Datasets:

alon-albalak

/

licensed_pubmed

like 0

PMC176545

Introduction {#s1} ============ Human malaria is caused by four species of the parasitic protozoan genus *Plasmodium*. Of these four species, Plasmodium falciparum is responsible for the vast majority of the 300--500 million episodes of malaria worldwide and accounts for 0.7--2.7 million annual deaths. In many endemic countries, malaria is responsible for economic stagnation, lowering the annual economic growth in some regions by up to 1.5% ([@pbio-0000005-Sachs1]). While isolated efforts to curb malaria with combinations of vector control, education, and drugs have proven successful, a global solution has not been reached. Currently, there are few antimalarial chemotherapeutics available that serve as both prophylaxis and treatment. Compounding this paucity of drugs is a worldwide increase in P. falciparum strains resistant to the mainstays of antimalarial treatment ([@pbio-0000005-Ridley1]). In addition, the search for a malaria vaccine has thus far been unsuccessful. Given the genetic flexibility and the immunogenic complexity of P. falciparum, a comprehensive understanding of *Plasmodium* molecular biology will be essential for the development of new chemotherapeutic and vaccine strategies. The 22.8 Mb genome of P. falciparum is comprised of 14 linear chromosomes, a circular plastid-like genome, and a linear mitochondrial genome. The malaria genome sequencing consortium estimates that more than 60% of the 5,409 predicted open reading frames (ORFs) lack sequence similarity to genes from any other known organism ([@pbio-0000005-Gardner1]). Although ascribing putative roles for these ORFs in the absence of sequence similarity remains challenging, their unique nature may be key to identifying *Plasmodium*-specific pathways as candidates for antimalarial strategies. The complete P. falciparum lifecycle encompasses three major developmental stages: the mosquito, liver, and blood stages. It has long been a goal to understand the regulation of gene expression throughout each developmental stage. Previous attempts to apply functional genomics methods to address these questions used various approaches, including DNA microarrays ([@pbio-0000005-Hayward1]; [@pbio-0000005-Ben1]; [@pbio-0000005-Le1]), serial analysis of gene expression ([@pbio-0000005-Patankar1]), and mass spectrometry ([@pbio-0000005-Florens1]; [@pbio-0000005-Lasonder1]) on a limited number of samples from different developmental stages. While all of these approaches have provided insight into the biology of this organism, there have been no comprehensive analyses of any single developmental stage. Here we present an examination of the full transcriptome of one of these stages, the asexual intraerythrocytic developmental cycle (IDC), at a 1-h timescale resolution. The 48-h P. falciparum IDC ([Figure 1](#pbio-0000005-g001){ref-type="fig"}A) initiates with merozoite invasion of red blood cells (RBCs) and is followed by the formation of the parasitophorous vacuole (PV) during the ring stage. The parasite then enters a highly metabolic maturation phase, the trophozoite stage, prior to parasite replication. In the schizont stage, the cell prepares for reinvasion of new RBCs by replicating and dividing to form up to 32 new merozoites. The IDC represents all of the stages in the development of P. falciparum responsible for the symptoms of malaria and is also the target for the vast majority of antimalarial drugs and vaccine strategies. ::: {#pbio-0000005-g001 .fig} Figure 1 ::: {.caption} ###### Parasite Culturing and Data Characteristics of the P. falciparum IDC Transcriptome Analysis \(A) Giemsa stains of the major morphological stages throughout the IDC are shown with the percent representation of ring-, trophozoite-, or schizont-stage parasites at every timepoint. The 2-h invasion window during the initiation of the bioreactor culture is indicated (gray area). (B--D) Example expression profiles for three genes, encoding EBA175, DHFR-TS, and ASL, are shown with a loess fit of the data (red line). \(E) MAL6P1.147, the largest predicted ORF in the *Plasmodium* genome, is represented by 14 unique DNA oligonucleotide elements. The location of each of the oligonucleotide elements within the predicted ORF and the corresponding individual expression profiles are indicated (oligo 1--14). A red/green colorimetric representation of the gene expression ratios for each oligonucleotide is shown below the graph. The pairwise Pearson correlation for these expression profiles is 0.98 ± 0.02. \(F) The percentage of the power in the maximum frequency of the FFT power spectrum was used as an indicator of periodicity. A histogram of these values reveals a strong bias toward single-frequency expression profiles, indicating that the majority of P. falciparum genes are regulated in a simple periodic manner. This bias is eliminated when the percent power was recalculated using random permutations of the same dataset (inset). For reference, the locations of EBA175 (peak B), DHFR-TS (peak C), and ASL (peak D) are shown. :::  ::: Our laboratory has developed a *P. falciparum--*specific DNA microarray using long (70 nt) oligonucleotides as representative elements for predicted ORFs in the sequenced genome (strain 3D7) ([@pbio-0000005-Bozdech1]). Using this DNA microarray, we have examined expression profiles across 48 individual 1-h timepoints from the IDC of P. falciparum. Our data suggest that not only does P. falciparum express the vast majority of its genes during this lifecycle stage, but also that greater than 75% of these genes are activated only once during the IDC. For genes of known function, we note that this activation correlates well with the timing for the respective protein\'s biological function, thus illustrating an intimate relationship between transcriptional regulation and the developmental progression of this highly specialized parasitic organism. We also demonstrate the potential of this analysis to elucidate the function of the many unknown gene products as well as for further understanding the general biology of this parasitic organism. Results {#s2} ======= Expression Profiling of the IDC {#s2a} ------------------------------- The genome-wide transcriptome of the P. falciparum IDC was generated by measuring relative mRNA abundance levels in samples collected from a highly synchronized in vitro culture of parasites. The strain used was the well-characterized Honduran chloroquine-sensitive HB3 strain, which was used in the only two experimental crosses carried out thus far with P. falciparum ([@pbio-0000005-Walliker1]; [@pbio-0000005-Wellems1]). To obtain sufficient quantities of parasitized RBCs and to ensure the homogeneity of the samples, a large-scale culturing technique was developed using a 4.5 l bioreactor (see [Materials and Methods](#s4){ref-type="sec"}). Samples were collected for a 48-h period beginning 1 h postinvasion (hpi). Culture synchronization was monitored every hour by Giemsa staining. We observed only the asexual form of the parasite in these stains. The culture was synchronous, with greater than 80% of the parasites invading fresh RBCs within 2 h prior to the harvesting of the first timepoint. Maintenance of synchrony throughout the IDC was demonstrated by sharp transitions between the ring-to-trophozoite and trophozoite-to-schizont stages at the 17- and 29-h timepoints, respectively ([Figure 1](#pbio-0000005-g001){ref-type="fig"}A). The DNA microarray used in this study consists of 7,462 individual 70mer oligonucleotides representing 4,488 of the 5,409 ORFs manually annotated by the malaria genome sequencing consortium ([@pbio-0000005-Bozdech1]). Of the 4,488 ORFs, 990 are represented by more than one oligonucleotide. Since our oligonucleotide design was based on partially assembled sequences periodically released by the sequencing consortium over the past several years, our set includes additional features representing 1,315 putative ORFs not part of the manually annotated collection. In this group, 394 oligonucleotides are no longer represented in the current assembled sequence. These latter ORFs likely fall into the gaps present in the published assembly available through the *Plasmodium* genome resource [PlasmoDB.org](PlasmoDB.org) ([@pbio-0000005-Gardner1]; [@pbio-0000005-Kissinger1]; [@pbio-0000005-Bahl1]). To measure the relative abundance of mRNAs throughout the IDC, total RNA from each timepoint was compared to an arbitrary reference pool of total RNA from all timepoints in a standard two-color competitive hybridization ([@pbio-0000005-Eisen1]). The transcriptional profile of each ORF is represented by the mean-centered series of ratio measurements for the corresponding oligonucleotide(s) ([Figure 1](#pbio-0000005-g001){ref-type="fig"}B--1E). Inspection of the entire dataset revealed a striking nonstochastic periodicity in the majority of expression profiles. The relative abundance of these mRNAs continuously varies throughout the IDC and is marked by a single maximum and a single minimum, as observed for the representative schizont-specific gene, erythrocyte-binding antigen 175 (*eba175*), and the trophozoite-specific gene, dihydrofolate reductase--thymidylate synthetase (*dhfr-ts*) ([Figure 1](#pbio-0000005-g001){ref-type="fig"}B and 1C). However, there is diversity in both the absolute magnitude of relative expression and in the timing of maximal expression (phase). In addition, a minority of genes, such as adenylosuccinate lyase (*asl*) ([Figure 1](#pbio-0000005-g001){ref-type="fig"}D), displayed a relatively constant expression profile. The accuracy of measurements from individual oligonucleotides was further verified by the ORFs that are represented by more than one oligonucleotide feature on the microarray. The calculated average pairwise Pearson correlation (*r*) is greater than 0.90 for 68% (0.75 for 86%) of the transcripts represented by multiple oligonucleotides with detectable expression during the IDC ([Table S1](#st001){ref-type="supplementary-material"}). Cases in which data from multiple oligonucleotides representing a single putative ORF disagree may represent incorrect annotation. The internal consistency of expression profile measurements for ORFs represented by more than one oligonucleotide sequence is graphically shown in [Figure 1](#pbio-0000005-g001){ref-type="fig"}E for the hypothetical protein MAL6P1.147, the largest predicted ORF in the genome (31 kb), which is represented by 14 oligonucleotide elements spanning the entire length of the coding sequence. The average pairwise correlation (*r*) for these features is 0.98±0.02. Periodicity in genome-wide gene expression datasets has been used to identify cell-cycle-regulated genes in both yeast and human cells ([@pbio-0000005-Spellman1]; [@pbio-0000005-Whitfield1]). Owing to the cyclical nature of the P. falciparum IDC dataset, a similar computational approach was taken. We performed simple Fourier analysis, which allowed us to calculate both the apparent phase and frequency of expression for each gene during the IDC (see [Materials and Methods](#s4){ref-type="sec"}). The fast Fourier transform (FFT) maps a function in a time domain (the expression profile) into a frequency domain such that when the mapped function is plotted (the power spectra), sharp peaks appear at frequencies where there is intrinsic periodicity. The calculated power spectra for each expression profile confirmed the observation that the data are highly periodic. The majority of profiles exhibited an overall expression period of 0.75--1.5 cycles per 48 h. We have used the FFT data for the purpose of filtering the expression profiles that are inherently noisy (i.e., that have low signal) or that lack differential expression throughout the IDC. Since the majority of the profiles display a single low-frequency peak in the power spectrum, we have taken advantage of this feature to classify profiles, similar to the application of a low-pass filter in signal processing. By measuring the power present in the peak frequency window (the main component plus two adjacent peaks) relative to the power present at all frequencies of the power spectrum, we were able to define a score (percent power) that we have used to stratify the dataset. The resulting distribution of expression profiles, scored in this way, is shown in [Figure 1](#pbio-0000005-g001){ref-type="fig"}F for all oligonucleotides. For reference, the positions of profiles corresponding to *eba175* (peak B), *dhfr-ts* (peak C), and *asl* (peak D) are indicated. It is striking that 79.5% of the expression profiles have a very high score (greater than 70%). For comparison, we applied our FFT analysis to the Saccharomyces cerevisiae cell cycle data, yielding only 194 profiles (3.8%) above a 70% score ([Figure S1](#sg001){ref-type="supplementary-material"}). In addition, we randomly permuted the columns of the complete dataset 1,000 times, each time recalculating the FFT, for a total of 5 million profiles (see inset in [Figure 1](#pbio-0000005-g001){ref-type="fig"}F). The randomized set exhibits essentially no periodicity: the probability of any random profile scoring above 70% is 1.3 × 10^−5^. P. falciparum Transcriptome Overview {#s2b} ------------------------------------ To provide an overview of the IDC transcriptome, we selected all 3,719 microarray elements whose profiles exhibited greater than 70% of the power in the maximum frequency window and that were also in the top 75% of the maximum frequency magnitudes. Although hierarchical clustering is extremely useful for comparing any set of expression data, regardless of the experimental variables, we sought to specifically address temporal order within the dataset. To accomplish this, the FFT phase was used to order the expression profiles to create a phaseogram of the IDC transcriptome of P. falciparum ([Figure 2](#pbio-0000005-g002){ref-type="fig"}A). The overview set represents 2,714 unique ORFs (3,395 oligonucleotides). An additional 324 oligonucleotides represent ORFs that are not currently part of the manually annotated collection. ::: {#pbio-0000005-g002 .fig} Figure 2 ::: {.caption} ###### Overview of the P. falciparum IDC Transcriptome \(A) A phaseogram of the IDC transcriptome was created by ordering the transcriptional profiles for 2,712 genes by phase of expression along the y-axis. The characteristic stages of intraerythrocytic parasite morphology are shown on the left, aligned with the corresponding phase of peak gene expression. (B--M) The temporal ordering of biochemical processes and functions is shown on the right. Each graph corresponds to the average expression profile for the genes in each set and the mean peak-to-trough amplitude is shown in parentheses. :::  ::: The IDC phaseogram depicts a cascade of continuous expression lacking clear boundaries or sharp transitions. During the first half of the IDC, a large number of genes involved in general eukaryotic cellular functions are induced with broad expression profiles. This gradual continuum includes the transition from the ring to the early trophozoite stage and the trophozoite to the early schizont stage, encompassing approximately 950 and 1,050 genes, respectively. Next, the mid- and late-schizont stages are marked by a rapid, large amplitude induction of approximately 550 genes, many of which appear to be continually expressed into the early-ring stage. However, owing to the level of synchrony in the culture, the ring-stage signal may be partially attributed to cross-contamination from residual schizonts. In the final hours of the IDC, approximately 300 genes corresponding to the early-ring stage are induced, indicating that reinvasion occurs without obvious interruptions to initiate the next cycle. The expression profiles for developmentally regulated genes in the P. falciparum IDC transcriptome reveal an orderly timing of key cellular functions. As indicated in [Figure 2](#pbio-0000005-g002){ref-type="fig"}B--2M, groups of functionally related genes share common expression profiles and demonstrate a programmed cascade of cellular processes that ensure the completion of the P. falciparum IDC. Ring and Early-Trophozoite Stage {#s2c} -------------------------------- In the following text, we have grouped the genes according to temporal expression phases based on their association with the common P. falciparum cytological stages. Following invasion, approximately 950 ORFs are induced during the ring and early trophozoite stage, including genes associated with the cytoplasmic transcriptional and translational machinery, glycolysis and ribonucleotide biosynthesis ([Figure 2](#pbio-0000005-g002){ref-type="fig"}B--2E). Represented in this group are 23 ORFs involved in transcription, including the four subunits of RNA polymerase I, nine subunits of RNA polymerase II, three subunits of RNA polymerase III, and four transcription factors. The average expression profile for this group is shown in [Figure 2](#pbio-0000005-g002){ref-type="fig"}B. (See [Table S2](#st002){ref-type="supplementary-material"} for all functional group details.) Also in this set are three previously identified P. falciparum RNA polymerase genes: the large subunits of P. falciparum RNA polymerase I ([@pbio-0000005-Fox1]) and RNA polymerase II ([@pbio-0000005-Li1]) and RNA polymerase III ([@pbio-0000005-Li2]). The cytoplasmic translation gene group ([Figure 2](#pbio-0000005-g002){ref-type="fig"}C) consists of 135 ORFs including homologues for 34 small and 40 large ribosomal subunits, 15 translation initiation factors, five translation elongation factors, 18 aminoacyl-tRNA synthetases, and 23 RNA helicases. In addition to the manually annotated ORFs, the translation gene group contains three ORFs predicted only by automated annotation including two ribosomal proteins (chr5.glm\_215, chr5.glm\_185) and a homologue of eIF-1A (chr11.glm\_489) ([PlasmoDB.org](PlasmoDB.org)). In one case, chr5.glm\_185 overlaps with the manually annotated ORF PFE0850w, which is found on the opposite strand. Oligonucleotide elements for both of these ORFs are present on the array. The oligonucleotide corresponding to the automated prediction yielded a robust FFT score and a phase consistent with the translation machinery, yet no PFE0850w expression was detected. These results suggest that the automated prediction for chr5.glm\_185 most likely represents the correct gene model for this genomic locus and illustrates the use of the IDC expression data for further verification of the P. falciparum genome annotations. Another set of 33 ORFs with homology to components of the translational machinery displayed an entirely distinct expression pattern, being induced during the late-trophozoite and early-schizont stage. This group includes 11 homologues of chloroplast ribosomal proteins, four mitochondrial/chloroplast elongation factors, and six amino acid tRNA synthetases ([Table S2](#st002){ref-type="supplementary-material"}). These ORFs also share a common pattern of expression, suggesting that these factors are components of the mitochondrial and/or the plastid translation machinery. This observation is supported by the presence of predicted apicoplast-targeting signals in 18 of these proteins ([PlasmoDB.org](PlasmoDB.org)). In addition, one of these factors, ribosomal protein S9, has been experimentally immunolocalized within the plastid ([@pbio-0000005-Waller1]). These data suggest that the peak of expression for the cytoplasmic translation machinery occurs in the first half of the IDC, whereas plastid and mitochondrial protein synthesis is synchronized with the maturation of these organelles during the second half of the IDC. In addition to transcription and translation, genes involved in several basic metabolic pathways were also induced during the ring and early-trophozoite stage, including glycolysis and ribonucleotide biosynthesis ([Figure 2](#pbio-0000005-g002){ref-type="fig"}D and 2E). Unlike the majority of P. falciparum biochemical processes, most of the enzymes involved in nucleotide metabolism and glycolysis have been identified ([@pbio-0000005-Reyes1]; [@pbio-0000005-Sherman1]). The glycolysis group ([Figure 2](#pbio-0000005-g002){ref-type="fig"}D) is tightly coregulated throughout the IDC and contains all of the 12 known enzymes. Expression initiates after reinvasion and continues to increase toward maximal expression during the trophozoite stage, when metabolism is at its peak. The glycolytic pathway is very well preserved in P. falciparum and exemplifies how data from this study can complement the homology-based interpretation of the genome. First, the genome contains two putative copies of pyruvate kinase on chromosomes 6 and 10, MAL6P1.160 and PF10\_0363, respectively ([@pbio-0000005-Gardner1]). However, only one of these genes, MAL6P1.160, has a similar expression profile to the other known glycolytic enzymes, suggesting that this enzyme is the main factor of this step in the glycolytic pathway. Interestingly, PF10\_0363 contains a putative apicoplast-targeting signal ([PlasmoDB.org](PlasmoDB.org)). In another case, the malaria genome sequencing consortium has predicted two homologues of triose phosphate isomerase, PF14\_0378 and PFC0381w. The latter is not detected by our analysis, suggesting that this gene is utilized in another developmental stage or may be a nonfunctional, redundant homologue. P. falciparum parasites generate pyrimidines through a de novo synthesis pathway while purines must be acquired by the organism through a salvage pathway ([@pbio-0000005-Gero1]). The mRNA levels of 16 enzymes corresponding to members of the pyrimidine ribonucleotide synthesis pathway, beginning with carbamoyl phosphate synthetase and ending with CTP synthetase, were uniformly induced immediately after invasion ([Figure 2](#pbio-0000005-g002){ref-type="fig"}E). The relative abundance of these transcripts peaked at approximately 18--22 hpi and then rapidly declined. Similar expression characteristics were detected for the enzymes of the purine salvage pathway, including the nucleoside conversion enzymes, hypoxanthine--guanine--xanthine phosphoribosyltransferase, and both guanylate and adenylate kinases ([Figure 2](#pbio-0000005-g002){ref-type="fig"}E; [Table S2](#st002){ref-type="supplementary-material"}). Trophozoite and Early-Schizont Stage {#s2d} ------------------------------------ The mRNA expression data indicate that ribonucleotide and deoxyribonucleotide production is clearly bifurcated into two distinct temporal classes. While ribonucleotide synthesis is required in the early stages of the IDC, deoxyribonucleotide metabolism is a trophozoite/early-schizont function. mRNA transcripts for enzymes that convert ribonucleotides into deoxyribonucleotides, including DHFR-TS and both subunits of ribonucleotide reductase, were induced approximately at 10 hpi, peaking at approximately 32 hpi ([Figure 2](#pbio-0000005-g002){ref-type="fig"}F). This represents a temporal shift from the induction of ribonucleotide synthesis of approximately 8--10 h. The expression of the deoxyribonucleotide biosynthesis is concomitant with the induction of DNA replication machinery transcripts, reflecting a tight relationship between DNA synthesis and production of precursors for this process. Thirty-two ORFs with homologies to various eukaryotic DNA replication machinery components are transcribed during the late-trophozoite and early-schizont stage. The timing of their transcription presages cell division. This functional gene group ([Figure 2](#pbio-0000005-g002){ref-type="fig"}G), with peak expression around 32 hpi, contains the previously characterized P. falciparum DNA Polα, DNA Polδ, and proliferating cell nuclear antigen, as well as the vast majority of the DNA replication components predicted by the malaria genome sequencing consortium ([@pbio-0000005-Gardner1]). These additional components include eight predicted DNA polymerase subunits, two putative origin recognition complex subunits, six minichromosome maintenance proteins, seven endo- and exonucleases, seven replication factor subunits, and two topoiosomerases. Interestingly, a number of proteins typically required for eukaryotic DNA replication, including the majority of the subunits of the origin recognition complex, have not yet been identified by conventional sequence similarity searches of the P. falciparum genome. All genes necessary for the completion of the tricarboxylic acid (TCA) cycle were detected in the *Plasmodium* genome ([@pbio-0000005-Gardner1]), although earlier studies indicate an unconventional function for this metabolic cycle. These studies suggest that the TCA cycle does not play a major role in the oxidation of glycolytic products. Instead, it is essential for the production of several metabolic intermediates, such as succinyl-CoA, a precursor of porphyrin biosynthesis ([@pbio-0000005-Sherman1]). The peak of expression for all TCA factors was detected during the late-trophozoite and early-schizont stage ([Figure 2](#pbio-0000005-g002){ref-type="fig"}H). Consistent with the model suggesting a disconnection of the TCA cycle from glycolysis during the IDC, no expression was detected for the subunits of the pyruvate dehydrogenase complex, including the α and β chains of pyruvate dehydrogenase and dihydrolipoamide S-acetyl transferase, the typical links between glycolysis and the TCA cycle. On the other hand, expression of TCA cycle genes is well synchronized with the expression of a large number of mitochondrial genes, including the three ORFs of the mitochondrial genome ([@pbio-0000005-Feagin1]), and several factors of electron transport ([Table S2](#st002){ref-type="supplementary-material"}). Although some of the TCA cycle proteins have been localized to the cytoplasm ([@pbio-0000005-Lang-Unnasch1]), the expression data suggest an association of this biochemical process with mitochondrial development and possibly with the abbreviated electron transport pathway detected in this organelle. Schizont Stage {#s2e} -------------- A transition from early to mid-schizont is marked by the maximal induction of 29 ORFs predicted to encode various subunits of the proteasome ([Figure 2](#pbio-0000005-g002){ref-type="fig"}I). Seven α and six β subunits of the 20S particle and 16 ORFs of the 19S regulatory particle were identified in this gene group. The common expression profile for the subunits of both of the 26S particle complexes suggests the involvement of ubiquitin-dependent protein degradation in the developmental progression of the parasite. The peak of proteasome expression coincides with a transition in the IDC transcriptome from metabolic and generic cellular machinery to specialized parasitic functions in the mid-schizont stage. This suggests an association between transcriptional regulation and protein turnover during this and possibly other transitions during the progression of the P. falciparum IDC. In the schizont stage, one of the first specialized processes induced was expression from the plastid genome ([Figure 2](#pbio-0000005-g002){ref-type="fig"}J). The essential extrachromosomal plastid (or apicoplast) genome contains 60 potentially expressed sequences, including ribosomal proteins, RNA polymerase subunits, ribosomal RNAs, tRNAs, and nine putative ORFs, including a ClpC homologue ([@pbio-0000005-Wilson2]). Very little is known about the regulation of gene expression in the plastid, but it is thought to be polycistronic ([@pbio-0000005-Wilson2]). In support of this observation, we find that 27 of the 41 plastid-specific elements present on our microarray displayed an identical expression pattern ([Figure 3](#pbio-0000005-g003){ref-type="fig"}C). The remaining elements correspond mainly to tRNAs and failed to detect appreciable signal. The highly coordinated expression of the plastid genome, whose gene products are maximally expressed in the late-schizont stage, is concomitant with the replicative stage of the plastid ([@pbio-0000005-Williamson1]). Note that not all plastid ORFs are represented on the microarray used in this study, and thus it is a formal possibility that the expression of the missing genes may differ from those shown in [Figure 3](#pbio-0000005-g003){ref-type="fig"}C. ::: {#pbio-0000005-g003 .fig} Figure 3 ::: {.caption} ###### Coregulation of Gene Expression along the Chromosomes of P. falciparum Is Rare, While Plastid Gene Expression Is Highly Coordinated Expression profiles for oligonucleotides are shown as a function of location for Chromosome 2 (\[A\], Oligo Map). With the exception of the SERA locus (B), coregulated clusters of adjacent ORFs are seldom observed, indicating that expression phase is largely independent of chromosomal position. (C) In contrast to the nuclear chromosomes, the polycistronic expression of the circular plastid genome is reflected in the tight coregulation of gene expression. This is an expanded view of the plastid-encoded genes from [Figure 2](#pbio-0000005-g002){ref-type="fig"}J. Genomic differences between strain 3D7, from which the complete genome was sequenced, and strain HB3 were measured by CGH. The relative hybridization between the gDNA derived from these two strains is shown as a percent reduction of the signal intensity for 3D7 (\[A\], CGH Data). Differences between the two strains are predominately located in the subtelomeric regions that contain the highly polymorphic *var, rifin,* and *stevor* gene families. Intrachromosomal variations, as observed for the *msp2* gene, were rare. :::  ::: Offset from the plastid by approximately 6 h, a set of approximately 500 ORFs exhibited peak expression during the late-schizont stage. Merozoite invasion of a new host cell is a complex process during which the parasite must recognize and dock onto the surface of the target erythrocyte, reorient with its apical tip toward the host cell, and internalize itself through invagination of the erythrocytic plasma membrane. The entire sequence of invasion events is facilitated by multiple receptor--ligand interactions with highly specialized plasmodial antigens ([@pbio-0000005-Cowman1]). The merozoite invasion group contains 58 ORFs, including 26 ORFs encoding antigens previously demonstrated to be important for the invasion process (see [Figure 2](#pbio-0000005-g002){ref-type="fig"}K). These include integral membrane proteins delivered to the merozoite surface from the micronemes (AMA1 and EBA175), GPI-anchored proteins of the merozoite membrane (MSP1, MSP4, and MSP5), proteins extrinsically associated with the merozoite surface during their maturation in the PV (MSP3 and MSP6), and soluble proteins secreted to the parasite--host cell interface (RAP1, RAP2, and RAP3). In addition, late-schizont-specific expression was observed for several antigens whose functions are not completely understood, but which have been associated with the invasion process. These ORFs include the merozoite-capping protein (MCP1), erythrocyte-binding-like protein 1 (EBL1), reticulocyte-binding proteins (RBP1 and RBP2), acid basic repeat antigen (ABRA), MSP7, and a homologue of the Plasmodium yoelii merozoite antigen 1. As expected, peak expression of these antigens coincides with the maturation of merozoites and development of several apical organelles, including rhoptries, micronemes, and dense granules. Many of these proteins have been considered as vaccine candidates since antibodies against these antigens were readily detected in the immune sera of both convalescent patients as well as individuals with naturally acquired immunity ([@pbio-0000005-Preiser1]). The sensitivity of invasion to protease and kinase inhibitors indicates an essential role for these activities in merozoite release as well as in the reinvasion process ([@pbio-0000005-Dluzewski1]; [@pbio-0000005-Blackman1]; [@pbio-0000005-Greenbaum1]). The merozoite invasion gene group contains three serine proteases, including PfSUB1, PfSUB2, and an additional homologue to plasmodial subtilases (PFE0355c), and two aspartyl proteases, plasmepsin (PM) IX and X. Peak expression during the mid-schizont stage was also observed for seven members of the serine repeat antigen (SERA) family, all of which contain putative cysteine protease domains. In addition to the proteases, expression of 12 serine/threonine protein kinases and three phophorylases was tightly synchronized with the genes of the invasion pathway, including six homologues of protein kinase C, three Ca^+^-dependent and two cAMP-dependent kinases, phosphatases 2A and 2B, and protein phosphatase J. Another functionally related gene group whose expression is sharply induced during the late-schizont stage includes components of actin--myosin motors (see [Figure 2](#pbio-0000005-g002){ref-type="fig"}L) ([@pbio-0000005-Pinder1]). As in other apicomplexa, actin and myosin have been implicated in host cell invasion ([@pbio-0000005-Opitz1]). Schizont-specific expression was observed for three previously described class XIV myosin genes, one associated light chain, two actin homologues, and three additional actin cytoskeletal proteins, including actin-depolymerizing factor/cofilin (two isoforms) and coronin (one isoform). Although the molecular details of plasmodial actin--myosin invasion are not completely understood, the tight transcriptional coregulation of the identified factors indicates that the examination of schizont-specific expression may help to identify additional, possibly unique elements of this pathway. Early-Ring Stage {#s2f} ---------------- The expression data are continuous throughout the invasion process, with no observable abrupt change in the expression program upon successful reinvasion. However, a set of approximately 300 ORFs whose expression is initiated in the late-schizont stage persists throughout the invasion process and peaks during the early-ring stages (see [Figure 2](#pbio-0000005-g002){ref-type="fig"}M). It was previously determined that immediately after invasion, a second round of exocytosis is triggered, ensuring successful establishment of the parasite within the host cell ([@pbio-0000005-Foley1]). One of the main P. falciparum virulence factors associated with this process is ring-infected surface antigen 1 (RESA1). RESA1 is secreted into the host cell cytoplasm at the final stages of the invasion process, where it binds to erythrocytic spectrin, possibly via its DnaJ-like chaperone domain ([@pbio-0000005-Foley1]). The early stages of the IDC contain a variety of putative molecular chaperones in addition to RESA1, including RESA2 and RESAH3, plus five additional proteins carrying DnaJ-like domains. However, the functional roles of these chaperones remain unclear. Despite the cytoplasmic role of RESA1, abundant antibodies specific for RESA1 are present in individuals infected with P. falciparum, indicating that RESA1 is also presented to the host immune system ([@pbio-0000005-Troye-Blomberg1]). Several genes encoding additional antigenic factors are found among the early ring gene group, including frequently interspersed repeat antigen (FIRA), octapeptide antigen, MSP8, and sporozoite threonine- and asparagine-rich protein (STARP). Like RESA1, antibodies against these antigens are also found in the sera of infected individuals, suggesting that the final stages of invasion might be a target of the immune response. Overall, the genes expressed during the mid- to late-schizont and early-ring stage encode proteins predominantly involved in highly parasite-specific functions facilitating various steps of host cell invasion. The expression profiles of these genes are unique in the IDC because of the large amplitudes and narrow peak widths observed. The sharp induction of a number of parasite-specific functions implies that they are crucial for parasite survival in the mammalian host and hence should serve as excellent targets for both chemotherapeutic and vaccine-based antimalarial strategies. IDC Transcriptional Regulation and Chromosomal Structure {#s2g} -------------------------------------------------------- Transcriptional regulation of chromosomal gene expression in P. falciparum is thought to be monocistronic, with transcriptional control of gene expression occurring through regulatory sequence elements upstream and downstream of the coding sequence ([@pbio-0000005-Horrocks1]). This is in contrast to several other parasites, such as Leishmania sp., in which polycistronic mRNA is synthesized from large arrays of coding sequences positioned unidirectionally along the arms of relatively short chromosomes ([@pbio-0000005-Myler1]). Recent proteomic analyses failed to detect any continuous chromosomal regions with common stage-specific gene expression in several stages of the P. falciparum lifecycle ([@pbio-0000005-Florens1]). However, transcriptional domains have previously been suggested for Chromosome 2 ([@pbio-0000005-Le1]). The availability of the complete P. falciparum genome coupled with the IDC transcriptome allows us to investigate the possibility of chromosomal clustering of gene expression (see [Figure 3](#pbio-0000005-g003){ref-type="fig"}A). To systematically explore the possibility of coregulated expression as a function of chromosomal location, we applied a Pearson correlation to identify similarities in expression profiles among adjacent ORFs. The pairwise Pearson correlation was calculated for every ORF pair within each chromosome ([Figure S2](#sg002){ref-type="supplementary-material"}). Gene groups in which the correlation of 70% of the possible pairs was greater than *r* = 0.75 were classified as putative transcriptionally coregulated groups. Using these criteria, we identified only 14 coregulation groups consisting of greater than three genes, with the total number of genes being 60 (1.4% of all represented genes) ([Table S3](#st003){ref-type="supplementary-material"}). In eight of the 14 groups, the coregulation of a pair of genes may be explained by the fact that they are divergently transcribed from the same promoter. A set of 1,000 randomized permutations of the dataset yielded 2.25 gene groups. Contrary to the nuclear chromosomes, there was a high correlation of gene expression along the plastid DNA element, consistent with polycistronic transcription (see [Figure 3](#pbio-0000005-g003){ref-type="fig"}C). The average pairwise Pearson correlation for a sliding window of seven ORFs along the plastid genome is 0.92±0.03. The largest group demonstrating coregulation on the nuclear chromosomes corresponds to seven genes of the SERA family found on Chromosome 2 (see [Figure 3](#pbio-0000005-g003){ref-type="fig"}B) ([@pbio-0000005-Miller1]). Besides the SERA gene cluster and a group containing three ribosomal protein genes, no additional functional relationship was found among the other chromosomally adjacent, transcriptionally coregulated gene groups. The limited grouping of regional chromosomal expression was independent of strand specificity and, with the exception of the SERA group, did not overlap with the groups of "recently duplicated genes" proposed by the malaria genome sequencing consortium ([@pbio-0000005-Gardner1]). Three major surface antigens, the *var*, *rifin*, and *stevor* families, have a high degree of genomic variability and are highly polymorphic between strains and even within a single strain ([@pbio-0000005-Cheng1]; [@pbio-0000005-Afonso1]; [@pbio-0000005-Gardner1]). Expression profiles for only a small subset of these genes were detected in the IDC transcriptome and were typically characterized by low-amplitude profiles. This could be due to two nonmutually exclusive possibilities: first, the HB3 DNA sequence for these genes may be substantially rearranged or completely deleted relative to the reference strain, 3D7; second, only a few of these genes may be selectively expressed, as has been proposed ([@pbio-0000005-Deitsch1]). To identify regions of genomic variability between 3D7 and HB3, we performed microarray-based comparative genomic hybridization (CGH) analysis. Array-based CGH has been performed with human cDNA and bacterial artificial chromosome-based microarrays to characterize DNA copy-number changes associated with tumorigenesis ([@pbio-0000005-Gray1]; [@pbio-0000005-Pollack2]). Using a similar protocol, CGH analysis revealed that the majority of genetic variation between HB3 and 3D7 is confined to the subtelomeric chromosomal regions containing the aforementioned gene families ([Figure 3](#pbio-0000005-g003){ref-type="fig"}A; [Figure S3](#sg003){ref-type="supplementary-material"}). Only 28.3% of *rifin*, 47.1% of *var*, and 51.0% of *stevor* genes predicted for the 3D7 strain were detected for the HB3 genomic DNA (gDNA) when hybridized to the 3D7-based microarray. Thus, the underrepresentation of these gene families in the HB3 IDC transcriptome is likely due to the high degree of sequence variation present in these genes. Excluding the three surface antigen families in the subtelomeric regions, 97% of the remaining oligonucleotide microarray elements exhibit an equivalent signal in the CGH analysis. However, 144 of the differences detected by CGH reside in internal chromosomal regions and include several previously identified plasmodial antigens: MSP1, MSP2 ([Figure 3](#pbio-0000005-g003){ref-type="fig"}A), S antigen, EBL1, cytoadherence-linked asexual gene 3.1 (CLAG3.1), glutamine-rich protein (GLURP), erythrocyte membrane protein 3 (PfEMP3), knob-associated histidine-rich protein (KAHRP), and gametocyte-specific antigen Pfg377 ([Table S4](#st004){ref-type="supplementary-material"}). These results demonstrate a high degree of genetic variation within the genes considered to be crucial for antigenic variation between these two commonly used laboratory strains of P. falciparum. Implications for Drug Discovery {#s2h} ------------------------------- The majority of the nuclear-encoded proteins targeted to the plastid are of prokaryotic origin, making them excellent drug targets ([@pbio-0000005-McFadden1]). Moreover, inhibitors of plastid-associated isoprenoid biosynthesis, DNA replication, and translation have been shown to kill the P. falciparum parasite, demonstrating that the plastid is an essential organelle ([@pbio-0000005-Fichera1]; [@pbio-0000005-Jomaa1]). The plastid has been implicated in various metabolic functions, including fatty acid metabolism, heme biosynthesis, isoprenoid biosynthesis, and iron--sulfur cluster formation ([@pbio-0000005-Wilson1]). It is clear that, within the plastid, functional ribosomes are assembled to express the ORFs encoded by the plastid genome ([@pbio-0000005-Roy1]). However, nuclear-encoded components are required to complete the translational machinery as well as for all other plastid metabolic functions. A bipartite signal sequence is required for efficient transport of these nuclear proteins from the cytoplasm to the plastid via the endoplasmic reticulum ([@pbio-0000005-Waller2]). Computational predictions suggest that the P. falciparum genome may contain over 550 nuclear-encoded proteins with putative transit peptides ([@pbio-0000005-Zuegge1]; [@pbio-0000005-Foth1]). Given that over 10% of the ORFs in the P. falciparum genome are predicted to contain an apicoplast-targeting sequence, we sought to use the IDC transcriptome as a means to narrow the search space for candidate apicoplast-targeted genes. As mentioned above, the expression profiles for genes encoded on the plastid genome are tightly coordinated (see [Figure 3](#pbio-0000005-g003){ref-type="fig"}C). We reasoned that genes targeted to the plastid would be expressed slightly before or coincidentally with the plastid genome. Therefore, we utilized the FFT phase information to identify ORFs in phase with expression of the plastid genome (see [Materials and Methods](#s4){ref-type="sec"}) ([Table S5](#st005){ref-type="supplementary-material"}). Because the genes of the plastid genome are maximally expressed between 33 and 36 hpi, we searched for all genes in the dataset with an FFT phase in this time window and then cross-referenced the list of predicted apicoplast-targeted sequences ([PlasmoDB.org](PlasmoDB.org)), resulting in a list of 124 in-phase apicoplast genes ([Figure 4](#pbio-0000005-g004){ref-type="fig"}A). Within this list are two ORFs that have been directly visualized in the apicoplast, acyl carrier protein and the ribosomal subunit S9 ([@pbio-0000005-Waller1]), as well as many ORFs associated with the putative plastid ribosomal machinery, enzymes involved in the nonmevalonate pathway, additional caseineolytic proteases (Clps), the reductant ferredoxin, and replication/transcriptional machinery components. However, this list contains only 14 of the 43 proteins categorized in the Gene Ontology (GO) assignments at [PlasmoDB.org](PlasmoDB.org) as apicoplast proteins by inference from direct assay (IDA). In addition, 30% of the nuclear-encoded translational genes that are not coexpressed with the known cytoplasmic machinery are found within this small group of genes. More importantly, 76 ORFs (62%) are of unknown function, with little or no homology to other genes. This limited subgroup of putative plastid-targeted ORFs are likely excellent candidates for further studies in the ongoing search for malaria-specific functions as putative drug targets. ::: {#pbio-0000005-g004 .fig} Figure 4 ::: {.caption} ###### Temporal Distribution of the Apicoplast-Targeted Proteins and P. falciparum Proteases, Potential Antimalarial Drug Candidates \(A) The expression profiles of all putative plastid-targeted genes represented on our microarray are shown. The yellow box encompasses a highly synchronized group of genes, which are in-phase with plastid genome expression. The average expression profile for this in-phase group of genes is shown and includes most of the known apicoplast-targeted genes as well as many hypothetical genes. For reference, the average expression profile for the plastid genome is shown (dashed gray line). \(B) Proteases represent an attractive target for chemotherapeutic development. The broad range of temporal expression for various classes of proteases and their putative functions are displayed. Abbreviations: HAP, histo-aspartyl protease (PM III); Clp, caseineolytic protease; sub1, 2, subtilisin-like protease 1 and 2. :::  ::: Similarly, P. falciparum proteases have received much attention, since they are candidates as drug targets and have been shown to play important roles in regulation as well as metabolism throughout the IDC ([@pbio-0000005-Rosenthal1]). A temporal ordering of expression profiles for several well-characterized P. falciparum proteases is shown in [Figure 4](#pbio-0000005-g004){ref-type="fig"}B, demonstrating the broad significance of these enzymes throughout the IDC. One of the principal proteolytic functions is considered to be the degradation of host cell hemoglobin in the food vacuole (FV) to produce amino acids essential for protein synthesis. This elaborate process is carried out by a series of aspartyl proteases, cysteine proteases, metalloproteases, and aminopeptidases ([@pbio-0000005-Francis1]). A family of ten aspartyl proteases, the plasmepsins (PMs), has been identified in the P. falciparum genome, four of which have been characterized as bona fide hemoglobinases: PM I, II, III (a histo-aspartic protease \[HAP\]), and IV ([@pbio-0000005-Coombs1]). Our data reveal that the PMs are expressed at different times throughout the lifecycle, suggesting that they are involved in different processes throughout the IDC. PM I, II, HAP, and PM IV are adjacent to one another on Chromosome 14 and have been localized to the FV. While HAP and PM II are expressed in the mid-trophozoite stage, during peak hemoglobin catabolism, PMI and IV are maximally expressed in the ring stage along with the cysteine protease falcipain-1 (FP-1). FP-1 has recently been implicated in merozoite invasion and has been localized to the interior of the PV ([@pbio-0000005-Greenbaum1]). The coincident expression of these proteases implies that the development of the PV and the FV occurs during the very early-ring stage. This observation is corroborated by similar expression profiles for the PV-associated protein RESA1 and the FV protein PGH1. Subsequently, a second group of hemoglobinases, including the m1-family aminopeptidase, FP-2, and falcilysin, is expressed simultaneously with HAP and PM II during the trophozoite stage of the IDC. The expression of PM V and the newly identified FP- 2 homologue during this stage suggests they are also important in the trophozoite stage. The other known falcipain, FP-3, does not show a marked induction in expression throughout the IDC. We fail to detect any transcripts for PM VI, VII, and VIII during the IDC. These genes may have roles in any of the other sexual, liver, or mosquito stages of development. In addition to the hemoglobinases, P. falciparum contains a variety of proteases involved in cellular processing, including a group of Clps and signal peptidases that are all expressed maximally at the late-trophozoite stage ([Figure 4](#pbio-0000005-g004){ref-type="fig"}B). The timing of these genes may play a key role in protein maturation during trafficking to various compartments, including the plastid. The three Clps contain putative leader peptides and may actually function within the plastid. Finally, a group of proteases are expressed in the schizont stage and include the P. falciparum subtilisin-like proteases PfSUB1 and PfSUB2 as well as PMs IX and X. PfSUB1 and PfSUB2 are believed to be involved in merozoite invasion and have been localized apically in the dense granules. Interestingly, there are two PfSUB1 protease homologues (PFE0355c and PFE0370c); PM X parallels the expression of PfSUB1 (PFE0370c), suggesting that aspartyl proteases may also be involved in merozoite invasion. In addition, the phase of the PfSUB1 homologue suggests a concomitant role, with PM IX slightly preceding merozoite invasion. In total, we have detected gene expression for over 80 putative proteases throughout the entire IDC ([Table S6](#st006){ref-type="supplementary-material"}). This set includes over 65 proteases from a group of recently predicted proteases ([@pbio-0000005-Wu1]). The differing temporal expression of these proteases may allow for a multifaceted approach toward identifying protease inhibitors with efficacy at all stages of the IDC. Implications for New Vaccine Therapies {#s2i} -------------------------------------- Merozoite invasion is one of the most promising target areas for antimalarial vaccine development ([@pbio-0000005-Good1]). Many vaccine efforts thus far have focused primarily on a set of plasmodial antigens that facilitate receptor--ligand interaction between the parasite and the host cell during the invasion process ([@pbio-0000005-Preiser1]) (see [Figure 2](#pbio-0000005-g002){ref-type="fig"}K and 2M). Merozoite invasion antigens are contributing factors to naturally acquired immunity, triggering both humoral and antibody-independent cell-mediated responses ([@pbio-0000005-Good2]). Antibodies against these antigens have been demonstrated to effectively block the merozoite invasion process in vitro and in animal models ([@pbio-0000005-Ramasamy1]). Owing to the highly unique character of merozoite surface antigens, homology-based searches have yielded only a limited set of additional invasion factors. We utilized the IDC transcriptome to predict a set of likely invasion proteins by identifying expression profiles with characteristics similar to previously studied merozoite invasion proteins. The expression profiles for all known invasion factors undergo a sharp induction during the mid- to late-schizont stage and are characterized by large expression amplitudes (see [Figure 2](#pbio-0000005-g002){ref-type="fig"}A). Among these proteins are seven of the best-known malaria vaccine candidates, including AMA1, MSP1, MSP3, MSP5, EBA175, RAP1, and RESA1. To identify ORFs with a possible involvement in the merozoite invasion process, we have calculated the similarity, by Euclidian distance, between the expression profiles of these seven vaccine candidates and the rest of the IDC transcriptome. A histogram of the distance values reveals a bimodal distribution with 262 ORFs in the first peak of the distribution ([Figure S4](#sg018){ref-type="supplementary-material"}). This represents the top 5% of expression profiles when ranked by increasing Euclidian distance ([Table S7](#st007){ref-type="supplementary-material"}). In addition to the seven vaccine candidate genes used for the search, essentially all predicted P. falciparum merozoite-associated antigens were identified in this gene set ([Figure 5](#pbio-0000005-g005){ref-type="fig"}). These include the GPI-anchored MSP4; several integral merozoite membrane proteins, such as EBA140 and EBL1; three RBPs (RBP1, RBP2a, RBP2b); and a previously unknown RBP homologue. In addition, components of two proteins secreted from the rhoptries to the host cell membranes, RhopH1 and RhopH3, or to the PVs RAP1, RAP2, and RAP3 were found in the selected set. Surprisingly, CLAG2 and CLAG9 were also classified into the merozoite invasion group. Although the biological function of these genes is believed to be associated with cytoadherence of the infected erythrocyte to the vascular endothelium, a highly related homologue, CLAG3.1 (RhopH1), was recently detected in the rhoptries, suggesting a possible secondary role for these genes in merozoites ([@pbio-0000005-Kaneko1]). ::: {#pbio-0000005-g005 .fig} Figure 5 ::: {.caption} ###### Phaseogram of Putative Vaccine Targets The similarity of all expression profiles to seven known vaccine candidates (boxed) was calculated. The top 5% of similar profiles correspond to 262 ORFs, 28 of which have been previously associated with plasmodial antigenicity and the process of merozoite invasion. :::  ::: A number of antigens are presently in various stages of clinical trials and are yielding encouraging results ([@pbio-0000005-Good3]). However, many single-antigen vaccine studies indicate that the most promising approach will require a combination of antigenic determinants from multiple stages of the complex plasmodial lifecycle ([@pbio-0000005-Kumar1]). Searches for new target antigens in the P. falciparum genome are thus vital to the development of future vaccines, since no fully protective vaccine has been assembled thus far. Of the 262 ORFs whose expression profiles were closest to the profiles of the seven major vaccine candidates, 189 are of unknown function. These ORFs represent a candidate list for new vaccine targets. Discussion {#s3} ========== The transcriptome of the IDC of P. falciparum constitutes an essential tool and baseline foundation for the analysis of all future gene expression studies in this organism, including response to drugs, growth conditions, environmental perturbations, and genetic alterations. Essentially all experiments involving asexual intraerythrocytic-stage parasites must be interpreted within the context of the ongoing cascade of IDC-regulated genes. In our global analysis of the P. falciparum transcriptome, over 80% of the ORFs revealed changes in transcript abundance during the maturation of the parasite within RBCs. The P. falciparum IDC significantly differs from the cell cycles of the yeast S. cerevisiae ([@pbio-0000005-Spellman1]) and human HeLa ([@pbio-0000005-Whitfield1]) cells, during which only 15% of the total genome is periodically regulated. Instead, the P. falciparum IDC resembles the transcriptome of the early stages of Drosophila melanogaster development, which incorporates the expression of over 80% of its genome as well ([@pbio-0000005-Arbeitman1]). Unlike the development of multicellular eukaryotes, there is no terminal differentiation and, with the exception of gametocytogenesis, the parasite is locked into a repeating cycle. In this respect, the P. falciparum IDC mirrors a viral-like lifecycle, in which a relatively rigid program of transcriptional regulation governs the progress of the course of infection. The lack of continuous chromosomal domains with common expression characteristics suggests that the genes are regulated individually, presumably via distinct sets of *cis*- and *trans*-acting elements. However, the extent and the simple mechanical character of transcriptional control observed in the IDC suggest a fundamentally different mode of regulation than what has been observed in other eukaryotes. It is plausible that a comparatively small number of transcription factors with overlapping binding site specificities could account for the entire cascade. While further experiments are ongoing, it may be the case that P. falciparum gene regulation is streamlined to the extent that it has lost the degree of dynamic flexibility observed in other unicellular organisms, from Escherichia coli to yeast. This observation also implies that disruption of a key transcriptional regulator, as opposed to a metabolic process, may have profound inhibitory properties. While a few putative transcription factors have been identified in the P. falciparum genome, no specific regulatory elements have been defined in basepair-level detail. A further analysis of the upstream regions of genes with similar phases should facilitate the elucidation of regulatory regions and their corresponding regulatory proteins. In general, the timing of mRNA expression for a given gene during the IDC correlates well with the function of the resultant protein. For example, replication of the genome occurs in the early-schizont stage and correlates well with the peak expression of all factors of DNA replication and DNA synthesis. Also, organellar biogenesis of several intracellular compartments such as mitochondria, the plastid, or the apical invasion organelles is concomitant with the maximal induction of mRNAs encoding proteins specific to these organelles. In addition, our data are generally in good agreement with proteomic analyses that have detected intraerythrocytic-stage proteins from the merozoite, trophozoite, and schizont stages. More than 85% of the 1,588 proteins detected in these studies were also expressed in our analysis ([@pbio-0000005-Florens1]; [@pbio-0000005-Lasonder1]). However, a more detailed proteomic analysis at different stages of the IDC will be needed to ascertain the temporal changes of these proteins. We initially expected that a high percentage of the genome would be specialized for each lifecycle stage (mosquito, liver, blood), yet this was not observed; the mRNA transcripts for 75% of proteins determined to be gamete-, gametocyte-, or sporozoite-specific by mass spectrometry are also transcribed in the plasmodial IDC. These findings confirm previous studies demonstrating that not only genes used for generic cellular processes are present in multiple developmental stages, but also factors of highly specialized *Plasmodium* functions ([@pbio-0000005-Gruner1]). This may indicate that only a small portion of the genome may actually be truly specific to a particular developmental stage and that the majority of the genome is utilized throughout the full lifecycle of this parasite. It is also feasible to speculate that a multilayer regulatory network is employed in the progression of the entire P. falciparum lifecycle. In this model, the same *cis*- and *trans*-acting regulatory elements driving the actual mRNA production in IDC are utilized in other developmental stages. These elements are then controlled by an alternate subset of factors determining the status of the lifecycle progression. These findings also outline two contrasting properties of the P. falciparum genome. The *Plasmodium* parasite devotes 3.9% of its genome to a complex system of antigenic determinants essential for host immune evasion during a single developmental stage ([@pbio-0000005-Gardner1]). On the other hand, large portions of the genome encode proteins used in multiple stages of the entire lifecycle. Such broad-scope proteins might be excellent targets for both vaccine and chemotherapeutic antimalarial strategies, since they would target several developmental stages simultaneously. While there are certainly proteins specific to these nonerythrocytic stages, a complementary analysis of both proteomic and genomic datasets will facilitate the search. With malaria continuing to be a major worldwide disease, advances toward understanding the basic biology of P. falciparum remain essential. Our analysis of the IDC transcriptome provides a first step toward a comprehensive functional analysis of the genome of P. falciparum. The genome-wide transcriptome will be useful not only for the further annotation of many uncharacterized genes, but also for defining the biological processes utilized by this highly specialized parasitic organism. Importantly, candidate groups of genes can be identified that are both functionally and transcriptionally related and thus provide focused starting points for the further elucidation of genetic and mechanistic aspects of P. falciparum. Such biological characterizations are presently a major objective in the search for novel antimalarial strategies. The public availability of the dataset presented in this study is intended to provide a resource for the entire research community to extend the exploration of P. falciparum beyond the scope of this publication. All data will be freely accessible at two sites: <http://plasmodb.org> and <http://malaria.ucsf.edu>. Materials and Methods {#s4} ===================== {#s4a} ### Cell culture. {#s4a1} A large-scale culture of P. falciparum (HB3 strain) was grown in a standard 4.5 l microbial bioreactor (Aplikon, Brauwweg, Netherlands) equipped with a Bio Controller unit ADI 1030 (Aplikon, Brauwweg, Netherlands). Cells were initially grown in a 2% suspension of purified human RBCs and RPMI 1640 media supplemented with 0.25% Albumax II (GIBCO, Life Technologies, San Diego, California, United States), 2 g/l sodium bicarbonate, 0.1 mM hypoxanthine, 25 mM HEPES (pH 7.4), and 50 μg/l gentamycin, at 37°C, 5% O~2~, and 6% CO~2~. Cells were synchronized by two consecutive sorbitol treatments for three generations, for a total of six treatments. Large-scale cultures contained 32.5 mM HEPES (pH 7.4). The bioreactor culture was initiated by mixing 25.0 ml of parasitized RBCs (20% late schizonts, approximately 45 hpi) with an additional 115.0 ml of purified RBC in a total of 1.0 l of media (14% hematocrit). Invasion of fresh RBCs occurred during the next 2 h, raising the total parasitemia from an initial 5% to 16%. After this period, the volume of the culture was adjusted to 4.5 l, bringing the final RBC concentration to approximately 3.3% to reduce the invasion of remaining cells. Immediately after the invasion period, greater than 80% of the parasites were in the ring stage. Temperature and gas conditions were managed by the Bio Controller unit. Over the course of 48 h, 3--4 ml of parasitized RBCs was collected every hour, washed with prewarmed PBS, and flash-frozen in liquid nitrogen. ### RNA preparation and reference pool. {#s4a2} P. falciparum RNA sample isolation, cDNA synthesis, labeling, and DNA microarray hybridizations were performed as described by [@pbio-0000005-Bozdech1]). Samples for individual timepoints (coupled to Cy5) were hybridized against a reference pool (coupled to Cy3). The reference pool was comprised of RNA samples representing all developmental stages of the parasite. From this pool, sufficient cDNA synthesis reactions, using 12 μg of pooled reference RNA, were performed for all hybridizations. After completing cDNA synthesis, all reference pool cDNAs were combined into one large pool and then split into individual aliquots for subsequent labeling and hybridization. Microarray hybridizations were incubated for 14--18 h. ### DNA microarray hybridizations and quality control. {#s4a3} In total, 55 DNA microarray hybridizations covering 46 timepoints were performed. Timepoints 1, 7, 11, 14, 18, 20, 27, and 31 were represented by more than one array hybridization. Data were acquired and analyzed by GenePix Pro 3 (Axon Instruments, Union City, California, United States). Array data were stored and normalized using the NOMAD microarray database system (<http://ucsf-nomad.sourceforge.net/>). In brief, a scalar normalization factor was calculated for each array using unflagged features with median intensities greater than zero for each channel and a pixel regression correlation coefficient greater than or equal to 0.75. Quality spots were retained based on the following criteria. The log~2~(Cy5/Cy3) ratio for array features that were unflagged and had a sum of median intensities greater than the local background plus two times the standard deviation of the background were extracted from the database for further analysis. Subsequently, expression profiles consisting of 43 of 46 timepoints (approximately 95%) were selected. For those timepoints that were represented by multiple arrays, the ratio values were averaged. ### FFT analysis of the expression profiles. {#s4a4} Fourier analysis was performed on each profile in the quality-controlled set (5,081 oligonucleotides). Profiles were smoothed with missing values imputed using a locally weighted regression algorithm with local weighting restricted to 12% using R (<http://www.R-project.org>). Fourier analysis was performed on each profile using the *fft()* function of R, padded with zeros to 64 measurements. The power spectrum was calculated using the *spectrum()* function of R. The power at each frequency (*Power()*), the total power (P~tot~), and the frequency of maximum power (F~max~) were determined. The periodicity score was defined as *Power*\[(F~max−1~) + (F~max~) + (F~max+1~)\]/P~tot~. The most frequent value of F~max~ across all profiles was deemed the major frequency (m) and used in determining phase information. The phase of each profile was calculated as atan2\\\[−(I (m)\],R (m)\\, where atan2 is R\'s arctangent function and I and R are the imaginary and real parts of the FFT. Profiles were then ordered in increasing phase from −π to π. The loess smooth profiles were drawn through the raw expression data using the *loess()* function found in the modern regression library of R (version 1.5.1). The default parameters were used, with the exception that local weighting was reduced to 30%. For the averaged profiles of the functional groups (see [Figure 2](#pbio-0000005-g002){ref-type="fig"}B--2M), the loess smooth profiles were calculated for each expression profile individually and subsequently averaged to create the representative profile. These same methods were applied to both the randomized set (see the inset to [Figure 1](#pbio-0000005-g001){ref-type="fig"}F) and the yeast cell cycle dataset (see [Figure S1](#sg001){ref-type="supplementary-material"}). The raw results files ([Dataset S1](#sd001){ref-type="supplementary-material"}), the fully assembled raw dataset ([Dataset S2](#sd002){ref-type="supplementary-material"}, the overview dataset ([Dataset S3](#sd003){ref-type="supplementary-material"}, and the quality control dataset ([Dataset S4](#sd004){ref-type="supplementary-material"}) are available as downloads. ### Evaluation of coexpression along chromosomes. {#s4a5} The evaluation of coexpression of genes along chromosomes was carried out as follows. The Pearson correlation coefficient was calculated for each pair of profiles. For ORFs with multiple oligonucleotides, the average profile was calculated. The neighborhood of each ORF profile was defined as a window of between one and ten adjacent ORF profiles. If any window in an ORF profile\'s neighborhood displayed more than 70% pairwise correlation of greater than 0.75, it was flagged as enriched. The length of the window was then recorded as a region of coexpression. This process was repeated without strand separation of ORFs and with randomly permuted datasets. ### Comparative genomic hybridization. {#s4a6} P. falciparum strains 3D7 and HB3 were cultured as previously described at a concentration of 10% parasitaemia. Genomic DNA (gDNA) was isolated from a minimum of 500 ml of total culture for each P. falciparum strain, as previously described ([@pbio-0000005-Wang1]). Isolated gDNA from each strain was sheared by sonication to an average fragment size of approximately 1--1.5 kb and then was purified and concentrated using a DNA Clean and Concentrator kit (Zymo Research, Orange, California, United States). Amino-allyl-dUTP first was incorporated into the gDNA fragments with a Klenow reaction at 37°C for 6--8 h with random nonamer primers and 3 μg of sheared gDNA. After purification and concentration of the DNA from the Klenow reaction, CyScribe Cy3 and Cy5 dyes (Amersham Biosciences, Buckinghamshire, United Kingdom) were coupled to HB3 DNA and 3D7 DNA, respectively, as previously described ([@pbio-0000005-Pollack1]). Uncoupled fluorescent dye was removed using a DNA Clean and Concentrator kit. Labeled DNA fragments were hybridized to the oligonucleotide-based DNA microarrays. Fluorescence was detected and analyzed using an Axon Instruments scanner and GenePix Pro 3.0 software. Only features that had median intensities greater than the local background plus two times the standard deviation of the background in each channel were considered for further analysis. For each feature, the percent of the total intensity was determined using the signal in the 3D7 channel as the total amount of intensity for each oligonucleotide; intensity differences less than 50% were considered to be significant for subsequence analysis. ### Calculation for in-phase plastid-targeted genes. {#s4a7} The range of FFT-based phases for the expression profiles of the plastid genome is between 0.32 and 1.05 (or roughly π/9 −π/3). Using the list of 551 apicoplast-targeted genes available at [PlasmoDB.org](PlasmoDB.org), we first ordered these genes by phase and then grouped all genes with a phase range between 0.00 and 1.40 (0--4π/9), resulting in 124 genes represented by 128 oligonucleotides on the microarray. This select group represents the in-phase plastid targeted genes (see [Table S6](#st006){ref-type="supplementary-material"}). ### Calculation for vaccine targets. {#s4a8} To select the expression profiles most related to the AMA1, MSP1, MSP3, MSP5, EBA175, RAP1, and RESA1 vaccine candidates, we calculated the similarity of all expression profiles in the dataset to those of these antigens by Euclidian distance. The minimum Euclidian distance calculated for every profile was then binned into 60 bins and plotted as a histogram. A natural break in the histogram was seen that included the set of 262 ORFs (see [Figure S2](#sg002){ref-type="supplementary-material"}). Supporting Information {#s5} ====================== Dataset S1 ::: {.caption} ###### Raw GenePix Results (29.5 MB ZIP). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S2 ::: {.caption} ###### Complete Dataset (3.7 MB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S3 ::: {.caption} ###### Overview Dataset (2.4 MB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S4 ::: {.caption} ###### Quality Control Set (3.1 MB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S1 ::: {.caption} ###### Histogram of the Percent Power at Peak Frequencies for the Yeast Cell Cycle Data The percent of power in the maximum frequency of the FFT power spectrum was used to determine periodicity of the yeast cell cycle data from [@pbio-0000005-Spellman1]). The histogram reveals periodic regulation of gene expression for only a small subset of genes (% power \>70%). (223 KB EPS). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S2 ::: {.caption} ###### Pearson Correlation Maps for the P. falciparum Chromosomes A matrix of the pairwise Pearson correlations was calculated for every expression profile along the chromosomes. The analysis included all annotated ORFs. The gray areas correspond to a Pearson correlation d(x, y) = 0 and indicate ORFs with no detectable IDC expression or ORFs not represented on the microarray. The starting point (left) and the end point (right) of the chromosomes and the ORF order along the chromosomes are identical to the order in [PlasmoDB.org](PlasmoDB.org). (30.9 MB EPS). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S3 ::: {.caption} ###### CGH of 3D7 versus HB3 for All Chromosomes Genomic differences between strain 3D7 and strain HB3 were measured by CGH. The relative hybridization between the gDNA derived from these two strains is shown as a percent reduction of the signal intensity for 3D7 along individual chromosomes. (1.7 MB ZIP). ::: ::: {.caption} ###### Click here for additional data file. ::: (A) ::: {.caption} ###### (216 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: (B) ::: {.caption} ###### (232 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: (C) ::: {.caption} ###### (237 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: (D) ::: {.caption} ###### (240 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: (E) ::: {.caption} ###### (252 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: (F) ::: {.caption} ###### (232 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: (G) ::: {.caption} ###### (235 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: (H) ::: {.caption} ###### (235 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: (I) ::: {.caption} ###### (249 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: (J) ::: {.caption} ###### (265 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: (K) ::: {.caption} ###### (283 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: (L) ::: {.caption} ###### (270 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: (M) ::: {.caption} ###### (305 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: (N) ::: {.caption} ###### (332 KB EPS) ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S4 ::: {.caption} ###### Distribution of Euclidian Distances between Expression Profiles of the IDC Genes and Seven Vaccine Candidates The similarity between each IDC expression profile and the profiles of the seven selected vaccine candidate genes was evaluated by Euclidian distance calculations, d(x,y) = Σ(x~i~ − y~i~)^2^. The Euclidian distance value to the closest vaccine homologue was selected for each IDC profile and used to generate this plot. Genes with d(x,y) \< 20 were selected for the phaseogram of putative vaccine targets (see [Figure 5](#pbio-0000005-g005){ref-type="fig"}). (494.02 KB EPS). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S1 ::: {.caption} ###### Pearson Correlation for ORFs Represented by Multiple Oligonucleotides This table contains all of the ORFs in the analyzed dataset that are represented by multiple oligonucleotides on the DNA microarray. The average Pearson correlation value has been calculated for the expression profiles of all oligonucleotides for each given ORF. (44 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S2 ::: {.caption} ###### P. falciparum Functional Gene Groups This table contains all of the P. falciparum groups discussed. The groups include the following: transcription machinery, cytoplasmic translation machinery, the glycolytic pathway, ribonucleotide synthesis, deoxyribonucleotide synthesis, DNA replication machinery, the TCA cycle, the proteaseome, the plastid genome, merozoite invasion, actin--myosin motility, early-ring transcripts, mitochondrial genes, and the organellar translational group. (291 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S3 ::: {.caption} ###### Coregulation along the Chromosomes of P. falciparum This table contains the regions of coregulation found in the chromosomes of P. falciparum determined by calculating the Pearson correlation between expression profiles for contiguous ORFs. The cutoff was 70% pairwise correlation of greater than 0.75 for each group. Only groups of two ORFs or more are listed. (6 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S4 ::: {.caption} ###### 3D7 versus HB3 CGH Data This table contains all of the intensity data from CGH of gDNA derived from the 3D7 and HB3 strains of P. falciparum. The averaged intensities from three microarray hybridization experiments are listed. (414 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S5 ::: {.caption} ###### Putative Apicoplast-Targeted Genes and Expression Profiles This table contains all of the predicted apicoplast-targeted ORFs from [PlasmoDB.org](PlasmoDB.org). The presence of each ORF on the DNA microarray is tabulated, as well as whether each ORF is present in the overview set. Finally, the plastid ORFs in-phase with plastid genome expression are listed, as well as the corresponding oligonucleotide identifiers. (147 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S6 ::: {.caption} ###### Putative P. falciparum Proteases and Their Expression Data The table was constructed by searching the database for any putative protease annotations and contains all of the 92 proteases identified by [@pbio-0000005-Wu1]). (59 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S7 ::: {.caption} ###### Vaccine Candidate Correlation Table The similarity of all expression profiles to seven known vaccine candidates was evaluated by a Euclidian distance calculation to all expression profiles measured. These 262 ORFs constitute the top 5% of genes in the IDC with minimum distance to these seven ORFs. The seven candidates used are AMA1, MSP1, MSP3, MSP5, EBA175, RAP1, and RESA1. (204 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: We would like to thank Ashwini Jambhekar, Pradip K. Rathod, David S. Roos, Phil J. Rosenthal, Anita Sil, Akhil Vaidya, and Dave Wang for critical comments. For technical assistance, we thank Takeshi Irie, Terry Minn, and Samara L. Reck-Peterson. This work was supported by the Burroughs-Wellcome Fund, the Kinship Foundation, a Sandler Opportunity Grant, and National Institute of Allergy and Infectious Diseases grant AI53862. **Conflicts of Interest.** The authors have declared that no conflicts of interest exist. **Author Contributions.** ZB, ML, and JLD conceived and designed the experiments. ZB, ML, and EDW performed the experiments. ZB, ML, BLP, EDW, JZ, and JLD analyzed the data. BLP and JZ contributed reagents/materials/analysis tools. ZB, ML, BLP, EDW, and JLD wrote the paper. Academic Editor: Gary Ward, University of Vermont. ASL : adenylosuccinate lyase CGH : comparative genomic hybridization CLAG : cytoadherence-linked asexual gene Clp : caseineolytic protease DHFR-TS : dihydrofolate reductase--thymidylate synthetase EBA : erythrocyte-binding antigen EBL : erythrocyte-binding-like protein FFT : fast Fourier transform FP : falcipain FV : food vacuole gDNA : genomic DNA HAP : histo-aspartyl protease hpi : hours postinvasion IDC : intraerythrocytic developmental cycle MSP : merozoite surface protein ORF : open reading frame PM : plasmepsin PV : parasitophorous vacuole RBC : red blood cell RBP : reticulocyte-binding protein RESA : ring-infected surface antigen SERA : serine repeat antigen TCA : tricarboxylic acid.

PubMed Central

2024-06-05T03:55:45.923570

2003-8-18

PMC176546

Introduction {#s1} ============ Elephants have a very limited distribution in Borneo, being restricted to approximately 5% of the island in the extreme northeast ([Figure 1](#pbio-0000006-g001){ref-type="fig"}). There are no historical records of elephants outside of this range. Fossil evidence for the prehistoric presence of elephants on Borneo is limited to a single specimen of a tooth from a cave in Brunei ([@pbio-0000006-Hooijer1]). ::: {#pbio-0000006-g001 .fig} Figure 1 ::: {.caption} ###### Asian Elephant Range and Sampling Locations in Borneo Solid lines demarcate country borders and the dotted line the boundary between the Malaysian states of Sabah and Sarawak. Black dots indicate areas of sample collection. :::  ::: Popular belief holds that elephants presented to the Sultan of Sulu in 1750 by the East India Trading Company and subsequently transported to Borneo founded the current population ([@pbio-0000006-Harrisson1]; [@pbio-0000006-Medway1]). These animals presumably originated in India ([@pbio-0000006-Shoshani1]), where company operations and trade in domesticated elephants were centred. Alternatively, considering the geographic proximity to Borneo, the elephant trade that flourished in Sumatra and peninsular Malaysia during the 16th--18th centuries ([@pbio-0000006-Andaya1]; [@pbio-0000006-Marsden1]\[1811\]) may have been the source. Thus, if elephants were introduced to Borneo, the source population could have been India, Sumatra, or peninsular Malaysia, and as a feral population, Borneo\'s elephants would have low conservation importance. Conversely, if elephants occurred naturally on Borneo, they would have colonised the island during Pleistocene glaciations, when much of the Sunda shelf was exposed ([Figure 2](#pbio-0000006-g002){ref-type="fig"}) and the western Indo-Malayan archipelago formed a single landmass designated as Sundaland ([@pbio-0000006-MacKinnon1]). Thus, the isolation of Borneo\'s elephants from other conspecific populations would minimally date from the last glacial maximum, 18,000 years ago, when land bridges last linked the Sunda Islands and the mainland ([@pbio-0000006-MacKinnon1]). If Borneo\'s elephants are of indigenous origin, this would push the natural range of Asian elephants 1300 km to the east, and as a unique population at an extreme of the species\' range, Borneo elephants\' in situ conservation would be a priority and ex situ cross-breeding with other populations would be contraindicated. ::: {#pbio-0000006-g002 .fig} Figure 2 ::: {.caption} ###### Asian Elephant Range and Sampling Locations Central sampling locations denote the countries sampled and represent a number of actual sampling locations within each country. 1. Sri Lanka, 2. India, 3. Bhutan, 4. Bangladesh, 5. Thailand, 6. Laos, 7. Vietnam, 8. Cambodia, 9. Peninsular Malaysia, 10. Sumatra (Indonesia) 11. Borneo (Sabah--Malaysia). :::  ::: Initially, Borneo elephants were classified as a unique subspecies (Elephas maximus borneensis) based on morphological differences from other populations ([@pbio-0000006-Deraniyagala1], 1955). Subsequently, they were subsumed under the Indian Elephas maximus indicus ([@pbio-0000006-Shoshani1]) or the Sumatran Elephas maximus sumatrensis ([@pbio-0000006-Medway1]) subspecies, based on an assumption of their introduction to the region or on the reasoning that morphological divergence was insufficient to warrant separate status. While unique subspecific status would highlight their conservation importance, evaluation of their status in terms of evolutionary significant units (ESUs) and management units (MUs) ([@pbio-0000006-Ryder1]; [@pbio-0000006-Moritz1]) would be more relevant to conservation management. Results {#s2} ======= We PCR-amplified and sequenced a 630 bp fragment of mitochondrial DNA (mtDNA), including the hypervariable left domain of the d-loop ([@pbio-0000006-Fernando3]), from 20 Borneo elephants and compared them with 317 sequences we generated for elephants across ten of the 13 Asian elephant range states ([Figure 2](#pbio-0000006-g002){ref-type="fig"}). Asian elephant haplotypes segregated into two distinct clades, α and β ([@pbio-0000006-Fernando3]). All 'Sundaland' (peninsular Malaysia, Sumatra, and Borneo) haplotypes fell in clade β, while α and β clades were observed in Sri Lanka and mainland populations ([Figures 3](#pbio-0000006-g003){ref-type="fig"} and [4](#pbio-0000006-g004){ref-type="fig"}). The Borneo population was fixed for the unique β-haplotype BD. Similar tree topologies were obtained by maximum parsimony, neighbour joining, and maximum-likelihood methods of phylogenetic analyses, with some minor rearrangements of the terminal branches. In all trees, Bornean and other haplotypes unique to 'Sundaland\' (Borneo: BD; peninsular Malaysia: BQ, BV; Sumatra: BS, BU, BT, BR) occupied basal positions in the β-clade phylogeny ([Figure 3](#pbio-0000006-g003){ref-type="fig"}) and were derived from internal nodes in a parsimony network of haplotypes ([Figure 4](#pbio-0000006-g004){ref-type="fig"}). Uncorrected *p* distances between the Borneo haplotype and other β-haplotypes ranged from 0.012 (haplotypes BQ, BP, BO, BS, BU) to 0.020 (haplotype BE), with a mean of 0.014. Assuming a nucleotide substitution rate of 3.5% per million years for the elephant mtDNA d-loop ([@pbio-0000006-Fleischer1]), the observed genetic distance indicates divergence of the Borneo haplotype BD and its closest relative from a common ancestor approximately 300,000 years ago. Owing to stochastic coalescent processes, the use of a single gene to infer population parameters is prone to error. Despite any such error, the magnitude of the genetic difference between Borneo and other Asian elephant haplotypes is such that it indisputably excludes divergence since introduction; the observed divergence is so great that even if there was some error it would not have any influence on the conclusion that places the Borneo haplotype in a timeframe supporting a Pleistocene colonisation rather than introduction by humans. ::: {#pbio-0000006-g003 .fig} Figure 3 ::: {.caption} ###### A Neighbour-Joining Phylogram of Asian Elephant Haplotypes Rooted with an African Elephant Out-Group Sunda Region haplotypes are in bold. :::  ::: ::: {#pbio-0000006-g004 .fig} Figure 4 ::: {.caption} ###### Network of Asian Elephant Haplotypes Based on Statistical Parsimony Grey circles with letters denote haplotypes unique to the Sunda region (BD: Borneo; BQ, BV: peninsular Malaysia; BR, BS, BT, BU: Sumatra). White circles with letters denote haplotypes found in mainland Asia (excluding peninsular Malaysia) and Sri Lanka. The small open circles denote hypothetical haplotypes. Haplotypes beginning with the letters A and B belong to the two clades α and β, respectively. :::  ::: We also genotyped 15 Borneo elephants for five polymorphic autosomal microsatellite loci ([@pbio-0000006-Nyakaana1]; [@pbio-0000006-Fernando4]) and compared them to 136 five-locus genotypes we generated for Asian elephants from nine range states. Tests of Hardy--Weinberg equilibrium and linkage disequilibrium in all populations indicated simple Mendelian inheritance of five unlinked, selectively neutral loci. The total number of alleles per locus across populations in the Asian elephant ranged from 2.0 (*EMX-2*) to 11.0 (*LafMS03*) (*x¯*, SE = 4.60, 1.51); the average number of alleles across loci, per population (excluding Borneo), from 2.0 (Sumatra) to 3.6 (Sri Lanka) (*x¯*, SE = 2.93, 0.155); the observed heterozygosity H~0~ across all populations (excluding Borneo) from 0.38 (*EMX-4*) to 0.63 (*LafMS03*) (*x¯*, SE = 0.44, 0.041); and gene diversity from 0.39 (*EMX-4*) to 0.69 (*LafMS03*) (*x¯*, SE = 0.47, 0.050). Comparatively, all indices demonstrated very low genetic diversity in the Borneo population: proportion of polymorphic loci, 0.4; number of alleles per locus, 1--2 (*x¯*, SE = 1.40, 0.219); gene diversity, 0--0.13 (*x¯*, SE = 0.04, 0.024); heterozygosity H~0~ = 0--0.07 (*x¯*, SE = 0.01, 0.013). The number of alleles, observed heterozygosity, and gene diversity, averaged across Asian elephant populations, were all higher than those in Borneo, at all loci ([Table 1](#pbio-0000006-t001){ref-type="table"}). Similarly, in all populations, the number of alleles and observed heterozygosity, averaged across loci, were higher than in Borneo ([Table 2](#pbio-0000006-t002){ref-type="table"}). Five unique genotypes were identified in the 15 Borneo elephants sampled. In tests of population subdivision, all pairwise comparisons between Borneo and other populations demonstrated highly significant differentiation, *F~ST~* 0.32--0.63 (*x¯*, SE = 0.44, 0.034) ([Table 3](#pbio-0000006-t003){ref-type="table"}). In tests of a recent bottleneck, no heterozygote excess ([@pbio-0000006-Maruyama1]) or mode-shift distortion of allele frequency distributions ([@pbio-0000006-Luikart1]), characteristic of a recent bottleneck, was observed in the Borneo population. In assignment tests indicating the distinctness of a population\'s genotypes, all five Borneo genotypes were assigned with maximum likelihood to Borneo (likelihoods ranging from 0.004 to 0.80, *x¯*, SE = 0.51, 0.175), and maximum-likelihood ratios of the most-likely (Borneo) to the next-most-likely population ranged from 2.97 to 48.20 (*x¯*, SE = 25.02, 8.795). Borneo was significantly more likely to be the source than any other population for all five genotypes, since each of the assignment likelihoods to Borneo fell outside the upper end of the corresponding distribution of assignment likelihoods to the other populations. Assignment likelihoods to the putative Indian, Sumatran, and peninsular Malaysian source populations were very small (India: 0--0.0004, *x¯*, SE = 0.000126, 0.000065; Sumatra: 0--0.0355, *x¯*, SE = 0.007146, 0.006336; peninsular Malaysia: 0.0003--0.1195, *x¯*, SE = 0.0301, 0.0201), indicating that Borneo\'s genotypes were highly unlikely to have originated from any of these populations. ::: {#pbio-0000006-t001 .table-wrap} Table 1 ::: {.caption} ###### Comparison of Measures of Genetic Variation at Individual Loci in Borneo with Those of the Other Populations :::  ::: ::: {#pbio-0000006-t002 .table-wrap} Table 2 ::: {.caption} ###### Measures of Genetic Variation Using Five Loci, in Asian Elephant Populations from across the Range :::  ::: ::: {#pbio-0000006-t003 .table-wrap} Table 3 ::: {.caption} ###### F*~ST~* Values in Pairwise Comparison of Borneo with Other Populations :::  ::: Discussion {#s3} ========== mtDNA evidence supports an indigenous hypothesis in three ways. First, this hypothesis assumes an ancient, independent evolution of Borneo\'s elephants, resulting in the unique, divergent Borneo haplotype(s), as we observed. Conversely, the introduction hypothesis assumes an introduction at 500 years ago or less, which approximates zero time on a scale of mtDNA d-loop evolution, and hence requires Borneo and source population haplotypes to be identical. This was not observed. Second, the estimated divergence time between the Borneo haplotype and other Asian elephant haplotypes is concordant with a mid- to late-Pleistocene isolation of elephants on Borneo and the vicariant history of the island ([@pbio-0000006-MacKinnon1]). Third, all observed 'Sundaland\' haplotypes, including Borneo\'s, were of the β clade, had basal relationships to that clade in a phylogenetic tree, and were independently derived from internal nodes in a haplotype network, suggesting an ancient isolation of these lineages on Borneo, Sumatra, and peninsular Malaysia. Thus, the Borneo haplotype fits a pattern of distribution and relatedness to other 'Sundaland\' haplotypes that is congruent with an ancient colonisation of the Sunda region by β clade and subsequent allopatric divergence of populations on its larger landmasses. Microsatellite data also support the indigenous hypothesis. If the Borneo population originated from animals introduced in the 16th--18th centuries, it would have reached its mid-20th-century size of approximately 2,000 individuals ([@pbio-0000006-deSilva1]) in fewer than 30 generations, assuming an Asian elephant generation time of 15--20 years ([@pbio-0000006-Sukumar1]). Thus, the Borneo population would have experienced a rapid demographic expansion after the 'recent' bottleneck caused by the founder-event of introduction. We did not observe a heterozygote excess or a mode-shift distortion in allele frequency distribution in the Borneo population, suggesting that the population did not undergo a recent bottleneck and hence did not arise from a few introduced animals. However, this result by itself is not conclusive, since with a sample size of 15 and five loci, the test for heterozygosity excess has low power and bottlenecks may not be detected ([@pbio-0000006-Luikart2]). We observed extremely low genetic diversity at Borneo elephant microsatellite loci, including fixation at three of the five loci. Sequential founder-events or persistent small population size, as would be expected in a small population isolated since the Pleistocene, would lead to substantial loss of genetic variation ([@pbio-0000006-Nei1]) and hence is consistent with the data. Successful founding of a population by a very few individuals from a single introduction could also result in a severe bottleneck. However, given the adversities faced by translocated elephants ([@pbio-0000006-Fernando1]) and the importance of social structure in the reproduction and survival of elephants ([@pbio-0000006-Fernando2]; [@pbio-0000006-McComb1]), such an explanation is unlikely. In the assignment tests, all five Borneo genotypes, which included free-ranging as well as captive animals, were assigned to Borneo with significantly higher likelihoods than to other populations and with extremely low likelihoods to the putative source populations. An introduced population may be highly divergent from the source population in terms of *F* statistics ([@pbio-0000006-Williams1]) due to allelic loss from founder-events. However, the probability of loss for a particular allele is inversely proportional to its frequency in the founder and hence the source population. Thus, genotypes in an introduced population would retain a high likelihood of assignment to the source population, enabling its identification from among a number of candidate populations. Therefore, the assignment tests strongly suggest that the Borneo elephants were not derived from another population in the recent past. Thus, microsatellite data strongly suggest a Pleistocene colonisation, independent evolution through a long period of isolation, and long-term small population size for the Borneo population. It strongly rejects a recent origin from any of the putative source populations. Mitochondrial and microsatellite analyses indicate that Borneo\'s elephants are indigenous to Borneo, have undergone independent evolution since a Pleistocene colonisation, and are not descended from animals introduced by humans. The evolutionary history of Borneo\'s elephants warrants their recognition as a separate ESU ([@pbio-0000006-Moritz1]). Thus, they should not be cross-bred with other Asian elephants in ex situ management. The genetic distinctiveness and evolutionary history of Borneo elephants support their recognition as a unique subspecies. However, one of the reasons E. maximus borneensis was subsumed under E. m. indicus and E. m. sumatrensis was the inadequacy of the original description of E. m. borneensis in terms of the morphological characters assessed and sample size. Therefore, we suggest that a formal reinstatement of the E. m. borneensis taxa await a detailed morphological analysis of Borneo elephants and their comparison with other populations. While Borneo\'s elephants appear to be genetically depauperate, through a long history of isolation and inbreeding, they may have purged deleterious recessive alleles from their genome and decreased their genetic load, thus becoming less susceptible to inbreeding depression. We recommend research on reproductive rates, juvenile survival, and other indicators of detrimental effects of inbreeding such as sperm deformities, sperm mobility, and genetic diversity at MHC loci. While increasing genetic diversity by introducing a small number of elephants from other populations ([@pbio-0000006-Whitehouse1]) may have to be considered if deleterious inbreeding effects are evident, in the absence of such findings Borneo\'s elephants should be managed separately from other Asian elephants. Materials and Methods {#s4} ===================== {#s4a} ### Samples. {#s4a1} Samples consisted of dung from free-ranging and dung or blood from captive elephants. Sample collection, storage, and DNA extraction followed published protocols ([@pbio-0000006-Fernando3], [@pbio-0000006-Fernando5]). For mitochondrial and microsatellite analysis, respectively, 20 and 15 samples from Borneo (nine blood samples from elephants captured for management purposes---eight from the Kretam area and one individual originating from around Lahad Datu---and the rest from dung samples from free-ranging elephants collected during a survey of the Kinabatangan watershed) were compared with 317 and 136 samples from across the current Asian elephant range, Sri Lanka (n = 81, 20), India (n = 81, 20), Bhutan (n = 13, 13), Bangladesh (n = 30, 20), Thailand (n = 8, 8), Cambodia (n = 30, 20), Vietnam (n = 5, 0), Laos (n = 20, 6), Indonesia (Sumatra) (n = 40, 20), and peninsular Malaysia (n = 9, 9). Vietnam was excluded from the microsatellite analysis owing to nonamplification of a number of samples. ### mtDNA amplification and sequencing. {#s4a2} Approximately 630 bp of mtDNA, including the left domain of the d-loop, were amplified using published primers ([@pbio-0000006-Fernando3]). PCR products were sequenced in both directions, using internal sequencing primers MDLseq-1 (CCTACAYCATTATYGGCCAAA) and MDLseq-2 (AGAAGAGGGACACGAAGATGG), and resolved in 4% polyacrylamide gels in an ABI 377 automated sequencer (Perkin-Elmer, Wellesley, Massachusetts, United States). ### mtDNA phylogenetic analysis. {#s4a3} We used 600 bp of the amplified segment in the analysis. Sequences were aligned and edited using SEQUENCHER version 3.1.1 (GeneCodes Corporation, Ann Arbor, Michigan, United States). Sequences were deposited in GenBank (accession numbers AY245538 and AY245802 to AY245827). Phylogenetic analyses were conducted using PAUP\* version 4.0 ([@pbio-0000006-Swofford1]). Three African elephant (Loxodonta africana) sequences from zoo animals in the United States were used as an out-group. Genetic distances among sequences were calculated using uncorrected *p* distances. Maximum-parsimony analysis was conducted using a heuristic search with random stepwise addition of taxa, tree bisection/reconnection branch swapping, and equal weighting; neighbour joining, with Kimura two-parameter distances; and maximum likelihood, using empirical base frequencies and estimated values for the shape parameter for among-site rate variation and transition/transversion ratios. A network of haplotypes was created using statistical parsimony in the software TCS version 1.13 ([@pbio-0000006-Clement1]). ### Microsatellite amplification. {#s4a4} Samples were screened with five published microsatellite loci, *EMX-1* to *EMX-4* ([@pbio-0000006-Fernando4]) and *LafMS03* ([@pbio-0000006-Nyakaana1]). Forward primers were fluorescent labelled (FAM, HEX, or TET), samples were amplified in 12.5 μl volumes with relevant cycling profiles ([@pbio-0000006-Fernando4]), and 1 μl of PCR product was mixed with 0.2 μl of loading-dye and 0.5 μl of Tamra 500 size standard (Applied Biosystems, Foster City, California, United States) and was resolved in 4% polyacrylamide gels in an ABI 377 automated sequencer. Alleles were scored using GENESCAN software (Applied Biosystems) and published guidelines ([@pbio-0000006-Fernando5]). ### Microsatellite data analysis. {#s4a5} Deviations from Hardy--Weinberg equilibrium for each locus and population were tested using the exact Hardy--Weinberg test as implemented in GENEPOP 3.2 ([@pbio-0000006-Raymond1]), with the complete enumeration method ([@pbio-0000006-Louis1]) for loci with fewer than four alleles and with the Markov chain method ([@pbio-0000006-Guo1]) (dememorization: 1000; batches: 100; iterations per batch: 1000) for loci with more than four alleles. GENEPOP was also used to test for linkage disequilibrium between loci, using the Markov chain method. Population differentiation was tested with estimates of Wright\'s fixation index ([@pbio-0000006-Weir1]), *F~ST~*, using the program Arlequin version 2 ([@pbio-0000006-Schneider1]). Evidence for a recent bottleneck in the Borneo population in terms of a heterozygote excess ([@pbio-0000006-Cornuet1]) or a mode-shift distortion in allele frequencies ([@pbio-0000006-Luikart1]) was conducted using the program BOTTLENECK version 1.2.02 ([@pbio-0000006-Piry1]) and a graphical method ([@pbio-0000006-Luikart1]). Assignment tests were performed using WHICHRUN version 4.1 ([@pbio-0000006-Banks1]). Assuming Hardy--Weinberg equilibrium in each baseline population and linkage equilibrium between loci, the likelihood that an individual originates from a particular population is the Hardy--Weinberg frequency of the individual\'s genotype at that locus, in that population. This likelihood was multiplied across loci to obtain a multilocus assignment likelihood of the test individual to each population, and the population with the highest value was identified as the 'most-likely' source population. To test for statistical significance of the most-likely source population, this assignment likelihood was compared with the distribution of assignment likelihoods of the other populations. Maximum-likelihood ratios were calculated as the ratio between the likelihood of assignment to the most-likely population to that for a particular population. Supporting Information {#s5} ====================== {#s5a} ### Accession Numbers {#s5a1} The GenBank accession numbers for the sequences reported in this paper are AY245538 and AY245802 to AY245827. We would like to thank Susan Mikota, Peter Malim, Eric Wickramakayake, Richard Lair, Jayantha Jayewardene, L. K. A. Jayasinghe, Manori Gunawardene, H. K. Janaka, Chandana Rajapakse, Ashoka Dangolla, Raman Sukumar, Ajay Desai, Christy Williams, Ainun Nishat, Mohsinuzzman Chowdhury, Mike Keele, Jeff Briscoe, Steve Osofsky, Karl Stromayer, Andrew Maxwell, Ou Ratanak, Lic Vuthy, Joe Heffernen, Rob Tizard, Tom Dillon, Vongphet, Buntjome, Kari Johnson, Heidi Riddle, Simon Hedges, Martin Tyson, Joshua Ginsberg, the Sabah Wildlife Department, Fauna and Flora International, the Department of Wildlife Conservation Sri Lanka, the International Union for Conservation of Nature and Natural Resources Bangladesh, the Zoological Gardens and Wildlife Rescue Centre Cambodia, Angkor Village Resort Elephant Farm (Siem Reap, Cambodia), the Seblat Elephant Training Centre (Bengkulu, Sumatra), Have Trunk Will Travel, the Department of Forests and Wildlife Cambodia, the Wildlife Trust, the World Wildlife Fund (WWF) Malaysia, WWF Vietnam, the Wildlife Conservation Society, the Portland (Oregon) Zoo, the Los Angeles Zoo, the Singapore Zoo, and the Pinnawela Elephant Orphanage Sri Lanka for help in obtaining samples; and Jennifer Pastorini, Ajay Desai, and two anonymous reviewers for comments on an earlier version of the manuscript. This study was conducted in partnership with the WWF\'s Asian Rhino and Elephant Action Strategy (AREAS) Programme and through additional collaboration with Wildlife Trust\'s Indian and Sri Lankan Elephant Programs and the Wildlife Conservation Society\'s Indonesia--Sumatran Elephant Project and Field Veterinary Program. It was made possible by grants from Ms. Nancy Abraham, the WWF United States, WWF for Nature (WWF International), the United States Fish and Wildlife Service\'s Asian Elephant Conservation Fund, and the Center for Environmental Research and Conservation Seed Grant Program and by additional support from the Laboratory of Genetic Investigation and Conservation, Columbia University. We dedicate this paper to the memory of our coauthor Annelisa Kilbourn, whose untimely death during her work in Gabon is a great loss to conservation. **Conflicts of Interest.** The authors have declared that no conflicts of interest exist. **Author Contributions.** This project is part of an ongoing multicentre collaboration on elephant conservation. All authors on this manuscript contributed substantively to the work described herein. Academic Editor: Craig Moritz, University of California, Berkeley. Δ In the online version of this article published on August 18, Annelisa Kilbourn\'s affiliation was incorrectly identified as the Sabah Wildlife Department. Her proper affiliation is shown here. ESU : evolutionary significant unit mtDNA : mitochondrial DNA MU : management unit.

PubMed Central

2024-06-05T03:55:45.929301

2003-8-18

PMC176547

A new study settles a long-standing dispute about the genesis of an endangered species. With scant fossil evidence supporting a prehistoric presence, scientists could not say for sure where Borneo\'s elephants came from. Did they descend from ancient prototypes of the Pleistocene era or from modern relatives introduced just 300--500 years ago? That question, as Fernando et al. report in this issue, is no longer subject to debate. Applying DNA analysis and dating techniques to investigate the elephants\' evolutionary path, researchers from the United States, India, and Malaysia, led by Don Melnick of the Center for Environmental Research and Conservation at Columbia, demonstrate that Borneo\'s elephants are not recent arrivals. They are genetically distinct from other Asian elephants and may have parted ways with their closest Asian cousins when Borneo separated from the mainland, effectively isolating the Borneo elephants some 300,000 years ago. In the 1950s, Borneo elephants had been classified as a subspecies of Asian elephants (either Indian or Sumatran) based on anatomical differences, such as smaller skull size and tusk variations. This classification was later changed, partly because of the popular view that these animals had descended from imported domesticated elephants. Until now, there was no solid evidence to refute this belief and no reason to prioritize the conservation of Borneo elephants. Their new status, as revealed by this study, has profound implications for the fate of Borneo\'s largest mammals. Wild Asian elephant populations are disappearing as expanding human development disrupts their migration routes, depletes their food sources, and destroys their habitat. Recognizing these elephants as native to Borneo makes their conservation a high priority and gives biologists important clues about how to manage them. ::: {#pbio-0000007-g001 .fig} ::: {.caption} ###### Borneo elephant :::  :::

PubMed Central

2024-06-05T03:55:45.932000

2003-8-18

PMC176548

Every year, malaria kills as many as 2.5 million people. Of these deaths, 90% occur in sub-Saharan Africa, and most are children. While four species of the single-celled organism Plasmodium cause malaria, Plasmodium falciparum is the deadliest. Harbored in mosquito saliva, the parasite infects its human host as the mosquito feeds on the victim\'s blood. Efforts to control the disease have taken on an increased sense of urgency, as more P. falciparum strains show resistance to antimalarial drugs. To develop new drugs and vaccines that disable the parasite, researchers need a better understanding of the regulatory mechanisms that drive the malarial life cycle. Joseph DeRisi and colleagues now report significant progress toward this goal by providing the first comprehensive molecular analysis of a key phase of the parasite\'s life cycle. While P. falciparum is a single-celled eukaryotic (nucleated) organism, it leads a fairly complicated life, assuming one form in the mosquito, another when it invades the human liver, and still another in human red blood cells (erythrocytes). The intraerythrocytic developmental cycle (IDC) is the stage of the P. falciparum lifecycle associated with the clinical symptoms of malaria. Using data from the recently sequenced P. falciparum genome, the researchers have tracked the expression of all of the parasite\'s genes during the IDC. The pattern of gene expression (which can be thought of as the internal operating system of the cell) during the IDC is strikingly simple. Its continuous and clock-like progression of gene activation is reminiscent of much simple life forms---such as a virus or phage---while unprecedented for a free living organism. Virus and phage behave like a "just-in-time" assembly line: components are made only as needed, and only in the amount that is needed. In this respect, malaria resembles a glorified virus. Given the remarkable coupling of the timing of gene activation with gene function, as shown in this paper, this understanding could help identify the biological function of the 60% of genes in P. falciparum that encode proteins of unknown function. P. falciparum appears to be ultra-streamlined and exquisitely tuned to perform a single job: consume, replicate, and invade. The simple program regulating the life of P. falciparum may hold the key to its downfall as any perturbation of the regulatory program will likely have dire consequences for the parasite. This offers renewed hope for the design of inhibitory drugs targeted at the regulatory machinery that would irreparably foul the parasite\'s regulatory program, ultimately resulting in its death. ::: {#pbio-0000011-g001 .fig} ::: {.caption} ###### Gene expression profile of P. falciparum :::  :::

PubMed Central

2024-06-05T03:55:45.932467

2003-8-18

PMC193604

Introduction {#s1} ============ Circadian rhythms of diverse organisms are based on similar intracellular molecular feedback loops ([@pbio-0000013-Dunlap1]; [@pbio-0000013-Allada2]; [@pbio-0000013-Panda1]). Based on this view, it is believed that one or a small number of clock cells are sufficient for self-sustained rhythms ([@pbio-0000013-Dunlap1]). This is despite the complex cellular organizations of many tissues, organisms, and systems ([@pbio-0000013-Kaneko1]; [@pbio-0000013-Schibler1]). In Drosophila, circadian clocks have been identified in a diverse range of cell types throughout the head and the body ([@pbio-0000013-Glossop1]; [@pbio-0000013-Hall1]). However, the clocks in different cells are considered nonidentical ([@pbio-0000013-Krishnan1]; [@pbio-0000013-Glossop1]; [@pbio-0000013-Levine1]; [@pbio-0000013-Schibler1]). In many tissues, molecular oscillations undergo rapid damping without environmental timing cues ([@pbio-0000013-Hardin1]; [@pbio-0000013-Plautz1]; [@pbio-0000013-Stanewsky1]; [@pbio-0000013-Giebultowicz1]). This is similar to the damping of in vitro rhythms in some mammalian tissues ([@pbio-0000013-Balsalobre1]; [@pbio-0000013-Schibler1]). In contrast, the Drosophila "core pacemaker" is believed to maintain robust oscillations for a long time in constant darkness (DD) with little or no damping, such that circadian behaviors can persist under such conditions ([@pbio-0000013-Dowse1]). Indeed, self-sustaining oscillations are a defining characteristic of true circadian rhythms and are believed to be required of a fully functional rhythmic cell. The differences between the "core pacemaker" and the clock machinery within damping cells or systems are unknown. The six clusters of approximately 100 clock neurons in the adult Drosophila brain are well characterized ([@pbio-0000013-Kaneko1]). Recent studies have focused principally on one of these groups, the small ventral lateral neurons (s-LN~v~s), as the best "core pacemaker" candidate for the following reasons: (1) in the developmental mutant *disco*, the presence of LN~v~s correlates with the maintenance of behavior rhythmicity ([@pbio-0000013-Helfrich-Forster2]); (2) LN~v~s specifically express the neuropeptide pigment-dispersing factor (PDF), and the *Pdf^01^*-null mutant loses behavioral rhythmicity under DD conditions ([@pbio-0000013-Renn1]); (3) genetic ablation of the LN~v~s by expressing proapoptotic genes causes the loss of rhythmicity in DD ([@pbio-0000013-Renn1]); and (4) the s-LN~v~s maintain robust molecular oscillations for at least for 2 days in DD ([@pbio-0000013-Yang1]; [@pbio-0000013-Shafer1]), in contrast to at least some other brain neurons and nonneuronal tissues. This final property suggests that these cells might fulfill the self-sustaining criterion for the "core pacemaker." Indeed, the s-LN~v~s have been proposed to the primary pacemaker cells that generate locomotor activity rhythms ([@pbio-0000013-Helfrich-Forster2]; [@pbio-0000013-Renn1]; [@pbio-0000013-Emery2]). Consistent with this cell-autonomous view of circadian rhythmicity, it has been shown that the LN~v~s possess all components of a fully functional, independent circadian clock: the photoreceptor cryptochrome, the rhythm-generating feedback loops, and a putative output factor, the neuropeptide PDF ([@pbio-0000013-Emery2]). Our pursuit of the self-sustaining "core pacemaker" of the Drosophila circadian system began with a test of the s-LN~v~ cell-autonomous clock hypothesis. Results {#s2} ======= LN~v~s Cannot Support Circadian Behavior Independently {#s2a} ------------------------------------------------------ To test whether the LN~v~s can support free-running circadian locomotor activity rhythms independently of other functional clock cells, we restricted pacemaker activity to these few PDF-expressing cells. CYCLE (CYC) is a bHLH--PAS protein ([@pbio-0000013-Rutila1]) and forms a heterodimeric transcription factor with CLOCK (CLK), another bHLH--PAS protein ([@pbio-0000013-Allada1]). CYC is an essential component of the Drosophila circadian oscillator transcriptional feedback loop ([@pbio-0000013-Glossop2]). The *cyc^01^* nonsense mutation completely eliminates molecular oscillations, and the direct target genes *period (per)* and *timeless (tim)* mRNAs are essentially undetectable ([@pbio-0000013-Rutila1]). Behavioral rhythms are also absent in the *cyc^01^* homozygous mutant strain ([@pbio-0000013-Rutila1]). We rescued *cyc^01^* specifically in the LN~v~s, by using a well-characterized *pdf--GAL4* driver ([@pbio-0000013-Renn1]) in combination with a *UAS--CYC* transgene to express ectopically wild-type CYC. Since CYC is apparently not a rate-limiting component of active dCLK--CYC complexes ([@pbio-0000013-Bae1]) and does not undergo molecular oscillations itself ([@pbio-0000013-Rutila1]), we expected that CYC overexpression would not cause circadian oscillator dysfunction. Indeed, the presence of the two transgenes did not affect locomotor activity rhythms in a wild-type background ([Figure 1](#pbio-0000013-g001){ref-type="fig"}C, right panel). ::: {#pbio-0000013-g001 .fig} Figure 1 ::: {.caption} ###### Rescuing Molecular Oscillations within the LN~v~s Is Not Sufficient to Rescue Locomotor Activity Rhythms The rescued mutant genotype is *y w*;*pdf--GAL4;UAS--CYC*,*cyc^01^*/*cyc^01^*. The flies were entrained in standard LD conditions and timepoints taken. Molecular oscillations were examined by whole-mount in situ hybridization of the *tim* gene. Double staining with a *Pdf* probe was used to label the LN~v~s neuronal group. (A and B) These show representative duplicate experiments. No *tim* mRNA signal is detectable in the dorsal region of the brain. The lower arrows point to the s-LN~v~s and the upper arrows to the l-LN~v~s. (A) Brain taken at timepoint ZT3. Panels shown from left to right are *Pdf* (green, FITC labeled), *tim* (red, Cy3 labeled), and an image overlay. (B) Brain taken at timepoint ZT15. Panels shown from left to right are *Pdf* (green, FITC labeled), *tim* (red, Cy3 labeled), and an image overlay. \(C) The double-plotted actograms of rescue mutant and control flies in a standard LD:DD behavior assay. The colors on the background indicate the lighting conditions of the behavior monitors (white, lights on; light blue, lights off). In the actogram, the average locomotor activity of the group of flies is plotted as a function of time. The left panel shows the actogram of the rescued mutant flies (*y w;pdf--GAL4/+;UAS--CYC,cyc^01^*/*cyc^01^*, n = 30). RI (rhythm index; [@pbio-0000013-Levine1]) = 0.14. The right panel shows the actogram for the rescued wild-type (control) flies (*y w;pdf--GAL4/+;UAS--CYC/+*, n = 32, RI = 0.61). :::  ::: The rescued mutant flies (*pdf--GAL4;UAS--CYC,cyc^01^/cyc^01^*) were examined by two independent criteria. First, molecular oscillations were assayed by in situ hybridization with a *tim* probe ([Figure 1](#pbio-0000013-g001){ref-type="fig"}A and 1B). *tim* RNA levels undergo robust cycling in wild-type flies, with a trough at ZT3 and a peak at ZT15 ([@pbio-0000013-Sehgal1]). This is also true within all individual clock neurons ([@pbio-0000013-Zhao1]). *tim* mRNA cycled in the LN~v~s ([Figure 1](#pbio-0000013-g001){ref-type="fig"}A and B), indicating successful rescue of the molecular oscillator within these cells. The fact that other clock neurons were still *tim* mRNA-negative ([Figure 1](#pbio-0000013-g001){ref-type="fig"}A and B) suggests that CYC and the rest of the molecular machinery can function cell autonomously, at least in the LN~v~s under these light--dark (LD) conditions. The observed oscillations are also not passively driven by light, since they persisted in DD, at least in the s-LN~v~s ([Figure S1](#sg001){ref-type="supplementary-material"}). Second, locomotor activity rhythms were examined by standard behavioral criteria. The transgenic flies were completely arrhythmic in DD. They were also arrhythmic under LD conditions, as the flies failed to anticipate the discontinuous transitions from light to dark or from dark to light (see [Figure 1](#pbio-0000013-g001){ref-type="fig"}C, left panel; [@pbio-0000013-Rutila1]). In summary, the behavioral phenotypes were indistinguishable from those of the parental *cyc^01^* mutant strain. Brain Clock Neurons Manifest Robust Molecular Oscillations in DD {#s2b} ---------------------------------------------------------------- The insufficiency of LN~v~ molecular rhythmicity indicates that one or more additional groups of rhythmic clock neurons are required for behavioral rhythmicity. We considered that robust molecular cycling under extended constant darkness conditions might be a good criterion for identifying these cell groups, because prior biochemical studies showed that some head and brain locations undergo damping of molecular oscillations under free-running conditions ([@pbio-0000013-Hardin1]; [@pbio-0000013-Stanewsky1]). This conclusion has been extended by more recent immunohistochemical observations ([@pbio-0000013-Yang1]; [@pbio-0000013-Shafer1]). The criterion of maintaining persistent and robust molecular rhythms in DD therefore suggests that only a limited set of brain locations are likely to be free-running pacemaker candidates. In order to identify these neurons, we assayed fly brains by *tim* in situ hybridization after 8 days in DD. To our surprise, we found that all *tim-*expressing brain cell groups (including both large ventral lateral neurons \[l-LN~v~s\] and small ventral lateral neurons \[s-LN~v~s\], doral lateral neurons \[LN~d~s\], and all three groups of dorsal neurons \[DNs\]) still cycle robustly at this time ([Figure 2](#pbio-0000013-g002){ref-type="fig"}). Previous studies have reported that the l-LN~v~s fail to maintain oscillations at the beginning of DD ([@pbio-0000013-Yang1]; [@pbio-0000013-Shafer1]). We have reproduced these observations, but noticed that the l-LN~v~s "adapt" to constant conditions by becoming rhythmic once again after about 2 days in DD (data not shown). These results clearly distinguish the brain from the eyes and other peripheral tissues, which rapidly lose coherent molecular oscillations under free-running conditions ([@pbio-0000013-Hardin1]; [@pbio-0000013-Plautz1]; [@pbio-0000013-Stanewsky1]; [@pbio-0000013-Giebultowicz1]). Although this approach failed to identify the additional neuronal groups necessary for behavioral rhythms, it suggests that many of these brain neuronal groups might act together in a network to support robust rhythms. ::: {#pbio-0000013-g002 .fig} Figure 2 ::: {.caption} ###### All Brain Clock Neuronal Groups Maintain Robust Oscillations of *tim* RNA Levels in DD Wild-type flies were entrained for at least 3 days and then released into DD. *tim* RNA was assayed at trough (left panels) and peak (right panels) timepoints by whole-mount in situ hybridization. Wild-type flies in LD (A) were compared with the eighth day of DD (B). On the eighth day of DD, the locomotor activities of the fly population were still in close synchrony, without any obvious phase spreading (data not shown). Left panels, brains at ZT3 (A) or CT3 (B); right panels, brains from ZT15 (A) or CT15 (B). Both (A) and (B) are representative of three replicate experiments. :::  ::: Sustained Molecular Oscillation in Constant Darkness Requires PDF {#s2c} ----------------------------------------------------------------- This association between robust molecular oscillations in all brain clock cells and behavioral rhythms in DD also made us consider the role of the neuropeptide PDF. The *Pdf^01^* mutant strain is unique among identified Drosophila circadian mutants, as it has little effect under LD conditions, but loses behavioral rhythmicity gradually and specifically in DD ([@pbio-0000013-Renn1]). This phenotype might reflect a disassociation between behavioral rhythmicity and the underlying molecular oscillations, as predicted from the role of PDF as a circadian output signal; it is proposed to connect the molecular oscillation in the LN~v~s to locomotor activity ([@pbio-0000013-Renn1]). We considered a completely different interpretation, namely, that PDF contributes to the functional integration of several brain clock neuronal groups, which is necessary to sustain molecular as well as behavioral rhythmicity under constant conditions. This fits well with previous studies of PDF in other organisms ([@pbio-0000013-Rao1]; [@pbio-0000013-Petri1]). In contrast to the canonical output model, this possibility suggests that the *Pdf^01^* mutant might manifest unusual molecular oscillations within clock neurons, especially under DD conditions. To address this issue experimentally, we examined *Pdf^01^* mutant flies by *tim* in situ hybridization. In *Pdf^01^* flies, all clock neurons had robust *tim* RNA oscillations in LD, and the cycling phase and amplitude were comparable to those of wide-type flies ([Figure 3](#pbio-0000013-g003){ref-type="fig"}A). The mutant flies were then released into DD and assayed at various times thereafter. In the first day of DD, cycling was similar to that observed in LD ([Figure 3](#pbio-0000013-g003){ref-type="fig"}B). By the fourth day of DD, however, the cycling amplitude was much reduced in all clock neurons ([Figure 3](#pbio-0000013-g003){ref-type="fig"}C and 3D). This was most evident from the unusually high signal in the CT2 sample; in wild-type flies, no *tim* signal was detected in any clock neuron at this timepoint ([Figure 3](#pbio-0000013-g003){ref-type="fig"}C, left panels). There was also a reduced signal strength at the peak time, CT14 ([Figure 3C](#pbio-0000013-g003){ref-type="fig"}, fourth panel from the left). The result parallels the damping of behavioral rhythms in the *Pdf^01^* mutant strain ([@pbio-0000013-Renn1]). ::: {#pbio-0000013-g003 .fig} Figure 3 ::: {.caption} ###### Molecular Oscillations of *tim* RNA Damp in DD in the *Pdf^01^* Mutant *tim* RNA oscillations were examined in the *Pdf^01^* mutant under both LD (A) and different days in DD (\[B\] and \[C\]), by whole-mount in situ hybridization. (A), (B), and (C) are representative images from replicas of three experiments. \(A) The left panel is from ZT3, and the right panel is from ZT15. A normal *tim* oscillation profile is observed compared to that of wild-type (see [Figure 2](#pbio-0000013-g002){ref-type="fig"}A). \(B) Brains from the *Pdf^01^* mutant in the first day of DD. Left panel, CT3; right panel, CT15. Oscillations are comparable to those in LD. \(C) Brains taken in the fourth day of DD. Six timepoints were taken throughout the circadian day. The sequence of panels from left to right is CT2, 6, 10, 14, 18, and 20, respectively. Wild-type brains (top row) were assayed in parallel with those from the *Pdf^01^* mutant (bottom row). See text for details. \(D) Quantification of (C). Relative intensities are taken from normalized mean pixel intensities. Different clock neuronal groups were quantified independently and compared between wild-type (blue curves) and *Pdf^01^* mutant (purple curves). The panels from left to right are quantification of *tim* RNA oscillation in the DNs, in the LN~d~s, and in the LN~v~s. Reduced cycling amplitude and a significant advanced phase were observed in the fourth day of DD. See text for details. :::  ::: Despite the gradual fading of locomotor activity rhythms in DD, a significant fraction of *Pdf^01^* mutant flies is still weakly rhythmic after 4 d of DD ([@pbio-0000013-Renn1]). By tracking their locomotor activity phases, we observed that most of them had accumulated an approximately 4-hour phase advance relative to wild-type flies by the fourth day in DD. This is consistent with the measured ca. 23-hour periods of these weakly rhythmic flies (1-hour phase advanced per day for 4 days) as well as their advanced evening activity peak in LD ([@pbio-0000013-Renn1]). Quantitation of the *tim* in situ hybridization signal showed that there was a comparable one-point (4 h) advance in the peak of *tim* RNA and also confirmed the reduced cycling amplitude ([Figure 3](#pbio-0000013-g003){ref-type="fig"}D). In order to eliminate the possibility that the observed damping is caused by the asynchrony of the *Pdf^01^* fly population, locomotor activities were tracked in real time. Individual flies were then removed from the monitors to assay *tim* RNA levels. Identical damped molecular oscillations were also observed in this case (data not shown). Taken together, the results indicate an excellent quantitative correspondence in phase and amplitude between the *tim* RNA rhythms and the behavioral rhythms in all clock neurons of the *Pdf^01^* strain. To extend these observations, we also assayed *cryptochrome (cry)* mRNA oscillations by in situ hybridization. *cry* is expressed in a similar clock neuron pattern to *tim*, but it has a peak expression at ZT2 and a trough at ZT14 ([@pbio-0000013-Emery1]; [@pbio-0000013-Zhao1]). This phase is opposite to that of *tim* and other CLK--CYC direct target genes and reflects the fact that *cry* is only indirectly regulated by this heterodimeric transcription factor; CLK--CYC directly regulates the transcription factors PDP1 and VRILLE, which then regulate *cry* ([@pbio-0000013-Cyran1]; [@pbio-0000013-Glossop3]). Despite these differences between *tim* and *cry*, a similar result was obtained for *cry* in the *Pdf^01^* strain in the fourth day of DD ([Figure 4](#pbio-0000013-g004){ref-type="fig"}), i.e., a reduced cycling amplitude compared to the fourth day of DD in a wild-type strain. This is suggested by the in situ pictures and is strongly indicated by the quantitation ([Figure 4](#pbio-0000013-g004){ref-type="fig"}). The correspondence between the *tim* and *cry* mRNA patterns indicates that the entire circadian transcriptional program damps in the mutant strain in DD, which underlies the behavioral damping. ::: {#pbio-0000013-g004 .fig} Figure 4 ::: {.caption} ###### *cry* RNA Oscillation Amplitude Is Also Reduced by the Fourth Day of DD in the *Pdf^01^* Mutant *cry* RNA expression in the brain was examined at the fourth day of DD by whole-mount in situ hybridization using a *cry* probe. Timepoints were taken every 4 hours throughout the circadian day. The sequence of panels from left to right is CT2, 6, 10, 14, 18, and 20, respectively. Wild-type brains (top row) were analyzed in parallel with those from the *Pdf^01^* mutant (bottom row). Shown are representative images from duplicate experiments. Quantification of *cry* RNA oscillations in different cell groups is as shown in [Figure 3](#pbio-0000013-g003){ref-type="fig"}. Ubiquitous damping of the cycling amplitude in the different cell groups was observed in the *Pdf^01^* mutant. :::  ::: PDF Is Likely to Act upon Clock Neurons Directly {#s2d} ------------------------------------------------ It is noteworthy that the mRNA oscillations damp uniformly in the *Pdf^01^* mutant strain, including the PDF-expressing LN~v~s (see [Figures 3](#pbio-0000013-g003){ref-type="fig"} and [4](#pbio-0000013-g004){ref-type="fig"}). Since PDF is a neuropeptide ([@pbio-0000013-Rao1]), it is unlikely to exert a direct intracellular effect on the LN~v~ transcriptional machinery. A more conservative interpretation is that PDF maintains intercellular communication between individual LN~v~ neurons ([@pbio-0000013-Petri1]) and/or between the LN~v~s and other cells; the communication is essential for self-sustained molecular rhythms within the LN~v~s. Although this "feedback" could be quite indirect, the l-LN~v~s project to the contralateral LN~v~s through the posterior optic tract. Moreover, the s-LN~v~s project dorsally to the superior protocerebrum, the location of the DNs. ([@pbio-0000013-Helfrich-Forster1]). These anatomic features suggest that PDF might bind directly to clock neurons. To test this hypothesis, in vitro biotinylated PDF peptide was incubated with fixed adult brains under near physiological conditions. The bound peptide was then detected in situ with a streptavidin-conjugated enzymatic amplification reaction. The vast majority of the signal localized with numerous cells at the periphery of medulla ([Figure 5](#pbio-0000013-g005){ref-type="fig"}A). This is exactly where the l-LN~v~s send large arborizations as their centrifugal projections ([@pbio-0000013-Helfrich-Forster1]). Importantly, signal was also detected coincident with the LN~v~s ([Figure 5](#pbio-0000013-g005){ref-type="fig"}B) and likely DN3 clock neurons ([Figure 5](#pbio-0000013-g005){ref-type="fig"}C) within the superior protocerebrum region, i.e., the bound peptide colocalized with GFP when the brains were from a strain with GFP-labeled clock neurons. Staining intensity was temporally constant; i.e., there was no systematic variation in signal intensity with circadian time. Although we obtained identical results with two differently biotinylated PDF peptides and there was no staining with two other biotinylated control peptides, we had difficulty to compete specifically the signal with nonbiotinylated PDF (see Materials and Methods). Moreover, PDF peptide staining of clock neurons was not reliably detected in every brain, in contrast to optic lobe staining. Nonetheless, we never detected peptide staining of other neurons in the vicinity of the LN~v~s; i.e., signal in this region of the brain was always coincident with the GFP-labeled LN~v~s. The peptide staining therefore suggests that PDF acts on the LN~v~s in an autocrine or paracrine fashion as well as on other clock neurons, but the results do not exclude additional, more indirect modes of action. ::: {#pbio-0000013-g005 .fig} Figure 5 ::: {.caption} ###### A PDF Peptide Binds to Many Cells, Including Several Clock Neuronal Groups In vitro biontinylated PDF peptide was used to visualize the peptide binding locations (middle panels, with Cy3) in the brain (see Materials and Methods for details). We used membrane-bound GFP (green panels on the left) to label specific circadian neurons as well as their projections (right panels show the overlay of both channels). \(A) The brain is from flies with labeled LN~v~s *(y w,UAS--mCD8iGFP;pdf--GAL4)*. Numerous cells at the periphery of the medulla have the vast majority of the bound PDF peptide signal within the brain. This region receives widespread dendritic arborizations from the l-LN~v~s. \(B) Bound PDF peptide was also detected on the surface of LN~v~s at a lower intensity. LN~v~ cell bodies were labeled using *UAS--mCD8iGFP;pdf--GAL4*. Since the signal from the Cy3 channel was much weaker than the GFP signal, we reduced the output gain from the GFP channel. Sequential scanning was used to prevent cross-talk between the two channels. \(C) *y w,UAS--mCD8iGFP;tim--GAL4/+* flies were used to label all circadian neurons. In the dorsal region shown in this series, the arrow points to a group of DN3 neurons. :::  ::: Discussion {#s3} ========== The strong behavioral phenotype of the *Pdf^01^* mutant strain in DD indicates that PDF makes an important contribution to free-running circadian rhythms. It was, however, unanticipated that the *Pdf^01^* mutant would have an additional effect on transcriptional oscillations within most if not all clock neurons. This observation extends the tight parallel between strong behavioral rhythms and robust transcriptional rhythms and suggests that the behavioral damping is due to the transcriptional damping ([@pbio-0000013-Marrus1]). In contrast to this strong effect of the *Pdf^01^* mutation on free-running rhythms, the molecular as well as behavioral rhythms of these mutant flies are nearly normal under LD conditions. We now interpret this difference to indicate that intercellular communication among different clock cells and neuronal groups is less important when they can independently receive photic information via cryptochrome. This probably serves not only to synchronize clock neurons but also to reinforce and strengthen the molecular oscillation ([@pbio-0000013-Emery1]; [@pbio-0000013-Stanewsky2]). The damping phenotype includes the LN~v~s, which have been proposed to be the principal pacemaker neurons in Drosophila ([@pbio-0000013-Helfrich-Forster2]; [@pbio-0000013-Renn1]). Their counterparts in mammals, the suprachiasmatic nucleus (SCN) neurons, can support circadian rhythms independently (e.g., [@pbio-0000013-Sujino1]). However, our data indicate that the LN~v~s cannot support locomotor activity rhythms without other clock cell groups (see [Figure 1](#pbio-0000013-g001){ref-type="fig"}). A similar attempt to rescue behavioral rhythms of an arrhythmic *Clk* mutant also failed ([@pbio-0000013-Allada3]). Although the negative result shown here might be due to developmental defects of the *cyc^01^* mutation ([@pbio-0000013-Park1]), the conclusion fits well with a role for PDF in functional cooperation between individual neuronal groups. Indeed, it appears that PDF secretion comprises much of what the LN~v~s contribute to rhythms, as the phenotype of flies missing the LN~v~s is virtually identical to that of the *Pdf^01^* strain ([@pbio-0000013-Renn1]). There is less known about the roles of other clock neurons, although they do have specific wiring properties ([@pbio-0000013-Kaneko1]) as well as specific sets of gene expression profiles (unpublished data). An additional indication that other clock neurons contribute to locomotor activity rhythms is that LD behavioral rhythms do not require the LN~v~s ([@pbio-0000013-Hardin2]; [@pbio-0000013-Renn1]). As the *Pdf^01^* strain also has a strong effect on geotaxis ([@pbio-0000013-Toma1]), clock neurons may even contribute to other behavioral modalities. The staining pattern suggests that the PDF ligand contacts a receptor on the surface of clock neurons, including the LN~v~s themselves. This is consistent with the notion that PDF acts as an important intercellular cell communication molecule within the Drosophila circadian system. The dorsal projections of the s-LN~v~s stain rhythmically with anti-PDF antibodies, and it has been suggested that released PDF affects dorsal clock neurons ([@pbio-0000013-Helfrich-Forster3]). Indeed, ectopic expression of PDF in neurons that project to the dorsal brain region causes severe rhythm defects, suggesting that misregulation of this signaling causes circadian system dysfunction ([@pbio-0000013-Helfrich-Forster3]). Our staining with a PDF peptide indicates that the PDF signaling to the DNs may be direct. Although rhythmic PDF staining is restricted to the s-LN~v~ terminals ([@pbio-0000013-Park1]), this could be because a smaller fraction of PDF is released from the l-LN~v~ terminals. Some of these processes follow the posterior optic track to the opposite side of the brain. Taken together with the LN~v~ peptide staining, it is likely that PDF from the l-LN~v~s signals contralaterally and positively influences clock cells on the opposite side of the brain. A very recent study of the Drosophila prothoracic gland (PG) clock and eclosion rhythms suggests that the LN~v~s also control the PG clock via PDF signaling ([@pbio-0000013-Myers1]). This raises the possibility that PDF not only synchronizes brain clock neurons, but also keeps peripheral clocks in pace with the core brain network. The *Pdf^01^* molecular phenotype implies that the wild-type organization of the system normally supports the individual clock cells as well as the entire circadian program in DD. Although we do not know that all molecular aspects of rhythms damp in DD in *Pdf^01^* flies, we suggest that damped transcriptional rhythms are the intracellular default state in Drosophila and are manifest without the driving and entraining LD cycle or without a functionally integrated clock network. This view is also consistent with recent studies showing that electrical silencing of clock neurons eliminates free-running molecular as well as behavioral rhythms ([@pbio-0000013-Nitabach1]). It will be interesting to learn how PDF signaling connects to the intracellular transcriptional machinery. We note that communication among clock neurons is likely to be important in other organisms. The ability of PDF to phase-shift the cockroach circadian clock ([@pbio-0000013-Petri1]) is more consistent with our proposal than with a simple role in clock output. A recent study of VPAC(2) receptor knock-out mice ([@pbio-0000013-Harmar1]) showed that these mice fail to sustain behavioral rhythms and have molecular rhythms defects within the SCN. This raises the intriguing possibility that SCN neurons as well as Drosophila clock neurons may require network integration to sustain free-running intracellular oscillations. Materials and Methods {#s4} ===================== {#s4a} ### Drosophila genetics. {#s4a1} Full-length *cyc* cDNA was obtained from BDGP cDNA clone GM02625 and was tagged with hemagglutinin (HA) epitope by PCR cloning. CYC--HA was subsequently cloned into pUAST to generate pUAS--CYC--HA. The transformation plasmid was used to generate transgenetic flies. A third chromosome insertion line *(UAS--CYC--HA15)* was used subsequently. All wild-type flies and specimens were taken from a Canton-S stock. The circadian driver lines *pdf--GAL4* ([@pbio-0000013-Renn1]), *tim--GAL4* ([@pbio-0000013-Kaneko1]), as well as the *cyc^01^* ([@pbio-0000013-Rutila1]) and *Pdf^01^* ([@pbio-0000013-Renn1]) mutant strains have been previously described. All molecular and behavioral analyses were conducted on flies entrained at 25°C. ### GFP expression analysis. {#s4a2} To visualize the axon projections from circadian neurons, a *UAS--mCD8GFP* line labeling the cell membrane was crossed with various circadian GAL4 drivers. The progeny brains were dissected in PBS and fixed in 3.7% paraformaldehyde in PEM. After rinses in PBS plus 0.3% Triton and PBS, brains were mounted in Vectashield mounting medium (Vector Laboratories, Burlingame, California, United States) and imaged on a Leica laser scanning confocal microscope. Optical sections were taken at 1--2 μm intervals and used to construct a maximum projection image for each brain. ### In situ mRNA hybridization on adult brain whole mounts. {#s4a3} In situ hybridization of *tim* and *cry* was done as described previously ([@pbio-0000013-Zhao1]). The maximum projection images taken from a Leica laser scanning confocal microscope were used for the quantification. The quantification was done using three brain images per sample with Leica confocal software. The mean pixel intensities of cell groups were normalized by subtracting the average of two general background areas in the brain. ### Behavioral analysis. {#s4a4} Flies were entrained for 3--5 d in 12 h light:12 h dark (LD) conditions before release into DD. Locomotor activities of individual flies were monitored using Trikinetics Drosophila Activity Monitors (Waltham, Massachusetts, United States). The analysis was done by using a signal processing toolbox ([@pbio-0000013-Levine2]). Autocorrelation and spectral analysis were used to assess rhythmicity and to estimate the period. The phase information was extracted using circular statistics ([@pbio-0000013-Levine2]). In some cases, the phases of individual *Pdf^01^* flies were also examined by inspection. ### In vitro peptide binding assay. {#s4a5} Biotinylation of the PDF peptide was with EZ-Link Sulfo--NHS--LC--Biotin reagent (Pierce Biotechnology, Rockford, Illinois, United States), following the manufacturer\'s instruction. Excess biotinylation reagent was removed by prolonged incubation in Tris--HCl buffer (1 M \[pH 7.5\]) followed by protein purification through a Polyacrylamide 1800 desalting column (Pierce Biotechnology). A control neuropeptide, allatostatin I (Sigma-Aldrich, St. Louis, Missouri, United States), was biotinylated using the same method. A second control was a synthetic, biotinylated peptide derived from the Drosophila PER protein (a gift from P. Nawathean). In addition, a new N-terminus biotinylated PDF peptide was chemically synthesized de novo (Sigma-Aldrich). Identical results were obtained with the two PDF peptides, and no specific signal was obtained with the two control peptides. To detect the binding of the neuropeptide in the CNS of Drosophila, brains were dissected in PBS and fixed in 3.7% paraformaldehyde in PEM for 30 min. After they were rinsed in PBS plus 0.3% Triton and blocked using 1% FBS or BSA, biotinylated peptide was incubated with the brains at a final concentration of 0.2 μg/ml. The brains were washed thoroughly with TNT (0.1 M Tris--HCl \[pH 7.5\], 0.15 M NaCl, 0.05% Tween 20). The bound peptide was subsequently detected through the biotin label using streptavidin--HRP (NEN LifeScience, now Perkin-Elmer, Torrance, California, United States) and fluorescent tyramides (NEN LifeScience). A detailed protocol is provided as [Protocol S1](#sd001){ref-type="supplementary-material"}. For the competition assay, unlabeled peptide was added at a 200- to 5000-fold concentration increase in the blocking step; subsequent steps were as described above. Supporting Information {#s5} ====================== Figure S1 ::: {.caption} ###### Rescued Molecular Oscillations Persist during DD in the s-LN~v~s The "rescued" mutant *y w; pdf--GAL4;UAS--CYC*,*cyc^01^*/*cyc^01^* was released into DD after entrainment and assayed by *tim* whole-mount in situ hybridization on the fourth day of DD. A *Pdf* probe was used to label the LN~v~ group. Brains were taken at two opposite timepoints, CT3 (top panels) and CT15 (bottom panels). From left to right are *Pdf* (green, FITC labeled), *tim* (red, Cy3 labeled), and an image overlay. The lower arrows point to the s-LN~v~s and the upper arrows to l-LN~v~s. Whereas the l-LN~v~s show barely visible *tim* RNA oscillations under these conditions, the s-LN~v~s are obviously cycling. This difference suggests that the l-LN~v~s might damp more rapidly or be more light-dependent than the s-LN~v~s in this unusual genotype. (7.1 MB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Protocol S1 ::: {.caption} ###### Short Protocol for Neuropeptide Biotinylation and Receptor Detection (23 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: We thank our colleagues Joan Rutila for making the *UAS--CYC--HA* transgenetic flies; Jie Zhao for help with the whole-mount in situ hybridization; Patrick Emery and Mike McDonald for inspiration and helpful discussions; Paul Taghert, Orie Shafer, Ravi Allada, and Ralf Stanewsky for critical readings of the manuscript and exchanging unpublished results. We also thank Ed Dougherty and National Institutes of Health (NIH) grant S10 RR16780 for assistance in confocal microscopy and Heather Felton for administrative assistance. The work was supported in part by NIH grants GM33205 and NS44232 to MR and JCH. **Conflicts of Interest.** The authors have declared that no conflicts of interest exist. **Author Contributions.** YP, DS, JCH, and MR conceived and designed the experiments. YP, DS, and JDL performed the experiments. YP, DS, JDL, and MR analyzed the data. YP, DS, JDL, and MR contributed reagents/materials/analysis tools. YP, DS, and MR wrote the paper. Academic Editor: Ueli Schibler, University of Geneva. *clk* : *clock* *cry* : *cryptochrome* *cyc* : *cycle* CT : circadian time DD : constant darkness DN : dorsal neuron HA : hemagglutinin LD : light--dark l-LN~v~ : large ventral lateral neuron LN~d~ : dorsal lateral neuron LN~v~ : ventral lateral neuron PDF : pigment-dispersing factor *per* : *period* PG : prothoracic gland SCN : suprachiasmatic nucleus s-LN~v~ : small ventral lateral neuron *tim* : *timeless* ZT : Zeitgeber time.

PubMed Central

2024-06-05T03:55:45.933052

2003-9-15

PMC193605

Introduction {#s1} ============ The availability of complete sequences of genomes for clusters of related organisms presents the first opportunity to reconstruct events of genomic evolution. By comparing related genomes and inferring ancestral ones, we can identify events, such as specific chromosomal rearrangements, gene acquisitions, duplications, and deletions, that have produced the observed diversity in genome content and organization. The Bacteria offer the most immediate opportunities for such reconstruction, because many clusters of related genomes are now available and because the genomes are small and contain relatively little repetitive sequence, reducing computational complexity. Among bacterial groups, the γ-Proteobacteria presents the most intensively studied and sequenced cluster of genomes with varying degrees of relatedness. Intertwined with the problem of reconstructing genomic change is the problem of inferring phylogeny. Evading this issue is particularly difficult in the Bacteria. First, using complete genomes to obtain a robust phylogeny for all bacteria has presented problems due to the age of the group and the resulting loss of phylogenetic signal. Furthermore, lateral gene transfer (LGT) occurs in bacteria and has been claimed to be rampant for all classes of genes, potentially resulting in a diversity of phylogenetic histories across genes and complicating, or completely defeating, attempts to reconstruct bacterial evolution at both deep and more recent evolutionary depths ([@pbio-0000019-Doolittle1]; [@pbio-0000019-Nesbo1]; [@pbio-0000019-Gogarten1]; [@pbio-0000019-Wolf1]; [@pbio-0000019-Zhaxybayeva1]). Although the existence of substantial levels of LGT in bacterial genomes is not disputed, the existence of a core of genes resistant to LGT has been proposed ([@pbio-0000019-Jain1]) and has received some support from recent studies using relatively intensive taxon sampling ([@pbio-0000019-Brochier2]; [@pbio-0000019-Daubin1]). For purposes of reconstructing genomic change, what we seek is the organismal phylogeny---that is, the topology that traces the history of the replicating cell lineages that transmit genes and genomes to successive generations. The organismal phylogeny provides the backdrop against which events of genomic change, including LGT, have occurred. High incidence of LGT may cause the organismal phylogeny to be elusive, because we do not know which genes represent the true history of the cell lineages. The gene most used for reconstructing organismal phylogeny is the small subunit ribosomal RNA (SSU rRNA), which has been argued to rarely undergo transfer among genomes ([@pbio-0000019-Woese1]; [@pbio-0000019-Jain1]). But even this gene may undergo occasional LGT or recombination ([@pbio-0000019-Ueda1]; [@pbio-0000019-Yap1]). Furthermore, by itself, it provides insufficient information to resolve phylogenies, particularly for cases of heterogeneous rates and patterns of substitution. Thus, building conclusions about organismal phylogeny on the basis of SSU rRNA alone is unsatisfactory. The availability of complete genome sequences presents us with the potential to exploit the much greater set of genes that are expected to share the same history of transmission along the branches of the organismal phylogeny. A robust phylogeny based on more sequences could then be used to reconstruct genome-scale events, including LGT and rearrangements. But, while complete genome sequences have enormous potential for addressing phylogenetic issues, their utility for reconstructing bacterial phylogeny is initially quite limited due to the requirement of thorough taxon sampling within a clade for accurate reconstruction of phylogenies ([@pbio-0000019-Zwickl1]; [@pbio-0000019-Hillis1]). Only now, with the continuing increase in numbers of fully sequenced bacterial species, is it becoming possible to obtain sufficiently dense taxon sampling to exploit the large amount of genomic sequence data for the purpose of phylogeny reconstruction. We have chosen one group of Bacteria, the γ-Proteobacteria, to address the problem of whether complete genome sequences can be used for robust reconstruction of the organismal phylogeny, despite high levels of LGT. The γ-Proteobacteria, distinguished on the basis of sequence signatures and structural differences in the SSU rRNA ([@pbio-0000019-Woese1]), is an ideal choice for this purpose. This group represents a model of bacterial diversification and includes free-living and commensal species, intracellular symbionts, and plant and animal pathogens. The sequence divergence of certain of its members ([@pbio-0000019-Clark1]) suggests an age of at least 500 million years. At the same time, members are sufficiently closely related to enable us to reduce the problem of lack of phylogenetic signal and to identify a large set of unambiguous orthologs. Currently, the γ-Proteobacteria contains the highest density of fully sequenced genomes, including those of species (Escherichia coli and Salmonella sp.) for which knowledge of gene function is more complete than for any other cellular organisms. The potential obstacles to phylogenetic inference that are found across the Bacteria are certainly present in the γ-Proteobacteria. In particular, LGT is known to be extensive in this group, based on studies of genome composition ([@pbio-0000019-Lawrence1]; [@pbio-0000019-Parkhill1], [@pbio-0000019-Parkhill2]; [@pbio-0000019-Stover1]). Symbiotic lineages present particular issues for phylogeny reconstruction owing to huge losses of genes ([@pbio-0000019-Shigenobu1]; [@pbio-0000019-Akman1]), accelerated sequence evolution, and shifts in base composition ([@pbio-0000019-Moran1]). These features create phylogenetic artifacts and make the use of additional data from genome sequences particularly desirable. Here we aim to use complete genome sequences to reconstruct the organismal phylogeny for the γ-Proteobacteria by first selecting a set of probable ortholog families and then determining whether most agree on a common topology. A major implication of our results is that the replacement of single-copy orthologous genes is extremely rare, even within phyla. Instead, LGT most often involves uptake of genes assuming functions that are not represented in the recipient and arriving from distantly related bacteria or from phage ([@pbio-0000019-Daubin2]; [@pbio-0000019-Pedulla1]). A consequence is that most single-copy orthologous genes show broad phylogenetic agreement that reflects the organismal relationships and that provides a foundation for reconstructing events of genome evolution. Results {#s2} ======= Gene Families and Identification of Orthologous Genes {#s2a} ----------------------------------------------------- The proteins of 13 complete γ-proteobacterial genomes were classified into an initial set of 14,158 homolog families, using the procedures described in [Materials and Methods](#s4){ref-type="sec"}. [Figure 1](#pbio-0000019-g001){ref-type="fig"}A shows the distribution of the number of genes per family. A majority (7,655) of the families contain only one gene. As the criteria we applied for grouping genes into families are stringent, this number is expected to exceed the number of real orphan genes; indeed, annotations for many of these genes do claim homology with other genes in the included genomes. As a result, values for most of the genomes ([Figure 1](#pbio-0000019-g001){ref-type="fig"}B) are higher than the number of genes annotated as orphans; for example, the number of this type of gene identified for Buchnera was 24, but the annotation indicates only four genes unique to this species ([@pbio-0000019-Shigenobu1]). Moreover, our comparison was made only within this group of 13 bacteria, and some single-gene families may have homologs in other, more distant bacterial species. Pseudomonas aeruginosa yielded the highest number of unique genes, which represent nearly 41% of proteins of this genome. This is congruent with the result obtained during the original annotation of this genome: the authors were unable to identify relatives for 32% of the ORFs ([@pbio-0000019-Stover1]). ::: {#pbio-0000019-g001 .fig} Figure 1 ::: {.caption} ###### Number of Genes and Species within the Gene Families \(A) Distribution of the number of genes contained in the homolog families. \(B) Number of orphan genes in each species in parentheses. Abbreviations: Ba, Buchnera aphidicola; Ec, Escherichia coli; Hi, Haemophilus influenzae; Pa, Pseudomonas aeruginosa; Pm, Pasteurella multocida; St, Salmonella typhimurium; Vc, Vibrio cholerae; Wb, Wigglesworthia brevipalpis; Xa, Xanthomonas axonopodis; Xc, Xanthomonas campestris; Xf, Xylella fastidiosa; Yp CO92, Yersinia pestis CO\_92; Yp KIM, Yersinia pestis KIM. \(C) Distribution of the number of species contained in the homolog families. :::  ::: At the other extreme, some families group large numbers of genes. The largest family contains 544 genes and corresponds to the ABC transporter family, known to be the largest protein family ([@pbio-0000019-Tatusov1]; [@pbio-0000019-Tomii1]). The second-largest family, with 404 genes, corresponds to the histidine kinase response regulators ([@pbio-0000019-Wolanin1]). [Figure 1](#pbio-0000019-g001){ref-type="fig"}C shows the distribution of number of species per gene family. Note that a large majority of families group only one or two species (8,035 and 2,693 families, respectively). In the families comprising only one species, Pseudomonas and Vibrio are heavily represented, with 2,397 and 1,474 families, respectively. The families containing two species often group two closely related genomes, such as the two Xanthomonas, the two Yersinia, Escherichia and Salmonella, and Haemophilus and Pasteurella. A total of 275 families are represented in all 13 species. Among these, 205 contain exactly one gene per species. We consider these 205 genes to represent likely orthologs and, consequently, to be good candidates for use in inferring the organismal phylogeny and the extent of LGT. The Extent of Conflict among Gene Families {#s2b} ------------------------------------------ We constructed trees based on several combinations of data and methods (see [Materials and Methods](#s4){ref-type="sec"}), with the aim of generating a set of candidate topologies for the organismal phylogeny. These seven analyses produced a total of six topologies (numbered 1--6 in [Figure 2](#pbio-0000019-g002){ref-type="fig"}). (The identical topology was obtained for the consensus tree and the tree based on the protein concatenation using the neighbor-joining \[NJ\] method and the γ-based method for correcting the rate heterogeneity among sites.) The trees differ, in particular, with regard to the positions of Wigglesworthia, Buchnera, and Vibrio. All topologies, except number 4 (that one resulting from the Galtier and Gouy distance method with the SSU rRNA), tend to place Wigglesworthia and Buchnera as sister taxa ([Figure 2](#pbio-0000019-g002){ref-type="fig"}). The sister relationship of Wigglesworthia and Buchnera was of particular interest because it would suggest a shared origin of symbiosis in an ancestor of these two species. Thus, we tested seven additional topologies (numbered 7--13 on [Figure 2](#pbio-0000019-g002){ref-type="fig"}) that did not place these two species as sister taxa, but that otherwise resembled topologies 1--6. ::: {#pbio-0000019-g002 .fig} Figure 2 ::: {.caption} ###### The 13 Candidate Topologies Topologies 1--4 correspond to tree reconstructions based on SSU rRNA. Topologies 5 and 6 correspond to the trees based on the concatenation of the proteins. Topologies 7--13 correspond to additional topologies constructed to test the sister relationship of the two symbiont species. Species abbreviations as in [Figure 1](#pbio-0000019-g001){ref-type="fig"}. Abbreviations: ML, maximum likelihood; NJ, neighbor joining; K, Kimura distance; G&G, Galtier and Gouy distance; γ, gamma-based method for correcting the rate heterogeneity among sites. The position of the root corresponds to the one obtained repeatedly using SSU rRNA. :::  ::: For each alignment, we tested the likelihood of the 13 topologies against the maximum-likelihood (ML) topology, using the Shimodaira--Hasegawa (SH) test, as recommended by [@pbio-0000019-Goldman1]). The question asked here was whether the tested topologies could be considered equally good explanations of the data. [Figure 3](#pbio-0000019-g003){ref-type="fig"} shows the result of this test. One topology (number 5) is in agreement with 203 of 205 alignments ([Figure 3](#pbio-0000019-g003){ref-type="fig"}). Three other slightly different topologies can be considered nearly as good on the basis of agreement with a large majority of alignments (for topologies 2, 3, and 6 agreement was with 197, 196, and 186 alignments, respectively; [Figure 3](#pbio-0000019-g003){ref-type="fig"}). ::: {#pbio-0000019-g003 .fig} Figure 3 ::: {.caption} ###### Result of the SH Test The graph shows the number of alignments accepting or rejecting each topology. The "Other Topologies" are those built to test the sister relationship of Wigglesworthia and Buchnera. The "Proteins" topologies are those obtained using both the protein concatenation and the consensus of trees from all 205 alignments. The "SSU rRNA" topologies were obtained using the SSU rRNA sequences with different methods. :::  ::: Cases of Lateral Transfer {#s2c} ------------------------- Of the 205 alignments, two were found to strongly reject the most accepted tree (topology 5) as well as all other topologies tested. We have investigated these two genes to determine whether the incongruence is likely to result from LGT. The proteins correspond to biotin synthase (BioB) and to the virulence factor MviN. The trees obtained using ML for each alignment are shown in [Figure 4](#pbio-0000019-g004){ref-type="fig"}A. In both cases, the position of Pseudomonas conflicts with all widely supported topologies; it is placed as sister-group to Vibrio (BioB) or as sister-group to the enterics Escherichia, Salmonella, and Yersinia (MviN). Although initial examination of the topologies obtained from these genes suggests more than a single LGT (comparing trees of [Figure 4](#pbio-0000019-g004){ref-type="fig"}A to topology 5 of [Figure 2](#pbio-0000019-g002){ref-type="fig"}), the hypothesis of a single transfer in an ancestor of Pseudomonas could not be rejected for either gene based on the results of the SH test after removal of Pseudomonas from the alignments and from topology 5 and other widely supported phylogenies. The implication is that a single transfer event in an ancestor of Pseudomonas is sufficient to explain the conflict of *bioB* and *mviN* with trees derived from other genes. In addition, we searched GenBank for homologous genes in other species of Pseudomonas and built trees using NJ and the Poisson correction ([Figure 4](#pbio-0000019-g004){ref-type="fig"}B). In each case, Pseudomonas species are grouped and display the same position as in the trees, based only on the 13 sequenced genomes ([Figure 4](#pbio-0000019-g004){ref-type="fig"}A). Moreover, the bootstrap support was high for the grouping of Pseudomonas with Vibrio in the BioB tree and for the grouping of Pseudomonas with the enteric bacteria in the MviN tree. Thus, the phylogeny of each of these two genes can be explained as the result of a single LGT event, from different donors within the γ-Proteobacteria to a shared ancestor of these Pseudomonas species. ::: {#pbio-0000019-g004 .fig} Figure 4 ::: {.caption} ###### Phylogenies for the Laterally Transferred Genes \(A) ML trees obtained for BioB (left) and MviN (right). (B) NJ trees obtained for BioB (left) and MviN (right). Abbreviations: Pf, Pseudomonas fluorescens; Pp, Pseudomonas putida; Ps, Pseudomonas syringae. Other species abbreviations as in [Figure 1](#pbio-0000019-g001){ref-type="fig"}. :::  ::: In Escherichia, Vibrio, Salmonella, Yersinia, and Pseudomonas, *bioB* is flanked by *bioF*, also involved in biotin biosynthesis. To determine whether *bioF* could have been transferred with the *bioB* gene, we built a tree based on the protein translation of *bioF* using all species except Buchnera, Haemophilus, and Pasteurella, which lack this gene. The tree obtained did not show any unexpected position of Pseudomonas, indicating that only *bioB* has been horizontally transmitted. A possible explanation, consistent with the flanking position of *bioB* and *bioF* in the Pseudomonas genome, is that the original *bioB* gene was replaced through homologous recombination in a common ancestor of the included Pseudomonas species. Similar comparisons for *mviN* did not illuminate its history in Pseudomonas, as the flanking genes differed from those in other species. The observations for *mviN* are consistent with a transfer event from an enteric species to a new genomic position in a Pseudomonas ancestor. Robustness of the Inferred Organismal Phylogeny {#s2d} ----------------------------------------------- The general lack of conflict observed among the 203 remaining families was not due to the absence of phylogenetic signal in the gene alignments because most genes did conflict with several other topologies (see [Figure 3](#pbio-0000019-g003){ref-type="fig"}). We interpreted this congruence as a reflection of shared history and a lack of LGT. Therefore, we chose these genes as the basis for inferring the true organismal phylogeny for these 13 species. The resulting tree was the same as that for the concatenation of all of the 205 genes and for the consensus of the trees obtained from all protein families (topology 5 in [Figure 2](#pbio-0000019-g002){ref-type="fig"} and tree presented in [Figure 5](#pbio-0000019-g005){ref-type="fig"}). It differed only slightly from other tested topologies (see [Figure 2](#pbio-0000019-g002){ref-type="fig"}) that also are not rejected by many individual alignments (see [Figure 3](#pbio-0000019-g003){ref-type="fig"}). Finally, an SH test performed using the complete concatenated alignment shows that this topology is significantly more likely than all alternative hypotheses. ::: {#pbio-0000019-g005 .fig} Figure 5 ::: {.caption} ###### Tree Based on the Concatenation of the 205 Proteins (NJ) The topology shown agrees with almost all individual gene alignments (topology 5 of [Figure 2](#pbio-0000019-g002){ref-type="fig"}). The same tree is obtained after removing the two genes showing evidence for LGT. The position of the root corresponds to the one obtained repeatedly using SSU rRNA. :::  ::: All topologies separating Wigglesworthia and Buchnera were rejected by the majority of the alignments. In the best-supported topology (see [Figure 5](#pbio-0000019-g005){ref-type="fig"}), Wigglesworthia and Buchnera are grouped and comprise the sister-group to the enteric bacteria Yersinia, Salmonella, and Escherichia. Previously published phylogenies, based on SSU rRNA, gave conflicting results for the positions of these symbionts, sometimes placing Buchnera as sister-group to Escherichia and Salmonella ([@pbio-0000019-van1]; [@pbio-0000019-Spaulding1]; [@pbio-0000019-Moya1]). Because the genomes of these endosymbionts present a strong bias toward A+T relative to other genomes in the set, their grouping could reflect convergence at some nucleotide sites. This convergence could affect both the SSU rRNA, which is enriched in A+T ([@pbio-0000019-Moran1]), and also the protein sequences, which are enriched in amino acids with A+T-rich codon families ([@pbio-0000019-Clark1]). To test this hypothesis, we removed from the alignment of the protein concatenation all sites at which Buchnera and Wigglesworthia contain amino acids encoded by A+T-rich codons (phenylalanine, tyrosine, methione, isoleucine, asparagine, and lysine) ([@pbio-0000019-Singer1]). Using the resulting alignment (about 30,000 residues), we have reconstructed two trees, one with the NJ method and the polyacid-modified (PAM) matrix; the other with the NJ method and the γ-based method for correcting the rate heterogeneity among sites and bootstrap. The trees obtained (data not shown) were identical and gave strong support to the grouping of Buchnera and Wigglesworthia and to their position as the sister-group of enteric bacteria (Escherichia, Salmonella, and Yersinia). Thus, this grouping is probably not an artifact of the biased composition of the endosymbiont genomes. Discussion {#s3} ========== The most striking result is the almost complete lack of conflict among the set of genes selected as likely orthologs. Only two of 205 ortholog families showed such disagreement, both involving the P. aeruginosa genome. Because the γ-Proteobacteria has been the bacterial group most often cited as showing high rates of LGT, this finding is unexpected. However, we note that the evidence for LGT from sequence features and comparisons of genome content ([@pbio-0000019-Lawrence1]; [@pbio-0000019-Ochman1]; [@pbio-0000019-Parkhill2]; [@pbio-0000019-Perna1]; [@pbio-0000019-Daubin2]) primarily implicate genes that are absent from related bacteria; such genes would not have been retained in our set of putative orthologs. Furthermore, such genes are not candidates for phylogeny reconstruction since they are missing from most taxa. We also eliminated the large set of homolog families present as more than one sequence within even one of the genomes. If families containing paralogs show relatively high susceptibility to LGT, the proportion of genes undergoing LGT would be underestimated by considering only the set with one sequence per genome. Our aim was to locate a set of genes giving strong and consistent signal regarding the organismal phylogeny, and our results do not imply a lack of LGT in genes other than the widespread, single-copy orthologs that we selected. By streamlining the dataset for our primary goal, we have excluded genes that undergo more frequent transfer. Phylogenetic evidence for LGT mostly involves transfer between distantly related organisms ([@pbio-0000019-Nelson1]; [@pbio-0000019-Brown2]; [@pbio-0000019-Brown1]; [@pbio-0000019-Xie1]), and most clear-cut cases involve genes that are sporadically distributed (e.g., [@pbio-0000019-Parkhill1]; [@pbio-0000019-Singer1]) and thus excluded from our selection of families. The selected set includes genes that are distributed across a wide set of bacteria and includes about 100 universally distributed genes, such as those encoding ribosomal proteins, DNA polymerase subunits, and transfer RNA synthetases ([Table S1](#st001){ref-type="supplementary-material"}). Thus, if LGT is affecting most categories of genes, it should be detectable in our set, resulting in discordance of phylogenies whether it occurs between related genomes (within the γ-Proteobacteria) or between very dissimilar genomes. Such discordance was extremely rare, affecting only two (1%) of our families. Several previous studies have provided evidence that a core of genes may resist LGT and give a consistent phylogenetic signal ([@pbio-0000019-Jain1]; [@pbio-0000019-Brochier2]; [@pbio-0000019-Daubin1]). However, the same studies have noted high incidence of genes showing incongruence, and, because they involved deeper trees and incorporated a much less dense sampling of genes or of taxa, this incongruence has not been firmly identified as due to LGT or to phylogenetic artifacts. Furthermore, recent analyses based on other sets of taxa have led to the proposal that all sets of genes, including orthologous genes, are subject to high rates of LGT ([@pbio-0000019-Nesbo1]; [@pbio-0000019-Gogarten1]; [@pbio-0000019-Zhaxybayeva1]), thereby casting doubt on the idea that we can identify a core set of orthologs that reflect the organismal phylogeny. Our analysis indicates that LGT is unusual for single-copy orthologous genes; that is, a gene copy from one species usually does not replace its ortholog in another species. The apparent discrepancy is not due to a relative lack of LGT in this particular group of bacteria, which is known to acquire foreign genes frequently ([@pbio-0000019-Lawrence1]; [@pbio-0000019-Parkhill1]; [@pbio-0000019-Perna1]). More likely explanations are that (1) our criterion for orthology was more stringent in ruling out undetected paralogy; (2) the use of quartet phylogenies ([@pbio-0000019-Zhaxybayeva1]) can be misleading owing to artifacts linked to taxon sampling ([@pbio-0000019-Zwickl1]); and (3) our focus on a relatively closely related group of bacteria minimizes the problem of loss of phylogenetic signal and reconstruction artifacts in deep divergences. This result thus provides further evidence that, though bacterial genomes constantly acquire and lose significant amounts of DNA, the incidence of LGT for widespread orthologous genes is relatively low ([@pbio-0000019-Daubin3]). Although we have likely excluded many actual orthologs, the set of retained genes provides a dataset that is sufficiently informative to give a highly resolved and well-supported phylogeny for these taxa. This study thus defines a minimal core of genes that show both wide representation and congruent phylogenetic signal in γ-Proteobacteria. We note that this core includes numerous genes in both "informational" and "operational" functional categories ([Table S1](#st001){ref-type="supplementary-material"}); thus, our results do not fit closely with the "complexity hypothesis," that only informational genes avoid LGT ([@pbio-0000019-Jain1]), although they do not exclude such a trend. Our set of 203 genes should not be considered as representative of all genes resisting LGT, since we did not explore the other gene families. The main functional feature distinguishing the set is likely to be essentiality, owing to the requirement of presence in all 13 genomes, including the reduced symbiont genomes. For the goal of selecting genes that reflect organismal phylogeny through vertical descent, our criteria (single copy and ubiquitous) appear to be more reliable than criteria based on functional information (informational genes, translational genes, etc.). Indeed, cases of LGT are known for informational genes (e.g., [@pbio-0000019-Brochier1]). One possible explanation for the lack of observed events of orthologous replacement might be that these are sufficiently rare that significant frequencies are encountered only when considering deeper phylogenetic levels. However, the group studied here, though recent enough to allow accurate phylogenetic reconstruction, is old. Indeed, the divergence of different Buchnera species has been dated to approximately 200 million years based on the host fossil record ([@pbio-0000019-Clark1]), and the clade we have studied must be much more ancient. A conservative molecular clock estimate, based on rRNA and dating the divergence of Escherichia and Buchnera at 200 million years, places the origin of the group at more than 500 million years (calculations not shown). Thus, our finding that very few orthologs have been exchanged within the group and that none show evidence of having been imported from other bacterial lineages is relevant for the understanding of long-term bacterial evolution. It has been proposed that LGT may be more frequent within clusters of related bacteria and even that phylogenetic groupings, such as the γ-Proteobacteria, may reflect boundaries to LGT rather than recent shared ancestry of lineages ([@pbio-0000019-Gogarten1]). Such a model, which is consistent with apparent concordance among ortholog families in studies with poor taxon sampling but predicts rampant discordance within a well-sampled bacterial cluster, is strongly rejected by our results. Our findings favor the view that the cohesion of major phylogenetic groups within the Bacteria is due to vertical transmission and common ancestry rather than to preferential lateral transfer of genes. However, the results presented here do not eliminate the possibility of nonrandom patterns of LGT for gene families that are more sporadically distributed. A robust phylogenetic framework for the organismal lineages provides the foundation for reconstructing the events of genome evolution. An example of the kind of biological inference that can be built upon a well-supported phylogeny is provided by the two endosymbionts included in our set. Wigglesworthia and Buchnera have sometimes been considered as closely related and sometimes not, based on relatively weak phylogenetic evidence provided by the SSU rRNA alone. Our confirmation of their close relationship raises the question of whether their common ancestor was an endosymbiont with a reduced genome or a free-living bacterium (perhaps one with a host-associated lifestyle that promoted formation of intimate symbiosis). Because Buchnera and Wigglesworthia do not share any genes absent from the other species, no particular genes can be implicated as conferring a predisposition to symbiosis, a result that eliminates some hypotheses about how symbiosis originates. Furthermore, although emphasis has previously been placed on the close relationship of Buchnera with E. coli, our results shows that the phylogenetic relationship is equally high with other enterics, such as Yersinia pestis, which indeed shares as many ortholog families with Buchnera as does E. coli. This knowledge of relationships to other genomes allows more accurate reconstruction of ancestral genome content and of the chromosomal deletions and rearrangements occurring during the evolution of reduced symbiotic genomes ([@pbio-0000019-Moran2]). One biological interpretation of our findings is that the immediate retention of an acquired gene within a lineage depends upon strong positive selection for its function ([@pbio-0000019-Ochman2]) and that such selection is unlikely if a homologous gene is already present in the recipient genome. An implication, from the perspective of phylogeny reconstruction, is that single-copy homologs with widespread distribution are a source of reliable information for inferring organismal phylogeny. The existence of many other gene families with multiple members per genome or with erratic distributions across the set of genomes (see [Figure 1](#pbio-0000019-g001){ref-type="fig"}) is consistent with a major role of LGT, gene loss, and gene duplication in the evolution of this bacterial clade. Combined with chromosomal rearrangements, these events are the major sources of genomic, and ultimately ecological, diversification of bacterial groups. By demonstrating the potential to establish robust organismal phylogenies using genome sequence data, our results provide a foundation for examining the rates and frequencies of LGT and other large-scale events in evolving genomes. Materials and Methods {#s4} ===================== {#s4a} ### Data. {#s4a1} The genomes chosen for this study correspond to 13 γ-Proteobacterial taxa that show different degrees of relatedness based on divergence of SSU rRNA and that include two symbionts having undergone large-scale genomic reduction ([@pbio-0000019-Shigenobu1]; [@pbio-0000019-Akman1]). The protein sequences of the 13 complete genomes were retrieved from the GenBank database ([@pbio-0000019-Benson1]). The species used were Escherichia coli K12 (accession number NC\_000913; [@pbio-0000019-Blattner1]), Buchnera aphidicola APS (NC\_002528; [@pbio-0000019-Shigenobu1]), Haemophilus influenzae Rd (NC\_000907; [@pbio-0000019-Fleischmann1]), Pasteurella multocida Pm70 (NC\_002663; [@pbio-0000019-May1]), Salmonella typhimurium LT2 (NC\_003197; [@pbio-0000019-McClelland1]), Yersinia pestis CO\_92 (NC\_003143; [@pbio-0000019-Parkhill1]), Yersinia pestis KIM5 P12 (NC\_004088; [@pbio-0000019-Deng1]), Vibrio cholerae (NC\_002505 for chromosome 1 and NC\_002506 for chromosome 2; [@pbio-0000019-Heidelberg1]), Xanthomonas axonopodis pv. citri 306 (NC\_003919; [@pbio-0000019-da1]), Xanthomonas campestris (NC\_003902; [@pbio-0000019-da1]), Xylella fastidiosa 9a5c (NC\_002488; [@pbio-0000019-Simpson1]), Pseudomonas aeruginosa PA01 (NC\_002516; [@pbio-0000019-Stover1]), and Wigglesworthia glossinidia brevipalpis (NC\_004344; [@pbio-0000019-Akman1]). To identify genes likely to have been transmitted vertically through the history of the γ-Proteobacteria, we first eliminated proteins annotated as elements of insertion sequences or as bacteriophage sequences, since they are likely to be subject to lateral transfer. Such sequences were present in most genomes but lacking in a few (B. aphidicola, W. brevipalpis, and P. multocida). [Table 1](#pbio-0000019-t001){ref-type="table"} shows the number of proteins that remain in each genome after such elimination. ::: {#pbio-0000019-t001 .table-wrap} Table 1 ::: {.caption} ###### Number of Protein-Coding Genes per Genome after Elimination of the Insertion and Bacteriophage Sequences :::  ^a^ Chromosome 1 ^b^ Chromosome 2 ::: ### Construction of the gene families. {#s4a2} We applied a stringent criterion for eliminating nonhomologous sequences and paralogous sequences, since both are likely to lead to false conclusions regarding the organismal phylogeny and frequency of LGT. In particular, the criterion of "best reciprocal hits" between sequences for a genome pair can lead to false conclusions of orthology because the resulting gene pairs are not always closest relatives phylogenetically ([@pbio-0000019-Koski1]). Instead, we used a cutoff for the degree of similarity as reflected in the BLASTP bit scores ([@pbio-0000019-Altschul1]). The bit score is dependent upon the scoring system (substitution matrix and gap costs) employed and takes into account both the degree of similarity and the length of the alignment between the query and the match sequences. We used it to detect homologous genes, described as follows. A bank of all annotated protein sequences of all included species was created. A BLASTP ([@pbio-0000019-Altschul1]) search was performed for all the proteins in each genome against the protein bank. This implies that all proteins were searched against both their resident genome and those from the 12 other species. The match of a given protein against itself gives a maximal bit score. To determine a threshold to group genes into a family, we examined the distribution of the ratio of the bit score to the maximal (self) bit score based on the proteins of E. coli compared against proteins of the 12 genomes ([Figure 6](#pbio-0000019-g006){ref-type="fig"}). In each case, the distribution showed a clear bimodal pattern with a first peak of low similarity values, which is constant among comparisons and therefore probably represents random matches, and a second peak of higher similarity values, representing true homologous genes. For comparisons of E. coli proteins with those of the most distant species in our set, such as Vibrio, Xanthomonas, Xylella, and Pseudomonas, the separation of the two portions of the distribution occurs at about 30% of the maximal bit score. Thus, in order to apply a stringent criterion for homology, we inferred as homologous genes those presenting a bit score value higher or equal to 30% of the maximal bit score. A protein was included in a family if this criterion was satisfied for at least one member. Our cutoff was chosen to minimize inclusion of nonhomologous sequences within a family; consequently, it may exclude some homologs, especially fast-evolving ones. ::: {#pbio-0000019-g006 .fig} Figure 6 ::: {.caption} ###### Similarity Levels for Pairwise Comparisons of Genes from Two Representative Genome Pairs Frequency distribution of the ratio (bit score/maximal bit score) in a BLASTP query of the proteins from E. coli on the proteins from the genomes of Salmonella enterica (solid line) and Vibrio cholerae (dashed line). The ratio of 0.3 allows identification of most homologs but excludes probable nonspecific matches (NS). :::  ::: After establishing homolog families, we selected the set that contained a single sequence in each represented genome and regarded these as likely orthologs that could give information about the organismal phylogeny and the frequency of LGT affecting orthologs in this bacterial group. ### Phylogenies. {#s4a3} The alignments for each identified gene family were created using the CLUSTALW 1.8 program ([@pbio-0000019-Thompson1]). We corrected the concatenated proteins alignment by removing ambiguous parts using the SEAVIEW sequence editor ([@pbio-0000019-Galtier1]). The TREE-PUZZLE 5.1 program ([@pbio-0000019-Schmidt1]) was used in order to determine the α parameter from the datasets for the γ-based method for correcting the rate heterogeneity among sites. We wished to generate a set of reasonable candidate topologies that could be tested against the alignments for individual genes. These topologies were generated based on the consensus of the 205 trees from individual protein families (one method, yielding topology 5 of [Figure 2](#pbio-0000019-g002){ref-type="fig"}), on the concatenation of all the proteins (over 75,000 amino acids) (two methods, yielding topologies 5 and 6), and on the SSU rRNA (four methods, yielding topologies 1--4). In the case of the reconstruction of the trees based on the SSU rRNA, we used the DNAML module of the PHYLIP package version 3.6 ([@pbio-0000019-Felsenstein1]), which performs ML reconstruction using the γ-based method for correcting the rate heterogeneity among sites; the PHYLO\_WIN program ([@pbio-0000019-Galtier1]) using the NJ method with bootstrap and using two different distances, Kimura 2P distance and Galtier and Gouy distance, designed to reduce bias due to base composition ([@pbio-0000019-Galtier1]); and the MEGA program ([@pbio-0000019-Kumar1]) using the NJ method with bootstrap and with the γ-based method for correcting the rate heterogeneity among sites. We used the PROML module of the PHYLIP package version 3.6 ([@pbio-0000019-Felsenstein1]) to conduct a ML reconstruction using the Jones, Taylor, and Thornton (JTT) model of substitution ([@pbio-0000019-Jones1]) and the γ-based method for correcting the rate heterogeneity among sites, on each of the 205 families of single-copy, orthologous proteins. The consensus of the trees of the 205 protein alignments was obtained using the CONSENSE module of the PHYLIP package version 3.6 ([@pbio-0000019-Felsenstein1]). As there are no missing data, we also concatenated all the proteins and used the PHYLO\_WIN program ([@pbio-0000019-Galtier1]), using the NJ method and the PAM distance matrix, and the MEGA program ([@pbio-0000019-Kumar1]), using the NJ method with bootstrap and with the γ-based method for correcting the rate heterogeneity among sites, on the protein concatenation. For each of the 205 alignments, a comparison of the likelihood of the best topology with the likelihood of the candidate topologies shown in [Figure 2](#pbio-0000019-g002){ref-type="fig"} were performed with the SH test ([@pbio-0000019-Shimodaira1]) implemented in TREE-PUZZLE 5.1 ([@pbio-0000019-Schmidt1]). This test determines whether these potential organismal phylogenies are significantly rejected by the alignment and thus whether an event of LGT must be invoked. Finally, we used the SH test to determine whether more than one LGT event was required to explain the lack of congruence between the favored topology and two gene alignments that rejected that topology. For each case, we observed which taxon showed the most evident discordance in the topology derived from the exceptional gene alignment. We then removed the corresponding sequence from the alignment and the corresponding taxon from the widely favored topologies. Using an SH test, we determined whether the alignment continued to show significant conflict with the favored topologies. Supporting Information {#s5} ====================== Table S1 ::: {.caption} ###### Names and Functional Categories of the 205 Genes Used to Reconstruct the Phylogenetical Relationship of γ-Proteobacteria (123 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Accession Numbers {#s5a} ----------------- The GenBank accession numbers discussed in this paper are NC\_000907, NC\_000913, NC\_002488, NC\_002505, NC\_002506, NC\_002516, NC\_002528, NC\_002663, NC\_003143, NC\_003197, NC\_003902, NC\_003919, NC\_004088, and NC\_004344. We thank Howard Ochman for comments. Support came from National Science Foundation Biocomplexity grant number 9978518. **Conflicts of Interest.** The authors have declared that no conflicts of interest exist. **Author Contributions.** EL, VD, and NAM conceived and designed the experiments. EL analyzed the data. EL, VD, and NAM wrote the paper. Academic Editor: David Penny, Massey University. BioB : biotin synthase LGT : lateral gene transfer ML : maximum likelihood NJ : neighbor joining PAM : polyacid modified SH test : Shimodaira--Hasegawa test SSU rRNA : small subunit ribosomal RNA.

