci towards resistance to either penicillins or cephalosporins ., Given the known pneumococcal population structure in both the Maela and Massachusetts datasets 20 , 21 , we sought to explore and compare the prevalence of candidate beta-lactam resistance alleles identified as loci co-detected in the two populations ., The two populations are composed of multiple pneumococcal lineages , many of which are present in one population but absent in the other ., This difference in population structure has a large influence on the types of resistance alleles detected in each setting ., Therefore , an unbiased comparison between the Maela and Massachusetts populations can only be made using the lineages common to both locations ., PMEN-14 , the globally dispersed multidrug resistant lineage , was detected in both populations ( Maela: 2007–2010 , Massachusetts 2001–2007 ) , and thus allows a comparative view between the two datasets ., PMEN-14 isolates from Maela and Massachusetts have significantly different beta-lactam resistance allelic profiles ( Mann-Whitney U test , p value\u200a=\u200a4 . 68×10−12 ) ., Though the local beta-lactam resistance profiles are different , the pattern of their distribution across the Maela and Massachusetts pneumococci is similar ., In both populations , the distribution of resistance alleles is not uniform ( Figure 4 ) ., The multidrug resistant lineages PMEN-14 and PMEN-1 , along with other vaccine target lineages appear to carry predicted resistance alleles at a higher frequency ., This reflects the vaccines design to target serotypes associated with antibiotic resistance 21 , 44 ., However , levels of beta-lactam resistance have generally remained stable post-vaccine introduction 21 , 45 , 46 , which has resulted from the success of resistant non-vaccine lineages with a high frequency of resistance alleles ( e . g . 35B in Massachusetts , NT in Maela; Figure 4 ) and serotype switching by resistant-vaccine type lineages such as PMEN-14 21 ., In pneumococcal populations dominated by NT lineages such as in Maela , the higher rate of recombination observed in these lineages 20 , and the fact that they are not targeted by current vaccines , may allow them to act as both source and sink for resistance alleles , generating more combinations that are then seeded into the wider population ., The power of phenotype-genotype association studies in bacteria is limited by the clonal population structure and limited recombination within these organisms ., One approach to overcoming this is to explicitly account for population structure in the analysis , and this was recently attempted for studies of host-association in Campylobacter 16 and Staphylococcus aureus 17 , but the sensitivity of both was limited by a relatively small sample size ., Our analysis used a S . pneumoniae data set of much larger size , enhancing statistical power in detecting associated variants ., Our approach also took advantage of the higher level of recombination in the genes participating in the peptidoglycan biosynthesis pathway , some of which are known to be recombination hotspots 19 , 20 significantly reducing the effect of long haplotype blocks within these important genes ., Together , this enabled us to identify specific nucleotide variants underlying beta-lactam resistance in this organism , some of which were previously known , but many of which are novel ., Our analysis allows the refinement of the understanding of resistance beyond “mosaic genes” to identify likely causative variants , and shows that there are multiple loci which may contribute to resistance ., We have also been able to show that , while some loci likely contribute universally to all beta-lactam resistance , some can demonstrate a stronger association with resistance to certain classes of antibiotics more than others ., We used this resistance variant dataset to examine the allele distribution within the sampled population ., While specific lineages can vary between populations in the resistance loci present , a general finding was that a high frequency of resistance alleles could be found in both vaccine and non-vaccine lineages , a potential explanation for why vaccination has not reduced beta-lactam resistance within the population ., Some non-vaccine target lineages with a high frequency of resistant alleles can act as both a source and a sink of resistance alleles within the population ., A limitation of our approach is that it is sensitive to recombination frequency and requires a non-clonal population and a large sample size ., Although the recombination frequency of different bacteria is relatively fixed , current sequencing technologies do now allow very high sample sizes within bacterial populations , and this may increase the applicability of this approach in the future ., The sensitivity of detection of association will also be enhanced by the occurrence of de-novo mutations conferring resistance ( homoplasy; 19 ) , representing convergent evolution ., However , this will only be possible using whole-genome sequencing , and these could not be detected by eukaryotic-style marker-SNP-based association studies ., Future use of whole-genome sequencing for antibiotic resistance/sensitivity prediction in clinical practice will rely on the ability to assign function to specific variants , rather than mosaic blocks , and this kind of study will be essential to enable these future applications ., Nevertheless , results reported from this genome-wide association study are hypothesis-generating and will require further functional validation ., The test populations represent the largest datasets for which whole genome sequences and antibiotic-resistance phenotypes are available - Maela 20 and Massachusetts 21 ., Beta-lactam susceptibilities were determined in both datasets by disk diffusion following the CLSI 2008 guidelines 47 ., Our analyses contained 1 , 501 non-susceptible , 1 , 568 susceptible and 16 unknown phenotypes in the Maela data; 228 non-susceptible , 383 susceptible and 5 unknown phenotypes from the Massachusetts data ., The minimum inhibitory concentrations ( MIC ) of non-susceptible isolates were confirmed by the E-test method ( bioMerieux , Marcy LEtoile , France ) ., Strain names and a full list of MICs from the Maela dataset 20 are given in Table S4 ., Strains and metadata for the Massachusetts dataset were given as supplementary data in ref 21 ., Samples were previously sequenced as multiplexed libraries on Illumina Hiseq 2000 machines using 75-nt or 100 nt paired-end runs as described in 20 , 21 ., Short reads from both studies have previously been deposited in the European Nucleotide Archive under study number: Maela data - ERP000435 , ERP000483 , ERP000485 , ERP000487 , ERP000598 and ERP000599; and Massachusetts data as listed in Table S1 of 21 ., Reads from both datasets were mapped onto a single reference genome , S . pneumoniae ATCC 700669 ( European Molecular Laboratory ( EMBL ) accession FM211187 ) 48 using SMALT 0 . 5 . 7 ., ( http://www . sanger . ac . uk/resources/software/smalt/ ) ., Bases were called from mapped sequences using the methods described in 49 , resulting in 392 , 524 and 198 , 248 SNP calls from the Maela and Massachusetts data respectively ., Genomes from Maela were de novo assembled using Velvet 50 with combinations of SSPACE 51 , GapFiller 52 , BWA 53 and Bowtie 54 as in 20 and genomes from Massachusetts were assembled with Velvet exclusively as described in 21 ., Assembled sequences allowed variations from insertions and deletions ( indels ) to be incorporated for a deeper analysis at each locus ., The Maela and Massachusetts populations represent species-wide data sets; they respectively consist of 65 and 46 different capsule types , and at least 277 and 154 known multilocus sequence types ., The population substructures as determined in 20 21 were used in this analysis ., Briefly , whole genome-mapped sequences and concatenated core genome sequences were used in the Maela 20 and Massachusetts data 21 , respectively , as input to the BAPS software 55–57 ., BAPS was used to define the clonal population structure by estimating the structure based on non-reversible stochastic optimization ., The method has successfully been applied to bacterial populations of several different species 58 , 59 ., Individual strains in Maela and Massachusetts data were first partitioned into clusters based on multiple runs of the estimation algorithm ( Methods in 2021 ) ., This resulted in 33 and 16 initial clusters for the Maela and Massachusetts data , respectively ., Due to the large sample size of the Maela dataset , BAPS was additionally run in a hierarchical manner ., As described in 20 , data from each of the primary clusters identified in the Maela data were re-analyzed to obtain secondary clusters within each primary cluster , and these were used to represent the population structure of Maela pneumococci ., The haploid bacterial SNP information was treated as human mitochondrial sequence in PLINK v . 1 . 07 60 and controlled for missing rate and allele frequency ., We excluded variants with minor allele frequency <0 . 01 , missingness by strain >0 . 1 and missingness by variants >0 . 1 ., For each site , the top two most common variants were parsed to the next analysis to reduce complexity in the test statistic ., Intrinsic noise from genetic variation alone can lead to false positive signals ., To estimate basal false positive rates and decide a suitable cut-off for each dataset , we separately ran 100 GWAS permutations with true genotypes but randomized binary phenotypes ( Figure S3 ) ., None of the permutations of either the Maela or Massachusetts datasets achieved any significant association at p-value 0 . 01 with a Bonferroni correction for multiple testing , therefore validating a Bonferroni-adjusted cut-off at p-value 0 . 01 as our conservative threshold ., We first determined SNPs associated with beta-lactam resistance with binary phenotypes: susceptible or non-susceptible ., However , the intrinsic clonal population structure of bacteria can result in high false positive rate in GWAS ., The tests were thus performed conditioned on the population structure generated by BAPS in previous publications 20 , 21 and controlled for genomic inflation factor ., Based on known cluster information , the Cochran-Mantel-Haenszel ( CMH ) test for 2×2×K binary phenotype x variants | population cluster was employed with sites corrected for multiple testing using the Bonferroni correction at a p-value of 0 . 01 ., The application of the CMH test reduced the genomic inflation factor from 80 . 16 ( mean chi-squared statistic\u200a=\u200a68 . 99 ) to 2 . 56 ( mean chi-squared statistic\u200a=\u200a3 . 05 ) in the Maela data , and 13 . 18 ( mean chi-squared statistic\u200a=\u200a14 . 17 ) to 3 . 76 ( mean chi-squared statistic\u200a=\u200a4 . 73 ) in the Massachusetts data ., The reductions in genomic inflation factor seen in both datasets suggest a decrease in false positive rates due to underlying population structure ., However , the genomic inflation factors observed here are relatively high compared to those observed in human nuclear chromosome GWAS , suggesting that intrinsic clonal population structure is still an issue for bacterial association studies ., Genome wide association studies are sensitive to population stratification ., While a stringent stratification helps reduce false positives , it potentially increases false negatives ., Due to the size of the Maela dataset , we had available the ( more relaxed ) primary BAPS clusters , and the ( more stringent ) secondary BAPS clusters , and we therefore used these to investigate the effect of clustering size with respect to the number of discovered variants and false positive rate in our data ., We separately repeated the Cochran-Mantel-Haenszel ( CMH ) test as described above using information on primary and secondary BAPS clusters as previously defined 20 ., We detected greater numbers of variants with significant associations when stratified by primary clusters compared to secondary clusters ( 10 , 451 SNPs compared to 858 SNPs ) ., Also , a higher false positive rate was observed in the analyses using the primary clusters than the secondary clusters ( genomic inflation factor of 6 . 58 compared to 2 . 56 ) ., This result is consistent with what is expected , reflecting a trade-off between false positives and false negatives , and will be dependent on the sample size and underlying population structure ., A high genomic inflation factor indicated that some of the candidate alleles were influenced by population structure and were likely to be hitchhikers ., We explicitly tested for linkage disequilibrium between candidate SNPs using Haploview version 4 . 2 61 ., The information was treated as male human X-chromosome to retain its haploidy ., Haploview was devised for human genetics where linkages between distant sites are disrupted by crossing-over ., Unlike human , bacterial recombination does not necessarily break long distance linkage ., We therefore set Haploview to consider all pairwise comparisons under 2 , 200 kb , which is the size of the whole S . pneumoniae genome , thus incorporating all possible linkage predictions into our analysis ., Using 95% confidence bounds as described in 22 , a haplo | Introduction, Results, Discussion, Materials and Methods | Traditional genetic association studies are very difficult in bacteria , as the generally limited recombination leads to large linked haplotype blocks , confounding the identification of causative variants ., Beta-lactam antibiotic resistance in Streptococcus pneumoniae arises readily as the bacteria can quickly incorporate DNA fragments encompassing variants that make the transformed strains resistant ., However , the causative mutations themselves are embedded within larger recombined blocks , and previous studies have only analysed a limited number of isolates , leading to the description of “mosaic genes” as being responsible for resistance ., By comparing a large number of genomes of beta-lactam susceptible and non-susceptible strains , the high frequency of recombination should break up these haplotype blocks and allow the use of genetic association approaches to identify individual causative variants ., Here , we performed a genome-wide association study to identify single nucleotide polymorphisms ( SNPs ) and indels that could confer beta-lactam non-susceptibility using 3 , 085 Thai and 616 USA pneumococcal isolates as independent datasets for the variant discovery ., The large sample sizes allowed us to narrow the source of beta-lactam non-susceptibility from long recombinant fragments down to much smaller loci comprised of discrete or linked SNPs ., While some loci appear to be universal resistance determinants , contributing equally to non-susceptibility for at least two classes of beta-lactam antibiotics , some play a larger role in resistance to particular antibiotics ., All of the identified loci have a highly non-uniform distribution in the populations ., They are enriched not only in vaccine-targeted , but also non-vaccine-targeted lineages , which may raise clinical concerns ., Identification of single nucleotide polymorphisms underlying resistance will be essential for future use of genome sequencing to predict antibiotic sensitivity in clinical microbiology . | Streptococcus pneumoniae is carried asymptomatically in the nasopharyngeal tract ., However , it is capable of causing multiple diseases , including pneumonia , bacteraemia and meningitis , which are common causes of morbidity and mortality in young children ., Antibiotic treatment has become more difficult , especially that involving the group of beta-lactam antibiotics where resistance has developed rapidly ., The organism is known to be highly recombinogenic , and this allows variants conferring beta-lactam resistance to be readily introduced into the genome ., Identification of the specific genetic determinants of beta-lactam resistance is essential to understand both the mechanism of resistance and the spread of resistant variants in the pneumococcal population ., Here , we performed a genome-wide association study on 3 , 701 isolates collected from two different locations and identified candidate variants that may explain beta-lactam resistance as well as discriminating potential genetic hitchhiking variants from potential causative variants ., We report 51 loci , containing 301 SNPs , that are associated with beta-lactam non-susceptibility ., 71 out of 301 polymorphic changes result in amino acid alterations , 28 of which have been reported previously ., Understanding the determinants of resistance at the single nucleotide level will be important for the future use of sequence data to predict resistance in the clinical setting . | genetics, biology and life sciences, population biology, microbiology | null |
journal.pntd.0004139 | 2,015 | Loss of Glycosaminoglycan Receptor Binding after Mosquito Cell Passage Reduces Chikungunya Virus Infectivity | Chikungunya virus ( CHIKV ) is a mosquito-transmitted , single-stranded RNA virus belonging to the genus Alphavirus of the family Togaviridae ., In humans , CHIKV infection can cause fever , headache , maculopapular rashes , myalgia , acute joint swelling , persistent arthritis , and even life-threatening neurological or cardiovascular complications 1–5 ., CHIKV was first identified in Africa in 1952 and has been endemic in the tropical Indian Ocean countries for decades 6 ., In recent years , this virus has caused more widespread and noticeable outbreaks ., From 2004 to 2011 , approximately six million cases of CHIKV infection were reported from nearly forty countries in Africa , Asia , and Europe 6–9 ., In recent years , CHIKV mosquito transmission vectors Aedes aegypti and Ae ., albopictus have spread from tropical to temperate climates , making CHIKV an emerging pathogen within these climate zones 10 , 11 ., In line with this , CHIKV cases have been recently reported from more than twenty-five countries in the Caribbean islands , thereby posing a potential threat to North America 12 ., Unfortunately , CHIKV pathogenesis is not well understood , and there is no vaccine or specific antiviral treatment currently available for CHIKV infection 13–15 ., CHIKV circulates between mammalian and mosquito hosts and this cyclical transmission may provide a suitable environment for increased viral fitness and the emergence of more pathogenic strains 16 , 17 ., Interestingly , re-emergence of CHIKV during the 2005–2006 epidemic on Reunion Island was associated with a single point mutation in its genome , which increased CHIKV fitness within its mosquito vector Ae ., albopictus 18 ., Additionally , CHIKV and other alphaviruses differ in their ability to infect mammalian and mosquito cells ., For example , alphaviruses can cause cytopathic effects in mammalian cells and can also shut-down the mammalian macromolecular machinery involved in cellular protein synthesis at both the transcription and translational levels 19–21 ., In contrast , alphavirus infection of mosquito cells causes little to no cytopathic effects and does not affect the cellular transcription and translational processes 21–24 ., Mammalian and mosquito cells have distinct cellular enzymatic systems for protein glycosylation; therefore , different post-translational processing of viral surface proteins are possible in these host cells 25 , which can influence replication 26–28 , pathogenesis 28 , 29 , transmission 30 , and evolution 17 of mosquito-transmitted viruses ., In line with this , mammalian- and mosquito-generated arboviruses can bind to different receptors expressed on the surface of host cells ., For instance , differential glycosylation of viral receptor-binding proteins in mammalian- and mosquito-generated Sindbis virus 31 , West Nile virus ( WNV ) 32 , and dengue virus 33 , can affect binding of these virus to host cell receptors ., Similarly , mammalian cell-generated Ross River virus ( RRV ) , Venezuelan equine encephalitis virus ( VEEV ) , and WNV can induce more potent interferon responses compared to their mosquito cell-generated counterparts 34 , 35 ., However , it remains unclear whether CHIKV generation in mosquito and mammalian cells can affect its infectivity and virulence ., Glycosaminoglycans ( GAGs ) are highly sulfated polysaccharides that are ubiquitously expressed on the cell surface and the extracellular matrix of mammalian cells 36 , 37 ., Many viruses including CHIKV can utilize GAGs as receptors to infect host cells 38 ., However , research on the role of GAG receptor binding in CHIKV and other alphaviruses has been inconclusive ., The GAG receptor binding of CHIKV and other alphaviruses can be acquired through acquisition of basic amino acids in viral receptor-binding proteins via mutations during their continuous passage in cell culture 36 , 38 ., Although such dependence on GAG receptor binding increases viral infectivity in vitro , it can potentially decreases viral fitness in vivo 39 ., In contrast , GAG binding properties have also been described in non-cell culture adapted alphaviruses , including a clinical strain of CHIKV 40 and a wild-type strain of eastern equine encephalitis virus ( EEEV ) 41 , suggesting that some other mechanisms that are independent of cell culture adaptation may also control GAG binding and virulence of CHIKV and other alphaviruses ., Biochemically , GAG receptors possess a negative charge that enables virus-GAG receptor interaction 40–42 ., In addition , mosquito and mammalian cells have different N-glycosylation mechanisms that can generate different configurations of viral glycoproteins 43 and modulate charge dependent interaction of viruses to host cell receptors ., Thus , virus generation in these different host cells can potentially influence receptor binding and infectivity of CHIKV ., However , the role of mosquito- and mammalian cells on GAG binding capability of CHIKV and other alphaviruses is unclear ., Herein , we report that CHIKV generated in mammalian cells replicates more efficiently during its subsequent infection of both human and murine cells in vitro and is more virulent in a mouse model of CHIKV arthritis , when compared to its mosquito cell-generated counterpart ., We further demonstrate that the reduced replication of mosquito cell-generated CHIKV is associated with its failure to bind to cell surface GAG receptors on mammalian cells due to differential glycosylation of viral proteins in mosquito cells ., This study was carried out in strict accordance with the recommendations described in the Guide for the Care and Use of Laboratory Animals of the National Research Council of The National Academies ., The Institutional Animal Care and Use Committee at the University of Southern Mississippi ( Animal Welfare Assurance # A3851-01 ) reviewed and approved all the animal care and use procedures under the protocol #12041201 ., All in vitro experiments and animal studies involving CHIKV were performed by certified personnel in biosafety level 3 ( BSL3 ) laboratories , following standard biosafety protocols approved by the University of Southern Mississippi Institutional Biosafety Committee ., Low-passaged , Vero cell-generated CHIKV Ross strain ( provided by Dr . John F . Anderson , Connecticut Agricultural Experiment Station ) and LR OPY1 2006 strain ( provided by Dr . Robert B . Tesh , University of Texas Medical Branch ) were used as parental viral stocks in this study ., Majority of experiments were performed using the Ross strain and to test strain specificity , some experiments were repeated using the LR OPY1 2006 strain ., The viral stocks used in this study were prepared by a single passage of parental viruses in C6/36 ( ATCC , CRL-1660 ) or Vero cells ( ATCC , CCL-81 ) and designated as CHIKVmos and CHIKVvero , respectively ., All viral stocks were titered in Vero cells by a plaque-forming assay ., C6/36 cells were cultured at 28°C with 5% CO2 in Eagle’s minimum essential media ( EMEM , ATCC ) supplemented with 10% fetal bovine serum ( FBS ) ., Vero cells were cultured at 37°C with 5% CO2 in Dulbecco’s modified Eagle’s medium ( DMEM , Life Technologies ) supplemented with 10% FBS ., L929 ( ATCC , CCL-1 ) , NIH3T3 ( ATCC , CRL-1658 ) , Raw 264 . 7 ( ATCC , TIB-71 ) , human foreskin fibroblasts ( HFF , ATCC , CRL-2522 ) , human THP-1 cells ( ATCC , TIB-202 ) and human dermal fibroblasts were cultured at 37°C with 5% CO2 in DMEM supplemented with 10% FBS ., To generate murine bone marrow-derived dendritic cells ( mBMDC ) , healthy C57BL/6J mice ( 7 weeks old ) were euthanized and bone marrow cells were recovered from both femurs ., After red blood cells were lysed , the bone marrow cells were cultured in R10 medium supplemented with 10% J558L cell supernatant ( as a source of granulocyte-macrophage colony-stimulating factor ) at a final density of 1 × 106 cells/ml at 37°C with 5% CO2 ., The medium was changed every 3 days and mBMDCs were ready for infection at day 10 ., Dermatan sulfate ( from porcine intestinal mucosa ) , chondroitin sulfate A ( from bovine trachea ) , heparin ( from porcine intestinal mucosa ) , sepharose CL-4B , heparin-sepharose , yeast mannan , tunicamycin , and neutral red were all purchased from Sigma ., Cells were plated 24 h before infection in 6- , 12- or 24-well plates to 60–80% confluence ., CHIKVmos or CHIKVvero ( MOI = 1 ) were added to the cells and incubated at 37°C for 1 h to allow for viral adsorption and penetration ., The inoculation medium was then replaced with fresh medium to remove unadsorbed viruses ., Cells were washed once with fresh medium and further incubated at 37°C with 5% CO2 and collected at selected time points for analysis of viral genome replication and host’s gene expression ., Cells were infected with CHIKV ( MOI = 1 or 5 ) for 48 h and fixed with 4% paraformaldehyde ( PFA , Electron Microscopy Science ) ., Phase contrast images were acquired using Zeiss LSM510 META confocal imaging system ( Carl Zeiss Microscopy , NY ) ., Cell viability was quantified by toluidine blue ( TB ) staining , according to the previously published method 44 ., Viable cells were assayed by measuring the absorbance of TB at 630 nm using a microplate reader ( BIO-TEC ) ., Percentage of viable cells was calculated after normalization to uninfected controls ., CHIKV infected cells were subjected to total RNA extraction using TRI-reagent ( Molecular Research Center , Inc . ) ., For RNA isolation from mouse blood samples , RNeasy mini kit ( Qiagen ) was used ., The first-strand complementary DNA ( cDNA ) was synthesized using the iSCRIPT cDNA synthesis kit ( Bio-Rad ) ., RT-qPCR assays were performed in a CFX96 Real-Time system ( Bio-Rad ) using SYBR Green supermix ( Bio-Rad ) ., Viral RNA copy numbers were expressed as the ratio of CHIKV envelope-1 ( CHIKV E1 ) to cellular β-actin ., For cytokine RT-qPCR assay , data were presented as relative fold change ( RFC ) in expression by the ΔΔCT method after normalized to cellular β-actin ., Primer sequences for β-actin of mice 45 and human 46 were previously described ., Primers for CHIKV E1 gene ( Forward: TCC GGG AAG CTG AGA TAG AA; Reverse: ACG CCG GGT AGT TGA CTA TG ) , and Ae ., albopictus ribosomal protein 7 gene ( Forward: CTC TGA CCG CTG TGT ACG AT; Reverse: CAA TGG TGG TCT GCT GGT TC ) were designed using NCBI online primer designing tool ., Primer sequences for host immune genes ( Ifn-α , Ifn-β , Tlr3 , Rig-I , Mda-5 , and Il-1β ) were described in a previous report 44 ., All primers were synthesized by Integrated DNA Technologies ., Plaque assays were performed according to our previous report with some modifications 47 ., Briefly , Vero , L929 or NIH3T3 cells were plated at 5 × 105 cells/well in 6-well plates one day before infection ., Virus-containing samples were added to cell monolayers to allow viral adsorption/penetration at 37°C with 5% CO2 for 1 h ., After removing unadsorbed viruses , cells were overlaid with 1% SeaPlaque agarose ( Lonza ) containing medium and further incubated at 37°C with 5% CO2 for an additional 48 h ., Plaques were counted after staining with 0 . 3% neutral red ., CHIKV particles in viral stocks were also quantified by RT-qPCR , as described previously 38 ., Briefly , 200 μl of viral stocks were treated with 50 units ( U ) of RNase A ( Affymetrix ) for 1 h at 37°C ., TRI-reagent was added to inactivate RNase and lyse viral particles , and viral RNA was isolated after adding 5 μg of tRNA as carrier ., First strand cDNA synthesis and CHIKV E1 gene quantification by RT-qPCR were performed as described above ., Five week old , sex-matched C57BL/6J mice ( The Jackson Laboratory ) were subcutaneously inoculated on the ventral side of the right hind footpad toward the ankle with 105 plaque forming units ( PFUs ) of CHIKVvero or CHIKVmos ( Ross strain or LR 2006 OPY 1 ) in 50 μl phosphate buffer saline ( PBS ) , or with 50 μl PBS for mock controls , according to previous publications 48–50 ., Blood samples were collected in 0 . 5M EDTA by retro-orbital bleeding and viral RNA in these samples were quantified by RT-qPCR ., The height ( thickness ) and breath ( width ) of the perimetatarsal area of inoculated feet were measured daily from day 0 to day 10 post infection ( d . p . i . ) by using a digital caliper ( Electron Microscopy Science ) , and the relative increase in swelling was calculated as previously described 50 ., Briefly , footpad swelling was expressed as the relative increase in swelling compared to pre-infection ( x d . p . i . – 0 d . p . i . ) /0 d . p . i . ) ., Mice were euthanized and inoculated footpad tissues were collected at 6 d . p . i . and fixed overnight in 4% PFA , followed by decalcification in 10% EDTA for over 10 days ., Tissues were then dehydrated , paraffin embedded , and sectioned ( 10 μm ) with a microtome ( American Optical Spencer 820 ) , followed by staining with hematoxylin and eosin ( H&E ) ., The images were acquired using a bright-field microscope ( Olympus BH2 ) ., CHIKVs ( MOI = 1 or 5 ) were added to cell monolayer for attachment at 4°C for 1 h followed by washing with fresh medium to remove unattached viruses ., The attached viruses were quantified either by RT-qPCR or by a plaque assay in the same cells , as described above ., In some experiments , media containing the unattached viruses were also collected and the unattached viruses were quantified by a plaque assay in Vero cells to confirm equal numbers of virions were added to each sample ., In addition , CHIKV attachment was analyzed by flow cytometry ., CHIKVvero or CHIKVmos ( MOI = 2 . 5 ) were added to NIH3T3 cells in triplicates in PBS supplemented with 2% FBS ( staining buffer ) and were incubated at 4°C for 45 min ., Unbound viruses were removed by washing twice with the staining buffer and the cells were fixed with 2% PFA ( Electron Microscopy Science ) for 15 min at room temperature ( RT ) ., After washing , the cells were probed with mouse monoclonal anti-CHIKV antibody ( Abcam ) and Cy5 conjugated goat anti-mouse IgG ( KPL ) secondary antibody , both for 1 h at RT ., The cells were then washed twice and re-suspended in the staining buffer and analysed in a BD LSRFortessa ( BD Biosciences ) using FACSDiva version 6 . 0 software ( BD Biosciences ) ., To assess CHIKVvero and CHIKVmos entry into the host cells , we blocked the endosome acidification process using a lysomotrophic agent alone or in combination with a low pH medium ( pH 5 . 5 ) , the latter mediated viral envelope and cytoplasmic membrane fusion , as previously described 51 , 52 ., Briefly , infection was carried out in the medium containing 20 mM NH4Cl to block endosomal acidification ., For the direct membrane fusion assay , viruses were allowed to attach onto cells and then immediately treated with low pH medium for 2 minutes ., The internalized viruses were quantified by RT-qPCR and plaque assays ., To analyze virus binding to GAG receptors , we performed a GAG neutralization assay , in which viruses were pre-incubated with soluble GAGs to inhibit their attachment to cell surface GAG receptors ., Briefly , viruses ( 2 . 5 x 106 PFU/ml ) were pre-incubated with heparin , chondroitin sulfate A or dermatan sulfate ( concentration indicated in figures ) in DMEM containing 2% FBS at 37°C for 1 h ., The virus-GAGs mixtures were then added to cells ( MOI = 1 ) at 4°C for 1 h to allow attachment ., The unattached virus-GAGs mixtures were removed and cells were washed once with fresh culture medium ., The viruses attached to cells were quantified by RT-qPCR and plaque assays ., In some experiments , the effect of heparin-pretreatment on viral replication was measured at 24 hours post-infection ( h . p . i . ) by RT-qPCR ., To measure virus binding to lectin receptors such as DC-SIGN and L-SIGN , we performed a blocking assay in the presence of yeast mannan that disrupts interaction of viruses to cell surface lectin receptors ., Briefly , cells were pretreated with different concentrations of yeast mannan for 30 minutes at room temperature ., Viruses were then added to the cells ( MOI = 1 ) and incubated at 4°C for 1 h ., The viruses attached on cells were quantified by RT-qPCR ., For flow cytometric analysis of GAG neutralization , CHIKVvero or CHIKVmos ( 2 . 5 × 106 PFU/ml ) were pre-incubated with different concentrations of GAGs in DMEM containing 2% FBS at 37°C for 1 h ., Virus-GAGs mixtures were added to NIH3T3 cells in PBS supplemented with 2% FBS ( MOI = 2 . 5 ) and incubated at 4°C for 45 min ., Cells were washed twice at 4°C to remove unbound virus and immediately fixed with 2% PFA for 15 min ., Cells were then probed with anti-CHIKV antibody and analyzed by flow cytometry , as described above ., Heparin-conjugated sepharose beads or unconjugated control beads were purchased from Sigma ., The beads ( 60 μl ) were washed twice in 200 μl DMEM , and mixed with 105 PFUs of CHIKV in a total of 60 μl DMEM containing 2% FBS , and incubated at 4°C for 30 min ., The beads were then washed three times in DMEM containing 2% FBS and the washed solution was collected for subsequent plaque assays to quantify the unbound viruses ., Viruses bound to beads were lysed in 50 μl of Laemmli sample buffer ( Bio- Rad ) , and viral proteins were separated by 10% SDS-polyacrylamide gel electrophoresis and transferred to a nitrocellulose membrane ( Bio-Rad ) ., After blocked with 5% bovine serum albumin ( BSA ) for 1 h at RT , the membranes were probed with mouse monoclonal anti-CHIKV primary antibody ( Abcam ) at 4°C for overnight on a rocker ., The membranes were then washed five times ( 5 min each ) with Tris-buffered saline with Tween 20 ( TBS-T ) buffer and reacted with horseradish peroxidase conjugated goat anti-mouse IgG secondary antibody ( Jackson Immunoresearch ) for 1 h at RT ., The membranes were then washed and developed using SuperSignal West Pico Chemiluminiscence Substrate ( Thermo Scientific ) and images were acquired using a ChemiDoc MP system ( Bio-Rad ) ., Parental CHIKV viruses ( original stocks received from suppliers ) , single-passaged CHIKV in Vero cells ( CHIKVvero ) or mosquito cells ( CHIKVmos ) , and single-passaged CHIKVvero and CHIKVmos in NIH3T3 cells were subjected to RNA isolation using RNeasy Mini Kit ( Qiagen ) ., cDNA was prepared using the iScript cDNA synthesis kit ( Bio-Rad ) ., Complete CHIKV E2 gene was amplified using a Q5 high fidelity polymerase ( New England Biolab ) ., The PCR primers were used according to a previous report 38 ., The PCR fragments were purified by PureLink quick PCR Purification Kit ( Life Technologies ) and sequenced by Functional Biolab ., Stocks of CHIKV ( Ross strain ) were prepared in Vero and C6/36 cells , UV-inactivated , and viruses were concentrated by pelleting with 20% sucrose at 28 , 000 rpm for 2 h in an ultracentrifuge ( Beckman Coulter ) ., Deglycosylation of viral proteins were carried out using peptide-N-glycosidase F ( PNGase F , Sigma ) treatment following the manufacturer’s instruction ., Viral proteins were separated in a 10% Mini-PROTEAN Precast Gels ( Bio-Rad ) and imaged in a ChemiDoc MP system ( Bio-Rad ) after coomassie brilliant blue staining ., Tunicamycin ( TM ) was purchased from Sigma and dissolved ( 10 mg/ml ) in cell culture grade dimethyl sulfoxide ( DMSO , ATCC ) ., Vero cells and C6/36 cells were plated for 24 h and infected with a 0 . 1 MOI of parental CHIKV ( Ross strain ) ., Viruses were allowed to adsorb and penetrate for 1 h at 37°C ., After unadsorbed viruses were removed , the cells were further cultured with medium containing 0 . 1 μg/ml of TM or the same final concentration of DMSO ( 0 . 001% ) as vehicle controls ., Cell culture media were collected at 24 h for virus quantification by plaque assays and RT-qPCR assays ., Virus stocks generated in Vero and C6/36 cells in the presence of TM or DMSO were used to infect NIH3T3 cells ( MOI = 0 . 1 ) and viral genome replication was analyzed at 24 h by RT-qPCR ., Data were analyzed using GraphPad Prism ( version 6 . 0 , GraphPad software ) and p < 0 . 05 was considered statistically significant ., Data were compared using the two-tailed students t-test or analysis of variance ( ANOVA ) ., Previous reports have suggested that passage of virus through mosquito and mammalian cells can modulate arboviral infectivity 29 , 31 , 43 ., To investigate the difference between mosquito and mammalian cells generated CHIKV , we prepared CHIKV stocks ( Ross strain ) by infecting African green monkey ( mammal ) kidney cell line ( Vero cells ) or an Ae ., albopictus ( mosquito ) cell line ( C6/36 cells ) ., Thus generated CHIKV stocks in Vero or C6/36 cells were titered by plaque assay in Vero cells and designed as CHIKVvero and CHIKVmos , respectively ., CHIKV replicates more efficiently in fibroblastic cells compared to hematopoietic cells 53 , 54 , therefore we infected mouse embryonic fibroblasts ( NIH3T3 cells ) with CHIKVmos or CHIKVvero at a multiplicity of infection ( MOI ) of 1 ., At 24 h . p . i . , the cells were collected for total RNA extraction and the first-strand complementary DNA ( cDNA ) synthesis ., CHIKV envelope-1 ( E1 ) gene RNA copy numbers were quantified by reverse transcription quantitative polymerase chain reaction ( RT-qPCR ) and cellular β-actin was used as an internal control ., The RT-qPCR results showed that the level of CHIKVmos replication was significantly lower ( approximately 25-folds ) than CHIKVvero at 24 h . p . i . ( Fig 1A , p < 0 . 01 ) ., In addition , we also confirmed that CHIKVmos had lower replication in mouse subcutaneous fibroblasts ( L929 cells , Fig 1B , p < 0 . 05 ) , human foreskin fibroblastic cells ( HFF cells , Fig 1C , p < 0 . 05 ) and human dermal fibroblasts ( HDF cells , Fig 1D , p < 0 . 01 ) at 24 h . p . i . by RT-qPCR assay ., To further test whether CHIKVmos had lower replication over CHIKVvero in cells other than fibroblasts , we compared their replication in a mouse macrophage cell line ( Raw 264 . 7 cells ) , primary mouse bone marrow derived dendritic cells ( mBMDC ) , and a human monocyte cell line ( THP-1 ) ., Although CHIKV replication levels were relatively lower in these immune cells when compared to fibroblasts , similar reduction of CHIKVmos replication over CHIKVvero was also observed in Raw 264 . 7 cells ( Fig 1E , p < 0 . 005 ) , mBMDC ( Fig 1F , p < 0 . 05 ) , and THP-1 cells ( Fig 1G , p < 0 . 0005 ) ., In contrast to murine and human cells , both CHIKVvero and CHIKVmos replicated similarly when their gene copy numbers were compared in mosquito ( C6/36 ) cells ( Fig 1H ) ., To further study the kinetics of CHIKVvero and CHIKVmos replication over a longer infection period , we infected NIH3T3 cells with CHIKVvero or CHIKVmos ( MOI = 1 ) and cells were collected at various time points to analyze CHIKV E1 gene replication by RT-qPCR ., In NIH3T3 cells , levels of CHIKVmos replication was about 30-fold lower than CHIKVvero at 12 and 24 h . p . i ( Fig 1I , p < 0 . 0001 ) , but both viruses replicated at comparable levels at the later time points ( 36 , 48 and 60 h . p . i ) ., Both CHIKVvero and CHIKVmos also replicated at comparable levels at 48 h when assayed in L929 cells ( Fig 1J ) ., These observations suggest that CHIKVvero replicates more efficiently than CHIKVmos in murine and human cells during the early time points ., However , over the course of infection in mammalian cells , CHIKVmos may gain its infectivity and replicates similarly to CHIKVvero ., Besides using PFU/ml as a standard titer for infection assays , we also determined viral titers by measuring viral genome copies in our viral stocks using RT-qPCR ., We infected NIH3T3 cells with equal amounts of genome copies of CHIKVvero and CHIKVmos , and compared their replication levels by RT-qPCR ., Similarly , we observed a significantly lower replication of CHIKVmos compared to CHIKVvero ( S1A Fig ) , which suggests that the lower replication of CHIKVmos over CHIKVvero was not due to the difference in unencapsidated viral genome present in our viral stocks ., To test whether our results were specific to the CHIKV Ross strain , we also compared the replication levels of Vero cell-generated and C6/36 cell-generated CHIKV-LR OPY1 strain ., Similarly , we observed a lower replication of C6/36 cell-generated CHIKV-LR OPY1 when NIH3T3 cells were infected ( S1B Fig ) ., Collectively , these results demonstrate that mosquito cell-generated CHIKV has reduced levels of replication in both murine and human cells during early stage of infection , when compared to Vero cell-generated CHIKV ., CHIKV infection can cause cytopathic effects and lysis of mammalian cells 21 ., Since CHIKVmos has a much slower replication than CHIKVvero in both mouse and human cells at the early time points post infection , we expected that CHIKVmos might also cause less cytopathic effects in these cells ., To test this , we infected NIH3T3 , L929 , HFF , and C6/36 cells with CHIKVmos and CHIKVvero ( Ross strain , MOI = 1 or 5 ) for 48 h , a time point when cytopathic effects were clearly visible under a microscope ., The microscopy results showed that CHIKVvero caused more morphological distress and cell death than CHIKVmos in both human and mouse fibroblasts , but no cytopathic effect was observed in C6/36 cells ( Fig 2A ) ., This observation was further confirmed by a cell viability assay using toluidine blue staining , which showed that CHIKVmos only caused moderate cytopathic effects compared to CHIKVvero in NIH3T3 ( Fig 2B , p < 0 . 005 ) , L929 ( Fig 2C , p < 0 . 005 ) and HFF cells ( Fig 2D , p < 0 . 005 ) ., In contrast to murine and human cells , both CHIKVmos and CHIKVvero did not cause any apparent cytopathic effects in C6/36 cells ( Fig 2E ) ., To rule out the possibility that the differences in cytopathic effects were not due to Vero and mosquito cell-specific proteins that could be released in culture supernatant and might be present in our virus stocks , we examined the cytopathic effects of UV-inactivated CHIKVvero and CHIKVmos stocks in L929 cells ., We did not observe any cytopathic effects until 72 h post-treatment ( S1C Fig ) , suggesting that the observed cytopathic effects were specific to CHIKVvero and CHIKVmos infection ., Some mosquito cell-derived viruses including RRV , VEEV and WNV have been reported to exhibit enhanced infection in primary myeloid dendritic cells due to their inhibition of type I interferon production when compared to corresponding mammalian cell-derived viral preparations 34 , 35 , 55 ., While our results of CHIKVvero and CHIKVmos were opposite to those of RRV , VEEV and WNV in terms of replication 34 , 35 , 55 , we asked whether the difference in CHIKVmos and CHIKVvero replication was due to differential induction of cellular antiviral or inflammatory responses by these viruses ., To test this , we measured the expression profiles of selected pattern recognition receptors ( PRRs ) and inflammatory cytokines in CHIKVmos or CHIKVvero ( Ross strain ) infected Raw 264 . 7 , L929 , NIH3T3 , and mBMDC ( MOI = 1 ) by RT-qPCR assay ., In consistent with its lower replication , CHIKVmos induced significantly lower levels of antiviral cytokines ( Ifn-α and Ifn-β ) , proinflammatory cytokine ( Il-1β ) , and PRRs ( Tlr3 , Rig-I , and Mda-5 ) in all of the tested cell types at 24 h . p . i . ( Fig 3 , p < 0 . 