PubMed Central

2024-06-05T03:55:45.935792

2003-9-15

PMC193606

How do people subjected to the endless dark days of winter in the far northern latitudes maintain normal daily rhythms? Though many might feel like hibernating, a highly regulated internal system keeps such impractical yearnings in check. From fruit flies to humans, nearly every living organism depends on an internal clock to regulate basic biological cycles such as sleep patterns, metabolism, and body temperature. And that clock runs on similar molecular mechanisms. Specific clusters of neurons in the brain are known to control the biological clock. Scientists believed these brain "clock cells" function as independent units. But new research described in this issue shows that the neurons do not act in isolation; rather, they collaborate with other neurons in a cell-communication network to sustain the repeating circadian rhythm cycles. Clock cells within the brain maintain an organism\'s circadian rhythms, even in the absence of cyclical environmental signals like light, in a state scientists call "free running." Though it has long been clear that the circadian rhythms of an organism persist under such free-running conditions (for example, constant darkness), it was thought that the gene-expression patterns within the cells governing these biorhythms did not require any external, or extracellular, signals to continue ticking. In experiments described here, Michael Rosbash and his colleagues show that the key brain clock cells in fruit flies (*Drosophila*), called *ventral lateral neurons*, do indeed support the fly\'s circadian rhythms during periods of constant darkness and that the molecular expression patterns associated with these rhythms continue to cycle as well within other clock cells. These sustained expression patterns, however, require intercellular communication between different groups of brain clock cells. In other words, the ventral lateral neurons do not act alone. When the molecular clock machinery was manipulated so that only the ventral lateral neurons were active, the fly\'s circadian rhythms were not sustained, suggesting the rhythms depend on other neuronal groups as well. The researchers also demonstrate that the persistence of normal cycling during constant darkness depends on a protein (called PDF) secreted by the ventral lateral cells. The PDF neuropeptide protein was thought to connect the molecular expression pattern of the ventral lateral neurons with the manifestation of circadian rhythms, but the researchers found evidence of a larger influence. When mutant flies lacking a functional PDF gene were exposed to constant darkness, the molecular expression patterns gradually stopped. The scientists say this suggests that the ventral lateral neurons and the PDF protein it produces help coordinate the entire neural network that underlies circadian rhythms. ::: {#pbio-0000023-g001 .fig} ::: {.caption} ###### Drosophila lateral neuron (green) :::  :::

PubMed Central

2024-06-05T03:55:45.939512

2003-9-15

PMC193607

Bacteria are an indiscriminate lot. While most organisms tend to pass their genes on to the next generation of their species, bacteria often exchange genetic material with totally unrelated species. That is why skeptics doubted that bacteria researchers could ever hope to map a reliable history of cell lineages in bacteria over time. But now, thanks to the availability of sequenced genomes for groups of related bacteria, researchers at the University of Arizona demonstrate that constructing a bacterial family tree is indeed possible. Previous efforts to trace the ancestry of bacteria were constrained by a dearth of related bacterial genomes, which, among other things, prevented scientists from successfully accounting for bacteria\'s tendency to exchange genes with unrelated organisms. In this process, called *lateral gene transfer*, organisms acquire genetic material not from their ancestors, the most prevalent route, but from unrelated organisms. Lateral gene transfer greatly complicates the issue of who descended from whom, because two organisms could appear closely related based on the similarity of some genes but distantly related based on other genes. The problem is to determine which genes have been faithfully vertically transmitted---from parent cell to offspring---and thus reflect the history of the bacterial cell lineages. In this issue, Nancy Moran, Emmanuelle Lerat, and Vincent Daubin propose an approach that solves this problem by identifying a set of genes that serve as reliable indicators of the vertical transfer of bacterial cell lineages. This method has important implications for biologists studying the evolutionary history of organisms by establishing a foundation for charting the evolutionary events, such as lateral gene transfer, that shape the structure and substance of genomes. With this method, scientists can begin to understand how bacteria have evolved and how their genomes have changed. Bacteria promise to reveal the most information about genomic evolution, because so many clusters of related bacterial genomes have been sequenced---allowing for broad comparative analysis among species---and their genomes are small and relatively compact. In this study, the researchers chose the ancient bacteria group Proteobacteria, an ecologically diverse group (including Escherichia coli and Salmonella species) with the most documented cases of lateral gene transfer and the highest number of species with sequenced genomes. The researchers identified a set of likely single-copy orthologs (homologous genes that diverged due to the speciation of ancestral lineages) with widespread distribution in the different species of Proteobacteria that could be used to trace the history of the cell lineages. Surprisingly, they found that almost all of the 205 ortholog gene families they selected supported the same evolutionary branching pattern. Only two did not, which the researchers then investigated and found to be the result of lateral gene transfer. These results, the researchers say, support the ability of their method to reconstruct the important evolutionary events affecting genomes. By mapping out the evolutionary path of genetic information on a genomic level, their approach promises to elucidate not only the evolution of bacterial genomes but also the diversification of species. ::: {#pbio-0000031-g001 .fig} ::: {.caption} ###### Electron micrograph of Proteobacteria in eukaryotic cell :::  :::

PubMed Central

2024-06-05T03:55:45.940002

2003-9-15

PMC212319

pmc Background {#Sec1} ========== Small GTPases of the ADP-ribosylation factor (ARF) family play a major role in membrane trafficking in eukaryotic cells \[[@CR1]\]. ARF activation is facilitated by specific guanine nucleotide exchange factors (ARF-GEFs). Several ARF-GEFs have been identified, varying in size, structure and subcellular distribution \[[@CR2]--[@CR6]\]. Of particular interest in signaling events are the members of the cytohesin/ARNO family of ARF-GEFs. These proteins have been found to associate with the plasma membrane under certain conditions, and consist of three well-defined motifs: an N-terminal coiled-coil domain (CC domain), a central domain with homology to the yeast protein Sec7 (Sec7 domain), and a C-terminal pleckstrin homology domain (PH domain) (Fig. [1](#Fig1){ref-type="fig"}). The catalytic activity of ARNO for guanine nucleotide exchange is localized in the Sec7 domain and appears to be regulated through the interaction of the PH domain with phosphatidylinositol (PtdIns) (3,4,5)-P3 \[[@CR7], [@CR8]\], an intermediate in signaling cascades regulated by insulin and other agonists \[[@CR3]\].Figure 1**Schematic structure of ARNO constructs.** Full length of wild type ARNO and ΔPH-ARNO were subcloned either in pCMV-myc or pEGFP-C1. PH-ARNO and ΔCC-ARNO were subcloned in pEGFP-C1. CC-ARNO was subcloned in pEGFP-N1. Phospholipase D (PLD) catalyzes the hydrolysis of phosphatidylcholine (PC) to produce phosphatidic acid (PA). It is involved in a variety of signaling pathways and membrane traffic processes \[[@CR9], [@CR10]\]. Many hormones, neurotransmitters, and growth factors, including insulin, have been shown to induce the activation of PLD \[[@CR11], [@CR12]\]. Several factors are involved in the regulation of cellular PLD activity, such as Ca^2+^, protein kinase C, tyrosine kinases, and G proteins \[[@CR13]--[@CR17]\]. Among these, the members of the ARF and Rho families of GTPases appear to be the most potent physiological activators \[[@CR18]--[@CR24]\]. However, the mechanism of the activation of PLD by ARF and Rho has not yet been fully established. This study was designed to investigate the role of ARNO in the regulation of PLD activity by insulin in HIRcB cells, a Rat-1 fibroblast cell line that overexpresses human insulin receptors. The objectives were: 1) to test if insulin induces the translocation of wild type ARNO to the plasma membrane in transiently transfected HIRcB cells; 2) to determine whether ARNO translocation is accompanied by activation and subcellular translocation of ARF; 3) to explore if overexpression of wild type ARNO in HIRcB cells alters insulin-dependent PLD activity; and 4) to investigate the function of individual domains of ARNO in insulin-dependent PLD and ARF activation. Results {#Sec2} ======= Insulin--dependent binding of ARNO to cell membranes {#Sec3} ---------------------------------------------------- The translocation of ARNO and ARNO mutants to the membranes was studied in HIRcB cells using a digitonin permeabilization assay. For these experiments, HIRcB cells were transiently transfected with myc-tagged wild type ARNO and the following mutants: ΔPH-ARNO, PH-ARNO, ΔCC-ARNO, CC-ARNO, E156K-ARNO and R280D-ARNO. This assay is based on the formation of pores in the plasma membrane induced by digitonin to allow cytosolic proteins to leak out of treated cells upon centrifugation. Fig. [2](#Fig2){ref-type="fig"} shows that, after digitonin permeablization, a significant fraction of ARNO proteins leaked out of serum-starved HIRcB cells that transiently overexpressed the wild type ARNO and its mutants. Since these proteins were mostly recovered from the supernatant fractions, suggesting that wild type ARNO and the mutants tested are predominantly cytosolic in non-stimulated cells. In contrast, when digitonin permeablization was performed in the presence of insulin (100 nM), most of wt-ARNO, E156K-ARNO, and ΔCC-ARNO as well as a part of CC-ARNO were recovered from the particulate membrane fraction, suggesting that these ARNO proteins can be recruited to the membrane by insulin to various degrees. However, neither R280D-ARNO nor ΔPH-ARNO was recovered from the particulate fraction after insulin stimulation, suggesting that the translocation of ARNO to the membrane requires an intact PH-domain. It should be noted that, although the CC domain alone binds to the membranes under stimulation conditions, the degree of the binding is much less than that of wild type ARNO (Fig. [2](#Fig2){ref-type="fig"}). Surprisingly, a construct containing only the PH domain of ARNO could not be recruited to the membranes by insulin, indicating that the PH domain is essential but not sufficient for the translocation of ARNO.Figure 2**Insulin promotes the translocation of ARNO to cell membranes.** HIRcB cells were transfected with myc-wt-ARNO, myc-E156K-ARNO, myc-R280D-ARNO, myc-ΔPH-ARNO, EGFP-PH-ARNO, EGFP-ΔCC-ARNO, and CC-ARNO-EGFP. The cells were treated with/without (Control) 10 μM digitonin (Dig). Where indicated, 100 nM insulin, 1 mM ATP, and 100 μM GTPγS were present during permeablization reaction. Pellets and supernatants were separated by centrifugation and the presence of myc-ARNO and its mutants or ARNO-EGFP in each fraction was determined by immunoblotting. ARNO recruits ARF1 to the plasma membrane in an insulin-dependent manner {#Sec4} ------------------------------------------------------------------------ Since ARNO is an activation factor of ARF, we tested the hypothesis that agonist-dependent ARNO translocation facilitates the local binding of ARF proteins to the membrane. An initial set of real-time studies was done using HeLa cells that had been stably transfected with an ARF1-GFP construct \[[@CR25]\]. These cells were transfected with myc-ARNO, serum-starved overnight, and imaged with a confocal microscope equipped with a constant-temperature microperfusion incubator to maintain the temperature at 37°C. Time-lapse images were collected at 30-second intervals. A representative experiment was shown in Fig. [3A](#Fig3){ref-type="fig"}. Prior to insulin stimulation, ARF1-GFP protein was mostly cytosolic or bound to the Golgi apparatus, although a small amount of ARF-GFP was localized on the surface of the cells. Ten minutes after the insulin stimulation, most of the ARF1-GFP was found on the plasma membrane. Similar results were obtained with HIRcB cells co-transfected with ARNO-myc and ARF1-GFP (Fig. [3B](#Fig3){ref-type="fig"}). It should be noted that a significant accumulation of ARF1-GFP on the plasma membrane was not observed in the cells that had not been transfected with ARNO (not shown), or that had been transfected with the inactive mutant E156K-ARNO (Fig. [3B](#Fig3){ref-type="fig"}). Since the endogenous levels of ARNO in HeLa cells were so low that the protein could not be detected in Western blots, it is reasonable to assume that under physiological conditions only a very small fraction of ARF1 translocates to the plasma membrane in response to extracellular agonists.Figure 3**A. Real time image of the translocation of ARF1-GFP to the plasma membrane.** HeLa cells that had been stably transfected with ARF1-GFP were transiently transfected with myc-ARNO, serum starved overnight, and treated with 100 nM insulin. Images were collected every 30 seconds using a Molecular Dynamics 2001 confocal microscope. The time intervals that were indicated on the upper right hand corner of each panel represent the time after the addition of insulin. **B. The translocation of ARF1-GFP to the plasma membrane by the effects of insulin requires ARNO.** ARF1-GFP/HeLa cells were transfected with myc-ARNO, treated, fixed, and stained for myc-epitope as described in the Materials and Methods section. Images displaying ARF1-GFP (green) and myc-ARNO (red) were merged us ing Adobe Photoshop software. ARNO interacts directly with the insulin receptor {#Sec5} ------------------------------------------------- Our previous work has shown that the insulin receptor co-immunoprecipitates with ARF in an agonist-dependent manner \[[@CR23]\]. Furthermore, we have also shown that an ARF-GEF activity is associated with the insulin receptor and that this activity is not a function of the receptor itself \[[@CR23]\]. Given that many receptor tyrosine kinases form complexes with their target proteins, we tested the hypothesis that ARNO binds the insulin receptor. Figure [4](#Fig4){ref-type="fig"} shows that insulin receptors that were immunoprecipitated in the presence of insulin were associated with an ARF-GEF activity (Fig [4](#Fig4){ref-type="fig"} ●), and that the ARF-GEF activity that was co-immunoprecipitated with the insulin receptor was significantly increased in the cells that had been transiently transfected with myc-ARNO (Fig. [4](#Fig4){ref-type="fig"} ■). Insulin receptors that were immunoprecipitated in the absence of insulin did not accelerate the binding of GTPγS to the recombinant ARF1 as much as those obtained in the presence of insulin (Fig. [4](#Fig4){ref-type="fig"} ○), indicating that the association of ARF-GEF activity with the insulin receptor was dependent on the presence of insulin.Figure 4**The ARF-GDP exchange activity of the coimmunoprecipitates with the insulin receptor.** The exchange activity was determined as described in Materials and Methods. (○,□) Receptors were immunoprecipitated in the absence of insulin from cells transfected with empty vector (○) or with myc-ARNO (□). (●,■) Receptors were immunoprecipitated in the presence of insulin from cells transfected with empty vector (●) or with myc-ARNO (■). We then transfected HIRcB cells with myc-tagged ARNO constructs. Fig. [5](#Fig5){ref-type="fig"} shows that the wild type ARNO co-immunoprecipitated with the insulin receptor in an insulin-dependent manner. E156K-ARNO was also co-immunoprecipitated with the insulin receptor upon insulin stimulation. However, none of the deletion mutants, including ΔPH-ARNO, PH-ARNO, ΔCC-ARNO, and CC-ARNO, as well as a site-directed mutant R280D-ARNO, was found co-immunoprecipitated with the insulin receptor. These data suggest that ARNO directly interacts with the insulin receptor and that the interaction requires intact PH and CC domains, but the catalytic activity of the Sec7 domain does not alter the interaction.Figure 5**Immunoprecipitation of the insulin receptor with ARNO and its mutants.** Immunoprecipitated proteins were resolved by SDS-PAGE and myc-ARNO, myc-E156K-ARNO, myc-R280D-ARNO and myc-ΔPH-ARNO were detected by immunoblotting with a monoclonal anti-myc epitope antibody. PH-ARNO-EGFP, ΔCC-ARNO-EGFP, and CC-ARNO-EGFP were detected by immunoblotting with a polyclonal antibody against EGFP. Effects of the overexpression of ARNO or its mutants on insulin-dependent PLD activity {#Sec6} -------------------------------------------------------------------------------------- We have shown so far that ARNO mediates the translocation of ARF proteins to the plasma membrane with insulin stimulation. Since ARF proteins mediate the activation of PLD by insulin \[[@CR23]\], we tested the hypothesis that ARNO may play a role in the regulation of PLD activitiy upon insulin stimulation. To prove this point, the PLD activity of HIRcB cells that had been transiently transfected with the wild type ARNO, and mutant ARNO constructs. Fig. [6](#Fig6){ref-type="fig"} shows that the overexpression of the wild type ARNO significantly increased insulin-induced PLD activity when compared with that of non-transfected cells. In contrast, the overexpression of the indicated ARNO mutants significantly decreased the ability of insulin to stimulate PLD. We conclude, therefore, that members of the cytohesin/ARNO family of ARF GEFs play an important role in the regulation of PLD activity by insulin.Figure 6**Effects of overexpression of the wild type and mutant ARNO constructs on the activation of phospholipase D by insulin.** HIRcB cells were trans fected with empty vector, myc-wt-ARNO, myc-E156K-ARNO, myc-R280D-ARNO, and myc-ΔPH ARNO, PH-ARNO-EGFP, ΔCC-ARNO-EGFP, and CC-ARNO-EGFP. PLD activity was determined by a transphosphatidylation assay as described in Materials and Methods. Discussion {#Sec7} ========== Several studies have demonstrated that ARF proteins may mediate receptor-dependent activation of PLD. Stimulation of cell surface receptors with agonists, such as insulin, promotes the translocation of ARF proteins to the cell membranes and the activation of ARF proteins and the subsequent activation of PLD \[[@CR16], [@CR18], [@CR21], [@CR23]\]. However, the mechanisms by which ARF proteins are activated by cell surface receptors remain obscure. ARF GEFs of the cytohesin/ARNO family have been shown to be recruited to cell membranes by mechanisms that are influenced by extracellular agonists \[[@CR7], [@CR26]\]. These GEFs have been implicated in the regulation of many cellular processes, ranging from the regulation of cell motility \[[@CR27]\] to cell adhesion \[[@CR28]\] and, more recently, oncogenesis \[[@CR29]\]. It has been speculated that PLD activation may mediate several of the cellular events regulated by cytohesin/ARNO GEFs \[[@CR30]\]. However, a direct proof of a role for these factors in the regulation of the receptor-mediated PLD activation is still lacking. To address these and other related issues, we have studied in detail some of the mechanistic aspects of this pathway using a fibroblast cell line that overexpresses human insulin receptors as a model. This model and other similar ones have been used in our laboratory and others to examine specific aspects of insulin receptor function, such as receptor phosphorylation and traffic \[[@CR23], [@CR31]--[@CR33]\] and the regulation of the MAPK pathway \[[@CR34]\]. Our studies showed that insulin promoted the translocation of myc-tagged ARNO constructs to the plasma membrane. This result is in agreement with data previously published by Venkateswarlu et al \[[@CR7]\] and Langille et al \[[@CR35]\] who demonstrated the insulin-dependent translocation of ARNO and the related protein GRP-1 to the plasma membrane, respectively. A detailed analysis of ARNO deletion and point mutants demonstrated that: 1) the translocation of ARNO to the membrane is independent of its ARF-GEF activity; 2) ARNO translocation to the plasma membrane requires an intact PH domain; 3) the CC domain of ARNO plays a role in targeting ARNO to the plasma membrane; 4) neither the PH domain of ARNO nor its CC domain alone sufice to target the protein to the plasma membrane; and 5) the plasma membrane translocation of ARNO is strongly regulated by insulin and, perhaps, other extracellular agonists. The linkage between ARNO translocation to specific subcellular fractions and ARF activation was studied using myc-tagged ARNO and ARF-GFP constructs in two different cell types. Our data showed conclusively that insulin promoted the co-localization of wild type myc-ARNO and ARF1-GFP on the surface of HIRcB and HeLa cells. Interestingly, insulin, acting through ARNO, promoted the translocation of ARF1-GFP to the plasma membrane. ARF1, like most members of the ARF family, is primarily a cytosolic protein that exerts its function on specific membranes to which it is recruited by specific activators that promote the binding of GTP. However, ARF1 seems to act primarily at the Golgi, promoting the binding of coatomer proteins to the Golgi membrane \[[@CR36], [@CR37]\]. Nevertheless, the fact remains that ARF1 is primarily cytosolic, and that only a small fraction of it is bound to the Golgi membrane at any time \[[@CR36]\]. It is not surprising, therefore, that some ARF1 may bind to the plasma membrane after being locally activated by ARNO, which is in turn recruited to the cell surface by the action of insulin. It should be remembered that our cells overexpress ARF1-GFP. Whether ARF1 does in fact work at the plasma membrane under physiological conditions or not remains to be established. Our data simply establish the fact that a receptor-dependent mechanism to recruit ARF1 to the plasma membrane does exist. On the other hand, ARF6 is normally found associated with the plasma membrane \[[@CR36], [@CR38]\], and there is evidence that ARF6 might be the primary target for ARF-GEFs of the cytohesin/ARNO family \[[@CR27]\]. However, when ARF dominant negative mutants were tested for their ability to inhibit agonist-dependent PLD activation, the data showed that ARF1 dominant negative mutants (T31N-ARF1) were as efficient as ARF6 mutants (T27N-ARF6) \[[@CR23]\]. These observations strongly support the idea that ARF-GEFs of the cytohesin/ARNO family have full access to the cytosolic ARF proteins. Therefore, although ARF6 might be the primary intermediate for ARNO-regulated PLD activation, other ARF proteins may as well play an important role in the pathway. The ability of insulin to promote the translocation of ARNO and ARF to the plasma membrane correlated well with the ability of insulin to promote the activation of PLD. Therefore, our data support the hypothesis that the activation of PLD by insulin is mediated by ARF-GEFs of the cytohesin/ARNO family by a mechanism that involves the interaction of the PH and CC domains of these GEFs with some specific cellular targets. This conclusion is based on the demonstration that ARNO constructs with catalytically inactive domain or the mutants with defective PH and CC domains acted as dominant inhibitors of insulin-dependent PLD activation. The dominant negative effects of E156K-ARNO were not unexpected, since this mutant contains the intact PH and the CC domains and is therefore likely to compete with endogenous ARNO. The dominant negative effect of the PH and the CC domain deletion mutants on PLD activation was of particular interest. These mutants were at best partially translocated to the membrane but blocked the ability of insulin to promote ARF and PLD activation. This result was somewhat surprising since these deletion mutants contain an intact Sec7 domain and, therefore, would have been expected to support ARF and PLD activity. However, this was not the case, suggesting that all regions of ARNO play an important role in the regulation of this protein. Moreover, the failure of the ΔCC mutant to activate ARF and PLD indicates that other cellular targets that bind to the CC doma in of ARNO and regulate the subcellular location or the function of the signaling protein complex may exist. In fact, some proteins that interact strongly with the CC domain of members of the ARNO family, such as CASP and GRASP, have already been identified \[[@CR39], [@CR40]\]. Consistent with these ideas was the observation that the overexpression of either the PH or the CC domain alone was sufficient to block insulin-dependent PLD activation. Therefore, we propose that cellular targets that recognize both the PH and CC domains of ARNO are important for the regulation of the function of this protein by cell surface receptors. On the other hand, our data also strongly support the hypothesis that the regulation of ARNO activity by insulin involves, at least transiently, a direct interaction of the insulin receptor with ARNO. Consistently, the presence of an ARNO-like activity and ARNO in the immunoprecipitated materials was confirmed by biochemical experiments. Finally, ARNO constructs lacking either the CC or the PH domain, or with a defective PH domain, failed to co-immunoprecipitate with the insulin receptor. These findings suggest a mechanism of the activation in which the binding of ARNO to the membrane is regulated by the insulin receptor at two different levels: 1) ARNO must interact with the receptor; and 2) ARNO must interact with the membrane, either via binding to polyphosphoinositides or through the interaction with specific protein targets. Our data strongly support the idea that both CC and PH domains play a crucial role in this phenomenon. Conclusions {#Sec8} =========== This study suggests a general model for the activation of PLD with insulin stimulation. Insulin, upon binding to its receptor, promotes the phosphorylation of IRS-1 and the activation of PI3 kinase. This results in the accumulation of polyphosphoinositides on the plasma membrane. In parallel, the insulin-bound receptor promotes the recruitment of ARNO (and/or other members of the ARNO family, such as GRP-1) to the plasma membrane, either by direct interaction with their CC and PH domains or by promoting the interaction of ARNO with other as yet unidentified targets. The binding of ARF-GEFs to the plasma membrane is stabilized by the interactions of their PH domain with polyphosphoinositides generated by the action of PI3 kinase. Once on the membrane, the ARF-GEFs catalyze the activation of membrane-bound ARF6 or cytosolic ARF proteins that are then recruited to the membrane where they may activate PLD. Cell culture {#Sec9} ============ Rat-1 fibroblasts overexpressing the human insulin receptors (HIRcB cells) were cultured in Dulbecco\'s modified Eagle\'s medium (DMEM)/Ham\'s F-12, supplemented with 10% fetal bovine serum, antibiotics, and 100 nM methotrexate, as previously described \[[@CR20]\]. Cells were subcultured, transfected as indicated in the figure legends, and serum starved for overnight (approximately 20 hrs) prior to insulin stimulation. HeLa cells were cultured in DMEM supplemented with 10% fetal bovine serum and antibiotics. HeLa-ARF1-GFP stable transfectants were obtained by using G418 as a selection agent as described elsewhere \[[@CR25]\]. Clonal populations were obtained and used in the assays described here. Transient Transfection {#Sec10} ====================== Subconfluent (70--90%) HIRcB cells were transfected with LipofectAMINE (Invitrogen) for biochemical analyses or Superfect (QIAGEN) for imaging analyses. Transfection was performed according to the manufacturer\'s instructions. Transfection efficiencies were 70--90% for LipofectAMINE and 40--50% for Superfect transfection reagent as previously described \[[@CR41]\]. Generation of fusion proteins {#Sec11} ============================= It has been reported that the members of the cytohesin/ARNO family of ARF-GEFs each exist in two isoforms in terms of existence of extra G (glycine) in PH domain \[[@CR42]\]. In this study, we used the isoform of ARNO with GGG (tri-glycine), which has similar binding affinities for both PI-(3,4,5)-P~3~ and PI-(4,5)-P~3~. The following myc-tagged ARNO constructs were generated: wt-ARNO, ΔPH-ARNO, PH-ARNO, ΔCC-ARNO, E156K-ARNO, and R250D-ARNO. wt-ARNO, ΔPH-ARNO (amino acids 1 to 269), PH-ARNO (amino acids 262--399), and ΔCC-ARNO (amino acids 51--399) (Fig. [1](#Fig1){ref-type="fig"}) were amplified by PCR and subcloned in the multiple cloning site of the vector pEGFP-C1 (CLONTECH) and fused to green fluorescent protein (GFP) as described by Venkateswarlu and coworkers \[[@CR7]\]. The CC domain of ARNO (amino acids 1 to 55) (Fig. [1](#Fig1){ref-type="fig"}) was PCR out of wt-ARNO and subcloned into pEGFP-N1 using BglII and EcoRI restriction sites. E156K-ARNO (inactive Sec7 domain) was generated by site-directed mutagenesis as described by Frank and coworkers \[[@CR43]\]. R280D-ARNO was designed on the basis of that a mutation on an analogous arginine impairs the binding of cytohesin-1 to polyphosphoinositides \[[@CR26]\]. The sequences of the constructs were verified by direct sequencing and the expression of appropriate fusion proteins was examined by Western blotting. The level of expression of all constructs was found to be comparable. Immunoprecipitation assay {#Sec12} ========================= Transfected and serum-starved HIRcB cells were washed with ice-cold PBS, scraped, and collected by centrifugation. The cell pellets were solubilized on ice for 1 hr in a solution of 50 mM Hepes, pH 7.45, containing 100 mM NaCl, 1.5% sodium cholate, 1 mM EDTA, 1 mM EGTA, 5 ug/ml leupeptin, 1 mM PMSF, and 1 mg/ml soybean trypsin inhibitor. Insoluble materials were removed by centrifugation. The cell lysate was immunoprecipitated with anti-mouse IgG agarose that had been equilibrated with a monoclonal antibody 83.7 (which recognizes the α subunit of the human insulin receptor). Immunoprecipitation was carried out overnight (approximately 20 hrs) at 4°C. The immunoprecipitates were washed with lysis buffer, resuspended in SDS-PAGE sample buffer, and subjected to Western blotting analysis. Immunoblotting {#Sec13} ============== Proteins were separated by SDS-PAGE, transferred to a nitrocellulose membrane, and blocked with 5% non-fat milk in PBS containing 0.1% Tween at room temperature for 2 hrs. The membrane was then cut in half horizontally. The upper part was used to detect the β subunit of the insulin receptor with a monoclonal antibody, CT-1, that recognizes the carboxyl terminus of the β subunit of the human insulin receptor. The lower part was used to detect ARNO proteins with a monoclonal antibody anti-myc or a polyclonal antibody anti-GFP. PLD activity assay {#Sec14} ================== Serum-starved HIRcB cells were labeled overnight with ^3^H-palmitate (5 μCi/ml) in serum-free medium. The cells were stimulated with insulin (100 nM) in the presence of 0.5--1% ethanol for 20 min. The reaction was stopped by addition of chloroform: methanol (1:1). The lipid phase was extracted and developed by thin layer chromatography (TLC) on silica gel 60 plates using ethyl acetate: trimethylpentane: acetic acid (9: 5: 2) as a solvent. The position of major phospholipids was determined using true standards (Avanti Biochemicals) and autoradiography. The TLC plates were scraped and the total amount of radioactivity associated with each lipid species was determined by liquid scintillation counting. The data were expressed as the number of counts associated with the phosphatidylethanol (PtdEtOH) spot normalized by the total number of counts of lipid. Digitonin treatment {#Sec15} =================== Serum-starved HIRcB cells were collected, resuspended in PBS, and treated with 10 μM digitonin in the presence or absence of insulin (100 nM), ATP (1 mM), and GTPγS (100 μM) at 37°C for 15 min. To release intracellular proteins, the digitonin-treated cells were centrifuged in a microcentrifuge for 20 min. The supernatants and the cell pellets were collected separately, and subjected to SDS-PAGE. ARNO proteins were detected by immunoblotting as described above. In vitro ARF activation assay {#Sec16} ============================= ARF activation was determined by the binding of GTPγS to the purified, myristoylated recombinant human ARF1 (mhARF1), as described by Shome and coworkers \[[@CR23]\]. The insulin receptor was immunoprecipitated in the presence or absence of 100 nM insulin as described above. Four to 8 μg mhARF1 and the immunoprecipitated insulin receptors were incubated with 100 nM GTPγ\[^35^S\] (1 μCi) in 20 mM Hepes buffer containing 2 mM MgCl~2~/ 0.1% Na-cholate / 1 mM ATP. At the indicated time points, the reaction was quenched by addition of 100 μM ice-cold, unlabeled GTPγS and the protein-bound nucleotide was determined by filtration through nitrocellulose filters as described \[[@CR23]\]. Confocal microscopy {#Sec17} =================== HIRcB cells were plated on poly-L-lysine coated glass coverslips and transfected with the constructs as indicated above. Cells were serum starved overnight and stimulated with 100 nM insulin. Live cells were imaged in a LSM5 Zeiss laser scanning confocal microscope equipped with a 63X oil immersion objective. For ARF and ARNO colocalization experiments, HIRcB cells were plated on poly-L-lysine coated coverslips as described above and co-transfected with myc-ARNO and ARF-GFP constructs using Superfect transfection reagent according to the manufacturer\'s instructions. Following insulin stimulation, the cells were fixed with 4% fresh paraformaldehyde in PBS at 4°C for 30 min, and permeabilized in 0.1% Triton X-100 at room temperature for 2 min. After permeabilization, the cells were blocked with 3% bovine serum albumin in PBS at room temperature for 30 min, and immunostained with a monoclonal antibody 9E10 (Upstate Biotechnology) that recognizes the myc epitope. After extensively washing, the cells were incubated with a Cy5-conjugated donkey anti-mouse secondary antibody (Jackson Immunoresearch) and imaged using a Zeiss laser scanning confocal microscope with filters appropriate for the detection of GFP and Cy5. Authors' original submitted files for images {#Sec18} ============================================ Below are the links to the authors' original submitted files for images.Authors' original file for figure 1Authors' original file for figure 2Authors' original file for figure 3Authors' original file for figure 4Authors' original file for figure 5Authors' original file for figure 6 **Authors\' contributions** Kuntala Shome carried out some *in vitro* ARF activation, ARNO translocation and PLD assays. Raúl Rojas contributed with the initial studies of ARNO/ARF translocation. Megan Rizzo and Chandrasekaran Vasudevan performed most of the color imaging studies. Eric Fluharty, Lorraine C Santy and James Casanova made ARNO mutants. Hai-Sheng Li made a CC-ARNO mutant; carried out imaging analysis; and participated in the ARNO/ARF translocation and PLD assays. Guillermo Romero coordinated the study and participated in the imaging studies. This research was supported by the NIH (R01 DK 51183 and R01 DK 54782). GR is a recipient of an independent investigator Award from NIDDK (K02 DK02465). MAR and RR were supported by NIH pre-Doctoral Training Grant 5T32-GM08424. CV was supported by a Grant from the American Heart Association (PA Affiliate).

PubMed Central

2024-06-05T03:55:45.940651

2003-9-11

PMC212687

Introduction {#s1} ============ During B- and T-lymphocyte development, the immunoglobulin (Ig) and T-cell receptor (TCR) genes are assembled from discrete V, D, and J gene elements via a process of genomic rearrangements known as V(D)J recombination ([@pbio-0000001-Fugmann2]; [@pbio-0000001-Hesslein1]). V(D)J recombination occurs in two steps: a cleavage phase, in which DNA double-strand breaks are created, followed by a joining phase ([@pbio-0000001-Fugmann2]). During cleavage, the lymphoid-specific recombinase proteins, RAG1 and RAG2, presumably together with the accessory DNA-binding factor HMG-1/2, bind recombination signal sequences (RSSs) located adjacent to each rearranging gene element. A complex consisting of RAG and HMG proteins bound to a single RSS is then thought to capture a second RSS ([@pbio-0000001-Jones1]; [@pbio-0000001-Mundy1]); within this synaptic complex, the RAG proteins introduce double-strand breaks at the junctions between each RSS and its associated gene element ([@pbio-0000001-Hiom2]). In the joining phase, ubiquitous DNA repair factors involved in nonhomologous end joining, in the presence of the RAG proteins, ligate the cleaved ends, generating two types of recombinant junctions: precise signal joints (SJs) and imprecise coding joints (CJs) ([@pbio-0000001-Bassing1]). RSSs are an essential part of V(D)J recombination, as their presence is both necessary and sufficient to direct RAG-mediated recombination on artificial substrates. Sequence alignments of RSSs suggested that each signal can be dissected into three components: a conserved heptamer (consensus: 5′-CACAGTG) and a conserved nonamer (consensus: 5′-ACAAAAACC), separated by a poorly conserved spacer of either 12 ± 1 or 23 ± 1 bp ([@pbio-0000001-Tonegawa1]; [@pbio-0000001-Akira1]; [@pbio-0000001-Ramsden2]). The heptamer is the site of DNA cleavage ([@pbio-0000001-Roth1]), while the nonamer provides a major binding surface for RAG1 ([@pbio-0000001-Difilippantonio1]; [@pbio-0000001-Spanopoulou1]; [@pbio-0000001-Nagawa1]; [@pbio-0000001-Swanson2]). Spacer length restricts recombination according to the "12/23 rule"; efficient recombination occurs between two gene elements only when one element is flanked by an RSS with a 12 bp spacer (12-RSS) and the other by an RSS with a 23 bp spacer (23-RSS) ([@pbio-0000001-Tonegawa1]). Despite the enormous specificity that RSSs confer on the recombination process, the recombination signals themselves demonstrate a remarkable degree of sequence heterogeneity. Only the first three nucleotides of the heptamer and the fifth and sixth positions of the nonamer show almost perfect conservation ([@pbio-0000001-Ramsden2]) and are therefore thought to be the major determinants of RSS specificity and function. Mutations in any of these five "critical" nucleotides, alone or in combination, essentially abolish recombination ([@pbio-0000001-Tonegawa1]; [@pbio-0000001-Akira1]; [@pbio-0000001-Hesse2]). The roles of the remaining "noncritical" heptamer and nonamer nucleotides are less understood. Some studies observed that mutations in these lesser-conserved residues have comparatively milder phenotypes unless present in combination ([@pbio-0000001-Tonegawa1]; [@pbio-0000001-Hesse2]). Others, however, reported that nonconsensus deviations of noncritical residues lead to vastly different recombination efficiencies, resulting in significant differences in gene element usage in the unselected antigen receptor repertoire ([@pbio-0000001-Ramsden1]; [@pbio-0000001-Suzuki1]; [@pbio-0000001-Connor1]; [@pbio-0000001-Larijani1]). Our current knowledge about the functional role of the spacer is that its length is crucial in directing V(D)J recombination ([@pbio-0000001-Tonegawa1]; [@pbio-0000001-Hesse2]). Comprehensive sequence alignments show that the spacer possesses some degree of sequence conservation, albeit at a level much lower than that of the heptamer or nonamer ([@pbio-0000001-Ramsden2]). This suggests that there is little or no selective pressure for spacers to adopt a given sequence. Studies examining the effects of different spacer sequences on recombination activity have yielded seemingly conflicting results. An early report found up to a 15-fold effect of different spacer sequences ([@pbio-0000001-Akira1]), while follow-up studies observed either no effect ([@pbio-0000001-Wei1]; [@pbio-0000001-Akamatsu1]) or up to 6-fold effects ([@pbio-0000001-Fanning1]; [@pbio-0000001-Nadel1]; [@pbio-0000001-Larijani1]). This suggests that spacer sequence may affect recombination activity, but a comprehensive picture of the rules that govern how it does so is lacking. One limitation inherent in many prior RSS studies is that they have often been performed in the context of RSSs with a preponderance of consensus nucleotides. While such analyses have been useful in characterizing the most conserved or critical determinants of RSS function, the contributions of other nucleotides are potentially masked in RSSs with high consensus nucleotide representation. That most endogenous RSSs do not contain consensus heptamer and/or nonamer motifs further suggests the need for a careful study of individual RSS nucleotides in the context of physiologically relevant RSSs. We have performed an extensive analysis of the functional properties of RSS elements in the context of endogenous recombination signals. To explore the nature of the complex relationships that might exist among different elements and positions in the RSS, we started with the nonfunctional RSS associated with the murine Jβ2.6 pseudogene element of the TCRβ locus (Jβ2.6 RSS). While most such pseudogene elements are flanked by RSSs with crippling mutations ([@pbio-0000001-Akira1]), Jβ2.6 is unique in that the sequence of its flanking RSS suggests no obvious explanation for its complete lack of activity ([Figure 1](#pbio-0000001-g001){ref-type="fig"}). All of the critical residues are conserved, and each nonconsensus nucleotide in the heptamer and nonamer is represented in at least one other functional RSS in the TCRβ locus ([Figure 1](#pbio-0000001-g001){ref-type="fig"}). A systematic analysis of Jβ2.6/consensus hybrid RSSs revealed that the nonamer, by itself, is the biggest determinant of Jβ2.6 RSS activity and that the lack of Jβ2.6 RSS function is due to the concerted action of nonconsensus nucleotides throughout the entire RSS, including the spacer. Surprisingly, we found that in combination with other consensus elements, an artificial consensus spacer can markedly boost recombination activity, while an anticonsensus spacer strongly impairs activity. Furthermore, in a genetic screen for functional spacer sequences, we observe a selective pressure for substrates with an increased representation of consensus nucleotides. Our results provide strong support for the model that RSS activity is a summation of numerous complex interactions between the RAG proteins and the RSS, involving not only the heptamer and nonamer but also most (if not all) basepairs of the spacer. ::: {#pbio-0000001-g001 .fig} Figure 1 ::: {.caption} ###### Recombination Signal Sequences Heptamer, spacer, and nonamer elements of 12-RSSs referred to in this study are shown. "Cons." and "Anti-Cons." denote the consensus and anticonsensus 12-RSSs, respectively. VκL8, Jβ2.6, and Jβ2.2 are murine 12-RSSs. "Jβ Cons." denotes the consensus RSS compiled for all functional 12-RSSs in the murine Jβ1 and Jβ2 clusters. Where more than one nucleotide is listed at any given position, this indicates a shared preponderance of those nucleotides. For consensus RSSs, nucleotides in bold indicate almost absolute conservation; for the anticonsensus RSS, bold nucleotides are almost completely absent. Nucleotides in lowercase italics appear at slightly reduced frequencies compared to the other nucleotides listed. "Jβ-G/-A/-T/-C" and the corresponding numbers indicate the number of functional RSSs in the murine Jβ1/Jβ2 clusters at which the respective nucleotide appears at the designated position. At the top of the figure, the position of each nucleotide is labeled with respect to the first position of the respective element. :::  ::: Results {#s2} ======= In Vivo Assay for Recombination {#s2a} ------------------------------- We generated a series of recombination substrates to measure the ability of various hybrid Jβ2.6/consensus 12-RSSs to rearrange to a "standard" 23-RSS (consisting of consensus heptamer and nonamer elements flanking a spacer from the functional Ig Jκ1 RSS). This standard 23-RSS was used instead of the natural Jβ2.6 RSS partner (the 23-RSS flanking Dβ2), since the substrates containing the Dβ2 23-RSS showed much lower levels of recombination in our hands (data not shown). The 12-RSS coding flank was the same for all constructs, namely that of Jβ2.6. For our study, a polymerase chain reaction (PCR)-based assay ([Figure 2](#pbio-0000001-g002){ref-type="fig"}, top) was employed, which allowed us to visualize recombination efficiencies across a \>1,000-fold range. The recombination substrates were transfected into the human embryonic kidney cell line 293T along with constructs expressing full-length RAG1 and RAG2 proteins, and recombination frequencies were measured by PCR using primers that amplify SJs. To confirm that the amplified products in our PCR assay were bona fide SJs, we demonstrated that they could be cleaved efficiently with *Apa*LI restriction endonuclease, which cuts precise RSS--RSS junctions (data not shown). The amount of recombination substrate recovered from each transfection was measured by PCR and used to normalize the recombination activity. Although we assayed primarily for SJ formation, analyses of CJ formation yielded parallel results (data not shown). As a reference, we used a substrate containing the 12-RSS from the TCR Jβ2.2 gene element (see [Figure 1](#pbio-0000001-g001){ref-type="fig"}), which recombines at low but detectable levels, as measured both in our system and during T-lymphocyte development ([Figure 2](#pbio-0000001-g002){ref-type="fig"}, lanes 1--4) ([@pbio-0000001-Livak2]). ::: {#pbio-0000001-g002 .fig} Figure 2 ::: {.caption} ###### Recombination Activities on Hybrid Jβ2.6/Consensus RSSs A diagram of the recombination assay (SJ formation) is shown (top). Activities were measured on substrates containing the indicated hybrid 12-RSS and a standard 23-RSS. H, Sk, Sc, or N denotes the consensus heptamer, VκL8 spacer, consensus spacer, or consensus nonamer, respectively; each 12-RSS bears the indicated combination of consensus/VκL8 elements, with the remaining elements belonging to Jβ2.6 RSS. To determine relative recombination efficiencies, the amount of SJs was first corrected for DNA recovery, then normalized to the values obtained for the substrate containing the Jβ2.2 RSS. Relative recombination efficiencies for each of three experiments are shown as bar graphs; the average value is shown below each sample. The gels shown here correspond to Experiment 3 and represent products of PCRs on 10-fold dilutions of recovered plasmid DNA. :::  ::: Consensus Heptamer, Spacer, and Nonamer Replacements {#s2b} ---------------------------------------------------- Recombination of Jβ2.6 RSS is below the level of detection of our assay ([Figure 2](#pbio-0000001-g002){ref-type="fig"}). Substitution of a consensus heptamer (H) into the Jβ2.6 RSS elevates the recombination frequency to levels just above background ([Figure 2](#pbio-0000001-g002){ref-type="fig"}, lanes 13--16). Similarly, substitution of a spacer from a standard, functional 12-RSS (recombination signal sequence spacer \[Sk\], from Ig VκL8; see [Figure 1](#pbio-0000001-g001){ref-type="fig"}) or of an artificial consensus spacer (Sc) only marginally restores recombination ([Figure 2](#pbio-0000001-g002){ref-type="fig"}, lanes 17--24). By contrast, substitution of a consensus nonamer (N) boosts recombination activity to the level of Jβ2.2 RSS ([Figure 2](#pbio-0000001-g002){ref-type="fig"}; compare lanes 1--4 to 25--28), approximately 20-fold higher than substitution of H, Sk, or Sc alone and at least two orders of magnitude above Jβ2.6 RSS. Therefore, the nonamer, by itself, is the biggest single determinant of Jβ2.6 RSS activity. The combination of a consensus heptamer and nonamer (H--N) further increases activity approximately 10-fold above N alone ([Figure 2](#pbio-0000001-g002){ref-type="fig"}, lanes 45--48). Hence, the cumulative effects of nonconsensus mutations in the heptamer and nonamer elements of Jβ2.6 RSS are quite large. In combination with a consensus heptamer and/or a consensus nonamer, the presence of either the VκL8 or the consensus spacer markedly enhances recombination activities above those observed with the Jβ2.6 RSS spacer ([Figure 2](#pbio-0000001-g002){ref-type="fig"}, lanes 29--44). Although there is some fluctuation between experiments, in each replicate the greatest enhancement by the Sk or Sc spacer is seen in combination with a consensus heptamer: on average, H--Sk and H--Sc are 30- to 50-fold higher than H alone. By comparison, Sk--N and Sc--N are 3- to 8-fold higher than N, while H--Sk--N and H--Sc--N are 3- to 9-fold higher than H--N. Thus, a functional spacer can, in most cases, "rescue" the effects of a nonconsensus nonamer more fully than the effects of a nonconsensus heptamer, suggesting that the spacer has greater functional overlap with the nonamer than with the heptamer. Single-Nucleotide Consensus Replacements {#s2c} ---------------------------------------- The heptamer and nonamer of Jβ2.6 RSS differ from the consensus in only five positions (see [Figure 1](#pbio-0000001-g001){ref-type="fig"}): the last three nucleotides of the heptamer and the second and fourth nucleotides of the nonamer. To determine which of these nucleotides make the greatest contributions to Jβ2.6 RSS activity, we introduced the respective consensus nucleotides individually at each of these positions. Since substitution of a consensus heptamer alone yields very low recombination levels ([Figure 2](#pbio-0000001-g002){ref-type="fig"}), we assayed single-nucleotide heptamer replacements (H\[5\], H\[6\], and H\[7\]) in combination with a consensus spacer. We also assayed substrates containing H(5) combined with a consensus nonamer or with both consensus spacer and nonamer elements. All single-nucleotide heptamer replacements result in significant partial restoration of activity, to levels at least 50% of those obtained with the full consensus heptamer (data not shown). This suggests that the low activity of the Jβ2.6 RSS heptamer is due to contributions of all three nonconsensus nucleotides. Substitution of a consensus nucleotide at either the second or fourth position of the nonamer (N\[2\] or N\[4\], respectively), alone or in combination with a consensus heptamer and/or spacer, partially reproduces the effects of the full consensus nonamer ([Figure 3](#pbio-0000001-g003){ref-type="fig"}A). Interestingly, in each set of constructs, N(2) confers a greater restoration of activity than N(4): on average, constructs containing N(2) recombine at 50% the level of N, while constructs containing N(4) recombine at roughly 10% of N. This suggests that the recombination process has a greater preference for preserving a consensus C at the second position of the nonamer than a consensus A at the fourth position. ::: {#pbio-0000001-g003 .fig} Figure 3 ::: {.caption} ###### In Vivo Recombination Activities on Hybrid 12-RSSs with Nonamer Point Mutations or with the Anticonsensus Spacer The plots, error bars, and values listed below each sample represent the averages of three experiments. Note that all recombination efficiencies presented in this figure were obtained from transfections/PCRs that were completely independent from those shown in [Figure 2](#pbio-0000001-g002){ref-type="fig"}. Abbreviations are identical to those used in [Figure 2](#pbio-0000001-g002){ref-type="fig"}. \(A) N(2) or N(4) denotes point substitution of the consensus nucleotide at the second or fourth position of the nonamer, respectively. \(B) Sac indicates substrates that contain an anticonsensus 12-RSS spacer. :::  ::: Anticonsensus Spacer Replacements {#s2d} --------------------------------- In the presence of a consensus heptamer and/or nonamer, a consensus spacer markedly enhances recombination levels over the Jβ2.6 RSS spacer. We therefore wondered whether the presence of an artificial anticonsensus spacer (Sac) (see [Figure 1](#pbio-0000001-g001){ref-type="fig"}), containing the least-conserved nucleotide at each position ([@pbio-0000001-Ramsden2]), would impair recombination. In all cases, Sac reduced recombination levels 10- to 20-fold compared to the already inefficient Jβ2.6 RSS spacer ([Figure 3](#pbio-0000001-g003){ref-type="fig"}B; compare N to Sac--N, and H--N to H--Sac--N). In our experimental system, the consensus and anticonsensus spacer sequences are therefore capable of specifying a surprisingly large range of recombination efficiencies of up to two orders of magnitude. Coupled Cleavage In Vitro {#s2e} ------------------------- Two important questions arise from the results of these in vivo assays. First, do the differences in the RSS nucleotide sequences affect the cleavage or the joining phase of the reaction? Second, are the RAG proteins by themselves the only proteins that mediate the discrimination between various RSSs? To address these questions, we performed standard 12--23 coupled cleavage reactions using purified, truncated (core) RAG proteins ([Figure 4](#pbio-0000001-g004){ref-type="fig"}A). The linear substrates for these reactions were amplified by PCR from the plasmids used in the transient recombination assay. The amount of coupled cleavage products from three independent sets of reactions was quantified ([Figure 4](#pbio-0000001-g004){ref-type="fig"}C). While the consensus RSS (H--Sc--N) promotes efficient cleavage of up to 23% of the input substrate, the Jβ2.6 RSS is cleaved at extremely low levels, at or below the limit of detection ([Figure 4](#pbio-0000001-g004){ref-type="fig"}A, lane 2). As expected from the in vivo experiments, Jβ2.2 is sufficient for low but clearly detectable cleavage ([Figure 4](#pbio-0000001-g004){ref-type="fig"}A, lane 26). In agreement with the SJ formation data, the consensus nonamer substitution (N) boosts the level of cleavage significantly ([Figure 4](#pbio-0000001-g004){ref-type="fig"}A, lane 6), while the introduction of Sk or Sc has less effect ([Figure 4](#pbio-0000001-g004){ref-type="fig"}A, lanes 8 and 10). In contrast to our findings on SJ formation, the substrate containing a consensus heptamer (H) is as efficiently cleaved as that containing N ([Figure 4](#pbio-0000001-g004){ref-type="fig"}A; compare lanes 4 and 6). Interestingly, all substrates containing a consensus nonamer (and to a lesser extent those harboring a consensus spacer) show a high level of single-site cleavage at the 12-RSS ([Figure 4](#pbio-0000001-g004){ref-type="fig"}A, lanes 6, 10, 12, 18, and 20); such products, which are only rarely generated on extrachromosomal substrates in vivo ([@pbio-0000001-Steen1]), could account for a reduced level of coupled cleavage compared to the recombination efficiencies obtained for the respective constructs in our SJ assays. The underlying mechanism of this phenomenon is the topic of ongoing studies. ::: {#pbio-0000001-g004 .fig} Figure 4 ::: {.caption} ###### In Vitro Cleavage Reaction (A and B) Coupled cleavage was performed using body-labeled DNA substrates containing a standard 23-RSS (filled triangle) and different 12-RSSs (open triangle) as indicated above the lanes. Reaction products were separated on 4% polyacrylamide gels. The identity of the bands is indicated by symbols located between the gels; an arrow indicates the double cleavage product, while an asterisk marks single-site cleavage products. The gels shown here correspond to Experiment 2. (C and D) The intensity of the bands from three individual experiments (see legend) was quantified and the average cleavage efficiency calculated for each individual substrate (indicated below the chart). The efficiencies are displayed as relative to those obtained for Jβ2.2, which were arbitrarily set to 1. :::  ::: Interestingly, a favorable spacer sequence (Sk or Sc), when paired with H or N, boosts cleavage over the Jβ2.6 RSS spacer ([Figure 4](#pbio-0000001-g004){ref-type="fig"}A, lanes 12, 14, 16, and 18). The levels of cleavage for H--Sk or H--Sc are reproducibly higher than those for Sk--N or Sc--N; although the effect is less striking than for SJ formation, the limits of detection in the coupled cleavage assay dictate that this assay spans a much narrower range of activities than the SJ formation assay. To further address the role of spacer sequences in our coupled cleavage system, we performed another set of experiments using the substrates containing the anticonsensus spacer (Sac) ([Figure 4](#pbio-0000001-g004){ref-type="fig"}B and 4D). In conjunction with either consensus heptamer (H--Sac) or consensus nonamer (Sac--N), the anticonsensus spacer reduces cleavage 5- to 10-fold compared to the consensus spacer (H--Sc or Sc--N) ([Figure 4](#pbio-0000001-g004){ref-type="fig"}C and 4D) and 3-fold compared to the Jβ2.6 RSS spacer (H or N) ([Figure 4](#pbio-0000001-g004){ref-type="fig"}B; compare lanes 4 and 8 to lanes 6 and 10, respectively). This suggests that the Jβ2.6 RSS spacer, although "poor" compared to Sk or Sc, is still more proficient for cleavage than Sac. RSS Binding {#s2f} ----------- It is likely that differences in the nucleotide sequences of the RSS lead to variations in the stability of RAG--RSS complexes ([@pbio-0000001-Hiom1]; [@pbio-0000001-Akamatsu1]; [@pbio-0000001-Swanson2]). This idea provides one obvious explanation for the observed differences in SJ formation and cleavage efficiency among the various analyzed 12-RSSs. To address this possibility, we analyzed binding of the RAG proteins to individual isolated 12-RSSs, since the 23-RSS remained identical in all experiments described above. Binding was assessed in standard gel-shift assays using oligonucleotide substrates containing the respective 12-RSSs ([Figure 5](#pbio-0000001-g005){ref-type="fig"}A). All binding assays were performed three times; the quantitation of binding for each RSS relative to Jβ2.2 is displayed in [Figure 5](#pbio-0000001-g005){ref-type="fig"}B. (Note that the amount of shifted complex has been normalized for the amount of free probe, which contributes to the fact that, between some samples, visual assessment of relative binding activities are less striking than quantitative measurements.) As expected, the consensus 12-RSS (H--Sc--N) shows the highest binding efficiency, while binding to the endogenous Jβ2.6 RSS is weak, about 2-fold reduced compared to our standard, the functional Jβ2.2 12-RSS. Given that, as with the coupled cleavage assay, the range of activities in the binding assay is much narrower than in the SJ formation assay, these results correlate well with those obtained in the other assays. Substitution of the individual consensus elements H, Sc, and N, however, led to surprising results. While the consensus nonamer (N) sequence, as expected, increases the level of binding (up to that of Jβ2.2), the consensus spacer (Sc) alone has no effect on binding at all, and the consensus heptamer (H) consistently reduces the level of binding. The consensus spacer boosts binding only in the context of a consensus nonamer (the ratios of Sc--N:N and H--Sc--N:H--N are greater than H--Sc:H), and the consensus heptamer contributes significantly to RAG--RSS interactions in this assay only when both spacer and nonamer are consensus sequences (H--Sc--N:Sc--N \> H--N:N or H:Jβ2.6 RSS). This indicates that the nonamer is the predominant element determining the stability of the initial RAG--HMG--RSS complex while the heptamer makes additional important contributions to cleavage and recombination not reflected in this binding assay. ::: {#pbio-0000001-g005 .fig} Figure 5 ::: {.caption} ###### In Vitro Binding \(A) Binding assays were performed using the 5′-end-labeled 12-RSS substrates indicated above the lanes. Each reaction contained identical amounts of DNA substrate. Owing to differences in the end-labeling efficiencies, the quantitation (shown in \[B\]) is required to make quantitative comparisons. The gels shown here correspond to Experiment 3. \(B) The relative amount of substrate in the shifted complex was determined. The binding efficiencies from three independent experiments were calculated relative to the binding seen for Jβ2.2 oligonucleotides (which were arbitrarily set to 1). The average value is displayed below the chart. :::  ::: In the context of a consensus nonamer, the consensus spacer reproducibly enhances binding more than a consensus heptamer (Sc--N \> H--N). In contrast, the anticonsensus spacer (H--Sac--N) reduces binding about 3-fold compared to H--Sc--N ([Figure 5](#pbio-0000001-g005){ref-type="fig"}A and 5B). The effects of Sc--N compared to Sac--N are also clearly visible. Interestingly, the levels of binding in the presence of Sac are very similar to those obtained for the respective RSSs containing the original Jβ2.6 RSS spacer, in contrast to the comparative effects of the two spacers on cleavage (see [Figure 4](#pbio-0000001-g004){ref-type="fig"}). Taken together, the results of our binding studies underline clearly that the reduced ability of the Jβ2.6 RSS to participate in the initial interaction with the RAG complex, and hence the subsequent steps of V(D)J recombination, is caused not solely by the Jβ2.6 RSS nonamer but also by the "inefficient" spacer sequence. This indicates that the spacer helps the nonamer to efficiently lock the RAG proteins onto the RSS. The heptamer can contribute to this only when interactions with the other two elements are favorable. Genetic Screen for Functional Spacer Sequences {#s2g} ---------------------------------------------- Although the RSS spacer is poorly conserved and no naturally occurring RSS has yet been identified that bears the published consensus spacer sequence, our results show that the presence of the most- or least-conserved nucleotides at all positions of the spacer dramatically alters recombination activities of RSSs that contain a consensus heptamer and/or nonamer. This suggests that a functional preference exists for certain spacer sequences over others. We therefore established a genetic screen for functional spacer sequences in which each position of the spacer was randomized to contain either a consensus or an anticonsensus nucleotide (Sc/Sac). Because the greatest effect of the consensus spacer in our experiments is seen in combination with a consensus heptamer (H--Sc), the randomized spacer was analyzed in the context of 12-RSSs containing a consensus heptamer and the Jβ2.6 RSS nonamer (H--Sc/Sac). The H--Sc/Sac library contained roughly 80,000 clones, sufficient to represent each of the 4,096 possible spacer sequences multiple times (data not shown). We transfected the H--Sc/Sac library into 293T cells together with vectors expressing full-length *RAG1* and *RAG2*, and we cloned and sequenced PCR-amplified SJs. As a control, we analyzed PCR products corresponding to unrearranged substrates from library pools transfected in the absence of *RAG1* and *RAG2* ([Figure 6](#pbio-0000001-g006){ref-type="fig"}). This control pool shows a bias toward the presence of C nucleotides (the consensus nucleotide at positions 4 and 7--9 of the spacer, and the anticonsensus nucleotide at positions 1 and 6), such that the overall bias of the unselected library is slightly toward the consensus spacer (total consensus/total anticonsensus nucleotides = 1.19), consistent with sequence analysis of untransfected library clones (data not shown). Sequence analysis of amplified SJs reveals an overall enrichment for consensus spacer nucleotides over the unrearranged control (total consensus/total anticonsensus nucleotides = 1.73 for SJs, versus 1.19 for control). Spacer positions 1--5 (adjacent to the heptamer) and 8--11 all show a preference for the consensus nucleotide; the remaining positions show little or no preference for the consensus or in one case (position 7) even an enrichment for the anticonsensus nucleotide ([Figure 6](#pbio-0000001-g006){ref-type="fig"}, white bars). The strongest preference for consensus is seen at position 5, which shows almost a 3-fold enrichment over the unrearranged control; interestingly, previous mutation analyses have implicated this spacer position as having a role in affecting recombination levels ([@pbio-0000001-Fanning1]; [@pbio-0000001-Larijani1]). In general, the degree of enrichment at any given position reflects the degree to which the consensus nucleotide is represented among the endogenous RSS repertoire ([Figure 6](#pbio-0000001-g006){ref-type="fig"}) ([@pbio-0000001-Ramsden2]). ::: {#pbio-0000001-g006 .fig} Figure 6 ::: {.caption} ###### Genetic Screen for Preferred Spacer Sequences A plasmid library containing 12-RSSs with a consensus heptamer and either consensus or anticonsensus nucleotides at each position of the spacer was screened for spacers with higher activity using either in vivo recombination or in vitro coupled cleavage assays (see text for details). The number of library clones screened was \>10^5^. In total, 240 sequences from two independent in vivo experiments and 205 sequences from two in vitro screens were analyzed. The relative enrichment for a consensus over an anticonsensus nucleotide at each position was calculated (taking the bias in the starting library into account). The average from two experiments is displayed in the bar graph and the values are displayed above or below the bars. The log~2~ of the ratio of the frequency of consensus and anticonsensus nucleotides at each position is displayed; hence, a value of one indicates that the respective nucleotide occurs two times more frequently in the selected population than in the starting library. In addition, the degree of conservation of each nucleotide is indicated ([@pbio-0000001-Ramsden2]). :::  ::: To determine whether the preferred spacer sequences for SJ formation and cleavage differ, the library screen was also performed in vitro. To obtain artificial SJs from our biochemical cleavage assays, T4 ligase was added to the deproteinized cleavage products, which circularized the cleavage product containing two signal ends. The sequence analysis of such artificial SJs from two independent cleavage reactions showed that positions 2--5 as well as positions 8--11 of the spacer are enriched for consensus over anticonsensus sequences ([Figure 6](#pbio-0000001-g006){ref-type="fig"}, black bars). While these observations mirror the SJ formation data, the nucleotide located at position 1 (and to some extent position 3) seems less important for coupled cleavage than for recombination in vivo. Similar to the in vivo experiment, position 5 shows the highest magnitude of enrichment for the consensus (about 4-fold). The differences between the results of the two experimental systems (SJ formation in vivo and cleavage in vitro) could be a reflection of the number of sequences obtained in each type of analysis (200--250) or could represent differences in the nucleotide requirements of spacer participation in cleavage versus SJ formation. Overall, our experiments indicate that spacer effects are largely mediated by the RAG proteins and occur, at least in part, in the first phase of V(D)J recombination: the recognition of the RSSs, their synapsis, and the cleavage step. Correlation with a Computational Model for RSS Function {#s2h} ------------------------------------------------------- The observation that an RSS spacer can act in concert with the noncritical residues of the heptamer and nonamer to drastically modulate RSS activity suggests the need for models of RSS function that take into account complex functional relationships among the different nucleotides. A predictive algorithm for quantitatively assessing the potential of a given DNA sequence to undergo V(D)J recombination has recently been developed ([@pbio-0000001-Cowell1], [@pbio-0000001-Cowell2]). This algorithm calculates the theoretical recombination potential, or RSS information content (RIC) score, by examining internucleotide relationships within a given DNA sequence. We calculated RIC scores for the hybrid Jβ2.6/consensus RSSs used in this study, and we compared them to the experimental binding, cleavage, and recombination values ([Figure 7](#pbio-0000001-g007){ref-type="fig"}A and 7B; data not shown). The correlation between RIC scores and our experimental data is striking. The RIC score for Jβ2.6 RSS is below the threshold (−40) for sequences that would be expected to recombine. The addition of consensus heptamer and/or nonamer elements boosts RIC scores, mirroring the increases in binding, cleavage, and SJ formation. Of particular interest is the fact that effects of consensus and anticonsensus spacers on binding/cleavage/recombination are prominently reflected in the RIC scores as well. Intriguingly, RIC scores appear to be more strongly correlated with cleavage (*r~S~* = 0.90) than with binding (*r~S~* = 0.86) and most correlated with SJ formation (*r~S~* = 0.96). The correlations between our experimental data and RIC scores suggest that the failure of Jβ2.6 RSS to recombine and the ability of consensus heptamer, spacer, and nonamer elements to rescue Jβ2.6 RSS activity are functions of how well RSS structure corresponds to that of a preferred sequence. In this case, the selective advantage of the consensus RSS is not limited to a few critical nucleotides in the heptamer or nonamer but, rather, extends throughout the length of the RSS, even in regions (e.g., the spacer) that were previously thought to be unimportant. ::: {#pbio-0000001-g007 .fig} Figure 7 ::: {.caption} ###### Theoretical Predictions of RSS Qualities The average recombination/cleavage efficiencies obtained in the in vivo experiments (A) and in vitro assays (B) are plotted against the RIC scores for the 12-RSS in the respective recombination substrates. Note that the values obtained from the in vitro cleavage assays were normalized to account for differences in the detection range of individual experiments. :::  ::: Further support for the potential of the RIC score as a theoretical measure for RSS activity arises from our genetic screen. For both the in vivo and the in vitro screens, the mean RIC score of the 12-RSSs in the enriched population is higher than that of the starting pool (data not shown), and those differences are statistically significant (Student\'s *t* test and the Mann--Whitney test, p\<0.0002 for all tests). This indicates that the RIC score is able to predict the quality of RSSs and that this ability is not limited to the well-conserved heptamer and nonamer but also applies to the far more diverse spacer. Discussion {#s3} ========== RSSs are the DNA elements that direct and control the V(D)J recombination reaction. In the TCR loci, differences in the abilities of individual RSSs to recombine with each other are a significant determinant of variations in the frequencies with which gene elements appear in the mature TCR population ([@pbio-0000001-Livak1] and references therein). The molecular basis of such differences in intrinsic recombination activities lies in the remarkable sequence diversity of endogenous RSSs. Previous studies using consensus or nearly consensus RSSs suggested that only a handful of absolutely conserved nucleotides in the heptamer and nonamer serve as the major determinants of RSS specificity and function. These studies, however, did not take into account the fact that the vast majority of endogenous RSSs do not contain fully consensus elements; hence, the physiologic roles of lesser-conserved RSS nucleotides are likely of much greater significance than previously estimated. Contributions of Individual Elements {#s3a} ------------------------------------ Starting from the nonfunctional Jβ2.6 RSS, we asked the following question: what effects do a perfect heptamer, nonamer, or spacer and combinations thereof have in an inactive or poorly active RSS? We show that a number of mutations in noncritical RSS positions are required to convert Jβ2.6 RSS into a highly active 12-RSS or to convert a highly active RSS (H--Sk--N or H--Sc--N) into a completely nonfunctional, pseudogene-type RSS. Our experiments demonstrate that all RSS nucleotides, including the spacer element and the noncritical positions of the heptamer and nonamer, have some sequence-directive roles. In general, we observe that the magnitude of the effects of unfavorable nucleotides in noncritical RSS positions is dependent on the presence of other unfavorable nucleotides. This explains why, in previous studies using largely consensus RSSs, the effects of nonconsensus nucleotides at the noncritical positions were concluded to be less significant ([@pbio-0000001-Tonegawa1]; [@pbio-0000001-Hesse2]). Contributions of Individual Nucleotides in Jβ2.6 RSS {#s3b} ---------------------------------------------------- The Jβ2.6 RSS heptamer differs from the consensus in the fifth, sixth, and seventh positions; none of these is drastically more important than any other in specifying overall heptamer function (data not shown). The Jβ2.6 RSS nonamer differs from the consensus in the second and fourth positions (see [Figure 1](#pbio-0000001-g001){ref-type="fig"}), and the G at the fourth position disrupts the poly(A) tract present in the consensus nonamer. Previous footprint analyses and studies on the homologous DNA-binding domain of the bacterial *Hin* recombinase ([@pbio-0000001-Feng1]) suggest that RAG1 may bind the nonamer in the minor groove of this poly(A) tract ([@pbio-0000001-Spanopoulou1]; [@pbio-0000001-Akamatsu1]; [@pbio-0000001-Nagawa1]). Hence, we expected that restoration of the poly(A) tract of the nonamer would have a greater boosting effect on recombination levels than a consensus substitution at the second position. Instead, the opposite is true, regardless of the sequences in the remainder of the RSS (see [Figure 3](#pbio-0000001-g003){ref-type="fig"}). Having the consensus cytidine at position 2 creates a CA step within the nonamer. Such CA steps have been implicated in alternative DNA structures ([@pbio-0000001-Gorin1]); while previous discussion has focused on the CA steps present at the site of cleavage in the heptamer, it is possible that a single CA step in the nonamer is important for the RAG complex to identify the subsequent downstream poly(A) tract. Defects in RAG Binding to Jβ2.6 RSS {#s3c} ----------------------------------- Previous binding studies have shown that the nonamer is the key element for initial RAG--RSS interactions and that mutations within the nonamer can strongly reduce or even completely abolish formation of the 12-SC (signal complex) ([@pbio-0000001-Hiom1]; [@pbio-0000001-Akamatsu1]). In contrast, mutating the entire heptamer leads only to a partial decrease in 12-SC formation, and, importantly, the absolutely conserved "CAC" triplet contributes only as much to binding as the last four nucleotides of the heptamer ([@pbio-0000001-Akamatsu1]). Our gel-shift studies recapitulate these observations with the Jβ2.6 RSS heptamer and nonamer (see [Figure 5](#pbio-0000001-g005){ref-type="fig"}). Moreover, a hybrid Jβ2.6/consensus RSS containing a consensus nonamer can promote 12-SC formation as efficiently as the functional Jβ2.2 RSS (see [Figure 5](#pbio-0000001-g005){ref-type="fig"}). This explains why replacement of the Jβ2.6 RSS nonamer with a consensus nonamer can restore recombination to low but physiologically relevant levels (see [Figure 2](#pbio-0000001-g002){ref-type="fig"}). The effect of a consensus spacer on 12-SC formation exhibits striking plasticity (see [Figures 2--5](#pbio-0000001-g002){ref-type="fig"}). Additionally, in our in vitro screen, the areas of the 12-RSS spacer most highly enriched for consensus nucleotides (see [Figure 6](#pbio-0000001-g006){ref-type="fig"}) correlate with sites of spacer contacts identified in previous footprinting studies (spacer positions 2--5 and 9--11) ([@pbio-0000001-Akamatsu1]; [@pbio-0000001-Nagawa1]; [@pbio-0000001-Swanson2]; [@pbio-0000001-Swanson1]). Given that the nonamer provides the most important contact surfaces, if strong interactions with the nonamer can form, then the presence of a consensus spacer may allow additional favorable contacts to be established, not only in the spacer itself, but even farther away, in the heptamer. By contrast, an unfavorable spacer (e.g., the Jβ2.6 RSS spacer or Sac) may structurally "insulate" protein--DNA contacts seen in the nonamer, such that potential heptamer contact surfaces that could otherwise contribute to overall 12-SC stability remain hidden. This may explain why a consensus heptamer, in the absence of a good nonamer, is unable to promote formation of a stable 12-SC complex. Our in vitro cleavage assay integrates the effects of RSS binding, pairing, and actual DNA cleavage. Hence, the differences between the results of binding and cleavage assays suggest that the steps following initial binding (paired complex \[PC\] formation and DNA cleavage) are also regulated by spacer sequences. PC formation requires the recognition of the partner RSS with respect to its spacer length, and thus it is plausible that the sequence of spacers influences the protein--DNA contacts required for this compatibility test. Since it is within the PC that coordinated, synchronous DNA cleavage takes place ([@pbio-0000001-Hiom2]; [@pbio-0000001-West1]), it is conceivable that RSSs "communicate" with each other and that their spacer sequences therefore may affect the alignment of the cleavage site with respect to the recombinase active site. Such structural changes may underlie the phenomenon of the "beyond 12/23 rule" that restricts V(D)J recombination of the TCRβ locus, preventing recombination of certain 12--23 RSS pairs and favoring recombination of others ([@pbio-0000001-Jung1]). The 23 bp spacer of the Vβ RSSs is the critical element in dictating the strong preference of Vβ RSSs for the 12-RSS flanking the D segments as compared to the 12-RSS flanking the J segments, and this preference is regulated before or at the cleavage step ([@pbio-0000001-Jung1]). These intriguing findings, however, did not provide experimental insight into how a DNA motif whose sequence had previously been deemed unimportant could paradoxically play such an important role. Our findings provide a framework with which to understand how such an unexpected phenomenon might occur. Finally, the differences between the in vitro cleavage and in vivo recombination assays indicate an additional role of the spacer sequence in the joining phase of the reaction. This seems plausible, since joining is thought to start with the controlled disassembly of the postcleavage complex in which the four DNA ends, including the RSSs, are held in intimate contact with each other, presumably by the RAG proteins ([@pbio-0000001-Hiom2]; [@pbio-0000001-Tsai1]). Spacer sequences might thus be involved in controlling the structure and stability of such complexes. Relationship between Spacer Sequence Conservation and Recombination Activity {#s3d} ---------------------------------------------------------------------------- Based on comprehensive sequence alignments showing a small but significant degree of spacer sequence conservation ([@pbio-0000001-Ramsden2]), a few studies demonstrated reproducible effects of up to 6-fold of naturally occurring spacers on recombination levels ([@pbio-0000001-Fanning1]; [@pbio-0000001-Nadel1]). In transient transfection assays, we infer a much wider range of recombination efficiencies solely due to differences in spacer sequence. Strikingly, we observe that spacer sequence variably affects RSS activity depending on the extent to which each nucleotide of the spacer matches either the most- or the least-conserved nucleotide. This observation resolves some of the apparent discrepancies observed among previously published studies. For example, a poly(G) spacer, which reduces recombination 15-fold compared to a highly active control ([@pbio-0000001-Akira1]), contains one consensus and five anticonsensus residues; by contrast, a spacer containing intermixed G and C residues, which has no effect on recombination activity ([@pbio-0000001-Wei1]), contains five consensus and four anticonsensus residues. A Structural Basis for the Ability of RAG Proteins to Recombine Highly Diverse RSSs {#s3e} ----------------------------------------------------------------------------------- We find that progressive accumulation of nonconsensus nucleotides within an RSS progressively impairs recombination activity and that, at the less-conserved positions of an RSS, a multitude of nonconsensus nucleotides acting in concert can render the RSS completely inactive. This suggests that the RAG--RSS complex can tolerate or correct for a considerable amount of sequence and/or structural diversity. UV--cross-linking studies previously demonstrated RAG1 and RAG2 cross-linking to the heptamer, particularly near the site of cleavage ([@pbio-0000001-Eastman2]; [@pbio-0000001-Mo1]; [@pbio-0000001-Swanson3]). Footprint analyses of the 12-SC show that complex formation is at least partly blocked by base or phosphate group modification on the spacer side of the heptamer, on both the heptamer- and nonamer-proximal sides of the spacer, and throughout the nonamer ([@pbio-0000001-Akamatsu1]; [@pbio-0000001-Nagawa1]; [@pbio-0000001-Swanson2]; [@pbio-0000001-Swanson1]). The identified contact sites in the spacer coincide with the areas of the spacer that were preferentially found to be consensus type in our genetic screen (see [Figure 6](#pbio-0000001-g006){ref-type="fig"}). Moreover, the observed recombination efficiencies of our hybrid substrates correlate well with the predicted recombination efficiencies from RIC analyses (see [Figure 7](#pbio-0000001-g007){ref-type="fig"}A and 7B). Together, these findings support a unifying model in which the RAG proteins establish multiple contacts throughout the length of an RSS (including the spacer) that allow for fine-tuning of activity. Such an extensive network of RAG--RSS contacts within the recombinase complex would create a "structural buffer," in which unfavorable nucleotides at only a few noncritical positions might be compensated for by favorable protein--DNA interactions at other positions. Conceptually similar models exist for the I-*Ppo*I and I-*Cre*I homing endonucleases, which cleave at recognition sites approximately 20 bp in length ([@pbio-0000001-Argast1]; [@pbio-0000001-Jurica1]), and which can tolerate sequence heterogeneity in cleavage sites. Both I-*Ppo*I and I-*Cre*I form direct sidechain interactions with most of the nucleotides in their recognition sites, and it is believed that the extensive protein--DNA contacts contribute to tolerance of sequence diversity. Based on our in vivo, in vitro, and in silico analyses, we propose that the RAG--RSS complex contains two distinct types of protein--DNA interactions: "digital" (or binary) interactions of a strictly sequence-specific nature, and "analog" (or multiplicative) contacts that fine-tune the strength of the digital contacts ([@pbio-0000001-Travers1]). Digital interactions are established with those nucleotides for which proper sequence is absolutely critical for activity (e.g., the first three nucleotides of the heptamer and positions 5 and 6 of the nonamer). Analog interactions describe local structural variations brought about by different sequences along the rest of the RSS. Disruption of digital interactions completely precludes complex formation (e.g., a single mutation of a critical residue in the consensus RSS can render it entirely inactive), yet digital interactions alone are not sufficient to establish complex formation (e.g., the critical residues by themselves cannot confer activity to the Jβ2.6 RSS). This duality in the nature of protein--DNA contacts present within the RAG--RSS recombinase may be applicable to other biological systems, including other transposases, transcription factors, and DNA-binding proteins. In most protein--DNA interaction systems, the target sequence to which a protein binds contains some nucleotides that are absolutely critical, and others that are noncritical. Digital interactions are established with the absolutely conserved nucleotides in the form of sequence-specific binding, conferring a binary specificity; the digital contacts therefore determine whether a protein will bind (+1) or not (0). Analog contacts are then established with the lesser-conserved nucleotides; the analog interactions act as functional multipliers that determine the efficiency of complex stability, yielding a spectrum of binding efficiencies ranging from full activity (1 × A~max~, where A = effect on binding efficiency due to analog interactions) to no activity (0 × A~min~). Hence, the noncritical residues are crucial for determining how well a protein complex can exert its biological function. By including so many nucleotides as requirements for RSS function, the V(D)J recombination system may have evolved to avoid random cleavage of DNA and translocation errors. If only the critical heptamer and nonamer nucleotides were required for activity, the frequency of cleavage at inappropriate or "cryptic" sites in the genome would be expected to be quite high. By contrast, the required participation of noncritical nucleotides in complex stability safeguards the reaction against uncontrolled cleavage. Hence, from the standpoint of controlled diversification of reaction specificity, it is beneficial for the recombinase to have evolved a spacer with a high degree of sequence heterogeneity, while maintaining intimate contact with the spacer nucleotides via analog interactions. The complex multiplier effect of analog contacts throughout the length of the RSS, superimposed onto specific digital contacts in the heptamer and nonamer, therefore confers upon the recombinase the critical ability to distinguish between inappropriate sites that happen to contain the requisite absolutely conserved nucleotides (e.g., the Jβ2.6 RSS) versus true binding sites whose sequences diverge markedly from the consensus (e.g., most endogenous RSSs). Theoretical Predictions of RSS Quality {#s3f} -------------------------------------- RIC scores provide a powerful tool for the prediction of RSS quality based on nucleotide sequence. This method generates statistical predictions of RSS function based on the physiologic 12- and 23-RSSs in the mouse antigen receptor gene loci. In our study, RIC scores accurately predicted the relative efficiencies with which RSSs were bound, cleaved, and rearranged (see [Figure 7](#pbio-0000001-g007){ref-type="fig"}; data not shown). Interestingly, the capacity of RIC models to predict RSS quality is not restricted to sequence variability in the conserved RSS heptamer and nonamer; RIC scores also predict the effects of the RSS spacer sequence on RSS function with considerable accuracy. It is striking that RIC scores correlate so well with SJ formation, less well with cleavage, and less well still with RSS binding. This supports the idea that individual nucleotides (and groups thereof) make distinct contributions to the different steps of the V(D)J recombination reaction. This concept is consistent with previous findings showing that the nonamer is a major determinant of binding while the influence of the heptamer becomes most apparent at the level of cleavage. Hence, the efficiency with which an RSS recombines represents an integration of its protein--DNA interactions throughout all steps of the reaction, and RIC scores provide a remarkably accurate prediction of this. RIC models should be useful not only in guiding RSS mutation studies, but also in identifying potential cryptic RSSs in the genome, whose usage could lead to genomic alterations as an initial event leading to chromosomal translocations and cancer ([@pbio-0000001-Cowell1], [@pbio-0000001-Cowell2]). Furthermore, an identical mathematical approach could be useful for predicting binding sites for DNA-binding complexes (e.g., transcription factors) in general, since the algorithm incorporates the combination of both the digital and the analog DNA--protein interactions that determine the biological function of a given protein complex on a potential DNA target. Materials and Methods {#s4} ===================== {#s4a} ### Oligonucleotides and plasmids. {#s4a1} The sequence of oligonucleotides used for cloning of recombination substrates and libraries are presented in [Table S1](#st001){ref-type="supplementary-material"}. The oligonucleotides used in the gel-shift experiments are listed in [Table S2](#st002){ref-type="supplementary-material"}, and the sequences of oligonucleotides used for PCR (*INNE1*, *CIT4A*, *TL1*, *TL2*, *TL3*, *TL4*, *TL5*, and *TL6*) have been described previously ([@pbio-0000001-Eastman1]; [@pbio-0000001-Leu1]). The pSJΔ series of substrates for the in vivo recombination and in vitro cleavage assays was created as follows: pSF299 ([@pbio-0000001-Fugmann1]) was modified to create p299-Jβ2.6 by replacing the original 12-RSS with a Jβ2.6 12-RSS such that the 12/23-RSS pair is in deletional orientation; for all other substrates, the 12-RSS of p299-Jβ2.6, flanked by *Hind*III and *Sal*I sites, was replaced with the respective annealed oligonucleotides (see [Table S1](#st001){ref-type="supplementary-material"}). To generate the library for the genetic screen, the oligonucleotide *HSCSAC1* was synthesized that contained a 1:1 molar ratio of consensus:anticonsensus nucleotides at each position of the spacer and an additional randomized trinucleotide sequence downstream of the nonamer. The oligonucleotide *SJLIBREV* was annealed, the overhang was filled in using Klenow fragment (New England Biolabs, Beverly, Massachusetts), and the double-stranded fragment was digested with *Hind*III and *Sal*I and ligated into the linearized p299-Jβ2.6 vector. Ligation reactions were transformed into DH5α, colonies were harvested into 120 ml of Luria broth (containing 100 μg/ml ampicillin), and plasmid DNA was prepared after an additional incubation at 37°C at 250 rpm for 15 min. pEBB, pEBB-RAG1, and pEBB-RAG2 expression constructs have been described elsewhere ([@pbio-0000001-Roman1]). ### Recombination assays. {#s4a2} Human embryonic kidney 293T cells were transfected with 6 μg of recombination substrate and 3 μg each of pEBB-RAG1 and pEBB-RAG2 using calcium phosphate as described previously ([@pbio-0000001-Fugmann1]); for control samples without RAG expression constructs, 6 μg of pEBB was substituted. After 48 h, DNA was recovered by rapid alkaline lysis preparation (RAP) ([@pbio-0000001-Hesse1]). PCR was performed on 10-fold serial dilutions in 20 μl reaction volumes containing 1× Taq buffer (Invitrogen, Carlsbad, California), 2 mM MgCl~2~, 0.1 mM each dNTP, 0.5 μM each oligo, and 0.2 U Taq (Invitrogen). To quantify DNA recovery, the oligonucleotide pair *TL5/TL6* was used for the PCR (94°C for 15 s, 60°C for 15 s, 72°C for 30 s, for 18 cycles). To detect SJs, DNA samples were treated with *Dpn*I, *Mlu*I, and *Xho*I to remove unreplicated and unrecombined plasmids. Oligonucleotides *INNE1* and *CIT4A* were used to amplify SJs (94°C for 15 s, 60°C for 15 s, 72°C for 30 s, for 28 cycles). To detect CJs, RAP samples were treated with *Dpn*I and CJs were amplified using primers *TL2* and *TL3*. All PCR products were electrophoresed on native 4.5% polyacrylamide gels, stained with SYBR green, visualized using a Fluoroimager 595 (Molecular Dynamics, Sunnyvale, California), and quantified using ImageQuant software (Molecular Dynamics). ### Genetic screen for functional spacer sequences. {#s4a3} 293T cells were transfected with the plasmid library and RAG or pEBB constructs as described in the Results. Extrachromosomal DNA was extracted and samples were digested with either *Dpn*I/*Mlu*I/*Xho*I (for cloning of SJs) or *Dpn*I only (for cloning of unrearranged bands in no-RAG controls). PCR was performed using *INNE1* and *CIT4A* primers, and samples were electrophoresed and stained as indicated above. The products corresponding to the appropriate SJ or unrearranged bands were excised, purified, and cloned into pCR2.1 using a TOPO-T/A cloning kit (Invitrogen). DNA was prepared from individual transformed colonies and sequenced. The in vitro screen was performed using the plasmid library as the substrate in a standard coupled cleavage reaction. After proteinase K digestion, the products were precipitated and dissolved in 100 μl of 1× ligase buffer. T4 DNA ligase (1 μl) (New England Biolabs) was added and the mixture incubated at 16 °C for 4 h to create artificial SJs. The resulting plasmids were treated identically to the plasmids recovered after transfection in the in vivo screen. ### Protein expression. {#s4a4} Recombinant GST-RAG2, MBP-RAG1, and HMG2 were expressed and purified as described previously ([@pbio-0000001-Spanopoulou1]; [@pbio-0000001-Eastman2]; [@pbio-0000001-Rodgers1]). ### DNA-binding and cleavage assays. {#s4a5} The body-labeled DNA substrates for the cleavage assay were generated by PCR using the oligonucleotides *TL1*, *TL4*, and the respective recombination substrate as a template. The 12-RSS oligonucleotide substrates used in EMSA were generated by annealing the 5′-end-labeled top strand with an equimolar amount of the unlabeled respective bottom strand (see [Table S2](#st002){ref-type="supplementary-material"}). Binding and cleavage reactions were performed as reported previously ([@pbio-0000001-Fugmann3]), and gels were quantified using a Storm 820 PhosphorImager and ImageQuant software (Molecular Dynamics). ### RIC score calculation and other computational analysis. {#s4a6} Statistical models of RSS correlation structure have been previously reported ([@pbio-0000001-Cowell1]) ([Data S1](#sd001){ref-type="supplementary-material"}). Supporting Information {#s5} ====================== Data S1 ::: {.caption} ###### RIC Score Calculation and Other Computational Analysis (23 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S1 ::: {.caption} ###### Oligonucleotides for Cloning of Recombination Substrates (31 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S2 ::: {.caption} ###### Oligonucleotides for Gel Shift Experiments (23 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: We would like to thank N. H. Ruddle, N. D. Grindley, C. M. Radding, and all Schatz lab members for helpful discussions and suggestions. We also thank the W. M. Keck Foundation Biotechnology Resource Laboratory at Yale University for oligonucleotide synthesis and DNA sequencing. AIL was supported by a predoctoral fellowship from the Howard Hughes Medical Institute (HHMI). SDF is a postdoctoral fellow of the Irvington Institute for Immunological Research, and DGS is an Investigator of the HHMI. LGC received a Bioinformatics and Genome Technology Postdoctoral Fellowship from Duke University and is supported by a National Institutes of Health (NIH) Basic Immunology Training Grant (T32 AI52077-01) to Duke University Medical Center. This research was supported by grant RO1 AI-32524 to DGS and in part by grants AI-24335 and AI-49326 to GK from the NIH. **Conflicts of Interest.** The authors have declared that no conflicts of interest exist. **Author Contributions.** AIL, SDF, LGC, LMP, and DGS conceived and designed the experiments. AIL, SDF, and LGC performed the experiments. AIL, SDF, LGC, GK, and DGS analyzed the data. AIL, SDF, LGC, and LMP contributed reagents/materials/analysis tools. AIL, SDF, LGC, and DGS wrote the paper. Academic Editor: Shizuo Akira, Osaka University. CJ : coding joint H : heptamer Ig : immunoglobulin N : nonamer PC : paired complex PCR : polymerase chain reaction RAP : rapid alkaline lysis preparation RIC : recombination signal sequence information content RSS : recombination signal sequence Sac : recombination signal sequence spacer Sc : recombination signal sequence spacer SC : signal complex SJ : signal joint Sk : recombination signal sequence spacer from the VκL8 segment TCR : T-cell receptor.

PubMed Central

2024-06-05T03:55:45.943420

2003-10-13

PMC212688

Based on sheer numbers, microbes should rule the world. Most don\'t cause disease, but those that do have the advantage of multiplying and mutating at a much faster rate than any multicellular organism can. So how does a slowly reproducing, trillion-celled organism like a human protect itself? By having the right weapon for the job---and that requires an incredibly diverse arsenal. A new study by a team of researchers from Yale University School of Medicine, Duke University Medical Center, and Mount Sinai School of Medicine demonstrates how the creation of that arsenal depends on a complex series of interactions between key genetic elements and proteins during the formation of the white blood cells called lymphocytes. Two heavy hitters of the immune system---B and T cells---each produce unique protein receptors that specifically recognize and mediate the killing of the variety of potential foreign invaders, or antigens, such as bacteria, viruses, and parasites. (B cells make immunoglobulin, or antibodies, and T cells make T-cell receptors.) But these lymphocytes are unlike other cells: instead of making proteins from genes they inherited, they custom-make their genes by recombining fragments of their genes into new configurations. This genetic reshuffling process, called V(D)J recombination, yields the diversity of molecules necessary to combat the billions of different antigens they might encounter. The V, D, and J refer to different clusters of DNA sequences that follow specific rules of recombination. While the products of recombination vary, the method does not. The fragments are spliced and then reassembled in a highly regulated process directed and controlled by a stretch of DNA (called a recombination signal sequence, or RSS) next to the gene fragment. The recombination process, the researchers show, relies on complex interactions among different parts of the signal sequences and the proteins that regulate them at key steps along the recombination pathway. Each RSS is made up of three components: the nonamer, which controls the ability of proteins to bind to the gene fragments and initiate recombination; the heptamer, which directs the splicing of the gene fragment; and the spacer, which regulates how the gene fragments are recombined. Mutations in the DNA sequence of each of the three RSS components show that all play a critical role in the ability of the gene fragments to recombine appropriately. While it has been established that spacers, as their name suggests, ensure that the space between the nonamer and heptamer is correct, the researchers show that spacers also regulate recombination activity by providing protein-binding sites along the DNA sequences that affect recombination. While the nonamer is the most important determinant of recombination, changes in the spacer, these researchers demonstrate, produced dramatic changes in the ability of the gene fragments to recombine. Past studies have shown that recombination depends on the presence and sequence of specific nucleotides, but the quality of that recombination, the researchers say, can\'t be understood simply by analyzing those nucleotides in isolation. Generally speaking, highly conserved sequences have functional importance. But it would be a mistake, they suggest, to think that just because a nucleotide sequence isn\'t highly conserved, it\'s not biologically important. Using a computer model to predict how different protein-gene interactions affect recombination, the researchers demonstrate that a fuller understanding of the process depends on observing how all these elements---including those that aren\'t highly conserved---interact throughout the recombination process.

PubMed Central

2024-06-05T03:55:45.948376

2003-10-13

PMC212689

Corn is one of humankind\'s earliest innovations. It was domesticated 10,000 years ago when humans learned to cross-pollinate plants and slowly turned a scraggly nondescript grass called teosinte into plump, productive modern corn ([Figure 1](#pbio-0000008-g001){ref-type="fig"}). As needs change, so does plant breeding. Today, while biotech super-giants manipulate corn genetics to satisfy farmer desires and a global market, indigenous Mexican farmers do so to fulfill individual needs. Although the tools differ, the goal remains the same---to cultivate desirable traits. ::: {#pbio-0000008-g001 .fig} Figure 1 ::: {.caption} ###### Crossing for Kernels Over time, selective breeding modifies teosinte\'s few fruitcases (left) into modern corn\'s rows of exposed kernels (right). (Photo courtesy of John Doebley.). :::  ::: Plant breeding was once restricted to sexually compatible plants, and generations of offspring were selectively bred to create unique varieties. In fact, corn, along with rice and wheat---today\'s global crop staples---would not exist without such techniques. With the goal of ever-widening the pool of genetic diversity, conventional plant breeding has gotten more technologically savvy in recent years. For example, realizing that natural mutants often introduce valuable traits, scientists turned to chemicals and irradiation to speed the creation of mutants. From test-tube plants derived from sexually incompatible crosses to the use of molecular genetic markers to identify interesting hereditary traits, the divide between engineering and genetics was narrowing long before kingdom boundaries were crossed. But when geneticists began to explore microorganisms for traits of interest---such as Bacillus thuringiensis (Bt) genes that produce a protein lethal to some crop pests---they triggered an uproar over ethical, scientific, and environmental concerns that continues today. (See [Box 1](#box1){ref-type="boxed-text"}.) ::: {.caption} ###### Box 1. Bt Technology ::: Bacillus thuringiensis, a soil bacterium, produces several crystal (Cry) protein toxins that destroy the gut of invading pests, such as larval caterpillars. So far, over 50 *cry* genes have been identified and found to affect insect orders differently. Considered safe to humans, mammals, and most insects, Bt has been a popular pesticidal spray since the 1960s because it had little chance of unintended effects. Engineering the gene into corn, however, caused an unexpected public backlash. "We thought it was going to be the greatest thing since sliced bread," says Guy Cardineau, agricultural biotechnologist at Arizona State University. "Here\'s a way to withstand insect pressure, eliminate the use of pesticides, and Bt spray was widely used in organic agriculture," he adds. The Bt wrangle illustrates how differently a product and a process can be regarded. After the expensive development process, today\'s concern is that broad-scale planting of Bt corn will render the toxin ineffective over time. Pests can gradually build resistance to any pesticide, and so the United States Environmental Protection Agency (EPA) requires that 20% of Bt field areas be planted to non-Bt corn to avoid such pressures. But humans have to follow the rules. A recent report from the Center for Science in the Public Interest shows that almost 20% of farmers in the United States Corn Belt are violating EPA standards by overplanting Bt corn, causing some to question the regulations and enforcement that will be necessary for certain GM crops. Despite such discord, genetically modified (GM) crops have the fastest adoption rate of any new technology in global agriculture simply because farmers benefit directly from higher yields and lowered production costs. (See [Table 1](#pbio-0000008-t001){ref-type="table"}.) To date, the two most prevalent GM crops traits are Btderived insect resistance and herbicide resistance. ::: {#pbio-0000008-t001 .table-wrap} Table 1 ::: {.caption} ###### Worldwide production of GM crops :::  Four crops account for most GM plantings: herbicide-tolerant soybeans (62%), insect-resistant corn (12.4%), insect-resistant cotton (6.8%), and canola (3%). Source: Summary Report on the Global Status of GM Crops by the International Service for the Acquisition of Agri-Biotech Applications (2002) ::: Since 1987, over 9,000 United States Animal and Plant Health Inspection Service (APHIS) permits have been issued to field-test GM crops. According to APHIS, corn is the most tested plant. The International Service for the Acquisition of Agri-Biotech Applications confirms that biotech corn is the second-most common GM crop (after soybean), with 12.4 million hectares planted in 2002. GM corn starch and soybean lecithin are just two of the ingredients already found in 70% of the processed food supply. With future incarnations on the horizon, GM corn remains a lightening rod for debate. Embroiled in numerous controversies, corn has become biotech\'s boon and bane. Benefits Emerging {#s2} ================= As Danforth Center President Roger Beachy, the first to develop a virus-resistant tomato, describes it, the first-generation GM crops were intended to help farmers reduce not only the impact of pests, but also the use of agrochemicals in modern crop production--a legacy of the Green Revolution. After a decade of cultivation, environmental benefits are emerging. Bt corn reduces the need for pesticides, and while the primary benefit comes largely during a heavy corn-borer infestation, an unpredictable event, a secondary effect is that beneficial insects fare much better under these conditions. The numbers are particularly impressive for Bt cotton: the spraying of almost 2 million pounds of pesticides---roughly 50% of previous usage---has been spared since the large-scale adoption of Bt cotton. According to Leonard Gianessi, senior research associate at the National Center for Food and Agricultural Policy, farmers who adopt GM crops make more money in tougher times. Indeed, insect- and virus-resistance traits have already saved several industries. Bt cotton is credited with reviving the Alabama cotton industry, hard hit by uncontrollable bollworm infestations. Likewise, genetically engineered papaya, made resistant to the papaya ringspot virus, salvaged Hawaii\'s fifth largest crop industry. Herbicide-resistant crops engendered a different reception. While GM critics acknowledge that the use of a more benign herbicide, called by its trade name Roundup, can have environmental benefits, the creation of a market monopoly is a key criticism. However, the increased planting of herbicide-resistant soybeans is an integral, but not sole, factor in the increased adoption of no-till farming--- a strategy that reduces soil erosion. Surprise benefits have also occurred. According to the recent International Council for Science (ICSU) review of GM crops, disease-resistant corn crops may have lower levels of mycotoxins, potentially carcinogenic compounds to humans. They result from fungal activity in insect-infested corn crops. With fewer insect holes in plant tissue, associated fungi are not able to invade and produce toxins. While there is a growing amount of data documenting the intended environmental benefits of GM crops, the potential risks are more elusive. Risky Business {#s3} ============== After seven years of GM crop production and no apparent health effects, potential environmental risks---particularly gene flow into other species---have eclipsed food safety as a primary concern. As pollen and seeds move in the environment, they can transmit genetic traits to nearby crops or wild relatives. Many self-pollinating crops, such as wheat, barley, and potatoes, have a low frequency of gene flow, but the more promiscuous, such as sugar beets and corn, merit greater concern. Determining where genes flow is a thriving research avenue, but the real question becomes "so what?" The risks associated with gene flow---such as creating weeds from introduced traits, reducing biodiversity, or harming nontarget species---are similar to those from conventionally bred crops. "I wouldn\'t dismiss the ecological concerns out of hand, but I think they can be exaggerated," says Gabrielle Persley, the ICSU report author. There are few instances of crop plants becoming weeds. Bred so intensely for hundreds of years, most crops cannot survive without human intervention. Increased weediness could be conveyed, however, if the plants are more fit or able to out-compete other crop species by producing more seed, by dispersing pollen or seed further, or by growing more vigorously in a specific environment. In fact, transgenic sunflowers can produce over 50% more seed than traditional varieties. Additionally, recent work shows that, compared to pollen, seeds are more likely to spread GM sugar beet genes into wild relatives in western Europe. Norman Ellstrand, plant geneticist at the University of California at Riverside, has shown that gene flow from many conventionally bred crops increases the weediness of nearby wild relatives. For many domesticated crops, wild varieties do not exist in current areas of cultivation. Nevertheless, regions where crop species originated are particularly vulnerable to transgenic gene flow into local varieties, or landraces. Some fear that transgenic varieties with a competitive advantage might gradually displace valuable genetic diversity. For these reasons, transgenic corn is prohibited in Mexico, home to over 100 unique varieties. Despite the ban, transgenes have been found in Mexican corn. "We have in several instances confirmed that there are transgenes in landraces of maize in Oaxaca," says Ariel Alvarez-Morales, plant geneticist at the Mexican Center for Research and Advanced Studies (CINVESTAV) in Irapuato. The ramifications of this will not be known for some time, but Luis Herrera-Estrella, CINVESTAV\'s Director of Plant Biotechnology, is convinced that these single gene traits will be of little consequence to native Mexican varieties. "If Bt genes give an advantage to the farmer, they will keep growing it. In that case it will not be bad," he says of dynamically changing landraces. "Gene flow has been occurring for 50 years from commercial lines to landraces." While admitting this, Ellstrand points out that "if there are genes that you don\'t want to get into landraces---this shows how easily they can get there." (See [Box 2](#box2){ref-type="boxed-text"}.) ::: {.caption} ###### Box 2. Pharma Corn ::: "The gene flow risk that keeps me awake at night is the possibility of hybridization between crops engineered to manufacture poisons and related crops intended for human consumption," says plant geneticist Norman Ellstrand. Indeed, this application of GM crops seeks to turn corn into cost-effective pharmaceutical factories and may bear the mark of unacceptable risk. It is currently the subject of intense debate. An open-pollinated crop, corn is known for its promiscuity---making it more prone to gene flow risks than other crops. Genetic contamination takes on a whole new meaning when the escapable trait could produce proteins to treat diabetes or a hepatitis B vaccine. Given that pharma corn demands multiple safety measures---including production in remote areas, separate farm equipment, delayed planting to offset pollination---many ask, "Why use corn?" "We know so much about corn genetics," explains agricultural biotechnologist Guy Cardineau, "and it naturally lends itself to production with kernel packets of protein that can be stored indefinitely." A number of scientists and United States food makers are not yet convinced that the benefits outweigh the risks and have joined environmental groups in questioning the use of pharma corn. Over 130 acres of pharma crop field-tests were planted in 2002. Several products have moved on to clinical trials. Aware of concerns, the members of the influential Biotechnology Industry Organization decided last December to overturn its initial decision to remove pharma crops from the United States Corn Belt states and voluntarily police their use. A Colorado trial of corn producing a protein to treat cystic fibrosis recently began. Indeed, unintended impacts are a primary concern. The potential risk to nontarget organisms took center stage when a 1999 paper in *Nature* suggested monarch butterfly populations might be adversely affected by Bt transgenes. Corrected by subsequent publications, the field experiments did not support original laboratory results. But effects on other nontarget organisms, such as soil microbes, remain a concern. When microbial genetics research uncovered how genes could be transferred between species in ways other than reproduction, so-called horizontal gene transfer, it not only explained why microorganisms were so diverse, but that microbes could potentially be endowed with GM plant DNA found in the soil. "Although a theoretical possibility, there is no evidence that it happens at any degree of frequency to be meaningful," says Persley. Opinions differ on this, however, and seem to follow the United States--European Union divide over the use of GM crops. Kaare Nielsen, microbial geneticist at Norway\'s University of Tromsø, is one of the few scientists to find examples of horizontal gene transfer. "There are actually very few studies and most of the ones conducted have been on first-generation plants," Nielsen explains. Given that plant DNA can last in soil for over two years, Nielsen does not believe the possibility can be dismissed and argues that long-term studies are necessary. Work continues in this area in Europe. The lack of baseline ecological data---even agreeing on what an appropriate baseline is---presents a substantial knowledge gap to environmental impact assessments. Scientists, including Nielsen, wonder whether there could be unexpected risk factors. Allison Snow, weed expert at Ohio State University, agrees with what many feel is the most important risk---the inability to anticipate all the effects. "Do we know all of the right questions we should be asking?" she wonders, adding, "Genes are complicated and can interact." For these reasons, identifying factors that regulate weed and pest populations and determining how microbial community changes affect larger ecosystems are important areas of research. Do Risks Differ for Developing Nations? {#s4} ======================================= To two academicians that kindled the biotech revolution, the real GM risks lie in how science is misinterpreted and misused. In fact, much of the currently conducted basic research is not likely to be applied in the near future. Public concerns coupled with corporate consolidation created huge roadblocks, especially in getting the technology to developing nations. While Beachy blames the skyrocketing regulatory costs that "are due to regulators who have not put into context this technology and its relative safety," Richard Jefferson, chairman and chief executive officer of the Center for the Application of Molecular Biology to International Agriculture in Australia, fears that innovation has been stifled by corporate short-sightedness. "The biggest risk is that \[biotechnology\] maintains itself as a monolithic, expensive industry with untenable entry barriers for smaller enterprises," he says. Indeed, when does the risk of not using available technology factor into the debate? (See [Box 3](#box3){ref-type="boxed-text"}.) Many scientists argue that genetic modification can help to ensure food security in developing countries, specifically in Africa. While more than 25% of the 2002 global biotech acreage was grown in countries such as Argentina, China, and India, there is little applied research on crops relevant to famine-prone African countries. ::: {.caption} ###### Box 3. Golden Rice ::: Current regulatory constraints have a choke-hold on innovations for genetic modifications that seek to improve subsistence crops, such as rice. Golden rice, yellowed in appearance because it is infused with the vitamin A precursor beta-carotene, could save thousands of malnourished people each year from blindness and the other vitamin A--deficiency diseases prevalent in Southeast Asia. Intellectual property issues and opposition from anti-GM activists have confounded the development for years. Faced with patent issues and regulatory hurdles and costs, developer and academic researcher Ingo Potrykus formed an alliance with Syngenta (then AstraZeneca Corporation) to allow the free licensing of the patents to public research institutions for humanitarian use. In addition, farmers making less than US\$10,000 will receive free golden rice seed. After over a decade of work, golden rice is still not on the market. The retired Potrykus is determined to bring this technology to farmers once it passes regulatory field testing---an additional four-year delay that he feels is scientifically unnecessary. "Nobody can construct even a hypothetical risk to the environment from golden rice," he says, adding that nutritional risks are nonexistent as well. He acknowledges, however, that field tests will be beneficial for acceptance of this and future bio-fortified products. "I have experienced so much support in these countries, I don\'t think it \[the anti-GM lobby\] will be able to stop this project once it passes regulation," he says. "Food security is not going to come from crops being marketed outside Africa, like wheat or rice," says John Kilama, Uganda native and president of the Global Bioscience Development Institute. He points out that of traditional staple crops such as cow peas and millet, only cassava has merited some publicly-funded research. Beachy estimates that it takes between US\$5 million and US\$10 million to bring a GM crop to market. Given regulatory costs, neither industry nor universities can afford to develop products that do not have mass appeal. "If the crop is not worth that much to the seed market, it\'s likely that we\'ll never see the varieties because of the cost of regulation," he says. To ensure a return on research investments, with the regulatory costs often the biggest ticket item, developing blockbuster traits is a priority. "Given the diversity of environments and cropping systems, there are not many more blockbusters such as Roundup resistance in the wings," says Jefferson. The alternative, he adds, is to make it cheaper to innovate local varieties in ways that are likely to gain public acceptance. (See [Box 4](#box4){ref-type="boxed-text"}.) ::: {.caption} ###### Box 4. Apomixis ::: One way to minimize the problems associated with gene flow is to introduce sterility, such that pollen cannot transmit information. Richard Jefferson has high hopes for an accessible, cheap way for farmers to produce genetically superior seeds, called apomixis. But similar concepts have been floated before. The controversial terminator technology prevented gene flow, but it also outraged activists because it kept farmers from reusing seed. Unlike terminator, apomixis is "germinator" technology---avoiding fertilization altogether by producing seeds without pollination. In effect, seeds can be natural clones of the mother, instead of a genetic exchange between mother and father. Therefore, hybrid quality can be maintained as farmers use seed year after year. Although apomixis occurs naturally in about 400 plant species, Jefferson believes that it can be successfully developed as a useful trait in other crop plants. To ensure its widespread availability, Jefferson and collaborators pledged not to create restrictive patent rights that could block the development of apomixis. "The Green Revolution largely bypassed Africa," says Josette Lewis, biotechnology advisor for the United States Agency for International Development. Given monetary constraints that prevent access to many biotechnologies, many scientists worry that the Gene Revolution might as well. Looming trade issues coupled with food insecurity shape the debate in Africa. Caught between the United States and European Union trade disputes, sub-Saharan countries are eager to use any technology that will enhance production without jeopardizing trade. Increasingly, industry is responding to the developing nations\' needs. Both newly formed, the industry-focused African Agricultural Technology Foundation and the Public-Sector Intellectual Property Resource for Agriculture are attempting to ease cost restrictions and encourage access to current and future patents. By entering into such agreements, industries will be able to protect patent rights and commercially important markets. Such partnerships are already working. The Syngenta Foundation for Sustainable Agriculture is working together with the International Maize and Wheat Improvement Center (CIMMYT) and the Kenyan Agricultural Research Institute to overcome corn stemborer infestations in Kenya ([Figure 2](#pbio-0000008-g002){ref-type="fig"}). "CIMMYT hopes to have a handful of local Bt corn varieties in farmers\' fields by 2008," says the admittedly ambitious Dave Hoisington, director of CIMMYT\'s Applied Biotechnology Center. Collaborations between public and private sectors may be the only way to navigate thorny patent issues and research crop varieties of interest to developing countries. ::: {#pbio-0000008-g002 .fig} Figure 2 ::: {.caption} ###### Biotech Bridge to Africa In an effort to reduce corn stem-borer infestations, corporate and public researchers partner to develop local Bt corn varieties suitable for Kenya. (Photo courtesy of Dave Hoisington/CIMMYT.). :::  ::: Conclusion {#s5} ========== "Agricultural biotechnology is here to stay" read a recent opinion piece by Gianessi. No doubt he is correct. As genetic engineering continues to evolve, transgenic methods will become just one of many tools. In fact, some researchers are currently focusing their work on manipulating an organism\'s own genetic code to achieve desired traits. Scientific inquiry will continue to weigh the risks and benefits of such technologies, realizing that there may never be enough evidence to ensure zero risk. Only with data will tolerable levels of environmental risks be determined---case by case. Indeed, the level of risks and benefits may differ for developing nations, where decisions must be made in the face of food security concerns. To many scientists, the risks associated with forgoing genetic engineering far surpass any environmental risk associated with its use and further development. However, all stakeholders must have access to the tools in order to realize future benefits. In the quest to feed the world, a few things are clear. Public outcries will continue to vet the need and use of genetic engineering. Private organizations will necessarily focus on research for profit, while exploring collaborative prospects. Public initiatives, however, will provide the critical bridge between science and global food security. Although genetic engineering cannot be summarily accepted or rejected, any lack of scientific risk now doesn\'t negate future concerns. And, no matter what direction future research takes, corn will continue to be a bellwether crop. Virginia Gewin is a freelance science journalist in Corvallis, Oregon, United States of America. E-mail: <gewin@nasw.org>. APHIS : United States Animal and Plant Health Inspection Service Bt : Bacillus thuringiensis CIMMYT : International Maize and Wheat Improvement Center CINVESTAV : Center for Research and Advanced Studies EPA : Environmental Protection Agency GM : genetically modified ICSU : International Council for Science.

PubMed Central

2024-06-05T03:55:45.949456

2003-10-13

PMC212690

When did calling a lawyer become part of the scientific process? It hasn\'t officially, of course. But as a generation of researchers has grudgingly come to know, navigating the legal red tape of universities, corporations, and publishers is an inevitable part of the practice. Whether seeking access to information or sharing one\'s own findings with others, scientists increasingly find themselves having to ask an intermediary\'s permission. This "ask first" culture has developed at just the moment when technology has opened vast new possibilities for collaboration and information-sharing. The timing is not coincidental. Policymakers, under the influence of lobbies defending pre-digital business models, have reacted to new technology with ever more extreme intellectual property laws. The result is a legal regime tailored to a powerful minority but ill-suited to a number of other constituencies---scientists and scholars chief among them---that thrive on openness. Worries over broad notions of "piracy" and "asset management" have insinuated themselves into fields where those terms, until recently, held no meaning. The Public Library of Science (PLoS) is at the vanguard of a growing cross-disciplinary movement to counteract this trend by demonstrating that voluntary models of open publishing are not only viable, but crucial to scientific innovation. Yet PLoS' goal of "immediate, unrestricted access to scientific ideas, methods, results, and conclusions" is not immediately compatible with the stringent rules of copyright, which apply fully and automatically to all published works, by default. The exercise of something less than full copyright requires, oddly, some legal tinkering---which is where Creative Commons, the organization I help manage, comes in. Creative Commons, a 501 (c) (3) nonprofit corporation based at Stanford University in California, is led by a board of expert legal and technical thinkers. (Its chairman, Lawrence Lessig, a law professor at Stanford and a recipient of the *Scientific American* 50 Award in 2003, recently joined the PLoS board of directors.) Creative Commons was founded on the idea that some people prefer to share their works on more generous terms than standard copyright provides. The organization offers such authors an easy and clear way to announce these preferences. The goal is to help endeavors like PLoS, as well as individual authors, expand access to quality content online while reducing the legal friction and uncertainty of copyright law. In other words, Creative Commons offers legal tools to help clear permissions, once and for all. We help get the law out of science\'s way. These tools are the Creative Commons licenses, a suite of form legal documents available for free on the Creative Commons website. Each license allows an author to retain his or her copyright while permitting certain uses of the work, on certain conditions: to declare "some rights reserved" as opposed to "all rights reserved." From a simple menu, copyright holders mix and match their preferences: an attribution requirement; a prohibition on commercial reuse; a restriction on derivative works; or a "share-alike" provision that obligates licensees to offer any derivative works to the public on the terms they received. (PLoS has chosen the simplest and least restrictive of the licenses, permitting copying, as well as free commercial reuse and transformation, in exchange for simple attribution.) Since the licenses' launch in December 2002, nearly 800,000 Web pages (well over 1,000,000 discrete works) have been made available under Creative Commons licenses. Because they\'re free and can apply to any kind of copyrighted work, the licenses have been popular with Webloggers, teachers, novelists, musicians, photographers, and hobbyists. Many institutional adopters, too, have used the licenses to facilitate innovative publishing techniques, particularly in the sciences. The Massachusetts Institute of Technology\'s Open Courseware project publishes materials from its university courses under a version of the licenses, inviting students and educators from around the world to reuse them royalty-free. Rice University\'s Connexions project, an interactive tool that helps instructors build courses and texts from a collective knowledge repository, requires authors to license their contributions for free reuse in return for authorial attribution. The American Museum of Natural History\'s Biodiversity Commons will soon use the licenses to facilitate search across a broad collection of conservation databases and websites. Like PLoS, all of these projects use Creative Commons licenses to simplify and streamline the process of rights clearance. But the licenses also serve another critical function: they formalize the collaborative ethos of the scientific and academic communities in a language that legal intermediaries cannot quarrel with. This standardization also helps otherwise disparate communities, whether across disciplines or geographic boundaries, to agree in advance on the rules for sharing. Creative Commons is now considering expanding into other fields where the law has begun to restrict open research: scientific data and patents, in particular. With a portion of a new US\$1 million grant from the Hewlett Foundation (putting our total of funding received at over US\$3 million), we hope to build the Science Commons, a branch of the organization dedicated to bringing a measure of reason, and restraint, to the legal thicket that has grown around scientific research. Like PLoS, Creative Commons' goals and methods are designed to make the most of the opportunities created by new communications technology. But, also like PLoS, our inspiration reflects the wisdom and optimism of the Enlightenment as much as that of the Digital Age. We are trying to restore the sense of legal moderation that policymakers of a bygone era, heavily influenced by the philosophy of the first scientific revolution, understood would "promote the progress of science and the useful arts," as the U.S. Constitution puts it. "Knowledge \[is\] not the personal property of its discoverer, but the common property of all," wrote Benjamin Franklin, the great cosmopolitan, polymath, and patron saint of innovation. As we enjoy great advantages from the inventions of others, we should be glad of an opportunity to serve others by any invention of ours, and this we should do freely and generously." Franklin, who knew a thing or two about both publishing and science, never practiced law. ::: {#pbio-0000009-g001 .fig}  ::: Glenn Otis Brown is the executive director of Creative Commons (<http://www.creativecommons.org>), located in Palo Alto, California, United States of America. E-mail: <glenn@creativecommons.org>.