05 ) ., Difference in expression of these genes in CHIKVvero and CHIKVmos infected cells correspond with replication levels of these viruses in respective cells , suggesting that higher replication of CHIKVvero over CHIKVmos may not be due to an inhibition of antiviral or inflammatory cytokine expression by these cells ., Thus , the slower replication of CHIKVmos in the early stage of infection might be due to a mechanism that is independent of the host cell antiviral responses ., Differences in in vitro replication of mammalian and mosquito-generated viruses may not always produce the similar clinical symptoms in a mouse model , as previously reported with WNV infection 56 ., Therefore , we asked whether CHIKVmos and CHIKVvero also differed in their virulence in vivo ., To test this , we infected five-week-old , sex-matched C57BL/6J mice subcutaneously via footpad inoculations with 1 × 105 PFUs of CHIKVmos or CHIKVvero ( Ross strain ) or PBS as a vehicle control ( mock ) , according to the previous reports 48–50 ., Blood samples were collected on day 1 , 2 , 4 and 6 post-infection ( d . p . i . ) for viremia measurement by RT-qPCR , and footpad swelling was measured daily from 0 to 10 d . p . i . ., CHIKVmos produced lower viremia ( presented as CHIKV E1 / β-actin ) in mice over the course of infection when compared to CHIKVvero , which reached statistical significance at 2 d . p . i . ( Fig 4A , p < 0 . 05 ) ., These results suggest that CHIKVmos displays reduced infectivity in mice ., Consistent with the viremia results , CHIKVmos induced milder footpad swelling than CHIKVvero throughout the experiment ( Fig 4B and 4D ) ., Similar results were also obtained when footpad swelling was compared in mice infected with mosquito cell- and Vero cell-generated CHIKV LR OPY1 strain ( Fig 4C and 4D ) ., To further test whether CHIKVmos causes less pathology in mice compared to CHIKVvero , we collected inflamed footpad tissue at 6 d . p . i . and performed a histological analysis ., We found that CHIKVmos induced less leukocyte infiltration and limited subcutaneous necrosis in the inflamed foot when compared to CHIKVvero ( Fig 4E ) ., Since type I interferons have been shown to play important roles in CHIKV pathogenesis 53 , 57 , we also measured expression of Ifn-α and Ifn-β in the blood of CHIKVvero and CHIKVmos infected mice at 1 , 2 and 4 d . p . i . by RT-qPCR ., ., No significant difference in expression of these genes was observed between mice infected with CHIKVvero and CHIKVmos ( S1D and S1E Fig ) , suggesting that lower level of swelling in mice infected with CHIKVmos may not be due to difference in IFN expression , but due to lower infectivity of this virus ., All these in vivo data suggest that CHIKVmos displays lower virulence than CHIKVvero in a mouse model of footpad swelling ., To further dissect the mechanism by which CHIKVmos has reduced replication and virulence , we next compared the plaque-forming phenotypes of CHIKVvero and CHIKVmos in various cells by plaque assays and counted plaques at 48 h post infection ., Consistent with RT-qPCR results ( Fig 1 ) , the plaque assays showed that CHIKVmos had an approximately 25-fold reduction in PFUs over CHIKVvero in both NIH3T3 and L929 cells ( Fig 5A and 5B , p < 0 . 01 ) when equal amounts of virus ( ~70 PFUs ) were used for plaque development ., In contrast to the numbers of PFUs , both CHIKVvero and CHIKVmos formed plaques at 48 h and no difference in plaque size was observed in all the tested cells ( Fig 5A ) ., These results suggest that the lower replication of CHIKVmos in murine and human cells may be due to its reduced ability to attach or enter into these cells ., To test whether CHIKVmos attaches to cell receptors at a lower affinity compared to CHIKVvero , we measured the attachment of CHIKVmos and CHIKVvero ( Ross strain , MOI = 5 ) on L929 , NIH3T3 and HFF cells ., Viruses were allowed to bind to the target cells for 1 h at 4°C , a condition at which most of the viruses attach to cell surfaces but do not enter into cells 58 ., Unattached viruses were removed by washing with fresh medium and the viruses attached to cells were quantified by measurement of CHIKV E1 RNA copies by RT-qPCR ., The results showed that CHIKVmos had significantly reduced attachment to L929 , NIH3T3 and HFF cells ( Fig 5C ) when compared to CHIKVvero ., Similar results were also obtained when mosquito- and Vero cell-generated CHIKV-LR OPY1 viruses were assayed for their attachment to these cells by RT-qPCR ( Fig 5D ) ., In contrast to murine and human cells , no difference in attachment between CHIKVvero and CHIKVmos was observed in C6/36 cells ( S1F Fig ) ., To further confirm lower attachment of CHIKVmos , we incubated NIH3T3 cells with CHIK | Introduction, Materials and Methods, Results, Discussion | Chikungunya virus ( CHIKV ) is a mosquito-transmitted alphavirus that can cause fever and chronic arthritis in humans ., CHIKV that is generated in mosquito or mammalian cells differs in glycosylation patterns of viral proteins , which may affect its replication and virulence ., Herein , we compare replication , pathogenicity , and receptor binding of CHIKV generated in Vero cells ( mammal ) or C6/36 cells ( mosquito ) through a single passage ., We demonstrate that mosquito cell-derived CHIKV ( CHIKVmos ) has slower replication than mammalian cell-derived CHIKV ( CHIKVvero ) , when tested in both human and murine cell lines ., Consistent with this , CHIKVmos infection in both cell lines produce less cytopathic effects and reduced antiviral responses ., In addition , infection in mice show that CHIKVmos produces a lower level of viremia and less severe footpad swelling when compared with CHIKVvero ., Interestingly , CHIKVmos has impaired ability to bind to glycosaminoglycan ( GAG ) receptors on mammalian cells ., However , sequencing analysis shows that this impairment is not due to a mutation in the CHIKV E2 gene , which encodes for the viral receptor binding protein ., Moreover , CHIKVmos progenies can regain GAG receptor binding capability and can replicate similarly to CHIKVvero after a single passage in mammalian cells ., Furthermore , CHIKVvero and CHIKVmos no longer differ in replication when N-glycosylation of viral proteins was inhibited by growing these viruses in the presence of tunicamycin ., Collectively , these results suggest that N-glycosylation of viral proteins within mosquito cells can result in loss of GAG receptor binding capability of CHIKV and reduction of its infectivity in mammalian cells . | Chikungunya virus ( CHIKV ) is a chronic arthritis-causing pathogen in humans , for which no licensed vaccine or specific antiviral drug is currently available ., Due to the global spread of its mosquito vectors , CHIKV is now becoming a public health threat worldwide ., CHIKV can replicate in both mammalian and mosquito cells , however it does not cause apparent damage to mosquito cells , yet it rapidly kills mammalian cells within a day after infection ., In addition , mosquito and mammalian cells have different mechanism of protein glycosylation , which can result in different glycan structures of viral glycoproteins ., In this study , we report that mosquito cell-generated CHIKV has lower infectivity in cell culture and causes less severe disease in mice , when compared to mammalian cell-generated CHIKV ., We demonstrate that only mammalian cell-generated CHIKV , but not mosquito-cell generated CHIKV , binds to mammalian cell surface glycosaminoglycan receptors ., Interestingly , mosquito-cell generated CHIKV can re-acquire glycosaminoglycan receptor binding capability after a single passage in mammalian cells and replicate at similar levels with mammalian cell-generated CHIKV , suggesting that passage of CHIKV in mosquito cells can reduce its infectivity . | null | null |
journal.pcbi.1000929 | 2,010 | Instantaneous Non-Linear Processing by Pulse-Coupled Threshold Units | Understanding the dynamics of single neurons , recurrent networks of neurons , and spike-timing dependent synaptic plasticity requires the quantification of how a single neuron transfers synaptic input into outgoing spiking activity ., If the incoming activity has a slowly varying or constant rate , the membrane potential distribution of the neuron is quasi stationary and its steady state properties characterize how the input is mapped to the output rate ., For fast transients in the input , time-dependent neural dynamics gains importance ., The integrate-and-fire neuron model 1 can efficiently be simulated 2 , 3 and well approximates the properties of mammalian neurons 4–6 and more detailed models 7 ., It captures the gross features of neural dynamics: The membrane potential is driven by synaptic impulses , each of which causes a small deflection that in the absence of further input relaxes back to a resting level ., If the potential reaches a threshold , the neuron emits an action potential and the membrane potential is reset , mimicking the after-hyperpolarization ., The analytical treatment of the threshold process is hampered by the pulsed nature of the input ., A frequently applied approximation treats synaptic inputs in the diffusion limit , in which postsynaptic potentials are vanishingly small while their rate of arrival is high ., In this limit , the summed input can be replaced by a Gaussian white noise current , which enables the application of Fokker-Planck theory 8 , 9 ., For this approximation the stationary membrane potential distribution and the firing rate are known exactly 8 , 10 , 11 ., The important effect of synaptic filtering has been studied in this limit as well; modelling synaptic currents as low-pass filtered Gaussian white noise with non-vanishing temporal correlations 12–15 ., Again , these results are strictly valid only if the synaptic amplitudes tend to zero and their rate of arrival goes to infinity ., For finite incoming synaptic events which are excitatory only , the steady state solution can still be obtained analytically 16 , 17 and also the transient solution can efficiently be obtained by numerical solution of a population equation 18 ., A different approach takes into account non-zero synaptic amplitudes to first calculate the free membrane potential distribution and then obtain the firing rate by solving the first passage time problem numerically 19 ., This approach may be extendable to conductance based synapses 20 ., Exact results for the steady state have so far only been presented for the case of exponentially distributed synaptic amplitudes 21 ., The spike threshold renders the model an extremely non-linear unit ., However , if the synaptic input signal under consideration is small compared to the total synaptic barrage , a linear approximation captures the main characteristics of the evoked response ., In this scenario all remaining inputs to the neuron are treated as background noise ( see Figure 1A ) ., Calculations of the linear response kernel in the diffusion limit suggested that the integrate-and-fire model acts as a low-pass filter 22 ., Here spectrum and amplitude of the synaptic background input are decisive for the transient properties of the integrate-and-fire model: in contrast to white noise , low-pass filtered synaptic noise leads to a fast response in the conserved linear term 12 ., Linear response theory predicts an optimal level of noise that promotes the response 23 ., In the framework of spike-response models , an immediate response depending on the temporal derivative of the postsynaptic potential has been demonstrated in the regime of low background noise 24 ., The maximization of the input-output correlation at a finite amplitude of additional noise is called stochastic resonance and has been found experimentally in mechanoreceptors of crayfish 25 , in the cercal sensory system of crickets 26 , and in human muscle spindles 27 ., The relevance and diversity of stochastic resonance in neurobiology was recently highlighted in a review article 28 ., Linear response theory enables the characterization of the recurrent dynamics in random networks by a phase diagram 22 , 29 ., It also yields approximations for the transmission of correlated activity by pairs of neurons in feed-forward networks 30 , 31 ., Furthermore , spike-timing dependent synaptic plasticity is sensitive to correlations between the incoming synaptic spike train and the firing of the neuron ( see Figure 1 ) , captured up to first order by the linear response kernel 32–38 ., For neuron models with non-linear membrane potential dynamics , the linear response properties 39 , 40 and the time-dependent dynamics can be obtained numerically 41 ., Afferent synchronized activity , as it occurs e . g . in primary sensory cortex 42 , easily drives a neuron beyond the range of validity of the linear response ., In order to understand transmission of correlated activity , the response of a neuron to fast transients with a multiple of a single synaptic amplitude 43 hence needs to be quantified ., In simulations of neuron models with realistic amplitudes for the postsynaptic potentials , we observed a systematic deviation of the output spike rate and the membrane potential distribution from the predictions by the Fokker-Planck theory modeling synaptic currents by Gaussian white noise ., We excluded any artifacts of the numerics by employing a dedicated high accuracy integration algorithm 44 , 45 ., The novel theory developed here explains these observations and lead us to the discovery of a new early component in the response of the neuron model which linear response theory fails to predict ., In order to quantify our observations , we extend the existing Fokker-Planck theory 46 and hereby obtain the mean time at which the membrane potential first reaches the threshold; the mean first-passage time ., The advantage of the Fokker-Planck approach over alternative techniques has been demonstrated 47 ., For non-Gaussian noise , however , the treatment of appropriate boundary conditions for the membrane potential distribution is of utmost importance 48 ., In the results section we develop the Fokker-Planck formalism to treat an absorbing boundary ( the spiking threshold ) in the presence of non-zero jumps ( postsynaptic potentials ) ., For the special case of simulated systems propagated in time steps , an analog theory has recently been published by the same authors 49 , which allows to assess artifacts introduced by time-discretization ., Our theory applied to the integrate-and-fire model with small but finite synaptic amplitudes 1 , introduced in section “The leaky integrate-and-fire model” , quantitatively explains the deviations of the classical theory for Gaussian white noise input ., After reviewing the diffusion approximation of a general first order stochastic differential equation we derive a novel boundary condition in section “Diffusion with finite increments and absorbing boundary” ., We then demonstrate in section “Application to the leaky integrate-and-fire neuron” how the steady state properties of the model are influenced: the density just below threshold is increased and the firing rate is reduced , correcting the preexisting mean first-passage time solution 10 for the case of finite jumps ., Turning to the dynamic properties , in section “Response to fast transients” we investigate the consequences for transient responses of the firing rate to a synaptic impulse ., We find an instantaneous , non-linear response that is not captured by linear perturbation theory in the diffusion limit and that displays marked stochastic resonance ., On the network level , we demonstrate in section “Dominance of the non-linear component on the network level” that the non-linear fast response becomes the most important component in case of feed-forward inhibition ., In the discussion we consider the limitations of our approach , mention possible extensions and speculate about implications for neural processing and learning ., Consider a leaky integrate-and-fire model 1 with membrane time constant and resistance receiving excitatory and inhibitory synaptic inputs , as they occur in balanced neural networks 50 ., We aim to obtain the mean firing rate and the steady state membrane potential distribution ., The input current is modeled by point events , drawn from homogeneous Poisson processes with rates and , respectively ., The membrane potential is governed by the differential equation ., An excitatory spike causes a jump of the membrane potential by , an inhibitory spike by , so , where is a constant background current ., Whenever reaches the threshold , the neuron emits a spike and the membrane potential is reset to , where it remains clamped for the absolute refractory time ., The approach we take is to modify the existing Fokker-Planck theory in order to capture the major effects of the finite jumps ., To this end , we derive a novel boundary condition at the firing threshold for the steady state membrane potential distribution of the neuron ., We then solve the Fokker-Planck equation obtained from the standard diffusion approximation 8 , 10 , 11 , 22 , 23 given this new condition ., The membrane potential of the model neuron follows a first order stochastic differential equation ., Therefore , in this section we consider a general first order stochastic differential equation driven by point events ., In order to distinguish the dimensionless quantities in this section from their counterparts in the leaky integrate-and-fire model , we denote the rates of the two incoming Poisson processes by ( excitation ) and ( inhibition ) ., Each incoming event causes a finite jump ( the excitatory synaptic weight ) for an increasing event and ( the inhibitory synaptic weight ) for a decreasing event ., The stochastic differential equation takes the form ( 1 ) where captures the deterministic time evolution of the system ( with for the leaky integrate-and-fire neuron ) ., We follow the notation in 46 and employ the Kramers-Moyal expansion with the infinitesimal moments ., The first and second infinitesimal moment evaluate to and , where we introduced the shorthand and ., The time evolution of the probability density is then governed by the Kramers-Moyal expansion , which we truncate after the second term to obtain the Fokker-Planck equation ( 2 ) where denotes the probability flux operator ., In the presence of an absorbing boundary at , we need to determine the resulting boundary condition for the stationary solution of ( 2 ) ., Without loss of generality , we assume the absorbing boundary at to be the right end of the domain ., A stationary solution exists , if the probability flux exiting at the absorbing boundary is reinserted into the system ., For the example of an integrate-and-fire neuron , reinsertion takes place due to resetting the neuron to the same potential after each threshold crossing ., This implies a constant flux through the system between the point of insertion and threshold ., Rescaling the density by this flux as results in the stationary Focker-Planck equation , which is a linear inhomogeneous differential equation of first order ( 3 ) with ., First we consider the diffusion limit , in which the rate of incoming events diverges , while the amplitude of jumps goes to zero , such that mean and fluctuations remain constant ., In this limit , the Kramers-Moyal expansion truncated after the second term becomes exact 51 ., This route has been taken before by several authors 8 , 22 , 23 , here we review these results to consistently present our extension of the theory ., In the above limit equation ( 3 ) needs to be solved with the boundary conditionsMoreover , a finite probability flux demands the density to be a continuous function , because of the derivative in the flux operator ., In particular , the solution must be continuous at the point of flux insertion ( however , the first derivative is non-continuous at due to the step function in the right hand side of ( 3 ) ) ., Continuity especially implies a vanishing density at threshold ., Once the solution of ( 3 ) is found , the normalization condition determines the stationary flux ., Now we return to the problem of finite jumps ., We proceed along the same lines as in the diffusion limit , seeking the stationary solution of the Fokker-Planck equation ( 2 ) ., We keep the boundary conditions at and at as well as the normalization condition as before , but we need to find a new self-consistent condition at threshold , because the density does not necessarily have to vanish if the rate of incoming jumps is finite ., The main assumption of our work is that the steady state solution satisfies the stationary Fokker-Planck equation ( 3 ) based on the diffusion approximation within the interval , but not necessarily at the absorbing boundary , where the solution might be non-continuous ., To obtain the boundary condition , we note that the flux over the threshold has two contributions , the deterministic drift and the positive stochastic jumps crossing the boundary ( 4 ) ( 5 ) with ., To evaluate the integral in ( 5 ) , for small we expand into a Taylor series around ., This is where our main assumption enters: we assume that the stationary Fokker-Planck equation ( 3 ) for is a sufficiently accurate characterization of the jump diffusion process ., We solve this equation for It is easy to see by induction , that the function and all its higher derivatives , can be written in the form , whose coefficients for obey the recurrence relation ( 6 ) with the additional values and , as denotes the function itself ., Inserting the Taylor series into ( 5 ) and performing the integration results in ( 7 ) which is the probability mass moved across threshold by a perturbation of size and hence also quantifies the instantaneous response of the system ., After dividing ( 4 ) by we solve for to obtain the Dirichlet boundary condition ( 8 ) If is small compared to the length scale on which the probability density function varies , the probability density near the threshold is well approximated by a Taylor polynomial of low degree; throughout this work , we truncate ( 7 ) and ( 12 ) at ., The boundary condition ( 8 ) is consistent with in the diffusion limit , in which the rate of incoming jumps diverges , while their amplitude goes to zero , such that the first ( ) and second moment ( ) stay finite ., This can be seen by scaling , , with such that the mean is kept constant 51 ., Inserting this limit in ( 8 ) , we find ( 9 ) since , and vanishes for , is bounded and ., The general solution of the stationary Fokker-Planck equation ( 3 ) is a sum of a homogeneous solution that satisfies and a particular solution with ., The homogeneous solution is , where we fixed the integration constant by chosing ., The particular solution can be obtained by variation of constants and we chose it to vanish at the threshold as ., The complete solution is a linear combination , where the prefactor is determined by the boundary condition ( 8 ) in the case of finite jumps , or by for Gaussian white noise The normalization condition determines the as yet unknown constant probability flux through the system ., We now apply the theory developed in the previous section to the leaky integrate-and-fire neuron with finite postsynaptic potentials ., Due to synaptic impulses , the membrane potential drifts towards and fluctuates with the diffusion constant ., This suggests to choose the natural units for the time and for the voltage to obtain the simple expressions for the drift- and for the diffusion-term in the Fokker-Planck operator ( 2 ) ., The probability flux operator ( 2 ) is then given as ., In the same units the stationary probability density scaled by the flux reads where is the flux corresponding to the firing rate in units of ., As is already scaled by the flux , application of the flux operator yields unity between reset and threshold and zero outside ( 10 ) The steady state solution of this stationary Fokker-Planck equation ( 11 ) is a linear superposition of the homogeneous solution and the particular solution ., The latter is chosen to be continuous at and to vanish at ., Using the recurrence ( 6 ) for the coeffcients of the Taylor expansion of the membrane potential density , we obtain and , where starts from ., The first important result of this section is the boundary value of the density at the threshold following from ( 8 ) as ( 12 ) The constant in ( 11 ) follows from ., The second result is the steady state firing rate of the neuron ., With being the fraction of neurons which are currently refractory , we obtain the rate from the normalization condition of the density as ( 13 ) The normalized steady state solution Figure 2A therefore has the complete form ( 14 ) Figure 2B , D shows the steady state solution near the threshold obtained by direct simulation to agree much better with our analytical approximation than with the theory for Gaussian white noise input ., Even for synaptic amplitudes ( here ) which are considerably smaller than the noise fluctuations ( here ) , the effect is still well visible ., The oscillatory deviations with periodicity close to reset observable in Figure 2A are due to the higher occupation probability of voltages that are integer multiples of a synaptic jump away from reset ., The modulation washes out due to coupling of adjacent voltages by the deterministic drift as one moves away from reset ., The oscillations at lower frequencies apparent in Figure 2A are due to aliasing caused by the finite bin width of the histogram ( ) ., The synaptic weight is typically small compared to the length scale on which the probability density function varies ., So the probability density near the threshold is well approximated by a Taylor polynomial of low degree; throughout this work , we truncate the series in ( 12 ) at ., A comparison of this approximation to the full solution is shown in Figure 2E ., For small synaptic amplitudes ( shown ) , below threshold and outside the reset region ( Figure 2A , C ) the approximation agrees with the simulation within its fluctuation ., At the threshold ( Figure 2B , D ) our analytical solution assumes a finite value whereas the direct simulation only drops to zero on a very short voltage scale on the order of the synaptic amplitude ., For larger synaptic weights ( , see Figure 2F ) , the density obtained from direct simulation exhibits a modulation on the corresponding scale ., The reason is the rectifying nature of the absorbing boundary: A positive fluctuation easily leads to a threshold crossing and absorption of the state in contrast to negative fluctuations ., Effectively , this results in a net drift to lower voltages within the width of the jump distribution caused by synaptic input , visible as the depletion of density directly below the threshold and an accumulation further away , as observed in Figure 2F ., The second term ( proportional to ) appearing in ( 13 ) is a correction to the well known firing rate equation of the integrate-and-fire model driven by Gaussian white noise 10 ., Figure 3 compares the firing rate predicted by the new theory to direct simulation and to the classical theory ., The classical theory consistently overestimates the firing rate , while our theory yields better accuracy ., Our correction resulting from the new boundary condition becomes visible at moderate firing rates when the density slightly below threshold is sufficiently high ., At low mean firing rates , the truncation of the Kramers-Moyal expansion employed in the Fokker-Planck description may contribute comparably to the error ., Our approximation captures the dependence on the synaptic amplitude correctly for synaptic amplitudes of up to ( Figure 3B ) ., The insets in Figure 3C , D show the relative error of the firing rate as a function of the noise amplitude ., As expected , the error increases with the ratio of the, synaptic effect compared to the amplitude of the noise fluctuations ., For low noise , our theory reduces the relative error by a factor of compared to the classical diffusion approximation ., We now proceed to obtain the response of the firing rate to an additional -shaped input current ., Such a current can be due to a single synaptic event or due to the synchronized arrival of several synaptic pulses ., In the latter case , the effective amplitude of the summed inputs can easily exceed that of a single synapse ., The fast current transient causes a jump of the membrane potential at and ( 2 ) suggests to treat the incident as a time dependent perturbation of the mean input ., First , we are interested in the integral response of the excess firing rate ., Since the perturbation has a flat spectrum , up to linear order in the spectrum of the excess rate is , where is the linear transfer function with respect to perturbing at Laplace frequency ., In particular , ., As is the DC susceptibility of the system , we can express it up to linear order as ., Hence , ( 15 ) We also take into account the dependence of on to calculate from ( 13 ) and obtain ( 16 ) Figure 4D shows the integral response to be in good agreement with the linear approximation ., This expression is consistent with the result in the diffusion limit : Here the last term becomes , where we used , following from ( 10 ) with ., This results in , which can equivalently be obtained directly as the derivative of ( 13 ) with respect to setting ., Taking the limit , however , does not change significantly the integral response compared to the case of finite synaptic amplitudes ( Figure 4D , Figure 5A ) ., The instantaneous response of the firing rate to an impulse-like perturbation can be quantified without further approximation ., The perturbation shifts the probability density by so that neurons with immediately fire ., This results in the finite firing probability of the single neuron within infinitesimal time ( 5 ) , which is zero for ., This instantaneous response has several interesting properties: For small it can be approximated in terms of the value and the slope of the membrane potential distribution below the threshold ( using ( 7 ) for ) , so it has a linear and a quadratic contribution in ., Figure 4A shows a typical response of the firing rate to a perturbation ., The peak value for a positive perturbation agrees well with the analytical approximation ( 7 ) ( Figure 4C ) ., Even in the diffusion limit , replacing the background input by Gaussian white noise , the instantaneous response persists ., Using the boundary condition our theory is applicable to this case as well ., Since the density just below threshold is reduced , ( 5 ) yields a smaller instantaneous response ( Figure 4C , Figure 5B ) which for positive still exhibits a quadratic , but no linear , dependence ., The increasing and convex dependence of the response probability on the amplitude of the perturbation is a generic feature of neurons with subthreshold mean input that also persists in the case of finite synaptic rise time ., In this regime , the membrane potential distribution has a mono-modal shape centered around the mean input , which is inherited from the underlying superposition of a large number of small synaptic impulses ., The decay of the density towards the threshold is further enhanced by the probability flux over the threshold: a positive synaptic fluctuation easily leads to the emission of a spike and therefore to the absorption of the state at the threshold , depleting the density there ., Consequently , the response probability of the neuron is increasing and convex as long as the peak amplitude of the postsynaptic potential is smaller than the distance of the peak of the density to the threshold ., It is increasing and concave beyond this point ., At present the integrate-and-fire model is the simplest analytically tractable model with this feature ., The integral response ( 15 ) as well as the instantaneous response ( 5 ) both exhibit stochastic resonance; an optimal level of synaptic background noise enhances the transient ., Figure 5A shows this noise level to be at about for the integral response ., The responses to positive and negative perturbations are symmetric and the maximum is relatively broad ., The instantaneous response in Figure 5B displays a pronounced peak at a similar value of ., This non-linear response only exists for positive perturbations; the response is zero for negative ones ., Though the amplitude is reduced in the case of Gaussian white noise background , the behavior is qualitatively the same as for noise with finite jumps ., Stochastic resonance has been reported for the linear response to sinusoidal periodic stimulation 23 ., Also for non-periodic signals that are slow compared to the neurons dynamics an adiabatic approximation reveals stochastic resonance 52 ., In contrast to the latter study , the rate transient observed in our work is the instantaneous response to a fast ( Dirac ) synaptic current ., Due to the convex nature of the instantaneous response ( Figure 4C ) its relative contribution to the integral response increases with ., For realistic synaptic weights the contribution reaches percent ., An example network in which the linear non-instantaneous response cancels completely and the instantaneous response becomes dominant is shown in Figure 6A ., At two populations of neurons simultaneously receive a perturbation of size and respectively ., This activity may , for example , originate from a third pool of synchronous excitatory and inhibitory neurons ., It may thus be interpreted as feed-forward inhibition ., The linear contributions to the pooled firing rate response of the former two populations hence is zero ., The instantaneous response , however , causes a very brief overshoot at ( Figure 6B ) ., Figure 6C reveals that the response returns to baseline within ., Figure 6D shows that the dependence of peak height on still exhibits the supra-linearity ., The quite exact cancellation of the response for originates from the symmetry of the response functions for positive and negative perturbations in this interval ( shown in Figure 4A , B ) ., The pooled firing rate of the network is the sum of the full responses: the instantaneous response at does not share the symmetry and hence does not cancel ., This demonstrates that the result of linear perturbation theory is a good approximation for and that the instantaneous response at the single time point completes the characterization of the neuronal response ., In this work we investigate the effect of small , but non-zero synaptic impulses on the steady state and response properties of the integrate-and-fire neuron model ., We obtain a more accurate description of the firing rate and the membrane potential distribution in the steady state than provided by the classical approximation of Gaussian white noise input currents 10 ., Technically this is achieved by a novel hybrid approach combining a diffusive description of the membrane potential dynamics far away from the spiking threshold with an explicit treatment of threshold crossings by synaptic transients ., This allows us to obtain a boundary condition for the membrane potential density at threshold that captures the observed elevation of density ., Our work demonstrates that in addition to synaptic filtering , the granularity of the noise due to finite non-zero amplitudes does affect the steady state and the transient response properties of the neuron ., Here , we study the effect of granularity using the example of a simple neuron model with only one dynamic variable ., The quantitatively similar increase of the density close to threshold observed if low-pass filtered Gaussian white noise is used as a model for the synaptic current has a different origin ., It is due to the absence of a diffusion term in the dynamics of the membrane potential 12 , 13 , 15 ., The analytical treatment of finite synaptic amplitudes further allows us to characterize the probability of spike emission in response to synaptic inputs for neuron models with a single dynamical variable and renewal ., Alternatively , this response can be obtained numerically from population descriptions 18 , 39–41 or , for models with one or more dynamic variables and gradually changing inputs , in the framework of the refractory density approximation 15 ., Here , we find that the response can be decomposed into a fast , non-linear and a slow linear contribution , as observed experimentally about a quarter of a century ago 53 in motor neurons of cat cortex in the presence of background noise ., The existence of a fast contribution proportional to the temporal change of the membrane potential was predicted theoretically 54 ., In the framework of the refractory density approach 15 , the effective hazard function of an integrate-and-fire neuron also exhibits contributions to spike emission due to two distinct causes: the diffusive flow through the threshold and the movement of density towards the threshold ., The latter contribution is proportional to the temporal change of the membrane potential and is corresponding to the instantaneous response reported here , but for the case of a gradually increasing membrane potential ., Contemporary theory of recurrent networks so far has neglected the transient non-linear component of the neural response , an experimentally observed feature 53 that is generic to threshold units in the presence of noise ., The infinitely fast rise of the postsynaptic potential in the integrate-and-fire model leads to the immediate emission of a spike with finite probability ., For excitatory inputs , this probability depends supra-linearly on the amplitude of the synaptic impulse and it is zero for inhibitory impulses ., The supra-linear increase for small positive impulse amplitudes relates to the fact that the membrane potential density decreases towards threshold: the probability to instantaneously emit a spike equals the integral of the density shifted over the threshold ., The detailed shape of the density below threshold therefore determines the response properties ., For Gaussian white noise synaptic background , the model still displays an instantaneous response ., However , since in this case the density vanishes at threshold , the response probability to lowest order grows quadratically in the amplitude of a synaptic impulse ., This is the reason why previous work based on linear response theory did not report on the existence of an instantaneous component when modulating the mean input and on the contrary characterized the nerve cell as a low-pass in this case 22 , 23 ., Modulation of the noise amplitude , however , has been shown to cause an instantaneous response in linear approximation in the diffusion limit 23 , confirmed experimentally in real neurons 55 ., While linear response theory has proven extremely useful to understand recurrent neural networks 29 , the categorization of the integrate-and-fire neurons response kernel as a low-pass is misleading , because it suggests the absence of an immediate response ., Furthermore we find that in addition to the nature of the background noise , response properties also depend on its amplitude: a certain level of noise optimally promotes the spiking response ., Hence noise facilitates the transmission of the input to the output of the neuron ., This is stochastic resonance in the general sense of the term as recently suggested 28 ., As noted in the introduction , stochastic resonance of the linear response kernel has previously been demonstrated for sinusoidal input currents and Gaussian white background noise 23 ., Furthermore , also slow aperiodic transients are facilitated by stochastic resonance in the integrate-and-fire neuron 52 ., We extend the known results in two respects ., Firstly , we show that the linear response shows aperiodic stochastic resonance also for fast transients ., Secondly , we demonstrate tha | Introduction, Model, Results, Discussion | Contemporary theory of spiking neuronal networks is based on the linear response of the integrate-and-fire neuron model derived in the diffusion limit ., We find that for non-zero synaptic weights , the response to transient inputs differs qualitatively from this approximation ., The response is instantaneous rather than exhibiting low-pass characteristics , non-linearly dependent on the input amplitude , asymmetric for excitation and inhibition , and is promoted by a characteristic level of synaptic background noise ., We show that at threshold the probability density of the potential drops to zero within the range of one synaptic weight and explain how this shapes the response ., The novel mechanism is exhibited on the network level and is a generic property of pulse-coupled networks of threshold units . | Our work demonstrates a fast-firing response of nerve cells that remained unconsidered in network analysis , because it is inaccessible by the otherwise successful linear response theory ., For the sake of analytic tractability , this theory assumes infinitesimally weak synaptic coupling ., However , realistic synaptic impulses cause a measurable deflection of the membrane potential ., Here we quantify the effect of this pulse-coupling on the firing rate and the membrane-potential distribution ., We demonstrate how the postsynaptic potentials give rise to a fast , non-linear rate transient present for excitatory , but not for inhibitory , inputs ., It is particularly pronounced in the presence of a characteristic level of synaptic background noise ., We show that feed-forward inhibition enhances the fast response on the network level ., This enables a mode of information processing based on short-lived activity transients ., Moreover , the non-linear neural response appears on a time scale that critically interacts with spike-timing dependent synaptic plasticity rules ., Our results are derived for biologically realistic synaptic amplitudes , but also extend earlier work based on Gaussian white noise ., The novel theoretical framework is generically applicable to any threshold unit governed by a stochastic differential equation driven by finite jumps ., Therefore , our results are relevant for a wide range of biological , physical , and technical systems . | biophysics/theory and simulation, neuroscience/theoretical neuroscience, computational biology/computational neuroscience | null |