PubMed Central

2024-06-05T03:55:45.951655

2003-10-13

PMC212691

Introduction {#s1} ============ Considerable effort employing biochemistry, genetics, and pharmacology has been invested in identifying the web of interactions that characterize signal transduction pathways in metazoan organisms. Several conclusions can be drawn from these efforts. Despite the large number of receptors, ligands, and downstream targets, the number of signal transduction pathways in metazoans is relatively small, arguably less than 20 ([@pbio-0000010-Gerhart1]). This limited diversity occurs despite large numbers of different organisms, cell types, states of growth, and differentiation, as well as sexual dimorphism in biology. Remarkably, these pathways are highly conserved, some among all eukaryotes, most among all metazoans. Whereas signaling pathways differ in detail, it is not clear whether these differences are functionally significant. Conservation in the face of diversity of function raises the question of whether the behaviors of the pathway are in reality as similar as they seem when one compares more quantitative aspects of the signals and responses, such as amplitude, duration, and flux ([@pbio-0000010-Heinrich3]). Finally, the structure and design of the pathways are themselves a mystery. Is the structure of these conserved pathways so deeply embedded in other conserved process that it is difficult to change any interaction, or does conservation imply continuous selection for function ([@pbio-0000010-Gerhart2])? Many of these questions require a more quantitative understanding of the behavior of signaling pathways. Such information is rarely available. Most mathematical models have to be satisfied with a general conceptual understanding and are seldom testable, as most parameters must be assumed or inferred. It is partly for this reason that such theoretical efforts up to now have had limited impact on experimentalists, who prefer powerful qualitative tools to construct logical and formal models of pathway structures. Mathematical modeling is more advanced for metabolic networks, where the pathways have been known for more than a half-century and where more kinetic data have been available, including more recent data on in vivo dynamics ([@pbio-0000010-Heinrich2]). To develop a better quantitative understanding of a signal transduction pathway, we have recreated a more accessible system for biochemical study. The Wnt signaling pathway downstream of its immediate cytoplasmic mediator, Dishevelled (Dsh), can be reconstituted in frog egg extracts. The readout of the pathway is the rate of degradation of the transcriptional coactivator, β-catenin. We chose the Wnt pathway because it is active in the early Xenopus embryo, it is widely used in many different contexts in development, and it is also very important in human cancer. Although the design features of the Wnt pathway are highly conserved in evolution, it is not clear what purposes those features serve. This paper is in part an answer to that question. The pivotal player in Wnt signaling is the scaffold protein axin, which is required for the constitutive degradation of β-catenin. Axin coordinates the assembly of a large complex that includes the glycogen synthase kinase 3β (GSK3β); another scaffold protein, adenomatous polyposis coli (APC); and the negative regulators Dsh and GSK3β-binding protein (glycogen synthase kinase-binding protein \[GBP\]/Frat). Binding of Wnt to its receptor, Frizzled, activates Dsh through an as-yet-unknown process. In the absence of Wnt, GSK3β bound to axin phosphorylates β-catenin bound to both axin and APC. Phosphorylated β-catenin is a substrate for ubiquitination and subsequent degradation through the F-box protein β-TRCP, which is part of an SCF ubiquitin ligase complex. In the presence of the Wnt signal, the activated Dsh protein binds to axin and, through bound GBP, inhibits β-catenin phosphorylation; this inhibits its ubiquitination and consequent degradation. The buildup in β-catenin in the presence of a Wnt signal leads to transcription of specific genes. Numerous questions arise from this general model. Why are two scaffold proteins, APC and axin, both necessary? Do their roles differ? Recently it has been discovered that axin, like β-catenin, is an unstable protein ([@pbio-0000010-Yamamoto1]; [@pbio-0000010-Tolwinski1]). In recent work (unpublished data), we have further described the conditions under which axin is unstable. We ask here what role axin instability plays in the behavior of the Wnt pathway and in the responsiveness of the pathway to the Wnt signal? Beyond these mechanistic questions are important biological ones that lie beyond the scope of this work but that may be raised by some of the findings here. For example, mutations in APC seem to play a particularly important role in colorectal cancer; is the peculiar sensitivity to APC mutations in the colonic epithelium understandable in terms of how the pathway performs in that tissue? Similarly, GSK3β is also essential but not commonly mutated in colorectal cancer; why is that the case? No one has purified a discrete complex containing all the major players in the Wnt pathway arrayed on the axin--APC scaffolds, suggesting that the pathway might be affected by nonproductive titration of components by subcomplexes. If this is a problem, how is it avoided? The Wnt pathway is likely present in all cell types, and yet several of its constituents are used in other pathways; how is crosstalk or interference in other pathways avoided? As we set out to produce a realistic kinetic model of the Wnt pathway, we encountered other questions. For example, quantitative measurements led to the unusual finding that axin is present at very low concentrations. Is there a satisfactory explanation of this fact and of other, previously unexplained, features of the pathway? We are aware that some of the answers to specific questions could lie in unknown components or unknown interactions among known components. We were under no illusions that we could accommodate all known interactions in a model at this time or that we already knew all we need to know to construct such a model. For this reason we have asked a more modest question---whether the properties of the core pathway, as presently understood, can provide insight into important questions not yet answered or not yet clearly raised. To answer such questions, we collected what we initially thought were sufficient quantitative data on rates, affinities, and fluxes to derive a reference state model of the Wnt pathway in this system. The provisional reference state model reflected most of the known properties of the system, but from this model we identified several rates and affinity constants whose values were critical to the behavior of the model. We then went back and measured these. Thus this paper contains a reference model, a large number of experimental measurements needed to define this model, theoretical predictions, and experimental tests of those predictions, where possible. A general test of the validity of the model is its predictive ability under a wide range of conditions. From this analysis, several unexpected properties emerged with significance for understanding the biological function of the Wnt pathway. Results {#s2} ======= A Proposed Kinetic Pathway {#s2a} -------------------------- The model was based on the reaction scheme shown in [Figure 1](#pbio-0000010-g001){ref-type="fig"}. A few steps are labeled, such as the synthesis of axin and β-catenin, the degradation of axin, the axin-independent (basal) and axin-dependent degradation of β-catenin, as well as the critical cycle involved in the phosphorylation of β-catenin for degradation (Destruction Core Cycle). The output is the formation of the β-catenin/T-cell factor (TCF) complex and the input is the Wnt signal. Although many proteins interact with the Wnt pathway, we have focused only on core components known to be necessary for mediating a Wnt signal in most contexts. These core proteins include GSK3β, protein phosphatase 2A (PP2A), β-catenin, APC, axin, Dsh, TCF, and Wnt. The reactions incorporated into our model include protein synthesis/degradation, protein phosphorylation/dephosphorylation, and the assembly/disassembly of protein complexes ([Figure 1](#pbio-0000010-g001){ref-type="fig"}, solid arrows). Reactions mediated by proteins that activate a process are represented with broken arrows: (1) activation of Dsh by Wnt (step 1), (2) activation of the release of GSK3β from APC/axin/GSK3β by Dsh (step 3), and (3) activation of APC-dependent axin degradation (step 15). The reactions and components in blue are concerned with additional features of the pathway, as discussed below. ::: {#pbio-0000010-g001 .fig} Figure 1 ::: {.caption} ###### Reaction Scheme for Wnt Signaling The reaction steps of the Wnt pathway are numbered 1 to 19. Protein complexes are denoted by the names of their components, separated by a slash and enclosed in brackets. Phosphorylated components are marked by an asterisk. Single-headed solid arrows characterize reactions taking place only in the indicated direction. Double-headed arrows denote binding equilibria. Blue arrows mark reactions that have only been taken into account when studying the effect of high axin concentrations. Broken arrows represent activation of Dsh by the Wnt ligand (step 1), Dsh-mediated initiation of the release of GSK3β from the destruction complex (step 3), and APC-mediated degradation of axin (step 15). The broken arrows indicate that the components mediate but do not participate stoichiometrically in the reaction scheme. The irreversible reactions 2, 4, 5, 9--11, and 13 are unimolecular, and reactions 6, 7, 8, 16, and 17 are reversible binding steps. The individual reactions and their role in the Wnt pathway are explained in the text. :::  ::: The centerpiece of the model is the formation of the unstable core complexes involved in β-catenin phosphorylation and subsequent destruction. In addition to β-catenin, this set of complexes contains GSK3β and the scaffold proteins APC and axin. The complexes assemble in several steps: (1) binding of axin to APC (forward reaction of step 7); (2) binding of GSK3β (forward reaction of step 6); (3) phosphorylation of axin and APC by GSK3β (step 4). Dephosphorylation of the core complex (step 5) is mediated by PP2A. The first step in β-catenin degradation is its binding to APC\*/axin\*/GSK3β (step 8), after which it is phosphorylated by GSK3β (step 9) and released from the complex (step 10). Our model assumes that the phosphorylation of β-catenin by GSK3β is negligible in the absence of axin. Indeed, recent work indicates that axin stimulates the phosphorylation of β-catenin by GSK3β at least 24,000-fold ([@pbio-0000010-Dajani1]). Free, phosphorylated β-catenin is rapidly polyubiquitinated and degraded by the SCF complex and the proteasome, respectively (step 11). The dynamic properties of the model, such as the flux through the pathway, are also affected by binding of β-catenin to other partners, such as TCF (step 16) and free APC (step 17). In special cases (high axin concentrations), the flux through the system is affected by the binding of axin to GSK3β (step 19) as well as β-catenin (step 18). We have previously shown experimentally that TCF reduces the rate of β-catenin degradation ([@pbio-0000010-Lee1]). Turnover of β-catenin (steps 11, 12, and 13) and axin (steps 14 and 15) are included in our model, but since biochemical experiments in Xenopus egg extracts indicate that the turnover of GSK3β, Dsh, and TCF is relatively slow (no detectable degradation after 3 h at room temperature; unpublished data), the synthesis and degradation of these proteins are not explicitly modeled. The activation of the pathway in vivo, which results in the stabilization of β-catenin, is initiated by binding of Wnt ligands to Frizzled receptors and the subsequent transition of Dsh from its inactive form (Dsh~i~) to its active form (Dsh~a~). Since these events are still poorly defined, both processes have been combined in step 1. Interaction of Dsh~a~ with the nonphosphorylated complex APC/axin/GSK3β (step 3) activates the release of GSK3β . This latter process requires the activity of the GBP/Frat (not shown on our diagram). Deactivation of Dsh~a~ occurs through an as-yet-unidentified mechanism (step 2). Analytical Description {#s2b} ---------------------- The mathematical analysis is based on a series of balance equations that describe the concentrations and complexes of proteins in the Wnt pathway, as depicted in [Figure 1](#pbio-0000010-g001){ref-type="fig"}. The set of variables and the set of 15 differential equations we obtained are given in [Table S1](#st001){ref-type="supplementary-material"}, and [Dataset S1](#sd001){ref-type="supplementary-material"}, respectively (Equations \[A-1\]--\[A-15\]). Stimulation of the pathway by Wnt is described by a time-dependent function, *Wnt(t)*. Since Dsh, TCF, and GSK3β are degraded very slowly, we assume that their concentrations remain constant throughout the timecourse of a Wnt signaling event. The conservation equations for Dsh, TCF, and GSK3β are as follows: Symbols with the superscript \"0\" denote total concentrations. Since the concentration of axin is very low (see below) compared to the concentration of GSK3β, we replaced Equation (3) with the simple relationship *GSK3β* ^0^ = *GSK3β*. Similarly, we omitted the concentration of complexes containing axin in the conservation relationship for APC, which leads to the following equation: We will, however, take into account the contribution of axin-containing complexes for GSK3β and APC conservation equations when we later consider the effect of large increases in axin concentration. The simplest possible equation was chosen to describe the kinetics of each individual reaction. Synthesis of β-catenin and axin are described by constant rates **ν** *~i~*. Unimolecular reactions are assumed to be irreversible and are described by linear rate equations **ν** *~i~* = *k~i~* · *X~j~*, where *k~i~* denotes the first-order rate constant and *X~j~* denotes the concentration of the reactants. Reversible binding steps (double-headed arrows in [Figure 1](#pbio-0000010-g001){ref-type="fig"}) are described by the equation **ν** *~i~* = *k~+i~X~j~Y~l~* − *k~−i~*(*X~j~* **.** *Y~l~*), where *X~j~* and *Y~l~* denote the concentrations of the binding partners and (*X~j~* **.** *Y~l~*) the concentration of their complex. The Dsh-mediated release of GSK3β from the destruction complex is described by an irreversible reaction that is bimolecular in the concentrations of Dsh and the degradation complex. The model is simplified by assuming that the reversible binding steps between axin, β-catenin, APC, and TCF are very fast, such that the corresponding protein complexes are in rapid equilibrium, so that only the dissociation constants *K~i~* = *k* ~−*i*~/*k* ~+*i*~ are considered in the kinetic description of these steps. The conservation equations and the binding equilibria reduce the number of independent dynamic variables. Accordingly, the original set of 15 differential equations is transformed into a set of only seven ordinary differential equations coupled to four conservation equations and four relationships for binding equilibria. For a detailed mathematical description of the model and the procedure for reducing the number of systems variables, see [Dataset S1](#sd001){ref-type="supplementary-material"}. Experimental Evaluation of the Reference and Stimulated States {#s2c} -------------------------------------------------------------- We define the reference state as the absence of Wnt signaling (*Wnt* = 0). In this unstimulated stationary state, Dsh is inactive and does not affect the degradation complex. β-Catenin concentration is kept low by continuous phosphorylation and degradation. The reference state can be characterized by the special values for its rate constants, its equilibrium constants, and its conservation quantities. If one can obtain values for all of these system parameters, the model equations should allow for a straightforward calculation of the variables in the reference state. Currently, we have experimental data for many of these parameters (see below). For the remaining system parameters that were not directly measured, we were able to derive numbers based on experimental data of steady-state concentrations and fluxes. A number of parameters were set such that the results of the model were in agreement with previous experimental data, specifically with the experimentally determined rate of β-catenin degradation ([@pbio-0000010-Salic1]; [@pbio-0000010-Lee1]). Finally, a few parameters had to be estimated; the constraint used was that the resulting model should be compatible with the steady-state and flux values. [Table 1](#pbio-0000010-t001){ref-type="table"} lists the numeric values of all of the input quantities of the model. These quantities are either specific parameters, such as dissociation constants, or systemic properties, such as steady-state concentrations or fluxes, from which the other parameters have been derived. Both types of input quantities include experimental data as well as estimated values. The specific numerical values affect the model to differing degrees. In a later section, we analyze the effects of changing the values of the parameters around their reference numbers. The types of input data used for our modeling can be divided into five groups. ::: {#pbio-0000010-t001 .table-wrap} Table 1 ::: {.caption} ###### Numeric Values of Input Quantities of the Model for the Reference State :::  The data are grouped into concentrations of pathway components, dissociation constants of protein complexes, concentration ratios, fluxes and flux ratios, and characteristic times of selected processes. Experimental evidence for these data is discussed in the text. From these data, the following rates and rate constants are calculated: **ν** ~12~ = 0.42 nM · min^−1^ (rate of β-catenin synthesis), **ν** ~14~ = 8.2 · 10^−5^ · nM min^−1^ (rate of axin synthesis), *k* ~4~ = 0.27 min^−1^, *k* ~5~ = 0.13 min^−1^, *k* ~6~ = 9.1 · 10^−2^ nM^−1^ · min^−1^, *k~−~* ~6~ = 0.91 · nM^−1^ · min^−1^, *k* ~9~ = 210 min^−1^, *k* ~10~ = 210 min^−1^, *k* ~11~ = 0.42 min^−1^, *k* ~13~ = 2.6 · 10^−4^ min^−1^, *k* ~15~ = 0.17 · min^−1^. See Table S2, found at <http://dx.doi.org/10.1371/journal.pbio.0000010.st002>, for more precise numbers used in the calculations **Bold: Measured values,** *Italics*: Estimated values ::: The first group of input data contains both total concentrations (*Dsh* ^0^, *APC* ^0^, *TCF* ^0^, and *GSK3β* ^0^) and steady-state concentrations (*Axin* ^0^, *β-catenin* ^0^, *β-catenin*\*). The total concentrations of Dsh, TCF, GSK3β, axin, β-catenin, and APC in Xenopus egg extract were determined experimentally using Western blot analysis by comparing the intensity of the signal to that of known amounts of recombinant protein. The concentration of phosphorylated β-catenin w as estimated because we have not been able to directly determine its level in extracts. However, we estimate that this value is small compared to that of nonphosphorylated β-catenin for the following reasons: (1) Addition of axin to Xenopus extracts dramatically increases the rate of β-catenin degradation. Since the role of axin is to promote phosphorylation of β-catenin, which is subsequently degraded, this would suggest that normally a significant pool of β-catenin exists in the nonphosphorylated form. (2) Western blot analysis of Xenopus extracts demonstrates that only a small percentage (\<10%) of total β-catenin can be detected as migrating with a slower mobility, which likely represents the phosphorylated form of β-catenin. The second group of input data was experimentally obtained from measurements of rates of dissociation of protein complexes. Binding constants were calculated based on the assumption that association rates approached that of the diffusion limits for a typical protein in an aqueous solution. The ratio *K* ~17~/*K* ~8~ = 10 of the dissociation constants characterizing the binding of β-catenin to APC and APC\*/axin\*/GSK3β, respectively, is based on previous experimental results demonstrating that β-catenin has a 10-fold lower affinity for nonphosphorylated compared to phosphorylated APC ([@pbio-0000010-Salic1]). In addition, we have shown experimentally (unpublished data; see Materials and Methods) that phosphorylated β-catenin dissociates from axin more rapidly (reaction 10) than nonphosphorylated β-catenin. Once phosphorylated, β-catenin will thus dissociate from the axin complex and undergo polyubiquitination and proteolysis. The third group of input data consists of the two concentration ratios in the Destruction Core Cycle for complexes that either contain or lack β-catenin. The concentration ratio for the complexes that lack β-catenin is represented by the ratio of its phosphorylated versus nonphosphorylated forms and reflects the relative activities of its kinase(s) and phosphatase(s), respectively. By contrast, the concentration ratio of the two β-catenin-containing degradation complexes represents the relative activities of β-catenin phosphorylation and the rate of release of phosphorylated β-catenin from the complex. These parameters were chosen rather arbitrarily to indicate roughly equal kinase and phosphatase activities and yielded realistic values for the overall fluxes, given the known concentrations and kinetic rate constants. The fourth group of data includes the steady state flux **ν** ~11~ for the degradation of β-catenin via the Wnt pathway and the flux ratio **ν** ~13~ */* **ν** ~11~ describing the extent to which β-catenin is degraded via non-Wnt mechanisms (e.g., via Siah-1 and presenilin \[[@pbio-0000010-Liu1]; [@pbio-0000010-Matsuzawa1]; [@pbio-0000010-Kang1]\]). We have now measured this Wnt pathway--independent degradation in Xenopus extracts (see Materials and Methods; value shown in [Table 1](#pbio-0000010-t001){ref-type="table"}). The final group of input data consists of the characteristic time constant (τ) of selected processes. This is the time it takes for the concentration to drop to 1/e of its initial value. These characteristic times include τ*~K~* **~.~** *~P~* = 1/(*k* ~4~ + *k* ~5~) for the kinase/phosphatase cycle that mediates phosphorylation/dephosphorylation of both APC and axin in the degradation complex (steps 4 and 5), τ*~GSK3β~* **~.~** *~ass~* = 1 */* (*k* ~6~ *GSK3β* + *k* ~−6~) for the binding equilibrium of GSK3β with the APC/axin complex (step 6), and τ*~ax~* **~.~** *~deg~* for axin degradation (step 15). Values for the rate of axin degradation were determined directly from experiments performed in Xenopus egg extracts (unpublished data). Experiments to determine the rate of APC and axin dephosphorylation (τ*~K~* **~.~** *~P~* ≈ 2.5 min) were performed using in vitro ^32^P-labeled recombinant APC and axin. Radiolabeled proteins were added to Xenopus egg extracts, and the loss of radioactivity over time was assessed by SDS-PAGE and autoradiography ([@pbio-0000010-Salic1]). The legend to [Table 1](#pbio-0000010-t001){ref-type="table"} contains the values of rate constants calculated from the input quantities using the described system of equations. The values of all variables in the reference state are listed in the first column of [Table 2](#pbio-0000010-t002){ref-type="table"}. These values represent the steady state solutions of system equations using the data in [Table 1](#pbio-0000010-t001){ref-type="table"} as input quantities with the value of Wnt set at *Wnt* = 0. ::: {#pbio-0000010-t002 .table-wrap} Table 2 ::: {.caption} ###### Steady-State Concentrations of Pathway Compounds in the Reference State and in the Standard Stimulated State :::  Calculation of the concentrations is performed by the systems equations (see Dataset S1) under steady-state conditions. For the reference state (*W* = 0), the corresponding parameters given in [Table 1](#pbio-0000010-t001){ref-type="table"} are used. Values marked by "\#" represent steady-state concentrations, which also appear in [Table 1](#pbio-0000010-t001){ref-type="table"}. The additional input quantities required for the calculation of the standard stimulated state (*W =* 1) are taken as follows: *Dsh ~a~ / Dsh ~i~ = k* ~1~ */ k* ~2~= 10, **ν** ~−6~ */* **ν** ~3~ *= k* ~−6~ */* (*k* ~3~ *Dsh~a~*) = 0.2, and τ*~Dsh~* **~.~** *~act~* = 1 */* (*k* ~1~ *+ k* ~2~) = 5 min. This yields the following values for rate constants: *k* ~1~ = 0.18 min^−1^, *k* ~2~ = 1.8 · 10^−2^ min^−1^, *k* ~3~ = 5.0 · 10^−2^ nM^−1^ · s^−1^ ::: Using the reference state as a starting point, we consider other stationary states that are attained when the pathway is permanently stimulated. To describe the strength of Wnt stimulation, we introduce a dimensionless quantity *W = Wnt/Wnt* ^0^ that represents the ratio of the Wnt concentration with respect to its concentration *Wnt* ^0^ in a "standard" stimulated (signaling) state. *W =* 0 and *W =* 1 characterize the reference state and a standard stimulated state, respectively, with the hyperstimulated state defined as *W* \> 1. In order to calculate concentrations in the standard stimulated state, additional input quantities are required. These include the ratio of the active and inactive forms of Dsh (*Dsh~a~/Dsh~i~*), the relation between non-Dsh-mediated and Dsh-mediated release of GSK3β from the destruction complex (the flux ratio **ν** *~−~* ~6~/**ν** ~3~), and the characteristic time for the Dsh activation/inactivation cycle (τ*~Dsh~* **~.~** *~act~*). These values are not at present measurable. The values for these input quantities are listed in the legend of [Table 2](#pbio-0000010-t002){ref-type="table"}. In a later section, we analyze the effects of changes in these additional input quantities. By setting *W* = 1 and fixing all other parameters, we arrive at steady-state solutions of the systems equations (see [Dataset S1](#sd001){ref-type="supplementary-material"}, Equations. \[A-1\]--\[A-15\]), which yield the numerical variables for the standard stimulated state (listed in the second column of [Table 2](#pbio-0000010-t002){ref-type="table"}). A comparison of this state with the reference state shows that the concentration of free nonphosphorylated β-catenin increases by a factor of approximately 6, from 25 to 153 nM. Upon Wnt stimulation, the free phosphorylated β-catenin concentration decreases by 8%, from 1 nM to 0.92 nM. The increase in β-catenin levels reflects the decrease in its degradation caused by the reduction in the ability of GSK3β to phosphorylate it. The concentration of the β-catenin/TCF complex increases by a factor of 1.8. The large increase in β-catenin concentration shifts the binding equilibrium between APC and β-catenin and the concentration of free APC falls slightly. Total axin concentration decreased by a factor of 2.7 upon Wnt stimulation since addition of Dsh decreases the concentration of the various axin containing complexes. Remarkably, the steady-state concentration of free axin is constant during the transition from *W* = 0 to *W* = 1. This is due to the fact that under steady-state conditions, the rate of axin synthesis equals its degradation; the rate of synthesis (**ν** ~14~) is a fixed value and the rate of degradation depends solely on the concentration of free axin (and independent of other parameters such as binding constants and strength of the Wnt signal). As expected, simulations of increasing Wnt activation (0 ≤ *W* ≤ 1.4) on the steady-state concentrations of β-catenin and axin reveal a nearly hyperbolic saturation of increasing concentrations of nonphosphorylated and total β-catenin with increasing strength of Wnt stimulation. Furthermore, Wnt stimulation affects the steady-state concentrations of axin and β-catenin in an opposite direction (see [Figure S1](#sg001){ref-type="supplementary-material"}). β-Catenin Degradation: Comparison of Theory and Experiment {#s2d} ---------------------------------------------------------- To test whether the mathematical model represented the Wnt pathway under a variety of conditions, we ran through a series of simulations, all of which used the same set of parameters. From these we calculated simulated timecourses for β-catenin degradation under a range of different conditions (increased axin concentration, increased Dsh^a^ concentration, inhibition of GSK3β, increased TCF concentration) ([Figure 2](#pbio-0000010-g002){ref-type="fig"}A). We then tested the results using the previously described biochemical system ([@pbio-0000010-Salic1]; [@pbio-0000010-Lee1]), adding purified proteins or compounds at *t =* 0 ([Figure 2](#pbio-0000010-g002){ref-type="fig"}B). Simulations and experimental results are each shown as plots of total β-catenin concentration versus time. The agreement between theory and experiment is excellent. ::: {#pbio-0000010-g002 .fig} Figure 2 ::: {.caption} ###### Kinetics of β-Catenin Degradation: Simulation and Experimental Results \(A) Simulated timecourses of β-catenin degradation. The straight line for *t \<* 0 corresponds to the reference state of β-catenin using the parameters given in the legends of [Table 1](#pbio-0000010-t001){ref-type="table"} and 2. In vitro conditions are simulated by switching off synthesis of β-catenin and axin (**ν** ~12~ = 0, **ν** ~14~ = 0 for *t* ≥ 0). Curve a: reference case (no addition of further compounds); curve b: addition of 0.2 nM axin; curve c: addition of 1 μM activated Dsh (deactivation of Dsh was neglected, *k* ~2~ = 0); curve d: inhibition of GSK3β (simulated by setting *k* ~4~ = 0, *k* ~9~ = 0); curve e: addition of 1μM TCF. Addition of compounds (axin, Dsh, TCF) and inhibition of GSK3β was performed at *t* = 0. \(B) Experimental timecourse of β-catenin degradation in Xenopus egg extracts in the presence of buffer (curve a′), axin (curve b′: 10 nM), Dsh (curve c′: 1 μM), Li^+^ (curve d′: 25 mM), or Tcf3 (curve e′: 1 μM). :::  ::: The straight line for *t \<* 0 represents the reference state. The simulated reference state curve ([Figure 2](#pbio-0000010-g002){ref-type="fig"}A, curve a) for β-catenin degradation is calculated for *t \>* 0, at which there is an absence of protein synthesis for axin (**ν** ~14~ = 0) and β-catenin (**ν** ~12~ = 0). This reference curve is in close agreement with our experimental data ([Figure 2](#pbio-0000010-g002){ref-type="fig"}B, curve a′) with identical half-lives for β-catenin degradation (theoretical value of *t* ~½~ = 60.2 min versus experimental value of *t* ~½~ = 60 min). We examined a new state, where we have increased the amount of endogenous axin (0.02 nM) by 0.2 nM. As shown in [Figure 2](#pbio-0000010-g002){ref-type="fig"}A, curve b, the additional axin markedly accelerated β-catenin degradation (*t* ~½~ = 11.8 min) in agreement with the experimentally obtained values ([Figure 2](#pbio-0000010-g002){ref-type="fig"}B, curve b′; *t* ~½~ = 12 min). Theoretically, the effect of axin on β-catenin degradation is primarily due to the large concentration difference between the two scaffold proteins, APC and axin. Owing to the high concentration of APC, an increase in axin concentration results in a sharp increase in the concentration of the APC/axin complex, thereby accelerating β-catenin binding to the destruction complex. Curve d in [Figure 2](#pbio-0000010-g002){ref-type="fig"} shows the effect of inhibiting GSK3β on β-catenin degradation. This effect is produced in the simulation by inhibiting GSK3β activity (steps 4 and 9). Only a small fraction of β-catenin (phosphorylated β-catenin) is available for degradation after complete inhibition of β-catenin phosphorylation (step 9), so inhibition is rapid. This is in complete agreement with our experimental data in which degradation is essentially blocked after inhibiting GSK3β activity by lithium ([Figure 2](#pbio-0000010-g002){ref-type="fig"}B, curve d′). Curve e in [Figure 2](#pbio-0000010-g002){ref-type="fig"}A predicts that β-catenin degradation is strongly inhibited after the addition of 1 μM TCF. Previously we have shown that β-catenin is sequestered by TCF, thereby resulting in a significant decrease in free β-catenin ([@pbio-0000010-Lee1]). The addition of TCF would be expected to decrease the rate of β-catenin phosphorylation (step 9) and subsequently β-catenin degradation. This is also seen experimentally ([Figure 2](#pbio-0000010-g002){ref-type="fig"}B, curve e′). The immediate inhibition by LiCl is in contrast with the action of Dsh that inhibits only after a significant delay. We were intrigued by the theoretical biphasic degradation curves of β-catenin in the presence of Dsh~a~, as well as the experimental support for it ([Figure 2](#pbio-0000010-g002){ref-type="fig"}A and 2B, curves c and c′). In both cases, there is an initial rapid decrease in β-catenin in the first 30 min to 1 h, followed by a much slower decrease. Such a feature should allow us to distinguish mechanistic details of complex formation. Experimentally, the biphasic nature of Dsh activity is not due to a delay in Dsh activation upon its addition to the Xenopus extracts since we see the same effect with Dsh protein that has been "activated" with extracts prior to its use in our degradation assay. As shown in [Table 1](#pbio-0000010-t001){ref-type="table"}, the characteristic time τ*~K~* **~.~** *~P~* of phosphorylation and dephosphorylation of APC and axin in the destruction complex is relatively slow (2.5 min), and it therefore takes 5 min for 75% of the complex to be dephosphorylated. If Dsh~a~ acted only on the dephosphorylated complex (through step 3) to remove GSK3β and thus block phosphorylation of the complex, then we would predict the biphasic kinetics shown in [Figure 2](#pbio-0000010-g002){ref-type="fig"}A, curve c. These data suggest that Dsh inhibits the phosphorylation of the scaffold complex by GSK3β, but does not inhibit the phosphorylation of β-catenin. When Dsh binds, the complex can go around many times binding and phosphorylating β-catenin before it dissociates and is inhibited by Dsh. One hour after the addition of Dsh, β-catenin degradation is significantly inhibited due to the removal of a significant pool of GSK3β from the degradation complex over time (through the action of Dsh). As a result, the scaffold protein axin is dephosphorylated by the phosphatase (step 5) that remains bound to the degradation complex. Dephosphorylated axin is rapidly ubiquitinated and degraded when the β-catenin degradation normally stops. The small decrease in β-catenin levels in [Figure 2](#pbio-0000010-g002){ref-type="fig"}, curve c, after a 1 h incubation with Dsh, is due to degradation of β-catenin via nonWnt pathway mechanisms (see [Table 1](#pbio-0000010-t001){ref-type="table"}) that we have incorporated into our model. To test this prediction beyond consistency with experimental data, we performed an experiment in which Dsh was either preincubated with extract before or added at the same time as radiolabeled β-catenin ([Figure 3](#pbio-0000010-g003){ref-type="fig"}). If β-catenin and Dsh are added at the same time, there is an initial rapid loss of β-catenin ([Figure 3](#pbio-0000010-g003){ref-type="fig"}, curve b) followed by pronounced inhibition of degradation after 1 h. This initial rapid loss is consistent with Dsh acting on a subpopulation of degradation complexes (presumably the unphosphorylated forms). Strikingly, preincubation with Dsh prior to the addition of radiolabeled β-catenin ([Figure 3](#pbio-0000010-g003){ref-type="fig"}, curve a) results in immediate action of Dsh. We interpret this result to simply reflect the fact that over time in the preincubated extract Dsh can remove GSK3β from the degradation complexes, thereby enhancing the activity of the phosphatase and, as a result, promoting the degradation of axin and inhibition of β-catenin degradation. The small decrease in β-catenin levels at *t* \> 2 h in both curves a and b again suggests the existence of a slow degradation process mediated by non-Wnt pathway mechanisms. ::: {#pbio-0000010-g003 .fig} Figure 3 ::: {.caption} ###### Preincubation of Dsh in Xenopus Egg Extracts Abolishes the Lag in Dsh Activity Labeled β-catenin was incubated in Xenopus extracts on ice 30 min prior to (B) or 30 min after (A) the extract had been preincubated with 1 μM Dsh. No degradation of the labeled β-catenin was detected while the reactions were on ice. The reactions were started by shifting to 20°C. :::  ::: Clues to Axin Activity from Its Very Low Cellular Concentration {#s2e} --------------------------------------------------------------- In establishing quantities for our model in [Table 1](#pbio-0000010-t001){ref-type="table"}, we found that the axin concentration (20 pM) is much lower than the concentration of the other major components (β-catenin, 35 nM; APC, 100 nM; Dsh, 100 nM; and GSK3β, 50 nM). This unusual finding suggests that the function of the Wnt signaling system may actually depend on a low axin concentration. Our theoretical predictions for the effects of axin, GSK3β, and Dsh on the half-lives of β-catenin are shown in [Figure 4](#pbio-0000010-g004){ref-type="fig"}A and 4B, respectively. At zero concentration of Dsh, doubling the concentration of axin (from the reference state, indicated as 0, to a state where the concentration has been increased by 0.02 nM) causes a 50% drop in the half-life of β-catenin. By contrast, a doubling of the GSK3β concentration only decreases the half-life of β-catenin by 10%. The small effect of GSK3β is predicted to be due to the fact that only a limited amount of axin can be recruited to the degradation complex through binding to additional GSK3β. On the other hand, increased axin concentrations are immediately translated into an increased concentration of the destruction complex, because the concentrations of APC and GSK3β are high. Changing the concentration of either GSK3β or of axin should also change the amount of Dsh~a~ required to inhibit β-catenin degradation, but the pathway is much more sensitive to axin concentration than it is to GSK3β concentration. In the presence of high concentrations of axin, the effect of Dsh~a~ should be blocked; high concentrations of axin will lead to high concentrations of the phosphorylated destruction complex no matter what level of Dsh~a~ activity is present. High levels of the destruction complex will require even higher levels of Dsh to overcome the inhibition. The interaction between Dsh~a~ and GSK3β is similar in principle: Dsh-mediated release of GSK3β (step 3) from the degradation complex can simply be reversed by sufficiently high concentrations of GSK3β (step 6). In this case, however, the effect is small. Thus, axin blocks the action of Dsh so effectively that it renders the Dsh pathway inoperable. ::: {#pbio-0000010-g004 .fig} Figure 4 ::: {.caption} ###### The Effect of Dsh versus Axin or GSK3β on the Half-Life of β-Catenin in Xenopus Extracts (A and B) Predicted effects of Dsh, axin, and GSK3β on the half-life of β-catenin degradation. The half-lives are calculated from simulated degradation curves. Data are plotted as function of added Dsh (logarithmic scale) for various concentrations of axin (A) and GSK3β (B). (C and D) Measured effects of Dsh, axin, and GSK3β on the half-life of β-catenin degradation. Stimulation of β-catenin degradation by axin occurs throughout the range of Dsh concentrations tested. (C) Axin increases the rate of β-catenin degradation even in the absence of added Dsh. (D) Stimulation of β-catenin degradation by GSK3β is detected only at high concentrations of Dsh. No effects of GSK3β on β-catenin degradation can be detected at less than 30 nM added Dsh. There is a disparity between the concentrations of axin in the experimental and theoretical curves. We assume that this is most likely due to the specific activity of the expressed axin protein. :::  ::: In [Figure 4](#pbio-0000010-g004){ref-type="fig"}C and 4D, we studied experimentally the dose-dependent effects of Dsh, GSK3β, and axin on β-catenin degradation. These curves represent β-catenin half-lives for various concentrations of axin ([Figure 4](#pbio-0000010-g004){ref-type="fig"}C) and GSK3β ([Figure 4](#pbio-0000010-g004){ref-type="fig"}D) with varying concentrations of Dsh. The results are qualitatively similar to those predicted by the model. As expected, β-catenin degradation is inhibited by increasing Dsh concentration and stimulated by increasing the concentration of either axin or GSK3β. There are, however, two pronounced differences in the effects of axin and GSK3β on Dsh inhibition. Whereas axin activates β-catenin degradation over a wide range of Dsh concentrations ([Figure 4](#pbio-0000010-g004){ref-type="fig"}C), the effect of GSK3β becomes significant only at high concentrations of Dsh ([Figure 4](#pbio-0000010-g004){ref-type="fig"}D). Furthermore, the inhibitory effect of Dsh can be almost completely blocked by high concentrations of axin (10 nM). In contrast, GSK3β (1 μM) can only partially inhibit the strong inhibitory effect of Dsh on β-catenin degradation. Our experimental results, however, show a smaller effect on the half-life of β-catenin degradation at high concentrations of Dsh as GSK3β levels are increased. Also, the concentrations of added axin in the theoretical curve and the experimental curves are very different. The quantitative difference between the model and experimental may simply reflect the fact that the specific activity of our GSK3β and axin preparations (purified from *Sf9* cells and bacteria, respectively) may be low and that a significant fraction of the recombinant proteins may not be active. Alternatively, the low activity of GSK3β may point to an unidentified inhibitory activity present in our Xenopus egg extracts. Effects of Dynamic Changes in Protein Concentrations {#s2f} ---------------------------------------------------- The dependence of flux on the concentration of a pathway component is a measure of how much the flux is sensitively controlled by that component. In metabolic control theory, the normalized concentration-dependent parameters of the total flux known as control coefficients have been very useful in defining the characteristics of pathways ([@pbio-0000010-Heinrich1]; [@pbio-0000010-Fell1]). Similarly, in the analysis of bacterial chemotaxis, the response of a behavioral parameter as a function of changes in specific kinetic rates has been termed robustness ([@pbio-0000010-Alon1]). Such terms are rarely measured in signal transduction. To determine the effects of changes in the levels of Wnt pathway components, we analyzed how the flux (β-catenin degradation) changes with changes in the concentrations of *APC* ^0^, *GSK3β* ^0^, *Dsh* ^0^, and *TCF* ^0^ (see [Figure S2](#sg002){ref-type="supplementary-material"}). We chose to focus on the effects of changes in the concentrations of pathway components in the reference state, because similar effects were also seen for the stimulated state. Recently, we investigated a new and important property of the Wnt pathway, namely that the degradation of axin (reaction 15) is dependent on APC (unpublished data). The degradation rate of axin is mathematically expressed in the following manner: where *K~M~* represents a half-saturation constant for the activating effect of APC. The theoretical effect of APC on the concentrations of both β-catenin and axin is shown in [Figure 5](#pbio-0000010-g005){ref-type="fig"}, where we considered independently the effect of APC-mediated degradation of axin ("with regulatory loop" where Equation \[5\] is applied) or the absence of such an effect (where the linear rate equation **ν** ~15~ *= k* ~15~ *axin* is applied). With APC-mediated axin degradation, β-catenin degradation is affected very little by changes in the concentration of APC (25% decrease with a 2-fold increase in APC concentration). This resistance to changes of β-catenin levels upon changes in APC concentration is due to the APC-dependence of axin degradation (see [Figure 1](#pbio-0000010-g001){ref-type="fig"} and Equation \[5\]). Decreasing the concentration of APC inhibits the degradation of axin, thereby promoting the formation of the degradation complex. As shown in [Figure 5](#pbio-0000010-g005){ref-type="fig"}, in the absence of the regulatory loop, axin degradation is APC independent, homeostasis is lost, and β-catenin levels are greatly upregulated with decreasing APC concentrations. A comparison of the curves that represent the dependence and independence of axin degradation on APC (dashed lines in [Figure 5](#pbio-0000010-g005){ref-type="fig"}) indicates that the regulatory loop acts in such a way that the normally inhibitory effect on β-catenin degradation as a result of lowering the concentration of APC is counteracted by an increase in axin levels. ::: {#pbio-0000010-g005 .fig} Figure 5 ::: {.caption} ###### Effect of the Regulatory Loop for Axin Degradation The case "with regulatory loop" takes into account that axin degradation is APC-dependent (black curves). Alternatively, the case without this regulatory loop is considered (red curves). For the regulatory loop, the rate law (5) is used assuming that in the reference state the APC activation is half of its maximum (*K~M~* = 98.0 nM). The value of *k*′~15~ was chosen such that in the reference state both cases, with and without regulatory loop, yield the same degradation rate of axin (*k*′~15~ = 0.33 min^−1^). :::  ::: We have also simulated the effects of changes in the rate of β-catenin (**ν** ~12~) and axin (**ν** ~14~) synthesis on both β-catenin and axin levels (see [Figure S3](#sg003){ref-type="supplementary-material"}). Interestingly, changing the level of axin or β-catenin affects the concentration of the other component in different ways. An increase in the synthesis of axin results in a decrease in β-catenin, whereas increasing β-catenin synthesis leads to an increase in axin levels. This latter effect contrasts with effects observed upon changes of other parameters (see [Figure S2](#sg002){ref-type="supplementary-material"}) that affect the concentrations of axin and β-catenin in opposite directions. Transient Stimulation of the Pathway {#s2g} ------------------------------------ Wnt stimulation in vivo is transient, likely due to receptor inactivation/internalization and/or other downregulatory processes. We model transient Wnt stimulation by an exponential decay:  where the reciprocal of *λ* represents the characteristic lifetime τ*~W~* of receptor stimulation and *t* ~0~ denotes the onset of signaling. The concentration changes of all other pathway compounds resulting from Wnt stimulation can be calculated by numerical solution of the system equations (see [Dataset S1](#sd001){ref-type="supplementary-material"}), with initial values of the variables corresponding to the reference state. Regulating axin turnover is important for Wnt signaling. Wnt-stimulated axin turnover has been reported in cultured mammalian cells ([@pbio-0000010-Yamamoto1]) and in *Drosophila* ([@pbio-0000010-Tolwinski1]). In a future paper we will show that axin turnover is affected inversely to β-catenin turnover by phosphorylation by GSK3β. Here we show theoretically that this regulated axin turnover sharply affects the dynamics of the response. [Figure 6](#pbio-0000010-g006){ref-type="fig"} shows the time-dependent behavior of the total concentration of β-catenin and the total concentration of axin upon transient Wnt stimulation. The concentration of β-catenin increases transiently and then returns to its initial value. In contrast, the concentration of axin is temporarily downregulated. Further analysis of [Figure 6](#pbio-0000010-g006){ref-type="fig"} reveals that the amplitude of the β-catenin signal upon transient stimulation is significantly lower than the steady-state concentration upon permanent stimulation (*W =* 1; see [Figure S1](#sg001){ref-type="supplementary-material"}). The curves a and a′ in [Figure 6](#pbio-0000010-g006){ref-type="fig"} are calculated for the reference values of the rate of axin synthesis and of the rate constant of axin degradation, whereas the curves b and b′ and the curves c and c′ are obtained for the case where both parameters are increased by a factor of 5 and decreased by a factor of 5, respectively. Under these conditions, both the degradation rate and the synthesis rate are altered by the same factor, thus maintaining essentially identical steady-state concentrations of axin. As a result, the steady-state concentrations of axin are the same in the unstimulated condition (*W =* 0) and after diminution of the Wnt signal; however, during active signaling, the differences in the dynamic nature of signal output at differing rates of axin turnover are dramatically revealed. ::: {#pbio-0000010-g006 .fig} Figure 6 ::: {.caption} ###### Timecourse of β-Catenin and Axin Concentrations Following a Transient Wnt Stimulation Transient activation of the pathway is modeled assuming a Wnt stimulus that decays exponentially (Equation \[6\] with τ*~W~* = 1/*λ* = 20 min) starting at *t* ~0~ = 0. The straight line for *t* \< 0 corresponds to the steady state before pathway stimulation. The curves are obtained by numerical integration of the differential equation system (see Dataset S1). The various curves for β-catenin and for axin differ in the turnover rate of axin determined by the parameters **ν** ~14~and *k* ~15~ (curves a: reference values of these parameters; curves b: increase by a factor of 5; curves c: reduction by a factor of 5). All other parameters are given in the legend of [Table 2](#pbio-0000010-t002){ref-type="table"}. :::  ::: Interestingly, an increase in the turnover rate of axin leads to higher amplitudes and shorter durations of the β-catenin signal. This can be explained by the faster degradation of axin after its Dsh-mediated release from the destruction complex.Thus, β-catenin degradation is effectively inhibited for a certain time period due to a reduced availability of the scaffold axin. Since the steady-state concentration of free axin remains unchanged (rate of axin synthesis equals the rate of its degradation) during the transition from *W =* 0 to *W =* 1, a fast axin turnover favors rapid replenishment of the axin pool after the decline of the Wnt stimulus and, in this way, fast recovery of the destruction complex. This explains why the β-catenin signal is not only amplified, but becomes more spike-like. Increasing the turnover rate of axin affects the response of axin to temporary Wnt stimulation in a similar way as the response of β-catenin; i.e., the signal is amplified and sharpened ([Figure 6](#pbio-0000010-g006){ref-type="fig"}). Closer inspection of [Figure 6](#pbio-0000010-g006){ref-type="fig"} reveals that the axin response precedes the β-catenin response. For example, in the reference case, the β-catenin concentration reaches its maximum at about 260 min (curve a), whereas the minimum of the axin concentration is reached at 130 min (curve a′). This effect can be understood by observing that it is the lowering of the axin concentration that decreases the concentration of the destruction core complexes; in turn, this stabilizes β-catenin. Mechanistic Differences between APC and Axin as Scaffolds {#s2h} --------------------------------------------------------- As the axin concentration is several orders of magnitude lower than that of the other components in the degradation pathway (see [Table 1](#pbio-0000010-t001){ref-type="table"}), we decided to test the effect of increasing axin levels (up to, equal to, and greater than the concentrations of other components in the pathway). To do this, we had to extend the model to include additional reactions, marked in blue in [Figure 1](#pbio-0000010-g001){ref-type="fig"}; these had previously been neglected due to the very low axin concentrations. High axin concentrations affect most prominently the formation of the β-catenin/axin complex. Assuming a realistic value for the β-catenin--axin dissociation constant (*K* ~18~ *=* 1 nM), a moderate increase in axin concentration should theoretically accelerate β-catenin degradation, whereas a much higher concentration should result in inhibition of β-catenin degradation, due to the formation of partial complexes on axin. A more extensive analysis of β-catenin half-lives over a range of axin concentrations shows such a biphasic curve ([Figure 7](#pbio-0000010-g007){ref-type="fig"}A, curve b). These effects can also be seen experimentally in extracts ([Figure 7](#pbio-0000010-g007){ref-type="fig"}B), where 10 nM axin accelerates and 300 nM axin inhibits β-catenin degradation. The *t* ~½~decrease for low amounts of added axin can be easily explained by the fact that greater amounts of APC and GSK3β can be recruited to form the destruction complex. As a result, the *t* ~½~ decreases from 60 min to *t* ~½~ = 3 − 4 min. The inhibitory effect of axin becomes apparent only for axin concentrations approaching that of the other components. As shown in [Figure 7](#pbio-0000010-g007){ref-type="fig"}A, the effect of axin binding only to GSK3β (*K* ~19~ = 1 nM, *K* ~18~ → ∞) only becomes inhibitory at higher than micromolar concentration (curve c), whereas the combined effect of binding to both β-catenin and GSK3β (*K* ~18~ = 1 nM, *K* ~19~ = 1 nM) shows inhibition at less than 500 nM (curve d). If, however, we model an ordered process of binding to axin, then abortive inhibitory complexes cannot form. We show this in [Figure 7](#pbio-0000010-g007){ref-type="fig"}A. Here there is no separate binding of axin to β-catenin or GSK3β. In this case, there is no increase in the t~½~ at high axin concentrations ([Figure 7](#pbio-0000010-g007){ref-type="fig"}A, curve a). ::: {#pbio-0000010-g007 .fig} Figure 7 ::: {.caption} ###### Effects of Increasing Axin Concentration on β-Catenin Degradation \(A) Effect of axin concentration on β-catenin half-life. Curve a: reference case (*K* ~18~, *K* ~19~ \> 1 nM, ordered mechanism); curve b: *K* ~18~ = 1 nM, *K* ~19~ \> 1 nM; curve c: *K* ~18~ \> 1 nM, *K* ~19~ = 1 nM; curve d: *K* ~18~ = 1 nM. \(B) High concentration of axin inhibits β-catenin degradation in Xenopus egg extracts. Labeled β-catenin was incubated in Xenopus extracts in the absence (0 nM) or presence of moderate (10 nM) and high (300 nM) concentrations of axin. Moderate concentrations of axin greatly accelerate, whereas high concentrations inhibit β-catenin degradation. :::  ::: We also examined theoretically the effects of increasing APC concentration on the half-life of β-catenin, as shown in [Figure 8](#pbio-0000010-g008){ref-type="fig"}. The black curve corresponds to a nonordered mechanism, such as that found in axin, in which the β-catenin--APC dissociation constant (reaction 17) is low. The inhibitory effect of APC at high concentrations is due to its β-catenin buffering activity. The green curve corresponds to an ordered mechanism and reflects a high β-catenin--APC dissociation constant (high *K* ~17~). In this case, increasing concentrations of APC greater than the reference concentrations does not lead to inhibition of β-catenin degradation even at very high concentrations of APC. In cultured cells, overexpression of APC stimulates β-catenin degradation ([@pbio-0000010-Munemitsu1]; [@pbio-0000010-Papkoff1]). Unfortunately, we are presently unable to express full-length APC in Xenopus egg extracts to measure the effects of high levels in the extract system. ::: {#pbio-0000010-g008 .fig} Figure 8 ::: {.caption} ###### Effects of APC Concentrations on β-Catenin Degradation Effect of APC concentration on β-catenin half-life assuming an ordered (curve a) or nonordered mechanism (curve b: *K* ~17~ = 1,200 nM), respectively. :::  ::: β-Catenin can also be degraded by nonaxin-dependent mechanisms, which include Siah-1 and presenilin-mediated degradation. Though they are expected to contribute very little to the total flux through the pathway, the nonaxin-dependent processes may have very important influences under certain conditions. In our Xenopus system, these alternative pathways do not contribute greatly to the half-life of β-catenin. Experimentally, we have measured only a 1.5% contribution to total β-catenin degradation such that the half-life of β-catenin is 45 h when the axin-dependent processes are inhibited. If in some situations the nonaxin-dependent degradation contributed 10% to the flux, the half-life would be 6.3 h (*k* ~13~ = 1.83 · 10^−3^ min^−1^). The alternative pathways have very little effect on the half-life of β-catenin at normal and supranormal concentrations of APC. However, the effect of these alternative pathways becomes much more prominent when the APC concentration is lowered, a situation that may be significant under pathological conditions. As seen in [Figure 9](#pbio-0000010-g009){ref-type="fig"}A, when APC levels are at 50% of their normal concentration, there are dramatic differences in β-catenin concentration, depending on whether the alternative degradation pathway contributes to 1.5% or 10% of the total β-catenin degradation activity. The importance of the regulatory loop involving APC-mediated axin degradation is shown in [Figure 9](#pbio-0000010-g009){ref-type="fig"}B. In the absence of the regulatory loop, a significant inhibition of APC levels would strongly inhibit axin degradation, leading to a large increase in β-catenin levels. The control of β-catenin would be very brittle in this circumstance. However, by making axin degradation dependent on APC, a loss of APC would not stabilize axin levels, and the high axin levels would support continued degradation of β-catenin. This is the situation labeled "with regulatory loop" shown in [Figure 9](#pbio-0000010-g009){ref-type="fig"}B. The control of axin degradation could be a decisive factor in the response of the system to genetic or environmental effects on APC. ::: {#pbio-0000010-g009 .fig} Figure 9 ::: {.caption} ###### Effects of the Alternative β-Catenin Degradation Pathway and of Axin Degradation at Low Concentrations of APC \(A) The alternative β-catenin degradation pathway (axin independent) can have profound effects on β-catenin levels at low APC concentrations. Variations of β-catenin and axin resulting from changes in APC concentration were calculated from the standard stimulated state. Relative variations were plotted since variation in the share of alternative degradation (1%, 5%, and 10%) results in changes of the standard stimulated state (all parameters are constant). β-Catenin and axin levels for varying contributions of the alternative degradation pathway are as follows: 1.5%, β-catenin 178 nM, axin 0.00728 nM; 5%, β-catenin 151 nM, axin 0.00679 nM; 10%, β-catenin 125 nM, axin 0.00629 nM. \(B) Inhibition of axin degradation reduces β-catenin concentration after loss of APC. Plotted is the concentration of a potential proteasome inhibitor I (scaled to its inhibition constant, *K* ~I~) necessary to reduce β-catenin concentration to its original level, depending on the concentration of APC. :::  ::: Control, Modular Composition, and Robustness of the Wnt Pathway {#s2i} --------------------------------------------------------------- The model contains many parameters that affect the system behavior in different ways and to various extents. We can systematically investigate these parameters and look for those whose perturbation the system is most sensitive or most robust against. We focus on the concentrations of β-catenin and axin and calculate the responses in the total concentrations of these two compounds upon changes in the rates of the individual processes. For quantifying the effects of the rate constants *k~+i~* and *k~−i~*, we use control coefficients for the total concentration of β-catenin  and corresponding definitions for the control coefficients *C^axin^~±i~* for the total axin concentration. These coefficients, originally proposed for quantifying control in metabolic networks (for reviews, see [@pbio-0000010-Heinrich2]; [@pbio-0000010-Fell1]), describe the relative changes of the concentrations of the given compounds to relative changes of the rate constants. The control coefficients for the reference state are listed in [Table 3](#pbio-0000010-t003){ref-type="table"}. It should be remembered that the following discussion refers to small perturbations of the reference state. ::: {#pbio-0000010-t003 .table-wrap} Table 3 ::: {.caption} ###### Control Coefficients for the Total Concentrations of β-Catenin and Axin and Parameters Quantifying the Sensitivity and the Robustness of the Wnt/β-Catenin Pathway :::  The control coefficients (Equation 7) were obtained by numerical determination of the response to a change of the rate constants of all steps by 1%. Using relative changes of rate constants less than 1% does not lead to a significant improvement of the precision of the *C* values. Coefficients are given for the reference state. Horizontal lines separate the coefficients for distinct modules of the pathway. The last block contains the coefficients for parameters that enter the systems equations as binding rate constants *k~+j~* and dissociation rate constants *k~−j~*via dissociation constants *K~j~* = *k~−j~ / k~+j~*. The upper signs of these coefficients refer to changes in *k~+j~* and the lower sign to changes in *k~−j~*. The sum of the control coefficients in each column is zero. Additional summation rules hold true for the rate constants within each module as well as for the two rate constants of each binding equilibrium. The standard deviation *σ* of the concentration control coefficients and the robustness *ρ* for β-catenin and axin are calculated by applying Equations (8) and (9) ::: For the reference state, there are six steps exerting strong negative control on the total β-catenin concentration (*C^βcat^~i~* ≅ −1). This group includes the reactions participating in assembling the destruction complex APC\*/axin\*/GSK3β. The corresponding parameters involve the rate constants *k* ~7~ for the binding of axin to APC, *k* ~6~ for the association of GSK3β to the APC/axin complex, and *k* ~4~ for the phosphosphorylation of axin and APC in the destruction complex. Similar strong negative control is exerted by β-catenin binding to the phosphorylated destruction complex (rate constant: *k* ~8~), the phosphorylation of β-catenin in the destruction complex (rate constant: *k* ~9~), and the synthesis of axin (**ν** ~14~). Six other reactions exert strong positive control in the reference state on the total concentration of β-catenin (concentration (*C^βcat^~i~* ≅ 1). To this group belong the reactions participating in the disassembly of the destruction complex APC\*/axin\*/GSK3β, which are described by the rate constants *k~−~* ~7~ for the dissociation of the APC/axin complex, *k~−~* ~6~ for the dissociation of GSK3β from the destruction complex, and *k* ~5~ for the dephosphorylation of the APC and axin in the destruction complex. Other steps with a high positive control are the dissociation of β-catenin from the destruction complex (rate constant: *k~−~* ~8~), axin degradation (rate constant: *k* ~15~), and β-catenin synthesis (**ν** ~12~). There are many reactions exerting almost no control on β-catenin levels in the reference state. This group includes binding of β-catenin to TCF and APC (*k* ~16~ and *k* ~17~), and the corresponding dissociation processes (*k~−~* ~16~ and *k~−~* ~17~; again only valid for small perturbations). Interestingly, the effects of the two β-catenin degradation processes (rate constants: *k* ~11~ and *k* ~13~) are also small. Calculation of control coefficients for the standard stimulated state reveals that some steps that exert no control in the reference state become important. These are the activation and inactivation of Dsh (rate constants: *k* ~1~ and *k* ~2~) and, more pronounced, the Dsh-mediated release of GSK3β from the destruction complex (*k* ~3~). For all other processes, the signs of the control coefficients for β-catenin and axin do not change at the transition from the reference state to the standard stimulated state. The effects of parameter changes on axin are generally opposite to those on β-catenin; i.e., processes with a positive control coefficient for β-catenin have negative control coefficients for axin and vice versa. A significant exception is the synthesis of β-catenin, which exerts a positive control not only on β-catenin but also on axin, as expected from the results obtained in the last section. Closer inspection of [Table 3](#pbio-0000010-t003){ref-type="table"} reveals that the values of the control coefficients for the rate constants sum up to zero. This fact is known as the summation theorem for concentration control ([@pbio-0000010-Heinrich1]) and is valid for all reaction networks at steady state. This result finds its explanation in the invariance of the steady-state concentrations against simultaneous change of all rate constants by the same factor. Interestingly, in the present case there are subgroups of processes whose control coefficients separately sum up to zero, indicating a modular structure of the pathway. In [Table 3](#pbio-0000010-t003){ref-type="table"}, the control coefficients of the different modules are separated by horizontal lines. The main four subgroups are the Dsh module (not shown in [Table 3](#pbio-0000010-t003){ref-type="table"}), the kinase/phosphatase module, the β-catenin module, and the axin module. A subgroup is defined by a set of reactions where the control coefficients of the binding reactions are opposite to those of the corresponding dissociation reactions (*C~+i~* = −*C~−i~* for *i* = 6, 7, 8, 16, 17). For those more familiar with genetic manipulation, it is more common to vary the concentrations of individual components rather than vary the rate constant of a reaction. [Table 4](#pbio-0000010-t004){ref-type="table"} shows the control coefficients for β-catenin and axin calculated for changes in the total concentrations of pathway components instead of the rate constants. Using the values of [Table 4](#pbio-0000010-t004){ref-type="table"}, the potential tumor-supressing effects (of APC, GSK3β, and axin) and potential oncogenic effects (of PP2A, TCF, Dsh, β-catenin) can be explained and quantified. It may be worth mentioning that there is no summation theorem for the control coefficients when calculated by changing total concentrations instead of rate constants. For practical reasons, it may be easier to discuss the coefficients with respect to concentration changes ([Table 4](#pbio-0000010-t004){ref-type="table"}); for theoretical reasons, changing rate constants are simpler to handle. We think that eventually it will also be clearest to speak about oncogenic reactions instead of oncogenic genes, especially if we are thinking of oncogenesis in response to pharmacologic or environmental perturbations. Genetic defects then can be considered in terms of changes in activity, transcription, translation, or proteolysis. ::: {#pbio-0000010-t004 .table-wrap} Table 4 ::: {.caption} ###### Concentration Control Coefficients for the Total Concentrations of β-Catenin and Axin Relative to Changes in the Concentrations of Pathway Components :::  The control coefficients were obtained by numerical determination of the response to a change of total concentrations by 1%. Coefficients are given for the reference state and for the standard stimulated state ::: Clearly, the robustness of a variable towards parameter changes is higher the lower the corresponding concentration control coefficient. To arrive at an estimation of the overall effects of parameter perturbations on the system as a whole, we consider first the standard deviation *σ* of the control coefficients from their mean value. According to the summation theorem, the mean value of all control coefficients for a given variable is zero. Thus, we get for the standard deviation for the control coefficients of β-catenin: where the summation is performed over all reactions, including forward and backward steps of fast equilibria. High values of *σ* indicate that the given variable is on average very sensitive towards changes of rate constants. We propose to introduce a measure for the robustness *ρ* of a variable towards changes of all parameters in the following way: As *σ* may vary between zero and infinity, the range of *ρ* is confined to the interval 1≥ *ρ* ≥ 0. High values of *ρ* resulting from low *σ* values for the control coefficients indicate that the variable is robust against parameter perturbations. The standard deviations *σ* of the control coefficients and the *ρ* values for β-catenin and axin are presented in the last two rows of [Table 3](#pbio-0000010-t003){ref-type="table"}. Because many control coefficients are close to zero and the absolute values of the others hardly exceed unity, the *σ* values for β-catenin as well as for axin are rather small. Since all values for *σ* are lower than unity, a 1% change in a rate constant leads, on average, to a response of \<1% in the overall level of β-catenin. The total concentration of axin is more robust against parameter perturbations than the total concentration of β-catenin, particularly in the standard stimulated state. A transition from the reference to the standard stimulated state results in a lower robustness for β-catenin and a higher robustness for axin. Discussion {#s3} ========== Theory and quantitation are mutually dependent activities. It would seem unlikely that one would go to the trouble to measure detailed kinetic quantities without a specific model to test, and it is equally unlikely that realistic models can be constructed without the constraints of quantitative experimental data. Our intent in trying to reproduce a substantial part of the Wnt pathway in Xenopus egg extracts was to acquire the kind of detailed kinetic data required to build a realistic model. There are several unusual advantages to the extract system that contributed to this effort. The Xenopus egg extract is essentially neat cytoplasm; it reproduces the in vivo rate of β-catenin degradation and responds to known regulators as expected from in vivo experiments. Kinetic experiments with high time resolution are possible in this system, since a well-stirred extract is presumably synchronous in ways in which collections of cells may not be. In extracts it is possible to precisely set the level of components by depletion or addition. The direct output of the canonical Wnt pathway is an easily measured cytoplasmic event, the degradation of β-catenin. Thus, in this unusual system it is possible to acquire quantitative information about signaling pathways, not achievable in vivo. At the same time, these extracts have limitations. We have not considered the receptor events, and it is likely that reactions at the plasma membrane contribute to dynamic features. Also, our analysis is incomplete, as there are other components of Wnt signaling, such as casein kinase I*δ*, casein kinase I*ɛ*, and PAR1, as well as cross-talk from other pathways, that influence the behavior of the system. We have also oversimplified the multiple phosphorylation steps. We have assumed a simple interconversion of the phosphorylated and unphosphorylated complex of axin, APC, and GSK3β, whereas in reality multiple phosphorylation states exist within the complex; the states may be random or sequential. We simply do not have the information to provide a much more specific model of phosphorylation interconversions at this time, although the model could easily be extended. Finally, there is the question of what Wnt process we are studying. We are looking at events in the cytoplasm of unfertilized eggs. Though endowed with all of the core components of the Wnt pathway, the egg is, as far as we know, transcriptionally silent and not involved in Wnt signaling, though this system is active very soon in embryogenesis. Thus, there is no biological in vivo behavior with which to compare the in vitro behavior. Nevertheless, the basic core circuitry is intact and is presumably prepared for the early Wnt events in the embryo. All the properties of the egg extract system are very similar to that circuitry in vertebrate somatic cells. To build a mathematical description of the Wnt signaling system, we started with the basic circuitry discerned from previous studies in Xenopus embryos and mammalian cells, whose similarity to the in vitro system we had already confirmed. We derived a system of differential equations that described the time-dependent variations of the system variables, i.e., the concentrations of the pathway components and their complexes. Parameters of the model are binding constants of proteins, rate constants of phosphorylations and dephosphorylations, rate constants of protein degradation, and rates of protein synthesis. Model reduction was achieved by considering conservation relations and by applying rapid equilibrium approximations for selected binding processes. Despite these simplifications, the model consists of a nonlinear system of differential equations whose solution requires the use of computers. Not all of these parameters were accessible to measurement. To circumvent this problem, we used as primary inputs not only kinetic parameters characterizing individual steps, but quantities that are more easily accessible from experiment, such as the overall flux of β-catenin degradation. This allowed us to derive rate constants as well as protein concentrations in a reference state, where there was no Wnt signal. This state serves as a starting point for predicting the system behavior during Wnt signaling as well as after experimental perturbations. The basic model reproduced quantitatively the behavior of the reference state, including perturbations of this state achieved by varying the concentration of axin, GSK3β, and TCF. It also reproduced extensions of this to the signaling state. A wide variety of different sets of experimental data could be simulated by the same model, employing the same sets of kinetic parameters. We approached this process iteratively. For example, the early model did not include nonaxin-dependent degradation of β-catenin, but inclusion of this process improved the fit to the experimental data. More significantly, addition of this process had interesting biological implications, which we discuss. In many ways, one of the most peculiar findings was the very low concentration of axin in the Xenopus extracts. Axin levels in other organisms may similarly be very low: *Drosophila* axin can be detected by Western blotting only following its immunoprecipitation ([@pbio-0000010-Willert1]). Although our theoretical and experimental studies have shown that axin is inhibitory at high concentrations, both indicate that axin is not present at the optimal concentration for the highest rate of β-catenin degradation. Therefore, axin levels are not set for optimality of β-catenin degradation, but are presumably optimized for some other purpose. Theoretically, axin levels must be held below the very sharp threshold of Dsh inhibition. Experimentally, these thresholds, which blunt Wnt signaling, are observed but are not as sharp as expected, and this may indicate some other compensatory effects. These thresholds would limit axin concentration to well below 1 nM if activated Dsh were constrained to concentrations of below 1 μM. Under these circumstances, we can expect that axin would never be found at concentrations approaching those of other Wnt pathway components (50--100 nM). The low concentration of axin relative to other components (such as GSK3β, Dsh, and APC) has another design feature potentially very general and important for the modularity of metazoan signaling pathways. Axin is a critical node point for controlling β-catenin levels, but it also interacts with components shared with several other important pathways. The interaction of these components with axin fluctuates due to Wnt signals (reflecting changes in binding as well as changes in axin levels), yet because the concentration of axin is so low, there will be no appreciable change in the overall levels of GSK3β, Dsh, or APC (all these components important in other pathways would otherwise be driven to fluctuate). The very low axin concentration thus isolates the Wnt pathway from perturbing other systems, a simple mechanism to achieve modularity. Other scaffold proteins may serve similar functions in other pathways. These insights follow from a very simple measurement of axin concentration and suggest the utility of measuring the levels of signaling pathway components in different cell types and circumstances. Since quantitative and kinetic features may be important in defining modules, it suggests that qualitative circuit diagrams of signal transduction may overlook very important design features. Modularity within the Wnt pathway can be defined by an extension of the summation theorem of [@pbio-0000010-Heinrich1]), which argues that the steady state of an entire pathway would have control coefficients that added to zero. When the Wnt pathway is broken down to several subpathways, we find that within these subpathways the control coefficients would sum to zero at steady state. While some of this subdivision is obvious (i.e., the kinase phosphatase module involving the phosporylation of APC and axin complexed to GSK3β), in other cases, such as the β-catenin module, it is much less obvious. Here the reactions include the phosphorylation of β-catenin in the APC/axin/GSK3β complex, the release and degradation of β-catenin, and the synthesis and nonaxin-dependent degradation of β-catenin. Balanced perturbation of these subpathways as a whole will not affect the overall flux of β-catenin degradation. It is not clear whether this concept of modularity might be extended usefully in two other directions: modularity in systems not at steady state, i.e., with transients, and estimates of linkage between pathways by some definition of nonzero summations expressing the degree of independence or modularity. In addition to work by [@pbio-0000010-Kholodenko1]), this paper marks one of the first extensions of metabolic control theory to signal transduction. Metabolism and signal transduction seem very different, the former involving the transfer of mass and the latter the transfer of information. In addition, metabolic pathways generally involve dedicated components and the specificity of interaction of substrates and enzymes is very high. Signaling pathways share many components; interactions are often weak. Metabolism, which has had a long history of quantitative study, was a natural field for the development of control theory, and this theory has been successful in converting the specific information about the behavior of enzymes in a pathway to the overall behavior of metabolic circuits. Control coefficients are useful measures of the impact of a process or quantity on another. In its application to metabolism, it allowed us to dispose of erroneous concepts, such as the notion of a rate-determining step. In signal transduction, control coefficients might play a similar role. Here they can be used to indicate quantitively the effects of a particular reaction on some other property, such as flux through the pathway or concentration of another component. By this definition, certain rate constants, such as the phosphorylation and dephosphorylation of APC and axin, have a major influence on the levels of β-catenin, while others, such as the degradation rate of phosphorylated β-catenin, have little effect. The sign and magnitude of these control coefficients give some indication what gene products could be oncogenes or tumor suppressors. As shown in [Table 4](#pbio-0000010-t004){ref-type="table"}, by this criterion APC, GSK3β, and axin are potent tumor suppressors, whereas β-catenin is an oncogene. Dsh would be expected to exert only moderate oncogenic effects. Clearly the effects of certain gene products are dependent on context, including their rate of synthesis and steady-state concentration. As our understanding of pathways improve, the effect of mutation or pharmacologic inhibition could be estimated quantitatively using control coefficients. The differences between cell types and organisms could be exploited to better predict mutagenic and pharmacologic impact on signal propagation. Despite considerable progress in identifying components of the Wnt pathway, many important mechanistic details are still lacking. In this analysis we have shown that Dsh seems to act to prevent the phosphorylation of the axin/APC complex, not the phosphorylation of β-catenin. Dsh (complexed to GBP) does not seem to be a general GSK3β inhibitor, like Li^+^, but rather is focused on the two scaffolding proteins. This was apparent from the biphasic nature of both the theoretical and experimental curves, which suggested that Dsh inhibited the rephosphorylation of axin/APC, but still allowed many cycles of β-catenin phosphorylation, ubiquitination, and degradation. This mechanism was further proven by a timing-of-addition experiment. It needs to be further confirmed and extended by looking specifically at individual phosphorylation sites on all the components of the complex. Another insight into the mechanisms of complex formation and control of β-catenin degradation concerns the inhibition of β-catenin degradation at concentrations of axin approaching those of other components. This suggests that axin binds APC, GSK3β, and β-catenin in random order. As discussed above, the axin concentration is limited by other factors; owing to the low concentration of axin, random binding is not likely ever to be a problem. The situation for APC seems very different. The concentration of APC is comparable to that of the other components. Overexpression studies show no inhibitory effects. These theoretical and experimental observations suggest that APC as a scaffold must be very different from axin as a scaffold. Most likely, APC binds components in an ordered manner. Metabolic pathways are understandable in terms of the familiar logic of chemical synthesis; signal transduction pathways, by contrast, often do not seem to conform to simple design principles. It is not clear at all whether signaling pathways have been optimized for a specific function or instead whether they are remnants of some early and rather arbitrary evolutionary experiment, now embedded in other processes that are difficult to change. Systems analysis, along with experiment, offers some hope of uncovering latent principles of design. For example, the modeling of the Wnt pathway gave a theoretical insight into the function of axin degradation in Wnt signaling. The degree of axin instability dramatically affects the amplitude and duration of the β-catenin response to a transient Wnt signal. If axin turnover were designed to be slow, then β-catenin would rise slowly to a low amplitude and persist for many hours. In a system where axin turnover was more rapid, the amplitude could increase several-fold and would persist a shorter time. The duration and amplitude of the response are likely to be important factors in developmental systems, which may respond differently to different amplitudes and durations of a signal. In addition, some developmental processes occur with such rapidity that the same signal would be interpreted differently at different times, hence the need to quickly terminate a signal. The very different effects of transient and persistent signals in the same pathway have been studied in PC12 cells in the MAP kinase pathway activated by EGF or NGF ([@pbio-0000010-Marshall1]). Finally, APC-dependent axin degradation stabilizes the Wnt pathway to variations in the APC concentration. Viewed from this perspective, the regulatory loop involving APC and axin degradation is an important design feature of the Wnt pathway. Robustness has also been considered an important design principle for signaling processes ([@pbio-0000010-Alon1]), and control coefficients can be a good measure of this robustness. We present here a general measure of robustness of the entire pathway, using a measure that sums the variation in every individual reaction. Though it is generally thought that "robust" is good, the complement of robust is adaptability, and it may be that some aspects of a signaling pathway are designed to be responsive to changes in some parameters so that the same pathway can be used differently in different circumstances by altering key parameters. Since quantitative measures of the concentration or posttranslational modification of signaling proteins are rare in the literature, we have very little information on whether the organism varies certain key components to achieve different behavior of the signaling systems. Another aspect of robustness is susceptibility to mutation or pharmacologic or environmental perturbation. The unexpected minimal phenotypes observed in numerous mouse knockout experiments have underscored our ignorance of the adaptable responses of organisms and in particular the adaptable nature of signaling pathways. One unexpected theoretical observation in this paper was the potential importance of the nonaxin-dependent degradation for β-catenin, under conditions where the APC levels are reduced. In our model, the nonaxin-dependent degradation contributed only a few percent to the overall flux of β-catenin degradation. Yet if the APC levels fall only 50%, the exact level of the alternative pathway made a large difference in the steady-state level of β-catenin. If the activity of the alternative pathways varied in different tissues, then this simple but largely silent effect could explain the tissue specificity of APC mutations. Similarly, variations in the alternative pathway might also explain some aspect of individual risk to loss of a single copy of the APC gene. An experiment can better be judged by how many questions it raises than by how many it answers. The same may be said about theoretical analysis. Such analysis is always a work in progress, in that the experimental basis is continually changing to some degree. Some of the experimental changes, though significant mechanistically, may have little effect on the model and its interpretation. Some may require major revision. In the case of the Wnt pathway, the theoretical analysis and modeling have already raised several interesting questions of biological importance. They have already stimulated further experimentation. More than anything, the modeling has increased the urgency for obtaining accurate quantitative information about both steady-state and transient processes in Wnt signaling and for obtaining information about the differences in parameters in different tissues and in different organisms. Materials and Methods {#s4} ===================== {#s4a} ### Egg extracts and degradation assays. {#s4a1} Xenopus egg extracts were prepared as described previously ([@pbio-0000010-Salic1]). Extracts were used either used immediately or stored at --80°C after being snap-frozen in liquid nitrogen. β-Catenin degradation assays were performed as described before ([@pbio-0000010-Salic1]). ### Measurement of β-catenin synthesis. {#s4a2} Freshly prepared Xenopus egg extracts were either supplemented with 25 mM LiCl in Xenopus buffer (XB) or XB and an aliquot were withdrawn for β-catenin degradation assays. The free methionine pool in Xenopus embryos is approximately 90 μM, and based on this number, \[^35^S\]methionine was added to the extract to a give a final activity of 1,000 counts per picomole methionine. Extracts were incubated at 20°C, and aliquots were withdrawn at the indicated times for SDS-PAGE, trichloroacetic acid (TCA) precipitation, and β-catenin pull-downs. To assay protein synthesis in the extract, total methionine incorporation was measured by TCA precipitation. In brief, Xenopus extracts (2 μl), metabolically labeled with \[^35^S\]methionine as above, were diluted to 100 μl with PBS and supplemented with 2 μl of 2% deoxycholate and TCA to 5%. The reaction was pelleted at 20,000 × g for 10 min at 4°C. After washing with ice-cold acetone and air drying, the radioactivity of the pellet was measured in a scintillation counter. To isolate metabolically labeled β-catenin from Xenopus extracts, we used His-tagged APCm3 cross-linked to Ultralink beads (Pierce, Woburn, Massachusetts, United States). Since phosphorylated APC has a much higher affinity for β-catenin, APCm3 beads were first phosphorylated with 300 nM His-tagged GSK3β in 25 mM HEPES (pH 7.7), 1 mM EDTA, 300 mM NaCl, 10 mM MgCl~2~, 1 mM DTT, and 1 mM ATP for 1 h at room temperature with shaking. The beads were washed three times with 25 mM HEPES (pH 7.7), 1 mM EDTA, 300 mM NaCl, 1% Tween, and 1 mM DTT and were then used to pull down β-catenin. Extracts (50 μl) labeled with \[^35^S\]methionine (see above) were diluted 5× with XB containing 1% Tween and protease inhibitors. The diluted extracts were incubated with phosphorylated APCm3 beads (20 μl) at 4°C for 2 h. The beads were washed and bound protein was eluted by boiling in SDS-PAGE loading buffer. Labeled β-catenin was detected following SDS-PAGE and autoradiography. ### Measuring dissociation rates of phosphorylated/nonphosphorylated β-catenin from axin. {#s4a3} Radiolabeled β-catenin (5 μl) was phosphorylated using 300 nM His-tagged GSK3β and 100 nM maltose-binding protein (MBP)--axin in 20 μl of XB containing 10 mM ATP, 20 mM MgSO~4,~ and 50 mM NaCl. For the nonphorphorylated control, β-catenin was incubated as above, except that 50 mM LiCl was used instead of NaCl to inhibit GSK3β. The kinase reactions were incubated in a shaker for 30 min at 20°C and then added to 50 μl of MBP--axin, bound to amylose beads (1 mg of protein per milliliter of beads), and brought up to 250 μl with XB containing 50 mM LiCl. After phorsphorylated and nonphosphorylated β-catenin, respectively, bound to axin beads, the beads were washed three times with 500 μl of XB containing 50 mM LiCl. Dissociation of the bound β-catenin was initiated by adding 1 μM unlabeled recombinant β-catenin (His-tagged, from *Sf9* cells) and incubated at 20°C in a shaker. At the appropriate times, 5 μl aliquots of beads were quickly removed, filtered through Wizard minicolumns (Promega, Madison, Wisconsin, United States), and washed with ice-cold XB (3 ml). Proteins bound to beads were eluted from the minicolumns with 20 μl of hot sample buffer, followed by SDS-PAGE and autoradiography. ### Recombinant proteins. {#s4a4} The expression and purification of all recombinant proteins have been previously described ([@pbio-0000010-Salic1]). Dsh was expressed as an MBP fusion in bacteria. His-tagged GSK3β, His-tagged Tcf3, His-tagged APCm3, and MBP--axin were expressed in *Sf9* cells. Supporting Information {#s5} ====================== Dataset S1 ::: {.caption} ###### The Roles of APC and Axin Derived from Experimental and Theoretical Analysis of the Wnt Pathway (287 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S1 ::: {.caption} ###### Effect of Wnt Stimulation on the Concentrations of β-Catenin and Axin The curves represent steady-state concentrations of β-catenin (solid lines) and axin (broken lines) as functions of the strength *W* of Wnt stimulation. Curve a: free unphosphorylated β-catenin; curve b: free phosphorylated β-catenin; curve c: total β-catenin; curve d: total axin. All concentrations are scaled with respect to their values in the reference state. It is worth mentioning that in the model "without regulatory loop," the steady-state concentration of free axin is determined by the condition *X* ~12~ *=* **ν** ~14~ */k* ~15~, and is, therefore, independent of Wnt stimulation. (2,954 KB TIFF). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S2 ::: {.caption} ###### Effects of the Amounts of Pathway Components on the Concentrations of β-Catenin and Axin This figure gives additional information with respect to the effects of Dsh, TCF, and GSK3β on the steady-state concentrations of total β-catenin (solid lines) and total axin (dashed lines) for the case of permanent Wnt-stimulation, *W* = 1. All concentrations and synthesis rates are scaled with respect to their values in the stimulated stationary state. (3,472 KB TIFF). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S3 ::: {.caption} ###### Effects of Synthesis Rates on the Concentrations of β-catenin and Axin The curves represent steady-state values of total concentrations of β-catenin (solid lines) and axin (dashed lines), depending on the rates of synthesis of β-catenin and axin. All concentrations and synthesis rates are scaled with respect to their values in the stimulated stationary state. (3,483 KB TIFF). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S1 ::: {.caption} ###### Mathematical Notation for Model Variables as Subdivided into Independent and Dependent Variables (45 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S2 ::: {.caption} ###### Complete List of Model Parameters of the Wnt Signal Transduction Model The rate constants marked with "\#" play a role only in stimulated states where *W* ≠ 0. Note that some of the numerical values are given in a higher precision compared to [Table 1](#pbio-0000010-t001){ref-type="table"}. (111 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: We thank Tom Rapoport, Rebecca Ward, and Yinon Ben-Neriah for thoughtful suggestions on the work. We thank the National Institute of Child Health and Development (grant HD037277 to MWK) for support and the National Institute of General Medical Sciences and the Harvard-Armenise Foundation for supplemental support (for RH). RH is also supported by the German Research Foundation (grant He 2049/2-1). EL is a Special Fellow of the Leukemia and Lymphoma Society. AS is supported by a Damon-Runyon Cancer Fund postdoctoral fellowship. RK is supported by the German Research Foundation (SFB 555). **Conflicts of Interest.** The authors have declared that no conflicts of interest exist. **Author Contributions.** EL, AS, RK, RH, and MWK conceived and designed the experiments. EL, AS, RK, and RH performed the experiments. EL, AS, RK, RH, and MWK analyzed the data. EL, AS, RK, RH, and MWK contributed reagents/materials/analysis tools. EL, AS, RK, RH, and MWK wrote the paper. Academic Editor: Roel Nusse, Stanford University School of Medicine. APC : adenomatous polyposis coli Dsh : Dishevelled EDTA : ethylene diamine tetraacetic acid GBP : glycogen synthase kinase-binding protein GSK3β : glycogen synthase kinase 3β MBP : maltose-binding protein PP2A : protein phosphatase 2A *Sf9* : *Spodoptera frugiperda* TCA : trichloroacetic acid TCF : T-cell factor XB : Xenopus buffer.

PubMed Central

2024-06-05T03:55:45.952727

2003-10-13

PMC212692

Introduction {#s1} ============ RNA interference (RNAi) is targeted gene silencing via double-stranded RNA (dsRNA); a gene is inactivated by specific breakdown of the mRNA ([@pbio-0000012-Fire1]; [@pbio-0000012-Montgomery1]). It is an ideal method for rapid identification of in vivo gene function. Initial studies on RNAi used microinjection to deliver dsRNA ([@pbio-0000012-Fire1]), but it was subsequently shown that dsRNA can be introduced very easily by feeding worms with bacteria that express dsRNA ([@pbio-0000012-Timmons1]). Using this technique on a global scale, an RNAi feeding library consisting of 16,757 bacterial clones that correspond to 87% of the predicted genes in Caenorhabditis elegans was constructed ([@pbio-0000012-Fraser1]; [@pbio-0000012-Kamath2]). Upon feeding to worms, these clones will give transient loss-of-function phenotypes for many genes by inactivating the target genes via RNAi. By feeding the clones in this library to wild-type Bristol N2 worms, loss-of-function phenotypes were assigned to about 10% of genes. However, RNAi phenotypes were missed for about 30% of essential genes and 60% of genes required for postembryonic development, probably because RNAi is not completely effective ([@pbio-0000012-Kamath2]). Other global RNAi screens have been recently performed in C. elegans using this RNAi library or other techniques ([@pbio-0000012-Gonczy1]; [@pbio-0000012-Maeda1]; [@pbio-0000012-Dillin1]; [@pbio-0000012-Piano1]; [@pbio-0000012-Ashrafi1]; [@pbio-0000012-Lee1]; [@pbio-0000012-Pothof1]). These screens were done using wild-type worms. We have already shown that mutation of *rrf-3*, a putative RNA-directed RNA polymerase (RdRP), resulted in increased sensitivity to RNAi ([@pbio-0000012-Sijen1]; [@pbio-0000012-Simmer1]). There are four RdRP-like genes in *C. elegans.* Two of these, *ego-1* and *rrf-1*, are required for efficient RNAi, as apparent from the fact that these mutants are resistant to RNAi against germline or somatically expressed genes, respectively ([@pbio-0000012-Smardon1]; [@pbio-0000012-Sijen1]). A third gene, *rrf-2,* appears to have no role in RNAi. The *rrf-3* strain, mutated in the fourth RdRP homolog, shows an opposite response to dsRNA; this mutant has increased sensitivity to RNAi ([@pbio-0000012-Sijen1]). A more detailed study of RNAi sensitivity of *rrf-3* mutants using a set of 80 genes showed that *rrf-3* is generally more sensitive to RNAi than wild-type worms ([@pbio-0000012-Simmer1]). RNAi phenotypes in *rrf-3* animals are often stronger, and they more closely approximate a null phenotype, when compared to wild-type. In addition, loss-of-function RNAi phenotypes were detected for a number of genes using *rrf-3* that were missed in a wild-type background. For example, known phenotypes were detected for many more neuronally expressed genes in the *rrf-3* background. These features suggest that the *rrf-3* strain could be used to improve and extend functional information associated with C. elegans genes. We have conducted a genome-wide RNAi screen using the *rrf-3* strain. In total, we found reproducible RNAi phenotypes for 423 clones that previously did not induce a phenotype (corresponding to 393 additional genes). To explore the variability of global RNAi screens, we performed the *rrf-3* screen twice for Chromosome I and carried out a Chromosome I screen with wild-type. These were cross-compared and also compared to the results of the wild-type screen of [@pbio-0000012-Fraser1]). From this, we find that *rrf-3* consistently allowed detection of more phenotypes than wild-type. In addition, we found that there is a significant screen-to-screenvariability (10%--30%). Results {#s2} ======= Comparative Analysis of RNAi for Chromosome I with Wild-Type and *rrf-3* {#s2a} ------------------------------------------------------------------------ We first conducted a pilot screen of Chromosome I using *rrf-3* and found RNAi phenotypes for 456 bacterial clones. We compared these data to those obtained by [@pbio-0000012-Fraser1]) for a screen in the wild-type Bristol N2 strain. For 153 of these 456 clones, no phenotypes were reported by [@pbio-0000012-Fraser1]) and phenotypes were observed for 303 clones in both screens. The N2 screen done by [@pbio-0000012-Fraser1]) resulted in RNAi phenotypes for 40 clones for which no phenotypes were found using *rrf-3* ([Figure 1](#pbio-0000012-g001){ref-type="fig"}A). These results indicate that *rrf-3* can be used in a global screen to identify loss-of-function phenotypes for additional genes. However, some phenotypes were missed in the *rrf-3* screen. To explore the reproducibility and variability of RNAi screens, we next screened the clones of Chromosome I using N2 and *rrf-3* side by side. We detected phenotypes for 447 clones: 140 were found only in *rrf-3*, 11 only in N2, and 296 in both strains ([Figure 1](#pbio-0000012-g001){ref-type="fig"}B). These data confirm that *rrf-3* is more sensitive to RNAi and, in addition, these data indicate that global RNAi screens with *rrf-3* will result in more clones with a detectable phenotype. ::: {#pbio-0000012-g001 .fig} Figure 1 ::: {.caption} ###### Comparison of Different RNAi Experiments of Chromosome I Using Wild-Type Bristol N2 and *rrf-3* Differences between different laboratories or investigators and between experiments done within the same laboratory and by the same investigators are observed. Ovals represent the amount of bacterial clones that gave an RNAi phenotype in an experiment. Areas that overlap represent clones for which in both experiments an RNAi phenotype was detected. Differences and overlap between an RNAi experiment done with the *rrf-3* mutant strain and the data obtained by [@pbio-0000012-Fraser1]) done with the standard laboratory strain, Bristol N2 (A); N2 and *rrf-3* tested at the same time within our laboratory (B); experiments done with N2 in two different laboratories: this study ('NL\') and [@pbio-0000012-Fraser1]) (C); two experiments done with the same strain, *rrf-3*, within our laboratory (D). :::  ::: Variability of the RNAi Effect {#s2b} ------------------------------ When we compared the RNAi results that we obtained using N2 with the [@pbio-0000012-Fraser1]) data, we were surprised to find significant differences: we only detected phenotypes for 75% of the clones that gave a phenotype in [@pbio-0000012-Fraser1]), and these researchers reported phenotypes for 84% of clones for which we found a phenotype ([Figure 1](#pbio-0000012-g001){ref-type="fig"}C). The differences do not appear to be due to false positives. For example, [@pbio-0000012-Fraser1]) detected the predicted phenotype for *goa-1* and *unc-73*, whereas we did not detect a mutant phenotype. Similarly, we detected the known mutant phenotype for *egl-30* and *cdc-25.1*, which were not detected by [@pbio-0000012-Fraser1]). In addition, we found that the false-positive rate is negligible (see below). It is possible that different laboratories or investigators have slightly different results. However, when we compare the results that we obtained with two independent screens of Chromosome I using *rrf-3* in our laboratory, we also see differences. For 394 clones we detected a phenotype in both experiments, 54 are specific for the first experiment, and 34 for the second ([Figure 1](#pbio-0000012-g001){ref-type="fig"}D). Among the clones that only gave an RNAi phenotype in one of the experiments are again clones that induced the predicted phenotype based on the phenotypes of genetic mutants (*unc-40*, *gpc-2*, and *sur-2*). These data show that large-scale RNAi screens done within the same laboratory and by the same investigators also give variable results. A few examples of variable RNAi results are shown in [Table 1](#pbio-0000012-t001){ref-type="table"}. ::: {#pbio-0000012-t001 .table-wrap} Table 1 ::: {.caption} ###### Variable RNAi Effects :::  Selection of clones that induced variable RNAi results in this study ('NL\') and or in the study by [@pbio-0000012-Fraser1]). In this subset of bacterial clones, each corresponds to a gene for which a mutant phenotype is known. The expected phenotypes are detected with RNAi, but not in each experiment, indicating false-negative results. The bacterial clones are indicated by 'GenePairs Name\' (name of genepair used to PCR-amplify a genomic fragment) and 'Predicted Gene\' (predicted gene targeted by the named genepair). 'Locus\' gives the genetic locus; 'Known Mutant Phenotype\' gives the mutant phenotype for the indicated gene described in the literature. The RNAi phenotypes are defined in the Materials and Methods section ::: In conclusion, we find that RNAi results from different laboratories and from experiments done in the same laboratory vary from 10% to 30%. This appears to be due to a high frequency of false negatives in each RNAi screen, even when the same method is used in the same laboratory. The Genome-Wide RNAi Screen {#s2c} --------------------------- Based on the positive results of the Chromosome I screen using the *rrf-3* strain, we next screened the complete RNAi library with *rrf-3* mutant animals. We obtained results for 16,401 clones and detected phenotypes for 2,079 (12.7%). Of these, we identified phenotypes for 625 clones for which no phenotype was reported in the [@pbio-0000012-Fraser1]) or [@pbio-0000012-Kamath2]) screens using N2, with the remaining 1,454generating phenotypes in both screens ([Table S1](#st001){ref-type="supplementary-material"}). In addition, there are 287 clones for which only [@pbio-0000012-Fraser1]) or [@pbio-0000012-Kamath2]) found phenotypes (23 of these were not done in our screen). The clones for which we only detected an RNAi phenotype once and that were specific for the *rrf-3* screen were retested. Subsequently, the phenotypes of the clones corresponding to Chromosomes II to X that were not confirmed by this repetition were tested once more. In this way, the clones specific for the *rrf-3* screen had two chances to be confirmed. Of the 625 clones for which no phenotype was found in the [@pbio-0000012-Fraser1]) and [@pbio-0000012-Kamath2]) N2 screens, the phenotypes of 423 clones were confirmed and 202 remained unconfirmed ([Table 2](#pbio-0000012-t002){ref-type="table"}; see [Table S1](#st001){ref-type="supplementary-material"}). Combining the N2 screens and these 423 clones, the percentage of clones with a phenotype increases from 10.3% to 12.8%. ::: {#pbio-0000012-t002 .table-wrap} Table 2 ::: {.caption} ###### Genome-Wide RNAi :::  Summary of the bacterial clones that induced detectable RNAi phenotypes ('Positive Clones\'). For 423 clones, RNAi phenotypes were reproducibly detected in our laboratory using *rrf-3*, but no RNAi phenotypes were reported in the N2 screens; 1,454 clones induced phenotypes in both laboratories; 264 were specifically detected by [@pbio-0000012-Fraser1]) or [@pbio-0000012-Kamath2]). For 202 clones, RNAi phenotypes were detected with *rrf-3* and no RNAi phenotypes were reported in the N2 screens, but this result could not be repeated. In addition, there are 23 clones for which we did not obtain results that gave a phenotype with N2. In the column with the overlapping clones, the *rrf-3* data are mainly from one experiment, whereas the N2 data reported by [@pbio-0000012-Fraser1]) and [@pbio-0000012-Kamath2]) are from repeated experiments. The phenotypes that were scored are described in the Materials and Methods section ::: Some of the RNAi phenotypes only found with *rrf-3* that remained unconfirmed could be confirmed by RNAi phenotypes detected with other clones of the RNAi library corresponding to the same gene or by other laboratories using different RNAi methods. For example, for the clones corresponding to the predicted genes *F56D1.1* (a member of the zinc finger C2H2-type protein family) and *F27C8.6* (a member of the esterase-like protein family), we detected sterile progeny (Stp) and embryonic lethality (Emb), respectively; these were also found by [@pbio-0000012-Piano1]). In addition, some unconfirmed RNAi phenotypes are confirmed by comparing to phenotypes of genetic mutants such as *gpc-2*, *hlh-8*, and *unc-84*. This suggests that many of the unconfirmed phenotypes reflect true gene functions. Analysis of the *rrf-3* Results {#s2d} ------------------------------- To validate the results obtained using *rrf-3*, we first assayed the rate of false positives in the total dataset (all RNAi results obtained with *rrf-3* for the 16,401 clones tested). In the assay used by [@pbio-0000012-Kamath2]), a set of genes for which it is known that genetic mutants display no lethality was selected. A false positive in the RNAi data is then defined as detecting a lethal RNAi phenotype for any of these genes. In the N2 screen, the false-positive rate was 0.4%. We find that the false-positive rate in the *rrf-3* data is similarly low (0 of 152 genes). To further determine the effectiveness of the screen, we compared the RNAi phenotypes with loss-of-function phenotypes of genetic mutants. For all chromosomes except for Chromosome I, the *rrf-3* data were confirmed by refeeding only if there was no phenotype detected in the N2 screens by [@pbio-0000012-Fraser1]) or [@pbio-0000012-Kamath2]). Therefore, to compare the difference in detection of known phenotypes between the *rrf-3* and the N2 screens, we used the Chromosome I datasets, where phenotypes were confirmed independently for the two strains. Of 75 genetic loci on Chromosome I, [@pbio-0000012-Fraser1]) detected 48% of published phenotypes, compared to 59% for *rrf-3* ([Table S2](#st002){ref-type="supplementary-material"}). Using the genome-wide *rrf-3* dataset (excluding the 202 unconfirmed phenotypes), we detected the published phenotype for 54% of 397 selected loci, compared to 52% for N2 ([Table 3](#pbio-0000012-t003){ref-type="table"}; see [Table S2](#st002){ref-type="supplementary-material"}). ::: {#pbio-0000012-t003 .table-wrap} Table 3 ::: {.caption} ###### Effectiveness of the *rrf-3* Screen :::  RNAi phenotypes obtained with *rrf-3* (confirmed using N2 data or *rrf-3* refeeding), and the N2 screens by [@pbio-0000012-Fraser1]) or [@pbio-0000012-Kamath2]) were compared with those of genes that have known loss-of-function phenotypes. 'Total Genetic Loci Scored\' denotes the number of genes that were analysed by RNAi. All loci have a loss-of-function phenotype that was detectable in our screen. 'RNAi Phenotype Detected\' gives the number of genes for which a phenotype was identified. 'Published Phenotype Detected\' gives the number of genes for which the RNAi phenotype matched the phenotype described in the literature ::: We next asked whether using the *rrf-3* strain improved general phenotype detection or whether certain types of phenotypes were particularly increased compared to the N2 screens by [@pbio-0000012-Fraser1]) and [@pbio-0000012-Kamath2]). To do this, we analysed the detection rate of different types of Chromosome I loci. First, we looked at a set of 23 loci with nonlethal postembryonic mutant phenotypes. Using *rrf-3*, we reproducibly detected the published phenotype for 11 of these compared to only two for N2. Of 50 loci required for viability (essential genes), we detected 31 using *rrf-3*, compared to 33 for N2. Thus, detection of essential genes was similar in the two strains, but detection of postembryonic phenotypes was improved with *rrf-3*. Finally, for the whole genome using *rrf-3*, we reproducibly detected the published phenotypes for 34 genetic mutants for which no RNAi phenotype was reported in the N2 screens (nine essential genes, 21 with postembryonic mutant phenotypes, and four with a slow-growth mutant phenotype). By comparison, published phenotypes were detected for 23 loci only with N2 (16 essential genes and seven with postembryonic mutant phenotypes) (see [Table S2](#st002){ref-type="supplementary-material"}). We conclude that *rrf-3* particularly improves detection of genes with postembryonic mutant phenotypes, a class that is poorly detected using wild-type N2. A striking feature of the *rrf-3* dataset is the high number of clones where a slow or arrested growth (Gro/Lva) defect was induced, without associated embryonic lethality or sterility. Overall, 619 clones induced a Gro/Lva defect using *rrf-3,* compared to 276 for N2, whereas the number of essential genes detected was similar (1,040 versus 1,170, respectively). In addition, in the confirmed set of 423 clones with *rrf-3*-specific phenotypes, Gro/Lva defects are the largest category (42%), whereas this is only 18% for N2, with the largest category being essential genes (49%). These data suggest that *rrf-3* might particularly enhance detection of genes that mutate to a slow-growth phenotype; we cannot easily test this hypothesis, as there are currently few known loci with this mutant phenotype. In some cases, a Gro/Lva phenotype was seen in *rrf-3*, whereas a different phenotype was seen in N2 (e.g., either lethality or a weak postembryonic phenotype). This suggests that some of the Gro/Lva phenotypes detected are due to incomplete RNAi of an essential gene (where lethality was seen in N2) or by a stronger RNAi effect (where no growth defect was seen in N2). In addition, it is possible that some of the Gro/Lva phenotypes detected are synthetic effects of using the *rrf-3* mutant strain. To summarise, using the *rrf-3* RNAi supersensitive strain in large-scale screens increases the percentage of clones for which it is possible to detect a phenotype. Detection of postembryonic phenotypes is particularly increased, whereas detection of essential genes is similar in *rrf-3* and N2. In addition, using *rrf-3*, there is a high rate of induction of Gro/Lva defects. Positional Cloning of Genetic Mutants with Visible Phenotypes {#s2e} ------------------------------------------------------------- Despite the advantages of RNAi, genetic mutants remain indispensable for many experiments. In the past decades, forward genetic screens identified a large number of genetic mutants, many of which are not yet linked to the physical map. We used the RNAi phenotypes obtained with the genome-wide screens to test whether we could systematically clone genes that are mutated in existing genetic mutants. First, the genetic map positions of all uncloned genetic mutants with visible phenotypes were checked using WormBase (<http://www.wormbase.org>, the Internet site for the genetics, genomics, and biology of C. elegans). Second, we searched for clones near the defined map positions that, when fed to N2, *rrf-3,* or both, gave phenotypes corresponding to the phenotypes of the genetic mutants. For most genetic mutants, more than ten clones with a similar phenotype were found in the interval to which the genetic mutant was mapped. However, for 21 genetic mutants, only one or a few candidate clones were found. The genes corresponding to these clones were subsequently sequenced in the genetic mutant to determine whether a mutation was present. In total, we sequenced 42 predicted genes for the 21 genetic mutants ([Table S3](#st003){ref-type="supplementary-material"}). For seven of these---*bli-3*, *bli-5*, *dpy-4*, *dpy-6*, *dpy-9*, *rol-3*, and *unc-108*---we found a mutation in one of the sequenced genes ([Table 4](#pbio-0000012-t004){ref-type="table"}). The mutated gene was confirmed by sequencing the same gene in a second or third allele (or both) of these genetic mutants ([Table 4](#pbio-0000012-t004){ref-type="table"}). ::: {#pbio-0000012-t004 .table-wrap} Table 4 ::: {.caption} ###### Properties of the Genetic Mutants Cloned Using the RNAi Phenotypic Data :::  Genetic mutants were linked to the physical map using RNAi phenotypes. The 'Genetic Map Position' is based on WormBase annotation. 'Mutated Gene' denotes the predicted gene, which is mutated in the genetic mutant. 'RNAi Phenotype' gives the loss-of-function phenotype either using *rrf-3* or N2 (the latter is based on findings of Kamath et al. \[2003\]). The phenotypes that were scored are described in the Materials and Methods section ^a^  *dpy-6(e2762)* has a deletion that removes the first six amino acid residues (aa) of the eighth exon and part of the seventh intron ^b^ Multiple mutations in *dpy-6(f11)* (5′-tcgAaaa\[G/T\]tt\[C/A\]aaccccacgccaact\[G/T\]cc); the AAA→AAAA mutation at position 2792 bp of the *F16F9.2* coding sequence causes a frameshift that results in a premature stop in the fifth exon ::: The identification of mutations in *unc-108* encoding the homolog of the small GTPase Rab2 is of particular interest. The RNAi phenotype of this gene gives a clue about the genetic property of the mutations in the mutants of *unc-108*. With *rrf-3*, we find that inactivation of Rab2 (F53F10.4) by RNAi causes uncoordinated movement ([Table 4](#pbio-0000012-t004){ref-type="table"}). Mutations in *unc-108* were isolated in a screen for dominant effects on behaviour; heterozygous *unc-108* mutants display dominant movement defects and are indistinguishable from homozygous mutants ([@pbio-0000012-Park1]). RNAi phenocopies a loss-of-function phenotype, suggesting that the dominant movement defects of *unc-108* mutants may be due to haplo-insufficiency. In eukaryotes, Rab2 is involved in regulating vesicular trafficking between the endoplasmic reticulum and Golgi. Based on the movement defects of *unc-108* mutants, UNC-108 might be involved in vesicle transport in neurons that regulate locomotion. Thus, the RNAi data are a powerful tool to facilitate rapid cloning of the genes identified by genetic mutants and will provide important starting points for further studies of their function. Discussion {#s3} ========== With this genome-wide RNAi screen using the hypersensitive strain *rrf-3*, we have significantly increased the functional information on the C. elegans genome, and we confirmed many RNAi phenotypes observed previously. We have assigned RNAi phenotypes for 406 genes (corresponding to the 423 extra clones) using *rrf-3*. For 13 genes, [@pbio-0000012-Kamath2]) or [@pbio-0000012-Fraser1]) had already found a phenotype using a different clone from the RNAi library that targeted the same gene, and for at least 44 genes a genetic mutant exists (see [Table S2](#st002){ref-type="supplementary-material"}). Other investigators have also found RNAi phenotypes for some of the genes using different methods. However, for most genes our result is to our knowledge the first hint about their biological function. Although we have identified new RNAi phenotypes for a substantial number of genes, others will have been missed in our screen for the following reasons. First, besides its increased sensitivity to RNAi, the *rrf-3* strain has an increased incidence of males (Him) and displays slightly increased embryonic lethality and a reduced brood size ([@pbio-0000012-Simmer1]). In our *rrf-3* experiments, we therefore made some minor adaptations to the original RNAi protocol described by [@pbio-0000012-Fraser1]). We did not score for the Him phenotype and had more stringent criteria for embryonic lethality and sterility. This may have reduced the number of extra clones identified with a phenotype. Moreover, the changes in the protocol can also account for some differences in the detection of RNAi phenotypes between *rrf-3* and N2. Second, when an RNAi phenotype is detected with N2 and not with *rrf-3*, the lack of a detectable phenotype may be the result of variability in the efficiency of RNAi. This is consistent with the fact that we observe differences between experiments done with the same strain. When an RNAi phenotype is detected with *rrf-3* and not with N2, this can be due to the increased sensitivity to RNAi of *rrf-3*. However, besides the higher sensitivity, we may also be observing synthetic effects with *rrf-3* (e.g., embryonic lethality, sterility, or developmental delay). In particular, a large number of clones induced a developmental delay phenotype using *rrf-3*. Synthetic effects cannot be excluded without investigating genetic mutants. Again, variability in the efficiency of RNAi will also contribute to these differences, and a small portion may be false positives. In general, the few false positives that occur in the screen are most likely due to experimental errors, whereas the false negatives are due to reduced efficiency of the RNAi. Finally, differences between *rrf-3* and N2 do not only involve the absence and presence of an RNAi phenotype, but also differences in the phenotypes for clones that did induce phenotypes in both screens (e.g., embryonic lethal in one screen and a postembryonic phenotype in the other). For example, we detected for *unc-112* a 100% embryonic lethal (Emb) phenotype with *rrf-3*, whereas [@pbio-0000012-Kamath2]) detected an adult lethal (Adl), uncoordinated (Unc), and paralyzed (Prz) phenotype with N2. Conversely, [@pbio-0000012-Kamath2]) detected for *gon-1* a 100% Emb phenotype and other phenotypes with N2, while we did not detect an Emb phenotype with *rrf-3*. What could be the source of the interexperimental variation of RNAi? Different phenotypes for the same gene can possibly occur owing to slight differences in the developmental stage at which the animals are exposed to dsRNA and owing to changes in temperature during the experiment. However, this probably does not account for the differences we see, as we always used animals of the same larval stage (L3/L4) and used incubators for constant temperature. It was shown previously that the level of induction of dsRNA production by isopropylthio-β-D-galactoside (IPTG) can modify the penetrance of the RNAi phenotype ([@pbio-0000012-Kamath1]). Therefore, differences in the induction of the dsRNA either by changes in the concentration of IPTG, temperature, timing, or the bacteria may be an important source of the variation in the outcome of RNAi. RNAi is starting to be used extensively in other systems experimentally, as well as therapeutically and agriculturally. The relative variability of the RNAi effect is an important fact to take in account also for the use of RNAi in other systems. The RNAi data can be a useful starting point for many new experiments, such as positional cloning of genetic mutants. By sequencing candidate genes based on the RNAi phenotypes, we identified the causal mutation in seven genetic mutants. Identification of these mutated genes gives insight into the biological process in which they are involved. In addition, cloning of these genes increases the resolution of the genetic map of C. elegans, since these mutants have been extensively used as visible markers in linkage studies. The complete set of RNAi phenotypes detected for the 2,079 clones using *rrf-3* will be submitted to WormBase, annotated as confirmed or unconfirmed. There the data can be evaluated in the context of information on gene structure, expression profiles, and other RNAi results. Materials and Methods {#s4} ===================== {#s4a} ### Nematode strains. {#s4a1} We used the following C. elegans strains: Bristol N2, NL4256 *rrf-3(pk1426)*, CB767 *bli-3(e767)*, MT1141 *bli-3(n259)*, CB518 *bli-5(e518)*, BC649 *bli-5(s277)*, CB1158 *dpy-4(e1158)*, CB1166 *dpy-4(e1166)*, CB14 *dpy-6(e14)*, CB4452, *dpy-6(e2762)*, F11 *dpy-6(f11)*, CB12 *dpy-9(e12)*, CB1164 *dpy-9(e1164)*, BC119 *dpy-24(s71)*, CB3497 *dpy-25(e817)*, MT1222 *egl-6(n592)*, MT1179 *egl-14(n549)*, MT1067 *egl-31(n472)*, MT151 *egl-33(n151)*, MT171 *egl-34(n171)*, *egl-34(e1452)*, MQ210 *mau-4(qm45)*, CB754 *rol-3(e754)*, BC3134 *srl-2(s2507dpy-18(e364)*; *unc-46(e177)rol-3(s1040)*, CB713 *unc-67(e713)*, CB950 *unc-75 (e950)*, HE177 *unc-94(su177)*, HE33 *unc-95(su33)*, HE151 *unc-96(su151)*, *unc-96(r291)*, HE115 *unc-100(su115)*, MT1093 *unc-108(n501)*, and MT1656 *unc-108(n777)*. ### RNAi by feeding. {#s4a2} RNAi was performed as described elsewhere ([@pbio-0000012-Fraser1]; [@pbio-0000012-Kamath1]) with minor adaptations when the *rrf-3* strain was used: after transferring L3- to L4-staged hermaphrodites onto the first plate, we left them for 48 h at 15°C instead of 72 h and then plated single adults onto other plates seeded with the same bacteria. Furthermore, we did not remove the mothers from the second plates. The phenotypes assayed are these: Emb (embryonic lethal), Ste (sterile), Stp (sterile progeny), Brd (low broodsize), Gro (slow postembryonic growth), Lva (larval arrest), Lvl (larval lethality), Adl (adult lethal), Bli (blistering of cuticle), Bmd (body morphological defects), Clr (clear), Dpy (dumpy), Egl (egg-laying defective), Lon (long), Mlt (molt defects), Muv (multivulva), Prz (paralyzed), Pvl (protruding vulva), Rol (roller), Rup (ruptured), Sck (sick), Unc (uncoordinated) Thin and Pale. Emb was defined as greater than 10% dead embryos for N2 and greater than 30% dead embryos for *rrf-3*. Ste required a brood size of fewer than ten among fed N2 worms and fewer than five among *rrf-3*. Each postembryonic phenotype was required to be present among at least 10% of the analysed worms. ### Sequencing of genetic mutants. {#s4a3} The coding sequence and the 5′- and 3′-untranslated region (about 500 bp upstream and downstream of the coding sequence) of the predicted genes, as annotated in WormBase, was analysed for mutations by sequencing amplified genomic DNA of the genetic mutants (see [Table S3](#st003){ref-type="supplementary-material"}). Nested primers were designed using a modification of the Primer3 program available on our website (<http://primers.niob.knaw.nl/>). Sequence reactions were done using the ABI PRISM Big Dye terminator sequencing kit (Applied Biosystems, Foster City, California, United States) and were analysed on the ABI 3700 DNA analyser. Sequences were compared to the genomic sequence of C. elegans using the BLAST program (<http://www.sanger.ac.uk/Projects/C_elegans/blast_server.shtml>) or analysed using the PolyPhred program (available from <http://droog.mbt.washington.edu/PolyPhred.html>). Supporting Information {#s5} ====================== Table S1 ::: {.caption} ###### RNAi Phenotypes for Bacterial Clones Using *rrf-3* (482 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S2 ::: {.caption} ###### Detailed Comparison of RNAi Phenotypes with Those of Known Loci (188 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S3 ::: {.caption} ###### Summary of Genes Sequenced in Several Genetic Mutants (25 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Accession Numbers {#s5a} ----------------- RNAi data from this study will be submitted to WormBase (<http://www.wormbase.org>). We thank the Caenorhabditis Genetics Stock Center for providing most of the strains used in this study; Mario de Bono for providing the *egl-34(e1452)*, *dpy-4(e1158)*, *dpy-9(e424)*, *dpy-9(e858)*, and *dpy-9(e1164)* strains; Guy Benian for providing the *unc-96(su151)* and *unc-96(r291)* strains; David Baillie and Domena Tu for providing the *bli-5(s277)* strain; and Robert Horvitz and An Na for providing the *unc-108(n777)* strain. This work was supported by the Netherlands Organization for Scientific Research (grants CW97045, MW90104094, and MW01480008). RSK was also supported by a Howard Hughes Medical Institute predoctoral fellowship. AGF was also supported by a United States Army Breast Cancer Research fellowship. JA was also supported by a Wellcome Trust Senior Research fellowship. **Conflicts of Interest.** The authors have declared that no conflicts of interest exist. **Author Contributions.** RHAP conceived and designed the experiments. FS, CM, AMvdL, EK, and PvdB performed the experiments. FS, CM, AMvdL, and JA analysed the data. RSK, AGF, and JA contributed reagents/materials/analysis tools. FS wrote the paper. Academic Editor: James Carrington, Oregon State University. Adl : adult lethal Bli : blistering of cuticle Bmd : body morphological defects Brd : low broodsize Clr : clear Dpy : dumpy dsRNA : double-stranded RNA Egl : egg-laying defective Emb : embryonic lethal Gro : growth defect/slow postembryonic growth IPTG : isopropylthio-β-D-galactoside Lon : long Lva : larval arrest Lvl : larval lethality Mlt : molt defects Muv : multivulva Prz : paralyzed Pvl : protruding vulva RdRP : RNA-directed RNA polymerase RNAi : RNA interference Rol : roller Rup : ruptured Sck : sick Ste : sterile Stp : sterile progeny Unc : uncoordinated.

PubMed Central

2024-06-05T03:55:45.961528

2003-10-13

PMC212693

In 1928 the German botanist Emil Heitz visualised in moss nuclei chromosomal regions that do not undergo postmitotic decondensation ([@pbio-0000014-Heitz1]). He termed these parts of the chromosomes heterochromatin, whereas fractions of the chromosome that decondense and spread out diffusely in the interphase nucleus are referred to as euchromatin. Further studies revealed that heterochromatin can be found in all higher eukaryotes, mainly covering regions with a low frequency of genes, such as pericentromeric regions and telomeres. Heitz proposed that heterochromatin reflects a functionally inactive state of the genome, and we now know that DNA in heterochromatic regions is less accessible to nucleases and less susceptible to recombination events. All these findings contributed to the current view that heterochromatin is a rigid nuclear compartment in which transcriptionally inactive regions of chromatin are densely packed and inaccessible to the transcription machinery ([@pbio-0000014-Grewal1]). This view was challenged earlier this year in two papers published back-to-back in *Science* ([@pbio-0000014-Cheutin1]; [@pbio-0000014-Festenstein1]). Certain proteins are specifically associated with heterochromatin--- notably, the family of heterochromatin protein 1 (HP1) ([@pbio-0000014-Eissenberg1]; [@pbio-0000014-Singh1]). HP1 is thought to play a central role in creating a stable and inaccessible heterochromatic network by interacting with several other proteins, including histones, the major protein constituent of all chromatin. In particular, HP1 binds to the tail of the histone H3 when it has been modified by methylation of lysine 9. This histone modification is an important landmark of inactive chromatin regions. In the two articles in *Science*, both groups generated cell lines stably expressing HP1 fused to green fluorescent protein (GFP) so that they could watch the behaviour of HP1 in living cells. Specifically, they used photobleaching techniques to study the in vivo mobility of HP1. In a defined region of a cell, fluorescently tagged proteins are bleached by a laser pulse. Recovery of fluorescence in the bleached area can then only occur if bleached molecules are replaced with unbleached molecules from regions outside the bleached area. The technique is called fluorescence recovery after photobleaching (FRAP) and provides information about the mobility and stability of the cellular structures and proteins. For HP1--GFP, the speed at which fluorescence recovers depends on how tightly it is bound within heterochromatic regions. Heitz (and many others) might have expected that heterochromatin-bound HP1 shows little turnover and therefore recovery should take place very slowly. [@pbio-0000014-Cheutin1] first demonstrated that the heterochromatic regions visualised with HP1--GFP are stable in shape for at least 2 h. By contrast, subsequent FRAP experiments revealed that HP1 proteins have a surprisingly high turnover rate in heterochromatic clusters as well as in regions the authors define as euchromatic ([Figure 1](#pbio-0000014-g001){ref-type="fig"}). Recovery of 50% was reached after 2.5 s in heterochromatin and after 0.6 s in euchromatin. In contrast, for histone proteins, the structural protein components of chromatin, 50% recovery took more than 2 h ([@pbio-0000014-Kimura1]). Cheutin et al. found that complete recovery of bleached HP1 took 5 s in euchromatin and 60 s in heterochromatin. [@pbio-0000014-Festenstein1] report, however, that recovery only reaches 90% in euchromatin and 70% in heterochromatin. Incomplete recovery would point to an immobile population of HP1 that does not exchange rapidly. In fact, such a stable fraction could be indicative of a stable structural network made of a minor fraction of HP1 that could serve as a nucleation site for a more mobile fraction of HP1. In my opinion, this should be kept in mind, even if 100% recovery is observed. It might well be that a few stably associated HP1 molecules that remain undetected in FRAP studies exert an important structural function in heterochromatin formation. Consequently, this can be regarded as an important discrepancy between the two studies. Both studies also reported a number of other experiments in which the condensation state of chromatin was modified and was found to alter the mobility of HP1, such that relaxed condensation was associated with increased HP1 mobility. ::: {#pbio-0000014-g001 .fig} Figure 1 ::: {.caption} ###### FRAP of HP1--GFP Reveals a Dynamic Association with Heterochromatin A fraction of a heterochromatic cluster (arrowhead) was bleached by a laser pulse, and recovery of fluorescence was monitored by time-lapse imaging. Images were kindly provided by Thierry Cheutin and Tom Misteli. :::  ::: As discussed by the authors of both studies, several important conclusions can be drawn. In striking contrast to previous models, HP1 appears to be a very mobile molecule. The formation of heterochromatin appears not to be based on a stable oligomeric network of HP1 molecules. Furthermore, heterochromatin is accessible. There is no obvious constraint shielding these transcriptionally inactive compartments from factors residing outside. Given the rapid exchange of HP1 in heterochromatic clusters, any other soluble nuclear protein, such as a transcription factor, should be able to gain access, compete with silencing factors, and potentially activate genes located within heterochromatin. Taken together, heterochromatin appears to be a surprisingly dynamic compartment even though it forms morphologically stable entities. This dynamic situation could imply that heterochromatic silencing is not just a switch, but rather a continuous and active process. Although the new work suggests that heterochromatin is more dynamic than was thought, some caveats remain. It is still possible that a stable "mark" of heterochromatin does exist. As I discussed above, this mark might be an undetected fraction of immobile HP1 molecules. In addition, one cannot exclude that HP1 is a downstream factor that is dynamically tethered to a stable binding site, the most likely candidate being the methylated histone tail. Perhaps the role of HP1 in heterochromatic silencing has simply been overinterpreted. The formation of facultative heterochromatin by X inactivation in mammals, for example, does not involve HP1 even though appropriate histone methylation marks are set. This indicates that heterochromatin formation does not always follow the same rules and suggests that our definitions of heterochromatin must be refined. In any case, it could well be that a silent state is marked by signals that have a slow turnover. In fact, histone methylation is believed to generate a quite stable "code" ([@pbio-0000014-Jenuwein1]). Until this hypothesis has been tested in vivo, we should keep an open mind about the stability and dynamics of nuclear structure. Several other nuclear compartments (spliceosomes, nucleoli) have also been proposed to consist of dynamic collections of components ([@pbio-0000014-Misteli1]). Personally, I am intrigued by the fact that heterochromatin also might function as a steady-state association of molecules. Is the entire nucleus, the genome organization---irrespective of its functional state---in constant flux? Of course, such a situation would provide an appealing explanation for the plasticity of gene expression. On the other hand, dynamic control of gene expression complicates explanations of how established expression patterns are stably inherited. So far, genetic knockout experiments have been the most powerful tools to unravel the mechanisms of epigenetic regulation. Unfortunately, many of those investigations can only provide insight into the establishment of expression profiles. What happens if regulatory factors are knocked down after expression patterns are set up? Which signals will be erased and which ones will persist? I am working on the mechanism of dosage compensation in *Drosophila* ([@pbio-0000014-Lucchesi1]). This process involves stable changes of chromatin structure, which leads to lasting effects on X-chromosomal gene expression. To examine the generality of the new results on heterochromatin, it will be important to find out whether the proteins involved in defining the X chromosome as an epigenetic compartment have the same dynamic behaviour as HP1. In addition, systematic knockdown of these factors after establishment of dosage compensation might disclose a hierarchy in epigenetic maintenance that is different from the one affecting establishment. Tobias Straub is a postdoctoral research fellow in the laboratory of Peter B. Becker at the Adolf-Butenandt Institute of the Ludwig-Maximilians University in Munich, Germany. E-mail: <tstraub@mol-bio.med.uni-muenchen.de>. FRAP : fluorescence recovery after photobleaching GFP : green fluorescent protein HP1 : heterochromatin protein 1.

PubMed Central

2024-06-05T03:55:45.964213

2003-10-13

PMC212694

Microarrays are used to survey the expression of thousands of genes in a single experiment. Applied creatively, they can be used to test as well as generate new hypotheses. As the technology becomes more accessible, microarray analysis is finding applications in diverse areas of biology. Microarrays are simply a method for visualizing which genes are likely to be used in a particular tissue at a particular time under a particular set of conditions. The output of a microarray experiment is called a "gene expression profile." Gene expression profiling has moved well beyond the simple goal of identifying a few genes of interest. The notion that this is the major objective of microarray studies has engendered the oft-repeated criticism that the approach only amounts to "fishing expeditions." The sophistication of microarray analysis very much blurs the distinction between hypothesis testing and data gathering. Hypothesis generation is just as important as testing, and very often expression profiling provides the necessary shift in perspective that will fuel a new round of progress. In many gene expression profiling experiments, the hypotheses being addressed are genome-wide integrative ones rather than single-gene reductionist queries. In general, without a hypothesis only the most obvious features of a complex dataset will be seen, while clear formulation of the scientific question undoubtedly fuels better experimental design. And in some cases, the results of a microarray screen that was initially designed as an effort at cataloguing expression differences are so unexpected that they immediately suggest novel conclusions and areas of enquiry. Fundamental Microarray Technology {#s2} ================================= All microarray experiments rely on the core principle that transcript abundance can be deduced by measuring the amount of hybridization of labeled RNA to a complementary probe. The idea of a microarray is simply to lay down a field of thousands of these probes in perhaps a 5 sq cm area, where each probe represents the complement of at least a part of a transcript that might be expressed in a tissue. Once the microarray is constructed, the target mRNA population is labeled, typically with a fluorescent dye, so that hybridization to the probe spot can be detected when scanned with a laser. The intensity of the signal produced by 1,000 molecules of a particular labeled transcript should be twice as bright as the signal produced by 500 molecules and, similarly, that produced by 10,000 molecules half as bright as one produced by 20,000 molecules. So a microarray is a massively parallel way to survey the expression of thousands of genes from different populations of cells. Trivially, if fluorescence is observed for a gene in one population but not another, the gene can be inferred to be on or off, respectively. With appropriate replication, normalization, and statistics, though, quantitative differences in abundance as small as 1.2-fold can readily be detected. The output of all microarray hybridizations is ultimately a series of numbers, which covers a range of almost four orders of magnitude, from perhaps one transcript per ten cells to a few thousand transcripts per cell ([@pbio-0000015-Velculescu1]). It is the comparison of gene expression profiles that is usually of most interest. This is because the visualization is done at the level of transcript abundance, but just seeing a transcript does not guarantee that the protein is produced or functional. If, however, a difference in transcript abundance is observed between two or more conditions, it is natural to infer that the difference might point to an interesting biological phenomenon. A general approach to performing gene expression profiling experiments is indicated as a flow diagram in [Figure 1](#pbio-0000015-g001){ref-type="fig"}. Having performed the experiment, quality control checks, statistical analysis, and data-mining are performed. More and more, investigators are interested not just in asking how large the magnitude of an expression difference is, but whether it is significant, given the other sources of variation in the experiment. Similarly, we might want to evaluate whether some subset of genes show similar expression profiles and so form natural clusters of functionally related genes. Or we may combine expression studies with genotyping and surveys of regulatory sequences to investigate the mechanisms that are responsible for similar profiles of gene expression. Finally, all of the expression inferences must be integrated with everything else that is known about the genes, culled from text databases and proteomic experiments and from the investigator\'s own stores of biological insight. ::: {#pbio-0000015-g001 .fig} Figure 1 ::: {.caption} ###### Flow Diagram of Gene Expression Profiling :::  ::: Fishing for Hypotheses {#s3} ====================== The ability to survey transcript abundance across an ever-increasing range of conditions gives geneticists a fresh look at their cellular systems, in many cases providing a more holistic view of the biology, but at the same time feeding back into the classical hypothetico-deductive scientific framework. The technology has rapidly advanced beyond the simple application of fishing for candidate genes and now sees applications as diverse as clinical prediction, ecosystem monitoring, quantitative mapping, and dissection of evolutionary mechanisms. Two of the better-known examples of the interplay between microarray profiling and hypothesis testing are provided by the studies of [@pbio-0000015-Ideker1] and [@pbio-0000015-Toma1]. The latter authors profiled the difference in expression between strains of flies that had been divergently selected for positive and negative geotaxis, a supposedly complex behavior relating to whether flies prefer to climb or stay close to the ground. They identified two dozen differentially expressed genes, several of which were represented by mutant or transgenic stocks that allowed tests of the effect of gene dosage on behavior. At least four of the candidate genes indeed quantitatively affect geotaxis. [@pbio-0000015-Ideker1] took this approach a step further in arguing for a four-step iterative feedback between profiling, identifying candidate genes, knocking them out, and then profiling once more. They showed how thoughtful experimentation can considerably enhance our understanding of genetic regulatory pathways such as the yeast galactose response. Much excitement has been generated recently by the potential for clinical applications of gene expression profiling in relation to complex diseases such as cancer, diabetes, aging, and response to toxins. An early foray into this realm was provided by [@pbio-0000015-Alizadeh1], who demonstrated that diffuse large B-cell lymphomas have two major subtypes defined by molecular profiles. Whereas it is difficult to predict clinical outcome on the basis of histology, these profiles define a set of genes that provide quite a strong indicator of long-term survival. Similarly, [@pbio-0000015-vantVeer1] have described a "poor prognosis" signature in breast cancer biopsies from young women prior to the appearance of metastases in the lymph nodes. Much statistical and empirical work remains to be done before these tools see clinical application, but the idea that gene expression integrates signals from the genotype and environment provides potent motivation for studying disease with microarrays. A good example of the ability of microarray analyses to simply surprise us is provided by the study reported in this issue of *PLoS Biology* by DeRisi and colleagues ([@pbio-0000015-Bozdech1]). They reasoned that profiling transcript abundance throughout the erythrocyte phase of the lifecycle of the malaria parasite Plasmodium falciparum might identify a handful of genes that are induced at critical times and hence might be novel drug targets. Employing very careful staging, a platform with low experimental noise, and appropriate statistical procedures, they discovered an extremely tight molecular lifecycle within the organism. Families of functionally related genes are induced as a unit, one after another, in a tightly orchestrated rhythm that testifies to incredible integration of the physiology of the parasite. They show that with microarray analysis it is possible to model the physiology and biochemistry of the pathways instead of just targeting a few genes. In the coming years, expect to see microarrays developed for an extremely diverse range of organisms and applied to an even wider range of questions, from parasitology to nutritional genomics. Consensus on a core set of statistical options will likely emerge, as will agreement on data quality standards. Applications will encompass defining gene function; inferring functional networks and pathways; understanding how variation is distributed among individuals, populations, and species; and developing clinical protocols relating to cancer prognosis and detection of toxin exposure. Similar profiling methods for proteins and metabolites will attract just as much attention as functional genomics, building on the foundations laid by genome sequencing. Greg Gibson is at the Department of Genetics, North Carolina State University, Raleigh, North Carolina, United States of America. E-mail: <ggibson@unity.ncsu.edu>.

PubMed Central

2024-06-05T03:55:45.965328

2003-10-13

PMC212695

The immune system needs to be able to identify and ultimately destroy foreign invaders. To do so, it utilizes two major types of immune cells, T cells and B cells (or, collectively, lymphocytes). Lymphocytes display a large variety of cell surface receptors that can recognize and respond to an unlimited number of pathogens, a feature that is the hallmark of the "adaptive" immune system. To react to such a variety of invaders, the immune system needs to generate vast numbers of receptors. If the number of different types of receptors present on lymphocytes were encoded by individual genes, the entire human genome would have to be devoted to lymphocyte receptors. To establish the necessary level of diversity, B- and T-cell receptor (BCR and TCR, respectively) genes are created by recombining preexisting gene segments. Thus, different combinations of a finite set of gene segments give rise to receptors that can recognize unlimited numbers of foreign invaders. This is accomplished by a supremely well-coordinated set of reactions, starting with cleaving DNA within specific, well-conserved recombination signal sequences (RSSs). This highly regulated step is carried out by the lymphocyte-specific recombinationactivating genes (*RAG1* and *RAG2*). The segments are then reassembled using a common cellular repair mechanism. For foreign invaders and their proteins (antigens) that are not part of the host to elicit an immune response, the immune system must be able to recognize countless numbers of antigens. For obvious reasons, an unlimited number of unique antigen receptors cannot be genetically encoded. Rather, the necessary diversity in receptors is achieved by creating variations in the antigen-recognition regions of the receptors of both B cells and T cells. These regions are created by the pairing of two different protein segments, called polypeptide chains (heavy \[H\] and light \[L\] chains in the case of the BCR and α and β chains in the case of the TCR), which form a cleft that provides a binding site for the antigen. The mechanism that generates variation in the antigen-binding pockets of these receptors involves mixing and matching variable (V), diversity (D), and joining (J) gene segments in a process called V(D)J recombination. To assemble a single functional receptor, preexisting V, D, and J gene segments are rearranged to yield a contiguous V(D)J region, just upstream of another element of the receptor, the constant (C) region ([Figure 1A](#pbio-0000016-g001){ref-type="fig"}). ::: {#pbio-0000016-g001 .fig} Figure 1 ::: {.caption} ###### V(D)J Recombination Takes Place within the BCR and TCR Loci \(A) Schematic of a receptor locus.V, D, and J segments are found just upstream of the constant region. (B) A cartoon view of a VJ recombination reaction. V segments (red) are flanked by RSSs with 12 bp-long spacers (green), while the J segments are flanked by RSS with 23 bp-long spacers (orange). Breaks are introduced directly between the heptamer and the coding sequence, and a CJ is formed between a V and a J segment, while the RSS ends are put together to form an SJ within a circular DNA that is later lost. Symbols: P, promoter; E, enhancer. :::  ::: The BCR H chain and the TCR β chain consist of V, D, and J segments, while BCR L chains and the TCR α chain are comprised of only V and J segments. The number of each type of segments within the chains allows for a large but finite combinatorial possibility in rearrangement, a phenomenon termed combinatorial diversity ([Table 1](#pbio-0000016-t001){ref-type="table"}). However, variations generated by V(D)J recombination are uncountable because they do not simply rely on the number of gene segments. Further diversity is introduced because the junctions between rearranged gene segments contain small insertions and deletions (junctional diversity). Finally, both BCR and TCR are heterodimers (consisting of two unmatched polypeptides), so the possibilities of different pairing between the chains can also increase variation. Successful V(D)J rearrangement is clearly useful in terms of antigen recognition, and it is absolutely required for the development and survival of B and T cells. ::: {#pbio-0000016-t001 .table-wrap} Table 1 ::: {.caption} ###### Diversification of BCRs and TCRs :::  Number of V, D, and J segments contributes to combinatorial diversity. Further changes are introduced by junctional diversity, to give the total number of BCR and TCR repertoires ::: V(D)J Recombination: A Cut-and-Paste Reaction {#s2} ============================================= In the first part of the "cut-and-paste" reaction, breaks within both strands of the DNA helix (double-stranded breaks) are made within the RSS sites; in the second part, the newly created breaks are repaired by the cell\'s general DNA repair pathway. In the initial phase, two lymphocyte-specific proteins that are encoded by the recombinationactivating genes (*RAG1* and *RAG2*) work together to recognize and bind RSSs. The complex consisting of these two proteins, RAG1--RAG2 (henceforth RAG), cuts the DNA between the rearranging DNA segments and the adjacent RSS motifs ([Figure 1B](#pbio-0000016-g001){ref-type="fig"}). The second step of the reaction glues together the ends of the chromosome containing the rearranging segments, which will ultimately code for the receptor and are called coding joints (CJs). The portion of DNA between the rearranged segments is shed from the genome, but it too gets glued together in a minicircle (a signal joint \[SJ\]). Typically, SJs are rapidly and precisely fused, but CJs are ligated more slowly, in part because their fusion is not precise---small insertions are present quite often, and even deletions can be detected. (For detailed reviews, see [@pbio-0000016-Fugmann1]; [@pbio-0000016-Gellert1]). Both pasting reactions are necessary for creating the receptors as well as for preventing havoc within the genome. RAGs: Indispensable for V(D)J Recombination {#s3} =========================================== RAG proteins carry out the first enzymatic step of the reaction---site-specific cleavage of DNA ([@pbio-0000016-vanGent1]). Artificial expression of RAGs in mammalian cells other than B- or T-lymphocytes suggests that RAG is the only lymphocyte-specific factor required for this recombination event to occur ([@pbio-0000016-Schatz1]). Indeed, in mice whose *RAG* genes have been deleted (*RAG* ^−/−^), V(D)J recombination is completely abolished, and these mice have neither mature B nor T cells ([@pbio-0000016-Mombaerts1]; [@pbio-0000016-Shinkai1]). A similar type of immunodeficiency, called Omenn syndrome, is seen in people with mutations in their *RAG* genes ([@pbio-0000016-Villa1]). In test-tube experiments, purified RAG proteins are sufficient for cleavage of a synthetic DNA containing the appropriate RSS ([@pbio-0000016-McBlane1]). This reaction can be subdivided into two stages. First, a nick is made in the DNA, at a specific site within the RSS, leaving specific chemical modifications at the ends ([Figure 1B](#pbio-0000016-g001){ref-type="fig"}). Then, one free end with a specific (3′-hydroxyl) group forms a new chemical (diester) bond with a different chemical (phosphoryl) group on the complementary nucleotide of the opposite strand (this is called a transesterification reaction). This results in the formation of a hairpin at the coding end, while leaving the signal end blunt ([@pbio-0000016-McBlane1]). RSSs: The Targets of the Reaction {#s4} ================================= RSSs are found next to every variable (V), diversity (D), and joining (J) segment. They consist of three distinct elements: a heptamer and a nonamer sequence, separated by a spacer element---either 12 or 23 bp long ([Figure 2](#pbio-0000016-g002){ref-type="fig"}) ([@pbio-0000016-Tonegawa1]; [@pbio-0000016-Akira1]). Although the two RAG proteins work together in a protein complex, they do have unique functions. RAG1 binds both the 12- and 23-RSSs with equal affinities, while RAG2 does not bind either RSS sequence. This suggests that RAG1 forms the initial complex with DNA, which then recruits and is stabilized by RAG2 ([@pbio-0000016-Fugmann1]). ::: {#pbio-0000016-g002 .fig} Figure 2 ::: {.caption} ###### RSSs Consist of a Fairly Conserved Heptamer and Nonamer Sequence, Separated by a Spacer Element Heptamer is shown in red and nonamer in green. Conserved nucleotides are shown in bold. The spacer is either 12 or 23 bp long. :::  ::: Both the heptamer and the nonamer contain nucleotides that are absolutely required for efficient V(D)J recombination. The first three nucleotides within the heptamer are conserved, whereas mutations in the next four positions affect but do not abolish the reaction. Within the nonamer, positions 5 and 6 are conserved, while variations are tolerated elsewhere in the sequence ([Figure 2](#pbio-0000016-g002){ref-type="fig"}). The length of the spacer (but not its sequence) was thought to play an important role in regulating which elements could be recombined ([@pbio-0000016-Gellert1]): a 12mer differs from a 23mer by a single turn of the DNA helix, providing for the proteins that bind RSSs to remain in the same rotational phase ([@pbio-0000016-Gellert1]). The paper in this issue of *PLoS Biology* by [@pbio-0000016-Lee1] now challenges this paradigm. They report that not only the length of the spacer but also its sequence are important, so that changes in the spacer sequence directly affect rates of recombination. Based on these data, they have developed an algorithm that accurately predicts the relative efficiencies of RSS binding, cleavage, and rearrangement based on nucleotide sequence. Breaks: Necessary Precursors of Recombination {#s5} ============================================= After RAG cuts the DNA at the RSS, the two sides of the break are different. The coding ends are closed to resemble hairpins, while the RSS ends are open and blunt. These blunt RSS ends are rejoined rapidly, forming SJs, but before the coding ends can be fused, the hairpins must be opened ([@pbio-0000016-Gellert1]). Normally, hairpins are opened either at the tip or on the side. In either case, an enzyme called terminal deoxynucleotidyl transferase can add a small amount of random (nontemplated) nucleotides to freed ends, and this phenomenon of N-nucleotide addition contributes to junctional diversity of the CJs ([@pbio-0000016-Bassing1]). When the hairpin is opened on the side, the reaction leaves a short overhang on one strand. When the overhangs are filled, palindromic sequences are generated, and these modifications are referred to as "P-nucleotides" ([@pbio-0000016-Bassing1]). These, too, contribute to the diversity of the CJs and ultimately, when the rearranged gene produces protein, to more variability in the antigen-binding pocket of the receptor. Resolution: The Final Step {#s6} ========================== As mentioned, *RAG* ^−/−^ mice do not have mature B and T cells. This causes a severe combined immunodeficiency syndrome (SCID) characterized by a complete block in B- and T-cell development, but no other defects ([@pbio-0000016-Mombaerts1]; [@pbio-0000016-Shinkai1]). However, there are other molecular deficiencies that also have a SCID phenotype. One of these is mapped to the enzyme DNA protein kinase (DNA-PKcs), which is required for the proper joining of DNA ends ([@pbio-0000016-Bosma1]). Mice deficient in DNA-PKcs can initiate V(D)J recombination, but cannot form the CJs ([@pbio-0000016-Gao1]). These mice are also sensitive to processes that induce DNA double-stranded breaks, such as ionizing radiation ([@pbio-0000016-Gao1]). Hence, the repair pathway responsible for fixing DNA breaks caused by radiation also creates CJs. Indeed, along with DNA-PKcs, other proteins of this nonhomologous end-joining repair pathway are important for the completion of V(D)J recombination (such as Ku70 and Ku80, Artemis, XRCC4, and DNA ligase IV) ([@pbio-0000016-Bassing1]). An Evolutionary Model of V(D)J Recombination {#s7} ============================================ From the discovery of the *RAG* genes on, investigators have suspected that V(D)J recombination may be the result of the landing of a transposable genetic element (a "jumping gene" or transposon) into the vertebrate genome. The clues were many. Firstly, the compact organization of the *RAG* locus resembles a transposable element ([@pbio-0000016-Schatz2]). Secondly, RAGs cut the DNA after binding RSSs throughout the BCR and TCR loci ([@pbio-0000016-Gellert1]). RSSs resemble the ends of other transposable elements. Biochemically, the reaction shares characteristics with enzymes found in other transposable elements ([@pbio-0000016-Gellert1]). Finally, these genes appear abruptly in evolution: they are present in the jawed vertebrates (like the shark), but not in more ancient organisms ([@pbio-0000016-Schluter1]). A current model is that an ancient transposon containing the *RAG* genes flanked by RSS ends "jumped" into an area of the vertebrate lineage containing a primordial antigen receptor gene, separating it into pieces. When the transposon lifted off, it left the RSSs behind. Multiple rounds of transposition and duplication eventually gave rise to our TCR and BCR loci. Mechanistically, transposition could happen even now, if the DNA segment containing the SJs is not ligated into a minicircle. These unsealed SJs are able to insert into heterologous sequences in vitro, and if they are allowed to float free in the cell, they have the potential to invade the genome. This phenomenon could give rise to certain types of chromosomal translocations prevalent in some types of B- and T-cell cancers ([@pbio-0000016-Shih1]). Generally, repair processes in the cell actively suppress this phenomenon by rapidly ligating SJs. Conclusion {#s8} ========== V(D)J recombination is absolutely crucial for the adaptive immune response. In its absence, our immune system is compromised. When it is not properly controlled, it gives rise to chromosomal translocations and B-and T-cell cancers. The elucidation of all steps of the reaction and attempts to understand exactly how these steps are regulated to avoid disastrous side effects are areas of study that have occupied researchers in the past and will continue to do so in the future. Both authors are at Rockefeller University, New York, New York, United States of America. Eleanora Market is in the Graduate School. F. Nina Papavasiliou is in the Laboratory of Lymphocyte Biology. BCR : B-cell receptor C : constant CJ : coding joint D : diversity DNA-PKcs\[RRK1\] : DNA protein kinase H : heavy J : joining L : light RAG : recombination-activating gene RSS : recombination signal sequence SCID : severe combined immunodeficiency syndrome SJ : signal joint TCR : T-cell receptor V : variable.

PubMed Central

2024-06-05T03:55:45.966344

2003-10-13

PMC212696

The completion of the human genome project has ushered in a new era in which biology has become an information science. In this new era, sharing of information is quickly becoming a critical aspect of scientific discovery. As directors of National Institutes of Health (NIH) institutes dedicated to neuroscience, we recognize several areas of research where sharing of primary data will be necessary for us to reach our scientific goals, including brain-mapping, genetics, and clinical trials. Progress in each of these areas will require not only new tools for sharing information but a change in our scientific culture. Here we describe some of the recent progress in efforts to map the brain as an example of the potential and the challenge of sharing data in an era when neurobiology, like genomics, is becoming an information science. In parallel to the worldwide effort to map the human genome, investigators in neuroscience have used a range of techniques to map the brain. The efforts share some superficial similarities: the genome has 3 × 10^9^ bases and the human brain has roughly 100 × 10^9^ neurons; both the genome and the brain have embedded modules of functional units (genes versus circuits) that can be mapped in space; and localization of both genes and circuits requires computational power that can be distributed across laboratories. But the analogy breaks down quickly. Whereas fundamental genome data can be addressed as unidimensional text of four letters in varying order, a comprehensive map of the brain includes molecular, cellular, system, and behavioral data---all of which are dynamic, interacting, and interdependent. For example, brain circuitry is organized in three-dimensional space constantly changing in time, with each neuron having 10^3^--10^4^ synapses and with many of those synapses capable of plasticity that may, in turn, have significant functional consequences. "As we emerge from the "decade of the brain," we are entering a decade for which data-sharing will be the currency for progress in neuroscience." As a testament to the complexity of brain data, a century after the classic age of neurohistology, there are continuing arguments about the taxonomy of neurons, depending on location, morphology, neurochemistry, or RNA profile. For instance, a population of neurons in a small region of the brain, the dorsal raphe, is the main source of the neurochemical serotonin that has been implicated in stress responsiveness and mood disorders. These serotonergic neurons can be subdivided according to rostralcaudal location, axon thickness, or projections ([@pbio-0000017-Mamounas1]; [@pbio-0000017-Lowry1]). However, what we recognize by immunochemical stain as a single shared phenotype in an anatomically distinct region may consist of a heterogeneous population of cells with diverse RNA profiles. In this sense, the strategy for brain-mapping might borrow a page from astronomy, with its maps of galaxies with mixed elements, as well as from the experience of the genome project. Indeed, advances in human brain-mapping, like discoveries in astronomy, have until recently largely depended on the tools available. The postmortem studies of the early 20th century provided delineation of cortical areas through light microscopic histology and gross connectional information. Neurochemical techniques in the last three decades yielded maps with cellular and subcellular resolution, identifying populations of cells usually by one or two neurochemical phenotypes. During the same period, electrophysiological approaches revealed the exquisite distribution of function across the brain, within particular brain subdivisions, and within neurons themselves. In the past two decades, direct study of the intact, functioning human brain in healthy and disordered states has been made possible by a variety of neuroimaging modalities. These studies have provided both structural and functional topography at increasing resolution, as well as neurochemical data and, most recently, information regarding neural connectivity ([@pbio-0000017-Behrens1]). The advent of techniques for mapping RNA profiles now permits analysis of several thousand species of RNA even in a single neuron, resulting in exponential increases in information. As these approaches are combined with the experimental behavioral and clinical sciences, opportunities abound for understanding this complex organ and treating its pathologies. The challenge now is to integrate this information into a coherent, accessible form that permits hierarchical analysis from RNA to protein to morphology to connectivity to function in a universal language while preserving fidelity. While earlier comprehensive maps in simpler nervous systems, such as the classic lineage maps of invertebrates ([@pbio-0000017-Stern1]), could be completed by single labs, more ambitious projects like a transcriptional map of the mouse brain, the Human Genome Project, and other goal-directed or large-scale research endeavors ([@pbio-0000017-Nass1]) will require *collaboration* of scientists who add value to the enterprise by working in multidisciplinary teams; *coordination* of efforts to attain a goal; and *computation* through the use of informatics, models, and simulations. The keystone in this new paradigm is, of course, meaningful data-sharing. Several initiatives serving the brain and behavioral research communities are advancing cooperative research. The Gensat Project ([www.gensat.org](www.gensat.org)) will soon provide developmental and whole-brain maps of several hundred genes in the mouse nervous system using a bacterial artificial chromosome (BAC) transgenic strategy with fluorescent reporters to provide subcellular resolution. A digital atlas of the mouse brain and associated informatics tools have been developed to organize, visualize, and analyze such gene expression (and other spatial) data generated by researchers (<http://www.loni.ucla.edu/MAP/index.html>). We now have the capability to map the transcriptional expression of virtually the entire mouse genome in the adult and the developing mouse brain, registering these data to a common, digital atlas. Like the galaxy maps generated by the Hubble telescope, this transcriptional atlas will provide important temporal as well as spatial information, revealing genes that may be expressed only at critical stages of brain development. Similarly, the Human Brain Project (<http://www.nimh.nih.gov/neuroinformatics/index.cfm>) is an informatics effort funded through several federal agencies to develop databases, analytical and computational simulations, and other resources to assist human brain-mapping as well as other large-scale coordinated neuroscience programs. While there are several initiatives at NIH aimed at overcoming the informatics barriers to sharing data and facilitating collaboration, coordination, and computation, we recognize that not all of the impediments to data-sharing are technical. The advent of neurobiology as an information science also demonstrates that the academic culture in which our science develops and the publication culture in which our science is communicated will need to change. Promotion decisions at major universities largely depend on the quality and quantity of first-authored or senior-authored papers. Multidisciplinary studies require teams of investigators in which hierarchical schemes for authorship may fail to reflect accurately the magnitude of each individual\'s contributions. Similarly, contributions to a research database may represent important scientific and scholarly achievements, but generally are underrecognized by promotions committees counting peer-reviewed publications. Indeed, the nature of publication itself needs to change in an era when some of the most important contributions will emerge from comprehensive descriptions of new landscapes (analogous to new genomes and new galaxies) rather than tests of specific hypotheses. These cultural issues are not peculiar to brain and behavioral science, of course, and have recently been considered broadly at the NIH (<http://www.becon.nih.gov/symposium2003.htm>). Scientific publication, as we have known it in print, is slow and expensive, with access limited to those with either the funds to purchase an individual subscription or the proximity to a library with an institutional subscription. Data-sharing also means open-access publishing so that data, whether from mapping efforts or from hypothesis-driven experiments, become available quickly and freely to the scientific community. As we emerge from the "decade of the brain," we are entering a decade for which data-sharing will be the currency for progress in neuroscience. Efforts driven by collaboration, coordination, and computation should yield the data, tools, and resources that neuroscientists will need in the coming decades. We hope that new electronic publications with open access will accelerate this change and provide the vehicle for disseminating the most exciting discoveries in neuroscience in a rapid, respected, and ready format. ::: {#pbio-0000017-g001 .fig} ::: {.caption} ###### A Constellation of Neurons Image courtesy of Miles Herkenham. :::  ::: The authors are all at the National Institutes of Health in Bethesda, Maryland, United States of America. Thomas R. Insel is the director of the National Institute of Mental Health. Nora D. Volkow is the director of the National Institute on Drug Abuse. Ting-Kai Li is the director of the National Institute on Alcohol Abuse and Alcoholism. James F. Battey, Jr., is the director of the National Institute on Deafness and Other Communication Disorders. Story C. Landis is the director of the National Institute of Neurological Disorders and Stroke.

PubMed Central

2024-06-05T03:55:45.968183

2003-10-13

PMC212697

Rich Lenski decided he was onto a good thing from his very first encounter with digital evolution. It all began when he used the technology in which artificial organisms in the form of computer code evolve independently by self-replicating, mutating, and competing to re-examine an earlier study with bacteria. The original study had contradicted 'some influential theory' suggesting that random mutations show a systematic tendency towards synergistic interactions. His digital results, he discovered, matched his organic ones. 'It\'s great when these two powerful experimental systems agree, because it suggests some generality about the evolution of genetic architectures\', recalls Lenski, professor of microbial ecology at Michigan State University (MSU). 'But even if the digital and biological realms sometimes come into scientific conflict, it would only lead one to ask why and then probe the relevant factors more deeply'. Complex Challenges and the Virtue of Simplicity {#s2} =============================================== He can hardly contain himself. 'It\'s a win--win situation, leading towards increased generality, on the one hand, and further experiments to better understand specific outcomes, on the other'. For his part, Lenski has since gone much further with the technology ([Box 1](#box1){ref-type="boxed-text"}; [Figure 1](#pbio-0000018-g001){ref-type="fig"}) and also soon expects to be announcing results that could broaden digital evolution\'s appeal even more. ::: {.caption} ###### Box 1. Impossible Evolutionary Experiments ::: Richard Lenski is using digital organisms to do 'impossible' evolutionary experiments. In one, he says, 'we test every incipient mutation before it occurs in a population and then allow it or disallow it, depending on its fitness effect, to see how important neutral and deleterious mutations are for long-term adaptation'. Lenski, professor of microbial ecology at Michigan State University, says his mind boggles at how digital evolution opens up so many avenues for research. 'I sometimes feel like a kid in a candy store who might starve because he can\'t make up his mind what he wants'. These opportunities and, at the other end, the prospect of having too much data to analyse, which Lenski admits is a strange thing for an evolutionary biologist to complain about, enforce a discipline to prioritise and define objectives: 'What *exactly* is the hypothesis I want to test, and what *exactly* must I measure to test that hypothesis?' Such enthusiasm for the technology makes it difficult for him to understand why some biologists might dismiss digital evolution as 'very interesting but with no value' or turn their backs on it altogether. 'My own view', says Lenski, 'is that something that is very interesting is also worth thinking about and exploring more fully, especially when it offers the opportunity to examine complex problems in greater depth and with more precision than is otherwise possible'. But he cautions against mistaking his enthusiasm for studying digital organisms as a call to abandon other lines of research. 'There\'s obviously much of value for understanding evolution that comes from many different empirical and theoretical perspectives', he says. 'That\'s one reason that evolutionary biology is such a vibrant field right now'. Lenski still spends as much research time on bacteria as he does on digital organisms. 'Although it\'s sometimes frustrating not to be able to devote 100% to each system, each one is so interesting to me that I couldn\'t bear to drop either of them'. The two systems have different strengths and limitations, which Lenski tries to exploit in his research, he says. From his laboratory\'s studies on long-term E. coli populations, he and his colleagues showed earlier this year how they used gene-expression arrays to work backwards to a set of key mutations in a global regulatory gene. More recent work, currently being written up, 'has led us to some adaptive mutations in several other key loci', he notes. As for his digital research using the Avida software system, Lenski acknowledges that speed is an obvious advantage, but not the most significant one. 'An even more important advantage is the ability to observe the dynamics and dissect the outcomes of evolution with absolute precision. For example, there are no missing links in the digital world'. Nevertheless, he wryly highlights one shortcoming of Avida: 'We\'ll know that we have been successful once the Avidians have evolved the ability to design their own experiments and write the papers without us'. ::: {#pbio-0000018-g001 .fig} Figure 1 ::: {.caption} ###### Hybrid Graphic of Petri Dishes with Bacteria Blending into Digital Organisms Lenski spends as much research time with bacteria (left) as he does with digital organisms (right), balancing the strengths and limitations of the two systems in an effort to understand and explain the principles of evolutionary theory. (Hybrid graphic courtesy of Dusan Misevic, Michigan State University.). :::  ::: "This is a world where time-scales contract and, above all, where other constraints of 'wet' biology have no place." Earlier this year, he and Chris Adami, who heads the Digital Life Laboratory at the California Institute of Technology (Caltech), published some breathtaking findings from the field. Their collaboration brings together biologists and computer scientists, physicists and philosophers in an artificial world on a quest to understand how evolution works. Though they may still be some way from reaching that objective, their latest advance suggests that they are on the right track. The research confronts evolutionary theory\'s long-standing challenge to explain how an organism can develop complex features simply as a result of random mutation and natural selection. The challenge remains a controversial one, too. Supporters of intelligent design, a branch of the creationist movement, promote the notion of 'irreducible complexity' as evidence that Darwinian evolution is a flawed theory. The notion purports that a complex feature cannot evolve sequentially from its elements, and must have been designed in one step by some higher intelligence. Traditional investigations, based on molecular biology and palaeontology, have yielded much evidence about the incremental evolution of the eye or the brain, for instance. But continuing ignorance about many developmental processes and the absence of key fossil records mean that accounts without missing links, to endorse the theory, may never be realised. Which is what tempted Lenski and Adami to examine the challenge in their virtual world. This is a world where timescales contract and, above all, where other constraints of 'wet' biology have no place. 'It\'s not just the speed, by any means', says Lenski. 'It\'s also the power to manipulate almost any variable one can imagine, to measure variables with absolute precision, to store information that then allows one to trace back a complex chain of events, and to take evolved organisms and subject them to new sorts of analyses that one might not even have anticipated when first collecting the data'. It is a place where virtue is made of simplicity. 'The worlds we\'re dealing with here are extraordinarily simple compared with the real world', says Adami. 'Any of the biochemistry associated with transcription and translation, for example, anything more complex than relatively short viral types of genomes, that\'s out of our league', he notes. 'We can\'t see transcription and translation because we don\'t have transcription and translation---we go right from sequence to function'. But the principles of evolutionary theory make such restrictions unimportant, he says. 'Many of the \[theory\'s\] predictions don\'t depend on these little details of molecular biology', notes Adami. 'The principles are very, very general, and very simple, and in the end they are mostly responsible for the overall dynamics that you see in these simple systems'. Lenski goes further. These virtual realities, he says, 'offer us a window into an alternative world, and perhaps even a part of the future of our own, where the fundamental evolutionary mechanisms of mutation and natural selection play out in a novel physical realm'. Lenski is interested in watching evolution as it happens and has a track record in the study of evolving organic systems, primarily using Escherichia coli. 'We\'re making great strides elucidating the precise genetic bases of the adaptation that has occurred during tens of thousands of generations in our long-term E. coli populations', he reports. 'Even after more than 30,000 generations in a constant environment, we\'re still seeing some major phenotypic evolutionary changes', he adds. Evolution in Action {#s3} =================== Adami, who also works in theoretical physics at the Jet Propulsion Laboratory at Caltech, has developed a software platform, known as Avida, for research on evolving computer programs, the digital organisms that he terms 'Avidians'. The second version, Avida 2.0, became available for free public use (<http://dllab.caltech.edu/avida/>) earlier this year, a decade after work began. 'I came to Caltech in 1992 on a special fellowship', he recalls, 'which basically told me, "You can do whatever you want and we\'re not going to check on you for three years---just sit there and think of something"'. So he did---and discovered the pioneering work on evolving computer programs by Tom Ray, the computational ecologist who invented the Tierra software system. 'In a sense, Tom Ray\'s Tierra was a proof of concept---he showed that computer programs can evolve, and it was a watershed moment. Without his work, mine wouldn\'t have existed', acknowledges Adami. 'But I wanted this digital life system to be an experimental system just like, let\'s say, Rich Lenski and E. coli bacteria'. Adami worked quickly with the help of undergraduates to design and write code and soon had a beta-version ready: 'Sure, these kids can program', he laughs. But the programmers were human and errors crept in. The team would run the system overnight and discover 'weird things' the next morning: 'The path of evolution went in a strange way, not because the world dictated it, but because some bug dictated it', notes Adami. 'You need to know your system perfectly, at least at the beginning, and that was really the hard part for the next five years'. On the way, however, the work attracted the attention of Microsoft, the software company, which was eager to know how its designers could evolve computer programs instead of writing them and inevitably introducing bugs, too. Some software already stretches to more than 10 million lines of code, and Microsoft, concerned for its survival as the fittest, foresaw a problem. It predicted programs expanding so much that, sometime between 20 and 50 years into the future, they would reach what Adami calls the 'complexity wall', where the number of errors would make them unusable. The alternative of evolving programs looked like a great idea to Microsoft, especially the way Adami tells it. 'I know a piece of software that\'s 3 billion lines of code that controls all our actions', he says, referring to the human genome. 'There may be bugs, but they don\'t lead to a crash. It\'s very robust programming, with pieces taken from all kinds of different sources, and somehow it works. And the reason why it works is because it was evolved and not written'. For a year, the Caltech team explored the features of programming languages that make one language more evolvable than another, but moved on when Microsoft\'s interests switched to more directly applied science and Adami wanted to continue to focus on the fundamental principles underpinning evolution. Avida was ready to run and beginning to offer a much more versatile platform than Tierra, with advances that have since been honed even further. 'We can exchange not only the \[processor\'s\] instruction set on the fly, we can also change the entire structure of the CPU \[central processing unit\] on the fly', says Adami. 'If you want to test different physics or chemistry, the flexibility of Avida compared with Tierra is like the difference between driving a modern Porsche and a Model-T Ford. They\'re both cars, but ...' The most important difference, insists Adami, 'is the possibility of rewards to programs if they accomplish interesting things, in this case computations'. He draws a parallel between the way replicating micro-organisms exploit chemical reactions to yield energy and the way evolving Avidians perform computations to secure extra CPU time. 'It\'s a one-to-one analogy', notes Adami, 'and the fact that it works so well may tell you something very, very fundamental about the duality between computational chemistries and biochemical chemistries'. In Adami\'s collaboration with Lenski to show how complex features can evolve sequentially, the Avidian genome is a circular sequence of instructions in computer code. At the start of its computational existence, an Avidian can only replicate. If it evolves logic functions in the process, however, the system rewards it with energy, in the form of time on the CPU. This reward enables the evolving Avidian to execute instructions that in turn help it to mature to secure more rewards, and so on, to safeguard its future. The results thrilled the experimenters. Teams at Caltech and MSU were able to trace the genealogy of Avidians, without any missing links, from simple self-replicator through unexpected transitional form to complex performer of many logic functions, with random mutation and natural selection alone responsible for the evolution. 'Many biologists are delighted to see such a clear demonstration of the evolution from scratch of demonstrably complex features', says Lenski, 'and in a way that accords so well with the hypothesis first voiced by Darwin and nowadays supported by a large body of comparative data that complex new features arise by co-opting existing structures that previously served other functions'. He also notes much interest in the way that damaging mutations sometimes proved to be essential stepping stones in the evolution of new functions. To opponents of evolutionary theory, Lenski is eager to emphasise that the study 'does not address the origin of life, nor whether the universe itself was designed to allow the evolution of complex organisms. Rather, our study shows that random mutation and natural selection can produce quite complex features, via many pathways, provided that the environment also favours some (but not all) transitional forms, even when the transitional forms are favoured for performing different functions from those that evolve later'. The Limits to Truth {#s4} =================== For many other biologists, however, digital evolution seems to have very little relevance. One eminent British evolutionary biologist dismissed the research in just eight words, according to the field\'s godfather, Tom Ray. 'His comment: "It\'s just not biology. Period. End of discussion". That\'s the whole story right there', recalls Ray. Less strident reservations concern the limits on complexity that the virtual world imposes and suspicions about the ability of digital processing to mirror evolutionary principles accurately. For Francisco Ayala, professor of biological sciences at the University of California, Irvine, it appears to be simply a question of trust in the natural world. 'Computers can give you only what you put in', he says. 'With natural models, you\'re not putting anything in---you\'re segregating a small region as an aspect of reality'. There are also more mundane worries over the technical skills needed for the computational operations, a fear acknowledged by Lenski. 'Computational skills are certainly opening up some exciting new directions \[in evolutionary biology\]', he says, 'but there are of course many other useful skills and fascinating directions'. At Caltech, meanwhile, Adami\'s team is trying to make Avida easier to use, backed by the National Institutes of Health\'s first-ever funding for digital-life work. Misunderstandings about the technology arise over whether the research is an 'instance' or a 'model' of evolution, suggests Ray, who now divides his time between the Advanced Telecommunications Research Laboratories in Kyoto and the University of Oklahoma, where he holds chairs in zoology and computer science. 'I never intended \[Tierra\] as a model, but that\'s the way a lot of people saw it because they weren\'t really prepared for this new idea, this different perspective of another instance of life', he says. 'They had a more traditional view of what you do with a computer, which is that you send e-mail, you process things, and you make models'. Levels of veracity determine limits of extrapolation, says Ray. 'Digital evolution is an abstraction, and it\'s not going to be able to tell us what humans will evolve into or why dinosaurs went extinct or what will be the next emerging disease.... You need the whole planet to do that kind of modelling'. But once you appreciate the constraints, 'it\'s a phenomenally good tool, because it\'s evolution in a bottle. You can instrument it 100%', he notes. 'I think Lenski and Adami have done a very good job of developing it that way'. Ray himself is now more interested in genomics and pharmacology and their application in a biologically inspired engineering project to design software agents, or 'virtual creatures', as he terms them. For Lenski, experiments with Avida provide 'both an "instance" and a "model" of evolution'. He says that 'populations of the digital organisms really do evolve and adapt, albeit in an unfamiliar physical realm. At the same time, they provide a sort of experimental model for testing and understanding the general principles of evolution'. And he agrees with Ray that digital evolution is not intended to explain how we got where we are today, 'in the sense of unravelling which species are more related to which other species, or what organismal features are adaptive for what purposes, and so forth'. The goal, says Lenski, is to examine evolutionary processes and dynamics in greater depth and detail than are otherwise possible. 'Watching a process as it occurs and being able to probe genetic details and manipulate environmental variables can provide new insights and evidence that one cannot get by comparative studies that typically require one to infer historical processes from present-day patterns'. The First Steps to Freedom {#s5} ========================== Such developments fascinate and enthral Paul Rainey, an evolutionary ecologist, even though he rarely needs any computing power for his research and recognises that digital evolution still lacks an ecological dimension. Rainey, who earlier this year moved from Oxford to become professor of ecology and evolution at the University of Auckland, uses bacterial populations of Pseudomonas fluorescens, which grow from single genotypes in pristine tubes, to test long-standing hypotheses about the causes of ecological diversification. 'The bottom line is that we\'re reducing the complexity we see in the real world to a much more manageable level', he says. 'The nice thing about bacterial populations is that ecological and evolutionary timescales coincide, so that you can actually see the ecological context of evolutionary change'. "'It\'s a phenomenally good tool, because it\'s evolution in a bottle.'" Rainey, a friend and colleague of Lenski\'s, would welcome the chance to take advantage of the speed, robustness, and flexibility of digital evolution to further his research, but doubts whether the technology will ever be able to match the performance of his 'wet' laboratory. Though his natural model is simple, it remains far too complex to program, he suspects. 'We try to understand how selection is working in this very complex ecological context, which includes interactions between genotypes and within genotypes and interactions with an environment that is constantly changing', he says. 'This sets the scene for selection, and the selective forces are constantly changing.... None of that complexity is really captured in Avida'. But Rainey is in for a surprise, according to Adami. 'The pace of development of Avida has accelerated', he says. 'More people are working on it because we have bigger grants. And Charles Ofria \[who helped to design the software as a postgraduate at Caltech\] is doing much of the development at Michigan State \[University, where he is now assistant professor of computer science and engineering\] with his students'. The result is that Avidians have made their first steps towards sexual freedom within ecologically diverse environments or, more accurately, code recombinations in a multi-niche virtual world. For almost a decade, says Adami, Avida has been a single-niche world in which every organism in the population sees exactly the same world and only a single species inhabits that world. But Avida has now been expanded, he continues, 'in such a manner that populations can see different types of worlds and they can adapt independently to different resources'. A research paper is being finalised on how the software is making its first steps towards incorporating the notion of evolutionary ecology. 'We show what pressures are necessary to make a population that is homogenous branch out and speciate into a stable system', notes Adami. 'Now we want to explore recombination, which we\'ve always shied away from.' With asexual reproduction virtually understood, the researchers are ready to tackle sexual reproduction in the digital world, says Adami. 'Some people are furiously working at implementing that.' 'Our goal is not to mimic natural systems in detail, but rather to expand Avida to give digital organisms access to more of the basic processes of life', says Lenski. 'Our goal is not so much to endow the ancestral organisms with additional capabilities, but rather we want to see how digital organisms will evolve if they are placed in an altered world where such things as sex and communication are physically possible. I see many years of interesting research along these lines'. Reflecting on future applications for the research, Lenski suggests it highlights how the traffic in computational biology is now becoming a significant and little recognised twoway exchange. Computer scientists are not only helping biologists to organise and analyse their vast datasets, says Lenski, but 'biological principles, from evolution and genetics to neurobiology and ecology, are informing computer scientists and engineers in designing software and hardware ... and that holds tremendous promise for the future'. Bill 0\'Neill is a freelance journalist from London, United Kingdom. E-mail: <billoneill@cix.co.uk>.

PubMed Central

2024-06-05T03:55:45.969435

2003-10-13

PMC212698

Introduction {#s1} ============ Type 2 diabetes is a serious metabolic disease associated with an increased risk of premature death and substantial disability, largely mediated through its adverse effects on the vasculature. The prevalence of the disease is increasing, and the World Health Organisation estimates suggest that by 2025 there will be 300 million affected individuals worldwide ([@pbio-0000020-King1]). The disorder is characterised by a combination of impaired insulin secretion and insulin action, both of which precede and predict the onset of disease ([@pbio-0000020-Weyer1]). Through its adverse impact on insulin action, obesity is a major risk factor for the disease. Although environmental factors, both post- and prenatal, play an important role in determining the risk of disease, a substantial body of evidence supports the notion that disease susceptibility is influenced by inherited factors ([@pbio-0000020-Zimmet1]). While the molecular basis for several uncommon Mendelian forms of Type 2 diabetes have been defined ([@pbio-0000020-Vionnet1]; [@pbio-0000020-Yamagata1], [@pbio-0000020-Yamagata2]; [@pbio-0000020-Horikawa1]; [@pbio-0000020-Stoffers1]; [@pbio-0000020-Barroso1]; [@pbio-0000020-Malecki2]; [@pbio-0000020-Savage1]), the nature and range of allelic variants conferring susceptibility to the more common forms of this disorder remain poorly defined. Many investigators have embarked on attempts to identify diabetes susceptibility genes through genome-wide linkage-based approaches using multigenerational pedigrees and/or large numbers of affected sibpairs. Regions of significant linkage, some of which have been replicated in more than one study, have been identified. To date, however, only Calpain 10 (*CAPN10*; LocusLink ID 11132) has emerged from such studies as a new putative diabetogene ([@pbio-0000020-Horikawa2]). While some subsequent studies have supported a role for the *CAPN10* alleles originally described as susceptibility alleles, others have found associations with different alleles and some have found no association with this gene ([@pbio-0000020-Baier2]; [@pbio-0000020-Cox1]; [@pbio-0000020-Evans1]; [@pbio-0000020-Hegele2]; [@pbio-0000020-Tsai1]; [@pbio-0000020-Daimon1]; [@pbio-0000020-Elbein1]; [@pbio-0000020-Garant1]). The positional cloning effort has been supplemented by a large number of studies examining specific candidate genes using a variety of methodologies, mostly of the case-control association design. Although many positive reports have emerged, few have been consistently replicated. Of these candidates, the most compelling evidence to date, generated from a meta-analysis of multiple published studies, is that a common amino acid variant in the N-terminus of the nuclear receptor peroxisome proliferator-activated receptor γ (*PPARG*; LocusLink ID 5468) confers significant protection against the development of Type 2 diabetes ([@pbio-0000020-Altshuler1]). More recently, evidence has accumulated supporting a role for the E23K variant of *KCNJ11* (LocusLink ID 3767) in Type 2 diabetes predisposition ([@pbio-0000020-Hani1]; [@pbio-0000020-Gloyn1], [@pbio-0000020-Gloyn2]; [@pbio-0000020-Love-Gregory1]; [@pbio-0000020-Nielsen1]). Whole-genome association studies in large case-control populations may ultimately have the greatest power to detect alleles of small but significant effects on the susceptibility to common diseases such as Type 2 diabetes. As yet, however, the resource implications of such an approach are prohibitive. In the meantime, knowledge of both mammalian biology and disease pathogenesis is progressing rapidly, and it is possible to identify a large panel of known genes, the dysfunction of which might reasonably be considered likely to contribute to Type 2 diabetes. In this study we have identified 152 informative single nucleotide polymorphisms (SNPs) in 71 such genes and, using these, have examined their association with Type 2 diabetes and related intermediate phenotypes in Caucasian subjects from the United Kingdom. Results/Discussion {#s2} ================== Overall Study Design {#s2a} -------------------- Candidate gene studies are based on selection of genes with a known or inferred biological function whose role makes it plausible that they may predispose to disease or the observed phenotype. These types of studies are similar to traditional epidemiological approaches in which an a priori hypothesis between exposure to a given factor (in this case, a genotype at a given locus) and disease is formulated. To date, most Type 2 diabetes candidate gene studies have largely lacked thoroughness and sensitivity, as they have tested a limited number of genes and variants in small populations or in populations that were poorly matched or phenotyped, frequently resulting in a lack of replication of the weak associations detected ([@pbio-0000020-Altshuler1]). Our strategy aimed to address these problems by a unique combination of features, including comprehensive SNP discovery in a large number of candidate genes, testing of a large number of SNPs, use of two independent populations, and analysis of haplotypes in addition to individual SNPs where possible. [Figure 1](#pbio-0000020-g001){ref-type="fig"} illustrates the overall design of the study. On the basis of their known or putative role in glucose metabolism, 71 candidate genes were selected for study ([Table 1](#pbio-0000020-t101){ref-type="table"}). These were subdivided into three broad groups: (1) genes primarily involved in pancreatic β-cell function; (2) genes primarily influencing insulin action and glucose metabolism in the main target tissues, muscle, liver, and fat; and (3) other genes. This group includes genes that influence processes potentially relevant to diabetes, such as energy intake, energy expenditure, and lipid metabolism. A de novo search for common SNPs was undertaken using fluorescent single-stranded conformation polymorphism (fSSCP) examination of all coding regions and splice junctions in a variety of human populations. All genes were minimally screened against 47 samples of mixed ethnicity, providing 0.99 probability of detecting variants with a minor allele frequency of 0.05. Our 'in-house\' polymorphism detection programme identified 954 SNPs in the 71 genes, with a range of allele frequencies from 0.003 to 0.50. Of the 152 SNPs chosen for further study ([Table S1](#st001){ref-type="supplementary-material"}), the great majority had a minor allele frequency of greater than 5%, but in a few instances less frequent variants were typed when the candidate gene had strong biological plausibility and there were no known polymorphisms of higher frequency at the time of SNP selection. ::: {#pbio-0000020-g001 .fig} Figure 1 ::: {.caption} ###### Study Design Candidate genes were selected based on known or putative function. A de novo polymorphism discovery step was undertaken to identify novel variants for association studies. We selected 152 SNPs and tested them in a case-control study and a QT study. Association analysis with Type 2 diabetes was done for SNPs and haplotypes under multiple genetic models. Only SNPs and haplotypes associated with disease were evaluated for association with five diabetes-related QTs under the same model in the QT study. :::  ::: ::: {#pbio-0000020-t101 .table-wrap} Table 1 ::: {.caption} ###### Genes with SNPs Genotyped in This Study :::  Candidate genes, identified by official HUGO Gene Nomenclature Committee symbols, are grouped by known or putative biological function, with the number of genotyped polymorphisms per gene shown ::: ::: {#pbio-0000020-t102 .table-wrap} Table 1 ::: {.caption} ###### Continued :::  ::: The 152 SNPs were genotyped in a population-based cohort of 517 unrelated Caucasians in the United Kingdom with Type 2 diabetes and an equal number of controls with normal glycated haemoglobin (HbA1c) levels, individually matched to cases by age, sex, and geographical location. A second independent population was also genotyped for the same 152 SNPs. This consisted of 1,100 middle-aged Caucasian subjects in the United Kingdom who had been extensively and serially phenotyped for glucose tolerance and variables related to insulin secretion, insulin action, and adiposity. In the first stage of data analysis, all SNPs (and haplotypes when multiple SNPs were present at the same gene) were examined for their association with diabetes in the case-control study using multiple models of inheritance. In the second phase of analysis, all SNPs and haplotypes showing statistically significant association with diabetes status in the first phase were examined for association with glucose levels and other intermediate phenotypes in the quantitative trait (QT) study population. The intermediate phenotypes chosen for study were fasting and 2-h post-glucose load plasma glucose levels (measures of glucose tolerance), fasting insulin (a measure of insulin sensitivity), 30-min insulin incremental response (a measure of β-cell function), and body mass index (BMI) (a measure of adiposity). Power to detect an association is dependent on several factors: the frequency of the 'predisposing\' allele, genotype, or haplotype; the accepted false-positive or Type 1 error rate (α); and the odds ratio (OR) or effect size. Rarer alleles, genotypes, or haplotypes with small effects require larger sample sizes to attain the same power to detect an association, as compared to more frequent alleles or alleles with larger effects. At the time that we collected the populations and designed this study, our power calculations had shown that a sample size of 500 cases and 500 matched controls would have 80% power to detected effect sizes as small as 1.3--1.7 OR, depending on the frequency of the predisposing allele, with a 5% Type 1 error rate ([Figure 2](#pbio-0000020-g002){ref-type="fig"}). ::: {#pbio-0000020-g002 .fig} Figure 2 ::: {.caption} ###### Power Calculations Power of the current Cambridgeshire Case-Control Study to detect associations with risk allele of varying frequencies and with a Type 1 error rate of 5%. Abbreviations: p0, frequency of the predisposing allele; chr, number of chromosomes. Graphs were plotted with the PS power and sample-size program (available at <http://www.mc.vanderbilt.edu/prevmed/ps>; [@pbio-0000020-DuPont1]). :::  ::: Overview of Results of Association Studies {#s2b} ------------------------------------------ [Table S1](#st001){ref-type="supplementary-material"} shows the genotype counts for all 152 SNPs in the case-control and QT studies. In the control subjects, 16 SNPs (10.6%) had a minor allele frequency below 5%; 19 (12.5%) had a minor allele frequency between 5% and 10%; and 117 (76.9%) had a minor allele frequency greater than or equal to 10%. Each variant was tested for association with disease status under several genetic models**.**Twenty SNPs in 11 different genes showed statistically significant association with disease status (*p* \< 0.05) under at least one model ([Table 2](#pbio-0000020-t002){ref-type="table"}). The strongest statistical evidence for disease association was for genes *SOS1* (LocusLink ID 6654), *SLC2A2* (LocusLink ID 6514), *PIK3R1 (*LocusLink ID 5295), *ABCC8* and *KCNJ11* (LocusLink ID 6833), and *INSR* (LocusLink ID 3643). Of the 29 loci with multiple SNPs, only three---*HNF4A* (LocusLink ID 3172) *INSR*, and *ABCC8*--*KCNJ11*---showed significant association of particular haplotypes with disease status ([Figure 3](#pbio-0000020-g003){ref-type="fig"}; [Table 3](#pbio-0000020-t003){ref-type="table"}). In only one case (*HNF4A*) was a haplotype significantly associated with disease risk (see below) when no significant association was seen with any individual SNP in that gene. [Table S2](#st002){ref-type="supplementary-material"}, shows the results of association studies undertaken in the QT population, further examining the SNPs that had shown significant association in the case-control study. [Table 3](#pbio-0000020-t003){ref-type="table"} shows the relationship between disease-associated haplotypes at *ABCC8*--*KCNJ11*, *HNF4A*, and *INSR* with intermediate phenotypes in the QT study. We now consider in more detail, the data for those genes where the strongest and most consistent effects were seen. ::: {#pbio-0000020-g003 .fig} Figure 3 ::: {.caption} ###### Genes with Haplotypes Associated with Type 2 Diabetes Genomic organization with exons (black boxes or vertical lines) and with genotyped SNPs and SNPs utilised in the haplotype reconstructions (in blue boxes) is shown. The most common haplotypes with population prevalence greater than 5% in the control population are shown, and the measure of LD (*r^2^*) is shown for a subset of the SNPs. (A) *ABCC8--KCNJ11*. (B) *HNF4A*. (C) *INSR*. :::  ::: ::: {#pbio-0000020-t002 .table-wrap} Table 2 ::: {.caption} ###### Genes with Variants Significantly Associated with Type 2 Diabetes Status :::  SNP identifiers (SNPID), OR, significance level (*p* value), and genetic model are shown. *p* values for the additive effect are for the test for a linear trend across the genotypes, which were coded as 0 = 11, 1 = 12, 2 = 22. Allele 2 dominant refers to a combination of 12 + 22 and allele 2 recessive refers to combination of 11 + 12 ::: ::: {#pbio-0000020-t003 .table-wrap} Table 3 ::: {.caption} ###### Association of *ABCC8--KCNJ11*, *HNF4A*, and *INSR* Haplotypes with Diabetes and QTs :::  For case control, OR and 95% CI of haplotypes are shown. For QT, means and 95% CI of haplotypes are shown. Means were adjusted for age and sex, but not for BMI. Abbreviations: BMI, body mass index (kg/m^2^); PG0, fasting plasma glucose (mmol/l); 2hPG, 2-h plasma glucose (mmol/l); INS0, fasting insulin (pmol/l); Ins inc, 30-min insulin increment (pmol/mmol). Associations significant at 0.10 and below are in***italics bold***, with association 0.05**** or below in**** **red bold** ***.*** ::: Genes Primarily Affecting β-Cell Function {#s2c} ----------------------------------------- ### *ABCC8* and *KCNJ11* (encoding, respectively, the sulphonylurea receptor and inwardly rectifying potassium channel KIR 6.2). {#s2c1} The genes encoding the two molecular components of the voltage-gated potassium channel of the pancreatic β-cell are located within 4.5 kb of each other on Chromosome 11. *KCNJ11* encodes the channel protein KIR6.2 and *ABCC8* encodes an ATP-binding cassette (ABC) transporter-containing transmembrane protein (SUR1) that is thought to regulate the activity of the channel and that also contains the site to which sulphonylurea antidiabetic drugs bind. Three SNPs in *KCNJ11* were associated with disease under multiple genetic models. The strongest statistical association in this gene was with a 3′-SNP (SNP74; OR 0.59, *p* = 0.0027 under recessive model for allele 2) (see [Table 2](#pbio-0000020-t002){ref-type="table"}). In *ABCC8*, five SNPs were associated with disease status under multiple models; the strongest evidence for association with disease were with SNP79 and SNP81, respectively an intronic variant (OR 1.68, *p* = 0.0043; see [Table 2](#pbio-0000020-t002){ref-type="table"}) and a missense variant A1369S (OR 1.68, *p* = 0.0048; see [Table 2](#pbio-0000020-t002){ref-type="table"}). Although neither of these two SNPs was significantly associated with any trait in the QT study, two other SNPs showed effects in the QT study (see [Table S2](#st002){ref-type="supplementary-material"}). SNP84 (IVS18--36) from *ABCC8*, which associated with increased disease risk (OR 3.43, *p* = 0.0163; see [Table 2](#pbio-0000020-t002){ref-type="table"}) also associated with increased BMI (mean 28.2 kg/m^2^, 95% confidence interval \[CI\] \[26.6, 29.9\] for homozygous 22 versus 26.2 kg/m^2^, 95% CI \[25.9, 26.5\] for homozygous 11 and 26.3 kg/m^2^, 95% CI \[25.7, 26.8\] for heterozygous subjects; *p* = 0.016) and associated with borderline significance with higher fasting glucose (5.53 mmol/l, 95% CI \[5.29, 5.77\] for homozygous 22 versus 5.30 mmol/l, 95% CI \[5.26, 5.35\] for homozygous 11 and 5.27 mmol/l, 95% CI \[5. 19, 5.35\] for heterozygous subjects; *p* = 0.057) under a recessive model for allele 2 (see [Table S2](#st002){ref-type="supplementary-material"}). SNP87 (K649), which was also significantly associated with increased disease risk (OR 3.90, *p* = 0.0157; see [Table 2](#pbio-0000020-t002){ref-type="table"}), also showed borderline significant association with decreased insulin secretion (23.6 pmol/mmol, 95% CI \[18.6, 30.1\] for homozygous 22 versus 29.8 pmol/mmol, 95% CI \[28.4, 31.1\] for homozygous 11 and 30.9 pmol/mmol, 95% CI \[28.6, 33.4\] for heterozygous subjects; *p* = 0.054; see [Table S2](#st002){ref-type="supplementary-material"}), consistent with a role for this gene in insulin secretion. Given the close physical proximity of *ABCC8* and *KCNJ11* and their role in the same functional unit, we performed haplotype reconstructions with data from both genes combined (see [Figure 3](#pbio-0000020-g003){ref-type="fig"}A). Haplotype B was associated with increased disease risk (OR 1.46, 95% CI \[1.14, 1.85\]; data not shown), but did not show any significant association in the QT study (see [Table 3](#pbio-0000020-t003){ref-type="table"}). Mutations in each of these genes have been associated with familial persistent hyperinsulinaemia hypoglycaemia of infancy (PHHI), a rare disorder of glucose homeostasis characterised by up-regulated insulin secretion despite severe hypoglycaemia. In addition, evidence for association of *KCNJ11* DNA variants with Type 2 diabetes has been evaluated in multiple studies, and until recently these data have been conflicting. Several recent studies have, however, suggested a role for the aminoacid variant E23K in Type 2 diabetes susceptibility ([@pbio-0000020-Hani1]; [@pbio-0000020-Gloyn1], [@pbio-0000020-Gloyn2]; [@pbio-0000020-Schwanstecher1]; [@pbio-0000020-Love-Gregory1]; [@pbio-0000020-Nielsen1]). In total we tested four SNPs at the *KCNJ11* locus for association with disease status; of these, three were tightly linked (data not shown) and all three had a statistically significant association with disease status (see [Table 2](#pbio-0000020-t002){ref-type="table"}). In our study we replicated the effect of the E23K polymorphism in Type 2 diabetes predisposition (KK homozygous, OR 1.49, *p* = 0.0333; see [Table 2](#pbio-0000020-t002){ref-type="table"}) with an OR estimate in agreement with that demonstrated by the meta-analysis of [@pbio-0000020-Nielsen1]); in addition, two other *KCNJ11* SNPs associated with disease risk (SNP74 and SNP76). The recent evidence from multiple studies and from meta-analysis for association of the E23K SNP with Type 2 diabetes, along with in vitro studies using cell lines expressing the E23K mutation showing an increased stimulation threshold of insulinsecretion ([@pbio-0000020-Schwanstecher2]), suggests that E23K is the functional variant leading to increased disease risk. Given our finding of high levels of linkage disequilibrium (LD) between SNP74 and SNP76 with E23K (data not shown), we adjusted the measures of association at these sites for the E23K genotype. These data suggest that these SNPs are independently associated with diabetes ([Table 4](#pbio-0000020-t004){ref-type="table"}). ::: {#pbio-0000020-t004 .table-wrap} Table 4 ::: {.caption} ###### Association of *KCNJ11* and *ABCC8* Variants with Type 2 Diabetes Status Adjusted for E23K Genotype :::  SNP identifiers (SNPID), OR, significance level (*p* value), and genetic model are shown. *p* Values for the additive effect are for the test for a linear trend across the genotypes, which were coded as 0 = 11, 1 = 12, 2 = 22. Allele 2 dominant refers to a combination of 12 + 22 and the allele 2 recessive refers to combination of 11 + 12 ::: *ABCC8* variants have been associated with Type 2 diabetes in multiple studies ([@pbio-0000020-Inoue1]; [@pbio-0000020-Hansen5]; [@pbio-0000020-Hart3]). However, a recent large study failed to replicate previous associations with Type 2 diabetes ([@pbio-0000020-Altshuler1]). In our study we found evidence for association with Type 2 diabetes in five of 16 *ABCC8* SNPs tested. Owing to the physical mapping of *ABCC8* in close proximity to *KCNJ11*, we further investigated whether the associations at the *ABCC8* locus could be completely explained through LD between *ABCC8* SNPs and the E23K variant at *KCNJ11*. After adjustment for E23K, two *ABCC8* SNPs (SNP79 and SNP81) that were significantly associated with diabetes (*p* = 0.0043 and *p* = 0.0048 for the recessive model; see [Table 2](#pbio-0000020-t002){ref-type="table"}) prior to adjustment were no longer significantly associated with diabetes (*p* = 0.0536 and *p* = 0.1339 for the recessive model; see [Table 4](#pbio-0000020-t004){ref-type="table"}). However, for the remaining three SNPs (SNP84, SNP87, and SNP89), although the significance levels were reduced, they remained statistically significant (*p* = 0.0401, *p* = 0.0214, and *p* = 0.0445 for the recessive model; see [Table 4](#pbio-0000020-t004){ref-type="table"}). Moreover, the OR for two of these SNPs increased to 4.36 and 3.16, respectively. This suggests that there are effects at the *ABCC8* locus that are independent from the E23K *KCNJ11* variant. The lowered significance levels are likely due to loss of power resulting from the adjustment. Our data and that from at least nine other independent association and linkage studies ([@pbio-0000020-Hansen6]) have shown some evidence for *ABCC8* involvement in Type 2 diabetes and related phenotypes. ### *SLC2A2* (encoding GLUT2). {#s2c2} *SLC2A2* encodes the glucose transporter GLUT2, a member of the facilitative glucose transporter family that is highly expressed in pancreatic β-cells and liver. We typed six SNPs in *SLC2A2*, three of which (SNP21, SNP23, and SNP24)were significantly associated with diabetes status with an OR of approximately 1.4--1.5 (see [Table 2](#pbio-0000020-t002){ref-type="table"}). The most highly significant association was with a T110I substitution (OR 1.49, *p* = 0.0059) under a dominant model for the minor allele. In the reduction process prior to haplotype estimations (see Materials and Methods), only one SNP (SNP21) contributed significantly to disease association. Therefore, haplotype reconstructions were not performed. In the QT study, all three disease-associated SNPs were also associated with lower levels of fasting plasma insulin. Rather surprisingly allele 2 (A) at T198, which was associated with increased disease risk, was associated with lower 2-h plasma glucose. No other significant associations with intermediate phenotypes were seen. Multiple previous studies have sought evidence for association or linkage between *SLC2A2* variants and Type 2 diabetes, and most have reported negative results. However, all studies have been small and were insufficiently powered to detect effects of modest size ([@pbio-0000020-Li1]; [@pbio-0000020-Baroni1]; [@pbio-0000020-Tanizawa1]; [@pbio-0000020-Moller2]). *SLC2A2* is a highly plausible candidate gene for Type 2 diabetes, as it is a high K~m~ transporter that regulates entry of glucose into the pancreatic β-cell, thus initiating the cascade of events leading to insulin secretion. GLUT2 is also highly expressed in the liver, where it is involved in the regulation of both glucose uptake and output. It is notable that the alleles that associated with increased diabetes risk were also all associated with lower fasting insulin levels, suggesting that these may influence basal insulin secretion. However, interpretation is complex, as (1) fasting insulin is strongly influenced by insulin sensitivity and (2) the potential risk alleles were not associated with any impairment of insulin secretion in response to a glucose load. Finally, allele 2 (A) at T198, which associated with increased risk of diabetes in the case-control study, was associated with **lower** 2-h glucose in the QT study. Clearly, more detailed genetic mapping combined with functional studies (which will be challenging in humans owing to the inaccessibility of the pancreatic β-cell) will be needed to identify the mechanism whereby variants in this gene influence diabetes risk. ### *HNF4A* (encoding hepatic nucleotide factor 4α). {#s2c3} *HNF4A* (the *MODY1* gene) encodes an orphan hormone nuclear receptor that, together with *TCF1* (LocusLink ID 6927), encoding HNF1α, *TCF2* (LocusLink ID 6928), encoding HNF1β, and *FOXA2* (LocusLink ID 3170), encoding HNF3β, constitutes part of a network of transcription factors controlling gene expression in pancreatic β-cells, liver, and other tissues. In β-cells, these transcription factors regulate expression of the insulin gene as well as genes encoding proteins involved in glucose transport and metabolism and in mitochondrial metabolism, all of which are linked to insulin secretion ([@pbio-0000020-Fajans1]). While no individual SNP in *HNF4A* was significantly associated with disease status, we identified a haplotype (haplotype B in [Figure 3](#pbio-0000020-g003){ref-type="fig"}B) that was significantly associated with reduced disease risk (OR 0.83, 95% CI \[0.68, 1.00\]; data not shown). In the QT study, this 'reduced-risk\' haplotype wassignificantly associated with increased insulin secretion (mean = 31.5 pmol/mmol, 95% CI \[29.9, 33.3\] versus 29.3 pmol/mmol, 95% CI \[28.0, 30.6\] for haplotype A and 30.9 pmol/mmol, 95% CI \[28.1, 34.0\] for haplotype C\]. Carriers of this haplotype also showed a trend towards lower fasting and 2-h plasma glucose, compared to the subjects with the other haplotypes (see [Table 3](#pbio-0000020-t003){ref-type="table"}). *HNF4A* maps to Chromosome 20 ([@pbio-0000020-Argyrokastritis1]) in a region that has been linked to Type 2 diabetes in multiple studies ([@pbio-0000020-Bowden1]; [@pbio-0000020-Ji1]; [@pbio-0000020-Zouali1]; [@pbio-0000020-Ghosh1]; [@pbio-0000020-Klupa1]; [@pbio-0000020-Permutt1]). This positional information, combined with the known role of major mutations at this gene in the causation of autosomal-dominant maturity-onset diabetes of the young (MODY), has led to *HNF4A* being considered as a strong candidate for involvement in Type 2 diabetes. However, most studies to date have failed to identify an association between variants at this locus and disease susceptibility ([@pbio-0000020-Moller1]; [@pbio-0000020-Malecki1]; [@pbio-0000020-Ghosh1]; [@pbio-0000020-Price1]). This study differs from all other previous reports in its examination of haplotypes, as well as in the fact that it included several SNPs not previously examined. Our findings lead us to speculate as to how a particular *HNF4A* haplotype might be associated with lower risk of diabetes and increased insulin secretory capacity. The fact that a multiplicity of heterozygous nonsense and missense mutations in HNF4α lead to an insulinopaenic form of MODY strongly suggests that β-cell dysfunction is sensitive to the amount of HNF4α in the β-cell and that haploinsufficiency is the likely mode of molecular pathogenesis in that condition ([@pbio-0000020-Stoffel1]; [@pbio-0000020-Shih1]). It is plausible, therefore, that variants in this gene that enhance expression levels of the protein might lead to increased insulin secretory capacity and protection against diabetes. ### *INS* (encoding insulin). {#s2c4} The *INS* (LocusLink ID 3630) gene encodes the hormone preproinsulin, which upon proteolytic cleavage generates mature insulin and C-peptide. We tested for association of a single SNP in the 3′-UTR (SNP72) of the insulin gene with disease status. This SNP was significantly associated with increased Type 2 diabetes risk under a recessive model for the T allele (OR 2.02, *p* = 0.0258) (see [Table 2](#pbio-0000020-t002){ref-type="table"}). In the QT study this SNP did not associate with any of the intermediate phenotypes studied. The insulin gene variable number tandem repeat (*INS*--VNTR) has been extensively studied and is proposed to exert pleiotropic effects on birth weight and diabetes susceptibility ([@pbio-0000020-Huxtable1]). However, evidence for this has been conflicting and a role for *INS* in Type 2 diabetes predisposition has not been definitively established. The data for the single SNP we tested suggest that either the insulin gene or other loci in LD may be involved in Type 2 diabetes risk. Genes Primarily Affecting Insulin Action {#s2d} ---------------------------------------- ### *INSR* (encoding the insulin receptor). {#s2d1} At the *INSR* locus of the seven SNPs genotyped, we detected a single intronic SNP (SNP131) that was significantly associated with increased disease risk (OR 1.48, *p* = 0.0039 for the dominant model for allele 2) (see [Table 2](#pbio-0000020-t002){ref-type="table"}). In the QT study, this SNP also had a nonsignificant association with increased 2-h glucose under a dominant model for allele 2 (see [Table S2](#st002){ref-type="supplementary-material"}). Haplotype C (see [Figure 3](#pbio-0000020-g003){ref-type="fig"}C) for *INSR* was associated with increased disease risk (1.34 mmol/l, 95% CI \[1.05, 1.71\]; data not shown); in the QT study, there was a nonsignificant trend for subjects carrying this haplotype to have increased values for fasting glucose (5.32 mmol/l, 95% CI \[5.24, 5.39\] versus 5.27 mmol/l, 95% CI \[5.21,5.33\] for haplotype B and 5.27 mmol/l, 95% CI \[5.23, 5.32\] for haplotype A), 2-h glucose (6.00 mmol/l, 95% CI \[5.80, 6. 20\] versus 5.78 mmol/l, 95% CI \[5.61, 5.96\] for haplotype B and 5.87 mmol/l, 95% CI \[5.75, 5.99\] for haplotype A), and fasting insulin (41.8 pmol/l, 95% CI \[39.2, 44.5\] versus 40.4 pmol/l, 95% CI \[38.4, 42.5\] for haplotype B and 41.3 pmol/l, 95% CI \[39.6, 43.1\] for haplotype A) (see [Table 3](#pbio-0000020-t003){ref-type="table"}). A role for *INSR* in Type 2 diabetes and related phenotypes has long been sought. Many studies initiated over the past decade have explored the possibility that DNA variants at this locus would not only cause rare syndromes of extreme insulin resistance, but would also associate with increased Type 2 diabetes risk. In particular, the role of the Val985Met in disease predisposition has been analysed in many different populations, but the data remain inconclusive, with some studies suggesting a role for this variant ([@pbio-0000020-Hart1], [@pbio-0000020-Hart3]), while others do not support this finding ([@pbio-0000020-OaRahilly1], [@pbio-0000020-OaRahilly2]; [@pbio-0000020-Hansen4]). In this study we provide preliminary evidence for a role of *INSR* in diabetes susceptibility through genotyping of a previously untested SNP in case-control studies and via haplotype analysis using multiple SNPs in the gene. ### *PIK3R1* and *SOS1*. {#s2d2} The gene *PIK3R1,* encoding the p85α regulatory subunit of the phosphatidylinositol 3-kinase, is a logical candidate gene for involvement in the development of Type 2 diabetes owing to its role in insulin signal transduction. An intronic variant, SNP42, was associated with increased disease risk under two genetic models (OR 1.41, *p* = 0.0090 for the allele 2 dominant and OR1.34, *p* = 0.0088 for the additive model; see [Table 2](#pbio-0000020-t002){ref-type="table"}). In the QT study, SNP42 was significantly associated with increased BMI and showed a borderline significance with increased fasting insulin (measure of insulin resistance) under a dominant model for allele 2 (see [Table S2](#st002){ref-type="supplementary-material"}). Obesity is a major risk factor for insulin resistance, and the observed increase in BMI coupled with increased insulin resistance in carriers of allele G at SNP42 suggests that variation at this gene may be increasing Type 2 diabetes risk through impaired insulin action. Other association studies at this locus have focussed on investigating the Met326Ile variant in disease predisposition, with mostly negative results ([@pbio-0000020-Hansen4], [@pbio-0000020-Hansen2]; [@pbio-0000020-Kawanishi1]). One study did describe an association with disease status and with QTs underlying Type 2 diabetes ([@pbio-0000020-Baier1]). However, functional data for this polymorphism have suggested that the Ile326 variant may have only minor impact on signalling events ([@pbio-0000020-Baynes1]; [@pbio-0000020-Almind1]). Our data suggest that variation in this gene is a risk factor for the development of Type 2 diabetes, although further detailed studies will be required to elucidate the precise functional variants and mechanisms that lead to increased disease risk. The gene *SOS1* (*son of sevenless homolog 1* in Drosophila) encodes a guanine nucleotide exchange factor that functions in the transduction of signals that control cell growth and differentiation. We analysed two SNPs for association with disease status, a nonsynonymous SNP (N1011S) and an intronic variant (SNP8). While the nonsynonymous S1011 variant, which was very rare (minor allele, 0.003), did not associate with disease status, the intronic SNP was highly significantly associated with decreased disease risk (OR 0.58, *p* = 0.0032) (see [Table 2](#pbio-0000020-t002){ref-type="table"}), despite not showing any effects in the QT study. To our knowledge, this is the first investigation into the role of *SOS1* in Type 2 diabetes risk. The data presented here suggest that further investigation into the potential role of common variants at this gene and diabetes risk is warranted. Other Genes {#s2e} ----------- *PPARGC1* (LocusLink ID 10891) encodes a transcriptional coactivator of nuclear receptors with a critical role in regulating multiple aspects of energy metabolism, including adaptive thermogenesis ([@pbio-0000020-Puigserver1]), mitochondrial biogenesis ([@pbio-0000020-Wu1]), fatty acid β-oxidation ([@pbio-0000020-Vega1]), the control of hepatic gluconeogenesis ([@pbio-0000020-Herzig1]; [@pbio-0000020-Yoon1]), and the control of glucose uptake ([@pbio-0000020-Michael1]). *PPARGC1* SNP30 (Thr528Thr), which was associated with decreased disease risk (see [Table 2](#pbio-0000020-t002){ref-type="table"}), was rather surprisingly associated with decreased insulin secretion in the QT study (see [Table S2](#st002){ref-type="supplementary-material"}). In this locus, Thr528Thr has not been previously associated with diabetes, and our data most likely reflect stochastic variation at this site. The Gly482Ser has in some studies been shown to be associated with increased Type 2 diabetes risk ([@pbio-0000020-Ek1]), but not in others ([@pbio-0000020-Hara1]; [@pbio-0000020-Lacquemant1]; [@pbio-0000020-Muller1]), and has additionally been associated with insulin resistance ([@pbio-0000020-Hara1]), obesity indices in women ([@pbio-0000020-Esterbauer1]), and mean insulin secretory response and lipid oxidation ([@pbio-0000020-Muller1]). In our study, this allele was not associated with increased diabetes risk, but rather was associated with a lower risk of diabetes under a recessive model (OR 0.67, *p* = 0.0295) (see [Table 2](#pbio-0000020-t002){ref-type="table"}). The opposing results for this polymorphism and the fact that the amino acid change Gly482Ser is unlikely to be a major functional change ([@pbio-0000020-Esterbauer1]) may indicate that the contributing functional polymorphism may be an unidentified variant in LD with the Gly482Ser. Amongst the remaining genes tested, of particular interest are the results observed in *PYY* (encoding polypeptide YY; LocusLink ID 5697). An intronic variant, IVS3+68, showed a significant association with increased Type 2 diabetes risk under two genetic models (OR 1.47, *p* = 0.0240 in the allele 2 dominant; OR 1.47, *p* = 0.0157 in the additive effect allele 2) (see [Table 2](#pbio-0000020-t002){ref-type="table"}), but no evidence of association with underlying traits was observed in the QT study. Early functional studies suggested an inhibitory role of *PYY* in glucose-stimulated insulin secretion ([@pbio-0000020-Bertrand1]; [@pbio-0000020-Nieuwenhuizen1]), which led us to evaluate the potential role of variants at this gene in Type 2 diabetes predisposition. Recent data have shown that the peptide PYY~3--36~ encoded by this gene inhibits food intake and reduces weight gain when injected in rats, while physiological infusions of PYY~3--36~in humans decreased food intake by 33% ([@pbio-0000020-Batterham1]). Although our data do not show an association between the intronic variant SNP122 with BMI, they suggest a putative role for *PYY* in Type 2 diabetes predisposition. As we only tested a noncoding variant in *PYY*, it is possible that the association is due to other contributing variants within the gene and that a link between those and BMI is still plausible. In the genes *ABCC9* (LocusLink ID 10060) and *LIPC* (LocusLink ID 3990), single SNPs of modest significance were associated with disease status in the case control and therefore are not discussed further here. Examination of Other Previously Reported Associations {#s2f} ----------------------------------------------------- We were unable to confirm some associations observed in other studies. The *PPARG* Pro12Ala Pro allele has previously been shown to confer susceptibility to Type 2 diabetes, with the Ala allele providing a decreased risk ([@pbio-0000020-Altshuler1]). Our results for this polymorphism show the same direction and magnitude of effect for Ala/Ala versus Ala/Pro and Pro/Pro genotypes (OR 0.54; derived from data in [Table 2](#pbio-0000020-t002){ref-type="table"}), but the association was not statistically significant (*p* = 0.2269). The lower limit of the 95% CI for the protective effect of the Ala allele (OR 1.08, 95% CI \[0.82,1.42\], *p* = 0.583) is still consistent with the results of the meta-analysis by [@pbio-0000020-Altshuler1]). Our study was not sufficiently powered to detect the small effects expected for the predisposing Pro allele. The study only had 25.4% power to detect an OR of 1.25 for the Pro allele that occurred in 89.4% of our control population. It is also possible that the Pro12Ala variant does not affect diabetes susceptibility in this population, because of the dependence of the allele effect on environmental factors such as dietary fat composition ([@pbio-0000020-Luan1]). In the *ENPP1* (LocusLink ID 5167) gene (commonly known as *PC-1*) the K121Q polymorphism has variably been found to be both associated with increased Type 2 diabetes risk ([@pbio-0000020-Gu1]; [@pbio-0000020-Rasmussen1]; [@pbio-0000020-Hegele2]) and with insulin resistance QTs ([@pbio-0000020-Pizzuti1]; [@pbio-0000020-Gu1]; [@pbio-0000020-Rasmussen1]). In our study we did not find evidence for association between the K121Q polymorphism and Type 2 diabetes (OR 1.10, *p* = 0.5277 for the dominant model for allele 2; OR 1.07, *p* = 0.6290 for the additive effect for allele 2; OR 0.86, *p* = *0.7*496 for a recessive model for allele 2; data derived from [Table 2](#pbio-0000020-t002){ref-type="table"}). Analysis of this allele in our QT study showed that QQ individuals have higher mean BMI levels compared to carriers of the K121 allele (28.3 kg/m^2^ \[26.4, 30.2\] versus 26.1 \[25.6, 26.7\] in KQ subjects, 26.3 \[25.9, 26.6\] in KK subjects; data not shown). *PPP1R3A* (protein phosphatase 1 regulatory subunit 3; LocusLink ID 5506), which encodes the muscle-specific regulatory subunit of PP1, has been investigated as a potential diabetogene. Evidence for a role of the *PPP1R3A* D905Y polymorphism in Type 2 diabetes risk has also been conflicting ([@pbio-0000020-Hansen1], 2000; [@pbio-0000020-Hegele1]; [@pbio-0000020-Xia1]). While in this study we did not find an association between the D905Y variant and disease risk, we have previously described an association between a rare frameshift and premature stop variant with Type 2 diabetes risk under a dominant model (OR 5.03, *p* = 0.0110) in this population ([@pbio-0000020-Savage1]). Concluding Remarks {#s3} ================== This study, which to our knowledge is the largest of its kind yet reported in Type 2 diabetes, has provided evidence for the existence of variants in certain key candidate genes that influence the risk of Type 2 diabetes and, in some cases, has afforded clues as to the pathophysiological mechanism whereby those effects on disease risk might be mediated. By its very nature, any study of candidate genes, however large, is restricted in scope, and it is likely that other variants (namely in regulatory regions, which we did not cover) in the genes that we have considered, as well as ones that we have not, may exist and have effects equal to or greater than those we have demonstrated. In addition, this study is not intended to be an 'exclusion study,\' as many issues that relate to coverage of any given gene, environmental risk factors, and power in our populations do not allow us to definitively assert that negative findings correspond to genes that truly do not play a role in Type 2 diabetes predisposition. The power of our study to detect small effects in uncommon variants was low. Evidence from many recent studies now suggests that in Type 2 diabetes the effects are likely in the range of OR 1.15--1.5. It is clear that much larger studies than that reported here are required for such effects ([Figure 4](#pbio-0000020-g004){ref-type="fig"}), in particular when adjusting to a lower Type 1 error rate of 0.01% to compensate for multiple testing. The significance of the associations we report have been described without adjustment for the number of tests undertaken, and thus the group of positive associations is likely to contain a proportion that is falsely positive. There is no consensus about the ideal method for adjusting the probability of an observation occurring by chance for multiple testing. The simple Bonferroni correction would constitute overadjustment because the 152 genetic markers in this study are not independent. In addition, in the false-discovery rate method ([@pbio-0000020-Benjamini1]), it is assumed that all N tests are carried out simultaneously, which may not correspond to reality if groups genotype one set of SNPs, as in this study, but then report results for additional SNPs at a later date. It is not clear whether the number of tests N should reflect the number to date or the number one might potentially undertake by continuing working through projects like these. An alternative Bayesian approach leading to a 'genome-wide' significance level for association, such as has been done for whole-genome linkage studies ([@pbio-0000020-Lander1]), might be preferable. However, this also runs into difficulties. In studies that are not based on fine-mapping of linkage intervals, but rather on candidate genes selected on the basis of data from other studies, including previous reports of association, it is unclear what level of prior probability of association should be used. As a result of this uncertainty about the appropriate method of correction for multiple testing, our preferred strategy is to report the number of tests done and to encourage readers to interpret the significance tests in that light, acknowledging that the results will require replication in other cohorts. ::: {#pbio-0000020-g004 .fig} Figure 4 ::: {.caption} ###### Size of Case-Control Study Required to Detect Small Risk Effects The number is shown of the case chromosomes (assuming an equal number of control chromosomes) required to attain 80% power to detect associations with the OR varying between 1.0 and 1.5 and with a Type 1 error rate of 0.01%. Abbreviations: p0, frequency of the predisposing allele; chr, number of chromosomes. Graphs were plotted with the PS power and sample-size program ([@pbio-0000020-DuPont1]). :::  ::: Although we have not undertaken a formal replication of the case-control study, we have used a complimentary QT population to examine the association of the variants studied with continuously distributed measures of glucose tolerance, insulin secretion, and insulin action. This provides different information to a replication case-control study, as it may identify pathophysiological mechanisms by which the association with diabetes arises. We are cautious about putting forward particular variants as established 'diabetogenes\' and enthusiastically invite researchers to examine these candidate variants in their own particular populations. Indeed, as the genetic architecture of Type 2 diabetes may vary between populations, it is critical that such variants are examined in multiple diverse ethnic groups. As with the Pro12Ala *PPARG* example, it is likely that meta-analysis of several studies will be required to narrow the CIs around the point estimates of association seen in any single study. This will be especially important when the association is weak, as it is for Pro12Ala, because few individual case-control studies, including the one reported here, are currently powered to detect very small increases in risk. It is, however, important that such meta-analyses include all studies of variants examined, rather than only those that are individually published, to avoid publication bias. The associations we describe are highly biologically plausible and in many of the genes are supported by associations with multiple SNPs at the same locus. These include genes affecting both insulin secretion and insulin action. Given the importance of both insulin resistance and defective β-cell function to the pathogenesis of Type 2 diabetes, it is intriguing that we have found a disproportionate representation of genes affecting pancreatic β-cell function among those that were found to be associated with diabetes risk. This contrasts with the impact of known environmental factors and their correlates (e.g., high-fat diet, lower physical activity, obesity, and central fat distribution), all of which are thought to have their major influence on diabetes risk through impairment of insulin action. While it would be premature to put forward any definitive model for the causation of Type 2 diabetes, it is tempting to speculate that the 'insulin resistance\' component of the disease may have a substantial environmental influence modulated by polygenic effects, some of which may relate to molecules identified in this and other studies. On the other hand, the ability of the pancreatic β-cell to continue to secrete sufficient insulin to maintain life-long normoglycaemia may be more profoundly influenced by genetic factors, some of which are reported herein. It will be critical to examine the functional consequences of such variants, a task that will be particularly challenging when it comes to genes influencing human β-cell function, as it is entirely possible that this disproportionate representation of β-cell genes may be a reflection of our success in choosing diabetes genes in each of the candidate genes in the major groupings. The success of the approach presented here, although necessarily limited in scope, suggests that thesystematic examination of panels of biological candidate genes in large, well-characterised populations may usefully complement positional approaches to the identification of allelic variants conferring susceptibility to complex polygenic disease. The detection of multiple small gene effects accentuates the need for larger populations in order to reliably identify the types of effects (OR 1.15--1.5) we now expect for complex diseases. Materials and Methods {#s4} ===================== {#s4a} ### Methods for SNP discovery and SNP selection for genotyping. {#s4a1} SNP discovery was performed by a high-fSSCP-based analysis, as previously described ([@pbio-0000020-Thorpe1]). Genomic structure was determined for all genes, and primers were designed to span the exons and splice junctions. To detect common variants, genes were screened against one or more of a variety of different DNA panels, which included a 47-member multiethnic human diversity panel (comprised of 17 Europid, seven Hispanic, 13 East Asian, and ten African-American subjects), our 129-member severe insulin-resistant cohort ([@pbio-0000020-Barroso1]), a panel of 47 European-American samples, a panel of 47 African-American samples, and in some cases a panel of 94 samples (half European and half Asian Indian). Some genes had only partially screened coding sequence and splice junctions at the time of SNP selection for genotyping. In addition, we had access to an internal database of in silico SNPs that had been validated against 100 samples. Choice of polymorphisms for further testing in association studies was not constrained by the type of variant (e.g., nonsynonymous, silent, noncoding), although higher priority was given to variants with a likely effect on protein structure and function. Polymorphisms with a minor allele frequency greater than or equal to 5% were selected for further testing in population-based studies. In some instances, polymorphisms of lower allele frequency were genotyped to examine whether lower frequency variants with high penetrance might account for some cases of polygenic disease. Polymorphisms of lower frequency were also genotyped when, at the time of selection for genotyping, no other variants of known frequency were identified in the gene to test. ### Populations for SNP genotyping. {#s4a2} The Cambridgeshire Case-Control Population ([@pbio-0000020-Poulton1]; [@pbio-0000020-Halsall1]) consists of a collection of 517 Type 2 diabetics and 517 matched controls. The cases were a random sample of Europid men and women with Type 2 diabetes, aged 47--75 years, from a population-based diabetes register in a geographically defined region in Cambridgeshire, United Kingdom. The presence of Type 2 diabetes in these subjects was defined as onset of diabetes after the age of 30 years without use of insulin therapy in the first year after diagnosis. The control subjects were individually age-, gender-, and geographical location-matched to each of the cases. Controls were not matched by BMI to cases. Potential controls that had HbA1c levels greater than 6% were excluded, as this group may contain a higher proportion of individuals with previously undiagnosed diabetics*.* HbA1c was assayed using high performance liquid chromatography on a BioRad Diamat 33 (Hercules, California, United States), according to the method of [@pbio-0000020-Standing1]). The coefficient of variation (CV) was 3.6% at the lower end of the range (mean = 4.94%) and 3.0% at the upper end (mean = 9.76%). Further details on the characteristics of the subjects are shown in [Table 5](#pbio-0000020-t005){ref-type="table"}. ::: {#pbio-0000020-t005 .table-wrap} Table 5 ::: {.caption} ###### Study Subjects in the Cambridgeshire Case-Control Study :::  Data are means and standard error is in parentheses ::: The QT study population is a collection of 1,100 samples collected for the Ely Study, a prospective population-based cohort study of the aetiology and pathogenesis of Type 2 diabetes and related metabolic disorders ([@pbio-0000020-Wareham1]). Height was measured using rigid stadiometer, and weight was measured on Seca-calibrated scales with participants in light clothing. BMI was estimated as weight (kg) divided by height (m) squared. Plasma glucose was measured in the routine National Health Service Laboratory at Addenbrooke\'s Hospital, using the hexokinase method ([@pbio-0000020-Kunst1]). Plasma insulin was measured by two-site immunometric assays with either ^125^I or alkaline phosphatase labels ([@pbio-0000020-Sobey1]; [@pbio-0000020-Alpha1]). Cross-reactivity with intact proinsulin was less than 0.2% and CVs were less than 7%. ### Methods for genotyping. {#s4a3} Genotyping was performed using an adaptation of the fluorescence polarisation template-directed incorporation (FP-TDI) method described by Chen et al.(1999). In short, PEP-amplified DNA samples were PCR-amplified in 8 μl reactions with primers flanking the variant site; unincorporated dNTP and remaining unused primer were degraded by exonuclease I and shrimp alkaline phosphatase at 37°C for 45 min before the enzymes were heat-inactivated at 95°C for 15 min. At the end of the reaction, the samples were held at 4°C. Single base primer extension reactions were performed as previously described ([@pbio-0000020-Chen1]), and allele detection was performed by measuring fluorescence polarisation on an LJL Analyst fluorescent reader (Molecular Devices, Sunnyvale, California, United States). The PEP protocol was specifically developed and tested to ensure that allele bias was not introduced during the amplification process. A minimum of 12% internal replicate samples within each population (case control and QT) were included in all genotyping tests to assess genotyping accuracy. Only assays that provided 100% concordance between replicates were analysed for association. The genotyping pass rate was greater than 90% once a working assay had been established. There was an 85% success rate for an SNP to be converted into a working assay at the first attempt, with a number of failed assays recovered by designing an assay to the reverse strand. ### Statistical analysis. {#s4a4} All analyses used SAS 8.02 (SAS Institute, Cary, North Carolina, United States) or Stata 7.0 (Stata Corporation, College Station, Texas, United States) statistical programs, unless otherwise stated. Agreement with Hardy--Weinberg equilibrium was tested using a χ^2^ 'goodness-of-fit\' test. The disequilibrium coefficient for the controls (*r^2^*) was calculated ([@pbio-0000020-Lewontin1]). For the case-control study, tests for association with disease status under dominant, additive, and recessive models were undertaken using univariate logistic regression analysis. Dominance was defined in terms of allele 2 effects; in the dominant allele 2 model, homozygous subjects for allele 1 were compared with carriers of allele 2; in the recessive allele 2 model, carriers of allele 1 were compared with homozygous subjects for allele 2. In some cases, a large number of polymorphisms within a gene were typed. To reduce complexity, a subset of markers within a gene associated with diabetes status was identified using backward logistic regression. Any polymorphism that had a *p* value greater than 0.1 was removed from the model. The genotypes were assumed as having additive effects. *p* values for the additive effect are for the test for a linear trend across the genotypes, which were coded as 0 = 11, 1 = 12, 2 = 22. Where the subset consisted of more than one polymorphism within a gene, haplotype analysis was performed. To avoid possible errors due to either genotyping or the estimation process, only haplotypes that had a frequency greater than 5% were considered for further analysis. Haplotype frequencies were estimated using maximum-likelihood methods. A log-linear model embedded with the expectation-maximization algorithm was fitted to a frequency table ([@pbio-0000020-Chiano1]; [@pbio-0000020-Mander1]). Differences in haplotype distributions between the diabetic and nondiabetic groups were examined using a likelihood-ratio statistic ([@pbio-0000020-Mander1]). To obtain separate ORs for each haplotype, the most common haplotype was used as the reference category. CIs were obtained using a profile-likelihood approach ([@pbio-0000020-Mander1]). For the QT study, the distributions of fasting plasma glucose, 2-h plasma glucose, fasting plasma insulin, and insulin increment were skewed and were thus normalised by logarithmic transformation. Baseline and follow-up measurements of BMI, fasting and 2-h plasma glucose, fasting plasma insulin, and 30-min insulin increment during an oral glucose tolerance test were collected. Where two measures were available, the mean was used. Otherwise, a single measure (either baseline or follow-up) was used for further analysis. The subset of SNPs identified in the case-control study was used. In separate dominant, additive, and recessive models, adjusting for age and sex, genotype differences in these measurements were modelled using the General Linear Model procedure in the statistical package SAS. For each individual, a list of possible haplotypes and their probabilities was obtained using Snphap software (<http://www-gene.cimr.cam.ac.uk/clayton/software/>). Haplotypes with a frequency greater than 5% were the same as those reconstructed in the case-control study. Only haplotypes that had a frequency greater than 5% and individuals that had at least one marker typed were considered for analysis. As currently haplotype analysis software cannot handle repeated measurements, the average of two measurements was used for further analysis. Associations of haplotypes (adjusted for age and sex) with the QTs were determined by cluster-weighted regression analysis, thereby taking into account nonindependent multiple observations from an individual ([@pbio-0000020-Huber1]; [@pbio-0000020-White1], [@pbio-0000020-White2]). QT means and their 95% CI were estimated for each haplotype. Supporting Information {#s5} ====================== Table S1 ::: {.caption} ###### Genotype Counts and Frequencies for All SNPs Genotyped in This Study (79 KB XLS). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S2 ::: {.caption} ###### Single SNP Associations with QTs (174 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Accession Numbers {#s5a} ----------------- The LocusLink accession numbers discussed in this paper are 3170, 3172, 3630, 3643, 3767, 3990, 5167, 5295, 5468, 5506, 5697, 6514, 6654, 6833, 6927, 6928, 10060, 10891, and 11132. NJW is a Wellcome Trust Senior Clinical Fellow. The Ely Study is supported by the Medical Research Council. SO is supported by the Wellcome Trust and the Medical Research Council. DC is supported by the Wellcome Trust and the Juvenile Diabetes Research Foundation. This study drew upon the combined efforts of many individuals at Incyte Genomics Cambridge, to whom we extend our grateful appreciation. We also thank Suzannah Bumpstead, Bill Bottomley, and Stephan Collins for technical assistance. We are grateful to anonymous reviewers for helpful comments and suggestions. **Conflicts of Interest.** Portions of the research were supported by Incyte Corporation (formerly Incyte Genomics). This collaboration provided access to certain technologies and scientific expertise for basic research in the genetics of diabetes in return for possible shared intellectual property. The work described here is not subject to any patent filings or intellectual property protection and restrictions from free use of the information. IB and AJS were employees of Incyte at the time the research took place. **Author Contributions.** IB, AJS, SO, and NW conceived and designed the experiments. IB, AJS, and NW performed the experiments or the fieldwork in the studies contributing data. IB, JL, RPSM, A-HH, PWF, RWJ, and NW analysed the data. DC contributed analysis tools and advice. All authors wrote the paper. Academic Editor: Philippe Froguel, Centre National de la Recherche Scientifique--Institut de Biologie de Lille. ¤ Present Address: The Wellcome Trust Sanger Institute, Cambridge, United Kingdom. ABC : ATP-binding cassette BMI : body mass index CI : confidence interval CV : coefficient of variation fSSCP : fluorescent single-stranded conformation polymorphism HbA1c : glycated haemoglobin INS0 : fasting insulin IVS : intervening sequence LD : linkage disequilibrium MODY : maturity-onset diabetes of the young OR : odds ratio PG0 : fasting plasma glucose levels PHHI : persistent hyperinsulinaemia hypoglycaemia of infancy PPARG : peroxisome proliferator-activated receptor γ PP1 : protein phosphatase 1 QT : quantitative trait SNP : single nucleotide polymorphism VNTR : variable number tandem repeat.