journal.pcbi.1004833 | 2,016 | A Theoretical Model of Jigsaw-Puzzle Pattern Formation by Plant Leaf Epidermal Cells | Throughout growth and differentiation , plant cells display various shapes that are primarily determined by the cell wall 1 , 2 ., Leaf epidermal cells in dicotyledonous plants have jigsaw puzzle–like shapes with winding cell wall 3 , 4 ., Prior to leaf expansion , epidermal cells have a simple rectangular shape ., The cell wall begins to wind during leaf expansion , forming interdigitated cell patterns 5–7 ., In the cotyledons of Arabidopsis thaliana , significant winding is observed for approximately one week after seed sowing ( Fig 1 ) ., Both the cell volume and total cell wall length of an epidermal cell increase as it changes in shape ., During this process , the thickness of the cell wall remains mostly unchanged , but traditional transmission electron microscopic observations suggest that the cell wall becomes slightly thicker as an accompaniment to cortical microtubule accumulation in the winding zone 8 ., Cell wall interdigitation is regulated by two Rho-like GTPases from plants ( ROPs ) , ROP2 and ROP6 5 , 9 , 10 ., ROP2 and ROP6 have opposing activities; the activity of ROP2 dominates under low auxin concentrations , whereas ROP6 activity becomes dominant under high auxin concentrations 10 ., ROP2 localizes to protrusions of epidermal cells and promotes localization of diffuse F-actin , which enhances outgrowth via targeted exocytosis 5 , 11 as observed in tip growth 12 ., In contrast , ROP6 localizes to the concave region and accumulates cortical microtubules 13 that likely restrict cell expansion via cell wall reinforcement 8 , 14 ., The general mechanisms underlying the formation of similar biological patterns have been examined using a reaction–diffusion framework ., Pattern formation by the reaction–diffusion system has been widely investigated in the field of mathematical biology 15 , and the findings obtained were recently used by developmental biologists 16 ., The dynamics of winding of a band-like structure , which are similar to the dynamics of interdigitation of the plant cell wall , have been modeled using the FitzHugh–Nagumo equation 17 , 18 ., The dynamics of winding are controlled by the combination of two mechanisms: maintenance of the band-like shape by the interaction of two interfaces and formation of curvature due to interface instability ., This mechanism has been applied to the interdigitation of the junctions between bones in human skulls 19 ., In the present study , we formulate a theoretical model to reproduce pattern formation by plant leaf epidermal cells ., This model assumes that the interdigitating pattern arises as a result of cell wall remodeling , and reproduces the maintenance of cell wall thickness and formation of a jigsaw puzzle–like pattern in vivo ., To monitor epidermal cell morphogenesis , time-lapse imaging of the cotyledon surface was performed with A . thaliana seedlings as described previously 20 ., Sterilized seeds expressing the plasma membrane marker GFP-PIP2a 21 were immersed in distilled water at 4°C for 2 days , and the seed coats were then carefully removed under a stereo microscope ( SZX12 , Olympus , Tokyo , Japan ) ., The naked cotyledons were mounted on a chamber slide ( Iwaki Co . , Ltd , Tokyo , Japan ) and covered with 1/2-strength Murashige–Skoog medium agar gel ( 2 . 3 g L−1 Murashige and Skoog Plant Salt Mixture , pH 5 . 8 from Wako Pure Chemical Industries , Osaka , Japan ) ., The chamber slides were placed in growth chambers at 23 . 5°C , with a 12-h light/12-h dark cycle , using 100 μmol m−2 s−1 white light ., For acquiring images , the chamber slide was placed onto the inverted platform of a fluorescence microscope ( IX70 , Olympus ) equipped with a UPlanFl 20×/0 . 50 objective lens and spinning disc confocal unit ( CSU10 , Yokogawa Electric Co . , Ltd , Tokyo , Japan ) , together with a cooled CCD camera head system ( CoolSNAP HQ; Photometrics , Huntington Beach , Canada ) ., Sterilized A . thaliana seeds expressing GFP-PIP2a 21 were immersed in 1/2-strength Murashige-Skoog media solution ( 2 . 3 g L−1 Murashige and Skoog Plant Salt Mixture , pH 5 . 8 from Wako Pure Chemical Industries ) supplemented with or without 1 . 0% cellulase ( Cellulase Y-C; Kyowa Chemical Products Co . , Ltd , Osaka , Japan ) in 24-well plates ( Sumitomo Bakelite Co . , Ltd , Tokyo , Japan ) ., The seeds were cultured for one week in growth chambers at 23 . 5°C , with a 12-h light/12-h dark cycle using 100 μmol m−2 s−1 white light , and then observed with a confocal laser scanning microscope ( FV300 , Olympus ) ., To observe the cell wall ultrastructure , we observed the lateral cell wall of cotyledon epidermal cells with transmission electron microscopy ., Cotyledon samples were fixed with 2% paraformaldehyde and 2% glutaraldehyde in 0 . 05 M cacodylate buffer ( pH 7 . 4 ) at 4°C overnight ., After fixation , the samples were rinsed three times with 0 . 05 M cacodylate buffer for 30 min each , followed by post fixation with 2% osmium tetroxide in 0 . 05 M cacodylate buffer at 4°C for 3 hours ., The samples were dehydrated through a graded ethanol series ( 50% ethanol for 30 min at 4°C , 70% ethanol for 30 min at 4°C , 90% for 30 min at room temperature , and 4 changes of 100% for 30 min each at room temperature ) ., Afterwards , the samples were continuously dehydrated with 100% ethanol at room temperature overnight ., The samples were infiltrated with propylene oxide twice for 30 min each and then placed into a 70:30 mixture of propylene oxide and resin ( Quetol-651; Nisshin EM Co . , Tokyo , Japan ) for 1 hour ., The cap of the tube was left open and propylene oxide was evaporated overnight ., The samples were transferred to fresh 100% resin , and polymerized at 60°C for 48 hours ., 80 nm sections were sliced from the blocks using an ultramicrotome equipped with a diamond knife ( ULTRACUT UCT; Leica , Tokyo , Japan ) , and sections were placed on copper grids ., They were stained with 2% uranyl acetate at room temperature for 15 minutes , rinsed with distilled water , and counter-stained with lead stain solution ( Sigma-Aldrich Co . , Tokyo , Japan ) at room temperature for 3 minutes ., The grids were observed under a transmission electron microscope ( JEM-1400Plus; JEOL , Ltd . , Tokyo , Japan ) at an acceleration voltage of 80 kV ., Images were taken with a CCD camera ( VELETA; Olympus ) ., Lateral cell wall thickness was measured at the thinnest point between two three-way junctions to avoid errors due to the direction of the cuts ., We obtained cell contour images of GFP-PIP2a-expressing plants 21 or rsw2/kor1 mutant lines 22 , 23 stained with the fluorescent dye FM4-64 ., The images obtained were thresholded by pixel intensity and skeletonized to segment the cell wall pattern ., As inhomogeneous fluorescence signal was occasionally observed , we manually corrected defects in the cell wall pattern in the segmented images ., We then extracted all cells in the upper leaf regions and measured the cell area ., To quantify the ratio of the wavenumber of the cell wall , we used a G-type Fourier descriptor , which generates a power spectrum from a closed curved shape , such as a two-dimensional representation of a leaf ( S1 Fig ) ., The angles of three-way junctions in each cell were measured at points that were 12 pixels from the central pixel of the junction , as shown in S2 Fig . Deviation of the angle from 120° was evaluated by the root-mean-square deviation ( RMSD ) of each cell as follows:, RMSD=\u20091NΣi=1N ( θi−120o ) 2, ( 1 ), where θi is the angle of the i-th three-way junction in the cell and N is the number of junctions in the cell ., To evaluate the density of anticlinal cortical microtubules in epidermal cells , we observed the cotyledon surfaces of transgenic A . thaliana plants expressing GFP-tubulin 24 , 25 8 days after sowing in 1/2-strength Murashige-Skoog solution ., The three-way junctions and points of interdigitation were manually determined from GFP-tubulin maximum intensity projection images obtained from serial optical sections at 0 . 5-μm steps along Z-axis ., Following this , GFP intensity peaks were semi-automatically detected as anticlinal microtubules within circles centered at the manually assigned points with a radius of 5 μm ., The density of the anticlinal microtubules was calculated as the number of GFP intensity peaks per cell surface length ., We formulated a model that incorporated local remodeling of the cell wall in an attempt to understand pattern formation within the cell wall during interdigitation ( Fig 2a ) ., Our model includes the following assumptions: We initially defined the indicator variable u ( x , y ) , which represents the structure at a certain location ( x , y ) ., We defined the u = 1 region as the cytoplasm and the u = 0 region as the cell wall ., v represented the local signaling molecule concentration ., We then defined the interface speed V as follows:, V\u2009=\u2009f ( v ) −\u2009σκ, ( 2 ), This equation means that local remodeling of the cell wall is a function of the local signaling molecule concentration and curvature of the cell wall ., We represented the effects of the signaling molecule as f ( v ) , where v ( x , y ) is the spatial distribution of signaling molecule ., At interface points where the local signaling molecule concentration is high , ROP6 becomes active and the cell wall is degraded , resulting in V < 0 ., Conversely , if the local concentration of signaling molecule is low , ROP2 becomes dominant and cell wall is produced , resulting in V > 0 ., As the cell wall is elastic , we also introduced the surface tension term σk , which inhibits the formation of pointy structures at the interface ., This type of interface equation can be calculated using the phase field method ., We then used a convolution kernel to implement the effects of signaling molecule on cell wall remodeling ., Cell wall interdigitation is relatively slow compared to the diffusion and degradation of typical signaling molecules , taking approximately one week to produce the final jigsaw puzzle–like pattern ( Fig 1 ) ., Therefore , we assumed that the distribution of signaling molecule was in a quasi-steady state and calculated its distribution separately ., We can calculate the distribution of signaling molecule by solving the diffusion equation; however , to simplify the model , we used a convolution kernel ., We defined the convolution kernel k , which represents the effects of a small piece of the cytoplasm on the distribution of signaling molecule ., The distribution v ( x , y ) was calculated by, v\u2009=\u2009k\u2009⊗\u2009u, ( 3 ), where ⊗ represents convolution ., For simplicity , we used the kernel shape, k=1/ ( πr2 ) ( x2+y2<r02 ), ( 4 ), k=0 ( x2+y2≥r02 ), ( 5 ), where r0 represents the effective range of signaling molecule influence ., The interface equation model was implemented using the phase field method 26 ., In this method , the interface equation, V\u2009=\u2009f ( v ) –\u2009σκ, ( 6 ), was calculated by the Allen–Cahn equation, u′\u2009=\u2009a\u2009u ( 1−u ) ( u−1/2\u2009+\u2009b\u2009f ( v ) ) +\u2009duΔu ., ( 7 ), A numerical simulation was implemented in Mathematica ( Wolfram Research Inc . , Champaign , U . S . A . , S1 Code . ) ., An implicit method was used to calculate the diffusion term of the model ., Convolution was calculated using Fourier transformation ., The source code of all numerical simulations is available as electronic supplemental data ., To observe pattern formation in interdigitating cotyledon epidermal cells , we used A . thaliana plants expressing a fluorescently tagged plasma membrane protein , GFP-PIP2a ., Two days after sowing , epidermal cells were rectangular ( Fig 1a; day 2 ) ., As the plant developed over 2–5 days and cells grew , cell wall gradually became bent , resulting in a jigsaw puzzle–like pattern ( Fig 1a; days 4 , 7 ) ., Cell wall interdigitation was confirmed by time-lapse imaging ( Fig 1b ) ., We performed a numerical simulation with our model using an appropriate initial distribution of cell wall ( Fig 2b ) ., Our simulation recapitulated interdigitation ( Fig 2c and 2d ) ., Most parts of the cell wall retain a uniform thickness , but the thickness is variable in interdigitated regions ( Fig 2c ) ., Winding is not observed in the cell wall of small stomatal lineage cells ( Fig 2c and 2d , asterisk ) ., According to our model , the convex interfaces of interdigitated cells had a lower concentration of signaling molecule , while the concave interfaces had a higher concentration ( Fig 2d ) ., As the concentration of signaling molecule is expected to influence the ratio of ROP2 and ROP6 signaling activities , the observed differences in our simulation may reflect in vivo modulation of ROP activity ., We performed a linear stability analysis of the model ., The growth speed λ and wavenumber k are represented as, λ=aπr2 ( ς ( k , r ) +\u2009ϕ ( k ) cos\u2009ψ−2σ ) −bk2, ( 8 ), Where, ϕ ( k ) =\u20092\u2009sink\u2009σk, ( 9 ), and, ς ( k , r ) =\u20092\u2009r ( 1\u2009–sin\u2009krkr ), ( 10 ), 27 ., ψ represents the phase difference between two interphases , and λ achieves maximum when ψ = 0 ., We plotted the relationship between λ and k , which shows a vault-like shape ( Fig 2e ) ., This shape indicates that a specific wavenumber was selected at the onset of pattern formation ., The maximum value of this curve corresponds to the fastest growing wavenumber of the pattern ., This analysis allows us to predict the size of the structure ( as determined by the fastest growing wavenumber ) , but phase of the pattern is dependent on initial small perturbations and thus not predictable ., To directly compare the predictions of the simulation with our experimental observations , we calculated the ratio of the power spectra at two different time points of growth ., We defined the time interval between the two stages as Δt , such that the ratio of power spectra was anticipated to be, u0eλ ( k ) ( t+Δt ) u0eλ ( k ) t=\u2009eλ ( k ) Δt, ( 11 ), We calculated the G-type Fourier descriptor of each cell , and calculated the ratio between 2-day-old and 4-day-old seedlings , and 2-day-old and 7-day-old seedlings ( red and green in Fig 2f , respectively ) ., In the G-type descriptor , the spectrum at k = 1 represents growth of the cell and was not similar to the result described in Fig 2e ., In other wavenumbers , both distributions were vault-like , indicating that the dynamics observed in vivo are qualitatively similar to those predicted by the model ., To examine the influence of cell wall metabolism on pattern formation , we investigated the effects of cellulose synthase dysfunction in rsw2/kor1 mutant plants 22 , 23 or the enzymatic degradation of cellulose in wild-type plants treated with 1 . 0% cellulase for 7 days ., A thicker lateral cell wall was observed in cotyledon epidermal cells from rsw2/kor1 seedlings ( Fig 3a ) ., Cellulase treatment also resulted in a thicker cell wall ( Fig 3c ) ., We confirmed this with a statistical analysis ( Fig 3b and 3d ) ., These observations are seemingly counterintuitive but agree with previous reports about rsw2/kor1 mutants 23 ., It is previously suggested that deposition of other cell wall components such as pectin is dramatically increased to compensate for decreased cellulose 28 , 29 ., To model the effect of impaired cellulose deposition on pattern formation , we assumed that decreased cellulose in the cell wall permits increased diffusion of signaling molecule and thus increases its range of action ., By increasing the kernel radius r0 , our model reproduced increased cell wall thickness ( Fig 3e ) ., An intuitive explanation of this phenomenon is as follows: the thickness of the cell wall becomes stable when V = f ( v ) = 0 ., f ( v ) approaches 0 when the effects of adjacent cells are balanced by the intrinsic effects of cell wall degradation ., Therefore , if the distance at which the signaling molecule is effective increases , the cell wall becomes thicker ( Fig 3f ) ., Cell wall curvature was decreased in both rsw2/kor1 mutants and cellulase-treated seedlings ( Fig 4a and 4b ) ., We also reproduced this decrease in cell wall curvature and increase in the characteristic length of the pattern by increasing the range of the effect r0 ( Fig 4c ) ., The cell wall in young epidermis displayed a brick wall–like pattern with T-shaped three-way junctions ., As development proceeded , the angles between each segment of cell wall at a three-way junction gradually approached 120° ( Fig 5a ) ., To quantify this , we automatically detected three-way junctions in cell wall and calculated the RMSD from 120° in the angles of junctions ., RMSD is at a minimum when all three angles at a three-way junction are 120° ., As development proceeds , cells become larger ( S3 Fig ) ., The RMSD of three-way junction angles in a population of smaller cells is biphasic , which represents three-way junctions with angles of 90° , 90° , and 180° ., The RMSD of three-way junction angles is smaller in larger cells , indicating that the angles at three-way junctions all gradually approach 120° during development ( Fig 5b ) ., We also observed this trend in our numerical simulation ( Fig 5c ) ., The angles at three-way junctions gradually became 120° over the course of the numerical simulation ., This symmetry can be intuitively explained by the fact that the effects of the cytoplasm are equivalent from all three adjacent cells , and , as a result , the patterns of equal angles are expected to be the most stable ., Based on this model , we were able to predict that three-way junctions were forced protrusions of the cytoplasm , resulting in decreased ROP6 activity in these regions ( Fig 6e ) ., To experimentally verify this prediction , we observed the distribution of cortical microtubules , which are known to be stabilized by ROP6 10 ( Fig 6a ) ., We used transgenic A . thaliana plants expressing GFP-tubulin to observe anticlinal cortical microtubules in interdigitated cells ., We semi-automatically detected GFP-tubulin signal intensity peaks and counted them in interdigitated regions ( Fig 6b ) and three-way junctions ( Fig 6c ) separately ., The density of anticlinal cortical microtubules was decreased in three-way junctions ( Fig 6d ) , consistent with the model prediction ., Our model postulates that the range at which a signaling molecule can exert its effects underlies the maintenance of cell wall thickness ., If a cell wall is too thick , the effects of signaling molecule on one side of the cell wall cannot reach the other side , resulting in a relatively ROP2 dominant state at that region ., The wall then retracts , resulting in a thinning of the cell wall ., In contrast , if the cell wall is too thin , the cell wall is strongly affected by signaling molecule produced in an adjacent cell , resulting in a ROP6 dominant state and thickening of the cell wall ., Cell wall thickness is thus kept constant by the balance of these two opposed mechanisms ., Therefore , the thickness of the wall is similar to the range at which signaling molecule can function ., Cellulase treatment may change the range of action of signaling molecule by changing the composition and thus diffusion coefficient of cell wall ., Compensatory production of other cell wall materials such as pectin may also result in thickening of cell wall , and experimental verification is necessary to distinguish these two mechanisms ., Our model also reproduces the formation of cell wall interdigitation ., We considered the case in which the cell wall is slightly bent ., Protruding cytoplasm near a concave region of the cell wall may be exposed to a lower concentration of signaling molecule because it is surrounded by less signaling molecule-producing cytoplasm ., Therefore , ROP2 becomes dominant at that point , resulting in further lobing of the cytoplasm ., In contrast , a convex area is exposed to a higher concentration of signaling molecule than a concave area because the area is surrounded by more signaling molecule-producing cytoplasm ., Therefore , ROP6 becomes dominant at that point , resulting in further retraction of the cytoplasm ., In a curved region of cell wall , both sides of the cell wall-cytoplasm interface tend to generate the same curvature , and as a result it , is difficult to retain the same cell wall thickness ., Therefore , cell wall thickness tends to vary in the curved regions ., Because only a specific wavenumber component grew in our model , large structures cannot grow in a small domain ., Stomatal guard cells are generally smaller than other epidermal cells and do not show a jigsaw puzzle–like pattern ., One possibility is that stomatal guard cells have a different cell wall composition that resists curving ., However , we observed jigsaw puzzle–like stomatal guard cells in a mutant line with giant stomatal guard cells 30 ., Our model does not generate the pattern when the domain size was below a certain threshold ., Therefore , the reason we do not observe pattern formation in wild-type stomatal guard cells is that these cells are too small to generate a pattern ., A mechanical factor may be also involved in pattern formation ., A previous study attempted to explain plant cell shape from a purely mechanical perspective 31 ., Pattern formation of suture tissue has also been previously explained from a primarily mechanical point of view 32 , 33 ., Our model included the surface tension term σk that represents the mechanical aspect of cell wall ., We also postulated that pattern formation by the cell wall may be due to buckling instability ( Takigawa-Imamura et al . , in prep ) ., It is possible to reproduce pattern formation with a mechanical model , but , in this case , we required a biological mechanism to increase the cell wall area while maintaining cell wall thickness ., In cell membranes , the thickness of the membrane is automatically determined by the unit size , whereas cell wall may have variable thickness ., Because cell size increases during development ( Fig 5b ) , we cannot assume that buckling is caused by a relative decrease in cell volume ., ROP GTPases self-organize to form patterns within a cell , even without significant changes in cell shape 34 ., Our model used cell geometry for interface instability , and did not include this mechanism ., If we assume that intracellular processes are the primary mechanism generating cell wall patterns , we need to implement a mechanism to keep cell wall thickness constant ., This mechanism and our model are not mutually exclusive ., A model that includes both mechanisms has been proposed in a different context 35 ., In this case , the intrinsic pattern formation mechanism modified the basic branched structure of Drosophila sensory neurons ., Our model explains interdigitation and maintenance of cell wall thickness , but it does not explain all aspects of pattern formation ., Therefore , continuous refinement of the model is necessary ., For example , the effect of the top wall is not considered in our model ., Locations of cell wall remodeling are likely correlated with the type of adjacent cytoskeleton , but we do not have direct experimental evidence to support this ., In addition , the molecular nature of the signaling molecule remains to be elucidated ., Auxin is a good candidate , but there are some inconsistent observations between its properties and the predictions of our model ., For example , our interface equation model assumed that signaling molecule acted at a very short range ( approximately 1 μm ) ., Auxin is , however , regarded as a long-range signal ., Imaging of morphogen diffusion dynamics during development has recently become possible in animal models 36 , 37 ., The observation of auxin diffusion dynamics is necessary to experimentally verify the interface equation model . | Introduction, Materials and Methods, Results, Discussion | Plant leaf epidermal cells exhibit a jigsaw puzzle–like pattern that is generated by interdigitation of the cell wall during leaf development ., The contribution of two ROP GTPases , ROP2 and ROP6 , to the cytoskeletal dynamics that regulate epidermal cell wall interdigitation has already been examined; however , how interactions between these molecules result in pattern formation remains to be elucidated ., Here , we propose a simple interface equation model that incorporates both the cell wall remodeling activity of ROP GTPases and the diffusible signaling molecules by which they are regulated ., This model successfully reproduces pattern formation observed in vivo , and explains the counterintuitive experimental results of decreased cellulose production and increased thickness ., Our model also reproduces the dynamics of three-way cell wall junctions ., Therefore , this model provides a possible mechanism for cell wall interdigitation formation in vivo . | It is well known that plant epidermal cells show beautiful jigsaw-puzzle like pattern ., However , mechanism of this pattern formation is not well understood ., In this study , we integrated known experimental information and mathematical modeling to reproduce the main features of the pattern formation—maintenance of cell wall thickness and formation of interdigitation ., Interestingly , the model is mathematically equivalent to the model of human skull suture interdigitation . | cell walls, plant anatomy, microtubules, plant growth and development, plant cell biology, plant embryo anatomy, organisms, plant cell walls, developmental biology, plant science, seedlings, plants, morphogenesis, cellular structures and organelles, pattern formation, cytoskeleton, plant embryogenesis, plant development, leaves, signal transduction, plant cells, cell biology, embryogenesis, signaling molecules, biology and life sciences, cellular types, cell signaling, cotyledons (botany), fruit and seed anatomy | null |
journal.pcbi.1004707 | 2,016 | Optimal Population-Level Infection Detection Strategies for Malaria Control and Elimination in a Spatial Model of Malaria Transmission | Malaria is a widespread infectious disease caused by Plasmodium parasites and leads to over half a million deaths each year , mostly in children under five years of age 1 ., As global burden has decreased dramatically over the past decade , local elimination of malaria is within sight for more and more endemic areas ., Regional elimination of malaria requires interrupting transmission between humans and mosquito vectors , and understanding the requirements for elimination is crucial for avoiding costly operations that are unlikely to succeed 2 , 3 ., Because the reservoir of malaria parasites lies in the human population , campaigns with antimalarial drugs can interrupt transmission under certain conditions 4 ., Testing such campaigns in the field is resource-intensive , and computational models have been used to describe how factors such as campaign coverage , local malaria transmission intensity , and individual compliance with drug regimens affect campaign outcomes 5–9 ., Mathematical modeling has shown that repeated annual campaigns of antimalarial drugs with high coverage can achieve local elimination in low- to moderate-transmission settings ., Modeling has also confirmed that mass drug administration ( MDA ) , where all individuals in a population are presumptively treated , can be substantially more effective than requiring positive diagnosis prior to treatment , as subpatent infections can constitute a substantial portion of the infectious reservoir 10 , 11 ., Although drug campaigns can be effective , large-scale interventions with antimalarials pose several potential drawbacks ., Dosing a large population will accelerate the emergence of drug-resistant parasites 12 , 13 ., Parasite resistance to both artemisinin and partner drug in artemisinin-based combination therapies has been observed in Southeast Asia , and spread of resistance to Africa would be catastrophic 14 ., Repeated rounds of campaigns can lead to community fatigue and widespread unnecessary suffering of drug side effects , and high community coverage has been shown to be vital to campaign success 4 , 5 ., Lastly , treating people who are uninfected and not at risk for infection is a waste of valuable resources ., Malaria transmission can be highly heterogeneous between neighboring villages and within the same village 15 ., Selective targeting of hotspots of transmission has been predicted to improve results of vector control interventions 16 ., However , it remains unknown whether current field diagnostics are adequate tools for defining hotspots , whether targeting hotspots with drug campaigns can achieve elimination , and how outcomes of targeted approaches compare with non-targeted approaches such as MDA ., Individuals living in malaria-endemic areas develop partial immunity to malaria , leading to asymptomatic , low-density infections that are difficult to detect but continue to infect mosquitoes 17 ., Mass screen-and-treat ( MSAT ) campaigns , where only individuals who test positive are treated with antimalarial drugs , have historically failed to achieve elimination due to the limited sensitivity of rapid diagnostic tests ( RDTs ) currently used in the field 18–21 ., In an ideal scenario , a cheap , fast , and sensitive field diagnostic would increase effectiveness of MSAT campaigns to near parity with MDA campaigns while avoiding overtreatment of uninfected individuals 22–24 ., In the absence of such a diagnostic , smart campaigns should be designed to treat as many subpatent infections as possible while simultaneously treating few uninfected individuals 25 ., Observed patterns in spatial heterogeneity in infection status may allow campaigns to effectively target subpatent infections based on proximity to an index case ., Since members of the same household and close neighbors likely experience similar exposure 21 , 26–28 , conducting a focal MDA ( fMDA ) around a confirmed-positive case may be a sound strategy for detecting subpatent infections in the face of limited sensitivity of RDTs 29 , 30 ., It remains unknown whether fMDAs can approach MDA campaigns’ effectiveness at interrupting transmission and how the size of the fMDA treatment area should be selected ., The amount of overtreatment that can be averted by conducting fMDA campaigns rather than MDA campaigns is also unknown ., Predicting both of these effects requires coupling spatial knowledge of regional heterogeneity in malaria exposure with a validated model of immunity acquisition in humans and transmission between humans and vectors ., Here we present for the first time modeling of malaria transmission on an operationally relevant scale accounting for household-scale levels of heterogeneity in transmission intensity and immunity ., We use a previously described model of malaria transmission , including a within-host model of immunity calibrated to age-stratified prevalence , incidence , and parasite density data from endemic settings , to estimate household exposure based on a spatial dataset of individual infection status from Southern Province , Zambia , where operations teams are currently carrying out mass drug campaigns 19 ., Campaigns with antimalarial drugs are simulated employing a variety of infection detection strategies , including MDA , MSAT , fMDA , reactive case detection ( RCD ) where fMDA is carried out around index cases of clinical malaria , pooled polymerase chain reaction ( PCR ) , and a hypothetical serological diagnostic ., We compare the strategies’ ability to avert clinical cases and interrupt transmission with minimal overtreatment of uninfected individuals ., A detailed survey of RDT prevalence by age and household was conducted in Southern Province , Zambia , in June-July 2012 at the end of the transmission season ( Fig 1A ) 19 ., Four representative health facility catchment areas ( HFCAs ) experiencing a wide range of malaria transmission intensity were selected for this analysis ., Overall RDT prevalence in the four HFCAs spanned 1 . 4 to 49% , but varied widely between households in the same HFCA , particularly in the Bbondo and Munyumbwe HFCAs where prevalence was intermediate ( Fig 1B ) ., In all four HFCAs , we observed clustering of RDT positive cases within households—individuals were more likely to be RDT positive if someone else in their household was RDT positive ( Fig 1C ) ., Individuals were also more likely to be RDT positive if someone within 50m , but not in their household , was RDT positive ., The clustering of RDT positivity within 50m held even in the Gwembe HFCA , suggesting that a small amount of endemic transmission persists in Gwembe and not all infections are imported ., To predict the outcome of various infection detection strategies on reducing transmission , we constructed a set of synthetic households for each HFCA made up of simulated individuals that reflected the geography and demographics of RDT positivity observed in the reference data for the HFCA ( Fig 1 , Table 1 , S1 Fig ) ., We assumed that members of each household experienced the same transmission intensity , but households within an HFCA could experience a different transmission intensity ., Transmission intensity of each household was determined by comparing the RDT prevalence by age of the household’s neighborhood to reference curves from simulations of known transmission intensity ( see Methods , S2 Fig ) ., In Gwembe and Sinamalima HFCAs , nearly all households experienced very low or very high transmission respectively ( Fig 2A ) ., Households in Bbondo and especially Munyumbwe HFCAs were more heterogeneous , and spatial patterns of high and low transmission intensity mirrored the household RDT positive rates ( Fig 2B ) ., The spatial clustering of RDT positivity within and near households suggested that fMDAs may be a good strategy for infection detection ., Since the asexual parasite density and infectiousness of each simulated individual was known , the true parasite prevalence and infectious potential of each HFCA could be estimated from the simulated households ( Fig 2C ) ., Infectious potential , a proxy for the infectious reservoir of malaria parasites in a human population , was defined as the number of mosquitoes that would be infected if 1000 mosquitoes were to feed on a village of 1000 people and accounts for heterogeneity in individual infectiousness due to parasite density and preference for mosquitoes to bite larger people ., We found that all four HFCAs had substantial rates of subpatent infection ., Low-density infections were four times as common as RDT-detectable infections in the Gwembe HFCA , where prevalence was the lowest , consistent with previous observations of low-density infections in low-prevalence seasonal settings 31 , 32 ., Under low-transmission conditions in our model , infections observed during June-July were three to six months old and past peak parasite levels ., In contrast , June-July infections under high transmission were more likely to be newer and composed of multiple infections , leading to higher parasite density ., Higher density infections acquired in low-transmission settings during the rainy season were also more likely to have triggered symptoms and hence treatment due to weaker host immunity ., Although low-density infections are less infectious than high-density infections , these subpatent infections comprised a substantial portion of the infectious potential in all four HFCAs ., Targeting infections with an RDT-based MSAT would thus be highly unlikely to lead to elimination at any level of transmission intensity ., Even improvement of RDT sensitivity by an order of magnitude from 100 parasites per μL to 10 parasites per μL would still leave a nontrivial amount of remaining infectious potential after an MSAT campaign ., Depleting the infectious reservoir was highly dependent on coverage and somewhat dependent on infection-detection strategy ( Fig 3A , Table 2 ) ., For all HFCAs , MDA was the most successful at depleting the infectious reservoir , MSAT the least successful , and other infection detection strategies fell in between MDA and MSAT ., Since dry season infections were more likely to be low-density in low-transmission settings than in high-transmission settings , MSAT was comparatively least effective at depleting the infectious reservoir at low prevalence , achieving only 40% of MDA’s effect in Gwembe HFCA , and most effective at high prevalence , achieving 70% of MDA’s effect in Sinamalima HFCA ., To evaluate the reduction in transmission after deploying drug campaigns , we estimated the expected number of new infections that would be seeded in humans due to vectors becoming infected in the first 30 days post-campaign ( Fig 3B ) ., We assumed that vectors tended to bite in the same neighborhood and that individuals who had received treatment during the campaign were protected from reinfection ( see Methods ) ., Compared with outcomes from a non-prophylactic drug ( S4 Fig ) , campaigns with a long-lasting prophylactic averted more new infections at moderate coverage , especially for MDA and other scenarios where a large fraction of the population was treated ., RDT-positive infections were more infectious than subpatent infections and , during the dry season , more likely to occur in households with a history of higher exposure ., MSAT campaigns and other RDT-dependent infection detection strategies were therefore more effective at averting new infections than might be predicted from their effectiveness at depleting the infectious reservoir relative to MDA campaigns ., At higher prevalence and higher coverage , fMDA strategies were just as effective as MDA at reducing onward transmission ., An order of magnitude improvement of RDT sensitivity from 100 parasites per μL to 10 parasites per μL was insufficient for increasing the efficacy of RDT-dependent infection detection strategies up to levels seen with MDA in low-prevalence areas ( S5 Fig ) ., In a control scenario where drug campaigns aim to reduce clinical incidence , we imagined that overtreatment was especially to be avoided , particularly in low-transmission settings , as it confers little benefit and may accelerate the rate of parasite resistance to antimalarial drugs ., Fig 4A shows receiver operating characteristic ( ROC ) curves of fraction of infected individuals treated ( true positive rate ) vs fraction of uninfected individuals treated ( false positive rate ) for each of the infection detection strategies ., MDA was agnostic to individual infection status , treating infected and uninfected individuals at the same rate as coverage increased ., MSAT could not treat uninfected individuals , and limited sensitivity of the RDT diagnostic resulted in at most 50% of infected individuals receiving treatment with an MSAT campaign; MSAT found the highest fraction of infected individuals in Sinamalima HFCA where prevalence was high ., The remaining infection detection strategies , the fMDAs and snowball RCD , fell between MSAT and MDA and in some cases exhibited favorable ROC curves , indicating a high rate of treating positives while minimizing treating negatives ., For fMDAs , ROC curves decreased in favorability with increasing HFCA prevalence ., Focal MDAs were successful at avoiding overtreatment in all but the highest-prevalence HFCAs ., Within-household fMDA and within-50m fMDA showed similar behavior , as households were sparse at 50m ( S1 Fig ) , while expanding the treatment radius to 200m resulted in much more overtreatment without capturing nearly as many additional infections ., Snowball reactive case detection resulted in more overtreatment per infection detected than within-200m fMDA in all four HFCAs ., To compare rates of clinical case prevention and overtreatment across HFCAs , we normalized populations to 1000 people and fixed coverage at 80% , a high but achievable rate ( Fig 4B ) ., Because transmission was so low in Gwembe HFCA , any mass campaign would avert only a handful of clinical cases: MSAT averted on average two clinical cases and MDA averted five , with the remaining infection detection strategies falling in between ., Yet an MDA campaign would result in treating over 700 people who were uninfected , and those individuals derived little benefit from prophylactic effects given the low rate of transmission ., Within-household fMDA , the infection-detection strategy that resulted in the least overtreatment next to MSAT , required overtreatment of nearly 50 individuals to avert less than one clinical case ., These high rates of overtreatment suggested that MSAT might be the only reasonable option for mass treatment for malaria control in low-prevalence areas despite MSAT’s relative inability to deplete the infectious reservoir ., In the Bbondo HFCA , within-household fMDA averted seven more clinical cases and overtreated 125 more people than MSAT ., Averting another five more clinical cases would require an MDA campaign that overtreated 475 more people ., In the Munyumbwe and Sinamalima HFCAs , within-household fMDA performed nearly as well as MDA at averting clinical cases , and for Munyumbwe , fMDA resulted in much lower numbers of overtreated people than MDA did ., In Sinamalima , rates of overtreatment with fMDA were nearly comparable to those of MDA , and within-household fMDA resulted in 20 more cases averted than MSAT ., In a high-prevalence site like Sinamalima , other factors such as costs or logistics would help decide whether fMDA or MDA is the best course of action ., The Gwembe and Bbondo HFCAs were considered for elimination scenarios , where drug campaigns aim to deplete the infectious reservoir such that transmission from humans to mosquitoes is interrupted ., In addition to the six infection detection strategies discussed above , we simulated pooled PCR , serological MSAT , and serological within-household fMDA to test strategies more appropriate for low-transmission regions ., The serological tests were modeled as hypothetical diagnostics that report whether an individual has experienced infection at any point in the previous twelve months ., We measured the probability of less than 1 new infection per 1000 people arising from vectors infected during the 30 days post-campaign as a proxy for elimination ., MDA was the most effective strategy for elimination , leading to high probability of less than 1 onward infection at lower coverage levels than the other infection detection strategies ( Fig 5 ) ., However , pooled PCR and serological diagnostics could also be highly effective as long as coverage was high ., MSAT with a serological diagnostic was especially promising as we predicted it to be efficient at avoiding overtreatment ., For pooled PCR , we grouped each HFCA into neighborhood pools consisting of 60–220 individuals per pool ( S3 Fig ) ., Individuals contributed a dried blood spot to a pooled sample , and MDA within the pool was triggered if the pooled sample tested positive ., In the Gwembe HFCA , pooled PCR led to lower overtreatment than MDA at the same level of coverage ., However , even at 100% coverage , pooled PCR could not reliably find all infections due to the detection limit of pooled PCR ., If a particular pool contained infected individuals but was not triggered for MDA , neighbors within the pool were vulnerable to onward transmission as no one in the pool received cure or prophylaxis ., When we relaxed the assumption that vectors tend to transmit in households close to their site of infection , and instead allowed infected vectors to bite individuals anywhere within the HFCA , pooled PCR was able to achieve high probability of interrupting onward transmission , requiring higher population coverage than MDA but less coverage than serological-based fMDA ( S6 Fig ) ., In Bbondo HFCA , all PCR pools always contained enough parasites to trigger MDA within the pool ., Pooled PCR became de facto MDA , indicating that performing pooled PCR would be a waste of resources as MDA is cheaper and easier to administer ., Unless a region is extremely heterogeneous , where a subregion experiences no transmission at all while another experiences a moderate amount , and vectors cannot migrate between heterogeneous areas , we anticipate that pooled PCR is an inferior strategy to MDA ., Neither of the lower-transmission HFCAs in this study showed such stark heterogeneity , but ongoing control efforts may push these regions into a regime where pooled PCR would be highly effective ., Of the RDT-based strategies , only within-200m fMDA showed any promise for elimination , and only with very small probability for the Bbondo HFCA at 100% coverage ., As parasite densities were slightly higher in Bbondo than in the Gwembe HFCA due to higher levels of transmission , RDTs were more likely to identify infected individuals to seed the fMDA foci ., However , fMDA at 200m did lead to substantial overtreatment compared to serology-dependent infection detection strategies ., For all infection detection strategies , a long-lasting prophylactic improved the chances of no onward transmission ( S4 Fig ) , and strategies such as pooled PCR that led to more overtreatment could outperform serological-based strategies at promoting elimination due to herd protection effects ., The selection of infection-detection strategy for a mass drug campaign depends on many factors , among them local transmission intensity , cost , operational feasibility , and population receptiveness ., In this study , we compare the effectiveness of MDA , MSAT , fMDA , RCD , pooled PCR , and hypothetical serological diagnostics at averting clinical cases and reducing onward transmission with minimal overtreatment of uninfected individuals ( Table 3 ) ., The spatial clustering of malaria infections means that fMDA strategies outperform MDA at selective targeting of infected individuals ., Shared household exposure can arise from both features of geography—local availability of larval habitat—and of human behavior—household preference for insecticide-treated net ( ITN ) use and shared travel history ., Due to the absence of data on individual histories of ITN use and travel , we assumed all infections were due to locally-acquired infections , and ITN usage was implicitly accounted for in each household’s selected transmission intensity ., How crucial is avoiding overtreatment with antimalarial drugs ?, MDA is the most effective infection-detection strategy for both control and elimination , yet MDA also leads to the most overtreatment ., When a drug campaign is a last push toward elimination and unlikely to be repeated many times , overtreatment may be less of an issue , especially if the campaign is set up for success with high coverage and a long-lasting prophylactic ., Given an excellent prophylactic , overtreatment is an irrelevant concern for elimination , particularly if vectors can migrate considerable distances ., If the drug campaign is for control purposes , for instance as a stopgap measure when health systems are temporarily broken as during the recent Ebola outbreak 33 , or as an ongoing program for gradual reduction in burden , minimizing overtreatment should be more of a priority ., Our recommendations for optimal infection detection strategies prioritize avoiding overtreatment for control recommendations and use overtreatment as a secondary consideration for elimination campaigns ., Local prevalence , household density , and heterogeneity of RDT positivity all influence the optimal infection-detection strategy ., While prevalence and population density may be known or estimated prior to a campaign , describing regional heterogeneity in exposure often requires more investment of resources through ongoing longitudinal surveys , multi-antigen serology , or sequencing of parasite genomes 15 , 34–36 ., Local population density and entomology can guide planners’ choice of fMDA radius if fMDA is under consideration ., In all but the lowest-prevalence settings , coverage is a stronger determinant of campaign success than the choice of infection-detection strategy ., However , when transmission is very low , the limited sensitivity of current diagnostics means that index cases are unlikely to be discovered by RDTs , and MDA , highly sensitive techniques like pooled PCR , or serological diagnostics that integrate history of infection are required to significantly reduce onward transmission ., Greater coverage cannot compensate for an insensitive diagnostic ., Simulation of serology-based diagnostics suggest that it is indeed possible to interrupt transmission in low-prevalence regions without distributing prophylactics to all individuals in the elimination area , although this finding may vary widely according to local entomology ., Under moderate to high prevalence , achieving high coverage is more important than selecting the optimal campaign type ., Of the non-MSAT strategies , all are equally efficacious at high prevalence , and within-household fMDA results in the least overtreatment ., When transmission is moderate , both MSAT and within-household fMDA are viable options , and other considerations such as cost , feasibility , and local culture will play a larger role in identifying the optimal infection detections strategy ., Compared with fMDAs , snowball RCD is a poor infection-detection strategy at moderate prevalence ., In snowball RCD , an initial clinical case serves as the primary trigger for a 200m-fMDA , and each RDT positive in that 200m radius triggers a secondary round of 200m-fMDA ., In low-transmission settings ( Gwembe HFCA ) , new infection is likely to lead to a clinical case , and thus a primary trigger , and although secondary triggers are uncommon due to low prevalence and old infections , there are enough primary triggers to achieve good spatial coverage in local areas of transmission ., In high-transmission settings ( Sinamalima HFCA ) , new infection is unlikely to lead to a primary trigger , but secondary triggers are common and thus the snowball effect leads to good spatial coverage and behaves like fMDA ., Under moderate transmission , infections are unlikely to lead to primary triggers due to immunity to clinical symptoms , and secondary triggers are less common than in high-transmission settings because infections are older and less likely to be superinfections ., In addition , symptomatic individuals are often less likely to seek care if they live further from a clinic 37 , leading to spatial dependence in detecting primary triggers ., In this study we have approximated interruption of transmission as the probability of less than one new infection per 1000 people arising from untreated infections in the 30 days post-campaign ., In a more realistic scenario , multiple campaign rounds per year are carried out , and campaigns may last for several years ., Thus we expect that all infection detection strategies are potentially more effective for elimination than predicted in the single-round analysis , but their relative efficacy will be as described above ., Other modeling studies have suggested that multiple rounds of drug campaigns in moderate-prevalence settings such as Munyumbwe HFCA may successfully interrupt transmission 5 ., In addition , programs are also likely to adapt campaign strategies as more data is collected , local pockets of transmission are identified , and overall prevalence declines ., A full dynamical model of malaria transmission at the household scale , with detailed simulation of vector feeding behavior and movement at the individual vector level , is necessary to fully explore the elimination power of various infection detection strategies ., Selective targeting of hotspots of malaria transmission has been proposed as a control measure , yet correctly identifying hotspots remains a challenge with current tools , particularly in seasonal settings where drug campaigns are likely to be deployed during the low-transmission season ., We predict that selective targeting via MSAT or fMDA strategies , where hotspots are identified by RDTs during the campaign , will not succeed in elimination until a new generation of diagnostics is ready for field use ., While our study does not rule out the possibility that repeated targeting of hotspots over many years may eventually lead to elimination , such extended campaigns pose significant feasibility challenges to communities , programs , and donors ., While single-timepoint measurements of infection status are insufficient for defining foci of transmission , serological diagnostics that report on many months of infection history appear to be very promising , and other sources such as longitudinal prevalence surveys and multi-year geotagging of clinical cases are also likely to be very informative for identifying areas for targeted MDA ., Size of the local population and patterns of human migration affect the likelihood of elimination , as elimination is easier in smaller population pools with less human mobility ., Human mobility and spatial heterogeneity also interact to inform local prevalence in complex ways 38 ., A study of RCD in a low-transmission region of Senegal found that most index cases reported recent travel 27 ., If every RDT positive case identified in the Gwembe HFCA were due to household inhabitants migrating from higher transmission regions , all RDT-dependent infection detection strategies would be less successful ( S7 Fig ) ., People arriving from higher-transmission areas will have relatively stronger immune responses to infection , making those infections more difficult to detect ., Infections detected in Gwembe HFCA could also be due to Gwembe inhabitants traveling elsewhere , acquiring infections , and returning to Gwembe; in this scenario , these individuals would be more likely to harbor recent , high-density infections amenable to detection by RDT ., Understanding regional demographics of mobility and inter-connectedness of elimination candidate areas can lead planners to decide whether non-MDA infection detection strategies are viable alternatives , and whether MSAT or MDA at border crossings would be effective policies 39 ., This study ignores campaign cost and feasibility as considerations for selecting an infection-detection strategy , yet these factors are important drivers in the real world ., Operational limitations make MDA and MSAT more attractive options than fMDAs , RCD , and strategies that require sensitive but expensive diagnostics ., In elimination scenarios , achievability may overrule cost as a consideration for determining campaign strategy ., Fortunately we find that MDA and MSAT are already the best strategies for elimination and control respectively in low-prevalence settings where drug campaigns are most likely to be deployed ., Reference data for household location , age structure , and RDT positivity by age were derived from a 2012 June-July survey performed in Gwembe and Sinazongwe districts in Southern Province , Zambia 19 ., Malaria transmission is heterogeneous and seasonal , with peak transmission between March and May ., Households in four HFCAs—Gwembe , Bbondo , and Munyumbwe in Gwembe district as well as Sinamalima in Sinazongwe district—were selected for inclusion in the reference dataset ( Fig 1 , Table 1 , S1 Fig , S1 Dataset ) ., Households without geolocation data and individuals without an RDT result were excluded: from Gwembe , 6 households and 985 individuals were excluded; from Bbondo , 9 households and 70 individuals; from Munyumbwe , 18 households and 652 individuals; and from Sinamalima , 14 households and 92 individuals ., Each household’s exposure to infectious bites was determined as follows ., An agent-based mechanistic model of malaria transmission , including exposure-dependent host immunity , was used to generate simulated populations experiencing endemic transmission ( EMOD DTK v2 . 0 ) 40–43 ., Twelve simulations of 10 , 000 people were run , where each simulation experienced the same southern Zambia seasonal temperature and rainfall patterns but supported different amounts of vectors ( S2 Fig ) ., These twelve simulations spanned a range of annual entomological inoculation rates ( EIRs ) from 0 . 003 to 120 infectious bites per person per year and included 10 imported cases per year ., All simulations incorporated case management as 30% treatment rate of clinical malaria and 50% treatment rate of severe malaria with artemether-lumefantrine ., Vector control was implicitly modeled in household exposure , and within-household heterogeneity in use of ITNs was ignored ., Simulations recorded daily asexual parasite density , infectiousness , and fever temperature for all individuals ., Infectiousness was defined as the fraction of mosquitoes feeding on the individual that would become infected and develop at least one oocyst ., Asexual parasite density and infectiousness were previously calibrated to age-stratified data from Burkina Faso 22 ., Each simulation was run for 90 years , allowing births and deaths but holding total population fixed , and RDT prevalence by age was measured on June 15 of year 90 , with RDT sensitivity at 50 asexual parasites/μL ., A higher RDT sensitivity was used here compared to later simulations as community health workers who gathered the reference dataset were highly trained in RDT use ., The relative probability Pij that a household i experiences exposure modeled by simulation j was calculated as follows for each household in the reference dataset and each of the twelve reference simulations ., All individuals k within 50m of the household were assumed to experience similar transmission intensity and aggregated to better inform household exposure ., The fraction of people of k’s age ak in simulation j with k’s RDT positivity , Rj± ( ak ) , is multiplied over all k within 50m of the household to find Pij:, Pij=∏individualskwithin50m of householdiRj± ( ak ), ( 1 ), Household transmission intensity is then determined by random selection from the j simulations according to weights Pij ., After selecting household transmission intensity , individuals were drawn from the 10 , 000 individuals simulated at that transmission intensity to form the age and RDT | Introduction, Results, Discussion, Methods | Mass campaigns with antimalarial drugs are potentially a powerful tool for local elimination of malaria , yet current diagnostic technologies are insufficiently sensitive to identify all individuals who harbor infections ., At the same time , overtreatment of uninfected individuals increases the risk of accelerating emergence of drug resistance and losing community acceptance ., Local heterogeneity in transmission intensity may allow campaign strategies that respond to index cases to successfully target subpatent infections while simultaneously limiting overtreatment ., While selective targeting of hotspots of transmission has been proposed as a strategy for malaria control , such targeting has not been tested in the context of malaria elimination ., Using household locations , demographics , and prevalence data from a survey of four health facility catchment areas in southern Zambia and an agent-based model of malaria transmission and immunity acquisition , a transmission intensity was fit to each household based on neighborhood age-dependent malaria prevalence ., A set of individual infection trajectories was constructed for every household in each catchment area , accounting for heterogeneous exposure and immunity ., Various campaign strategies—mass drug administration , mass screen and treat , focal mass drug administration , snowball reactive case detection , pooled sampling , and a hypothetical serological diagnostic—were simulated and evaluated for performance at finding infections , minimizing overtreatment , reducing clinical case counts , and interrupting transmission ., For malaria control , presumptive treatment leads to substantial overtreatment without additional morbidity reduction under all but the highest transmission conditions ., Compared with untargeted approaches , selective targeting of hotspots with drug campaigns is an ineffective tool for elimination due to limited sensitivity of available field diagnostics ., Serological diagnosis is potentially an effective tool for malaria elimination but requires higher coverage to achieve similar results to mass distribution of presumptive treatment . | Millions of people worldwide live at risk for malaria , a parasitic infectious disease transmitted by mosquitoes ., Great progress has been made in reducing malaria burden in recent years , and many regions are now devising strategies for elimination ., One way to eliminate malaria is to deplete the reservoir of parasites in human hosts by treating large groups of people with antimalarial drugs ., However , current field diagnostics are not sensitive enough to correctly identify all infected individuals ., Presumptively administering antimalarial drugs to whole populations will effectively clear infections but can also lead to substantial overtreatment and encourage the evolution of drug resistance in parasites ., We might be able to predict which individuals who test negative are actually infected based on whether their household members and neighbors are testing positive ., Using a mathematical model of malaria immunity acquisition and a spatial dataset of malaria prevalence in southern Zambia , we simulate strategies of identifying infected individuals and compare each strategy’s ability to deplete the infectious reservoir and avoid overtreatment ., We make different recommendations for optimal strategies depending on a region’s malaria prevalence . | null | null |
journal.pbio.3000159 | 2,019 | Microglia exit the CNS in spinal root avulsion | Microglia are the surveying phagocytic cells of the central nervous system ( CNS ) 1–3 ., They enter the CNS during embryonic development 4 , 5 ., Once in the CNS , their roles include a growing list of functions , including synaptic pruning and clearance of debris from both developmental and injured cells 6–9 ., During these processes , microglia transition from a surveying to activated state , leading to increased cellular migration and phagocytic activity 10–12 ., This microglia activation can have lasting impacts on the nervous system ., For example , pollutants in pregnant mice can lead to altered microglia in their progeny; such embryonic changes are implicated to autism-like phenotypes 13 ., Meanwhile , activation after spinal cord injury is linked to neuropathic pain 14 ., Despite these growing contributions of microglia to the CNS , little is known about their role outside of the CNS domain ., There is growing evidence that cells can override their domain-specific nature 15–18 ., CNS-resident cells like oligodendrocytes can reside in the peripheral nervous system ( PNS ) in peripheral neuropathy 15 ., Similarly , both oligodendrocytes and astrocytes can populate the PNS when boundary cells are disrupted 15–18 ., These ectopically localized CNS cells migrate to the PNS from the CNS instead of differentiating from resident PNS cells 16 , 17 ., Despite these examples of CNS cell intrusion , microglial emigration to the PNS is not understood ., In each example of ectopic CNS cell residence , the ability of the emigrated cells to return to their respective domains is not known ., Given the highly migratory nature of microglia , their emerging roles in circuit formation and maintenance in both healthy and disease states and their inclusions in disorders such as autism , spinal cord injury , neuropathic pain , and multiple sclerosis 12 , 13 , it is imperative to investigate not only the full capacity of microglia to migrate to specific domains but the consequence of such movements ., Complicating this issue , microglia and macrophages are labeled with similar molecular markers and when located outside of their resident domain , can express the limited number of specific markers that label either microglia or macrophages ., Here , we exploit time-lapse imaging of zebrafish in a laser-induced model of obstetrical brachial plexus injury ( OBPI ) to dissect the capacity of microglia in the PNS ., Spinal root avulsion can occur developmentally during the birth process; OBPI is a complication in an estimated 3 out of every 1 , 000 births 19 ., We first demonstrate that pu1+ cells display stable residency in distinct CNS and PNS domains at 4 days post fertilization ( dpf ) in zebrafish , a time comparable to when OBPI occurs ., We then show that CNS-resident microglia exit their domain to clear PNS debris in these injuries ., Despite that both macrophages and microglia respond to these injuries , microglia function as the debris-clearing cell in both the PNS and CNS ., This emigration to the PNS alters the microglia as they re-enter the CNS and migrate to distal areas from the injury site , including the brain ., We show that the exit of microglia to the PNS is mediated by opposing mechanisms; N-methyl-D-aspartate ( NMDA ) receptor and glutamate induce emigration , whereas contact-dependent repulsion prevents intrusion ., The observation that contact-dependent interactions of microglia with macrophages impact the cells’ anatomical position could provide insight into pathologies of diseases that contain both macrophages and microglia in the same domain 20 ., Together , these data provide evidence that microglia function expands beyond their textbook-defined CNS-resident domain ., To model obstetrical root avulsion , we created injuries in 4 dpf zebrafish ., At 4 dpf , zebrafish have an established anatomical organization of neurons in the brain and spinal cord but myelination is ongoing , similar to newborn children ( Fig 1A ) ., Cells that compose the spinal sensory root are organized by 2–3 dpf , before our avulsion model at 4 dpf 21 , 22 ., These injuries were created by exposing a 4 μm region of the spinal cord sensory root nerve to pulses of a laser in Tg ( ngn1:gfp ) zebrafish , which use regulatory sequences of ngn1 to express green fluorescent protein ( GFP ) in dorsal root ganglia ( DRG ) neurons 21 , 23 ( S1 Movie ) ., To confirm that the laser induced root avulsion , we first created intensity surface plots along DRG projections and measured an absence of intensity in the afferent projection specifically where the laser was exposed ( Fig 1B , S1A and S1B Fig ) ., The absence of signal was initially restricted to a small region until the axonal region degraded , leaving a DRG cell soma without a central projection ( Fig 1B ) ., This decrease was specific to the lesion site , persisted for hours , and was not created when the peripheral projection was injured ( S1C and S1D Fig ) ., In this avulsion model , as with obstetrical avulsions , we created injuries of varying severity ( S2A–S2D Fig ) ., We also fixed and stained Tg ( ngn1:gfp ) animals at 4 dpf with anti-Sox10 and anti-GFAP post-avulsion to assess the integrity of the GFAP+ glial limitans and Sox10+ Schwann cells and oligodendrocytes in each injury category immediately following injury ( S3A Fig ) ., Glial fibrillary acidic protein ( GFAP ) fluorescently labels the glial limitans , or the radial glial boundary ., DRG cell bodies and supporting Sox10+ cell nuclei were present and intact in uninjured animals and category I and II injuries ( S3A Fig ) ., The GFAP+ boundary was disrupted in category III injuries with little damage in category I–II ( S3B–S3E Fig ) ., These data recapitulate characteristics of spinal avulsion with varying degrees of severity but demonstrate that category I and II injuries lack massive damage of the spinal interface ., In movies of these injuries , we noted that neural debris of both glial and neuronal identity was present in both CNS and PNS regions ., To investigate debris clearance , we visualized phagocytic cells in response to injury 24 ( Fig 1C , S2 Movie ) ., To do this , we imaged Tg ( pu1:gfp ) ; Tg ( sox10:mrfp ) animals , which use regulatory sequences of pu1 to express GFP in microglial and macrophages and sox10 to label glial and neuronal cells of sensory and spinal nerves with mRFP 25 , 26 ., With this imaging , we first distinguished between pu1+ cells like PNS macrophages and CNS microglia based on their stable starting location without injury ( Fig 1C , S4A and S4B Fig ) ., By definition , microglia are stable CNS-resident cells that are labeled with pu1 and anti-4C4 and are sensitive to csf1r inhibitors ( S4C Fig ) ., To confirm their identity , before injuries , we stained for 4C4 , which labels microglia but not macrophages in zebrafish 27 , and detected 4C4 overlap only with CNS-located pu1+ cells and not peripheral cells ( Fig 1D ) ., These spinal cord-located 4C4+;pu1+ cells resembled well-defined microglia in the brain 5 , 27 ( Fig 1D ) ., Additionally , we assayed for the microglia-specific transcript tmem119 28 with a smFISH probes and detected colocalization with pu1+ microglia in the spinal cord ( S4F Fig ) ., To confirm microglia in the spinal cord at 4 dpf , we then tested that in noninjured animals , pu1+ cells in the CNS and PNS maintain their domain-specific residency ., In confocal images of the spinal cord , pu1+ cells in the spinal cord could first be detected at 3 . 5 dpf ., The number of pu1+ cells in the spinal cord then increased during development ., To determine their stable residency , time-lapse imaging from 4–5 dpf was used in noninjured animals ., Consistent with their identity as microglia , pu1+ cells in the spinal cord at 4 dpf remained resident in the CNS ( 100% of microglia remained in the spinal cord , n = 5 animals ) ., These movies revealed that the majority of pu1+ cells that colonize the spinal cord at 4 dpf originate from the anterior , potentionally the brain region , which colonizes microglia by 2 . 5 dpf 26 ., Not only were CNS cells marked by pu1+ , 4C4 , and tmem119 , they were also located in the spinal cord proper , distal from the spinal meninges space and the spinal vascular network , and were sensitive to csf1r inhibitors , consistent with their classification as microglia ( S4C , S4E and S4G Fig ) 2 ., PNS-located pu1+ cells were similarly stable in their PNS anatomical domain in 24-hour movies ., These data are consistent with the hypothesis that the domain residency of pu1+ cells like CNS-resident microglia and PNS-resident macrophages are established by 4 dpf in zebrafish , a comparable developmental time to when OPBIs occur ., To dissect phagocytic cellular response in OBPI , we then created injuries in Tg ( pu1:gfp ) ; Tg ( sox10:mrfp ) animals ., After injury , we imaged z-stacks that spanned the root and spinal cord every 2 . 5 minutes for 24 hours ., Following spinal avulsion and consistent with a microglial injury response , we visualized microglia migrated to the injury site within the first hour following injury ( Fig 1E , S5A Fig ) 29 , 30 ., In these movies , we also visualized macrophages responding immediately to the injury , traveling at an average velocity of 151 . 34 μm/h compared to microglia , which respond at a velocity of 203 . 58 μm/h ( S5B Fig ) ., To ask whether these responses were correlated with the size of the injury , we categorized injuries based on the lesion size ( S5C and S5D Fig ) ., However , all size injuries provoked microglial response ( S5C and S5D Fig ) ., To determine the migration path of these cells , we tracked individual cells and found that the migration path of microglia and macrophages was direct ( S5E Fig ) ., Microglia traveled 72 . 72 μm to injury compared to macrophages traveling 58 . 60 μm on average , with different velocities ( S5B and S5F Fig ) ., Given that both CNS and PNS cells responded , we next asked which cells responded first by quantifying the percentage of injuries for which each cell was a first responder ., Despite the difference in distance traveled , microglia and macrophages each were first responders 50% of the time ( S5G Fig ) ., Although macrophages outnumbered microglia at the injury site nearly 3-fold ., These data are consistent with the hypothesis that both microglia and macrophages respond to spinal root avulsion ., Because both cells responded to the injury , we next tested whether , while at the injury site , microglia and macrophages each clear debris ., To do this , we injured Tg ( pu1:gfp ) ; Tg ( sox10:mrfp ) animals , created 24-hour movies , identified CNS and PNS GFP+ cells based on their pre-injury location , and then scored mRFP debris within GFP+ cells , a result consistent with clearance of debris from GFP+ cells 26 , 31 , 32 ., Although both macrophages and microglia responded to injury , mRFP+ debris was present in 92% of microglia , whereas mRFP+ debris was present in 8% of macrophages ( Fig 1F and 1G ) , consistent with the hypothesis that microglia primarily clear debris ., We confirmed that the phagocytosis of mRFP+ debris from microglia was specific to the injury site by scoring an average of 1 . 2 mRFP+ debris puncta before microglia arrived to the injury site , which increased to 4 . 3 mRFP+ debris after arrival ( Fig 1H ) ., Although mRFP+ debris was primarily in microglia , phagocytic vacuoles could be seen in both macrophages and microglia , consistent with the previous hypotheses that both cells are capable of phagocytosis ( S6A and S6B Fig ) ., These vacuoles were specific to the injury site , because after arrival to injury , on average 2 . 18 vacuoles were present compared to the 0 vacuoles before arrival ( S6A and S6B Fig ) ., Although the vacuoles lacked mRFP+ debris in macrophages , every microglial cell with vacuoles contained mRFP+ debris ., If microglia are the primary debris-clearing cells in OBPI , we hypothesized that they would clear debris longer at the injury side than macrophages ., To test this , we tracked the distinct pu1+ cell populations in time-lapse movies and scored microglia spend on average 9 . 14 hours at the injury site compared to macrophages , which spend 5 . 97 hours ( S6C and S6D Fig ) , a result that is consistent with the hypothesis that microglia could function as the primary debris-clearing cells following obstetrical avulsion ., To determine the role of phagocytic cells in clearance of these domains , we tracked both cells in uninjured cases and following injury in lateral views in which we could distinguish the location of the spinal cord and the PNS ( Fig 2A and 2B , S7A and S7B Fig ) ., In these movies , we visualized GFP+ cells that originated in the CNS , migrate to the injury site , squeeze at the spinal cord boundary , and potentially migrate outside of its CNS domain ( S7C and S7D Fig , S3 Movie ) ., This hourglass-like morphology is typical of cells that must squeeze through a space-restricted area to leave the spinal cord 17 , 33 ., We confirmed the ectopic migration of these cells by rotating our z-stack 90° in the movies; the distance of this migration also extended laterally beyond the normal range of the gfap+ spinal cord glial limitans ( Fig 2A and 2B , S7E Fig ) ., We measured these cells extended on average 14 . 7 μm beyond the glial limitans boundary; in comparision , the DRG was measured on average 4 . 8 μm outside the glial limitans ( S7F Fig ) ., We also visualized microglia in contact with the PNS-located DRG resident cells , supporting the hypothesis that microglia clearly exited the CNS ( S7G–S7H Fig ) ., We quantified this behavior in our 24-hour movies after injury and microglia were detected laterally outside of the mRFP region in 40% of the movies , they displayed an hourglass-like morphology , and in 3D reconstructions were located outside the normal region of the curvature of the glial limitans ( Fig 2C , S7E Fig ) ., In these emigration events , typically one to two microglia exited ( S7I Fig ) ., This is in contrast to noninjured spinal cords in which microglia never were present in these regions ., On average , microglia emigrated 12 . 87 μm outside of the sox10+ CNS region following spinal root avulsion , remained there on average 4 . 13 hours , and displayed microglia-like morphology while positioned there ( S7J and S7K Fig ) ., In these emigration events , a continuous GFP+ process that remained in the CNS could not be detected , suggesting that the entirety of the microglia cell left the CNS ., After emigration , individual microglia did return to the CNS ( Fig 2B and 2E ) ., It is possible that this microglia emigration occurs from a massive disruption of the spinal cord boundary ., However , despite disruption of the glial limitans in category III injury cases , microglia emigration was still observed in category I and II injury types with limited glial limitans disruption ( S7L Fig ) ., Also inconsistent with this idea , microglia emigration was preceeded by a squeezing of the microglia , suggesting it migrates through a space-restricted area ( S7C and S7D Fig ) ., If a general boundary disruption was present , oligodendrocytes and neurons would potentially ectopically exit as they do when boundary cap cells and/or Schwann cells are disrupted 15–18 ., However , we also did not detect emigration of oligodendrocytes ( S7L Fig ) ., We next asked whether emigration was specific to root injuries by injuring the mixed PNS nerve ( analogous to the sciatic nerve ) and did not detect emigration ( S8A–S8C Fig ) ., Instead , and as previously reported , macrophages were the responding cell at mixed nerves 34 ( S8D and S8E Fig ) ., Microglia also did not emigrate when we created a CNS-specific injury ( S8C and S8E Fig ) ., To rule out the possibility that this was specific to developmental properties at 4 dpf , we also observed microglia emigration to avulsions at 7 dpf ( Fig 2D ) ., Based on these data , we propose that microglia can exit the spinal cord , at least to the PNS-located spinal root and DRG , following OBPI-like injuries ., Such an observation is in contrast to current textbook definitions of microglia ., To gain further insight into this emigration , we tracked the trajectory of individual microglia after injury ., Not only did microglia exit following injury , but this tracing analysis showed that individual microglia traverse the spinal cord boundary an average 6 . 25 times throughout their response to injury ( Fig 2E , S9A–S9I Fig ) ., We next considered the hypothesis that , although microglia ectopically migrated , their phagocytic properties could be different between their resident domain and the ectopically located PNS domain ., To test this , we took advantage of our imaging approach and tracked individual microglia after injury , then scored and tracked the individual debris concentrates within those microglia ., In these movies , we could identify pu1+ cells that originated in the CNS contain mRFP+ debris ( Fig 2F , S9J–S9M Fig ) ., To test whether debris could be carried across the CNS boundary , we scored the appearance of individual mRFP+ clusters in the GFP+ cells in the CNS ., During their emigration to the spinal root , on average two mRFP+ particles from the CNS were carried within the microglia to the PNS ., And while in the PNS , an additional one mRFP+ debris appeared in the cell ( Fig 2F , S9N Fig ) ., This appearance of mRFP+ debris particles while in the PNS is unlikely from already present CNS debris particle fission because the area of individual particles increases from 2 . 05 μm2 to 2 . 83 μm2 while the microglia cell is migrating ( S9O Fig ) ., As the cells entered back into the CNS , mRFP+ debris particles that appeared while the cell was in the PNS continued to be present ( Fig 2F ) ., These data are consistent with the possibility that pu1+ microglia not only migrate to the PNS-located roots but also clear debris while there ., Additionally , their entry back into the CNS with PNS debris introduces the CNS to PNS debris ., Given this movement of microglia , we next sought to determine where microglia and macrophages eventually reside following their injury response ., To do this , we identified CNS versus PNS pu1+ cells before injury , created injuries , imaged those injury sites for 24 hours , and tracked the individual cells with tracking software ., In this analysis of the microglia that exited to the PNS , 44 . 44% of them migrated back to the CNS , where they continued to reside until the end of the 24-hour imaging window ( Fig 2G ) ., This phenomonon is distinct from other CNS cells that have been shown to emigrate because microglia also re-enter 15 , 17 ., In the other 55 . 56% , microglia continued to be present in the PNS at the injury site at the end of the 24-hour imaging window ( Fig 2G ) ., These PNS-located microglia did not leave the injury site during this time ., Macrophages migrated into and out of the injury site , sometimes entering the CNS ., However , at the end of the imaging window , macrophages were rarely seen in the CNS ., The simplest explanation for this data is that microglia can leave the CNS to respond to avulsions and can return to CNS residency after clearing debris ., To dissect the consequence of this emigration , we tracked individual microglia that emigrated and then returned to the CNS ., During this process , we measured numerous cellular properties that were previously described across species to indicate altered microglia 9 ., We first tested whether the morphology of single microglia changed as they progressed through their emigration and re-entry ( Fig 2H , S10A–S10D Fig ) ., To determine morphological changes , we measured four shape descriptors ., Microglia showed a signficant shift in aspect ratio ( an average measure of 8 . 57 ) and cell roundness ( an average measure of 0 . 21 ) as cells were leaving the CNS ( Fig 2I and 2J , S10B and S10C Fig ) ., These differences remained while the cell returned to the CNS ., To address whether this was a result of their location at the injury as apposed to emigration , we compared emigrating microglia to microglia that responded to and actively cleared debris at the injury site but never exited the CNS ( S10E–S10H Fig ) ., Again , this analysis showed that aspect ratio and roundness were different from cells that emigrated compared to nonexiting microglia ., We could not detect any differences in the cells before emigration ( Fig 2H–2J , S10I and S10J Fig ) , inconsistent with the hypothesis that emigrating microglia are distinct , at least morphologically , before emigration ., To further dissect whether microglia are altered , we tested whether emigrated microglia were physiologically changed by scoring their phagocytic activity ., Using movies , we could score the number of new debris particles within individual microglia ., We scored that individual microglia before their emigration increase the number of new mRFP+ particles once they return to the CNS ( Fig 2K , S10K Fig ) ., As a third indicator of altered microglia , we also scored the number of secondary projections that are used in the phagocytic process 9 , 23 ( S10L Fig ) ., Consistent with the conclusion that emigrated microglia return in an altered state , they increase their secondary projections while in the PNS and remain elevated as they re-enter ( S10L Fig ) ., Emigrated microglia also were distinct from microglia that responded to injuries of CNS tissue only ( S10M Fig ) ., Together , these data are consistent with the idea that emigration itself could induce a unique microglial state ., Because CNS-resident pu1+ cells migrated out of the spinal cord but then returned to the CNS , we next asked whether those ectopically migrated and altered cells moved to distal areas of the CNS following their emigration ., To do this , we created injuries in animals , tiled the animal from the brain to tail with confocal positions , and time-lapse imaged each of these positions for 24 hours ., With this whole spinal cord analysis , we could visualize CNS-derived pu1+ cells migrate to the injury , squeeze into the PNS , relocate to the CNS through the spinal cord , and then migrate anteriorly toward the brain and then caudally to the tail , surveying 743 . 56 μm ( 56 . 61% ) of the spinal cord on average ( Fig 2L , S11A–S11E Fig , S4 Movie ) ., These movies demonstrated PNS-primed microglia that carried debris from the injuries to the brain ( S11A Fig ) ., This tiling analysis also allowed us to see interactions between individual cells along the length of the animal ., We commonly visualized emigrated microglia interacting with other CNS-resident microglia in the spinal cord ( S11F–S11H Fig , S4 Movie ) ., Based on these data , we conclude that not only can microglia migrate out of the CNS but they can re-enter the CNS as PNS-primed cells and migrate , in an altered state , to distal areas from the injury site ., These migration sites include the brain ., To begin to understand the potential functional consequence of microglia emigration , we created a primary avulsion injury in Tg ( pu1:gfp ) ;Tg ( sox10:mrfp ) animals at 4 dpf and observed emigration of microglia to the PNS ., Then we created a distal secondary CNS-specific injury ., Upon secondary injuries , we observed PNS-primed microglia immediately re-enter the CNS and migrate to the secondary injury site ( S12A Fig ) ., We quantified the amount of new debris PNS-primed microglia collected at the secondary injury site , which was greater than the amount of new debris naïve CNS microglia collected at the secondary injury site ( Fig 3A ) ., Naïve microglia were defined as never contacting the initial primary injury site ., Additionally , PNS-primed microglia created more secondary projections at the secondary injury and spent more time there ( S12B–S12D Fig ) ., Together , these data demonstrate that PNS-primed microglia are more phagocytically active when they return to the CNS and could present an altered response to other injuries that occur after the avulsion ., To dissect the molecular mechanism of this emigration , we screened through small molecules that could disrupt emigration ., In this , we identified that NMDA inhibitors , MK-801 and D-AP5 30 , disrupted the emigration of the microglia to the PNS ( Fig 3B , S13A Fig ) ., To dissect this mechanism further , we tracked individual microglia following injury in DMSO and NMDA inhibitor exposure ., We first hypothesized that lack of NMDA signaling disrupted the response of the microglia to spinal sensory roots ., However , by tracking individual pu1+ cells after NMDA inhibition , microglia and macrophages still responded to the injury , ruling out the possibility that blocking NMDA receptors prevented an initial injury response ( S13B and S13C Fig ) ., Instead , microglia responded to injury but did not emigrate ( Fig 3B ) ., This lack of emigration when NMDA signaling is disrupted perturbed the alteration of microglia at the injury that occurs as they emigrate ( S14 Fig ) ., Together , these data are consistent with the hypothesis that microglia exit of the CNS following injury is dependent on NMDA ., To test this mechanism further , we asked whether glutamate , an activator of NMDA , could induce microglia emigration ., We did this by soaking animals in caged glutamate 35 , created brachial plexus injury ( BPI ) -like injuries , waited for a pu1+ cellular response to the injury , uncaged a 4 μm region in the PNS by exposing it to 405 nm laser , and then imaged for 2 hours after uncaging ( Fig 3C and 3D , S5 Movie ) ., As a control , soaking of caged glutamate did not alter the initial response time to the injury ( Fig 3C ) ., However , on average , the uncaging of glutamate induced exit of microglia to the PNS in 30 minutes ., This was significnalty faster than controls: untreated injuries showed average exit in 7 . 29 hours and caged-glutamate soaked animals that were exposed to a 4 μm region in the PNS of 641 nm light did not exit before 2 hours ( Fig 3E , S15A Fig ) ., Uncaging glutamate induced emigration in the first 2 hours in 80% of injuries compared to 0% in mock-activated controls ( S15A Fig ) ., Microglia in both cases traveled the same distance , ruling out the possibility that these significant response times were caused from varying travel distances ( S15B and S15C Fig ) ., Consistent with their emigration , uncaging glutamate also caused morphological changes that occur in emigration states ( S14 Fig , S15D–S15F Fig ) ., To ask whether glutamate was sufficient without injury to induce emigration , glutamate was uncaged in the absence of injury ( S15G and S15H Fig ) ., Consistent with NMDA signaling inhibitors not reducing the initial injury response but specifically emigration , glutamate was not sufficient to induce emigration without injury ., These data are consistent with the hypothesis that the mechanism of microglia emigration is NMDA- and glutamate-dependent ., We hypothesized that a balance between NMDA induction and an emigration restriction mechanism determined emigration efficiency ., Domain-specific cells can be restrictive to cells outside and inside of their domain 16 , 17 , 25 , 36–38 ., Given that both macrophages and microglia responded to injury , we hypothesized that the presence of specific cells at the injury site , like macrophages , could prevent microglia from performing their full debris-clearing potential ., We tested this mechanism in our imaging set-up by initially scoring interactions between the distinct pu1+ cells ( S6 Movie ) ., We first quantified the number of times each cell type displayed homotypic versus heterotypic contact ( Fig 4A and 4B , S16A–S16C Fig ) ., Then , we asked whether those interactions induced directional changes ., In this analysis , we identified that homotypic interactions between microglia induced migration of microglia 83 . 33% of the time and heterotypic interactions between microglia and macrophages induced migration of microglia away from the contacting cell 88 . 88% of the time ( Fig 4C , S6 , S7 and S8 Movies ) ., Macrophages did not respond to either homotypic or heterotypic contact ( Fig 4A and 4C , S16B–S16D Fig , S9 Movie ) ., We could quantify this by measuring distance traveled over time and measured that microglia travel on average 157 . 25 μm after contact whereas macrophages travel 9 . 20 μm after contact ( S16D Fig ) ., We next hypothesized that these heterotypic and homotypic interactions could impact the ability of microglia to emigrate ., To first test this possibility , we visualized microglial migration in the absence of macrophages ., To do this , we tailored the laser parameters to produce spinal root avulsions with limited peripheral injury ., In these injuries , microglia responded , but macrophages did not ( S17A Fig , S10 Movie ) ., We then traced the behavior of individual microglia ., In contrast to injuries with macrophage responses , microglia migrated to the PNS quicker to these injuries , remained in the PNS , and extended long cellular processes into the PNS that were not visualized when macrophages were present ( Fig 4D , S17 Fig ) ., These data are consistent with the cellular mechanism that macrophages inhibit microglia emigration ., We tested this potential mechanism with a second approach by reducing macrophages with focal laser ablations ., In these experiments , we created avulsions , then after pu1+ cells migrated to the injury , we laser ablated single macrophages in close proximity to microglia at the injury site ( Fig 4E , S18A and S18B Fig , S11 Movie ) ., Following these single macrophage ablations , microglia migrated into the empty space that was created from macrophage ablations within seconds ( Fig 4F and 4G , S18C–S18E Fig ) ., Although we cannot completely rule out that macrophage debris attracts microglia , as a control identical laser ablation exposure to adjacent nonmacrophage space did not provoke migration into that area ., This is consistent with the idea that ablation does not itself—or debris it creates—induce this microglial response ( Fig 4G , S19 Fig ) ., The simplest explanation for this is that dynamic interactions between microglia and macrophages deter microglia from occupying the PNS-located injury site ., To test whether this cell–cell contact mechanism has functional implications on debris clearance , we again took advantage of the simplicity of the zebrafish system; during early developmental stages , macrophages are abundant , but microglia are limited in number in the spinal cord and approximately 5% of animals display no CNS pu1+ cells in the spinal cord at 4 dpf ., In these rare microglia-less spinal cords , we injured the spinal root and scored the percentage of injuries containing the mRFP+ debris that concentrated into puncta by phagocytic cells ( Fig 4H , S20 Fig ) ., As mentioned above , mRFP+ particles appear as the microglia migrate to the injury ( Fig 1F–1H , S6 Fig ) ., In contrast to injuries in animals with microglia , mRFP+ debris puncta was not detected in injuries without microglia , suggesting that microglia are required for debris clearance ( Fig 4I ) ., In contrast , for injuries with no macrophages , debris clustered quicker at the injury site ( Fig 4F–4G ) ., And consistent with the role of microglia , this debris was present only within microglia ., These results are consistent with the cellular mechanism that microglia are responsible for debris clearance following injury but their access is inhibited to the injury site by macrophages ., This stymies their ability to arrive and efficiently clear debris in both the CNS and PNS following spinal root avulsion ., Here , we demonstrate that microglia , despite their textbook definition , can alter their domain-specific residency ., In this model of OBPI , the emigrating microglia serve as the debris-clearing cell ( S21 Fig ) ., Their potential , however , is stymied by macrophages ., Macrophages restrict access to the area via a contact-dependent mechanism; a principle that appears to be consistently utilized across neural cell | Introduction, Results, Discussion, Materials and methods | Microglia are central nervous system ( CNS ) -resident cells ., Their ability to migrate outside of the CNS , however , is not understood ., Using time-lapse imaging in an obstetrical brachial plexus injury ( OBPI ) model , we show that microglia squeeze through the spinal boundary and emigrate to peripheral spinal roots ., Although both macrophages and microglia respond , microglia are the debris-clearing cell ., Once outside the CNS , microglia re-enter the spinal cord in an altered state ., These peripheral nervous system ( PNS ) -experienced microglia can travel to distal CNS areas from the injury site , including the brain , with debris ., This emigration is balanced by two mechanisms—induced emigration via N-methyl-D-aspartate receptor ( NMDA ) dependence and restriction via contact-dependent cellular repulsion with macrophages ., These discoveries open the possibility that microglia can migrate outside of their textbook-defined regions in disease states . | Cells are precisely organized in specific anatomical domains to ensure normal functioning of the nervous system ., One such cell type , microglia , is usually considered to be confined to the central nervous system ( CNS ) ., Using time-lapse imaging to capture microglia as they migrate , we show that their characteristic CNS-residency can be altered after spinal root injury ., After such injury , the microglia exit the spinal root to the periphery , where they clear debris at the injury site and then carry that debris back into the CNS ., In addition , microglia that leave the CNS after spinal root injury become distinct from those that remain within the CNS ., This emigration event of microglia after injury is driven by two mechanisms—dependence on glutamatergic signaling that induces their emigration to the injury and interactions with macrophages that prevent their ectopic exit from the spinal cord ., Together , these discoveries raise the possibility that microglia could override their CNS-residency in certain disease contexts . | blood cells, traumatic injury, medicine and health sciences, fish, neurochemistry, immune cells, nervous system, statistics, immunology, vertebrates, neuroscience, animals, microglial cells, animal models, organisms, osteichthyes, model organisms, mathematics, experimental organism systems, neurotransmitters, research and analysis methods, spinal cord, white blood cells, animal cells, animal studies, neurotrauma, glial cells, short reports, critical care and emergency medicine, trauma medicine, glutamate, biochemistry, zebrafish, eukaryota, neuroanatomy, cell biology, anatomy, central nervous system, spinal cord injury, neurology, biology and life sciences, cellular types, physical sciences, macrophages, statistical data | Microglia are normally assumed to be confined to the central nervous system (CNS), but this study shows show that after spinal root injury, microglia can exit the CNS to clear debris. Upon re-entry, the emigrated microglia are altered and can travel to distal areas such as the brain. |
journal.pcbi.1001113 | 2,011 | Consensus-Phenotype Integration of Transcriptomic and Metabolomic Data Implies a Role for Metabolism in the Chemosensitivity of Tumour Cells | In the quest to understand complex biological systems at multiple levels of, biological organization , the need arises to combine knowledge from experiments of, different types to create a full picture of a systems behavior ., Modern, molecular profiling ( “omics” ) methods , such as transcriptomics ,, proteomics and metabolomics allow one to build up a global picture of system, characteristics , and to search for interactions and coordinated behavior between the, different levels ., While each level can be studied separately , greater statistical, and explanatory power can be gained by integrating this knowledge into a single, coherent model of the system ., This is currently one of the greatest challenges in, systems biology ., Inter-omic data integration can be performed at different levels 1 , the simplest of, which is conceptual integration ., At this level , each omics data set is analysed, separately and a coherent biological rationale is constructed which explains, phenomena observed in the separate molecular profiles ., For example , changes in, levels of both enzyme transcripts and metabolites from the same pathway could be, explained by the hypothesis of differential regulation of that pathway ., However this, subjective approach can lead to plausible biological explanations that arise through, spurious statistical associations and conversely some potentially novel mechanisms, may be overlooked ., The statistical level of integration is more objective ., In this, approach , links between data sets are made using rigorous statistical procedures, such as correlation , regression or more sophisticated techniques ., To date , much, inter-omic data integration has been performed at the conceptual level 2 , 3 , 4 while various, methods have been proposed and demonstrated for statistical integration 5 , 6 , 7 , 8 , 9 ., Many researchers have found that interpretation of omics data at the level of, individual molecular entities can be difficult and have opted for an analysis at the, pathway or functional level 10 ., This is mainly because particular changes in biochemical, pathways , associated with phenotypic conditions such as disease can often arise from, a range of different alterations in a pathway ., A common method for performing, pathway-level analysis on single omic data is over-representation ( OR ) analysis, 11 , 12 , in which a set, of molecular elements ( e . g . genes ) that are differentially expressed or correlated, with the phenotype of interest are first selected ., The set is then compared against, molecular sets defined a priori ( e . g . genes in established, pathways ) to identify those sets that show greater overlap with the, phenotype-associated genes than would be expected by chance ., The final list of, significantly over-represented or ‘enriched’ sets/pathways is used to, aid biological interpretation of the data ., As well as performing OR with genes ,, Metabolite Set Enrichment Analysis ( MSEA ) 13 and other metabolite, over-representation techniques 14 have also been developed ., In this work we contrast the, application of the OR analysis approach to transcript and metabolite data, individually to the alternative of considering them simultaneously , using, established pathways to guide an integrated analysis of the two data sets ., In addition to the inter-omic integration of metabolomic and transcriptomic data , our, approach involves a further type of data integration that we call, consensus-phenotype integration ., In this approach , several examples of the same, phenotype , achieved in different ways , are used within the experimental design ., For, example , one may study a particular mechanism of toxicity via the use of different, chemical treatments that have a similar mode of action ., One can thus identify, features that are central to the phenotype in question across different types of, “omics” data , as opposed to features that are specific to a single, instance of the phenotype being studied ., In this work , we aim to elucidate mechanisms of drug sensitivity through the use of, inter-omic statistical data integration using drug sensitivity , transcriptomic and, metabolomic data from the NCI60 cell line panel 15 ., The NCI60 is a panel of tumor, derived cell lines corresponding to diverse tissue types , which has been subject to, extensive molecular phenotypic and pharmacological characterization ., We used, baseline ( untreated ) metabolic and transcriptional profiles readily available for 58, lines as well as growth inhibition data from an array of 118 drugs 15 , 16 ., We correlate, growth inhibition to the molecular profiles to identify pathways related to drug, sensitivity ., We first focus on platinum sensitivity as it is a well-defined, phenotype , linked to a well-investigated mode of action that has important clinical, implications ., Many chemotherapeutic regimes are based on platinum compounds , and, resistance to these drugs is a major obstacle in successful treatment of some, cancers ., The mechanisms that cause variation in response to therapy are not well, understood , and the ability to predict sensitivity from a baseline profile of the, tumor would help to improve therapy selection and thereby potentially reduce patient, morbidity and mortality ., We then expand our analysis to a larger set of 118 drugs to, investigate whether the method is able to associate drugs with similar modes of, action ., We show that statistical integration conducted through a joint analysis of, the data gives specific advantages in terms of sensitivity and confidence of pathway, associations ., For each drug we ranked all probe sets by their absolute Pearson correlation, ( |r| ) to the −log ( GI50 ), values across all cell lines ., Setting the false discovery rate ( FDR ) 19 at, 60% we then selected genes considered to be significantly associated to, chemosensitivity ., A high FDR was tolerated at this stage of the analysis to, ensure that subsequent pathway analysis was adequately powered ., Repeating this, process for the metabolite data we obtained separate panels of genes and, metabolites that were deemed to be associated with the sensitivity to each drug, ( see Table, S1 ) ., In total 3 , 33 , 37 and 92 metabolites and 915 , 1620 , 5035 &, 6533 genes were identified as associated with sensitivity to carboplatin ,, cisplain , iproplatin and tetraplatin treatment respectively ., To assess which pathways characterized the drug sensitivity phenotype we then, performed OR analysis with pathways from the ConsensusPathDB 20 ., The, ConsensusPathDB collates pathways from several public databases of protein, interactions , signaling and metabolic pathways as well as gene regulation in, humans ., We restricted our analysis to sources covering biochemical reactions:, KEGG 21 ,, Reactome 22 , Netpath ( http://www . netpath . org ) ,, Biocarta ( http://www . biocarta . com ) , HumanCyc 23 and the pathway interaction, database ( PID ) 24 ., The use of multiple databases reduces bias by, enhancing coverage ., At the time of analysis the ConsensusPathDB contained 1875, pathways from the selected sources , of which 1651 contain at least one gene and, 581 contain at least one metabolite measured in the NCI60 data ( excluding the, highly prevalent ‘currency’ metabolites phosphate , diphosphate and, NADP+ ) ., OR analysis of the phenotype-associated gene panels indicated that, 63 , 74 , 233 and 242 pathways were associated with cisplatin , carboplatin ,, iproplatin and tetraplatin sensitivity respectively ( p<0 . 05 ) ., The equivalent, analysis for metabolite panels indicated that 24 , 13 , 4 , & 5 pathways were, associated with these phenotypes ., To highlight pathways relevant to general platinum sensitivity , as opposed to, particular platinum compounds , we looked for pathways that were associated with, more than one drug response phenotype ( ‘consensus-phenotype, integration’; Figure 2 A, & B ) ., Within the gene transcript analysis ( Figure 2A ) , the drugs appeared to divide into, two pairs that shared many pathways in common ., Iproplatin and tetraplatin were, most similar , sharing 143 ( 133+4+4+2 ) of the 330, ( 75+5+5+4+143+92+3+1+2 ) , ie ., 43%, of the pathways associated with either drug ( Figure 2A ) ., Carboplatin and cisplatin also, show a high level of similarity ( 32 of 103 pathways , 31% ) ., In the, metabolic analysis ( Figure, 2B ) the similarity between iproplatin and tetraplatin was much lower ,, while carboplatin and cisplatin retained a high level of similarity with 7 of, the 30 pathways being shared ( 23% ) ., The gene analysis highlights many, more pathways than metabolite analysis due to both the higher number of pathways, with sufficient numbers of quantified transcripts and the limited number of, quantified and identified metabolites ., We next combined the transcriptomic and metabolomic data into a joint inter-omic, OR analysis ( Figure 2C ) by, estimating the joint probability of association of each pathway with the drug, sensitivity phenotype assuming independence between the probability of, association from the gene and metabolite data separately ( see Methods ) ., 35 pathways were found to be, significant for at least one drug in the joint analysis that did not feature in, either of the separate analyses of gene expression or metabolite levels ., To, confirm the significance of the increase in pathway detection after integration, of the metabolic and transcriptomic data , we estimated the null distribution of, the joint analysis probabilities by permuting the gene analysis pathway, probabilities relative to the metabolite analysis pathway probabilities ., For, carboplatin only 3 of the 100 permutations produced more pathways than the real, data and for cisplatin no permutations produced as many pathways as the real, data ., However , for iproplatin and tetraplatin , the number of pathways detected, was not significantly enhanced by the joint OR analysis , suggesting that the, combined analysis may be most advantageous when the numbers of significantly, associated genes or metabolites are relatively low ., To examine the significance of the numbers of pathways in the joint OR analysis, that were commonly associated to the effect of multiple drug treatments , two, null models were generated ., Null model I assumed that genes and metabolites, identified as significantly associated to a phenotype were randomly selected, whereas null model II correspondingly assumed that pathways are selected, randomly ., Table 1, summarizes the pathway coincidence between the output of joint OR analysis, across the four platinum drugs for these two null models compared to the real, data and reports the associated FDR in each analysis ., We observed that by, setting our criterion of significant association between a pathway and platinum, sensitivity at requiring a majority of the drugs to be associated with that, pathway ( i . e . at least 3/4 ) we achieved acceptable FDRs of 0 . 2% ( null, model I ) and 16 . 9% ( null model II , the most extreme scenario ) ., Using the majority overlap criterion we compared the number of pathways, consistently associated with platinum sensitivity between the individual and, joint analyses ( Figure 2D ) ., The joint OR analysis identified all pathways highlighted by the individual, –omic OR analyses combined ( 17 in total ) , but also indicated a further 13, pathways that were consistently associated ( +76% ) ., No pathways were, found to be common between both the separate gene and metabolite analyses ., Overall 30 pathways met the majority criterion of association with sensitivity at, least 3 platinum drugs and hence general platinum cytotoxicity ( Table 2; Figure 2C & D ) ., All the, databases used to compile the ConsensusPathDB contributed pathways to the final, selected consensus pathways , highlighting the value of the ConsensusPathDB, strategy in pathway analysis ., While this subset of pathways included those with, established relationships to platinum sensitivity and general chemosensitivity ,, such as DNA repair and Akt regulation of nuclear transcription , there were also, several pathways related to metabolic processes not previously reported as, determinants of platinum sensitivity ., These included nucleotide metabolism ,, fatty acid , triglyceride and lipid metabolism ., The added value of the inter-omic OR analysis prior to consensus phenotype, integration can be more clearly discerned at the individual pathway level ., Figure 3 is a network, representation of the base excision repair ( BER ) pathway from Reactome and, depicts both the detected entities and the drugs with which each detected entity, is associated ., While the majority of entities were significantly associated to, the effect of at least one of the four platinum agents , there was significant, variation in the pattern of association and no gene or metabolite was, significantly associated to all four treatments ., Accordingly the pathway was, only significantly associated to tetraplatin and iproplatin sensitivity using, the transcriptome data alone , and to carboplatin and cisplatin sensitivity using, the metabolite data in isolation ., Using the joint OR analysis the BER pathway, was significantly associated to the sensitivity to all four platinum compounds, ( Table 2 ) and the, evidence for association with each drug was increased , due to the added, information from the alternative data type ., Of the 12 pathways for which, inter-omic OR analysis improved the consensus between the drugs , 10 refer to, metabolic processes ., In order to validate and to test the generalisability of our findings we then, examined GI50 data from a test compound , diaminocyclohexyl-Pt ( II ) ., After conducting the same inter-omic OR analysis as described previously , we, observed that the effects of this compound on the NCI60 panel was associated, with 5 of the 6 pathways common to all 4 other platinum drugs along with a, further 12 pathways from Table, 2 and 90% ( 220/245 ) of the pathways associated with, diaminocyclohexyl-Pt ( II ) were Associated with at least one of the other platinum, drugs ., In particular there were 138/152 pathways commonly associated between, diaminocyclohexyl-Pt ( II ) , iproplatin and tetraplatin sensitivity ., Since OR, analysis makes no distinction between positive and negative molecule/sensitivity, correlations , we also examined the direction of associations between the, metabolites detected in the consensus pathways and the GI50 of all, platinum drugs ( Table 3 ) ., In total , a panel of 22 metabolites were associated with the consensus metabolic, pathways from analysis of the four training compounds ., While there was variation, in the metabolites associated with specific treatments , where a significant, association was observed the direction of correlation was consistent across the, training set ., The GI50 values of our test compound ,, diaminocyclohexyl-Pt ( II ) , was significantly correlated to 19/22 metabolites in, this panel , with complete consistency in the direction of association with the, training set data ., To explore more broadly the relationships between chemosensitivity and biological, pathways across a range of agents , and to ascertain the specificity of the, consensus phenotype analysis for platinum sensitivity pathways , inter-omic OR, analysis was performed using GI50 data for all 118 compounds, available within the NCI 60 dataset ., In total 1262 pathways were significantly, associated with the drug sensitivity of at least one compound , while 82, compounds gave at least one significant pathway ., Figure 4 shows the clustered heat-map of the, binary association matrix in which each element is set to one if a pathway is, significantly associated with sensitivity to a given drug and zero otherwise, ( see Table, S2 ) ., Significant clustering of the drugs according to mode of action, is visible ., For example the dihhryofolate reductase inhibitor methotrexate, co-clustered with related compounds aminopterin , trimetrexate , and, Bakers-soluble-antifolate ( triazinate ) ( Figure 4 , blue asterisks ) ; the sensitivities, to all four compounds were associated with 91 common pathways ., While one might, expect structural analogues such as these to produce a similar pattern of, sensitivity and hence similar pathway associations , structurally unrelated, compounds that share a common molecular target also co-clustered in certain, instances ., One interesting observation was the similarity in pathway, association , reflected by common membership of a cluster , of several structural, analogues , the anthracycline-based compounds , ( doxorubicin , zorubicin ,, danorubicin hydrochloride and deoxydoxorubicin ) with the podophyllotoxin-based, etoposide and teniposide ( Figure, 4 , green asterisks ) ., The four anthracyclines in the cluster share a, large proportion of associated pathways: 63 of the 401 pathways associated with, any of the anthracyclines are commonly associated to the effect of all four, compounds , and of these , 60 are also associated to the chemosensitivity to, either teniposide or etoposide ., Etoposide and its derivatives directly inhibit, topoisomerase II activity , followed by induction of DNA strand breaks and, selective cytotoxicity in tumour cells 25 whereas anthracyclines, intercalate DNA , indirectly inhibiting the progression of topoisomerase II and, blocking replication 26 ., Thus , the inter-omic pathway analysis is apparently, able to associate chemosensitivity phenotypes on the basis of a common, pathophysiological link independent of whether the key molecular targets are, affected directly , or indirectly by an upstream process ., While such mechanistic relationships were readily observable , the most prominent, division between the compounds , visible as the two largest clusters in Figure 4 , appeared to be, separating on the overall frequency of pathways associated with, chemosensitivity , with the top cluster in the diagram possessing on average 2 . 95, times the number of positive associations of the lower cluster ., While each of, the five platinum compounds in the dataset were most similar in pathway, associations to another platinum compound , they were separated across the two, largest clusters with cisplatin and carboplatin forming one group and, tetraplatin , iroplatin and diaminocyclohexyl-Pt II another ., This separation, agreed with the low numbers of common chemosensitivity pathways between members, of these two groups in earlier analyses ( Figure 2 ) ., Thus , the clustering structure did, not describe associations common across the platinum compounds , illustrating the, difficulty of using clustering approaches alone to identify pathways that may, determine class-specific chemosensitivity and the advantages of the consensus, phenotype approach ., To assess the specificity of the identified consensus platinum-sensitivity, pathways we compared these to the most frequently associated pathways in the, global inter-omic OR analysis ( Table S2 ) ., Of the 54 ( top 50 including ties ), most frequently associated pathways ( Table S3 ) , just seven intersect with pathways, identified by consensus phenotype integration , mostly related to, immunoregulatory processes ( “T-cell receptor” – Netpath;, “B-Cell receptor” – Netpath; “Rho GTPase cycle”, – Reactome; “lCK and FYN tyrosine kinases in initiation of TCR, activation” – BioCarta; “AMB2 integrin signalling”, – PID; “Immunoregulatory interactions between a Lymphoid and a, non-Lymphoid cell” – Reactome; and “TCR signalling in naive, CD8 T cells” – PID ) ., Hence the remaining 23/30 consensus, platinum-sensitivity pathways , dominated by metabolic processes , are not, associated with sensitivity to a wide range of chemotherapeutic agents and are, more likely to be specific to platinum sensitivity ., Our results show that an inter-omic , consensus phenotype approach to integration of, molecular profiles can reveal a cellular metabolic phenotype robustly associated, with platinum chemosensitivity across the NCI-60 cell line panel ., Many of the, specific aspects of this phenotype are consistent with the perturbations described, across many studies of tumour cell metabolism , and several of these have been, associated with the development , or likely acquisition , of drug resistance, phenotypes ., The classic hallmark of tumour cell metabolism is the Warburg effect: an, increase in glucose uptake and glycolysis to lactate even in normal oxygen, conditions ., In addition to the Warburg effect tumour cells are frequently reported, as exhibiting higher rates of glutaminolysis , fatty acid and lipid metabolism , and, nucleotide synthesis 27 ., Our observations from the NCI-60 molecular profiles, suggest a positive correlation between all of these phenotypes and platinum, chemosensitivity ., Figure 5 summarises some of the, key correlations observed between gene transcription , metabolite levels and platinum, sensitivity from the consensus pathways indicated by our analysis ., The relatively, higher levels of citrate and phosphoenolpyruvate ( PEP ) , observed in more sensitive, cell lines ( Figure 5A ) , are, consistent with low TCA cycle activity ( via product inhibition ) and increased, diversion of glycolytic intermediates into anabolic pathways such as the pentose, phosphate which feeds nucleotide synthesis 28 ., Under these conditions tumour, cells increase the uptake of glutamine and its conversion to oxaloacetate via, glutamate and 2-oxoglutarate ( 2-OG ) in order to replace TCA cycle intermediates and, NADPH 29 ., Both glutamate and 2-OG levels were also higher in more sensitive cell lines ., Thus, more ‘Warburg–like’ cells appear more sensitive to platinum, treatment than less metabolically transformed lines ., The selection of TCA cycle and, pyruvate metabolism as a sensitivity pathway in our analysis is likely to reflect, these associations ., The dependency of tumour cells on glycolysis for synthetic intermediates could be, exploited in platinum chemotherapy; for example the clinically-relevant glycolysis, inhibitor 2-deoxy-glucose ( 2-DG ) has been shown to enhance cisplatin cytotoxicity in, head and neck cancer cells 30 ., Interestingly , this synergy appeared to be mediated in, part via oxidative stress , a process that would lead to DNA lesions ( e . g ., 8-oxo-2′-deoxyguanosine ) requiring base excision repair ( BER ) which was one of, the key consensus sensitivity pathways selected by our analysis ( Figure 3 ) ., While it is clear that, nucleotide excision repair ( NER ) capacity is linked to cisplatin resistance 31 , 32 , 33; it is becoming, evident that BER is also important in the effect of cisplatin derived drugs 34 ., Cross-linking, of DNA via platinum derived drugs can increase the production of free radicals by, disrupting the cellular redox balance 35 ., We suggest that the, association of the BER pathway with four platinum drugs observed in the present, study is related to increased ROS production and not adduct formation ( repaired by, NER ) ., Intracellular levels of ROS seem vital to the cytotoxic effect of the platinum, derived drugs , further evidenced by the fact that oxaliplatin ( a later generation of, Pt drug ) is highly cytotoxic but forms less platinum-DNA adducts compared to equal, amounts of cisplatin 35 ., A particularly high degree of coordination between gene transcript and metabolite, levels was observed in nucleotide metabolism , revealing a robust association between, increased nucleotide synthesis , both de novo and via recovery of, catabolic intermediates , and tumour cell Pt sensitivity ( Figure 5B ) ., For example , in the de, novo pathway , we observed a positive correlation between levels of dUTP, ( a precursor to dTMP ) , expression of dUTP pyrophosphatase ( DUT, r\u200a=\u200a0 . 38 ) , expression of thymidylate synthase ( TYSY, r\u200a=\u200a0 . 27 ) and platinum sensitivity ., dUTP has to be hydrolysed, to dUMP by DUT to prevent the incorporation of uracils into DNA and suppression of, DUT has been shown to sensitize cells to other chemotherapeutics such as pyrimidine, anti-metabolites 36 ., Increased expression of nucleotide salvage pathway enzymes ( e . g . uracil, phosphoribosyl transferase or UPP ( r\u200a=\u200a0 . 20 ) ,, hypoxanthine-guanine phosphoribosyl transferase or HPRT ,, r\u200a=\u200a0 . 27 ) in sensitive cell lines was accompanied by decreases, in several intermediates of purine and pyrimidine catabolism ( namely guanine ,, guanosine , hypoxanthine , inosine , uracil , uridine and urea ) and increase in CMP , the, nucleotide product of HPRT ., Kowalski et al . 37 have shown clear links, between inactivation of salvage pathway enzymes such as HGPRT or loss of feedback, inhibition to AMP and GMP de novo synthesis and cisplatin, resistance in yeast ., Interestingly in the same study the addition of low, concentrations of extracellular purines also abolished cisplatin cytotoxicity; thus, the metabolome may have a causal influence on platinum sensitivity and not just, represent epiphenomena that is a passive consequence of aberrant cell division ., Our pathway analysis also predicts that lipid metabolism has a direct impact on, chemosensitivity ., We observed lower cholesterol , glycerol , and hexadecanoic acid, ( palmitate ) in more sensitive cell lines , together with negative correlations, between expression of apolipoprotein E ( APOE; mean, R\u200a=\u200a−0 . 21 ) , LDL receptor ( LDLR; mean, R\u200a=\u200a−0 . 27 ) and platinum sensitivity ( Figure 5C ) ., All these observations are consistent, with a hypothesis that increased uptake of lipoproteins and constituent, triglycerides , fatty acids and cholesterols can confer resistance to platinum , a, phenomenon previously shown in drug resistant leukemic cell lines 38 ., A related, pathway highlighted as associated with sensitivity was phosphatidylcholine, biosynthesis ., We observed a positive correlation between choline kinase ( CK ,, r\u200a=\u200a−0 . 28 , correlation to −log ( GI50 ) ) expression, and resistance to platinum ., Recent work by Shah et al . 39 in breast cancer cells have shown, that CK regulates pro-survival MAPk and PI3K/Akt signaling via phosphatidic acid ,, and that overexpression leads to drug resistance ., While previous pathway analysis was conducted on gene expression profiles alone from, the NCI60 dataset 40 , the use of correlation analysis and the combination of, metabolite and gene transcription measures in our study provides an unprecedented, level of detail into the contribution of metabolic pathways to drug sensitivity ., Using gene set enrichment analysis ( GSEA ) , Reidel et al . 40 suggested that , in addition to, a number of cell signaling and survival networks , methionine metabolism may, contribute to chemotherapeutic resistance to multiple agents , while fatty acid and, β-alanine metabolism were specifically associated with platinum-resistance ., In, the context of fatty acid metabolism we show here that lipid uptake and processing, may in fact be the driving factor in this association ., It is also interesting to, note that although we did not observe over-representation of β-alanine and, methionine metabolic pathways , both β-alanine and S-adenosylmethionine levels, were significantly positively correlated to platinum sensitivity , adding functional, evidence in support of these earlier findings ., At present , our study is one of very few that presents a strategy for simultaneous, interpretation of gene expression data , metabolic profiles and physiological, endpoints using biological pathway analysis , and has several advantages over other, approaches ., Multivariate analysis using pattern recognition algorithms such as PCA ,, 41 , PLS, 42 and, Kohonen Networks ( Self-Organising maps ) 43 , have been shown to be useful in, revealing novel associations between “-omics” datasets , but fail to take, into account prior biological knowledge relevant to the phenomenon at hand - a, feature which is clearly present in pathway-based techniques ., Gene and metabolite, coregulation at the pathway level has been previously studied using OR analysis, 44 , 45 ., Transcripts, significantly correlated to metabolite levels were examined for over-representation, of Gene Ontology terms 46 or pathways ( defined by MapMan BINS ) ., Bradley and, Gibons work reveals a degree of coordination present between transcriptional, and metabolic measurements at a pathway level , a necessary prerequisite for our, approach to be successful ., Importantly none of these examples use a function, physiological endpoint ( cytotoxicity ) as driver in pathway selection , leading to a, consensus phenotype description of the phenomenon of interest ., We show here that, such an approach is critical in reducing false positive selection of pathways ., All OR techniques share the limitation that they rely on a database containing, pre-defined pathways , and therefore cannot identify novel pathways or functional, modules ., In this work we have tried to overcome this limitation somewhat through, deconstructing the pathways which were significantly associated and then, functionally interpreting the elements of the pathways which showed significant, associations ( Figure 5 ) ., However , even this requires that the elements of the process are sufficiently, grouped in existing pathways to allow for those pathways to be significantly, associated ., Ultimately , a systems biology approach , such as the inter-omic pathway analysis, presented in our study , could assist the development of anti-resistance, chemotherapeutic strategies , and better individualization of treatment , i . e ., personalized medicine ., Using gene expression models ( GEMs ) based on cytotoxicity in, the NCI-60 panel , Williams et al . 47 were able to stratify tumour response and/or patient, survival in seven independent cohorts of patients with breast , bladder and ovarian, cancer ., Crucially , the in vitro derived GEMs outperformed those, derived directly from in vivo data ., Recently it has also been shown, that pre-treatment metabolic profiles can be used to predict the metabolic fate or, effect of drugs in rodents 48 , healthy humans 49 , 50 and breast cancer patients, 51 ., Given that, the metabolic phenotype of cancer is already the basis of imaging techniques such as, FDG-PET that are currently used to detect early responses to therapy , there is, potentially great value in combing such pharmaco-metabonomic studies with other, characterization of the patient or tumour genome and it is our belief that the, integration of molecular profile data yields more than the sum of its parts ., It, remains to be seen if the combination of “-omics” data provides a, competitive advantage over targeted biomarker studies for prognosis and prediction, of drug response in oncology ., It remains to be seen if the combination of, “-omics” data provides a competitive advantage over targeted biomarker, studies for prognosis and prediction of drug response in oncology ., Several major, challenges to such approaches and translation from in vitro studies , in particular, tumour heterogeneity , require further study ., However , irrespective of biomarker, development , the knowledge that chemotherapeutic sensitivity is in part determined, by the metabolic phenotype suggests that metabolic enzymes may be potential targets, in oncology for both d | Introduction, Results, Discussion, Methods | Using transcriptomic and metabolomic measurements from the NCI60 cell line panel ,, together with a novel approach to integration of molecular profile data , we show, that the biochemical pathways associated with tumour cell chemosensitivity to, platinum-based drugs are highly coincident , i . e . they describe a consensus, phenotype ., Direct integration of metabolome and transcriptome data at the point, of pathway analysis improved the detection of consensus pathways by 76% ,, and revealed associations between platinum sensitivity and several metabolic, pathways that were not visible from transcriptome analysis alone ., These pathways, included the TCA cycle and pyruvate metabolism , lipoprotein uptake and, nucleotide synthesis by both salvage and de novo pathways ., Extending the, approach across a wide panel of chemotherapeutics , we confirmed the specificity, of the metabolic pathway associations to platinum sensitivity ., We conclude that, metabolic phenotyping could play a role in predicting response to platinum, chemotherapy and that consensus-phenotype integration of molecular profiling, data is a powerful and versatile tool for both biomarker discovery and for, exploring the complex relationships between biological pathways and drug, response . | Resistance to chemotherapy drugs in cancer sufferers is very common ., Using a, panel of 59 cell lines obtained from different types of cancer we study the, links between the genes and metabolites measured in these cells and the, resistance the cells show to common cancer drugs containing platinum ., In order, to combine the information given by the genes and metabolites we introduce a new, pathway-based approach , which allows us to explore synergy between the different, types of data ., We then extend the procedure to look at a wider panel of drugs, and show that the pathways we found were associated with platinum are not just, the pathways which are frequently selected for a large number of drugs ., Given, the increasing use of multiple sets of measurements ( genes , metabolites ,, proteins etc ., ) in biological studies , we demonstrate a powerful , yet, straightforward method for dealing with the resulting large datasets and, integrating their knowledge ., We believe that this work could contribute to, developing a personalised medicine approach to treating tumours , where the, genetic and metabolic changes in the tumour are measured and then used for, prediction of the optimal treatment regime . | oncology/oncology agents, computational biology, computational biology/systems biology | null |