PubMed Central

2024-06-05T03:55:45.972275

2003-10-13

PMC212699

Introduction {#s1} ============ Endosymbiosis is common in insects, with more than 10% of insect species relying upon intracellular bacteria for their development and survival ([@pbio-0000021-Baumann1]). Full genome sequencing of the endosymbiotic bacteria, Buchnera aphidicola, of several species of aphids has revealed extensive gene loss ([@pbio-0000021-Shigenobu1]; [@pbio-0000021-Tamas1]; [@pbio-0000021-van1]), but has failed to reveal the genetic basis for the interaction between the bacteria and host cells. The key adaptations that allow incorporation of the bacteria into host cells may therefore be encoded by the host genome. The symbiotic bacteria of aphids, B. aphidicola, live within large polyploid cells, called bacteriocytes, that are grouped into organ-like structures, called bacteriomes, located adjacent to the ovarioles. During most of the aphid lifecycle, embryos develop parthenogenetically from unfertilized diploid oocytes, and multiple embryos develop serially within a single ovariole ([@pbio-0000021-Dixon1]) ([Figure 1](#pbio-0000021-g001){ref-type="fig"}A). Maternal bacteria are transferred directly to the developing blastoderm-stage embryos through an opening in the posterior of the embryo ([@pbio-0000021-Buchner1]; [@pbio-0000021-Miura1]) ([Figure 1](#pbio-0000021-g001){ref-type="fig"}B). Several researchers have described this transovarial transfer of bacteria (e.g., [@pbio-0000021-Uichanco1]; [@pbio-0000021-Klevenhusen1]; [@pbio-0000021-Toth1], [@pbio-0000021-Toth2]; [@pbio-0000021-Lampel1]; [@pbio-0000021-Buchner1]), but the details of bacteriocyte development have remained unclear. ::: {#pbio-0000021-g001 .fig} Figure 1 ::: {.caption} ###### Expression of Three Transcription Factors during Early Bacteriocyte Development \(A) Drawings of some stages of pea aphid embryonic development, approximately to scale. Embryos develop viviparously within a follicular epithelium of the ovariole (data not shown). For a complete description, see [@pbio-0000021-Miura1]). Bacteria are transferred at stage 7. Embryos are labeled with bacteria (b), head (h), thoracic (t), and abdominal (a) regions. The three thoracic segments (t1, t2, t2) and germ cells (gc) are indicated in the stage 14 embryo. \(B) A drawing of a stage 7 embryo illustrates transovarial transfer of the bacteria (red arrowhead) to the embryo and the presumptive bacteriocyte nuclei (arrow). \(C) Confocal micrograph of a stage 6 embryo stained with anti-Dll antibody (red, indicated by arrow). Anti-Dll labels syncytial nuclei (presumptive bacteriocyte nuclei) in the posterior of the embryo. \(D) Confocal micrograph of stage 7 embryo stained with anti-Dll and FP6.87 antibodies. Soon after the bacteria begin to invade the embryo, we observe staining with the FP6.87 antibody localized to the nucleoli (blue), which recognizes both Ubx and Abd-A in diverse arthropods, in the same nuclei that are already expressing Dll (red). The region outlined with a broken white box is enlarged in (D′) to show the bacteria, and only the green channel is shown in monochrome. The red arrow indicates one bacterium. (E and F) In these two panels of the same focal plane from the same stage 9 embryo, Ubx/Abd-A staining (blue) is observed throughout the entire nucleus of all nuclei that also express Dll (red). \(G) Confocal micrograph of a stage 8 embryo stained with anti-En (yellow). As the transfer of bacteria (arrowhead) is being completed, the bacteriocyte nuclei begin to express En (yellow, indicated with arrow). In (C)--(G), confocal micrographs show only one focal plane of the embryo, so not all bacteriocyte nuclei in each embryo can be seen. In all figures, F-actin is stained with phalloidin (green). Embryos in all figures, except [Figure 2](#pbio-0000021-g002){ref-type="fig"}, are oriented with anterior of the entire embryo (towards the germarium) to the left. :::  ::: We have identified bacteriocyte-specific markers that allow us to track the proliferation of bacteriocytes throughout the development of the pea aphid Acyrthosiphon pisum (Harris) (Hemiptera: Aphididae). Using these markers, we aimed to determine the developmental origin of bacteriocytes and to what extent bacteria are required for the formation of the bacteriocytes. We also tested whether the observed patterns of bacteriocyte development are evolutionarily conserved among distantly related aphid species. We show that three transcription factors are expressed in a specific temporal order during early bacteriocyte development of the pea aphid. The final population of bacteriocytes originates from two distinct populations of nuclei recruited at different times of development. Furthermore, we experimentally demonstrate that the specification and proliferation of bacteriocytes occur independently of B. aphidicola. In distant relatives of the pea aphid, we found that the two-step determination of bacteriocytes is conserved. We also investigated two cases involving the loss of B. aphidicola. In the first case, in which the bacterial symbionts have been replaced with extracellular, eukaryotic symbionts, bacteriocyte development appears to proceed normally. In a second case, in which males do not inherit B. aphidicola, the bacteriocytes have been lost. Results {#s2} ======= Three Transcription Factors Are Expressed in a Specific Temporal Order during Early Bacteriocyte Development {#s2a} ------------------------------------------------------------------------------------------------------------ We tested five cross-reacting antibodies (see [Materials and Methods](#s4){ref-type="sec"}) for their expression patterns in aphids. In every case, we observed antibody staining in the expected population of cells in the developing embryo (also see [@pbio-0000021-Miura1]). In addition, we found that three of the antibodies stained nuclei that form bacteriocytes of A. pisum. We infer that these antibodies are recognizing the homologues, or possibly paralogues, of their respective target proteins. The three proteins are expressed in a specific temporal order. We first observe expression of the Distal-less (Dll) protein (FlyBase ID: FBgn0000157) ([@pbio-0000021-Panganiban2]) in syncytial nuclei at the posterior of the blastoderm embryo just prior to the invasion of bacteria into the embryo ([Figure 1](#pbio-0000021-g001){ref-type="fig"}C). As the bacteria enter the embryo, these nuclei associate with the bacteria and start to express a second protein, Ultrabithorax (Ubx) (FBgn0003944) or Abdominal-A (Abd-A) (FBgn0000014) or both, detected by the FP6.87 antibody ([@pbio-0000021-Kelsh1]) ([Figure 1](#pbio-0000021-g001){ref-type="fig"}D--1F). The bacteria can be easily observed as spheres 2--4 μm in diameter ([@pbio-0000021-Buchner1]) that stain with phalloidin ([Figure 1](#pbio-0000021-g001){ref-type="fig"}D′). As the transfer of bacteria to the embryo is being completed, expression of the Engrailed (En) protein (FBgn0000577) ([@pbio-0000021-Patel1]) is detected ([Figure 1](#pbio-0000021-g001){ref-type="fig"}G). Two Populations of Cells Are Recruited to the Bacteriocyte Fate at Different Times in Development {#s2b} ------------------------------------------------------------------------------------------------- The early embryo contains approximately eight bacteriocyte nuclei that express Dll ([Figure 1](#pbio-0000021-g001){ref-type="fig"}C), whereas the adult aphid contains 60--90 uninucleate polyploid bacteriocytes ([@pbio-0000021-Baumann1]) that also express Dll (data not shown). We found that the increase in bacteriocyte number occurs through two mechanisms. First, we infer that the original bacteriocyte nuclei divide, apparently in a syncytium and perhaps synchronously, through two rounds of division because we observe that the number of Dll-expressing nuclei increases from approximately eight to 16 by stage 12 and then to approximately 32 by stage 13 (data not shown). By stage 14, these original bacteriocytes have formed cell membranes and become polyploid ([Figure 2](#pbio-0000021-g002){ref-type="fig"}A). At stage 13, a second population of approximately 40--60 cells located near the posterior end of the dorsal germband begins to express Dll ([Figure 2](#pbio-0000021-g002){ref-type="fig"}B)*.* The nuclei of these cells are visibly smaller than those of the original bacteriocytes ([Figure 2](#pbio-0000021-g002){ref-type="fig"}A--2E). Based on observations of multiple fixed specimens, we infer that these cells then migrate across the germband ([Figure 2](#pbio-0000021-g002){ref-type="fig"}E) and intercalate between the original bacteriocytes ([Figure 2](#pbio-0000021-g002){ref-type="fig"}C and 2D). The bacteria are presumably then subdivided among all of the Dll-expressing nuclei and the final bacteriocytes are formed. ::: {#pbio-0000021-g002 .fig} Figure 2 ::: {.caption} ###### The Second Wave of Bacteriocyte Determination In (A)--(D), the embryos, which are normally folded in upon themselves in a pretzel shape within the ovariole ([@pbio-0000021-Miura1]), have been dissected flat, stained with anti-Dll antibody (red) and phalloidin (green), and examined with a confocal microscope. \(A) Dll expression (red) in a stage 14 embryo is detected in the labrum (La) and all developing limbs on the ventral surface except the mandibular segment (Mn). (Other abbreviations: An, antenna; Mx, maxilla; Lb, labium; T1, T2, T3, first, second, and third thoracic leg, respectively.) The dorsal surface of the abdomen of the same embryo is shown illustrating Dll expression in the original bacteriocytes (white arrow) and in a more posterior population of nuclei or cells (blue arrow). Germ cells (gc) are labeled. \(B) Dll expression is first observed in the new bacteriocyte nuclei at stage 13. \(C) By stage 15, many of the new bacteriocytes have migrated to and begun intercalating between the original bacteriocytes. \(D) By stage 16, all of the new bacteriocytes have intercalated between the original bacteriocytes. \(E) The migration of the new bacteriocytes is seen in a confocal section of an undissected stage 14 embryo. Embryos in (A)--(D) are oriented with the anterior of the germband towards the left. :::  ::: Bacteriocytes Are Specified and Maintained When the Bacteria Have Been Experimentally Removed {#s2c} --------------------------------------------------------------------------------------------- The observations described in the first section suggest that the initial specification of the bacteriocyte may occur independently of B. aphidicola. We tested this idea by eliminating B. aphidicola from pea aphids by feeding aphids on an artificial diet containing antibiotics. We found that the embryos within these aposymbiotic aphids specify the bacteriocyte cell fate, as revealed by Dll expression, and maintain the bacteriocyte cell fate in the absence of bacteria ([Figure 3](#pbio-0000021-g003){ref-type="fig"}). In addition, we have observed that the number of bacteriocytes in aposymbiotic embryos increases precisely as described for symbiotic embryos, including the second wave of bacteriocytes ([Figure 3](#pbio-0000021-g003){ref-type="fig"}F; data not shown). ::: {#pbio-0000021-g003 .fig} Figure 3 ::: {.caption} ###### Elimination of B. aphidicola by Treatment with Antibiotics Has No Effect on the Determination and Maintenance of the Bacteriocyte Cell Fate in A. pisum (A--C) Confocal micrographs of control embryos stained with anti-Dll antibody (red) show expression of Dll, as described in [Figure 1](#pbio-0000021-g001){ref-type="fig"}. Enlarged views of the bacteria within the broken white boxes in each embryo are shown in (A′)--(C′). (D--F) Embryos within aposymbiotic aphids at comparable stages as the controls in (A)--(C) express Dll in bacteriocyte nuclei. No bacteria are observed within these embryos, as seen in the enlarged views of (D′)--(F′). :::  ::: The Two-Step Determination of Bacteriocytes Is Evolutionarily Conserved {#s2d} ----------------------------------------------------------------------- The two-step determination of bacteriocytes described in the previous sections appears to be a conserved feature of the aphids. Using the anti-Dll antibody, we examined development of the bacteriocytes in two species of aphids that diverged from A. pisum (subfamily Aphidinae) approximately 80--150 million years ago ([@pbio-0000021-von1]): Pemphigus spyrothecae (Eriosomatinae) and Tuberaphis styraci (Hormaphidinae) (discussed below). In both cases, Dll is expressed in a small number of bacteriocyte nuclei of the blastoderm-stage embryo and additional Dll-expressing cells are recruited later. In *P. spyrothecae,* one or two nuclei are originally determined as bacteriocytes, as suggested by [@pbio-0000021-Lampel1]) ([Figure 4](#pbio-0000021-g004){ref-type="fig"}A). These nuclei become highly polyploid prior to bacterial invasion and do not divide ([Figure 4](#pbio-0000021-g004){ref-type="fig"}B and 4C). A second population of bacteriocytes is determined at approximately stage 14 ([Figure 4](#pbio-0000021-g004){ref-type="fig"}D). These surround the original bacteriocyte ([Figure 4](#pbio-0000021-g004){ref-type="fig"}E) and appear to divide the bacteria into independent bacteriocytes. ::: {#pbio-0000021-g004 .fig} Figure 4 ::: {.caption} ###### Expression of Dll in Bacteriocytes and the Pattern of Bacteriocyte Development Are Conserved in Parthenogenetic Females of P. spyrothecae Confocal micrographs of P. spyrothecae parthenogenetic embryos stained with anti-Dll antibody (red). \(A) Dll is first detected in stage 6 embryos in one or two nuclei posterior to the cellular blastoderm (arrow). \(B) By stage 8, the bacteria have been transferred to and entirely fill the embryo (red arrowhead). The Dll-expressing nuclei (arrow) have become highly polyploid. (C and D) At stage 12, only the original bacteriocyte nuclei are observed expressing Dll (white arrow), but by stage 14 (D) additional nuclei (blue arrow) closely apposed to the dorsal germband express Dll. \(E) By stage 15, these new nuclei surround the original bacteriocyte, and at later stages the bacteria are divided into individual cells. :::  ::: Bacteriocytes Develop in Aphids in Which the Bacteria Have Been Replaced with Extracellular Eukaryotic Symbionts {#s2e} ---------------------------------------------------------------------------------------------------------------- B. aphidicola has been lost in the lineage leading to T. styraci and has been replaced by a yeast-like symbiont ([@pbio-0000021-Buchner1]; [@pbio-0000021-Fukatsu1]; [@pbio-0000021-Fukatsu3]). These symbionts live in the hemolymph and occasionally invade cells of the fat body ([@pbio-0000021-Buchner1]). Previous studies have therefore claimed that these species lack bacteriocytes ([@pbio-0000021-Buchner1]; [@pbio-0000021-Fukatsu1]). We found that these aphids contain one or two nuclei in the posterior of the blastoderm embryo that express Dll ([Figure 5](#pbio-0000021-g005){ref-type="fig"}A). These nuclei divide once or twice and then become polyploid. At approximately stage 14, we observed a second population of Dll-expressing cells that migrate to the original Dll-expressing cells ([Figure 5](#pbio-0000021-g005){ref-type="fig"}B). Therefore, T. styraci appears to retain the bacteriocyte cell fate although these cells do not apparently house the novel symbionts. ::: {#pbio-0000021-g005 .fig} Figure 5 ::: {.caption} ###### Bacteriocytes Are Retained in One Species That Has Evolutionarily Lost Bacteria, but Not in Males of Another Species That Do Not Inherit Bacteria (A and B) Confocal micrographs of embryos of T. styraci stained with anti-Dll antibody (red). In T. styraci, in which B. aphidicola has been evolutionarily lost ([@pbio-0000021-Fukatsu1]), embryos still contain nuclei that express Dll in the correct time and place to be bacteriocyte nuclei. (A) Dll expression is first detected in posterior nuclei at blastoderm at approximately stage 6 (arrow). (B) By stage 14, the original nuclei have divided once or twice and become polyploid (original bacteriocytes), and new cells begin to express Dll (new bacteriocytes; blue arrow) and migrate towards the original bacteriocytes. (C--F) Confocal micrographs of embryos of P. spyrothecae stained with anti-Dll antibody (red). (C) Stage 16 male embryos of P. spyrothecae do not contain *B. aphidicola,* and no Dll-expressing cells are observed in the expected location for bacteriocytes. We believe that the cells in this location are sperm (marked with an asterisk). Sexual female embryos within the same ovary do contain Dll-expressing original and new bacteriocyte nuclei (white and blue arrows, respectively). (D and E) Transient expression of Dll in putative bacteriocytes is observed in stage 7 male embryos (arrow in male embryo of \[D\]), but this expression does not persist into stage 10 male embryos (E), where no Dll-expressing nuclei are observed. By contrast, stage 6 female embryos (D) contain polyploid Dll-expressing nuclei (arrow in female embryo of \[D\]). The sex of each embryo could be determined because males develop synchronously and earlier than females ([@pbio-0000021-Lampel1], 1968). (F) In stage 14 male embryos, we observe transient Dll expression in nuclei (blue arrow) adjacent to the germ cells (gc) in the correct location to be the second wave of bacteriocyte nuclei. This Dll expression does not persist (see stage 16 male in \[C\]), and the fate of the cells is unknown. :::  ::: The Bacteriocyte Fate Has Been Lost in Male Eriosomatine Aphids That Do Not Harbor B. aphidicola {#s2f} ------------------------------------------------------------------------------------------------ Males of some species in the subfamily Eriosomatinae do not harbor B. aphidicola ([@pbio-0000021-Toth1]; [@pbio-0000021-Buchner1]; [@pbio-0000021-Fukatsu2]). As these males lack mouthparts and do not feed, B. aphidicola are not required for growth. In addition, inheritance of B. aphidicola is strictly maternal, so males do not require symbionts for passage to their offspring. We did not detect any putative bacteriocyte cells that persist in male embryos of P. spyrothecae, although we observed them in female sexual embryos ([Figure 5](#pbio-0000021-g005){ref-type="fig"}C and 5D). In stage 7 male embryos, we detected weak Dll expression in a few nuclei ([Figure 5](#pbio-0000021-g005){ref-type="fig"}D), although this expression does not persist ([Figure 5](#pbio-0000021-g005){ref-type="fig"}E). In addition, in stage 14 males we detected weak expression in cells that are in the correct location to be the second population of bacteriocytes ([Figure 5](#pbio-0000021-g005){ref-type="fig"}F), but this expression also does not persist (see the stage 16 male in [Figure 5](#pbio-0000021-g005){ref-type="fig"}C). Discussion {#s3} ========== The aphid bacteriocyte expresses three transcription factors: Dll, En, and Ubx or Abd-A. These transcription factors play important roles during later stages of development in insects. For example, Dll is required for limb development, En is required for segmentation, and Ubx and Abd-A are the products of *Hox* genes, required for patterning thoracic and abdominal body regions ([@pbio-0000021-Kuner1]; [@pbio-0000021-Hidalgo1]; [@pbio-0000021-Weatherbee1]; [@pbio-0000021-Panganiban1]). We know of no other cases in other insects in which any of these three transcription factors are expressed at such early stages of development as we have observed in the bacteriocytes (approximately cellular blastoderm). We cannot exclude the possibility that bacteriocytes evolved from a cell type that expressed this combination of transcription factors, but there are no obvious candidate cell types, such as fat cells or vitellophages, in other insects that fulfill this criterion. We do not yet know whether these genes are involved in the determination of bacteriocytes. However, bacteriocytes may require a novel combination of transcription factors to regulate the symbiont population and to mediate transovarial transmission. We have demonstrated that two cell populations express Dll in spatially and temporally distinct patterns before incorporating bacteria. Our observation of the initial putative bacteriocytes in the blastoderm embryo is consistent with observations of earlier researchers, who suggested---based on morphological observations---that the nuclei located at the posterior of the embryo constitute the future bacteriocyte nuclei ([@pbio-0000021-Lampel1]; [@pbio-0000021-Buchner1]). In addition, we have found that the second population of presumptive bacteriocytes appears to migrate across the germband to the original bacteriocytes, where they take up bacteria. This is an unusual process that has not to our knowledge been described previously. In contrast, earlier studies indicated that bacteriocyte proliferation occurs solely by cell division or by budding of small nuclei from an existing polyploid bacteriocyte nucleus (e.g., [@pbio-0000021-Lampel1]). We have not yet performed experiments that would allow us to positively identify the embryonic origin of this second population of cells. Based on their position---posterior to the germ cells and dorsal---these cells may be the descendants of the nuclei of the central syncytium (syncytial nuclei in the center of the blastoderm embryo) (see [@pbio-0000021-Miura1]). Our results suggest that B. aphidicola is required for neither bacteriocyte induction nor for the origin and migration of the second population of bacteriocytes. While bacteria do not seem to be required for the developmental maintenance of this cell type, the bacteria may provide signals to the cells that are involved in mediating the symbiosis at the physiological level. Nonetheless, the absence of an effect of the bacteria on bacteriocyte development contrasts with other symbioses where the bacteria induce specific developmental changes in host tissues ([@pbio-0000021-McFall-Ngai1]). We investigated two cases in which B. aphidicola have been lost during the evolution of aphids. Given our observations that bacteria are not required for the developmental maintenance of bacteriocytes, it is possible that the bacteriocyte cell type might be lost if it had no other function. This does not appear to be the case. In the lineage including T. styraci, B. aphidicola was lost and a eukaryotic "yeast-like" symbiont has been gained ([@pbio-0000021-Buchner1]; [@pbio-0000021-Fukatsu1]; [@pbio-0000021-Fukatsu3]). [@pbio-0000021-Buchner1]) suggested that the bacteriocytes of Cerataphis freycinetiae, another species in the same lineage, are originally specified, become polyploid and then degenerate. We found Dll-expressing putative bacteriocyte nuclei to be specified and maintained over extensive periods of embryonic development in T. styraci. Buchner documented considerable variation in the details of symbiotic transmission and bacteriocyte development, and it is possible that bacteriocyte development proceeds along different paths in these two species. We also examined the development of bacteriocytes in males of P. spyrothecae. The males do not have bacteria and we have observed, consistent with observations of earlier researchers ([@pbio-0000021-Lampel1]; [@pbio-0000021-Buchner1]), that bacteriocytes are not maintained in this morph. We found that bacteriocytes initially express Dll, but this expression is not maintained, which is consistent with Lampel\'s and Buchner\'s observations that the original bacteriocytes appear to be present but are not maintained. In addition, we found that the second wave of bacteriocytes is also initiated, as shown by brief, weak Dll expression. It is not clear whether these cells are subsequently respecified or are eliminated. B. aphidicola are derived from free-living bacteria ([@pbio-0000021-Baumann1]), and both the bacteriocyte and the symbiont must have evolved mechanisms for integrating the bacteria into the workings of the cell. The aphid--*Buchnera* symbiosis represents a particularly intimate form of symbiosis. In some symbioses, the bacteria reside both intra- and intercellularly and actively invade the host cell ([@pbio-0000021-Dale1]). In contrast, B. aphidicola always exist either within host cells, within a membrane-bound maternal package, or with host nuclei in a syncytium. This advanced stage of symbiosis is similar to the presumptive early stages of plastid evolution. Materials and Methods {#s4} ===================== {#s4a} ### Aphid rearing and collecting. {#s4a1} Colonies of A. pisum were reared on broad bean (Vicia faba) or alfalfa (Medicago sativa) ([@pbio-0000021-Miura1]). P. spyrothecae were collected from galls on Populus nigra var*. italica* in Cambridge and London, United Kingdom. T. styraci were collected from galls on Styrax obassia in Gunma Prefecture, Japan. Asexual aphid embryos of various developmental stages were dissected and fixed as described previously ([@pbio-0000021-Miura1]). ### Antibody staining. {#s4a2} A limited number of antibodies recognize the homologues of their target proteins across insects. We tested five of these antibodies in aphids and found that three stained the bacteriocyte nuclei: rabbit anti-Dll ([@pbio-0000021-Panganiban2]), mouse anti-En (4D9) ([@pbio-0000021-Patel1]), and mouse anti-Ubx /Abd-A (FP.6.87) ([@pbio-0000021-Kelsh1]), kindly provided by G. Panganiban, N. Patel, and R. White, respectively. Two antibodies, rabbit anti-Vasa (FBgn0000606**)** (a gift of C.-C. Chang \[[@pbio-0000021-Chang1]\]) and mouse anti-Even-skipped (Eve) (2B8) (FBgn0003970) ([@pbio-0000021-Patel2]) did not stain bacteriocytes, but, as expected, anti-Vasa stained the germ cells ([@pbio-0000021-Chang1]) and anti-Eve stained cells in the nervous system ([@pbio-0000021-Patel1]). Secondary antibodies conjugated with fluorescent moieties (Jackson ImmunoResearch, West Grove, Pennsylvania, United States) were tested for cross-reactivity to aphid cells by staining embryos with secondary antibodies alone. No cross-reactivity was detected. We further tested whether an additional mouse antibody (mouse anti-digoxigenin; Jackson ImmunoResearch) cross-reacted with bacteriocyte nuclei, and it did not stain any parts of the aphid embryo. In addition, the anti-Dll, anti-En, and anti-Ubx/Abd-A all stained the expected cells ([@pbio-0000021-Patel1]; [@pbio-0000021-Kelsh1]; [@pbio-0000021-Panganiban2]; [@pbio-0000021-Miura1]) in other regions of the embryos, indicating that the antibodies were working as expected. Cell outlines were visualized by staining for F-actin with fluorescein-conjugated phalloidin. Embryos were stained using standard protocols ([@pbio-0000021-Miura1]) and visualized on Leica SP and Zeiss confocal microscopes. ### Antibiotic treatment. {#s4a3} In the pea aphid, *Buchnera* can be eliminated by treating animals with antibiotics ([@pbio-0000021-Wilkinson1]). First- or second-instar aphids were fed on an artificial diet containing 50 μg/ml of the antibiotic rifampicin for 72 h (e.g., [@pbio-0000021-Caillaud1]). Aphids were then transferred to leaves of Medicago arborea in Petri-dish cultures ([@pbio-0000021-Miura1]). Control aphids were treated identically, except that the antibiotic was omitted from the artificial diet. Embryos that were less than 4 d old ([@pbio-0000021-Miura1]) were dissected from aposymbiotic aphids within 2--4 d after the end of the antibiotic treatment and stained with anti-Dll and FP6.87 antibodies and fluorescein-conjugated phalloidin. The absence of bacteria in aposymbiotic aphids was confirmed by observation with a confocal microscope (see [Figure 3](#pbio-0000021-g003){ref-type="fig"}). Supporting Information {#s5} ====================== {#s5a} ### Accession Numbers {#s5a1} The FlyBase accession numbers discussed in this paper are FBgn0000014, FBgn0000157, FBgn0000577, FBgn0000606, FBgn0003944, and FBgn0003970. We thank J. Truman for preparing the specimens in [Figure 2](#pbio-0000021-g002){ref-type="fig"}; M. Caillaud for help generating the aposymbiotic aphids; Y. Rahbé for providing the artificial aphid diet; H. Shibao, B. Olynyk, and V. Olynyk for help collecting aphids; and N. Moran, H. Frydman, R. Blackman, C. Dale, and anonymous referees for comments on the paper. We acknowledge financial support for CB (Boehringer Ingelheim Fonds, the Roche Research Foundation, and the Janggen-Poehn-Stiftung), TM (Grants-in-Aid from the Ministry of Education, Culture, Sports, Science, and Technology of Japan and from the Research for the Future Program of the Japan Society for the Promotion of Science), SK (United States Department of Agriculture), and DLS (Biotechnology and Biological Sciences Research Council and National Institutes of Health). **Conflicts of Interest.** The authors have declared that no conflicts of interest exist. **Author Contributions.**CB, TM, and DLS conceived and designed the experiments. CB, TM, RB, AWS, SK and DLS performed the experiments. CB, TM, and DLS analyzed the data. CB, TM and DLS wrote the paper. Academic Editor: Benjamin Normark, University of Massachusetts at Amherst. ¤ Present Address: Institut Jacques Monod, Paris, France. Abd-A : Abdominal-A Dll : Distal-less En : Engrailed Eve : Even-skipped Hox : homeotic complex Ubx : Ultrabithorax.

PubMed Central

2024-06-05T03:55:45.977055

2003-10-13

PMC212700

When it comes to exploiting a niche, endosymbionts take the prize. In endosymbiosis, one organism---the *endosymbiont*---invades the cells of another, in some cases taking up residence in a way that actually benefits the host. Bacteria are particularly adept at making themselves indispensable by insinuating themselves into some fundamental aspect of an organism\'s biology. The endosymbiotic hypothesis proposes that this is how certain eukaryotic organelles evolved from endosymbiotic bacteria. Insights into the mechanisms governing endosymbiosis will help biologists understand how this mutually beneficial relationship evolved and provide clues to one of the fundamental questions in biology: How did the eukaryotic cell evolve? Over 10% of insect species rely on endosymbionts for their development and survival. In this issue, David Stern and colleagues look at one of the most studied pairs, the pea aphid and Buchnera aphidicola, and discover clues to the molecular foundation of their shared fate. (Buchnera, which can no longer survive outside its host cell, is thought to produce essential amino acids that the aphid cannot get on its own.) While it is known that Buchnera are transferred from clusters of bacteriocytes in the mother to the adjacent early-stage embryo, it has been unclear how the bacteriocytes develop. Previous studies of the bacteria\'s genome have failed to explain the genetic basis of Buchnera\'s ability to invade aphid cells. Consequently, Stern and colleagues have focused on the bacteriocytes, the specialized insect cells that house Buchnera, shedding light on the development of these cells as well as on the evolutionary adaptations in the aphid that made the bacteriocytes hospitable to Buchnera. The researchers show that bacteriocytes differentiate and proliferate independently of Buchnera\'s presence in the cell, and they identify three aphid transcription factors (proteins that regulate gene expression) that are expressed in three distinct stages during early-bacteriocyte development in the aphid embryo. The first protein is expressed just before Buchnera enters the embryo; a second, as the bacteria invades; and a third, after the transfer is nearly complete. A second wave of the same transcription factors occurs at a later stage in aphid embryo development and increases the population of bacteriocytes. This two-step specification of bacteriocytes, which occurs in related Buchnera-carrying aphid species, appears to be an evolutionarily conserved feature of aphids. It even occurs in an aphid species that once had a Buchnera endosymbiont and now has a yeast-like symbiont that lives outside the bacteriocytes. But this process is not observed in males of another aphid species that do not carry Buchnera. While traces of the first transcription factor activated in bacteriocytes are evident, the characteristic gene-expression pattern is not, and the aphids have no mature bacteriocytes. While it seems that the aphid has evolved new domains of expression in the bacteriocyte for these transcription factors---none of these transcription factors is expressed at a similar stage in other insects---the researchers cannot yet say whether these genes direct the specification of bacteriocytes. Still, these transcription factors are likely to play important roles in the bacteriocyte, suggesting that the union of aphids and Buchnera involved significant adaptations by the host. ::: {#pbio-0000024-g001 .fig} ::: {.caption} ###### Aphid host of Buchnera endosymbionts :::  :::

PubMed Central

2024-06-05T03:55:45.979401

2003-10-13

PMC212701

The past ten years saw great progress in the field of molecular genetics, as new tools gave scientists the ability to investigate entire genomes instead of just one or two genes at a time. In this paper, Ronald Plasterk and colleagues developed a systemic approach using Caenorhabditis elegans, a tiny nematode and the first animal to have its genome sequenced, to gather functional information on nearly 400 genes. Many of the systemic approaches to discovering gene function involve either measuring or deleting messenger RNA (mRNA), the molecule that helps translate genes into proteins. The method used here, called RNA interference, or RNAi, follows the deletion approach by taking advantage of a cellular process bearing the same name. In nature, RNAi is thought to be an important part of the innate defense machinery in plants and animals, protecting them from invaders like viruses by interrupting the manufacture of viral proteins. To do this, short double-stranded RNA molecules with complementary sequences to the target gene inhibit the gene\'s function by disabling mRNA, which effectively shuts down the gene. By mimicking this natural process to turn off selected genes, scientists can find clues to how those genes might normally function by watching what happens when they are taken out of the picture. With the fully sequenced worm genome, it is possible to create interfering RNAs for all of its 20,000 or so genes. And because worms eat bacteria---which can themselves be used to deliver interfering RNAs---worms are the perfect RNAi model organism. The researchers fed the worms RNAi-producing bacteria, then observed the effects on the worm or its offspring to infer the function of the targeted gene. As previously reported, repeating this experiment for every gene in the worm genome, yields about 10% of the worms displaying abnormalities ranging from embryonic death to uncoordinated movement, suggesting defects in genes controlling development or muscle control, respectively. Having previously identified an RNAi-hypersensitive mutant worm strain, Plasterk and his colleagues repeated the experiment in the mutants and report proposed functions for 393 previously unknown genes. The types of abnormalities observed in the short-lived mutations induced by RNAi, they say, resemble the more stable mutations seen in the collection of worm mutations cataloged by worm researchers over the years. Though the DNA alterations for many of these mutations are not yet known, researchers know roughly where they occur in the genome. And the researchers show here that they can use their RNAi experimental results along with what is known about the mutants to identify several of the sequence alterations. They also performed what is believed to be the first analysis in which independently generated large-scale RNAi results were systematically compared to see how variable such RNAi results are, and the results have implications for similar approaches not just in worms but in plants and other animals. ::: {#pbio-0000026-g001 .fig} ::: {.caption} ###### Caenorhabditis elegans worms :::  :::

PubMed Central

2024-06-05T03:55:45.979927

2003-10-13

PMC212702

Diversity of human appearance and form has intrigued biologists for centuries, but nearly 100 years after the term "genetics" was coined by William Bateson in 1906, the genes that underlie this diversity are an unsolved mystery. One of the most obvious phenotypes that distinguish members of our species, differences in skin pigmentation, is also one of the most enigmatic. There is a tremendous range of human skin color in which variation can be correlated with climates, continents, and/or cultures, yet we know very little about the underlying genetic architecture. Is the number of common skin color genes closer to five, 50, or 500? Do gain- and loss-of-function alleles for a small set of genes give rise to phenotypes at opposite ends of the pigmentary spectrum? Has the effect of natural selection on similar pigmentation phenotypes proceeded independently via similar pathways? And, finally, should we care about the genetics of human pigmentation if it is only skin-deep? Why Should We Care? {#s2} =================== From a clinical perspective, inadequate protection from sunlight has a major impact on human health ([@pbio-0000027-Armstrong1]; [@pbio-0000027-Diepgen1]). In Australia, the lifetime cumulative incidence of skin cancer approaches 50%, yet the oxymoronic "smart tanning" industry continues to grow, and there is controversy over the extent to which different types of melanin can influence susceptibility to ultraviolet (UV) radiation ([@pbio-0000027-Schmitz1]; [@pbio-0000027-Wenczl1]). At the other end of the spectrum, inadequate exposure to sunlight, leading to vitamin D deficiency and rickets, has been mostly cured by nutritional advances made in the early 1900s. In both cases, understanding the genetic architecture of human skin color is likely to provide a greater appreciation of underlying biological mechanisms, much in the same way that mutational hotspots in the gene *TP53* have helped to educate society about the risks of tobacco ([@pbio-0000027-Takahashi1]; [@pbio-0000027-Toyooka1]). From a basic science perspective, variation in human skin color represents an unparalleled opportunity for cell biologists, geneticists, and anthropologists to learn more about the biogenesis and movement of subcellular organelles, to better characterize the relationship between genotypic and phenotypic diversity, to further investigate human origins, and to understand how recent human evolution may have been shaped by natural selection. The Color Variation Toolbox {#s3} =========================== Historically, measurement of human skin color is often based on subjective categories, e.g., "moderate brown, rarely burns, tans very easily." More recently, quantitative methods based on reflectance spectrophotometry have been applied, which allow reddening caused by inflammation and increased hemoglobin to be distinguished from darkening caused by increased melanin ([@pbio-0000027-Alaluf2]; [@pbio-0000027-Shriver1]; [@pbio-0000027-Wagner1]). Melanin itself is an organic polymer built from oxidative tyrosine derivatives and comes in two types, a cysteine-rich red--yellow form known as pheomelanin and a less-soluble black\--brown form known as eumelanin ([Figure 1A](#pbio-0000027-g001){ref-type="fig"}). Discriminating among pigment types in biological samples requires chemical extraction, but is worth the effort, since the little we do know about common variation in human pigmentation involves pigment type-switching. The characteristic phenotype of fair skin, freckling, and carrot-red hair is associated with large amounts of pheomelanin and small amounts of eumelanin and is caused by loss-of-function alleles in a single gene, the melanocortin 1 receptor *(MC1R)* ([@pbio-0000027-Sturm1]; [@pbio-0000027-Rees1]) However, *MC1R* variation has a significant effect on pigmentation only in populations where red hair and fair skin are common ([@pbio-0000027-Rana1]; [@pbio-0000027-Harding1]), and its primary effects---to promote eumelanin synthesis at the expense of pheomelanin synthesis, or vice versa--- contribute little to variation of skin reflectance among or between major ethnic groups ([@pbio-0000027-Alaluf1]). ::: {#pbio-0000027-g001 .fig} Figure 1 ::: {.caption} ###### Biochemistry and Histology of Different Skin Types \(A) Activation of the melanocortin 1 receptor (MC1R) promotes the synthesis of eumelanin at the expense of pheomelanin, although oxidation of tyrosine by tyrosinase (TYR) is required for synthesis of both pigment types. The membrane-associated transport protein (MATP) and the pink-eyed dilution protein (P) are melanosomal membrane components that contribute to the extent of pigment synthesis within melanosomes. (B) There is a gradient of melanosome size and number in dark, intermediate, and light skin; in addition, melanosomes of dark skin are more widely dispersed. This diagram is based on one published by [@pbio-0000027-Sturm1] and summarizes data from [@pbio-0000027-Szabo1], [@pbio-0000027-Toda1], and [@pbio-0000027-Konrad1] based on individuals whose recent ancestors were from Africa, Asia, or Europe. :::  ::: More important than the ratio of melanin types is the total amount of melanin produced. In addition, histological characteristics of different-colored skin provide some clues as to cellular mechanisms that are likely to drive pigmentary variation ([Figure 1B](#pbio-0000027-g001){ref-type="fig"}). For the same body region, light- and dark-skinned individuals have similar numbers of melanocytes (there is considerable variation between different body regions), but pigment-containing organelles, called melanosomes, are larger, more numerous, and more pigmented in dark compared to intermediate compared to light skin, corresponding to individuals whose recent ancestors were from Africa, Asia, or Europe, respectively ([@pbio-0000027-Szabo1]; [@pbio-0000027-Toda1]; [@pbio-0000027-Konrad1]). From these perspectives, oxidative enzymes like tyrosinase (TYR), which catalyzes the formation of dopaquinone from tyrosine, or melanosomal membrane components like the pink-eyed dilution protein (P) or the membrane-associated transporter protein (MATP), which affect substrate availability and activity of TYR ([@pbio-0000027-Orlow1]; [@pbio-0000027-Brilliant1]; [@pbio-0000027-Newton1]; [@pbio-0000027-Costin1]), are logical candidates upon which genetic variation could contribute to the diversity of human skin color. Of equal importance to what happens inside melanocytes is what happens outside. Each pigment cell actively transfers its melanosomes to about 40 basal keratinocytes; ultimately, skin reflectance is determined by the amount and distribution of pigment granules within keratinocytes rather than melanocytes. In general, melanosomes of African skin are larger and dispersed more widely than in Asian or European skin ([Figure 1](#pbio-0000027-g001){ref-type="fig"}). Remarkably, keratinocytes from dark skin cocultured with melanocytes from light skin give rise to a melanosome distribution pattern characteristic of dark skin, and vice versa ([@pbio-0000027-Minwalla1]). Thus, at least one component of skin color variation represents a gene or genes whose expression and action affect the pigment cell environment rather than the pigment cell itself. Genetics of Skin Color {#s4} ====================== For any quantitative trait with multiple contributing factors, the most important questions are the overall heritability, the number of genes likely to be involved, and the best strategies for identifying those genes. For skin color, the broad sense heritability (defined as the overall effect of genetic vs. nongenetic factors) is very high ([@pbio-0000027-Clark1]), provided one is able to control for the most important nongenetic factor, exposure to sunlight. Statements regarding the number of human skin color genes are attributed to several studies; one of the most complete is by [@pbio-0000027-Harrison1]. In that study, skin reflectance measurements were obtained from 70 residents of Liverpool whose parents, grandparents, or both were of European ("with a large Irish component") or West African ("mostly from coastal regions of Ghana and Nigeria") descent and who were roughly classified into "hybrid" and "backcross" groups on this basis. An attempt to partition and analyze the variance of the backcross groups led to minimal estimates of three to four "effective factors," in this case, independently segregating genes. Aside from the key word *minimal* (Harrison and Owen\'s data could also be explained by 30--40 genes), one of the more interesting findings was that skin reflectance appeared to be mainly additive. In other words, mean skin reflectance of "F1 hybrid" or "backcross hybrid" groups is intermediate between their respective parental groups. An alternative approach for considering the number of potential human pigmentation genes is based on mouse coat color genetics, one of the original models to define and study gene action and interaction, for which nearly 100 different genes have been recognized ([@pbio-0000027-Bennett1]; [@pbio-0000027-Jackson1]). Setting aside mouse mutations that cause white spotting or predominant effects outside the pigmentary system, no more than 15 or 20 mutations remain, many of which have been identified and characterized, and most of which have human homologs in which null mutations cause albinism. This brings us to the question of candidate genes for skin color, since, like any quantitative trait, a reasonable place to start is with rare mutations known to cause an extreme phenotype, in this case Mendelian forms of albinism. The underlying assumption is that if a rare null allele causes a complete loss of pigment, then a set of polymorphic, i.e., more frequent, alleles with subtle effects on gene expression will contribute to a spectrum of skin colors. The *TYR, P*, and *MATP* genes discussed earlier are well-known causes of albinism whose primary effects are limited to pigment cells ([@pbio-0000027-Oetting1]); among these, the *P* gene is highly polymorphic but the phenotypic consequences of *P* gene polymorphisms are not yet known. Independent of phenotype, a gene responsible for selection of different skin colors should exhibit a population signature with a large number of alleles and rates of sequence substitution that are greater for nonsynonymous (which change an amino acid in the protein) than synonymous (which do not change any amino acid) alterations. Data have been collected only for *MC1R*, in which the most notable finding is a dearth of allelic diversity in African samples, which is remarkable given that polymorphism for most genes is greater in Africa than in other geographic regions ([@pbio-0000027-Rana1]; [@pbio-0000027-Harding1]). Thus, while *MC1R* sequence variation does not contribute significantly to variation in human skin color around the world, a functional *MC1R* is probably important for dark skin. Selection for Skin Color? {#s5} ========================= Credit for describing the relationship between latitude and skin color in modern humans is usually ascribed to an Italian geographer, Renato Basutti, whose widely reproduced "skin color maps" illustrate the correlation of darker skin with equatorial proximity ([Figure 2](#pbio-0000027-g002){ref-type="fig"}). More recent studies by physical anthropologists have substantiated and extended these observations; a recent review and analysis of data from more than 100 populations ([@pbio-0000027-Relethford1]) found that skin reflectance is lowest at the equator, then gradually increases, about 8% per 10° of latitude in the Northern Hemisphere and about 4% per 10° of latitude in the Southern Hemisphere. This pattern is inversely correlated with levels of UV irradiation, which are greater in the Southern than in the Northern Hemisphere. An important caveat is that we do not know how patterns of UV irradiation have changed over time; more importantly, we do not know when skin color is likely to have evolved, with multiple migrations out of Africa and extensive genetic interchange over the last 500,000 years ([@pbio-0000027-Templeton1]). ::: {#pbio-0000027-g002 .fig} Figure 2 ::: {.caption} ###### Relationship of Skin Color to Latitude \(A) A traditional skin color map based on the data of Biasutti. Reproduced from <http://anthro.palomar.edu/vary/> with permission from Dennis O\'Neil. [^1] (B) Summary of 102 skin reflectance samples for males as a function of latitude, redrawn from [@pbio-0000027-Relethford1]. :::  ::: Regardless, most anthropologists accept the notion that differences in UV irradiation have driven selection for dark human skin at the equator and for light human skin at greater latitudes. What remains controversial are the exact mechanisms of selection. The most popular theory posits that protection offered by dark skin from UV irradiation becomes a liability in more polar latitudes due to vitamin D deficiency ([@pbio-0000027-Murray1]). UVB (short-wavelength UV) converts 7-dehydrocholesterol into an essential precursor of cholecaliferol (vitamin D~3~); when not otherwise provided by dietary supplements, deficiency for vitamin D causes rickets, a characteristic pattern of growth abnormalities and bony deformities. An oft-cited anecdote in support of the vitamin D hypothesis is that Arctic populations whose skin is relatively dark given their latitude, such as the Inuit and the Lapp, have had a diet that is historically rich in vitamin D. Sensitivity of modern humans to vitamin D deficiency is evident from the widespread occurrence of rickets in 19th-century industrial Europe, but whether dark-skinned humans migrating to polar latitudes tens or hundreds of thousands of years ago experienced similar problems is open to question. In any case, a risk for vitamin D deficiency can only explain selection for light skin. Among several mechanisms suggested to provide a selective advantage for dark skin in conditions of high UV irradiation ([@pbio-0000027-Loomis1]; [@pbio-0000027-Robins1]; [@pbio-0000027-Jablonski1]), the most tenable are protection from sunburn and skin cancer due to the physical barrier imposed by epidermal melanin. Solving the Mystery {#s6} =================== Recent developments in several areas provide a tremendous opportunity to better understand the diversity of human pigmentation. Improved spectrophotometric tools, advances in epidemiology and statistics, a wealth of genome sequences, and efficient techniques for assaying sequence variation offer the chance to replace misunderstanding and myths about skin color with education and scientific insight. The same approaches used to investigate traits such as hypertension and obesity---genetic linkage and association studies---can be applied in a more powerful way to study human pigmentation, since the sources of environmental variation can be controlled and we have a deeper knowledge of the underlying biochemistry and cell biology. This approach is especially appealing given the dismal success rate in molecular identification of complex genetic diseases. In fact, understanding more about the genetic architecture of skin color may prove helpful in designing studies to investigate other quantitative traits. Current debates in the human genetics community involve strategies for selecting populations and candidate genes to study, the characteristics of sequence polymorphisms worth pursuing as potential disease mutations, and the extent to which common diseases are caused by common (and presumably ancient) alleles. While specific answers will be different for every phenotype, there may be common themes, and some answers are better than none. [@pbio-0000027-Harrison1] concluded their 1964 study of human skin color by stating, "The deficiencies in the data in this study are keenly appreciated by the writers, but since there appear at present to be no opportunities for improving the data, it seems justifiable to take the analysis as far as possible." Nearly 40 years later, opportunities abound, and the mystery of human skin color is ready to be solved. I am grateful to members of my laboratory and colleagues who study pigment cells in a variety of different experimental organisms for useful discussions and to Sophie Candille for helpful comments on the manuscript. Many of the ideas presented here emerged during a discussion series on Unsolved Mysteries in Biomedical Research that was initiated by Mark Krasnow and the Medical Scientist Training Program at Stanford University. Gregory S. Barsh is an associate professor of Departments of Genetics and Pediatrics and an associate investigator at the Howard Hughes Medical Institute, Stanford University School of Medicine, Stanford, California, United States. E-mail: <gbarsh@cmgm.stanford.edu>. [^1]: Erratum note: The source of this image was incorrectly acknowledged. Corrected 12/19/03.

PubMed Central

2024-06-05T03:55:45.980639

2003-10-13

PMC212703

A leading cause of death and disability, diabetes affects some 16 million Americans and up to 135 million people worldwide. While environmental factors such as diet and lifestyle are known to influence an individual\'s risk of getting adult-onset, or Type 2, diabetes, there is also a substantial inherited component, though many of the genetic pathways involved remain unidentified. The challenge of defining these genetic pathways lies in the fact that diabetes is what is known as a "complex trait": not only is it likely that variations in many different genes or some combination of genes contribute to an increased risk, but there are probably different genes associated with diabetes in different populations. Tackling the monumental task of deciphering this genetic puzzle in the largest known study of its kind, Inês Barroso and colleagues confirm the genetic complexity of the disease and clearly demonstrate that untangling the genetics will require even larger studies. In diabetes, defects in both the secretion and function of insulin---which is produced by the pancreas---impair the body\'s ability to metabolize glucose. Based on what is known about the biology of pancreatic function and diabetes, the researchers chose 71 potential suspect genes that could reasonably be expected to contribute to the disease if they malfunctioned. Some of these genes are involved in the function of pancreatic beta-cells, which secrete insulin; a second group influences the function of insulin and glucose metabolism; and a third plays a broader role in energy metabolism. The results show that none of the genes on their own had a large effect, though a number of gene variations increased risk slightly. They also suggest the existence of variations in several genes that influence the risk of Type 2 diabetes. The dataset will be valuable for future studies of diabetes and supports the view that variation in genes affecting insulin production as well as insulin action can influence the risk of Type 2 diabetes. However, the genetic complexity of the disease---with a number of genes conveying a slightly increased risk--- demands additional studies of larger populations to reliably identify the genes involved and the genetic variations that, alone or in combination, increase or lower an individual\'s risk of developing the disease.

PubMed Central

2024-06-05T03:55:45.982120

2003-10-13

PMC212704

From the moment its life begins, the fate of a multicellular organism depends on how well its cells communicate. Proteins act as molecular switchboard operators to keep the lines of communication open and the flow of cellular messages on track. But charting the protein interactions, signaling pathways, and other elements that regulate these networks is no small feat. Previous efforts have been hampered by the lack of quantitative data---measurements of signal duration, amplitude, and fluctuation---on these regulatory pathways. Hoping to fill in some of the quantitative gaps, Marc Kirschner of Harvard Medical School, Reinhart Heinrich of Humboldt University Berlin, and colleagues developed a mathematical model as a framework for understanding the quantitative relationships among signaling proteins. To do this, they focused on a well-studied signaling pathway, the Wnt pathway, which plays a role both in various stages of embryonic development and in carcinogenesis. The researchers chose the Wnt pathway in part because a lot is known about it and in part because they could collect enough of the additional measurements they needed to build a solid model from experiments. And like most signaling pathways, Wnt is highly conserved. Consequently, developing tools that elucidate the Wnt pathway will not only provide insights into this important pathway, but have implications for understanding other communication pathways in animals from jellyfish to humans. To get the additional measurements needed to build their model, the researchers reproduced aspects of the Wnt pathway in the cytoplasm of unfertilized frog eggs. Among the new data collected from these experiments were measurements of the concentrations of scaffold proteins, which bring other components in a pathway together by providing an interaction surface. Strikingly, they found that the principal scaffold proteins involved in the pathway, axin and adenomatous polyposis coli (APC), occur in dramatically different concentrations and perform their jobs in different ways. After a series of refinements based on additional experiments, the model could not only simulate the behavior of the main players in the pathway---both in the absence and presence of a Wnt signal---it also suggested why the two scaffold proteins are present in different concentrations. Axin occurs at very low concentrations relative to the other proteins in the pathway and is likely to bind with them randomly, while APC occurs in similar concentrations and probably binds with the other components in an ordered manner. Because the proteins axin interacts with are also involved in other signaling pathways, the authors propose that the low level of axin here may help the pathways retain their modularity, preventing the Wnt pathway from interfering with the other pathways. These findings demonstrate that modeling can offer powerful new insights into the workings of complex signaling systems, cutting through the static to pick up important signals even in those pathways that are well understood. The results have important implications for developmental biology and human disease: The Wnt pathway is often activated during carcinogenesis---and mutations in several of these signaling proteins have been linked to colon cancer---suggesting that cancer can develop when signals in the Wnt circuitry somehow get crossed. By predicting how quantitative factors may influence the behavior of signaling networks, mathematical models such as this could shed light on the role that breakdowns in cellular communication play in carcinogenesis. The researchers argue that future attempts to characterize these complex networks must incorporate quantification measurements, and their modeling efforts suggest ways to do that. ::: {#pbio-0000032-g001 .fig} ::: {.caption} ###### Understanding Wnt signaling through molecular modeling :::  :::

PubMed Central

2024-06-05T03:55:45.982522

2003-10-13

PMC212705

Welcome to *PLoS Biology*. We would like to introduce you to your journal, one that is run by and for the scientific community in the broadest sense: researchers, teachers, students, physicians, and the public. One could argue whether scientists need more journals, but we believe there is a global need for greater access to scientific and medical information and that open-access journals can meet this need by removing subscription barriers to the written scientific record. As professional editors, each of us entered the publishing world from the research community with the desire to promote the effective communication and dissemination of science. Offered the opportunity to help spark the transition to open-access publishing by creating an open-access journal that would compete successfully with the most prestigious existing journals, we jumped at the chance. What you see in this issue is the result of a collaborative effort among the founders, the journal\'s editorial board, and its professional editors. A glance at the table of contents and the list of outstanding scientists on our editorial board will give you a sense for the scope of the journal, which ranges from molecules to ecosystems and spans the experimental and theoretical disciplines that help to explain our biological world. We aim to publish original articles that address an important question, that challenge our assumptions, that drive science forward. Our editorial and peer-review process combines the expertise of both professional editors, who are available on a full-time basis to survey the broad landscape of science as well as to engage reviewers and communicate decisions, and academic editors, who understand deeply the strengths and limitations of their area of research. Readers benefit from a selection of exciting and important scientific advances in diverse disciplines. Authors benefit from decisions in which constructive advice is offered and in which papers are not automatically sent back for multiple rounds of review when it is clear to an academic editor who works in the field that concerns raised by reviewers have been satisfied. Every paper published in *PLoS Biology* is read not only by carefully selected reviewers, but also by an academic editor and a professional editor, who work together throughout the editorial process. Many scientists, both from our editorial board and beyond it, have been our partners in making editorial decisions on the research articles chosen for publication in this and subsequent issues. You will see them listed as academic editors on the individual papers. Research articles in *PLoS Biology* have no strict length limits; we are keen to provide authors with the opportunity to tell their story in a clear manner and in their own voice. We do not expect, however, that every reader of *PLoS Biology* will be able to appreciate the primary research articles in full, that, for example, a computational neuroscientist will have the specialized vocabulary and working knowledge of an immunologist, or vice versa. What we do expect is that all interested readers will be able to understand and appreciate the Synopses that accompany every article. We also hope that readers will take advantage of our Primers to learn about the tools and topics that are current to the scientific enterprise and of our other magazine content to explore the larger world of science. We hope that you will lead the open-access revolution by publishing your most exciting research in *PLoS Biology*. And we welcome your ideas for our magazine section. We invite students and scientists at early stages in their careers to present in our Journal Club their critical perspectives and insights on research articles that have captured their imaginations or aroused their skeptical instincts. And we invite other organizations interested in the dissemination and understanding of scientific information to tell us about their activities in our Community Pages. *PLoS Biology* will be published monthly in print and online. We view our Web site as the primary form of publication for *PLoS Biology,* and we offer different formats to meet the needs of individual readers, such as a version that will help readers with low bandwidth connections and a separate view of the figures and tables in each paper. The Public Library of Science is committed to making the scientific literature an open resource. The most tangible evidence of this will be the lack of barriers and the ease of navigation that you will experience as you explore our Web site. Many of you with institutional site licenses to other journals might not readily appreciate this ease of access (unless perhaps you are away from your home institution and need that critical paper). But for many of you around the world, it will make the difference between reading the abstract or reading the entire paper. Our goal, however, is not simply to provide a free online journal. Our goal is to create a potent scientific and public resource. As more open-access articles become available, we and many others will be working to develop new tools for integrating, interlinking, organizing, searching, and annotating this expanding collection of information. We invite the community to share these tools as they become available on our site, as well as to propose tools, links, and techniques that would make this treasury of scientific information more useful. *PLoS Biology* is yours. Download it, copy it, incorporate it in your own database, post or reprint any article; use your imagination. We ask only that you give fair credit to the authors of any work that you use. If you like what you see, e-mail an article to a friend and encourage others to submit their best work. We also invite you to tell us how we can improve the journal and help it evolve into a resource that optimizes the international investment of money, time, energy, and intelligence that goes into the scientific process. You can make a difference.

PubMed Central

2024-06-05T03:55:45.983191

2003-10-13

PMC212706

Communication among scientists has undergone a revolution in the last decade, with the movement of scientific publication to a digital medium and the emergence of the Internet as the primary means for distributing information. Millions of articles are, in principle, just a mouse click away from our computers. For many of us, PDFs have replaced printed journals as the primary form in which we read about the work of our colleagues. Yet we have barely begun to realize the potential of this technological change. For practicing scientists, it provides myriad opportunities to expand and improve the ways we can use the scientific literature. Equally important, it is now possible to make our treasury of scientific information available to a much wider audience, including millions of students, teachers, physicians, scientists, and other potential readers, who do not have access to a research library that can afford to pay for journal subscriptions. We founded the Public Library of Science three years ago to work toward realizing these opportunities. We began as a grassroots organization of scientists, advocating the establishment and growth of online public libraries of science, such as the National Institutes of Health\'s PubMed Central, to provide free and unrestricted access to the scientific literature. Today, with the launch of *PLoS Biology*, we take on a new role as publishers, to demonstrate that high-quality journals can flourish without charging for access. Open Access {#s2} =========== *PLoS Biology*, and every PLoS journal to follow, will be an open-access publication--everything we publish will immediately be freely available to anyone, anywhere, to download, print, distribute, read, and use without charge or other restrictions, as long as proper attribution of authorship is maintained. Our open-access journals will retain all of the qualities we value in scientific journals---high standards of quality and integrity, rigorous and fair peer-review, expert editorial oversight, high production standards, a distinctive identity, and independence. Although most readers will be satisfied with the free and unrestricted use of the online edition (including the right to print their own copies), a printed edition of *PLoS Biology* will be made available, for the cost of printing and distribution, to readers who prefer the convenience and browseability of the traditional paper format. And the full contents of every issue will immediately be placed in the National Library of Medicine\'s public online archive, PubMed Central, guaranteeing their permanent preservation and free accessibility. Our aim is to catalyze a revolution in scientific publishing by providing a compelling demonstration of the value and feasibility of open-access publication. If we succeed, everyone who has access to a computer and an Internet connection will be a keystroke away from our living treasury of scientific and medical knowledge. This online public library of science will form a valuable resource for science education, lead to more informed healthcare decisions by doctors and patients, level the playing field for scientists in smaller or less wealthy institutions, and ensure that no one will be unable to read an important paper just because his or her institution does not subscribe to a particular journal. Open access will also enable scientists to begin transforming the scientific literature into something far more useful than the electronic equivalent of millions of individual articles in rows of journals on library shelves. The ability to search, in an instant, an entire scientific library for particular terms or concepts, for methods, data, and images---and instantly retrieve the results---is only the beginning. Freeing the information in the scientific literature from the fixed sequence of pages and the arbitrary boundaries drawn by journals or publishers--- the electronic vestiges of paper publication---opens up myriad new possibilities for navigating, integrating, "mining," annotating, and mapping connections in the high-dimensional space of scientific knowledge. Consider how the open availability and freedom to use the complete archive of published DNA sequences in the GenBank, EMBL, and DDBJ databases inspired and enabled scientists to transform a collection of individual sequences into something incomparably richer. With great foresight, it was decided in the early 1980s that published DNA sequences should be deposited in a central repository, in a common format, where they could be freely accessed and used by anyone. Simply giving scientists free and unrestricted access to the raw sequences led them to develop the powerful methods, tools, and resources that have made the whole much greater than the sum of the individual sequences. Just one of the resulting software tools---BLAST---performs 500 trillion sequence comparisons annually! Imagine how impoverished biology and medicine would be today if published DNA sequences were treated like virtually every other kind of research publication---with no comprehensive database searches and no ability to freely download, reorganize, and reanalyze sequences. Now imagine the possibilities if the same creative explosion that was fueled by open access to DNA sequences were to occur for the much larger body of published scientific results. Paying the Bill for Open Access {#s3} =============================== The benefits of open access are incontestable. The questions and concerns that remain focus on finances. As everyone acknowledges, publishing a scientific journal costs money---the more rigorous the peer review, the more efficient and expert the editorial oversight, the more added features and the higher the production standards, the greater the cost to publishers. Most journals today depend on subscriptions and site-licensing fees for most of their revenue. Since these access tolls are incompatible with open access, how will newly formed open-access journals pay their bills, and how will the traditional journals that have served the scientific community for many years survive in an open-access world? Because publishing is an integral part of the research process, a natural alternative to the subscription model is to consider the significant but relatively small costs of open-access publication as one of the fundamental costs of doing research. The institutions that sponsor research intend for the results to be made available to the scientific community and the public. If these research sponsors also paid the essential costs of publication---amounting, by most estimates, to less than 1% of the total spent on sponsored research (statistics found at <http://dx.doi.org/10.1371/journal.pbio.0000036.sd001>)---we would retain a robust and competitive publishing industry and gain the benefit of universal open access. The subscription model---in which the publishers own the works they publish and dictate the conditions under which they can be accessed or used---is sometimes presented as the only possible way to pay for scientific publishing. This pay-for-access model was well suited to a world in which the most efficient way to record and transmit scientific information on a large scale was by printing and distributing scientific journals. When each incremental copy represented a significant expense to the publisher, any sustainable business model depended on recovering the cost for each copy---the recipients of the information had to pay for access. But the essential rationale of the pay-for-access model has disappeared, now that electronic publication and Internet distribution have become routine. Instead, this business model is what stands in the way of all the benefits of open access. Asking research sponsors to pay for publication of the research they support may seem to impose new financial burdens on the government agencies, foundations, universities, and companies that sponsor research. But these organizations already pay most of the costs of scientific publishing---a huge fraction of the US\$9 billion annual revenue of scientific, medical, and technology journals comes from subscriptions, site licenses, and publication fees ultimately billed to grants or employers. Much of the rest is borne by society in the form of increments to university tuitions; healthcare costs, including drug prices; and state and federal taxes that subsidize healthcare, libraries, and education. Surely the cost of open-access digital publishing cannot, in total, be more than we are already paying under the subscription and licensing model. By simply changing the way we support the scientific publishing enterprise, the scientific community and public would preserve everything we value in scientific publishing and gain all of the benefits of open access. There are reasons to believe that open-access publishing would cost significantly less than the current system. Today, each journal has a monopoly on a resource vital to scientists---the unique collection of articles it has published. Anyone who depends on the information in a specific article has no choice but to pay whatever price the publisher asks (or find a colleague or library that has done so). Because scientists are so dependent on ready access to previously published work, publishers are able to set their prices with little fear of subscription cancellations. Indeed, journal prices have been rising at a rate well in excess of inflation, straining the budgets of universities, hospitals, and research institutions. Open access would eliminate monopolies over essential published results, diminishing profit margins and creating a more efficient market for scientific publishing---a market in which publishers would compete to provide the best value to authors (high quality, selectivity, prestige, a large and appreciative readership) at the best price. Joining Forces {#s4} ============== In recent months, we have witnessed a remarkable surge of awareness and support for open-access publication, both within the scientific community and in the public at large, exemplified by recent newspaper articles and editorials supporting PLoS and open access; by the recent introduction of the Public Access to Science Act in the United States Congress; by the Bethesda Workshop on Open Access; and by public statements of support from organizations as diverse as the NIH Council of Public Representatives, the Association of Research Libraries, and the Susan G. Komen Breast Cancer Foundation. Achieving universal open access will require action from funding agencies and institutions. The Howard Hughes Medical Institute, the largest private sponsor of biomedical research in the United States, has already taken a leading role in promoting open access. They will provide each of their investigators with supplemental funds to cover the costs of publishing in open-access journals like *PLoS Biology.* Other major institutional sponsors of biomedical research are actively considering similar policies. Private foundations with a commitment to science and education have contributed generously to this cause. Like any new business, PLoS needed to raise funds to cover our startup costs. A generous grant from the Gordon and Betty Moore Foundation enabled PLoS to launch our nonprofit publishing venture. Other individuals and organizations, notably the Irving A. Hansen Foundation, also provided generous and welcome support. These start-up funds made it possible for us to assemble an outstanding editorial board and staff, who have today accomplished the extraordinary feat of launching a new publisher and a premiere journal from scratch in less than nine months. The opposition of most established journals to open access has left it to new journals like *PLoS Biology* and BioMed Central\'s *Journal of Biology* to lead the way. These new journals face a double challenge. First, we are introducing an unfamiliar model---open-access publication. Second, any new journal, even one with the stringent standards and the extraordinary editorial team of *PLoS Biology*, must begin without the established "brand name" of the older journals, which, like a designer logo, elevates the perceived status of the articles that bear it. With all that is at stake in the choice of a journal in which to publish---career advancement, grant support, attracting good students and fellows---scientists who believe in the principle of open access and wish to support it are confronted with a difficult dilemma. We applaud the courage and pioneering spirit of the authors who have chosen to send to a fledgling journal the outstanding articles you will read in the premiere issues of *PLoS Biology.* In the end, it\'s the contributions of these authors that will make *PLoS Biology* a success.

PubMed Central

2024-06-05T03:55:45.983989

2003-10-13

PMC261870

Introduction {#s1} ============ Over the past two decades, technological advances in molecular biology have fuelled progress in the discovery of new pathogens associated with human diseases. The identification of novel viruses such as hepatitis C virus ([@pbio.0000002-Choo1]), sin nombre virus ([@pbio.0000002-Nichol1]), and Kaposi\'s sarcoma herpesvirus ([@pbio.0000002-Chang1]) has relied upon a diverse range of modern molecular methods such as immunoscreening of cDNA libraries, degenerate PCR, and representational difference analysis, respectively. In spite of these successes, there remain numerous syndromes with suspected infectious etiologies that continue to escape identification efforts, in part due to limitations of existing methodologies for viral discovery ([@pbio.0000002-Muerhoff1]; [@pbio.0000002-Kellam1]). These limitations, coupled with the constant threat posed by newly emerging infectious diseases of unknown origin, necessitate that new approaches be developed to augment the repertoire of available tools for pathogen discovery. We have previously described a prototype DNA microarray designed for highly parallel viral detection with the potential to detect novel members of known viral families ([@pbio.0000002-Wang1]). This microarray contained approximately 1600 oligonucleotides representing 140 viruses. Building upon this foundation, a more comprehensive second-generation DNA microarray consisting of 70mer oligonucleotides derived from every fully sequenced reference viral genome in GenBank (as of August 15, 2002) was constructed. The most highly conserved 70mers from each virus were selected as described by [@pbio.0000002-Wang1]) to maximize the probability of detecting unknown and unsequenced members of existing families by cross-hybridization to these array elements. On average, ten 70mers were selected for each virus, totaling approximately 10,000 oligonucleotides from approximately 1,000 viruses. The objective was to create a microarray with the capability of detecting the widest possible range of both known and unknown viruses. This pan-viral microarray was used as part of the global effort to identify a novel virus associated with severe acute respiratory syndrome (SARS) in March 2003, as reported by [@pbio.0000002-Ksiazek1]). We describe here the experimental details of the microarray methodology for novel virus identification, using the SARS outbreak as an example. Results {#s2} ======= During the initial phase of research into the etiology of SARS, an unknown virus was cultured in Vero cells from a patient suffering from SARS ([@pbio.0000002-Ksiazek1]). Total nucleic acid purified from this viral culture, as well as a control culture, was obtained from the Centers for Disease Control and Prevention on March 22, 2003. These two samples, along with additional controls (HeLa cell RNA and water alone), were amplified and hybridized within 24 h to the virus DNA microarray. The strongest hybridizing array elements from the infected culture were derived from two families: astroviridae and coronaviridae. [Table 1](#pbio.0000002-t001){ref-type="table"} lists the oligonucleotides from these families with the greatest hybridization intensity. By comparison, these oligonucleotides yielded essentially background levels of hybridization in the various control arrays performed in parallel. The initial suggestion from this hybridization pattern was that members of both of these viral families might be present. However, alignment of the oligonucleotides using ClustalX revealed that all four hybridizing oligonucleotides from the astroviridae and one oligonucleotide from avian infectious bronchitis virus (IBV) (GenBank NC\_001451), an avian coronavirus, shared a core consensus motif spanning 33 nucleotides (data not shown); thus, these five oligonucleotides behaved essentially as multiple redundant probes for the same sequence. This motif is known to be present in the 3′ UTR of all astroviruses and the avian coronaviruses ([@pbio.0000002-Jonassen1]), but appears to be absent in the available sequenced mammalian coronaviruses (bovine coronavirus, murine hepatitis virus \[MHV\], human coronavirus 229E, porcine epidemic diarrhea virus, and transmissible gastroenteritis virus). The other three hybridizing oligonucleotides were derived from three conserved regions within the ORF1AB polyprotein common to all coronaviruses ([Figure 1](#pbio.0000002-g001){ref-type="fig"}). Based on the aggregate hybridization pattern, the virus appeared to be a novel member of the coronavirus family. ::: {#pbio.0000002-g001 .fig} Figure 1 ::: {.caption} ###### Prototypical Coronavirus Genome Structure Red bars indicate physical location of virus microarray DNA elements mapped to a generic coronavirus genome. Portions of the coronavirus genome sequenced by physical recovery and PCR methods are highlighted with homologies to known coronaviruses. Abbreviations: aa, amino acid; nt, nucleotide :::  ::: ::: {#pbio.0000002-t001 .table-wrap} Table 1 ::: {.caption} ###### Oligonucleotides Hybridizing to Viral Sample :::  Underlined nucleotides represent regions of identity to the SARS coronavirus. Does not include reverse complement oligos ^a^ BLAST identities to the SARS coronavirus genome (NC\_004718) ::: To further characterize this virus, we sequenced fragments of the viral genome using two complementary approaches. First, BLAST alignment of two of the hybridizing viral oligonucleotides, one each from bovine coronavirus and human coronavirus 229E, to the IBV genome indicated that the oligonucleotides possessed homology to distinct conserved regions within the *NSP11* gene (BLAST identity matches of 42/47 and 26/27, respectively). A pair of PCR primers was designed to amplify the intervening sequences between the two conserved regions, and a fragment that possessed 89% identity over 37 amino acids to MHV, a murine coronavirus, was obtained ([Figure 1](#pbio.0000002-g001){ref-type="fig"}; sequence available as [Data S1](#sd001){ref-type="supplementary-material"}). In a parallel approach, we directly recovered hybridized viral sequences from the surface of the microarray. This procedure took advantage of the physical separation achieved during microarray hybridization, which effectively purified the viral nucleic acid from other nucleic acid species present in the sample**.** Using a tungsten needle, the DNA microarray spot corresponding to the conserved 3′ UTR motif was repeatedly scraped and the hybridized nucleic acid was recovered. This material was subsequently amplified, cloned, and sequenced ([Figure 2](#pbio.0000002-g002){ref-type="fig"}). The largest clone spanned almost 1.1 kb; this fragment encompassed the 3′ UTR conserved motif and extended into the most 3′ coding region of the viral genome. BLAST analysis revealed 33% identity over 157 amino acids to MHV nucleocapsid, thus confirming the presence of a novel coronavirus (see [Figure 1](#pbio.0000002-g001){ref-type="fig"}; see [Data S1](#sd001){ref-type="supplementary-material"}). We subsequently confirmed results obtained from both strategies described above by using a random-primed RT-PCR shotgun sequencing approach that generated contigs totaling approximately 25 kb of viral genome sequence (see [Data S1](#sd001){ref-type="supplementary-material"}). ::: {#pbio.0000002-g002 .fig} Figure 2 ::: {.caption} ###### Viral DNA Recovery and Sequencing Scheme Hybridized viral sequences were physically scraped from a DNA microarray spot, amplified, cloned, and subsequently sequenced. :::  ::: Discussion {#s3} ========== In this report, we have demonstrated the viability of detecting novel pathogens via cross-hybridization to highly conserved sequence motifs. With the recent sequencing of the complete SARS coronavirus genome (GenBank NC\_004718) ([@pbio.0000002-Marra1]; [@pbio.0000002-Rota1]), we were able to retrospectively determine the degree of nucleotide identity shared between the hybridizing oligonucleotides and the new coronavirus genome (see [Table 1](#pbio.0000002-t001){ref-type="table"}). Stretches of relatively uninterrupted nucleotide identity as short as 25 nucleotides yielded clearly detectable hybridization signal, confirming that novel viruses with only limited homology to known viruses can be successfully detected by this strategy. A key feature of this approach is that direct recovery of hybridized material from the microarray provides a rapid route for obtaining sequences of novel viruses. By contrast, conventional strategies for subsequent sequence identification would require time-consuming steps such as library screening or additional rounds of PCR primer design and synthesis. In the case of SARS, we were able to ascertain within 24 h that a novel coronavirus was present in the unknown sample, and partial genome sequences of this virus were obtained over the next few days without the need for specific primer design. To our knowledge, this is the first demonstration of the feasibility and utility of directly recovering nucleic acid sequences from a hybridized DNA microarray. In light of the continuous threat of emerging infectious diseases, this overall approach will greatly facilitate the rapid identification and characterization of novel viruses. Materials and Methods {#s4} ===================== {#s4a} ### Nucleic acid isolation {#s4a1} Total nucleic acid was purified using the automated NucliSens extraction system (BioMerieux, Durham, North Carolina). Following the manufacturer\'s instructions, 100 μl of each specimen was added to tubes containing 900 μl of prewarmed NucliSens lysis buffer and incubated at 37°C for 30 min with intermittent mixing. Fifty microliters of silica suspension provided in the extraction kit was added to each tube and mixed. The mixtures were then transferred to a nucleic acid extraction cartridge and loaded onto the extractor workstation for processing. Approximately 50 μl of total nucleic acid eluate was recovered. ### Amplification {#s4a2} For the culture supernatants, 450 ng of nucleic acid was used as input for the amplification protocol. In parallel, 50 ng of HeLa cell RNA was used as a positive amplification control and water was used for a negative control. Samples were amplified using a random-primer protocol as described by [@pbio.0000002-Wang1]), with the following modifications: first- and second-strand synthesis were primed using primer-A (5′-GTTTCCCAGTCACGATCNNNNNNNNN) followed by PCR amplification using primer-B (5′-GTTTCCCAGTCACGATC) for 40 cycles. Aminoallyl-dUTP was incorporated into the PCR product using an additional 20 cycles of thermocycling. A detailed protocol is available as [Protocol S1](#sd002){ref-type="supplementary-material"}. ### Microarray hybridization and analysis {#s4a3} DNA microarrays were printed and hybridized essentially as described by [@pbio.0000002-Wang1]), with the following modifications: for array printing, a single-defined 70mer (spike-70) was mixed with each viral oligonucleotide in a 1:50 ratio. Array hybridizations used Cy5-labeled amplified probe from either virally infected cultures or controls (mock-infected culture, HeLa RNA, or water); a reference signal for every spot on each array was generated by using a Cy3-labeled version of the reverse complement of spike-70. Oligonucleotides were assessed by Cy5 intensity. Oligonucleotides from the astrovirus and coronavirus families that passed a conservative, arbitrarily set cutoff of (Cy5~infection~-Cy5~mock~) \> 1500 intensity units are listed in [Table 1](#pbio.0000002-t001){ref-type="table"}. Additional oligonucleotides from these families and their homology to the SARS coronavirus are listed in [Table S1](#st001){ref-type="supplementary-material"}. Array data has been deposited in the Gene Expression Omnibus (GEO) database (accession number GSE546). A complete list of the viral oligonucleotide sequences on the microarray is also available as [Table S2](#st002){ref-type="supplementary-material"}. ### Conventional PCR using array element sequences {#s4a4} PCR primers were designed by aligning the hybridizing oligonucleotides (Oligo IDs 15081544\_766 and 12175745\_728) to the IBV genome (Fwd: 5′-TGTTTTGGAATTGTAATGTGGAT; Rev: 5′-TACAAACTACCTCCATTACAGCC) and selecting stretches of near-identity. Primer-B-amplified material was used as the template for 35 cycles of thermocycling using the following program: 94°C for 30 s, 56°C for 30 s, and 72°C for 60 s. ### Direct sequence recovery from the microarray {#s4a5} Amplified viral sequences hybridized to individual microarray spots were recovered by scraping a 100 μm area of the microarray using a tungsten wire probe (Omega Engineering, Inc.) mounted on a micromanipulator while visualized by fluorescence microscopy (Nikon TE300). Recovered material was PCR amplified using primer-B, cloned into pCR2.1TOPO (Invitrogen), and sequenced. A detailed protocol is available as [Protocol S2](#sd003){ref-type="supplementary-material"}. ### Shotgun sequencing {#s4a6} Primer-B-amplified nucleic acid (see above) was cloned in pCR2.1TOPO, plated on 2xYT/kan plates, and grown overnight at 37°C. White colonies were picked into 384-well plates containing 2xYT/kan plus 8% glycerol and incubated overnight at 37°C. DNA was purified by magnetic bead isolation. DNA sequencing involved adding 3 μl of water to each bead pellet, followed by 3 μl of Big Dye terminator (v3.1) sequencing cocktail, and incubation for 35 cycles of 95°C for 5 s, 50°C for 5 s, and 60°C for 2 min. Reaction products were ethanol precipitated, resuspended in 25 μl of water, and loaded onto the ABI 3730xl sequencer. The resulting sequence reads were trimmed to remove primer sequences from the RT-PCR step and then assembled by Phrap (P. Green, unpublished data). Resulting contigs were screened by blast to remove any contigs with high human or monkey sequence similarity. The remaining contigs were edited to high quality, making any obvious joins. (Sequences are available as [Data S1](#sd001){ref-type="supplementary-material"}.) Supporting Information {#s5} ====================== Data S1 ::: {.caption} ###### Supporting Data (91.5 KB DOC) ::: ::: {.caption} ###### Click here for additional data file. ::: Protocol S1 ::: {.caption} ###### Supporting Protocol (28 KB DOC) ::: ::: {.caption} ###### Click here for additional data file. ::: Protocol S2 ::: {.caption} ###### Supporting Protocol (39.5 KB DOC) ::: ::: {.caption} ###### Click here for additional data file. ::: Table S1 ::: {.caption} ###### Supporting Table (97 KB DOC) ::: ::: {.caption} ###### Click here for additional data file. ::: Table S2 ::: {.caption} ###### Supporting Table (2.2 MB XLS) ::: ::: {.caption} ###### Click here for additional data file. ::: Accession Numbers {#s5a} ----------------- The Gene Expression Omnibus accession number for the array sequence is GSE546. This work was supported by a grant from the Sandler Program for Asthma Research (to JLD). **Conflicts of Interest.** The authors have declared that no conflicts of interest exist. **Author Contributions.** DW, DG, and JLD conceived and designed the experiments. DW, AU, Y-TL, MS, MH, VM, and JLD performed the experiments. DW, AU, JE, JPL, and JLD analyzed the data. DW, AU, DDE, TGK, ERM, RKW, and JLD contributed reagents/materials/analysis tools. DW and JLD wrote the paper. **Note Added in Proof**  The correct sequences for Primer A and Primer B are 5′-GTTTCCCAGTCACGATANNNNNNNNN and 5′-GTTTCCCAGTCACGATA, respectively. Academic Editor: Herbert Virgin, Washington University School of Medicine IBV : avian infectious bronchitis virus MHV : murine hepatitis virus SARS : severe acute respiratory syndrome

PubMed Central

2024-06-05T03:55:45.985560

2003-11-17

PMC300675

Iron is an essential nutrient for sustaining life-forms as diverse as plankton and humans. But too much iron, or too little, can spell trouble. Mammalian cells maintain the proper balance partly with the help of a specialized cell surface protein called the transferrin receptor (TfR). TfRs bind to the iron-carrying transferrin protein (Fe-Tf) and escort their cargo to the cell\'s interior. (To learn more about iron metabolism, see the primer by Tracey A. Rouault in this issue \[DOI: [10.3171/journal.pbio.0000079](10.3171/journal.pbio.0000079)\].) This receptor also binds the hereditary hemochromatosis protein (HFE), which is mutated in individuals who have the common iron-overload disorder hereditary hemochromatosis. While the molecular pathway that mediates cellular intake of iron through the TfR is known, it was not clear just how TfR assists in iron release to the cell and how it binds HFE and transferrin. By introducing multiple mutations in human TfRs, Pamela Bjorkman and colleagues identified functional binding sites for transferrin in both its iron-loaded and iron-free (apo-Tf) forms and for HFE. From these data, the researchers mapped out a scenario of the dynamic interactions between receptor and ligands (the bound molecule) and worked out a structure-based model for the mechanism of TfR-assisted iron release from Fe-Tf. Bjorkman\'s lab, which had previously solved the structures of both HFE and HFE bound to the TfR, used their structural information to investigate how the proteins interact, which amino acid residues are required for binding, whether the two ligands bind differently to the receptor, and how HFE binding affects transferrin binding. They found that Fe-Tf and HFE occupy the same or an overlapping site on the receptor, but since transferrin is much larger than the HFE protein, it appeared that transferrin could also interact with other parts of TfR. And it remained to be seen whether TfR discriminated between the iron-loaded and iron-free states of transferrin. In this study, Bjorkman and colleagues expanded their library of TfR mutants to clarify the transferrin binding signature on TfR and to see how the TfR mutations affect the way apo-Tf and Fe-Tf interact with the receptor. They characterized the binding affinities of 30 TfR mutants to HFE and Fe-Tf and to apo-Tf, and they report that mutations in 11 of the TfR residues interfere with either one or both forms of transferrin. Four of these residues are essential for transferrin binding and are conserved in all known TfR DNA sequences. Since residues that didn\'t have much impact are not conserved, the scientists say the results are likely to describe transferrin--TfR interactions for other species as well. As expected, the most critical residues required for transferrin binding fall within the receptor\'s helical domain and have significant physical overlap with residues required for HFE binding; though some residues that are required for apo-Tf binding do not affect Fe-Tf binding. Bjorkman et al. also identify additional residues in another domain on TfR (called the protease-like domain) that support Fe-Tf but not apo-Tf binding, confirming that the receptor binding footprints of the two metal-binding states of transferrin are indeed different. With a structural model showing *where* Fe-Tf and apo-Tf bind to the receptor, they could evaluate *how* they bind and thus explain how the receptor mediates iron release. By suggesting a mechanism through which TfR binding regulates iron release, this structural model of the transferrin--TfR complex will bolster efforts to elucidate the molecular details of this process. Confirmation that transferrin and HFE do indeed compete for docking privileges reveals a possible role for HFE in maintaining iron homeostasis and will provide valuable insights into the dysregulation that leads to the warehousing of iron and resulting tissue and organ damage associated with hemochromatosis. ::: {#pbio.0000029-g001 .fig} ::: {.caption} ###### Ribbon diagram of transferrin receptor homodimer :::  :::

PubMed Central

2024-06-05T03:55:45.987417

2003-12-22

PMC300700

One of the variants associated with increased diabetes risk was incorrectly indicated throughout this article. The A1369S variant in the gene ABCCB should have been written S1369A. The alanine variant is associated with increased risk. This mistake affects Tables 2 and 4, the text of the article in the section entitled \"*ABCCB* and *KCNJ11*\" on page 45, and the Supporting Information Tables S1 and S2. The full text XML and HTML versions of the article, and the supporting Tables S1 and S2 have been corrected online.

PubMed Central

2024-06-05T03:55:45.988154

2003-12-22

PMC300881

Introduction {#s1} ============ Many cellular proteins are degraded by the proteasome after they become covalently modified with a multiubiquitin chain. The 26S proteasome is a massive protein composed of a 20S core and two 19S regulatory particles ([@pbio-0020002-Voges1]). The 20S core can be subdivided into a dimer of heptameric rings of β subunits---which contain the proteolytic active sites responsible for the protein degradation activity of the proteasome---flanked by heptameric rings of α subunits. The 19S regulatory particle can be divided into a base thought to comprise a hexameric ring of AAA ATPases and a lid composed of eight or more distinct subunits. Whereas 20S core particles and AAA ATPase rings have been found in compartmentalized proteases in prokaryotes, the lid domain of the 19S regulatory particle is unique to eukaryotes and provides the specificity of 26S proteasomes for ubiquitinated substrates ([@pbio-0020002-Glickman1]). Ubiquitin (Ub), an 8 kD protein, is conjugated by Ub ligases to proteasome substrates via an isopeptide bond that links its carboxyl terminus to the amino sidechain of a lysine residue in the substrate. Ub-like proteins (Ubls), of which there are several, are conjugated to their target proteins in a similar manner. Ubls typically do not promote degradation of their targets by the proteasome, but rather regulate target activity in a more subtle manner reminiscent of protein phosphorylation ([@pbio-0020002-Hershko1]; [@pbio-0020002-Peters1]). As is the case for protein phosphorylation, the attachment of Ub and Ubls to target proteins is opposed by isopeptidase enzymes that undo the handiwork of Ub ligases. For example, removal of the Ubl Nedd8 (neural precursor cell expressed, developmentally downregulated 8) regulatory modification from the Cullin 1 (Cul1) subunit of the SCF (Skp1/Cdc53/Cullin/F-box receptor) Ub ligase is catalyzed by the COP9 signalosome (CSN) ([@pbio-0020002-Lyapina1]). The CSN was identified in Arabidopsis thaliana from genetic studies of constitutively photomorphogenic mutant plants ([@pbio-0020002-Osterlund1]). It later became evident that CSN and the proteasome lid are paralogous complexes ([@pbio-0020002-Glickman1]; [@pbio-0020002-Seeger1]; [@pbio-0020002-Wei1]). Csn5 of CSN and Rpn11 (regulatory particle number 11) of the proteasome lid are the subunits that are most closely related between the two complexes. CSN-dependent isopeptidase activity is sensitive to metal ion chelators, and Csn5 contains a conserved, putative metal-binding motif (EX~n~HS/THX~7~SXXD), referred to as the JAMM motif, that is embedded within the larger JAB1/MPN/Mov34 domain (hereafter referred to as the JAMM domain) and is critical for Csn5 function in vivo ([@pbio-0020002-Cope1]). Removal of Ub from proteasome substrates is also promoted by a metal ion-dependent isopeptidase activity associated with the proteasome ([@pbio-0020002-Verma1]; [@pbio-0020002-Yao1]). The JAMM/MPN^+^ motif of Rpn11 is critical for its function in vivo ([@pbio-0020002-Maytal-Kivity1]; [@pbio-0020002-Verma1]; [@pbio-0020002-Yao1]), and proteasomes that contain Rpn11 bearing a mutated JAMM motif are unable to promote deubiquitination and degradation of the proteasome substrate Sic1 ([@pbio-0020002-Verma1]). Taken together, these observations suggested that the JAMM motif specifies a catalytic center that in turn defines a novel family of metalloisopeptidases. Interestingly, the JAMM motif is found in proteins from all three domains of life ([@pbio-0020002-Cope1]; [@pbio-0020002-Maytal-Kivity1]), indicating that it has functions beyond the Ub system. In this study, we present the crystal structure of the *Archaeoglobus fulgidus AF2198* gene product AfJAMM and explore the implications of its novel metalloprotease architecture. Results and Discussion {#s2} ====================== We proposed that the subset of JAMM domain proteins that contain a JAMM motif comprise a novel family of metallopeptidases ([@pbio-0020002-Cope1]). To gain a clearer understanding of these putative enzymes---in particular the pertinent subunits of the proteasome lid and signalosome ([Figure 1](#pbio-0020002-g001){ref-type="fig"})---we cloned and expressed in Escherichia coli a variety of JAMM motif-containing proteins to find a suitable candidate for crystallographic analysis. The expression of all candidates except for AfJAMM led to insoluble aggregates. Unlike many JAMM proteins that contain an additional domain, the AfJAMM protein consists entirely of the JAMM domain. We were able to purify and crystallize native and selenomethionine-substituted AfJAMM; the latter was used for phasing by employing the multiwavelength anomalous diffraction (MAD) technique (see [Table 1](#pbio-0020002-t001){ref-type="table"} for statistics). ::: {#pbio-0020002-g001 .fig} Figure 1 ::: {.caption} ###### Alignment of Eukaryotic JAMM Domains with AfJAMM Eukaryotic JAMM domain proteins were aligned with AfJAMM using ClustalX and manually refined. Sequences are named with a two-letter code corresponding to the genus and species of the respective organism followed by the name of the protein (see Supporting Information for accession numbers), and 'hyp' is an abbreviation for hypothetical. The JAMM motif comprises the residues highlighted in green (E22, H67, H69, S77, and D80), and the active site core is surrounded by a red box. Conserved residues are highlighted in gray. The disulfide cysteine residues are highlighted in yellow (C74, C95). Active site residues that were mutated in S. pombe Csn5 are marked with an asterisk beneath the alignment. The secondary structure of AfJAMM is indicated above the sequence; helices are blue, sheets are red arrows, and loops are yellow lines. The dashed yellow line indicates a loop (F42--G58) that is disordered in the crystal. :::  ::: ::: {#pbio-0020002-t001 .table-wrap} Table 1 ::: {.caption} ###### Data Collection Statistics :::  ^a^Owing to pseudocentering, reflections with l values such that cos^2^ (0.54πl) \< ½ are systematically weak, leading to an R-factor higher than would be expected for a nonpseudocentered crystal structure. R~N~ are the R-factors calculated with only the reflections with cos^2^ (0.54πl) \> ½ (see Materials and Methods). Barring rearrangements of sidechains in the vicinity of the zinc atom, no significant changes were seen between the native and selenomethionine forms ::: AfJAMM consists of an eight-stranded β sheet (β1--β8), flanked by a long α helix (α1) between the first and second strand, and a short α helix (α2) between the fourth and fifth strand. This β sheet resembles a β barrel halved longitudinally and curled around α1 ([Figure 2](#pbio-0020002-g002){ref-type="fig"}A). The α2 helix is oriented lengthwise on the convex surface of the β sheet. The zinc-binding site is adjacent to a loop that spans the end of β4 to the beginning of α2 and is stabilized by a disulfide bond between C74 from this loop to C95 on β5. Although disulfide bonds are scarce in intracellular proteins, they are often present in homologous proteins found in hyperthermophiles ([@pbio-0020002-Mallick1]). The overall fold resembles that of the zinc metalloenzyme cytidine deaminase (CDA). CDA from Bacillus subtilis ([@pbio-0020002-Johansson1]) can be superimposed onto AfJAMM with a root-mean squared (RMS) deviation of 3.0 Å over 79 α carbons, despite only 9% sequence identity over structurally aligned residues. The catalytic zinc ions of AfJAMM and CDA, 4.9 Å apart in the superposition, occupy the same general vicinity in the tertiary structures but are coordinated by entirely different protein ligands, two histidines and an aspartic acid in AfJAMM compared to three cysteines in CDA, located at different positions in the sequence ([Figure 2](#pbio-0020002-g002){ref-type="fig"}A). Consequently, the JAMM fold represents a departure from the papain-like cysteine protease architecture that underlies the deubiquitinating activity of the most thoroughly characterized deubiquitinating enzymes (DUBs), the Ub carboxy-terminal hydrolases (UCHs) ([@pbio-0020002-Johnston1]) and Ub-specific proteases (UBPs) ([@pbio-0020002-Hu1]). ::: {#pbio-0020002-g002 .fig} Figure 2 ::: {.caption} ###### Crystal Structure of AfJAMM \(A) On the left, the AfJAMM protomer is presented. The amino and carboxyl termini are marked by N and C. The catalytic zinc atom is depicted as a gray sphere. The zinc ligands (H67, H69, and D80) are colored in green. Secondary structure elements are numbered α1--α2 and β1--β8. The amino acids that mark the beginning and end of the disordered loop (P41--M60) are labeled. On the right, the crystal structure of the cytidine deaminase protomer is shown in the same orientation as AfJAMM to highlight the fold likeness as well as the similarly situated zinc-binding sites. The zinc ligands (C53, C86, and C89) are colored in green. \(B) The dimer in the asymmetric unit of AfJAMM crystals. The side view is obtained by rotating the monomer in (A) by 90° as indicated by the quarter-arrow around the y-axis. The gold protomer is related to the green protomer by a 180° rotation around the crystallographic c-axis (shown as a black bar in the side view) and a translation of 3.38 Å. :::  ::: The two AfJAMM subunits in the asymmetric unit are connected through a parallel β sheet formed at the dimer interface ([Figure 2](#pbio-0020002-g002){ref-type="fig"}B). The subunits are related by a 2-fold screw axis along the crystallographic c-axis with a translation of 3.38 Å, corresponding to a displacement of one residue along the β3 strand. AfJAMM behaves as a monomer during size exclusion chromatography, suggesting that the dimer observed in the asymmetric unit is an artifact of crystallization. Yet the residues of β3 are highly conserved among JAMM proteins (see [Figure 1](#pbio-0020002-g001){ref-type="fig"}) and predominantly hydrophobic, which makes it difficult to regard the observed interaction as completely insignificant. Flanking β3 to the carboxy-terminal side, there is a striking covariation of residues, MPQSGTG in Rpn11 orthologues and LPVEGTE in Csn5 orthologues. The potential of β3 and the flanking region to mediate specific protein--protein interactions, such as the assembly of Rpn11 and Csn5 into their respective complexes or their specificity towards Ub or Nedd8, warrants further investigation. The zinc-binding site of AfJAMM is located in a furrow formed by the convex surface of the β2--β4 sheet and α2. The catalytic zinc has a tetrahedral coordination sphere ([Figure 3](#pbio-0020002-g003){ref-type="fig"}A), with ligands provided by N^ɛ2^ of H67 and H69 on β4, the carboxylate of D80 on α2, and a water molecule. The latter hydrogen-bonds to the sidechain of E22 on β2. Thus, the crystal structure confirms previous predictions that the histidine and aspartic acid residues in the JAMM motif are ligands for a metal ([@pbio-0020002-Cope1]; [@pbio-0020002-Verma1]; [@pbio-0020002-Yao1]). It must be noted that the identity of the physiological metal in AfJAMM and eukaryotic JAMM homologues is still unknown. The majority of metalloproteases naturally employ zinc but show altered activities with other substituted metals ([@pbio-0020002-Auld1]). ::: {#pbio-0020002-g003 .fig} Figure 3 ::: {.caption} ###### Metalloprotease-Like Active Site of AfJAMM \(A) The active site of AfJAMM is shown centered around the catalytic zinc ion, which is represented as a dark gray sphere surrounded by anomalous cross Fourier difference density (contoured at 9.5 σ) colored in red. The aqua ligand, which lies at 2.9 Å from the zinc, is shown as a red sphere surrounded by purple density (contoured at 3 σ) of an F~obs~ -- F~calc~ map, in which the aqua ligand was omitted from the calculation. Residues that underlie isopeptide bond cleavage are shown in green. The carboxylate oxygen atoms of D80 lie 2.2 Å from the zinc. The N^ɛ2^ atoms of H67 and H69 lie 2.1 Å from the zinc. The carboxylate oxygen atoms of E22 lie 3.2--3.5 Å from the aqua ligand and 4.5--5.0 Å from the zinc. Ancillary active site residues and the backbone (ribbon diagram) are shown in grey. The disulfide bond that links C74 to C95 is shown in yellow. The JAMM motif is shown in the upper lefthand corner for reference. \(B) Superimposition of active site residues in ScNP, thermolysin, and AfJAMM. AfJAMM is in green, ScNP in blue, and thermolysin in red. For clarity only, the sidechains from the residues that bind the zinc or aqua ligands are shown in their entirety. In addition, atoms that stabilize the putative tetrahedral intermediate are shown. These include O^γ^ of S77 in AfJAMM, O^η^ of Y95 in ScNP, and the N^ɛ2^ of H231 in thermolysin. :::  ::: The arrangement of zinc ligands in AfJAMM resembles that found in thermolysin, the Streptomyces caespitosus zinc endoprotease (ScNP), and neurolysin, a mammalian metalloprotease ([@pbio-0020002-Kurisu1]; [@pbio-0020002-Brown1]; [@pbio-0020002-English1]). Thermolysin, neurolysin, and ScNP are homologues that have the classical HEXXH metalloprotease motif and adopt the same core fold. In contrast, the sequence, zinc-binding motif, and fold adopted by AfJAMM are entirely distinct. Nonetheless, the active site metal and ligand atoms of thermolysin and ScNP can be superimposed on those of AfJAMM with an RMS deviation of approximately 0.4--0.5 Å ([Figure 3](#pbio-0020002-g003){ref-type="fig"}B). While this manuscript was under revision, an independent report of a crystal structure of the *AF2198* gene product appeared ([@pbio-0020002-Tran1]). These authors used the fold similarity to CDA as a framework to evaluate the function of the JAMM motif. Given the biochemical data supporting the JAMM motif\'s role in proteolysis, the common active site architecture seen in AfJAMM and thermolysin, and the similarity of zinc ligands between thermolysin and AfJAMM, we believe that the extensive body of mechanistic studies on thermolysin and related metalloproteases provide a better framework for the analysis of JAMM function than CDA. In addition to the correspondence between zinc ligands, the glutamic acid residue (E166) downstream of the HEXXH motif of thermolysin is functionally equivalent to the aspartic acid ligand of AfJAMM (D80). E22 in AfJAMM is functionally equivalent to the glutamic acid in thermolysin\'s HEXXH motif, which serves as the general acid-base catalyst. The conserved serine between the histidine ligands interacts with E22 through a sidechain--main chain hydrogen bond. In more distant JAMM relatives, the serine is replaced by a threonine or asparagine ([@pbio-0020002-Aravind1]), both of which are capable of the same bracing function. Meanwhile, the γ-hydroxyl group of the highly conserved S77 in AfJAMM occupies a position similar to N^ɛ2^ of H231 in thermolysin. This atom flanks the 'oxyanion hole' and is implicated in stabilizing the tetrahedral intermediate formed during hydrolysis of the scissile bond ([@pbio-0020002-Matthews1]; [@pbio-0020002-Lipscomb1]). AfJAMM was tested for the ability to hydrolyze a number of substrates, including Ub derivatives, resofurin-labeled casein, and D-alanine compounds. Unfortunately, none of the in vitro assays yielded positive results. As nothing is known about AfJAMM in the context of A. fulgidus biology, these negative results do not rule out the possibility that AfJAMM functions as a peptide hydrolase in vivo. To validate the suitability of the AfJAMM structure as a basic model for eukaryotic JAMM proteins, we performed site-directed mutagenesis of *Schizosaccharomyces pombe csn5^+^*. The zinc ligands of Csn5 were previously established to be essential for its role in sustaining cleavage of the isopeptide bond that links Nedd8 to Cul1 ([@pbio-0020002-Cope1]). Alanine substitutions for the putative general acid-base catalyst (E56A) and the catalytic serine (S128) in the JAMM motif of Csn5 likewise abolished its ability to remove the Nedd8 moiety from Cul1 in a *csn5* ^+^ background ([Figure 4](#pbio-0020002-g004){ref-type="fig"}A). The E56A mutation had no effect on the assembly of Csn5 with Csn1*^myc13^*, while assembly with S128A was slightly hindered ([Figure 4](#pbio-0020002-g004){ref-type="fig"}A). Mutation of the equivalent serine codon in *RPN11* destroyed complementing activity without altering assembly of Rpn11 into the lid. However, the effect of this mutation on Rpn11 isopeptidase activity was not evaluated ([@pbio-0020002-Maytal-Kivity1]). Alanine substitutions for a catalytic residue (E56) or zinc ligands (H118A, D131N) exerted a modest dominant-negative phenotype in *csn5^+^* cells ([Figure 4](#pbio-0020002-g004){ref-type="fig"}B). ::: {#pbio-0020002-g004 .fig} Figure 4 ::: {.caption} ###### Mutations in the JAMM Motif of Csn5 Abrogate the Deneddylating Activity of the CSN \(A) Mutations in the glutamic acid (E56A) that positions the aqua ligand and in the proposed catalytic serine (S128A) of Csn5 disrupt deneddylation of Cul1 by CSN but have no effect on assembly with Csn1. A *csn5Δ* strain of S. pombe was transformed with an empty pREP-41 plasmid (lane 1) or with the plasmid encoding FLAG tagged: Csn5 (lane 2), Csn5^E56A^ (lane 3), or Csn5^S128A^ (lane 4). Whole-cell lysates were used for Western blot analysis with anti-Cul1 antibodies (top gel) and anti-FLAG antibodies (second from top). A strain with a *myc13*-tagged Csn1 was transformed with the above plasmids, and whole-cell lysates were used for Western blot analysis. Antibodies to the Myc tag were used to detect Csn1*^myc13^* (third from top), and were used to pull down Csn1*^myc13^* and subsequently blot with anti-FLAG antibodies to detect coprecipitated Csn5 mutant proteins (bottom gel). \(B) Mutations in the JAMM motif display a modest dominant-negative phenotype. Western blot analysis of crude cell lysates was performed as described in (A). \(C) Selected JAMM motifs from proteins of diverse functions. The canonical JAMM motif residues are highlighted in green. The conserved proline is highlighted in blue, and semiconserved cysteine is highlighted in yellow. :::  ::: We have been able to assign biochemical functions to Csn5 and Rpn11 ([@pbio-0020002-Cope1]; [@pbio-0020002-Verma1]; [@pbio-0020002-Yao1]), but the functions of other eukaryotic JAMM proteins ([Figure 4](#pbio-0020002-g004){ref-type="fig"}C) such as AMSH and C6.1A, as well as the prokaryotic protein RadC and the viral phage λ tail assembly protein K, remain unknown. The structure of AfJAMM provides a useful tool for dissecting the functions of JAMM motifs in these varied contexts and inspires the search for specific JAMM active site inhibitors. The mechanistic implications of the AfJAMM structure explain why the deubiquitinating activity of the lid was unaffected by inhibitors of classical DUBs, the UCHs and UBPs. In classical DUBs, the nucleophile that attacks the carbon of the scissile bond is provided by a cysteine residue in the active site. This property is exploited by using the irreversible inhibitor Ub--aldehyde, which forms a nonhydrolyzable bond to the nucleophilic cysteine ([@pbio-0020002-Johnston2]). In contrast, JAMM proteins likely hydrolyze Ub conjugates in a manner similar to thermolysin, in which the zinc-polarized aqua ligand serves as the nucleophile ([@pbio-0020002-Lipscomb1]). In the case of thermolysin, metal chelators and phosphonamidate peptides are effective inhibitors ([@pbio-0020002-Bartlett1]), whereas other zinc metalloproteases are sensitive to peptidomimetic substrates bearing a hydroxamate group ([@pbio-0020002-Skiles1]). Metal chelators have been shown to be effective inhibitors of JAMM proteins ([@pbio-0020002-Cope1]; [@pbio-0020002-Verma1]); it would be interesting to see whether phosphonamidate and hydroxamate peptide mimics of Ub conjugate isopeptides would be equally effective. The proteasome inhibitor PS-341 has gained attention for its novelty and effectiveness in treating various forms of cancer ([@pbio-0020002-Adams1]). PS-341 was recently approved by the United States\' Food and Drug Administration for treatment of relapsed multiple myeloma, thereby validating the proteasome as a target for anticancer therapies. The active site of JAMM proteins is an intriguing target for second-generation therapeutics targeted at the Ub--proteasome pathway for two reasons: the JAMM motif in the proteasome lid is essential for the proteasome to function and the JAMM motif in the CSN specifically regulates the activity of a critical family of E3 Ub ligases ([@pbio-0020002-Nalepa1]). Inhibition of SCF and other Cullin-based ligases by way of the JAMM motif may be a more specific means of modulating levels of key proteasome substrates in cancer cells. Materials and Methods {#s3} ===================== The gene for A. fulgidus JAMM ([@pbio-0020002-Ponting1]), open reading frame *AF2198*, was cloned from genomic DNA (ATCC \#49558D; American Type Culture Collection, Manassas, Virginia, United States) into the pCRT7 vectors (Invitrogen, Carlsbad, California, United States). During cloning, the alternate start codon, GTG, was replaced with the canonical start codon, ATG. The construct was expressed in BL21(DE3)pLysS cells (Novagen, Madison, Wisconsin, United States). The cells were grown to midlog phase in terrific broth media and induced with 0.5 mM IPTG. The cells were lysed by sonication and the protein was isolated by immobilized metal ion chromatography using a Ni-NTA resin (Qiagen, Valencia, California, United States). The protein was further purified by gel filtration on a Sephacryl S100 column (Amersham Pharmacia Biotech, Chalfont St Giles, United Kingdom) and concentrated. The amino-terminal tag was removed by limited digestion with trypsin. Mass spectrometry analysis revealed that trypsin only cut AfJAMM in the amino-terminal tag region, and only a single band was evident on a Coomassie-stained polyacrylamide gel. The tag and uncut protein were removed with Ni-NTA resin followed by anion-exchange chromatography with SOURCE 30Q resin (Amersham Pharmacia Biotech). The processed protein was then concentrated to approximately 30 mg/ml by ultrafiltration. The selenomethionine protein was produced as described elsewhere ([@pbio-0020002-Van1]) and purified using the same protocol as for the native protein. Protein crystals were obtained in 100 mM NH~4~H~2~PO~4~, 200 mM sodium citrate (pH 5) using vapor diffusion with sitting drops and hanging drops. Crystals were incubated for approximately 1 min in a cryo-solution of equal volumes of reservoir solution and 35% meso-erythritol for the selenomethionine crystals and supplemented with 5 mM ZnCl~2~ for the native crystals. The crystals belonged to the space group P6~5~, with cell dimensions of a = b = 76 Å, c = 94 Å and two subunits per asymmetric unit. Data for the selenomethionine crystals were collected on Beamline 9.2 at the Stanford Synchrotron Radiation Laboratory (SSRL) (Stanford, California, United States) and data for the native crystals were collected on Beamline 8.2.1 at the Advanced Light Source (ALS) (Lawrence Berkeley National Laboratory, Berkeley, California, United States) (see [Table 1](#pbio-0020002-t001){ref-type="table"}). Phases were obtained by the MAD technique using data collected from selenomethionine-substituted crystals (see [Table 1](#pbio-0020002-t001){ref-type="table"}). Three Se atoms were located by SOLVE ([@pbio-0020002-Terwilliger2]) and used to calculate the initial phases. Phasing was subsequently improved by noncrystallographic symmetry averaging, using operators derived from the Se positions, and solvent flattening in RESOLVE ([@pbio-0020002-Terwilliger1]). The polypeptide model was built in O ([@pbio-0020002-Jones1]) and refined with CNS ([@pbio-0020002-Brunger1]). Since two monomers in the unit cell are related by a fractional translation along c of approximately 0.54, the intensities of the diffraction pattern are modulated by a factor of cos^2^ (0.54πl). As a result, reflections with l-indices such that cos^2^ (0.54πl) \< ½ are systematically weak, leading to an R-factor higher than would be expected for a nonpseudocentered crystal structure. However, when only the reflections with cos^2^ (0.54πl) \> ½ (which will have a more normal intensity distribution) are used for the R-factor calculation, reasonable values for R are obtained. The geometry of the final model was analyzed with PROCHECK ([@pbio-0020002-Morris1]). The Ramachandran plot shows 98.9% of the residues in the allowed regions and 1.1% in the disallowed regions. The main chain of K66, which constitutes the residue in the disallowed region, was modeled on segments taken from well-refined, high-resolution structures. The Protein Data Bank was searched for structural neighbors of AfJAMM using the DALI server ([@pbio-0020002-Holm1]). The superpositions with cytidine deaminase (1JTK), thermolysin (1FJQ), and ScNP (1C7K) were done using the LSQKAB program of the CCP4 distribution (CCP4 1994). All structural figures were made with PyMOL ([@pbio-0020002-DeLano1]).The experiments with S. pombe were performed as previously described by [@pbio-0020002-Cope1]). Supporting Information {#s4} ====================== Accession Numbers {#s4a} ----------------- The accession numbers for the proteins discussed in this paper are 20S proteasomes (PDB ID 1RYP), AfJAMM (Entrez Protein ID NP\_071023; PDB ID 1R5X), AMSH (Entrez Protein ID NP\_006454), AtCSN5/AJH1 (Entrez Protein ID NP\_173705), AtRpn11 (Entrez Protein ID NP\_197745), C6.1A (Entrez Protein ID NP\_077308), CeCSN5 (Entrez Protein ID NP\_500841), CeRpn11 (Entrez Protein ID NP\_494712), Csn5 (Entrez Protein ID NP\_593131), Cul1 (Entrez Protein ID NP\_594259), cytidine deaminase (PDB ID 1JTK), DmCsn5/CH5 (Entrez Protein ID NP\_477442), DmRpn11/p37b (Entrez Protein ID AAF08394), EcRadC (Entrez Protein ID NP\_418095), HsAMSH (Entrez Protein ID NP\_006454), HsC6.1A (Entrez Protein ID NP\_077308), HsCsn5 (Entrez Protein ID NP\_006828), HsRpn11/POH1 (Entrez Protein ID NP\_005796), JAB1 (Entrez Protein ID AAC17179), lambdaK (Entrez Protein ID AAA96551), Mov34 (Entrez Protein ID NP\_034947), Mpr1p (Entrez Protein ID AAN77865), Nedd8 (Swiss-Prot ID Q15843), neurolysin (PDB ID 1I1I), Pad1p (Entrez Protein ID NP\_594014), phage λ tail assembly protein K (Entrez Protein ID AAA96551), RadC (Entrez Protein ID NP\_418095), Rpn11 (Entrez Protein ID AAN77865), SCF (PDB ID 1LDK), ScNP (PDB ID 1C7K), ScRpn11 (Entrez Protein ID AAN77865), Sic1 (Entrez Protein ID 1360441), SpCsn5 (Entrez Protein ID NP\_593131), SpRpn11/Pad1 (Entrez Protein ID NP\_594014), thermolysin (PDB ID 1FJQ), ubiquitin (Swiss-Prot ID P04838), UBP (PDB ID 1NB8), and UCH (PDB ID 1UCH). These databases may be found at <http://www.ncbi.nlm.nih.gov/entrez/> (Entrez Protein), <http://www.rcsb.org/pdb/> (Protein Data Bank \[PDB\]), and <http://us.expasy.org/sprot/> (Swiss-Prot). This work was supported by the National Science Foundation Graduate Research Fellowship and the Gordon Moore Foundation (to XIA), as well as the Howard Hughes Medical Institute (to DCR and RJD). We would like to thank the staff at the Stanford Synchrotron Radiation Laboratory, a national user facility operated by Stanford University on behalf of the United States Department of Energy, Office of Basic Energy Sciences, and the Advanced Light Source, which is supported by the Director of the Office of Science, Office of Basic Energy Sciences, Materials Sciences Division of the United States Department of Energy under contract number DE-AC03-76SF00098 at Lawrence Berkeley National Laboratory. Special thanks go to R. Verma and G. Cope for assistance with the associated biochemistry, T. Yeates and O. Einsle for insights concerning the treatment of the pseudocentered crystals, K. Locher and P. Strop for helpful discussions, and J. Ambroggio for back massages and constant support. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** XIA, DCR, and RJD conceived and designed the experiments. XIA performed the experiments. XIA, DCR, and RJD analyzed the data. XIA, DCR, and RJD contributed reagents/materials/analysis tools. XIA wrote the paper. Academic Editor: Hidde L. Ploegh, Harvard Medical School AfJAMM : A. fulgidus JAMM protein AMSH : associated molecule with SH3 domain of STAM CDA : cytidine deaminase CSN : COP9 signalosome Cul1 : Cullin 1 DUB : deubiquitinating enzyme JAMM : JAB1/MPN/Mov34 metalloenzyme MAD : multiwavelength anomalous diffraction MPN : Mpr1p Pad1p N-terminal domain Nedd8 : neural precursor cell expressed RMS : root-mean squared Rpn11 : regulatory particle number 11 SCF : Skp1/Cdc53/Cullin/F-box receptor ScNP : S. caespitosus zinc endoprotease Ub : ubiquitin Ubls : ubiquitin-like proteins UBP : ubiquitin-specific protease UCH : ubiquitin C-terminal hydrolase

PubMed Central

2024-06-05T03:55:45.988446

2003-11-24

PMC300882

Introduction {#s1} ============ Microarray experiments are now being used to address a large diversity of biological issues. The large datasets obtained by pooling those experiments together contain a wealth of biological information beyond the insights gained by individual measurements. For example, it was demonstrated that diverse datasets of genome-wide expression profiles can be applied for facilitating functional assignment of uncharacterized ORFs and for identification of *cis*-regulatory elements ([@pbio-0020009-Eisen1]; [@pbio-0020009-Kim1]; [@pbio-0020009-Ihmels1]). Comparing the genomic sequences of different organisms presents an alternative prominent approach for gene annotation and identification of regulatory elements ([@pbio-0020009-Chervitz1]; [@pbio-0020009-Lynch1]; [@pbio-0020009-Rubin1]; [@pbio-0020009-Yanai1]; [@pbio-0020009-Frazer1]). Sequenced-based comparative analyses also proved crucial for deciphering evolutionary principles. As evolutionary changes frequently also involve modifications of the gene regulatory program ([@pbio-0020009-Carroll1]; [@pbio-0020009-True1]; [@pbio-0020009-Wray1]), integration of expression data into interspecies comparative analyses could potentially provide new insights into the relation between genomic sequence and organismal form and function. So far, however, such an approach has been mostly applied to small numbers of genes ([@pbio-0020009-Carroll1]; [@pbio-0020009-True1]; [@pbio-0020009-Wray1]) or has been restricted to variations in the genome-wide expression profiles during the development of closely related species ([@pbio-0020009-Rifkin1]). With the accumulation of large-scale expression data for a number of diverse species, the time may be ripe for a macro-evolutionary comparison of gene expression. Expression data differ from sequence data in two main aspects, which make their integration into comparative analysis challenging. First, unlike sequence information, which is direct and accurate, expression profiles provide only indirect and noisy information about the regulatory relationships between genes. Second, while the genomic sequence is essentially complete, expression profiles only cover a subset of all possible cellular conditions and thus provide only partial information about the underlying regulatory program. Moreover, this subset is typically very different for each organism, reflecting distinct physiologies as well as different research foci. One way to circumvent this problem is to restrict the data to a small subset of similar conditions, such as timepoints along the cell cycle ([@pbio-0020009-Alter1]). Such an approach, however, drastically reduces the size of the dataset and limits the scope of comparison. Here, we present a comparative analysis of large sets of expression data from six evolutionarily distant organisms ([Table 1](#pbio-0020009-t001){ref-type="table"}). We integrate the expression data with genomic sequence information to address three biological issues. First, we verify that coexpression is often conserved among organisms and propose a method for improving functional gene annotations using this conservation. We provide a Web-based application suitable for this purpose. Second, we compare the regulatory relationships between particular functional groups in the different organisms, giving initial insights into the extent of conservation of the gene regulatory architecture. Interestingly, we find that while functionally related genes are frequently coexpressed in several organisms, their organization and relative contribution to the overall expression program differ. Finally, we compare global topological properties of the transcription networks derived from the expression data, using a graph theoretical approach. This analysis reveals that despite the differences in the regulation of individual gene groups, the expression data of all organisms share large-scale properties. ::: {#pbio-0020009-t001 .table-wrap} Table 1 ::: {.caption} ###### Large-Scale Expression Data Used in This Study :::  Publicly available large-scale expression data were obtained from different sources (see Materials and Methods for references). We excluded genes or conditions with more than 90% missing datapoints, resulting in expression matrices of the dimensions shown. The data for Saccharomyces cerevisiae, Caenorhabditis elegans, and Escherichia coli are genome-wide, while those for Arabidopsis thaliana, Drosophila melanogaster, and Homo sapiens contain a large fraction of the respective genomes. The datasets comprise diverse experimental conditions, including environmental changes, time courses, tissues, and mutants. The *Drosophila* data consist of 75 timepoints during development ::: Results and Discussion {#s2} ====================== Combining Sequence and Expression Data for Improving Functional Gene Annotations {#s2a} -------------------------------------------------------------------------------- With the rapid increase in the number of sequenced genomes, assigning function to novel ORFs has become a major computational challenge. Functional links are often imputed based on sequence similarity with genes of known functions. Despite the large success of this approach, it has several well-recognized limitations. Foremost, an ORF can have several close homologues, some of which may be related to different functions. Furthermore, the sequence of an ORF may have diverged beyond recognition although the gene maintained its function. Gene expression analysis can provide functional links for new ORFs based on their coexpression with known genes. However, in this case, only links between genes of the same organism can be established. Moreover, owing to biological interference and the noise in the expression data, the inferred coexpression could be accidental and may not necessarily reflect similar function. Combining expression and sequence data may help to overcome the abovementioned limitations. Specifically, homologous genes whose function has been preserved are expected to be coregulated with genes related to that function. Conserved coexpression could thus distinguish them from homologues whose function diverged. This can be done, for example, by focusing on a group of functionally related genes in a characterized genome, identifying simultaneously all the respective homologues in a second genome, and then examining which of the homologues are indeed coexpressed ([Figure 1](#pbio-0020009-g001){ref-type="fig"}A). Importantly, restricting the search for coexpressed genes to a limited set of candidates provides an effective mean to overcome the noise in the expression data ([@pbio-0020009-Ihmels1]). ::: {#pbio-0020009-g001 .fig} Figure 1 ::: {.caption} ###### Using Expression Data to Identify and Refine Sequence-Based Functional Assignments \(A) Starting from a set of coexpressed genes (yellow dots in left box) associated with a particular function in organism A, we first identify the homologues in organism B using BLAST (middle box). Only some of these homologues are coexpressed while others are not (blue dots). The signature algorithm selects this coexpressed subset and adds further genes (light yellow) that were not identified based on sequence, but share similar expression profiles (right box). \(B) The 15 coexpressed genes associated with heat shock in yeast (center) have eight homologues in E. coli (left) and 14 in C. elegans (right). Among the ten genes whose expression profiles are the most similar to these homologues (bottom), many are known to be associated with heat-shock response (boldface). \(C) For each of the six organisms, the distribution of the *Z*-scores for the average gene--gene correlation of all the "homologue modules" ([see Materials and Methods](#s4){ref-type="sec"}) obtained from the yeast modules is shown (top). Rejecting the homologues that are not coexpressed gives rise to the "purified modules," whose *Z*-scores generally are larger (except for the yeast modules, which contain only coexpressed genes from the beginning). Adding further coexpressed genes yields the "refined modules," which have significantly larger *Z*-scores (bottom). :::  ::: ### Conserved coregulation of functionally related genes {#s2a1} To explore systematically the utility of this approach, we first examined to what extent coexpression is conserved among different organisms. We performed a statistical analysis comparing the pairwise correlations between genes in one organism to the correlations between their respective homologues. Indeed, a significant fraction of such correlations were similar (see [Figure S8](#sg002){ref-type="supplementary-material"}). The strongest conservation of coexpression was found between pairs of genes associated with particular cellular processes, such as core metabolic functions or central complexes (e.g., ribosome and proteasome) (lists of gene pairs with conserved coexpression are available at <http://barkai-serv.weizmann.ac.il/ComparativeAnalysis>). Next, we examined whether coexpression is conserved among groups of genes that are associated with the same cellular function. To this end, we used as a benchmark coexpressed groups of genes (termed transcription modules; [see Materials and Methods](#s4){ref-type="sec"} for a precise definition) that we extracted from the Saccharomyces cerevisiae expression data ([@pbio-0020009-Ihmels1]; J. Ihmels, unpublished data). (The yeast data are the most comprehensive and best annotated, resulting in a large number of transcription modules that can be associated with a specific cellular function.) For each yeast module, we constructed five "homologue modules," which contain the respective S. cerevisiae homologues in the other organisms, and measured the correlation between the genes of these homologue modules. The average correlation between the genes of the homologue modules was indeed statistically significant (see the top panel of [Figure 1](#pbio-0020009-g001){ref-type="fig"}C), indicating that coexpression of functionally linked genes is often conserved among organisms. ### Coexpression can be used for refining homologue modules {#s2a2} Examining the pairwise correlations themselves, however, revealed that usually only a fraction of the genes are correlated with each other (see [Figure S9](#sg003){ref-type="supplementary-material"}). Such lack of correlation probably reflects the inadequacy of defining function solely based on homology. To search for a coexpressed subset within each homologue module, we applied the signature algorithm we proposed recently ([@pbio-0020009-Ihmels1]). The algorithm identifies those homologues that are coexpressed under a subset of the experimental conditions. Furthermore, it reveals additional genes that are not homologous with any of the original genes, but display a similar expression pattern under those conditions ([see Materials and Methods](#s4){ref-type="sec"}). Studying the output of the algorithm, we found that the rejected homologues are usually not associated with the original function, while many of the added genes are. For example, from the 15 coexpressed yeast genes involved in heat-shock response, we identified eight homologues in Escherichia coli and 16 in Caenorhabditis elegans. While only some of these homologues are highly coexpressed, they are sufficient to retrieve additional genes known to be involved in heat shock ([Figure 1](#pbio-0020009-g001){ref-type="fig"}B; see [Figure S10](#sg004){ref-type="supplementary-material"} for other modules). A statistical analysis using all yeast modules revealed that many homologue modules are significantly coexpressed. The extent of coregulation increases drastically upon removing uncorrelated homologues and adding related genes ([Figure 1](#pbio-0020009-g001){ref-type="fig"}C). We note that in some cases such a "purified module" may contain two or more distinct coexpressed groups. Such substructures are identified by clustering all pairwise gene correlations (see [Data S3](#sd003){ref-type="supplementary-material"}). We conclude that sequence-based functional annotation can be significantly improved through the integration of expression data. We provide an interactive tool for this purpose on our Web site at <http://barkai-serv.weizmann.ac.il/ComparativeAnalysis> (see also [Figure S7](#sg001){ref-type="supplementary-material"}). We note that while this paper was in review, the possibility of enhancing functional assignment based on the conservation of coexpression was reported independently by [@pbio-0020009-Stuart1]). Higher-Order Regulatory Structures {#s2b} ---------------------------------- ### Regulatory relations between functional groups vary among organisms {#s2b1} The observation that groups of functionally related genes are often coexpressed in multiple organisms prompted us to ask whether also the higher-order regulatory relationships between these groups have been conserved ([see Materials and Methods](#s4){ref-type="sec"}). To address this question, we focused on eight representative yeast modules related to cellular core processes. Several of the regulatory relations among the homologues of these modules have been conserved ([Figure 2](#pbio-0020009-g002){ref-type="fig"}A). For example, in all organisms the modules associated with protein synthesis and protein secretion are positively correlated, while the rRNA synthesis and the peroxide modules are anticorrelated. Interestingly, however, most of the relations between modules differ among organisms. In particular, one of the prominent features of the yeast transcription program, namely the strong anticorrelation between heat-shock and protein-synthesis modules ([@pbio-0020009-Ihmels1]), was observed only in the yeast and *Drosophila* data. In contrast, those two modules displayed a significant positive correlation in the expression data of all other organisms. We note that both types of regulation are consistent with the role of heat-shock proteins as chaperones; it appears that in yeast their primary role is to assist in protein folding during stress conditions (when ribosomal protein genes are repressed), while in the other organisms they may be required to accelerate folding during cell growth. ::: {#pbio-0020009-g002 .fig} Figure 2 ::: {.caption} ###### Regulatory Relations between Modules A selection of eight transcription modules whose function is known in yeast was used to generate the corresponding (refined) homologue modules in the other five organisms. Each module is associated with a "condition profile" generated by the signature algorithm based on the expression data. \(A) Correlations between these profiles were calculated for all pairs of modules in each organism. Note that for E. coli there is no proteasome and that the mitochondrial ribosomal proteins (MRPs) correspond to ribosomal genes. Modules are represented by circles (legend). Significantly correlated or significantly anticorrelated modules are connected by colored lines indicating their correlation (color bar). Positively correlated modules are placed close to each other, while a large distance reflects anticorrelation. See Figure S11 for a numerical tabulation of all pairwise correlations. (B and C) Correlations between pairs of modules according to the cell-cycle data as a function their correlation in the full data. Each circle corresponds to a pair of S. cerevisiae modules (B) or human modules (C). \(D) To check the sensitivity of our results with respect to the size of the dataset, we reevaluated the correlations between the sets of conditions for randomly selected subsets of the data. Shown are the mean and standard deviation of the correlation coefficient between the heat-shock and protein-synthesis modules as a function of the fraction of removed conditions (see Figures S4 and S5 for correlations between other module pairs). :::  ::: In order to test whether the variations in the regulatory relations among functional groups in different organisms are due to the use of unrelated sets of experimental conditions, we restricted both the human and the yeast expression data to the cell cycle experiments. We found that the correlations between modules did not change qualitatively due to this restriction ([Figure 2](#pbio-0020009-g002){ref-type="fig"}B and 2C). We also examined the sensitivity of our results to the number of conditions used ([see Materials and Methods](#s4){ref-type="sec"}). Removal of up to 50% of all conditions did not considerably change the gene content of most refined modules (see [Data S2](#sd002){ref-type="supplementary-material"}). Importantly, this analysis also revealed that the correlations between modules are insensitive to the subset of conditions used ([Figure 2](#pbio-0020009-g002){ref-type="fig"}D; see also [Figure S2](#sd001){ref-type="supplementary-material"}). Note, for example, that for the largest datasets (yeast and C. elegans), the standard deviations of the correlation coefficients do not exceed 0.1, even when removing half of the expression profiles. Taken together, these results indicate that, despite the sparseness of the data, our findings reflect real properties of the expression networks and not the specific subset of experimental conditions used. ### Global decomposition of the expression data of different organisms {#s2b2} To compare the higher-order regulatory structures more systematically, we decomposed the expression data of each organism into a set of transcription modules using the iterative signature algorithm (ISA) we proposed recently ([@pbio-0020009-Bergmann1]; J. Ihmels, unpublished data). A transcription module consists of coexpressed genes and the conditions that induce their coregulation. Importantly, the stringency of coregulation is determined by a threshold parameter, which allows for a modular decomposition at different resolutions. At low resolution, a few relatively large transcription modules are identified. At higher resolution, the data are usually decomposed into a large number of modules, which contain fewer but more tightly regulated genes. We visualize the modular decomposition by a module tree ([Figure 3](#pbio-0020009-g003){ref-type="fig"}A and 3B). Highly similar modules, identified at adjacent thresholds, are connected by lines and define the branches of the tree. In contrast to the common dendrograms used to summarize the results of hierarchical clustering, here distinct branches may share common genes, and when two branches merge, the resulting branch is not necessarily their union. ::: {#pbio-0020009-g003 .fig} Figure 3 ::: {.caption} ###### Properties of Transcription Modules (A and B) Module trees summarize the transcription modules identified by the ISA at different resolutions. Branches represent modules (rectangles) that remain fixed points over a range of thresholds. Fixed points that emerge at a higher threshold converge into an existing module when iterated at a lower threshold (thin transversal lines). Modules are colored according to the fraction of homologues they possess in the other organism (see the color bar). Among the yeast modules, those associated with protein synthesis (arrow) have the largest fraction of worm homologues. Searchable trees for all six organisms are available at <http://barkai-serv.weizmann.ac.il/ComparativeAnalysis>. \(C) Histogram for the number of yeast modules with a given fraction of genes possessing a homologue in C. elegans (black bars). The distribution indicates that a significant number of modules have either much less or much more homologues than expected; indicated *p*-value were computed according to Kolmogorov--Smirnov test against control distribution (gray) generated from random sets of modules preserving their size. \(D) Same as in (C) for C. elegans modules considering yeast homologues (see Figure S12 for other organisms). :::  ::: ### Modular architectures of the transcription programs are distinct {#s2b3} Modular architectures, as reflected by the structure of the associated module trees, vary greatly among organisms. Differences were observed in the total number of modules, the threshold ranges over which modules are stable, and the overall hierarchical organizations. For example, in yeast the data were composed into just five transcription modules at low resolution, which remained stable for a wide range of thresholds ([Figure 3](#pbio-0020009-g003){ref-type="fig"}A). As we reported previously ([@pbio-0020009-Bergmann1]), these modules correspond to the central yeast functions (protein synthesis, stress, amino-acid biosynthesis, cell cycle, and mating). At high resolution, a large number of modules with specific cellular functions were identified. The corresponding module tree reveals a clear hierarchy in the transcriptional network, with gradually increasing complexity. In contrast, the C. elegans tree exhibits a sharp transition between a regime dominated by a single branch (from which only few less-stable modules branch off) to a part of the tree that rapidly bifurcates into many branches at higher thresholds ([Figure 3](#pbio-0020009-g003){ref-type="fig"}B). Interestingly, the functional groups that dominate the transcription program of each organism are also distinct. For example, in S. cerevisiae and E. coli, genes coding for ribosomal proteins are associated with a central branch that persists over a wide range of thresholds, reflecting the large number of the experimental perturbations that induce the coregulation of these genes. In contrast, although ribosomal proteins are also coregulated in higher organisms, they are associated with short branches that extend only over a small range of thresholds. This suggests that transcriptional regulation of genes involved in protein synthesis plays a major role in the transcription program of unicellular organisms, but a less dominant role in multicellular organisms. ### Conserved and organism-specific transcription modules {#s2b4} We observed that several functional groups were repeatedly identified as coexpressed in several organisms. This includes modules related to core biological functions such as protein synthesis, rRNA processing, the proteasome, and oxidative phosphorylation. Still, most of the transcription modules were observed in just one organism. In order to distinguish more systematically between generic modules and those that are involved in an organism-specific function, we determined for each module the fraction of genes that possess at least one homologue in a second organism ([see Materials and Methods](#s4){ref-type="sec"}). For S. cerevisiae and C. elegans (the two largest datasets), most modules have either significantly less or significantly more homologues than expected ([Figure 3](#pbio-0020009-g003){ref-type="fig"}C and 3D). This indicates that while a number of generic modules have been conserved under evolution, each transcriptome also contains more recently evolved modules that are associated with organism-specific functions. Comparing Global Features of Gene Expression Networks {#s2c} ----------------------------------------------------- ### Power-law connectivity distribution {#s2c1} We next sought to compare global topological properties of the expression data. To this end, we represented the data by an undirected "expression network," whose nodes correspond to genes. Two genes are connected by an edge if their expression profiles are sufficiently correlated ([see Materials and Methods](#s4){ref-type="sec"}). We use this mapping to explore the global structure of the expression data using tools of graph theory. A well-established indicator of the network topology is the distribution *n(k)* of the connectivity *k* (the number of edges of a particular gene). We find that for all organisms, the connectivity is distributed as a power-law, *n(k)* ∼ *k* ^−γ^, with similar exponents γ ≈ 1.1--1.8 (see [Figure 4](#pbio-0020009-g004){ref-type="fig"}A). The expression networks thus belong to the class of scale-free networks, which comprises many real-world networks ([@pbio-0020009-Albert1]). Power-law distributions have been attributed to dynamically evolving networks (Barabasi and Albert 1999) and to systems that are optimized to provide robust performance in uncertain environments ([@pbio-0020009-Doyle1]). In the present context, a power-law connectivity distribution indicates that there is no typical size for sets of coexpressed genes and that there is a significant enrichment of highly connected genes as compared to random networks (see also [@pbio-0020009-Guelzim1]; [@pbio-0020009-Lee1]; [@pbio-0020009-Shen-Orr1]). ::: {#pbio-0020009-g004 .fig} Figure 4 ::: {.caption} ###### Global Properties of Transcription Networks \(A) The number of genes *n(k)* with connectivity *k* is plotted as a function of *k* ([see Materials and Methods](#s4){ref-type="sec"}). For each of the six organisms *n(k)* is distributed as a power-law, *n(k)* ∼ *k* ^−γ^, with similar exponents γ ≈ 1.1--1.8 (see Figure S13). \(B) The fraction of lethal genes is shown as a function of *k* for S. cerevisiae, E. coli, and C. elegans. The control (gray line) is obtained from 10,000 random choices for the lethal genes (preserving their total number). The dashed lines indicate standard deviations. \(C) The fraction of genes with at least one yeast homologue is shown as a function of *k* for all six organisms. Control (gray) as in (B). \(D) *Z*-score quantifying the deviation of the number of connections between genes with connectivities *k* and *k*′ from that expected by randomly rewired networks (see [@pbio-0020009-Maslov1]). Note that connections between genes of similar connectivity are enhanced (red regions), while those between highly and weakly connected genes are suppressed (blue). \(E) The clustering coefficient *C* is plotted against *k*. Each dot corresponds to a single gene and is colored according to the transcription module it is associated with (see also [Figure 2](#pbio-0020009-g002){ref-type="fig"}). Note that genes associated with the same module correspond to a specific band in the *k*--*C* plane. Several genes with high connectivity belong to more than one module (green dots superimposed on orange ones). :::  ::: ### Highly connected genes are often essential and evolutionarily conserved {#s2c2} To see whether higher-order features of the connectivity distribution are also conserved, we calculated the likelihood *P(k, k′)* that two genes of connectivity *k* and *k*′, respectively, are connected with each other ([@pbio-0020009-Maslov1]). In all expression networks, connections between genes with similar connectivity occur much more often than expected, while connections between highly and weakly connected genes are suppressed ([Figure 4](#pbio-0020009-g004){ref-type="fig"}D). The common topology of the expression networks is thus different from the topology of the yeast protein--protein interaction network, although both exhibit a scale-free connectivity distribution ([@pbio-0020009-Maslov1]). We next examined whether highly connected genes are involved in central biological functions. In yeast, most of such genes are associated with protein synthesis, in particular rRNA processing. In the other organisms, the functional role of the highly connected genes is different and less coherent (lists of these "hub" genes are available at <http://barkai-serv.weizmann.ac.il/ComparativeAnalysis>). Interestingly, in the three organisms in which large-scale knockout information is available ([see Materials and Methods](#s4){ref-type="sec"}), the likelihood of a gene to be essential increases with its connectivity ([Figure 4](#pbio-0020009-g004){ref-type="fig"}B). Similar results were recently reported for the yeast expression and protein--protein interaction networks ([@pbio-0020009-Jeong1]; [@pbio-0020009-Farkas1]). We also observed that the highly connected genes are more likely to have homologues in the other organisms ([Figure 4](#pbio-0020009-g004){ref-type="fig"}C). This finding is consistent with the framework of dynamically evolving networks, where nodes that were added at an early stage (and may thus correspond to highly conserved genes) are more likely to develop many connections. ### Expression networks are highly clustered {#s2c3} A further indicator of the network structure is the clustering coefficient *C*, which quantifies the degree of modularity ([@pbio-0020009-Watts1]). For expression networks, *C~g~* measures to what extent the genes connected to a specific gene *g* are also connected with each other ([see Materials and Methods](#s4){ref-type="sec"}). The networks of all organisms exhibit a high modularity with 〈*C~g~*〉 ≈ ½, several orders of magnitude higher than what would be expected for random networks ([@pbio-0020009-Albert1]). We also examined the relation between the clustering coefficient and the connectivity of each gene. For all six organisms, we observed an approximately triangular region in the *k--C* plane where genes clustered into several localized elongated regions ([Figure 4](#pbio-0020009-g004){ref-type="fig"}E). Within these "bands," the clustering coefficient decreases monotonically as a function of the connectivity. Recently, a similar monotonic relation was observed in metabolic networks as well as in several nonbiological networks ([@pbio-0020009-Ravasz2]; Ravasz and Barabasi 2003). For random networks and for simple dynamically evolving networks, it was shown that *C* is independent of *k*. However, deterministic models that lead to a hierarchical organization of modularity predict *C* ∼ *k* ^--1^ ([@pbio-0020009-Dorogovtsev1]; [@pbio-0020009-Ravasz1]). Intriguingly, we found that genes belonging to the same band are often coexpressed and associated with one of the dominating coexpressed units (transcription modules) identified by our modular analysis. The decrease of *C* as a function of *k* may reflect overlap between modules. Genes that are associated with only one module have a connectivity reflecting the size of the module and a large clustering coefficient. In contrast, genes that belong to several modules are correlated with a larger number of genes, but many of these genes are not connected with each other, leading to a smaller clustering coefficient. In support of this, we found that highly connected genes with a small clustering coefficient are often associated with several modules ([Figure 4](#pbio-0020009-g004){ref-type="fig"}E). Thus, the band-like structures we observed may reflect the combinatorial regulation of gene expression. Conclusions {#s3} =========== Comparing genomic properties of different organisms is of fundamental importance in the study of biological and evolutionary principles. Although much of the differences among organisms is attributed to different gene expression, comparative analysis thus far has been based primarily on genomic sequence information. The potential of including functional genomic properties in a comparison analysis was demonstrated in recent studies that compared protein--protein interaction networks of different organisms ([@pbio-0020009-Matthews1]; [@pbio-0020009-Kelley1]). In this paper we presented a comparative analysis of large datasets of expression profiles from six evolutionarily distant organisms. We showed that all expression networks share common topological properties, such as a scale-free connectivity distribution and a high degree of modularity. While these common global properties may reflect universal principles underlying the evolution or robustness of these networks, they do not imply similarity in the details of the regulatory programs. Rather, with a few exceptions, the modular components of each transcription program as well as their higher-order organization appear to vary significantly between organisms and are likely to reflect organism-specific requirements. Nevertheless, coexpression of functionally linked genes is often conserved among several organisms. Based on this finding, we proposed an efficient method that uses coexpression analysis for improving sequence-based functional annotation. An interactive implementation of this algorithm is available at <http://barkai-serv.weizmann.ac.il/ComparativeAnalysis/>. Our analysis was based on the available expression data, which are still sparse for most organisms. It is likely that the modular decompositions we obtained are partial, so additional modules can be identified as more expression data become available. Nevertheless, by analyzing the sensitivity of our results to the number of conditions, we concluded that the composition of the modules themselves is rather robust. Moreover, the higher-order correlations between modules are only slightly affected by the number of conditions. The absence of a large set of common experimental conditions, however, does limit the scope of the present analysis and reduces the possibility of addressing particular evolutionary issues. It would be interesting, for example, to compare how different organisms respond to a variety of stress conditions, which were found to induce a unified transcription program in S. cerevisiae ([@pbio-0020009-Gasch1]). Similarly, it would be intriguing to examine whether knockouts of homologous genes induce a similar transcriptional response in the different organisms. Comparative studies of gene expression pattern could be largely facilitated by unified datasets, which examine the genome-wide expression profiles of diverse as well as related species, under comparable experimental conditions. Materials and Methods {#s4} ===================== {#s4a} ### Expression data {#s4a1} Preprocessed expression data from E. coli, Arabidopsis thaliana, and Homo sapiens were downloaded from the Stanford Microarray Database ([@pbio-0020009-Sherlock1]) using default parameters and selecting data from all experimenters and categories. For technical reasons (see [Data S5](#sd005){ref-type="supplementary-material"}), we only used the first 720 experimental conditions for the human dataset or all conditions related to the cell cycle. C. elegans expression data were obtained from [@pbio-0020009-Kim1]) and Drosophila melanogaster data from [@pbio-0020009-Arbeitman1]). The yeast expression data ([@pbio-0020009-Gasch1]; [@pbio-0020009-Hughes1]; [@pbio-0020009-Causton1]) contain more than 1,000 experiments (see <http://barkai-serv.weizmann.ac.il/modules/page/references.html> for a complete list of references). We excluded genes or conditions with more than 90% missing datapoints, resulting in expression matrices of the dimensions shown in [Table 1](#pbio-0020009-t001){ref-type="table"} (see [Data S4](#sd004){ref-type="supplementary-material"} for comment on missing values in the expression data). ### Sequence data {#s4a2} FASTA files for amino-acid sequences of coding regions were downloaded from the sources detailed in [Table 2](#pbio-0020009-t002){ref-type="table"}. We ran the BLASTP 2.2.2 ([@pbio-0020009-Altschul1]) locally in order to determine the sequence similarity among all coding regions. Gene/ORF identifiers were used to link the sequence data with the expression profiles. ::: {#pbio-0020009-t002 .table-wrap} Table 2 ::: {.caption} ###### The Sources of the Sequence Data Used in This Study :::  ::: ### Knockout data {#s4a3} Data for deletion mutants (S. cerevisiae and E. coli) and RNAi experiments (C. elegans) were obtained from the sources indicated in [Table 3](#pbio-0020009-t003){ref-type="table"}. Note that the fraction of the genome that was tested for viability varies among the three organisms. ::: {#pbio-0020009-t003 .table-wrap} Table 3 ::: {.caption} ###### Sources of Knockout Data Used in This Study :::  ::: ### Module definition {#s4a4} A transcription modules consist of a set of coregulated genes (a subset *G~m~* of the genome *G*) and an associated set of regulating conditions (a subset *C~m~* of all conditions *C*). The defining property of a transcriptional module is self-consistency, which is achieved as follows. First, we assign scores to both genes and conditions that reflect their degree of association with the module. The gene score is the average expression of each gene over the module conditions, weighted by the condition score: . Analogously, the condition score is the weighted average over the module genes, . Here, and are the log-expression ratio of gene *g* in condition *c* normalized over genes and conditions, respectively, such that , and , . Self-consistency denotes the property that the genes of the module are those genes of the genome that receive the highest scores *s~g~*, while the module conditions are those conditions in the dataset with the highest scores *s~c~*. The ISA identifies transcription modules through iterative refinement of a large number of random gene scores. ### Module analysis {#s4a5} For the analysis of the fraction of homologues (see [Figure 3](#pbio-0020009-g003){ref-type="fig"}C and [3](#pbio-0020009-g003){ref-type="fig"}D) as well as the average pairwise correlations (see [Figure 1](#pbio-0020009-g001){ref-type="fig"}C), we used most of the transcription modules identified by the ISA. In order to avoid bias from similar modules identified at adjacent thresholds, we considered only modules with less than 70% similarity to any module identified at a lower threshold. Two sequences were considered homologues if they could be aligned along at least 40% of the shorter sequence by the BLAST algorithm and obtained an E-value smaller than 10^--5^. The precise parameter values have only a minor effect on our results (see [Data S1](#sd001){ref-type="supplementary-material"} for detailed statistical analysis). We only considered modules with at least five homologues. ### Module purification and refinement {#s4a6} A "homologue module" consists of the genes homologous to a transcription module in another organism. We used the signature algorithm to purify and refine these homologue modules (see [@pbio-0020009-Ihmels1] for details of the algorithm). A "purified module" is the intersection between the homologue module, used as input for the signature algorithm, and the resulting output. It contains only genes that are coexpressed. A "refined module" is obtained by applying the signature algorithm again, this time using the purified module as input. The output consists both of the coexpressed genes and the conditions inducing their coexpression. This twofold application of the signature algorithm usually provides a more accurate determination of the coexpressed genes related to the original transcription module than a single application. In order to also capture weakly coexpressed modules, we used relatively low thresholds (*t~g~* = *t~c~* = 1.5) in the present analysis, but retained only genes whose score is not less than 70% of the most significant gene ([@pbio-0020009-Ihmels1]). ### Correlations between modules {#s4a7} Both a transcription modules and the refined homologue module derived from it are associated with a set of coregulating experimental conditions ([@pbio-0020009-Ihmels1]). The significance of each condition is characterized by a score *s~c~*. The sets of scores can be used to compute the regulatory relation between two modules of the same organism. We use as the correlation coefficient between two modules with score sets and , respectively. Note that, unlike for the Pearson correlation, this definition of *C~ij~* does not center the scores. ### Network analysis {#s4a8} Each expression network can be described by a symmetric adjacency matrix *A~ij~*, whose elements are 1 if the expression of gene *i* and gene *j* are sufficiently similar and 0 otherwise. Similarity was measures by the Pearson correlation coefficient between the expression profiles. Owing to the very different sizes of the respective sets of expression data, we demanded that the average connectivity \<*k*\> (rather than the minimal correlation) is identical in all expression networks and fixed it to \<*k*\> = 0.001. Using the top 0.1% of all possible correlations corresponds to a lower limit on the correlation coefficients between 0.63 for S. cerevisae and 0.85 for D. melanogaster. The results are insensitive to the precise threshold value (see [Figure S2](#sd001){ref-type="supplementary-material"} for detailed analysis). The connectivity of gene *i* is *k* = Σ~*j*≠*i*~ *A~ij~*. In order to obtain the connectivity distributions *n(k)*, we used logarithmic binning. The edges of the bins were powers of 2, and we counted the number of genes with *k~i~* between two edges and normalized by the bin width. We applied a linear fit to the log values of the bin centers against the normalized counts. We note that the resulting connectivity distributions scale as a power-law for a wide range of thresholds and the exponents only depend weakly on the choice of the threshold. The clustering coefficient of gene *i* is *C~i~* = (Σ~*k*\>*j*≠*i*~ *A~ik~* *A~kj~* *A~ji~*)/\[*k~i~*(*k~i~* −1)/2\]. ### Web site {#s4a9} Interactive applications for the refinement of sets of homologous genes and the exploration of our modular decompositions of the expression data are available online. We also present details about the highly connected genes in each organism, the pairs of genes that are significantly correlated in two organisms, and the eight modules related to core processes in yeast (and their homologue modules before and after refinement) on our website at <http://barkai-serv.weizmann.ac.il/ComparativeAnalysis>. Supporting Information {#s5} ====================== Data S1 ::: {.caption} ###### Testing the Robustness of Our Analyses with Respect to the Precise Values of Threshold Parameters This note includes **Figure S1** and **Figure S2**. (38 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Data S2 ::: {.caption} ###### Controls to Verify That Our Results Are Not Impaired by the Sparseness of the Available Expression Data This note includes **Figure S3**, **Figure S4**, and **Figure S5**. (59 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Data S3 ::: {.caption} ###### Testing for Coregulated Subsets within the Homologue Modules This note includes **Figure S6**. (11 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Data S4 ::: {.caption} ###### Comment on Missing Values in the Expression Data (3 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Data S5 ::: {.caption} ###### Comment on the Size of the Human Dataset Used in This Work After this work was completed, we succeeded in processing the more than 2,000 human chip experiments deposited at the SMD. Removing genes and conditions with more than 90% missing values resulted in 1,474 expression profiles for 24,795 genes. Our Web tools ("GeneHopping" and "ModuleTree") allow researchers to use also this updated dataset. (3 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S7 ::: {.caption} ###### The Interactive Web Tool (137 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S8 ::: {.caption} ###### Statistical Analysis Comparing the Pairwise Correlations between Genes in One Organism to the Correlations between Their Respective Homologues (16 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S9 ::: {.caption} ###### Pairwise Correlations of C. elegans Homologues to the Yeast Heat-Shock Module (15 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S10 ::: {.caption} ###### Correlations between the Genes of Eight Representative Yeast Modules and Their Homologue Modules, Purified Modules, and Refined Modules (33 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S11 ::: {.caption} ###### Pairwise Correlations between Eight Transcription Modules of Known Function in Yeast and Their Refined Homologue Modules in the Five Other Organisms (11 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S12 ::: {.caption} ###### Histograms Showing the Number of Modules of One Organism with a Given Fraction of Homologues in Another Organism (29 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S13 ::: {.caption} ###### Connectivity Distributions for the Six Organisms in Separate Plots (19 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: We thank I. Yanai and the members of our lab for discussions and R. Milo and S. Shen-Orr for comments on the manuscript. This work was supported by the National Institutes of Health grant A150562, the Israeli Science Ministry, and the Y. Leon Benoziyo Institute for Molecular Medicine. SB is a Koshland fellow. NB is the incumbent of the Soretta and Henry Shapiro Career Development Chair. SB performed the numerical experiments. SB analyzed the data. SB and JI contributed reagents/materials/analysis tools. SB and NB wrote the paper. JI programmed the online applications. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** SB and NB conceived and designed the experiments. Academic Editor: Michael Eisen, Lawrence Berkeley National Laboratory ISA : iterative signature algorithm

PubMed Central

2024-06-05T03:55:45.991279

2003-12-15

PMC300883

Introduction {#s1} ============ Most, if not all, cells have the potential to carry out the apoptotic cell death program ([@pbio-0020015-Jacobson1]). Key players in this process are caspase family proteases. Apical caspases are activated through interactions with adapter molecules in response to death signals arising from cellular compartments such as the mitochondria and plasma membrane death receptors. These caspases transduce death signals by cleaving and activating effector caspases. Effector caspases then cleave and alter the function of a number of cellular proteins, leading to the morphological and biochemical events associated with apoptosis ([@pbio-0020015-Kumar1]). Proteolysis is an irreversible protein modification. Therefore, caspase activation is normally kept under tight control in viable cells. However, in *Drosophila* the apoptotic effector caspase DRICE is cleaved and activated throughout the length of elongated spermatids, and testis-specific expression of the baculovirus caspase inhibitor p35 results in male sterility, despite the fact that apoptosis is not an obligate step in spermatogenesis ([@pbio-0020015-Arama1]). These observations demonstrate that caspase activity is important for male fertility, but leave a number of questions unanswered: For what events in spermatid differentiation are caspases required? Which caspases mediate this requirement? How are they activated and where do they act? And how do these cells avoid apoptosis? Spermatid development in *Drosophila* takes place within a syncytium (cyst), in which 64 haploid spermatid nuclei descended from a diploid primary spermatogonial cell are connected by abundant cytoplasmic bridges (reviewed in [@pbio-0020015-Lindsley1]). In mammals, similar bridges facilitate the sharing of haploid gene products between spermatids, thereby allowing spermatid development to be directed by the products of both sets of parental chromosomes ([@pbio-0020015-Erickson1]; [@pbio-0020015-Braun1]). It is presumed that intercellular bridges play a similar role in *Drosophila*. Ultimately, these bridges must be eliminated in a process known as individualization in order to form freely swimming sperm. At the end of male meiosis, each cyst contains 64 haploid spermatids, each approximately 2 mm long, encapsulated by two somatic cyst cells. The 64 nuclei are located at the basal end of the testis, near the seminal vesicle, and the flagellar tails extend apically, throughout the length of the testis. Individualization in *Drosophila* initiates when an actin-based structure known as an investment cone assembles around each spermatid nucleus ([@pbio-0020015-Tokuyasu1]). These assemble into a macroscopic structure known as the individualization complex ([@pbio-0020015-Fabrizio1]), which moves along the length of the cyst toward the sperm tails. The individualization complex is the site at which the cyst membrane is remodeled to enclose each sperm. Cytoplasm and organelles are extruded from between the sperm tails and pushed ahead of the individualization complex, forming a visible bulge known as the cystic bulge. When the cystic bulge reaches the sperm tails, it is detached and becomes known as the waste bag ([@pbio-0020015-Tokuyasu1]). A similar process, involving encapsulation of syncytial spermatids within individual plasma membranes and elimination of excess cytoplasm, also occurs during mammalian spermatogenesis ([@pbio-0020015-de1]). The importance of cytoplasm elimination for human fertility is suggested by the fact that many conditions or treatments resulting in infertility disrupt this process ([@pbio-0020015-Russell1]; [@pbio-0020015-Keating1]; [@pbio-0020015-Akbarsha1]). Cytoplasm elimination during spermatogenesis may also represent a strategy by which male gametes eliminate cytoplasmic parasites, thereby preventing their transmission to the zygote ([@pbio-0020015-Randerson1]). Results {#s2} ======= Caspase Activity Is Required for Spermatid Individualization {#s2a} ------------------------------------------------------------ To determine whether caspase activity is required for spermatid individualization, we examined cysts from flies in which caspase activity in the male germline was inhibited. We generated flies that expressed the broad-specificity *Drosophila* caspase inhibitor DIAP1 or the baculovirus caspase inhibitor p35 under the control of the male germline-specific β~2~-tubulin promoter (β2tub-DIAP1 and β2tub-p35 flies, respectively) ([@pbio-0020015-Hay1]; [@pbio-0020015-Santel1]). Cysts undergoing individualization contain activated versions of the effector caspase DRICE, as visualized with an anti-active DRICE-specific antibody ([@pbio-0020015-Arama1]). Testis from wild-type animals always contained active DRICE-positive cysts with prominent cystic bulges and waste bags ([Figure 1](#pbio-0020015-g001){ref-type="fig"}A). In contrast, while elongated cysts from β2tub-DIAP1 and β2tub-p35 flies remained active DRICE-positive, cystic bulges and waste bags were largely absent and reduced in size when present ([Figure 1](#pbio-0020015-g001){ref-type="fig"}B and [1](#pbio-0020015-g001){ref-type="fig"}C). In addition, the normally coordinated tailward movement of investment cones in active DRICE-positive wild-type cysts ([Figure 1](#pbio-0020015-g001){ref-type="fig"}D) was dramatically disrupted in β2tub-DIAP1 males ([Figure 1](#pbio-0020015-g001){ref-type="fig"}E). Cysts from β2tub-p35 males showed milder defects in investment cone movement ([Figure 1](#pbio-0020015-g001){ref-type="fig"}F). These phenotypes, in conjunction with related observations by [@pbio-0020015-Arama1]), suggest, but do not prove, that caspase inhibition results in defects in individualization. To further test this hypothesis, we examined spermatids for individualization defects directly, using transmission electron microscopy (EM). In cysts from wild-type animals in which individualization had occurred, spermatid tails consisted largely of a flagellar axoneme and major and minor mitochondrial derivatives, all of which were tightly surrounded by a unit plasma membrane ([Figure 1](#pbio-0020015-g001){ref-type="fig"}G and [1](#pbio-0020015-g001){ref-type="fig"}J). In contrast, in many cysts from β2tub-DIAP1 and β2tub-p35 flies, spermatids failed to separate from each other and contained excess cytoplasm, often including an enlarged minor mitochondrial derivative ([Figure 1](#pbio-0020015-g001){ref-type="fig"}H and [1](#pbio-0020015-g001){ref-type="fig"}K and [Figure 1](#pbio-0020015-g001){ref-type="fig"}I and [1](#pbio-0020015-g001){ref-type="fig"}L, respectively). Phenotypes similar to those seen in cysts from β2tub-DIAP1 and β2tub-p35 flies were also observed in cysts from flies in which levels of the caspase *Drosophila* caspase-1 (DCP-1) were decreased specifically in the male germline using RNA interference (RNAi) (β2tub-Dcp-1-RNAi flies) ([Figure S1](#sg001){ref-type="supplementary-material"}). Short prodomain caspases such as DCP-1 and DRICE are activated in response to cleavage by upstream signal-transducing caspases ([@pbio-0020015-Hawkins2]; [@pbio-0020015-Hay1]; [@pbio-0020015-Meier1]; [@pbio-0020015-Shi1]). Together, these observations demonstrate that caspase activity is required for individualization and suggest that DCP-1 (but perhaps not DRICE; see [Discussion](#s3){ref-type="sec"} below) is an important downstream target caspase. ::: {#pbio-0020015-g001 .fig} Figure 1 ::: {.caption} ###### Caspase Activity Is Required for Spermatid Individualization (A--C) Testis of different genotypes were visualized with antibodies specific for activated Drice (green). (A) Wild-type testis. Active DRICE is present in multiple elongated cysts. Cystic bulges (cb) and waste bags (wb) are indicated by arrows. (B and C) Testes from β2tub-DIAP1 and β2tub-p35 males, respectively. Active DRICE is present in elongated cysts, but cystic bulges and waste bags are reduced in number and size. (D--F) Phalloidin-stained investment cones from testes of different genotypes (red). Spermatid axonemes in (D)--(F) are highlighted by the AXO49 antibody, which recognizes polyglycylated β2tub ([@pbio-0020015-Bressac1]) (blue). (D) In wild-type testes, investment cones move as a coordinated group. (E and F) Coordinated investment cone movement is disrupted in cysts from β2tub-DIAP1 and β2tub-p35 males, respectively. (G--L) EM sections of elongated cysts of different genotypes. (G) A cyst from a wild-type male that has undergone individualization. The boxed region is shown at higher magnification in (J), along with the locations of the major mitochondrial derivative (mj), minor mitochondrial derivative (mi), and axoneme (ax). A single spermatid unit is outlined with a dashed line. (H and I) In cysts from β2tub-DIAP1 and β2tub-p35 males, respectively, many spermatid units are present in a common cytoplasm that contains organelles, often including an enlarged minor mitochondrial derivative. Boxed regions of β2tub-DIAP1 and β2tub-p35 cysts shown in (H) and (I) are shown at higher magnification in (K) and (L), respectively. Several examples of multiple spermatids present in a common cytoplasm are outlined by the dashed line in (K) and (L). Scale bar for EM micrographs = 1 μm. :::  ::: Ark and Dronc Participate in Spermatid Individualization {#s2b} -------------------------------------------------------- What are the pathways that lead to caspase activation during individualization? Cell death in many contexts in the fly requires the activity of the *Drosophila* APAF-1 homolog ARK, which promotes activation of the apical caspase DRONC ([@pbio-0020015-Dorstyn1]; [@pbio-0020015-Igaki1]; [@pbio-0020015-Muro1]; [@pbio-0020015-Zimmermann1]). DRONC, in turn, can cleave and activate the downstream caspases DCP-1 and DRICE ([@pbio-0020015-Hawkins2]; [@pbio-0020015-Meier1]; [@pbio-0020015-Muro1]). Genetic and biochemical evidence implicates all three of these caspases as apoptosis inducers ([@pbio-0020015-Kumar1]). Animals homozygous for a hypomorphic *Ark* allele (*Ark^CD4^*) showed a high level of male sterility ([@pbio-0020015-Rodriguez1]), despite the fact that cell death is not an obligate step in spermatogenesis ([@pbio-0020015-Fuller1]). This suggested to us that ARK-dependent DRONC activity might be important. To test this hypothesis, we decreased ARK levels specifically in the male germline by expressing double-stranded RNA homologous to *Ark* under the control of the β2tub promoter (β2tub-Ark-RNAi flies) ([Figure 2](#pbio-0020015-g002){ref-type="fig"}I). To decrease levels of active DRONC, we generated flies that expressed a dominant-negative version of DRONC (Dn-DRONC) under the control of the β2tub promoter (β2tub-Dn-DRONC flies). Similar versions of DRONC are potent suppressors of DRONC-dependent cell death in other contexts ([@pbio-0020015-Hawkins2]; [@pbio-0020015-Meier1]). DRICE was still activated in elongated cysts from β2tub-Ark-RNAi and β2tub-Dn-DRONC males ([Figure 2](#pbio-0020015-g002){ref-type="fig"}A and [2](#pbio-0020015-g002){ref-type="fig"}C). However, as with active DRICE-positive cysts from β2tub-DIAP1 and β2tub-p35 flies, cystic bulges and waste bags were largely absent, and coordinated investment cone movement was disrupted ([Figure 2](#pbio-0020015-g002){ref-type="fig"}B and [2](#pbio-0020015-g002){ref-type="fig"}D). Examination of β2tub-ARK-RNAi and β2tub-Dn-DRONC spermatids using EM showed that inhibition of ARK ([Figure 2](#pbio-0020015-g002){ref-type="fig"}E, [2](#pbio-0020015-g002){ref-type="fig"}G, and [2](#pbio-0020015-g002){ref-type="fig"}H) and DRONC ([Figure 2](#pbio-0020015-g002){ref-type="fig"}F) function resulted in individualization failure in many cysts. In addition, many single spermatid units that were surrounded by a unit plasma membrane still contained large fingers of excess cytoplasm ([Figure 2](#pbio-0020015-g002){ref-type="fig"}E and [2](#pbio-0020015-g002){ref-type="fig"}G). (See [Discussion](#s3){ref-type="sec"} below.) ::: {#pbio-0020015-g002 .fig} Figure 2 ::: {.caption} ###### ARK and DRONC Are Required for Spermatid Individualization (A and C) Testis from β2tub-Ark-RNAi and β2tub-Dn-DRONC males, respectively. Active DRICE-positive cysts are present, but cystic bulges and waste bags are largely absent. (B and D) Investment cone movements in testis from β2tub-Ark-RNAi and β2tub-Dn-DRONC, respectively, are uncoordinated. (E, G, and H) EM images of an elongated cyst from a β2tub-Ark-RNAi male. Some individualization failures are observed (E, G, and H), two of which are highlighted by the dashed lines in (G) and (H). In addition, many spermatids that have apparently undergone individualization still contain large amounts of excess cytoplasm (E and G). \(F) EM image of a cyst from a β2ub-Dn-DRONC male. A large region in which individualization did not occur is outlined. \(I) Western blot from wild-type (Wt) and β2tub-Ark-RNAi (DArki) testis probed with anti-ARK and anti-DRICE antibodies. ARK, but not DRICE, levels are greatly reduced in β2tub-Ark-RNAi testis. :::  ::: ARK-Dependent Activation of DRONC Occurs at Sites of Individualization and Requires the Apoptosis Inducer HID {#s2c} ------------------------------------------------------------------------------------------------------------- To determine where active DRONC is localized and thus where DRONC is likely to be functioning during individualization, we generated an antibody that recognized versions of DRONC that had undergone autoactivation-associated cleavage at glutamate-352 (TQT*E*) ([Figure S2](#sg002){ref-type="supplementary-material"}). In contrast to active DRICE, which appeared uniformly throughout the cyst, just as the individualization complex began its apical movement away from the spermatid nuclei ([Figure 3](#pbio-0020015-g003){ref-type="fig"}A and [3](#pbio-0020015-g003){ref-type="fig"}B), active DRONC showed a dynamic pattern of localization. It was initially observed in a punctate pattern just apical to the juxtanuclear individualization complex (arrowhead in [Figure 3](#pbio-0020015-g003){ref-type="fig"}C). The individualization complex moved through this region (arrow in [Figure 3](#pbio-0020015-g003){ref-type="fig"}C), and active DRONC then trailed the individualization complex for the remainder of its apical movement through the cyst ([Figure 3](#pbio-0020015-g003){ref-type="fig"}D). As expected, DRONC activation required ARK and was eliminated in testis in which ARK levels were decreased ([Figure 3](#pbio-0020015-g003){ref-type="fig"}E) or in which access of wild-type DRONC to ARK was inhibited by expression of inactive Dn-DRONC ([Figure 3](#pbio-0020015-g003){ref-type="fig"}F). ::: {#pbio-0020015-g003 .fig} Figure 3 ::: {.caption} ###### DRONC Activation Occurs in Association with Individualization Complexes and Is ARK-Dependent (A and B) Wild-type testis stained for active DRICE (green), phalloidin-stained filamentous actin (red), and TOTO-3-stained DNA (blue). (A) Active DRICE is present throughout the length of cysts undergoing individualization. (B) Higher magnification of the testis in (A). The arrowhead points to a cyst in which the individualization complex has assembled around the spermatid nuclei, but DRICE activation has not occurred. The arrow points to a neighboring cyst in which the individualization complex has just begun to move away from the spermatid nuclei. Active DRICE is now present throughout the length of this cyst, indicating that DRICE activation within a cyst occurs rapidly and globally. \(C) Active DRONC (green) is initially present in a punctate pattern, apical to the individualization complex (red) at the base of the testis (arrowheads). The individualization complex then moves through the region containing active DRONC (arrow). \(D) Subsequently, active DRONC is found associated with the trailing edge of the individualization complex as it moves apical within the cyst. A higher magnification view of active DRONC staining in the left-most cyst is shown in the inset. (E and F) Active DRONC is eliminated in cysts from β2tub-Ark-RNAi and β2tub-Dn-DRONC testis, respectively. :::  ::: How is DRONC activation during individualization regulated? DRONC undergoes continuous ARK-dependent activation in many viable cells ([@pbio-0020015-Dorstyn1]; [@pbio-0020015-Igaki1]; [@pbio-0020015-Muro1]; [@pbio-0020015-Rodriguez2]; [@pbio-0020015-Zimmermann1]). DIAP1 promotes the survival of these cells by ubiquitylating DRONC ([@pbio-0020015-Wilson1]; [@pbio-0020015-Chai1]) and by inhibiting the activity of caspases activated by DRONC ([@pbio-0020015-Hawkins1]; [@pbio-0020015-Wang1]). REAPER/HID/GRIM family proteins promote DRONC activity and apoptosis by disrupting DIAP1--caspase interactions, thereby preventing DIAP1-dependent ubiquitylation of DRONC and inhibition of caspases activated by DRONC ([@pbio-0020015-Wang1]; [@pbio-0020015-Goyal1]; [@pbio-0020015-Lisi1]; [@pbio-0020015-Wilson1]; [@pbio-0020015-Chai1]). To determine whether the REAPER/HID/GRIM family proteins played a similar role during individualization, we examined available mutants for these genes. Cysts from flies lacking *reaper* (*XR38*/*H99*) ([@pbio-0020015-Peterson1]) showed normal investment cone movement. In contrast, the coordinated movement of investment cones was disrupted in *hid^ 05014^*/*H99* cysts ([Figure 4](#pbio-0020015-g004){ref-type="fig"}D, compared with [Figure 4](#pbio-0020015-g004){ref-type="fig"}C), indicating a requirement for HID in spermatogenesis. In addition, HID protein was enriched in the cystic bulge region of wild-type cysts ([Figure 4](#pbio-0020015-g004){ref-type="fig"}A), but not those from animals that lacked HID (*hid^ 05014^*/*H99*) ([Figure 4](#pbio-0020015-g004){ref-type="fig"}B). These observations suggested that HID participates in DRONC activation, stabilization, or both and thereby in spermatid individualization. ::: {#pbio-0020015-g004 .fig} Figure 4 ::: {.caption} ###### HID, dFADD, and DREDD Participate in Individualization \(A) HID protein (green) is concentrated in the region of the cystic bulge, which is marked by the presence of the phalloidin-stained individualization complex (red). \(B) HID immunoreactivity is absent in testis from *hid ^05014^/H99* flies. \(C) Active DRONC (green) is associated with the trailing edge of the individualization complex in a wild-type cyst. \(D) Active DRONC is absent from the individualization complex in cysts from *hid ^05014^/H99* males. \(E) EM section from *hid ^05014^/H99* testis. Essentially all spermatids have failed to individualize. \(F) Higher magnification view of boxed area in (E). Multiple spermatid units sharing a common cytoplasm are outlined by the dashed line. \(G) Representative EM section of cyst from *dFadd ^f02804^*/*dFadd ^f02804^* testis. Essentially all spermatids have failed to individualize. \(H) EM section of cyst from *Dredd ^B118^*/*Dredd ^B118^* testis in which individualization has failed to occur. In some other cysts from this same male, individualization proceeded apparently normally (data not shown). :::  ::: Several observations support this hypothesis. First, cysts from two different *hid* allelic combinations, *hid^ A329^*/*hid^ A329^* (data not shown) and *hid^ 05014^*/*H99*, showed defects in individualization similar to those observed in β2tub-DIAP1 or β2tub-p35 males, demonstrating a requirement for HID in this process ([Figure 4](#pbio-0020015-g004){ref-type="fig"}E and [4](#pbio-0020015-g004){ref-type="fig"}F). Second, localized active DRONC was eliminated in *hid^ 05014^*/*H99* (and *hid^ A329^*/*hid^ A329^*; data not shown) flies, consistent with the idea that HID promotes individualization, at least in part, by promoting DRONC activity ([Figure 4](#pbio-0020015-g004){ref-type="fig"}D, compared with [Figure 4](#pbio-0020015-g004){ref-type="fig"}C). HID, by virtue of its ability to disrupt IAP (inhibitor of apoptosis)--caspase interactions, may also regulate the activation of other caspase cascades during spermatid individualization (see below). Spermatid Individualization Utilizes Multiple Pathways of Caspase Activation {#s2d} ---------------------------------------------------------------------------- Together the above observations demonstrate that components of a canonical apoptosis-inducing pathway involving ARK, DRONC, and HID are required for spermatid individualization. However, it is important to note that the individualization defects observed in testis from β2tub-Ark-RNAi and β2tub-Dn-DRONC males (see [Figure 2](#pbio-0020015-g002){ref-type="fig"}) were less severe then those seen in β2tub-DIAP1, β2tub-p35, or *hid^ 05014^*/*H99* males (see Figures [1](#pbio-0020015-g001){ref-type="fig"} and [4](#pbio-0020015-g004){ref-type="fig"}). These differences may reflect incomplete inactivation of ARK and DRONC. Alternatively, they may reflect roles for ARK- and DRONC-independent caspase activities. DREDD is an interesting candidate to mediate such an activity since it is an apical caspase that can promote cell death in some contexts ([@pbio-0020015-Chen1]; [@pbio-0020015-Hu1]). Its activation is stimulated through interactions with dFADD, the *Drosophila* homolog of mammalian FADD, an adaptor that mediates recruitment of apical caspases to ligand-bound death receptors, thereby promoting caspase activation ([@pbio-0020015-Hu1]). Elongated cysts from *dFadd^ f02804^*/*dFadd^ f02804^* and *Dredd^ B118^*/*Dredd^ B118^* males (both are genetic null mutations) contained active DRICE, but often showed uncoordinated investment cone movement (data not shown). At the EM level, elongated cysts from testis of single *dFadd^ f02804^*/*dFadd ^f02804^* and *Dredd^ B118^*/*Dredd^ B118^* males showed a range of phenotypes. About 50% of cysts from *Dredd^ B118^*/*Dredd^ B118^* males and almost all cysts from *dFadd^ f02804^*/*dFadd^ f02804^* males (greater than 90%) displayed defects in individualization similar to those of β2tub-DIAP1 and β2tub-p35 flies ([Figure 4](#pbio-0020015-g004){ref-type="fig"}G and [Figure 4](#pbio-0020015-g004){ref-type="fig"}H, respectively). In other cysts, individualization occurred apparently normally (data not shown). Together these observations argue that dFADD and DREDD participate in individualization. The fact that loss of *dFadd* resulted in phenotypes more severe than those due to loss of *Dredd* suggests that dFADD has functions in individualization independent of promoting DREDD activation. Finally, we noted that DRICE activation was insensitive to inhibition (but perhaps not to complete elimination) of ARK and DRONC; to complete loss of HID, DREDD, or FADD; and to expression of the potent general caspase inhibitors DIAP1, p35, or p49 (see [Figures 1--4](#pbio-0020015-g001){ref-type="fig"}; data not shown). This, together with the observation that DRONC and DRICE were activated in distinct spatial and temporal patterns (see [Figure 3](#pbio-0020015-g003){ref-type="fig"}A--[3](#pbio-0020015-g003){ref-type="fig"}D), suggests that DRICE activation occurs through an unknown HID-, ARK-, DRONC-, dFADD-, and DREDD-independent mechanism. It has been proposed that DRICE activation in spermatids is essential for fertility and that DRICE activation is mediated by an isoform of cytochrome c, cytochrome c-d (cyt-c-d), based on the observation that males homozygous for a P-element insertion (*bln^1^*) in the *cyt-c-d* gene were sterile and lacked active DRICE staining in testis ([@pbio-0020015-Arama1]). However, as illustrated in [Figure 5](#pbio-0020015-g005){ref-type="fig"}, the region surrounding the *bln^1^* insertion contains multiple transcription units. In addition, cysts from *bln^1^* males showed multiple defects in spermatogenesis prior to individualization, including failure to carry out polyglycylation of axonemal microtubules ([Figure 5](#pbio-0020015-g005){ref-type="fig"}C and [5](#pbio-0020015-g005){ref-type="fig"}E), and aberrant development of the major and mitochondrial derivatives ([Figure 5](#pbio-0020015-g005){ref-type="fig"}F--[5](#pbio-0020015-g005){ref-type="fig"}H). These observations leave it unclear whether cyt-c-d is in any direct sense required for DRICE activation or whether DRICE is required for fertility. We serendipitously identified a line of flies carrying an X chromosome mutation (*driceless*) in which DRICE activation during spermatid individualization was completely eliminated ([Figure 6](#pbio-0020015-g006){ref-type="fig"}A) (see [Materials and Methods](#s4){ref-type="sec"} for details). Testis from these flies contained large cystic bulges in which individualization complexes were present as a coordinated front, as in wild-type ([Figure 6](#pbio-0020015-g006){ref-type="fig"}B). In contrast to *bln^1^* males, *driceless* males were fertile and investment cones moved apically. As expected from this phenotype, some cysts from *driceless* males underwent individualization normally (approximately 50%) ([Figure 6](#pbio-0020015-g006){ref-type="fig"}C). However, in others, individualization failed completely ([Figure 6](#pbio-0020015-g006){ref-type="fig"}D and [6](#pbio-0020015-g006){ref-type="fig"}E). ::: {#pbio-0020015-g005 .fig} Figure 5 ::: {.caption} ###### The *bln^1^* P-Element Insertion, Which Inhibits Cyt-c-d Expression, Results in Pleiotropic Defects in Spermatogenesis \(A) Genomic organization of the cyt-c-d region. Upper half of the panel illustrates the structure of the region, as described by [@pbio-0020015-Arama1]). The lower half of the panel indicates the relative locations of several other genes in the region, as annotated by the Berkeley Drosophila Genome Project (<http://flybase.bio.indiana.edu/search/>) as of August 2002. The *bln^1^* P element is inserted within the cyt-c-d transcription unit. This P element is also inserted within the transcription unit of a second gene, *CR31808-RA* (*RE70695*). Both of these genes and the *bln^1^* P element reside within the intron of a third gene, *CG31782*. (B and D) Wild-type and *bln^1^* testis, respectively, stained with anti-active DRICE antibodies. Active DRICE immunoreactivity is eliminated in *bln^1^* testis, as described in [@pbio-0020015-Arama1]). (C and E) Wild-type and *bln^1^* testis, respectively, stained with AXO49 antibodies (blue), which recognize polyglycylated β2tub present in axonemal microtubules, and phalloidin (red). Polyglycylation occurs prior to individualization ([@pbio-0020015-Bressac1]). Axonemes of elongated cysts from wild-type flies stain with AXO49 (C), while those from *bln^1^* males do not (E). (F--I) EMs of cysts of different developmental stages from wild-type (F and G) and *bln^1^* (H) testis. (F) Wild-type cyst prior to individualization. Note the structures of the major and minor mitochondrial derivatives, in particular the fact that the major mitochondrial derivative is increased in size and is electron dense. (G) Wild-type cyst following individualization. (H) Representative example of the most mature cysts found in *bln^1^* testis. Note the dramatically increased cell size and the lack of differentiation of the major and mitochondrial derivatives, as compared to wild-type. :::  ::: ::: {#pbio-0020015-g006 .fig} Figure 6 ::: {.caption} ###### *driceless* Males Lack Active Drice Staining and Show Defects in Individualization \(A) Testis from *driceless* male stained with active DRICE. Active DRICE staining is eliminated. \(B) Elongated cysts from *driceless* male. AXO49 staining (blue) outlines the location of three cystic bulges. Individualization complexes (arrows) are marked with phalloidin (red). \(C) Example of a cyst from a *driceless* male in which individualization has proceeded normally. \(D) Example of a cyst from a *driceless* male in which individualization has failed to occur. \(E) Boxed area in (D) shown at higher magnification. A region in which individualization has failed is outlined with a dashed line. :::  ::: The above observations indicate that DRICELESS promotes individualization, but leave the role of DRICE (which we have thus far been unable to effectively inactivate with RNAi) unclear. Interestingly, cysts from *driceless* males also showed reduced levels of localized active DRONC staining (data not shown), raising the possibility that DRICELESS has at least some of its effects on individualization through regulation of DRONC activity. We do not favor a simple linear model in which DRICELESS mediates its effects on individualization only by promoting DRONC-dependent activation of DRICE. This is because removal of HID or inhibition of ARK or DRONC, each of which inhibited individualization, had no significant effect on DRICE activation. An attractive alternative is that DRICELESS-dependent activation of DRICE promotes individualization, at least in part, by indirectly facilitating local activation of DRONC and perhaps other caspases, such as DREDD (see [Discussion](#s3){ref-type="sec"} below), that are themselves activated through distinct pathways. Positive feedback pathways that perform a similar caspase-activating function have been described in a number of apoptotic contexts ([@pbio-0020015-Adams1]). DRICE can cleave DIAP1 near its N-terminus. This promotes DIAP1 degradation through the N-end rule ubiquitylation pathway ([@pbio-0020015-Ditzel1]), providing one possible mechanism by which active DRICE could facilitate the activation of other caspases. Characterization of *driceless* should provide insight into the functional relationships between these caspases in spermatogenesis. Discussion {#s3} ========== All together, our observations demonstrate that multiple caspases and caspase regulators, acting at distinct points in space and time, are utilized to promote spermatid individualization. In one pathway, whose mechanism of activation is unknown, active DRICE appears throughout elongated spermatids just as individualization begins. DRICELESS, which promotes individualization, is required for DRICE activation. But whether active DRICE mediates the requirement for DRICELESS is unknown. In a second pathway, HID, concentrated through unknown mechanisms in the cystic bulge, promotes the local ARK-dependent activation of the apical caspase DRONC, presumably at least in part through disruption of complexes between DRONC and DIAP1. As discussed above, active DRICE may facilitate this activation. Components of a second pathway for apical caspase activation, dFADD and DREDD, are also important for individualization. These proteins bind each other ([@pbio-0020015-Hu1]; [@pbio-0020015-Horng1]), and dFADD expression promotes DREDD activation ([@pbio-0020015-Hu1]). Adaptors such as mammalian FADD mediate recruitment of apical caspases to ligand-bound death receptors, thereby promoting caspase activation. Interestingly, dFADD and DREDD are absolutely required for the innate immune response to gram-negative bacterial infection ([@pbio-0020015-Hultmark1]). In this pathway, dFADD-dependent activation of DREDD promotes cleavage and activation of the transcription factor RELISH. DREDD activation is mediated by homophilic death domain interactions between dFADD and IMD (an immune deficiency gene) that occur downstream of the peptidoglyclan recognition protein PGRP-LC receptor binding to bacterial cell wall components ([@pbio-0020015-Hultmark1]). Homophilic death domain interactions also mediate binding of dFADD to the adaptor dMyD88, a component of the Toll receptor-dependent immune response to fungal infection ([@pbio-0020015-Horng1]). It will be interesting to determine whether these or other receptor pathways mediate the requirements for dFADD and DREDD during spermatid individualization. How do caspases contribute to spermatid individualization? Testis from flies mutant for any one of the above pathways (*Ark*, *Dronc*, and *Hid*; *dFadd* and *Dredd*; and *Driceless*) contained cysts in which individualization failed to occur. Interestingly, however, other cysts in the same testis, or from testis of sibling males, carried out individualization apparently normally. Thus, these flies were fertile, though in some cases at a reduced frequency (β2tub-Ark-RNAi and *dFadd^ f02804^*/*dFadd^ f02804^*; *hid* mutants have defects in external genitalia that prevent mating). These observations suggest that no one of these caspase pathways is absolutely required for individualization. The stochastic nature of the defects observed complete failure of individualization in some mutant cysts and apparently normal individualization in others may reflect a requirement for a threshold level of caspase activity, which can be achieved through multiple pathways, or as a result of positive feedback between pathways, in order for a cyst to initiate individualization. Consistent with these possibilities, double mutants between components of the *Ark* and *Dronc* caspase cascade and mutants in the *dFadd* and *Dredd* cascade were almost completely sterile (*Dredd^ B118^*/*Dredd ^B118^*; *Ark^CD4^*/*Ark^CD4^*, 8% fertile, *n* = 12) or completely sterile (*Ark^CD4^*/*Ark^CD4^*; *dFadd ^f02804^*/*dFadd^ f02804^*, *n* = 12), while single mutants for any of these components showed significant fertility (*Dredd^ B118^*/*Dredd^ B118^*, 79% fertile, *n* = 24; *Ark^CD4^*/*Ark^CD4^*, 70% fertile, *n* = 20; *dFadd^ f02804^*/*dFadd^ f02804^*, 71% fertile, *n* = 14). Caspase activity may also participate more directly in processes that mediate encapsulation or cytoplasm elimination. Several observations suggest a role for caspases in at least the latter process. First, in contrast to the situation in wild-type cysts, active DRICE was not effectively swept up into the stunted cystic bulges formed in the presence of caspase inhibitors such as p35 ([Figure 7](#pbio-0020015-g007){ref-type="fig"}) or in other contexts in which caspase activity was inhibited (β2tub-DIAP1, β2tub-Dcp-1-RNAi, β2tub-Ark-RNAi, β2tub-Dn-DRONC, *hid^ 05014^*/*H99*, *dFadd^ f02804^*/*dFadd^ f02804^*; data not shown). Second, spermatids in cysts with decreased levels of ARK often contained large fingers of excess cytoplasm despite the fact that in some cysts membrane encapsulation occurred apparently normally (see [Figure 2](#pbio-0020015-g002){ref-type="fig"}). Together these observations are interesting because they also suggest that the processes of investment cone movement and spermatid encapsulation can be separated from that of cytoplasm elimination. Investment cones carry out a daunting task. They move apically within a cyst for more than 2 mm, sieving and sweeping an ever-increasing body of cytoplasmic organelles, components of the nuclear membrane, nucleoplasm, and bulk cytoplasm in front of them. Little is known about how investment cones function other than that movement is actin-based and that a number of actin-binding proteins are located in or around these structures ([@pbio-0020015-Hicks1]; [@pbio-0020015-Noguchi1]). It is tempting to speculate that spermatid caspase activity functions, at least in part, to free organelles from preexisting attachments, thus facilitating their apical transport. In this way, caspase activity would provide a permissive environment for investment cone movement and cytoplasm removal. More active roles in promoting membrane remodeling or investment cone-dependent force generation or movement, based on spatially restricted cleavage of cytoskeletal components or other proteins, can also be imagined. The identification of caspase substrates will be important in understanding how caspases regulate this process. ::: {#pbio-0020015-g007 .fig} Figure 7 ::: {.caption} ###### Active DRICE Is Eliminated from the Cytoplasm of Wild-Type Spermatids Following Passage of the Individualization Complex, but Not from Spermatids in Which Caspase Activity Has Been Inhibited \(A) Cystic bulge from a wild-type cyst stained with active DRICE (red). The cystic bulge (arrowhead) is moving to the left. Active DRICE staining is absent in areas of the spermatid bundle that the individualization complex has passed through and in which excess cytoplasm has been eliminated (arrow). \(B) Cystic bulge from a β2tub-p35 cyst. The cystic bulge (arrowhead) is decreased in size, and active DRICE is present in areas of the spermatid bundle through which the individualization complex has moved (arrows). These observations suggest caspase inhibition results in at least a partial failure to eliminate excess cytoplasm, but that this is not necessarily associated with lack of movement of the individualization complex. :::  ::: What is the relationship of our observations in *Drosophila* to spermatid differentiation in mammals? During step 18 of murine spermatid differentiation, a lobe of cytoplasm accumulates around the spermatid head. It then separates from the spermatid body and is ultimately phagocytosed by the associated Sertoli cell ([@pbio-0020015-de1]). Separation of this mass, known as the residual body, removes a large volume of spermatid cytoplasm. It also brings about the encapsulation of each spermatid within a single plasma membrane, since the cytoplasmic bridges linking spermatids are between the membrane compartments defined by the residual bodies. Finally, it severs the connection between the spermatid and the Sertoli cell that supported and anchored it, thereby freeing the now-individualized spermatozoa to enter the seminiferous tubule. Residual bodies show several features commonly associated with apoptosis: their plasma membrane binds Annexin V, and they are phagocytosed by Sertoli cells ([@pbio-0020015-Blanco-Rodriguez1]), which also phagocytose apoptotic germ cells (compare [@pbio-0020015-Shiratsuchi1] and references therein). In addition, residual body cytoplasm is condensed and contains elevated levels of CASPASE-1 ([@pbio-0020015-Blanco-Rodriguez1]) and the proapoptotic BCL-2 family member BAK ([@pbio-0020015-Krajewski1]). These observations suggest that, as in *Drosophila*, local activation of apoptotic caspase cascades within late-stage spermatids promotes their individualization and elimination of excess cytoplasm. Mice lacking the proapoptotic proteins APAF-1 or the BLC-2 family member BAX are infertile and have dramatic defects in spermatogenesis ([@pbio-0020015-Knudson1]; [@pbio-0020015-Honarpour1]; [@pbio-0020015-Russell2]). However, these phenotypes are thought to be an indirect consequence of a failure in an earlier, normally occurring postnatal spermatogonial cell death. A test of the importance of caspase activity in mammalian spermatid differentiation will be most directly achieved by determining the consequences of caspase inhibition specifically in these cells. Finally, how is it that elongated spermatids avoid apoptosis in the presence of activated apoptotic caspases for prolonged periods of time? Perhaps the caspase substrates are different from those targeted during apoptosis. But, if so, then what is the basis for the selective targeting? If the targets are the same as those activated during apoptosis, then how is the caspase cascade kept from promoting an apoptotic cell fate? Tight control over the subcellular site of caspase activation (or stabilization of the active caspase), such as we observed with DRONC, provides one possible solution. Others may also exist. In particular, it is important to recognize that while active caspase-specific antibodies recognize caspases that are in the cleaved and therefore activated conformation, these caspases may be kept inactive through interactions with other proteins or as a result of posttranslational modification. *Drosophila* is a powerful system in which to isolate male-sterile mutants (compare [@pbio-0020015-Castrillon1]; [@pbio-0020015-Fuller1]; [@pbio-0020015-Fabrizio1]). It is likely that an exploration of the relationship between the genes identified by these mutations and the apoptotic regulators described here will provide insight into these questions. Materials and Methods {#s4} ===================== {#s4a} ### Fly strains and constructs {#s4a1} All crosses and stocks were maintained at 25°C. The following fly stocks were used: *w1118, Ark^CD4^/Cyo* ([@pbio-0020015-Rodriguez1]), *H99/TM3* ([@pbio-0020015-White1]), *hid^05014^/TM3* ([@pbio-0020015-Grether1]), *dFadd ^f02804^/TM6B* ([@pbio-0020015-Naitza1]), *Dredd ^B118^/FM7* ([@pbio-0020015-Leulier1]), *GMR-Dronc^ F118E^* ([@pbio-0020015-Chai1]), and *bln^1^/Cyo* ([@pbio-0020015-Castrillon1]). *Dronc^ F118E^* contains a mutation that prevents interaction between DRONC and DIAP1. Thus, *Dronc ^F118E^* has enhanced activity in vivo ([@pbio-0020015-Chai1]). The P-element vector pβ2Tub contains sequences from the β2tub locus (85D) sufficient to direct testis germline-specific expression. It was generated by removing an XhoI--EcoRI promoter fragment from pGMR ([@pbio-0020015-Hay2]) and introducing in its place a 340-bp fragment from the β2tub locus ([@pbio-0020015-Santel1]), amplified by PCR with the primers 5′-gcg ctc gag atc ctc tat tgc ttc caa ggc acc and 5′-gcg gaa ttc agt tag ggc ccc ttt ttc aca ccg. Coding region fragments for Dn-DRONC ([@pbio-0020015-Hawkins2]) and DIAP1C422Y (which results in stabilization of DIAP1 by blocking its ability to autoubiquitinate \[[@pbio-0020015-Yoo1]\]) were introduced into pβ2Tub to produce pTub-Dn-DRONC and pTub-DIAP1, respectively. A vector to express double-stranded RNA for ARK was generated as follows. A 900-bp fragment of *Ark* genomic DNA corresponding to the first exon and intron was amplified using primers 5′-gcg gaa ttc ccg aag agg cat cgc gag cat ata cg and 5′-cgc aga tct ata agg ggt gag tgc tcc cag cgg ctc. This was introduced into pβ2Tub using EcoRI and BglII. A second fragment corresponding to the first exon, but in reverse orientation, was amplified using primers 5′-gcg gcg gcc gc gct aac gca ggg tcc ttc gga ggc and 5′-cgc agg cct aag agg cat cgc gag cat ata cgc. This was introduced into the intermediate described above using NotI and StuI, generating pTub-Ark-RNAi. A similar strategy was used to generate pTub-Dcp-1-RNAi. A 540-bp fragment of *Dcp-1* genomic DNA corresponding to the first exon and intron was amplified using primers 5′-ctg ccg gaa ttc ttc gac ata ccc tcg ctg and 5′-cgc gga aga tct gtt gcg cca gga gaa gta g. A second fragment corresponding to the first exon, but in reverse orientation, was amplified using primers 5′-aag gaa aaa a gcg gcc gc cgg aat ggt cga gta gga gaa g and 5′-cgc gga agg cct ttg aaa acc tgg gat tc. Germline transformants of pTub-Dn-DRONC, pTub-DIAP1, and pTub-Ark-RNAi were created using standard procedures. Testis characterized in this paper carried multiple copies of the relevant β2tub expression transgene. These were β2tub-DIAP1, β2tub-p35, and β2tub-Dn-DRONC (four copies); β2tub-Ark-RNAi (three copies); β2tub-Dcp-1-RNAi (six copies). ### Isolation of the *driceless* mutant {#s4a2} We stained testis from *puc^ E69^/TM6B* males ([@pbio-0020015-Martin-Blanco1]) with active DRICE antibodies. These males lacked active DRICE staining, but fully elongated axonemes were present, as visualized by staining with AXO49 antibody. The mutation was mapped to the X chromosome using standard procedures. ### Immunocytochemistry {#s4a3} Conditions for immunocytochemistry and confocal microscopy were as described in [@pbio-0020015-Yoo1]). Palloidin-Alexafluor488 (Molecular Probes Inc., Eugene, Oregon, United States) was used at 1:40 concentration to label filamentous actin; TOTO-3 was used for DNA labeling at 1:10,000 (Molecular Probes Inc.). Antibodies were used at the following concentrations: purified rabbit anti-active DRICE (1:50) ([@pbio-0020015-Yoo1]); purified rabbit anti-DRONC (1:100) (this paper); mouse anti-DIAP1 (1:400) ([@pbio-0020015-Yoo1]); mouse anti-AXO49 (1:5,000) ([@pbio-0020015-Bressac1]), rabbit anti-HID (1:1,000) ([@pbio-0020015-Yoo1]), and purified rabbit anti-active DRONC peptide (1:50) and anti-DCP-1 (1:100) (this paper). Anti-DCP-1 antibodies were produced in rabbits and purified using a C-terminal 6× His-tagged version of the DCP-1 p20 subunit as the immunogen. Anti-DRONC antibodies were raised against the C-terminal fragment of the DRONC large subunit (amino acid residues 336--352; EPVYTAQEEKWPDTQTE), and anti-active DRONC-specific antibodies were raised in rabbits using a synthetic nonapeptide corresponding to residues just N-terminal to the DRONC autoactivation cleavage site E352 (EKWPD*TQTE*), both of which were conjugated with keyhole limpet hemocyanin as the immunogen (Covance Research Products Inc., Richmond, California, United States). Active DRONC-specific antibodies were purified by sequential protein affinity purification. Antisera were first applied to a column bound with full-length inactive DRONC (*Dronc^C318A^*) to eliminate antibodies reactive with uncleaved DRONC. The flowthrough was applied to a DRONC large subunit (residues 1--352) affinity column. Bound proteins were eluted using 100 mM glycine (pH 2.5). These antibodies detect the large fragment of active DRONC (cleaved after E352), but do not recognize full-length DRONC (see [Figure S2](#sg002){ref-type="supplementary-material"}). Anti-DRONC antibodies were purified using full-length inactive DRONC (*Dronc^C318A^*). Western blot analysis to demonstrate binding specificity was carried out with 100 ng of full-length *Dronc^C318A^* and *Dronc^1--352^*. These were detected using purified anti-DRONC peptide (1:100) or purified anti-active DRONC peptide (1:100) antibodies. ### Male fertility tests {#s4a4} Individual male flies were placed with 4- to 5-d-old virgin females in vials for 3 d at 25°C. They were then transferred to fresh vials with four new females and allowed to mate for another 3 d. Males were scored as sterile if they failed to produce progeny by day 6. ### Western blotting of adult testis {#s4a5} Testes extracts were prepared in 50 μl of cell lysis buffer (20 mM HEPES--KOH \[pH 7.6\], 150 mM NaCl, 10% glycerol, 1% Triton X-100, 2 mM EDTA, 1× protease inhibitor cocktail \[Roche, Basel, Switzerland\], and 1 mM DTT) from 30--50 adults of the appropriate genotype. Total protein (70 μg) was used for Western blot analysis using rabbit anti-ARK (1:1,000) (generously provided by Lai Wang and Xiaodong Wang) or purified rabbit anti-DCP-1 (1:100). Filters were stripped using Restore Western blot stripping buffer (Pierce Biotechnology, Rockford, Illinois, United States) and reprobed with rabbit anti-full-length DRICE (1:1,000) ([@pbio-0020015-Dorstyn1]) as a loading control. ### Electron microscopy {#s4a6} Testes were dissected from adult 2- to 4-d-old males raised at 25°C and prepared for EM as described by [@pbio-0020015-Tokuyasu1]). Thin sections were observed and photographed using a Philips 201 transmission electron microscope (Royal Philips Electronics, Eindhoven, The Netherlands) at 80 kV accelerating voltage. Elongated cysts in which spermatids should have been undergoing or have undergone individualization were identified by their central position in the testis as well as the stage of differentiation of major and mitochondrial derivatives ([@pbio-0020015-Tokuyasu1]). At least two to three testes of each genotype were examined. Supporting Information {#s5} ====================== Figure S1 ::: {.caption} ###### Inhibition of *Dcp-1* Prevents Spermatid Individualization \(A) EM section from β2tub-Dcp-1-RNAi testis. Individualization has failed to occur throughout the cyst. \(B) The boxed area in M is shown at a higher magnification. Spermatid units sharing a common cytoplasm are outlined by the dashed line. \(C) Western blot from wild-type (Wt) and β2tub-Dcp-1-RNAi (Dcpi) testis probed with anti-DCP-1 and anti-DRICE antibodies. DCP-1, but not DRICE, levels are greatly reduced in β2tub-Dcp-1-RNAi testis. (552 KB JPEG). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S2 ::: {.caption} ###### Antibodies Specific for Active DRONC \(A) Third instar eye imaginal disc from *GMR-Dronc^ F118E^* larvae stained with purified anti-DRONC peptide antiserum (green). All cells posterior to the morphogenetic furrow labeled with this antiserum, as expected based on the pattern of GMR (glass multimer reporter)-dependent gene expression ([@pbio-0020015-Hay2]). Eye discs from wild-type larvae showed only very low, uniform levels of staining (data not shown). The inset shows a Western blot probed with purified anti-DRONC peptide antiserum. The first lane was loaded with full-length DRONC mutated in its active site (*Dronc^C318A^*). The second lane was loaded with a version of DRONC consisting of only residues 1--352. This protein terminates following glutamate-352, the DRONC autoactivation cleavage site, and is equivalent to the large subunit of cleaved and active Dronc. The anti-DRONC antibodies react well with both proteins. \(B) Third instar eye imaginal disc from *GMR-Dronc ^F118E^* larvae stained with anti-active DRONC antiserum extensively purified to select for antibodies that react only with versions of DRONC that have been cleaved at glutamate-352, as described in the Materials and Methods. Only cells in the most posterior region of the eye disc, which are presumably undergoing apoptosis, react with these purified antibodies. The inset shows a Western blot, similar to that in (A), which was probed with the purified active DRONC-specific antibodies. These antibodies react with the glutamate-352-cleaved version of DRONC, but not with full-length DRONC. (374 KB JPEG). ::: ::: {.caption} ###### Click here for additional data file. ::: Accession Numbers {#s5a} ----------------- The National Center for Biotechnology Information (NCBI) (<http://www.ncbi.nlm.nih.gov/>) accession number for p35 is P08160. The FlyBase (<http://flybase.bio.indiana.edu/search/>) accession numbers of the sequences discussed in this paper are *Ark* (FBgn0024252), cyt-c-d (FBgn0000408), *Dcp-1* (FBgn0010501), *dFadd* (FBgn0038928), *Diap1* (FBgn0003691), *Dredd* (FBgn0020381), *Drice* (FBgn0019972), *Dronc* (FBgn0026404), *Grim* (FBgn0015946), *Hid* (FBgn0003997), and *Reaper* (FBgn0011706). We thank Jules Hoffmann, Bruno Lemaitre, Marco Di Fruscio, Kristen White, John Abrams, Alfonso Martinez-Arias, and Takashi Adachi-Yamada for fly stocks; M. E. Bre for AXO49; Lai Wang and Xiaodong Wang for anti-ARK antibodies; Pat Koen and Jean Edens for assistance with EM; and David Anderson for use of his confocal microscope. This work was supported by grants to BAH from the Ellison Medical Foundation and The Burroughs Wellcome Fund (New Investigators Award) and by National Institutes of Health grant R01 GM057422. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** JRH and BAH conceived and designed the experiments. JRH and SYV performed the experiments. JRH and BAH analyzed the data. JRH, HY, NY, YS, MG, and BAH contributed reagents/materials/analysis tools. JRH and BAH wrote the paper. Academic Editor: Kathryn Miller, Washington University \*To whom correspondence should be addressed. E-mail: <haybruce@its.caltech.edu> β2tub : β~2~-tubulin cyt-c-d : cytochrome c-d DCP-1 : *Drosophila* caspase-1 Dn : dominant negative EM : electron microscopy GMR : glass multimer reporter IAP : inhibitor of apoptosis IMD : immune deficiency PGRP : peptidoglyclan recognition protein RNAi : RNA interference

PubMed Central

2024-06-05T03:55:45.994883

2003-12-15

PMC300884

Proteins may be the workhorse of the cell, but when a cell can synthesize one protein in a matter of minutes, chances are some will become obsolete. Though many proteins put in years of productive service, others quickly outlive their usefulness and can even damage the cell. Proteins that help form bone and muscle, for example, function for years while regulators of mitosis and cell proliferation might finish their jobs in seconds. Such short-timers are soon tagged as superfluous by a chain of small proteins called ubiquitin, which marks the proteins for degradation in an enzyme called the proteasome. Once in the proteasome, these proteins are broken down and can then be recycled for more productive ventures. A massive structure by cellular standards, the proteasome consists of multiple subunits, including a cylindrical core particle called 20S, which catalyzes degradation, and regulatory complexes called 19S caps, which form lid and base structures at both ends of the core. While the structure and biomechanics of the 20S core have been well characterized, much less is known about the functional mechanics of the regulatory complexes. The lid\--base complex recognizes only ubiquitin-tagged proteins, which are then unfolded so they can enter the proteasome. But first ubiquitin chains must be detached from the protein, a task performed by an enzyme in the proteasome called Rpn11 isopeptidase. How the lid--base complex removes the ubiquitin tag, unfolds the protein, and shuttles it into the proteasome\'s core is not clear. Now Raymond Deshaies and colleagues present the structure of a homolog of the 19S lid\'s isopeptidase enzymatic center and provide new insights into these questions. The proteasome Rpn11 subunit contains a key region called the JAMM motif, which Deshaies\' lab has shown previously is required for the proteasome to remove ubiquitin tags. For the work discussed in this paper, the researchers set out to understand how the proteasome strips off ubiquitin tags from proteins about to be destroyed by determining the three-dimensional structure of the JAMM motif. The researchers tested many genes to look for a JAMM-containing protein that would crystallize properly and found one in the heat-loving prokaryote Archaeoglobus fulgidus. After determining the structure of the JAMM protein (called AfJAMM), the researchers discovered that AfJAMM looks nothing like the well-known deubiquitinating enzymes. But the arrangement of a set of amino acids that binds a zinc ion and forms the proposed active site of AfJAMM does resemble that found in a well-known protein-degrading metalloprotease called thermolysin, even though in other respects AfJAMM and thermolysin have very different features. The researchers mutated amino acid residues in another JAMM protein called Csn5 (they expected these residues to be critical for isopeptidase activity as well, based on comparisons of the AfJAMM and thermolysin structures) and found that the residues are indeed important for Csn5 function. These results suggest that JAMM does indeed represent a novel family of metalloproteases. As for the wider function of JAMM proteins, the researchers speculate that these proteins are likely to be involved in a variety of important regulatory systems since they appear in life forms that lack ubiquitin and ubiquitin-like proteins. The crystal structure reported in this paper will provide a valuable tool for investigations into the underlying structural and functional mechanisms of these enzymes. And it may have important therapeutic implications. Proteasome inhibitors are promising anticancer therapies---fighting cancer by blocking machinery required by rapidly dividing cells. In the hopes of developing more targeted therapies, scientists are trying to fine-tune their control of the ubiquitin system and the proteasome. Inhibiting the JAMM domain of enzymes like Csn5, which remove ubiquitin-like tags from proteins upstream of the proteasome, for example, might just do the trick. ::: {#pbio-0020017-g001 .fig} ::: {.caption} ###### The active site of JAMM :::  :::

PubMed Central

2024-06-05T03:55:45.998969

2003-11-24

PMC300885

Bioinformatics and microarrays have given scientists powerful new tools to investigate the structure and activity of genes on a global scale. Rather than studying just a few genes, scientists can analyze tens of thousands within and across species. Microarrays flag which genes are expressed under particular cellular conditions in an organism, while genome sequencing offers clues to gene function and regulation. By comparing the genomic properties of different species, scientists can spot patterns that help them identify functional and regulatory elements, learn about genome structure and organization, and gain a better understanding of the evolutionary forces that shape life on Earth. The potential of these technologies to reveal insights into the fundamental structure and function of biological systems continues to grow along with the wealth of gene sequence and expression data---but the ability to interpret and merge these datasets lags behind the ability to collect them. In an effort to overcome these limitations, Sven Bergmann, Jan Ihmels, and Naama Barkai developed a comparative model that integrates gene expression data with genomic sequence information. Because functionally related genes are expected to be coexpressed in different organisms and because the sequence of some of these functionally related genes may also be conserved between organisms, Bergmann and colleagues hypothesized that "conserved coexpression" could serve as an indicator of gene function on a genomic level. (Conserved genes are those that have changed little since they first evolved. Conserved coexpression describes functionally related genes that are activated together in different species.) But first they had to determine whether coexpression was conserved among species. Analyzing the gene expression profiles of six distantly related organisms---bacteria, yeast, plant, worm, fruitfly, and human---the researchers found that functionally related genes were indeed coexpressed in each species. The most strongly conserved sets of coexpressed genes are associated with core cellular processes or organelles. These results indicate that conserved coexpression can improve the interpretation of genome sequence data by providing another functional indicator for homologous sequences. Since functionally related genes are expressed together in different organisms, it would be reasonable to think their regulatory networks are also conserved. To explore this idea, the researchers grouped coexpressed genes and their regulatory elements into "transcription modules" for each organism. They found significant variation in the number, organization, and relative importance of these modular components. Which components contributed most to an organism\'s global transcription program, for example, depended on the organism. But they also found that the transcription networks are highly clustered---meaning that genes connected to a specific gene are also connected to each other. This finding indicates that gene expression programs, regardless of their size or individual components, are highly modular. Each transcriptome contains modules that have been conserved over time along with "add-on" modules that reflect the needs of a particular species. This modularity supports the notion that variation between and among species arises from the diversity of gene expression programs. Although the regulatory details of individual gene groups varied, the researchers found common ground in the overall landscape of the expression data. The transcription programs exhibit properties typical of dynamically evolving "real-world" networks that are designed to perform in uncertain environments and to maintain connections between elements independent of scale. These properties were originally identified in studies of social networks and the World Wide Web, but they aptly describe the real-world challenges of the cell. Studies of dynamically evolving networks show that nodes (i.e., genes and proteins) added at an early stage (much like highly conserved genes) are more likely to develop many connections, acting as a hub. Following these organizational principles, transcription networks would have a relatively small number of highly connected "hub genes"---though a much higher number than one would expect in a random network. And that is what the authors observed: the networks they constructed from the expression data had the expected number of highly connected hub genes, which tend to be essential and conserved among organisms. Since these highly connected genes are likely to have homologues in other organisms, they can serve as powerful and efficient tools for assigning function to the thousands of uncharacterized sequences found in sequence databases. This model presents a framework to explore the underlying properties that govern the design and function of the cell and provides important clues---in the form of conserved transcription modules---to the evolutionary building blocks that generate diversity. ::: {#pbio-0020018-g001 .fig} ::: {.caption} ###### Regulatory relations among transcription modules :::  :::

PubMed Central

2024-06-05T03:55:45.999543

2003-12-15

PMC300886

You might say that caspases are obsessed with death. The primary agents of programmed cell death, or apoptosis, caspases kill cells by destroying proteins that sustain cellular processes. Apoptosis, a highly controlled sequence of events that eliminates dangerous or unnecessary cells, contributes to a wide variety of developmental and physiological processes---in a developing embryo, apoptosis creates the space between fingers and adjusts nerve cell populations to match the number of cells they target; in an adult, apoptosis counters cell proliferation to maintain tissue size and density. Now it appears that caspases may also play a role in creating life. As Bruce Hay, Jun Huh, and colleagues report, multiple caspases and caspase regulators are required for the proper formation of free-swimming sperm in the fruitfly Drosophila. Caspases, which typically exist in a quiescent state in nearly all cells, are regulated through a complex network of activators and inhibitors. Once activated, a "caspase cascade" ultimately cleaves and irreversibly alters the function of essential cellular proteins, leading to apoptosis. A few of the dozen-plus known caspases appear to contribute to inflammation responses, but the vast majority are enlisted to kill cells. Not surprisingly, cells keep caspase activation under tight wraps. That\'s why it\'s intriguing that multiple caspases normally associated with the induction of cell death participate in this nonapoptotic process. During spermatogenesis, germline precursor cells---the cells that generate sex cells---give rise to 64 haploid spermatids. (Sex cells are haploid, containing half the chromosomes found in body cells.) Spermatids are connected by intracellular "bridges" that, along with most other cytoplasmic components, must be expelled in a process called "individualization" to create terminally differentiated free-swimming sperm. Protein structures known as investment cones surround each spermatid nucleus and sweep out the neighboring cytoplasm, bridges, and organelles, forming a bulge that eventually detaches as a "waste bag" as it reaches the sperm tail. This process---elimination of cytoplasm and membrane packaging of individual spermatids---also occurs in mammals. Many types of human infertility result when it is disrupted. To explore how caspases affect this process, Hay\'s group studied the consequences of inhibiting caspase activity (or the activity of specific caspase activators) in the male germline cells of fruitflies. In both cases, they observed that the bulges and waste bags were either abnormally small or absent and that the normal path of investment cone movement was disrupted. The researchers also inspected the flies to look for structural differences and found that spermatids in both mutant strains remained connected by cytoplasmic bridges and retained residual cytoplasm. Together, the authors conclude, these results demonstrate that individualization depends on caspase activity. Hay\'s team went on to characterize the pathways that activate caspases during sperm individualization. They found that in one pathway, two key activators of caspase-dependent cell death---Ark and Hid (both of which have mammalian counterparts)---promote the activity or stabilization of the caspase Dronc. A second caspase, Dredd, and its activator Fadd (which also have mammalian counterparts) were also found to be important. Double mutants that removed both Dronc and Dredd activity had more severe defects in individualization than mutants that removed only one or the other, suggesting that these caspases have distinct roles in this process. Interestingly, Drice---the downstream caspase activated just as individualization begins (downstream caspases are typically activated by upstream caspases such as Dronc and Dredd)---was not affected by inhibition of Dronc and Dredd. This result, along with the fact that Dronc and Drice were activated at different times and places, suggests that some other mechanism activates Drice. Different apoptosis-related caspases and caspase regulators, the authors conclude, are recruited through different pathways at distinct points in time and space to create individually packaged, free-swimming sperm, a distinctly nonapoptotic process. Studies in mice suggest that individualization may occur similarly in mammals, with activation of apoptotic caspase cascades resulting in free-swimming sperm and loss of specific caspase activators causing infertility and defective spermatogenesis. The abnormal differentiation and residual cytoplasm seen in caspase-inhibited Drosophila mutants, for example, resemble "cytoplasmic droplet sperm," a condition seen in infertile men. Insights into the molecular basis of caspase activation in sperm individuation could provide clues to male infertility and suggest possible treatments. Given the widespread role of programmed cell death in supporting processes fundamental to life, perhaps it\'s not surprising that the agents of apoptosis also support the creation of life. ::: {#pbio-0020034-g001 .fig} ::: {.caption} ###### Developing spermatids in a normal Drosophila testis :::  :::

PubMed Central

2024-06-05T03:55:46.000149

2003-12-15

PMC314300

Introduction {#s1} ============ Since the classic observations of the many histologic similarities between the tumor microenvironment and normal wound healing, it has been proposed that tumor stroma is "normal wound healing gone awry" ([@pbio-0020007-Dvorak1]). During normal wound healing, coagulation of extravasated blood initiates a complex cascade of signals that recruit inflammatory cells, stimulate fibroblast and epithelial cell proliferation, direct cell migration, and induce angiogenesis to restore tissue integrity. Many of these normally reparative processes may be constitutively active in the tumor milieu and critical for tumor engraftment, local invasion, and metastasis to distant organs ([@pbio-0020007-Bissell1]). Indeed, keratinocytes from the wound edge transiently exhibit many similarities to their transformed counterparts in squamous cell carcinomas ([@pbio-0020007-Pedersen1]). Epidemiologically, chronic wound and inflammatory states are well-known risk factors for cancer development: the connection between cirrhosis and liver cancer, gastric ulcers and gastric carcinoma, and burn wounds and subsequent squa-mous cell carcinoma (so-called Majorlin\'s ulcer) are but a few examples. In the genetic blistering disorder recessive dystrophic epidermolysis bullosa, nearly 80% of the patients develop aggressive squamous cell carcinoma in their lifetime ([@pbio-0020007-Mallipeddi1]), attesting to the powerful inductive environment of wounds for cancer development. In recent years, the roles of angiogenesis, extracellular matrix remodeling, and directed cell motility in cancer progression have been intensely studied ([@pbio-0020007-Bissell1]). Nonetheless, a comprehensive molecular view of wound healing and its relationship to human cancer is still lacking. Thus, there is currently no established method to quantify the risk of cancer from wounds diagnostically or to intervene therapeutically. The complete sequence of the human genome and the advent of microarray technology have spurred a revolution in the classification and diagnosis of human cancers ([@pbio-0020007-Golub1]; [@pbio-0020007-Alizadeh1]; [@pbio-0020007-Perou1]; [@pbio-0020007-Sorlie1]; [@pbio-0020007-Van1]; [@pbio-0020007-Ramaswamy1]). By detailing the expression level of thousands of genes simultaneously in tumor cells and their surrounding stroma, gene expression profiles of tumors can provide "molecular portraits" of human cancers. The variations in gene expression patterns in human cancers are multidimensional and typically represent the contributions and interactions of numerous distinct cells and diverse physiological, regulatory, and genetic factors. Although gene expression patterns that correlate with different clinical outcomes can be identified from microarray data, the biological processes that the genes represent and thus the appropriate therapeutic interventions are generally not obvious. In this study, we explore an alternative strategy to infer physiologic mechanisms in human cancers. We began with a gene expression profile derived from a cell culture model of a physiological process. The in vitro expression profile is used to guide interpretation of publicly available gene expression data from human cancers and thereby test a specific hypothesis. In principle, this strategy allows one to connect the controlled and dynamic molecular perturbations possible in vitro with the complex biology of human clinical samples in a comprehensive and quantitative fashion. Fibroblasts are ubiquitous mesenchymal cells in the stroma of all epithelial organs and play important roles in organ development, wound healing, inflammation, and fibrosis. Fibroblasts from each anatomic site of the body are differentiated in a site-specific fashion and thus may play a key role in establishing and maintaining positional identity in tissues and organs ([@pbio-0020007-Chang1]). Tumor-associated fibroblasts have previously been shown to promote the engraftment and metastasis of orthotopic tumor cells of many epithelial lineages ([@pbio-0020007-Elenbaas1]). We previously observed that the genomic response of foreskin fibroblasts to serum, the soluble fraction of coagulated blood, represents a broadly coordinated and multifaceted wound-healing program that includes regulation of hemostasis, cell cycle progression, epithelial cell migration, inflammation, and angiogenesis ([@pbio-0020007-Iyer1]). We hypothesized that if one could identify a canonical gene expression signature of the fibroblast serum response, this signature might provide a molecular gauge for the presence and physiologic significance of the wound-healing process in human cancers. Results {#s2} ======= Identification of a Stereotyped Genomic Response of Fibroblasts to Serum {#s2a} ------------------------------------------------------------------------ We previously observed that the global transcriptional response of fibroblasts to serum integrates many processes involved in wound healing ([@pbio-0020007-Iyer1]). Because fibroblasts from different anatomic sites are distinct differentiated cells with characteristic gene expression profiles ([@pbio-0020007-Chang1]), we investigated whether the genomic responses to serum varied significantly among fibroblasts cultured from different anatomic sites. Fifty fibroblast cultures derived from ten anatomic sites were cultured asynchronously in 10% fetal bovine serum (FBS) or in media containing only 0.1% FBS. Analysis of the global gene expression patterns, using human cDNA microarrays containing approximately 36,000 genes, revealed that although fibroblasts from different sites have distinctly different gene expression programs, they share a stereotyped gene expression program in response to serum ([Figure 1](#pbio-0020007-g001){ref-type="fig"}A). Selection for genes that were concordantly induced or repressed by most types of fibroblasts yielded 677 genes, represented by 772 cDNA probes, of which 611 are uniquely identified by UniGene (<http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=unigene>). This common genomic response to serum includes induction of genes that represent entry into and progression through the cell cycle (e.g., *E2F1*, *FOXM1*, *PTTG1*), induction of cell motility (e.g., *CORO1C*, *FLNC*), extracellular matrix remodeling (*LOXL2*, *PLOD2*, *PLAUR*), cell--cell signaling (*SDFR1*, *ESDN*, *MIF*), and acquisition of a myofibroblast phenotype (e.g., *TAGLN*, *TPM2*, *MYL6*). Analysis of the public Gene Ontology (GO) annotation of the fibroblast serum response genes confirmed a significant enrichment of genes involved in cell proliferation, blood coagulation, complement activation, secretory protein synthesis, angiogenesis, and proteolysis, reflecting the diverse roles that fibroblasts may play during wound healing (Worksheet 9 in [Dataset S2](#sd002){ref-type="supplementary-material"}). ::: {#pbio-0020007-g001 .fig} Figure 1 ::: {.caption} ###### Identification and Annotation of a Common Serum Response in Fibroblasts \(A) The fibroblast common serum response. Genes with expression changes that demonstrate coordinate induction or repression by serum in fibroblasts from ten anatomic sites are shown. Each row represents a gene; each column represents a sample. The level of expression of each gene in each sample, relative to the mean level of expression of that gene across all the samples, is represented using a red--green color scale as shown in the key; gray indicates missing data. Representative genes with probable function in cell cycle progression (orange), matrix remodeling (blue), cytoskeletal rearrangement (red), and cell--cell signaling (black) are highlighted by colored text on the right. Three fetal lung fibroblast samples, cultured in low serum, which showed the most divergent expression patterns among these samples (in part due to altered regulation of lipid biosynthetic genes \[[@pbio-0020007-Chang1]\]), are indicated by blue branches. \(B) Identification of cell cycle-regulated genes in the common serum response signature. The expression pattern of each of the genes in (A) during HeLa cell cycle over 46 h after synchronization by double thymidine block is shown ([@pbio-0020007-Whitfield1]). Transit of cells through S and M phases during the timecourse, verified by flow cytometry, is indicated below. Approximately one-quarter of genes demonstrate a periodic expression patterns and are therefore operationally annotated as cell cycle genes; the remainder of the genes are used in further analyses to define the CSR. \(C) Validation of annotation by temporal expression profiles. Timecourse of gene expression changes in a foreskin fibroblast culture after shifting from 0.1% to 10% FBS is shown. Global gene expression patterns were determined using cDNA microarrays containing 36,000 genes; genes whose transcript levels changed by at least 3-fold during the timecourse and those in (A) are displayed. The cell cycle genes identified in the analysis illustrated in (B) were found to have a distinct temporal expression pattern with coordinate upregulation at 12 h. :::  ::: One of the most consistent and important responses of human cells to serum is proliferation. Abnormal cell proliferation is also a consistent characteristic of cancer cells, irrespective of any possible involvement of a wound-healing response. We therefore sought to eliminate the contributions of genes directly related to cell proliferation, to improve the specificity of a genomic signature of the fibroblast serum response. To identify features directly related to cell cycle progression, we examined the expression pattern of these 677 genes during the cell cycle (in HeLa cells) ([@pbio-0020007-Whitfield1]). Despite the well-known role of serum as a mitogen, only one-quarter (165 out of 677 genes) of the fibroblast serum response genes showed periodic expression during the cell cycle ([Figure 1](#pbio-0020007-g001){ref-type="fig"}B). The majority of the genes whose expression levels in fibroblasts showed the most consistent response to serum exposure do not appear simply to reflect cell growth or division; these 512 serum-responsive and cell cycle-independent genes are operationally defined as the fibroblast core serum response (CSR). Comparison of the common fibroblast serum response with a detailed analysis of the temporal program of gene expression following serum exposure in foreskin fibroblasts confirmed that the cell cycle genes and the CSR have distinct temporal profiles during serum stimulation and are thus distinguishable biological processes ([Figure 1](#pbio-0020007-g001){ref-type="fig"}C). Expression of Fibroblast CSR in Human Cancers {#s2b} --------------------------------------------- Because serum (as distinct from plasma and normal extracellular fluid) is encountered in vivo only at sites of tissue injury or remodeling and induces in fibroblasts a gene expression response suggestive of wound healing, we reasoned that expression of fibroblast CSR genes in tumors might gauge the extent to which the tumor microenvironment recapitulates normal wound healing. We examined the expression of genes comprising the fibroblast CSR in publicly available microarray data from a variety of human cancers and their corresponding normal tissues. To facilitate visualization and analysis, we organized the gene expression patterns and samples by hierarchical clustering ([@pbio-0020007-Eisen1]). Remarkably, we observed a predominantly biphasic pattern of expression for the fibroblast CSR in diverse cancers, including breast cancers, lung cancers, gastric cancers, prostate cancers, and hepatocellular carcinoma. Expression levels of genes that were activated by serum in fibroblasts varied coordinately in tumors, and genes that were repressed by serum in fibroblasts were mostly expressed in a reciprocal pattern ([Figure 2](#pbio-0020007-g002){ref-type="fig"}). ::: {#pbio-0020007-g002 .fig} Figure 2 ::: {.caption} ###### Survey of Fibroblast CSR Gene Expression in Human Cancers Expression patterns of available CSR genes in over 500 tumors and corresponding normal tissues were extracted, filtered as described in [Materials and Methods](#s4){ref-type="sec"}, and organized by hierarchical clustering. The response of each gene in the fibroblast serum response is shown on the right bar (red shows activated; green shows repressed by serum). The strong clustering of the genes induced or repressed, respectively, in fibroblasts in response to serum exposure, based solely on their expression patterns in the tumor samples, highlights their coordinate regulation in tumors. The dendrograms at the top of each data display represent the similarities among the samples in their expression of the fibroblast CSR genes; tumors are indicated by black branches, normal tissue by green branches. :::  ::: In each of the tumor types examined, the expression pattern of the fibroblast CSR genes in normal tissues closely approximated that seen in quiescent fibroblasts cultured in the absence of serum ([Figure 2](#pbio-0020007-g002){ref-type="fig"}). In prostate and hepatocellular carcinomas, all of the normal tissue samples had the serum-repressed signature and almost all of the tumors had the serum-induced signature, albeit with varying amplitude. In breast, lung, and gastric carcinomas, the common fibroblast serum response signature was clearly evident in some of the tumors and apparently absent in others, suggesting that a "wound-healing phenotype" was a variable feature of these cancers. We therefore classified breast, lung, and gastric cancer samples based on the pattern of expression of the genes that comprise the fibroblast CSR. Link between the Gene Expression Signature of Fibroblast Serum Response and Cancer Progression {#s2c} ---------------------------------------------------------------------------------------------- To investigate the stability and consistency of the serum response signature in individual tumors and to explore its clinical implications, we examined CSR gene expression in a group of locally advanced breast cancers with extensive clinical and molecular data ([@pbio-0020007-Perou1]; [@pbio-0020007-Geisler1]; [@pbio-0020007-Sorlie1]). As shown in [Figure 3](#pbio-0020007-g003){ref-type="fig"}A, the expression profiles of the CSR genes were biphasic, allowing a natural separation of these tumors into two classes. Interestingly, in 18 out of 20 paired tumor samples obtained from the same patients before and after excisional biopsy and chemotherapy, the CSR expression phenotypes were consistent between the two samples. Thus, the wound-related expression program appears to be an intrinsic property of each tumor and not easily extinguished. In a set of 51 patients with clinically matched disease and equivalent treatment ([@pbio-0020007-Sorlie1]), primary tumors with the activated CSR signature were significantly more likely to progress to metastasis and death in a 5-y follow-up period (*p* = 0.013 and 0.041, respectively) ([Figure 3](#pbio-0020007-g003){ref-type="fig"}B). Using an alternative analytic approach, classifying each sample by the Pearson correlation between tumor and fibroblast expression patterns of the fibroblast CSR genes, also reproduced the identification of two classes of samples with differing clinical outcomes (Worksheet 2 in [Dataset S2](#sd002){ref-type="supplementary-material"}). A gene expression pattern similar to the serum-activated program of fibroblasts is thus a powerful predictor of prognosis. Other significant prognostic factors in these same patients include tumor grade, estrogen receptor status, and tumor subtype based on gene expression profile ([@pbio-0020007-Geisler1]; [@pbio-0020007-Sorlie1]). Tumor stage, lymph-node status, and p53 status were not statistically significant predictors of survival in these patients (*p* = 0.13, 0.79, 0.05, respectively). A "basal-like" subtype of breast cancer, characterized by molecular similarities of the tumor cells to basal epithelial cells of the normal mammary duct and associated with a particularly unfavorable prognosis ([@pbio-0020007-Sorlie1]), was significantly associated with a gene expression pattern resembling the fibroblast CSR: six of seven basal-like breast cancers had the "serum-activated" gene expression signature (*p* = 0.0075, Fisher\'s exact test). Thus, the presence or absence of the wound-like phenotype may be linked to intrinsic features of the tumor cells. ::: {#pbio-0020007-g003 .fig} Figure 3 ::: {.caption} ###### Context, Stability, and Prognostic Value of Fibroblast CSR in Breast Cancer \(A) Expression patterns of CSR genes in a group of breast carcinomas and normal breast tissue previously described in [@pbio-0020007-Perou1]). Genes and samples were organized by hierarchical clustering. The serum response of each gene is indicated on the right bar (red shows induced; green shows repressed by serum). Note the biphasic pattern of expression that allows each tumor sample to be classified as "activated" or "quiescent" based on the expression of the CSR genes. The previously identified tumor phenotype (color code) and p53 status (solid black box shows mutated; white box shows wild-type) are shown. Pairs of tumor samples from the same patient, obtained before and after surgery and chemotherapy, are connected by black lines under the dendrogram. Two primary tumor--lymph node metastasis pairs from the same patient are connected by purple lines. \(B) Kaplan--Meier survival curves for the two classes of tumors. Tumors with serum-activated CSR signature had worse disease-specific survival and relapse-free survival compared to tumors with quiescent CSR signature. Similar results were obtained whether performing classification using all breast tumors in this dataset or just the 58 tumors from the same clinical trial ([@pbio-0020007-Sorlie1]). :::  ::: We considered the possibility that the observed phenomenon may be simply a reflection of the number of fibroblasts in tumor samples. Perhaps tumors that are infiltrative or otherwise worrisome clinically would demand a wide margin of excision that would include more fibroblasts in the resultant samples. However, classification of breast cancers using the top 1% most highly expressed fibroblast genes (which include a number of extracellular matrix genes and have been previous observed as the "stroma signature" \[[@pbio-0020007-Perou1]\]) showed no relationship between the generic fibroblast signature and clinical outcome (*p* = 0.75; Worksheet 1 in [Dataset S2](#sd002){ref-type="supplementary-material"}). Thus, the prognostic value of the fibroblast CSR likely reflects the physiologic state of the tumor microenvironment and not just the number of fibroblasts in tumor stroma. Similarly, although the mitotic index is an established criterion of tumor grade, classification of these tumors based on expression of cell cycle genes (specifically, all S and G2/M phase genes identified by Whitfield et al. \[2002\]) only had moderate prognostic value (*p* = 0.08; Worksheet 1 in [Dataset S2](#sd002){ref-type="supplementary-material"}). This result also suggests that the prognostic value of the fibroblast CSR is unlikely to be accounted for by the incomplete annotation and removal of genes representing cell growth or division. To extend and validate these results, we tested the prognostic power of the fibroblast CSR signature in independent datasets and different kinds of human cancer ([Figure 4](#pbio-0020007-g004){ref-type="fig"}). Using published DNA microarray data from a study of gene expression patterns in a group of 78 early (tumor smaller than 5 cm, stage I and IIA) breast cancer patients (van \'t Veer et al. 2002), we could segregate the patients into two groups based on expression of the fibroblast CSR genes in the biopsy samples. Tumors with the serum-induced signature had a significantly increased risk of metastasis over 5 y (*p* = 0.00046) ([Figure 4](#pbio-0020007-g004){ref-type="fig"}A). Multivariate Cox proportional hazard analysis confirmed that the CSR classification is a significant independent predictor (*p* = 0.009); the serum-induced gene expression signature was associated with a 3.3-fold relative risk of breast cancer metastasis within 5 y of diagnosis. In the two breast cancer datasets examined, approximately 50% of the CSR genes demonstrated significant differences in expression between the activated and quiescent groups of samples, but permutation and 10-fold balanced leave-one-out analyses revealed that the correct classification can be accomplished using as few as 6% of CSR genes (Worksheets 10--12 in [Dataset S2](#sd002){ref-type="supplementary-material"}). Thus, the expression pattern of the CSR genes provides a robust basis for predicting tumor behavior. Similarly, in analysis of published DNA microarray data from 62 patients with stage I and II lung adenocarcinomas ([@pbio-0020007-Bhattacharjee1]), tumors with the serum-induced signature were associated with significantly higher risk of death compared to tumors with the serum-repressed signature (*p* = 0.021) ([Figure 4](#pbio-0020007-g004){ref-type="fig"}B). These results suggest that presence or absence of a wound-like phenotype in these cancers, with its prognostic implication for their metastatic potential, may be determined at an early stage in their development. In a second, independent group of lung adenocarcinomas of all stages ([@pbio-0020007-Garber1]), tumors with the fibroblast serum-induced signature were associated with a significantly worse prognosis (*p* = 0.0014) ([Figure 4](#pbio-0020007-g004){ref-type="fig"}C). A significant correlation between advanced stage and the serum-induced signature was also apparent in this dataset. Finally, in 42 patients with stage III gastric carcinomas, all treated with gastrectomy alone ([@pbio-0020007-Leung1]), tumors with the activated CSR signature were again associated with shorter survival (*p* = 0.02) ([Figure 4](#pbio-0020007-g004){ref-type="fig"}D). These results suggest that a wound-healing phenotype, reflected in the expression of a set of serum-inducible genes in fibroblasts, is strongly linked to progression of diverse human carcinomas and can provide valuable prognostic information even at an early stage in the natural history of a cancer. ::: {#pbio-0020007-g004 .fig} Figure 4 ::: {.caption} ###### Prognostic Value of Fibroblast CSR in Epithelial Tumors Kaplan--Meier survival curves of tumors stratified into two classes using the fibroblast CSR are shown for stage I and IIA breast cancer (van \'t Veer et al. 2002) (A), stage I and II lung adenocarcinoma ([@pbio-0020007-Bhattacharjee1]) (B), lung adenocarcinoma of all stages ([@pbio-0020007-Garber1]) (C), and stage III gastric carcinoma ([@pbio-0020007-Leung1]) (D). :::  ::: For many other cancers, simple stratification based on expression of genes in the fibroblast CSR gene set is unlikely to be predictive of outcome. The dramatic differences in cellular composition and architecture among the tissues in which cancers can arise may influence the role that a wound-healing response can play in their progression. For example, lymphoma cells proliferate in the specialized microenvironment of lymph nodes and bone marrow, and the "stromal" cells in the central nervous system, predominantly astrocytes and microglia, are markedly different from those associated with most epithelial tissues. Indeed, in our initial analysis, the pattern of expression of the fibroblast CSR genes failed to stratify the outcomes in diffuse large B-cell lymphoma ([@pbio-0020007-Rosenwald1]), medulloblastoma ([@pbio-0020007-Pomeroy1]), and glioblastoma multiforme (M. Diehn and P. O. Brown, unpublished data). Histological Architecture of CSR Gene Expression in Tumors {#s2d} ---------------------------------------------------------- Both to validate the DNA microarray results and to investigate the histological architecture of CSR gene expression in tumors, we examined the expression patterns of five CSR genes implicated in extracellular matrix remodeling and cell--cell interaction, using tissue microarrays containing hundreds of breast carcinoma tissues. *PLAUR*, also known as urokinase-type plasminogen activator receptor, is a well-characterized receptor for matrix-degrading proteases that has been implicated in tumor cell invasion ([@pbio-0020007-Blasi1]; [@pbio-0020007-Sidenius1]). *LOXL2* is a member of a family of extracellular lysyl oxidases that modify and cross-link collagen and elastin fibers ([@pbio-0020007-Akiri1]). *PLOD2* is a member of the lysyl hydroxylase family that plays important roles in matrix cross-linking and fibrosis ([@pbio-0020007-Van1]). *SDFR1*, previously named gp55 and gp65, encodes a cell surface protein of the immunglobulin superfamily that regulates cell adhesion and process outgrowth ([@pbio-0020007-Clarke1]; [@pbio-0020007-Wilson1]). *ESDN* is a neuropilin-like cell surface receptor that was also previously found to be upregulated in metastatic lung cancers ([@pbio-0020007-Koshikawa1]). All five of these genes were included in the fibroblast CSR gene set by virtue of their induction by serum in fibroblasts (see [Figure 1](#pbio-0020007-g001){ref-type="fig"}). Anti-PLAUR antibody is commercially available and served as a positive control. We prepared specific riboprobes for *LOXL2* and *SDFR1* and generated affinity-purified anti-peptide antibodies to PLOD2 and ESDN to detect the predicted protein products. As shown in [Figure 5](#pbio-0020007-g005){ref-type="fig"}, PLAUR, *LOXL2,* PLOD2, and ESDN were not detectably expressed in normal breast tissue; *SDFR1* was expressed at a low level in normal breast epithelial cells (*n* = 11). In contrast, all five genes were induced in a significant fraction of invasive ductal carcinomas of the breast. As previously reported ([@pbio-0020007-Costantini1]), PLAUR protein is expressed in both tumor cells and peritumoral stroma (70 out of 96, 73% positive) ([Figure 5](#pbio-0020007-g005){ref-type="fig"}). PLOD2 protein and *SDFR1* mRNA were detected in breast carcinoma cells and in a small but consistent fraction of peritumor stroma cells (78 out of 100, 78% positive, and 55 out of 79, 70% positive, respectively). ESDN protein was detected exclusively in breast carcinoma cells (69 out of 112, 62% positive). In contrast, *LOXL2* mRNA was abundant in peritumoral fibroblasts around invasive carcinomas (45 out of 106, 42% positive). LOXL2 protein has been previously reported to be expressed in normal mammary ducts and increased in invasive breast carcinoma cells ([@pbio-0020007-Akiri1]). Our data suggest that LOXL2 is primarily synthesized by peritumoral fibroblasts, but may act on or in the vicinity of epithelial cells during tissue remodeling. Collectively, these results suggest that the pathophysiology represented by expression of the fibroblast CSR genes in cancers represents a multicellular program in which the tumor cells themselves, tumor-associated fibroblasts, and perhaps diverse other cells in the tumor microenvironment are active participants. ::: {#pbio-0020007-g005 .fig} Figure 5 ::: {.caption} ###### Histological Architecture of CSR Gene Expression in Breast Cancer Representative ISH of *LOXL2* and *SDFR1* and IHC of PLOD2, PLAUR, and ESDN are shown (magnification, 200×). Panels for *LOXL2*, PLAUR, PLOD2, and ESDN represent cores of normal and invasive ductal breast carcinoma from different patients on the same tissue microarray. Panels for *SDFR1* demonstrate staining in adjacent normal and carcinoma cells on the same tissue section. Arrows highlight spindle-shaped stromal cells that stain positive for *SDFR1* and PLOD2. No signal was detected for the sense probe for ISH or for control IHC without the primary antibody. :::  ::: Discussion {#s3} ========== The remarkable ability of a single physiological fluid---serum---to promote the growth and survival of diverse normal and cancer cells in culture suggests that there may be a conserved, programmed response to the molecular signals that serum provides. In vivo, serum as a physiological signal has a very specific meaning: cells encounter serum---the soluble fraction of coagulated blood---only in the context of a local injury. In virtually any tissue, a rapid, concerted multicellular response, with distinct physiological exigencies that evolve over minutes, hours, and days, is required to preserve the integrity of the tissue and often the survival of the organism. In response to a wound, many of the normal differentiated characteristics of the cells in the wounded tissue are temporarily set aside in favor of an emergency response. In wound repair, as in cancer, cells that ordinarily divide infrequently are induced to proliferate rapidly, extracellular matrix and connective tissues are invaded and remodeled, epithelial cells and stromal cells migrate, and new blood vessels are recruited. In all these respects, a wound response---and the characteristic physiological response to serum---would appear to provide a highly favorable milieu for cancer progression. We defined a stereotyped genomic expression response of fibroblasts to serum, which reflects many features of the physiology of wound healing. When we examined the expression of these genes in human tumors, we found strong evidence that a wound-like phenotype was variably present in many common human cancers (including many that are not known to be preceded by chronic wounds) and was a remarkably powerful predictor of metastasis and death in several different carcinomas. The proposed link between the fibroblast serum response signature and cancer progression raises many questions for additional studies. Perhaps most importantly, our results do not allow us to distinguish whether the wound-like phenotype has a functionally important role in tumor progression or merely serves as a marker for the underlying propensity of a cancer to progress and metastasize. However, at least three genes induced in the fibroblast serum response, *PLAUR*, *LOXL2*, and *MIF*, have been previously shown to increase cancer invasiveness or angiogenesis in animal xenograft models; each of these three genes has also been shown to play an important role in wound healing ([@pbio-0020007-Akiri1]; [@pbio-0020007-Nishihira1]; [@pbio-0020007-Sidenius1]). Thus, we are inclined to believe that coordinate induction of a wound-healing program in carcinomas contributes to tumor invasion and metastasis. Several potential mechanisms might contribute to the wound-like gene expression pattern in cancers. In some cancers, ongoing local tissue injury, resulting from growth and dysfunctional behavior of the tumor cells, could continuously trigger a normal wound-healing response. The classic observation of deposited fibrin products in human tumors is consistent with this model ([@pbio-0020007-Dvorak1]). Inflammatory cells, presumably recruited by tissue disorder, may amplify the wound response and contribute to tumor invasion in part by expression of metalloproteinases ([@pbio-0020007-Coussens1]; [@pbio-0020007-Daniel1]). The wound response might also be initiated directly by signals from the tumor cells ([@pbio-0020007-Fukumura1]), whose ability to activate an inappropriate wound-healing response---favorable to cell proliferation, invasion, and angiogenesis---might be strongly selected during cancer progression. The possibility that stromal cells might play a primary role in promoting a wound-like phenotype in some cancers is raised by studies showing that tumor-associated fibroblasts can enhance tumor engraftment and metastasis in animal models ([@pbio-0020007-Elenbaas1]) and the demonstration in some cancers of genotypic abnormalities in tumor-associated fibroblasts ([@pbio-0020007-Kurose1]). Heterotopic interaction experiments, genetic models, and cell-culture models should enable these and other possible mechanisms to be investigated. Our results illustrate the power of using gene expression data from specific cells or physiological and genetic manipulations to build an interpretive framework for the complex gene expression profiles of clinical samples ([@pbio-0020007-Lamb1]). Several prognostic models based on gene expression patterns have previously been identified from systematic DNA microarray profiles of gene expression in human cancers. Some of these prognostic gene expression profiles appear to reflect the developmental lineage of the cancer cells ([@pbio-0020007-Alizadeh1]; [@pbio-0020007-Sorlie1]; [@pbio-0020007-Pomeroy1]), some appear to reflect the activity of specific molecular determinants of tumor behavior (e.g., the activity of *PLA2G2A* in gastric cancer \[[@pbio-0020007-Leung1]\]), while still others represent the mechanistically agnostic results of machine-assisted learning ([@pbio-0020007-Van1]; [@pbio-0020007-Ramaswamy1]). Although they serve to identify many of the same tumors with unfavorable prognosis, the genes that define the fibroblast CSR overlap minimally with the genes previously used to predict outcome in the same cancers. For example, the fibroblast CSR involves only 20 out of 456 genes in an "intrinsic gene list" that can serve to segregate breast cancers into prognostically distinct groups ([@pbio-0020007-Perou1]) and four out of 128 genes that define the general metastasis signature reported by [@pbio-0020007-Ramaswamy1]). Only 11 genes are in common between the 231 gene van\'t Veer poor prognosis signature for breast cancer ([@pbio-0020007-Van1]) and the fibroblast CSR genes. The prognostic power of these different sets of genes illustrates the multidimensional variation in the gene expression programs in cancers and the complex interplay of many distinct genetic and physiological factors in determining the distinctive biology of each individual tumor. Our success in discovering a significant new determinant of cancer progression, using previously published and publicly available data, illustrates the richness of the data as a continuing source for future discoveries and the importance of unrestricted access to published research data ([@pbio-0020007-Roberts1]). The signals and regulatory systems that normally initiate, sustain, and eventually shut down the physiological response to a wound remain to be identified and understood. Identification of the molecular control mechanisms in this pathway may pave the way to new cancer therapies or chemopreventative agents. For example, cyclooxygenase 2 is strongly induced in the response of fibroblasts to serum ([@pbio-0020007-Iyer1]), and platelet-derived growth factor is one of the principal molecular signals and mitogenic factors in serum. Platelet-derived growth factor receptor and cyclooxygenase 2 are inhibited by imatinib mesylate and nonsteroidal anti-inflammatory agents, respectively---two drugs with established efficacy in treating or preventing cancer ([@pbio-0020007-Bergers1]; [@pbio-0020007-Huls1]). Whether these or other small molecules might derive significant activity against cancer from their ability to inhibit a dysregulated wound-healing response will be an important question for future investigation. Materials and Methods {#s4} ===================== {#s4a} ### Cells and tissue culture {#s4a1} Human primary fibroblasts from ten anatomic sites were cultured in 0.1% versus 10% FBS, as previously described ([@pbio-0020007-Chang1]). For the serum induction timecourse, foreskin fibroblasts CRL 2091 (American Type Culture Collection \[ATCC\], Manassas, Virginia, United States) were serum-starved for 48 h and harvested at the indicated timepoints after switching to media with 10% FBS, essentially as described in [@pbio-0020007-Iyer1]). ### Microarray procedures {#s4a2} Construction of human cDNA microarrays containing approximately 43,000 elements, representing approximately 36,000 different genes, and array hybridizations were as previously described ([@pbio-0020007-Perou1]). mRNA was purified using FastTrack according to the manufacturer\'s instructions (Invitrogen, Carlsbad, California, United States). For the serum timecourse, RNA from all of the sampled timepoints were pooled as reference RNA to compare with RNA from individual timepoints as described in [@pbio-0020007-Iyer1]). ### Data analysis {#s4a3} For defining a common serum response program in fibroblasts, global gene expression patterns in 50 fibroblast cultures derived from ten anatomic sites, cultured in the presence of 10% or 0.1% FBS, were characterized by DNA microarray hybridization ([@pbio-0020007-Chang1]). We selected for further analysis genes for which the corresponding array elements had fluorescent hybridization signals at least 1.5-fold greater than the local background fluorescence in the reference channel, and we further restricted our analyses to genes for which technically adequate data were obtained in at least 80% of experiments. These filtered genes were then analyzed by the multiclass Significance Analysis of Microarrays (SAM) algorithm ([@pbio-0020007-Tusher1]) to select a set of genes whose expression levels had a significant correlation with the presence of serum in the medium, with a false discovery rate (FDR) of less than 0.02%. The corresponding expression patterns were organized by hierarchical clustering ([@pbio-0020007-Eisen1]). Genes that were coordinately induced or repressed in response to serum in most samples (Pearson correlation, greater than 90%) were identified. This set of 677 genes, represented by 772 cDNA probes, of which 611 are uniquely identified by UniGene (<http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=unigene>), was termed the common fibroblast serum response gene set. To identify the subset of these 677 genes whose variation in expression was directly related to cell cycle progression, we compared this set of genes to a published set of genes periodically expressed during the HeLa cell cycle ([@pbio-0020007-Whitfield1]). Because both datasets were generated using similar cDNA microarrays, we tracked genes by the IMAGE number of the cDNA clones on the microarrays. The majority of the genes in the fibroblast serum response gene set showed no evidence of periodic expression during the HeLa cell cycle. One hundred sixty-five genes, represented by 199 cDNA clones, overlapped with the cell cycle gene list; the remaining 512 genes, represented by 573 clones, of which 459 are uniquely identified in UniGene, was termed the CSR gene set. The patterns of expression in human tumors of the 512 genes of the fibroblast CSR gene set were analyzed using data from published tumor expression profiles. Detailed methods and primary datasets are available as Datasets [S1](#sd001){ref-type="supplementary-material"} and [S2](#sd002){ref-type="supplementary-material"} and on our Web site (<http://microarray-pubs.stanford.edu/wound>). We used the Unigene unique identifier (build 158, release date January18, 2003) to match genes represented in different microarray platforms. For cDNA microarrays, genes with fluorescent hybridization signals at least 1.5-fold greater than the local background fluorescent signal in the reference channel (Cy3) were considered adequately measured and were selected for further analyses. For Affymetrix data, signal intensity values were first transformed into ratios, using for each gene the mean values of the normalized fluorescence signals across all the samples analyzed as the denominators ([@pbio-0020007-Bhattacharjee1]). The genes for which technically adequate measurements were obtained from at least 80% of the samples in a given dataset were centered by mean value within each dataset, and average linkage clustering was carried out using the Cluster software ([@pbio-0020007-Eisen1]). In each set of patient samples, the samples were segregated into two classes based on the first bifurcation in the hierarchical clustering dendrogram. For the datasets shown, the clustering and reciprocal expression of serum-induced and serum-repressed genes in the tumor expression data allowed two classes to be unambiguously assigned. Samples with generally high levels of expression of the serum-induced genes and low levels of expression of the serum-repressed genes were classified as "activated"; conversely, samples with generally high levels of expression of serum-repressed genes and low levels of expression of the serum-induced genes were classified as "quiescent." Survival analysis by a Cox--Mantel test was performed in the program Winstat (R. Fitch Software). ### In situ hybridization and immunohistochemistry {#s4a4} Digoxigenin-labeled sense and antisense riboprobes for *LOXL2* and *SDFR1* were synthesized using T7 polymerase-directed in vitro transcription ([@pbio-0020007-Iacobuzio-Donahue1]). Sense and antisense riboprobes for *SDFR1* were made from nucleotides 51--478 of IMAGE clone 586731 (ATCC \#745139), corresponding to the last 388 nucleotides of the 3′ end of the coding sequence and 39 nucleotides of the 3′ untranslated region. Sense and antisense riboprobes for *LOXL2* were made from nucleotides 41--441 of IMAGE clone 882506 (ATCC \#1139012), corresponding to the 3′ end of the coding sequence. In situ hybridization (ISH) results were considered to have appropriate specificity when we observed a strong, consistent pattern of hybridization of the antisense probe and little or no hybridization of the corresponding sense probe. Immunohistochemical (IHC) staining was performed using Dako (Glostrup, Denmark) Envision Plus following the manufacturer\'s instructions. Anti-PLAUR antibody against whole purified human uPA--receptor protein (AB8903; Chemicon, Temecula, California, United States) was used at 1:200 dilution. Affinity-purified polyclonal antibody to PLOD2 was produced by immunizing rabbits with peptides EFDTVDLSAVDVHPN, coupled to keyhole limpet hemocyanin (KLH) (Applied Genomics, Inc., Sunnyvale, California, United States); affinity-purified antiserum was used for IHC at 1:25,000 dilution. Similarly, affinity-purified polyclonal antibody to ESDN was produced by immunizing rabbits with peptide DHTGQENSWKPKKARLKK coupled to KLH (Applied Genomics, Inc.) and used for IHC at 1:12,500 dilution. High-density tissue microarrays containing tumor samples were constructed as described in [@pbio-0020007-Kononen1]). ISH ([@pbio-0020007-Iacobuzio-Donahue1]) and IHC ([@pbio-0020007-Perou1]) were as reported. ISH and IHC images and data were archived as described in [@pbio-0020007-Liu1]). Supporting Information {#s5} ====================== [Figure 1](#pbio-0020007-g001){ref-type="fig"}A can be interactively explored at <http://microarray-pubs.stanford.edu/wound/>. Raw datasets and all supporting data are also available at <http://microarray-pubs.stanford.edu/wound/>. Dataset S1 ::: {.caption} ###### Detailed Bioinformatic Methods Provides a description of microarray datasets, cross-platform mapping and data normalization, classification of cancers by fibroblast CSR genes and correlated clinical outcomes (Worksheets 1--8 in [Dataset S2](#sd002){ref-type="supplementary-material"}), the top 1% fibroblast genes in breast cancer prognosis (see Worksheet 1 in [Dataset S2](#sd002){ref-type="supplementary-material"}), cell cycle S and G2/M genes in breast cancer prognosis (see Worksheet 1 in [Dataset S2](#sd002){ref-type="supplementary-material"}), analysis of GO annotations of fibroblast serum response genes (see Worksheet 9 in [Dataset S2](#sd002){ref-type="supplementary-material"}), and the minimum number of CSR genes necessary for tumor classification (see Worksheets 10--12 in [Dataset S2](#sd002){ref-type="supplementary-material"}). (120 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S2 ::: {.caption} ###### Supporting Data Excel Worksheets of clinical and microarray data, as described in [Dataset S1](#sd001){ref-type="supplementary-material"}. (736 KB XLS). ::: ::: {.caption} ###### Click here for additional data file. ::: Accession Numbers {#s5a} ----------------- The Locus Link (<http://www.ncbi.nlm.nih.gov/LocusLink/>) accession numbers for the genes discussed in this paper are *CORO1C* (Locus Link ID 23603), *E2F1* (Locus Link ID 1869), *ESDN* (Locus Link ID 131566), *FLNC* (Locus Link ID 2318), *FOXM1* (Locus Link ID 2305), *LOXL2* (Locus Link ID 4017), *MIF* (Locus Link ID 4282), *MYL6* (Locus Link ID 4637), *PLAUR* (Locus Link ID 5329), *PLOD2* (Locus Link ID 5352), *PTTG1* (Locus Link ID 9232), *SDFR1* (Locus Link ID 27020), *TAGLN* (Locus Link ID 6876), and *TPM2* (Locus Link ID 7169). The accession numbers of the Gene Ontology (GO) (<http://www.geneontology.org/>) terms that appear in Dataset S1 are angiogensis (GO:0001525), blood coagulation (GO:0007596), complement activation (GO:0006956), immune response (GO:0006955), N-linked glycosylation (GO:0006487), protein translation (GO:0006445), and proteolysis and peptidolysis (GO:0006508). We thank M. Diehn and J. Lapointe for sharing unpublished data; M. Whitfield for help with cell cycle data; T. Hastie for assistance with statistical analysis; members of Brown and Botstein labs for thoughtful discussions; D. Ross and Applied Genomics, Inc., for antisera; and the staff of the Stanford Microarray Database and Stanford Functional Genomic Facility for support. This work was supported by National Institutes of Health grants T32 AR07422 (HYC), CA77097 (POB), and CA85129 (POB and DB) and by a National Science Foundation Predoctoral Fellowship (JBS); POB is an investigator of the Howard Hughes Medical Institute. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** HYC, DB, and POB conceived and designed the experiments. HYC, JBS, RS, and KM performed the experiments. HYC, JBS, AAA, RBW, JTC, MVDR, DB, and POB analyzed the data. HYC and POB wrote the paper. Academic Editor: Edison T. Liu, Genome Institute of Singapore ¤1 Current address: Department of Microbiology and Immunology, University of California at San Francisco, San Francisco, California, United States of America ¤2 Current address: Lewis--Sigler Institute for Integrative Genomics, Princeton University, Princeton, New Jersey, United States of America CSR : core serum response FBS : fetal bovine serum FDR : false discovery rate GO : Gene Ontology IHC : immunohistochemistry ISH : in situ hybridization KLH : keyhole limpet hemocyanin SAM : significance analysis of microarrays

PubMed Central

2024-06-05T03:55:46.000791

2004-1-13

PMC314301

The idea that cancer cells go through a fateful transition that turns them into fast-growing, invasive, metastasizing tumors first surfaced in the early 1970s. During this conversion, blood vessels form around the tumor, providing a dedicated supply of blood to fuel the tumor\'s aggressive behavior. By the mid-1980s histological analysis revealed a similarity between the tumor "microenvironment" and that of a healing wound, prompting Harvard pathologist Harold Dvorak to describe cancer as a wound that does not heal. When the body sustains a wound, it coordinates an emergency response defined by rapid cell proliferation, invasion and "remodeling" of connective tissues and extracellular matrix (the network of proteins and molecules around cells), cell migration, and blood vessel formation (angiogenesis). These processes, which are restorative in normal wound healing, may promote cancer by supporting tumor formation, invasion, and metastasis. With no systematic method to measure the "wound-like" features in cancer, however, scientists have no way to evaluate the risk that a wound-healing genetic program may pose in cancer progression. A molecular understanding of the wound-healing process and its connection to cancer would provide insight into the nature of these similarities and perhaps provide molecular indicators of tumor progression. In an effort to create a framework for evaluating this relationship, Howard Chang and his colleagues at Stanford University developed a model to predict cancer progression based on the gene expression profile of a cellular response to serum in cell culture. Part of the problem with evaluating the physiological status of a tumor based on its genetic profile is that current techniques indicate only the expression, not the effect, of genes. To develop a strategy for interpreting biological outcomes from a gene expression profile, Brown\'s team modeled a physiological process by exposing cultured fibroblasts to serum---the soluble fraction of coagulated blood---and tracking gene expression. Serum is encountered in the body where blood leaks out of blood vessels (in essence, all the sites of injury) and is thought to be a major initiator of the wound response. Fibroblasts exist in the connective tissue of epithelial organs (which include the digestive tract, lungs, and mammary glands) and contribute to organ development, wound healing, inflammation, and a condition called fibrosis. (Fibrosis involves the same type of extracellular matrix remodeling seen in wound healing and cancer.) And fibroblasts can promote tumor formation and metastasis when mixed with epithelial cancer cells. Though fibroblasts from different sites in the body differ in their properties and gene expression profiles, Chang et al. found that they share a common expression pattern in response to serum. From this expression profile, the researchers identified a core group of genes---a genetic signature---associated with a serum response. Because many of the genes in the signature were known to be involved in various wound-healing processes---such as matrix remodeling, cell motility, and angiogenesis---Chang et al. used this signature as a surrogate marker to measure how much tumors may be like wounds. When they compared the wound-like genetic signature with the expression profiles of various clinical tumor samples, they found the signature was always present in certain cancers---prostate and liver-cell carcinomas---and occurred variably in others---breast, lung, and gastric carcinomas. In each of these three latter types of tumors, patients with tumors carrying the serum-activated wound-like genetic signature had a significantly increased risk of metastasis and death compared to patients with tumors that lacked the signature. Therefore, Chang et al. conclude that a wound-like phenotype is a general risk factor for metastasis and the aggressive behavior in many of the most common cancers. These results reveal a robust and useful similarity between the molecular programs in normal wound healing and tumor progression and metastasis. Although Chang et al. point out that their results do not indicate whether this fibroblast "fingerprint" is merely a marker for cancer progression or plays a role in orchestrating this pathway, they conclude that the genetic program activated in response to serum also contributes to tumor invasion and metastasis. This serum-response expression profile, the authors propose, provides a valuable new tool for predicting tumor behavior and determining a patient\'s prognosis. ::: {#pbio-0020039-g001 .fig} ::: {.caption} ###### Genomics predicts tumor behavior :::  :::

PubMed Central

2024-06-05T03:55:46.004773

2004-1-13

PMC314462

Kofi Annan, the Secretary-General of the United Nations, recently called attention to the clear inequalities in science between developing and developed countries and to the challenges of building bridges across these gaps that should bring the United Nations and the world scientific community closer to each other ([@pbio-0020001-Annan1]). Mr. Annan stressed the importance of reducing the inequalities in science between developed and developing countries, asserting that "This unbalanced distribution of scientific activity generates serious problems not only for the scientific community in the developing countries, but for development itself." Indeed, Mr. Annan\'s sentiments have also been echoed recently by several scientists, who present overwhelming evidence for the disparity in scientific output between the developing and already developed countries ([@pbio-0020001-Gibbs1]; [@pbio-0020001-May1]; [@pbio-0020001-Goldemberg1]; [@pbio-0020001-Riddoch1]). For example, recent United Nations Educational, Scientific, and Cultural Organization (UNESCO) estimates ([@pbio-0020001-UNESCO1]) indicate that, in 1997, the developed countries accounted for some 84% of the global investment in scientific research and development, had approximately 72% of the world researchers, and produced approximately 88% of all scientific and technical publications registered by the Science Citation Index (SCI). North America and Europe clearly dominate the number of scientific publications produced annually, with 36.6% and 37.5%, respectively, worldwide ([@pbio-0020001-UNESCO1]). > *North America and Europe clearly dominate the number of scientific publications produced annually.* It is rather obvious that richer countries are able to invest more resources in science and therefore account for the largest number of publications. It is also likely that there is a statistical bias on the part of the SCI as a bibliometric database, since it represents North American and European publications far better than those of the rest of the world ([@pbio-0020001-Gibbs1]; [@pbio-0020001-May1]; [@pbio-0020001-Alonso1]; [@pbio-0020001-Vohora1]). But is the disparity in scientific contributions between the developed and developing worlds actually remaining unchanged or even increasing, as Mr. Annan has implied? A closer look at the trends over the last decade reveals important advances in developing countries. For example, Latin America and China, although representing, respectively, only 1.8% and 2% of scientific publications worldwide, have increased the number of their publications between 1990 and 1997 by 36% and 70%, respectively, which is a much higher percentage than the increments reached by Europe (10%) and industrial Asia (26%). The percentage of global scientific publications from North America actually decreased by 8% over the same period ([@pbio-0020001-UNESCO1]). Publishing Trends in the Americas {#s2} ================================= Using the SCI databases produced by the Institute for Scientific Information (ISI), as well as data compiled by the Red Iberoamericana de Indicadores de Ciencia y Tecnología (RICYT), we examined the differences in the number and proportion of scientific publications between the developed world and the developing world from 1990 until 2000, focusing on the Americas as a case study. Not surprisingly, there was a huge disparity in the number of publications from 1990 until 2000, with the United States contributing the lion\'s share (84.2%), followed by Canada (10.35%). Latin America as a whole contributed only 5.45% to the total number of scientific publications in these ten years ([@pbio-0020001-RICYT1]). The total number of publications, however, is not necessarily the best measure for assessing scientific productivity or technical advances ([@pbio-0020001-May1]). More relevant measurements for these factors include the proportional change in the number of publications and the total number of publications when corrected for investment in research and development ([@pbio-0020001-May1]). The proportional change in the number of publications, using 1990 as a comparison, revealed that scientific publishing in Latin America increased the most rapidly in the Americas, far outpacing the United States and Canada ([Figure 1](#pbio-0020001-g001){ref-type="fig"}). Further analyses, correcting the number of overall publications for the amount of money invested in research and development for each region, also show that, in contrast to both Canada and United States, the trend in Latin America has been an increase in relative output throughout the 1990s ([Figure 2](#pbio-0020001-g002){ref-type="fig"}). Moreover, when taking into account the amount of research money available to researchers, Latin America actually out-published the United States and Canada by the year 2000 ([Figure 2](#pbio-0020001-g002){ref-type="fig"}). Although the cost of research is undoubtedly cheaper in the developing world due to relatively low researcher salaries, overhead and other work standards, these factors do not explain the substantial increase in the number of publications per amount of money allocated to research and development in Latin America, particularly from 1995 until 2000 ([Figure 2](#pbio-0020001-g002){ref-type="fig"}). ::: {#pbio-0020001-g001 .fig} Figure 1 ::: {.caption} ###### Relative Increase in Scientific Publications in the Americas This figure shows the relative increase in publication in the Americas measured as the proportional change (%) in the number of SCI publications compared with the number of publications in 1990 ([@pbio-0020001-RICYT1]). :::  ::: ::: {#pbio-0020001-g002 .fig} Figure 2 ::: {.caption} ###### Number of SCI Publications per Million Dollars This figure shows the number of SCI publications per million dollars that are invested in research and development in the Americas ([@pbio-0020001-RICYT1]). :::  ::: Other relative indicators of scientific productivity, such as the number of publications picked up by the SCI in relation to the number of scientists in a particular country, also demonstrate that such developing regions as Latin America are making substantial contributions to science, despite the fact that the average proportion of gross domestic product (GDP) invested in science in Latin America throughout this 10-year period was only 21% of the amount invested in United States ([@pbio-0020001-RICYT1]). Indeed, this scientific productivity is remarkable when we compare it with the relatively low investment in science itself as compared with the GDP of Latin America as a whole. In fact, [@pbio-0020001-Albornoz1] concluded that, as a group, Latin America could afford to invest a much higher proportion of its resources in scientific research and development. Latin American investment in research and development represented only 0.59% of the regional GDP in 1998, a very weak effort compared with that of the United States (2.84%) and Canada (1.5%). Among Latin American countries, there is a high degree of variability in publication rate as well as in financial investment in science and technology. Some countries have performed particularly well. For example, Uruguay, Chile, Panama, and Cuba averaged, respectively, 6.8, 5.3, 5.2, and 3.4 publications per million dollars of research and development investment in the 10 years studied, which is notoriously high compared with United States (1.5) and even Canada (3.3) ([@pbio-0020001-RICYT1]). Other countries, such as Costa Rica, Cuba, Brazil, and Chile, have invested a much greater proportion of their GDP in research and development than the other countries of this region ([@pbio-0020001-Albornoz1]). > *Why has the number of publications per dollar invested in research and development been increasing in Latin America while decreasing in United States and Canada?* Explaining the Increase in Publishing Productivity in Latin America {#s3} =================================================================== One potential explanation for the increase in scientific productivity in Latin America is that scientific development during the 1990s was particularly strong for many countries of this region. Indeed, this would explain the rapid rise in the number of publications in Latin America compared with the relatively flat increases in the United States and Canada, which were publishing just as well at the beginning of the decade. A potentially more important question, however, is why the number of publications per dollar invested in research and development has been increasing in Latin America while decreasing in the United States and Canada. This pattern could be the result of a variety of factors, none of which are mutually exclusive. It is possible that publishing in international journals as a measure of scientific productivity is becoming more important in Latin America. Increased funding to the most productive scientists from the national science development programs might have been an important stimulus. International cooperation resulting in more scientific collaborations among scientists in Latin America, Europe, and the United States may also have increased the relative number of publications in Latin America. In contrast, the decreasing trends in the number of publications per investment dollar in Canada and United States could reflect a trend towards more costly research in larger scientific programs. Scientific Impact from Latin America {#s4} ==================================== What, exactly, is the relative impact of such developing regions as Latin America on the scientific community? We used SCI 2001 data to examine the proportion of publications in the area of ecology (including the fields of evolutionary biology, conservation biology, and global change biology) between 1990 and 2002 in both the two top general science journals (*Nature* and *Science*; with impact factors of 27.96 and 23.33, respectively) and in the 20 top ecological journals (with impact factors of 10.51--2.47) ([@pbio-0020001-ISI1]). We credited a region with a publication if any of the authors were affiliated with institutions from that region. Thus, more than one region would receive credit for a single publication if that publication had been written by multiple authors from institutions of different regions. For the top 20 ecological journals, the American subcontinents of South, Central, and North America accounted for 62% of the publications worldwide. Within the Americas, however, Latin America represented only 6%, while Canada and United States accounted, respectively, for 13% and 82% of the top 20 ecological publications. When we examined the data as contributions to the top 10 ecological journals (impact factors 10.51--3.31) versus the top 11--20 (impact factors 3.28--2.47), the Latin American countries contributed nearly twice as many publications to journals in the second category (8% in the top 11--20 compared with 4% in the top 10). These findings suggest that publications from such developing regions as Latin America are falling short of reaching the top journals. In contrast, the United States contributed somewhat more publications to the top 10 journals (84%) than the top 11--20 journals (79%). The difference in the proportion of publications contributed by the United States to the top 10 and top 20 journals was even more pronounced when we examined it in respect to worldwide publications. In this case, the United States contributed 60% of the publications to the top 10 journals and only 40% of the publications to the top 11--20 journals. Interestingly, the proportion of publications from Latin America, the United States, and Canada across all subject areas in *Science* and *Nature* were nearly identical to those of the top 20 ecological journals. In *Science* and *Nature*, Latin America had 7% of the publications within the Americas versus 6% in the top 20 ecological journals, whereas the United States and Canada had 81% versus 82% and 12% versus 13%, respectively. These similarities suggest that the Latin American researchers are not shying away from the two top-ranked general science journals. However, publishing in *Science* and *Nature* was not enough to gain prominence, as evidenced by the number of citations of these researchers. The latest list of the 247 most-cited researchers in ecology and environmental sciences emphasizes the overwhelming contributions of authors from North America (73%) and Europe (21%) ([@pbio-0020001-ISI2]). No researcher working in a Latin American institution was included in the remaining 6%. Overall, these data indicate that the scientific output in the field of ecology in Latin America is having a relatively low impact in the international scientific community and is underrepresented in the top international journals, despite its robust productivity as measured by the number of publications per researcher funding amount. Similar findings were also reported for Asia ([@pbio-0020001-Swinbanks1]) and thus could be a general phenomenon in the developing world. Although there are outstanding scientific researchers in the developing world who independently are making important contributions to the international scientific community, they are the exception. Why, in general, do Latin American scientists often fail to reach the top journals or become amongst the most cited researchers in their fields? One possibility is that the main research agendas between both regions are somewhat different and that the top journals, which are published in the developed world, respond more to the scientific mainstream of the developed regions. This is not to suggest any sort of conspiracy, but rather it implies that the perception of the most important science is linked to the region and that because the major funding agencies as well as most prominent journals share a similar economic region, they also share the same perception of what science is most interesting to them. Another consideration is that more local journals from developed regions are listed by the SCI than similar journals from developing regions ([@pbio-0020001-Gibbs1]). Consequently, there are more high-profile regional publication opportunities available to scientists from the developed region, whereas much of the research published locally in the developing world is overlooked. But it takes more than publishing good papers to become a highly cited scientist. It requires attending international meetings and introducing novel research findings in multiple scientific forums. Funding these activities, however, requires a greater proportion of research money being spent on meetings for researchers in the developing compared with the developed world. A Long Road Yet to Travel {#s5} ========================= The positive trends in scientific productivity in Latin America should not be misinterpreted as a reason to be unconcerned about the existing gap highlighted by Mr. Annan. There are many compelling reasons for the push to increase scientific input from the developing world ([@pbio-0020001-Goldemberg1]; [@pbio-0020001-Annan1]). One is that science, as a discipline, would benefit from the contributions of many disparate groups around the world, rather than being dominated by two geographic regions. Many scientific problems could be solved much more readily with the cooperation and scientific insight of scientists from developing regions. Climate change and biodiversity research, for example, urgently need the scientific input from those developing regions that are so important for these global processes. It is also critical for the developing world to promote, through research and publications, those areas of concern that are having a proportionally greater scientific and social impact upon them. There are now examples in which research on priority areas for the developing nations can actually become pioneering work in areas neglected by the research agenda of the industrialized world. This has been the case for research on renewable energy sources in Brazil ([@pbio-0020001-Goldemberg1]) and biomedical sciences in Cuba ([@pbio-0020001-CastroDiaz-Balart1]). These examples are important not only for those regions of the developing world, but are also in themselves scientific innovations that can greatly advance the knowledge of the rest of the world. > *Climate change and biodiversity research urgently need the scientific input from those developing regions that are so important for global processes.* Although the evidence presented here demonstrates that there is a long way to go before developing countries contribute a more equitable share to the international scientific community, there are also reasons to be optimistic. The relative increase in the number of publications, especially when corrected for the amount of money available in research and development, demonstrates that many developing countries are heading in the right direction. The extremely high scientific productivity of many developing nations, corrected for and despite the rather limited availability of funds, suggests that increased funding to the sciences will be an excellent investment by developing nations in terms of publications as a measure of scientific output, particularly if these publications can target the journals that have the greatest impact. Although there may still be a long road to travel, we feel optimistic that the bridges mentioned by Mr. Annan are slowly being built. We would like to thank A. Ercoli, P. A. Jansen, M. Scheffer, and two anonymous reviewers for comments. This work was partially supported by the European Union Specific Research and Technological Development Programme in the Field of Cooperation with Third Countries and International Organisations (INCO) project grant ICA4-CT-2001-10051, by the University of Wisconsin--Milwaukee, and the by Ecosyn project of Wageningen University. Milena Holmgren is a Research Scientist for the Forest Ecology and Forest Management Group at Wageningen University in The Netherlands. Stefan A. Schnitzer is an Assistant Professor in the Department of Biological Sciences at the University of Wisconsin--Milwaukee in the United States of America and also works with the Forest Ecology and Forest Management Group at Wageningen University in The Netherlands. GDP : gross domestic product ISI : Institute for Scientific Information RICYT : Red Iberoamericana de Indicadores de Ciencia y Tecnología SCI : Science Citation Index UNESCO : United Nations Educational