journal.pcbi.1007067 | 2,019 | Simulation of single-protein nanopore sensing shows feasibility for whole-proteome identification | Modern DNA sequencing techniques have revolutionized genomics 1 , but extending these methods to routine proteome analysis , and specifically to single-cell proteomics , remains a global unmet challenge ., This is attributed to the fundamental complexity of the proteome: protein expression level spans several orders of magnitude , from a single copy to tens of thousands of copies per cell; and the total number of proteins in each cell is staggering 2 ., Given the lack of in-vitro protein amplification assays the ability to accurately quantify both abundant and rare proteins hinges on the development of single-protein identification methods that also feature extraordinary-high sensing throughput ., To date , however , protein sequencing techniques , such as mass-spectrometry , have not reached single-molecule resolution , and rely on bulk averaging from hundreds of cells or more 3 ., Affinity-based method can reach single protein sensitivity 4 , but depend on limited repertoires of antibodies , thus severely hindering their applicability for proteome-wide analyses ., Consequently , in the past few years single-molecule approaches for proteome analysis based on Edman degradation 5 or FRET 6 have been proposed ., To date , however , profiling of the entire proteome of individual cells remains the ultimate challenge in proteomics 7 ., Nanopores are single-molecule biosensors adapted for DNA sequencing , as well as other biosensing applications 8 , 9 ., Recent nanopore studies extended nucleic-acid detection to proteins , demonstrating that ion current traces contain information about protein size , charge and structure 10–17 ., However , to date , the challenge of deconvolving the electrical ion-current trace to determine the protein’s amino-acid sequence from the time-dependent electrical signal has remained elusive ., In an analogy to the field of transcriptomics , in many practical cases it is sufficient to identify and quantify each protein among the repertoire of known proteins , instead of re-sequencing it ., Yao and co-workers showed theoretically that most proteins in the human proteome database can be uniquely identified by the order of appearance of just two amino-acids , lysine and cysteine ( K and C , respectively ) 18 ., But taking into account experimental errors , for example due to false calling of an amino-acid , or an unlabeled amino-acid , sharply reduces the ID accuracy ., Motivated by recent experiments suggesting the ability to translocate SDS-denatured proteins through either small nanopores ( ~0 . 5 nm ) 19 , or large nanopores 20 ( ~10 nm ) , and the possibility to differentiate among polypeptides based on optical sensing in nanopore 21 , we here introduce a protein ID method that according to simulation remains robust against the expected experimental errors ., We show that relatively low-resolution , tri-color , optical fingerprints produced during the passage of proteins through a nanopore , preserve sufficient information to allow a deep-learning classification algorithm to accurately identify the entire human proteome with >95% accuracy ., Even in cases where the apparent spatial and temporal resolutions of the optical system appear to be prohibitively low , and the amino-acids labelling efficiency is incomplete , whole proteome ID efficiency remains high and robust ., Particularly , the expected protein ID efficiency is of an extremely high clinical relevancy ., We illustrate the broad applicability of the method by analyzing the human plasma proteome , as well as commercially-available cytokine identification panel based on antibodies , showing that our antibody-free method can readily surpass current techniques in a number of key parameters , while displaying a near perfect accuracy ., In our method , proteins extracted from any source ( serum , tissue or cells ) , are denatured using urea and SDS ( Fig 1A ) ., Three amino-acids lysine ( K ) , cysteine ( C ) and methionine ( M ) are labeled with three different fluorophores using three orthogonal chemistries: the primary-amines in lysines are targeted with NHS esters; thiols in cysteines are targeted with maleimide groups , and methionines are labeled using the two-step redox-activated chemical tagging 22 ., The negatively charged SDS-denatured polypeptides are electrophoretically threaded , one at a time , through a sub-5 nanometer pore fabricated in a thin insulating membrane to ensure single file threading of the SDS-coated polypeptide ., The voltage , nanopore diameter and other factors , such as solution viscosity are used to regulate the protein translocations speed ., The nanopore is illuminated using laser beams for multi-color excitation 23 ., The excitation volume ( Fig 1A , yellow highlighted region ) is centered with the nanopore , and importantly , its axial depth is confined by plasmonic focusing of the incident electromagnetic field 24 ., Consequently , depending on the excitation depth , either a single or multiple labeled amino-acids will be simultaneously illuminated , during the passage of the protein ., Three-color fluorescence time traces ( “fingerprints” ) are recorded for each protein passage and are classified using deep-learning ( Fig 1B ) ., The theoretical likelihood of protein ID can be tested by calculating the percentages of unique matches of all proteins in the human Swiss-Prot database 25 based on the number and the order of appearance of three amino-acids only ., Simply counting the number of K , C and M residues in each protein identifies 72% of the total proteins uniquely , and another 14% identified as either one of two proteins in which one of them is the correct match ( online methods ) ., Moreover , the percentage of uniquely identified proteins is close to 99% with the determination of the KCM order of appearance along all proteins in the human proteome database ( Fig 1C ) ., Thus , in principle , the boundaries for the expected ID accuracies fundamentally permit whole-proteome , single-protein , identification ., The theoretical analysis shown in Fig 1C may be considered as an upper limit for the accuracy of a protein ID method based on a three amino-acid labelling , which neglects inter-dye distances ., However , it ignores experimental limitations , such as the sensing spatial and temporal constraints , the labelling efficiency and the photophysical properties of fluorophores ., These factors are likely to impact the accuracy of the protein ID method , and hence must be considered ., To this end we developed a detailed photophysical model to numerically calculate the time-dependent photon emission during the passage of each SDS-denatured protein through a solid-state nanopore ., Our model consists of three layers: first , we used Finite Difference Time Domain ( FDTD ) computations to evaluate the expected electromagnetic field distribution for a simple plasmonic structure fabricated on top of the nanopore ( Materials and Methods ) ., Second , an amino-acid labelling simulation was applied to each protein , in order to generate partial labelling of each of the three target amino-acids ., Finally , SDS-denatured proteins were allowed to slide through the plasmonic nanopore complex while illuminated at three distinct wavelengths ., The expected detected photon emissions were calculated at each step of the protein translocation taking into account the photophysical properties of the fluorophores , as well as energy transfer ( FRET ) , bleaching kinetics and collection efficiencies ., This allowed us to generate detailed photon emission time traces for each and every protein translocation ., To illustrate our method , we schematically show in Fig 2A snap-shots of the system at two time points during the passage of the PSD protein ., This figure is plotted in scale to illustrate the relative dimensions of the plasmonic field , the nanopore and the SDS-coated polypeptide chain ( marked as orange layer around the chain ) ., Specifically , the axial FWHM of the plasmonic field is 20 nm calculated from the FDTD field distribution , and the nanopore diameter is 3 nm ., Each protein was modeled as a fully-denatured , SDS-coated , wormlike polymer 26 , translocating across the nanopore at an instantaneous velocity ui = 〈u〉+δui where 〈u〉 is its average velocity , and the random term δui accounts for thermal fluctuations in its motion ., Since the SDS-coated biopolymers have a Kuhn length of approximately 7 nm 26 , they can be assumed to be partially-stretched ( unfolded ) wormlike polymers during translocation through a sub ~5 nm pore ., Moreover , when threaded through a 3 nm pore , the roughly 2 nm wide SDS-coated proteins are confined laterally in a small volume in the nanopore proximity where the electromagnetic field remains nearly constant ., Hence , in this study the protein translocations can be treated as one dimensional 27 ., The excitation profile calculated from the FDTD simulations was approximated by a one-dimensional Gaussian function as shown in S1 Fig . The fluorescence emission rate of each labeled amino-acid while passing through the excitation zone was modeled as a two-state system ( Fig 2C ) , as described in the Materials and Methods section ., Triplet state transition rates , which may result in microsecond-long dark-states were also investigated ( equations not shown ) based on literature values of three specific fluorophores 28–30 ., We explicitly took into consideration energy transfer rates ( Fig 2B and 2C ) , which directly depend on the amino-acid sequence , as well as photo-bleaching rates ( indicated by dotted yellow lines and solid grey arrows in Fig 2 , respectively ) ., At each time step of the simulation the emitted light from all fluorophores residing in the excitation zone were split to three spectrally-resolved , photon-counter channels as shown in Fig 2D ., In addition to the collection and detection efficiency of each channel , we also considered photon statistics by incorporating shot-noise ., The labeling efficiency was modeled by randomly positioning fluorophores at the K , C and M amino-acid , such that in each protein only a fraction Γj of them ( j represents K , C or M ) was actually labelled ( indicated by purple arrows in Fig 2A ) ., In all the following computational results presented the three amino-acids , K , C and M were labelled by Atto488 , Atto565 and Atto647N fluorophores , and the fluorophores properties were taken into account when simulating the photon emission rates ., Additionally , we introduced cross-labelling efficiency ( green arrows in Fig 2A ) , although this is known to be negligible 31 ., In order to estimate the translocation velocity of SDS-denatured polypeptides we performed electrical translocation measurements using SDS-denatured albumin ( 585 amino-acids ) proteins using ~4 nm-wide solid-state nanopores , as described in the Materials and Methods section ., Representative translocation events measured at a bias voltage of V = 300 mV , in which a single blockage current level is observed , are shown in Fig 3A ., Examining a statistical set of >900 translocation events showed a single blockade current level ( IB = 0 . 7 ) indicative of single-file polypeptide translocations ., This experiment supports the assumption that proteins are likely to be fully denatured as they thread through the narrow nanopore , in agreement with a previous publication 20 ., Fig 3B displays an overlay of the scatter plot of the fractional blockade current IB versus the translocation dwell-time tD , with its corresponding density map ., The area delimited by the dashed red lines approximate the typical full-width-half-maximum of a Gaussian centered on the characteristic dwell-time ( 94 . 3±7 . 2 μs as determined by the histogram shown in the inlet panel ) ., Accordingly , we estimate the mean translocation velocity by 0 . 2 cm/s ., Notably , this velocity is slower than the previous report , presumably due to the fact that in our experiments a much smaller nanopore was used ., We first focus on the simulated optical signals calculated for two proteins having nearly the same length: the EGF precursor , and its receptor EGFR ( 1208 and 1210 amino-acids , respectively ) ., Under near-ideal experimental conditions ( 100% labelling , 0 . 5 nm resolution , and velocity of 0 . 035 cm/s ) their tri-color fingerprints were readily distinguishable from each other , despite similar K , C and M compositions , and followed the actual K , C , M amino-acid order in each protein ( Fig 4A ) ., We then extended our protein translocation simulations under much lower spatial resolutions , lower labelling efficiencies and higher translocation velocities ., As expected , in the more realistic conditions we no longer can resolve individual fluorophore photon bursts , associated to single K , C or M residues ., Instead , the resulting signals appear as continuous tri-color fingerprints of each protein translocation ., Importantly , however , the fingerprints , even at the poorest resolution of 50 nm maintain an overall pattern characteristic of each protein ( Fig 4B ) ., Analyzing >5·107 single protein translocations events , under different conditions suggest that even at 100 nm resolution some characteristic features of each protein are preserved ( S2 Fig ) ., Moreover , we expect that small variations in the nanopore size would result in different translocation velocities ., To evaluate this effect , we repeated the translocation simulation experiments at mean values of 0 . 035 , 0 . 2 and 2 cm/s and increasing the translocation velocity fluctuations ( 20% , 30% and 40% of the mean velocity ) ., Our result presented in ( S3 , S4 & S8 Figs ) suggest that as long as the velocity is in the order of ~ 0 . 2 cm/s ( or below ) in accordance with our experimental result ( Fig 3 ) , the identification accuracy remains sufficiently high ., We tested the similarity among repeated translocations of the same proteins , which were subject to different labeling and random velocity fluctuations , by evaluating the Pearson correlation coefficients between all pairs of 50 translocation repeats of the same protein ., The results , showed in all cases high values ( 0 . 85–0 . 97 ) when considering auto-correlation ( Fig 5 , diagonal values ) ., In contrast , attempting to cross-correlate among 5 different , randomly-chosen , proteins produced in most cases much lower Pearson coefficient values ( 0 . 03–0 . 35 ) ., Obviously , this is just a small fraction of all possible cross-correlations ., However , even as is , this sample of data suggests that the protein translocation simulator generates highly-reproducible signals ., Next we vastly scaled-up our simulations to include thousands of different proteins , each one repeated hundreds of times under different labeling efficiencies , translocation velocities and spatial resolutions ., The accurate classification of noisy , low-resolution , time-dependent signals is often encountered in areas such as image and speech recognition and is effectively handled by Convolutional Neural Networks ( CNN ) approaches 33 , 34 ., We postulated that provided sufficient training , the CNN would be able to identify most proteins based on the tri-color fingerprints ., To check this hypothesis , we set up deep-learning whole-proteome analyses ., First , we trained the CNN network using a large data set containing at least 80 individual nanopore passages of each protein in the Swiss-Prot database ., Then the CNN was presented with new protein translocation events and queried as to the protein identity ., This procedure was repeated at least 5 times for whole-proteome analysis allowing us to establish the mean ID accuracy and its standard deviation , for 16 different experimental conditions ( Fig 6A ) ., Starting with the highest labelling efficiency ( 90% , right-hand set ) we observed that 96%-97% of all protein translocations were correctly identified , as long as the spatial resolution was ≤50nm ., The correctly identified protein fraction dropped down to 92% using a 100 nm resolution ., A similar pattern can be observed for the other labelling efficiencies with somewhat lower numbers ., In the worst-case scenario considered here ( 100 nm resolution and only 60% labeling efficiency ) the CNN nevertheless was able to correctly classify 68% of all translocation events , similar to the ideal case considered in Fig 1C , ( C , K , M counts only ) ., In other words , despite the fact that 40% of the target amino-acids were not labeled , and the resolution of the probing was about a third of the optical diffraction limit , the pattern recognition algorithm identified correctly nearly 70% of all protein translocation events ., When the labelling efficiency was improved to the expected standards ( between 70%-90% ) 22 , 35 , and the sensing resolution assumed to be in the 20–30 nm , the correct identification of all translocation was roughly 95% ., Increasing the translocation speed of proteins by nearly two orders of magnitude to 2 cm/s ( an order of magnitude higher than the mean measured velocity in Fig 3 ) , reduced the ID accuracy ( S8 Fig ) ., However , for high labeling efficiencies ( 80% and 90% ) the ID accuracy was high ( 72% and 81% , respectively ) ., In addition to the mean accuracies , the CNN algorithm produces a “confusion matrix” , which presents the number of times each and every protein x was identified as protein y ( where x and y could be any of the proteins in the set ) ., We used this information to calculate the probability density function ( pdf ) of correct ID for each and every classification set , namely the likelihood that a given protein is correctly identified with probability p ., The pdf of correct ID calculated for the case of 30 nm resolution and 80% labelling efficiency ( Fig 6A right panel ) indicates that 51% , 71% and 89 . 2% of proteins were correctly identified with probability of 1 . 0 , 0 . 98–1 . 0 and 0 . 9–1 . 0 , respectively ., The probability distributions for all other conditions are shown in SI S5 and S6 Figs ., We also analyzed the results for misclassified proteins ., Specifically , we were interested to know whether a misclassified protein is likely to be deterministically or randomly misclassified ., To investigate the degree of randomness in misclassification , we first selected proteins that had at least 10% misclassified events ., Then , we determined the fraction of identical mismatch ri = maxjnij/Ni for each protein i , where nij is the number of translocation events misidentified to protein j and Ni the total number of misclassified translocation events ., With this a high ri was characteristic of a deterministic misidentification , i . e . protein i is consistently mistaken with another specific protein j , and conversely a low ri was indicative of a rather random misidentification ., As shown in the right panel of Fig 6A , proteins were often confused with several others , suggesting a relatively high degree of randomness in misclassification , while only 10% were consistently mis-identified , that is with the same partner ., The distributions for all other conditions are shown in SI S5 and S7 Figs ., We further evaluated the performance of our approach for clinically-relevant applications including whole human plasma proteome and a cytokine panel ., In both studies , we kept the CNN training at the whole human proteome , rather than restricting it to the clinical sub-set ., Then we presented nanopore translocation traces of the plasma/cytokines proteins and evaluated the classification accuracy as before ., Interestingly for the high-spatial resolutions ( 20 nm and 30 nm ) the correct ID of the 3852 plasma proteins was only slightly larger than the whole proteome accuracy at the different labelling efficiencies , reflecting the fact that there is a small set of proteins that are hard to be classified in both cases ( Fig 6A and 6B right panel ) ., However , at the lower resolutions , especially for the 100 nm case in which we observed a significant drop in the ID accuracy for the whole proteome results , we still obtained very high scores for the plasma proteome ., Even at the lowest labelling efficiency of 60% at 100 nm resolution the CNN classified correctly 93% of all translocations ( Fig 6B ) ., In addition , the fraction of proteins correctly identified with probability between 0 . 9–1 . 0 improved over that of the whole-proteome classification , reaching 96 . 8% for the case of 30nm resolution and 80% labeling efficiency ., Finally , close to 30% of mis-identified proteins were consistently mistaken with another specific partner , suggesting that the accuracy of classification could be further significantly improved by relaxing the requirements of correct ID for selected proteins ., These results indicate that single-molecule plasma proteome application , which holds great clinical value , does not require extremely-stringent experimental resolutions or super-efficient labelling chemistries ( S9–S11 Figs ) ., The cytokine panel ( CytokineMAP 36 ) contains 16 proteins involved in inflammation , immune response and repair ., We evaluated the CNN classification under 16 different experimental conditions ( Fig 6C ) ., At the lowest labelling efficiency of 60% the ID accuracy drops between 43% - 85% , and at the realistic 80% labelling we obtain correct ID in the range of 73% - 97% ., However , despite the functional similarity between the candidate cytokines , and the wide range of conditions tested , each was distinguishable from all other cytokines within the commercial test panel ., This indicates that our approach has the potential to meet the requirements of a broad range of clinically relevant applications–that are less demanding than whole-proteome identification–with extremely high accuracies and yet very poor experimental conditions ( S12 Fig ) ., Single-molecule protein ID and quantification techniques are on the verge of revolutionizing the field of proteomics by enabling researches to achieve single-cell proteomics and to identify low abundance proteins that are essential biomarkers in biomedical and clinical research 7 ., Specifically , nanopore discrimination among poly-peptides based solely on two color labeling of C and K residues has recently been demonstrated 21 ., Here , we have proposed and simulated the feasibility and limits of a novel method for single-molecule protein ID and quantification using tri-color amino-acid tags and a plasmonic nanopore device ., Specifically , we designed a simulator that incorporates a range of physical phenomena to predict and model the behavior of our proposed device and performed a computational analysis taking into account a broad range of experimental conditions to characterize its performance ., Importantly , we developed a whole-proteome single-molecule identification algorithm based on convolutional neural networks providing high accuracies ( >90% overall ) , reaching up to 95–97% in challenging but attainable experimental conditions ., To facilitate the computational efforts , in this study we approximated each protein translocation dwell-time using a Gaussian distribution function ., Notably , past studies 37 successfully utilized CNN to identify signals from exponentially-distributed time-dependent signals , which may better reflect the experimental dwell-time distribution ( Fig 3 ) ., However , further studies will be required to evaluate the full impact of the temporal distributions of proteins translocation dwell-time on the CNN identification accuracy ., In clinical samples lysine residues may be post-translationally modified hence reducing their labelling efficiency ., To account for this effect and for the limitations in the chemical labelling yield , we evaluated the protein identification accuracy under partial labelling conditions ., Our results ( Fig 6 ) show that our tri-color protein identification method nevertheless largely circumvents this potential issue , yielding very high accuracies for up to 40% of unlabeled residues ., This is attributed to a redundancy in the tri-color labelling scheme that provides a higher degree of robustness against partial labelling ., Solid-state nanopores can process tens of individual proteins per second , and importantly because our method does not rely exclusively on measurements of the ion-current through the pore , it lends itself for parallel readout of high-density nanopore arrays fabricated on a sub mm2 membranes , using multi-pixel single-photon sensors 38 ., The versatility and robustness of convolutional neural networks tremendously simplify any calibration procedures and even potentially allow protein ID based on partial reads 39 ., This ensures that the whole-proteome ID is reliable and compatible with a wide variety of systems , able to overcome real experimental challenges ., Furthermore , in many cases ( notably for the plasma proteome ) misidentified proteins were consistently confused with another specific protein , which in a broad range of applications such as identifying disease-specific biomarkers , may not pose a significant issue as only small-subsets of the proteome are considered , or since the quantification of proteins can be cross-examined with expected counts ( e . g . low , medium or high abundance ) ., Finally , we evaluated the expected efficacy of our approach with commercially available applications , even resolving functionally similar proteins in rather poor experimental conditions ., The theoretical identification values were calculated using the human proteome Swiss-Prot database , which contains 20 , 328 entries ., For each entry we extracted the number of the target amino-acids ( C , K and M ) , as well as their order of appearance ., For example , the p53 protein would either be characterized by its C , K , M counts ( 10 , 20 , 12 , respectively ) or by the sequence below: MKMMMKKCKMCKCMKMCCCMCCMMCCKKKKKKMKKKKKKKMK , in which all intervening amino-acids were deleted ., Proteins having identical characteristic sequences ( or C , K and M counts ) are grouped together ., A protein is identified when it is the sole member of a group ., In the case of p53 , both the C , K and M counts and the characteristic sequence gave a unique identification ., The pie charts ( Fig 1C ) distribute the proteins according to the size of the group in which they belong to ., Each protein primary sequence was transformed into a string ( B ( i ) ) to which we assigned a value of 1 , 2 or 3 corresponding to each of the three aa tags ( K , C , and M ) , respectively; and 0 for all other aa in the protein sequence ., To account for partial or nonspecific labelling a set of randomly selected labeled positions in the string were omitted according to a given labeling efficiency ( ηL ) , and a set of artificial labeled positions were inserted according to a given nonspecific labeling efficiency ( ηNS ) ., It is important to note that nonspecific labeling did not affect all aa equally ., For instance , in generating a barcode for lysine ( K ) positions , nonspecific labeling could only be inserted at positions of either threonine , serine and tyrosine ( amino-acids which have been shown to compete with NHS-ester-based labeling ) with a probability of typically 1% 31 ., The strings were generated for the entire Swiss-Prot data base , and were re-generated each time to simulate an uneven labelling of the same protein data sets , as well as whenever we used different values of ηL and ηNS ., The three-dimensional near field enhancement of the plasmonic structure ( 2D vertical cross-section shown in Fig 2A ) was determined using a finite difference time domain ( FDTD ) 40 method solving for Maxwell’s time-dependent electromagnetic equations ., The architecture over which the FDTD computations were performed comprised a 10 nm-thick silicon ( Si ) membrane–exhibiting a 3 nm-wide nanopore–on top of which a gold ( Au ) plasmonic structure was deposited ( Fig 2D ) ., An additional 2 nm-thick titanium oxide ( TiO2 ) adhesive layer was inserted in between the Au structures and underlying Si membrane ., The plasmonic structure consisted of a gold ring ( inner and outer diameter of 12 and 32 nm , respectively , and a height of 40 nm ) centered at the nanopore and embedded inside a gold nanowell ( diameter of 120 nm and a height of 100 nm ) ., Water was used as the immersion media ., The excitation field was modeled as a total-field scattering-field source ( TFSFS ) 41 and the spatial sampling frequency was set to 5 nm-1 ( taking 60 frequency points over the 500–800 nm wavelength range ) ., The FDTD boundary conditions consisted of 8-layer PMLs ( perfectly matched layers ) symmetric in the x axis and antisymmetric in the y axis thus minimizing the reflections and the computational cost , respectively ., Frequency domain power monitors only were incorporated in the simulation to determine the near field enhancement in the vicinity of the nanopore ., All numerical simulations were performed using Lumerical FDTD Solutions ( Lumerical , Inc ) ., To simulate the translocation of the linearized protein through the nanopore , we assumed a unidirectional motion with steps of a single aa length ( Δ≈0 . 35 nm ) and an average velocity u ( cm/s ) ., To account for thermal fluctuations in this process we added a random noise term δu at each step ( δu can be positive or negative ) ., Hence the simulation step time of the i-th aa was defined as τi = Δ / ( u + δu ) ., The average protein velocity value was typically ~0 . 2 cm/s , based on experiments using SDS denatured proteins in solid-state nanopores as shown in Fig 3 ., Additionally , we tested faster translocations ( 2 cm/s ) ., The fluorescence emission rate of each fluorophore n in our system Kfl , j , n ( t ) was modeled as a two-state system:, Kfl , j , n ( t ) =kfl , jPj , n ( t ), Eq 1, where j = 1 . . 3 correspond to each of the three excitation/emission channels , kfl the fluorescence transition rate and Pn ( t ) the occupation probability of the excited molecular state S1 ., The fluorophores are excited by up to three laser lines corresponding to the three channels , that form sub-wavelength excitation volumes by means of a plasmonic nanostructure or total internal reflection ., The axial full width at half maximum of our Gaussian excitation volume Iex is defined as ξ and is allowed to vary from 5 nm to 200 nm in order to account for broad possible experimental conditions ., The emitted light from the three-color channels is assumed to be acquired with given efficiencies ηj , which include both the optical transmission efficiencies and the photodetector efficiencies ., The photon counts Iij at each channel j during each step i of the protein translocation is then determined by summing the emissions of all the fluorophores n that resides within the excitation volume ., Namely:, Iij=ηj∑nKfl , j , n ( ti ) +kbgτi=ηj∑nkfl , jPj , n ( ti ) +kbgτi, Eq 2, {Pj , n ( ti ) =Pj , n ( ti−1 ) + ( kex , j, ( n ) kj, ( n ) −Pj , n ( ti−1 ) ) ( 1−e−kj, ( n ) ti ) kj, ( n ) =kex , j, ( n ) +kS1 , j=σex , jIex , j, ( n ) λex , jhc0+τS1 , j−1, Eq 3 , Eq 4, where kbg is the background emission rate , ti the time at which step the translocation occurred such that ti−ti-1 = τi , kex , j, ( n ) is the excitation rate of the fluorophore n of channel j , σex , j is its absorption coefficient , λex , j is the excitation wavelength and τS1 , j is its excited state lifetime ., The number of cycles ( S0→S1→S0 ) undergone by each fluorophore was capped to account for photobleaching according to a decaying exponential distribution ., Specifically , the maximum number of cycles performed by each fluorophore before photobleaching was given by a random number drawn from a decaying exponential distribution with a characteristic decay of ~106 ., Finally , we applied a Poisson distribution to the photon counts Iij to simulate shot noise ., To include energy transfer ( such as Förster Energy Transfer and homo-transfer ) in our system we calculated a 2D distance matrix for each fluorophore in our system ., The distances between the labelled aa’s ( or fluorophores ) in each li | Introduction, Results, Discussion, Methods | Single-molecule techniques for protein sequencing are making headway towards single-cell proteomics and are projected to propel our understanding of cellular biology and disease ., Yet , single cell proteomics presents a substantial unmet challenge due to the unavailability of protein amplification techniques , and the vast dynamic-range of protein expression in cells ., Here , we describe and computationally investigate the feasibility of a novel approach for single-protein identification using tri-color fluorescence and plasmonic-nanopore devices ., Comprehensive computer simulations of denatured protein translocation processes through the nanopores show that the tri-color fluorescence time-traces retain sufficient information to permit pattern-recognition algorithms to correctly identify the vast majority of proteins in the human proteome ., Importantly , even when taking into account realistic experimental conditions , which restrict the spatial and temporal resolutions as well as the labeling efficiency , and add substantial noise , a deep-learning protein classifier achieves 97% whole-proteome accuracies ., Applying our approach for protein datasets of clinical relevancy , such as the plasma proteome or cytokine panels , we obtain ~98% correct protein identification ., This study suggests the feasibility of a method for accurate and high-throughput protein identification , which is highly versatile and applicable . | Macromolecules identification methods are central for most biological and biomedical studies , and while the field of genomics advanced to single-molecule resolution , the proteomic field still relies on bulk and costly techniques ., We describe a solution for single protein identification , based on the analysis of optical traces obtained from fluorescently-labeled proteins threaded through a nanopore and processed by a pattern recognition algorithm ., To evaluate the feasibility of our method we constructed computer simulations of the system , producing and analyzing nearly 108 individual protein translocations from the human Swiss-Prot database ., Our results suggest protein identification of >95% for the whole human proteome , even under non-ideal conditions ., These results constitute the basis for a novel whole proteome identification method , with single molecule resolution . | protein transport, innate immune system, medicine and health sciences, immune physiology, cytokines, chemical compounds, particle physics, energy transfer, cell processes, immunology, organic compounds, developmental biology, photons, molecular development, basic amino acids, amino acids, proteins, chemistry, proteomics, immune system, physics, biochemistry, biochemical simulations, cell biology, organic chemistry, proteomes, physiology, elementary particles, biology and life sciences, lysine, physical sciences, computational biology | null |
journal.pcbi.1005078 | 2,016 | Higher-Order Synaptic Interactions Coordinate Dynamics in Recurrent Networks | Understanding any complex system requires a mechanistic account of how dynamics arise from underlying architecture ., Patterns of connections shape dynamics in diverse settings ranging from electric power grids to gene transcription networks1–5 ., It is critical to establish how synaptic connectivity orchestrates the dynamics of propagating activity in neocortical circuitry , since dynamics are closely tied to cortical computation ., For example , trial-to-trial differences in network dynamics6–9 can be used to decode sensory inputs and behavioral choice10 , 11 ., It is particularly important to understand the transformation from connectivity to activity within local populations of neurons since this is the scale at which the majority of connections arise ., Locally , neocortical neurons are highly interconnected , and their connectivity schemes are characterized by the prevalence of specific motifs12 ., At the level of local populations , functional coordination has been demonstrated in diverse ways , e . g . on the basis of active neurons13 , 14 and their correlation patterns15 ., Yet predicting population responses on the basis of pairwise connections alone has proven to be difficult ., Establishing a mechanistic link between connectivity and dynamics in neocortical networks is intricate and non-trivial because individual neurons themselves are complex computational units16–20 ., Fundamentally , neurons are state dependent non-linear integrators of synaptic input21–23 ., When neurons in neocortex process information , they are generally subjected to numerous synaptic inputs which activate diverse receptors , and concomitant gating of voltage-dependent channels24–26 ., In consequence , neocortical neurons tend to operate in a high-conductance state , which lessens the impact of any one synaptic input21 , 27 ., Because inputs are weak individually , collective synaptic bombardments are necessary to depolarize a neuron to threshold for action potential generation ., As a result , it is difficult to predict the flow of activity through a synaptic network based solely on knowledge of single connections , without the context of ongoing activity in the entirety of the system ., Network models are an important tool for linking synaptic connectivity to dynamics in neocortex because they enable precise measurement and manipulation of simulated connectivity ., In this work , we generate networks comprised of leaky integrate-and-fire model neurons with naturalistic dynamics that mimic recordings from superficial neocortical layers ., Despite random synaptic topology in the model network , we find that small-world topological organization emerges in maps of propagating activity ., This paradoxical divergence of dynamics from synaptic connectivity is not explained by coactivity alone ., Rather , recruitment preferentially occurs in a selective subset of active connected pairs ., In the model , activity is preferentially routed through clustered fan-in triangles , despite their statistical scarcity ., Because they result in coordinated presynaptic timing , fan-in triangle motifs are particularly effective for spike generation ., By comparison , among neurons converging on a common target but lacking presynaptic interconnectivity , presynaptic timing is less synchronous on average , and postsynaptic recruitment is less likely ., Moreover , when we decrease the need for cooperative presynaptic action , by doubling synaptic weights in network models , the fan-in triangle motif becomes significantly less prevalent ., We evaluate the prediction of our model using high speed two-photon imaging of emergent network activity ex vivo , in somatosensory cortex ., We verify that propagating activity in real neuronal networks has small-world characteristics and elevated clustering , Decomposing this clustering , we discover that neocortical circuitry also manifests propagating activity that is dominated by the fan-in triangle motif ., These results suggest a mechanistic account for the widespread findings of clustered activity in neuronal populations 14 , 28–31 ., We suggest that clustered fan-in triangles are a canonical building block for reliable cortical dynamics ., Multineuronal dynamics are the computational substrate for sensation and behavior , implemented by synaptic architectures ., Propagating multineuronal activity arises from three main sources: the underlying connectivity itself , recent network history , and the non-linear integrative properties of individual neurons ., Here , multineuronal activity was modeled using conductance-based leaky integrate-and-fire neurons , stimulated with brief periods of Poisson input and recorded during self-sustained firing ( Fig 1A ) ., Model neurons were connected with heterogeneous synaptic weights drawn from a heavy-tailed distribution , in a random arrangement ( Erdős-Rényi; pee = 0 . 2 ) ., Simulated dynamics were asynchronous , irregular , and sparse , with critical branching ( see Methods ) ., A synaptic network was constructed for each simulation , consisting of excitatory model neurons and their synaptic connectivity ., For each structural iteration of the model we generated three distinct maps of activity ( and in two of the cases , multiplex connectivity and activity ) : a functional network , the active subnetwork , and a recruitment network ( Fig 2 ) ., Edges in the functional network summarized network dynamics and represented frequency of lagged firing between every pair of nodes ( with maximum interspike interval T = 25 ms; see Methods ) ., The active subnetwork was a subgraph of the synaptic network and consisted of model neurons active at least once and all their interconnections ( regardless of lagged firing relationships ) ., Finally , the recruitment network was a subgraph of the functional network defined by its intersection with the synaptic network , to map the routing of activity through synaptic interactions ., In this way , non-zero edges in the recruitment network linked synaptically connected nodes that also spiked sequentially in the interval T at least once ., For T = 25 ms , 10 . 9 ± 3 . 52 excitatory presynaptic input spikes immediately preceded each postsynaptic spike ( mean±std ) ., Surprisingly , although underlying synaptic connectivity was Erdős-Rényi ( i . e . random ) , functional activity networks were small world ( Fig 1B ) 32 ., To judge the small world character of these networks , global clustering coefficient and characteristic path were normalized by their respective abundances in density-matched Erdős-Rényi networks and combined as a quotient33 ., Comparison with density-matches was important given that sparseness itself results in enhanced smallworldness34 ., Functional networks were marked by significantly increased small world scores ( functional network: 2 . 8±0 . 23; synaptic network: 1 . 0±0 . 035; n = 5 , p = 0 . 0079 , Wilcoxon rank-sum ) resulting from increased clustering ( function: 2 . 8±0 . 23; synaptic network: 1 . 0±0 . 035 , n = 5 , p = 0 . 0079 ) , with characteristic path lengths similar to random-matches ( function: 1 . 0±6 . 4x10-4; synaptic network: 0 . 99±0 . 033; n = 5 , p = 0 . 69 ) ., The lag interval T was chosen to encompass important network timescales for synaptic plasticity and integration35 , 36 ., We also generated functional networks using intervals of 10 and 50 ms , which showed that the emergence of non-random features does not depend strongly on choice of T ( functional network for T = 10ms: small world ratio 3 . 2±0 . 24 , n = 5 , p = 0 . 0079; functional network for T = 50ms: small word ratio 2 . 6±0 . 22 , n = 5 , p = 0 . 0079 ) ., Given modest sampling conditions ( e . g . binning near timescales of synaptic integration ) , functional relationships can indicate locations of probable synaptic recruitment35 ., However , a subset of edges in functional networks are false positives—they reflect polysynaptic relationships and other combined statistical dependencies rather than monosynaptic connectivity and recruitment35 , 37 ., To determine whether these measurement artifacts were responsible for the statistical differences between functional and synaptic networks , we turned to recruitment networks ., Pruned of false positives , recruitment networks were significantly more small world than functional networks constructed from the same activity ( 4 . 6±0 . 87; n = 5 , p = 0 . 0079 ) , with even shorter characteristic paths ( recruitment: 0 . 65±0 . 072 , n = 5 , p = 0 . 0079 compared to function , Wilcoxon rank-sum ) and a similar elevation in clustering ( recruitment: 3 . 0±0 . 26; n = 5 , p = 0 . 22 ) ., Thus , emergent statistical structure in the functional networks reflected coordinated timing among multiple synaptically connected neurons ., As demonstrated by non-random recruitment , i . e . clustering in the recruitment network , activity did not propagate homogeneously through the random topology ., However , it remained a possibility that the seemingly non-random routing of activity was simply the byproduct of shared activity , without being selective on the basis of connectivity ., As a control , the active subnetwork establishes the role of interactions among neurons with elevated firing rates ( including pairs of neurons which never recruited one another within the interval T ) ., Compared to functional networks , the corresponding active subnetwork exhibited reduced small world ratio ( active network: 2 . 2±0 . 26 , n = 5 , p = 0 . 0159 ) and reduced clustering ( 1 . 3±0 . 041 , p = 0 . 0079 ) , despite somewhat shorter characteristic paths ( 0 . 60±0 . 055 , n = 5 , p = 0 . 0079 ) ., If directed connections that never fired sequentially were pruned from the active subnetwork , it would attain the same binary topology as the recruitment network ., Comparing the active network with the recruitment network , global clustering ratio was significantly increased ( from 1 . 3±0 . 041 to 3 . 0±0 . 26 , n = 5 , p = 0 . 0079 , Wilcoxon rank-sum ) ., Thus , the select connections which were directly involved in propagation of spiking activity were more clustered than activated connections as a whole ( Fig 1C ) ., We next evaluated whether neuronal pairs that never fired sequentially differed from those that did ., Comparisons were performed between in-degree matched samples ., Connected neurons that never fired in succession shared significantly fewer neighbors than those that did fire sequentially at least once ( n = 500 pairs , p = 3 . 1 x 10−17 , Wilcoxon rank-sum ) ., In the model , activity was selectively routed through interconnected neighborhoods ., Connectivity within a triplet is the simplest way two nodes can share a common neighbor and be clustered ., However , this measure fails to account for the direction of connection ., Since direction is crucial in synaptic communication , we turned to a formulation which differentiates directed triangle motifs38 ., From the perspective of a reference postsynaptic neuron , clustered neighbors can be arranged into four kinds of three-edge triangle motifs: fan-in , fan-out , middleman , and cycle arrangements ( Fig 3A ) ., Taken in isolation , fan-in , middle-node , and cycle triangles are isomorphic to one another through rotation , i . e . dependent on labeling the reference node ( which is necessary to compute local clustering ) ., Measures of undirected clustering can be decomposed fractionally into these four components ., Because the underlying model synaptic connectivity was random , none of the four triangle motifs were more prevalent than the others , and each contributed equally to synaptic clustering ( Fig 3B ) ., By contrast , in recruitment networks , fan-in triangle motifs were highly overrepresented ( Fig 3C ) ., The overrepresentation of fan-in triangle motif was also present in the functional network: for example , iterative Bayesian inference35 was sensitive to asymmetric directed clustering in model activity ( fan-in: 0 . 38±0 . 052 , fan-out: 0 . 29±0 . 032 , middleman: 0 . 19±0 . 016 , cycle: 0 . 15±0 . 0076; mean±std , threshold at the 95th percentile ) ., To understand whether these higher order asymmetric features emerge from chance correlations tied to firing rates , we generated Poisson populations that were rate-matched on a neuron-by-neuron and trial-by-trial basis ., This resulted in an inhomogeneous distribution of firing rates across all trails ., Our Poisson null populations had identical expected spike counts as model activity in each 100ms bin but no synaptic interactions and no causal propagation of activity ., Undirected clustering was significantly lower in iterative Bayesian maps of uncoupled Poisson rate-matched activity compared to connected network models ( Poisson rate-match: 0 . 0052±3 . 6x10-4; simulated activity: 0 . 024±0 . 013; Wilcoxon rank-sum p = 0 . 036; n = 3 ) , and the fan-in triangle motif was not elevated relative to other clustering patterns ( Fig 3D ) ., The Poisson populations demonstrated that elevated fan-in triangle motifs do not result trivially from the analysis procedure but instead are the result of synaptic interactions between neurons ., Interestingly , we found that model neurons with high fan-out clustering were characterized by elevated firing rates ( Fig 4A and 4B ) , but model neurons which comprised the fan-in triangle motif actually contracted towards low firing rates ( Fig 4C and 4D ) ., Fan-in triangles were more abundant in propagating activity than would be expected from their frequency in the synaptic network or component firing rates alone ., Like undirected clustering , the emergence of fan-in clustering in maps of propagating activity was robust to choice of T . Fan-in clustering was highly elevated in recruitment maps for T = 10 ms ( undirected: 0 . 0068±0 . 0007; fan-in 0 . 011±0 . 0017; fan-out: 0 . 0028±0 . 0001; middle-node: 0 . 0068±0 . 0007; cycle: 0 . 0052±0 . 0004; mean±std for 5 simulations ) and T = 50 ms ( undirected: 0 . 019±0 . 0015; fan-in 0 . 027±0 . 0027; fan-out: 0 . 0077±0 . 0003; middle-node: 0 . 019±0 . 0013; cycle: 0 . 015±0 . 0007; mean±std for 5 simulations ) ., Because of the different levels of sparseness in the numbers of connections these values should not be compared across values of T . Instead these analyses demonstrate that the over-representation of fan-in triangles is robust across a number of timescales ., To investigate the mechanism for overrepresentation of fan-in triangles in recruitment networks , we measured spike timing at their locations ., The signature of fan-in triangle motifs is convergence from interconnected presynaptic neurons , a motif that could potentially facilitate cooperative summation of synaptic inputs ., Consistent with this postulate , presynaptic neurons in fan-in triangle motifs were marked by increased probability of firing in the 10 ms prior to postsynaptic spiking ( Fig 5A and 5B ) ., We next compared differences in presynaptic timing relationships at loci of fan-in triangle motifs compared to loci of simple convergence , to assess the role of presynaptic interconnectivity ., For this analysis , random samples were obtained from epochs of coincident firing: 50 ms windows where every neuron in a triplet was active , centered on a spike in the postsynaptic reference neuron ., To avoid confounds from juxtaposing multiple motifs , neuron triplets with any additional connections , including recurrent loops , were excluded for this specific analysis alone ., As a result only fan-in triangles with exactly three interconnections were analyzed in this case ., We found fan-in presynaptic neurons were stereotypically ordered in a manner consistent with the direction of their interconnection , resulting in an asymmetric distribution of intervals between their firing ( Fig 5C ) ., In addition to the temporal structure imposed by this asymmetry , mean absolute timing difference between presynaptic neurons in clustered fan-in motifs was modestly but significantly more temporally precise than were neurons in simple convergence motifs ( 13 . 5±10 . 2 ms compared to 14 . 9±10 . 7 ms; Wilcoxon rank-sum on mean-absolute timing difference , p = 0 . 0035 , n = 1000 samples ) ., Moreover , we found that coincidence in fan-in triangle motifs occurred nearly twice as frequently as in motifs of simple convergence ( 1 . 9 ± 0 . 17 times more frequent , mean ± std; Wilcoxon rank-sum , p = 0 . 0079 , n = 5 model datasets ) ., Accounting for expected frequency of the two connection patterns in the underlying synaptic network , coincident activity is far more common at sites of fan-in triangles than at sites of simple convergence ( linear regression: slope 3 . 0 , y-intercept 0 . 00075 , n = 5 simulations , r2 = 0 . 91 , p = 0 . 011 ) ( Fig 5D ) ., We postulated that clustering is efficacious for synaptic integration and examined whether the prevalence of clustering was predictive of postsynaptic membrane potentials ., Pooling over all neurons and time bins , we binned the distribution of membrane voltages into segments that contained equal numbers of samples ( Fig 6A ) ., On average , because the model was active in the analyzed simulations , membrane voltages were depolarized from the resting equilibrium potential of -65 mV ( median: -60 . 2 mV; lower quartile: -63 . 6 mV; upper quartile: -56 . 8 mV ) ., To test our hypothesis , we generated functional networks that related recent presynaptic activity ( within a 25 ms interval ) to postsynaptic voltage ( Fig 6B; see Methods ) , yielding one network for each division of the voltage distribution ( Fig 6C ) ., These networks can be viewed as reverse correlograms conditioned on postsynaptic voltage , and differed in the statistics of their topologies across different voltage regimes ., At more negative membrane potentials , the active neurons which connected to the postsynaptic reference neuron ( and accounted for its recent excitatory synaptic drive ) were only modestly more clustered than random sparseness-matched controls ., As the postsynaptic neuron depolarized , the presynaptic nodes driving that depolarization became increasingly clustered , peaking at the threshold for firing ( Fig 6D ) ., Characteristic paths were similar to random graphs at all subthreshold voltages ., As a result of elevated clustering during membrane depolarization , small world ratios peaked at the most depolarized voltages corresponding to threshold for action potential generation ., These data support the hypothesis that activity among clustered presynaptic neurons is particularly effective for recruiting the postsynaptic neuron to spike ., The statistical incongruence of function and synaptic connectivity indicates that spiking activity does not flow in an egalitarian fashion through the synaptic network ., Instead , patterns of local clustering influence and direct where propagating activity occurs most frequently ., That is , patterns of activity are shaped by higher-order patterns in synaptic connectivity and not just pairwise couplings ., To further explore the dependence of activity flow on higher order synaptic connections we evaluated postsynaptic recruitment in a network model with a modest increase in mean synaptic strength ., Synaptic connections were twice as strong on average compared to the network models used throughout the remainder of this study but remained too weak to drive spiking alone ( Fig 7A ) ., The two network designs did not differ in connection density ., After synaptic weights were doubled , functional networks became more similar in topology to synaptic networks ( small world ratio decreased; Wilcoxon rank-sum , p = 0 . 0079 , n = 5 ) ( Fig 7B ) ., The double-strength models were less clustered ( Fig 7C ) ( Wilcoxon rank-sum , p = 0 . 0079 , n = 5 ) , and exhibited longer average path lengths ( Wilcoxon rank-sum , p = 0 . 0079 , n = 5 ) ., Directed clustering was compared across the two families of models ., Recruitment networks were analyzed with binary edges to control for their distinct mean synaptic weights ., In addition to their decreased overall clustering , the fan-in triangle motif was significantly rarer in double-strength recruitment networks ( Fig 7D ) ( from 0 . 030±0 . 0051 to 0 . 022±0 . 0025 , p = 0 . 030 , n = 6 ) , while the fan-out triangle motif showed a small but significant increase in abundance ( from 0 . 0040±2 . 0x10-4 to 0 . 0046±3 . 2x10-4 , p = 0 . 0043 , n = 6 ) ., Stronger presynaptic inputs reduced the need for extensive postsynaptic integration , allowing individual presynaptic cells to have a more independent impact on their postsynaptic partners ., As a result , statistics of propagating activity were more faithful to underlying pairwise connections in the models with increased synaptic strength ., In model simulations , fan-in triangle motifs were abundant in maps of function and recruitment ., We next evaluated whether the preponderance of fan-in triangle motifs would be robust to additional complexity in single-neurons and their connections ., Unlike the simple model neurons that we used for simulation , real neurons are complex elements16 and the connections between them are structured12 , 39 ., If clustered fan-in triangle motifs are a general feature of high-conductance nodes in a complex system , where coordinated inputs drive integration , the fan-in triangle will be overabundant in experimental dynamics ., This postulate would be falsified if all directed clustering motifs were equally common in functional networks ., To investigate , we analyzed high speed imaging data ( 20 Hz ) of spontaneous circuit activity collected ex vivo in mouse somatosensory cortex ( Fig 8A ) ( following 40 ) ., We generated functional networks from the imaged experimental data using an iterative Bayesian approach which is robust to relatively small numbers of observations 33 ., We then measured the prevalence of fan-in motifs in the functional topology ( Fig 8B ) ., Importantly , iterative Bayesian inference was not biased toward detection of fan-in triangle motifs , as demonstrated with rate-matched Poisson spiking ( see Fig 3D ) ., Though imperfect indicators , functional weights probabilistically identify the likelihood of true monosynaptic excitatory connectivity35 ., As a result , expected error rate for inferred connections can be adjusted with a sliding threshold on functional weight ., Stricter thresholds yield a more accurate approximation of the underlying recruitment network at the cost of restricted sampling ., Using inferred recruitment networks , beginning at the top quartile of inferred weights , directed clustering was computed in five-percentile increments ., Confidence intervals were obtained using bootstrap resampling under the assumption of a 30% false-positive rate ., As confidence of synaptic connectivity increased , the fan-in triangle motif became increasingly abundant and fan-out triangles less so ( Fig 8C ) ., Differences between the two motifs were significant ( threshold at 95th percentile , p = 4 . 8x10-34 , n = 100 bootstrap resampled functional networks , Wilcoxon ranksum ) ., We next measured whether strong functionally coupled neurons were more spatially proximal than random pairs ., We defined strong functional connections as those exceeding a 95% threshold on non-zero weights since previous work has indicated that these particular functional connections are more likely to reflect a causal synaptic connection35 ., We found that the median pairwise distance separating strong functionally connected cells was 249 μm , whereas randomly chosen pairs of neurons were separated by a median 263 μm ( Wilcoxon-ranksum p = 0 . 0336 , nfunctional = 638 , nrandom = 10000 ) ., We then measured triplets of neurons with functional connections that form triangles to determine whether these neurons were more spatially proximal to one another than randomly chosen triplets of neurons ., To investigate , proximity was quantified as the perimeter around the triangle formed by vertices at the spatial location of each neuron ., Neurons in functional triangles with mutual connectivity and at least three functional connections were inscribed by perimeters of median length 807 μm , compared to median perimeter of 823 μm for randomly selected triplets that were unconstrained by direction and number of edges ( Wilcoxon rank-sum p = 0 . 0097 , ntriangles = 2556 , nrandom = 10 , 000 ) ., Interestingly , triplets of neurons connected into arrangements of either simple divergence or simple convergence ( i . e . neurons in wedges , lacking interconnectedness between the common neighbors ) , were even more distant , inscribed by a perimeter of median 839 μm ( Wilcoxon rank-sum , ntriangles = 2556 , nwedges = 14 , 882 ) ., Thus , clustered triplets ( triangles ) tended to be arranged significantly more locally than simple convergent or simple divergent triplets ( wedges ) ., We then compared measures of clustering between the model , which was comprised of random connections , and the experimental data which almost certainly contained structured connectivity 12 , 39 to evaluate how the measure of fan in and fan-out triangles depend on the underlying structural topology ., To do so we used a measure of clustering propensity41 which allowed us to make comparisons of networks which have very different connection densities ., Clustering propensity ( 1-ΔCfan-in and 1-ΔCfan-out ) results in a normalized value where 1 is extreme clustering as seen in lattices , and 0 indicates no clustering above that expected in Erdős-Rényi random networks ., For the model , fan-in clustering was scored at 0 . 18 ± 0 . 019; and for the experimental data , fan-in clustering was scored at 0 . 20 ± 2 . 0x10-4 ( Wilcoxon ranksum p = 1 . 74x10-4 , nmodel = 5 simulations; ndata = 100 bootstrap samples ) ., Thus , fan-in clustering was modestly but significantly more abundant in maps of propagating activity based on experimental recordings ., We note that we compared thresholded graphs at the 80%-level ( i . e . top 20% of non-zero edges ) for this measure because the experimentally derived functional networks were not well-matched by regular lattices below this density ., Finally , we measured timing relationships among imaged active neurons ., Reliable timing relationships were measured independent of other functional analyses , using cross-correlations on the normalized fluorescence traces ( Methods ) ., Presynaptic coactivity was assessed as the product of the two z-scored presynaptic traces and compared to postsynaptic fluorescence as a straightforward cross correlation ., The resulting average cross-correlogram for fan-in triangles was stronger and more asymmetric than those measured from simple-convergence motifs ( Fig 8D ) ., Thus , presynaptic activity in fan-in triangles was more predictive of postsynaptic firing than presynaptic activity in motifs of simple convergence ., These results are consistent with fan-in triangles supporting coincident input and favoring reliable propagation of activity ., Results from the model indicated that the fan-in triangle motif temporally coordinates presynaptic inputs , rendering them more capable of driving recipient neurons to threshold ., Supporting our prediction of its fundamental importance for reliable recruitment , in acutely dissected neocortical tissue with more complex patterns of connectivity and intrinsic neuronal properties , we find a robust elevation of the same directed motif ., Using a model composed of random connections among leaky integrate-and-fire neurons with conductance-based synapses , we found that maps of propagating activity were structured and non-random ., Small-world patterning in the dynamics emerged because a specific higher-order connection pattern was particularly effective for postsynaptic integration: convergence of synaptic input from connected neighbors ., Synaptic connections between neighbors favored coincident timing of inputs onto their targets ., This coincident activation led to efficient postsynaptic integration ., As a consequence , clustering among active presynaptic cells tracked depolarization of model postsynaptic neurons ., Thus , activity was preferentially routed through fan-in triangle motifs ., In experimental recordings of emergent activity in hundreds of neurons ex vivo , after mapping inferred recruitment patterns 33 , we found that fan-in triangles were even more dramatically overrepresented than in the model ., These results are contextualized by increasing recognition of non-random functional structure in networks of neurons: Rich club structure has been reported ex vivo and in vivo31 ., Clustered30 , small world functional networks28 , and nucleation of dynamics29 have also been observed in neuronal cultures ., Since cultured populations differ from neocortex in the details of their topological makeup , these findings across model systems further suggest that clustering in general and the fan-in triangle motif in particular may be a canonical feature of propagating activity among interconnected neurons ., Despite differences in details of connectivity and neuronal intrinsic properties , dynamics are constrained by the requirement for coincident summation of individually weak inputs ., Constraining dynamics with beyond-pairwise relationships can be helpful for cortical computation ., Theoretical work has shown that non-uniform features of connection topology impact information transfer42 , and higher-order correlations were particularly impactful in low spike-rate regimes43 ., These complementary results from complex networks , statistical physics and network biology suggest that , by shaping feasible dynamics , the fan-in triangle motif could enhance information transfer from inputs to outputs ., We hypothesize that local circuits are organized around fan-in triangle motifs , promoting cooperative patterns of firing and stabilizing44 the propagation of activity despite individually unreliable neurons ., This canonical mechanism provides the coordination necessary to propagate signal despite weak synaptic connections ., Indeed , reliable sequential firing was associated with number of fan-in triangles even after controlling for overall in-degree ., Although clustering among fan-in triangles has not been tested directly until now , paired patch clamp recordings have shown that local neocortical circuitry is characterized structurally by abundant triplet motifs12 , 39 ., Our data and modeling suggest a functional consequence for a subset of these synaptic motifs: connected presynaptic neurons help establish coordinated timing among convergent inputs , leading to cooperative summation at the postsynaptic membrane ., Such cooperativity has been shown to be one potential mechanism capable of generating spike trains that are consistent with experimental observations in vivo45 ., While there are certainly explicit developmental rules that govern neuron to neuron connectivity , our results suggest that higher-order connectivity need not require specification a priori ., It could emerge autonomously if fan-in triangle motifs within a random network were stabilized and magnified during network development , e . g . by pruning non-recruiting connections through activity-dependent plasticity ., Thus , higher-order synaptic motifs that are particularly effective for postsynaptic recruitment could potentially self-organize46 ., These results do not indicate a complete schism between synaptic connectivity and dynamics—one clearly depends on the other ., However , their relationship is complicated by the integrative properties of single neurons ., Synaptic integration constrains feasible dynamics , and distributed synaptic motifs route the propagation of activity ., These interactions are a source of higher-order dynamical structure ., The routing of information is coordinated by higher-order synaptic patterns and the context of ongoing activity because the routing of spikes is determined by relative timing and collective interactions ., Simulations were implemented using the Brian Brain Simulator47 ., Model populations consisted of 1000 excitatory neurons , 200 inhibitory neurons | Introduction, Results, Discussion, Materials and Methods | Linking synaptic connectivity to dynamics is key to understanding information processing in neocortex ., Circuit dynamics emerge from complex interactions of interconnected neurons , necessitating that links between connectivity and dynamics be evaluated at the network level ., Here we map propagating activity in large neuronal ensembles from mouse neocortex and compare it to a recurrent network model , where connectivity can be precisely measured and manipulated ., We find that a dynamical feature dominates statistical descriptions of propagating activity for both neocortex and the model: convergent clusters comprised of fan-in triangle motifs , where two input neurons are themselves connected ., Fan-in triangles coordinate the timing of presynaptic inputs during ongoing activity to effectively generate postsynaptic spiking ., As a result , paradoxically , fan-in triangles dominate the statistics of spike propagation even in randomly connected recurrent networks ., Interplay between higher-order synaptic connectivity and the integrative properties of neurons constrains the structure of network dynamics and shapes the routing of information in neocortex . | Active networks of neurons exhibit beyond-pairwise dynamical features ., In this work , we identify a canonical higher-order correlation in network dynamics and trace its emergence to synaptic integration ., We find that temporally coordinated firing preferentially occurs at sites of fan-in triangles—a synaptic motif which coordinates presynaptic timing , leading to greater likelihood of postsynaptic spiking ., The influence of fan-in clustering leads to the surprising emergence of non-random routing of spiking in random synaptic networks ., When synaptic weights are made artificially stronger in simulation , so that cooperative input is less crucial , dynamics are no longer dominated by fan-in triangles but instead more closely reflect the random synaptic network ., Thus , the emergence of fan-in clustering in maps of synaptic recruitment is a collective property of individually weak connections in neuronal networks ., Because higher-order interactions are necessary to shape the timing of presynaptic inputs , activity does not propagate uniformly through the synaptic network ., Like water finding the deepest channels as it flows downhill , spiking activity follows the path of least resistance and is routed through triplet motifs of connectivity ., These results argue that clustered fan-in triangles are a canonical network motif and mechanism for spike routing in local neocortical circuitry . | action potentials, medicine and health sciences, neural networks, membrane potential, brain, electrophysiology, neuroscience, simulation and modeling, network analysis, research and analysis methods, computer and information sciences, network motifs, animal cells, neocortex, cellular neuroscience, cell biology, anatomy, physiology, neurons, biology and life sciences, cellular types, neurophysiology | null |
journal.ppat.1004835 | 2,015 | Conserved Streptococcus pneumoniae Spirosomes Suggest a Single Type of Transformation Pilus in Competence | Despite medical advances and vaccination campaigns , respiratory tract invasion by Streptococcus pneumoniae remains a leading mortality cause worldwide 1–3 ., A particular challenge in the prevention and treatment of pneumococcal infections lies in the bacterium’s striking genomic plasticity , as it allows for efficient antibiotic resistance development , capsular serotype switching and vaccine escape 4 ., Horizontal gene transfer and chromosomal rearrangements typically result from the avid uptake and recombination of exogenous DNA known as natural transformation ., A strictly regulated event , it occurs during a transitory state of the bacterium’s life cycle—competence—and requires the timed expression of a dedicated set of genes 5 ., Among these are the genes of the comG operon , which are conserved among naturally competent Gram-positive bacteria and are homologous to the ones encoding Type 4 pili ( T4P ) and Type 2 secretion system ( T2SS ) pseudo-pili components in Gram-negative bacteria 6 , 7 ., Although mechanistic studies of structural determinants for DNA uptake—such as putative transformation-specific cellular appendages—hold promise for the development of novel antiinfectives and helper compounds , there have been only limited and contradictory reports on the initial steps of this important biological process 8–10 ., As until recently no pilus-like structure had been observed in any transformable Gram-positive bacterium , it had been postulated that the pneumococcal comG operon encodes a short T2SS-like pseudo-pilus that serves to destabilize the cell wall peptidoglycan for DNA entry 6 , 9 , 11 ., The main experimental evidence for this model comes from a different transformable organism , Bacillus subtilis , where pilus length was indirectly deduced from biochemical data 9 ., Our team identified a long , micrometer-sized , T4P-like pilus protruding on the surface of competent cells from different pneumococcal strains with wild-type genotype ( Fig 1A ) 10 ., Among these are two highly transformable laboratory strains of different genetic background—R6 and TCP1251—as well as a capsulated clinical isolate—the G54 strain 10 ., We showed that major constituent of the transformation pilus is the ComGC pilin and that the pilus is sensitive to mechanical stress , which can lead to its detachment from the cell ( shearing ) 10 , 12 ., Finally , we showed that this transformation pilus binds extracellular DNA and proposed that it acts as the initial DNA receptor on the surface of competent pneumococci 10 ., A subsequent study visualized completely different structures—short , ‘plaited’ polymers—in the medium of competence-induced S . pneumoniae 8 ., Biochemical observation of significant ComGC release in the medium during competence convinced the authors that the ‘plaited’ structures corresponded to secreted transformation pili ., After failing to immunolabel these structures , they expressed heterologously the whole comG operon in Escherichia coli and visualized the release of similar polymers 8 ., Finally , they proposed a model , which is consistent with the classical but speculative model of transformation pseudo-pili: rather than acting as a DNA receptor , the pneumococcal transformation pilus acts as a peptidoglycan-drilling device whose release leaves a gateway for transforming DNA to find the uptake machinery 8 , 10 ., Here we show definitive experimental evidence that the short ‘plaited’ filaments are not transformation pili or other structural determinants of natural transformation ., We further identify the structures as fermentative spirosomes , or macromolecular complexes of the acetaldehyde-alcohol dehydrogenase enzyme AdhE , which is widely conserved across the bacterial kingdom ., Being aware of the limited view and resolution that observation by electron microscopy provides , we underscore the need for thorough validation by orthogonal approaches ., Finally , we briefly synthesize the present-day published collective knowledge by proposing an updated model of pneumococcal transformation ., Perhaps the most intriguing aspect of the Balaban et al . study is the distinctive morphology of the reported ‘plaited’ filaments themselves 8 ., As the authors point out , the genetic makeup of the comG operon resembles significantly that of operons encoding T4P or T2SS components in Gram-negative bacteria 5 , 7 , 8 , 13 ., This includes from sequence homology of the individual genes through their intraoperon organization to the putative bioassembly platform and post-translational modifications of the encoded components ., The structure of both T4 pili and T2SS pseudo-pili has been extensively studied 14 , 15 ., Generally T4 pilins pack tightly into thin but extremely strong and several micrometers long surface-attached helical filaments 14 ., Typical dimensions vary from 5–6 nm width for the T4aP of many bacteria ( Pseudomonas aeurginosa , Neisseria gonorrhoeae and others ) to the thicker , about 8 nm wide T4bP of enteropathogens such as Vibrio cholerae and Salmonella enterica serovar Typhi 14 ., Conversely , structural models of T2SS pseudopili , which normally act as short , protein ejecting pistons in the periplasm , present an architecture that is very similar to that of gonococcal T4aP ( Fig 1B ) 15 ., Both T4P and T2SS pseudo-pili feature a grooved surface with relatively small protuberations characteristic of the pilin helical packing 14 , 15 ., The characteristic structural features of T4P were conserved in the transformation pili that we visualized previously on the surface of competence-induced pneumococci from several different wild-type strains 10 ., In contrast , the ‘plaited’ structures visualized subsequently represent short , 40–200 nm long structures that are significantly thicker ( ~ 10 nm ) and present large protuberant domains on their helical surface ( Fig 1C ) 8 ., The authors proposed that the filaments are ‘plaited’ , i . e . composed of two interlaced transformation pili 8 ., Given the tight pilin packing in known T4P structures , however , we found it quite striking that homologous pili could sustain such a significant deformation to form an interlaced dimer ., Although molecular dynamics simulations revealed that T2SS pseudo-pili can adopt a wide range of helical twist angles , they were not shown to have a propensity for short-range bending 16 ., Moreover , while many T4P can form bundles , those can be hundreds of nanometers thick , contain a large number of pili and present much more limited distortion at the level of individual filaments 14 ., Intrigued by the striking new features of the ‘plaited’ filaments we wanted to see whether the transformation pili we previously reported could form similar structures ., We were able to visualize the ‘plaited’ polymers along with the T4P-like long transformation pili in wild-type pneumococci ( Fig 1D ) ., However , in a strain with an additional FLAG-tagged ectopic copy of comGC gene for the major pilin 10 ( S1 Table ) , we were able to immunolabel only the long , surface-attached pili using an anti-FLAG antibody ( Fig 1E ) ., Similar results were reported by Balaban and colleagues who failed to immunolabel the ‘plaited’ structures with a different antibody raised against ComGC 8 ., As negative immunolabeling results are difficult to interpret and can be due to a variety of technical and structural factors , we proceeded to investigate the role of the plaited filaments in competence ., We tested three negative control strains carrying either single-gene deletions for essential pilus biogenesis components ( S1 Table ) —the assembly platform protein ComGB 17 or the associated powering ATP-ase ComGA—or expressing a preprocessing incompetent ComGC variant ( ComGCE20V , or ComGCE5V in mature pilin residue numbering ) ., As shown previously , although these mutants can express monomeric pilins , they cannot assemble surface exposed pili and are transformation-incompetent 8 , 10 , 17 ., In all three ΔcomGA , ΔcomGB and comGCE20V strains we still detected release of ‘plaited’ filaments while expression of the long , T4P-like transformation pilus was abolished ( Fig 1F and Fig 1G and S1 ) ., Finally , we observed significant release of these structures even in the absence of competence induction ( Fig 2A ) , further confirming that they are not related to pilus biogenesis during natural transformation ., To identify the building subunits of the ‘plaited’ filaments , we developed an enrichment and purification protocol based on differential ultracentrifugation , microfiltration and size exclusion chromatography ., Electron microscopy imaging of the purified polymers showed a homogeneous sample composed primarily of the characteristic coiled polymers with an average length of 100–300 nm ( Fig 2A ) ., SDS-PAGE analysis of the corresponding fraction showed the presence of a predominant protein species with a molecular weight of ~100 kDa ., LC-MS/MS proteomic analyses on the excised , trypsin-digested gel band , as well as on the total purified filaments fraction ( S2 Table ) , unambiguously identified the predominant protein as bifunctional acetaldehyde-alcohol dehydrogenase AdhE , and the result was validated biochemically by Western blot detection using an anti-AdhE antibody ( Fig 2B ) ., Heterologous expression of the S . pneumoniae adhE gene in Escherichia coli and affinity purification of the expressed protein showed spontaneous coiled filament formation in the eluted fraction ( Fig 2C ) ., Finally , the protein composition of the ‘plaited’ polymers was validated by affinity pull-down with anti-FLAG antibody-conjugated resin on samples from a S . pneumoniae strain carrying an additional ectopic adhE gene copy for competence-inducible expression of a C-terminally FLAG-tagged protein ( Fig 2D ) ., AdhE is a 98 kDa protein with an N-terminal acetylating aldehyde dehydrogenase domain ( AldDH ) and a C-terminal Fe-dependent alcohol dehydrogenase domain ( Fe-ADH ) ( Fig 2E ) 18 ., Homologous dual domain proteins are common among fermentative bacteria and are reported to catalyze the NADH-dependent conversion of acetyl-CoA to ethanol via an aldehyde intermediate ., Most importantly , in many species AdhE has been shown to polymerize into fine helical filaments called spirosomes 19–26 that are morphologically consistent with the ‘plaited’ filaments discussed here and reported as self-secreting , ‘plaited’ transformation pili by Balaban and colleagues 8 ., A high resolution structural model of spirosome assembly by the closely homologous AdhE of Geobacillus thermoglucosidasius shows multimeric arrangement of the individual subunits into a right-handed spiral filament with six protomers per helical turn and overall pitch and width parameters consistent with negatively stained class averages of its pneumococcal counterpart ( Fig 2E and 2F ) 19 ., It is also important to note that the proposed spirosome structure—which is based on crystallographic and in-solution biophysical data , homology modeling and in silico macromolecular docking—corresponds to a single-start helix rather than a ‘plaited’ polymer of two or more interlaced filaments 19 ., To examine a putative role of AdhE in natural transformation , we first followed the protein’s expression over the course of competence induction that was verified by the detection of a competence-inducible FLAG-tagged ectopic copy of ComGC ., While we have shown competence-specific ComGC expression in wild-type genetic background previously 10 , AdhE protein levels remained stable over the course of the experiment ( Fig 3A ) ., We next constructed an adhE-null Streptococcus pneumoniae R6 mutant ( ΔadhE ) and examined its transformation efficiency for uptake of resistance-encoding DNA cassette under challenge with the corresponding antibiotic ., While the ΔadhE mutant shows slightly decreased transformation efficiency ( ~ 2-fold ) , this change is negligible compared to typical results under comGC disruption ( ~ 10 000-fold ) and can be due to reduced metabolic fitness under the microaerobic conditions of the experiment ( Fig 3B ) ., Our data are consistent with a previous genome-wide study aiming to identify genes essential for natural transformation in Streptococcus pneumoniae , which have failed to identify AdhE as a requirement for DNA uptake 27 ., Finally , while no spirosome release was detected for the competence-induced ΔadhE pneumococci , typical T4P-like transformation pili were observed ( Fig 3C and 3D ) ., Formation of spirosomes has been reported in a variety of Gram-positive and Gram-negative bacteria , with the first studies dating back several decades and refering to the building protein , AdhE , as spirosin 19–26 ., AdhE conservation across representative species with confirmed spirosome formation shows significant sequence homology even among relatively distant taxa ( Table 1 ) ., Nevertheless , sequence similarity mapping along the AdhEG . thermoglucosidasius structural model reveals that highly conserved residues cluster in only few surface-exposed patches 19 , 28 ., These correspond to the deep active site clefts of the two dehydrogenase domains , as well as sites at or near the interdomain linker ., The latter would likely remain buried in the context of mature spirosomes , as they stabilize embrace-like interactions between AdhE monomers in the high-resolution structural model of G . thermoglucosidasius spirosomes 19 ( S2 Fig ) ., Thus the exposed spirosome surface would retain significant variability , which in turn could translate into differences in spirosome morphology and stability across species ., Moreover , earlier reports have demonstrated that spirosome helix parameters can vary significantly depending on the presence and type of small molecule and metal ion cofactors 20 , 21 ., In addition to S . pneumoniae , we observed spirosome release in cultures of Clostridium difficile , Streptococcus sanguinis , and E . coli ( Fig 4A–4C and Table 1 ) 29 , 30 ., An adhE null strain of S . sanguinis 30 showed no release of morphologically consistent filaments , serving as an additional control for correct target identification ., For the two Gram-positive species , C . difficile and S . sanguinis , spirosome morphology was practically indistinguishable from that of S . pneumoniae ., The helical filaments we observed in E . coli cultures , on the other hand , were visibly more tightly coiled ( Fig 4D ) To verify that those correspond to AdhE macromolecular complexes we purified an enriched spirosome fraction and validated its major component as the bifunctional dehydrogenase using proteomic and biochemical methods ( Fig 4D and S3 Table ) ., Thus , although we expect morphological variations to be commonplace across species and even sample handling protocols , we are confident that spirosome assembly and extracellular release can be detected in many more environmental , clinically isolated , or genetically engineered bacterial strains ., The major horizontal gene transfer mechanism in S . pneumoniae—natural transformation—requires regulated expression of the comG pilus biogenesis operon , homologous to operons encoding T4P and T2SS pseudopili in Gram-negative bacteria 5 , 7 , 8 , 13 ., Recent studies have reported conflicting results regarding the morphology and function of pneumococcal transformation pili ., One proposed mechanism is that S . pneumoniae expresses a long , DNA-binding , T4P-like cell surface appendage to ‘fish’ extracellular transforming DNA 10 , while an alternative hypothesis argues that competent pneumococci express short , self-secreting T2SS plaited pili that perforate the cell wall peptidoglycan to allow for DNA entry 8 ., Here we show that the ‘plaited’ filamentous polymers are not related to natural transformation or pilus biogenesis but are instead widely conserved and well documented macromolecular complexes of the fermentative enzyme bifunctional acetaldehyde-alcohol dehydrogenase AdhE ., Its tandem domain architecture secures the two-step NADH-dependent reduction of acetyl-CoA to ethanol via an aldehyde intermediate 19–25 , 31 ., Although the biological significance of AdhE polymerization in such massive structures remains enigmatic , it is plausible that spirosome assembly delivers spatially localized metabolic flux to limit diffusion of the highly reactive aldehyde species and secure optimized conversion kinetics 32 ., In addition , the high resolution structural model of G . thermoglucosidasius spirosomes shows that polymer assembly buries ~ 6 500 Å2 of surface area per monomer , which could have dramatic effects on protein stability and function 19 , 31 ., Extracellular spirosome release by cultured bacteria appears to be the result of random cell lysis as no biological function or secretion mechanism can be assigned to the phenomenon ., As expected for a fermentative enzyme and consistent with reports in the literature , AdhE expression and spirosome assembly is expected to increase under microaerobic and anaerobic conditions as opposed to aerobic cultures 23 , 33 ., Anaerobic growth and increased cell lysis are both common in cultures of competence-induced pneumococci , where the signaling process of fratricide kills non-competent cells to release extracellular DNA available for uptake 17 , 34 ., This can explain why the spirosomes were initially associated to natural transformation in S . pneumoniae and reported to represent ‘plaited’ transformation pili 8 ., Conversely , in their study Balaban and colleagues attempted to reconstitute expression of pneumococcal transformation pili in E . coli by heterologous expression of the entire comG operon 8 ., As a result , extracellular release of spirosomes was readily detected and the AdhE polymers were again labeled erroneously ., It remains unclear why the authors failed to detect spirosomes in their negative control culture ., One possibility is that the structures were omitted in the limited observation field that high-magnification electron microscopy experiments provide ., Conversely , it is possible that overexpression of several non-native proteins—among which a macromolecular complex targeted to the inner membrane—could have had destabilizing effects on the expression strain , leading to increased cell lysis and spirosome release 35 , 36 ., In our quest to identify the structures observed by Balaban and colleagues 8 , we initially hypothesized that they were randomly released RecA nucleofilaments due to their striking similarity to polymerized RecA homologs from other bacterial and eukaryotic species ., Such structures have been extensively documented in the literature: forming characteristic helical coils , RecA filaments can be more or less extended depending on the presence and type of DNA and small-molecule ligands 37–39 ., Essential but not exclusive to natural transformation , cytosolic RecA is massively expressed during competence and its polymerization on the incoming single-stranded DNA is key to DNA protection and subsequent integration in the genome ( Fig 5 ) 40 ., Nevertheless , while ΔrecA cells display normal DNA uptake during competence , RecA-based recombination plays pivotal role in DNA repair throughout the bacterial life cycle 40 , 41 ., This indicates that pilus-dependent DNA uptake and RecA polymerization are intrinsically uncoupled and could have explained the continuous release of ‘plaited’ filaments in the pilus-defficient strains ., We were indeed able to detect RecA release in the medium of competence-induced cultures of both wild-type and ΔcomGB cells ( S3 Fig panel A ) ., Also , although with slightly different parameters in terms of helical width and pitch , the ‘plaited’ filaments were structurally similar to both the coiled structure of a eukaryotic RecA homolog ( S3 Fig panel B ) 37 , as well as to in vitro reconstituted RecAS . pneumoniae nucleofilaments ( S3 Fig panel C ) ., Nevertheless , release of the characteristic polymers persisted in a ΔrecA strain ( S3 Fig panel D ) and we were unable to detect the protein in the filament-enriched fractions following purification ( S1 Table ) ., It is therefore important to note that macromolecular organization in helical filaments is not uncommon among proteins from both the bacterial and other taxa ., These include but are not limited to nucleic acid-binding proteins , cytoskeletal elements , building blocks of cell surface appendages and phage capsid subunits 39 , 42–44 ., This , together with the markedly different helical parameters of the highly conserved E . coli spirosomes shown here underscores the fact that limited-view , low-resolution morphology imaging and bulk biochemical experiments alone are often insufficient to deduce the nature of macromolecular assemblies ., Rather , a combination of orthogonal approaches that spans the different resolution levels and integrates genetic , biochemical and structural data in a meaningful way is generally warranted to avoid false-positive or otherwise erroneous results ., Taken together , our data rule out the existence of short ‘plaited’ transformation pili in competent pneumococci and reassert the expression of a long , 5–6 nanometer wide appendage , structurally and compositionally similar to T4P in Gram-negative bacteria 10 ., This finding bridges Gram-negative and Gram-positives DNA uptake systems and provides a comprehensive picture of this major lateral gene transfer event ., Indeed , a recent study showed the existence of a T4-pilus on competent V . cholerae , which shares many features with the pneumococcal transformation pilus: competence-induced expression , prerequisite for DNA uptake , and roughly a single copy per cell 45 ., Apart from morphology alone , however , it is interesting to discuss the probable mechanism through which the transformation pili secure DNA entry into competent pneumococci ., Although expression of any type of pilus would require overcoming the physical barriers of cell-wall peptidoglycan and overlaying capsule—and thus possibly facilitate DNA entry—the similarities among transformation pili of Gram-positive and Gram-negative bacteria suggest that naturally transformable species might have evolved a conserved and more sophisticated mechanism of pilus function than simple cell-wall destabilization ., In agreement with a long-standing ‘pseudo-pilus’ hypothesis , Balaban and colleagues proposed a model in which the transformation pili self-secrete in the medium of competent S . pneumoniae , thus opening gateways in the cell wall peptidoglycan for passive exogenous DNA entry 8 , 9 ., Their hypothesis was supported by the observation that ComGC found in the supernatant of different S . pneumoniae strains after centrifugation correlates with the peak of transformation efficiency 8 ., Since we previously showed that transformation pilus expression is absolutely required for DNA uptake , it is not surprising to observe correlation between extracellular ComGC and transformation efficiency 10 ., However , ComGC release in culture supernatants can be a result from both cell lysis and/or compromised pilus integrity ., As we showed previously , pneumococcal transformation pili are fairly sensitive to mechanical stress and short vortexing and centrifugation are routinely used for their shearing and isolation 10 , 12 ., Such mechanical forces , however , are unlikely to be exerted in nature , where competent pneumococci are typically cushioned in protective biofilm matrix 46 ., As we have conducted only single time point visualization experiments , it is theoretically possible that the expresses transformation pili eventually detach from the cell to open entry pores for transforming DNA ( Fig 5 ) ., However , the sheer size and ATP-dependent assembly of the transformation pilus makes such self-secretion hypothesis unlikely: the observed long native pili would be energetically taxing on the cells if their sole function were to be ejected prior to DNA uptake ., Finally , it has been previously reported that native transformation pili bind and co-purify with DNA already present in the cell culture and that DNA binding at the surface of competent pneumococci is abolished in a pilus-deficient strain 10 , 47 ., DNA-binding is also conserved in the homologous T4P of Gram-negative bacteria 14 , 48 , 49 ., In such a DNA-binding context , pilus release would actually inhibit transformation by titrating out DNA available for uptake ( Fig 5 ) ., This once again argues against a self-secreting mechanism of function and reinforces a cell-surface attached role for the pilus in transformation ., While no mechanistic or quantitative data on DNA binding by the pilus are available , electron microscopy showed extensive contact interfaces between the long transformation pili and DNA chains 10 ., It is therefore plausible that multiple weak interactions along the helical pilus lattice stabilize this interaction and allow its reversal upon DNA uptake ., Such a scenario would also explain why no DNA binding to a non-polymerizing ComGC truncation has ever been detected 8 , 50 ., Even more interesting , however , is the question of how pilus-bound DNA gets brought to the DNA uptake machinery in the cell membrane ., In Gram-negative bacteria , T4P-bound DNA is proposed to be actively hauled to the cell by rapid bottom-up pilus depolymerization powered by a dedicated retraction ATPase 14 ., Although a similar mechanism has been proposed for S . pneumoniae and other transformable Gram-positive bacteria , pneumococci lack homologous retraction ATPase and are likely to utilize a distinct mechanism for DNA entry ., In addition , transforming DNA uptake occurs at much lower speeds in Gram-positive bacteria than Gram-negative T4P retraction 51 , 52 ., Many sequence-specific DNA binding proteins can scan DNA for their target sites at speeds several orders of magnitude higher than the upper limit for a three-dimensional diffusion-controlled process 53 ., This can generally be achieved by at least two passive mechanisms , which involve sequence non-specific DNA binding and subsequent translocation of the protein along the DNA:, 1 ) charge-based protein sliding , where the protein engages in a one-dimensional random walk along the DNA in search of its target , and, 2 ) direct intersegment transfer , where the protein can bind and hop between two remote regions on the DNA without losing the non-specifically bound state 53 ., Although we can not exclude the involvement of an unidentified retraction ATPase or additional receptor proteins in exogenous DNA uptake , we favor a model where the pneumococcal transformation pilus provides a similar facilitated diffusion framework ( Fig 5 ) ., By preserving multiple dynamic non-specific interactions with the pilus , transforming DNA would overcome the thermodynamic limitations of a three-dimensional diffusion process until it passively finds the membrane associated uptake machinery and becomes actively pumped in the cell ( Fig 5 ) ., Streptococcus pneumoniae spirosomes were observed in both competent and non-competent cells ., For competence induction cells were grown in microaerobic conditions , without agitation , at 37°C in Casamino Acid-Tryptone ( CAT ) medium supplemented with 0 . 