PubMed Central

2024-06-05T03:55:46.005576

2004-1-20

PMC314463

Introduction {#s1} ============ A fundamental question in developmental biology is how adjacent regions of the vertebrate body acquire differences in their appearance or morphology. Mechanisms that establish the general body plan make use of a relatively small number of signaling pathways shared among all animals (reviewed in [@pbio-0020003-Pires-daSilva1]), but the extent to which these pathways control finer differences between body regions is not clear. Among vertebrates, differences in the shape or number of skeletal elements, altered morphology of epidermal appendages, and variation in pigment distribution combine to produce the majority of what distinguishes one animal from another. Among these, pigment patterns are an excellent system to investigate how morphological differences arise, both for different regions of the body within a species and for different animals from closely related species. In natural environments, color variation is a nearly universal mechanism for recognition, camouflage, or both; consequently, a large number of pigment patterns have been characterized from an evolutionary and ecological perspective ([@pbio-0020003-Boughman1]; [@pbio-0020003-Jiggins1]). In the laboratory, color variation has been the subject of vertebrate genetics for more than a century ([@pbio-0020003-Searle1]; [@pbio-0020003-Silvers1]), and many pigmentary components have been identified whose actions are understood in a cellular or organ-based context (reviewed in [@pbio-0020003-Bennett1]). Several mechanisms may contribute to regional differences in vertebrate pigmentation. In the embryo, alterations in the determination or migration of melanoblasts from the neural crest affect the number or distribution of pigment cells in the skin (reviewed in [@pbio-0020003-Reedy1]). Within hair follicles, paracrine signals control the type of pigment made in specific regions of the body or at specific times during the hair cycle (reviewed in [@pbio-0020003-Furumura1]; [@pbio-0020003-Barsh1]). Finally, movement of pigment granules within melanocytes or from melanocytes to keratinocytes makes use of cellular machinery that is shared by a variety of cell types, but that can vary in different regions of the body (reviewed in [@pbio-0020003-Marks1]). However, for all of these mechanisms---white spotting, pigment-type switching, and melanosome biogenesis---more is known about the identity of the molecular components than their spatial and temporal control. One of the most obvious aspects of regional color variation in vertebrates is a dark dorsal surface juxtaposed to a light ventral surface, apparent in the color of skin, scales, feathers, or hair, in which the boundary between dorsal and ventral compartments is often sharp and lies in register with the limbs. In rodents and probably other mammals, this dorsoventral difference in hair color is brought about by differences in pigment type as determined by allelic variation of the *Agouti* gene ([@pbio-0020003-Bultman1]; [@pbio-0020003-Miller1]). Secreted by dermal papilla cells within each hair follicle ([@pbio-0020003-Millar1]), Agouti protein causes melanocytes in that follicle to switch from the production of brown/black eumelanin to red/yellow pheomelanin. Agouti protein has a short radius of action ([@pbio-0020003-Silvers2]) and can be switched on and off during a single hair cycle ([@pbio-0020003-Bultman1], [@pbio-0020003-Bultman2]; [@pbio-0020003-Miller1]; [@pbio-0020003-Vrieling1]); thus, its regulated expression is thought to be responsible for the cream-colored or yellow ventral surface of mice carrying the *black-and-tan* (*a^t^*) allele and for the yellow markings around the feet, ears, or head, i.e., tan points or head spots, of certain dog breeds. In laboratory mice, previous studies from our group and others identified two predominant *Agouti* mRNA isoforms that differ by virtue of their transcriptional initiation site and 5′ untranslated exons. A "hair cycle-specific" transcript is expressed in both dorsal and ventral skin for 2--3 days during early hair growth, while a "ventral-specific" transcript is expressed throughout the entire period of active hair growth, but only in ventral skin ([@pbio-0020003-Bultman2]; [@pbio-0020003-Vrieling1]). Animals carrying the *a^t^* allele express only the ventral-specific *Agouti* transcript ([@pbio-0020003-Bultman2]; [@pbio-0020003-Vrieling1]) and have black dorsal hairs and cream-colored to yellow ventral hairs, with a sharp boundary at the level of the limb--body wall articulations and in the middle of the whisker pad. Ventral-specific *Agouti* isoforms are also expressed in developing skin from embryonic day 10.5 (E10.5) and beyond and may play a role in pigment cell differentiation ([@pbio-0020003-Millar1]). Thus, regulatory elements for ventral-specific *Agouti* isoforms are responsive to dorsoventral positional cues established in the embryo and whose effects persist after birth. The boundary between dorsal and ventral color compartments in *a^t^*/*a^t^* mice bears superficial resemblance to dorsoventral boundaries apparent for many other mammals, but morphogenetic differences between dorsal and ventral skin seem likely to include more elements than the type of pigment made by hair follicle melanocytes. In particular, dermis of the flank has at least two distinct origins: dermatomal derivatives of somites and loose mesenchyme derived from the lateral plate mesoderm ([@pbio-0020003-Mauger1]; [@pbio-0020003-Christ1]; [@pbio-0020003-Olivera-Martinez1]; [@pbio-0020003-Nowicki1]); these lineages are established early in development and could, in principle, set up compartments whose identity contributes to dorsoventral differences in adult skin. To better understand the mechanisms that give rise to differences between dorsal and ventral skin and to the boundary between them, we have determined how several morphologic characteristics vary along the dorsoventral axis of the mouse and how these characteristics correspond to ventral-specific *Agouti* expression and the lineage boundary that distinguishes somite from lateral plate derivatives. Our results indicate that the apparent uniformity of the dorsoventral boundary represents the sum of independent mechanisms that affect melanocyte density and/or differentiation, pigment-type synthesis, and hair length; surprisingly, none of these coincide with the somite--lateral plate lineage boundary. We also make use of a classical mouse mutation, *droopy ear* ([@pbio-0020003-Curry1]), that produces a dorsal-to-ventral transformation of flank coat color by allowing expansion of the ventral-specific *Agouti* transcript. By positional cloning and gene targeting, we identify an allele of *droopy ear*, *de^H^*, as a loss of function for *Tbx15*, which encodes a T-box transcription factor expressed in a dynamic and spatially restricted manner in the developing skin and musculoskeletal system. Embryonic expression and transplantation studies suggest that *Tbx15* is required to establish certain characteristics of dorsal patterning in mesenchymal cells of the developing flank. These results identify a previously unappreciated aspect of dorsoventral patterning that is widely represented in furred mammals and provide insight into the mechanisms that underlie region-specific differences in body morphology. Results {#s2} ======= Morphological Components of Dorsoventral Skin Differences {#s2a} --------------------------------------------------------- Besides the obvious change in hair color that frequently distinguishes dorsal from ventral skin, casual observation suggests there are additional differences in hair length, distribution of hair type, and skin thickness. Furthermore, dorsoventral differences in pigmentation can represent differences in the number and/or differentiated state of pigment cells, as well as the type of pigment synthesized in response to expression of *Agouti*. In particular, ventral hair of *a^t^*/*a^t^* animals can vary from cream-colored to reddish-yellow depending on age, strain background, and position along the dorsoventral axis. To evaluate the relationship among these components, we compared their features among mice of different *Agouti* genotypes. Semiquantitative measurements of hair length plotted as a function of dorsoventral position reveal that the apparent sharp boundary between dorsal and ventral pigment compartments in *a^t^*/*a^t^* mice coincides with a more gradual change in both hair color and hair length ([Figure 1](#pbio-0020003-g001){ref-type="fig"}A--[1](#pbio-0020003-g001){ref-type="fig"}D). Within the region of transition from dorsum to ventrum ([Figure 1](#pbio-0020003-g001){ref-type="fig"}B), flank hairs from *a^t^*/*a^t^* mice become progressively shorter and exhibit increasing amounts of pheomelanin deposition progressing from the tip to the base of the hair. However, the region of transition for hair length is considerably broader than that for pigmentation and independent of *Agouti* genotype. Although hair-cycle timing varies along the rostrocaudal axis, measurements of absolute hair length for mice matched for age and rostrocaudal level are remarkably similar ([Figure 1](#pbio-0020003-g001){ref-type="fig"}D). Furthermore, measurements of relative hair length for animals of different age, size, and *Agouti* genotype also are very similar when normalized to body circumference ([Figure 1](#pbio-0020003-g001){ref-type="fig"}C). Taken together, these observations indicate that variation of hair length along the dorsoventral axis is stereotyped and maintained through multiple hair cycles, with a transition in hair length that is gradual and encompasses the pigment-type transition in *a^t^*/*a^t^* mice. ::: {#pbio-0020003-g001 .fig} Figure 1 ::: {.caption} ###### Dorsoventral Skin Characteristics \(A) Skin slices from animals of different age and genotype demonstrate similar patterns of hair-length variation along the dorsoventral axis (scale bar = 1 cm). \(B) Enlarged area from (A), demonstrating the transition in hair length and color in *a^t^*/*a^t^* mice (scale bar = 0.375 cm). \(C) Proportional hair length for (A) plotted as a function of relative position along the dorsoventral axis. \(D) Hair length plotted as a function of absolute position along the dorsoventral axis for 8-wk-old BA strain mice. \(E) Proportion of zigzag hairs (± SEM) differs slightly between dorsum and ventrum of inbred mice (*p* \< 0.0001, χ^2^ test, *n* = 1,958, 1,477, 1,579, 1,502). \(F) Differences in dorsal and ventral skin development at P4.5 (scale bar = 1 mm, upper; 200 μm, lower). \(G) Differences in hair melanin content and DOPA staining for dorsum (d), flank (f), and ventrum (v) in *a^e^*/*a^e^* and *a^t^*/*a^t^* mice. The upper panel also demonstrates a cream-colored appearance of the *a^t^*/*a^t^* ventrum. The middle panel shows representative awls (scale bar = 100 μm). The lower panel shows DOPA-stained dermis (scale bar = 200 μm). :::  ::: Dorsal and ventral skin develop at different rates. Transverse sections of skin at postnatal day 4.5 (P4.5) exhibit dorsal hair follicles that are noticeably more developed than ventral hair follicles, along with a gradual dorsoventral decrease in dermal thickness ([Figure 1](#pbio-0020003-g001){ref-type="fig"}F). However, differences in skin thickness disappear by 3--4 wk of age ([@pbio-0020003-Forsthoefel1]), and, overall, the proportion of different hair types is also similar in dorsa and ventra of adult mice. In age-matched inbred mice, we observed a small decrease in the ratio of undercoat hairs (zigzags) to overcoat hairs (auchenes, awls, and guard hairs) in dorsum compared to ventrum ([Figure 1](#pbio-0020003-g001){ref-type="fig"}E), but there was no consistent difference in hair-type distribution for outbred mice (data not shown). Differences between dorsal and ventral pigmentation of *a^t^*/*a^t^* mice are usually attributed to pigment-type differences caused by ventral-specific expression of *Agouti*, but animals homozygous for a null allele of *Agouti*, *extreme nonagouti* (*a^e^*), have ventral hairs that contain less melanin than dorsal hairs, giving a slightly paler appearance to the ventral coat ([Figure 1](#pbio-0020003-g001){ref-type="fig"}G). Using DOPA staining as an indicator of tyrosinase activity, we observed a gradual dorsoventral transition in isolated dermis preparations from P4.5 *a^e^*/*a^e^* mice ([Figure 1](#pbio-0020003-g001){ref-type="fig"}G). By contrast, skin from *a^t^*/*a^t^* mice reveal an abrupt dorsoventral transition of DOPA staining, which probably reflects the additive effects of reduced melanin content (as in *a^e^*/*a^e^* mice) and downregulation of tyrosinase activity induced by *Agouti*. Melanin content of individual hairs is likely to be influenced both by the number of pigment cells and their follicular environment. Regardless, dorsoventral differences in hair pigment content of *a^e^*/*a^e^* mice persist throughout multiple hair cycles into adulthood, similar to hair length (but unlike skin thickness). Thus, at least three characteristics distinguish dorsal from ventral skin: differences in pigment-type synthesis (depending on *Agouti* genotype), differences in hair length, and differences in melanin content. Ventralization of Skin Morphology by the *droopy ear* Mutation {#s2b} -------------------------------------------------------------- Named after its effects on craniofacial morphology, *droopy ear* is a recessive mutation on mouse Chromosome 3; the original allele described more than 40 years ago by [@pbio-0020003-Curry1] is extinct, but a spontaneous remutation that occurred in Harwell, *de^H^*, is available through The Jackson Laboratory (Bar Harbor, Maine, United States). External craniofacial malformations are the most obvious characteristic of *de^H^*/*de^H^* animals, including widely spaced eyes, small palpebral fissures, a broad nasal area, and a shortened skull held in an elevated position, which presumably causes or contributes to the abnormal position of the ears. We became interested in *droopy ear* because the original allele was described to affect pigment pattern in a way that suggests a possible dorsal to ventral transformation: "On a genetic background (*a^t^* and *A^W^*) which causes the belly hair to be lighter than the back hair, the belly hair comes up farther round the sides of the body and face" ([@pbio-0020003-Curry1]). An abnormal dorsoventral pigment pattern is readily apparent in *a^t^*/*a^t^*; *de^H^*/*de^H^* mice, but comparison to nonmutant animals is more accurately described in terms of ventral, lateral, and dorsal regions (Figures [1](#pbio-0020003-g001){ref-type="fig"}G and [2](#pbio-0020003-g002){ref-type="fig"}A). The ventral region has short hairs with a gray base and cream-colored tip whose boundary coincides with the limb--body wall junction; both the appearance of this region and position of the boundary are approximately similar in *a^t^*/*a^t^* compared to *a^t^*/*a^t^*; *de^H^*/*de^H^* mice. The lateral region contains yellow hairs of progressively increasing length; in *a^t^*/*a^t^* mice, the lateral region appears as a thin yellow stripe along the flank, but in *a^t^*/*a^t^*; *de^H^*/*de^H^* mice, the lateral region is considerably expanded with a diffuse boundary along the dorsal flank, and a dorsal eumelanic region whose size is correspondingly reduced ([Figure 2](#pbio-0020003-g002){ref-type="fig"}A and [2](#pbio-0020003-g002){ref-type="fig"}B). Total body size is smaller in mutant compared to nonmutant animals, but the proportion of body circumference occupied by the lateral region in mutant animals is increased about 2-fold, from 11.9% to 22.2% ([Figure 2](#pbio-0020003-g002){ref-type="fig"}C). The proportion of the ventral cream-colored region is also expanded a small amount, 47.9% in mutant compared to 37.8% in nonmutant animals, but expansion of the lateral region, which occurs at all levels of the body, including the limbs and the cranium (but not the whisker pad), is the major feature responsible for the ventralized appearance caused by *de^H^*. ::: {#pbio-0020003-g002 .fig} Figure 2 ::: {.caption} ###### The *de^H^* Pigmentation Phenotype \(A) 10-wk-old *de^H^/de^H^* and nonmutant animals on a *a^t^* background. A thin stripe of yellow hair normally separates the dorsal black hairs from the ventral cream hairs. In *de^H^*, the yellow stripe is extended dorsally, and the boundary between the yellow and the black hairs is fuzzier. \(B) Skin slices taken from 1.5-mo-old *de^H^/de^H^* and nonmutant littermates (scale bar = 0.5 cm). \(C) Proportion of total skin area as determined by observation of pelts taken from the interlimb region. The proportion occupied by the yellow lateral compartment (± SEM) differs between mutant and nonmutant littermate flanks (*p* \< 0.0005, paired *t*-test, *n* = 6 pairs). There is also (data not shown) a small increase in the proportion of total skin area occupied by the ventral cream-colored compartment, 47.9 % in mutant compared to 37.8% in nonmutant (*p* \< 0.005, paired *t*-test, *n* = 6 pairs). \(D) On an *a^e^/a^e^* background, the extent of dorsal skin pigmentation is reduced in *de^H^/de^H^* neonates (P3.5). \(E) Hair length in a representative pair of 1.5-mo-old *de^H^/de^H^* and nonmutant littermates, averaged over three skin slices at different rostrocaudal levels, and plotted as a function of the absolute distance from middorsum or the percentage of total slice length. :::  ::: To investigate whether *de^H^* affects other dorsoventral skin characteristics besides pigment-type switching, we examined its effects on hair length and pigmentation in an *a^e^*/*a^e^* background. Overall, *de^H^* causes a small but consistent reduction in hair length in both dorsum and ventrum; when mutant and nonmutant animals are normalized for body circumference, reduced hair length is most apparent in the lateral region ([Figure 2](#pbio-0020003-g002){ref-type="fig"}E). Adult *a^e^*/*a^e^*; *de^H^*/*de^H^* animals exhibit body-size reduction and skeletal abnormalities, but display no coat-color phenotype (data not shown). However, *a^e^*/*a^e^* and *a^e^*/*a^e^*; *de^H^*/*de^H^* neonates are clearly distinguishable in the first few days after birth, when a dorsoventral gradient of melanogenic activity is apparent under the skin ([Figure 2](#pbio-0020003-g002){ref-type="fig"}D). At this stage, melanoblast migration from the neural crest is mostly complete, but there is a dorsoventral gradient in melanocyte differentiation and pigment synthesis. The skin of *a^e^*/*a^e^* neonates appears uniformly dark over the entire dorsum, but in *a^e^*/*a^e^*; *de^H^*/*de^H^* neonates, the area of dark skin is more restricted, particularly above the limbs, and resembles the pattern of dorsal eumelanin in *a^t^*/*a^t^*; *de^H^*/*de^H^* adult animals. Taken together, these observations suggest that *de^H^* interferes with the establishment of dorsoventral patterning during skin development by causing dorsal expansion of a lateral region that is normally 3--5 mm in width. This same region may serve as a boundary between dorsal and ventral skin by inhibiting melanocyte differentiation, by promoting pheomelanin synthesis, and by supporting a progressive increase in hair growth from ventrum to dorsum. As described below, the gene defective in *de^H^*, *Tbx15*, is normally expressed in the dorsal region and therefore is likely to play a role in establishing the size and dorsal extent of the lateral region. Positional Cloning of *de^H^* {#s2c} ----------------------------- As a visible marker, early linkage studies with the original *droopy ear* allele or the *de^H^* allele identified a map position in the middle of Chromosome 3, distal to *matted* and proximal to *Varitint-waddler* ([@pbio-0020003-Carter1]; [@pbio-0020003-Curry1]; [@pbio-0020003-Lane1]; [@pbio-0020003-Holmes1]). We used an F~2~ intercross with CAST/Ei mice to localize *de^H^* to a 0.1 cM interval between D3Mit213 and 16.MMHAP32FLF1, which was refined by development of a bacterial artificial chromosome (BAC) contig and additional markers to a 1.4 Mb region that contained eight genes, including *Tbx15* ([Figure 3](#pbio-0020003-g003){ref-type="fig"}A). We considered *Tbx15* as an excellent candidate for the skeletal abnormalities caused by *de^H^*, based on studies by [@pbio-0020003-Agulnik1], who described its embryonic expression in the craniofacial region and developing limbs. ::: {#pbio-0020003-g003 .fig} Figure 3 ::: {.caption} ###### Molecular Genetics of *de^H^* and *Tbx15* \(A) Genetic and physical map, as described in the text. Markers M1 to M3 are SSCP markers generated from a BAC contig of the region; marker M4 is STS 16.MMHAP32FLF1 and was also used as an SSCP marker. M2 and M3, which flank the *Tbx15* and *M6pr-ps* on the UCSC genome browser map and lie 634 kb apart, were nonrecombinant with *de^H^* in 2340 meioses. \(B) The *de^H^* mutation is a deletion that starts in *Tbx15* intron 1 and ends in the *M6pr-ps*. \(C) Sequence of deletion breakpoints. \(D) Diagram of *Tbx15^LacZ^* allele constructed by gene targeting. As described in the text, this allele is predicted to give rise to a protein truncated after approximately 154 codons and is lacking critical residues of the T box. Heterozygotes for the targeted allele exhibit normal size, morphology, and hair-color patterns, but homozygotes and *Tbx15^LacZ^*/*de^H^* compound heterozygotes are identical to *de^H^* homozygotes. :::  ::: Using sequence information from [@pbio-0020003-Agulnik1] and the partially completed mouse genome sequence, we found that portions of several *Tbx15* exons could not be amplified from *de^H^*/*de^H^* genomic DNA. The same gene was initially referred to as *Tbx8* ([@pbio-0020003-Wattler1]) and then later renamed *Tbx14*, but is currently referred to in several vertebrate genomes as *Tbx15* ([@pbio-0020003-Agulnik1]; [@pbio-0020003-Begemann1]). By comparing the sequence of a 1.3 kb junction fragment amplified from *de^H^*/*de^H^* genomic DNA to publicly available mouse genome sequence, we identified a 216 kb deletion that extends from *Tbx15* intron 1 to 148 kb downstream of the polyadenylation sequence in a region annotated as a mannose-6-phosphate receptor pseudogene, *M6pr-ps* ([Figure 3](#pbio-0020003-g003){ref-type="fig"}B and [3](#pbio-0020003-g003){ref-type="fig"}C). ([@pbio-0020003-Ludwig1]). By Northern blot analysis, we identified a fusion transcript produced from the *de^H^* chromosome (data not shown). However, the deletion removes 534 of the 602 amino acids encoded by *Tbx15* (including the T-box DNA-binding domain), *de^H^*/+ animals are grossly normal, and the phenotype of *de^H^*/*de^H^* animals is identical to that described for the original allele. In addition, other than *M6pr-ps*, no other genes or transcripts have been annotated to the 216 kb deletion. While the positional cloning work was underway, one of us (A. Russ) generated an independent mutation of *Tbx15* by gene targeting in embryonic stem cells. The targeted allele, *Tbx15^LacZ^*, carries an *IRES-LacZ-neo* cDNA cassette that disrupts the open reading frame at codon 154 early in the T-box domain ([Figure 3](#pbio-0020003-g003){ref-type="fig"}D). Animals heterozygous for the targeted allele are completely normal with regard to size, skeletal morphology, and hair-color distribution, but *Tbx15^LacZ^*/*Tbx15^LacZ^* homozygotes were noted to exhibit reduced body size and an abnormal craniofacial appearance identical to that caused by *de^H^*. We generated *Tbx15^LacZ^*/*de^H^* compound heterozygotes; on an *A^w^*/*a^t^* background, these animals exhibited the same abnormal restriction of dorsal pigmentation at P3.5 and expanded yellow flank area as described above for *de^H^*/*de^H^* animals (see [Figure 2](#pbio-0020003-g002){ref-type="fig"}). These observations demonstrate that the pigmentary and craniofacial characteristics of *de^H^* are caused by loss of function for *Tbx15*. Expression of *Tbx15* and *Agouti* {#s2d} ---------------------------------- Previous studies by [@pbio-0020003-Agulnik1] using whole-mount in situ hybridization described expression of *Tbx15* as first detectable at E9.5 in the limb buds, progressing to the branchial arches, flanks, and craniofacial regions through E12.5. To investigate this pattern in more detail, we hybridized a *Tbx15* mRNA probe to a series of transverse sections at E12.5 and observed expression in multiple mesenchymal tissues of the head, trunk, and developing limbs ([Figure 4](#pbio-0020003-g004){ref-type="fig"}A), much of which is consistent with the skull, cervical vertebrae, and limb malformations reported for mice carrying the original *droopy ear* allele. ::: {#pbio-0020003-g004 .fig} Figure 4 ::: {.caption} ###### Developmental Expression of *Tbx15* \(A) At E12.5, transverse sections at different levels show expression in head mesenchyme (a and b); myotome, occipital, and periocular mesenchyme (b); palatal shelf, cervical sclerotome, and nasal cartilage (c); maxillary and mandibular processes (d); limbs (e); and myotome and lateral mesenchyme (e and f) (scale bars = 500 μm). \(B) Transverse sections through the flank at different times show expression in lateral mesenchyme (E11.5), expanding dorsally at E12.5, and both ventrally and dorsally at E13.5, detectable in loose mesenchyme underlying the dermis and the abdominal and subcutaneous muscles (scale bar = 500 μm). At P3.5, *Tbx15* is expressed in the entire dermis and is most strongly expressed in dermal sheaths (scale bar = 200 μm). :::  ::: We were particularly interested in determining the exact nature of the embryonic flank expression relative to the ventralized phenotype of adult *de^H^*/*de^H^* mice. Transverse abdominal sections from different times during development reveal a dorsolateral band of expression in the superficial mesenchyme at E11.5 that broadens both dorsally and ventrally over the next several days ([Figure 4](#pbio-0020003-g004){ref-type="fig"}B). By E13.5, the developing dermis has become separated from the loose mesenchyme by a subcutaneous muscle layer; *Tbx15* is expressed in all of these layers as well as the underlying abdominal muscles. In P3.5 skin, *Tbx15* is expressed in both dorsal and ventral skin, most strongly in the condensed upper dermis and developing dermal sheaths of hair follicles; faint expression can also be detected in rare dermal papillae cells ([Figure 4](#pbio-0020003-g004){ref-type="fig"}B). Although the effects of *Agouti* on pigment-type switching occur during postnatal hair growth, the ventral-specific isoform of *Agouti* is expressed in developing skin beginning at E11.5. We compared adjacent sections hybridized with probes for *Tbx15* and *Agouti* and observed complementary patterns at E12.5, with expression of *Agouti* in ventral skin and expression of *Tbx15* in dorsal skin ([Figure 5](#pbio-0020003-g005){ref-type="fig"}A and [5](#pbio-0020003-g005){ref-type="fig"}B). The junction between expression domains is indistinct, and by E14.5, *Tbx15* expression extends ventrally and overlaps extensively with *Agouti* expression ([Figure 5](#pbio-0020003-g005){ref-type="fig"}C and [5](#pbio-0020003-g005){ref-type="fig"}D). ::: {#pbio-0020003-g005 .fig} Figure 5 ::: {.caption} ###### Embryonic Expression of *Tbx15* Compared to *Agouti* in *a^t^*/*a^t^* Mice (A and C) *Tbx15*. (B and D) *Agouti*. At E12.5, expression of *Tbx15* in dorsal skin is approximately complementary to that of *Agouti* in ventral skin. At E14.5, the levels of expression for both genes are lower, but *Tbx15* expression has expanded ventrally and overlaps extensively with that of *Agouti*. In all four panels, arrows mark the approximate ventral limit of *Tbx15* and the approximate dorsal limit of *Agouti* (scale bars = 500 μm). :::  ::: We also examined the effect of *de^H^* on expression of *Agouti* and found no difference between mutant and nonmutant at E12.5 or E13.5 (data not shown). However, at E14.5, the normal ventral-to-dorsal gradient of *Agouti* expression appeared to extend more dorsally in *de^H^*/*de^H^* embryos ([Figure 6](#pbio-0020003-g006){ref-type="fig"}A). In P4.5 skin, expression of *Agouti* is also extended dorsally in *de^H^*/*de^H^* animals and is most apparent in the midflank region within the upper dermis and dermal papillae cells ([Figure 6](#pbio-0020003-g006){ref-type="fig"}B). Thus, while the pigmentation phenotype of *de^H^*/*de^H^* mice can be explained, not surprisingly, by dorsal extension of *Agouti* expression after birth, patterned expression of *Tbx15* and *Agouti* are apparent some 10 days earlier, between E12.5 and E13.5, and the effects of *Tbx15* deficiency on expression of *Agouti* can be detected by E14.5. ::: {#pbio-0020003-g006 .fig} Figure 6 ::: {.caption} ###### Effect of *de^H^* on *Agouti* Expression Comparable sections from *a^t^/a^t^*; *de^H^*/*de^H^* and *a^t^/a^t^*; +/+ littermates. \(A) At E14.5, *de^H^*/*de^H^* embryos have a smaller body cavity and loose skin within which *Agouti* expression appears to be shifted dorsally, as marked by arrows (scale bars = 500 μm). \(B) At P4.5, *Agouti* expression in both dorsal and ventral skin is similar in *de^H^*/*de^H^* compared to nonmutant, but in the midflank region, there is increased *Agouti* expression in *de^H^*/*de^H^*, especially in the upper dermis (scale bars = 200 μm). Sections shown are representative of two mutant and two nonmutant samples examined at each time. :::  ::: Relationship of Embryonic *Tbx15* Expression to Dorsal and Ventral Pigmentation Domains {#s2e} --------------------------------------------------------------------------------------- The observations described above are consistent with a model in which transient expression of *Tbx15* in the embryonic dorsal flank is required to establish positional identity of the future dermis, at least with respect to pigment-type synthesis caused by the ventral-specific *Agouti* isoform. To further investigate this hypothesis, we carried out transplantation experiments in which pieces of embryonic skin were isolated from different dorsoventral positions. We evaluated the embryonic skin fragments for their potential to give rise to different hair colors and for their expression of *Tbx15* and *Agouti*. Previous studies by Silvers and colleagues ([@pbio-0020003-Poole1]) showed that dorsal and ventral skin isolated from *a^t^*/*a^t^* embryos gives rise to black and yellow hair, respectively, when transplanted into testis and allowed to develop for several weeks. Furthermore, dermal--epidermal recombination experiments carried out at E14.5 demonstrated that positional identity is carried by the embryonic dermis. In a variation on this experiment, we divided embryonic skin from *a^t^*/*a* embryos into dorsal, flank, and ventral pieces and analyzed the different pieces for their ability to give rise to black or yellow hair after testis transplantation, and, in parallel, for gene expression using in situ hybridization. For the purposes of a reproducible morphologic boundary, we divided flank from ventral skin based on a change in skin thickness and divided dorsal from flank skin at the level of an ectodermal notch that lies at the same level as the ventral extent of the myotome ([Figure 7](#pbio-0020003-g007){ref-type="fig"}) ([@pbio-0020003-Huang1]; [@pbio-0020003-Olivera-Martinez1]; [@pbio-0020003-Sudo1]; [@pbio-0020003-Burke1]; [@pbio-0020003-Nowicki1]). ::: {#pbio-0020003-g007 .fig} Figure 7 ::: {.caption} ###### Embryonic Establishment of Dorsoventral Skin Patterning Pieces of skin from dorsal, flank, and ventral regions of *a^t^/a* E12.5 embryos were transplanted into the testes of congenic animals as described in the text. Hair color of the grafts was examined 3 wk later. Grafts of ventral embryonic skin (*n* = 3) produced yellow hairs, dorsal embryonic skin (*n* = 4) produced black hairs, and flank embryonic skin produced mostly (13 out of 15) black and yellow hairs in distinct regions as shown. In parallel, in situ hybridization studies revealed that the embryonic flank contains the boundary of expression between *Agouti* and *Tbx15* (scale bars = 1 mm for hairs and 200 μm for in situ hybridization results). :::  ::: We found that E12.5 is the earliest time at which embryonic ventral skin is able to produce hair when transplanted to the testis. Of the grafts that produced hair, ventral skin gave rise to yellow hair (*n* = 3), and dorsal skin gave rise to black hair (*n* = 4). Transplantation of flank skin gave rise to a patch of yellow hair juxtaposed against a patch of black hair in 85% of the successful grafts (*n* = 13); the remaining two flank grafts produced solely black or yellow hair. In no case did we observe intermingling of black and yellow hairs. As predicted from the experiments using tissue sections (see Figures [5](#pbio-0020003-g005){ref-type="fig"} and [6](#pbio-0020003-g006){ref-type="fig"}), dorsal pieces expressed *Tbx15* but not *Agouti*, while flank pieces expressed both genes (see [Figure 7](#pbio-0020003-g007){ref-type="fig"}). Thus, dorsoventral identity for adult pigmentation is established by the time when patterned expression becomes apparent for *Tbx15* and *Agouti* (E11.5--E12.5); furthermore, positional identity is maintained throughout later stages of skin development, even though expression of *Tbx15* broadens to include ventral as well as dorsal skin. Relationship of the Dorsoventral Pigment Boundary to Lineage Compartments and the Lateral Somitic Frontier {#s2f} ---------------------------------------------------------------------------------------------------------- The ectodermal notch that we used to mark the boundary between embryonic dorsum and embryonic flank is a characteristic feature in vertebrate embryos. In cell lineage studies carried out in the chick system, the notch serves as a landmark for the boundary between dermis derived from somitic mesoderm and dermis derived from lateral plate mesoderm and has been termed the "lateral somitic frontier" ([@pbio-0020003-Olivera-Martinez1]; [@pbio-0020003-Sudo1]; [@pbio-0020003-Burke1]; [@pbio-0020003-Nowicki1]). Although fate-mapping studies have not been carried out in mammalian embryos, somite- and lateral plate-derived mesoderm could give rise to precursors for dermis dorsal and ventral to the limb--body wall junction, respectively. However, this notion conflicts with our observation that the future pigmentation boundary lies ventral to the ectodermal notch (see [Figure 7](#pbio-0020003-g007){ref-type="fig"}). To examine directly the relationship between the pigmentation boundary and dermis derived from lateral plate mesoderm, we made use of a *Cre* transgene driven by the *Hoxb6* promoter that was developed by Kuehn and colleagues ([@pbio-0020003-Lowe1]). As described by [@pbio-0020003-Lowe1], midgestation embryos carrying both the *Hoxb6-Cre* transgene and the *R26R lacZ* reporter gene ([@pbio-0020003-Soriano1]) exhibit X-Gal staining in lateral plate mesoderm but not somite-derived mesoderm of the trunk. In whole-mount skin preparations from P1.5 or P4.5 neonatal animals, we observed a ventral band of dark X-Gal staining corresponding to lateral plate-derived dermis, which represents 63% of the total circumference ([Figure 8](#pbio-0020003-g008){ref-type="fig"}A). However, in parallel preparations from *a^t^*/*a^t^* mice, the ventral pheomelanin domain represents 47% of the total skin circumference; therefore, the proportions of total skin circumference occupied by dorsal eumelanin and somite-derived dermis are 53% and 37%, respectively ([Figure 8](#pbio-0020003-g008){ref-type="fig"}B). These results indicate that the pigmentation boundary is clearly distinct from, and more ventral to, the boundary between lateral plate- and somite-derived dermis. ::: {#pbio-0020003-g008 .fig} Figure 8 ::: {.caption} ###### Comparison of the Dorsoventral *a^t^/a^t^* Pigmentation Boundary to the Lateral Somitic Frontier \(A) Dorsoventral slices of skin from at the midtrunk region prepared such that the dorsal midline lies in the center of the slice. Sections were taken at P1.5 (a) or P4.5 (b--e) from *a^t^/a^t^* or *R26R*/+; *Tg.Hoxb6-Cre*/+ mice (the latter were stained with X-Gal), as described in Materials and Methods. For purposes of comparison, images were proportionally scaled. The boundary of X-Gal staining marks dermis derived from lateral plate versus dermis derived from mesoderm (the lateral somitic frontier) and lies more dorsal than the *a^t^/a^t^* pigmentation boundary. \(B) Quantitation of mean (± SEM) dorsal pigmentation area (*n* = 5) and somite-derived dermis area (*n* = 3) shows a significant difference (*p* \< 0.005, *t*-test). \(C) RNA in situ hybridization showing that *Tbx15* expression at E11.5 is complementary to *En1* expression on the flank (scale bars = 200 μm). The arrow indicates the boundary between the expression domains of the two genes. :::  ::: Because the pigmentation boundary lies in register with the limb--body wall junction (see [Figure 2](#pbio-0020003-g002){ref-type="fig"}), we wondered whether mechanisms used for dorsoventral limb patterning might be related to those used to establish the pigmentation boundary. In the developing limb, *Engrailed1* (*En1*), *Wnt7a*, and *Lmx1b* are part of a network whose restricted domains of expression help to establish dorsoventral identity (reviewed in [@pbio-0020003-Niswander1]). *En1* is transiently expressed in the developing flank; at E11.5, transverse abdominal sections reveal domains in the neural tube, somite-derived mesenchyme, and the ventral body wall ([Figure 8](#pbio-0020003-g008){ref-type="fig"}C). An adjacent section hybridized with *Tbx15* reveals a complementary pattern in the flank, which provides additional evidence for developmental mechanisms that establish a pigmentation boundary entirely within lateral plate mesoderm and independent of lineage restrictions imposed by the lateral somitic frontier. Discussion {#s3} ========== Several mutations and genes have been identified that affect the pattern of hair follicle development, but *Tbx15* is the only gene of which we are aware that affects the pattern of hair pigmentation in different body regions. Ventral areas that normally produce yellow hair in the trunk, limbs, and craniofacial regions are expanded in *de^H^*/*de^H^* mice and, in the trunk at least, represent inappropriate dorsal expression of an *Agouti* mRNA isoform that is normally restricted to ventral skin. The *de^H^* allele is caused by a large deletion that removes most of the *Tbx15* coding sequence, but the pleiotropic phenotype is caused by a simple loss of function for *Tbx15* rather than a dominant-negative or contiguous gene effect. In particular, there is no heterozygous phenotype, no other genes lie within or close to the deletion breakpoints, and the expression pattern of *Tbx15* is consistent with the spectrum of phenotypic abnormalities in both the original *de* allele and the *de^H^* allele. Finally, a *Tbx15* targeted allele has the same phenotype as *de^H^*. Our results suggest that patterned expression of *Tbx15* provides an instructional cue required to establish the future identity of dorsal dermis with regard to pigmentary and hair length patterning. The ventral edge of *Tbx15* expression in the developing flank does not correspond to a known lineage compartment, but, like limb development, occurs within lateral plate mesoderm. These findings represent a novel role for T-box gene action in embryonic development and provide evidence for a previously unappreciated complexity to acquisition of dorsoventral positional identity in mammalian skin. Distinct Morphologic Regions Represent the Sum of Different Gradients {#s3a} --------------------------------------------------------------------- The visual boundary between dorsal and ventral skin in *a^t^*/*a^t^* mice is reminiscent of other systems in which adjacent compartments enforce a binary choice between alternative patterns of gene expression and cell fate (reviewed in [@pbio-0020003-Dahmann1]). However, *Agouti* mRNA in both embryonic and postnatal skin is distributed along a gradient whose dorsal boundary is indistinct and overlaps with two additional gradients recognized by their effects on hair length and histochemical staining for melanocytes. The three gradients are close but not congruent, and it is their proximity that gives rise to the superficial distinction between dorsal and ventral skin of *a^t^*/*a^t^* mice. Indeed, slight differences between the regions of transition for pigment-type switching and pigment content give rise to a subtle yellow stripe along the flank (see Figures [1](#pbio-0020003-g001){ref-type="fig"}, [2](#pbio-0020003-g002){ref-type="fig"}, and [9](#pbio-0020003-g009){ref-type="fig"}A). Levels of *Agouti* mRNA remain high throughout the entire ventrum, but hair pigment content is reduced, giving rise to a cream-colored region in the ventrum that, depending on age and genetic backgrounds, may appear more or less distinct from the yellow flank stripe. ::: {#pbio-0020003-g009 .fig} Figure 9 ::: {.caption} ###### Model for Acquisition of Dorsoventral Patterning in the Trunk and the Role of *Tbx15* \(A) A tricolor pigmentation pattern is generated by the combination of distinct mechanisms that affect distribution of *Agouti* mRNA and histochemical staining for melanocytes; effects of the latter mechanism by itself are evident in *a^e^*/*a^e^* mice (see [Figure 1](#pbio-0020003-g001){ref-type="fig"}). In *a^t^/a^t^* mice, reduced hair melanocyte activity and high levels of *Agouti* mRNA in the ventrum lead to a cream color; as melanocyte activity gradually increases towards the dorsum, a lateral stripe is apparent on the flank. The distributions of *Agouti* mRNA and histochemical staining for melanocytes are both affected by *Tbx15* and are externally evident by a widening of the lateral stripe and an increased proportion of total skin occupied by the cream-colored area. \(B) The lateral yellow stripe in *a^t^/a^t^* mice lies at the same level as the limb dorsoventral boundary. As described in the text, we propose that distinct dorsoventral compartments in ectoderm of the trunk provide an instructional cue to the mesoderm, leading to expression of *Tbx15* in dorsal trunk mesenchyme and acquisition of dorsal dermis character. In the absence of *Tbx15*, dorsal mesenchyme assumes ventral characteristics instead. :::  ::: Loss of *Tbx15* affects dorsoventral transitions of hair length, pigment content, and expression of the ventral-specific *Agouti* isoform; however, the former two effects are subtle and contribute little, if at all, to the abnormal pigmentation of adult *de^H^*/*de^H^* mice. Thus, despite the abnormal pattern of dark skin in neonatal *de^H^*/*de^H^* mice (e.g., [Figure 2](#pbio-0020003-g002){ref-type="fig"}D), the most obvious feature in adults is dorsal displacement of the "boundary" between black and yellow hair ([Figure 9](#pbio-0020003-g009){ref-type="fig"}A). Genetics of *Tbx15* {#s3b} ------------------- Named for the presence of a DNA-binding domain first identified in the mouse *Brachyury* gene (haploinsufficiency causes a short tail), T box--containing genes have been identified as developmental regulators in a wide spectrum of tissues and multicellular organisms (reviewed in [@pbio-0020003-Papaioannou1]). The *Tbx15* subfamily, which also includes *Tbx18* and *Tbx22*, is likely to have arisen during early chordate evolution since there is a single gene in amphioxus but no obvious homolog in the fly genome ([@pbio-0020003-Ruvinsky1]). Consistent with this relationship, the three genes are expressed in partially overlapping patterns that include anterior somites (*Tbx18* and *Tbx22*), limb mesenchyme (*Tbx15* and *Tbx18*), and craniofacial mesenchyme (all three genes, *Tbx15* more broadly than *Tbx18* or *Tbx22*) ([@pbio-0020003-Agulnik1]; [@pbio-0020003-Kraus1]; [@pbio-0020003-Braybrook2]; [@pbio-0020003-Bush1]; [@pbio-0020003-Herr1]). These observations suggest that an ancestral gene for *Tbx15*, *Tbx18*, and *Tbx22* may have been important for craniofacial development in cephalochordates, with acquisition of additional expression patterns and developmental functions in the limb and the trunk during early vertebrate evolution. Expression of *Tbx18* and *Tbx22* has not been reported in embryonic flank mesenchyme, which suggests that *Tbx15* is the only family member involved in establishing the dorsoventral identity of the trunk. However, it would not be surprising to find some degree of functional redundancy in animals mutated for two or three of the subfamily members in other body regions, particularly the limbs and the head. For example, mutations in *Tbx22* cause the human syndrome X-linked cleft palate and ankyloglossia ([@pbio-0020003-Braybrook1]). Despite high levels of *Tbx22* expression in periocular embryonic mesenchyme ([@pbio-0020003-Braybrook2]; [@pbio-0020003-Bush1]; [@pbio-0020003-Herr1]), the condition does not affect the eye, perhaps because residual activity is provided by *Tbx15* in the same region. In an initial description of the expression and map location of mouse *Tbx15*, [@pbio-0020003-Agulnik1] suggested human *Tbx15* that lies on Chromosome 1p11.1 as a candidate for acromegaloid facial appearance (AFA) syndrome, for which there is a weak positive LOD score to Chromosome 1p ([@pbio-0020003-Hughes1]). Originally described as a rare autosomal-dominant syndrome with progressive facial coarsening, overgrowth of the intraoral mucosa, and large, doughy hands, more recent case reports describe macrosomia, macrocephaly, or both and generalized hypertrichosis with progressive coarsening ([@pbio-0020003-Dallapiccola1]; [@pbio-0020003-Irvine1]; [@pbio-0020003-da1]; [@pbio-0020003-Zelante1]). The *de^H^* phenotype exhibits little overlap with these features; instead, we suggest a more likely candidate for mutations of human *TBX15* would be frontofacionasal syndrome, an unmapped autosomal recessive condition characterized by brachycephaly, blepharophimosis, and midface hypoplasia ([@pbio-0020003-Reardon1]). Two of us (S. Kuijper and F. Meijlink) became interested in the *de^H^* mutation because of its effects on skeletal development ([@pbio-0020003-Curry1]) and the possibility that the *aristaless*-related gene *Alx3* might be allelic with *droopy ear* ([@pbio-0020003-ten1]). In spite of similarities between skeletal phenotypes of *de^H^* and *Alx3* or *Alx4* mutants, subsequent experiments (unpublished data) excluded allelism of *Alx3* and *de^H^*, and a full description of the *Tbx15* skeletal phenotype will be published elsewhere. Developmental Mechanism of *Tbx15* Expression and Action in the Skin {#s3c} -------------------------------------------------------------------- Our attention to the role of *Tbx15* in pigment patterning was motivated by the effects of *Agouti* in postnatal animals. However, *Agouti* is also expressed in the embryo, where it provides a convenient marker of ventral dermis identity. Because an expanded domain of embryonic *Agouti* expression in *de^H^*/*de^H^* animals is detectable by E14.5, the effects of *Tbx15* on dorsoventral patterning must occur prior to this time. Among other T-box genes whose developmental actions are at least partially understood, two general themes have emerged, one focused on the ability to specify alternative fates for an undifferentiated group of precursor cells and another focused on the ability to support proliferative expansion of a cell population whose fate is already determined (reviewed in [@pbio-0020003-Tada1]). Either mechanism may apply to the apparent dorsal-to-ventral transformation in *de^H^*/*de^H^* mice. For example, while the expanded domain of *Agouti* expression in postnatal *de^H^*/*de^H^* animals can be traced to events that occur between E11.5 and E13.5, the underlying cause may be that embryonic cells in dorsolateral mesenchyme acquire a ventral rather than dorsal identity or that those cells fail to proliferate normally, followed by compensatory expansion of ventral cells. Cell lineage studies should provide a definitive answer, but we favor the latter hypothesis, because measurements of dorsoventral regions according to hair color in *de^H^*/*de^H^* mice revealed a small increase of the cream-colored ventral region in addition to the approximate doubling of the yellow flank region (see [Figure 2](#pbio-0020003-g002){ref-type="fig"}). In embryonic mesenchyme, expression of *Tbx15* and *Agouti* are complementary, and it is possible that *Tbx15* acts directly to inhibit *Agouti* transcription in dorsolateral mesenchyme. However, the ability of *Tbx15* to suppress expression of the ventral-specific *Agouti* isoform in postnatal mice is likely to be indirect, since postnatal expression of *Tbx15* occurs broadly along the dorsoventral axis and overlaps extensively with that of *Agouti*. In either case, the targets of *Tbx15* action in the skin include genes in addition to *Agouti*, since hair length and melanocyte distribution exhibit a demonstrable, albeit subtle, alteration in animals that carry a null *Agouti* allele. One potential target is *Alx4*, which, like *Agouti*, is expressed in ventral embryonic mesenchyme, and, when mutated, affects hair-follicle as well as limb and craniofacial development ([@pbio-0020003-Qu1], [@pbio-0020003-Qu2]; [@pbio-0020003-Wu1]; [@pbio-0020003-Wuyts1]; [@pbio-0020003-Mavrogiannis1]). However, expression of ventral markers such as *Alx4*, as well as *Alx3* and *Msx2*, appears to be unaffected at E11.5 in *de^H^*/*de^H^* embryos (data not shown). Differences and Similarities to Dorsoventral Limb Patterning {#s3d} ------------------------------------------------------------ Loss of *Tbx15* also affects regional distribution of hair color in the limbs, with areas that would normally produce black hair giving rise to yellow hair instead. However, neither normal patterns of pigment-type synthesis in the limb nor their disruption in *de^H^*/*de^H^* mice correspond to obvious developmental compartments. Furthermore, losses of function for *En1* or *Wnt7a*, which cause a partial transformation of the distal limb from dorsum to ventrum ([@pbio-0020003-Loomis1]) or ventrum to dorsum ([@pbio-0020003-Parr1]), respectively, have no effect on regional patterns of *Agouti* expression or distribution of hair-color regions (Y. Chen, unpublished data). (Ectopic pigmentation of the ventral footpads that develops in *En1* mutant mice is unrelated to pigment-type synthesis and instead likely reflects a requirement for *En1*, independent of *Wnt7a*, to repress migration or proliferation (or both) of pigment cells in ventral epidermis \[[@pbio-0020003-Cygan1]; [@pbio-0020003-Loomis2]\].) These considerations notwithstanding, control of dorsoventral trunk pattern by *Tbx15* shares certain features with control of dorsoventral limb patterning by *Lmx1b*, a LIM domain transcription factor that acts downstream of *Wnt7a* and *En1* ([@pbio-0020003-Riddle1]; [@pbio-0020003-Vogel1]; [@pbio-0020003-Cygan1]; [@pbio-0020003-Logan1]; [@pbio-0020003-Loomis2]; [@pbio-0020003-Chen1]). Both *Tbx15* and *Lmx1b* act autonomously in mesenchymal cells to promote a dorsal identity, yet have expression domains that do not correspond to cell lineage compartments in the flank (*Tbx15*) or the limb (*Lmx1b*) ([@pbio-0020003-Altabef1]; [@pbio-0020003-Michaud1]). In the case of *Lmx1b*, its expression in the distal limb depends on *Wnt7a* produced in the overlying dorsal ectoderm ([@pbio-0020003-Riddle1]; [@pbio-0020003-Cygan1]; [@pbio-0020003-Loomis2]). *Wnt7a*, in turn, is restricted to dorsal ectoderm by *En1* in the ventral ectoderm ([@pbio-0020003-Loomis1]; [@pbio-0020003-Cygan1]; [@pbio-0020003-Logan1]), whose expression marks a lineage boundary coincident with the dorsoventral midline of the apical ectodermal ridge ([@pbio-0020003-Altabef1]; [@pbio-0020003-Michaud1]; [@pbio-0020003-Kimmel1]). As described above, *En1* or *Wnt7a* mutations have not been reported to affect patterns of hair-color distribution (C. Loomis, personal communication; [@pbio-0020003-Parr1]; [@pbio-0020003-Loomis1]). However, the essential theme that ectodermal lineage compartments control the fate of underlying mesenchyme in developing limbs may apply to the trunk as well as the limb. The mammary glands also develop at a stereotyped dorsoventral position and depend on epithelial--mesenchymal interactions. However, the number and apparent position of the mammary glands are normal in *de^H^*/*de^H^* animals, indicating the existence of additional mechanisms that control dorsoventral patterning in the trunk as well as in the limbs. These ideas are summarized in the model shown in [Figure 9](#pbio-0020003-g009){ref-type="fig"}B. We speculate that a diffusible signal from dorsal trunk ectoderm, at or prior to E11.5, promotes expression of *Tbx15* in dorsal trunk mesenchyme, which then establishes dorsal positional identity of those cells as manifested by differences in *Agouti* expression, pigment-cell development, and hair growth. Because the ventral limit of *Tbx15* expression corresponds to the dorsal limit of *En1* expression and because the normal position of the pigmentation boundary lies approximately in register with the limb-bud outgrowths, we depict the position of a putative dorsoventral boundary in trunk ectoderm as coincident with the limb dorsoventral boundary. This model is consistent with studies in the chick, where distinct dorsal and ventral lineage compartments exist for ectoderm in both the limb ([@pbio-0020003-Altabef1], [@pbio-0020003-Altabef2]; [@pbio-0020003-Michaud1]; [@pbio-0020003-Kimmel1]) and interlimb regions ([@pbio-0020003-Altabef1], [@pbio-0020003-Altabef2]), but not for limb mesoderm ([@pbio-0020003-Altabef1]; [@pbio-0020003-Michaud1]). In fact, the same mechanism that determines dorsoventral position of the limbs and the apical ectodermal ridge may also act on expression of *Tbx15* in the trunk, since ectopic limbs induced in the interlimb region by application of FGF beads develop along a single line that is coincident with normal limb buds (and the future pigmentation boundary) ([@pbio-0020003-Cohn1]; [@pbio-0020003-Crossley1]; [@pbio-0020003-Vogel2]; [@pbio-0020003-Altabef1], [@pbio-0020003-Altabef2]). Our model predicts that ectopic expression of *Tbx15* in ventral mesenchyme should give rise to a dorsalized pigmentation phenotype and could be tested with gain-of-function approaches. However, *Tbx15* expression is very dynamic and is restricted to dorsal mesoderm only from E11.5 to E13.5. It is possible that *Tbx15* influences skin patterning in a very narrow window of development; alternatively, establishment of dorsal identity by *Tbx15* may require another as-yet-unidentified factor that is only present in the mesenchyme underlying dorsal ectoderm. Pigmentation Patterns and *Tbx15* in Other Mammals {#s3e} -------------------------------------------------- The lateral somitic frontier, defined as the lineage boundary between somite-derived versus lateral plate-derived mesoderm, is established during somitogenesis early in development ([@pbio-0020003-Mauger1]; [@pbio-0020003-Christ1]; [@pbio-0020003-Olivera-Martinez1]; [@pbio-0020003-Nowicki1]), but remains distinct in postnatal animals despite the potential for extensive cell mixing (see [Figure 8](#pbio-0020003-g008){ref-type="fig"}). However, our transplantation and fate-mapping studies demonstrate that the lateral somitic frontier lies dorsal to the pigmentation boundary and does not obviously correlate with a difference in skin morphology. An additional dorsoventral domain that is not externally apparent has emerged from studies of *Msx1*, whose expression marks a subgroup of somite-derived mesenchymal cells that contribute to dermis in a narrow stripe along the paraspinal region ([@pbio-0020003-Houzelstein1]). Thus, there exist at least three distinct boundaries in postnatal mammalian skin that are parallel to the sagittal plane, marked by differences in pigment-type synthesis, differences in cell lineage, and differences in expression of *Msx1*. In rodents, only the pigmentation boundary is evident externally, but many mammals have more complicated patterns of hair type, length, and/or color that vary along the dorsoventral axis. Raccoons, squirrels, skunks, and many different ungulates exhibit lateral stripes whose developmental origins have not been investigated, but may correspond to the lateral somitic frontier, the paraspinal *Msx1* compartment, or an interaction between these domains. The effect of *Tbx15* on pigmentation in laboratory mice is reminiscent of coat-color patterns in both selected and natural populations of other mammals. Saddle markings are common in some dog breeds, such as German shepherds, and in certain populations of Peromyscus polionotus, in which a dorsal extension of ventral depigmentation provides an adaptive advantage to subspecies that live on white sand reefs ([@pbio-0020003-Blair1]; [@pbio-0020003-Kaufman1]; [@pbio-0020003-Belk1]). Neither German shepherds nor deer mice have craniofacial characteristics similar to the *de^H^* mutation, but the pigmentation patterns in these animals could represent alterations in the regulation or action of *Tbx15* activity. From the opposite perspective, the effects of *Tbx15* on coat color are only apparent in certain genetic backgrounds and may not be evident at all in mammals that lack dorsoventral pigmentation patterns. Studying the sequence and expression of *Tbx15* in other vertebrates may provide additional insight into patterns that affect the skeleton as well as the pigmentary system. Materials and Methods {#s4} ===================== {#s4a} ### Mice {#s4a1} All mice were obtained originally from The Jackson Laboratory (Bar Harbor, Maine, United States), except the BA strain (Stanford Veterinary Services Center, Stanford, California, United States), *Hoxb6-Cre* transgenic mice (kindly provided by M. Kuehn of the National Institutes of Health, Bethesda, Maryland, United States), mice carrying the *R26R lacZ* reporter allele (kindly provided by P. Soriano, Fred Hutchinson Cancer Research Center, Seattle, Washington, United States), and C57BL/6J (B6) *a^e^/a^e^* mice (kindly provided by L. Siracusa, Jefferson Medical College, Philadelphia, Pennsylvania, United States). The *de^H^* mutation arose in the 1960s in Harwell, probably on the BN strain background (C. Beechey, personal communication). We obtained *de^H^* on a B6/EiC3H background, introduced the *a^t^* allele from the BTBR strain, and have maintained the line as a mixed *de^H^*/+ × *de^H^*/+ intercross stock with periodic outcrossing to BTBR or B6. For timed matings, the morning of the plug was considered E0.5. Postnatally, the day of birth was considered to be P0.5. ### Phenotypic analysis {#s4a2} For measurements of hair length and color, the entire interlimb region of skin was first dissected with a single incision at the dorsal midline and preserved with powdered sodium bicarbonate. Slices 2--2.5 mm in width were then prepared parallel to the dorsoventral axis, hair length boundaries determined from electronic images with Adobe Photoshop (San Jose, California, United States), and measurements obtained using ImageJ (National Institutes of Health). This approach samples awls and auchenes, because they are much thicker and therefore visually more predominant than zigzag underhairs. To assess dorsoventral variation in hair-type distribution, several hundred hairs were plucked from the middorsum or midventrum of 8-wk-old male BA strain animals, then sorted and categorized using a dissection microscope. No attempt was made to distinguish between awls and auchenes. For skin histology, 12 μm sections from paraffin-embedded tissue were stained with hematoxylin and eosin. For DOPA staining, the dermis and epidermis were split after 3 h of incubation in 2 M sodium bromide at 37°C (this preparation causes most hair follicles to remain with the dermis), individually fixed for 1 h, then rinsed and stained with 0.1% L-DOPA (Sigma, St. Louis, Missouri, United States), 0.1 M sodium phosphate buffer (pH 6.8) for 5 h at 37°C in the dark, changing the staining solution after 1 h. The samples were then fixed overnight, dehydrated, and mounted. This method is sufficient to stain interfollicular melanocytes without creating a high background. The fixative used was always 4% paraformaldehyde. ### Positional cloning {#s4a3} A high-resolution map for *de^H^* was generated from an intersubspecific intercross between *de^H^*/*de^H^* and CAST/Ei mice. We initially localized *de^H^* to a 1 cM interval between D3Mit233 and D3Mit11. F~2~ animals carrying recombinant chromosomes between these markers whose genotype at *de* was indeterminate (*de^H^*/+ or +/+) were progeny-tested by crossing to *de^H^*/*de^H^* animals. Further genetic mapping established a minimal region of 0.1 cM between D3Mit213 and 16.MMHAP32FLF1; these markers were used to initiate construction of a physical map with BAC genomic clones (Research Genetics, Huntsville, Alabama, United States, and Genome Systems, St. Louis, Missouri, United States). End sequence from those BACs was used to develop SSCP markers M1 to M3, as depicted in [Figure 3](#pbio-0020003-g003){ref-type="fig"}, and to establish a minimal physical interval of 1.4 Mb. Primer pairs used were TTCCCTCCAATAAGTTCTGGGTACC and AAGCTTGCTGCTCTGGATTCCATTTGTAG for M1, CCTTCATTTTTTTTTCAAGTAAAA and AAGCTTGGCTTAGTCCCAGTGGC for M2, CCTCCAGGAAGATCTACTAGGCAC and ATGGAAAAAAAAAAGTAAGATTGAAAG for M3, and TGGTTATCGATCTGTGGACCATTC and AAGTGAGAGAGCAGGATGGACCAC for M4 (the M4 marker represents STS 16.MMHAP32FLF1). Genomic sequence and annotations were obtained from the UCSC Genome Browser February 2003 assembly version mm3 (<http://genome.ucsc.edu>); the 1.4 Mb interval between M1 and M4 contains eight genes: four hydroxysteroid dehydrogenase isomerases, *Hsd3b3, Hsd3b2*, *Hsd3b6*, and *Hsd3b1*; an hydroacid oxidase, *Hao3*; a tryptophanyl-tRNA synthetase, *Wars2*; a T-box gene, *Tbx15*; and a novel gene, *4931427F14Rik*. In the genome sequence, M1 primers correspond to AGGCCTCCAATAAGTTCTGGGTACC and AAGCTTGCTCTCTGGATTCCATTTGTAG, the M2 reverse primer corresponds to AAGCTTGGCTTTAGTCCCAGTGGGC, and the M3 primers correspond to CCTCCAGGAAGAATCTACTAGGCAC and AATGAAAAAAAAAAAAGTAAGATTGAAAG. Minor differences among the sequences of the primers we obtained from the BAC ends and the public genome sequence may represent strain differences or sequencing errors on the BAC DNA. A multiplex genotyping assay was developed to genotype for the *de^H^* deletion using primers GGAGCAGATCCAATTGCTTT, TCCATAGCCCATCTTCACAA, and CATGTCCACTTCTGCTTCCA. This PCR assay produces a 392 bp product from the *de^H^* chromosome and a 595 bp product from the nonmutant chromosome. ### Gene targeting {#s4a4} A targeted allele of *Tbx15* was constructed using the same approach described in [@pbio-0020003-Russ1]. In brief, an *IRES-LacZ-neo* cassette with 5′ and 3′ homology arms of 3.5 kb and 1.8 kb was inserted into a unique BamHI site that lies 479 nucleotides downstream of the transcriptional initiation site (relative to the mRNA sequence) in exon 3. Positive ES clones were injected into B6 blastocysts, and chimeric founders crossed to either B6 mice or to *de^H^*/+ animals. ### In situ hybridization {#s4a5} In situ hybridization was carried out on 12-μm paraffin sections using digoxigenin-labeled RNA probes (Roche Diagnostics, Indianapolis, Indiana, United States) according to standard protocols ([@pbio-0020003-Wilkinson1]). Embryos and postnatal skin samples were obtained from intercrosses of *de^H^*/*+* mice. Embryos E13.5 or younger were fixed for 24 h; those older than E13.5 and postnatal skin were fixed for 36--48 h prior to embedding. The *Tbx15* probe was generated by RT--PCR using primers GGCGGCTAAAATGAGTGAAC and TGCCTGCTTTGGTGATGAT (corresponds to exons 1 and 2), and the *En1* probe was generated by PCR from genomic DNA using primers ACGCACCAGGAAGCTAAAGA and AGCAACGAAAACGAAACTGG (located in the last exon). The *Agouti* probe corresponds to the protein-coding sequence. ### Embryonic skin transplantation {#s4a6} (BTBR-*a^t^*/*a^t^* × B6-*a*/*a*)F~1~ embryos at E12.5 were dissected in sterile Tyrode\'s solution, and embryonic skin was divided into dorsal, flank, and ventral pieces, each 1--2 mm^2^ in size, as shown in [Figure 7](#pbio-0020003-g007){ref-type="fig"}. Skin fragments were grafted to the testes of congenic animals as follows. After anesthetization with 2.5% Avertin, a 1.5-cm incision in the skin and body wall was made at a point level with the top of the limbs. The fat pads were pulled out and laid on the outside of the body, exposing the testes. Forceps were used to introduce a small hole in the testis capsule through which a piece of dissected embryonic skin was inserted, the testes were then replaced into the abdominal cavity, and the wound was closed in both the body wall and the skin. After 21 days, mice that received grafts were sacrificed and the resulting hair was dissected from the testes and examined. ### Fate-mapping the lateral somitic frontier {#s4a7} The *Hoxb6-Cre* transgene described by Kuehn and colleagues ([@pbio-0020003-Lowe1]) is expressed in the lateral plate but not the somitic mesoderm of the trunk, beginning at E9.5. Animals doubly heterozygous for this transgene and the *R26R* reporter gene were used as a source of whole skin at P1.5 or P4.5. Skin sections parallel to the dorsoventral axis were prepared with a single incision along the ventral midline and stained for β-galactosidase activity using standard protocols at room temperature. The P1.5 sample was stained overnight and the P4.5 samples were stained for 5.5 h. Similar nonstained skin sections were prepared from animals carrying the *a^t^* allele. Images of the different skin fragments were aligned and scaled, and the relative position of the somite--lateral plate and the pigmentation boundaries were measured using ImageJ. Supporting Information {#s5} ====================== The GenBank (<http://www.ncbi.nlm.nih.gov/Genbank/index.html>) accession numbers discussed in this paper are for *4931427F14Rik* (AK016477), *Agouti* gene (L06451), *Alx3* (U96109), *Alx4* (AF001465), *En1*(L12703), *M6pr-ps* (X64069), *Tbx14* (AF013282), *Tbx15* (AF041822), *Tbx18* (AF306666), and *Tbx22* (NM\_145224). The OMIM (<http://www.ncbi.nlm.nih.gov/omim/>) accession numbers discussed in this paper are for acromegaloid facial appearance (MIM 102150), frontofacionasal syndrome (MIM 229400) and human syndrome X-linked cleft palate and ankyloglossia (MIM 303400). We thank Colin Beechey and Bruce Cattanach for providing information about the origin of *de^H^*, Cindy Loomis for communicating unpublished results, and Michael Kuehn for providing *Hoxb6-Cre* transgenic mice. We are especially grateful to Hirotake Ono for discussion, to Hermie Manuel and to Carla Kroon-Veenboer for technical help, and to David Kingsley for first drawing our attention to the effects of *droopy ear* on pigmentation. SIC is supported by a grant from The Donald E. and Delia B. Baxter Foundation. GSB is an associate investigator of the Howard Hughes Medical Institute. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** SIC and GSB conceived and designed the experiments. SIC, CDVR, CC, AR, and SK performed the experiments. SIC, CDVR, CC, SK, and FM analyzed the data. YC and AR contributed reagents/materials/analysis tools. SIC and GSB wrote the paper. Academic Editor: Brigid L. M. Hogan, Duke University Medical Center *a^e^* : *extreme nonagouti* allele AFA : acromegaloid facial appearance *a^t^* : *black-and-tan* allele *A^W^* : *white-bellied Agouti* allele B6 : C57BL/6J mice BAC : bacterial artificial chromosome *de^H^* : *droopy ear* allele that arose in Harwell E10.5 : embryonic day 10.5 *En1* : *Engrailed1* *M6pr-ps* : mannose-6-phosphate receptor pseudogene P4.5 : postnatal day 4.5 *Tbx15* : T-box gene 15

PubMed Central

2024-06-05T03:55:46.007645

2004-1-20

PMC314464

Introduction {#s1} ============ Malaria is one of the most devastating infectious diseases in the world, killing more than 1 million people per year. Malaria is transmitted by bites of infected mosquitoes that inject sporozoites under the skin. The first obligatory step for these parasites to establish infection in humans is migration to hepatocytes, where they proliferate and develop into the erythrocyte-invasive form ([@pbio-0020004-Sinnis1]). This liver-invasive stage has been demonstrated as a promising target for antimalarial strategies that aim to establish sterile immunity against the malarial parasite ([@pbio-0020004-Nussenzweig1]; [@pbio-0020004-Hoffman1]). However, the mechanisms underlying the parasite\'s liver infection are largely unknown. In particular, it has been controversial how sporozoites reach the hepatocytes that are separated from blood circulation by the liver sinusoidal layer. The routes the sporozoites use to cross this layer, the modes of motility on which their migration is based, and the molecules of the parasite involved in this process are poorly understood. Malarial parasites develop into sporozoites in the mosquito midgut and then invade the salivary gland, where they wait to be transferred to the mammalian host ([@pbio-0020004-Menard1]). Once injected by mosquito bites under the skin, sporozoites enter the blood circulation and are carried to the liver by the bloodstream ([@pbio-0020004-Sinnis2]; [@pbio-0020004-Menard1]; [@pbio-0020004-Mota1]). In the liver, they are thought to be arrested on the inner surface of the liver sinusoidal vein and then leave the vein and infect the hepatocytes by crossing the sinusoidal wall ([@pbio-0020004-Sinnis2]). This wall is a single-cell layer mainly composed of sinusoidal endothelial cells and Kupffer cells, which are hepatic macrophages. Several models have been proposed to explain how sporozoites cross this layer. Some authors proposed that sporozoites infect hepatocytes after crossing the liver endothelial cell through fenestrations in this cell ([@pbio-0020004-Vanderberg1]), but these openings are too small for sporozoites to freely pass through ([@pbio-0020004-Mota1]). Other authors have suggested that Kupffer cells are gates for sporozoites to access hepatocytes, based on the ultrastructural observation that sporozoites were found inside Kupffer cells shortly after intravenous inoculation ([@pbio-0020004-Mota1]). This Kupffer cell hypothesis, however, has not been convincingly demonstrated, because other tools for probing into this event were lacking. Furthermore, the observation that the sporozoites in Kupffer cells sometimes have a vacuole around them makes the conclusion uncertain. Some authors have proposed that sporozoites are passively engulfed by Kupffer cells and then carried to the hepatocyte ([@pbio-0020004-Meis1]), and some have proposed that this migration is based on active motility accompanied by vacuole formation ([@pbio-0020004-Pradel1]). The malarial parasite has no locomotory organelles such as flagella or cilia. Motility of the host-invasive stages of the malarial parasite, including the sporozoite, is dependent on secretion of micronemes that are organelles occupying the cytoplasm of the parasite ([@pbio-0020004-Sultan1]; [@pbio-0020004-Menard2]). Micronemal components, which may include several attachment proteins, are secreted from the apical pore during parasite movement and are translocated backwards along the parasite cell surface by actomyosin motors of the parasite. This surface movement is believed to generate traction for parasite-invasive motility. Salivary gland sporozoites display three modes of motility in vitro dependent on secretion of micronemes ([@pbio-0020004-Mota1]). One mode is gliding motility on a solid surface, which can be observed as circular movement on a glass slide, probably representing gliding motility on the cell surface. The other two are cell-invasive motilities: cell-infection and cell-traversal motility ([@pbio-0020004-Mota2]; [@pbio-0020004-Kappe1]). Cell-infection motility is accompanied by vacuole formation and is followed by parasite development into exoerythrocytic forms (EEFs). Cell-traversal motility, on the other hand, involves plasma-membrane disruption and is followed by migration through the cytoplasm and eventual escape from the cell. Recently, [@pbio-0020004-Mota3]) revealed that this type of cell-invasion motility can be identified by conventional cell-wound assay. According to the observation that passage through some hepatocytes by this motility precedes hepatocyte infection, they proposed the hypothesis that this motility is necessary for sporozoites to be activated for hepatocyte infection ([@pbio-0020004-Mota3]). However, the role of this motility in liver infection remains unclear. Aiming at identification of molecules involved in sporozoite infection, we screened an expressed sequence tag (EST) database of the salivary gland sporozoite of a rodent malarial parasite, Plasmodium berghei. In this paper, we report a novel microneme protein, named SPECT (sporozoite microneme protein essential for cell traversal), which is specifically produced by the liver-infective sporozoite and is essential for the sporozoite\'s cell-passage ability. By using *spect*-disrupted parasites, we show that cell-passage ability of the sporozoite plays a critical role in malarial transmission to the vertebrate host and is required for sporozoites to access hepatocytes by traversal of the liver sinusoidal cell layer. In addition, we provide a model of sporozoite liver infection, which suggests an answer to the question of how sporozoites reach the hepatocytes. Results {#s2} ======= Identification of cDNA Encoding SPECT from P. berghei Salivary Gland Sporozoite EST Database {#s2a} -------------------------------------------------------------------------------------------- Sporozoites acquire the ability to infect the mammalian liver after infection of the mosquito salivary glands ([@pbio-0020004-Sultan2]), indicating that novel protein synthesis for liver infection begins in this stage ([@pbio-0020004-Matuschewski1]). To search for malarial genes involved in liver infection, we screened an EST database of P. berghei salivary gland sporozoites. We assembled 3,825 ESTs, obtained 502 contigs, and screened them for genes encoding secretory proteins or membrane-associated proteins, which may participate in host--parasite interactions. This screening was started from the contigs containing a high number of ESTs, since the number of ESTs may correlate with the expression levels of the respective genes. In this process, we identified a contig composed of ten ESTs, encoding a putative secretory protein of 241 amino acids ([Figure 1](#pbio-0020004-g001){ref-type="fig"}A). The expected molecular mass for the N-terminal signal sequence-processed form of this protein was 25 kDa. We named this protein SPECT (sporozoite microneme protein essential for cell traversal), since it is essential for sporozoite passage through a host cell, as described later. ::: {#pbio-0020004-g001 .fig} Figure 1 ::: {.caption} ###### Sequence Analysis of *spect* cDNA \(A) Nucleotide sequence of *spect* cDNA (top lane) and the deduced amino acid sequence (bottom lane) are shown. The predicted N-terminal signal sequence is underlined. The numbers indicate positions of the nucleotides starting from the 5′ end. The asterisks indicate the termination codon. \(B) A comparison of deduced amino acid sequences of *P. berghei spect* (top) and *P. falciparum spect* (bottom). Gaps are introduced to obtain optical matching by using GENETIX-MAC software. Asterisks or dots show conserved or similar residues, respectively. The amino acid numbers from the first Met residue are shown on the left of each line. :::  ::: Southern blot analysis showed that the *spect* gene is a single-copy gene (data not shown). Sequence analysis of the *spect* gene identified four introns (data not shown). A computer search of *Plasmodium* genome databases ([@pbio-0020004-Carlton1]; [@pbio-0020004-Gardner1]) revealed that this gene is conserved through several *Plasmodium* species. The orthologous protein in *P. falciparum,* the clinically most important human malaria parasite, shared 45.6% sequence identity with P. berghei SPECT ([Figure 1](#pbio-0020004-g001){ref-type="fig"}B). SPECT Is Produced Specifically by Salivary Gland Sporozoites and Localized in Micronemes {#s2b} ---------------------------------------------------------------------------------------- The expression profile of this gene in the malarial life cycle was investigated. Immunofluorescent analysis in all host-invasive stages showed that SPECT production was restricted to sporozoites in the salivary gland ([Figure 2](#pbio-0020004-g002){ref-type="fig"}A). It is noteworthy that SPECT is not detected in sporozoites in the midgut, because this expression profile strongly suggests that SPECT is specifically involved in liver infection. Western blot analysis revealed SPECT as a 22 kDa protein in salivary gland sporozoites, but not in midgut sporozoites ([Figure 2](#pbio-0020004-g002){ref-type="fig"}B), confirming that SPECT is produced after invasion into the salivary gland. Immunoelectron microscopy showed that SPECT is localized in the sporozoite to micronemes that are secretory organelles occupying the cytoplasm ([Figure 2](#pbio-0020004-g002){ref-type="fig"}C). Micronemes are common to motile stages of *Plasmodium* parasites and play a central role in host-invasive motility ([@pbio-0020004-Sultan1]; [@pbio-0020004-Menard2]). Taken together, these results indicate that SPECT plays a role in the liver-invasive motility of the sporozoite. ::: {#pbio-0020004-g002 .fig} Figure 2 ::: {.caption} ###### SPECT Is a Microneme Protein Specifically Produced in the Liver-Infective Sporozoite Stage \(A) Indirect immunofluorescence microscopy of all four invasive forms of the malarial parasite (indicated over the panel). Parasites were stained with primary antibodies against SPECT, followed by FITC-conjugated secondary antibodies. SPECT was detected only in the salivary gland sporozoite, the liver-infective stage. The corresponding phase-contrast or DAPI-stained image (Phase or DAPI) is shown under each image. Scale bars, 5 μm \(B) Western blot analysis of SPECT production in the midgut sporozoite (M) and the salivary gland sporozoite (S). Lysate of 500,000 sporozoites was loaded onto each lane and detected with the same antibody used in (A). SPECT was detected as a single band of 22 kDa (arrowhead) only in the salivary gland sporozoite. \(C) Immunoelectron microscopy of sporozoites in the salivary gland. Ultrathin sections of a mosquito salivary gland infected with sporozoites were incubated with the same antibody used in (A) followed by secondary antibodies conjugated with gold particles (15 nm). Particles were localized to micronemes (Mn) but not to rhoptories (Rh). Axial (inset) and vertical images are shown. Scale bars, 0.5 μm. :::  ::: SPECT Plays an Important Role in Sporozoite Infection of the Host Liver {#s2c} ----------------------------------------------------------------------- To investigate the function of SPECT protein, we generated *spect*-disrupted parasites by homologous recombination ([Figure 3](#pbio-0020004-g003){ref-type="fig"}A). The *spect* disruptants were selected by the antimalarial drug pyrimethamine and were separated from wild-type parasites by limiting dilution. Disruption of the *spect* locus was confirmed by Southern blot analysis ([Figure 3](#pbio-0020004-g003){ref-type="fig"}B). To exclude the possibility that the *spect*-disrupted populations obtained were derived from a single clone, two independently obtained *spect*-disrupted populations (*spect*(−)1 and *spect*(−)2) were used in the following experiments. ::: {#pbio-0020004-g003 .fig} Figure 3 ::: {.caption} ###### Targeted Disruption of the *spect* Gene \(A) Schematic representation of targeted disruption of the *spect* gene. The targeting vector (top) containing a selectable marker gene is integrated into the *spect* gene locus (middle) by double crossover. This recombination event resulted in the disruption of the *spect* gene and confers pyrimethamine resistance to disruptants (bottom). \(B) Genomic Southern blot hybridization of wild-type (WT) and *spect*(−) populations. Genomic DNA isolated from the respective parasite populations was digested with EcoT22I and hybridized with the probe indicated in (A) by a solid bar. By integration of the targeting construct, the size of detected fragments was decreased from 1.9 kbp to 1.2 kbp. The result is shown for two independently prepared populations, *spect*(−)1 and *spect*(−)2. \(C) Immunofluorescence microscopy of the wild-type (WT) and *spect*(−) parasite. Sporozoites were collected from the salivary gland and stained with primary antibody against SPECT followed by FITC-conjugated secondary antibodies. The apical end of each sporozoite is indicated by an arrowhead. :::  ::: In the intra-erythrocytic stage, SPECT gene disruption did not affect parasite proliferation, as the growth rates in rat blood were almost the same in the *spect*-disrupted and wild-type populations (data not shown). Furthermore, disruption of the gene did not affect parasite development in the mosquito vector, as numbers of sporozoites residing in the midgut and in the salivary glands were similar in the *spect*-disrupted and wild-type populations ([Table 1](#pbio-0020004-t001){ref-type="table"}). ::: {#pbio-0020004-t001 .table-wrap} Table 1 ::: {.caption} ###### SPECT Disrupted Parasites Develop Normally into Sporozoites and Invade the Salivary Gland in the Mosquito Vector :::  Mosquitoes were fed on mice infected with spect(−) parasite populations or wild-type polyclonal populations. Sporozoites were collected separately from the midgut and the salivary glands of mosquitoes 24--28 d after feeding and then counted. Each value is the mean of the number with its standard error from three independent experiments ::: Next, the liver infectivity of the *spect*-disrupted sporozoites was examined. Rats were intravenously inoculated with sporozoites, and the progress of parasitemia, the percentage of infected erythrocytes, was measured in the exponential growth period (from 3.5 d to 5 d after inoculation). It is thought that the parasitemias reflect the liver infectivity of the respective parasite populations, since the growth rates of their intraerythrocytic stages are similar (shown by the parallel slopes of the increase in parasitemia in [Figure 4](#pbio-0020004-g004){ref-type="fig"}). Based on the average parasitemia at 3.5 d after inoculation of 30,000 sporozoites, the liver infectivities of the two disruptant strains were estimated to be 15- and 28-fold lower, respectively, than that of the wild-type. These results are consistent with the observation that the parasitemias after injection of 30,000 disruptant sporozoites were lower than that from 3,000 wild-type sporozoites. ::: {#pbio-0020004-g004 .fig} Figure 4 ::: {.caption} ###### Targeted Disruption of *spect* Results in Reduction of Sporozoite Infectivity to the Liver \(A) The salivary gland sporozoites of each parasite population were injected intravenously into five rats. The parasitemia of each rat was checked by a Giemsa-stained blood smear after inoculation on the days indicated. The average parasitemia after inoculation of 30,000 sporozoites was significantly low in disruptant populations, whereas their growth rates in the blood were essentially the same as the wild-type. The numbers of parasites inoculated were as follows: 30,000 *spect*(−)1 (open circles), 30,000 *spect*(−)2 (open triangles), 30,000 wild-type (filled circles), and 3,000 wild-type (filled squares). Values shown represent the mean parasitemia (± SEM) of five rats. \(B) The salivary gland sporozoites (500,000) of wild-type or *spect*-disrupted parasites were inoculated intravenously into 3-wk-old rats. After 24 h, the livers were fixed with paraformaldehyde and frozen. The number of EEFs on each cryostat sections was estimated by indirect immunofluorescence analysis using anti-CS antiserum. Values shown represent the mean number of EEFs per square millimeter (± SEM) of at least three rats. :::  ::: The liver infectivity was also evaluated by the number of early EEFs. Frozen sections of the rat liver was prepared 24 h after sporozoite injection and EEFs were counted by immunofluorescence microscopy. As shown in [Figure 4](#pbio-0020004-g004){ref-type="fig"}B, EEFs were approximately 30-fold decreased by *spect* gene disruption. This reduction rate agrees well with that estimated by parasitemia. These results indicate that SPECT plays a role in the process of sporozoite invasion into the liver. SPECT Is Essential for Sporozoite Cell-Passage Ability {#s2d} ------------------------------------------------------ Localization of SPECT in micronemes indicates its involvement in the invasive motility of the sporozoite. The motility of *spect*-disrupted sporozoites was investigated by three in vitro assays corresponding to three modes of motility of the sporozoite. First, we checked gliding motility on a solid surface, which is essential for sporozoite infectivity. Most disruptants displayed a typical circular movement, and the proportion of motile sporozoites was almost identical in disruptant and wild-type parasites (63.6% and 67.5%, respectively), showing that their gliding motility is not affected by SPECT gene disruption. Second, we examined the ability of the sporozoites to infect hepatocytes. This was assayed by formation of EEFs in a human hepatoma cell line, HepG2 ([@pbio-0020004-Hollingdale1]). As shown in [Figure 5](#pbio-0020004-g005){ref-type="fig"}A, the disruptants formed EEFs in similar numbers to the wild-type, indicating that they retain normal infectivity to the hepatocyte. Third, we examined cell-traversal ability that takes place prior to hepatocyte infection. This was estimated by the number of membrane-wounded cultured cells that were labeled by uptake of fluorescein isothiocyanate (FITC)-conjugated dextran from the medium ([@pbio-0020004-Mota2]). As shown in [Figure 5](#pbio-0020004-g005){ref-type="fig"}B, the cell-wound assay using HeLa cells showed that the disruptants lost their cell-passage activity completely. The same results were obtained in HepG2 cells (data not shown). These results revealed that SPECT is specifically involved in cell-traversal ability and suggest that lack of this ability reduced liver infectivity of the disruptants. ::: {#pbio-0020004-g005 .fig} Figure 5 ::: {.caption} ###### *spect* Disruption Results in Loss of Cell-Passage Activity of the Sporozoite \(A) *spect* disruption does not affect sporozoite ability to infect hepatocytes. (Top panel) Comparison of EEF numbers between disruptants (*spect*(−)) and wild-type (WT) parasites. Salivary gland sporozoites were added to HepG2 cells and cultured for 48 h. EEFs formed were counted after immunofluorescence staining with an antiserum against CS protein. (Bottom panels) Representative fluorescence stained images. \(B) Sporozoites lacking SPECT cannot traverse HeLa cells. (Top) Comparison of cell-passage activity between disruptants and wild-type parasites. Salivary gland sporozoites were added to HeLa cells and incubated for 1 h with FITC-conjugated dextran (1 mg/ml). Cell-passage activity was estimated by the number of cells wounded by sporozoite passage, which were identified by cytosolic labeling with FITC-conjugated dextran. (Bottom) Representative fluorescence stained images. All data are mean numbers of EEFs or FITC-positive cells in a one-fifth area of an 8-well chamber slide with standard errors for at least three independent experiments. :::  ::: Cell Passage Ability Is Necessary for Sporozoites to Traverse the Sinusoidal Layer Cells and to Access Hepatocytes {#s2e} ------------------------------------------------------------------------------------------------------------------ To access the hepatocytes, sporozoites must cross the sinusoidal layer, which separates them from the circulation. We assumed that SPECT was necessary for this process. Since Kupffer cells are major components of this layer and have been reported as the main gates for sporozoite access to the hepatocyte, we prepared Kupffer cell-depleted rats by intravenous injection of liposome-encapsulated dichloromethylene diphosphonate (Cl~2~MDP) ([@pbio-0020004-Vreden1]; [@pbio-0020004-van1]) and tested them for infection by disruptant and wild-type sporozoites. As shown in [Figure 6](#pbio-0020004-g006){ref-type="fig"}A, infectivities of *spect*-disruptants assessed by parasitemia were increased by 22- and 53-fold by Kupffer cell depletion and, as a result, became equal to that of the wild-type. The numbers of early EEFs detected in the liver sections were also almost identical in wild-type and *spect*-disrupted parasites ([Figure 6](#pbio-0020004-g006){ref-type="fig"}B). These results show that the disruptants\' loss of infectivity is localized at the sinusoidal cell layer and that the cell-passage ability of the sporozoite is necessary to cross this layer and, specifically, the Kupffer cells. ::: {#pbio-0020004-g006 .fig} Figure 6 ::: {.caption} ###### Restoration of *spect*(−) Sporozoite Infectivity in Kupffer Cell-Depleted Rats \(A) Liposome-encapsulated Cl~2~MDP (filled points) or PBS (open) was injected intravenously into rats. After 48 h, 30,000 sporozoites of *spect*(−)1 (circles), *spect*(−)2 (triangles), or wild-type (squares) populations were inoculated intravenously. Parasitemia of each rat was checked by Giemsa-stained blood smears after inoculation on the days indicated. Values shown represent the mean parasitemia (± SEM) of five rats. \(B) Salivary gland sporozoites (500,000) of each parasite population were inoculated intravenously into Kupffer cell-depleted rats. After 24 h, the livers were fixed with paraformaldehyde and frozen. The number of EEFs on each cryostat section was estimated by indirect immunofluorescence analysis using anti-CS antiserum. Values shown represent the mean number of EEFs per square millimeter (± SEM) of at least three rats. :::  ::: Discussion {#s3} ========== It has been reported that the *Plasmodium* sporozoite has the ability to traverse cultured cells rapidly ([@pbio-0020004-Mota2]), but the role of this process in liver infection has remained unclear. On the other hand, it is poorly understood how the sporozoite migrates from the circulatory system to the hepatocyte. In this paper, we address these issues using a gene-targeting technique. We have shown that the cell-traversal activity of the sporozoite is necessary for it to leave the circulatory system by crossing the liver sinusoidal cell layer. These results are the first to reveal the role of cell-traversal activity in malarial transmission. In vitro cell invasion assays showed that *spect*-disrupted sporozoites completely lose cell passage activity, but preserve normal infectivity to the hepatocyte (see [Figure 5](#pbio-0020004-g005){ref-type="fig"}). These results clearly demonstrated that these two cell-invasion activities are independent of each other. This conclusion contradicts the hypothesis proposed by [@pbio-0020004-Mota3]) that cell passage activates the sporozoite for hepatocyte infection. They assumed that sporozoites traverse some hepatocytes before infecting a hepatocyte and that this passage alters their mode of cell invasion from passage to infection ([@pbio-0020004-Mota3]). Our results, however, demonstrated that lack of previous cell passage has no influence on the infectivity to hepatocytes. This independence was confirmed in vivo by the result that disruptants and wild-type showed the same liver infectivities in Kupffer cell-depleted rats (see [Figure 6](#pbio-0020004-g006){ref-type="fig"}). Therefore, sporozoites may change their mode of invasive motility according to other factors, which remain to be elucidated. We suppose that secretion of the micronemal contents during gliding on the cell surface might be one such factor, since this motility may precede hepatocyte infection as discussed below. Our results indicate that the liver sinusoidal barrier is not perfect, since a small proportion of the *spect*-disrupted sporozoites can infect the liver (see [Figure 4](#pbio-0020004-g004){ref-type="fig"}). It is supposed that this layer may have a few openings and the disruptants can migrate through them by gliding along the epithelial cell surface. In Kupffer cell-depleted rats, on the other hand, both disruptants and wild-type may migrate through the numerous gaps created among the endothelial cells, resulting in elimination of the phenotypic difference. Since Kupffer cells constitute approximately 30% of the sinusoidal cells ([@pbio-0020004-Bouwens1]), their depletion from this layer may leave many gaps that cannot readily be repaired. Supposedly, sporozoites cross these gaps in the same way as they migrate through the few gaps in normal rats. Experiments using Kupffer cell-depleted rats indicate that Kupffer cells are not involved in sporozoites targeting the liver, because the depletion did not reduce the susceptibility of rats to sporozoite infection. Thus, sporozoites seem to be first arrested on the endothelial cell surface or on the glycosaminoglycans extending through endothelial fenestrations and then migrate to Kupffer cells ([@pbio-0020004-Cerami1]; [@pbio-0020004-Pradel2]). If so, gliding motility on the cell surface would be necessary for the sporozoite to migrate from initial attachment sites to Kupffer cells (or to gaps) along the inner surface of the sinusoidal layer as well as for the sporozoite to migrate through gaps. These assumptions imply that after Kupffer cell depletion, sporozoites can arrive at the hepatocyte by gliding motility alone, in accord with the observation that the disruptants can infect Kupffer cell-depleted rats with the same infectivity as the wild-type. Our results strongly suggest that Kupffer cells are main gates for sporozoites to access hepatocytes. Previous electron microscopic studies have reported that sporozoites are observed in Kupffer cells after intravenous inoculation, and some of them are found within vacuoles ([@pbio-0020004-Meis1]; [@pbio-0020004-Pradel1]). Based on this observation, it has been speculated that sporozoites invade the Kupffer cell by a motility distinct from passage that does not involve parasitophorous vacuole formation. Our results, on the contrary, indicate that sporozoites cross the layer by the same cell-passage motility as observed in vitro. We think this discrepancy indicates the following two possibilities. One is that the vacuole formed in the Kupffer cell after rupture of its cell membrane is different from the parasitophorous vacuole formed in the hepatocyte, although their differences cannot be distinguished by electron microscopy. Another possibility is that the parasites seen in vacuoles were phagocytosed ones and not in the process of invasion. In fact, if their invasion mode is cell-traversal motility, as we believe, this event may be rapidly completed and difficult to catch by electron microscopy. Therefore, many phagocytosed parasites could be included among those seen. Taking the evidence together, we propose that the sporozoites access the hepatocyte through Kupffer cells by the same cell-traversal motility that has been identified in vitro, and we propose a model for sporozoite liver infection in [Figure 7](#pbio-0020004-g007){ref-type="fig"}. ::: {#pbio-0020004-g007 .fig} Figure 7 ::: {.caption} ###### Schematic Representation of Sporozoite Migration to and Infection of Hepatocytes (Left) Sporozoites migrate to the space of Disse through the Kupffer cells. \[1\] The sporozoite (Sp) in the circulatory system is sequestered to the sinusoidal endothelial cell (EC) by specific recognition of the cell surface or glycosaminoglycans extending from the hepatocytes (He) through fenestration. \[2\] The sporozoite begins to glide on the epithelial cell surface. \[3\] Encountering a Kupffer cell (KC), the sporozoite ruptures the plasma membrane, passes through the cell, and enters into the space of Disse. Thus, the sporozoite gains access to hepatocytes. This step requires SPECT. \[4\] The sporozoite infects a hepatocyte with formation of a vacuole and develops into EEF in the hepatocyte. (Right) An alternative route to the hepatocyte. A small number of sporozoites, which find gaps in the sinusoidal layer while gliding, migrate to hepatocytes directly through the openings without need for cell passage and infect the hepatocytes. Likewise, in Kupffer cell-depleted rats, both wild-type and *spect*(−) sporozoites can enter hepatocytes through numerous gaps present between the sinusoidal endothelial cells. :::  ::: In this study we have established the significance of cell-passage ability of the sporozoite in malaria transmission and have demonstrated that this ability is necessary for breaking through the liver sinusoidal barrier. Cell-traversal activity plays an important role in other invasive stages of the malarial parasite, including the ookinete, which migrates through the epithelial cells of the mosquito midgut, and the sporozoite in the oocyst, which is released from the mature oocyst and then migrates through the salivary gland cell. Our study revealed that another cellular barrier is present in the malarial life cycle and sporozoites must break through this barrier by cell-traversal activity. Our recent work has identified two other genes that are involved in the cell passage activity of the sporozoite. Like SPECT, the products of these genes have a secretory protein-like structure and are localized in the micronemes. Furthermore, sporozoites disrupted for these genes have similar phenotypic character to *spect*-disrupted ones, including impaired cell-passage ability, decreased liver infectivity with similar reduction rate, complete restoration of the infectivity in Kupffer cell-depleted rats, normal gliding motility, and normal hepatocyte infectivity (unpublished data). This suggests that the cell-traversal ability of the sporozoite is realized by cooperation of several microneme proteins. We suggest that these molecules could be targets for antimalarial strategies, since success in crossing this layer is critical for the malarial parasite to establish infection in humans. Elucidation of the molecular mechanisms of passage may lead to novel malaria transmission-blocking strategies that prevent sporozoites from gaining access to the hepatocyte. Materials and Methods {#s4} ===================== {#s4a} ### Parasite preparations {#s4a1} Female 6--10-wk-old BALB/c mice (Japan SLC, Inc., Hamamatsu, Japan) infected with the P. berghei ANKA strain were prepared by peritoneal injection of infected blood that was stored at −70°C. For the purification of sporozoites, infected mosquitoes were dissected 24--28 d after the infective blood meal. The salivary glands and midgut were separately collected in medium 199 on ice and then gently ground to release the sporozoites. Ookinetes and erythrocytic-stage parasites were prepared as described previously ([@pbio-0020004-Yuda1]; [@pbio-0020004-Kariu1]). ### Genomic Southern blot hybridization {#s4a2} Genomic DNA of P. berghei parasites (2 μg) was digested with ClaI, EcoRI, EcoT22I, HindIII, or XbaI, separated on 1.2% agarose gel and then transferred to a nylon membrane. DNA fragments were amplified by PCR using genomic DNA as template with the following primers: 5′-TGGGCAATTTTGCCTTTAAAAACG-3′ and 5′-TTTCATTGTGTTAAACGATAAGTG-3′. They were labeled with \[^32^P\]dCTP and used as probes. ### Antibody preparation and Western blot analysis {#s4a3} Recombinant SPECT without signal sequence was expressed as a glutathione S-transferase (GST)--fusion protein using the pGEX 6p-1 system (Amersham Bioscience, Uppsala, Sweden). The recombinant protein was purified with a GST column and used for immunization of rabbits. Specific antibodies were affinity purified using a N-hydroxysuccinimide-activated column (Amersham Bioscience) coupled with recombinant SPECT protein. For CS antiserum production, the peptide DPPPPNANDPAPPNAN, corresponding to the repeat region, was conjugated to keyhole limpet hemocyanin as a carrier and used for the immunization of rabbits. Western blot analysis was performed as described previously ([@pbio-0020004-Kariu1]). ### Immunofluorescence microscopy and immunoelectron microscopy {#s4a4} Immunofluorescence microscopy was performed as described previously ([@pbio-0020004-Kariu1]). Purified parasites were fixed in acetone for 2 min. The slides were incubated with anti-SPECT rabbit antibodies and then with FITC-conjugated secondary antibody (Zymed. South San Francisco, California, United States). For nuclear staining, 4′,6-diamidino-2-phenylindole (DAPI) (0.02 μg/ml final concentration) was added to the secondary antibody solution. Immunoelectron microscopy was performed as described previously ([@pbio-0020004-Yuda2]). In brief, purified parasites were fixed in 1% paraformaldehyde--0.1% glutaraldehyde for 15 min on ice. After embedding in LR Gold resin (London Resin Company Ltd., London, United Kingdom), ultrathin sections were incubated with anti-SPECT antibodies and then with secondary antibody conjugated to gold particles (15 nm diameter) (AuroProbe, Amersham Pharmacia Biotech, Uppsala, Sweden). The samples were examined with a Hitachi H-800 transmission electron microscope (Hitachi, Tokyo, Japan) at an acceleration voltage of 100 kV. ### Targeted disruption of the *spect* gene {#s4a5} For construction of the targeting vector, two fragments of the *spect* gene were amplified by PCR using genomic DNA as template with the primer pairs 5′-CGCGAGCTCGCAATATGGTATTAAATTTTGGGCTAGCCA-3′ and 5′-CGCGGATCCGGTATTTTCATTGTGTTAAACGATATGTGA-3′ and 5′-CCGCTCGAGGTCCTATTTATCATTTTAAAATGTGTTTTATC-3′ and 5′-CGGGGTACCAATCGTCATAAATAGGAGTTATGAAGT-3′. These fragments were cloned into either side of the selectable marker gene in pBluescript (Strategene, La Jolla, California, United States). The gene targeting experiment was performed as described previously ([@pbio-0020004-Yuda1]). ### Evaluation of sporozoite infectivity to rats {#s4a6} Sporozoites collected from mosquito salivary glands were suspended in medium 199 and then injected intravenously into 3-wk-old female Wistar rats (Japan SLC, Inc., Hamamatsu, Japan) (*n* = 5). Before each inoculation, sporozoites were checked for their ability to glide in vitro to confirm that they contained over 60% motile sporozoites. Parasitemia was checked every 12 h by a Giemsa-stained blood smear. ### Measurement of the number of EEFs in the infected liver {#s4a7} Sporozoites (5.0 × 10^5^) were intravenously inoculated into a 3-wk-old female Wistar rat. After 24 h, the liver was perfused with PBS followed by 4% paraformaldehyde. The liver was further fixed in 4% paraformaldehyde for 6 h and frozen in liquid nitrogen. Cryostat sections (20 μm) were prepared from the left lobe and fixed in acetone for 2 min on a glass slide. The EEFs were detected by immunofluorescence staining using rabbit anti-CS antiserum and FITC-conjugated secondary antibody. At least 12 sections were examined under an Olympus (Tokyo, Japan) BX60 fluorescence microscope (200×) and the number of EEFs per square millimeter was calculated. ### EEF development assay in vitro {#s4a8} The EEF formation assay was performed as described previously ([@pbio-0020004-Hollingdale1]) with minor modifications. HepG2 cells (5.0 × 10^5^) were plated in 8-well chamber slides. Sporozoites (5.0 × 10^3^ or 5.0 × 10^4^) were suspended in 100 μl of complete medium and added to this culture. After 2 h, the media were replaced with 400 μl of fresh complete medium supplemented with 3 μg/ml glucose. The slides were incubated for 2 d with medium changed twice a day and were fixed in acetone for 2 min. The EEFs were detected by immunofluorescence staining as described above. The number of EEFs in one-fifth of the area of each well was counted under an Olympus BX60 fluorescence microscope (200×). ### Cell-traversing activity assay {#s4a9} The traversing activity of the sporozoite was examined using a standard cell-wounding and membrane repair assay ([@pbio-0020004-Mota2]). HepG2 cells (2.5 × 10^5^) or HeLa cells (5.0 × 10^4^) were inoculated into 8-well chamber slides (Nunc Inc., Napierville, Illinois, United States). Sporozoites were added 2 d later to cells for 1 h in the presence of 1 mg/ml FITC-labeled dextran (10,000 MW, lysine-fixable; Molecular Probes, Inc., Eugene, Oregon, United States). The cells were incubated for an additional 3 h in complete culture medium and fixed with 4% paraformaldehyde in PBS. The number of FITC-positive cells was counted under a fluorescence microscope. ### Depletion of rat Kupffer cells {#s4a10} For depletion of Kupffer cells, 3-wk-old female Wistar rats were injected intravenously with 120 μl of liposome-encapsulated Cl~2~MDP or an equal volume of PBS as control. After 48 h, sporozoites were injected into a tail vein and the parasitemia was checked by Giemsa-stained blood smears. Cl~2~MDP liposomes were prepared as described elsewhere ([@pbio-0020004-van1]). Elimination of Kupffer cells was confirmed by immunoperoxidase staining after liver perfusion with PBS followed by fixation with 4% paraformaldehyde in PBS. Cl~2~MDP was a gift from Roche Diagnostics (Mannheim, Germany). This study was supported by a grant-in-aid for Scientific Research on Priority Areas to YC (15019042) and to MY (15019043) and for Scientific Research (A) to YC (14207011) of the Ministry of Education, Science, Culture, and Sports of Japan. It was also supported by a grant from the Research for the Future Program of the Japan Society for the Promotion of Science to YC and by a grant from the Core Research for Evolutional Science and Technology (CREST) of the Japan Science and Technology Agency to YC. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** TI, YC, and MY conceived and designed the experiments. TI, KY, and MY performed the experiments. TI and MY analyzed the data. TI and MY wrote the paper. Academic Editor: Gary Ward, University of Vermont Cl~2~MDP : dichloromethylene diphosphonate CS : circum sporozoite protein DAPI : 4′,6-diamidino-2-phenylindole EEF : exoerythrocytic form EST : expressed sequence tag FITC : fluorescein isothiocyanate GST : glutathione S-transferase SPECT : sporozoite microneme protein essential for cell traversal

PubMed Central

2024-06-05T03:55:46.013236

2004-1-20

PMC314465

Introduction {#s1} ============ Olfactory receptor (OR) genes provide the basis for the sense of smell ([@pbio-0020005-Buck1]) and, with more than 1,000 genes, comprise the largest gene superfamily in mammalian genomes ([@pbio-0020005-Glusman1]; [@pbio-0020005-Zozulya1]; [@pbio-0020005-Young1]; [@pbio-0020005-Zhang2]; [@pbio-0020005-Olender1]). OR genes are organized in clusters ([@pbio-0020005-Trask1]; [@pbio-0020005-Young1]) and in humans are found on every chromosome save the Y and 20 ([@pbio-0020005-Glusman1]; [@pbio-0020005-Zozulya1]). On the basis of sequence similarity, they are classified into two major classes and 17 families ([@pbio-0020005-Glusman1]). All OR genes have an approximately 1 kb coding region that is uninterrupted by introns ([@pbio-0020005-Ben-Arie1]; [@pbio-0020005-Gilad1]). Interestingly, approximately 60% of human OR genes carry one or more coding region disruptions and are therefore considered pseudogenes ([@pbio-0020005-Rouquier1]; [@pbio-0020005-Glusman1]; [@pbio-0020005-Zozulya1]). In nonhuman apes, the fraction of OR pseudogenes is only approximately 30% ([@pbio-0020005-Gilad2]). However, both humans and other apes have a significantly higher fraction of OR pseudogenes than do the mouse or the dog (approximately 20%) ([@pbio-0020005-Young2]; [@pbio-0020005-Zhang2]; [@pbio-0020005-Olender1]). Thus, there has been a decrease in the size of the intact OR repertoire in apes relative to other mammals, with a further deterioration in humans ([@pbio-0020005-Rouquier2]; [@pbio-0020005-Gilad2]). Although the causes are unclear, it seems reasonable to speculate that the high fraction of OR pseudogenes in apes reflects a decreased reliance on the sense of smell in species for whom auditory and visual cues may be more important (e.g., [@pbio-0020005-Dominy1]). We were therefore interested in investigating whether the high fraction of OR pseudogenes is characteristic of primates as a whole and, if not, to pinpoint when the proportion of OR pseudogenes increased. To this end, we randomly selected subsets of 100 OR genes in 19 primate species, including a human, four apes, six Old World monkeys (OWMs), seven New World monkeys (NWMs) and one prosimian. We find that a decrease in the size of the intact olfactory repertoire occurred independently in two evolutionary lineages: in the ancestor of OWMs and apes, and in the New World howler monkey. Results and Discussion {#s2} ====================== Owing to the high levels of DNA sequence divergence among the primate species in our sample, orthologous OR genes could not be amplified by primers designed based on human sequences ([@pbio-0020005-Gilad2]). Instead, we used two sets of degenerate primer pairs, constructed to amplify OR genes from all of the species studied (see [Materials and Methods](#s4){ref-type="sec"}). We then cloned the PCR products and determined the sequences of clones until we had identified 100 distinct OR genes from each species. A danger of this approach is that degenerate primers may bind preferentially to certain OR genes, thereby resulting in a biased representation of the OR repertoire. To safeguard against this, we tested the degenerate primers on human and mouse, for which the entire OR gene repertoire is known, by using them to amplify 100 OR genes from the two species. The sample thus obtained faithfully represented the composition of the full OR gene repertoire in human and mouse with respect to the 17 OR gene families ([Figure 1](#pbio-0020005-g001){ref-type="fig"}). Moreover, the sample estimates of the fractions of pseudogenes were accurate (see [Materials and Methods](#s4){ref-type="sec"}; [Figure 2](#pbio-0020005-g002){ref-type="fig"}). This pilot study demonstrates that the degenerate primers yield an unbiased representation of the OR gene repertoire, as measured by the family composition and pseudogene content of the human and mouse samples. Since the primers performed well both in human and a distantly related species, the mouse, there was no reason to assume that they would not do so in nonhuman primate species. ::: {#pbio-0020005-g001 .fig} Figure 1 ::: {.caption} ###### Results of the Pilot Study in Human and Mouse The percentage of OR genes from each family is given for the entire repertoire (filled bars) and a sample of 100 genes amplified using PC1 and PC2 degenerate primers (open bars). (A) OR genes in human. (B) OR genes in mouse. None of the differences between the full repertoires and the samples are significant at the 5% level. Only full-length OR genes (larger than 850 bp) were considered. :::  ::: ::: {#pbio-0020005-g002 .fig} Figure 2 ::: {.caption} ###### The Proportion of OR Pseudogenes in 20 Species Primate species are color-coded according to family. The arrow points to the howler monkey. Datapoints (from left to right) are for apes (green): human (Homo sapiens), chimpanzee (Pan troglodytes), gorilla (Gorilla gorilla), orangutan (Pongo pygmaeus), gibbon (Hylobates syndactylus); for OWMs (blue): Guinea baboon (Papio papio), rhesus macaque (Macaca mulatta), silver langur (Trachypithecus auratus), mona (Cercopithecus mona), agile mangabey (Cercocebus agilis), black-and-white colobus (Colobus guereza); for NWMs (red): brown capuchin monkey (Cebus apella), southern owl monkey (Aotus azarai), spider monkey (Ateles fusciceps), black howler monkey (Alouatta caraya), squirrel monkey (Saimiri sciureus), wooly monkey (Lagothrix lagotricha), common marmoset (Callithrix jacchus); for one prosimian primate (brown): crowed lemur (Eulemur mongoz); and for the mouse (Mus musculus) (grey). :::  ::: We therefore proceeded to sequence 100 genes from 18 nonhuman primates using these primer pairs. Since the genome sequence is not available for these species, we were not able to compare the familial composition of our samples of OR genes to that of the full OR repertoires. However, with the exception of OR families 3, 11, 12, and 55 (whose absence in a sample of 100 genes is not unlikely, as they represent less than 1.8% of human OR genes), we identified OR genes from all families in all species ([Table 1](#pbio-0020005-t001){ref-type="table"}). Moreover, the representation of the three largest OR gene families in the sample varied across species, again suggesting that there is no strong bias towards the amplification of specific families. ::: {#pbio-0020005-t001 .table-wrap} Table 1 ::: {.caption} ###### Distribution of OR Genes in Families across Species :::  ::: We then tabulated the proportion of OR pseudogenes in each species ([Figure 2](#pbio-0020005-g002){ref-type="fig"}). Consistent with previous results based on direct sequencing of full-length OR orthologs ([@pbio-0020005-Gilad2]), we found that the proportion of OR pseudogene in the great apes and rhesus macaque is approximately 30% ([Figure 2](#pbio-0020005-g002){ref-type="fig"}). Together, these findings confirm the validity of this degenerate primer approach. We further found that the proportion of OR pseudogenes in OWMs (29.3% ± 2.4%) is very similar to that of nonhuman apes (33.0% ± 0.8%), but notably higher than that of NWMs (18.4% ± 5.6%). One NWM species, the howler monkey, was a conspicuous exception, with an elevated proportion of OR pseudogenes, similar to that of OWMs and apes (31.0%) ([Figure 2](#pbio-0020005-g002){ref-type="fig"}) and significantly higher than any other NWM (one-tailed *p* \< 0.02 for the difference between the howler monkey and the NWM with the second highest proportion of pseudogenes, the Wooly monkey, as assessed by a Fisher\'s exact test \[FET\]). Thus, it appears that a deterioration of the olfactory repertoire occurred in all apes and OWMs as well as, independently, in the howler monkey lineage. Strikingly, a second phenotype is shared only by the howler monkey, OWMs, and apes: full (or "routine") trichromatic color vision. In primates, trichromatic color vision is accomplished by three opsin genes whose products are pigments sensitive to short, medium, or long wavelength ranges of visible light ([@pbio-0020005-Nathans1]). In OWMs and apes, the short-wavelength opsin gene is found on an autosome, while two distinct X-linked loci for medium and long wavelengths underlie full trichromatic color vision (and so are present in both males and females). In contrast, most NWM species carry an autosomal gene and only one X-linked gene, where different alleles encode for photopigment opsins that respond to medium or long wavelengths. Heterozygous females can therefore possess trichromatic vision, but males are dichromatic ([@pbio-0020005-Jacobs1]; [@pbio-0020005-Boissinot1]; [@pbio-0020005-Hunt1]). The sole exception among NWMs is the howler monkey ([@pbio-0020005-Jacobs3]; [@pbio-0020005-Jacobs2]; [@pbio-0020005-Surridge1]), which has a duplication of the opsin genes on the X chromosome ([@pbio-0020005-Goodman1]; [@pbio-0020005-Jacobs2]) ([Figure 3](#pbio-0020005-g003){ref-type="fig"}). Thus, full trichromatic vision arose twice in primates, once in the common ancestor of OWMs and apes and once in the howler monkey lineage. ::: {#pbio-0020005-g003 .fig} Figure 3 ::: {.caption} ###### Phylogenetic Tree of Primates Schematic phylogenetic tree of the primate species used in the current study. Phylogenetic relationships between species are based on [@pbio-0020005-Harada1]), [@pbio-0020005-Page1]), and [@pbio-0020005-Surridge1]). Arrows indicate on which lineages the acquisition of full trichromatic color vision occurred ([@pbio-0020005-Goodman1]; [@pbio-0020005-Jacobs2]). The red color highlights lineages with a high proportion of OR pseudogenes. :::  ::: While OWMs, apes, and the howler monkey carry 32.5% ± 6.3% OR pseudogenes in their OR gene repertoire, species without full trichromatic vision have 16.7% ± 1.0%, significantly fewer (*p* \< 10^−4^, or, excluding humans from the full trichromatic group, *p* \< 10^−3^, as assessed by a Mann--Whitney *U* test). This *p* value is only indicative since the species lineages are not all independent. However, if significance is instead assessed by a FET for all pairwise comparisons of species with full trichromatic color vision and without, the difference is again striking: 94 out of 96 comparisons are significant at the 5% level. Thus, the evolution of full trichromatic vision coincided with an increase in the fraction of OR pseudogenes, indicative of a deterioration of the sense of smell. Apes and OWMs acquired trichromatic color vision approximately 23 million years ago ([@pbio-0020005-Yokoyama1]), while the duplication of the opsin genes in the howler monkey occurred approximately 7--16 million years ago ([@pbio-0020005-Jacobs1]; [@pbio-0020005-Cortes-Ortiz1]). In spite of this difference in timing, the proportion of OR pseudogenes in species from both lineages is very similar. We estimated the rate of fixation of neutral gene disruptions for OR genes to be approximately 0.12 per gene per million years (Y. Gilad, S. Pääbo, and G. Glusman, unpublished data). This estimate implies that both apes, OWMs and the howler monkey could have a much higher proportion of OR pseudogenes than observed (data not shown), indicating that the process of functional OR gene loss has decreased or stopped in these species. A plausible explanation for the similar proportion of OR pseudogenes in the different lineages is that while full trichromatic vision relaxed the need for a sensitive sense of smell, it did not render olfaction unnecessary. Accordingly, while some OR genes can accumulate coding region disruptions, others are still evolving under evolutionary constraint. This model predicts that the possession of full trichromatic color vision alone allows for the loss of some but not all OR genes. A natural next step would then be to identify which OR genes or families were lost after the acquisition of full trichromatic vision. The answer to this question awaits sequence from a large number of orthologous OR genes. In this respect, it is interesting to note that the *TRP2* gene, a major component of the vomeronasal pheromone transduction pathway, was found to be intact in several NWM species, but is a pseudogene in OWMs and apes ([@pbio-0020005-Liman1]; [@pbio-0020005-Zhang1]). The authors raised the possibility of a connection between the acquisition of full trichromatic color vision and decreased pheromone perception, based on the difference between OWMs and apes on the one hand and NWMs on the other ([@pbio-0020005-Liman1]; [@pbio-0020005-Zhang1]). However, since many traits can potentially be mapped to the lineage that leads to OWMs and apes, the connection between full trichromatic vision and pheromone perception was tenuous. Furthermore, [@pbio-0020005-Liman1]) did not find a coding region disruption in four exons of *TRP2* in the howler monkey. An intact *TRP2* gene in the howler monkey would be inconsistent with the hypothesis that the enhancement of color vision replaced pheromone signaling in primates. In contrast, in the present study, we find that the deterioration of the olfactory repertoire occurred concomitant with the evolution of full trichromatic vision in two separate primate lineages. Thus, although at this point we are unable to demonstrate that the decline in the sense of smell is a direct result of the evolution of color vision, our results strongly suggest an exchange in the importance of these two senses in primate evolution. Future studies of the sensory cues involved in detection and selection of food (e.g., [@pbio-0020005-Smith1]), or the choice of a mate, may test this association directly. Materials and Methods {#s3} ===================== {#s3a} ### Design and test of degenerate primers. {#s3a1} OR genes have a coding region that is approximately 1 kb long and contains no introns. In order to test the performance of degenerate primers, we sequenced 30 genes amplified with each primer pair in human and mouse and compared the composition of the different OR families in the sample to that of the full OR gene repertoire of these two species ([@pbio-0020005-Glusman1]; [@pbio-0020005-Zhang2]). We also compared the sample estimates of the proportion of pseudogenes to the proportion in the entire OR repertoire of human and mouse. Since the degenerate primers amplify only 670 bp of the approximately 1 kb coding region of the OR gene, a subset of the coding region disruptions will fall in segments of OR genes not amplified by our primers. As a result, the true fraction of OR genes carrying coding region disruptions will be underestimated by our approach. We therefore determined the proportion of OR genes with at least one disruption within the corresponding 670 bp in the entire human and mouse OR gene repertoires (47.7% and 16.3% in humans and mouse, respectively). We first tested an existing set of primers, used by [@pbio-0020005-Rouquier2]), but found significant deviations from the family composition of the full OR repertoire in both species. As an illustration, among the 60 OR genes obtained in humans, 36.6% were of the subfamily 7E (all pseudogenes), significantly more than expected given the true proportion of the 7E subfamily in the full human OR gene repertoire (12.4%, *p* = 2 × 10^−6^, assessed by FET). As a consequence of these biases, estimates of the proportion of pseudogenes in human and mouse obtained with these primers ([@pbio-0020005-Rouquier2]) differ significantly from the true value (*p* \< 0.01, assessed by FET). We proceeded by designing new pairs of degenerate primers for the OR gene family by using the program HYDEN ([@pbio-0020005-Fuchs1]; [@pbio-0020005-Linhart1]). The first primer pair, PC1 (PC1--5′: CTSCAYSARCCCATGTWYHWYTTBCT, PC1--3′: GTYYTSAYDCHRTARAYRAYRGGGTT), was designed based on class 1 human OR sequences only. The second primer pair, PC2 (PC2--5′: YTNCAYWCHCCHATGTAYTTYTTBCT, PC2--3′: TTYCTNARGSTRTAGATNANDGGRTT), was designed based on solely class 2 human OR sequences, excluding all genes that belong to subfamily 7E. Both primer pairs were designed to amplify a 670-bp product that approximately covers the region from transmembrane domains 2--7 of the OR protein. As a first step, we used each primer pair to amplify and sequence (see below) 30 genes from human genomic DNA. We found that PC1 primer pairs amplify OR class 1 and OR class 2 genes in roughly equal proportions. PC2 primer pairs amplified only OR class 2 genes, including members of the 7E OR subfamily. Based on the OR family composition that we observed for the 60 genes, we estimated that if we constructed a sample containing 25% of genes amplified with PC1 and 75% of genes amplified with PC2, we would obtain an unbiased representation of the familial composition of the human OR gene repertoire. This approach was validated by amplifying and examining 100 genes collected in the same way from human as well as from mouse. ### PCR and DNA sequencing {#s3a2} Each primer pair was used to amplify a set of eight reactions in each species using a temperature-gradient PCR. The use of several annealing temperatures for each species yielded a greater diversity of amplified OR genes. PCR was performed in a total volume of 25 μl, containing 0.2 μM of each deoxynucleotide (Promega, Madison, Wisconsin, United States), 50 pmol of each primer, 1.5 mM MgCl~2~, 50 mM KCl, 10 mM Tris (pH 8.3), 2 U of Taq DNA polymerase, and 50 ng of genomic DNA. Conditions for the PCR amplification from all species were as follows: 35 cycles of denaturation at 94°C, annealing at a gradient temperature of 48°C to 60°C, and extension at 72°C, each step for 1 min. The first step of denaturation and the last step of extension were 3 min each. The PCR products were separated and visualized in a 1% agarose gel. From each amplification set (a given primer pair in a given species), all successful products were mixed and subjected to cloning using a TA cloning kit (Boehringer, Mannheim, Germany). Cloning was followed by a touchdown PCR using the vector primers for amplifications from isolated bacterial colonies. Products were purified using the High Pure PCR Product Purification Kit (Boehringer). Sequencing reactions were performed in both directions on PCR products, using the vector primers and the dye-terminator cycle sequencing kit (Perkin Elmer, Wellesley, Massachusetts, United States) on an ABI 3700 automated sequencer (Perkin Elmer). ### Sequence analysis {#s3a3} After base calling with the ABI Analysis Software (version 3.0), the data were edited and assembled using the Sequencher program, version 4.0 (GeneCodes Corporation, Ann Arbor, Michigan, United States). Assembly of the clones was done using a similarity cutoff of 98%. This cutoff ensures that Taq-generated mutations that may have been sequenced in individual clones are not counted as independent genes. Clones that were collapsed to the same contig by the assembly process were counted as one gene. Once 25 and 75 genes (independent contigs) were identified from PC1 and PC2 primer pairs, respectively, a majority consensus was generated for each gene. In order to confirm that only OR genes were amplified from all the species, we used the consensus sequences of all genes from all species as queries in a BLAT search against the human genome sequence (<http://genome.ucsc.edu/>). In every case, the best hit was a human OR gene. This analysis was also used to insure that none of the genes were an artifact of ("jumping") PCR fusion. Finally, each consensus sequence was searched for an uninterrupted open reading frame (ORF) in all six possible frames. If an uninterrupted ORF was found, the gene was annotated as intact. If no ORF was identified, the gene was annotated as a pseudogene. This approach probably results in an underestimate of the proportion of pseudogenes, as not all OR genes with an intact coding region are functional. Mutations in promoter or control regions of OR genes may lead to reduced or no expression. Similarly, radical missense mutations in highly conserved positions of the OR protein may result in dysfunction ([@pbio-0020005-Menashe1]). Although it is known that there are several highly conserved positions among OR genes, it is not always straightforward to ascertain which, if any, of these positions is necessary to retain function. We therefore chose the most straightforward definition of a pseudogene: a gene without a full ORF. Supporting Information {#s4} ====================== Accession Numbers {#s5a} ----------------- Sequences for all OR genes from all primate species were deposited to GenBank (<http://www.ncbi.nlm.nih.gov/Genbank/>) as accession numbers AY448037--AY449380 and AY454789--AY455274. We thank Christian Roos from the Primate Genetics German Primate Center in Göttingen for the primate DNA samples. We are also grateful to Chaim Linhart from Tel Aviv University for helping us to design the degenerate primer pairs. The experimental work was financed by the Bundesministerium für Bildung und Forschung (01KW9959-4) and by the Max Planck Gesellschaft. This research was completed while YG was supported by a Clore doctoral fellowship. DL holds the Ralph and Lois Silver Chair in Human Genomics and is supported by the Crown Human Genome Center at the Weizmann Institute of Science. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** YG and MP conceived and designed the experiments. YG and VW performed the experiments. YG and VW analyzed the data. SP contributed reagents/materials/analysis tools. YG, MP, DL, and SP wrote the paper. Academic Editor: David Hillis, University of Texas at Austin FET : Fisher\'s exact test OR : olfactory receptor ORF : open reading frame OWM : Old World monkey NWM : New World monkey

PubMed Central

2024-06-05T03:55:46.016348

2004-1-20

PMC314466

Introduction {#s1} ============ Primate lentiviruses persist in the host by active replication and can reemerge from latent reservoirs that are established in cells of the immune system ([@pbio-0020006-Finzi1]; [@pbio-0020006-Douek2]). The infectious cycle is intimately linked to interactions between circulating T cells and antigen-presenting cells ([@pbio-0020006-Stevenson1]; [@pbio-0020006-Bukrinsky1]; [@pbio-0020006-Embretson1]; [@pbio-0020006-Swingler1]; [@pbio-0020006-Geijtenbeek1]). These interactions involve T cell migration, adhesion, and antigen-initiated signaling, processes that are dependent on cytoskeletal dynamics regulated by the Rho subfamily of small GTPases ([@pbio-0020006-Hall1]; [@pbio-0020006-Schmitz1]). The lentiviral accessory protein Nef is a multifunctional regulator that is important for rapid progression to AIDS ([@pbio-0020006-Kestler1]; [@pbio-0020006-Piguet2]; [@pbio-0020006-Renkema1]). One key function of Nef is its ability to facilitate activation of infected cells and thus provide an environment that is conducive for viral replication ([@pbio-0020006-Skowronski1]; [@pbio-0020006-Baur1]; [@pbio-0020006-Du1]; [@pbio-0020006-Schrager1]; [@pbio-0020006-Simmons1]). Another important function of Nef is its ability to promote evasion of the antiviral immune response. This is accomplished by downregulation of class I MHC complexes from the surface of infected cells, which protects against detection by cytoxic T cells specific for viral antigens ([@pbio-0020006-Schwartz1]; [@pbio-0020006-Collins1]). The ability of Nef to facilitate T cell activation is well documented. Thymic and peripheral CD4^+^ T cells from transgenic mice are hypersensitive to stimulation via the T cell antigen receptor (TCR) ([@pbio-0020006-Skowronski1]; [@pbio-0020006-Hanna1]), as are resting primary human CD4^+^ T cells ([@pbio-0020006-Schrager1]; [@pbio-0020006-Wang1]) and cell lines transduced to express HIV-1 Nef ([@pbio-0020006-Alexander1]; [@pbio-0020006-Baur1]). Nef was reported to associate with molecules that play important roles in antigen-initiated signaling in T cells, including elements of signaling pathways involving small GTPases. Specifically, Nef was reported to associate with Vav ([@pbio-0020006-Fackler1]) and activate p21-activated serine--threonine kinases (PAKs), possibly though the activation of Rac or CDC42 ([@pbio-0020006-Lu1]). Recent observations that Nef can activate Rac in a glial cell line have strengthened the connection between Nef and these pathways ([@pbio-0020006-Vilhardt1]). The notion that effects of Nef on signaling machineries in T cells are mediated by small GTPases, their effectors, or both represents an attractive possibility, yet the exact mechanism resulting in activation of these pathways has remained elusive. Since Nef likely functions as an adaptor protein, we exploited a proteomic approach to directly identify the key molecules Nef uses to subvert the signaling machinery in T cells. Here we show that Nef targets a key activator of Rac GTPases that functions downstream of the TCR and chemokine receptors. Results {#s2} ======= Nef Binds DOCK2, ELMO1, and Rac in T Cells {#s2a} ------------------------------------------ To identify downstream effectors of Nef in T lymphocytes, we generated CD4^+^ Jurkat T cells that stably express the extensively studied patient-derived HIV-1 Nef protein NA7 ([@pbio-0020006-Mariani1]) tagged at its C-terminus with HA and FLAG epitopes (NA7-hf) ([Figure 1](#pbio-0020006-g001){ref-type="fig"}A). Nef and its associated proteins were purified by successive immunoprecipitations with anti-HA- and then anti-FLAG epitope antibodies, followed each time by elution with the respective peptide epitope and resolved by SDS-PAGE. Several polypeptides with apparent molecular weights ranging from approximately 20 kDa to 220 kDa copurified with HIV-1 Nef, but were absent in preparations from control cells that do not express Nef ([Figure 1](#pbio-0020006-g001){ref-type="fig"}B). Gel slices containing these polypeptides were digested with trypsin, and the resulting peptides were sequenced by liquid chromatography tandem mass spectroscopy (LC/MS/MS) and subjected to database searches ([@pbio-0020006-Hu1]). Two abundant Nef-associated proteins, DOCK2 and ELMO1, were thus identified. DOCK2 is a lymphocyte-specific CED5/DOCK180/Myoblast City (CDM) family protein that regulates the activity of Rac1 and Rac2 GTPases downstream of chemokine receptors and the TCR and is essential for lymphocyte migration and normal antigen-specific responses of T cells ([@pbio-0020006-Fukui1]; [@pbio-0020006-Reif1]; [@pbio-0020006-Sanui1]). Rac GTPases are members of the Rho subfamily of small GTP-binding proteins that control several processes, including cytoskeletal rearrangements during cell motility and T cell activation ([@pbio-0020006-Hall1]). Recent studies showed that ELMO1 functionally cooperates with CDM family proteins to activate Rac ([@pbio-0020006-Brugnera1]; [@pbio-0020006-Sanui2]). Significantly, our mass spectroscopic analyses of Nef-associated proteins also identified Rac2. Furthermore, in addition to Rac2-specific peptides, we also detected peptides shared by Rac1 and Rac2, raising the possibility that Rac1 also associates with HIV-1 Nef in T cells. The ubiquitously expressed Rac1 and hematopoietic cell-specific Rac2 are 95% identical, and both isoforms regulate cytoskeletal dynamics and gene expression in T lymphocytes ([@pbio-0020006-Yu1]; [@pbio-0020006-Croker1]). ::: {#pbio-0020006-g001 .fig} Figure 1 ::: {.caption} ###### DOCK2, ELMO1, and Rac Are Abundant Nef-Associated Proteins in T Cells \(A) Schematic representation of epitope-tagged HIV-1 Nef (NA7-hf). The structured regions of Nef are boxed and the disordered regions, as determined by X-ray crystallography and NMR studies, are shown by a thin line. The locations of the N-terminal myristoyl moiety, prolines P72 and P75 in the PP-II helix, arginine R106, leucines L164 and L165 (LL164), and the C-terminal HA-FLAG epitopes are indicated. \(B) DOCK2, ELMO1, and Rac2 specifically copurify with HIV-1 Nef from Jurkat T cells. Jurkat T cells (1.8 ×10^10^) stably expressing NA7-hf (lane 3) or control Jurkat cells (lane 2) were subjected to the two-step immunopurification procedure described in the text (see [Materials and Methods](#s4){ref-type="sec"}). Polypeptides present in purified immune complexes were resolved by SDS-PAGE and analyzed by LC/MS/MS. We identified 58 DOCK2-specific peptides covering 869 out of 1830 total amino acid residues (47.5% coverage, expectation value 6.0 × 10^--130^), 10 ELMO1-specific peptides covering 122 out of 727 total amino acid residues (16.8% coverage, expectation value 1.0 × 10^−10^), and three Rac-specific (two of which were Rac2-specific) peptides covering 26 out of 192 total amino acid residues (13.5% coverage, expectation value 4.6 × 10^−4^). Bands corresponding to DOCK2, ELMO1, Rac2 and their predicted molecular weights, NA7-hf Nef, and the FLAG peptide used for elution are indicated. :::  ::: Nef Binds the DOCK2--ELMO1--Rac Complex {#s2b} --------------------------------------- By analogy to previously described interactions among Rac, ELMO1, and CDM family proteins ([@pbio-0020006-Brugnera1]; [@pbio-0020006-Sanui2]), our finding that DOCK2, ELMO1, and Rac2 copurified with HIV-1 Nef suggested that DOCK2 forms a ternary complex with ELMO1 and Rac2 and that Nef binds this complex. To investigate these possibilities, we attempted to reconstitute these interactions in human embryonic kidney 293 (HEK 293) cells. Although HEK 293 cells express endogeneous ELMO1, our initial studies revealed that the association of Nef with DOCK2 and Rac2 was significantly enhanced by ectopic expression of ELMO1 (data not shown). Thus, to determine whether ELMO1 and Rac2 copurify with DOCK2, DOCK2-containing complexes were purified from HEK 293 cells transiently expressing His-tagged DOCK2, Myc-tagged ELMO1, and Myc-tagged Rac2 via DOCK2 using Ni--NTA resin and eluted with imidazole. Immunoblotting revealed that ELMO1 and Rac2 copurified with DOCK2 ([Figure 2](#pbio-0020006-g002){ref-type="fig"}A, lane 3), indicating that DOCK2 complexes with ELMO1 (DOCK2--ELMO1) and Rac2. ::: {#pbio-0020006-g002 .fig} Figure 2 ::: {.caption} ###### Lentiviral Nef Binds the DOCK2--ELMO1--Rac Complex \(A) HIV-1 Nef binds the DOCK2--ELMO1--Rac2 complex. His-DOCK2, Myc-ELMO1, and Myc-Rac2 alone (lanes 1, 3, and 5) or together with NA7-hf Nef (lanes 2, 4, and 6) were transiently expressed in HEK 293 cells as indicated. DOCK2 was precipitated from extracts (lanes 1 and 2) with Ni--NTA resin (lanes 3 and 4). Nef--DOCK2 was then precipitated with anti-FLAG affinity gel (lanes 5 and 6), and the epitope-tagged proteins were detected by immunoblotting and visualized by enhanced chemiluminescence. \(B) Rac1 associates with HIV-1 Nef. Nef and associated proteins were isolated from extracts of HEK 293 cells transiently expressing DOCK2, ELMO1, and Rac1 either alone (lanes 1 and 4), with NA7-hf (lanes 2 and 5), or with a Nef variant containing a disrupted myristoylation signal (lanes 3 and 6). Nef and associated proteins were detected in anti-FLAG immunoprecipitates (lanes 1--3) and in extracts (lanes 4--6) by immunoblotting. \(C) The interaction with DOCK2, ELMO1, and Rac2 is a conserved function of lentiviral Nef proteins. The ability of selected hf-tagged HIV-1 (lanes 1--3 and 5) and SIV mac239 (lane 4) Nef proteins to bind DOCK2, ELMO1, and Rac2 was determined as described in the legend to (B) above. The protein band in (C) indicated by the asterisk is the heavy chain of anti-FLAG mAb. :::  ::: Subsequently, we asked whether ELMO1 and Rac2 are subunits of DOCK2--Nef complexes. DOCK2--Nef-containing complexes were isolated from HEK 293 cells transiently expressing His-tagged DOCK2, Myc-tagged ELMO1, Myc-tagged Rac2, and HA-FLAG epitope-tagged HIV-1 Nef (NA7-hf) via DOCK2 using Ni--NTA resin and eluted with imidazole ([Figure 2](#pbio-0020006-g002){ref-type="fig"}A, lane 4). DOCK2--Nef complexes were then reisolated from this eluate via Nef by anti-FLAG immunoprecipitation. It is evident that ELMO1 and Rac2 also copurified with complexes containing both Nef and DOCK2 ([Figure 2](#pbio-0020006-g002){ref-type="fig"}A, lane 6), thus supporting the possibility that HIV-1 Nef binds DOCK2--ELMO1 complexes that contain Rac2. Nef Targets Rac1 and Rac 2 Isoforms {#s2c} ----------------------------------- Our mass spectroscopic analyses indicated that HIV-1 Nef associates with Rac2, but left open the possibility that it also targets Rac1. Therefore, we tested whether Nef can associate with Rac1 in the context of DOCK2 and ELMO1 using the same HEK 293 transient expression assay. Nef and its associated proteins were isolated from cell extracts by anti-FLAG immunoprecipitation and visualized by immunoblotting ([Figure 2](#pbio-0020006-g002){ref-type="fig"}B). Nef formed readily detectable complexes incorporating Rac1 ([Figure 2](#pbio-0020006-g002){ref-type="fig"}B, lane 2), while a mutant Nef protein unable to associate with membranes due to disruption of its N-terminal myristoylation signal (NA7~(G2~ ^~∇~^ ~HA)~), and therefore functionally defective, did not associate with DOCK2, ELMO1, or Rac1 ([Figure 2](#pbio-0020006-g002){ref-type="fig"}B, lane 3). These results indicate that myristoylated Nef targets the Rac1 and Rac2 isoforms. Nef proteins from well-characterized primate lentiviruses display considerable amino acid sequence variation. Therefore, we verified that Nef proteins from additional well-characterized laboratory HIV-1 strains (SF2 and NL4--3) bind DOCK2, ELMO1, and Rac2. We also tested a Nef protein from a strain of pathogenic SIV, mac239, that is important for rapid progression to AIDS in experimentally infected rhesus macaques ([@pbio-0020006-Kestler1]). Nef and its associated proteins were isolated from HEK 293 cell extracts by anti-FLAG immunoprecipitation and visualized by immunoblotting ([Figure 2](#pbio-0020006-g002){ref-type="fig"}C). Functional Nef proteins from all lentiviral strains tested associated with DOCK2, ELMO1, and Rac2. This indicates that the ability to associate with Rac and its upstream regulators is a conserved function of primate lentiviral Nef. Nef Activates Rac in Resting T Cells {#s2d} ------------------------------------ Since DOCK2, ELMO1, and Rac are major Nef-associated proteins in Jurkat T cells and since DOCK2 mediates Rac activation, we determined the effect of Nef on Rac activity in these cells. The active GTP-bound form of Rac (Rac~GTP~) binds the p21-binding domains (PBD) of PAKs directly ([@pbio-0020006-Burbelo1]). Hence, we used a PBD--GST fusion protein in pulldown assays to measure the fraction of activated Rac in vivo. Jurkat T cells were transduced with a lentiviral vector directing the expression of HIV-1 Nef (FUGWCNA7) or a control empty vector (FUGW). Extracts prepared from these cells were incubated with PBD--GST and the fraction of PBD-bound Rac was determined by immunoblotting ([Figure 3](#pbio-0020006-g003){ref-type="fig"}A). Notably, the expression of Nef resulted in a readily detectable increase in the steady-state level of PBD-bound Rac ([Figure 3](#pbio-0020006-g003){ref-type="fig"}A, lane 3), consistent with the possibility that the interaction of Nef with DOCK2--ELMO1 increases Rac activation. ::: {#pbio-0020006-g003 .fig} Figure 3 ::: {.caption} ###### Nef Activates Rac in Resting CD4^+^ T Lymphocytes \(A) HIV-1 Nef activates Rac in Jurkat T cells. Jurkat T cells (lane 1) were transduced with a control empty vector (FUGW; lane 2) or the same vector expressing HIV-1 NA7 Nef (FUGWCNA7; lane 3). Rac~GTP~ was precipitated from cell extracts with recombinant PAK1 PBD--GST. PBD--GST bound Rac~GTP~ (top), total Rac present in extracts (middle), and Nef (bottom) were detected by immunoblotting. \(B) Flow cytometric analysis of Gag and CD4 expression in resting CD4^+^ T lymphocytes transduced with HIV-1 derived vectors in the presence of IL-7. Percentages of cells productively infected with *nef*-deleted H-Δ vector (boxed area in middle panel) or with HIV-1 NA7 *nef* containing H-NA7 vector (right panel) are shown. Results obtained with uninfected control CD4^+^ T cells cultured in the presence of IL-7 are also shown (left panel). \(C) HIV-1 Nef specifically activates Rac in resting primary CD4^+^ T lymphocytes. Rac~GTP~ and CDC42~GTP~ were precipitated with PAK1 PBD--GST from extracts prepared from CD4^+^ T lymphocytes transduced with HIV-1 derived vectors, shown in (B), and analyzed as described in (A). :::  ::: In nontransformed T lymphocytes, Rac activation through DOCK2 is tied to chemotactic and antigenic stimuli. To assess whether Nef can uncouple these processes, we determined the effect of Nef on Rac activation in primary CD4^+^ T lymphocytes in the absence of stimulation with antigen and chemokines. While resting T cells are normally refractory to productive infection by lentiviruses and lentivirus-derived vectors, a sizable fraction becomes permissive for infection when cultured in the presence of cytokines such as IL-7 ([@pbio-0020006-Unutmaz1]). We used this procedure to infect primary resting CD4^+^ T lymphocytes with an HIV-1-derived vector expressing HIV-1 NA7 Nef (H-NA7) or a control *nef*-deleted vector (H-Δ). Since HIV-1 Env protein may activate DOCK2-controlled signaling pathways through binding to chemokine receptors such as CXCR4, *env*-defective, VSV-G-pseudotyped viruses were used in these experiments. We cultured 98% pure populations of CD4^+^ T cells isolated from the peripheral blood leukocytes of healthy donors for 5 d in the presence of IL-7 and then transduced them with Nef-expressing H-NA7 or control H-Δ virus. The purity of the infected populations and the efficiency of transduction were assessed 4 d later by flow cytometric analysis of CD4 expression on the cell surface and intracellular p24 Gag expression, respectively. As shown in [Figure 3](#pbio-0020006-g003){ref-type="fig"}B, between 13% and 15% of CD4^+^ T cells were productively infected. Notably, the unusually low level of CD4 on the surface of cells infected with H-NA7 virus was due to robust downregulation of cell surface CD4 by NA7 Nef ([@pbio-0020006-Mariani1]). Cell extracts were prepared from the infected populations, and PBD--GST pulldown assays were performed to determine the fraction of activated Rac. Strikingly, infection with H-NA7 resulted in a readily detectable increase in the steady state level of activated Rac ([Figure 3](#pbio-0020006-g003){ref-type="fig"}C). Based on direct quantitations of chemiluminescent signals of total and PBD--GST bound Rac, we estimated that approximately 1.2% of the total Rac in extracts from cells transduced with H-NA7 was bound to PBD--GST as compared to 0.2% in extracts from cells transduced with H-Δ. The activation of Rac was specifically due to the expression of Nef and not other viral gene products, as infection with the otherwise isogenic H-Δ virus did not increase PBD--GST-reactive Rac. To address the specificity of Nef effect towards Rac, we then asked whether Nef affects activity of CDC42 GTPase, which also uses PAK as a downstream effector ([@pbio-0020006-Burbelo1]). Direct quantitations of chemiluminescent signals for total and PBD--GST-bound CDC42 revealed that less than 0.2% of the total CDC42 in extracts from H-NA7 and H-Δ transduced cells was PBD--GST bound. Therefore, we concluded that Nef primarily activates Rac and not CDC42 in CD4^+^ T lymphocytes in the absence of antigenic stimuli. Nef Activates Rac through DOCK2--ELMO1 {#s2e} -------------------------------------- Next we asked whether Nef activates Rac through DOCK2--ELMO1. To determine whether ELMO1 is required for the effect of Nef, we measured Rac activation by Nef in NS1 lymphoma cells, which do not express the endogenous ELMO1 ([@pbio-0020006-Sanui2]) and in NS1 cells in which ELMO1 expression was restored by retrovirus-mediated transfer of *ELMO1* cDNA (NS1~ELMO1~). NS1 and NS1~ELMO1~ cells were infected with a lentiviral vector expressing HIV-1 NA7 Nef (FUGWCNA7) or with a control empty vector (FUGW). Cell extracts were prepared from the infected populations, and PBD--GST pulldown assays were performed to determine the fraction of activated Rac. In agreement with a previous report (Sanui at al. 2003b), NS1 cells contain a small but readily detectable pool of activated Rac in spite of the lack of detectable ELMO1 expression, which is most likely generated by ELMO1-independent mechanism(s) ([Figure 4](#pbio-0020006-g004){ref-type="fig"}A, lane 1). Notably, expression of Nef in the absence of ELMO1 and expression of ELMO1 in the absence of Nef did not increase the fraction of activated Rac (compare lane 2 to lane 1 and lane 3 to lane 1, respectively, in [Figure 4](#pbio-0020006-g004){ref-type="fig"}A). In contrast, expression of Nef in the presence of ELMO1 induced a readily detectably increase in the pool of activated Rac in the NS1~ELMO1~ cells ([Figure 4](#pbio-0020006-g004){ref-type="fig"}A, lane 4). These observations indicate that Nef activates Rac through an ELMO1-dependent mechanism. ::: {#pbio-0020006-g004 .fig} Figure 4 ::: {.caption} ###### ELMO1 and DOCK2 Mediate Rac Activation by HIV-1 Nef \(A) ELMO1 is required for Rac activation by Nef in NS1 cells. Rac~GTP~ and total Rac in the extracts prepared from ELMO1-deficient NS1 cells (lanes 1 and 2) and ELMO1-expressing NS1 cells (lanes 3 and 4) following transduction with a lentiviral vector expressing HIV-1 Nef (lanes 2 and 4) or a control empty vector (lanes 1 and 3) were visualized as described in the legend to [Figure 3](#pbio-0020006-g003){ref-type="fig"}. \(B) Nef activates Rac through DOCK2 and ELMO1 in HEK 293 cells. Rac~GTP~ and total Rac in the extracts prepared from HEK 293 cells coexpressing the indicated proteins were visualized as described above. :::  ::: Next, we studied Rac activation by Nef in HEK 293 cells, which do not express endogeneous DOCK2. Combinations of Nef, DOCK2, ELMO1, and either Rac1 or Rac2 were expressed in HEK 293 cells by transient transfection. The expression of Nef in the absence of DOCK2 had little effect on the activation of either Rac isoform ([Figure 4](#pbio-0020006-g004){ref-type="fig"}B, lanes 2 and 7). Ectopic expression of DOCK2 and ELMO1 increased the fraction of activated Rac1 and Rac2 by approximately 3- to 4-fold ([Figure 4](#pbio-0020006-g004){ref-type="fig"}B, lanes 3 and 8), which is in agreement with a previous report ([@pbio-0020006-Sanui2]). Significantly, coexpression of myristoylated, but not unmyristoylated, Nef ([Figure 4](#pbio-0020006-g004){ref-type="fig"}B, lanes 4 and 9 versus lanes 5 and 10) with DOCK2 and ELMO1 further enhanced the fraction of activated Rac isoforms by approximately 2-fold, and this effect of Nef was more pronounced for Rac2 (compare lane 9 with lane 4 in [Figure 4](#pbio-0020006-g004){ref-type="fig"}B). Together, these data indicate that myristoylated Nef stimulates Rac activation through the DOCK2--ELMO1 complex. Nef Activates Rac through Association with DOCK2 and ELMO1 {#s2f} ---------------------------------------------------------- We identified mutations in Nef that disrupt Rac activation. Nef was reported to associate with an active form of PAK, and this interaction was suggested to be important for Nef effects on the cytoskeleton and T cell activation ([@pbio-0020006-Fackler1]; [@pbio-0020006-Arora1]; [@pbio-0020006-Wang1]). Since PAK is an immediate downstream effector of Rac, we asked whether the abilities of Nef to activate Rac through DOCK2--ELMO1 and to associate with activated PAK are correlated. Two different Nef mutations that were previously reported to disrupt its association with activated PAK (P72A,P75A and R106A) ([@pbio-0020006-Sawai1]; [@pbio-0020006-Renkema1]), yet unlike the myristoylation signal mutation (G2~∇~HA) did not significantly affect Nef functions in other assays, were tested for their effects on Rac activation in Jurkat T cells ([Figure 5](#pbio-0020006-g005){ref-type="fig"}A) and in HEK 293 cells transiently expressing DOCK2, ELMO1, and Rac2 ([Figure 5](#pbio-0020006-g005){ref-type="fig"}B). Both mutations abolished the ability of Nef to stimulate Rac activation ([Figures 5](#pbio-0020006-g005){ref-type="fig"}A and [5](#pbio-0020006-g005){ref-type="fig"}B, lanes 4 and 5). As expected, the same was true for mutation of the myristoylation signal in Nef ([Figures 5](#pbio-0020006-g005){ref-type="fig"}A and [5](#pbio-0020006-g005){ref-type="fig"}B, lane 3). In contrast, a mutation that specifically abrogates the interaction of Nef with clathrin adaptor proteins (LL164AA; [@pbio-0020006-Greenberg2]) had little disruptive effect on Rac activation ([Figures 5](#pbio-0020006-g005){ref-type="fig"}A and [5](#pbio-0020006-g005){ref-type="fig"}B, lane 6). ::: {#pbio-0020006-g005 .fig} Figure 5 ::: {.caption} ###### Nef Potentiates Rac Activation through Association with DOCK2--ELMO1 (A and B) Myristoylation signal, P72,P75, and R106 in Nef are required for Rac activation. Rac~GTP~ and total Rac in the extracts prepared from Jurkat T cells transduced with lentiviral vectors expressing no Nef (−) or the indicated Nef proteins (A) and HEK 293 cells transiently coexpressing the indicated Nef mutants together with DOCK2, ELMO1, and Rac2 (B) were visualized as described in the legend to [Figure 3](#pbio-0020006-g003){ref-type="fig"} and quantified by direct imaging of chemiluminescent signals. The fraction of total Rac present in the extracts that was PBD--GST bound is shown in the histograms. Data in the histogram shown in (B) are averages of three independent experiments and error bars represent two standard deviations. \(C) Myristoylation signal, P72,P75, and R106 in Nef are required for association with DOCK2, ELMO1, and Rac2. The ability of selected Nef mutants to associate with DOCK2, ELMO1, and Rac2 was determined as described in [Figure 2](#pbio-0020006-g002){ref-type="fig"}. :::  ::: Next we asked whether mutations in Nef that disrupted Rac2 activation affected the association with DOCK2, ELMO1, and Rac2 ([Figure 5](#pbio-0020006-g005){ref-type="fig"}C). The LL164AA mutation, which did not significantly compromise the stimulation of Rac activation, did not affect the association of Nef with these proteins ([Figure 5](#pbio-0020006-g005){ref-type="fig"}C, lane 6). In contrast, mutations that reduced Rac2 activation by Nef also diminished its association with DOCK2, ELMO1, and Rac2 ([Figure 5](#pbio-0020006-g005){ref-type="fig"}C, lanes 3--5). Notably, the P72A,P75A and R106A mutations completely disrupted detectable association with ELMO1 and Rac2, but only weakened that with DOCK2, suggesting that Nef associates with both DOCK2 alone and DOCK2 complexed with ELMO1 and/or Rac2 and that the P72A,P75A and R106A mutations preferentially disrupt binding to the latter complex. Thus, robust stimulation of Rac2 activation by Nef requires its association with both DOCK2 and ELMO1. Functional Consequences of Nef Interactions with DOCK2--ELMO1 and Rac {#s2g} --------------------------------------------------------------------- The CD4^+^ T lymphocyte is a major target of infection by primate lentiviruses. Nef was reported to lower the threshold signal required for antigen-induced responses of T cells ([@pbio-0020006-Schrager1]; [@pbio-0020006-Wang1]), and this effect was proposed to be an important component to stimulation of viral replication by Nef in vivo ([@pbio-0020006-Alexander1]; [@pbio-0020006-Simmons1]). Since DOCK2 mediates Rac activation downstream of the TCR to modulate T cell responsiveness and downstream of chemokine receptors to mediate chemotactic responses ([@pbio-0020006-Fukui1]; [@pbio-0020006-Sanui1]), we studied effects of Nef on these processes in T lymphocytes. Purified populations of primary resting CD4^+^ T cells were transduced with VSV-G-pseudotyped H-NA7 or *nef*-deleted H-Δ in the presence of IL-7. Cells were stimulated 4--6 d following transduction with plate-bound anti-CD3 and anti-CD28 antibodies, mixed at various ratios, for various amounts of time. Intracellular IL-2 and p24 Gag were visualized and quantified by flow cytometry to provide a measure of cellular activation of the infected cells in response to the stimulation. In the absence of anti-CD3/anti-CD28 stimulation, neither uninfected (Gag-negative) nor the productively infected (Gag-positive) cells, produced detectable amounts of IL-2 ([Figure 6](#pbio-0020006-g006){ref-type="fig"}). In contrast, stimulation through CD3 and CD28 induced readily detectable accumulation of IL-2 in both H-Δ- and H-NA7-transduced populations. Notably, a larger fraction of productively infected CD4^+^ T lymphocytes typically proceeded to express IL-2 than uninfected cells. This phenomenon suggests that the permissive state for HIV-1 infection induced by IL-7 is associated with an increased responsiveness to activation via CD3 and CD28. However, no significant difference in the levels of IL-2 expression in cells infected with H-Δ compared to those infected with H-NA7 was detected across a wide range of stimulation conditions. These observations contrast with results from previous studies using T cell lymphoma and nontransformed CD4^+^ T lymphocytes expressing Nef alone ([@pbio-0020006-Schrager1]; [@pbio-0020006-Wang1]). This difference could, for example, reflect the modifying effect of other HIV-1 gene products that were not tested in the previous experiments. On the other hand, we cannot exclude the possibility that IL-7 treatment masks the effect of Nef. Since Nef deregulates DOCK2 function and DOCK2 is essential for proper signaling through the immunological synapse, further studies under a variety of conditions that modulate the formation and function of the immunological synapse may be required to reveal this effect of Nef in the context of HIV-1 infection. ::: {#pbio-0020006-g006 .fig} Figure 6 ::: {.caption} ###### Effect of Nef on IL-2 Expression in HIV-1-Infected CD4^+^ T Lymphocytes Stimulated through CD3 and CD28 CD4^+^ T lymphocytes transduced with H-Δ and H-NA7 HIV-1-derived vectors were not stimulated (unstimulated) or stimulated with immobilized anti-CD3 and anti-CD28 mAbs (anti-CD3, anti-CD28) in the presence of Golgi-Stop for 5 h and stained for intracellular IL-2 and p24 Gag. Percentages of IL-2-positive and IL-2-negative cells in the Gag-negative and Gag-positive populations are shown. :::  ::: DOCK2 regulates the activation of Rac proteins during lymphocyte migration in response to chemokine gradients ([@pbio-0020006-Fukui1]). Therefore, we also asked whether Nef affects lymphocyte chemotaxis. Jurkat T cells, which constitutively express CXCR4, a major coreceptor for T-cell tropic HIV and a receptor for stromal-derived factor 1 (SDF-1) ([@pbio-0020006-Deng1]; [@pbio-0020006-Feng1]), were transiently transfected with a control plasmid expressing enhanced green fluorescent protein (GFP) alone or with a plasmid that coexpresses Nef and a GFP marker protein from the same bicistronic transcription unit. We then measured the chemotaxis of transfected populations to SDF-1 using a transwell migration assay. The relative frequency of control and Nef-expressing cells in the migrated populations was determined by flow cytometric measurement of GFP expression. Approximately 30% of control cells migrated regardless of the level of GFP expression, indicating that the chemotaxis in this assay was robust ([Figure 7](#pbio-0020006-g007){ref-type="fig"}A). In contrast, the chemotaxis of cells coexpressing Nef and GFP was inhibited in a dose-dependent manner. ::: {#pbio-0020006-g007 .fig} Figure 7 ::: {.caption} ###### Nef Disrupts T Cell Migration to SDF-1 \(A) Migration of cell populations shown in (B) expressing GFP (open circle), ectopic CXCR4, and GFP (filled circle), HIV-1 NA7 Nef and GFP (open box), HIV-1 NA7 Nef, ectopic CXCR4 and GFP (filled box) to SDF-1 was measured in transwell assays. \(B) Transient expression of ectopic CXCR4 restores CXCR4 levels on the surface of Nef-expressing cells. Flow cytometric analysis of Jurkat T cells transiently expressing GFP (panel 1) or HIV-1 NA7 Nef and GFP (panel 2) and together with ectopic CXCR4 (panels 3 and 4, respectively). :::  ::: Control experiments revealed that Nef caused a modest decrease in cell surface expression of CXCR4 (compare panels 1 and 2 in [Figure 7](#pbio-0020006-g007){ref-type="fig"}B). This observation raised the possibility that Nef-expressing cells were unresponsive to SDF-1 due to abnormally low levels of CXCR4 at the cell surface rather than due to deregulation of the DOCK2--ELMO1 complex. To address this possibility, we restored CXCR4 on the surface of Nef-expressing cells to levels equal to and even higher than those seen in control cells by transiently expressing ectopic CXCR4 receptor and then performed migration assays using these cell populations ([Figure 7](#pbio-0020006-g007){ref-type="fig"}B, panels 3 and 4). Significantly, the migration of cells with restored CXCR4 levels was still inhibited by Nef expression ([Figure 7](#pbio-0020006-g007){ref-type="fig"}A). We concluded that HIV-1 Nef blocks lymphocyte migration to SDF-1 principally by interfering with CXCR4-controlled signaling cascades rather than by downregulating CXCR4 from the cell surface. We then asked whether inhibition of Jurkat T cell migration by Nef correlated with its ability to potentiate Rac activation by DOCK2 and ELMO1. As expected, we found that disruption of the myristoylation signal in Nef (NA7~(G2~ ~^∇^~ ~HA)~) abolished the inhibition of migration ([Figure 8](#pbio-0020006-g008){ref-type="fig"}A and [8](#pbio-0020006-g008){ref-type="fig"}B). Furthermore, the P72A,P75A, and R106A mutations diminished the ability of Nef to block migration, albeit to different extents. In contrast, the LL164AA mutation, which had little disruptive effect on enhancement of Rac activation, was fully functional in this assay. These observations suggested that deregulated activation of Rac GTPases is instrumental for the defective migration of Nef-expressing cells. To explore this possibility further, we ectopically expressed constitutively active Rac1 and Rac2 (Rac1~G12V~, Rac2~G12V~) and, as controls, wild-type Rac1 and Rac2 in Jurkat T cells and measured their migration to SDF-1 ([Figure 8](#pbio-0020006-g008){ref-type="fig"}C). Expression of wild-type Rac GTPases stimulated chemotaxis approximately 2- to 3-fold. In contrast, expression of the constitutively active forms of each Rac GTPase severely suppressed lymphocyte migration to SDF-1. Thus, deregulated Rac activation inhibits directional cell movement, most likely by disrupting spatially organized rearrangements of the cytoskeleton that are induced by chemokine gradients. These results further support a model in which Nef disrupts migration to SDF-1 by activating Rac through the DOCK2--ELMO1 module and thus uncoupling Rac activation from chemokine receptor signaling. ::: {#pbio-0020006-g008 .fig} Figure 8 ::: {.caption} ###### Nef Disrupts Chemotaxis by Activating Rac through DOCK2--ELMO1 (A and B) Jurkat T cells expressing wild-type or mutant HIV-1 Nef proteins and GFP reporter were used in transwell chemotaxis assays with SDF-1. Percentage of migrated cells expressing GFP alone (open circle), or together with HIV-1 NA7 (open square), NA7~(G2~ ~^∇^~ ~HA)~ (open diamond), NA7~(P72A,P75A)~ (open triangle), NA7~(R106A)~(filled circle), and NA7~(LL164AA)~ (filled triangle) is shown as a function of GFP fluorescence intensity in (A) and in (B) for the single GFP fluorescence intensity interval indicated by the shaded rectangle in (A). \(C) Constitutively active Rac GTPases disrupt lymphocyte migration to SDF-1. Migration of Jurkat T cells transiently expressing wild-type (Rac1, Rac2), constitutively active (Rac1~G12V~, Rac2~G12V~), or as a control HIV-1 Nef were also measured. Data shown are averages of three independent experiments and error bars represent two standard deviations. :::  ::: The above observations predicted that Nef likely causes a general migration defect. Therefore, we also studied the effect of Nef on Jurkat T cell migration to the MIP-1β chemokine. Since the MIP-1β receptor (CCR5) is not constitutively expressed in Jurkat T cells, we transiently expressed CCR5 and GFP marker from a bicistronic vector either alone or together with HIV-1 NA7 Nef. Using this vector, CCR5 expression levels were positively correlated with GFP marker protein expression ([Figure 9](#pbio-0020006-g009){ref-type="fig"}A, panels 2 and 3). Notably, flow cytometric analysis revealed that CCR5 cell-surface expression was not downregulated by Nef (compare panels 5 and 6 in [Figure 9](#pbio-0020006-g009){ref-type="fig"}A). We then measured the ability of these cells to migrate to MIP-1β. Migration of cells coexpressing CCR5 and HIV-1 Nef was impaired compared to control cells expressing CCR5 at comparable levels ([Figure 9](#pbio-0020006-g009){ref-type="fig"}B). These data confirm that Nef induces a general defect in lymphocyte migration by targeting DOCK2--ELMO1 and Rac. ::: {#pbio-0020006-g009 .fig} Figure 9 ::: {.caption} ###### HIV-1 Nef Disrupts CCR5-Mediated Migration \(A) HIV-1 Nef does not downregulate CCR5. Flow cytometric analysis of CCR5 and GFP in Jurkat T cells transiently expressing GFP alone (panel 1) or CCR5 and GFP in the absence (panel 2) and presence (panel 3) of HIV-1 Nef. Histograms of CCR5 expression for cell populations within a single GFP fluorescence intensity interval indicated by the rectangle in panel 1 are shown in panels 4 to 6, respectively. \(B) Percentage of cells migrated to MIP-1β and expressing GFP alone (open circle), GFP and CCR5 (open triangle), or GFP, CCR5 and HIV-1 Nef (filled triangle) is shown as a function of GFP fluorescence intensity. :::  ::: Discussion {#s3} ========== Nef is a multifunctional adaptor protein that modulates signal transduction and protein-sorting machineries. We purified to near homogeneity an abundant Nef-associated protein complex from T cells and identified by mass spectroscopy its major subunits as DOCK2 and ELMO1, a bipartite Rac activator ([@pbio-0020006-Sanui2]). Notably, the extensive large-scale biochemical purification and sensitive proteomic analyses described in this report did not detect several cellular proteins previously reported to associate with HIV-1 Nef and mediate the effects of Nef in the cell (data not shown). This includes cellular proteins such as Vav ([@pbio-0020006-Fackler1]), PAKs ([@pbio-0020006-Sawai1]), inositol triphosphate receptor ([@pbio-0020006-Manninen1]), Lck protein-tyrosine kinase ([@pbio-0020006-Baur2]), clathrin adaptors ([@pbio-0020006-Greenberg1]; [@pbio-0020006-Le1]; [@pbio-0020006-Piguet1]), and others. Why were proteins previously reported to associate with Nef not detected in our studies? One possible explanation is that previously reported associations are unstable under the biochemical purification conditions used in our studies. This likely explains the apparent absence of clathrin adaptors in Nef preparations purified in our studies, since we know that Nef binds AP-2/AP-1 clathrin adaptors only weakly in the salt and pH conditions used here (data not shown). It is also possible that in some cases, epitope tags may be buried and therefore inaccessible to the monoclonal antibodies (mAbs) used for immunoaffinity purification. Moreover, some of the previously reported associations, especially those with protein kinases, such as PAKs, were best detected by an ultrasensitive in vitro kinase assay ([@pbio-0020006-Sawai1]) and, as our data show, are likely of exceedingly low stoichiometry. (Of note, we did detect the presence of p62 phosphoprotein \[PAK\] by in vitro kinase assays of anti-Nef purifications \[data not shown\]). Finally, some of the described associations, such as that with thioesterase, are known to occur only with selected Nef variants ([@pbio-0020006-Cohen1]), while those we report here occur with all Nef variants tested. Nonetheless, the specific isolation of the Nef--DOCK2--ELMO1--Rac complex reported here provides strong biochemical evidence to reinforce predictions from previous genetic studies that Nef functions through multiple independent interactions with different sets of downstream effector proteins. Our observations strongly argue that DOCK2--ELMO1 is the major upstream regulator used by Nef to activate Rac in T cells and that through this interaction Nef can activate Rac in CD4^+^ T lymphocytes even in the absence of stimulation with antigen or chemokines. These data indicate that Nef targets a critical switch, DOCK2--ELMO1, that regulates Rac GTPases downstream of chemokine receptors and the TCR and uses it to modulate the downstream processes they control. Thus, the interaction of Nef with DOCK2 and ELMO1 provides an important mechanism by which Nef may impact pathogenesis by primate lentiviruses. Our data strongly suggest that DOCK2--ELMO1 is the major activator of Rac targeted by Nef in T lymphocytes. This model is supported by the observations that Nef physically associates with a complex that contains DOCK2--ELMO1 and Rac, that specific mutations in Nef simultaneously disrupt its ability to bind this complex and to activate Rac, and that Nef fails to activate Rac in the absence of ELMO1. Although previous reports implicated Vav, a Rac1 guanine nucleotide exchange factor (GEF), as the critical downstream effector that Nef binds directly to activate PAK ([@pbio-0020006-Fackler1]), our data do not support this possibility. Notably, we have been unable to detect the presence of Vav in anti-Nef immune complexes by both proteomic analyses and immunoblotting, indicating that these interactions are of low abundance relative to those with DOCK2 and ELMO1 and are therefore unlikely to mediate the bulk of Nef\'s impact on the Rac pathway. Significantly, ELMO1 is ubiquitously expressed ([@pbio-0020006-Gumienny1]) and can associate with Nef in nonlymphoid cells in the absence of DOCK2 (data not shown). Thus, we postulate a general mechanism in which ELMO1, possibly in complex with another CDM family protein, mediates Rac activation and PAK recruitment by Nef that is observed in nonlymphoid cells ([@pbio-0020006-Sawai1]; [@pbio-0020006-Fackler1]; [@pbio-0020006-Arora1]). Our observation that the expression of Nef from the integrated HIV-1 provirus in primary CD4^+^ T cells did not alter IL-2 production in standard activation protocols was unexpected, because previous genetic evidence linked Nef binding to DOCK2--ELMO1 and the ensuing activation of Rac GTPases to the the ability of Nef to facilitate T cell activation. Specifically, the same mutations that we observed to disrupt Rac activation through the DOCK2--ELMO1 complex were previously reported to disrupt the stimulatory effect of Nef on aspects of T cell activation ([@pbio-0020006-Sawai1]; [@pbio-0020006-Simmons1]), and the observed effects were in some cases dramatic ([@pbio-0020006-Wang1]). Moreover, Rac activation by DOCK2 facilitates T cell responsiveness to antigen, as disrupted Rac activation in *DOCK2*(−/−) and *Rac2*(−/−) mice is associated with defective immunological synapse formation and depressed antigen-specific responses ([@pbio-0020006-Yu1]; [@pbio-0020006-Sanui1]). Since Nef deregulates DOCK2 function and DOCK2 is essential for proper signaling through the immunological synapse, further studies under a variety of conditions that modulate the formation and function of the immunological synapse may be required to reveal the effect of Nef in the context of HIV-1 infection. Interestingly, the chemokine receptor system plays crucial roles in infection by primate lentiviruses. Previous studies revealed its essential role for virus entry into target cells ([@pbio-0020006-Deng1]; [@pbio-0020006-Feng1]). Primate lentiviruses also exploit this system to recruit uninfected target cells to sites of viral replication ([@pbio-0020006-Weissman1]; [@pbio-0020006-Swingler1]). Our observations reveal an additional level of complexity in primate lentivirus--chemokine receptor system interactions. Inhibiting chemotaxis of the infected T cells likely disrupts the generation of the immune response. During generation of the immune response, T cells are initially activated in the paracortex of lymph nodes and then migrate to the edges of follicles, where they interact with antigen receptor-activated B cells ([@pbio-0020006-Garside1]). This physical interaction is required to drive B cells towards antibody production, isotype switching, and the affinity maturation of the antibody response. Hence, the development and maturation of the immune response require the ordered migration of activated T cells to specific sites within lymphoid tissue ([@pbio-0020006-Delon1]). Notably, recent in vivo evidence documents Nef-dependent alterations in the distribution of SIV mac239-infected CD4^+^ T lymphocytes in the lymph nodes of experimentally infected rhesus macaques ([@pbio-0020006-Sugimoto1]). In lymph nodes of monkeys infected with *nef-*deleted SIV, most infected T cells were located in the B cell-rich follicles and in the border region between the paracortex and the follicles. In contrast, in monkeys infected with SIV harboring a functional *nef* gene, most productively infected T cells remained in the T cell-rich paracortex and were only infrequently present in proximity to B cell follicles. This evidence shows that Nef disrupts the ordered migration patterns of infected CD4^+^ T lymphocytes in vivo and reinforces the possibility that this disruption impairs the generation and maturation of the immune response to antigens. Since a large fraction of HIV-1-infected CD4^+^ T cells is specific for HIV-1 antigens ([@pbio-0020006-Douek1]), this effect of Nef provides another mechanism to suppress the antiviral immune response. Taken together, the functional interaction of Nef with DOCK2, ELMO1, and Rac enables Nef to modulate multiple aspects of T cell function. Materials and Methods {#s4} ===================== {#s4a} ### Construction of expression plasmids {#s4a1} Sequences encoding variant and mutant Nef proteins tagged at their C-termini with a peptide (hf) containing the HA and FLAG epitopes (DTYRYIYANATYPYDVPDYAGDYKDDDDK) were subcloned into pBABE-*puro* and pCG expression plasmids ([@pbio-0020006-Morgenstern1]; [@pbio-0020006-Greenberg2]). In NA7~(G2~ ~∇~ ~HA)~ Nef, the myristoylation signal was disrupted by insertion of the HA epitope (ANATYPYDVPDYAG) at glycine G2. The full-length human *ELMO1* cDNA (clone IMAGE:4521393; ResGen, Carlsbad, California, United States), *DOCK2* cDNA (KIAA0209, clone ha04649; Kazusa DNA Research Institute, Chiba, Japan), and cDNAs encoding wild-type and mutant forms of Rac1 and Rac2 (kindly provided by Linda Van Aelst, Cold Spring Harbor Laboratory, Cold Spring Harbor, New York, United States) were subcloned into pCG plasmids containing N-terminal c-Myc (EQKLISEEDL) or polyhistidine (HHHHHHH) epitope tags, using standard techniques. pBABE *ELMO1* contains an N-terminal c-Myc epitope-tagged *ELMO1* cDNA subcloned into the pBABE-*neo* vector. H-NA7 was constructed by substituting the HIV-1 NL4--3 *nef* coding sequence with that of HIV-1 NA7 *nef* ([@pbio-0020006-Mariani1]) in pNL4--3 carrying a frameshift mutation at the KpnI site at position 6463 in *env*, kindly provided by Klaus Strebel (National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland, United States). H-Δ is based on H-NA7 with a deletion that removes residues 1--34 of Nef and prevents expression of Nef protein. For chemotaxis assays, cDNAs encoding HIV-1 NA7 Nef, human CCR5 (kindly provided by Frank Kirchhoff, Universitätsklinikum, Ulm, Germany), and Rac proteins were subcloned into the pCGCG bicistronic vector that directs the expression of GFP from an internal ribosomal entry site element ([@pbio-0020006-Lock1]). ### Generation of stable cell lines {#s4a2} pBABE plasmids were introduced into the Phoenix amphotropic packaging cell line, kindly provided by G. P. Nolan (Stanford University Medical Center, Palo Alto, California, United States), by calcium phosphate coprecipitation, and viral supernatants were used to infect a Jurkat T cell subline ([@pbio-0020006-Greenberg2]), provided by Dan R. Littman (New York University School of Medicine, New York, New York, United States), or NS1 cells. Transduced Jurkat cells were selected and subsequently maintained in the presence of puromycin (0.4 μg/ml) (Sigma, St. Louis, Missouri, United States). Transduced NS1 cells were selected and maintained in the presence of G418 (1.0 mg/ml) (Invitrogen, Carlsbad, California, United States). ### Immunoaffinity purification of epitope-tagged Nef and associated proteins {#s4a3} Unless stated otherwise, all reactions were performed at 4°C. Approximately 1.8 × 10^10^ Jurkat T cells stably expressing NA7-hf (or control cells) were lysed for 1 h in 200 ml of LB buffer (150 mM NaCl, 50 mM Tris--HCl \[pH 7.5\], 1% Triton X-100, 10% glycerol) (Complete Protease Inhibitors, Roche, Basel, Switzerland). Extracts were precleared with protein G--agarose (Roche) for 1 h and incubated with 12CA5 mAb crosslinked to protein G--agarose beads for 4 h. The immunoprecipitate was washed extensively with LB, and bound proteins were eluted by competition with HA peptide (0.2 mg/ml) (ANATYPYDVPDYAG; Invitrogen) in LB for 45 min at 30°C. The eluate was incubated with anti-FLAG M2 affinity gel beads (Sigma) overnight, and the immunoprecipitate was washed extensively with LB and then LB modified to contain 0.1% Triton X-100 (FEB buffer). Proteins were eluted with FEB containing FLAG peptide (0.2 mg/ml) (Sigma) for 45 min at 30°C. The eluate was concentrated on Microcon centrifugal filter devices (Millipore, Billerica, Massachusetts, United States) with a cutoff of 8 kDa. ### Protein identification by mass spectrometry {#s4a4} Nef and associated proteins were separated by 8%--17% SDS-PAGE and visualized using SYPRO stain (Molecular Probes, Eugene, Oregon, United States). Visible bands were excised and processed for identification by mass spectrometry. The samples were analyzed by LC-MS/MS as described previously ([@pbio-0020006-Hu1]). Spectra resulting from LC/MS/MS were analyzed with the SONARS software package (ProteoMetrics LLC, New York, New York, United States). ### Lentiviral vectors and infections {#s4a5} FUGWCNA7 contains the wild-type HIV-1 NA7 *nef* coding sequence under control of the CMV promoter, subcloned downstream of the Woodchuck-responsive element in the FUGW lentiviral vector ([@pbio-0020006-Lois1]). FUGWC vectors containing amino acid substitutions in Nef have a similar structure. Supernatants containing infectious particles were produced by calcium phosphate cotransfection of HEK 293 cells, as described previously ([@pbio-0020006-Lois1]). For biochemical analyses of Rac activation, approximately 10^7^ Jurkat T cells or NS1 cells were infected with supernatants containing approximately 10^7^ infectious units of FUGW or FUGWC vectors encoding wild-type or mutant HIV-1 NA7 Nef proteins, in the absence of polycationic agents. Cell extracts were prepared 3--4 d following infection and used for PBD--GST pulldown assays. PBMCs were purified from buffy coats of healthy donors (New York Blood Bank, Hicksville, New York, United States) by density gradient separation on Lymphocyte Separation Medium (ICN Biomedicals, Inc., Irvine, California, United States), and CD4^+^ T lymphocytes were isolated using CD4^+^ T Cell Enrichment Columns (R&D Systems, Minneapolis, Minnesota, United States). Replication incompetent HIV-1 particles pseudotyped with VSV-G were produced by calcium phosphate transfection of HEK 293 cells and used to infect \>98% pure populations of CD4^+^ T lymphocytes that were cultured in the presence of IL-7 ([@pbio-0020006-Unutmaz1]). Cell-surface CD4 in the infected populations was revealed with FITC-conjugated SK3 mAb (Becton Dickinson, San Jose, California, United States). Cells stained for CD4 were permeabilized using Cytofix/Cytoperm Kit (BD PharMingen, San Jose, California, United States) and p24 Gag expression was revealed with PE-conjugated KC57-RD1 mAb (Beckman Coulter, Inc., San Diego, California, United States) as described elsewhere ([@pbio-0020006-Mascola1]). CD4 and Gag expression were quantitated simultaneously using an LSR-II flow cytometer (Becton Dickinson). ### Cell stimulations and IL-2/Gag assays {#s4a6} Anti-CD3 mAb and anti-CD28 mAb (MAB100 and MAB342; R&D Systems), alone or in combinations, were immobilized on 12-well microtiter plates (351143; Becton Dickinson) in PBS overnight at 4°C. Wells were washed three times with PBS, and 5 × 10^5^ CD4^+^ T lymphocytes were added 4--5 d after they were transduced with H-Δ or H-NA7 vectors in the presence of IL-7. In some experiments, transduced cells were cultured for additional 48 h in the absence of IL-7 before stimulations. Stimulations were performed for 4 h to 16 h in the presence and absence of Golgi-Stop (Becton Dickinson). Cells were recovered from wells by vigorous pipetting, fixed, and permeabilized using Cytofix/Cytoperm Kit (BD Bioscience PharMingen). Intracellular IL-2 and p24 Gag were revealed simultaneously with PE-conjugated rat anti-human IL-2 mAb, (559334; B&D Biosciences PharMingen) and FITC-conjugated KC57-RD1 mAb (Beckman Coulter, Inc.), respectively. ### Transient transfections of HEK 293 cells, immunoprecipitations, and immunoblotting {#s4a7} HEK 293 cells were transfected by calcium phosphate coprecipitation with 20 μg of pCG plasmids expressing epitope-tagged DOCK2, ELMO1, Rac2 or Rac1, and/or tagged Nef proteins and a control empty vector. Cells were lysed 48 h posttransfection in LB buffer. To isolate Nef and associated proteins, extracts were incubated overnight with anti-FLAG M2 Affinity Gel (Sigma), the immunoprecipitates were washed four times with LB buffer, once with LB containing 0.5 M LiCl, and proteins were eluted with FLAG peptide as described above. To isolate DOCK2 and associated proteins, extracts were incubated with Ni--NTA agarose (Qiagen, Valencia, California, United States) and proteins were eluted from the precipitate with 250 mM imidazole. DOCK2--Nef complexes were isolated from imidazole eluates with anti-FLAG M2 affinity gel as described above. Eluted proteins proteins were separated by 16% SDS-PAGE, electroblotted onto PVDF membrane (Millipore), and immunoblotted with the following antibodies: anti-c-Myc mAb (1:100; Oncogene, Tarrytown, New York, United States), anti-FLAG M2 mAb (1:5000; Sigma), 12CA5 mAb (1:5000), or rabbit serum raised to a DOCK2-specific peptide (GDKKTLTRKKVNQFFKTM). Immune complexes were revealed with HRP-conjugated antibodies specific for the Fc fragment of mouse or rabbit immunoglobulin G. (1:5,000; Jackson ImmunoResearch Laboratories, Inc., West Grove, Pennsylvania, United States) and ECL (Amersham, Little Chalfort, United Kingdom). ### Rac and CDC42 activity assays {#s4a8} Cells were lysed in LB modified to contain 500 mM NaCl, 0.5% sodium deoxycholate, 0.1% SDS, and 10 mM MgCl~2~ (RB buffer). Extracts were incubated for 2 h with 40 μg of recombinant PAK1 PBD--GST (kindly provided by Linda Van Aelst) bound to glutathione--agarose beads (8 mg/ml) (Roche), and beads were washed extensively with RB buffer. Bound proteins and aliquots of extracts were separated by 16% SDS-PAGE. Rac was immunoblotted with anti-c-Myc epitope mAb (1:100; Oncogene), or with anti-Rac mAb (1:1000; BD Transduction Laboratories), CDC42 was detected with sc-87 rabbit antibody (1:100; Santa Cruz Biotechnology, Santa Cruz, California, United States), and Nef was detected with rabbit serum raised against HIV-1 HxB3 Nef (1:300; [@pbio-0020006-Mariani1]). Immune complexes were visualized by chemiluminescence using Lumi-Lite Plus (Roche). Chemiluminescent signals were imaged and quantitated using the FluorChem Imaging System and software (Alpha Innotech, Cannock, United Kingdom). ### Chemotaxis assays {#s4a9} Jurkat T cells were transfected by electroporation with plasmids coexpressing Nef and GFP marker protein from a single bicistronic transcription unit (pCGCG NA7; [@pbio-0020006-Lock1]), except that in some experiments were also cotransfected with pCG CXCR4, expressing human CXCR4 receptor. Transfected populations were used 24--48 h later in transwell chemotaxis assays in the presence of 10 ng/ml of SDF-1 (R&D Research). For CCR5-dependent migration, Jurkat T cells were transfected with pCGCG CCR5, expressing human CCR5 and GFP, alone or in the presence pCG NA7, and migration to MIP-1β (10 ng/ml) was measured. Cells were allowed to migrate for 2 h, and the relative frequency of GFP-positive cells in the initial and migrated populations was determined by flow cytometry using an LSR-II flow cytometer (Becton Dickinson). Supporting Information {#s5} ====================== Accession Numbers {#s5a} ----------------- The LocusLink (<http://www.ncbi.nlm.nih.gov/LocusLink/>) accession numbers of the genetic loci discussed in this paper are *DOCK2* (LocusLink ID 1794), *ELMO1* (LocusLink ID 9844), *Rac1* (LocusLink ID 5879), and *Rac2* (LocusLink ID 5880). We thank Linda Van Aelst, Nouria Hernandez, Winship Herr, and Dan Littman for helpful discussions, sharing reagents, and critical reading of the manuscript. We also thank Frank Kirchhoff and Klaus Strebel for sharing reagents; Dan Littman for making his P3 facility available to us; Dan Bogenhagen, Robert Haltiwanger, and Bruce Stillman for support; and Takahiro Nagase from the Kazusa DNA Research Institute for providing the KIAA0209 cDNA clone. This work was supported by Public Health Service grant AI-42561 (to JS). **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** AJ, TS, BH, and JS conceived and designed the experiments. AJ, TS, MPM, and JS performed the experiments. AJ, TS, BH, MPM, and JS analyzed the data. AJ, TS, BH, and JS contributed reagents/materials/analysis tools. AJ and JS wrote the paper. Academic Editor: Michael Emerman, Fred Hutchinson Cancer Research Center ¤1 Current address: The Rockefeller University, New York, New York, United States of America ¤2 Current address: Pritzker School of Medicine, University of Chicago, Chicago, Illinois, United States of America CCR5 : chemokine (CC motif) receptor 5 CDM : CED-5/DOCK180/Myoblast City CXCR4 : chemokine (CXC motif) receptor 4 DOCK : dedicator of cytokinesis ELMO : engulfment and cell motility GEF : guanine nucleotide exchange factor GFP : green fluorescent protein GST : glutathione S-transferase HEK : human embryonic kidney LC/MS/MS : liquid chromatography tandem mass spectroscopy mAb : monoclonal antibody PAK : p21-activated kinase PBD : p21-binding domain SDF-1 : stromal-derived factor-1 TCR : T cell antigen receptor

PubMed Central

2024-06-05T03:55:46.018662

2004-1-20

PMC314467

JSTOR is successful for reasons its founders did not intend. Bill Bowen\'s inspired vision was of a solution to libraries\' ever-voracious demands for space to house paper volumes. The idea was that libraries could save space by removing volumes available in electronic format. Few libraries have discarded the volumes digitised in JSTOR, but many libraries without the paper volumes have been able to offer their users access to the important journal runs JSTOR has digitised. Paper holdings have not decreased dramatically, but electronic holdings have increased. So a space-saving service became an access service. As an access service, JSTOR is a creation of its time. Understandable though the decision to use page images may have been eight years ago, future user-friendly access requires searching capabilities across full-text, which page images cannot supply. Likewise, the decision to digitise the back-runs of around 100---now 218---paper journals was a bold decision at the time, but the future for access to journal literature lies in electronic versions of thousands rather than hundreds of titles, both current and retrospective. When we reach that point, JSTOR will still have a valued place in the content on offer, but it is difficult to see JSTOR providing thousands rather than a few hundred titles. Its technical solutions and financial models look dated as both subscription-based and open-access publishers improve their services to authors and to readers. As the number of journal articles accessible over the networks increases, JSTOR will be seen as a small-scale pioneer from which we learned valuable lessons. Roger Schonfeld ends his very detailed description of JSTOR with a chapter on 'Lessons Learned', many of which are relevant to current access initiatives. The need for grant funding to launch any such initiative has to be accompanied by a sound business plan to ensure long-term economic viability. JSTOR has achieved that transition, and its success provides a model for others. Much of the credit must go to JSTOR\'s enterprising president, Kevin Guthrie, who found the quickest way through the maze of conflicting advice---much of which could have resulted in JSTOR\'s reaching a deadend---and convinced the library and publishing communities to buy into a product that was only a promise. Meeting user needs for easy access to high-quality content was the key to the fulfilment of that promise. JSTOR\'s public image is of quality in depth---long runs of core journals---and that image has to become the hallmark of the new open-access initiatives as they develop. It is understandable that some mistakes were made on the way. The difficulty that JSTOR financial planning had in coming to terms with consortial purchases delayed its growth as an access service. Although selling to consortia of academic libraries may not have improved JSTOR\'s financial position in the short-term, consortia are a route to spreading access and therefore securing longer-term financial stability (as the major publishers have realised through their 'Big Deals' in selling hundreds of journals to hundreds of libraries in a consortium). Some opportunities were also delayed---not lost---through too slow an adaptation of the JSTOR purchasing model for selling outside the United States, the United Kingdom being the exception. The UK deal was with JISC, the Joint Information Systems Committee of the UK Higher Education Funding Councils, acting more as a negotiating agent than a consortium, and this model could have been applied in other countries. More countries would have valued access to JSTOR earlier, but the approach to non-US deals had to be imaginative. For all vendors, there has to be an understanding of the political, social, economic, and educational structure of the country into which the product is being sold, an understanding that takes time to acquire but that pays dividends. Open-access publishers do not have to sell their product to users of their journals, but local knowledge is essential in 'selling' their services to authors. The globalisation of publishing has combined with the globalisation of the networks and with the globalisation of research to provide opportunities for high-quality research conducted outside North America and Western Europe to be published in peer-reviewed open-access journals more readily than in the traditional subscription-based journals. Roger Schonfeld\'s book draws out many of the significant points about JSTOR\'s place in the history of electronic publication through a minute examination of the process leading to JSTOR as it is today. There is so much detail in the book that the reader may feel that its comprehensiveness cannot be questioned, but one small omission of which I have personal knowledge makes me question the value of so much detail. The omission concerns the interest by my institution, University College London, in joining JSTOR before the JISC deal was considered. Not a detail of world-shattering significance, but it does illustrate the fact that outside the United States, as well as within, the early interest in JSTOR came from individual institutions rather than from consortia. I sympathise with Roger Schonfeld in attempting to write such a comprehensive history, but what is the point of appearing to be comprehensive when comprehensiveness is an impossible goal? Would a briefer history have been just as valuable? Leaving aside quibbles and caveats about the book and about JSTOR, this remains a fascinating and instructive history of an important and ground-breaking initiative. Bill Bowen\'s vision may not have developed in quite the way he expected, but the 'bottom-line' is that the vision did become a successful reality. The problem of ever-expanding libraries has not gone away in the ten years since JSTOR was conceived, but the ultimate solution---the availability of electronic content---has become closer, and JSTOR\'s success has encouraged others to develop services that are more in accord with 2003 than 1993. One lesson Roger Schonfeld does not draw out is the pace of change in electronic publishing, and if so much has been achieved since 1993, what promise is held out by the next ten years\'! ::: {#pbio-0020010-g001 .fig}  ::: Frederick J. Friend is the Honorary Director of Scholarly Communication at University College London, United Kingdom. E-mail: <ucylfjf@ucl.ac.uk> **Book Reviewed** Schonfeld RC (2003) JSTOR: A history. Princeton, New Jersey: Princeton University Press. 412 pp. ISBN (hardcover) 0-691-11531-1. US\$29.95.

PubMed Central

2024-06-05T03:55:46.023208

2004-1-20

PMC314468

Inside eukaryotic cells there is a massive protein complex called the proteasome whose raison d\'être is to remove unnecessary proteins by breaking them down into short peptides. The proteasome is thus responsible for an important aspect of cellular regulation because the timely and controlled proteolysis of key cellular factors regulates numerous biological processes such as cell cycle, differentiation, stress response, neuronal morphogenesis, cell surface receptor modulation, secretion, DNA repair, transcriptional regulation, long-term memory, circadian rhythms, immune response, and biogenesis of organelles ([@pbio-0020013-Glickman1]). With the multitude of substrates targeted and the myriad processes involved, it is not surprising that aberrations in the pathway are implicated in the pathogenesis of many diseases, including cancer. With so many proteins to target for degradation, the activity of the proteasome is subject to multiple levels of regulation. In the overwhelming majority of cases, selected proteins are first "labeled" by the addition of several copies of a small protein tag called ubiquitin and are thus targeted for degradation in the proteasome ([Figure 1](#pbio-0020013-g001){ref-type="fig"}). The ubiquitination of proteins is regulated through precise selection of protein substrates by specific E3 ubiquitin ligases ([@pbio-0020013-Pickart1]). These enzyme complexes each recognize a subset of substrates and tag them by linking the carboxyl terminus of ubiquitin with an amino group on the target protein via an amide bond ([Figure 1](#pbio-0020013-g001){ref-type="fig"}). ::: {#pbio-0020013-g001 .fig} Figure 1 ::: {.caption} ###### Structure of an Ubiquitinated Protein Ubiquitin (light violet) is a small 76 amino acid protein that can be covalently attached to target proteins (green) by specific E3 ubiquitin ligases. Such conjugation takes the form of an isopeptide bond between the carboxyl terminus of ubiquitin (denoted as C) and a lysine amino sidechain (K) on the substrate, or in some cases, conjugation can be via a peptide bond between ubiquitin and the amino terminus of the protein (N). These amide bonds are indicated as blue links. Multiple ubiquitin moieties can link in a similar manner via lysine-48 (K48) to form a polyubiquitin chain. As symbolized, more than one such chain can assemble on a single target. The result is a branched fusion protein with multiple amino termini (seven in the depicted example) coalescing at a single carboxyl terminus. Polyubiquitination in this manner targets proteins to the proteasome, where they are hydrolyzed into short peptides (green stack). Deubiquitinating enzymes can hydrolyze the bond between one ubiquitin moiety and another or between ubiquitin and the target protein. :::  ::: Interestingly, ubiquitination is a reversible process. Even when a protein has been tagged with ubiquitin, its fate is not sealed---specific hydrolytic enzymes called deubiquitinases can remove the ubiquitin label intact ([Figure 1](#pbio-0020013-g001){ref-type="fig"}). By deubiquitinating their substrates, these enzymes compete with the proteasome, which acts on the polyubiquitined form. In the competition between proteolysis and deubiquitination, polyubiquitinated proteins rarely accumulate in the cytoplasm of "healthy" cells, as they are either irreversibly degraded or deubiquitinated and rescued. It is thought that this competition provides a certain level of stringency or quality control to the system. Based on sequence homology, deubiquitinating enzymes were traditionally classified into two families: ubiquitin-specific proteases (UBPs or USPs) and ubiquitin carboxy-terminal hydrolases (UCHs). Both enzyme families are classified as cysteine proteases that employ an active site thiol to cleave ubiquitin from its target ([@pbio-0020013-Kim1]; [@pbio-0020013-Wing1]). The proteasome itself is made up of a multiprotein core particle (CP) where proteolysis occurs and a separate multiprotein regulatory particle (RP) that recognizes and prepares substrates for degradation by the CP. A base subcomplex of the RP is pivotal in anchoring polyubiquitin chains during this process, either directly or via auxiliary ubiquitin-binding proteins ([@pbio-0020013-Lam2]; [@pbio-0020013-Hartmann-Petersen1]). The base attaches to the outer surface of the CP and uses energy to unravel the substrate, simultaneously with preparing the channel that leads into the proteolytic chamber of the CP ([@pbio-0020013-Forster1]). The lid subcomplex of the RP attaches to the base and is required for proteolysis of ubiquitin--protein conjugates, but not of unstructured polypeptides ([@pbio-0020013-Glickman2]; [@pbio-0020013-Guterman1]). The size and complexity of this protein-eating machine hints at the exquisite controls that must rgulate its function. An intriguing evolutionary and structural relationship between the proteasome lid and an independent complex, the COP9 signalosome (CSN), may shed light on their respective roles in regulated protein degradation. Both are made up of eight homologous protein subunits that contain similar structural and functional motifs. While a lot is still unknown, the CSN appears to mediate responses to signals (e.g., light, hormones, adhesion, nutrients, DNA damage) in a manner that is intimately linked to the ubiquitin--proteasome system. This is accomplished, for instance, by suppressing ubiquitin E3 ligase activity or interacting with various components of the pathway ([@pbio-0020013-Bech-Otschir1]; [@pbio-0020013-Cope1]; [@pbio-0020013-Li1]). In particular, one subunit---Csn5---moderates SCF (Skp1--cullin--F box) and other cullin-based E3 ubiquitin ligases by removal of the ubiquitin-like Rub1/Nedd8 molecule from the cullin subunit of the ligase complex. Further analysis of the CSN will no doubt uncover additional mechanisms whereby ubiquitin-mediated protein degradation is controlled. Surprisingly, the proteasome itself harbors intrinsic deubiquitination activity ([@pbio-0020013-Eytan1]). Moreover, both the lid and the base contribute independently to RP deubiquitination activity. The source of this activity has been attributed to a number of different subunits. These include the associated cysteine proteases Ubp6/USP14 ([@pbio-0020013-Borodovsky1]; [@pbio-0020013-Legget1]), UCH37/p37 ([@pbio-0020013-Lam1]; [@pbio-0020013-Hoelzl1]), and Doa4/Ubp4 ([@pbio-0020013-Papa1]), as well as the intrinsic proteasome subunit Rpn11/POH1 ([@pbio-0020013-Verma1]; [@pbio-0020013-Yao1]). The importance of these components to proteasome function is apparent in their partially overlapping properties. In groundbreaking work, an intrinsic "cryptic" deubiquitinating activity that is sensitive to metal chelators has been reported for the proteasome, in addition to "classic" cysteine protease behavior ([@pbio-0020013-Verma1]; [@pbio-0020013-Yao1]). This metalloprotease-like activity maps to the putative catalytic MPN+/JAMM motif of the lid subunit Rpn11 and lies at the heart of proteasome mechanism by linking deubiquitination with protein degradation. Notably, Rpn11 shares close homology with Csn5, which is also responsible for proteolytic activities in its respective complex. By defining a new family of putative metalloproteases that includes a proteasomal subunit, a CSN subunit, and additional proteins from all domains of life, the MPN^+^/JAMM motif garnered great attention. The trademark of the MPN^+^/JAMM motif is a consensus sequence E---HxHx~(7)~Sx~(2)~D that bears some resemblance to the active site of zinc metalloproteases. Members of this family were predicted to be hydrolytic enzymes, some of which are specific for removal of ubiquitin or ubiquitin-like domains from their targets ([@pbio-0020013-Maytal-Kivity1]; [@pbio-0020013-Verma1]; [@pbio-0020013-Yao1]). In a further development, two independent groups determined the molecular structure of an MPN^+^/JAMM protein from an archaebacterium ([@pbio-0020013-Ambroggio1]; [@pbio-0020013-Tran1]). The structures identify a zinc ion chelated to the two histidines and the aspartic residue of the MPN^+^/JAMM sequence. The fourth ligand appears to be a water molecule activated through interactions with the conserved glutamate to serve as the active site nucleophile. Overall, this protein certainly has properties consistent with a metallohydrolase and can serve as the prototype for the deubiquitinating enzymes in its class. This revelation adds an all-new enzymatic activity and, with it, an additional layer of regulation to the ubiquitin--proteasome system. Now that it is evident that the proteasome contains a member of a novel metalloprotease family, a fundamental question can be raised: why does a proteolytic enzyme like the proteasome need auxiliary proteases for hydrolysis of ubiquitin domains? At first glance, the delegation of tasks between the proteolytic subunits of the proteasome (situated in the proteolytic core particle) and the auxiliary deubiquitinating enzymes (situated in the regulatory particle) is clear-cut: the latter cleave between ubiquitin domains, while the core proteolytic subunits process the target protein itself ([Figure 1](#pbio-0020013-g001){ref-type="fig"}). However, this still does not explain the mechanistic rational for finding deubiquitination within the proteasome itself. In principle, deubiquitination could be used for (1) recycling of ubiquitin, (2) abetting degradation by removal of the tightly folded highly stable globular ubiquitin domain, or (3) mitigating degradation by removal of the ubiquitin anchor, without which the substrate is easily released and rescued. There is evidence that recycling of ubiquitin by the proteasome is indeed a crucial feature of deubiquitination in proper cellular maintenance ([@pbio-0020013-Legget1]). Distinguishing between options 2 and 3, however, depends to a large extent on the delicate balance between the two proteolytic activities associated with the proteasome: proteolysis and deubiquitination ([Figure 2](#pbio-0020013-g002){ref-type="fig"}). ::: {#pbio-0020013-g002 .fig} Figure 2 ::: {.caption} ###### Deubiquitination versus Proteolysis at the Proteasome Once recognized and anchored to the proteasome via its polyubiquitin tag (light violet), a substrate (green) can be unraveled, unfolded, and translocated by the 19S regulatory particle (red) into the proteolytic chamber of the 20S core particle (purple), where it is hydrolyzed into short peptides (left). A byproduct of proteolysis is the polyubiquitin anchor (that may still be linked to a residual peptide). Cytoplasmic deubiquitinating enzymes eventually process this chain and recycle ubiquitin. However, the proteasome can also directly deubiquitinate the substrate, with diverse outcomes. For example, the substrate can be "shaved" upon cleavage of the bond to the proximal ubiquitin (right). Without its anchor, the substrate is presumably released and rescued. A distinct deubiquitinating activity is "trimming" or removal of the distal ubiquitin moiety (middle). According to one hypothesis, trimming serves as a timer; extended or difficult-to-process chains allow ample time for substrate unfolding and irreversible proteolysis (left), while short or easy-to-process chains inevitably lead to substrate release and rescue (right). This delicate balance between destruction and rescue is fundamental to proteasome efficiency. :::  ::: Once bound to the proteasome, a polyubiquitinated substrate can be unfolded by the RP and irreversibly translocated into the CP. It has been proposed that long polyubiquitin chains commit a substrate to unfolding and degradation by the proteasome, whereas short chains are poor substrates because they are edited by deubiquitinating enzymes, resulting in premature substrate release ([@pbio-0020013-Eytan1]; [@pbio-0020013-Lam1]; [@pbio-0020013-Thrower1]; [@pbio-0020013-Guterman1]). Extended polyubiquitin chains could slow down chain disassembly, thereby allowing ample time for unfolding and proteolysis of the substrate ([Figure 2](#pbio-0020013-g002){ref-type="fig"}). Interestingly, both "trimming" and "shaving" deubiquitinating activities are associated with the proteasome, though the exact contribution of the various proteasome-associated deubiquitinating enzymes to each of these distinct activities has yet to be elucidated. It is expected that in order to obtain efficient proteolysis of the target, shaving of chains at their proximal ubiquitin should be slower than the rate of trimming at the distal moiety. As an outcome of this requirement, longer polyubiquitin tags would be preferential substrates for degradation by the proteasome. Thus, the uniqueness of ubiquitin as a label for degradation may lie in its being a reversible tag. Deubiquitinases, such as Rpn11, serve as proofreading devices for reversal of fortune at various stages of the process, right up to the final step before irreversible degradation by the proteasome. Identifying Rpn11 and Csn5 as members of a novel class of metallohydrolases immediately elevates them into promising "drugable" candidates. Undoubtedly, the molecular structures deciphered by the groups of Deshaies ([@pbio-0020013-Ambroggio1]) and Bycroft ([@pbio-0020013-Tran1]) will focus efforts to design novel site-specific inhibitors of the ubiquitin--proteasome pathway. While Csn5 is thought to impede the action of ubiquitin ligases through shaving cullins from their Rub1/Nedd8 modification (and possibly also by deubiquitinating substrates bound to the cullins), the outcome of Rpn11 inhibition will depend largely on whether Rpn11 participates primarily in shaving substrates from their chains, promoting release and rescue, or in trimming the polyubiquitin tag, allowing for proteolysis quality control ([Figure 2](#pbio-0020013-g002){ref-type="fig"}). Michael H. Glickman is in the Department of Biology and Noam Adir is in the Department of Chemistry at the Technion--Israel Institute of Technology in Haifa, Israel. E-mail: <glickman@tx.technion.ac.il> CP : core particle CSN : COP9 signalo-some RP : regulatory particle SCF : Skp1--cullin--F box UBP : ubiquitin-binding protein UCH : ubiquitin carboxy-terminal hydrolase UBP/USP : ubiquitin-specific protease

PubMed Central

2024-06-05T03:55:46.024235

2004-1-20

PMC314469

Introduction {#s1} ============ Protein--protein interactions govern cell physiology, and the disruption of some sensitive connections in the network can have pathological effects. Once a genome has been sequenced, one of the goals of functional genomics is the elucidation of the protein interaction network supporting biochemical and genetic pathways. Eventually, the aim is to study the consequences on cell physiology of disrupting the specific interaction between any two given proteins. Over the past few years, a number of high-throughput strategies have been proposed to achieve this goal ([@pbio-0020014-Uetz1]; [@pbio-0020014-Ito1]; [@pbio-0020014-Gavin1]; [@pbio-0020014-Ho1]). These endeavors demonstrated the feasibility of a proteomic approach to the protein interaction problem. However, the lack of a substantial overlap between the results of projects designed to cover the entire interactome of Saccharomyces cerevisiae emphasized the importance of confirming any interaction by different methods ([@pbio-0020014-von1]). An in vitro strategy that has received considerable attention is based on the production of proteins in a high-throughput fashion and on their analysis in an array format ([@pbio-0020014-Zhu1]). This approach is not limited to the study of protein interactions, and various other protein functions, including enzymatic activities, can be tested in the array format. However, although several experimental strategies are being explored, it is not yet clear which percentage of a eukaryotic proteome can be produced in a folded form in conventional expression systems and still remain functional once printed onto a solid support. High-density arrays of relatively short peptide chains, on the other hand, can be efficiently synthesized by a positionally addressable synthesis of peptides on cellulose membranes (SPOT synthesis) and have been used to facilitate mapping of antibody epitopes and more generally to study protein binding specificity ([@pbio-0020014-Frank1]; [@pbio-0020014-Kramer1]; [@pbio-0020014-Reineke1]). The clear advantage of the array format could then be fully exploited to study protein interaction in those cases in which one of the partners participates in complex formation by docking a relatively short peptide into a receptor protein. In fact, a fairly large set of protein interactions are mediated by families of protein-binding domains (SH2, WW, SH3, PDZ, etc.) that act as receptors to accommodate, in their binding pockets, short peptides in an extended conformation ([@pbio-0020014-Pawson2]; [@pbio-0020014-Pawson3]; [@pbio-0020014-Pawson1]). We have recently shown that the peptide sequences, obtained by panning phage-displayed random peptide libraries with SH3 domains, can be used to derive position-specific scoring matrices to computationally scan the entire proteome in search of putative partners ([@pbio-0020014-Tong1]). This approach is affected by relatively low accuracy and/or coverage, depending on the threshold score that is set in the predictive algorithm. As a consequence, reliable inferences are only achieved by considering the intersection of the network obtained by the phage display and the one obtained by a completely unrelated (orthogonal) technique, such as the yeast two-hybrid method ([@pbio-0020014-Tong1]). We reasoned that an alternative strategy whereby the domain of interest is challenged with the entire collection of peptides that the domain is likely to encounter in the cell could eliminate one source of error. However, this straightforward approach is technically not feasible because the number of short peptides, even in a proteome as simple as the one of baker\'s yeast, is in the order of 10^7^. This figure is far beyond the limits of the current technology for peptide synthesis. On the other hand, one could use the information obtained from screening random peptide repertoires to filter out the amino acid sequences that are highly unlikely to bind, thereby decreasing the peptide sequence space to be tested experimentally. We will refer to this approach with the acronym WISE (whole interactome scanning experiment) ([Figure 1](#pbio-0020014-g001){ref-type="fig"}). ::: {#pbio-0020014-g001 .fig} Figure 1 ::: {.caption} ###### Schematic Representation of the WISE Strategy :::  ::: It should be pointed out that WISE only addresses the problem of identifying natural peptides with the potential for binding to any given recognition domain. Although we use this information to infer the formation of protein complexes in vivo, there are a number of reasons why this inference could turn out to be incorrect. For instance, a peptide could be unavailable for interaction in the native protein structural context. Alternatively, the two inferred partners could be located in different cell compartments or expressed in different tissues or at different times during development. Finally, all the interactions that are mediated by an extended region of a protein surface and that cannot be supported by a relatively short peptide will be missed by this approach. Results {#s2} ======= To assess the feasibility of this strategy, we have chosen eight S. cerevisaie proteins that contain SH3 domains belonging to five different specificity classes, as determined by phage display experiments ([@pbio-0020014-Tong1]). The SH3 domains of Rvs167 (P39743), Yfr024c (P43603), and Ysc84 (P32793) bind to peptides that conform to typical class 1 (RxxPxxP) and class 2 (PxxPxR) motifs. The SH3 domains of Boi1 (P38041) and Boi2 (P39969) bind to peptides that also match class 1 or class 2 motifs but that display a somewhat higher complexity and variability. By contrast, the SH3 domain of Sho1 (P40073) and Myo5 (Q04439) were found to bind only to class 1 peptides, with a preference either for positively charged or for large hydrophobic sidechains at position P-2, respectively. The SH3 binding motifs\' residue nomenclature (P-0 being the first Pro in the PxxP motif) is according to [@pbio-0020014-Lim1]). Finally, the SH3 domain of Abp1 (P15891) was poorly defined by the phage display experiments, possibly because peptides longer than nine amino acids are required for efficient binding. For each SH3 domain, we have defined a "relaxed pattern," less selective than the pattern identified by comparing the most frequent ligands discovered by phage display. We have then used this to scan the whole S. cerevisiae proteome in search of peptides matching that pattern. The yeast proteome was searched with the program PatMatch at the *Saccharomyces* Genome Database (<http://genome-www.stanford.edu/Saccharomyces/>). A detailed description of the relaxed patterns can be found in [Table S1](#st001){ref-type="supplementary-material"}. For instance, for class 1 peptides bound by the Rvs167 SH3 domain, instead of using the strict consensus motif RxFPxPP, we have defined a relaxed consensus by allowing either Arg or Lys at position P-3 and any amino acid at P-1 and P+2 (R/KxxPxxP). Standard conventions are used for representing consensus sequences of peptide ligands and protein modules ([@pbio-0020014-Aasland1]) and for the nomenclature of residue positions in SH3 ligands ([@pbio-0020014-Mayer1]). We have to emphasize that this strategy is only suitable for identifying SH3 partners whose ligand domain can be confined to a short peptide detectable by the phage display approach. Although this is often the case, we need to realize that some SH3 domain interactions require more extended binding surfaces. As a consequence, they will not be identified by this approach ([@pbio-0020014-Barnett1]). This approach was repeated for the eight SH3 domains. For each domain, approximately 1,500 peptides, matching the relaxed patterns, were selected for synthesis (see [Datasets S1--S8](#st001){ref-type="supplementary-material"}). The peptides were synthesized at high density on cellulose membranes by SPOT synthesis technology, and the membranes were probed with the corresponding SH3 domain fused to glutathione S-transferase (GST). Finally, the bound domains were revealed by an anti-GST antibody and by a secondary anti-immunoglobulin G (IgG) antibody coupled to horseradish peroxidase (POD). The intensity of each SPOT was measured quantitatively in Boehringer light units (BLUs) (arbitrary light intensity units measured by a Lumi-Imager^TM^ instrument) [Figure 2](#pbio-0020014-g002){ref-type="fig"} and [Datasets S1--S8](#sd001){ref-type="supplementary-material"} report the results of these experiments in which the pattern of the reactive spot forms a sort of fingerprint that defines the recognition specificity of the specific SH3 domain. The differences and similarities in recognition specificity are better appreciated in the representation of [Figure 3](#pbio-0020014-g003){ref-type="fig"}, where the red hue of the small horizontal bars indicate the intensity of the binding reaction of a specific peptide for each SH3 domain column. As expected from the phage display experiment, the SH3 domains of Rvs167, Yfr024c, and Ysc84 have overlapping specificities, with Rvs167 proving more selective and Yfr024c more promiscuous. By contrast, the peptides that bind to the Abp1 (P15891) and Myo5 SH3 domains are characterized by different motifs. The results reported in [Figure 3](#pbio-0020014-g003){ref-type="fig"} also point out that peptide recognition patterns inferred from the phage display experiments (green in [Figure 3](#pbio-0020014-g003){ref-type="fig"}) overlap only partially with the SPOT recognition patterns. These differences are particularly apparent in the case of Boi1 and Boi2. For these domains, the data obtained by phage display have proven insufficient for the target peptides to be inferred with sufficient accuracy, since only three out of 15 peptides have been predicted correctly. This comparison underlines the danger of using regular expressions or position-specific scoring matrices, derived from a relatively small number of peptide sequences, for inferring new peptide targets. ::: {#pbio-0020014-g002 .fig} Figure 2 ::: {.caption} ###### WISE Screening of the Binding Potential of Yeast SH3 Domains Seven GST--SH3 domain fusion proteins were challenged with peptides that match different relaxed consensi: class 1 (R/K)xxPxxP and class 2 PxxPx(R/K) . The Myo5 SH3 domain was also tested with peptides matching (F/P/L/W/A/E)xx(W/Y/L/M/F/H)xxPxxP, while the Abp1 membrane contains peptides matching either xxPx(K/R)P or Pxxx(K/R)P. In the design of these relaxed patterns, we first aimed at defining regular expressions that could retrieve from the proteome all the peptides that had been demonstrated, to bind to the domain under consideration. Whenever the number of matching peptides did not exceed an arbitrary chosen threshold of 1,500, we used subjective considerations about sidechain similarities to further relax the search pattern. The three spots near the membrane corners contain peptides that bind to the anti-GST antibody. The intensity of these spots was used for normalization. :::  ::: ::: {#pbio-0020014-g003 .fig} Figure 3 ::: {.caption} ###### Comparison of the Phage Display Prediction and the Results of the SPOT Binding Test by the WISE Approach The quantitative results of the experiments in [Figure 1](#pbio-0020014-g001){ref-type="fig"} are visualized with a graphical representation obtained with the tool EPCLUST available at <http://ep.ebi.ac.uk/EP/EPCLUST>. The PepSpot data, represented in red in a semiquantitative scale, is compared to the phage display prediction. Only peptides with BLUs (measured on a Lumi-Imager^TM^) higher than 25K are included in the representation. The red intensity scale corresponds to BLU values in the ranges 25K--35K, 35K--45K, 45K--55K, 55K--85K, and larger than 85K, where higher BLU values correspond to a brighter red. Peptides that obtained a high score with the phage display-derived position-specific scoring matrix ([@pbio-0020014-Tong1]) are in brighter green. Peptides with a lower score are represented with a correspondingly lighter green according to an arbitrary linear scale. :::  ::: Correlation of the SPOT Quantitative Output and Dissociation Constant {#s2a} --------------------------------------------------------------------- The sensitivity of the SPOT interaction experiment is such that even peptides with a dissociation constant as high as 10^−4^ M or above give a positive signal in the assay ([@pbio-0020014-Kramer2]). To establish a correlation between affinity and the BLU signal, we have measured, by surface plasmon resonance, the dissociation constants of a number of peptides that were positive in the membrane assay ([Figure 4](#pbio-0020014-g004){ref-type="fig"}A). The dissociation constants ranged from 9.4 × 10^−7^ M to values that, being larger than 10^−4^ M, could not be confidently measured. As previously observed for antibodies ([@pbio-0020014-Kramer2]), in these experiments signal intensity also correlated inversely with the dissociation constant (correlation coefficient of --0.4; [Figure 4](#pbio-0020014-g004){ref-type="fig"}B). This correlation was obtained by comparing experiments performed with different probes and different membranes and can be further improved through more careful standardization (C. Landgraf and R. Volkmer-Engert, unpublished data). Thus, this approach, in contrast with other high-throughput approaches, is accompanied by a quantitative output that correlates, albeit partially, with the dissociation constant. As such, it can be used to assign figures to the edges of the inferred interaction network. This is illustrated in [Figure 5](#pbio-0020014-g005){ref-type="fig"}A, where the inferred SH3-mediated interaction network is represented by different colors to differentiate interaction mediated by different SH3 domains and different edge thicknesses in order to distinguish interactions with different affinities. ::: {#pbio-0020014-g004 .fig} Figure 4 ::: {.caption} ###### Measurement of Dissociation Constants and Correlation with SPOT Intensities \(A) Dissociation constants were measured with a BIAcoreX instrument as described in the Materials and Methods. The experiments with the Abp1 SH3 domain were carried out in triplicate. \(B) Normalized BLU intensities plotted as a function of the log of the dissociation constant. :::  ::: ::: {#pbio-0020014-g005 .fig} Figure 5 ::: {.caption} ###### Inferred Protein Interaction Networks \(A) Protein interaction network mediated by the SH3 domains of the proteins characterized in this study. The SH3-containing proteins are represented as blue dots, while the prey partner proteins are represented as black dots. The interactions mediated by each SH3 are represented in a different color, and the edge thicknesses are proportional to the BLU intensity of the corresponding interaction, according to the scale described in [Figure 3](#pbio-0020014-g003){ref-type="fig"}. \(B) The graph represents the interaction network mediated by the SH3 domains of Rvs167, Ysc84, Yfr024c, Abp1, Myo5, Sho1, Boi1, and Boi2 as determined by the two-hybrid approach ([@pbio-0020014-Tong1]). The interactions (edges) that were confirmed by our WISE method (BLU value higher than 25K) are colored in red or magenta. The interactions in magenta, differently from the ones in red, were not correctly inferred by the phage display approach. The interaction in orange was inferred by the phage display approach, but not confirmed by the WISE method. The network was visualized by the Pajek package (<http://vlado.fmf.uni-lj.si/pub/networks/pajek/>). :::  ::: Inferred Protein Ligands Share Common Functions {#s2b} ----------------------------------------------- Interacting proteins often share similar functions and participate in common processes. Hence, we examined whether the proteins, found in our approach to bind to a specific SH3 domain, could be preferentially associated to a biological process. For this analysis we considered, as putative ligands, all the proteins containing at least one peptide with an intensity higher than an arbitrarily chosen threshold of 20,000 (in BLUs, corresponding to a dissociation constant of approximately 100 μM). We then used the FunSpec software ([@pbio-0020014-Robinson1]) to identify the Gene Ontology (GO) terms significantly enriched in the list of proteins interacting with any specific SH3 domains. FunSpec uses a hypergeometric distribution to evaluate the probability (*p*) that the intersection of a protein list with any given functional category occurs by chance. The inferred ligands of the Ysc84, Yfr024c, and Rvs167 SH3 domains were significantly enriched for the GO biological process term "actin cytoskeleton organization and biogenesis" (*p* \< 5.46 × 10^−7^, *p* \< 7.06 × 10^−6^, and *p* \< 5.50 × 10^−5^, respectively). By contrast, the partners of Abp1 and Myo5 SH3 domains were found to be enriched for the GO terms "actin cortical patch assembly" (*p* \< 3.49 × 10^−7^) and "actin cytoskeleton" (*p* \< 7.14 × 10^−6^). These results are in accord with the available information about the participation of these bait proteins in the organization of the yeast cytoskeleton, whereas arbitrarily selected gene groups of similar size showed no comparable enrichments for any of the GO terms (best result, *p* \< 10^−3^). We have previously shown that the information obtained by panning random peptide libraries can be used to draw an interaction network that recapitulates a fraction of the SH3-mediated interaction network determined by the two-hybrid approach ([@pbio-0020014-Tong1]). By using inferred networks of comparable size, the intersection with the two-hybrid network was larger for the WISE than for the phage display network, including three more proteins and six new edges ([Figure 5](#pbio-0020014-g005){ref-type="fig"}B). Furthermore, as shown below, at least some of the WISE interactions, not present in the two-hybrid network, can be shown to occur in physiological conditions in yeast. The Tightest SH3 Peptide Ligands Mediate Complex Formation In Vivo {#s2c} ------------------------------------------------------------------ The Abp1 SH3 domain, compared to most of the remaining SH3 domains that we have studied, has a narrower peptide recognition specificity and, as a consequence, fewer inferred partners. Our analysis confirmed that Srv2 and Ark1, previously identified Abp1 SH3 functional partners ([@pbio-0020014-Lila1]; [@pbio-0020014-Fazi1]), contain peptides that bind with an affinity in the 1--100 μM range. Furthermore, fragments of Prk1 (P40494), Yir003w (P40563), and Ynl094w (P53933) were reported to bind to the Abp1 SH3 domain in vitro ([@pbio-0020014-Fazi1]). Surprisingly, we could not identify any tetradecapeptide in the Ynl094w protein with affinity better than 100 μM. We noticed, though, that if we extend the peptide RRPPPPPIPSTQKP (predicted to be a ligand of the Abp1 SH3 domain by a variety of approaches) to include three more residues at the C-terminus, the affinity rises to approximately 40 μM. Finally, we identified Scp1 (Q08873), the yeast homolog of calponin, as a putative new Abp1 partner. In order to assess how accurately peptide binding affinity permits us to infer physiological partners, we investigated whether the putative partners can be copurified with Abp1 in vivo. We used the tandem affinity purification (TAP) technology ([@pbio-0020014-Rigaut1]) to tag the Prk1, Ynl094w, Scp1, and Yir003w proteins, and we initially asked, by pulldown assays, whether the putative peptide targets of the Abp1 SH3 domain were available for interaction in the protein natural context. As seen in [Figure 6](#pbio-0020014-g006){ref-type="fig"}A, the four proteins identified by our approach can be affinity-purified on a sepharose resin containing the Abp1 SH3 domain, thus indicating that the target peptides can bind in their native protein context. ::: {#pbio-0020014-g006 .fig} Figure 6 ::: {.caption} ###### Characterization of Abp1 Ligands \(A) The dissociation constants of the 11 peptides that bound most efficiently to the Abp1 SH3 domain in the SPOT synthesis assay were measured by BIAcore experiments. (See also [Table S1](#st001){ref-type="supplementary-material"}.) The results of the experiments for the peptides with the highest affinity are reported here. \(B) The genes encoding the putative Abp1 ligands (Prk1, Yir003w, Scp1, and Ynl094w) were modified by the TAP technology to produce tagged proteins. A strain expressing the "tapped" Bmh1 protein is used as a control. Yeast extracts encoding the tagged proteins were used in pulldown experiments in the presence of 100 μg of GST--Abp1 SH3 or GST alone as a negative control. The "Ext." lane was loaded with 1/20 of the extract used in the pulldown experiment. \(C) The same extracts were affinity-purified on an IgG affinity resin and then the affinity tag, protein A, released by cutting with the TEV protease. The proteins that were copurified with the "tapped" baits were revealed with an anti-Abp1 serum. :::  ::: To establish whether Abp1 forms a complex with these proteins in vivo, we next affinity-purified the four tagged proteins on an IgG resin. We next probed the purified complexes with an anti-Abp1 antibody. As shown in [Figure 6](#pbio-0020014-g006){ref-type="fig"}B, Abp1 could be copurified with Prk1, Scp1, Ynl094w, and Yir003w, but not with Bmh1, used as a negative control. In conclusion, at least in the case of Abp1, the search for the tightest binding peptides in the whole yeast proteome led to the to identification of proteins that form a complex with the bait domain when expressed at physiological levels in vivo. We have also investigated whether the fraction of coimmunoprecpitated Abp1 protein correlates with either the BLU intensity or the dissociation constant of the corresponding SH3--peptide interaction. The observed lack of correlation indicates that other factors, as, for instance, local protein concentration (mediated by different interactions), are important in determining the efficiency of complex formation. WISE Scanning of the Human Proteome {#s2d} ----------------------------------- Finally, we asked whether this approach could be extended to the analysis of a mammalian proteome that is approximately five to six times more complex than the yeast one. To this end, we selected two proteins involved in membrane recycling, amphiphysin-1 (P49418) and endophilin-1 (Q99962), and whose SH3 domains we had previously characterized by phage display ([@pbio-0020014-Cestra1]). These two SH3 domains are also known to have overlapping recognition specificity, although their preferred target peptides are different and the overall recognition specificity differs from the ones of the yeast SH3 domains characterized so far by this approach (see [Figure 2](#pbio-0020014-g002){ref-type="fig"}). We have screened with the amphiphysin and endophilin SH3 domains all the peptides in the SwissProt/TREMBL database that contain the (P/F/l/I)XRPXX(R/K), the (P/F/l/I)(K/R)RP, or the (P/l/R/F/S/I/V/K/G)PX(R/K)PP motifs. Because of the redundancy of the SwissProt/TREMBL database and because the peptide families matching the three motifs overlap, some of the total 3,774 peptides were synthesized several times, thereby providing an internal control of the approach\'s reproducibility ([Figure 7](#pbio-0020014-g007){ref-type="fig"}; Datasets [S9](#sd009){ref-type="supplementary-material"} and [S10](#sd010){ref-type="supplementary-material"}). ::: {#pbio-0020014-g007 .fig} Figure 7 ::: {.caption} ###### Scanning of the Human Proteome in Search of Ligands for the Amphiphysin and Endophilin SH3 Domains The relaxed target peptide consensi (right) were derived from the available phage display experimental data and used to search the human proteins contained in the SwissProt/TREMBL database with the software ScanProsite, found at <http://us.expasy.org/tools/scanprosite/>. :::  ::: Dynamin and synaptojanin, two proteins involved in endocytosis, form an SH3-mediated complex with amphiphysin and endophilin in vivo ([@pbio-0020014-McPherson1]; [@pbio-0020014-de1]; [@pbio-0020014-Ringstad1]). Our approach identified in both proteins at least one peptide that is a ligand for the amphiphysin and endophilin SH3 domains. Interestingly, other proteins that have been already implicated in endocytosis and its control (but not yet described as physiological partners of amphiphysin and endophilin) contain peptides that are ranked among the highest affinity ligands by our approach. Several other proteins of unknown function are predicted to bind to the SH3 domain of these two proteins. Discussion {#s3} ========== The WISE strategy described here has the merit of combining the strengths of a selective approach (such as panning combinatorial peptide libraries displayed on phage) with a quantitative analysis that can be achieved by screening a large number of peptides arrayed at high density on a solid support. This makes it possible to identify rapidly and directly the tighest ligands of a peptide-binding receptor among all the peptides in an entire proteome. We have demonstrated the approach by applying it to the family of SH3 domains. However, WISE can also be extended to all those domain families (WW, PDZ, EH, GYF, VHS, SH2 PTB, 14-3-3, FHA, WD40, etc.) that mostly recognize short peptides in their partner proteins. Our approach, as any in vitro approach, suffers from some simplifications when it comes to inferring the physiological partners from the domain--peptide interaction data. According to a naive strategy, we would assimilate the cell to a cellulose membrane, where all the peptides are equally represented and accessible to the bait domain, and we would be tempted to conclude that all the proteins containing the identified peptide ligands were likely to be physiological partners. In the real cell, however, the target peptides may be hidden inside the core of the folded proteins, and the protein partners may not be equally represented. Furthermore, the partner proteins may be expressed in different cell types or segregated in different macromolecular complexes or cell compartments. In order to obtain more reliable inferences, the peptide interaction information obtained by a WISE approach should be complemented by information about peptide accessibility obtained by structural predictors ([@pbio-0020014-Garner1]; [@pbio-0020014-Linding1]) and data about mRNA and protein concentrations in different physiological and subcellular contexts ([@pbio-0020014-Simpson1]; [@pbio-0020014-Kumar1]). Nevertheless, the average number of peptides in the yeast proteome that have the potential to bind SH3 domains with an affinity that may have physiological relevance was found to be surprisingly high, ranging from a few peptides, in the case of the Abp1 and Boi2 SH3 domains, to several tenths, in the case, for instance, of the Yfr024w SH3 domain. Given the hypothesis that all (or most of) these peptides are equally expressed inside the cell and exposed to the solvent in the folded protein structure as most Pro-rich peptides are, these findings raise the question of whether the observed binding promiscuity has any physiological implication. Recent proteome-wide analyses of yeast protein complexes have revealed that many proteins are organized in discrete complexes ([@pbio-0020014-Gavin1]; [@pbio-0020014-Ho1]). Yet this approach has failed to identify a large number of interactions whose physiological relevance was validated by traditional single (or few) protein studies, implying that many physiologically relevant protein interactions do not lead to the formation of stable complexes. SH3-mediated interactions may belong to this latter class. This is consistent with the observation that SH3-containing proteins have a connectivity significantly lower than average (2.33; average, 4.00) in the yeast complexosome ([@pbio-0020014-Gavin1]; [@pbio-0020014-Ho1]), in contrast with the observed connectivity in the interaction network derived from high-throughput two-hybrid experiments (average connectivity of SH3-containing proteins, 5.05; average connectivity for all proteins, 1.53) ([@pbio-0020014-Uetz1]; [@pbio-0020014-Ito1]; L. Montecchi-Palazzi and G. Cesareni, unpublished data). SH3-mediated interactions are much less likely to be detected by coimmunoprecipitation assays than by solid-state (or two-hybrid) assays, because relatively weak interactions are almost certainly lost in the extensive washing needed for coimmunoprecipitation experiments. Our approach has made it possible to rediscover most of the SH3-mediated protein interactions that were previously described for these proteins. Admittedly, though, few clearly characterized protein interactions of this type have yet been reported in the literature. The few failures of our approach (false negatives) are due to weaknesses in the design of the relaxed consensus used to search for matching peptides in the protein databases. All the same, we have identified a larger number of target peptides that bind with affinities that are comparable to the ones of the validated physiological targets. Some of these peptides may never encounter the cognate SH3 domain, while some will only meet partners in specific physiological conditions. Others may contribute to add specificity to the formation of a complex by cooperating with other associated low-specificity binding domains. Finally, we have to consider a new scenario in which proteins, even when not forming stable complexes, are seldom isolated in solution, but navigate in the cell by moving from one weak partner to another. These weak interactions may be important in modulating cell architecture even when they are not instrumental in the nucleation of a stable complex. Although this is difficult to prove, the semiquantitative data provided by our approach, complemented with the results of large-scale expression and localization studies, may eventually allow one to model these different settings. The in vitro approach that we have described, albeit limited to interactions in which one of the partners can be reduced to a relatively short peptide, presents a number of interesting features that complement other strategies aimed at revealing the details of the protein interaction network within cells. First, it takes full advantage both of the genomic information that is being accumulated and of the array format in which all the possible targets are equally represented. Second, it is comprehensive and provides a high level of detail on the interaction topology. Third, it is not affected by protein concentrations inside the cell and is very sensitive (interactions up to 100 μM can be detected). Fourth, interactions that depend on peptide modifications, for instance, phosphorylation, can also be studied. Fifth, the output is semiquantitative. Finally, the identified target peptide can be used as a lead to develop tighter binding molecules in order to interfere with complex formation in vivo. We have shown that the current implementation of the SPOT synthesis technology is sufficient to carry out WISE screening of a proteome as complex as the one of a mammalian organism. Foreseeable technological improvements of the SPOT synthesis technology will permit the assembly of relatively cheap microarrays containing up to 15,000 peptides. This will extend the approach\'s power by relieving, in some cases, the requirement for an experimental filtering step, as performed here by the phage display approach, thereby allowing more freedom in the design of the relaxed pattern. Materials and Methods {#s4} ===================== {#s4a} ### Genome tagging {#s4a1} Yeast PJ694a strains expressing TAP-tagged ORFs were constructed as described ([@pbio-0020014-Rigaut1]). Primers bearing a sequence identical to the C-terminal part of the ORF were used to amplify the TAP cassette. Primers for Yir003 are forward: GACGTTGATTCTGCCTTACATTCAGAAGAAGCGTCTTTTCACTCCCTTTCCATGGAAAAGAGAAGATG and reverse: CCATTATTATTAATAACACCTCTAGTTTGCTCGTCATTCACATATTTCTACGACTCACTATAGGGCGA. Primers for Scp1 are forward: TCTCAGGCTACTGAAGGAGTGGTGTTAGGACAACGGAGAGATATAGTTCCATGGAAAAGAGAAGATG and reverse: GGAAAACTAAAATATATCAAAGGAACTTTGGTTGCGTATATAGGGTTCTACGACTCACTATAGGGCGA. Primers for Prk1 are forward: GTAGATGATTTAGAAGCCGATTTTAGAAAAAGGTTTCCCAGCAAAGTTTCCATGGAAAAGAGAAGATG and reverse: AAAAATTTCAAATGATTGACGAAAGAAAATTTGTACATTTTGTATGACTACGACTCACTATAGGGCGA. Primers for Ynl094w are forward: TTAAGTTTGGAAGACAGTATTCGCAGAATTAGGGAGAAGTATTCAAACTCCATGGAAAAGAGAAGATG and reverse: CACTCTAAAACGTTGAAAATGGCTCCAATTCATAAGGTCACTTTAGTGTACGACTCACTATAGGGCGA. The polymerase chain reaction (PCR) fragments were used to transform the yeast strain. Positive clones were selected on selective plates and checked by PCR analysis and Western blot analysis. For the PCR analysis, we used a new forward primer together with the reverse ones used for the construction: forward for Yir003, AGCAGATGGAGGACCAAATGGAGGTTG; forward for Scp1, CGGTTATATGAAAGGTGCATCTCAGGC; forward for Prk1, CGTTTACAATCAAAGAAACTGCCGATTG; and forward for Ynl094w, GGACTCAATTCAAAAATTGAGCAATCAAG. ### Pulldown assay {#s4a2} Yeast strains expressing TAP-tagged Yir003w, Scp1, Prk1, Ynl094c, or Bmh1 as a control, were cultured at 30°C in 5-l flasks containing 2 l of YPD medium, collected in the exponential growth phase, and lysed mechanically with glass beads in 5 ml of IPP-150 buffer (10 mM Tris--HCl \[pH 8.0\], 150 mM NaCl, 0.1% NP-40) in the presence of protease inhibitors (2 mM benzamidine, 0.5 mM PMSF, 1 mM leupeptina, 2.6 mM aprotinin). Half of the extract was incubated for 2 h at 4°C with 100 μg of GST--Abp1SH3 (bound to glutathione--sepharose), while the remaining half was incubated with 100 μg of GST as a control. The resins were washed four times with 5 ml of IPP-150 buffer and the bound proteins recovered (by boiling in SDS--BLU--dye) and analyzed on a 10% SDS--polyacrylamide gel. They were transferred onto nitrocellulose membranes. Filters were blocked overnight at 4°C in PBS containing 5% milk powder (blocking solution), and then incubated with peroxidase (POD)-conjugated anti-POD antibody (PAP) antibody (Sigma P-2026; Sigma, St. Louis, Missouri, United States) diluted 1:1,000 for 2 h at room temperature (RT), washed five times for 15 min with PBS--0.05% Tween, and revealed by chemoluminescence. GST fusion proteins were expressed and purified by standard procedures. ### Coimmunoprecipitation {#s4a3} Yeast cultures expressing TAP-tagged Yir003, Scp1, Prk1, Ynl094c, or Bmh1 were cultured, collected, and lysed as described for the pulldown experiments. Each extract was incubated with 500 μl of IgG--sepharose (Pharmacia Biotech 17--0969-01; Amersham Pharmacia, Uppsala, Sweden) for 2 h at 4°C. The resins were washed four times in 5 ml of IPP-150 buffer, resuspended in 300 ml of 50 mM Tris--HCl (pH 8), 0.5 mM EDTA, 5 mM DTT, transferred to Eppendorf tubes, and incubated with 30 U of recombinant TEV protease (Invitrogen 10127--017; Invitrogen, Carlsbad, California, United States) for 1 h at 20°C. After centrifugation for 2 min at 2,300 rpm, the supernatants were loaded on 10% SDS--polyacrylamide gels and then transferred onto nitrocellulose membranes. Filters were blocked overnight at 4°C in blocking solution, incubated for 2 h at RT with anti-Abp1 antibody (diluited 1:1,000), and washed five times with PBS--0.05% Tween. They were then incubated for 1 h at RT with anti-rabbit POD coniugated, washed ten times with PBS--0.05% Tween, and detected by chemoluminescence. ### Peptide array synthesis {#s4a4} Cellulose membrane-bound peptides were automatically prepared according to standard SPOT synthesis protocols ([@pbio-0020014-Frank1]) using a Spot synthesizer (Abimed, Langenfeld, Germany) as described in detail ([@pbio-0020014-Kramer1]). For generation of the sequence files, the software LISA (Jerini AG, Berlin, Germany) was used. To exclude false-positive spots in the incubation experiment, all Cys were replaced by Ser. The generated arrays of 15mer peptides were synthesized on cellulose-(3-amino-2-hydroxy-propyl)-ether (CAPE) membranes, because of a better signal-to-noise ratio in the incubation experiments. ### Preparation of CAPE membranes {#s4a5} A 18 × 28 cm Whatman 50 paper (Whatman, Maidstone, United Kingdom) was immersed in a stainless steel dish containing a solution of 400 mg of *p*-toluenesulfonic acid in methanol (50 ml) and shaken for 3 min. The membrane was removed from the tray and air-dried. Meanwhile a solution of 7.8 g of N-(2,3-epoxypropyl)-phathalimid in dioxane (60 ml) was heated up to 80°C in a covered stainless steel dish placed on a shaking platform using a heater mat. Then, a solution of 400 mg of *p*-toluenesulfonic acid in 5 ml of dioxane was added. Immediately, the membrane was placed in this solution and shaken at 80°C for 3--5 h. Afterwards, the membrane was washed three times with 50 ml of dioxane and ethanol (twice, 50 ml each) and subsequently incubated with a 10% (v/v) solution (50 ml) of hydrazine hydrate (80%) in ethanol for approximately 6 h. Finally, the membrane was washed twice with ethanol, three times with dimethylacetamide, and once again with ethanol (twice, 50 ml each), and dried. The loading of this type of amino-functionalized cellulose membrane is about 120--200 nmol/cm^2^. ### SH3 domain binding studies of cellulose-bound peptides {#s4a6} Generally, all incubations and washing steps were carried out under gentle shaking. After washing the membrane once with ethanol (10 min) and three times for 10 min with Tris-buffered saline (TBS: 50 mM Tris-(hydroxymethyl)-aminomethane, 137 mM NaCl, 2.7 mM KCl, adjusted to pH 8 with HCl), the membrane-bound peptide arrays were blocked (3 h) with blocking buffer. Blocking reagent (CRB, Northwich, United Kingdom) was diluted 1:10 in TBS containing 5% (w/v) sucrose. After washing with TBS (10 min), 10 μg/ml of the corresponding GST-fused SH3 domain (in blocking buffer) was added and incubated overnight at 4°C. After washing three times for 10 min with TBS, the anti-GST monoclonal antibody (mAb) (G1160; Sigma) was added at a concentration of 1 μg/ml in blocking buffer for 2 h at RT. Subsequently, the membrane was washed three times with TBS (10 min each) and the POD-labeled anti-mouse mAb (1 μg/ml in blocking buffer) was applied for 1.5 h at RT, followed by washing three times with TBS. Analysis and quantification of peptide-bound SH3 domains were carried out using a chemoluminescence substrate and the Lumi-Imager^TM^ instrument (Roche Diagnostics, Basel, Switzerland). For quantification, the SPOT signal intensities were measured in BLUs. To exclude false-positive results, in the SH3-incubation experiment, each membrane was preexamined with GST/anti-GST mAb/anti-mouse mAb. The data obtained with the different membranes were normalized by using as a reference the intensity of three control peptides that bind to the anti-GST antibody. The sequence of these peptides were QRALAKDLIVPRRP, LAKDLIVPRRPEWN, and DLVIRPPRPPKVLGL. ### BIAcore analysis {#s4a7} Surface plasmon resonance measurements were carried out with a BIAcoreX instrument (BIAcore, Uppsala, Sweden). Experiments were carried out on sensor chips with GST-fused SH3 domains and GST as a control. GST-fused SH3 domains and the GST were coupled to CM5 sensor chips using the EDC/NHS (N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide and N-hydroxysuccinimide)amine-coupling kit, yielding approximately 4,350 resonance units in the case of the GST-fused SH3 domain and 4,330 resonance units for GST. Interaction analysis was performed at 25°C with the peptides dissolved in 10 mM HEPES, 150 mM NaCl, 3.4 mM EDTA, and 0.005% surfactant P20 (pH 7.4), at 15 μl/min flow rate, using six to seven dilutions, ranging from 500 μM to 65 nM. Dissociation constant values were evaluated by applying the steady-state model using BIAcore evalution 3.1 software. Supporting Information {#s5} ====================== Dataset S1 ::: {.caption} ###### Results of the SPOT Analysis Experiments for the Abp1 SH3 Domain (102 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S2 ::: {.caption} ###### Results of the SPOT Analysis Experiments for the Boi1 SH3 Domain (104 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S3 ::: {.caption} ###### Results of the SPOT Analysis Experiments for the Boi2 SH3 Domain (103 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S4 ::: {.caption} ###### Results of the SPOT Analysis Experiments for the Myo5 SH3 Domain (86 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S5 ::: {.caption} ###### Results of the SPOT Analysis Experiments for the Rvs167 SH3 Domain (104 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S6 ::: {.caption} ###### Results of the SPOT Analysis Experiments for the Yfr024c SH3 Domain (104 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S7 ::: {.caption} ###### Results of the SPOT Analysis Experiments for the Yhr016c SH3 Domain (104 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S8 ::: {.caption} ###### Results of the SPOT Analysis Experiments for the Sho1 SH3 Domain (76 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S9 ::: {.caption} ###### Results of the SPOT Analysis Experiments for the Amphyphisin SH3 Domain (183 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S10 ::: {.caption} ###### Results of the SPOT Analysis Experiments for the Endophilin SH3 Domain (184 KB TXT). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S1 ::: {.caption} ###### Design of Relaxed Consensi (66 KB PDF). ::: ::: {.caption} ###### Click here for additional data file. ::: We thank B. Andrews for the Abp1 antibody and G. Bader and S. Nasi for critical reading of the manuscript. This work was supported by the following grants: an Associazione Italiana per la Ricerca sul Cancro (AIRC) and Fondo per gli Investimenti della Ricerca di Base (FIRB) grant (RBNE01KXC9) to GC; European Union grants (QLG1-1999-00739 and QLK-CT-2002--01956) to GC and Ministero dell\'Istruzione, dell\'Università e della Ricerca (MIUR) (Progetto Genomica Funzionale) grant to LC. SP was supported by a European Science Foundation Fellowship. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** LC, JS-M, RV-E, and GC conceived and designed the experiments. CL and SP performed the experiments. LM-P analyzed the data. JS-M and RV-E contributed reagents/materials/analysis tools. GC wrote the paper. Academic Editor: Gerald Joyce, Scripps Research Institute BLU : Boehringer light unit CAPE : cellulose-(3-amino-2-hydroxy-propyl)-ether EDC : N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide GO : Gene Ontology GST : glutathione S-transferase HNS : N-hydroxysuccinimide IgG : immunoglobulin G mAb : monoclonal antibody PAP : peroxidase-conjugated anti-peroxidase antibody PCR : polymerase chain reaction POD : peroxidase RT : room temperature SPOT synthesis : synthesis of peptides on cellulose membranes TAP : tandem affinity purification TBS : Tris-buffered saline WISE : whole interactome scanning experiment

PubMed Central

2024-06-05T03:55:46.025508

2004-1-20

PMC314470

Scientists have long known that variation in animal color patterns carry far more than cosmetic significance. Darwin first connected pigmentation with adaptive advantage, noting that male finches with bright red plumage enjoyed greater reproductive success than their drab competitors. Explaining why coloration confers such advantages, however, has proved somewhat easier than showing how it arises. Biologists studying how neighboring regions of the vertebrate body plan develop differences in appearance and form have identified a small number of signaling pathways common to all animals. How and whether these pathways also control the developmental expression and variation of surface attributes like hair color, hair density, and hair length are unclear. By studying an old mouse mutant called *droopy ear*, Gregory Barsh and colleagues show that a member of the well-known family of T-box genes is required for a key pigmentation pattern in mice. Many vertebrate species---be they fish, bird, or mammal---have a much lighter belly than back. Studies in mice indicate these dorsoventral pigment differences arise from differential expression of the Agouti gene in the ventral and dorsal regions of the developing mouse; Agouti produces a pale yellow color and thus mice with light bellies have Agouti expressed in their ventral but not dorsal region. *Droopy ear* was discovered more than 50 years ago by virtue of its effects on head and ear shape, but it also affects pigmentation patterns; mutant mice have expanded ventral-specific Agouti expression into the dorsal region. First, Sophie Candille, a graduate student in Barsh\'s laboratory, searched for the gene that underlies the defect in *droopy ear*. When the researchers homed in on the chromosomal region known to harbor *droopy ear*, they found *Tbx15*---a member of the T-box gene family. T-box genes are found in a wide range of species and play diverse roles during embryonic development. In the *droopy ear* mouse, *Tbx15* carries a mutation that makes the protein nonfunctional. The researchers made certain that *Tbx15* really is the *droopy ear* gene by deleting most of the gene\'s coding region and showing that this "knocked-out" gene produces the typical *droopy ear* mouse. The pattern of embryonic *Tbx15* expression---determined by observing messenger RNA transcripts in developing tissues of the head, trunk, and limbs---suggests that early expression of *Tbx15* in the dorsal flank sets coordinates for dorsoventral differences in hair length and pigmentation. Candille et al. demonstrate that the regional pigment differences characteristic of adults is indeed established soon after embryonic *Tbx15* expression. So this boundary in pigmentation is set up very early during development. Interestingly, the early coordinates of the future pigment boundary do not correspond to any other known developmental boundary. The *Tbx15* pigmentation effects seen in these lab mice, the researchers note, resembles coat variations in other mammals, including German shepherds and an endangered mouse whose lighter dorsal markings once gave it an adaptive advantage on the white sand reefs where it lives (sadly, such markings offer no protection against loss of habitat). T-box genes are also found in humans; mutations in *Tbx1*, *Tbx4*, *Tbx5*, and *Tbx22* can cause developmental abnormalities of the heart, limbs, or of the head and neck. Mutations of human *Tbx15* have not yet been identified, but could contribute to regional differences of pigmentation (in dorsal and ventral surfaces of the limbs, for example) or to development of the head and neck. The identification of *Tbx15* adds a new player to the genes that help pattern the developing embryo---attention now turns to the controls that regulate *Tbx15* and the *Tbx15* targets, which set up the pattern. ::: {#pbio-0020016-g001 .fig} ::: {.caption} ###### Dorsoventral pigment boundaries in mouse and human :::  :::

PubMed Central

2024-06-05T03:55:46.029150

2004-1-20

PMC314471

Individuals within a wild population show remarkably little morphological variation, given the amount of environmental variation they encounter during development and the amount of genetic variation within the population. This phenotypic constancy led to the proposal that individuals were somehow buffered, or canalized, against genetic and environmental variation ([@pbio-0020019-Waddington1]). Clearly, such a mechanism would have important evolutionary consequences; because natural selection acts upon phenotypic variation within a population, canalization first appears to reduce the evolvability of the trait upon which it is acting ([@pbio-0020019-Gibson1]). However, canalization also reduces the effects of new mutations (which may be deleterious), potentially allowing individuals to store this genetic variation without suffering the consequences. If canalization breaks down due to genetic or environmental circumstances, then the stored genetic variation will be released, providing an additional substrate for natural selection. In this way, individuals could potentially undergo large, rapid phenotypic changes. Experiments in both Drosophila and Arabidopsis have suggested that Hsp90 (heat shock protein 90), a member of a family of proteins expressed at high temperatures (heat shock), may be an excellent candidate for bringing about canalization ([@pbio-0020019-Rutherford1]; [@pbio-0020019-Queitsch1]). Several features of Hsp90 suggest that it is an evolutionary buffer, capable of hiding and then releasing genetic variation: (1) individuals heterozygous for mutations in *Hsp83* (the gene encoding Hsp90) show increased levels of morphological abnormalities; (2) individuals treated with a pharmacological inhibitor of Hsp90 show severe morphological abnormalities; (3) the normal function of Hsp90 is to stabilise the tertiary structure of signal transduction molecules involved in developmental pathways; and (4) this function may be compromised by environmental factors, e.g., heat shock. Gene Networks Generate Canalization {#s2} =================================== Hsp90 may not, however, be uniquely placed to act as an evolutionary buffer producing canalization. Recent theoretical work has suggested that canalization may be an emergent property of complex gene networks and may not require specific mechanisms of protein stabilisation and environmental coupling such as those provided by Hsp90 ([@pbio-0020019-Siegal1]). [@pbio-0020019-Siegal1] proposed that when a network is compromised by 'knocking out' one of several genes, buffering may be lost or compromised, releasing variation that was hidden in the intact network. To test this, [@pbio-0020019-Bergman1] used numerical simulations of a complex network of ten genes in which each gene is capable of influencing the expression of other genes as well as itself ([Figure 1](#pbio-0020019-g001){ref-type="fig"}A). This network essentially defines the genotype of the individuals within the population, and the amount of gene expression at equilibrium defines the phenotype. Comparison of populations founded by either wild-type individuals or those with a single gene 'knockout' revealed much higher levels of phenotypic variation in populations derived from the 'knockouts'. ::: {#pbio-0020019-g001 .fig} Figure 1 ::: {.caption} ###### Similarity between a Gene Network Acting as an Evolutionary Buffer and a Gene Network Regulating Neuronal Ion Channel Expression \(A) Each gene (horizontal arrow) is regulated by the products of the other genes by means of upstream enhancer elements (boxes). The strength and direction of regulation (depicted as different colour saturation levels) are a function of both the upstream element and the abundance of its corresponding gene product. \(B) A similar representation of a putative network for activity-dependent ion channel regulation in a neuron in which Ca^2+^ concentration acts as a feedback mechanism. \(C) The mechanism of ion channel compensation in a neuron. The activity of the neuron is dependent upon its synaptic inputs and the suite of ion channels it expresses. Mutation of a gene encoding an ion channel leads to a change in the properties of that channel (depicted as a change in colour saturation) and hence to an increase in activity and internal Ca^2+^ (purple). These changes induce a compensatory increase in the expression of another ion channel (red) to restore the original level of activity. :::  ::: Thus, populations derived from 'knockouts' express phenotypic variation that was not expressed by the wild-type network, suggesting that any of the genes within the network may buffer genetic variation. This suggests that at least one aspect of generating evolutionary buffering is not unique to Hsp90. But can genes that, unlike Hsp90, are not conditional upon the environment act as evolutionary buffers? To test this, [@pbio-0020019-Bergman1] simulated a gene network that incorporated a mutation process in which single genes may be 'knocked out' and then, at a later time, restored. The simulated populations were allowed to evolve whilst being selected for an optimum phenotype (i.e., the populations were exposed to an environment in which a particular phenotype was optimal). A new optimum phenotype was then specified in which the expression of three of the ten network genes changed from on to off or vice versa (i.e., there was a shift in the environmental conditions favouring a different phenotype). Populations evolving with the mutation process reached the new optimum before populations without the mutation process. Thus, the 'knockout' mutations were clearly beneficial because they sped up adaptation to a new phenotypic optimum by releasing hidden genetic variation, thereby providing a new substrate upon which natural selection may act. Yet these mutations were not coupled to the new environment, suggesting that the release of the hidden genetic variation does not have to be linked to an environmental change in order to be beneficial. The simulations described by [@pbio-0020019-Bergman1] suggest that the key properties of an evolutionary buffer, the ability to store and then release genetic variation in response to environmental or genetic change, are not unique to Hsp90. Indeed, the simulations suggest that evolutionary buffering may be a widespread property of gene networks. They also suggest that the hidden genetic variation does not have to be revealed by an environmental change, but can be produced by a gene 'knockout'. These results may go some way to explain the original observation by [@pbio-0020019-Waddington1] of phenotypic constancy, yet many questions remain ([@pbio-0020019-Stearns1]). One of the major outstanding questions must be whether it is possible to verify these results experimentally. [@pbio-0020019-Bergman1] used data from the yeast Sacchromyces cerevisiae, in which each gene may be 'knocked out' in turn and the expression of the remaining genes determined, to demonstrate that their simulations also had application to biological gene networks. Using these data, they showed that 'knockouts' show greater variability in gene expression than wild-type yeast, suggesting that buffering has been disrupted. Ion Channels as Evolutionary Buffers {#s3} ==================================== Given the results of [@pbio-0020019-Bergman1], it should be possible to find gene networks in which the elimination of single genes reveals variation in gene expression and hence in phenotype. One class of gene network that may conform to the structure outlined by [@pbio-0020019-Bergman1] is that of the gene networks regulating ion channel expression in neurons. Neurons contain an array of voltage-dependent Na^+^ and K^+^ channels as well as numerous Cl^−^, Ca^2+^, and voltage-independent leak channels. The electrical properties of a single neuron are dependent, though not exclusively, upon the suite of ion channels expressed within that neuron. The properties of a neural network, which generates behaviour, are determined both by the intrinsic expression patterns of ion channels within neurons and the connectivity between neurons. The nervous system develops as an interaction between experience and genetically programmed events. One mechanism by which this interaction is achieved is ion channel compensation ([@pbio-0020019-Turrigiano1]); individual neurons can change their sensitivity to inputs by altering the relative proportion of ion channels, enabling them to maintain stable properties in the face of changing experience ([@pbio-0020019-Turrigiano2]; [@pbio-0020019-Brickley1]; [@pbio-0020019-Maclean1]; [@pbio-0020019-Niven1]) ([Figure 1](#pbio-0020019-g001){ref-type="fig"}B). Many studies of ion channel 'knockouts' show relatively little change in overall neuronal activity, although predictions based upon pharmacological blockade of the ion channels suggest there should be a more severe phenotypic change ([@pbio-0020019-Marder1]). Subsequent work has shown that the loss of an ion channel may often be compensated by a change in the expression of other ion channels. For example, the neurons upon which I work are Drosophila photoreceptors. In these neurons, loss of the one particular ion channel leads to compensatory changes in other ion channels linked to the activity of the neuron to restore the ability to process visual information ([@pbio-0020019-Niven1], [@pbio-0020019-Niven2]). However, these changes do not restore the original phenotype completely, and the compensated photoreceptors still show a reduced ability to process visual information. In many neurons, it appears that the intracellular Ca^2+^ concentration acts as an internal sensor of neural activity ([@pbio-0020019-Marder1]). Ca^2+^, along with other second messengers, may influence the expression of genes encoding ion channels, allowing their expression to be coupled to neural activity ([@pbio-0020019-Berridge1]) ([Figure 1](#pbio-0020019-g001){ref-type="fig"}B and [1](#pbio-0020019-g001){ref-type="fig"}C). Additionally, activity-independent mechanisms of ion channel compensation have been described in which the expression of one ion channel is linked to the expression of other opposing ion channels within a neuron ([@pbio-0020019-Maclean1]). These two systems of activity-dependent and activity-independent ion channel compensation bear a close resemblance to the gene network simulated by [@pbio-0020019-Bergman1] in which each gene regulates its own expression and that of other network genes. It is possible, therefore, that the networks of genes regulating ion channel expression may act as evolutionary buffers. The relationship between neural activity and the network of ion channel encoding genes may stabilise the neural activity in relation to both the genetic and environmental variation. The stabilisation of neural activity may have consequences for the generation of adaptive behaviour, which is constructed from neural activity. It is possible that ion channels could canalize the evolution of the nervous system by reducing behavioural variation and therefore removing the substrate on which natural selection may act. For example, changes in voltage-dependent Na^+^ channel properties (such as the activation voltage) may be compensated for by regulating the expression of other ion channels. 'Knockout' of one of these compensating ion channels may reveal the change in voltage-dependent Na^+^ channel properties, resulting in a shift in the output of the neuron. This hypothesis has several testable predictions. For example, 'knocking out' an ion channel should increase the variation in the activity of particular neurons among individuals in a population. This variation in neural activity may produce an effect on the behaviour of the whole organism. Studying canalization in ion channel gene networks may have significant advantages over studying developmental gene networks because it is relatively straightforward to measure the amounts of ion channels expressed in single identified neurons, to alter the expression of individual ion channels, and to relate these alterations to behaviour. I am currently pursuing the impact of ion channel compensation in Drosophila photoreceptors ([@pbio-0020019-Niven1], [@pbio-0020019-Niven2], [@pbio-0020019-Niven3]). In this system, changes in ion channel expression produce changes in the coding of visual information, which may lead to behavioural differences. The possible role of ion channel compensation in canalizing the evolution of the nervous system may have important implications not just for understanding this system, but also for understanding the contribution of ion channel compensation to the function of the nervous system and its evolution. Thanks to Simon Laughlin, Adrian Friday, and Mike Bate for helpful comments and discussions on a previous version of this manuscript. During this work JEN was supported by a Biotechnology and Biological Sciences Research Council grant to Simon B. Laughlin. Jeremy E. Niven is postdoctoral researcher working with Simon B. Laughlin in the Department of Zoology at the University of Cambridge in Cambridge, United Kingdom. E-mail: <jen22@cam.ac.uk> Hsp90 : heat shock protein 90

PubMed Central

2024-06-05T03:55:46.030401

2004-1-20

PMC314472

Introduction {#s1} ============ Repair of double-strand breaks (DSBs) by homologous recombination involves the search for homology to locate an intact donor sequence. The search is successful when the broken DNA molecule basepairs with the homologous template, termed synapsis, and forms strand invasion intermediates of recombination. In budding yeast and other higher eukaryotes, this process requires both the Rad51 strand exchange protein and the single-stranded DNA (ssDNA)-binding protein replication protein A (RPA) ([@pbio-0020021-Alani1]; [@pbio-0020021-Shinohara2]; [@pbio-0020021-Ogawa1]; [@pbio-0020021-Sung1]; [@pbio-0020021-Symington1]). RPA was first discovered through its essential role in SV40 DNA replication in vitro as a ssDNA-binding protein ([@pbio-0020021-Wold2]). The RPA complex forms a heterotrimer, which consists of three subunits of 70, 34, and 14 kDa, encoded by the *RFA1*, *RFA2*, and *RFA3* genes, respectively ([@pbio-0020021-Wold1]). Deletion of any of these genes leads to lethality in yeast ([@pbio-0020021-Heyer2]; [@pbio-0020021-Brill1]). The biological function of RPA was further demonstrated to be important in homologous recombination in Saccharomyces cerevisiae ([@pbio-0020021-Alani1]; [@pbio-0020021-Firmenich1]; [@pbio-0020021-Umezu2]) and in other aspects of DNA metabolism. Cells carrying a point mutation (K45E) in the largest subunit of RPA (*rfa1-t11*) are proficient for DNA replication, but their ability to perform mating-type (*MAT*) gene switching, single-strand annealing, and meiotic recombination is severely impaired ([@pbio-0020021-Umezu2]; [@pbio-0020021-Soustelle1]). Because RPA is essential for DNA replication, a great deal about its role in recombination has been learned from in vitro studies of Rad51-mediated strand exchange ([@pbio-0020021-Bianco1]; [@pbio-0020021-Symington1]). These studies have shown that RPA facilitates the formation of continuous Rad51 filaments on ssDNA by removing inhibitory secondary structures ([@pbio-0020021-Alani1]; [@pbio-0020021-Sugiyama2], [@pbio-0020021-Sugiyama3]). A similar requirement is seen in bacteria, where the ssDNA-binding protein SSB apparently plays an analogous role to allow the Rad51 homologue RecA to polymerize across regions that contain secondary structures ([@pbio-0020021-Shibata1]; [@pbio-0020021-West1]; [@pbio-0020021-Kowalczykowski1]; [@pbio-0020021-Kuzminov1]). Rad51 further displaces RPA, while RecA displaces SSB, leading to the filament that facilitates the search for homologous double-stranded DNA (dsDNA) sequences and then catalyzes strand invasion and the formation of a displaced single strand ([@pbio-0020021-Kowalczykowski2]; [@pbio-0020021-New1]; [@pbio-0020021-Eggler1]; [@pbio-0020021-Sugiyama1]). However, order-of-addition experiments have suggested that if RPA/SSB is added to ssDNA prior to Rad51/RecA, successful displacement will not occur because RPA/SSB has higher affinity for ssDNA, unless mediator proteins, such as Rad52 and Rad55/Rad57 in yeast and RecO/RecR in bacteria, are present ([@pbio-0020021-Umezu1]; [@pbio-0020021-New1]; [@pbio-0020021-Shinohara1]; [@pbio-0020021-Kuzminov1]; [@pbio-0020021-Sugiyama1]; [@pbio-0020021-Symington1]). But if Rad51/RecA has polymerized onto ssDNA first under conditions that prevent the formation of secondary structures, further addition of RPA/SSB will stimulate in vitro strand exchange in a species-specific manner ([@pbio-0020021-Heyer1]; [@pbio-0020021-Morrical1]; [@pbio-0020021-Sung1]; [@pbio-0020021-Sugiyama2]). Using the same in vitro system, [@pbio-0020021-Kantake1]) have examined the effects of the *rfa1-t11* (Rfa1-K45E) mutation on strand exchange with Rad51. Although *rfa1-t11* protein bound to ssDNA identically to wild-type and could stimulate strand exchange if Rad51 was preloaded onto ssDNA, the mutant protein exhibited delayed and less-efficient strand exchange if it was first bound to ssDNA, especially at higher concentrations, even in the presence of Rad52. This defect was explained by the slow displacement of *rfa1-t11* from ssDNA by Rad51. Recently, immunostaining experiments have been carried out in S. cerevisiae as well as in higher eukaryotes to investigate the association of RPA to DSBs following γ-irradiation and during meiosis. These studies have suggested that RPA and Rad51 form subnuclear foci at sites of ssDNA after irradiation and during meiotic recombination ([@pbio-0020021-Gasior1]; [@pbio-0020021-Golub1]; [@pbio-0020021-Raderschall1]) and that RPA is recruited to these sites prior to Rad51 ([@pbio-0020021-Golub1]; [@pbio-0020021-Gasior2]). In order to understand better how RPA is involved in DSB repair in vivo, we have looked at its function in *MAT* switching in yeast, which is the well-studied example of DSB-induced homologous recombination ([@pbio-0020021-Haber3]). *MAT* switching is initiated when HO endonuclease creates a site-specific DSB at the *MAT* locus, which is then repaired by gene conversion using one of the two heterochromatic donor sequences, *HML* or *HMR* ([@pbio-0020021-Paques1]; [@pbio-0020021-Haber2], [@pbio-0020021-Haber3]). By using a galactose-inducible HO endonuclease gene ([@pbio-0020021-Jensen1]), the induction of the DSB and its repair occur synchronously in a population of cells so that the kinetics of DSB repair and the appearance of intermediates of recombination can be followed by physical monitoring of the process via Southern blot and PCR assays ([@pbio-0020021-Haber1], [@pbio-0020021-Haber3], [@pbio-0020021-Haber4]). To learn more precisely about how RPA participates in homologous recombination in vivo, we have used chromatin immunoprecipitation (ChIP) assays ([@pbio-0020021-Dedon1]; [@pbio-0020021-Sugawara1]) to analyze the association of RPA and Rad51 to DNA as it undergoes *MAT* switching. At the same time, the fate of recombining DNA was analyzed by Southern blot and PCR techniques ([@pbio-0020021-White1]; [@pbio-0020021-Haber1]). The combination of these approaches has enabled us to visualize the kinetics and extent of RPA binding to a DSB and the homologous template during recombination. We report that the biological function of RPA is also required during the strand invasion step of recombination. *Rfa1-t11* mutant cells are not defective in Rad51 nucleoprotein filament assembly, as observed by ChIP, but are incapable of performing the strand exchange and thus the completion of DSB repair. Results {#s2} ======= Kinetics and Extent of RPA Binding at DSB Ends in the Absence of DNA Repair {#s2a} --------------------------------------------------------------------------- In wild-type yeast cells, a DSB created at the *MAT* locus can be repaired by gene conversion with one of the two donor sequences, *HML* or *HMR*, or the DSB can be left unrepaired in most cells by deleting these donors ([@pbio-0020021-Haber3]). In order to monitor RPA binding to DSB ends, we first performed ChIP analysis on strains in which both of the donor loci were deleted so that the DSB at *MAT* could not be repaired and 5′ to 3′ exonuclease activity would generate resected ssDNA unimpeded for many hours ([@pbio-0020021-Lee1]). In these strains, nearly complete cutting of *MAT* by the galactose-induced HO endonuclease occurred within 20 min after induction (see below). In wild-type cells, after HO induction, significant RPA binding to sequences close to the HO cleavage site was seen by ChIP ([Figure 1](#pbio-0020021-g001){ref-type="fig"}A and [1](#pbio-0020021-g001){ref-type="fig"}B), using a pair of primers (P1 and P2) that amplify sequences 189 bp to 483 bp distal to the HO cut ([Figure 1](#pbio-0020021-g001){ref-type="fig"}A). As shown in [Figure 2](#pbio-0020021-g002){ref-type="fig"}, RPA was recruited to DSB ends as soon as the DSB could be detected on a Southern blot (20 min after induction). The binding of RPA increased for about 2--3 h, until presumably all sequences near *MAT* had been rendered single-stranded ([@pbio-0020021-Frank-Vaillant1]) (see [Figure 1](#pbio-0020021-g001){ref-type="fig"}B). At later times, one detects RPA binding at increasing distances from the cleavage site, as these regions were rendered single-stranded by 5′ to 3′ exonuclease activity ([@pbio-0020021-Lee1]) (see [Figure 1](#pbio-0020021-g001){ref-type="fig"}C). ::: {#pbio-0020021-g001 .fig} Figure 1 ::: {.caption} ###### Recruitment of RPA to a DSB in the Absence of DNA Repair A strain deleted for donors (yXW1), thus incapable of repairing a DSB by gene conversion, was pregrown in YP--lactate medium, and 2% galactose was added to the culture to induce a DSB at *MAT*. DNA was extracted at intervals after HO cutting, to which polyclonal antibody against Rfa1 was applied to immunoprecipitate RPA-bound chromatin. Another set of DNA samples were taken at the same time for Southern blot analysis. \(A) Map of *MAT* showing the locations of the HO-cut site as well as the StyI restriction sites and the primers (P1 and P2), 189 bp to 483 bp distal to the DSB, used to PCR-amplify RPA-associated *MAT* DNA from the immunoprecipitated extract. Purified genomic DNA was digested with StyI, separated on a 1.4% native gel, and probed with a ^32^P-labeled *MAT* distal fragment to monitor the appearance of the HO-cut fragment (see Materials and Methods). The 1-h timepoint represents 1 h after galactose induction of the HO endonuclease. \(B) PCR-amplified RPA-bound *MAT* DNA in a wild-type strain (yXW1). As controls, primers to an independent locus, *ARG5,6* (see Materials and Methods), were used to amplify DNA from the immunoprecipitated chromatin. PCR samples were run on ethidium bromide-stained gels (reverse images are shown). Quantitated signals were graphed for the wild-type strain. IP represents ratio of the *MAT* IP signal to *ARG5,6* IP signal. Error bars show one standard deviation. \(C) RPA-bound chromatin was PCR-amplified from sites located proximal and distal to the DSB and then quantitated and graphed as described in (B). The DSB is shown at 0 bp. \(D) Effect of formaldehyde cross-linking on RPA binding to ssDNA. In both the noncross-linked samples and the cross-linked samples, 4 ng of single-stranded heterologous β-lactamase (*AMP*) gene DNA was added during the extract preparation step of ChIP. The amount of input genomic and heterologous DNA was measured by PCR primers specific to the *ARG5,6* locus and to the *AMP* sequence, respectively. RPA-associated *ARG5,6* and *AMP* DNA were analyzed from the IP samples. PCR samples were run on ethidium bromide-stained gels (reverse images are shown). :::  ::: ::: {#pbio-0020021-g002 .fig} Figure 2 ::: {.caption} ###### Timing of Recruitment of RPA versus Rad51 to the DSB An unrepairable DSB was created in the wild-type strain (yXW1), and closer timepoints were harvested at 20 min and 30 min after the HO cut. DNA samples extracted at each timepoint were split. One half was applied with antibody against Rfa1 to immunoprecipitate RPA-associated DNA, while the other half was applied with anti-Rad51 antibody to immunoprecipitate Rad51-bound chromatin. RPA- or Rad51-associated *MAT* DNA was PCR-amplified and run on ethidium bromide-stained gels (reverse images are shown). DNA signals were quantitated and graphed as described in [Figure 1](#pbio-0020021-g001){ref-type="fig"} for RPA ChIP. PCR-amplified *ARG5,6* signals from the input DNA were used as controls for quantitation and graphing for Rad51 ChIP (see Materials and Methods). :::  ::: In carrying out the ChIP measurements, we were aware that RPA is both abundant within the cell and binds strongly and cooperatively to ssDNA in vitro ([@pbio-0020021-Heyer1]). It was possible that some of the RPA binding we measured by ChIP could have arisen after the cells were broken and could be independent of formaldehyde cross-linking. Indeed, in the absence of cross-linking, we found that there was substantial binding of RPA to the HO-cut *MAT* locus, which was resistant to both addition of 2 mg of ssDNA (equivalent to a 1,000-fold genome excess) at the time of cell breakage and washing with 4.7 M NaCl, although it greatly reduced background binding (data not shown). However, in formaldehyde cross-linked samples, there was no such adventitious binding of RPA to ssDNA regions, apparently because the formaldehyde-treated proteins are no longer able to bind. This was shown directly by adding 4 ng of purified single-stranded β-lactamase (*AMP*) gene DNA from plasmid pBR322 at the time of cell breakage. Whereas there was substantial ChIP of the *AMP* sequences in noncross-linked samples, there was almost no signal in cells that had first been treated with formaldehyde (see [Figure 1](#pbio-0020021-g001){ref-type="fig"}D). RPA Binding Precedes the Binding of the Strand Exchange Protein Rad51 {#s2b} --------------------------------------------------------------------- In vitro studies of the early steps of recombination have suggested that in order to make a continuous and functional nucleoprotein filament, RPA must bind before Rad51 to ssDNA to remove inhibitory secondary structures ([@pbio-0020021-Sugiyama2]). Indirect immunofluorescence experiments in S. cerevisiae have also suggested that RPA assembles before Rad51 at DSBs after γ-irradiation ([@pbio-0020021-Gasior2]). Therefore, the timing of recruitment of both RPA and Rad51 proteins to a DSB in vivo were compared by ChIP. RPA was detected at *MAT* 20 min after the HO cut, while Rad51 binding was not observed until the 30 min timepoint ([Figure 2](#pbio-0020021-g002){ref-type="fig"}). Similar results were obtained in strains that are able to carry out gene conversion (see [Figure 4](#pbio-0020021-g004){ref-type="fig"}). These observations strongly support the idea that RPA binding to HO-cut DNA precedes that of Rad51. ::: {#pbio-0020021-g004 .fig} Figure 4 ::: {.caption} ###### Localization of RPA and Rad51 to *HML* and *MAT* during DSB-Induced Gene Conversion A strain carrying an *HML*α donor (yXW2), thus able to repair the DSB at *MAT* by gene conversion, was treated with 2% galactose to induce HO endonuclease and then with 2% glucose after 1 h to repress further HO expression. DNA extracted at intervals after HO cutting was split. One half was applied with antibody against Rfa1 to immunoprecipitate RPA-associated DNA, while the other half was applied with anti-Rad51 antibody to immunoprecipitate Rad51-bound chromatin. Another set of DNA samples were taken at the same time for Southern blot analysis. \(A) Diagram of *MAT* and *HML* showing the locations of the primers, 189 bp to 483 bp distal to the DSB at *MAT* (P1 and P2) and 189 bp to 467 bp from the uncleaved HO recognition site at *HML* (P1 and P3), used to PCR-amplify RPA- and Rad51-associated *MAT* and *HML* DNA from the immunoprecipitated extract. \(B) Purified genomic DNA was digested with StyI, separated on a 1.4% native gel, and probed with a ^32^P-labeled *MAT* distal fragment to monitor the appearance of the HO-cut fragment and the repaired product Yα (see [Figure 1](#pbio-0020021-g001){ref-type="fig"}A; see Materials and Methods). The arrowhead indicates the switched product Yα. RPA- and Rad51-bound *MAT* and *HML* DNA were PCR-amplified with primers P1 and P2 and with P1 and P3, respectively. Samples were run on ethidium bromide-stained gels. (C and D) Reverse images are shown for RPA ChIP (C) and Rad51 ChIP (D). DNA signals were quantitated and graphed as described in [Figure 2](#pbio-0020021-g002){ref-type="fig"}. Error bars show one standard deviation. :::  ::: In Vivo Competition between RPA and Rad51 for ssDNA {#s2c} --------------------------------------------------- Studies of RPA in vitro would suggest that the amount of RPA bound to ssDNA may be limited by its displacement by Rad51, through the help of Rad52, and the Rad55/57 heterodimer ([@pbio-0020021-Sung2], [@pbio-0020021-Sung3]; [@pbio-0020021-New1]; [@pbio-0020021-Shinohara1]; [@pbio-0020021-Sugiyama1]). To test this idea, we deleted *RAD51* and measured RPA binding at *MAT*. The extent of RPA binding was approximately 5- to 6-fold higher in the *rad51*Δ strain, consistent with this expectation ([Figure 3](#pbio-0020021-g003){ref-type="fig"}). A similar result was found in a *rad52*Δ strain ([Figure 3](#pbio-0020021-g003){ref-type="fig"}), supporting the hypothesis that the displacement of RPA by Rad51 depends on Rad52, which acts as a mediator between these two ssDNA-binding proteins ([@pbio-0020021-Sung2]; [@pbio-0020021-New1]; [@pbio-0020021-Song1]; [@pbio-0020021-Sugiyama1]; [@pbio-0020021-Sugawara1]). ::: {#pbio-0020021-g003 .fig} Figure 3 ::: {.caption} ###### Effect of *rad51*Δ and *rad52*Δ on the Extent of RPA Binding to an Unrepairable DSB An unrepairable DSB was created in wild-type (yXW1), *rad51*Δ (ySL306), and *rad52*Δ (ySL177) strains and RPA-bound chromatin was immunoprecipitated using anti-Rfa1 antibody. PCR-amplified DNA from the *MAT* locus was run on ethidium bromide-stained gels (reverse images are shown). DNA signals were quantitated and graphed as described in [Figure 1](#pbio-0020021-g001){ref-type="fig"}. Error bars show one standard deviation. :::  ::: RPA Is Recruited to Both the Donor and the Recipient Sequences during Gene Conversion {#s2d} ------------------------------------------------------------------------------------- We then examined RPA in a strain in which the DSB at *MAT* **a** could be repaired by gene conversion, using *HML*α as the donor ([Figure 4](#pbio-0020021-g004){ref-type="fig"}A). As soon as the DSB was visible, an increase in RPA binding was seen (Figure [4](#pbio-0020021-g004){ref-type="fig"}B and [4](#pbio-0020021-g004){ref-type="fig"}C). RPA binding increased for about 1 h and then decreased nearly to the baseline level about the time that *MAT* switching was completed (Figure [4](#pbio-0020021-g004){ref-type="fig"}B and [4](#pbio-0020021-g004){ref-type="fig"}C). Importantly, RPA also appeared to become associated with the donor locus. This was detected by ChIP using the same primer P1 located in the *Z* region shared by *MAT* and *HML* and an *HML* sequence-specific primer P3 ([Figure 4](#pbio-0020021-g004){ref-type="fig"}A). Whereas RPA could be found associated with *MAT* 20 min after HO induction, its association with *HML* was seen only after 1 h ([Figure 4](#pbio-0020021-g004){ref-type="fig"}C). The association of RPA with *HML* came at about the same time as we saw synapsis between *HML* and *MAT* as revealed by ChIP with anti-Rad51 antibody ([Figure 4](#pbio-0020021-g004){ref-type="fig"}D; also [@pbio-0020021-Sugawara1]). The extent of RPA binding to *HML* was substantially less than seen at *MAT* ([Figure 4](#pbio-0020021-g004){ref-type="fig"}C), where ssDNA may extend further than the 320 bp of homology between *MAT* and *HML*; these more distal ssDNA sequences would not be involved directly in recombination. The lower amount of RPA binding at the donor locus may also arise from a lower concentration of RPA that is needed at the sites of strand invasion or a transient presence of RPA at those loci. But the fact that cross-linked RPA can immunoprecipitate the donor locus might indicate that RPA is recruited onto the single-stranded D-loop that is created by strand invasion. This would be consistent with in vitro studies of Rad51-mediated strand exchange that suggest that strand invasion per se can occur without RPA, but that the heteroduplex DNA is unstable unless the displaced strand is bound by RPA ([@pbio-0020021-Eggler1]). A similar requirement for SSB was suggested in RecA-mediated strand invasion ([@pbio-0020021-Lavery1]). We cannot entirely rule out the possibility that the apparent association of RPA with *HML* resulted from the cross-linking of synapsed *MAT* and *HML* sequences directly or through Rad51-containing cross-links and where RPA was bound to ssDNA sequences distal to the 320-bp homology shared by *MAT* and its donor. Further evidence of a role for RPA in synapsis will be presented below. To understand better the dynamics between RPA and Rad51 during gene conversion, we also examined Rad51 recruitment to *MAT* and *HML* relative to that of RPA ([Figure 4](#pbio-0020021-g004){ref-type="fig"}D). Rad51 was only detected at *MAT* 30 min postinduction (compared to 20 min for RPA). Rad51 binding increased for about 1 h and then remained bound for several hours. As reported previously ([@pbio-0020021-Sugawara1]), Rad51 showed a delayed association with the donor *HML*, reflecting the time required to form a functional filament and to search the genome for homologous sequences. These observations provide evidence of the time at which synapsis between *MAT* and *HML* is achieved. Here, too, Rad51 association with the donor remained for several hours, beyond the time when *MAT* switching is completed. *rfa1-t11* Mutant Cells Are Defective in the Synapsis Step of Gene Conversion {#s2e} ----------------------------------------------------------------------------- To learn more about RPA function during recombination, we investigated the behavior of the *rfa1-t11* (K45E) mutation in the largest subunit of RPA. This mutation has little effect on DNA replication per se, but severely impairs both gene conversion (*MAT* switching) and single-strand annealing pathways of homologous recombination ([@pbio-0020021-Umezu2]) (also [Figure 6](#pbio-0020021-g006){ref-type="fig"}A). Cells containing this mutation displayed hyperresection at meiotic DSB ends and defects in the repair of these DSBs ([@pbio-0020021-Soustelle1]). In vitro biochemical studies have shown that *rfa1-t11* is displaced from ssDNA by Rad51 more slowly than wild-type RPA, and as a consequence, Rad51-mediated strand exchange is inhibited when the ssDNA is complexed with the mutant RPA heterotrimer ([@pbio-0020021-Kantake1]). Here, we examined binding of Rfa1-K45E in vivo by ChIP and also its effect on Rad51 localization to an HO-induced DSB, using the same antibodies as against wild-type RPA. ::: {#pbio-0020021-g006 .fig} Figure 6 ::: {.caption} ###### *rfa1-t11* Was Not Able to Associate with the Donor Sequence during Gene Conversion The wild-type strain carrying the *HML*α donor (yXW2) and an isogenic strain carrying the *rfa1-t11* mutation (yXW3) were treated with 2% galactose to induce HO endonuclease and then with 2% glucose after 1 h to repress further HO expression. DNA extracted at intervals after HO cutting was split. One half was applied with antibody against Rfa1 to immunoprecipitate RPA-associated DNA, while the other half was applied with anti-Rad51 antibody to immunoprecipitate Rad51-bound chromatin. Another set of DNA samples was taken at the same time for Southern blot analysis. \(A) Purified genomic DNA was digested with StyI, separated on a 1.4% native gel, and probed with a ^32^P-labeled *MAT* distal fragment to monitor the appearance of the HO-cut fragment and the repaired product Yα (see [Figure 1](#pbio-0020021-g001){ref-type="fig"}A; see Materials and Methods). Arrowheads indicate the switched product Yα. \(B) RPA-bound *MAT* and *HML* DNA was PCR-amplified with primers P1 and P2 and with P1 and P3, respectively (see [Figure 4](#pbio-0020021-g004){ref-type="fig"}A). Samples were run on ethidium bromide-stained gels (reverse images are shown). DNA signals were quantitated and graphed as described in [Figure 1](#pbio-0020021-g001){ref-type="fig"}. :::  ::: In a strain lacking *HML* and *HMR*, Rfa1-K45E binding was nearly identical to wild-type, both in a *RAD51* and in a *rad51*Γ background ([Figure 5](#pbio-0020021-g005){ref-type="fig"}A). Moreover, binding of Rad51 to ssDNA at the HO-cut *MAT* locus was also comparable to that observed in wild-type cells ([Figure 5](#pbio-0020021-g005){ref-type="fig"}B). Thus, the Rfa1-t11 protein is neither impaired in loading onto ssDNA, nor does it affect the loading of Rad51 in vivo. ::: {#pbio-0020021-g005 .fig} Figure 5 ::: {.caption} ###### *rfa1-t11* Mutation Does Not Affect the Recruitment of Itself or Rad51 to an Unrepairable DSB \(A) An unrepairable DSB was created in wild-type (yXW1), *rfa1-t11* (ySL31), *rad51*Δ (ySL306), and *rfa1-t11 rad51*Δ (ySL351) strains, and half of the DNA sample was immunoprecipitated with anti-Rfa1 antibody to extract *rfa1-t11*-bound chromatin. \(B) For wild-type (yXW1) and *rfa1-t11* (ySL31) strains, the other half of the DNA sample was applied with anti-Rad51 antibody to extract Rad51-associated chromatin. PCR-amplified DNA from the *MAT* locus was run on ethidium bromide-stained gels (reverse images are shown). DNA signals were quantitated and graphed as described in [Figure 2](#pbio-0020021-g002){ref-type="fig"}. :::  ::: We then examined the effect of *rfa1-t11* during HO-induced switching of *MAT* **a** to *MAT*α, using *HML*α as the donor. As shown previously ([@pbio-0020021-Umezu2]), *rfa1-t11* strongly impaired *MAT* switching, with only 15% product evident after 5 h ([Figure 6](#pbio-0020021-g006){ref-type="fig"}A). RPA bound normally to the *MAT* locus, but unlike what occurs in wild-type strains, its binding remained undiminished at later times ([Figure 6](#pbio-0020021-g006){ref-type="fig"}B). Moreover, there was no increased association of RPA with *HML* over background levels ([Figure 6](#pbio-0020021-g006){ref-type="fig"}B). When we examined Rad51 binding in this mutant, we found that Rad51 immunoprecipitated with *MAT* DNA, but not with *HML* ([Figure 7](#pbio-0020021-g007){ref-type="fig"}A). In support of this important finding, we also used a PCR assay to show that *rfa1-t11* prevented the appearance of newly synthesized DNA using the 3′ end of the invading strand as a primer ([Figure 7](#pbio-0020021-g007){ref-type="fig"}B). In this assay, a primer specific for the Yα region in *HML* (pA) can only amplify a strand invasion product with a primer specific for *MAT*-distal sequences (pB) if the 3′ end of the strand-invading DNA is extended by DNA polymerase at least 35 nucleotides ([@pbio-0020021-White1]) ([Figure 7](#pbio-0020021-g007){ref-type="fig"}B). These data strongly raise the possibility that RPA is required during the process of strand invasion and synapsis and not merely to facilitate formation of a Rad51 filament, as the binding of Rad51 to ssDNA at *MAT* seems to be normal in both kinetics and extent. ::: {#pbio-0020021-g007 .fig} Figure 7 ::: {.caption} ###### *rfa1-t11* Mutants Are Defective in the Strand Invasion Step of Gene Conversion \(A) One half of the DNA extract collected from a typical timecourse experiment as described in [Figure 6](#pbio-0020021-g006){ref-type="fig"} was applied with anti-Rad51 antibody to immunoprecipitate Rad51-bound chromatin. Primers P1 and P2 and P1 and P3 were used to PCR-amplify Rad51-bound *MAT* and *HML* DNA, respectively (see [Figure 4](#pbio-0020021-g004){ref-type="fig"}A). Samples were run on ethidium bromide-stained gels (reverse images are shown). DNA signals were quantitated and graphed as described in [Figure 2](#pbio-0020021-g002){ref-type="fig"}. \(B) Input DNA was used to PCR-amplify strand invasion product using a unique primer distal to *MAT* (pB) and a primer within the Yα sequence from *HML* (pA) ([@pbio-0020021-White1]). PCR-amplified *ARG5,6* signals from the input DNA were used as loading control. :::  ::: Discussion {#s3} ========== ChIP analysis provides a powerful tool for studying in vivo protein--DNA and protein--protein interactions. Using ChIP and related assays, we have demonstrated important roles of RPA during homologous recombination in vivo that could not have been known with certainty from in vitro studies. RPA is recruited to the DSB ends as soon as the DSB is detected on a Southern blot, and its binding precedes that of Rad51 (see Figures [2](#pbio-0020021-g002){ref-type="fig"} and [4](#pbio-0020021-g004){ref-type="fig"}), which supports the idea that RPA is required to remove inhibitory secondary structures on ssDNA for Rad51 to polymerize across these regions ([@pbio-0020021-Sugiyama2], [@pbio-0020021-Sugiyama3]). This observation is also consistent with in vivo immunofluorescent staining results, suggesting that RPA foci appear earlier than Rad51 foci after irradiation ([@pbio-0020021-Golub1]; [@pbio-0020021-Gasior2]). Rad51 apparently displaces RPA from ssDNA, with the help of Rad52 (see [Figure 3](#pbio-0020021-g003){ref-type="fig"}) and perhaps the Rad55/Rad57 auxiliary proteins. We note that our results are different from those reported by [@pbio-0020021-Wolner1]), who observed initial binding of RPA only after 45 min, whereas Rad51 was detected 25 min earlier, although it is not clear whether there is a statistically significant increase in Rad51 binding at the earliest time. In that assay, *RFA1* was tagged with 13 Myc epitope tags, which may have altered its behavior. We believe our results are consistent with the fact that RPA has a higher-affinity constant for ssDNA and is present in much greater abundance in the cell ([@pbio-0020021-Heyer1]; [@pbio-0020021-Mazin1]; [@pbio-0020021-Sugawara1]). We noticed that when RPA ChIP was carried out in donorless strains as well as in *rfa1-t11* strains that carry the donor loci, there was a continued presence of some RPA near the ends of a DSB. This may occur for several reasons. First, it is likely that the formation and maintenance of the Rad51 filament are a dynamic process, with subunits coming off the end and perhaps being replaced by RPA before being in turn replaced by Rad51. Second, the Rad51 nucleoprotein filament may not be, in vivo, a fully continuous structure, given that there are only about 3,500 monomers of Rad51 in the cell and that Rad51 binding is not highly cooperative ([@pbio-0020021-Mazin1]; [@pbio-0020021-Sugawara1]). Finally, the very ends of the DSB can religate and be recleaved by HO in a cycle that lasts several hours in the absence of donor sequences (and hence in the absence of homologous recombination) to repair the DSB ([@pbio-0020021-Frank-Vaillant1]). Thus, a fraction of molecules will be newly generated and will show RPA binding before Rad51, as we saw for the initial DSB. The ChIP analysis presented here has shown that RPA is required for homologous recombination even after Rad51 has bound to ssDNA. First, RPA can immunoprecipitate donor sequences, the timing of which coincides with the loading of Rad51 at *HML* (see [Figure 4](#pbio-0020021-g004){ref-type="fig"}). Second, the replication-proficient but recombination-deficient mutant of the largest subunit of RPA (*rfa1-t11*) is able to allow Rad51 to bind to ssDNA, but is incapable of forming normal levels of strand invasion and primer extension products (see Figures [6](#pbio-0020021-g006){ref-type="fig"} and [7](#pbio-0020021-g007){ref-type="fig"}). We offer two possible explanations for this unexpected finding. First, whereas Rad51 can bind to ssDNA in *rfa1-t11*, it may not establish a functional filament capable of carrying out a search for homology and strand invasion, even though the association of Rad51 with ssDNA appears to be normal. But the defect in cells with *rfa1-t11* seems different from that seen in cells lacking Rad55 ([@pbio-0020021-Sugawara1]), where there was delayed and less-extensive binding of Rad51 to ssDNA; moreover, although Rad51 eventually bound, it was unable to catalyze synapsis between *MAT* ssDNA and *HML*. In *rad55*Γ cells, it is likely that the Rad51 filament is discontinuous and unable to function. However, with *rfa1-t11*, the loading of Rad51 onto ssDNA appears to be identical to that seen in wild-type cells (see Figures [5](#pbio-0020021-g005){ref-type="fig"}B and [7](#pbio-0020021-g007){ref-type="fig"}A). Alternatively, in *rfa1-t11* cells, the filament may indeed be functional, but RPA is needed to stabilize the strand invasion intermediate and *rfa1-t11* is unable to carry this out. RPA may be required to bind to the displaced D-loop, to prevent rapid reversal of the process, which is implicated by in vitro studies of strand exchange ([@pbio-0020021-Eggler1]). In that study, extensive heteroduplex could be formed without RPA, as revealed by psoralen cross-linking of joint molecule DNA before removal of Rad51 by deproteinization, but without cross-linking, the deproteinized joint molecule DNA fell apart into the original single-stranded and double-stranded substrates very quickly. An analogous role for SSB has been suggested in RecA-mediated strand invasion ([@pbio-0020021-Lavery1]), in which SSB prevents the reversal of DNA strand exchange by removing the displaced single strand. It is possible that the Rfa1-K45E mutation renders the mutant RPA complex unable to bind to the displaced ssDNA at *HML* and thus unable to carry out strand exchange, while binding to *MAT* ssDNA that has a free 3′ end tail is not affected. Both in vitro ([@pbio-0020021-Kantake1]) and in vivo analyses showed that *rfa1-t11* was able to bind to ssDNA very similarly to wild-type, but our in vivo data did not see any significant impairment of its Rad52-mediated displacement by Rad51. It should be noted that the inhibition of Rad51-mediated strand exchange by *rfa1-t11* in vitro was carried out with saturating amounts of Rad51 (whereas the amount of Rad51 in the cell is quite limited) and that the inhibition of Rad51-mediated strand exchange was impaired primarily when RPA was present in excess ([@pbio-0020021-Kantake1]). How these conditions relate to those prevailing in vivo remains unknown. In this regard, it is also noteworthy that in vitro studies did not see any impairment of single-strand annealing ([@pbio-0020021-Kantake1]), whereas in vivo, single-strand annealing is nearly eliminated in *rfa1-t11* strains ([@pbio-0020021-Umezu2]). Further comparisons of in vitro and in vivo data will be valuable in understanding how the more complex environment within the cell affects processes of recombination. Materials and Methods {#s4} ===================== {#s4a} ### Strains {#s4a1} Donorless strains are isogenic derivatives of JKM139, which has the genotype of *ho*Δ *hml*Δ::*ADE1 MAT* **a** *hmr*Δ::*ADE1 ura3--52 leu2--3,112 trp1*::*hisG lys5 ade1--100 ade3*::*GAL*::*HO*. The wild-type strain yXW1 was constructed by transforming JKM139 with pGI4 (*bar1*::*ADE3*) ([@pbio-0020021-Wach1]). ySL83 contains *yku80*Δ::*KAN* and *bar1*::*TRP1*. ySL306 and ySL177 contain *rad51*Δ::*URA3* and *rad52*Δ::*TRP1*, respectively ([@pbio-0020021-Lee1], 2001). ySL31 has the point mutation (K45E) in the largest subunit of RPA ([@pbio-0020021-Lee1]), and ySL351 was derived from ySL31 and contains *rad51*Δ::*LEU2.* Strains capable of undergoing DSB-induced gene conversion were derived from OAy470 ([@pbio-0020021-Aparicio1]), which has the genotype of *ho*Δ *MAT* **a** *ura3--1 trp1--1 leu2--3,112 his3--11,15 ade2--1 can1--100 bar1*::*hisG*. A galactose-inducible *GAL*::*HO* gene was integrated at *ADE3* of OAy470 using YIPade3HO constructed by L. L. Sandell ([@pbio-0020021-Sandell1]) to obtain the wild-type strain yXW2. yXW3 is an isogenic derivative of yXW2, into which the point mutation (K45E) of Rfa1 was introduced by integration and excision of a YIp5 (*URA3*-containing) plasmid ([@pbio-0020021-Lee1]). ### DNA analysis {#s4a2} When cells were harvested for ChIP at intervals after induction of HO (see below), a second set of DNA samples were collected for Southern blot analysis as described before ([@pbio-0020021-White1]). The strand invasion/primer extension assay in [Figure 7](#pbio-0020021-g007){ref-type="fig"}B was previously described ([@pbio-0020021-White1]). The primers used were 5′-GCAGCACGGAATATGGGACT-3′ (pA) and 5′-ATGTGAACCGCATGGGCAGT-3′ (pB). ### ChIP {#s4a3} ChIP was carried out as described previously with minor modifications ([@pbio-0020021-Dedon1]; [@pbio-0020021-Sugawara1]). Cells were pregrown to a density between 5 × 10^6^ and 1 × 10^7^ cells/ml in YP--lactate medium and HO endonuclease was induced by addition of 2% galactose. Strains undergoing DSB-induced gene conversion were treated with 2% glucose after 1 h to repress further cutting by HO. Proteins were cross-linked by addition of 1% (final concentration) formaldehyde to 45 ml of culture for 10 min, followed by quenching with 125 mM glycine (final concentration) for 5 min. Cells were lysed with glass beads, and the extracts were sonicated to shear the DNA to an average size of 0.5 kb. Extracts were then divided into IP and input samples (12:1 ratio). IP samples were split. Half of the extract was incubated with polyclonal anti-Rfa1 antibody (kindly provided by S. Brill) for 1 h at 4°C and bound to protein G--agarose beads for 1 h at 4°C. In the ChIP experiments described in [Figure 2](#pbio-0020021-g002){ref-type="fig"}, [Figure 4](#pbio-0020021-g004){ref-type="fig"}D, [Figure 5](#pbio-0020021-g005){ref-type="fig"}B, [Figure 6](#pbio-0020021-g006){ref-type="fig"}, and [Figure 7](#pbio-0020021-g007){ref-type="fig"}, the other half of the extract was incubated with affinity-purified anti-Rad51 antibody (provided by P. Sung) or unpurified antibody (provided by A. Shinohara) for 1 h at 4°C and bound to protein A--agarose beads for 1 h at 4°C. The protein-bound beads were carried through a series of washes, followed by elution of the proteins and reversal of cross-linking (6 h at 65°C). Samples were treated with proteinase K followed by phenol extraction and ethanol precipitation. In the control experiments described in [Figure 1](#pbio-0020021-g001){ref-type="fig"}D, 4 ng of purified single-stranded β-lactamase (*AMP*) gene DNA from plasmid pBR322 was added at the time of cell breakage. IP and input samples were further subject to PCR to test the presence of the *AMP* sequences. ### PCR amplification {#s4a4} The *MAT*-specific primers were 5′-TCCCCATCGTCTTGCTCT-3′ (P1) and 5′-GCATGGGCAGTTTACCTTTAC-3′ (P2), which amplifies a PCR product of 293 bp. The *HML*-specific primers were 5′-TCCCCATCGTCTTGCTCT-3′ (P1) and 5′-CCCAAGGCTTAGTATACACATCC-3′ (P3), which amplifies a PCR product of 280 bp. Primers used for the amplification of the sites proximal to the DSB (see [Figure 1](#pbio-0020021-g001){ref-type="fig"}C) were −29.8 kb, 5′-TCGTCGTCGCCATCATTTTC-3′ and 5′-GCCCAAGTTTGAGAGAGGTTGC-3′; −16.6 kb, 5′-CGTCTTCTCAGCGAACAACAGC-3′ and 5′-GCAATAACCCACGGAAACACTG-3′; −9.5 kb, 5′-TCAGGGTCTGGTGGAAGGAATG-3′ and 5′-CAAAGGTGGCAGTTGTTGAACC-3′; −5.3 kb, 5′-ATTGCGACAAGGCTTCACCC-3′ and 5′-CACATCACAGGTTTATTGGTTCCC-3′; −3.6 kb, 5′-ATTCTGCCATTCAGGGACAGCG-3′ and 5′-CGTGGGAAAAGTAATCCGATGC-3′; −1.6 kb, 5′-ATGTCCTGACTTCTTTTGACGAGG-3′ and 5′-ACGACCTATTTGTAACCGCACG-3′; and −0.2 kb, 5′-AAAGAAGAAGTTGCAAAGAAATGTGG-3′ and 5′-TGTTGCGGAAAGCTGAAACTAAAAG-3′. Oligos used for the sites distal to the DSB were 0.2 kb, 5′-CCTGGTTTTGGTTTTGTAGAGTGG-3′ and 5′-GAGCAAGACGATGGGGAGTTTC-3′; 2.1 kb, 5′-GCCTCTATGTCCCCATCTTGTCTC-3′ and 5′-GTGTTCCCGATTCAGTTTGACG-3′; 3.1 kb, 5′-TAACCAGCAATACCAAGACAGCAC-3′ and 5′-TTTTACCTACCGCACCTTCTAAGC-3′; 5.7 kb, 5′-CCAAGGAACTAATGATCTAAGCACA-3′ and 5′-ACCAGCAGTAATAAGTCGTCCTGA-3′; and 9.5 kb, 5′-TGGATCATGGACAAGGTCCTAC-3′ and 5′-GGCGAAAACAATGGCACTCT-3′. These PCR primers gave products of about 300 bp. Primers specific for the *ARG5,6* locus were either 5′-AGAAAGGGGGTATTATCAATGGCTC-3′ and 5′-AGGAAAATCACGGCGCAAAA-3′, which amplifies a PCR product of 533 bp, or 5′-CAAGGATCCAGCAAAGTTGGGTGAAGTATGGTA-3′ and 5′-GAAGGATCCAAATTTGTCTAGTGTGGGAACG-3′, which amplifies a PCR product of 381 bp. Normalization using these two different pairs of primers has been shown not to affect the final quantification results. Primers used for the amplification of the *AMP* sequences (see [Figure 1](#pbio-0020021-g001){ref-type="fig"}D) were 5′-GAAGACGAAAGGGCCTCGTG-3′ and 5′-GCTGCAGGCATCGTGGTGTC-3′, which amplifies a PCR product of 750 bp. All PCR assays were accompanied by reactions using dilutions of the 0-h input sample to assess the linearity of the PCR signal and to create a calibration curve, as described before ([@pbio-0020021-Sugawara1]). Samples were run on ethidium bromide-stained agarose gels (1.4%) and quantitated using an Innotech Alphaimager™ and Quantity One software™ (BioRad, Hercules, California, United States), which was also used to correct for minor deviations from a linear response in signal. Quantification and graphing were carried out as described previously with minor changes ([@pbio-0020021-Sugawara1]). For RPA ChIP analysis, all IP samples were first normalized to IP signals from an independent locus (*ARG5,6*) on chromosome V in a multiplex experiment, by using *ARG5,6* and *MAT* or *HML* primers in the same PCRs. This was accomplished by dividing each *MAT* or *HML* IP signal by the corresponding *ARG5,6* IP signal to correct for differing amounts of chromatin collected at each timepoint. Then *MAT* or *HML* IP signals at later timepoints were normalized and graphed to the 0-h IP signal to measure the relative increase. For Rad51 ChIP analysis, quantification and graphing were carried out as described before ([@pbio-0020021-Sugawara1]), in which all IP samples were normalized to the *ARG5,6* input signals at the respective time points. Graphing represents the average of at least three independent ChIP timecourse experiments for each strain. Supporting Information {#s5} ====================== Accession Numbers {#s5a} ----------------- The *Saccharomyces* Genome Database (<http://www.yeastgenome.org/>) ID accession numbers for the entities discussed in this paper are *ARG5,6* (S0000871), *HML* (L0000791), *HMR* (L0000792), *MAT* (L0001031), Rad51 (S0000897), Rad52 (S0004494), Rad55 (S0002483), Rad57 (S0002411), *RFA1* (S0000065), *RFA2* (S0005256), and *RFA3* (S0003709). We thank Neal Sugawara for technical support for ChIP analysis. We thank Neal Sugawara, Grzegorz Ira, and André Walther for providing valuable suggestions. We are grateful for the insightful comments of Richard Kolodner and Steve West concerning the manuscript. We thank Steven Brill, Patrick Sung, and Akira Shinohara for gifts of anti-Rfa1 antibody and anti-Rad51 antibodies. This work was supported by National Institutes of Health grants GM20056 and GM61766. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** XW and JEH conceived and designed the experiments. XW performed the experiments. XW and JEH analyzed the data. XW and JEH wrote the paper. Academic Editor: Steven West, Cancer Research UK ChIP : chromatin immunoprecipitation DSB : double-strand break dsDNA : double-stranded DNA IP : immunoprecipitation *MAT* : mating type RPA : replication protein A SSB : single-stranded binding protein ssDNA : single-stranded DNA

PubMed Central

2024-06-05T03:55:46.031640

2004-1-20

PMC314473

One of the cornerstones of immune system function is movement. When word spreads that a virus has entered the body, chemical signals tell lymphocytes to proliferate and travel to the site of infection. Efforts to combat HIV have focused on understanding how the virus disrupts this immune response in the hopes of developing drugs to block its replication as well as vaccines to control the virus itself. Toward this end, scientists are investigating how each of the virus\'s nine genes---which all appear to have multiple functions---contribute to HIV infection. When HIV infects a cell, viral enzymes copy its RNA genes into DNA, which can then invade the infected cell\'s chromosomes. The viral DNA might lay dormant or it might use the cell to reproduce more viruses, which go on to infect other cells. The course of infection is determined by interactions between circulating T cells and antigen-presenting cells (cells that present evidence of infection), like macrophages, which may unwittingly aid the virus by transferring it to the T cells. Macrophages, for example, produce proteins that tell T cells to come check out an infection. A viral protein called Nef sparked intensive research after observations that patients with a rare strain of HIV lacking Nef took a very long time to develop AIDS symptoms. Nef has been linked to molecules involved in macrophage- and other antigen-signaling pathways and may use the molecules to appropriate these pathways for its own ends---enhancing virulence by facilitating viral replication. How Nef does this is not entirely clear. Now Jacek Skowronski and his colleagues at Cold Spring Harbor Laboratory in New York have identified the key molecules that Nef enlists to coopt the signaling machinery of immune cells. To understand how this might happen, biochemically speaking, Skowronski\'s lab first needed to determine which molecules Nef associates with. An adaptor protein, Nef does not directly catalyze reactions, but binds to enzymes that do. The researchers identified two proteins, DOCK2 and ELMO1, that form a complex with Nef. DOCK2 regulates enzymes, called Rac1 and Rac2, that are required for normal lymphocyte migration and antigen-specific responses. ELMO1 has also been shown to help DOCK2 activate Rac. Because DOCK2 activates Rac as part of two different signaling pathways---one activated by the T cell receptor, which mediates T cell activation, and one by a chemokine receptor, which controls T cell migration---the researchers investigated whether Nef could affect these important pathways by modulating Rac activity. They found that Nef in fact activates Rac by binding to the DOCK2--ELMO1 complex. And they went on to show that HIV uses these components of the chemokine receptor pathway to disrupt T cell migration. To generate an effective immune response, it is crucial that T cells travel to sites within lymphatic tissues where they interact with other lymphocytes. By inhibiting T cell migration, the researchers propose, Nef prevents these critical interactions, thereby providing a mechanism for stifling the immune response. These results, the authors argue, provide the biochemical evidence that Nef targets a protein "switch" that can interfere with important aspects of T cell function. In this way, Nef subverts the immune response pathways controlled by receptors on the surface of T cells to effectively disarm the immune system and turn T cells into viral replication factories. Understanding how Nef interacts with these proteins to spread infection could lay the foundation for valuable new therapies aimed at inhibiting and arresting HIV infection by blocking Nef-mediated effects. ::: {#pbio-0020023-g001 .fig} ::: {.caption} ###### Pathway by which Nef disrupts T cell migration :::  :::

PubMed Central

2024-06-05T03:55:46.034687

2004-1-20

PMC314474

Introduction {#s1} ============ Sleep is important for the consolidation of newly acquired memories ([@pbio-0020024-Jenkins1]; [@pbio-0020024-Fishbein1]; [@pbio-0020024-Pearlman1]; [@pbio-0020024-Smith1]; [@pbio-0020024-Smith2]; [@pbio-0020024-Karni1]; [@pbio-0020024-Stickgold1]; [@pbio-0020024-Laureys1]; [@pbio-0020024-Fenn1]). The discovery of experience-dependent neuronal reactivation during sleep ([@pbio-0020024-Pavlides2]) corroborated the notion that novel memory traces, after successful encoding, must be replayed in their supporting neuronal networks until synaptic plasticity can effect trace consolidation ([@pbio-0020024-Hebb1]; [@pbio-0020024-Gutwein1]; [@pbio-0020024-Winson1]; [@pbio-0020024-Ribeiro1]). Postacquisition neuronal reactivation during sleep or quiet waking (WK) was found to preserve the temporal relationships of alert, exploratory WK in the hippocampus (HP) ([@pbio-0020024-Wilson1]; [@pbio-0020024-Skaggs1]; [@pbio-0020024-Nadasdy1]; [@pbio-0020024-Poe1]; [@pbio-0020024-Louie1]; [@pbio-0020024-Lee1]) and the cerebral cortex (CX) ([@pbio-0020024-Qin1]; [@pbio-0020024-Hoffman1]), causing a correlated replay of activity patterns across two-neuron ([@pbio-0020024-Wilson1]) or many-neuron ([@pbio-0020024-Louie1]) ensembles. To date, experience-dependent brain reactivation during sleep has been observed in rodents ([@pbio-0020024-Pavlides2]; [@pbio-0020024-Wilson1]; [@pbio-0020024-Skaggs1]; [@pbio-0020024-Qin1]; [@pbio-0020024-Nadasdy1]; [@pbio-0020024-Louie1]; [@pbio-0020024-Lee1]), nonhuman primates ([@pbio-0020024-Hoffman1]), humans ([@pbio-0020024-Maquet1]), and even songbirds ([@pbio-0020024-Dave1]), pointing to a very general biological phenomenon. Importantly, postacquisition brain reactivation during sleep has been shown to be proportional to memory acquisition in rats ([@pbio-0020024-Gerrard1]) and humans ([@pbio-0020024-Peigneux1]). In spite of the positive evidence, the brain reactivation hypothesis for memory consolidation during sleep faces several objections. First, the neocortical reactivation detected to date is extremely subtle and decays rapidly within less than 1 h of memory trace formation ([@pbio-0020024-Qin1]; [@pbio-0020024-Hoffman1]). Such transient reactivation falls short of explaining the disruption of memory traces by sleep deprivation several hours and even days after initial acquisition ([@pbio-0020024-Fishbein1]; [@pbio-0020024-Pearlman1]; [@pbio-0020024-Smith1]; [@pbio-0020024-Smith2]; [@pbio-0020024-Karni1]; [@pbio-0020024-Stickgold1]; [@pbio-0020024-Fenn1]). Second, *strictu sensu* neuronal reactivation during sleep in mammals has only been investigated in the hippocampocortical loop ([@pbio-0020024-Pavlides2]; [@pbio-0020024-Wilson1]; [@pbio-0020024-Skaggs1]; [@pbio-0020024-Qin1]; [@pbio-0020024-Nadasdy1]; [@pbio-0020024-Louie1]; [@pbio-0020024-Hoffman1]; [@pbio-0020024-Lee1]), making it difficult to ascertain whether the phenomenon is particular to this neural circuit or whether it represents global experience-dependent changes in the brain. Third, brain reactivation has mostly been observed in highly trained animal subjects ([@pbio-0020024-Wilson1]; [@pbio-0020024-Skaggs1]; [@pbio-0020024-Qin1]; [@pbio-0020024-Nadasdy1]; [@pbio-0020024-Dave1]; [@pbio-0020024-Louie1]; [@pbio-0020024-Hoffman1]; [@pbio-0020024-Lee1]), raising skepticism about its relevance for the acquisition and consolidation of novel information ([@pbio-0020024-Kudrimoti1]). Finally, experience-dependent brain reactivation has been reported to occur in all behavioral states ([@pbio-0020024-Pavlides2]; [@pbio-0020024-Wilson1]; [@pbio-0020024-Skaggs1]; [@pbio-0020024-Qin1]; [@pbio-0020024-Louie1]; [@pbio-0020024-Lee1]), including WK ([@pbio-0020024-Nadasdy1]; [@pbio-0020024-Hoffman1]). Although the first finding in this regard has hinted at a possible predominance of reactivation during slow-wave (SW) sleep ([@pbio-0020024-Pavlides2]), a comprehensive comparison of the relative contributions of WK, SW, and rapid eye-movement (REM) sleep for brain reactivation is still missing. To further complicate the issue, recent studies have raised the possibility that neuronal processing may occur at either slower or faster speed than normal physiological rates during REM ([@pbio-0020024-Louie1]) and SW ([@pbio-0020024-Nadasdy1]; [@pbio-0020024-Lee1]) sleep, respectively. Thus, it is uncertain at the moment how brain reactivation relates to different behavioral states. In order to address these objections, we set out to investigate the effects of a transient novel tactile experience on the long-term evolution of ongoing brain activity across the major behavioral states of the rat. In each of the five animals studied, extracellular activity of 59--159 neurons per animal and local field potentials (LFPs) representing larger-scale neural rhythms were simultaneously recorded from four different brain regions: HP, primary somatosensory "barrel field" CX, ventral posteromedial thalamic nucleus (TH), and putamen (PU) ([Figure 1](#pbio-0020024-g001){ref-type="fig"}A and [1](#pbio-0020024-g001){ref-type="fig"}B; see also Figures [S6](#sg006){ref-type="supplementary-material"} and [S7](#sg007){ref-type="supplementary-material"}). These brain regions were chosen because they comprise three major forebrain circuit loops essential for rodent species-specific behaviors. Rats are nocturnal gatherers that exhibit a variety of exploratory behaviors during the night, sleeping intermittently and mostly during the day ([@pbio-0020024-Timo-Iaria1]) (upper panel in [Figure 1](#pbio-0020024-g001){ref-type="fig"}C). In the wild, rats rely on spatial navigation and superb whisker-based tactile discrimination to explore new territories in search of food (Nowak 1999). The corticothalamic, corticohippocampal, and corticostriatal loops probed in this study have been implicated in tactile information processing ([@pbio-0020024-Simons1]; [@pbio-0020024-Ghazanfar1]), spatial navigation and memory formation ([@pbio-0020024-OaKeefe1]; [@pbio-0020024-Squire1]), and the execution of complex motor sequences ([@pbio-0020024-Dieckmann1]; [@pbio-0020024-Jog1]). ::: {#pbio-0020024-g001 .fig} Figure 1 ::: {.caption} ###### Methodology \(A) Neuroanatomical location of multielectrode implants, indicated on a schematic parasaggital section based on [@pbio-0020024-Paxinos1]). Indicated are the cerebral cortex (CX), the hippocampus (HP), the thalamus (TH), and the putamen (PU). \(B) Top view of a rat implanted with several multielectrode arrays. \(C) Experimental design. The upper panel shows a representative example of the strong circadian dynamics of the rat sleep--wake cycle (rat 5). Gray bands indicate lights-off; white bands indicate lights-on. Notice the fixed 12-h periods of darkness and light. The lower panels show animals continuously recorded for up to 96 h that were kept undisturbed except for a 1-h period of novel CSS (white segment) produced by the tactile exploration of four distinct novel objects placed at the corners of the recording box. Neural data from pre- and postnovelty periods (black and red segments, respectively, in the middle panel) were compared. \(D) Neuronal ensemble correlation method. Neuronal activity templates (red boxes) were compared with extensive recordings of neuronal action potentials (green ticks in upper panel) by way of an offline template-matching algorithm ([@pbio-0020024-Louie1]) that generalizes the notion of pairwise correlations to neuronal ensembles of any size. Templates and targets (white boxes) were binned, firing-rate normalized, and correlated (middle panel). This procedure yields a time series of neuronal ensemble correlations for each template--target sliding match (lower panel). \(E) Templates of interest (red boxes) were sampled around the origin of pre- and postnovelty periods during alert WK and slid against their corresponding neuronal targets so as to sample neuronal correlations every 30 s for up to 48 h. :::  ::: In our studies, neural signals were continuously recorded across the natural sleep--wake cycle for 48--96 h, with a single 1-h exposure to four complex objects placed in the four corners of the recording box ([Figure 1](#pbio-0020024-g001){ref-type="fig"}C). All objects were strictly novel to the subjects and were designed to maximize shape, texture, and behavioral value differences ([Figure S1](#sg001){ref-type="supplementary-material"}). Objects were presented half-way through the recording time (01:16 a.m. ± 00:51, mean ± SD) when lights were off and WK peaked (see [Figure 1](#pbio-0020024-g001){ref-type="fig"}C), so as to maximize the drive for whisker-based tactile exploration of the environment. The experiment, therefore, consisted of a naturalistic behavioral paradigm involving multiple novel sensory and spatial cues; it was designed to maximize changes induced by exposure to novel objects, as opposed to changes related to repeated behavioral training. As expected, this paradigm increased WK relative to sleep during the exposure time ([Figure S2](#sg002){ref-type="supplementary-material"}), leading to novel complex sensory stimulation (CSS). Other than novel stimulation and the periodic removal of waste and introduction of food pellets and water, animals were kept undisturbed in the same environment throughout the recordings. Our paradigm produced marked and acute exploratory behavior ([Figure S2](#sg002){ref-type="supplementary-material"}) without disrupting the large-scale sleep--wake structure across the many hours of recording (see upper panel in [Figure 1](#pbio-0020024-g001){ref-type="fig"}C). In order to investigate the long-term effects of novel stimulation on the spatiotemporal evolution of ongoing neuronal activity, we took advantage of a neuronal ensemble correlation method previously shown to detect experience-dependent reactivation of rodent hippocampal ensembles during SW and REM sleep ([@pbio-0020024-Louie1]). This method generalizes the concept of pairwise neuronal correlations ([@pbio-0020024-Qin1]; [@pbio-0020024-Hoffman1]) to an arbitrarily large number of neurons, quantifying the degree of similarity between spatiotemporal patterns of neuronal activity by way of a firing rate-normalized template-matching algorithm (see [Figure 1](#pbio-0020024-g001){ref-type="fig"}D). Templates of alert WK neuronal ensemble activity were selected from moments when animals made whisker contact with the novel objects (*n*= five templates per animal). Control templates were selected from epochs of alert WK 24 h or 48 h before novel stimulation (24 h for rats 1--3, 48 h for rats 4 and 5; *n* = five templates per animal), during which familiar tactile stimulation was produced by the contact of whiskers with the smooth walls of the recording box, to which animals had been habituated. Templates were matched against the entire record of neuronal activity using the neuronal ensemble correlation method (see [Figure 1](#pbio-0020024-g001){ref-type="fig"}E). The resulting correlation temporal profiles were averaged for each template set, aligned with reference to the light--darkness cycle to control for possible circadian effects, and compared. Results {#s2} ======= Novelty-Induced Patterns of Neuronal Activity Reverberate for up to 48 h {#s2a} ------------------------------------------------------------------------ First, we tested whether the neuronal ensemble correlation method could detect any trace of neuronal reverberation lasting at least more than 1 h after exposure to novel stimulation. For this, we examined correlation profiles obtained for all recorded neurons (three to four brain areas pooled together) in each animal. As shown in [Figure 2](#pbio-0020024-g002){ref-type="fig"}A, postnovelty average correlation distributions were significantly right-shifted relative to prenovelty distributions (ANOVA of mean pre- and postnovelty correlations over 24 h or 48 h; *n* = five animals, *F* = 9.5, *d.f.* 1, *p* = 0.016). This indicates that the neuronal firing patterns concomitant with novel stimulation persisted significantly more during the ensuing time than patterns sampled 24 h or 48 h before novel stimulation, when animals were in the same behavioral state (alert WK), but without novel objects to explore. The effect was independently observed, to a variable degree, in all the five animals studied (Bonferroni comparison, *p* \< 0.01). [Figure 2](#pbio-0020024-g002){ref-type="fig"}B shows the temporal evolution of neuronal ensemble correlations for 24 h (upper three panels in [Figure 2](#pbio-0020024-g002){ref-type="fig"}B) and 48 h (lower two panels in [Figure 2](#pbio-0020024-g002){ref-type="fig"}B). ::: {#pbio-0020024-g002 .fig} Figure 2 ::: {.caption} ###### Neuronal Ensemble Correlations Including up to Four Forebrain Regions Reveal Long-Lasting Reverberation \(A) Postnovelty neuronal correlations were significantly larger than prenovelty correlations in all animals studied. \(B) Temporal profiles of neuronal ensemble correlations. Gray bands indicate lights-off; white bands indicate lights-on. \(C) Temporal evolution of *p* values associated with pre- and postnovelty Bonferroni comparisons performed in intervals of 1 h (rats 1--3) or 2 h (rats 4 and 5). Significant experience-dependent neuronal reverberation was detected up to 48 h after novel stimulation. Color bar in linear scale; black denotes *p* \> 0.05. The minimum *p* values (MIN P) were, respectively, 10^−14^, 10^−3^, 10^−12^, 10^−23^, and 10^−22^. :::  ::: Despite the marked interanimal variability in the shapes and magnitudes of these profiles, a significant increase of neuronal correlations after exposure to novel stimulation was observed in most recording sites. Importantly, these increases lasted well above 1 h, as revealed by the temporal evolution of the *p* values generated by Bonferroni comparisons between post- and prenovelty correlation distributions ([Figure 2](#pbio-0020024-g002){ref-type="fig"}C). These results indicate that significant experience-dependent neuronal reverberation could be detected in the forebrain up to 48 h after exposure to novel stimulation. In order to assess the contributions of different neurons to total ensemble correlations, we ran the correlation analysis on a neuron-by-neuron basis. We found that no one subset of neurons was particularly responsible for the reverberation effect, since the contribution of individual neurons was highly variable in time (data not shown). This indicates that the neuronal changes associated with exposure to novel stimuli were highly distributed through the neuronal populations sampled. Furthermore, judging by the maximum neuronal ensemble correlations observed (rat 3, peak at 0.35; [Figure 2](#pbio-0020024-g002){ref-type="fig"}B), one would conclude that novel stimulation-specific neuronal activity was not perfectly repeated, but was rather loosely reverberated for several hours. Neuronal Reverberation Occurs in Multiple Forebrain Areas {#s2b} --------------------------------------------------------- In order to assess the anatomical distribution of experience-dependent neuronal reverberation, we performed the neuronal ensemble correlation analysis for each area separately ([Figure 3](#pbio-0020024-g003){ref-type="fig"}A). At first glance, differences between pre- and postnovelty correlation profiles were noticeable in all animals, with predominant effects in a different subset of areas for each animal. For example, rat 4 showed marked reverberation in the HP, but small changes in the CX, while rat 5 showed just the opposite. In the majority of the recorded sites, postnovelty traces (red in [Figure 3](#pbio-0020024-g003){ref-type="fig"}A) run above prenovelty traces (black in [Figure 3](#pbio-0020024-g003){ref-type="fig"}A), but the reverse also occurs, suggesting some sort of antireverberation. The most widespread reverberation was seen in rat 1, which showed sustained reverberation in the CX and decaying reverberation in the HP and TH. A somewhat similar pattern was seen in rat 5, while rat 3 showed strong reverberation only in the PU, and rat 4 in the HP and TH. Rat 2 showed the least reverberation of all, with somewhat stronger effects in the PU. ::: {#pbio-0020024-g003 .fig} Figure 3 ::: {.caption} ###### Long-Lasting Neuronal Reverberation Occurs in the CX, HP, PU, and TH \(A) Temporal profiles of neuronal ensemble correlations in all recording sites. Gray bands indicate lights-off; white bands indicate lights-on. Red and black traces indicate post- and prenovelty correlations, respectively. \(B) Temporal evolution of *p* values associated with pre- and postnovelty Bonferroni comparisons for individual brain areas, calculated as in [Figure 2](#pbio-0020024-g002){ref-type="fig"}C. Color bar in linear scale; black denotes *p* \> 0.05. Minimum *p* values (MIN P) in crescent "rat number" order, as follows: CX: 10^−10^, 10^−4^, 10^−5^, 10^−9^, 10^−9^; HP: 10^−12^, not significant, 10^−22^, 10^−3^ (both red and blue scales); PU: 10^−2^, 10^−12^, 10^−15^ (blue scale) and 10^−17^ (red scale), 10^−16^ (red scale) and 10^−34^ (blue scale); TH: 10^−18^, 10^−3^ (blue scale), not significant, 10^−30^, 10^−16^. \(C) Neuronal ensemble correlations for no-contact templates (taken from epochs within the novel stimulation period in which animals had no sensory contact with the novel objects) also show enhanced neuronal reverberation. :::  ::: Despite the interanimal and interarea differences in the magnitude and shape of correlation profiles, significant changes between pre- and postnovelty correlations were observed in all areas studied (CX, five of five rats; HP, three of four rats; PU, four of four rats; and TH, four of five rats; Bonferroni comparison, *p* \< 0.05). Indeed, experience-dependent changes were not statistically different across different forebrain areas (ANOVA, *F* = 0.24, *d.f.* 3, *p* = 0.86). The temporal evolution of *p* values (Bonferroni comparison) associated with single-area correlation profiles shows that significant reverberation was present in 16 of 18 recording sites for several hours after exposure to novel stimulation ([Figure 3](#pbio-0020024-g003){ref-type="fig"}B). It also confirms that neuronal ensemble reverberation (post-/prenovelty correlations) is not the only kind of experience-dependent change possible. Some animals showed significant long-lasting antireverberation (pre-/postnovelty correlations), i.e., patterns of activity that were statistically more dissimilar from novel stimulation templates than expected by chance. Antireverberation (indicated by blue hues in [Figure 3](#pbio-0020024-g003){ref-type="fig"}B) occurred in the HP (one of four rats), PU (two of four rats), and TH (one of five rats), but not in the CX. Single-area postnovelty average correlations showed peaks of the order of 0.4 (rat 5, PU; [Figure 3](#pbio-0020024-g003){ref-type="fig"}A), but typically ranged from 0.1 to 0.2. Therefore, high-fidelity replay of neuronal firing patterns was not observed even when single areas were considered. An intriguing observation came from the scrutiny of no-contact templates of neuronal activity, sampled within the novel stimulation 1-h period during alert WK, but excluding moments of contact between whiskers and objects. Surprisingly, no-contact templates yielded correlation profiles that were almost indistinguishable from those obtained when animals had tactile contact with the novel objects ([Figure 3](#pbio-0020024-g003){ref-type="fig"}C). Therefore, the exploration of the novel environment enhanced the reverberation of all the neuronal activity patterns concomitant with the experience and not just of those corresponding to moments in which animals received sensory inputs from the objects. This rules out the possibility that stimulus complexity, rather than novelty, is the underlying cause of the enhanced neuronal reverberation observed after exploration of the objects. Neuronal Reverberation Peaks during SW Sleep {#s2c} -------------------------------------------- Single-area results indicate that neuronal ensemble correlations often peak during discrete epochs that last a few hours. We also noticed marked oscillations of the correlation trace in several recorded sites (e.g., rat 5, CX). These observations suggest that some underlying biological process, with slow evolution but with sharp phase transitions, governs the long-term reverberation of neuronal firing patterns. To test whether transitions in the wake--sleep cycle could amount for these effects, we investigated how experience-dependent changes in neuronal correlations varied across the three major rat behavioral states: WK, SW sleep, and REM sleep. A comparison across states of post-/prenovelty correlation ratios calculated from averages of entire recordings indicated a significant state-specific effect (ANOVA, *F* = 9.289, *d.f.* 2, *p*= 0.0004), with SW ratios being significantly higher than those of both WK (Bonferroni comparison, *p* \< 0.05) and REM (Bonferroni comparison, *p* \< 0.003). Indeed, significant state-specific differences in post-/prenovelty correlation ratios were individually detected in four of five animals (ANOVAs, *d.f.* 2: rat 2, PU, *F* = 4.13, *p* = 0.039; rat 3, CX, *F* = 6.45, *p* = 0.026; rat 4, HP, *F* = 3.99, *p* = 0.029; rat 5, CX, *F* = 13.81, *p* \< 0.0001). The mean correlation values found in those recording sites for the three behavioral states reveal that SW correlations were systematically larger than WK correlations ([Figure 4](#pbio-0020024-g004){ref-type="fig"}A). Several other recorded sites displayed similar but nonsignificant trends. Meanwhile, the REM correlations mea-sured were variable and could not be consistently ranked in relation to WK and SW sleep. Comparable neuronal reverberation between SW and REM sleep was observed in only one animal (rat 5, CX). A major effect of SW sleep on neuronal reverberation was corroborated by the temporal evolution of successive state-specific Bonferroni comparison *p* values calculated for pre- and postnovelty 4-h average correlations across all animals and brain areas studied ([Figure 4](#pbio-0020024-g004){ref-type="fig"}B). The strongest contrast between pre- and postnovelty neuronal correlations was clearly seen during SW sleep, with less effect seen in WK and even less in REM sleep. [Figure 4](#pbio-0020024-g004){ref-type="fig"}C depicts the state-sorted pre- and postnovelty correlations for rat 5, CX, illustrating both the general SW effect and the much less-prevalent REM sleep changes. ::: {#pbio-0020024-g004 .fig} Figure 4 ::: {.caption} ###### Neuronal Reverberation Depends on Behavioral State \(A) Histograms (mean *±* SEM) of post- and prenovelty correlation ratios in the recording sites where significant state-related differences in neuronal ensemble correlations were detected. SW sleep post-/prenovelty correlation ratios were significantly higher than WK in all four cases (Bonferroni comparison *p* values as follows: rat 2, PU, SW\>WK 0.013; rat 3, CX, SW\>REM 0.017 and SW\>WK 0.022; rat 4, HP, SW\>REM 0.013 and SW\>WK 0.039; rat 5, CX, SW\>WK 0.0001 and REM\>WK 0.0002). \(B) Bonferroni comparison *p* values for post-/prenovelty comparisons in all animals according to behavioral state and brain area, calculated in intervals of 4 h. Animal order and time as in [Figure 3](#pbio-0020024-g003){ref-type="fig"}B. Color bar in linear scale. \(C) Neuronal ensemble correlations sorted by state for rat 5 CX. In comparison with WK, there is a clear increase in the contrast between pre- and postnovelty correlations during SW sleep. In this particular animal and brain area, increased correlations were also seen for REM sleep, but this was not the case in other animals (A). Furthermore, this REM effect was substantially weakened when expressed in *p* values (B), due to the very short duration of REM sleep episodes. In this respect, notice that REM sleep has much fewer datapoints, reflecting the short duration of this state relative to WK and SW sleep. Thus, even in a site where REM sleep showed results similar to SW sleep, the cumulative neuronal reverberation that takes place during REM is necessarily less than that of SW. \(D) Statistical comparison of matches between templates of neuronal activity sampled at WK normal speed with a range of targets spanning different temporal scales. Plotted are Bonferroni comparison *p* values for post- and prenovelty comparisons in all animals according to behavioral state and brain area, calculated in intervals of 4 h for speed factors ranging from 20 times faster to two times slower than the WK normal rate. No evidence for optimization at speeds different from the WK physiological rate (1×) was found. Color scale as in (B). :::  ::: Next, we tested the possibility reported in hippocampal place cells ([@pbio-0020024-Nadasdy1]; [@pbio-0020024-Louie1]; [@pbio-0020024-Lee1]) that experience-dependent replay of neuronal firing patterns during sleep can be slower (REM) or faster (SW) than during WK. Template-to-target matches at different speed factors were obtained by comparing 250 ms-binned templates with targets binned within a range of different bin sizes (from 12.5 ms to 500 ms). By temporally compressing and expanding "target" spike records before matching them to templates, we determined the magnitude of neuronal ensemble correlations for speed factors ranging from 0.5 to 20 times the physiological WK processing speed, which covers the reported optimum speed ranges for SW ([@pbio-0020024-Lee1]) and REM sleep ([@pbio-0020024-Louie1]). A predominance of neuronal reverberation during SW sleep was seen for all speed factors, as indicated by Bonferroni comparisons ([Figure 4](#pbio-0020024-g004){ref-type="fig"}D). However, no significant differences were seen when post- and prenovelty correlation ratios (calculated from averages of entire recordings) were compared across different speed factors (ANOVA, *F* = 1.496, *d.f.* 5, *p* = 0.19). [Figure 4](#pbio-0020024-g004){ref-type="fig"}D shows that within any given state or area, neuronal reverberation did not vary systematically with speed factor, and the temporal distribution of correlation hot-and-cold spots was largely insensitive to speed factor. Thus, we found no evidence that forebrain neuronal reverberation can be optimized assuming replay speeds different from the WK normal rate. Indeed, a subtle but consistent decrease of *p* values can be observed for speed factors 10 times and 20 times faster than normal WK rates, while speed factors near the physiological range (2×--0.5×) show stronger and similar effects. This was the case even in the HP, in contrast with previous findings in hippocampal place cells recorded in highly trained animals performing a spatial navigation task ([@pbio-0020024-Nadasdy1]; [@pbio-0020024-Louie1]; [@pbio-0020024-Lee1]). At present, it is unclear whether this discrepancy reflects differences in stimulus familiarity (novel versus habitual), stimulation modality (tactile exploration versus spatial navigation), the very low representation of place cells in our hippocampal samples (less than 5%), or possible differences in the analyses used in previous studies, based on the statistical boot-strapping of relatively small datasets ([@pbio-0020024-Nadasdy1]; [@pbio-0020024-Louie1]; [@pbio-0020024-Lee1]). All together, our results indicate that neuronal reverberation was consistently stronger during SW sleep, decreasing during WK. This is remarkably well-illustrated by a superimposition of behavioral state classification and neuronal ensemble correlations (middle panel in [Figure 5](#pbio-0020024-g005){ref-type="fig"}A), which reveals an exquisite long-term temporal match between epochs of increased neuronal correlations and SW episodes (red in [Figure 5](#pbio-0020024-g005){ref-type="fig"}A). Likewise, neuronal correlation troughs show a tight correspondence with WK episodes (blue in [Figure 5](#pbio-0020024-g005){ref-type="fig"}A). This characteristic state-dependency persisted throughout the 45 h of postnovelty recording (see [Figure 4](#pbio-0020024-g004){ref-type="fig"}B). Notice that REM sleep only showed SW-like results in one out of five animals (rat 5, CX; depicted in Figures [4](#pbio-0020024-g004){ref-type="fig"} and [5](#pbio-0020024-g005){ref-type="fig"}). In the remaining animals, REM correlations were either closer to WK levels than to SW levels or in between (see [Figure 4](#pbio-0020024-g004){ref-type="fig"}A). Given this marked variability and the very short duration of total REM sleep in comparison with total SW sleep (WK 52%, SW 40%, and REM 8% of total recording time for five animals), this indicates that REM sleep plays a minor role in neuronal reverberation. ::: {#pbio-0020024-g005 .fig} Figure 5 ::: {.caption} ###### Neuronal Reverberation Is Strongest during SW Sleep \(A) Rat 5 (CX) dramatically illustrates the state dependency of neuronal ensemble correlations, which are strongly increased by SW sleep but readily decreased by WK. The upper panel shows the firing rates of 38 cortical neurons for approximately 45 h after exposure to novel stimulation (indicated by an asterisk). The middle panel shows the superimposition of successive neuronal ensemble correlations and concurrent behavioral states. Nearly all correlation peaks correspond to SW episodes, while almost all troughs match WK epochs. The lower panel represents pooled LFP forebrain coherence ([@pbio-0020024-Amjad1]) over time, useful to discriminate between WK (strong coherence above 25 Hz and weak coherence under 5 Hz) and SW sleep (the opposite). Notice that in this particular example (rat 5, CX), REM episodes show correlations similar to those of SW sleep, but, as shown in [Figure 4](#pbio-0020024-g004){ref-type="fig"}, this was the exception and not the rule across several animals. \(B) State-dependent neuronal reverberation was sustained throughout the recording period, as shown by segments representing the beginning (3,200--3,300 min), middle (4,700--4,800 min), and end (5,200--5,250 min) of the experimental record. In the upper panel, notice the progressive increase of neuronal correlations across single SW sleep episodes. \(C) Blow-up of two selected data segments indicated by asterisks in (A). Despite having being sampled from moments of high neuronal firing rates (asterisk), novel stimulation templates reverberate most strongly during SW sleep when firing rates are low (single asterisk and double asterisks). The high firing rates that characterize WK correspond to decreased neuronal reverberation, probably due to sensory interference. :::  ::: Interestingly, a comparison of the correlation temporal profile with the concurrent neuronal firing record (upper panel in [Figure 5](#pbio-0020024-g005){ref-type="fig"}A) reveals that SW correlation peaks correspond to periods of decreased firing rate, while WK correlation troughs match epochs of increased neuronal activity. This is better shown in [Figure 5](#pbio-0020024-g005){ref-type="fig"}C, which depicts data segments approximately 2-h long, comprising the three major behavioral states studied. The first segment (shown by a single asterisk in [Figure 5](#pbio-0020024-g005){ref-type="fig"}A) corresponds to 60 min of novel stimulation and the immediately ensuing sleep--wake cycles, while the second segment (shown by double asterisks in [Figure 5](#pbio-0020024-g005){ref-type="fig"}A) illustrates sleep--wake episodes occurring approximately 15 h after the original experience. Thus, although novel stimulation templates of neuronal activity were selected from WK episodes characterized by high firing rates, ensuing reverberation of these neuronal firing patterns was most pronounced during SW sleep, under lower firing rates. Discussion {#s3} ========== In order to assess several objections to the replay hypothesis for memory consolidation during sleep, we conducted long-term continuous neuronal recordings on animals subjected to a naturalistic behavioral paradigm, which involved multiple novel sensory and spatial cues. Our results indicate that large-scale neuronal firing patterns generated during the exploration of novel objects can recur for several hours after the reference experience throughout most of the forebrain, while firing patterns associated with familiar stimulation (i.e., the walls of the recording box) are substantially less detectable over time. Our results fend off three major objections to the notion that neuronal reverberation during sleep may underlie memory consolidation. First, significant experience-depen-dent changes in neuronal ensemble correlations can be tracked as late as 48 h after the reference novel experience, being therefore compatible with memory impairment effects of sleep deprivation applied hours or days after training ([@pbio-0020024-Fishbein1]; [@pbio-0020024-Pearlman1]; [@pbio-0020024-Smith1]; [@pbio-0020024-Smith2]; [@pbio-0020024-Karni1]; [@pbio-0020024-Stickgold1]; [@pbio-0020024-Fenn1]). Second, these effects were observed in rats completely naïve with respect to the reference stimuli, ruling out the possibility that only the performance of highly trained behaviors would be followed by neuronal reverberation. Third, neuronal ensemble correlations were significantly enhanced during SW sleep and decreased during WK, while REM sleep produced variable results. The data indicate that novel experience caused sustained neuronal reverberation ([@pbio-0020024-Hebb1]) rather than discrete reactivation ([@pbio-0020024-Wilson1]; [@pbio-0020024-Kudrimoti1]), in the sense that reverberation decreased, but did not disappear, during WK (see [Figure 4](#pbio-0020024-g004){ref-type="fig"}C). The consistent increase in neuronal reverberation during SW sleep, the high interanimal variability of neuronal reverberation during REM sleep, and the small contribution of REM sleep to total sleep time indicate that the cognitive effects of experience-dependent neuronal reverberation ([@pbio-0020024-Gerrard1]; [@pbio-0020024-Peigneux1]) must be largely attributed to SW sleep. Therefore, our results suggest a major role for SW sleep in the reverberation of new memory traces. An important result of the present study is the inverse correlation between neuronal correlations and concurrent firing rates. Although all neuronal activity templates were taken from epochs of high arousal WK when firing rates were generally high, their neuronal reverberation during subsequent WK was not very prominent (see Figures [4](#pbio-0020024-g004){ref-type="fig"} and [5](#pbio-0020024-g005){ref-type="fig"}). In contrast, reverberation of the same activity templates peaked during SW sleep, when the firing rates of forebrain neurons are generally low (see Figures [4](#pbio-0020024-g004){ref-type="fig"} and [5](#pbio-0020024-g005){ref-type="fig"}). This suggests that reverberating patterns of neuronal activity associated with past novel experience are largely---but not completely---masked during WK by incoming sensory inputs unrelated to the reference experience. By the same token, peak neuronal correlations arise during SW sleep, when sensory interference ceases. Taken together, these observations corroborate the notion that the importance of sleep for memory consolidation stems from the offline processing of memory traces, i.e., from the absence of sensory interference ([@pbio-0020024-Jenkins1]; [@pbio-0020024-Melton1]; [@pbio-0020024-Winson1]). Were there differences in average firing rate before and after exposure to the novel environment, and could such differences account for the effects seen in neuronal correlations? The neuronal ensemble correlation method ([@pbio-0020024-Louie1]) involves a normalization of firing rates after binning, and therefore it is insensitive to moderate changes in the mean firing rate. In our experiment, firing rates for individual neurons varied substantially during exposure to novelty, with some neurons firing more and other neurons firing less than before exposure. This variability caused a moderate but nonsignificant increase in the average activity of the cells so that the firing rates within novelty templates were on average approximately 10% higher than the neuronal firing rates of preexposure templates. Neuronal firing rates increased during exposure to novel objects and persisted elevated for up to 1 h after removal of the objects, returning to baseline afterwards. This contrasts with the timecourse of neuronal correlations changes, which were increased for up to 48 h. Finally, as explained above, neuronal reverberation was inversely correlated with firing rates. Thus, mean firing-rate differences were not responsible for the reverberation effect, which should rather be attributed to specific firing-rate relationships across multiple neurons. Our results impose some clear constraints on future sleep and learning theories. First, no sign of neuroanatomical specificity was found in the correlations measured, and in particular no significant differences between hippocampal and extrahippocampal areas could be detected. Despite considerable interanimal variability in the magnitude of the correlations observed in the different brain structures, statistically significant neuronal reverberation produced by novel stimulation was observed in 16 out of 18 recorded brain sites, comprising the CX, HP, PU, and TH. This broad forebrain reverberation was related to the free exploration of four novel and complex objects, placed in four well-separated places and including the presence of novel food. Thus, novel experience involving tactile, gustatory, olfactory, spatial, and motor components is able to engage multiple forebrain structures, all similarly capable of reverberating neuronal patterns of activity after novel stimulation. Second, neuronal ensemble correlations measured across the forebrain were typically small (on the order of 0.1--0.3), agreeing with values previously reported for pairwise ([@pbio-0020024-Wilson1]; [@pbio-0020024-Skaggs1]; [@pbio-0020024-Qin1]; [@pbio-0020024-Hoffman1]) or many-neuron ([@pbio-0020024-Louie1]) correlations. Qualitatively similar results were observed for bins ranging from 5 ms to 1,000 ms, with higher correlation values for larger bin sizes. This suggests that neurons of multiple forebrain areas, once exposed to novel experience, do not accurately replay prior WK activity patterns longer than 5 ms. Instead, they show a mild but long-lasting bias towards (or against) the reference activity patterns. Indeed, not a single template-to-target match (out of 979,200 matches sampled) yielded correlation values higher than 0.45, indicating that novelty-induced neuronal reverberation occurs at low fidelity. It has been proposed that a high-fidelity replay of neuronal firing patterns during sleep may be achieved assuming that replayed patterns can undergo time compression and expansion ([@pbio-0020024-Nadasdy1]; [@pbio-0020024-Louie1]; [@pbio-0020024-Lee1]). We assessed this possibility thoroughly, but found no evidence of such effects in any of the forebrain sites recorded. Thus, in the face of consistently low neuronal correlation values, the "high-fidelity replay hypothesis" for timeperiods larger than 5 ms should be rejected, at least in mammals ([@pbio-0020024-Dave1]). It remains to be seen whether more precise patterns of spike-to-spike correlations may reverberate in intervals smaller than 5 ms. A third important point regards the observation that neural activity sampled when animals were aroused, but not touching the objects, yielded neuronal reverberation that was nearly identical to that obtained when animals made sensory contact with the objects. This indicates that the kind of experience-dependent neuronal reverberation detected by the neuronal ensemble correlation method ([@pbio-0020024-Louie1]) does not reflect the specific features of the stimuli, but is related to the overall behavioral salience of the novel stimulation period, i.e., is context- rather than stimulus-specific. In principle, these results are compatible with a slow and progressive process of memory consolidation ([@pbio-0020024-Bryson1]), proportional to the novelty of the experience, and able to bind together a multitude of contextual cues related to its core sensory elements ([@pbio-0020024-Kohler1]). It has been suggested that the neuronal reverberation of newly acquired synaptic changes during SW sleep may lead to the recall and storage of new memories by way of "calcium-mediated intracellular cascades" capable of opening the "molecular gates to plasticity" ([@pbio-0020024-Sejnowski1]). This hypothesis is partially contradicted by evidence that calcium-dependent gene expression related to synaptic plasticity is upregulated during REM sleep ([@pbio-0020024-Ribeiro1], [@pbio-0020024-Ribeiro2]), but not during SW sleep ([@pbio-0020024-Pompeiano1]). The present findings and the current literature suggest instead that SW and REM sleep play separate roles on memory consolidation, with memory recall occurring during SW sleep and memory storage taking place during REM sleep. According to this view, the deleterious effects of sleep deprivation on memory consolidation would be a consequence of the disruption of the underlying neuronal reverberation and gene expression during SW and REM sleep, respectively. The fact that neuronal reverberation is sustained for long epochs during SW sleep suggests that unconsolidated synaptic changes may not only be recalled, but also amplified over time during SW sleep. Indeed, a progressive increase of neuronal correlations across single SW sleep episodes was often observed (upper panel in [Figure 5](#pbio-0020024-g005){ref-type="fig"}B). A model of how such amplification may arise is presented in [Figure 6](#pbio-0020024-g006){ref-type="fig"}. We have recently proposed that the cyclical reiteration of trace amplification during SW sleep and trace storage during REM sleep promotes the postsynaptic propagation of memory traces ([@pbio-0020024-Pavlides1]), as suggested by the hippocampofugal pattern of gene expression during REM sleep ([@pbio-0020024-Ribeiro2]). Potentially, this propagation could cause memory traces to progressively reach farther and farther away from the original synaptic trajectory activated at initial encoding. Over time, this sleep-dependent propagation could lead to deeper encoding within the CX ([@pbio-0020024-Craik1]; [@pbio-0020024-Cermak1]), as well as hippocampal disengagement ([@pbio-0020024-Scoville1]; [@pbio-0020024-Mishkin1]; [@pbio-0020024-Kesner1]; [@pbio-0020024-Squire1]; [@pbio-0020024-Izquierdo1]; [@pbio-0020024-Bontempi1]). The notion that the two major phases of sleep play distinct and complementary roles in memory consolidation is in line with evidence that SW and REM sleep have synergistic effects on human procedural learning ([@pbio-0020024-Mednick1]). Put in historical perspective, our model argues that sleep separately harbors both mechanisms postulated in the past ([@pbio-0020024-Hebb1]) to be necessary for memory consolidation: postacquisition neuronal reverberation and structural synaptic plasticity. In conclusion, sustained neuronal reverberation during SW sleep, immediately followed by plasticity-related gene expression during REM sleep, may be sufficient to explain the beneficial role of sleep on the consolidation of new memories. ::: {#pbio-0020024-g006 .fig} Figure 6 ::: {.caption} ###### Conceptual Model of the Role of Sleep for Memory Consolidation Arrows indicate pathway activation by sensory inputs during WK, as well as intrinsic brain activity such as pontine waves during sleep ([@pbio-0020024-Datta1]); different arrow sizes indicate different magnitudes of pathway activation. Red indicates calcium-dependent pretranscriptional processes, with different hue intensities representing the progressive amplification of recently acquired synaptic changes. Green indicates plasticity-related transcriptional regulation. The initial state of the model (data not shown) consists of environmental habituation, during which ongoing activity patterns only repeat themselves by chance. (First panel) A novel WK experience encodes a memory trace across multiple forebrain areas, selectively activating functionally related synapses. This triggers calcium-dependent pretranscriptional cascades (red) and plasticity-related gene expression (green) that lead to the common strengthening of the activated synapses. (Second panel) The continuation of WK involves a succession of unrelated sensory experiences capable of producing interference, i.e., a progressive weakening of recently encoded synaptic changes. (Third panel) Upon entering SW sleep, intrinsic brain activation is biased towards previously potentiated synapses, causing neuronal firing patterns originally produced during the novel WK experience to reverberate significantly above chance levels. (Fourth panel) The periodic activation of calcium-dependent second-messenger cascades by large-amplitude SW oscillations may result in the progressive amplification of the synaptic changes that encode the novel memory trace. (Fifth panel) SW-amplified synaptic changes are stored during REM sleep by way of plasticity-related transcriptional regulation. :::  ::: Materials and Methods {#s4} ===================== {#s4a} ### Chronic neuronal recordings {#s4a1} Multielectrode arrays ([Figure S3](#sg003){ref-type="supplementary-material"}) were surgically implanted according to National Institutes of Health (NIH) guidelines in five adult male Long--Evans rats (250--300 g). The following coordinates in millimeters relative to Bregma ([@pbio-0020024-Paxinos1]) were used to center the arrays: HP (+2.8 anteroposterior \[AP\], +1.5 mediolateral \[ML\], −3.3 dorsoventral \[DV\]), CX (+3.0 AP, +5.5 ML, −1.5 DV), PU (−1.0 AP, +2.5 ML, −5.0 DV), and TH (+3.0 AP, +3.0 ML, −5.0 DV). Hippocampal data pool together neurons recorded with staggered electrodes from the CA1 field and the dentate gyrus. Locations of implants were histologically verified by comparing cresyl-stained frontal brain sections with reference anatomical planes ([@pbio-0020024-Paxinos1]) ([Figure S4](#sg004){ref-type="supplementary-material"}). A Multineuron Acquisition Processor (128 channels; Plexon Inc., Dallas, Texas, United States) was used to perform recordings of neuronal spikes and LFPs, as previously described ([@pbio-0020024-Nicolelis1]) (Figures [S5](#sg005){ref-type="supplementary-material"} and [S6](#sg006){ref-type="supplementary-material"}). A waveform-tracking technique involving periodic template adjustment was employed for the continuous recording of individual units (see [Figure S5](#sg005){ref-type="supplementary-material"}). Units that showed nonstationary waveforms, unstable firing-rate profiles, or both were discarded. The stability of firing rates within each behavioral state can be appreciated in the upper panel in [Figure 5](#pbio-0020024-g005){ref-type="fig"}. ### Behavior {#s4a2} Before the beginning of the experiment, animals were individually habituated to an empty recording box for 5--7 entire days (12 h:12 h light:dark schedule, lights on at 06:00 a.m.) so as to reach steady-state behavioral activity and baseline wake--sleep cycles. Behaviors were continuously recorded by way of two infrared-sensitive CCD video-cameras; infrared illumination was used to monitor behavior when visible lights were off. Behavioral states were coded as WK, SW sleep, or REM sleep based on a spectral analysis of LFPs and visual inspection of videotaped behaviors, according to previously described criteria ([@pbio-0020024-Ribeiro1]) ([Figure S7](#sg007){ref-type="supplementary-material"}). ### Data analysis {#s4a3} Data were processed and analyzed by custom-made MATLAB^TM^ (MathWorks, Natick, Massachusetts, United States) code running in a computer cluster comprising 32 CPU (Evolocity^TM^, LNXI, Sandy, Utah, United States). Neuronal ensemble correlations were calculated following the method in [@pbio-0020024-Louie1]). In brief, a stretch of data (target) is scanned for similarity to certain multineuron temporal patterns of activity (templates). Templates consist of *CxN* matrices corresponding to simultaneously recorded spike trains of *C* neurons binned into *N* intervals of a certain length. Our results were obtained using 9 s-long templates binned with 250 ms-long bins with as many neurons as were available for the brain areas in question. Hence, each template pattern yielded a *Cx36* matrix. Target data comprising 24 h or 48 h were sparsely sampled (every 30 s) and binned into matrices of the same dimensions. For a *C*×*N* template data matrix *x* and a target data matrix *y* sampled at time *t*, the ensemble correlation index C~t~ is obtained by: where (the same applies to Y~c~, y¯, and σ~y~). Templates were selected by careful scrutiny of behavior recorded on videotapes and of the corresponding spectral characteristics of hippocampal LFPs, so as to assure sampling during alert WK ([Figure S7](#sg007){ref-type="supplementary-material"}). Particular care was used to select template sets with a comparable prevalence of θ (5--8 Hz) over δ (2--4 Hz) frequencies (θ/δ hippocampal spectral ratios, ≥10). All templates and targets were individually sampled without overlap. Because binning may destroy higher-frequency phenomena, control analyses were run to assess the consequences of binning the data with bins of different sizes. Results obtained with bin sizes in the interval of 5--1,000 ms are qualitatively equivalent to the ones obtained using 250 ms bins ([Figure S8](#sg008){ref-type="supplementary-material"}). In order to detect temporally expanded or contracted repetitions of template patterns, a range of larger- or smaller-than-template bin sizes, respectively, was used when binning each target sample (12.5 ms, 25 ms, 125 ms, 250 ms, 375 ms, and 500 ms). This allowed us to compare targets spanning different temporal scales with templates of neuronal activity sampled at WK speed. The bin size range used allowed the detection of patterns temporally compressed by factors of 20, 10, and 2 (bin sizes of 12.5 ms, 25 ms, and 125 ms), temporally expanded by factors of 1.5 and 2 (bin sizes of 375 ms and 500 ms) as well as replayed at the same speed (250 ms). StatView^TM^ (SAS, Cary, North Carolina, United States) and MATLAB^TM^ software were used for descriptive statistics and hypothesis testing. Supporting Information {#s5} ====================== Figure S1 ::: {.caption} ###### Objects Four different objects were used to produce CSS. (9.7 MB PPT). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S2 ::: {.caption} ###### Behavior All animals were highly habituated to the recording box, so that exposure to novel complex objects caused a general increase in the animals\' arousal. Four of five animals showed an increase in time spent in WK with respect to SW and REM sleep during CSS (A), as compared to adjacent pre- and postnovelty periods of equal length (60 min). The only exception was rat 1, which showed nevertheless a marked exploratory drive, spending nearly 20% of the exposure period in direct whisker-contact with the objects (B). Individual object preferences were moderately varied, as indicated in (C). (746 KB PPT). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S3 ::: {.caption} ###### Multielectrode Arrays Teflon-coated tungsten wires (50 μm diameter, 300 μm between wires, 1.0--1.2 MΩ at 1 KHz; California Fine Wire Company, Grover Beach, California, United States) were assembled in multielectrode arrays shaped to fit different neuroanatomical targets. (282 KB PPT). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S4 ::: {.caption} ###### Location of Implants Frontal brain sections stained for cresyl-violet were used to determine the sites of electrode placement. Electrode tracks, tissue scars, and reference electrolytic lesions performed a few days before sacrifice were used to delimit the implant sites, indicated in red in the figure below. Numbers on the right represent standard AP coordinates ([@pbio-0020024-Paxinos1]) in millimeters from Bregma. (6.3 MB PPT). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S5 ::: {.caption} ###### Neuronal Recordings To record neuronal activity, differentiated neural signal was preamplified (2,000×--32,000×) and digitized at 40 KHz. Up to four neuronal action potentials per recording channel were sorted online (SortClient 2002, Plexon Inc.) and validated by offline analysis (Offline Sorter 2.3, Plexon Inc.) according to the following cumulative criteria: voltage thresholds greater than two standard deviations of amplitude distributions; signal-to-noise ratio greater than 2.5 (as verified on the oscilloscope screen); less than 1% of interspike intervals smaller than 1.2 ms; and stereotypy of waveform shapes, as determined by a waveform template algorithm and principal component analysis. In order to continuously record individual neurons for up to 96 h, we used an adaptive algorithm (available on SortClient 2002, Plexon Inc.) that adjusts waveform templates based on the recent accumulated mean shapes (1% of midline every 20 min). This allows for the same neuron to be tracked across consecutive days, as verified by the superimposition of waveforms acquired thoughout the experiment (Wavetracker software, Plexon Inc.). (3 MB PPT). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S6 ::: {.caption} ###### Neuronal Yield Up to 159 neurons were recorded from three to four different brain areas. (1.4 MB PPT). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S7 ::: {.caption} ###### Recording LFPs and Behaviors LFPs were recorded in parallel with spikes from the same electrodes. Neural signals were split, preamplified (1,000×), and filtered (0.5--400 Hz) by way of a Plexon LFP board. Signals were then fed to the MAP acquisition principal component through a NIDAQ card and digitized at 500 Hz. Behaviors were constantly recorded in videotape by two diametrically opposed infrared-sensitive CCD cameras (model WV-BP332, Panasonic, Laguna, Philippines). A millisecond-precision timer (model VTG-55, For-A Company, Tokyo, Japan) was used to synchronize the acquisition of spikes, LFPs, and videotape records. Behavioral states were identified by the combined inspection of videotapes and the spectral content (1--20 Hz) of LFPs. Behaviors were classified according to the following criteria: (1) alert WK: active exploration with whisking, plus strong hippocampal θ rhythm; (2) quiet WK: stillness or grooming, with eyes open and low-power hippocampal θ rhythm; (3) SW sleep: stillness with eyes closed, plus large-amplitude hippocampal δ rhythm; (4) REM sleep: overall stillness with intermittent whisking, eyes closed, strong hippocampal θ rhythm. The inspection of videotape records readily separates alert and quiet WK from sleep states, but the separation of SW and REM sleep relies strongly on LFP analysis. Hippocampal LFP is particularly useful to disambiguate SW and REM sleep: SW sleep has a strong δ band (2--4 Hz), while REM sleep shows increased θ band (5--8 Hz). The distinction between alert and quiet WK was used only for template selection (all templates taken from alert WK). For all other purposes, alert and quiet WK data were combined into a single WK category. The graphs depict the θ/δ hippocampal spectral ratios (mean ± SEM) of the three major behavioral states for rat 5 (entire recording). (2.1 MB PPT). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S8 ::: {.caption} ###### Bin Size Exploration The figure shows the effect of using different bin sizes to calculate neuronal ensemble correlations. We observed quantitative differences (the larger the bin size, the larger the correlations), but qualitatively the correlations profiles are equivalent, i.e., have very similar shapes. (404 KB PPT). ::: ::: {.caption} ###### Click here for additional data file. ::: This work was supported by a National Institutes of Health (NIH) grant (MALN) and by fellowships from the Pew Latin American program (SR), the Institut National de la Santé et de la Recherche Médicale (INSERM) (DG), the Fundação para a Ciência e a Tecnologia (FCT) (ESS), and the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) (JP). We thank G. Lehew and J. Meloy for manufacturing multielectrode arrays and for outstanding electronic support; C. Henriquez and J. Poorman for computer cluster administration; H. Wiggins for continuous technical support; R. Crist and J. Carmena for help in the initial phases; L. Oliveira, G. Wood, S. Halkiotis, and L. Hawkey for miscellaneous support; R. Costa and M. Engelhard for comments on the manuscript; and R. Stickgold and D. Schwartz for insightful discussions regarding deep-memory encoding. This paper is dedicated to the creation of the International Institute for Neuroscience of Natal, Brazil ([natalneuroscience.com](natalneuroscience.com)). **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** SR, DG, and MALN conceived and designed the experiments. SR, DG, ESS, YZ, S-CL, and JP performed the experiments. SR, ESS, YZ, and S-CL analyzed the data. ESS, YZ, S-CL, MLL, and MALN contributed reagents/materials/analysis tools. SR wrote the paper. Academic Editor: Wolfram Schultz, University of Cambridge AP : anteroposterior CSS : complex sensory stimulation CX : cerebral cortex DV : dorsoventral HP : hippocampus LFP : local field potential ML : mediolateral PU : putamen REM : rapid eye-movement SW : slow-wave TH : thalamus WK : waking

PubMed Central

2024-06-05T03:55:46.035777

2004-1-20

PMC314475

In the first month after our launch, the PDF of the "monkey-robot" article by Miguel Nicolelis received tens of thousands of downloads. We are not sure who downloaded the paper because we do not ask people to register at our site. We suspect, however, that its popularity is in part due to the widespread media coverage of the article (from *Time* magazine to Al-Jazeera and *Die Zeit*), which demonstrates a thirst for the original scientific paper and makes a strong argument for open access. We were even more surprised by the popularity of the "issue PDF," a whopping 72 MB file that contains the entire journal, cover to cover. The inaugural issue PDF was downloaded thousands of times within the first few weeks, even though we had printed 30,000 hard copies and distributed them widely. The Sky\'s the Limit (with Proper Attribution) {#s2} ============================================== But who are our users, and what do they do with the content? We know that *PLoS Biology* articles are used in a variety of educational settings. For example, the Nicolelis article has already been used for several high-school science projects and by a psychology student who compared the original research paper to its media derivatives. And the paper by Joseph DeRisi and colleagues on the malaria transcriptome has served as the basis for a continuing medical education drug discovery class and has been the topic of several undergraduate classes. Under the terms of the Creative Commons Attribution License, not only can *PLoS Biology* articles be reproduced and distributed without the need to obtain explicit permission, they can also be used for the publication of derivative works. Two *PLoS Biology* articles have already been entered in the *Internet Encyclopedia*. The source of the articles is clearly cited; it is also clear that they have been modified ed by the addition of extra links and information and that they are editable by users of the encyclopedia. Although this is an experiment in freely available and editable information, there will also be opportunities for entrepreneurs to produce derivative works with the type of added value that some users might wish to pay for. Open access provides free access to the research literature, but also provides publishers with new commercial opportunities. Once a significant body of full-text literature is available, it also becomes possible to use it for the development of new tools and resources for text- and data-mining and knowledge discovery. We plan to collaborate with developers of such tools. For their use---and for anybody else who likes their text "marked-up"---we make the XML version of our articles available. These are formatted according to the Journal Publishing DTD (Document Type Definition) from the United States government\'s National Library of Medicine, which provides a standard for archiving and exchanging XML versions of published documents. Overcoming Obstacles {#s3} ==================== A barrier for many potential users is that all our content---at least for now---is in English. We are delighted to hear that some of it is already being translated into other languages for local use. The feature article on the environmental benefits and risks of genetically modified crops, for example, will be republished (in Spanish) in the Argentinian environmental magazine *Gerencia Ambiental*. We hope that this will catch on---and urge anyone translating our content to let us know so that we can point others to the various language versions. We merely ask that the translators and their publishers acknowledge the authors and the source, by including a statement such as "this is a translation from the original article by Virginia Gewin published in *PLoS Biology*, DOI: [10.1371/journal.pbio.0000008](10.1371/journal.pbio.0000008)." Almost any translation is better than none for those excluded by language barriers, but quality control is a concern, and we are keen to collaborate with individuals or organizations who are interested in providing high-quality translations for some or all of our content on a regular basis. Internet connectivity is another obstacle. We know that some of the downloads of the issue PDF were transferred onto CD-ROMs that were copied and distributed in Uganda and Cambodia, areas in which Internet access is often slow and expensive. Other copies of the PDF were being used to create local hardcopies of the journal for communal use. We are happy to support these and related efforts to bring *PLoS Biology* content to readers by, for example, increasing the range of formats available at our Web site. Let us know what would help. Wanted: More Feedback {#s4} ===================== Besides the encouraging Web statistics, we have heard from many individual users since the launch of *PLoS Biology*. We\'d like to hear from even more. Tell us how you use *PLoS Biology*: your ideas might inspire others. As a way of building on the work and ideas of others, we have added a page on the PLoS Web site ([www.plos.org/creative\_uses](www.plos.org/creative_uses)) where we list some of the more creative and unusual uses of *PLoS Biology*. Let us know what you do with *PLoS Biology*, or what you\'d like to do, and we\'ll see what we can do to make it possible. Barbara Cohen can be contacted at E-mail: <bcohen@plos.org>.