2% glucose , 15mM dipotassium phosphate , 3mM sodium hydroxide and 1mM calcium chloride and adjusted to pH 7 . 8 ., Competence was triggered by the addition of Competence Stimulating Peptide ( CSP ) at OD600 = 0 . 15 for 10–30 min ., Non-competent pneumococci and Escherichia coli cells were grown similarly in LB to OD600 = 0 . 3 and OD600 = 0 . 6 , respectively ., Clostridium difficile cells were grown at 37°C under strict anaerobic conditions on Tryptone-Yeast extract-Glucose ( TYG ) plates supplemented with 0 . 1% thioglycolate ., Streptococcus sanguinis cells were grown anaerobically , without agitation , at 37°C in CAT medium to OD600 = 0 . 3 ., For spirosome visualization cells were scraped off the plates or pelleted by centrifugation and resuspended in TBS ( 50 mM Tris-HCl pH 7 . 6 , 150 mM NaCl ) at ~ 5 μl TBS per milliliter of culture at OD600 = 0 . 3 ., 5 μl drops of each suspension were then placed directly on glow discharged carbon coated grids ( EMS , USA ) for 1 minute ., The grids were then blot-dried on filter paper , washed on a drop of ultrapure water , and negatively stained with 2% uranyl acetate in water ., Specimens were examined on an FEI Tecnai T12 BioTWIN LaB6 electron microscope operating at 120 kV at nominal magnifications of 18500–68000 and 1–3 μm defocus ., Images were recorded on a Gatan Ultrascan 4000 CCD camera ., An adhE deletion ( strain AD001 ) was introduced in the R1501 genetic background by transformation with a DNA cassette carrying a kanamycin resistance gene inserted between two ~1000 base pair fragments corresponding to the S . pneumoniae genomic regions flanking adhE ., Briefly , the genomic region upstream from the AdhE open reading frame was amplified using forward and reverse primers 5’-ACA TGG CAA TCC GAT TCA TAA GGG G-3’ and 5’-GCC ATC TAT GTG TCG GAA CGA TAT CCT TTG TTA ATT TTT TCA CAA GTT TAT TAT AAC G-3’ , respectively , while the genomic region downstream of the adhE gene was amplified the following primer pair 5’-AAA ATG TGT TTT TCT TTG TTT TGT TTA TCA GTC TAG AAG CAA GAC AAA AAC TCA A-3’ and 5’-TTG CTA TTT ATG CAT GCA GAA GAC CAA ATG-3’ ., A third pcr reaction was used to amplify a kanamycin-resistance gene using the pR411 plasmid as template DNA 54 and forward and reverse primers 5’-AGG ATA TCG TTC CGA CAC ATA GAT GGC GTC GCT AGT-3’ and 5’-GCT TCT AGA CTG ATA AAC AAA ACA AAG AAA AAC ACA TTT TTT TGT CAA AAT TCG TTT-3’ , carrying complementarity to the 3’-end of the adhE-upstream and 5’-end of the adhE-downstream fragments , respectively ., The three pcr products were then assembled using overlap extension PCR and the purified DNA cassette was used for transformation of competence-induced S . pneumoniae R1501 cells ., adhE-null mutants ( strain AD001 ) were positively selected by growth in the presence of kanamycin ( 60 μg/ml ) and adhE deletion was confirmed independently by DNA sequencing and western blot detection using an anti-AdhE antibody ., For all transformation experiments , competence was triggered as above at OD600 = 0 . 15 for 10 minutes , followed by DNA addition and 20 minute incubation at 30°C ., Transformants were selected on Columbia Agar supplemented with 5% horse blood and appropriate antibiotics ., For the transformation efficiency assays , cells were transformed with 100 ng of a DNA cassette , amplified from S . pneumoniae R304 genomic DNA and containing the streptomycin resistance gene str41 ., Bacteria were plated in the presence and absence of streptomycin ( 100 μg/ml ) and incubated at 37°C overnight before colony counting ., All steps of the purification protocol were performed at 4°C ., 8L of S . pneumoniae culture grown in LB to OD600 = 0 . 3 or 4L of E . coli culture grown to OD600 = 0 . 6 were pelleted by centrifugation ( 20 min at 5000 g ) and resuspended in 6 ml of cold TBS , vortexed briefly and centrifuged to remove the bulk of intact cells and debris ( 10 min at 12000 g followed by 15 min at 50000 g ) ., Triton x-100 was added to the supernatant at final concentration of 0 . 25% ., Following 30 minute agitation for solubilization of remaining membrane fragments , the samples were filtered through a 0 . 45 μm cellulose acetate filter ( Corning ) and centrifuged for 1h at 125000 g for spirosome pelleting ., After careful removal of the supernatant , the pellet was resuspended in 50 μl TBS , re-filtered and loaded on a Superose 6 3 . 2/300 size exclusion column ( GE Healthcare ) ., Spirosome enriched fractions were found to elute with the void volume ., Sample preparation for electron microscopy was performed as above ., Trypsin digestion was performed as described previously 55 and the digests were analyzed under standard conditions on an LTQ-Orbitrap Velos ( Thermo Fisher , Bremen ) equipped with Ultimate 3000 nano-HPLC ( Dionex ) ., Briefly , tryptic peptides were desalted and separated on a C-18 nano-HPLC column under a gradient of acetonitr | Introduction, Results, Discussion, Materials and Methods | The success of S . pneumoniae as a major human pathogen is largely due to its remarkable genomic plasticity , allowing efficient escape from antimicrobials action and host immune response ., Natural transformation , or the active uptake and chromosomal integration of exogenous DNA during the transitory differentiated state competence , is the main mechanism for horizontal gene transfer and genomic makeover in pneumococci ., Although transforming DNA has been proposed to be captured by Type 4 pili ( T4P ) in Gram-negative bacteria , and a competence-inducible comG operon encoding proteins homologous to T4P-biogenesis components is present in transformable Gram-positive bacteria , a prevailing hypothesis has been that S . pneumoniae assembles only short pseudopili to destabilize the cell wall for DNA entry ., We recently identified a micrometer-sized T4P-like pilus on competent pneumococci , which likely serves as initial DNA receptor ., A subsequent study , however , visualized a different structure - short , ‘plaited’ polymers - released in the medium of competent S . pneumoniae ., Biochemical observation of concurrent pilin secretion led the authors to propose that the ‘plaited’ structures correspond to transformation pili acting as peptidoglycan drills that leave DNA entry pores upon secretion ., Here we show that the ‘plaited’ filaments are not related to natural transformation as they are released by non-competent pneumococci , as well as by cells with disrupted pilus biogenesis components ., Combining electron microscopy visualization with structural , biochemical and proteomic analyses , we further identify the ‘plaited’ polymers as spirosomes: macromolecular assemblies of the fermentative acetaldehyde-alcohol dehydrogenase enzyme AdhE that is well conserved in a broad range of Gram-positive and Gram-negative bacteria . | Streptococcus pneumoniae often escapes prevention and treatment through rapid horizontal gene transfer via natural transformation ., Uptake of exogenous DNA requires expression of a transformation pilus but two markedly different models for pilus assembly and function have been proposed ., We previously reported a long , Type 4 pilus-like appendage on the surface of competent pneumococci that binds extracellular DNA as initial receptor , while a separate study proposed that secreted short , ‘plaited’ transformation pili act simply as peptidoglycan drills to open DNA gateways ., Here we show that the ‘plaited’ structures are not competence-specific or related to transformation ., We further demonstrate that these are macromolecular assemblies of the metabolic enzyme acetaldehyde-alcohol dehydrogenase—or spirosomes—broadly conserved across the bacterial kingdom . | null | null |
journal.pgen.1008370 | 2,019 | Environmental and epigenetic regulation of Rider retrotransposons in tomato | Transposable elements ( TEs ) replicate and move within host genomes ., Based on their mechanisms of transposition , TEs are either DNA transposons that use a cut-and-paste mechanism or retrotransposons that transpose through an RNA intermediate via a copy-and-paste mechanism 1 ., TEs make up a significant part of eukaryotic chromosomes and are a major source of genetic instability that , when active , can induce deleterious mutations ., Various mechanisms have evolved that protect plant genomes , including the suppression of TE transcription by epigenetic silencing that restricts TE movement and accumulation 2–5 ., Chromosomal copies of transcriptionally silenced TEs are typically hypermethylated at cytosine residues and are associated with nucleosomes containing histone H3 di-methylated at lysine 9 ( H3K9me2 ) ., In addition , they are targeted by 24-nt small interfering RNAs ( 24-nt siRNAs ) that guide RNA-dependent DNA methylation ( RdDM ) , forming a self-reinforcing silencing loop 6–8 ., Interference with these mechanisms can result in the activation of transposons ., For example , loss of DNA METHYLTRANSFERASE 1 ( MET1 ) , the main methyltransferase maintaining methylation of cytosines preceding guanines ( CGs ) , results in the activation of various TE families in Arabidopsis 9–11 and in rice 12 ., Mutation of CHROMOMETHYLASE 3 ( CMT3 ) , mediating DNA methylation outside CGs , triggers the mobilization of several TE families , including CACTA elements in Arabidopsis 10 and Tos17 and Tos19 in rice 13 ., Interference with the activity of the chromatin remodelling factor DECREASE IN DNA METHYLATION 1 ( DDM1 ) , as well as various components of the RdDM pathway , leads to the activation of specific subsets of TEs in Arabidopsis ., These include DNA elements CACTA and MULE , as well as retrotransposons ATGP3 , COPIA13 , COPIA21 , VANDAL21 , EVADÉ and DODGER 14–17 ., Similarly , loss of OsDDM1 genes in rice results in the transcriptional activation of TE-derived sequences 18 ., In addition to interference with epigenetic silencing , TE activation can also be triggered by environmental stresses ., In her pioneering studies , Barbara McClintock denoted TEs as “controlling elements” , thus suggesting that they are activated by genomic stresses and are able to regulate the activities of genes 19 , 20 ., In the meantime , a plethora of stress-induced TEs have been described , including retrotransposons ., For example , the biotic stress-responsive Tnt1 and Tto1 families in tobacco 21 , 22 , the cold-responsive Tcs family in citrus 23 , the virus-induced Bs1 retrotransposon in maize 24 , the heat-responsive retrotransposons Go-on in rice 25 , and ONSEN in Arabidopsis 26 , 27 ., While heat-stress is sufficient to trigger ONSEN transcription and the formation of extrachromosomal DNA ( ecDNA ) , transposition was observed only after the loss of siRNAs , suggesting that the combination of impaired epigenetic control and environmental stress is a prerequisite for ONSEN transposition 28 ., Studies have further shown that stress-responsive TEs can affect the expression of surrounding genes , by providing novel regulatory elements and , in some cases , conferring stress-responsiveness 28–30 ., The availability of high-quality genomic sequences revealed that LTR ( Long Terminal Repeat ) retrotransposons make up a significant proportion of plant chromosomes , from approximately 10% in Arabidopsis , 25% in rice , 42% in soybean , and up to 75% in maize 31 ., In tomato ( Solanum lycopersicum ) , a model crop plant for research on fruit development , LTR retrotransposons make up about 60% of the genome 32 ., Despite the abundance of retrotransposons in the tomato genome , only a limited number of studies have linked TE activities causally to phenotypic alterations ., Remarkably , the most striking examples described so far involve the retrotransposon family Rider ., For example , fruit shape variation is based on copy number variation of the SUN gene , which underwent Rider-mediated trans-duplication from chromosome 10 to chromosome 7 ., The new insertion of the SUN gene into chromosome 7 in the variety “Sun1642” results in its overexpression and consequently in the elongated tomato fruits that were subsequently selected by breeders 33 , 34 ., The Rider element generated an additional SUN locus on chromosome 7 that encompassed more than 20 kb of the ancestral SUN locus present on chromosome 10 33 ., This large “hybrid” retroelement landed in the fruit-expressed gene DEFL1 , resulting in high and fruit-specific expression of the SUN gene containing the retroelement 34 ., The transposition event was estimated to have occurred within the last 200–500 years , suggesting that duplication of the SUN gene occurred after tomato domestication 35 ., Jointless pedicel is a further example of a Rider-induced tomato phenotype that has been selected during tomato breeding ., This phenotypic alteration reduces fruit dropping and thus facilitates mechanical harvesting ., Several independent jointless alleles were identified around 1960 36–38 ., One of them involves a new insertion of Rider into the first intron of the SEPALLATA MADS-Box gene , Solyc12g038510 , that provides an alternative transcription start site and results in an early nonsense mutation 39 ., Also , the ancestral yellow flesh mutation in tomato is due to Rider-mediated disruption of the PSY1 gene , which encodes a fruit-specific phytoene synthase involved in carotenoid biosynthesis 40 , 41 ., Similarly , the “potato leaf” mutation is due to a Rider insertion in the C locus controlling leaf complexity 42 ., Rider retrotransposition is also the cause of the chlorotic tomato mutant fer , identified in the 1960s 43 ., This phenotype has been linked to Rider-mediated disruption of the FER gene encoding a bHLH-transcription factor ., Rider landed in the first exon of the gene 44 , 45 ., Sequence analysis of the element revealed that the causative copy of Rider is identical to that involved in the SUN gene duplication 45 ., The Rider family belongs to the Copia superfamily and is ubiquitous in the tomato genome 34 , 45 ., Based on partial tomato genome sequences , the number of Rider copies was estimated to be approximately 2000 34 ., Previous DNA blots indicated that Rider is also present in wild tomato relatives but is absent from the genomes of potato , tobacco , and coffee , suggesting that amplification of Rider happened after the divergence of potato and tomato approximately 6 . 2 mya 45 , 46 ., The presence of Rider in unrelated plant species has also been suggested 47 ., However , incomplete sub-optimal sampling and the low quality of genomic sequence assemblies has hindered a comprehensive survey of Rider elements within the plant kingdom ., Considering that the Rider family is a major source of phenotypic variation in tomato , it is surprising that its members and their basic activities , as well as their responsiveness and the possible triggers of environmental super-activation , which explain the evolutionary success of this family , remain largely unknown ., Contrary to the majority of TEs characterized to date , previous analyses revealed that Rider is constitutively transcribed and produces full-length transcripts in tomato 34 , but the stimulatory conditions promoting reverse transcription of Rider transcripts that results in accumulation as extrachromosomal DNA are unknown ., To fill these gaps , we provide here a refined annotation of full-length Rider elements in tomato using the most recent genome release ( SL3 . 0 ) ., We reveal environmental conditions facilitating Rider activation and show that Rider transcription is enhanced by dehydration stress mediated by abscisic acid ( ABA ) signalling , which also triggers accumulation of extrachromosomal DNA ., Moreover , we provide evidence that RdDM controls Rider activity through siRNA production and partially through DNA methylation ., Finally , we have performed a comprehensive cross-species comparison of full-length Rider elements in 110 plant genomes , including diverse tomato relatives and major crop plants , in order to characterise species-specific Rider features in the plant kingdom ., Together , our findings suggest that Rider is a drought stress-induced retrotransposon ubiquitous in diverse plant species that may have contributed to phenotypic variation through the generation of genetic and epigenetic alterations induced by historical drought periods ., We used the most recent SL3 . 0 tomato genome release for de novo annotation of Rider elements ., First , we retrieved full-length , potentially autonomous retrotransposons using our functional annotation pipeline ( LTRpred , see Materials and Methods ) ., We detected a set of 5844 potentially intact LTR retrotransposons ( S1 Table ) ., Homology search among these elements identified 71 elements that share >85% sequence similarity over the entire element with the reference Rider sequence 45 and thus belong to the Rider family ., We then determined the distribution of these Rider elements along the tomato chromosomes ( Fig 1A ) and also estimated their age based on sequence divergence between 5’ and 3’ LTRs ( Fig 1A ) ., We classified these elements into three categories according to their LTR similarity: 80–95% , 95–98% and 98–100% ( S1A Fig ) ., While the first category contains relatively old copies ( last transposition between 10 . 5 and 3 . 5 mya ) , the 95–98% class represents Rider elements that moved between 3 . 5 and 1 . 4 mya , and the 98–100% category includes the youngest Rider copies that transposed within the last 1 . 4 my ( S1A Fig ) ., Out of 71 Rider family members , 14 were found in euchromatic chromosome arms ( 14/71 or 19 . 7% ) and 57 in heterochromatic regions ( 80 . 3% ) ( Table 1 ) ., In accordance with previous observations based on partial genomic sequences 34 , young Rider elements of the 98–100% class are more likely to reside in the proximity of genes , with 50% within 2 kb of a gene ., This was the case for only 37 . 5% of old Rider members ( 85–95% class ) ( Table 2 ) ., Such a distribution is consistent with the preferential presence of young elements within euchromatic chromosome arms ( 50% , 5/10 ) compared to old Rider elements ( 9 . 4% , 3/32 ) ( Table 2 and S1B Fig ) ., In addition , the phylogenetic distance between individual elements is moderately correlated to the age of each element ( Fig 1B ) ( S2 Table ) ., To better understand the activation triggers and , thus , the mechanisms involved in the accumulation of Rider elements in the tomato genome , we examined possible environmental stresses and host regulatory mechanisms influencing their activity ., Transcription of an LTR retroelement initiates in its 5’ LTR and is regulated by an adjacent promoter region that usually contains cis-regulatory elements ( CREs ) ( reviewed in 48 ) ., Therefore , we aligned the sequence of the Rider promoter region against sequences stored in the PLACE database ( www . dna . affrc . go . jp/PLACE/ ) containing known CREs and identified several dehydration-responsive elements ( DREs ) and sequence motifs linked to ABA signalling ( Fig 2A ) ., First , we tested whether these CREs were present in the LTR promoter sequences of the 71 de novo annotated Rider elements ( S3 Table ) ., Comparison of Rider LTRs to a set of gene promoter sequences of the same length revealed significant enrichment of several CREs in Rider LTRs ( Fisher’s exact test P<0 . 001 ) ( S4 Table ) ., It is known , for example , that the CGCG sequence motif at position 89–94 ( Fig 2A ) is recognized by transcriptional regulators binding calmodulin ., These are products of signal-responsive genes activated by various environmental stresses and phytohormones such as ABA 49 ., We also detected two MYB recognition sequence motifs ( CTGTTG at position 176–181 bp , and CTGTTA at position 204–209 bp ) ( Fig 2A ) ., MYB recognition sequences are usually enriched in the promoters of genes with transcriptional activation during water stress , elevated salinity , and ABA treatments 50 , 51 ., In addition , an ABA-responsive element-like ( ABRE-like ) was found at position 332–337 bp in the R region of Rider’s LTR , along with a coupling element ( CE3 ) located at position 357–372 bp ( Fig 2A ) ., The co-occurrence of ABRE-like and CE3 has often been found in ABA-responsive genes 52 , 53 ., The simultaneous presence of these five CREs in promoters of Rider elements suggests that Rider transcription may be induced by environmental stresses such as dehydration and salinity that involves ABA mediated signalling ., To test whether Rider transcription is stimulated by drought stress , glasshouse-grown tomato plants were subjected to water deprivation and levels of Rider transcripts quantified by RT-qPCR ( Fig 2B ) ., When compared to control plants , we observed a 4 . 4-fold increase in Rider transcript abundance in plants subjected to drought stress ., Thus , Rider transcription appears to be stimulated by drought ., To further test this finding , we re-measured levels of Rider transcripts in different experimental setups ., In vitro culture conditions with increasing levels of osmotic stress were used to mimic increasing drought severity ( Fig 2C ) ., Transcript levels of Rider increased in a dose-dependent fashion with increasing mannitol concentration , corroborating results obtained during direct drought stress in greenhouse conditions ., Interestingly , tomato seedlings treated with NaCl also exhibited increased levels of Rider transcripts ( Fig 2C ) ., ABA is a versatile phytohormone involved in plant development and abiotic stress responses , including drought stress 54 ., Therefore , we asked whether Rider transcriptional drought-responsiveness is mediated by ABA and whether increased ABA can directly stimulate Rider transcript accumulation ., To answer the first question , we exploited tomato mutants defective in ABA biosynthesis ., The lines flacca ( flc ) , notabilis ( not ) and sitiens ( sit ) have mutations in genes encoding a sulphurylase 55 , a 9-cis-epoxy-carotenoid dioxygenase ( SlNCED1 ) 56 , 57 , and an aldehyde oxidase 58 , respectively ., Both flc and sit are impaired in the conversion of ABA-aldehyde to ABA 55 , 58 , while not is unable to catalyse the cleavage of 9-cis-violaxanthin and/or 9-cis-neoxanthin to xanthoxin , an ABA precursor 57 ., Glasshouse-grown flc , not and sit mutants and control wild-type plants were subjected to water deprivation treatment and Rider transcript levels quantified by RT-qPCR ( Fig 2D ) ., Rider transcript levels were reduced in flc , not and sit by 43% , 26% and 56% , respectively ., To examine whether ABA stimulates accumulation of Rider transcripts , tomato seedlings were transferred to media supplemented with increasing concentrations of ABA ( Fig 2E ) ., The levels of Rider transcripts increased in a dose-dependent manner with increasing ABA concentrations ., This suggests that ABA is not only involved in signalling that results in hyper-activation of Rider transcription during drought , but it also directly promotes the accumulation of Rider transcripts ., The effectiveness of the treatments was verified by assaying expression of the stress- and ABA-responsive gene SlASR1 ( S2A–S2F Fig ) ., Identification in the U3 region of Rider LTRs of a binding domain for C-repeat binding factors ( CBF ) , which are regulators of the cold-induced transcriptional cascade 52 , 59 , led us to test Rider activation by cold stress ., However , Rider transcription was not affected by cold treatment , leaving drought and salinity as the predominant environmental stresses identified so far that stimulate accumulation of Rider transcripts ( S2G Fig ) ., The suppression of transposon-derived transcription by epigenetic mechanisms , which typically include DNA methylation , maintains genome integrity 2 , 3 , 5 ., We asked whether Rider transcription is also restricted by DNA methylation ., Tomato seedlings were grown on media supplemented with 5-azacytidine , an inhibitor of DNA methyltransferases ., Rider transcript steady-state levels increased in plants treated with 5-azacytidine compared to controls ( Fig 3A ) ., Comparison of Rider transcript accumulation in 5-azacytidine-treated and ABA-treated plants revealed similar levels of transcripts and the levels were similar when the treatments were applied together ( P <0 . 05; Fig 3A ) ., To further examine the role of DNA methylation in controlling Rider transcription , we took advantage of tomato mutants defective in crucial components of the RdDM pathway , namely SlNRPD1 and SlNRPE1 , the major subunits of RNA Pol IV and Pol V , respectively ., These mutants exhibit reduced cytosine methylation at CHG and CHH sites ( in which H is any base other than G ) residing mostly at the chromosome arms , with slnrpd1 showing a dramatic , genome-wide loss of 24-nt siRNAs 60 ., To evaluate the role of RdDM in Rider transcript accumulation , we first assessed the consequences of impaired RdDM on siRNA populations at full-length Rider elements ., Deficiency in SlNRPD1 resulted in a complete loss of 24-nt siRNAs that target Rider elements ( Fig 3B ) ., This loss was accompanied by a dramatic increase ( approximately 80-fold ) in 21-22-nt siRNAs at Rider loci ( Fig 3B ) ., In contrast , the mutation in SlNRPE1 triggered increases in both 21-22-nt and 24-nt siRNAs targeting Rider elements ( Fig 3B ) ., We then asked whether altered distribution of these siRNA classes is related to the age of the Rider elements and/or their chromosomal position , and thus local chromatin properties ., Compilation of the genomic positions and siRNA data in RdDM mutants didn’t reveal preferential accumulation of 21-22-nt siRNAs ( S3A Fig ) or 24-nt siRNAs ( S3B Fig ) over specific Rider classes ., Subsequently , we examined whether loss of SlNRPD1 or SlNRPE1 was sufficient to increase levels of Rider transcripts and observed increased accumulation of Rider transcripts in both slnrpd1 and slnrpe1 compared to WT ( Fig 3C ) ., We assessed whether this increase in Rider transcript levels is linked to changes in DNA methylation levels in Rider elements of RdDM mutants ., There was no significant change in global DNA methylation in the three sequence contexts in the 71 de novo annotated Rider elements ( S3C Fig ) , despite a tendency for young Rider elements to lose CHH in slnrpd1 and slnrpe1 ( S3D Fig ) ., Thus , the RdDM pathway affects the levels of Rider transcripts ., Also , features of Rider copies such as age and chromatin location alone cannot predict potential for activation based on DNA methylation levels ., The life cycle of LTR retrotransposons starts with transcription of the element , then the synthesis and maturation of accessory proteins including reverse transcriptase and integrase , reverse transcription , and the production of extrachromosomal linear ( ecl ) DNA that integrates into a new genomic location 61 ., In addition , eclDNA can be a target of DNA repair and can be circularised by a non-homologous end-joining mechanism or homologous recombination between LTRs , resulting in extrachromosomal circular DNA ( eccDNA ) 62–65 ., We searched for eccDNA to evaluate the consequences of increased Rider transcript accumulation due to drought stress or an impaired RdDM pathway on subsequent steps of the transposition cycle ., After exonuclease-mediated elimination of linear dsDNA and circular ssDNA , Rider eccDNA was amplified by sequence-specific inverse PCR ( Fig 4A ) ., Rider eccDNA was absent in plants grown in control conditions but was detected in plants subjected to drought stress ( Fig 4A ) ., Sanger sequencing of the inverse PCR products showed that the amplified eccDNA probably originates from the Rider_08_3 copy , which has 98 . 2% sequence homology of the 5’ and 3’ LTR sequences ( S4A Fig ) ., Residual sequence divergence may be due to genotypic differences between the reference genomic sequence and the genome of our experimental material ., Analysis of CREs in the LTR of the eccDNA revealed the presence of all elements identified previously with the exception of a single nucleotide mutation located in the CGCGBOXAT box ( S4A Fig ) ., Examination by quantitative PCR of the accumulation of Rider DNA , which included extrachromosomal and genomic copies , in drought-stressed plants also revealed an increase in Rider copy number due to eccDNA ( Fig 4B ) ., Importantly , Rider eccDNA was not detected in sit mutants subjected to drought stress ( Fig 4A ) , suggesting that induced transcription of Rider by drought stress triggers production of extrachromosomal DNA and this response requires ABA biosynthesis ., We also examined the accumulation of Rider eccDNA in plants impaired in RdDM ., Interestingly , Rider eccDNA was detected in slnrpd1 and slnrpe1 ( Fig 4C ) and increase in Rider DNA copy number due to eccDNA accumulation was confirmed by qPCR ( Fig 4D ) ., Absence of newly integrated genomic copies has been further validated by genome sequencing ., The eccDNA forms differed between the mutants ( Fig 4C ) ., Sequencing of Rider eccDNA in slnrpd1 showed a sequence identical to the Rider eccDNA of wild-type plants subjected to drought stress ., Thus , the Rider_08_3 copy is probably the main contributor to eccDNA in drought and in slnrpd1 ., In contrast , eccDNA recovered from slnrpe1 had a shorter LTR ( 287 bp ) and the highest sequence similarity with Rider_07_2 ( 89 . 2% ) ( S4B Fig ) ., Shortening of the LTR in this particular element results in the loss of the upstream MYBCORE as well as the CGCGBOXAT elements ( S4B Fig ) ., We then asked whether DNA methylation and siRNA distribution at these particular Rider copies had changed in the mutants ., DNA methylation at CHH sites , but not CG nor CHG , was drastically reduced at Rider_08_3 in slnrpd1 ( Fig 4E , S4C–S4E Fig and S5A Fig ) together with a complete loss of 24-nt siRNAs at this locus ( Fig 4F and S4F Fig ) but DNA methylation at Rider_07_2 was not affected , despite the deficiency of SlNRPD1 or SlNRPE1 ( Fig 4E , S4C–S4E Fig and S5B Fig ) ., Levels of 21-22-nt siRNAs in both mutants and 24-nt siRNA in slnrpe1 were increased ( Fig 4F and S4F and S4G Fig ) ., Altogether , this suggests that RdDM activity on Rider is highly copy-specific and that different components of the RdDM pathway differ in their effects on Rider silencing ., To examine the distribution of Rider retrotransposons in other plant species , we searched for Rider-related sequences across the genomes of further Solanaceae species , including wild tomatoes , potato ( Solanum tuberosum ) , and pepper ( Capsicum annuum ) ., We used the Rider reference sequence 45 as the query against genome sequences of Solanum arcanum , S . habrochaites , S . lycopersicum , S . pennellii , S . pimpinellifolium , S . tuberosum , and Capsicum annuum ( genome versions are listed in Materials and Methods ) ., Two BLAST searches were performed , one using the entire Rider sequence as the query and the other using only the Rider LTR ., Consistent with previous reports , Rider-like elements are present in wild relatives of tomato such as S . arcanum , S . pennellii and S . habrochaites; however , the homology levels and their lengths vary significantly between species ( Fig 5A ) ., While S . arcanum and S . habrochaites exhibit high peak densities at 55% and 61% homology , respectively , S . pennellii show a high peak density at 98% over the entire Rider reference sequence ( Fig 5A ) ., This suggests that the S . arcanum and S . habrochaites genomes harbour mostly Rider-like elements with relatively low sequence similarity , while S . pennellii retains full-length Rider elements ., To better visualize this situation , we aligned the BLAST hits to the reference Rider copy ( Fig 5B ) ., This confirmed that Rider elements in S . pennellii are indeed mostly full-length Rider homologs showing high density of hits throughout their lengths , while BLAST hits in the S . arcanum and S . habrochaites genomes showed only partial matches over the 4867 bp of the reference Rider sequence ( Fig 5B ) ., Unexpectedly , this approach failed to detect either full-length or truncated Rider homologs in the close relative of tomato , S . pimpinellifolium ., Extension of the same approaches to the genomes of the evolutionary more distant S . tuberosum and Capsicum annuum failed to detect substantial Rider homologs ( Fig 5A and 5B ) , confirming the absence of Rider in the potato and pepper genomes 45 ., As a control , we also analysed Arabidopsis thaliana , since previous studies reported the presence of Rider homologs in this model plant 45 ., Using the BLAST approach above , we repeated the results provided in 45 and found BLAST hits of high sequence homology to internal sequences of Rider in the Arabidopsis thaliana genome ., However , we did not detect sequence homologies to Rider LTRs ( Fig 5C and 5D ) ., Motivated by this finding and the possibility that Rider homologs in other species may have highly divergent LTRs , we screened for Rider LTRs that would have been missed in the analysis shown in Fig 5A and 5B due to the use of the full-length sequence of Rider as the query ., Using the Rider LTR as a query revealed that S . pennellii , S . arcanum and S . habrochaites retain intact Rider LTR homologs , but S . pimpinellifolium exhibits a high BLAST hit density exclusively at approximately 60% homology ., This suggests strong divergence of Rider LTRs in this species ( Fig 5C and 5D ) ., Overall , the results indicate intact Rider homologs in some Solanaceae species , whereas sequence similarities to Rider occur only within the coding area of the retrotransposons in more distant plants such as Arabidopsis thaliana ., Therefore , LTRs , which include the cis-regulatory elements conferring stress-responsiveness , diverge markedly between species ., Finally , we performed a reciprocal BLAST against tomato using Rider-like hits from all other species having sequence similarity over the entire element between 50% - 84% and confirmed that all Rider loci in tomato were among the top reciprocal BLAST hits ., To address the specificity of this divergence in Solanaceae species , we examined whether the CREs enriched in S . lycopersicum ( Fig 2A ) are present in LTR sequences of the Rider elements in S . pennellii , S . arcanum , S . habrochaites and S . pimpinellifolium ( Fig 5C ) ., While the LTRs identified in S . pennellii , S . arcanum and S . habrochaites retained all five previously identified CREs , more distant LTRs showed shortening of the U3 region associated with loss of the CGCG box ( S6 Fig and S5 Table ) ., This was observed already in S . pimpinellifolium , where all identified Rider LTRs lacked part of the U3 region containing the CGCG box ( S6 Fig ) ., Thus , Rider distribution and associated features differ even between closely related Solanaceae species , correlated with the occurrence of a truncated U3 region and family-wide loss of CREs ., Finally , to test the evolutionary conservation of Rider elements across the plant kingdom , we performed Rider BLAST searches against all 110 plant genomes available at the NCBI Reference Sequence ( RefSeq ) database ( www . ncbi . nlm . nih . gov/refseq ) ., Using the entire Rider sequence as the query to measure the abundance of Rider homologs throughout these genomes , we found Rider homologs in 14 diverse plant species ( S7 Fig ) ., The limited conservation of Rider LTR sequences in the same 14 species , revealed using the LTR sequence as the query , suggests that Rider LTRs are highly polymorphic and that drought-responsive CREs may nevertheless be restricted to Solanaceae ( S8 Fig ) ., Comprehensive analysis of individual LTR retrotransposon families in complex plant genomes has been facilitated and become more accurate with the increasing availability of high-quality genome assemblies ., Here , we took advantage of the most recent tomato genome release ( SL3 . 0 ) to characterize with improved resolution the high-copy-number Rider retrotransposon family ., Although Rider activity has been causally linked to the emergence of important agronomic phenotypes in tomato , the triggers of Rider have remained elusive ., Despite the relatively low proportion ( approximately 20% ) of euchromatic chromosomal regions in the tomato genome 32 ) , our de novo functional annotation of full-length Rider elements revealed preferential compartmentalization of recent Rider insertions within euchromatin compared to aged insertions ., Mapping analyses further revealed that recent rather than aged Rider transposition events are more likely to modify the close vicinity of genes ., However , Rider copies inserted into heterochromatin have been passively maintained for longer periods ., This differs significantly from other retrotransposon families in tomato such as Tnt1 , ToRTL1 and T135 , which show initial , preferential insertions into heterochromatic regions 66 ., TARE1 , a high-copy-number Copia-like element , is present predominantly in pericentromeric heterochromatin 67 ., Another high-copy-number retrotransposon , Jinling , is also enriched in heterochromatic regions , making up about 2 . 5% of the tomato nuclear genome 68 ., The Rider propensity to insert into gene-rich areas mirrors the insertional preferences of the ONSEN family in Arabidopsis ., Since new ONSEN insertions confer heat-responsiveness to neighbouring genes 28 , 69 , it is tempting to speculate that genes in the vicinity of new Rider insertions may acquire , at least transiently , drought-responsiveness ., We found that Rider transcript levels are elevated during dehydration stress mediated by ABA-dependent signalling ., The activation of retrotransposons upon environmental cues has been shown extensively to rely on the presence of cis-regulatory elements within the retrotransposon LTRs 48 ., The heat-responsiveness of ONSEN in Arabidopsis 26 , 27 , 70 , Go-on in rice 25 , and Copia in Drosophila 71 is conferred by the presence in their LTRs of consensus sequences found in the promoters of heat-shock responsive genes ., Thus , the host’s heat-stress signalling appears to induce transcriptional activation of the transposon and promote transposition 70 ., While ONSEN and Go-on are transcriptionally inert in the absence of a triggering stress , transcripts of Drosophila Copia are found in control conditions , resembling the regulatory situation in Rider ., Due to relatively high constitutive expression , increase in transcript levels of Drosophila Copia following stress appears modest compared to ONSEN or Go-on , which are virtually silent in control conditions 25–27 , 70 ., Regulation of Drosophila Copia mirrors that of Rider , where transcript levels during dehydration stress are very high but the relative increase compared to control conditions is rather modest ., The presence of MYB recognition sequences within Rider LTRs suggests that MYB transcription factors participate in transcriptional activation of Rider during dehydration ., Multiple MYB subfamilies are involved in ABA-dependent stress responses in tomato , but strong enrichment of the MYB core element CTGTTA within Rider LTRs suggests involvement of R2R3-MYB transcription factors , which are markedly amplified in Solanaceae 72 ., Members of this MYB subfamily are involved in the ABA signalling-mediated drought-stress response 73 and salt-stress signalling 74 ., This possible involvement of R2R3-MYBs in Rider is reminiscent of the transcriptional activation of the tobacco retrotransposon Tto1 by the R2R3-MYB , member NtMYB2 75 ., Drought-responsiveness has been observed for Rider_08_3 only , despite other individual Rider copies displaying intact MYB core element ( S3 Table ) ., This suggests that presence of this CRE is not the only feature required for drought-responsiveness , and other factors , such as genomic location , influence Rider activity ., Indeed , Rider_08_3 is located within a gene-rich area , with low TE content that might facilitate its activation ., This is striki | Introduction, Results, Discussion, Materials and methods | Transposable elements in crop plants are the powerful drivers of phenotypic variation that has been selected during domestication and breeding programs ., In tomato , transpositions of the LTR ( long terminal repeat ) retrotransposon family Rider have contributed to various phenotypes of agronomical interest , such as fruit shape and colour ., However , the mechanisms regulating Rider activity are largely unknown ., We have developed a bioinformatics pipeline for the functional annotation of retrotransposons containing LTRs and defined all full-length Rider elements in the tomato genome ., Subsequently , we showed that accumulation of Rider transcripts and transposition intermediates in the form of extrachromosomal DNA is triggered by drought stress and relies on abscisic acid signalling ., We provide evidence that residual activity of Rider is controlled by epigenetic mechanisms involving siRNAs and the RNA-dependent DNA methylation pathway ., Finally , we demonstrate the broad distribution of Rider-like elements in other plant species , including crops ., Our work identifies Rider as an environment-responsive element and a potential source of genetic and epigenetic variation in plants . | Transposons are major constituents of plant genomes and represent a powerful source of internal genetic and epigenetic variation ., For example , domestication of maize has been facilitated by a dramatic change in plant architecture , the consequence of a transposition event ., Insertion of transposons near genes often confers quantitative phenotypic variation linked to changes in transcriptional patterns , as documented for blood oranges and grapes ., In tomato , the most widely grown fruit crop and model for fleshy fruit biology , occurrences of several beneficial traits related to fruit shape and plant architecture are due to the activity of the transposon family Rider ., While Rider represents a unique endogenous source of genetic and epigenetic variation , mechanisms regulating Rider activity remain unexplored ., By achieving experimentally-controlled activation of the Rider family , we shed light on the regulation of these transposons by drought stress , signalling by phytohormones , as well as epigenetic pathways ., Furthermore , we reveal the presence of Rider-like elements in other economically important crops such as rapeseed , beetroot and quinoa ., This suggests that drought-inducible Rider activation could be further harnessed to generate genetic and epigenetic variation for crop breeding , and highlights the potential of transposon-directed mutagenesis for crop improvement . | biotechnology, retrotransposons, engineering and technology, gene regulation, plant science, genetic elements, plant genomics, sequence motif analysis, epigenetics, dna, plants, chromatin, dna methylation, research and analysis methods, bioengineering, sequence analysis, small interfering rnas, mobile genetic elements, sequence alignment, bioinformatics, chromosome biology, gene expression, plant genetics, chromatin modification, tomatoes, dna modification, fruits, biochemistry, rna, eukaryota, cell biology, nucleic acids, database and informatics methods, genetics, transposable elements, biology and life sciences, genomics, non-coding rna, plant biotechnology, organisms, solanum | null |
Subsets and Splits