PubMed Central

2024-06-05T03:55:46.040106

2004-1-20

PMC314476

As cells can only arise from cells that already exist, continuity of life depends on the highly regulated sequence of events that control cell division. This process is mediated by a complex macromolecular structure called the mitotic spindle. The most conspicuous components of the spindle are microtubules, which are made of tubulin and other associated proteins. In most animal cells---body cells and male germline cells (spermatocytes)---spindle assembly is orchestrated by organelles called centrosomes, which actively polymerize (that is, add tubulin subunits) and stabilize microtubules. The spindles found in these cells are known as *astral* because of the star-shaped asters---structures made of centrosome-anchored microtubules---that can be observed associating with each spindle pole. Some cells---such as the cells of the female germline (oocytes)---do not contain centrosomes, and the chromosomes themselves seem to arrange and stabilize the microtubules into spindles. These spindles are referred to as *anastral*. To gain insight into the mechanisms of spindle assembly, scientists are increasingly relying on techniques that allow them to directly observe dynamic, complex processes in the living cell. Using time-lapse microscopy of fluorescently labeled fruitfly (Drosophila melanogaster) spermatocytes, Cayetano Gonzalez and his colleagues at the European Molecular Biology Laboratory in Germany (and now at the Centro Nacional de Investigaciones Oncológicas in Spain) have been able to observe the assembly and sorting of microtubules of noncentrosomal origin in cells that contain centrosomes. The task of flagging such microtubules is complicated by the fact that centrosomes become quite active microtubule organizers once cell division begins. Thus, as soon as the membrane around the nucleus breaks down, microtubules from the centrosome invade the nuclear region, making it hard to identify any noncentrosomal microtubules that might appear. To get around this problem, Elena Rebollo in the Gonzalez lab set up two experimental conditions under which centrosomes remain functional but are kept affixed to the cell membrane---and, therefore, away from the nucleus---in Drosophila spermatocytes. One takes advantage of a genetic mutation (called *asp*, for abnormal spindle); the other uses a transient treatment with a drug (called colcemid) that depolymerizes microtubules. In these modified cells, microtubules can be seen growing not only over the membrane-bound centrosomes, as expected, but also over the nuclear region, away from the centrosomes. Nucleation, or formation, of such noncentrosomal microtubules has a relatively late onset, starting only once chromosomes are condensed, and takes place on the inner side of the remnants of the nuclear envelope. In a fraction of cells, these microtubules are sorted into bipolar spindle-shaped structures, highly reminiscent of the anastral spindles found in oocytes. Chromosome segregation---a critical stage of cell division---and cell division itself tend to be aberrant in these cells. These results, Rebollo et al. propose, strongly suggest that microtubules of noncentrosomal origin may significantly contribute to spindle assembly even in cells that contain active centrosomes. Moreover, by facilitating the nucleation of such noncentrosomal microtubules, the degraded nuclear envelope may play a previously unsuspected role in spindle assembly in Drosophila spermatocytes. It is unlikely, the researchers also conclude, that the anastral spindles they have observed can fill in as a backup to ensure successful cell division. More likely, they argue, both centrosomal and noncentrosomal microtubules are required for proper spindle assembly and robust cell division in cells with centrosomes. As the authors point out, Drosophila is a rich model system that should help scientists further investigate the intricacies of spindle assembly. The answers will help us understand how the cell executes one of its most important duties: safeguarding genomic stability for future generations. ::: {#pbio-0020026-g001 .fig} ::: {.caption} ###### Centrosome-independent spindle assembly :::  :::

PubMed Central

2024-06-05T03:55:46.040780

2004-1-20

PMC314477

RNA interference (RNAi) has been called "one of the most has exciting discoveries in biology in the last couple decades," and since it was first recognized by Andrew Fire et al. in 1998, it has quickly become one of the most powerful and indispensable tools in the molecular biologist\'s toolkit. Using short double-stranded RNA (dsRNA) molecules, RNAi can selectively silence essentially any gene in the genome. It is an ancient mechanism of gene regulation, found in eukaryotes as diverse as yeast and mammals, and probably plays a central role in controlling gene expression in all eukaryotes. In the lab, RNAi is routinely used to reveal the genetic secrets of development, intracellular signaling, cancer, infection, and a full range of other phenomena. But can the phenomenon hailed by the journal *Science* as the "Breakthrough of the Year" in 2002 break out of the lab and lead to novel therapies as well? Pharmaceutical giants are hoping so, and several biotech companies have bet their futures on it, but not everyone is sanguine about the future of RNAi therapy. At the heart of its promise as a powerful therapeutic drug lies the exquisite selectivity of RNAi: like the fabled "magic bullet," an RNAi sequence seeks out and destroys its target without affecting other genes. The clinical applications appear endless: any gene whose expression contributes to disease is a potential target, from viral genes to oncogenes to genes responsible for heart disease, Alzheimer\'s disease, diabetes, and more. But for all its promise, RNAi therapy is a long way from entering the clinic. While it is a proven wunderkind in the lab, to date no tests have been done in humans, and only the most modest and circumscribed successes have been demonstrated in animals. The road to clinical success is littered with great ideas that have come a cropper along the way, including two other RNA-based therapies, antisense and ribozymes, both of which showed promise at the bench but have largely stumbled before reaching the bedside. Is RNAi also likely to fall short? Or is it different enough to make this third try the charm? Clinical Naïveté, Mysterious Mechanisms {#s2} ======================================= To be a successful drug, a molecule must overcome a long set of hurdles. First, it must be able to be manufactured at reasonable cost and administered safely and conveniently. Then, and even more importantly, it must be stable enough to reach its target cells before it is degraded or excreted; it must get into those cells, link up with its intracellular target, and exert its effect; and it must exert enough of an effect to improve the health of the person taking it. And, finally, it must do all this without causing significant toxic effects in either target or nontarget tissues. No matter how good a compound looks in the lab, if it fails to clear any one of these hurdles, it is useless as a drug. For RNA-based therapies, manufacture has been seen as a soluble problem, while delivery, stability, and potency have been the most significant obstacles. "There was a lot of clinical naïveté" in the early days of antisense and ribozymes, according to Nassim Usman, Vice President for Research and Development at Sirna Therapeutics in Boulder, Colorado. "Compounds were pushed into the clinic prematurely." Sirna began as the biotech startup Ribozyme Pharmaceuticals, which tried to develop ribozymes to treat several conditions, including hepatitis C. A ribozyme is an RNA molecule whose sequence and structure allow it to cleave specific target RNA molecules (see [Figure 1](#pbio-0020028-g001){ref-type="fig"}). "The initial results with hepatitis C were not that inspiring," says Usman, because the molecule they used had low potency and a short half-life once in the body. Despite "enormous doses," the viral load was not significantly affected. "It just didn\'t have the characteristics to be a drug," he says. No ribozyme has yet been approved for use by the United States Food and Drug Administration (FDA). ::: {#pbio-0020028-g001 .fig} Figure 1 ::: {.caption} ###### Ribozymes A ribozyme binds to a specific mRNA, cleaves it, and thus prevents it from functioning. :::  ::: Similarly, despite much initial enthusiasm, attempts to develop antisense drugs have been largely disappointing. Antisense is a single strand of RNA or DNA, complementary to a target messenger RNA (mRNA) sequence; by pairing up with it, the antisense strand prevents translation of the mRNA (see [Figure 2](#pbio-0020028-g002){ref-type="fig"}). At least that was the theory, and early clinical results seemed to support the theory: antisense drugs effectively reduced tumor sizes in anticancer trials and viral loads in antiviral trials. But closer inspection revealed these results were largely due to an increase in production of interferons by the immune system in response to high doses of the foreign RNA, rather than to specific silencing of any target genes. (The relatively high proportion of C--G sequences in antisense mimics bacterial and viral genes, thus triggering the immune response.) When the antisense dose was lowered to prevent the interferon response, the clinical benefit largely disappeared as well. Thus, rather than being a highly specific therapy, antisense seemed to be a general immune system booster. ::: {#pbio-0020028-g002 .fig} Figure 2 ::: {.caption} ###### Antisense Antisense DNA or RNA binds to a specific mRNA and prevents it from being translated into protein. :::  ::: But as long as patients were getting better, does it matter what the mechanism was? "It doesn\'t matter if you are a patient, but it does matter if you are trying to develop the next drug," says Cy Stein, Associate Professor of Medicine and Pharmacology at Columbia University College of Physicians and Surgeons in New York City. Stein has researched antisense for more than a decade. "If you get the mechanism wrong, you\'re not going to be able to do it." To date, only one antisense drug has received FDA approval. Vitravene, from Isis Pharmaceuticals in Carlsbad, California, is used to treat cytomegalovirus infections in the eye for patients with HIV. Vitravene is actually a DNA antisense drug, which binds to viral DNA, though, says Usman, "it\'s unclear whether it actually works by an antisense mechanism." Stein expresses a similar skepticism about the mechanism of a second antisense drug, Genasense. Genasense is a DNA-based treatment that targets Bcl-2, a protein expressed in high levels in cancer cells, which is thought to protect them from standard chemotherapy. The FDA is currently reviewing an application for Genasense, based on promising results in the treatment of malignant melanoma. RNAi: A Natural Alternative {#s3} =========================== Growing disillusionment with antisense and ribozymes coincided with the discovery of RNAi and the realization that it was a far more potent way to silence gene expression. RNAi uses short dsRNA molecules whose sequence matches that of the gene of interest. Once in a cell, a dsRNA molecule is cleaved into segments approximately 22 nucleotides long, called short interfering RNAs (siRNAs) (see [Figure 3](#pbio-0020028-g003){ref-type="fig"}). siRNAs become bound to the RNA-induced silencing complex (RISC), which then also binds any matching mRNA sequence. Once this occurs, the mRNA is degraded, effectively silencing the gene from which it came. (Details of the structure and function of the RISC are still largely unknown, but it is thought to act as a true enzyme complex, requiring only one or several siRNA molecules to degrade many times that number of matching mRNAs.) ::: {#pbio-0020028-g003 .fig} Figure 3 ::: {.caption} ###### RNAi RNAi is a recently described naturally occurring process in which small RNA molecules activate a cellular process that results in the destruction of a specific mRNA. :::  ::: The extraordinary selectivity of RNAi, combined with its potency---in theory, only a few dsRNAs are needed per cell---quickly made it the tool of choice for functional genomics (determining what a gene product does and with what other products it interacts) and for drug target discovery and validation. By "knocking down" a gene with RNAi and determining how a cell responds, a researcher can, in the course of only a few days, develop significant insight into the function of the gene and determine whether reducing its expression is likely to be therapeutically useful. But does RNAi have a better chance to succeed as a drug than antisense or ribozymes? "The fundamental difference favoring RNAi is that we\'re harnessing an endogenous, natural pathway," says Nagesh Mahanthappa, Director of Corporate Development at Alnylam Pharmaceuticals in Cambridge, Massachusetts, the second of two major biotech company developing RNAi-based therapy. The exploitation of a pre-existing mechanism, he says, is the main reason RNAi is orders of magnitude more potent than either of the other two types of RNA strategies. Delivery, Delivery, Delivery {#s4} ============================ More potent in the test tube, at least. But stability and delivery are also the major obstacles to successful RNAi therapy, obstacles that are intrinsic to the biochemical nature of RNA itself, as well as the body\'s defenses against infection with foreign nucleotides. "For the strongest reasons, you can\'t get away from this," says Stein. "The problem is that a charged oligonucleotide will not pass through a lipid layer," which it must do in order to enter a cell. John Rossi, Director of the Department of Molecular Biology at City of Hope Hospital in Duarte, California, who has worked on RNA-based therapies for 15 years, concurs. "The cell doesn\'t want to take up RNA," he says, which makes evolutionary sense, since extracellular RNA usually signifies a viral infection. Injected into the bloodstream, unmodified RNA is rapidly excreted by the kidneys or degraded by enzymes. To solve the problem of cell penetration, most researchers have either complexed the RNA with a lipid or modified the RNA\'s phosphate backbone to minimize its charge. Mahanthappa thinks the complexing approach is unlikely to be a simple solution, since drug approval would require independent testing of the lipid delivery system as well. Instead, Alnylam is pursuing backbone modification. "Some minimal modification is going to be necessary" to increase cell uptake and to improve stability in the blood stream, Mahanthappa says. "What we have learned from the antisense field is that even without other delivery strategies, when you administer RNA at sufficient doses, if it\'s stable, it gets taken up by cells." "Anything that can be done to increase half-life in circulation would improve delivery," says Judy Lieberman, a Senior Investigator at the Center for Blood Research and Associate Professor of Pediatrics at Harvard Medical School in Cambridge, Massachusetts. But that may not be the only problem, she cautions. Lieberman\'s lab recently demonstrated the ability of RNAi to silence expression of the Fas gene in mice, protecting them from fulminant hepatitis. Fas triggers apoptosis, or programmed cell death, in response to a variety of cell insults. In her experiment, Lieberman delivered the RNA by high-pressure injection into the tail. The RNA got to the liver, silenced Fas, and protected the mice from hepatitis. However, a significant fraction of animals died of heart failure, brought on because the injection volume was about 20% of the mouse blood volume. Such a delivery scheme simply will not work in humans. "Delivery to the cell is still an obstacle," Lieberman explains. "Unless you really focus on how to solve that problem, you\'re not going to get very far." Unanswered Questions {#s5} ==================== Even assuming delivery problems can be solved, other questions remain, including that of whether therapeutic levels of RNAi may be toxic. Mahanthappa says, "The conservative answer is we just don\'t know. The more aggressive answer is there\'s no reason to think so." Rossi isn\'t so sure. "The target of interest may be in normal cells as well as cancer cells," he says. "That\'s where you get toxicity." But if small RNAs can be delivered to target cells efficiently and without significant toxicity, will they be effective medicines? Usman of Sirna is confident they will be. "If you can get it there, and if it\'s in one piece, there no doubt in our minds that it will work," he says. To date, numerous experiments in animal models suggest RNAi can downregulate a variety of target genes effectively. However, there are still two unanswered questions about whether that will translate into effective therapy. The first is whether RNAi\'s exquisite specificity is really an advantage beyond the bench. "It\'s unclear whether highly specific drugs give you a big therapeutic effect," says Cy Stein. For instance, he says, "most active antitumor medicines have multiple mechanisms of action. The more specific you make it, the less robust the therapeutic activity is likely to be." Rossi agrees: "Overspecificity has never worked," he says. The second question is what effect an excess of RNA from outside the cell will have on the normal function of the RISC, the complex at the heart of the RNAi mechanism. The number of RISCs in the cell is unknown, and one concern is that the amount of RNA needed to have a therapeutic effect may occupy all the available complexes. "We are usurping a natural cell system that is there for some other purpose, for knocking out endogenous gene function," says Rossi. With the introduction of foreign RNA, will the system continue to perform its normal function as well, or will it become saturated? "That\'s the big black box in the field," he says. Looking Ahead to the Clinic {#s6} =========================== Despite the questions and unsolved problems, Sirna, Alnylam, and several other companies are moving ahead to develop RNAi therapy; indeed, some outstanding questions are probably only likely to be answered in the process of therapeutic development. The first applications are likely to be in cancer (targeting out-of-control oncogenes) or viral infection (targeting viral genes). To avoid some of the problems of delivery, initial trials may deliver the RNA by direct injection into the target tissue (for a tumor, for instance) or ex vivo, treating white blood cells infected with HIV, for example. Having spent a decade trying to develop ribozymes, says Usman, Sirna is prepared for the rough road it faces. "We haven\'t solved all the problems, but we know how to proceed to work through them. It\'s no surprise to us---we\'ve seen this movie before." Usman expects Sirna to file an Investigational New Drug Application with the FDA by the end of 2004 and to have a human clinical trial in progress in 2005. "Without a doubt, there will be an RNAi-based drug within ten years," Usman predicts. Stein isn\'t so sure and thinks that too much is still to be learned about RNAi and the body\'s reaction to it to be confident that RNA-based therapies will ultimately be successful. "The whole field was founded on the belief it was rational, but there are huge gaps in our knowledge, and so you need a bit of luck to be successful," he says. "I think people are surprised at how complicated it is, but why should it be any other way? It\'s an intellectual conceit to think that nature is simple." Richard Robinson is a freelance science writer from Sherborn, Massachusetts, United States of America. E-mail: <rrobinson@nasw.org> dsRNA : double-stranded RNA FDA : Food and Drug Administration mRNA : messenger RNA RISC : RNA-induced silencing complex RNAi : RNA interference siRNA : short interfering RNA

PubMed Central

2024-06-05T03:55:46.041632

2004-1-20

PMC314478

Plasmodium, the microscopic parasite that causes malaria, passes through two hosts, two reproductive modes, four habitats, and over half-a-dozen distinct developmental stages in one lifecycle. When a Plasmodium-infected mosquito bites a human, it injects the parasite---sequestered in the mosquito\'s salivary glands in its sporozoite stage---into the victim\'s bloodstream. Within hours, the sporozoites invade the liver---a critical stage for establishing infection---and spend the next few weeks asexually dividing inside liver cells, eventually releasing thousands of merozoites into the bloodstream. Merozoites quickly invade red blood cells and begin a second round of asexual proliferation. The infected cells rupture and die, releasing more parasites and toxins. The toxins cause malaria\'s characteristic fever and chills, and the liberated merozoites initiate another cycle of red blood cell attacks. An unresolved question has been how the circulating sporozoites reach the liver cells in the first place, since liver cells are separated from the bloodstream by a layer of endothelial and Kupffer cells, which form the walls of the liver capillaries. (Kupffer cells project into the bloodstream and remove contaminants.) Having identified a protein required for sporozoite migration through the capillary lining, Tomoko Ishino, Masao Yuda, and their colleagues at Mie University School of Medicine in Japan may have found an answer. Only four of the roughly 150 vertebrate-infecting Plasmodium species affect humans. P. falciparum, the most pathogenic of the human-infecting species, is closely related to avian and rodent species. One rodent species,---P. berghei---shares fundamental aspects of structure, physiology, and lifecycle with P. falciparum and so serves as a model for the human parasites. Since sporozoites must infect mosquito salivary glands before they can infect the mammalian liver, Yuda\'s team searched for sporozoite genes that are predicted to encode secretory or membrane proteins and are expressed only in mosquito salivary glands. Their search revealed a coding region conserved in several species of Plasmodium. Tracing the gene\'s activity through the parasite\'s life cycle, Yuda\'s team confirmed that it was expressed only in sporozoites in the mosquito salivary gland---not in the mosquito midgut, where sporozoites are produced after mosquitoes feed on the blood of an infected person. The corresponding protein was localized to micronemes, specialized secretory organelles found at the front end of malaria parasites. Because micronemes are known to play a central role in Plasmodium motility and invasion, the researchers predicted this protein would also be important in migrating to or invading liver cells. They named the protein SPECT, for sporozoite microneme protein essential for cell traversal. Yuda\'s team tested SPECT\'s function by generating *spect*-disrupted mutants and observing how the altered parasites affected their hosts. spect disruption did not affect parasite proliferation in rat red blood cells or interfere with parasite development in the mosquito midgut or salivary glands, but it did have an effect on the parasite\'s ability to infect the liver. Rats injected with *spect*-disrupted parasites had significantly lower levels of liver infection than rats injected with nonmutant parasites. Since it was unclear whether the *spect*-disrupted mutants lost their infectivity or simply could not pass through the cell layer, the researchers inoculated human liver cells with the mutants and found that they infected the cells normally. Yuda\'s team also tested SPECT\'s impact on sporozoite cell-passage ability; if the mutants couldn\'t reach the liver cells, they couldn\'t infect them. *spect*-disrupted parasites completely lost their ability to pass through cells. Since traversal of the cellular barrier between liver cells and the circulatory system is a crucial step in malarial infection, the authors conclude, SPECT and other proteins involved in shuttling sporozoites into liver cells could be effective targets for malaria treatment and prevention. ::: {#journal.pbio.0020032-g001 .fig} ::: {.caption} ###### Sporozoite migration to hepatocytes :::  :::

PubMed Central

2024-06-05T03:55:46.043134

2004-1-20

PMC314479

Conventional wisdom says that people deficient in one sense---such as vision or hearing---often acquire heightened acuity in another. And some studies support this notion by showing that areas of the brain known to control vision can respond to other forms of sensory stimuli in persons without sight. These adjustments, of course, take place over the lifetime of an individual. Now it appears that similar adjustments may occur over evolutionary time. Investigating the deterioration of olfactory receptor (OR) genes in primates, Yoav Gilad and his colleagues at the Max Planck Institute for Evolutionary Anthropology in Germany and the Weizmann Institute in Israel found a correlation between the loss of OR genes and the acquisition of full trichromatic color vision. OR genes---the molecular basis for the sense of smell---form the largest gene superfamily in mammalian genomes. But a high percentage of these genes are "pseudogenes," DNA sequences that are remnants of genes that are no longer functional. Following an evolutionary "use-it-or-lose-it" rule, pseudogenes tend to evolve in larger gene families where there\'s no selective advantage in having, say, 100 versus 120 genes. While humans, nonhuman primates, and mice have roughly the same number of OR genes, in humans a much higher percentage of these are pseudogenes, at 60%, while nonhuman apes have about 30%, and the mouse has about 20%. Reliance on the sense of smell, it appears, decreases for animals that develop a dependence on other senses, such as hearing or sight, to survive. In characterizing this high proportion of pseudogenes, Yoav Gilad et al. asked: Is this characteristic of all primates? If not, at what point in primate evolution did the increase occur? Looking at 19 primate species---including one human, four apes, six Old World monkeys, seven New World monkeys, and one prosimian---Gilad et al. randomly sequenced 100 distinct OR genes from each of the species. The team found that Old World monkeys had roughly the same percentage of OR pseudogenes as nonhuman apes, but a much higher percentage than New World monkeys---except for one, the howler monkey. The percentage of OR pseudogenes in the howler monkey was much closer to that seen in the Old World monkeys and apes than in its New World cousins. The sense of smell, it appears, deteriorated both in the ape and Old World monkey lineage and in the howler monkey lineage. Since Old World monkeys, apes, and the howler monkey do not share an exclusive common ancestor, this deterioration must have evolved independently in both groups. Surprisingly, howler monkeys share another sensory feature with apes and Old World monkeys: trichromatic color vision. In trichromatic color vision, three retinal protein pigments, called opsins, absorb various wavelengths of light, which the brain processes to produce full-color images. Apes and Old World monkeys carry three opsin genes, and most New World monkeys carry only two, though females can sometimes have three. Only howler monkeys routinely have three genes occurring in both sexes. Thus, full trichromatic vision evolved twice in primates---once in the common ancestor of apes and Old World monkeys, about 23 million years ago, and once in the howler monkey lineage, about 7--16 million years ago. The evolution of color vision, the authors propose, coincided with a growing complement of OR pseudogenes and a deterioration of the sense of smell. Gilad et al. suggest that investigating the types of visual cues required for finding food may shed light on the nature of this connection. ::: {#pbio-0020033-g001 .fig} ::: {.caption} ###### Howler monkey :::  :::

PubMed Central

2024-06-05T03:55:46.043832

2004-1-20

PMC314480

How does your brain pass the time while you\'re sleeping? If you\'ve ever wrestled the demons of insomnia, you know what sleepless nights can do to your mental agility. Sleep cycles in mammals are characterized by two distinct, successive sleep stages: slow wave and rapid eye movement (REM). Both stages of sleep have uniquely associated electrical activity in the brain, which neuroscientists can measure by placing elec-trodes on the brain during sleeping and waking states. What neuroscientists can\'t easily measure is the purpose of these two sequential sleep stages. The notion that sleep helps to improve memory was introduced over 80 years ago. Since then, several studies have demonstrated that sleep deprivation following the acqui-sition of a new memory strongly impairs its consolidation. Insight into the mechanisms underlying this effect came from the observation that neuronal activity patterns detected during waking reappear during ensuing sleep, suggesting that newly acquired "memory traces" may be replayed in the brain to solidify neural connections and thus "consolidate" memory. These reverberating patterns of activity have been observed in both mammals and birds, pointing to a very general biological phenomenon. Still, the relationship between brain reverberation and memory consolidation remains unclear for a number of reasons. First, studies to date have observed only subtle, short-lived reverberations lasting less than an hour and can\'t explain the memory-disrupting effects of sleep deprivation applied several hours and even days after initial memory encoding. And since brain reverberation in mammals has only been investigated in the hippocampus and cerebral cortex, it is unclear whether the phenomenon is specific to this neural circuit or is a more general property of the brain. Furthermore, reverberation studies have so far relied on neural activity measured in animals that were highly trained on specific laboratory tasks and therefore may simply not be representative of the acquisition of new memories. Finally, experience-dependent neural reverberation has been detected in both phases of sleep as well as waking, but no quantitative comparison of the different contributions of each state has been established. In a study designed to address these concerns, Sidarta Ribeiro and his colleagues at Duke University in Durham, North Carolina, recorded over a hundred neurons continuously over the course of the normal sleep\--wake cycle in rats, focusing on four major forebrain areas that are essential for rodent-specific behaviors. Halfway through the recording time, animals were transiently allowed to explore four strictly novel objects, each of them designed to provide different spatial and sensory cues. The researchers found that in all the forebrain areas examined the neuronal firing patterns recorded when the rats initially explored the new objects reverberated for up to 48 hours after these objects were removed. The reverberation of neuronal activity sampled when rats explored familiar environs was insignificant. Reverberation was most significant during slow-wave sleep (a state that accounts for nearly 40% of a rat\'s life), decreased during waking periods, and was highly variable during REM sleep. In this study, Ribeiro et al. demonstrate that long-lasting neuronal reverberation following novel waking experiences can occur in several forebrain sites and is strongly enhanced during slow-wave sleep. Because neuronal reverberations are sustained for long periods, this may provide a mechanism to recall and amplify memories until they are effectively stored. On the basis of differences observed between REM and slow-wave sleep in this and previous studies, the authors propose that the two phases of sleep play separate and complementary roles in memory consolidation. Thus, the two stages of sleep give the brain a chance to process the novel events of the day in peace. ::: {#pbio-0020037-g001 .fig} ::: {.caption} ###### Sleeping rats :::  :::

PubMed Central

2024-06-05T03:55:46.044330

2004-1-20

PMC314481

Genes use a simple language---written in the molecules of DNA---to build thousands of proteins in a dizzying variety of sizes and shapes. With only four different nucleotide building blocks, DNA codes for the 20 different amino acids (each with their own structures and properties) that provide the foundation for the enormous diversity of protein form and function. This diversity makes the systematic study of all the proteins of a given organism (called proteomics), a challenging enterprise. Interactions between proteins underlie nearly every fundamental process within the cell. They can form higher-order multiprotein complexes like those involved in transcription and replication, help transport proteins to their proper location in the cell, and participate in signaling pathways. Because of their importance, disruption of these interactions can have disastrous consequences. For example, the loss of the ability of a normal cellular protein called Src to bind to certain other proteins can be associated with cancer progression. The determinants of these interactions are poorly understood, but in many cases these interactions are mediated by small pieces of the proteins, which are called peptides. Peptides serve as the starting point for the novel strategy reported in this issue. Gianni Cesareni and colleagues have added to the repertoire of proteomic analysis by devising a global strategy to investigate protein--protein interactions on an organismal level using yeast as a model organism. The authors select a protein of interest from yeast, which can be thought of as the "bait" for which they wish to identify protein-binding partners. They start by looking at a number of different previously identified peptides that bind the bait protein. Commonalities between the sequences of these peptides form the "consensus" binding sequence, a base framework of protein sequence from which many possible variations can be derived. Since the protein sequences of all proteins (the proteome) in yeast can be deduced from the sequenced genome, the authors can scan the proteome for proteins that contain the consensus, or a closely related, sequence. These proteins could potentially bind the bait protein. Peptide sequences from these identified proteins are synthesized chemically and arrayed on a membrane, which is bathed in a solution containing the bait protein. After washing off the excess bait protein, they can figure out where it remains on the membrane and therefore tell which peptides the bait protein has bound. The proteins corresponding to these peptides are candidate binding proteins that are validated by further experimentation. The protein--protein interactions identified by this approach can be used to extend the network of known interactions in the proteome. This will enable researchers to draw functional linkages between proteins, whether they are involved in a basic biological process or in human disease. By examining whole families of proteins, it may also aid in elucidating the underlying determinants of binding specificity, which would provide clues to the biomechanisms underlying cell processes. These insights could lead to methods for manipulating these interactions, perhaps even in cases of human disease, as in the case of Src and cancer. This approach can readily be applied to the proteomes of more complex organisms like humans and adds to the growing number of experimental strategies available to researchers in proteomics. ::: {#pbio-0020038-g001 .fig} ::: {.caption} ###### Protein interaction network :::  :::

PubMed Central

2024-06-05T03:55:46.044986

2004-1-20

PMC314482

One of the central problems for much of the 20th century was how to reconcile genetic stability with evolutionary change. Genomic fidelity was thought to arise from an inherent invariability in the DNA structure itself. Biologists now know that DNA constantly undergoes modifications as it unwinds, replicates, condenses, twists, and untwists. This dynamic interplay produces both stability and variation---and occasionally genetic damage. If DNA damage goes unrepaired, it can disrupt chromosomal integrity and may lead to cancer and other diseases. When the DNA double helix breaks, the cell must enlist a number of proteins to repair the broken DNA ends, but much remains to be learned about the molecular mechanisms involved. Tracking a protein that binds to single strands of DNA during replication and recombination in living yeast cells, Xuan Wang and James Haber report that this protein plays a role in at least two key steps in the repair of double-strand breaks in DNA. When double-strand breaks occur, the cell mounts a search for similar (homologous) sequences that can be used as a template to repair the damaged sequence. If successful, the broken DNA molecule basepairs with the homologous region and forms a complex, ultimately replacing the damaged sequence with a similar sequence. In yeast---which serves as a stand-in for higher eukaryotes, including humans---this "strand invasion" process requires both an exchange protein, called Rad51, and a single-stranded DNA-binding protein, called RPA (replication protein A). Single-stranded binding proteins bind to regions of DNA that are opened up during replication. They also bind to strands when broken ends of DNA are cut by enzymes that leave long single-stranded tails. RPA proteins are thought to facilitate the formation of Rad51 polymers, or filaments, on single-stranded DNA by clearing away structures that block Rad51\'s path. The growing filament searches for homologous DNA sequences and promotes the invasion of the single strand, preparing it to copy the homologous template by "repair DNA synthesis," which patches up the lesion. To investigate how RPA functions in double-strand break repair in a living organism, Wang and Haber created cells with a double-strand break at a specific site and monitored the activity of proteins recruited to repair the damage. With this approach, the researchers could observe these interactions in living yeast to determine what role RPA plays in repairing DNA damage and how it works with the Rad51 protein. The authors show that as soon as a double-strand break occurs, the RPA protein binds to the exposed strand ends, before the Rad51 protein does. This is not unexpected, because this binding order supports the model that RPA prepares the way for Rad51, perhaps by stabilizing the strand long enough for Rad51 filaments to establish themselves. The surprise was that RPA appears to be necessary even after Rad51 binds to the DNA strand, perhaps by stabilizing the interaction with homologous DNA sequences. That RPA is required for successful repair is supported by evidence that a particular mutated form of RPA can stimulate Rad51 DNA binding normally, but inhibits strand exchange and template copying, thus preventing repair of DNA damage. Wang and Haber\'s work highlights the complex repertoire of DNA--protein and protein--protein interactions that manage and manipulate the genome in the service of genomic stability. The study of DNA repair mechanisms in living cells---a daunting task---promises to lend valuable insights into the truly dynamic nature of maintaining genome stability. ::: {#pbio-0020074-g001 .fig} ::: {.caption} ###### Repair of double-strand breaks involves invasion of the homologous region, displacement, and DNA synthesis to fill in the gap :::  :::

PubMed Central

2024-06-05T03:55:46.045508

2004-1-20

PMC314483

Developing countries today face a wide range of needs, from more secure food supplies to cheap and effective medicines. One factor that almost all these needs have in common is that satisfying them adequately will not occur without the use of modern science. These same countries also face a range of political dilemmas---such as whether to accept the growing of genetically modified crops or how to adapt to the impact of climate change---that also require a knowledge of science, albeit of a slightly different nature. From both points of view, development can be characterised as the process of putting scientific and technical knowledge into practice. Conversely, it is important that building the capacity to absorb and make use of scientific and technical knowledge must be placed at the heart of the development aid efforts if these are to be successful in achieving their goals. But knowledge will not reach those who can benefit from it unless it has been effectively communicated to individuals with the power and skills to put it into practice, whether those are government officials and decision-makers, community groups and their representatives, or even nongovernmental organisations. Formal education, of course, has a key function here. But so does the informal education provided through the media. Furthermore, information and communications technologies (ICTs) have an important role to play in this process by reducing or eliminating the transactional (nonproduction) costs of communicating knowledge about science and technology. At the same time, it is important that the practice of science communication reflects the fact that it takes place in social context. In other words, it is not just a question of conveying information, but also of engaging the potential users of that information. The need is to encourage dialogue and eventually to empower those to whom the information is being provided so that this information can be applied in a practical and useful way. SciDev.Net {#s2} ========== It was with this in mind that the Science and Development Network (SciDev.Net) was launched in December 2001 as a source, through its Web site ([www.scidev.net](www.scidev.net)), of online news and analysis about the role of science and technology in meeting the needs of the developing world. Much of the material we use is taken from the science journals *Nature* and *Science*, both of which provide us with free access for up to four articles each week, the selection being based on a decision about which articles---ranging from news items or editorials to full scientific papers---are directly relevant to the needs of developing countries. In addition, other news articles are contributed by staff writers and a growing team of correspondents, including science journalists in South Africa, India, Tanzania, Brazil, Colombia, and China. We also summarise and link to relevant news stories, feature items, and opinion articles from media outlets around the world. An equally important part of the Web site are our dossiers. Dossiers are comprehensive, online guides to topical issues at the interface of science, technology, and society. Each dossier is compiled using advice from an international panel of experts and provides authoritative insight into a topic at the heart of international debate. So far, we have produced six dossiers on the following topics: intellectual property rights, climate change, indigenous knowledge, genetically modified crops, the ethics of biomedical research, and the brain drain. Those currently in the works focus on HIV/AIDS, research policy, genomics, and biodiversity. Target Audience {#s3} =============== Our target audience is not easy to define, as it is made up of individuals from many different groups, from university researchers and teachers to government officials and aid agency staff. In broad terms, however, it can be taken to include all those with a professional or personal interest in the interactions between science, technology, and development. The number of registrants has risen steadily since the Web site was launched in December 2001 and now stands at more than 7,000. About half of our registered users come from the developing countries themselves. Of these, 43% come from Latin America---a reflection of various factors, including the size of its scientific community and its relatively high level of Internet connectivity. Sub-Saharan Africa comprise about 25%, and South Asia (mainly India) makes up 20%. For those with an interest in science activities in a particular geographical region, we link together coverage into so-called regional gateways on the Web site. At present, there are four of these---covering Latin America, the Middle East, South and East Asia, and sub-Saharan Africa. Each gateway therefore has a strong regional identity, and in some cases this is reinforced by the translation of material---either in full text or as a summary---into local languages. The Latin American gateway, for example, carries three languages, and material can appear in one, two, or all three of these, with summaries in the others. Channels of Communication {#s4} ========================= Reflecting a commitment to the idea that scientific knowledge will only be effectively communicated if individuals are adequately trained to carry out this task, we are engaged not only in presenting information on our Web site, but also in helping to generate a capacity within developing countries to improve channels of communication for this knowledge. One way in which we do this is through workshops on science communication. Some of these have a broad focus, seeking to help stimulate a national or regional debate on the importance of science communication, how it can be best carried out, and the challenges it faces. Such, for example, was the goal of two roundtables on the theme of "science communication and sustainable development" organised in August 2002 during the World Summit on Sustainable Development in Johannesburg, South Africa. Similarly, a meeting launching our activities in Latin America was held in Sao Paulo in May this year under the title "Science Communication and Development in Latin America." Other events to encourage scientific communication have been more specific. In particular, we organised (jointly with the United Nations Educational, Scientific, and Cultural Organisation) a five-day workshop for African women communicators in Kampala, Uganda, in April 2003 on the use of ICTs to report on the science of HIV/AIDS ([Figure 1](#pbio-0020011-g001){ref-type="fig"}). A similar event took place in Chennai, India, in November 2003, with participants from seven Asian nations. ::: {#pbio-0020011-g001 .fig} Figure 1 ::: {.caption} ###### A Workshop on the Use of ICTs to Report on the Science of HIV/AIDS This workshop was held in Kampala, Uganda, in April 2003. (Photograph: SciDev.Net.) :::  ::: Partnerships {#s5} ============ Building the communication channels that will allow science to be fully integrated into the development process at all levels---from governments down to communities---is a major challenge that involves many participants. SciDev.Net does not pretend to be more than one actor and is keen to establish partnerships with others in this field, convinced both of the synergies that can be gained through collaboration and of the contribution that ICTs can make to this (for example, by sharing information and links between Web sites). One of the major challenges that we face is to build up our role as a platform for the voice of the developing world on the issues that we cover. We already do this to a certain degree through our correspondents, our regional networks of contributors, and the participation of scientists from developing countries who are on the advisory panels for our dossiers. But we are very aware that more effort is needed. Despite this, we already feel that we are beginning to make a mark. A staff member of the United States\' National Academy of Sciences recently informed us: "We hosted a group of visitors from seven African scientific academies here in Washington last week, and they spoke highly of the SciDev project ... you are becoming a truly global newsletter!" ::: {#pbio-0020011-g002 .fig}  ::: Further Information [www.scidev.net](www.scidev.net) David Dickson is the director of the Science and Development Network (SciDev.Net), located in London, United Kingdom. E-mail: <david.dickson@scidev.net> ICTs : information and communications technologies

PubMed Central

2024-06-05T03:55:46.045928

2004-1-20

PMC317275

Introduction {#s1} ============ Two different pathways of spindle assembly are known to operate in the animal kingdom. The first, observed in somatic as well as in male germline cells, requires the microtubule organising activity of centrosomes ([@pbio-0020008-Compton1]; [@pbio-0020008-Bornens1]). The second, restricted to female germline and some embryonic cells that lack centrosomes, is thought to depend upon the microtubule stabilisation and organisation activity of the chromosomes themselves ([@pbio-0020008-McKim1]; [@pbio-0020008-de1]; reviewed in [@pbio-0020008-Karsenti1]). Centrosome-independent microtubule growth and sorting into a bipolar spindle have been observed in vitro around chromatin-coated beads in *Xenopus* egg extracts ([@pbio-0020008-Heald1]). Moreover, some experimental data suggest that a centrosome-independent pathway for spindle assembly also exists in somatic cells ([@pbio-0020008-Bonaccorsi1], [@pbio-0020008-Bonaccorsi2]; [@pbio-0020008-Megraw1], [@pbio-0020008-Megraw2]; [@pbio-0020008-Vaizel-Ohayon1]; [@pbio-0020008-Khodjakov2]; [@pbio-0020008-Hinchcliffe1]; reviewed in [@pbio-0020008-Raff1]). It is generally assumed that the microtubules that build these acentrosomal spindles originate over the chromatin. However, so far, the actual process of centrosome-indepen dent microtubule nucleation, polymerisation, and sorting into a bipolar spindle has not been documented in any of the centrosome-containing cell lineages of a living animal. The problem in visualising such microtubules of noncentrosomal origin when centrosomes are present is a technical one. In *Drosophila*, as in most animal cells, at the onset of cell division the two segregated pairs of centrosomes have a strong microtubule organising activity ([@pbio-0020008-Tates1]; [@pbio-0020008-Church1]; [@pbio-0020008-Cenci1]). Consequently, as soon as the nuclear envelope (NE) breaks down, numerous microtubules invade the nuclear region, making it extremely difficult to single out any noncentrosomal microtubules that might be present. To circumvent this limitation, we have taken advantage of two experimental conditions that inhibit the natural process of centriole migration from the plasma membrane to the interior of the cell that takes place at the onset of meiosis in *Drosophila* spermatocytes. Under such conditions, the centrosomes organise asters, but these are kept at the plasma membrane, away from the nuclear region. In these cells, microtubules can be seen to grow from the remnants of the fenestrated NE and to assemble into anastral bipolar spindles in a centrosome-independent manner. We propose that these spindle-shaped arrays correspond to a subset of microtubules that are normally present in the spindles of wild-type cells. Results {#s2} ======= Centriole Migration towards the Nucleus in *Drosophila* Spermatocytes Requires Microtubules and the Function of *asp* {#s2a} --------------------------------------------------------------------------------------------------------------------- Studies based on electron microscopy ([@pbio-0020008-Tates1]) had shown that soon after the last round of mitotic divisions that precede meiosis in *Drosophila* spermatocytes, the centrioles migrate towards the periphery of the cell and position themselves underneath the plasma membrane. The same studies revealed that shortly before the onset of prometaphase I, the centrioles are found again close to the nuclear membrane, thus strongly suggesting that they migrate back near the nucleus in preparation for meiosis. Using an endogenously expressed centriolar green fluorescent protein (GFP) marker, we have been able to demonstrate such migration in living spermatocytes ([Figure 1](#pbio-0020008-g001){ref-type="fig"}A; [Video 1](#pbio-0020008-v001){ref-type="other"}). The entire process takes about 2 h. Initially, the two centriolar pairs move towards the nucleus and start to migrate apart as they approach the nuclear membrane. They finally position themselves at opposite sides of the nucleus, about 30 min before the onset of NE breakdown (NEB). ::: {#pbio-0020008-g001 .fig} Figure 1 ::: {.caption} ###### Centriole Migration in Primary Spermatocytes \(A) Time-lapse series of confocal images from a wild-type primary spermatocyte expressing GFP-PACT (centrioles) and His2AvD--GFP (chromosomes). The centrioles (arrows) can be seen moving away from the plasma membrane (0) towards the nucleus (N) and then migrating diametrically apart as the chromatin condenses. The chromosomes are fully condensed at timepoint 121 min. (B--D) The two centriole pairs (green) projected over the phase-contrast view (grey) can be seen close to the fenestrated NE and away from the plasma membrane (pm) in control cells (B), while they remain plasma membrane-bound in *asp* (C) and in colcemid-treated wild-type cells (D). In *asp* spermatocytes (C), the position of the membrane-bound centrioles correlates tightly with the pointed end of phase-dark protrusions (arrows) that are not present in colcemid-treated cells. These reflect the distribution of phase-contrast membranes known to overlap microtubules in these cells. (E--J) XY projections (E--G) and their corresponding optical sections (H--J) of control (E and H), *asp* (F and I), colcemid-treated spermatocytes (G and J) expressing an endogenous GFP--α-tubulin confirm that the two major MTOCs in control cells are close to the nucleus, but remain near the plasma membrane in the two experimental conditions. MTOC activity in colcemid-treated spermatocytes was assayed following a 1-s pulse of 350 nm light to inactivate the drug, thus allowing microtubule regrowth. The yellow bar in the XY projections (E--G) marks the position of the corresponding XZ optical sections (H--J). :::  ::: We have identified two experimental conditions that inhibit centriole movement, back from the plasma membrane in *Drosophila* spermatocytes. The first one is mutation in the gene abnormal spindle, *asp* ([@pbio-0020008-Ripoll1]; [@pbio-0020008-Casal1]; [@pbio-0020008-Gonzalez1]; [@pbio-0020008-Saunders1]; [@pbio-0020008-do1]; [@pbio-0020008-Wakefield1]; [@pbio-0020008-Riparbelli1]). In contrast to wild-type control cells ([Figure 1](#pbio-0020008-g001){ref-type="fig"}B), the two pairs of centrioles in *asp^E3^/asp^L1^* spermatocytes at late prophase are still located at the plasma membrane ([Figure 1](#pbio-0020008-g001){ref-type="fig"}C), where they remain throughout meiosis. Centriole migration back towards the NE can also be inhibited by microtubule depolymerisation. Like in *asp* mutant spermatocytes, the centrioles of wild-type spermatocytes exposed to the microtubule-depolymerising drug colcemid remain close to the plasma membrane throughout meiosis ([Figure 1](#pbio-0020008-g001){ref-type="fig"}D). Top ([Figure 1](#pbio-0020008-g001){ref-type="fig"}F) and lateral ([Figure 1](#pbio-0020008-g001){ref-type="fig"}I) views of living *asp* mutant spermatocytes expressing a GFP--α-tubulin fusion reveal that the microtubule organising centres (MTOCs) are found at the periphery of these cells. The MTOCs of control cells at this stage can be seen near the nuclear membrane ([Figure 1](#pbio-0020008-g001){ref-type="fig"}E and [1](#pbio-0020008-g001){ref-type="fig"}H). These observations strongly suggest that the membrane-bound centrioles observed in *asp* mutant spermatocytes are associated to active centrosomes that retain MTOC activity. This conclusion is further substantiated by the localisation of the pericentriolar material (PCM) marker γ-Tub23C ([@pbio-0020008-Zheng1]; [@pbio-0020008-Sunkel1]) around the membrane-bound centrioles, coinciding with the position of the MTOCs (data not shown). The same applies to cells in which centriole migration is inhibited by colcemid. Immediately after a short pulse of 350 nm UV light to inactivate the drug (W. E. Theurkauf, personal communication), two asters are organised around the membrane-bound centrioles in these cells ([Figure 1](#pbio-0020008-g001){ref-type="fig"}G and [1](#pbio-0020008-g001){ref-type="fig"}J). Anastral Spindles Are Assembled When the Centrosomes Are Kept Membrane Bound {#s2b} ---------------------------------------------------------------------------- Inhibition of centrosome migration back from the plasma membrane in *Drosophila* spermatocytes offers an unprecedented opportunity to assay centrosome-independent microtubule polymerisation during spindle assembly in the cells of a living animal that contain centrosomes. Therefore, we decided to follow microtubules by time-lapse confocal microscopy in *asp* and colcemid-treated cells that expressed a GFP--α-tubulin fusion, as they went through meiosis. At the onset of prometaphase, the NE becomes fenestrated, but does not disappear in *Drosophila* ([@pbio-0020008-Tates1]; [@pbio-0020008-Stafstrom1]; [@pbio-0020008-Church2]). This partial NEB can be readily identified by the sudden entry of GFP--α-tubulin into the nuclear region ([Figure 2](#pbio-0020008-g002){ref-type="fig"}, timepoint 0; [Video 2](#pbio-0020008-v002){ref-type="other"}). In control cells, microtubule polymerisation and organisation are largely concentrated around the centrosomes ([@pbio-0020008-Church1]; [@pbio-0020008-Cenci1]). Consequently, the abundance of these microtubules makes it extremely difficult to determine the possible contribution of any centrosome-independent microtubule polymerisation activity to spindle assembly ([Figure 2](#pbio-0020008-g002){ref-type="fig"}, control, 10 min to 32 min; [Video 2](#pbio-0020008-v002){ref-type="other"}). Microtubule organisation is significantly different in the case of *asp* mutant spermatocytes. At the time of NEB, the membrane-bound centrosomes can be seen organising the two asters at a significant distance from the nucleus, which is kept clear from astral microtubules ([Figure 2](#pbio-0020008-g002){ref-type="fig"}; [Video 3](#pbio-0020008-v003){ref-type="other"}). Around 10 min after NEB, a distinct focus of microtubule polymerisation appears within the nuclear region, away from the asters ([Figure 2](#pbio-0020008-g002){ref-type="fig"}, *asp*, 10 min; [Video 3](#pbio-0020008-v003){ref-type="other"}). It gives rise to a few bundles ([Figures 2](#pbio-0020008-g002){ref-type="fig"}, 15 min) that grow ([Figure 2](#pbio-0020008-g002){ref-type="fig"}, 22 min) and get organised into a bipolar spindle-shaped microtubule array that in 28% (*n* = 43) of the cells is anastral and establishes no contact with the membrane-bound centrosomes ([Figure 2](#pbio-0020008-g002){ref-type="fig"}, 39 min). The remaining 72% was accounted for by cells in which, despite the distance, microtubules from one or both asters reach the spindle so that spindle poles and asters were aligned. Although the acentrosomal origin of the spindle microtubules in these cells is fairly convincing, only those cells that assembled truly anastral spindles that remained so throughout meiosis were considered as cases of noncentrosomal spindle assembly. ::: {#pbio-0020008-g002 .fig} Figure 2 ::: {.caption} ###### Time-Lapse Series of Meiosis Progression in Control, *asp*, and Colcemid-Treated Spermatocytes Timepoint 0 coincides with the time of NEB revealed by the sudden entry of GFP signal into the nucleus. In control cells ([Video 2](#pbio-0020008-v002){ref-type="other"}), microtubules are mainly organised around the centrosomes (arrows). However, when centrosomes are kept away from the nuclear region by mutation in *asp* ([Video 3](#pbio-0020008-v003){ref-type="other"}) or colcemid treatment ([Video 4](#pbio-0020008-v004){ref-type="other"}), microtubule nucleation and growth are clearly revealed over the nuclear region (N), well away from the centrosomes. Such noncentrosomal microtubules may form bundles that eventually are sorted into spindlelike bipolar microtubule arrays. Microtubules were labelled with an endogenous GFP--α-tubulin fusion. :::  ::: These observations strongly suggested that microtubules can nucleate in a centrosome-independent manner and assemble a spindle-like array in *Drosophila* spermatocytes. The question remained open, however, as to whether such anastral structures could not simply be a consequence of mutation in *asp* itself. To rule out such a possibility, we followed spindle assembly in wild-type spermatocytes in which centrosomes had been kept membrane-bound by colcemid treatment. The results were strikingly similar to those observed in *asp* cells. Seconds after colcemid inactivation, the cortex-bound position of the centrosomes is revealed by the growing asters that were not visible before ([Figure 2](#pbio-0020008-g002){ref-type="fig"}, colcemid inactivation, 0 timepoint; [Video 4](#pbio-0020008-v004){ref-type="other"}). Like in *asp* mutant spermatocytes, microtubules can clearly be seen to nucleate over the nuclear region, well away from the asters ([Figure 2](#pbio-0020008-g002){ref-type="fig"}, 11 min), grow ([Figure 2](#pbio-0020008-g002){ref-type="fig"}, 15 min), and get sorted ([Figure 2](#pbio-0020008-g002){ref-type="fig"}, 37 min and 43 min) into anastral bipolar arrays. Of the colcemid-treated cells studied (*n* = 32), 22% behaved like the cell shown in [Figure 2](#pbio-0020008-g002){ref-type="fig"}. The remaining cells assembled more than one spindle, multipolar spindles, or spindles that were connected to one of the asters. The Nucleation of Noncentrosomal Microtubules in Spermatocytes with Membrane-Bound Centrosomes Has a Late Onset {#s2c} --------------------------------------------------------------------------------------------------------------- We then decided to time the onset of anastral microtubule nucleation. The timing of the main landmarks of meiosis progression in control, *asp*, and colcemid-treated spermatocytes is summarised in [Figure 3](#pbio-0020008-g003){ref-type="fig"}. In control spermatocytes, chromosome condensation starts within 2 min after NEB, and the first centrosomal microtubules enter the nuclear region shortly afterwards. Chromosome condensation is completed between 10 and 12 min after NEB. In agreement with previous reports, anaphase onset takes place between 32 and 47 min after NEB ([@pbio-0020008-Church2]; [@pbio-0020008-Rebollo1]; [@pbio-0020008-Savoian2]). Remarkably, the timing of meiosis progression from NEB to anaphase onset, both in *asp* and in colcemid-treated wild-type spermatocytes, seems to be largely unaffected, suggesting that the feeble spindle checkpoint of these cells ([@pbio-0020008-Rebollo1]; [@pbio-0020008-Savoian2]) is not triggered by the membrane-bound centrosomes\' condition. The timing of the onset of noncentrosomal microtubule growth within the nuclear region in *asp* and colcemid-treated cells is tightly controlled. It occurs between 9 and 13 min after NEB, at the same time or marginally later than the end of chromosome condensation. If this process occurs with the same timing in wild-type cells, the noncentrosomal microtubules will intermingle with numerous centrosomal microtubules that are already present at this stage. ::: {#pbio-0020008-g003 .fig} Figure 3 ::: {.caption} ###### The Timing of Noncentrosomal Microtubule Nucleation Referred to NEB (timepoint 0), the timing of chromosome condensation and of onset of chromosome segregation is essentially identical in control, *asp*, and colcemid-treated spermatocytes. In control spermatocytes, aster microtubules can be seen entering the nuclear region 3--6 min after NEB. They do not in *asp* or following colcemid treatment. In these two cases, however, centrosome-independent microtubule polymerisation can be seen over the nuclear region. It starts between 9 and 13 min after NEB, coinciding with or very shortly after the end of chromosome condensation. :::  ::: Acentrosomal Microtubules Are Nucleated on the Inner Side of the Remnants of the NE and Not around the Chromosomes {#s2d} ------------------------------------------------------------------------------------------------------------------ To determine the nucleation site of the microtubules organised over the nuclear region, we followed the initial stages of microtubule assembly by time-lapse microscopy, acquiring several Z series of XY confocal and phase-contrast sections at different timepoints. From these, we generated a time-lapse series of 3D reconstructions that allowed us to localise the foci of nucleation of anastral microtubules. We were able to draw the following three main conclusions that apply to both *asp* and colcemid-treated spermatocytes. Firstly, the foci from which microtubules grow may be clustered ([Figure 4](#pbio-0020008-g004){ref-type="fig"}A) or dispersed ([Figure 4](#pbio-0020008-g004){ref-type="fig"}B). Secondly, no significant correlation can be established between the site of microtubule nucleation and the chromosomes ([Figure 4](#pbio-0020008-g004){ref-type="fig"}A and [4](#pbio-0020008-g004){ref-type="fig"}B). Finally, nucleation takes place in close proximity to the remnants of the NE (ten out of ten cells reconstructed; [Figure 4](#pbio-0020008-g004){ref-type="fig"}A and [4](#pbio-0020008-g004){ref-type="fig"}B), which in *Drosophila* ruptures without disassembling completely ([@pbio-0020008-Tates1]; [@pbio-0020008-Stafstrom1]). ::: {#pbio-0020008-g004 .fig} Figure 4 ::: {.caption} ###### The Place of Noncentrosomal Microtubule Nucleation The initial stages of noncentrosomal microtubule nucleation revealed by an endogenous GFP--α-tubulin fusion (left) and phase contrast (right). Following the corresponding videos, it is possible to unmistakably tell the chromosomes (arrows) apart form the other phase-dark objects that are present over the nuclear region (asterisks). The cell in (A) is shown as a single timeframe and the cell in (B) as a time-lapse series. In both cells, noncentrosomal microtubule nucleation (arrowheads) takes place close to the remains on the NE and does not overlap with the major chromosomes. Nucleation sites can be clustered (A) or dispersed (B). In the time-lapse series (B), only the chromosomes that are in focus are labelled. Timepoint 0 min in these series corresponds to the first sign of noncentrosomal microtubule nucleation, around 11 min after NEB. A white bar marks the growing end of a microtubule bundle that at timepoint 93 min reaches one of the bivalents. :::  ::: The Anastral Spindles Organised in Cells with Membrane-Bound Centrosomes Can Sustain Some Degree of Chromosome Segregation {#s2e} -------------------------------------------------------------------------------------------------------------------------- We then decided to study in more detail the extent to which the anastral spindles organised in cells with membrane-bound centrosomes can mediate successful cell division. To this end, we produced transgenic flies carrying a GFP--α-tubulin fusion together with a His2AvD--YFP (yellow fluorescent protein) strain so that both chromosomes and microtubules could be visualised in the same cell ([Figure 5](#pbio-0020008-g005){ref-type="fig"}; [Video 5](#pbio-0020008-v005){ref-type="other"}). In *asp* cells, during prometaphase, the bivalents do not move to the extent that they do in control cells (data not shown). As mentioned before, congression occurs ([Figure 5](#pbio-0020008-g005){ref-type="fig"}; [Video 6](#pbio-0020008-v006){ref-type="other"}), but orientation is rarely bipolar. Homologue chromosomes separate at the onset of anaphase, but they barely move, remaining near the center of the spindle. Moreover, they tend to cosegregate ([Video 7](#pbio-0020008-v007){ref-type="other"}) and end up included in the same daughter nucleus. All together, these abnormalities result in high levels of aneuploidy in agreement with previous genetic analysis data ([@pbio-0020008-Ripoll1]). In contrast, in half (52%) of the anastral spindles assembled following transient colcemid treatment, homologue chromosomes could be seen to segregate from one another ([Figure 5](#pbio-0020008-g005){ref-type="fig"}; [Video 8](#pbio-0020008-v008){ref-type="other"}). Anaphase in these cells is not complete, however, because only the chromosome-to-pole movement (anaphase A) is observed. The further separation achieved in wild-type cells by the extension of the spindle (anaphase B) is very limited in these cells. ::: {#pbio-0020008-g005 .fig} Figure 5 ::: {.caption} ###### Chromosome Segregation in Anastral Spindles in *Drosophila* Spermatocytes (Control \[[Video 5](#pbio-0020008-v005){ref-type="other"}\]) At metaphase I (0), the bivalents (revealed by a His2Avd--YFP fusion, shown by double arrowheads) are aligned in the middle of the spindle (revealed by a GFP--α-tubulin fusion), at the metaphase plate. At the onset of anaphase (3 min), the homologue chromosomes start to migrate towards opposite poles (single arrowheads) and to decondense. During anaphase B (4 min and 6 min), the spindle poles move apart from each other and the two sets of decondensed chromosomes become further separated. (*asp* \[[Video 6](#pbio-0020008-v006){ref-type="other"}\]) At timepoint 0, the bivalents align at the metaphase plate. Homologue chromosomes split apart at the onset of anaphase I (4 min). However, anaphase A migration is highly impaired. By the time the chromosomes start to decondense, they have barely moved towards the spindle poles (8 min and 14 min), and often homologue chromosomes end up included in the same daughter nucleus. (Colcemid \[[Video 8](#pbio-0020008-v008){ref-type="other"}\]) As in *asp* spermatocytes, the asters (arrows) remain at the plasma membrane at metaphase I in colcemid-treated cells, and the bivalents align in a metaphase plate-like within the acentrosomal spindles (0 min). Homologue chromosomes split apart at the onset of anaphase (upper cell, 6 min) and significantly segregate from one another (upper cell, 8 min; lower cell, 3 min). Further separation of the daughter nuclei during anaphase B is very limited in these cells (8 min), and cytokinesis does not occur. :::  ::: The Orientation of the Cytokinesis Furrow Correlates with the Position of the Membrane-Bound Asters, Independently of Spindle Orientation {#s2f} ----------------------------------------------------------------------------------------------------------------------------------------- Following colcemid treatment, we have never observed complete cytokinesis. However, as reported before ([@pbio-0020008-Riparbelli1]), cytokinesis does proceed to completion in around half (47%, *n* = 19) of *asp* cells. These cells, which contain unconnected centrosomal asters and anastral spindles, provide a valuable experimental system to assess the contribution of asters and spindle to specifying the place of cleavage. To this end, we plotted the angles between the line defined by the two asters, the major spindle axis, and the plane of cleavage in wild-type and *asp* spermatocytes ([Figure 6](#pbio-0020008-g006){ref-type="fig"}). Two conclusions can be drawn from these data. Firstly, the anastral spindles assembled in *asp* cells can be observed at any angle, even up to 90°, with respect to the position of the two asters. Interestingly, in most such cases, furrow progression forces the spindle to rotate and align with the asters so that, at the end, a fairly normal cytokinesis takes place. Secondly, the orientation of the plane of cleavage keeps a tight 90° ± 10° with respect to the axis defined by the asters and does not correlate with the orientation of the anastral spindle. ::: {#pbio-0020008-g006 .fig} Figure 6 ::: {.caption} ###### Correlation between the Orientation of the Cytokinesis Furrow, the Asters, and the Spindle in *asp* Spermatocytes Schematic representation (A--C) of the relative position of the asters (red), the spindle (yellow), and the cytokinesis furrow (blue), corresponding to a control cell (D) and two examples of *asp* mutant spermatocytes (E and F), respectively. Asters (arrows) and spindles are labelled with a GFP--α-tubulin fusion. The position of the cleavage furrow (double-headed arrow) was determined by time-lapse imaging of these cells (data not shown). In wild-type cells (*n* = 10), plotting spindle and furrow orientation relative to the interastral axes shows that asters and spindle are tightly aligned, and cleavage occurs at an angle of 90° ± 10° with respect to them (G). In *asp* spermatocytes (*n* = 10), the plane of cleavage occurs at a 90° ± 10° angle with respect to the asters and does not correlate with the orientation of the anastral spindle. :::  ::: Discussion {#s3} ========== We have found that when the asters are kept near the plasma membrane during meiosis I in *Drosophila* spermatocytes, noncentrosomal microtubules appear over the nuclear region and, in a fraction of the cells, are sorted into anastral bipolar spindles (summarised in [Figure 7](#pbio-0020008-g007){ref-type="fig"}). Identical observations are derived whether centrosomes are forced to remain membrane-bound by mutation in *asp* or by transient colcemid treatment of wild-type cells. The very different nature of these two experimental conditions strongly argues against these spindles being assembled as a consequence of the experimental conditions themselves. It rather suggests that the observation of anastral spindles is due to the impaired ability of the plasma membrane-bound centrosomes to contribute to spindle assembly. Anastral spindles are also assembled following the removal of centrosomes by laser ablation or microdissection in cultured cells ([@pbio-0020008-Khodjakov2]; [@pbio-0020008-Hinchcliffe1]) or by inhibiting the formation of centrosomes in mutant *Drosophila* embryos ([@pbio-0020008-Megraw1]; [@pbio-0020008-Vaizel-Ohayon1]), thus reinforcing this argument. ::: {#pbio-0020008-g007 .fig} Figure 7 ::: {.caption} ###### Noncentrosomal Microtubules and Spindle Assembly (Central column) Spindle assembly in *Drosophila* spermatocytes with membrane-bound centrosomes. At the time of NEB, the chromatin (pale blue) starts to condense, and the membrane-bound centrosomes (red) organise asters (yellow) at a significant distance from the nuclear region. Around 12 min after NEB, the first noncentrosomal microtubules (green) start to nucleate near the remnants of the NE (grey), as the chromosomes achieve full condensation (dark blue). These microtubules then bundle, associate with the chromosomes, and eventually end up organised into a bipolar anastral array whose shape is reminiscent of the female meiotic spindle. (Left column) Spindle assembly in wild-type *Drosophila* oocytes ([@pbio-0020008-Theurkauf1]; [@pbio-0020008-Matthies1]). NEB starts at the beginning of stage 13 of oocyte development. At this stage, the oocyte does not contain centrosomes and the chromosomes (karyosome) are tightly condensed (dark blue). Microtubules (green) appear 11--15 min after NEB within the nuclear region in association with the karyosome. These microtubules form bundles and are sorted around the chromatin into a bipolar spindle. Evidence suggests that ER components may be required for spindle assembly in these cells ([@pbio-0020008-Kramer1]). At metaphase I, recombined bivalents are aligned at the spindle equator, while those that have not recombined are found closer to the spindle poles. Meiosis remains arrested at this point (stage 14) until oocyte activation. Despite the obvious morphological similitude, the equivalence between these and the anastral spindles organised in spermatocytes with membrane-bound centrosomes is unclear. (Right column) Hypothesis regarding the contribution of centrosomal and noncentrosomal microtubules to spindle assembly during meiosis I in wild-type *Drosophila* spermatocytes. Before NEB, the centrosomes are located at opposite positions near the nucleus. Shortly after NEB, astral microtubules enter the nuclear region and make the first contact with the condensing chromatin. No evidence of noncentrosomal microtubule polymerisation near the nuclear region at this stage has been found yet. Once chromosomes are fully condensed, microtubule bundles of centrosomal origin (yellow) connecting centrosomes to chromosomes already exist. At this stage, noncentrosomal microtubules (green) start to polymerise in association with the remnants of the NE. These microtubules form bundles that interact with the chromosomes and intermingle with the microtubules of centrosomal origin. The fully mature spindle in these cells would therefore contain a spindle-shaped structure made of microtubules of noncentrosomal origin (green) embedded in another spindle-shape array made of two overlapping asters (yellow). We propose that each of these subsets may perform to a certain extent some of the functions carried out by normal spindles, but neither of them can on its own mediate robust cell division. :::  ::: The Place of Noncentrosomal Microtubule Nucleation {#s3a} -------------------------------------------------- Noncentrosomal microtubule nucleation during cell division is thought to take place over the chromatin ([@pbio-0020008-Nachury1]). This assumption is largely based on the observations carried out in the few acentrosomal systems in which spindle assembly has been followed by time-lapse microscopy (reviewed in [@pbio-0020008-Karsenti1]). These include wild-type *Drosophila* female meiocytes ([@pbio-0020008-Theurkauf1]; [@pbio-0020008-Matthies1]), parthenogenetic *Sciara* embryos ([@pbio-0020008-de1]), and *Xenopus* egg extracts ([@pbio-0020008-Heald1]). It is also consistent with the active role of chromosomes in spindle organisation. For instance, it has been reported that bivalents micromanipulated away from the spindle in *Drosophila* spermatocytes induce the assembly of anastral minispindles in the cytoplasm ([@pbio-0020008-Church3]). Likewise, the removal of chromosomes before NEB has been shown to inhibit spindle assembly in grasshopper spermatocytes ([@pbio-0020008-Zhang1]), although the phenotype of *fusolo* mutants, recently described, seems to argue otherwise ([@pbio-0020008-Bucciarelli1]). Moreover, the chromosomal localisation of the RanGEF RCC1 is expected to result in a local enrichment of the GTP-bound form of Ran, known to facilitate spindle assembly ([@pbio-0020008-Nachury1]; [@pbio-0020008-Wilde1]; [@pbio-0020008-Gruss1]; reviewed in [@pbio-0020008-Hetzer1]), thus providing a mechanistic interpretation for the suspected role of chromatin in this process. In contrast, our observations reveal that nucleation of the noncentrosomal microtubules over the nuclear region occurs over the remnants of the NE, which in *Drosophila* are present throughout cell division despite extensive fenestration at the onset of prometaphase ([@pbio-0020008-Tates1]; [@pbio-0020008-Stafstrom1]; [@pbio-0020008-Church2]). However, upon closer examination, our observations may also be consistent with the literature quoted above. The single bivalents that organise minispindles when micromanipulated into the cytoplasm in *Drosophila* spermatocytes have actually been shown to be surrounded by masses of stacked membranes, whose contribution to microtubule nucleation/stabilisation, according to the authors themselves, cannot be ruled out ([@pbio-0020008-Church3]). In *Drosophila* oocytes, too, the meiotic spindle is ensheathed in a membrane structure derived from the endoplasmic reticulum (ER). Although the bulk of microtubules has been described by time-lapse confocal microscopy to form over the chromatin ([@pbio-0020008-Theurkauf1]; [@pbio-0020008-Matthies1]), the contribution of these membranes to the initial stages of microtubule nucleation cannot be discarded either. In this regard, the recent cloning of *Axs*, which encodes a transmembrane protein associated with the membranes that surround the spindle and is required for the segregation of achiasmate chromosomes, is very tantalising ([@pbio-0020008-Kramer1]). Asx is distributed within the ER of the germinal vesicle just before meiotic spindle assembly. Upon germinal vesicle breakdown (GVBD), Axs associates with the developing spindle through all stages of assembly. These observations have been taken as an indication that the ER may be organised into structures that impinge on spindle assembly during meiosis in *Drosophila* females ([@pbio-0020008-Kramer1]), very much in line with our observations in *Drosophila* spermatocytes. Indeed, our results do not discard the contribution of chromatin to microtubule stabilisation and sorting into a bipolar array, even if chromatin itself is not the place of initial microtubule nucleation, nor do they rule out the possibility of microtubules being polymerised over the chromatin at later stages. They merely show that in these cells, the remaining bits of the fenestrated NE provide a particularly favourable environment to sustain the initial stages of noncentrosomal microtubule nucleation. Moreover, these observations strongly suggest that, unlike centrosomes, the foci of microtubule nucleation over the nuclear region do not behave as stable MTOCs. As soon as the microtubule bundles acquire a certain length, they interact with the condensed chromosomes and are often sorted into a bipolar spindle, regardless of the initial number of nucleation sites. We have not been able to detect γ-Tub23C at these nucleation sites. We still do not know the actual contribution of Ran to spindle assembly in *Drosophila* spermatocytes, although, given its known conservation across distant species ([@pbio-0020008-Hetzer1]), it is likely to play a major role. Since orthologues of most of the known components of this pathway are known in *Drosophila*, it is technically possible to address this question both under normal conditions and in cells in which centrosomes cannot contribute to spindle assembly as described in this work. Experiments are underway in our laboratory to address these points. Functional Relevance of the Anastral Spindles {#s3b} --------------------------------------------- Perhaps the most fundamental question regarding the anastral spindles organised in cells that normally contain centrosomes is the extent to which they could provide a back-up, able to mediate successful and robust cell division when the centrosomes cannot contribute to spindle assembly. Our observations suggest that this is an unlikely scenario. Firstly, only a fraction of cells display a single bipolar array, the rest being accounted for by cases in which either the spindle is multipolar or there is more that one per cell or there is no spindle at all. Secondly, chromosome segregation is also significantly less efficient than in control cells. These two points, however, carry a caveat since they could reflect the effect of depleted *asp* function or residual traces of active colcemid, rather than the anastral nature of the spindle. Finally, cytokinesis is severely disrupted in these cells. Around half of *asp* spermatocytes containing anastral spindles go through and complete cytokinesis. However, in these cells, the orientation of the cleavage furrow correlates tightly with the position of the two asters and not at all with the orientation of the spindle. This situation gives rise to cases in which the plane of cleavage is nearly parallel to the spindle. In colcemid-treated cells that contain notoriously small asters, cytokinesis does not occur. These observations strongly argue that asters contribute to specify the place of furrow and may be required for cleavage. The contribution of centrosomes to ensure proper cytokinesis has been previously observed in vertebrate cell lines ([@pbio-0020008-Rieder1]; [@pbio-0020008-Savoian1]; [@pbio-0020008-Hinchcliffe1]; [@pbio-0020008-Khodjakov1]), human cell lines ([@pbio-0020008-Gromley1]), *Dyctiostelium* ([@pbio-0020008-Neujahr1]), or *Xenopus* ([@pbio-0020008-Takayama1]). This conclusion, however, is not consistent with the observation that cytokinesis is not inhibited in asterless *Drosophila* spermatocytes ([@pbio-0020008-Bonaccorsi1]). Therefore, the anastral spindles organised in spermatocytes with membrane-bound centrosomes seem able to provide only some of the functions required for cell division, with relatively low efficiency. The functionality of the anastral spindles assembled in embryos laid by *cnn* mutant females, which do not appear to contain centrosomes, is also compromised. These spindles are not always properly shaped, the chromosomes are not tightly aligned at the spindle equator, chromosome movements are nonsynchronous, and their segregation not always faithful ([@pbio-0020008-Megraw1]; [@pbio-0020008-Vaizel-Ohayon1]). Thus, in this instance, too, when centrosome function is abrogated in a syncytium that normally contains centrosomes and that does not naturally undergo parthenogenesis, anastral spindles can be assembled that are able to perform some of the functions of their wild-type counterparts, but in a rather inefficient manner. Origin of the Anastral Spindles: Neomorphic or Constitutive {#s3c} ----------------------------------------------------------- Two alternative interpretations can account for the origin of the anastral spindles that we have observed ([@pbio-0020008-Khodjakov2]). First, they could be neomorphic structures, assembled through a pathway normally repressed that is only triggered in response to the impaired contribution of centrosomal microtubules. Although we cannot at the moment discard this interpretation, we find it hard to envisage how such an alternative pathway could have evolved, given the extremely low frequency of centrosome loss or inactivation in wild-type populations. Moreover, it is also difficult to imagine what sort of signalling mechanism could trigger the alternative pathway in these cells since centrosomes are still present and active as MTOCs. Alternatively, these anastral microtubule arrays could be a constitutive component of wild-type spindles, normally masked by the abundance of centrosome-derived microtubules, but revealed when asters are kept away. This interpretation is summarised in [Figure 7](#pbio-0020008-g007){ref-type="fig"}. In wild-type spermatocytes under normal conditions, the first astral microtubules enter the nuclear area shortly after NEB and start to build a bipolar spindle as chromosome condensation progresses. By the time chromosome condensation is fully achieved, a distinct bipolar spindle can be observed in these cells. However, it is not yet fully mature, as the number of microtubules will still increase until anaphase onset. It is about this time that nucleation of the acentrosomal microtubules occurs in cells with plasma membrane-bound centrosomes. Therefore, if this process occurs at the same time in wild-type cells, the acentrosomal microtubules could significantly contribute to the maturation of the cell division spindle. This interpretation is consistent with the recent proposal put forward by [@pbio-0020008-Gruss1] to account for their observations regarding spindle assembly in HeLa cells. They found that when the function of the human homologue of TPX2 is inhibited by RNA interference, the centrosomal asters do not interact and do not form a spindle. From these observations, they concluded that, intermingled with microtubules of centrosomal origin, the mitotic spindle may contain noncentrosomal microtubules that are stabilised and organised by the chromatin and are essential for the assembly of functional spindles. In *Drosophila*, secondary spermatocytes\' mutation in *fusolo* seem to reveal the centrosome-derived component of the spindle ([@pbio-0020008-Bucciarelli1]). Forcing the asters away from the nucleus in *Drosophila* primary spermatocytes reveals the noncentrosomal component that, indeed, does not require asters to get organised into a spindle-like structure. We propose that both components are required to mediate robust cell division. The very recent finding of peripheral, noncentrosomal microtubules that contribute to spindle assembly in LLCPK1α cells provides additional evidence to substantiate this conclusion ([@pbio-0020008-Tulu1]). Regardless of their neomorphic or constitutive nature, the acentrosomal spindles that we have found in *asp* and colcemid-treated spermatocytes are, from a morphological point of view, closely reminiscent of the anastral female meiotic spindles found in many animal species, including *Drosophila* ([@pbio-0020008-Theurkauf1]). The same holds true for the anastral spindles assembled when centrosomes are removed from the cell or cannot be organised due to mutation in essential centrosomal components ([@pbio-0020008-Megraw1]; [@pbio-0020008-Khodjakov2]; [@pbio-0020008-Khodjakov1]; [@pbio-0020008-Hinchcliffe1]). In the case of the *Drosophila* female meiotic spindle, the timing of microtubule nucleation is also very similar: between 9 and 12 min after NEB in spermatocytes and 11 to 15 min in oocytes ([@pbio-0020008-Matthies1]). These similarities have led some to propose that experimentally induced anastral spindles could require the same motors and structural components that build the spindles in female meiocytes ([@pbio-0020008-Megraw1]; [@pbio-0020008-Khodjakov2]) (summarised in [Figure 6](#pbio-0020008-g006){ref-type="fig"}). In fact, it has been suggested that the absence of some of these components at the time syncytial divisions occur could explain the lack of robustness of the anastral spindles assembled in embryos derived from *cnn* mothers ([@pbio-0020008-Megraw1]). We still do not know to what extent the anastral spindles of spermatocytes share components with the oocyte spindle. Some essential ones cannot be shared, though, since they are only expressed in the female germline. Given the wealth of probes and mutants available in *Drosophila*, it should be possible to draw a clear picture of the situation regarding this fundamental question. Materials and Methods {#s4} ===================== {#s4a} ### Fly stocks {#s4a1} Flies from *w^1118^;e^11^ asp^E3^/TM6C* and *w^1118^;red asp^L1^/TM6C* stocks were crossed to generate *w^1118^;e^11^ asp^E3^/red asp^L1^* transheterozygous individuals. The viability of *asp^E3^/asp^L1^* males is high, but they are poorly fertile and produce high levels of aneuploid gametes. ### Transgenes {#s4a2} The chromosomes were labeled with transgenes expressing either a His2avD--GFP fusion ([@pbio-0020008-Clarkson1]) or its derivative, His2avD--EYFP, constructed by us under the control of the polyubiquitin promoter ([@pbio-0020008-Lee1]). To visualise centrioles, we used the transgene expressing GFP-PACT (pericentrin-AKAP450 centrosomal targeting) (kindly provided by J. Raff) that contains the predicted Drosophila homologue of the PACT domain described by [@pbio-0020008-Gillingham1]. To visualise microtubules, we constructed a transgene that contained the GFP--α-tub84B fusion as previously described ([@pbio-0020008-Grieder1]) under the control of the polyubiquitin promoter ([@pbio-0020008-Lee1]). ### Time-lapse recording {#s4a3} Live spermatocytes were recorded as previously described ([@pbio-0020008-Rebollo1], [@pbio-0020008-Rebollo2]). For most applications, we collected a series of timepoints at 15--30 s intervals, each containing four to eight XY sections at different depths along the Z axis and including both the phase-contrast and fluorescence channels. For more-detailed 3D reconstruction, stacks containing 20 sections were obtained. Laser intensity was always kept to a minimum, and only the excitation laser line 488 was utilised. GFP and YFP signals were distinguished by overlying the two recorded channels. Image processing was performed with NIH-Scion Image, Interactive Data Language (IDL), and huygens2. For 3D reconstructions, we wrote macros in NIH-Scion Image to navigate through the three dimensions of the cell stack. ### Colcemid treatment {#s4a4} Newly hatched adult males were fed for 8--12 h with a solution containing 32 μg/ml of colcemid (Sigma, St. Louis, Missouri, United States) in 1 M sucrose. Upon dissection, their testes were prepared for in vivo imaging as described above. Once under the microscope, microtubules were allowed to repolymerise by a 1-s pulse of 350 nm light that inactivates the drug. For simplicity, this entire procedure of exposure to colcemid followed by light inactivation of the drug is referred to through this manuscript as 'colcemid treatment.' Supporting Information {#s5} ====================== Accession Numbers {#s5a} ----------------- The FlyBase accession numbers discussed in this paper are α-tubulin 84B (CG1913), *asp* (CG6875), *Axs* (CG9703), *cnn* (CG4832), γ-Tub23C (CG3157), His2AvD (CG5499). The GeneBank accession numbers discussed in this paper areGFP (U57609.1), Ran (NM\_006325), RCC1 (D00679), TPX2 (BC020207), and YFP (U57609.1). ::: {#pbio-0020008-v001 .fig} Video 1 ::: {.caption} ###### Centriole Migration in Primary Spermatocytes (849 KB MOV) ::: ::: ::: {#pbio-0020008-v002 .fig} Video 2 ::: {.caption} ###### Spindle Assembly in Control Spermatocytes (442 KB MOV) ::: ::: ::: {#pbio-0020008-v003 .fig} Video 3 ::: {.caption} ###### Spindle Assembly in *asp* Spermatocytes (675 KB MOV) ::: ::: ::: {#pbio-0020008-v004 .fig} Video 4 ::: {.caption} ###### Spindle Assembly in Colcemid-Treated Spermatocytes (844 KB MOV) ::: ::: ::: {#pbio-0020008-v005 .fig} Video 5 ::: {.caption} ###### Chromosome Segregation in Control Spermatocytes (206 KB MOV) ::: ::: ::: {#pbio-0020008-v006 .fig} Video 6 ::: {.caption} ###### Chromosome Segregation in *asp* Spermatocytes I (452 KB MOV) ::: ::: ::: {#pbio-0020008-v007 .fig} Video 7 ::: {.caption} ###### Chromosome Segregation in *asp* Spermatocytes II (377 KB MOV) ::: ::: ::: {#pbio-0020008-v008 .fig} Video 8 ::: {.caption} ###### Chromosome Segregation in Colcemid-Treated Spermatocytes (465 KB MOV) ::: ::: We are grateful to W. Theurkauf for advice on the method to inactivate colcemid; to J. W. Raff for providing the Drosophila GFP-PACT construct; to T. Zimmerman for his help with image processing; to A. M. Voie for performing embryo injections; and to I. Vernos and E. Karsenti for very helpful comments on this manuscript. Our lab is financed by grants from the European Union and the Ramon Areces Foundation. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** ER and CG conceived and designed the experiments. ER performed the experiments. ER and CG analyzed the data. ER, SL, and JR contributed reagents/materials/analysis tools. ER and CG wrote the paper. Academic Editor: R. Scott Hawley, Stowers Institute for Medical Research ER : endoplasmic reticulum GFP : green fluorescent protein GVBD : germinal vesicle breakdown IDL : Interactive Data Language MTOC : microtubule organising centre NE : nuclear envelope NEB : nuclear envelope breakdown PACT : pericentrin-AKAP450 centrosomal targeting PCM : pericentriolar material YFP : yellow fluorescent protein

PubMed Central

2024-06-05T03:55:46.047356

2004-1-20

PMC322746

The pathologist makes do with red wine until an effective drug is available, the biochemist discards the bread from her sandwiches, and the mathematician indulges in designer chocolate with a clear conscience. The demographer sticks to vitamin supplements, and while the evolutionary biologist calculates the compensations of celibacy, the population biologist transplants gonads, but so far only those of his laboratory mice. Their common cause is to control and extend the healthy lifespan of humans. They want to cure ageing and the diseases that come with it. "I would take resveratrol if it were feasible," notes David Sinclair, assistant professor of pathology at Harvard Medical School in Boston, Massachusetts. In the meantime, he adds, "I do enjoy a glass of red wine about once a day." It was Sinclair\'s laboratory, in association with a commercial partner, that revealed last August how the team had identified for the first time a group of simple organic molecules capable of extending lifespan. The most proficient of the group is resveratrol, the plant polyphenol found in red wine, and its discovery as a potential elixir to combat ageing represents another extraordinary advance in a decade of discoveries that have revolutionised the field. > *"These molecules will be useful for treating diseases associated with ageing, like diabetes and Alzheimer\'s."* Extending Life {#s2} ============== Although the life-enhancing effects of Sinclair\'s polyphenols are so far confined to the baker\'s yeast Saccharomyces cerevisiae, the work suggests that researchers are only one small step from making a giant leap for humankind. "People imagined that it might have been possible, but few people thought that it was going to be possible so quickly to find such things," says Sinclair. The field of ageing research is buzzing. Resveratrol stimulated a known activator of increased longevity in yeast, the enzyme Sir-2, and thereby extended the organism\'s lifespan by 70% ([Box 1](#box1){ref-type="boxed-text"}). Sir-2 belongs to a family of proteins with members in higher organisms, including SIR-2.1, an enzyme that regulates lifespan in worms, and SIRT-1, the human enzyme that promotes cell survival ([Figure 1](#pbio-0020012-g001){ref-type="fig"}). Though researchers still do not know whether SIRT-1, or "Sir-2 in humans," as Sinclair puts it, has anything to do with longevity, there is a good chance that it does, judging by its pedigree. In any event, resveratrol proved to be a potent activator of the human enzyme. This might not be altogether surprising, at least not now, given that the polyphenol is already associated with health benefits in humans, notably the mitigation of such age-related defects as neurodegeneration, carcinogenesis, and atherosclerosis. ::: {#pbio-0020012-g001 .fig} Figure 1 ::: {.caption} ###### SIRT-1 Deacetylase---the Human Enzyme That Promotes Cell Survival---in a Dividing Human Cell The enzyme is marked in red, and the image is superimposed on acetylated proteins (green) and condensed chromosomes (blue). (Image courtesy of David Sinclair.) :::  ::: "The study came out from a pretty big gamble," recalls Sinclair, who used the human enzyme to screen and identify molecules that he expected would also stimulate those related enzymes in lower organisms. Unlike SIRT-1, these related enzymes are known to increase longevity when activated, usually by restricting the organism\'s calorie intake. Not only did they find "a whole collection of related polyphenols that activate 'Sir-2 from humans,' ... but we put them onto yeast, justbeing the simplest model, and amazingly \[they\] did what we were hoping \[they\] would do," recalls Sinclair. "But it was a real long shot." Now there\'s great eagerness in the Sinclair laboratory to complete and publish related research, notably by replicating the yeast work in higher organisms. "We have very promising results in Drosophila, which is a huge jump from a yeast cell," says Sinclair. "So we\'re very encouraged by that." Publication of these results is imminent. The team has also quickly broadened its horizons and is already testing the polyphenols on mouse disease models. "We think we may have tapped into a cell survival and defence programme \[and\] that these molecules will be useful for treating diseases associated with ageing, like diabetes and Alzheimer\'s," says Sinclair. He hopes to publish the diabetes results by mid-2004 and those for Alzheimer\'s by the end of the year. Harvard and BIOMOL Research Laboratories, its commercial partner based in Pennsylvania, have already filed a patent application for the use of "synthetic related molecules" to combat diseases of ageing---an application, Sinclair adds, "very much linked to the \[polyphenols\] paper." There\'s been a radical shift in attitude towards ageing, says Sinclair. Before the 1990s, "people thought that we were a lot like cars, that we would just rust and breakdown---nothing we could do about it. The new idea is that there are pathways that can boost our defences against ageing---the 'ageing-can-be-regulated' discovery ... that genes can control ageing \[and\] that there are pathways that \[we can use to\] slow down the process," he says. "If that\'s true---and it really seems to be true for a lot of organisms---if it\'s true for us, it really means that there is hope that we will be able, one day, to find small molecules that can alter these pathways." How Long Could We Live? {#s3} ======================= Sinclair expects to see such developments within his lifetime, but he ridicules the notion that humans will experience anything like the 70% extension to lifespan of his cultured yeast. "It\'ll be great if we can just give people an extra five years and have less disease in their old age and make it less painful," he says. "We won\'t be seeing any Methuselahs around," he insists. On his side are James Vaupel, one of Europe\'s leading demographers, and Marc Mangel, a mathematical modeller at the University of California at Santa Cruz. "Since 1840, life expectancy has been going up at 2.5 years per decade and will continue at this rate, maybe a little faster," says Vaupel, head of the Laboratory of Survival and Longevity at the Max Planck Institute for Demographic Research in Rostock, Germany. Women in Japan currently have the highest average life expectancy of 85, he notes: "So the figure could be 100 in six decades, but not 500." There\'s remarkably little people can do even if they want to live as long as possible, he says. "Give up smoking, lose weight, don\'t drive when drunk, install a smoke detector, take regular exercise," suggests Vaupel, who insists he does them all, as well as taking vitamin supplements. > *"You look at these worms and think, 'Oh my God, these worms should be dead.' But they\'re not. They\'re moving around."* Mangel sees the problem of assessing the limitations of ageing research as fairly straightforward. Mathematical models, he says, could solve it by linking demographic properties and physiological developments. "We\'ve had a separation of the biology of ageing and the demography of ageing, and they need to come together again," notes Mangel, whose personal anti-ageing regime involves taking "a dose of anti-oxidant chocolate with a good feeling." But Cythnia Kenyon, whose laboratory reported in October that it had generated a 6-fold increase in the lifespan of its nematodes, is not so sure about the limitations. "You look at these worms and think, 'Oh my God, these worms should be dead.' But they\'re not. They\'re moving around.... Once you get your brain wrapped around that ... then you start thinking, oh my goodness, so lifespan is something you can change---it\'s plastic. Then who knows what the limit is?" (Cynthia Kenyon has recorded video clips of the superstars of her lab, Caenorhabditis elegans, to show how long-lived mutant nematodes are as vigorous as normal young adults \[[Videos 1--4](#pbio-0020012-v001){ref-type="other"}\].) Warming to the theme, Kenyon hypothesises: "If you\'d asked me many generations ago, when we were actually common precursors of worms and flies, 'Cynthia, you have a two-week lifespan, do you think that you could \[live longer\]?' And if I\'d told you, 'Well, I think our descendants will live 1,000 times longer,' you\'d have said, 'Oh, come on!' But we do. It happened," she notes. "Who knows what you could do in people?" Kenyon muses. "I don\'t want to go on record saying that it\'s not possible in people because I don\'t see why it wouldn\'t be.... I\'m certainly not imagining that my company in the next few years is going to come up with a compound that can make people live to be 500. That seems just preposterous." So the timescale is millions of years? "No, not necessarily," she insists, "because once we understand the mechanism, then we can intervene and see what we can accomplish." ::: {.caption} ###### Box 1. Model Systems for Ageing ::: Yeast, as a model system for ageing, is at a distinct disadvantage. It lacks an endocrine system, and yet much research indicates that the key to longevity is control of hormones such as insulin and insulin-like growth factor 1 (IGF-1), as well as their downstream pathways and associated tissues, including the reproductive network. But David Sinclair, whose laboratory used yeast to show how an elixir might extend life, remains sanguine. "If it doesn\'t have an endocrine system, we can\'t understand cell-to-cell communication, but not all of ageing is just communication," he says. "There are things that occur inside the cells that provide longevity, and that\'s where yeast can be applied." Sinclair, assistant professor of pathology at Harvard Medical School, discovered a group of polyphenols that cause the human enzyme SIRT-1 and its homologues in lower organisms, including Sir-2 in yeast, to deacetylate the p53 protein and its homologues, notably the histones H3 and H4, in yeast. "Our findings are that the activation of the pathway downregulates p53\'s ability to cause cell death," he notes. Although p53, the tumour suppressor, is known to be involved in programmed cell death, it is not known whether SIRT-1 has any role in ageing. So Sinclair "went straight back to yeast to prove the principle of longevity extension." He found that deacetylisation of histones in yeast caused the DNA that\'s wrapped around them to become more compact and thus more stable. "DNA stability is key to longevity, and Sir-2 promotes that," says Sinclair. "We don\'t know yet whether it\'s the same in humans." Cynthia Kenyon, meanwhile, sees worms as the optimal model for helping to substantiate the links to humans. Besides the nematodes\' being multicellular organisms with endocrine systems, she also notes that their short lifespan of around 20 days is a big advantage: "You can do lots of experiments with them." Mice, which have short lifespans for mammals, still live two years, and long-lived mice for three or four years, she notes. And the advantage of fruit flies? "It\'s good to use more than one animal." Kenyon, professor of biochemistry and biophysics at the University of California at San Francisco, has focused on decoding the role of genes in ageing, notably *daf-2*, whose receptor is similar to those for insulin and IGF-1 in humans and inhibits ageing, and *daf-16*, which promotes it. "The DAF-2 receptor activates a highly conserved PI-3-kinase, the PDK/Akt pathway, and that pathway affects ageing, at least in part, by inhibiting the activity of the DAF-16 transcription factor," says Kenyon. "It does so by phosphorylating DAF-16 and inhibiting its entry into the nucleus." She adds: "We think that the DAF-2 pathway has another way of influencing ageing ... but we don\'t know what this other way is." In the long-lived mutants, which are defective in the *daf-2* receptor gene or in the genes encoding downstream signalling components, such as the PI-3-kinase, DAF-16\'s activity in the nucleus leads to the changes in expression of a wide variety of downstream genes, between 100 and 200, estimates Kenyon. Her studies show, she says, that a large number of those genes influence ageing. Signalling Life and Sweet 16 {#s4} ============================ Kenyon, professor of biochemistry and biophysics at the University of California at San Francisco, is among the key contributors responsible for showing that a single gene, and subsequently many genes, can change an organism\'s lifespan. > "It is inconceivable ... that a life-extending therapy will ever be developed that is able to extend life independent of every other change." In a seminal paper published a decade ago, Kenyon\'s laboratory showed that mutations in the *daf-2* gene doubled the lifespan of the nematode C. elegans. *daf-2* encodes a receptor that is similar to those for insulin and insulin-like growth factor-1 (IGF-1) in humans; this hormone receptor normally speeds up ageing in worms, but the mutations inhibit its action and enable the organisms to live longer. Before the results appeared, there was a "very negative attitude" towards ageing research, recalls Kenyon. Since then, and especially over the past few years in response to later findings, graduate students have been scrambling for a chance to work in her laboratory. "You can\'t believe the difference---there was such resistance to it," she says. "*daf-2* made a huge difference." But then so did her subsequent research in the field. Among her most significant findings is the identification of many more longevity genes; the results, published in July, derive directly from her early work on *daf-2*. "We discovered that in order for long-lived worms to live so long, they need another gene called *daf-16*," recalls Kenyon. "*daf-16* is kind of the opposite of *daf-2*, in the sense that it promotes longevity and youthfulness ... so we call it 'sweet 16.'" *daf-16* encodes a transcription factor that controls the expression of more than 100 genes. "They don\'t do just one thing, they do many things," says Kenyon. They can act as anti-oxidants (to prevent damage from oxygen radicals), as chaperones (to prevent misfolded proteins from forming aggregates), as antimicrobials (to protect against bacteria and fungi), and as metabolic agents. "So the picture that emerges is that the way the insulin/IGF-1 hormone system produces these enormous effects on lifespan is by coordinating the expression of many genes that do different things to affect lifespan, each of which on its own has only a small effect," notes Kenyon. "It\'s as though *daf-2* and *daf-16*, the regulators, would be the conductors of an orchestra. They bring together the flutes and the violins and the French horns, each of which do different things, and they make them all work together in concert." Kenyon is unequivocal about the bottom line: "Now we have a whole set of genes whose biochemical functions we can be working on to understand more about the actual mechanisms of ageing." Complementary results in flies and mammals persuade her to be more explicit. "The common ancestor of worms, flies, and mice must have had an insulin/IGF-1-like hormone system that controlled ageing. And that ability has been maintained. So the question is, has \[that ability\] been lost in humans? I think it\'s quite likely that it will also function in humans, but there isn\'t a direct demonstration yet that that\'s the case." Nevertheless, the discoveries about the role of the insulin/IGF-1 pathway in ageing have had a profound impact on her own lifestyle, which includes a tendency to discard the bread from sandwiches and eat only the toppings of pizzas ([Box 2](#box1){ref-type="boxed-text"}). "I\'m on a low-carb diet. I gave my worms glucose, and it shortened their lifespan. \[The diet\] makes sense because it keeps your insulin levels down," she says. "Caloric restriction extends lifespan of mice, and so does the insulin/IGF-1 pathway," she notes. Indeed, starting a low-calorie diet at any point in adulthood appears to help fruit flies live longer, according to research in Britain published last September. "What we don\'t know for sure in mice," Kenyon continues, "is whether the two pathways are different or the same." While much ageing research focuses on these two influences, she says that there are another two areas of investigation. Her laboratory reported in December 2002 that inhibiting the respiration of mitochondria in developing worms increased longevity, but that it had no effect in adult worms, for reasons still unexplained, she says. Further microarray analysis is underway to pinpoint whether the cause simply lies downstream of the insulin/IGF-1 pathway or whether it is something different altogether. The Price of Life {#s5} ================= Then there\'s research looking at the effects on lifespan of changes to an organism\'s reproductive system. For Kenyon, such work often involves a battle to convince sceptics that longevity is not a trade-off with fertility. Four years ago, her laboratory reported that killing germ cells increases the lifespan of worms by 60%, but only because, she stresses, it affects endocrine signalling and not because it prevents reproduction. Further research, published last year, showed quite clearly, she says, that ageing and reproduction are controlled independently of one another. And as for her recent work on infertile worms, which lived six times as long as normal following the removal of their entire reproductive systems, she says: "If we could intervene in the hormone signalling pathways directly, we think the animals would still live six times as long as normal, but would be fertile as well." Jim Carey is one of those "trade-off" sceptics. He is a population biologist at the University of California at Davis and his research, on the effect on life expectancy of replacing the ovaries of old mice with ovaries from younger mice, is intended to complement Kenyon\'s work. But he insists that "an honest discussion of lifespan extension must include consideration of tradeoffs." Many manipulations that extend lifespan in model systems, whether genetic or dietary, for example, ignore or gloss over the side effects, such as permanent sterility, huge weight loss, distorted organ-to-body ratios, or major behavioural aberrations, he notes. "It is inconceivable to me that a life-extending therapy will ever be developed that is able to extend life independent of every other change," he concludes. "All life systems are interlinked and hierarchically integrated at all levels, so to talk about life extension using analogies with a car warranty concept is wrong-headed." Another "trade-off" sceptic takes a different tack. As Armand Leroi puts it: "During occasional periods of involuntary celibacy I have thought, well, I may not be getting laid, but at least I shall live to a miserable and solitary old age." Leroi, an evolutionary biologist at Imperial College of Science, Technology, and Medicine in London, offers an optimistic appraisal of the chances of finding a cure for ageing in his new book about the effects of genetic variety on the human body. He sees ageing simply as a collection of curable diseases: "There is no obvious impediment to that advance, nothing to make us think that human beings have a fixed lifespan." ::: {.caption} ###### Box 2. Practise What You Preach ::: Cynthia Kenyon\'s eating habits are defined by her ageing research on worms. "There\'s a lot of these diets ... and what they all have in common is *low carb*---actually, low glycaemic index carbs," she says. "That\'s not eating the kind of carbohydrates where the sugar gets into your bloodstream very quickly \[and stimulates production of insulin\]." No desserts. No sweets. No potatoes. No rice. No bread. No pasta. "When I say 'no,' I mean 'no, or not much,'" she notes. "Instead, eat green vegetables. Eat the fruits that aren\'t the sweet fruits, like melon." Bananas? "Bananas are a little sweet." Meat? "Meat, yes, of course. Avocados. All vegetables. Nuts. Fish. Chicken. That\'s what I eat. Cheese. Eggs. And one glass of red wine a day." Kenyon, professor of biochemistry and biophysics at the University of California at San Francisco, has been on her diet for two-and-a-half years. "I did it because we fed our worms glucose and it shortened their lifespan." But the diet is unproven, she cautions, and she\'s not recommending it for all. Nevertheless, she\'s pleased with its performance for her. "I have a fabulous blood profile. My triglyceride level is only 30, and anything below 200 is good." Kenyon is angered by the general lack of nutritional knowledge: "It\'s a little bit embarrassing to say that scientists actually don\'t know what you should eat.... We can target particular oncogenes, but we don\'t know what you should eat. Crazy," she says. Does her dieting represent a return to scientists experimenting on themselves? "I don\'t think so---you have to eat something, and you just have to make your best judgement. And that\'s my best judgement. Plus, I feel better. Plus, I\'m thin---I weigh what I weighed when I was in college. I feel great ---you feel like you\'re a kid again. It\'s amazing." ::: {#pbio-0020012-v001 .fig} Video 1 ::: {.caption} ###### Normal Nematodes at Day 1 of Adulthood (Video used by permission from Cynthia Kenyon.) ::: ::: ::: {#pbio-0020012-v002 .fig} Video 2 ::: {.caption} ###### Long-Lived *daf-2* Mutants at Day 1 of Adulthood (Video used by permission from Cynthia Kenyon.) ::: ::: ::: {#pbio-0020012-v003 .fig} Video 3 ::: {.caption} ###### Normal Nematodes at Day 13 of Adulthood The worm on the left is dead. (Video used by permission from Cynthia Kenyon.) ::: ::: ::: {#pbio-0020012-v004 .fig} Video 4 ::: {.caption} ###### A Long-Lived *daf-2* Mutant at Day 13 of Adulthood (Video used by permission from Cynthia Kenyon.) ::: ::: Bill O\'Neill is a freelance journalist from London, United Kingdom. E-mail: <billoneill@cix.co.uk> IGF-1 : insulin-like growth factor-1

PubMed Central

2024-06-05T03:55:46.051503

2004-1-20

PMC338072

Introduction {#s1} ============ The Wnt signal transduction pathway has been studied extensively in both vertebrate and invertebrate systems. The *Drosophila* ortholog *wingless* (*wg*) is a segment polarity gene that defines posterior cell fates in each of the larval segments (for a review of the various functions of Wg, see [@pbio-0020095-Wodarz1]). The pathway is activated when the extracellular ligand Wg binds to the transmembrane receptors Frizzled and Arrow. These in turn activate Disheveled (Dsh), which inactivates a complex composed of Axin, adenomatous polyposis coli (APC), and Zeste-white 3 (Zw3) (the *Drosophila* homolog of glycogen synthase kinase \[GSK3β\]). This complex is responsible for the retention of Armadillo (Arm) in the cytoplasm, for its phosphorylation, and thus for its targeting for ubiquitination and destruction. When the complex is inactivated by Dsh, the intracellular levels of Arm increase, and Arm enters the nucleus, where in combination with the transcription factor TCF/Pangolin, it activates the transcription of genes such as cyclin D and c-*myc* ([@pbio-0020095-Wodarz1]). We have argued that Axin plays a key role in the Wnt signaling process, functioning both as an anchor for Arm and a scaffold for the degradation complex. Wnt signaling results in a visible reduction in Axin levels, and mutations in Axin cause a relocalization of Arm to the nucleus ([@pbio-0020095-Tolwinski1]; [@pbio-0020095-Tolwinski2]). The nuclear import and export of Arm are not clearly understood (for a review, see [@pbio-0020095-Henderson1]), but Arm can cross the nuclear membrane by interacting with the nuclear pore complex directly. Once in the nucleus, Arm interacts with a variety of nuclear factors, in particular the transcription factor TCF/LEF ([@pbio-0020095-Behrens1]; [@pbio-0020095-Molenaar1]; [@pbio-0020095-Brunner1]; [@pbio-0020095-van1]). The β-catenin--TCF complex releases repression and activates transcription ([@pbio-0020095-Cavallo1]). A recent study has challenged this view and has questioned the importance of nuclear localization of Arm protein ([@pbio-0020095-Chan1]). These authors\' conclusions were based primarily on the observation that a membrane-tethered, stabilized form of Arm (Arm^ΔArm^) causes activation of the Wnt pathway without entering the nucleus. However, this is not the first time that the controversy about the location of Arm/β-catenin function has arisen. Previously, a group working with amphibian embryos had found that membrane-tethered plakoglobin, a close relative of β-catenin, can activate Wnt signaling ([@pbio-0020095-Merriam1]). Another group showed, however, that expression of membrane-tethered forms of β-catenin leads to the nuclear localization of endogenous β-catenin ([@pbio-0020095-Miller1]). When the endogenous Arm/β-catenin gene was mutated, the activity of membrane-tethered forms was lost ([@pbio-0020095-Cox2]b). These experiments illustrate the importance of following the activity of the endogenous allele in evaluating the activity of membrane-tethered forms. Previously, we had expressed the same membrane-tethered form used by [@pbio-0020095-Chan1]) in embryos with various endogenous *arm* mutations and had concluded that it functions by titrating Axin to the membrane, releasing the endogenous Arm protein and allowing it to move freely into the nucleus ([@pbio-0020095-Tolwinski1]). These experiments are difficult, because none of the cell-viable alleles are absolute genetic nulls, as Arm plays essential roles in both Wnt signaling and cell adhesion. In this study, we reexamine Arm function using three classes of previously described *arm* alleles. We find that by manipulating their levels and localizations, many alleles believed to be signaling nulls can still activate transcription. When the cell-adhesive defects of the most severe class of alleles are rescued, however, the mutant protein still fails to signal, allowing us to assay the activity of membrane-tethered Arm in a true signaling-null background. We find that nuclear localization is necessary for pathway activation and that exclusively membrane-bound forms of Arm are insufficient for this. We use two novel missense mutations in *arm* to assess the nuclear activity of Arm and confirm that negative regulation by the transcriptional regulator Chibby (Cby) is required for patterning. Results {#s2} ======= Membrane-Tethered Arm Is Dependent upon the Endogenous *arm* Allele {#s2a} ------------------------------------------------------------------- The original mutants in the *arm* gene were classified into three groups based upon their phenotypes and the position of stop codons that result in truncated proteins. The "weak" class has the smallest truncations and is represented by *arm^XM19^*. In germline clones (where maternal and zygotic contribution of protein is removed; [@pbio-0020095-Chou1]), its phenotype is identical to loss-of-function *wg* mutations ([Figure 1](#pbio-0020095-g001){ref-type="fig"}B; [@pbio-0020095-Peifer1]). The "medium" class, represented here by *arm^O43A01^*, shows defects in adhesion as well as transcription. Here germline clones give embryos that fail to differentiate an intact cuticle ([Figure 1](#pbio-0020095-g001){ref-type="fig"}C; [@pbio-0020095-Tolwinski1]). The "strong" class (*arm^XK22^*) does not allow proper progression through oogenesis and germline clones do not make eggs ([Figure 1](#pbio-0020095-g001){ref-type="fig"}D; [@pbio-0020095-Peifer2]). [@pbio-0020095-Cox2]b) showed that the junctional defects of the "medium" alleles can be circumvented by coexpression of a membrane-tethered full-length form of Arm (Arm^S18^) ([Figure 2](#pbio-0020095-g002){ref-type="fig"}). We have confirmed their findings and extended them to the "strong" allele during oogenesis. We show that uniform expression of Arm^S18^ allows *arm^XK22^*germ cells to produce normal eggs and rescues the adhesive defects of both *arm^XK22^* and *arm^O43A01^* embryos. The membrane-tethered form does not, however, rescue the signaling defects associated with either of these alleles and the embryos show typical *wg* phenotypes (see [Figure 1](#pbio-0020095-g001){ref-type="fig"}G and [1](#pbio-0020095-g001){ref-type="fig"}H). ::: {#pbio-0020095-g001 .fig} Figure 1 ::: {.caption} ###### Arm^ΔArm^ Requires Endogenous Arm Endogenous allele indicated at top; ectopically expressed transgenes indicated at left. \(A) The wild-type cuticle of a *Drosophila* embryo. \(B) The *arm^XM19^* "weak" allele phenotype, similar to *wg* mutations in which the entire cuticle is covered with denticles. \(C) The *arm^O43A01^* "medium" allele phenotype shows disintegrated embryos in which cells delaminate owing to an inability to form adherens junctions. \(D) *arm^XK22^* "strong" allele does not produce embryos, owing to an oogenesis defect. \(E) A wild-type embryo expressing Arm^S18^ shows a wild-type cuticle. \(F) *arm^XM19^* mutant expressing Arm^S18^ is rescued to a wild-type cuticle. \(G) *arm^O43A01^* mutant expressing Arm^S18^ shows rescued adhesion, but a *wg* mutant signaling phenotype. \(H) *arm^XK22^* mutant expressing Arm^S18^ also shows rescued adhesion, as well as a *wg* mutant signaling phenotype. \(I) Coexpression of Arm^ΔArm^ and Arm^S18^ in wild-type embryos leads to naked cuticle or the uniform Wg active phenotype. \(J) Coexpression of Arm^ΔArm^ and Arm^S18^ leads to naked cuticle or the uniform Wg active phenotype in an *arm^XM19^* mutant background. \(K) Coexpression of Arm^ΔArm^ and Arm^S18^ in *arm^O43A01^* mutant embryos leads to naked cuticle or the uniform Wg active phenotype. \(L) However, coexpression of Arm^ΔArm^ and Arm^S18^ in "strong" mutant *arm^XK22^* background shifts embryos back to the *wg* mutant phenotype. Expression of the membrane-tethered, stabilized form of Arm (Arm^ΔArm^) leads to uniform activation of signaling in all cells. This effect is independent of whether the cell is exposed to Wg signal or not, because Arm^ΔArm^ functions independently of Wg ligand. The membrane-tethered, unstabilized form of Arm (Arm^S18^) leads to pathway activation only in cells that receive Wg signal, because this form of Arm is still subject to Wg-dependent phosphorylation and phosphorylation-dependent degradation. :::  ::: ::: {#pbio-0020095-g002 .fig} Figure 2 ::: {.caption} ###### Structure of Arm Protein and Alleles Arm protein consists of three regions. The N-terminus is required for transactivation, for phosphorylation-based and proteasome-mediated degradation, and for α-catenin binding. The central repeats region is a superhelical structure that contains the binding sites for most of Arm\'s binding partners, including APC, TCF, Cadherin, and Axin. The C-terminus is required for Cby and Teashirt binding and transactivation. The *arm^F1a^* mutation causes an arginine-to-histidine change within repeat six. The *arm^LM134^* mutation causes a serine-to-phenylalanine change in repeat five. The "weak" allele *arm^XM19^* removes the entire C-terminus. The "medium" allele *arm^O43A01^* causes early termination within repeat nine. The "strong" allele *arm^XK22^* causes early termination within repeat six. The Arm^ΔArm^ transgene consists of the entire repeats region and C-terminus fused to an HA tag and myristoylation sequence at the N-terminus under GAL4/UAS control. The Arm^S10^ transgene contains a small deletion in the N-terminus, which removes the four phosphorylation sites necessary for degradation and is under GAL4/UAS control. The Arm^S8^ transgene contains a deletion of approximately a third of the C-terminus and is under endogenous promoter control. The Arm^S18^ transgene contains the entire Arm sequence fused to the CAAX myristoylation sequence of Ras and is under endogenous promoter control. The UAS--Arm^XM19^ is the equivalent of the *arm^XM19^* allele in deletion, but is fused to an N-terminal HA tag and is under GAL4/UAS control. :::  ::: Expression of Arm^S18^ has no effect on the cuticle of wild-type embryos (compare [Figure 1](#pbio-0020095-g001){ref-type="fig"}E to [1](#pbio-0020095-g001){ref-type="fig"}A), but it does rescue the signaling defects of *arm* alleles, like *arm^XM19^*, that have only short C-terminal truncations ([@pbio-0020095-Cox2]b; see [Figure 1](#pbio-0020095-g001){ref-type="fig"}F). These alleles normally show very low levels of protein ([@pbio-0020095-Peifer1]), and [@pbio-0020095-Cox2]b) postulated that expression of a membrane-tethered Arm might "free up" the endogenous mutant protein, allowing the "weak" allele to signal. The low levels of *arm^XM19^* may reflect degradation of nonsense mRNAs triggered by the premature stop codon in this mutant (reviewed in [@pbio-0020095-Wagner1]). To eliminate this degradation, we expressed a cDNA version of the *arm^XM19^* allele under GAL4/UAS control ([@pbio-0020095-Brand1]) in embryos mutant for *arm^XM19^* ([Figure 3](#pbio-0020095-g003){ref-type="fig"}B). To avoid the possibility of overexpression artifacts, we also expressed a smaller C-terminal deletion from the endogenous promoter (Arm^S8^; [@pbio-0020095-Orsulic1]). In both experiments, the truncated protein from the transgene accumulated to levels approximating those observed in wild-type ([Figure 3](#pbio-0020095-g003){ref-type="fig"}G and [3](#pbio-0020095-g003){ref-type="fig"}H) and in the characteristic striped pattern indicative of response to the Wg signal ([@pbio-0020095-Peifer1]). The truncated protein rescued the *arm^XM19^* phenotype to a wild-type cuticle pattern and allowed hatching ([Figure 3](#pbio-0020095-g003){ref-type="fig"}C). When combined with a mutation in the kinase *zw3*, Arm^S8^ causes the cuticles of these embryos to appear uniformly naked (compare [Figure 3](#pbio-0020095-g003){ref-type="fig"}E to [3](#pbio-0020095-g003){ref-type="fig"}D), as would be expected since the Arm^S8^ protein is expressed to high uniform levels throughout the epidermis when Zw3 is removed ([Figure 3](#pbio-0020095-g003){ref-type="fig"}I). These experiments argue that the C-terminus is not essential for signaling or transcriptional activation of Wnt targets required for cuticle patterning. However, as we do not obtain adult flies containing exclusively the truncated alleles, it is very likely that the C-terminus is not entirely expendable and must have important functions later in development. ::: {#pbio-0020095-g003 .fig} Figure 3 ::: {.caption} ###### C-Terminally Truncated Arm Can Signal If Its Levels Are Increased \(A) *arm^XM19^* shows a *wg* mutant phenotype. \(B) Expression of GAL4/UAS-driven Arm^XM19^ protein in *arm^XM19^* mutant background rescues this to a wild-type pattern. \(C) The same is true of expression of an endogenous promoter-driven truncation Arm^S8^. \(D) Removal of Zw3 has no effect on *arm^XM19^* cuticle pattern. \(E) However, when Arm^S8^ is introduced into *arm^XM19^, zw3* mutants, the cuticle is naked. \(F) Wild-type embryo is shown for comparison. (G--I) Arm stainings reveal that expression of UAS--Arm^XM19^ (stained for the HA tag \[G\]) and Arm^S8^ (stained for Arm \[H\]) is present in stripes corresponding to Wg striping, whereas removal of Zw3, along with Arm^S8^ expression, leads to uniform and high levels of Arm throughout the epidermis (I). :::  ::: Null Allele Background Proves That Arm^ΔArm^ Cannot Signal on Its Own {#s2b} --------------------------------------------------------------------- The fact that *arm^XM19^* is able to signal when expressed at normal levels invalidates its use in tests for a direct activity of membrane-tethered Arm in Wnt signaling ([@pbio-0020095-Chan1]). Therefore, expression of Arm^ΔArm^ in a "weak" allele background cannot address whether membrane-tethered Arm activates transcription without ever entering the nucleus, since a membrane-untethered, signaling-competent form of Arm is also present. To directly address whether the Arm^ΔArm^ transgene can transmit Wg signal on its own, we turned to the "strong" and "medium" alleles. Although Arm^S18^ is not sufficient to restore signaling to these alleles, it raises the possibility that stronger expression of stabilized, membrane-tethered Arm (Arm^ΔArm^) might reveal some signaling capacity of those alleles as well. Experiments of this kind have been difficult with Arm^ΔArm^, given that it lacks the α-catenin-binding site and fails to rescue the junctional defect in "medium" and "strong" endogenous *arm* allele backgrounds. We have found that by expressing both Arm^ΔArm^ and Arm^S18^, we can recover intact embryos in all backgrounds tested. We find that "medium" and "weak" alleles can be induced to activate transcription, but the "strong" *arm* allele cannot (see [Figure 1](#pbio-0020095-g001){ref-type="fig"}J--[1](#pbio-0020095-g001){ref-type="fig"}L), consistent with the position of the "medium" alleles in the hypomorphic allelic series. These findings demonstrate that Arm^ΔArm^ is dependent upon the endogenous form of *arm*, as it cannot activate transcription in the "strong" allele background. Loss-of-Function Missense Mutations {#s2c} ----------------------------------- When Arm^ΔArm^ is expressed in a wild-type embryo, it strongly activates Wg signaling ([Figure 4](#pbio-0020095-g004){ref-type="fig"}C; [@pbio-0020095-Chan1]). [@pbio-0020095-Chan1]) suggest that this is because this membrane-tethered form of Arm can signal on its own. The results presented above argue, on the other hand, that it does so by stabilizing the endogenous protein. To further test this, we asked whether expression of Arm^ΔArm^ can induce Wg signaling when endogenous Arm is replaced by signaling-deficient Arm. We turned to two novel missense mutations where the rest of the *arm* coding region remains intact. Because these alleles do not produce truncations through stop codons, they are immune to nonsense mRNA degradation ([@pbio-0020095-Wagner1]). Both mutations result in amino acid substitutions close to repeat seven, a key hinge region postulated to be important in binding of TCF ([@pbio-0020095-Huber1]; [@pbio-0020095-Graham1]). Both mutants retain the phosphorylation sites required for degradation and therefore accumulate in stripes in response to Wg signal ([Figure 5](#pbio-0020095-g005){ref-type="fig"}I and [5](#pbio-0020095-g005){ref-type="fig"}J). They supply apparent wild-type junctional activity and accumulate to high levels in all cells when the kinase responsible for the degradation signal (Zw3) is removed ([Figure 5](#pbio-0020095-g005){ref-type="fig"}K and [5](#pbio-0020095-g005){ref-type="fig"}L). The primary phenotype of these alleles is a loss or reduction of Wnt transcriptional responses ([Figure 5](#pbio-0020095-g005){ref-type="fig"}A and [5](#pbio-0020095-g005){ref-type="fig"}B). The *arm^F1a^* allele produced a partial loss-of-function phenotype, and germline clone embryos show some residual naked cuticle. *arm^LM134^* produces a stronger phenotype comparable to a loss of *wg* function, although it may not be a signaling null (see below). ::: {#pbio-0020095-g004 .fig} Figure 4 ::: {.caption} ###### Expression of Arm^ΔArm^ Leads to the Nuclear Localization of Endogenous Arm Protein \(A) Wild-type Arm protein appears in stripes that correspond to cells responding to Wg signaling. \(B) Expression of Arm^ΔArm^ in an *arm^F1a^* background leads to the nuclear localization of endogenous Arm. (C and D) Dark-field images reveal that expression of both Arm^ΔArm^ and Arm^S10^ leads to similar naked cuticle phenotypes. \(E) An anti-Arm Western blot showing a faster-migrating band, which correlates with endogenous Arm\'s being active, and a slower-migrating band, which correlates with Arm\'s being inactive. :::  ::: ::: {#pbio-0020095-g005 .fig} Figure 5 ::: {.caption} ###### ΔArm Functions through Endogenous Arm \(A) Embryonic cuticle of *arm^F1a^* mutant showing a weak loss-of-function phenotype. \(B) Cuticle of *arm^LM134^* mutant embryo showing a strong loss-of-signaling phenotype. \(C) Embryo mutant for *arm^F1a^* expressing Arm^ΔArm^ showing relatively normal segment polarity. \(D) *arm^LM134^* mutant expressing Arm^ΔArm^ also shows segment polarity. (E and F) Both alleles in combination with a null *zw3* allele and expressing Arm^ΔArm^ show a complete lack of denticles. (G and H) Both alleles expressing the activated but nontethered form of stabilized Arm, Arm^S10^, show the naked cuticle phenotype. (I--L) In both missense alleles, the mutant protein is expressed in stripes (I and J), corresponding with Wg expression (data not shown), which is abolished when the key degradation kinase Zw3 is removed (K and L). :::  ::: We asked whether these signaling-deficient alleles could block the cell fate transformation and Wnt target activation observed when Arm^ΔArm^ is expressed in wild-type epidermis. If Arm^ΔArm^ functions independently of the endogenous protein, then all cells should assume the naked cell fate. However, this does not occur ([Figure 5](#pbio-0020095-g005){ref-type="fig"}C and [5](#pbio-0020095-g005){ref-type="fig"}D). Instead, both point mutants produce a cuticle pattern with periodic denticle belts and regions of intervening naked cuticle. This periodicity may reflect the fact that *arm^F1a^* and *arm^LM134^* can still be controlled by Wg even when Arm^ΔArm^ is expressed. This periodicity is, in fact, abolished when Zw3 activity is removed from such embryos (i.e., in triply mutant *zw3, arm^F1a^;*Arm^ΔArm^ embryos). Under these conditions, all cells in the cuticle take on the naked cell fate ([Figure 5](#pbio-0020095-g005){ref-type="fig"}E and [5](#pbio-0020095-g005){ref-type="fig"}F). Since Arm^ΔArm^ lacks the N-terminal sites that respond to Zw3, the sensitivity of the double-mutant phenotype confirms that the pattern of the double mutant is dependent on the endogenous Arm protein. The behavior of membrane-tethered Arm^ΔArm^ contrasts with that of other stabilized forms of Arm that would be predicted to move more freely between the cytoplasm and the nucleus. Arm^S10^, for example, contains a small N-terminal deletion that blocks Zw3 phosphorylation, but preserves binding sites for various nuclear proteins (see [Figure 2](#pbio-0020095-g002){ref-type="fig"}; [@pbio-0020095-Pai2]). Arm^S10^ is not membrane-tethered, but the cell fate transformations it produces are identical to those produced by Arm^ΔArm^ (compare [Figure 4](#pbio-0020095-g004){ref-type="fig"}C and [4](#pbio-0020095-g004){ref-type="fig"}D). They do not, however, depend on the endogenous allele and are still observed in an *arm^F1a^* or *arm^LM134^* germline clone background ([Figure 5](#pbio-0020095-g005){ref-type="fig"}G and [5](#pbio-0020095-g005){ref-type="fig"}H). Arm^ΔArm^ Causes Nuclear Localization and Mobility Shift of Endogenous Arm {#s2d} -------------------------------------------------------------------------- All of our experiments argue that Arm^ΔArm^ produces its effect on transcription by activating the endogenous alleles. To investigate the mechanism that underlies this effect, we looked at the in situ localization of the endogenous Arm protein and its migration pattern on Western blots. Expression of Arm^ΔArm^ is sufficient to drive both wild-type and the point mutant forms of Arm into nuclei (see [Figure 4](#pbio-0020095-g004){ref-type="fig"}A and [4](#pbio-0020095-g004){ref-type="fig"}B; [@pbio-0020095-Miller1]; [@pbio-0020095-Tolwinski1]). Generally, the most obvious feature observed upon removal of any of the negative factors of the Wg pathway is the rapid accumulation of Arm in cells. However, another feature is the phosphorylation state of the Arm protein. [@pbio-0020095-Peifer3]a) found that a fast-migrating band of Arm corresponds with active Wg signaling and that a slower-migrating band corresponds with Wg\'s being off. Therefore, it is the unphosphorylated band that corresponds with signaling. Here we show that, on Western blots, endogenous Arm protein responds to Arm^ΔArm^ expression in much the same way that it does to the removal of negative components of the pathway such as Axin and APC1 and APC2 (see [Figure 4](#pbio-0020095-g004){ref-type="fig"}E). We see a downshift of the protein, which is directly opposite to what is seen when a positive component of the pathway is removed (Dsh or Wg; see [Figure 4](#pbio-0020095-g004){ref-type="fig"}E). Wild-type embryos show the expected intermediate phenotype, as they have both active and inactive forms of Arm protein (see [Figure 4](#pbio-0020095-g004){ref-type="fig"}E). The observed shift is most likely the result of phosphorylation ([@pbio-0020095-Peifer3]a), though we do not address this directly in this study. The C-Terminus of Arm Is Necessary for Cby-Mediated Repression {#s2e} -------------------------------------------------------------- Although the missense mutations we have used in our studies produce (on average) weaker phenotypes, they are more effective at blocking the cell-fate transformation induced by Arm^ΔArm^ than the "medium" C-terminal truncation mutants (compare [Figure 1](#pbio-0020095-g001){ref-type="fig"}K with [Figure 5](#pbio-0020095-g005){ref-type="fig"}C and [5](#pbio-0020095-g005){ref-type="fig"}D). The comparison is somewhat indirect, owing to the necessity of expressing Arm^S18^ in the "medium" *arm* allele background in order to get intact embryos. However, we find that expression of Arm^S18^ in an *arm^F1a^* background has no visible effect on the cuticle (data not shown). Therefore, the activity of C-terminally truncated *arm* alleles in response to ΔArm expression suggests that, under certain conditions, removal of the C-terminus may actually enhance the transcriptional activity of Arm. One possibility is suggested by the recent discovery of Cby ([@pbio-0020095-Takemaru2]), a nuclear negative regulator of the Wg pathway that binds to the C-terminus of Arm. To test whether nuclear Cby affected the transformation produced by Arm^ΔArm^, we used RNA interference (RNAi) to reduce Cby levels in *arm^F1a^* embryos with and without Arm^ΔArm^. In the absence of Arm^ΔArm^, i.e., in embryos where most Arm^F1a^ protein is cytoplasmic, Cby RNAi has no effect ([Figure 6](#pbio-0020095-g006){ref-type="fig"}D). However, when Arm^ΔArm^ is present, lowering Cby levels leads to increased naked cuticle characteristic of Wnt pathway activation (compare [Figure 6](#pbio-0020095-g006){ref-type="fig"}B to [6](#pbio-0020095-g006){ref-type="fig"}C). We propose that Cby\'s effect on *arm^F1a^* protein is dependent on Arm^ΔArm^ relocalizing Arm to the nucleus. ::: {#pbio-0020095-g006 .fig} Figure 6 ::: {.caption} ###### Relief of C-Terminal Repression through the Elimination of Cby Leads to Uniform Activation of Signaling \(A) A wild-type cuticle shown for comparison. \(B) Expression of Arm^ΔArm^ in the *arm^F1a^* background. \(C) Expression of a Cby RNAi construct along with Arm^ΔArm^ in the *arm^F1a^* background. \(D) Expression of a Cby RNAi construct in an *arm^F1a^* background. :::  ::: Discussion {#s3} ========== In this study we offer genetic proof that the nuclear localization of Arm is important for the activation of the pathway. The dissenting view ([@pbio-0020095-Chan1]) relied on C-terminal truncations that we have shown retain their ability to signal if their levels are increased. These alleles also appear to bypass the normal nuclear regulation by Cby. We show that full-length loss-of-function forms of Arm provide a novel way of assessing the activity of the pathway. Finally, we show that in an approximate signaling-null condition, Arm^ΔArm^ cannot activate transcription on its own. Based on these findings, we propose that membrane-tethered Arm, whether wild-type or activated, cannot activate transcription on its own. It does, however, have a profound effect on the endogenous form, forcing both "weak" and "medium" alleles to translocate to the nucleus and activate transcription. Our findings extend and build upon the original nuclear localization of Arm model ([@pbio-0020095-Miller1]; [@pbio-0020095-Cox2]b). Further support for the nuclear localization of Arm model has recently been provided by the publication of a study that uses tissue culture experiments to show that nuclear localization of Arm is required ([@pbio-0020095-Cong1]). Our results also point to an unexpected feature of Arm, namely that the C-terminus, although it has been shown to supply transcriptional activation ([@pbio-0020095-Hsu1]), does not appear to be required for Wnt activation. [@pbio-0020095-Cox1]a) studied this aspect of Arm function and found that a C-terminally truncated form of Arm can significantly rescue the signaling defects of *arm* mutants, but is not as good as the wild-type form at transcriptional activation. Further, given that *arm* mutant flies expressing the transgene that lacks the C-terminus do not survive to adulthood, the C-terminus may not be entirely expendable. This may point to the requirement for Cby-based repression or Teashirt-mediated activation at a later stage of development, as both these proteins function by binding the C-terminus of Arm ([@pbio-0020095-Gallet1]; [@pbio-0020095-Takemaru2]). However, taken together with the finding that an N-terminally truncated Arm sent to the nucleus fails to activate transcription ([@pbio-0020095-Chan1]), it appears that it is the N-terminus that is most important for the nuclear transactivation and chromatin remodeling functions ascribed to β-catenin ([@pbio-0020095-Hsu1]; [@pbio-0020095-Hecht1]; [@pbio-0020095-Takemaru1]; [@pbio-0020095-Barker1]; [@pbio-0020095-Tutter1]; [@pbio-0020095-Bienz1]). We have previously shown that the "medium" *arm* mutant (*arm^O43A01^*, which creates a stop codon eliminating repeats 10 through 12 and the entire C-terminus) does not signal in the presence of uniform Arm^ΔArm^ ([@pbio-0020095-Tolwinski1]). [@pbio-0020095-Chan1]) found that *arm^O43A01^* embryos expressing high levels of Arm^ΔArm^ from the paired GAL4 driver were able to activate Wnt targets. But since neither Arm^ΔArm^ nor *arm^O43A01^* can provide junctional Arm activity, the abnormalities of these embryos make these experiments difficult to interpret. As an alternative, we used a membrane-tethered but otherwise wild-type form of Arm (Arm^S18^), which we expressed in *arm^O43A01^* mutant embryos (see [Figure 1](#pbio-0020095-g001){ref-type="fig"}G). The Arm^S18^ allele rescues the junctional defects, but does not allow signaling. Similar results have been obtained with another "medium" allele, *arm^XP33^* ([@pbio-0020095-Cox2]b). However, when combined with Arm^ΔArm^ and Arm^S18^, *arm^O43A01^* can now be clearly seen to activate naked cell fates. It thus appears that even the "medium" alleles of *arm* actually do retain some ability to function when Arm^ΔArm^ is present. This is not observed in the larger truncations ("strong" alleles), consistent with the "medium" alleles retaining the TCF-binding region ([@pbio-0020095-Graham1]). The question now becomes what is Arm^ΔArm^ doing at the membrane that causes such drastic change in the signaling kinetics of the pathway. We have previously argued that Arm^ΔArm^ may function by titrating the cytoplasmic anchoring activity of Axin and by therefore allowing rapid enrichment of Arm in the nucleus. We have in fact observed such an enrichment and have shown that it is counteracted by increasing the level of Axin ([@pbio-0020095-Tolwinski1]). Further work has pointed to the importance of controlling Axin stability in pathway activation ([@pbio-0020095-Salic1]; [@pbio-0020095-Mao1]; [@pbio-0020095-Lee1]; [@pbio-0020095-Tolwinski2]). Expression of large quantities of a stabilized, membrane-tethered form of Arm might also remove additional cytoplasmic inhibitory factors, preventing them from interacting with nontethered Arm. In turn, even lower-level or lower-activity alleles will now be able to activate transcription, simply owing to the complete lack of inhibiting factors. The missense mutations described here provide a glimpse of the in vivo activity of Arm protein. Structural studies of β-catenin found that although the central repeat region forms a uniformly repeating super helix, one α-helix was missing from repeat seven. The missing helix might allow a local flexibility in the structure and led the authors to define this region as a potential hinge ([@pbio-0020095-Huber1]). Further crystallographic analysis concluded that this region was important for TCF binding ([@pbio-0020095-Graham1]). Both our point mutations cluster around this repeat and would probably lead to structural consequences for this hinge. The apparent specificity of these alleles for the transcriptional response to Wnt signaling provides in vivo evidence that the postulated hinge may be very important for that aspect of Arm protein function. Note {#s4} ==== As the final version of this paper was being prepared, the paper by [@pbio-0020095-Chan1]) was retracted. Materials and Methods {#s5} ===================== Fly Strains {#s5a} ----------- The wild-type strain used was Oregon R. See Flybase (<http://flybase.bio.indiana.edu>) for details on mutants used. Hypomorphic mutants of *arm* are as follows: *arm^LM134^* TCC to TTC at nucleotide 2776, *arm^F1a^* CGC to CAC at nucleotide 2990, *arm^XM19^* stop codon at nucleotide 3850, *arm^O43A01^* stop codon at nucleotide 3404, *arm^XP33^* stop codon at nucleotide 3466, *arm^XK22^* stop codon at nucleotide 3013. Other alleles used were *axin^S044230^* , *zw3^M11--1^*, *dsh^V26^*, *apc1^Q8^*, *apc2^d40^*, and *wg^IG22^*. Crosses and Expression of UAS Constructs {#s5b} ---------------------------------------- ### *arm* mutants {#s5b1} As Arm and many other *Drosophila* proteins are contributed maternally, to fully evaluate the function of a mutant protein, one needs to make embryos maternally and zygotically mutant. Therefore, maternally mutant eggs were generated by the dominant female sterile technique ([@pbio-0020095-Chou1]). For all expression experiments, the Arm--GAL4 driver was used. All X-chromosome mutants use FRT 101. The *arm* mutants used were as follows: *arm^F1a^ zw3^M11--1^* (maternal)/Y (zygotic) *arm^LM134^ zw3^M11--1^* (maternal)/Y (zygotic) *arm^F1a^* (maternal)/Y (zygotic) *arm^LM134^* (maternal)/Y (zygotic) *arm^F1a^* (maternal)/Y (zygotic); Arm--GAL4/UAS--ΔArm (zygotic) *arm^F1a^ zw3^M11--1^* (maternal)/Y (zygotic); Arm--GAL4/UAS--ΔArm (zygotic) *arm^LM134^* (maternal)/Y (zygotic); Arm--GAL4/UAS--ΔArm (zygotic) *arm^LM134^ zw3^M11--1^* (maternal)/Y (zygotic); Arm--GAL4/UAS--ΔArm (zygotic) *arm^O43A01^* (maternal)/Y (zygotic) *arm^O43A01^* (maternal)/Y (zygotic); Arm^S18^ (zygotic) *arm^F1a^* (maternal)/Y (zygotic); Arm--GAL4/UAS--WIZ--Cby, UAS--ΔArm (zygotic) *arm^F1a^* (maternal)/Y (zygotic); Arm--GAL4/UAS--WIZ--Cby (zygotic) *arm^XK22^* (maternal)/Y (zygotic); Arm^S18^ (maternal) *arm^XK22^* (maternal)/Y (zygotic); MAT--GAL4/UAS--ΔArm (zygotic); Arm^S18^ (maternal) ### UAS transgenes and GAL4 driver lines {#s5b2} Previously published transgenes used were UAS--Arm\*\[S10\], a small deletion of the phosphorylation sites ([@pbio-0020095-Pai1]); the Arm--GAL4 driver ([@pbio-0020095-Sanson1]); Arm^S18^, a mirystoylated, membrane-tethered full-length Arm ([@pbio-0020095-Cox2]b); and UAS--WIZ--Cby for RNAi ([@pbio-0020095-Takemaru2]). ΔArm is a pUAST transgene that deletes the first 128 amino acids, including the GSK3β and CKII phosphorylation sites, the CBP acetylation site, the α-catenin-binding domain, and a transactivation domain ([@pbio-0020095-Zecca1]). Antibodies and Immunofluorescence {#s5c} --------------------------------- Embryos were treated and stained as described previously ([@pbio-0020095-Tolwinski1]), except that they were fixed with heptane/4% formaldehyde in phosphate buffer (0.1 M NaPO~4~ \[pH 7.4\]). The antibodies used were anti-Arm (monoclonal antibody \[mAb\] N2 7A1; Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, Iowa, United States), rabbit anti-Arm ([@pbio-0020095-Peifer4]b), rabbit anti-c-Myc (Santa Cruz Biotechnology, Santa Cruz, California, United States), and anti-Sexlethal (mAb M-14; Developmental Studies Hybridoma Bank). Staining, detection, and image processing were as described previously ([@pbio-0020095-Tolwinski1]). Though not shown, the Sexlethal antibody was used to sex embryos. This allows for the identification of male embryos laid by germline clone mothers, which are hemizygous and therefore maternally and zygotically mutant for X-chromosome genes. Western Blotting {#s5d} ---------------- Embryos were lysed in extract buffer (50 mM Tris \[pH 7.5\], 150 mM NaCl, 1% NP-40, 1 mM EDTA, 10% glycerol; Complete Mini Protease, Sigma, St. Louis, Missouri, United States), and the extracts were separated by 7.5% SDS-PAGE and were blotted as described elsewhere ([@pbio-0020095-Peifer3]a). Maternally and zygotically mutant embryos were hand-selected using standard GFP balancers (<http://flybase.bio.indiana.edu>). Supporting Information {#s6} ====================== **Accession Numbers** The GenBank (<http://www.ncbi.nlm.nih.gov/Genbank/>) accession numbers for the genes and alleles discussed in this paper are *apc1^Q8^* (U77947), *apc2^d40^* (AF091430), *arm* (X54468), *axin^S044230^* (AF086811), *dsh^V26^* (U02491), *wg^IG22^* (NM 164746), and *zw3^M11--1^* (X54005). We are indebted to members of the Wieschaus and Schüpbach labs for helpful discussions and to J. Zallen and A. Nouri for critical reading of the manuscript. We thank R. Carthew, G. Struhl, and M. Peifer for fly stocks. This work was supported by the Howard Hughes Medical Institute and by the National Institutes of Health grant PO1CA41086 to EW. NST was supported by a New Jersey Commission on Cancer Research predoctoral grant. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** NST and EW conceived and designed the experiments, performed the experiments, analyzed the data, contributed reagents/materials/analysis tools, and wrote the paper. Academic Editor: Alfonso Martinez Arias, Cambridge University APC : adenomatous polyposis coli Arm : Armadillo Cby : Chibby Dsh : Disheveled GSK3β : glycogen synthase kinase 3β mAb : monoclonal antibody RNAi : RNA interference Wg : Wingless Zw3 : Zeste-white 3

PubMed Central

2024-06-05T03:55:46.053819

2004-2-10

PMC340127

A couple of years ago, a paper was published in a high-profile journal that challenged a long-established model of cell signaling. While researchers in the field mostly greeted the results with skepticism, some went into the lab to investigate the discrepancy. Many elements of this pathway, called the Wnt pathway, have been well characterized. The standard model of Wnt signaling holds that when the Wnt protein binds to its receptor, it initiates a labyrinthine signaling cascade that sends a protein called β-catenin into the cell\'s nucleus where, together with a protein complex, it initiates transcription. In the absence of this signal, β-catenin binds to an inactivating complex in the cytoplasm and is targeted for degradation. The paper that disputed this view suggested that β-catenin can effect gene expression without entering the nucleus and that it can activate the Wnt pathway while tethered to the cell membrane. Before that paper was published, Nicholas Tolwinski and Eric Wieschaus had shown that β-catenin, also known as Armadillo (Arm) in the fruitfly, is sent into the nucleus in response to Wnt signaling. Upon entering the nucleus, Arm interacts with a second protein complex to activate transcription. Now Tolwinski and Wieschaus have reexamined the function of Arm in the fruitfly and have demonstrated that the pathway "in fact does depend on the nuclear localization of β-catenin." While their paper was in the final stages of acceptance, the dissenting paper was retracted, after it was learned that the results had been fabricated. Tolwinski and Wieschaus\' findings confirm what had already been known about Arm\'s role in Wnt signaling and also fill in important details about how it works. Multicellular organisms rely on elaborate communication networks of signaling proteins and enzymes to exchange information between cells. The Wnt signaling pathway regulates the expression of a host of different genes during embryogenesis to control body patterning and cell differentiation in organisms from fruitflies to mammals. Miscommunications in this tightly regulated pathway contribute to a variety of developmental defects and cancers. In the developing fruitfly, Wnt signaling is normally restricted to the front of each larval segment, where it produces a smooth surface; the rear of the segments, where Wnt signaling is absent, is hairy. If Wnt signaling is overexpressed, it produces fruitfly larvae with only smooth segments; lack of Wnt signaling produces hairy segments. Using the smooth phenotype as a measure of Wnt signaling, Tolwinski and Wieschaus delved deeper into the role of Arm in this signaling process. These experiments are complicated because Arm functions not just in Wnt signaling, but also in cell adhesion. The trick is to make the endogenous Arm (the version encoded by the fly genome) defective for signaling, while leaving the cell adhesion functions fairly normal. Set against this "background," an additional Arm protein is expressed that is tethered to the membrane; it still retains the protein domains required for signaling, but it\'s stuck on the inside of the membrane and can\'t move into the nucleus. In this background, any signaling in response to Wnt must mean that Arm can signal to the nucleus without actually having to get inside. Tolwinski and Wieschaus prove---again---that this is not the case. They do this by showing that the weak, medium, and strong endogenous Arm mutants---these classifications reflect the severity of the mutations\' effects---have different effects on signaling in the presence of the membrane-tethered Arm. It\'s clear, they argue, that tethered Arm cannot signal on its own and must somehow be helping the weaker mutants signal. To further investigate how the tethered Arm activates the endogenous mutants, Tolwinski and Wieschaus developed two new and "cleaner" Arm mutants that impair Arm\'s signaling ability but have no effect on its cell adhesion function. The tethered Arm could not produce a completely smooth phenotype with these nonsignaling endogenous mutants. These experiments, the authors conclude, indicate that the tethered form of Arm produces its transcriptional effects by activating the endogenous Arm protein. Normal activation of the pathway liberates Arm proteins from the inactivating complex, which allows them to enter the nucleus and activate transcription. Tethered Arm appears to accomplish this by sequestering the inactivating complex at the cell membrane, preventing it from interfering with endogenous Arm. Even though Tolwinski and Wieschaus started these experiments based on what turned out to be fabricated results, their investigations produced valuable contributions. They not only reaffirm the standard model of Wnt signaling, but reveal important new insights into the workings of a major player in the pathway.

PubMed Central

2024-06-05T03:55:46.056778

2004-2-10

PMC340937

Introduction {#s1} ============ During primary human immunodeficiency virus 1 (HIV-1) infection, the immune system appears to respond appropriately in order to prevent viral spread, with the mounting of a strong HIV-specific CD8^+^ T-cell response and a corresponding reduction in viraemia ([@pbio-0020020-Koup1]). In common with the majority of persistent viruses, HIV has developed a number of strategies to evade host immunity ([@pbio-0020020-Alcami1]). Continuous adaptive mutation ([@pbio-0020020-Borrow1]) and destruction or impairment of elements necessary for an optimal immune response (e.g., CD4^+^ T-cells and antigen-presenting cells) ([@pbio-0020020-Kalams1]) may explain the failure of antiviral immunity to eradicate the virus. However, unlike most other persistent viruses, HIV-1 progressively destroys the immune system, resulting in acquired immunodeficiency syndrome (AIDS) and death. The precise mechanisms by which immune function is lost remain the subject of considerable controversy. In addition to elevated T-cell turnover and an increase in the proportion of highly differentiated antigen-experienced CD8^+^ and CD4^+^ T-cells during HIV infection ([@pbio-0020020-Wolthers2]; [@pbio-0020020-Appay5]), HIV-infected individuals are characterised by decreased thymic output ([@pbio-0020020-Douek1]) and reduced naïve T-cell numbers ([@pbio-0020020-Roederer1]; [@pbio-0020020-Hellerstein1], [@pbio-0020020-Hellerstein2]), which reflect a diminished capacity to renew the pool of T-cells. Increasing evidence suggests an association between high levels of immune activation and poor outcome in HIV-infected individuals ([@pbio-0020020-Giorgi1]; [@pbio-0020020-Hazenberg1], [@pbio-0020020-Hazenberg3]; [@pbio-0020020-Grossman1]; [@pbio-0020020-Sousa1]), although the underlying mechanism remains unclear. This is supported by studies of sooty mangabeys and African green monkeys, the natural hosts of simian immunodeficiency virus (SIV), which survive SIV infection and are characterised by low immune activation, in striking contrast to rhesus macaques, for which SIV infection is fatal ([@pbio-0020020-Kaur1]; [@pbio-0020020-Broussard1]; [@pbio-0020020-Silvestri1]). To gain further insights into the mechanisms involved, we have studied the potential interplay among immune activation, CD8^+^ T-cell differentiation, and outcome in the context of HIV-1 pathogenesis. We report here that T-cell activation and differentiation are closely related, and that HIV-1 induces immune activation directly and indirectly, which results in differentiation of CD8^+^ T-cells towards replicative senescence. Results {#s2} ======= HIV-Infected Subjects {#s2a} --------------------- Our study involved the analysis of two distinct groups of HIV-1-infected individuals. On one hand, we performed a longitudinal analysis of T-cell subsets during acute HIV-1 infection and its resolution. To examine the effect on T-cells of elevated immune activation associated with an episode of vigourous HIV replication (particularly evident at time of high HIV-1 viraemia, such as the acute infection phase), T-cells were studied in individuals during HIV acute infection and later on---postacute---when viral replication was suppressed following the start of antiretroviral therapy (ART) ([Table 1](#pbio-0020020-t001){ref-type="table"}). These donors were diagnosed at an early stage of primary infection: before or at the time of HIV-1 seroconversion. On the other hand, we carried out a cross-sectional study of HIV-infected untreated individuals at different stages of infection, to draw a correlation between their T-cell characteristics and clinical status. For this purpose, untreated HIV-infected donors were classified into three different groups: acute infection, chronic infection with no sign of progression (infected for more than 10 y with a CD4^+^ count above 500 per milliliter and mean viral load of 10^4^ copies/ml), and chronic infection with signs of disease progression (with decreasing CD4^+^ count, 500 \< *x* \< 130 per milliliter, and mean viral load of 7 × 10^4^ copies/ml). In addition to analysing whole CD8^+^ T-cell populations in these individuals, we have used a panel of tetramers to study the phenotypic evolution of CD8^+^ T-cells specific for HIV, cytomegalovirus (CMV), Epstein--Barr virus (EBV), and influenza. Although this approach focuses on a limited number of viral epitopes (restricted by the number of tetramers available), it remains the only way to avoid stimulation of the cells in order to detect them (e.g., by interferon-γ \[IFN-γ\] secretion), which may alter cellular phenotype and does not enable the detection of all cells. ::: {#pbio-0020020-t001 .table-wrap} Table 1 ::: {.caption} ###### Clinical Characteristics and Percentages of Activated HIV-Nonspecific CD8^+^ T-Cells in Donors Studied during Both Acute and Postacute HIV-1 Infection Stages :::  ^a^ Sampled during the influenza season and low-positive titers for complement fixation antibody assays to both influenza A and influenza B (although these titers did not vary significantly after the first timepoint) ::: Direct and Indirect T-Cell Activation during Acute HIV-1 Infection {#s2b} ------------------------------------------------------------------ CD38 was used as a marker of activation; cells expressing high levels of CD38 ([@pbio-0020020-Appay4]) were considered as being activated. During acute HIV-1 infection, HIV-specific CD8^+^ T-cells were strongly activated, and, intriguingly, activation of the CD8^+^ T-cell compartment as a whole was particularly high, reaching to levels of 80%--90%, in contrast to CD4^+^ T-cells, which show much less activation ([Figure 1](#pbio-0020020-g001){ref-type="fig"}A). In order to shed light on the elevated level of activation experienced by the CD8^+^ T-cell population, we examined which CD8^+^ T-cell subsets were activated and whether all activated cells were HIV-specific. Naïve cells exhibited a slight increase in Ki67 (proliferation marker) expression during acute infection (*p* = 0.03), in keeping with activation-related proliferation of this subset, as previously described ([@pbio-0020020-Hazenberg2]). However, little or no difference in activation levels CD38^+^ between acute and postacute infection stages was observed within the naïve CD8^+^ T-cell subset (CD62L^+^/CD45RA^+^) and antigen-experienced CD45RA^+^ (quiescent \[[@pbio-0020020-Dunne1]; [@pbio-0020020-van3]\]) CD8^+^ T-cells, in contrast to the rest of antigen-experienced CD8^+^ T-cells ([Figure 1](#pbio-0020020-g001){ref-type="fig"}B). This indicates that most activated CD8^+^ T-cells are or have become antigen-experienced. According to the expression of the costimulatory receptors CD28 and CD27, antigen-experienced CD8^+^ T-cells can be positioned along a putative linear model of differentiation or post-thymic development: early (CD28^+^/CD27^+^), intermediate (CD28^−^/CD27^+^), and late (CD28^−^/CD27^−^) differentiated subsets ([@pbio-0020020-Appay3]). While both CD28^+^/CD27^+^ and CD28^−^/CD27^+^ T-cell subsets expressed high levels of CD38 and Ki67 during acute infection, CD28^−^/CD27^−^ T-cells exhibited little activation and proliferation despite increased proportions of these cells following acute infection ([Figure 1](#pbio-0020020-g001){ref-type="fig"}C), suggesting the differentiation into this subset of earlier differentiated cells following activation. ::: {#pbio-0020020-g001 .fig} Figure 1 ::: {.caption} ###### CD8^+^ T-Cell Activation during Acute HIV-1 Infection \(A) Percentages of activated CD38^+^ cells (gated on whole CD8^+^ T-cells, HIV tetramer-positive CD8^+^ T-cells, or whole CD4^+^ T-cells) in donors during acute HIV-1 infection and later postacute on ART (*n* = 12); healthy donors (*n* = 11) and untreated donors with nonprogressing chronic infection (*n* = 12) are also shown. (B and C) CD38 and Ki67 expression on CD8^+^ T-cell subsets defined by CD45RA/CD62L (B) or CD28/CD27 (C) expression, shown in one single donor from acute to postacute (on ART) HIV-1 infection. Percentages of positive cells are shown. Means (± SEM) of CD38^+^ and Ki67^+^ CD8^+^ T-cells for ten patients are also shown; statistics concern CD38 expression. \(D) Staining for the activation marker CD38 on CMV-, EBV-, or influenza A virus-specific CD8^+^ T-cells during acute and postacute (on ART) HIV-1 infection in a single donor. Percentages of CD38^+^ tetramer-positive CD8^+^ T-cells are shown. Data on all donors (see [Table 1](#pbio-0020020-t001){ref-type="table"}) are also shown. \(E) Activation (CD38 and Ki67 staining) of CMV-specific CD8^+^ T-cells or whole CD8^+^ T-cell population during acute and postacute (on ART) HIV-1 infection in a single donor. Percentages of cells present in quadrants are shown. Statistics: \* *p* \< 0.002, \*\* *p* \< 0.01, NS = nonsignificant, with the nonparametric Mann--Whitney test. :::  ::: Surprisingly, from the analysis of CD8^+^ T-cells specific for non-HIV viral antigens in donors with suitable human leukocyte antigen (HLA) type (HLA-A\*0201 for CMV, EBV, and influenza A virus; HLA-B\*0701 for CMV; and HLA-B\*0801 for EBV), both CMV- and EBV-specific CD8^+^ T-cells displayed significant levels of activation exclusively during acute HIV infection, compared to chronic infection (*p* \< 0.002) ([Figure 1](#pbio-0020020-g001){ref-type="fig"}D; see [Table 1](#pbio-0020020-t001){ref-type="table"}). Activated cells specific for non-HIV viral antigens also participated in the expansion of the CD8^+^ T-cell population observed in HIV primary infection, as shown by expression of the proliferation marker Ki67 ([Figure 1](#pbio-0020020-g001){ref-type="fig"}E). Plasma DNA levels of CMV and EBV in these study subjects were below detection limits of the assays and thus did not provide evidence of high levels (greater than 400 genomes per milliliter) of systemic reactivation (data not shown). However, the observation of nonactivated influenza A virus-specific CD8^+^ T-cells ([Figure 1](#pbio-0020020-g001){ref-type="fig"}D), in contrast to CMV- or EBV-specific CD8^+^ T-cells (*p* \< 0.01), strongly suggests that the stimulation of these cells associated with HIV-1 infection is due to reactivation of pathogens such as CMV and EBV, rather than as a result of bystander activation. Overall, these data show that HIV-1 infection leads to activation of antigen-experienced CD8^+^ T-cells at early stages of differentiation, both in direct (HIV-specific) and indirect (HIV-nonspecific) manners. Activation-Induced T-Cell Differentiation {#s2c} ----------------------------------------- The potential relationship between T-cell activation and differentiation was first studied using a system of in vitro priming of naïve CD8^+^ T-cells by dendritic cells (DCs), which represents a useful model to analyse the generation of antigen-experienced CD8^+^ T-cells. This system is based on the existence in normal human donors of a significant number of naive CD8^+^ T-cells (reactive for the HLA-A2-restricted melan-A antigen \[[@pbio-0020020-Dutoit1]; [@pbio-0020020-Zippelius1]\]), which can be primed by autologous matured DCs loaded with specific peptides to become antigen-experienced cells ([@pbio-0020020-Salio1]). Although we cannot with certainty extend our interpretation of data from this assay system beyond the in vitro conditions (i.e., signals involved in T-cell differentiation, apoptosis, or both, as well as homeostatic signals, may be absent or differ from the in vivo situation), this system represents a unique opportunity to study the priming of naïve CD8^+^ T-cells using human material. We used a range of concentrations of the melan-A antigen loaded onto professional antigen-presenting cells to generate different levels of stimulation. Mature DCs do not persist very long in culture (2--3 d); moreover, the half-life of class I MHC--peptide complexes on mature DCs is rather short ([@pbio-0020020-Cella1]); therefore, the results reflect increasing antigen doses from a single round of antigen exposure. We observed a close relationship between the level of stimulation induced and the size of the resulting antigen-specific CD8^+^ T-cell population ([Figure 2](#pbio-0020020-g002){ref-type="fig"}A). This relationship was steady, as maintained over time, following priming of naïve cells and following a second round of stimulation of the antigen-experienced cells with antigen-loaded matured DCs ([Figure 2](#pbio-0020020-g002){ref-type="fig"}B). The priming of naïve cells (granzyme A-negative) was successfully initiated at all antigen concentrations, as shown by the expression of the cytotoxic factor granzyme A in all melan-A-specific CD8^+^ T cells ([Figure 2](#pbio-0020020-g002){ref-type="fig"}C). Increasing concentrations of antigen were associated with increasing activation levels and proliferation, indicated by increased expression of Ki67 and declining expression of CD62L ([Figure 2](#pbio-0020020-g002){ref-type="fig"}C). The analysis of the differentiation phenotype (based on CD28 and CD27 expression) throughout the priming of the cells provided in vitro confirmation of the hypothetical model of CD8^+^ T-cell differentiation observed ex vivo ([@pbio-0020020-Hamann2]; [@pbio-0020020-Appay3]): starting from a population with naïve characteristics (CD28^+^/CD27^+^/CD62L^+^/CD45RA^+^/granzyme A^−^) at day 0 (data not shown), antigen-primed cells lost sequentially expression of CD28 and CD27 ([Figure 2](#pbio-0020020-g002){ref-type="fig"}D). Following priming, the differentiation phenotype of the melan-A-specific CD8^+^ T-cells varied according to the level of stimulation induced, with high antigen load resulting in further differentiation of the cells ([Figure 2](#pbio-0020020-g002){ref-type="fig"}E). These data show that there is a close correlation among the level of activation, size, and differentiation of the antigen-specific CD8^+^ T-cells. ::: {#pbio-0020020-g002 .fig} Figure 2 ::: {.caption} ###### In Vitro Priming of Antigen-Specific CD8^+^ T-Cells \(A) Representative stainings for melan-A-specific CD8^+^ T-cells following priming of naïve cells from healthy donor PBMCs by autologous mature DCs loaded with various concentrations of antigen. Cells are gated on lymphocytes 47 d after priming. Percentages of melan-A tetramer-positive CD8^+^ T-cells are shown. \(B) Percentages of melan-A-specific CD8^+^ T-cells over time following priming at day 0 with mature DCs loaded with various concentrations of antigen, with no restimulation or with restimulation using mature DCs at day 25. The legend indicates the concentration of melan-A--peptide used in microgram per milliliter; populations generated with 0 or 10^−3^ μg/ml of antigen are plotted on the right-hand side Y axis. \(C) Percentages of melan-A tetramer-positive CD8^+^ T-cells expressing granzyme A, Ki67, CD62L, or CD57 according to antigen concentration used, at day 30 following priming. Ki67 and CD57 expressions are plotted on the right-hand side Y axis. \(D) CD28 and CD27 expression on melan-A tetramer-positive CD8^+^ T-cells in PBMC (day 0), and over time following priming with 1 μg/ml of antigen. Percentages of cells present in quadrants are shown. The model of CD8^+^ T-cell differentiation based on CD28 and CD27 expression is illustrated (top left panel). \(E) Distribution of the melan-A-specific CD8^+^ T-cells into the distinct differentiated subsets according to antigen concentration used, at day 47 following priming. Similar observations were made whether the cells were subjected to a second round of stimulation or not. Data are representative of three independent experiments. :::  ::: This relationship was confirmed by ex vivo analysis of antigen-experienced CD8^+^ T-cells. Despite that the majority of HIV-specific CD8^+^ T-cells are usually found at an intermediate stage of differentiation ([@pbio-0020020-Appay3]), certain of these populations exhibit a significant percentage of late-differentiated CD8^+^ T-cells, as exemplified by the analysis of three HIV-1-specific CD8^+^ T-cell populations in a single individual ([Figure 3](#pbio-0020020-g003){ref-type="fig"}A). The examination of the differentiation state (percentage of CD27^−^ in the tetramer-positive cells) and the size (percentage of tetramer-positive cells in the whole CD8 population) of a variety of HIV-specific CD8^+^ T-cell populations in several donors revealed a correlation between these two parameters ([Figure 3](#pbio-0020020-g003){ref-type="fig"}B). A similar correlation was also found in the case of CMV-specific populations (although these cells are usually more differentiated, as previously described \[[@pbio-0020020-Appay3]\]), as well as in EBV- and influenza-specific CD8^+^ T-cells. The correlation between differentiation and population size becomes highly significant when data on all specificities are combined. Following acute HIV infection and related strong activation, HIV-specific CD8^+^ T-cells displayed increased percentages of CD28^−^/CD27^−^ cells (especially with larger populations) ([Figure 3](#pbio-0020020-g003){ref-type="fig"}C; [Figure 4](#pbio-0020020-g004){ref-type="fig"}B). The differentiation phenotype of non-HIV-specific CD8^+^ T-cells could also vary from acute to postacute HIV infection stages in relation to activation: while the differentiation phenotype of influenza A virus-specific cells remained unchanged, CMV- and (although less frequently) EBV-specific CD8^+^ T-cells became further differentiated ([Figure 3](#pbio-0020020-g003){ref-type="fig"}D; [Figure 4](#pbio-0020020-g004){ref-type="fig"}B). This is in keeping with a recent report, which shows increased differentiation of EBV-specific CD8^+^ T-cells during HIV-1 infection ([@pbio-0020020-van1]). Taken together, these data indicate that the immune activation induced in the context of HIV-1 infection can result in the differentiation of T-cells specific for HIV-1 as well as other pathogens such as CMV and EBV, which may explain the increase in the proportions of highly differentiated cells observed during HIV-1 infection. ::: {#pbio-0020020-g003 .fig} Figure 3 ::: {.caption} ###### Activation and Differentiation of Antigen-Specific CD8^+^ T-Cells during HIV-1 Infection \(A) Representative staining for the differentiation marker CD27 on three HIV-specific (HLA-B8 nef, HLA-A2 p17, and HLA-B8 p24) populations in a single HIV-1-infected donor. Numbers show percentages of tetramer-positive CD8^+^ T-cells (outside the quadrants) and percentages of CD27^−^ tetramer-positive cells (inside the quadrants). \(B) Correlation between size (percentage of tetramer-positive CD8^+^ T-cells) and differentiation (percentages of CD27^−^ tetramer-positive cells) of CD8^+^ T-cells specific for HIV antigens (including HLA-A2 p17, pol, HLA-B7 nef, gp41, HLA-B8 nef, p24, and HLA-B57 p24) (open circles), CMV antigens (including HLA-A2, B7, and B35 pp65) (filled circles), EBV (HLA-A2 BMLF1, HLA-B8 BZLF1, EBNA3A) (filled squares), and influenza (HLA-A2 matrix) (open squares) antigens or all antigens together. These populations were studied in individuals with chronic infection for HIV, CMV, or EBV (independently from clinical status). *P* values were obtained using the nonparametric Spearman rank correlation test. \(C) CD28 and CD27 expression on whole, HIV nef-, or p24-specific CD8^+^ T-cells during acute and postacute (on ART) HIV-1 infection in a single donor. \(D) CD28 and CD27 expression on CMV-, EBV-, or influenza-specific CD8^+^ T-cells during acute and postacute (on ART) HIV-1 infection in a single donor. Percentages of cells present in quadrants are shown. :::  ::: ::: {#pbio-0020020-g004 .fig} Figure 4 ::: {.caption} ###### CD8^+^ T-Cell Differentiation and HIV-1 Disease Progression \(A) Distribution of the CD8^+^ T-cell population in differentiated subsets (CD28^+^/CD27^+^ early, CD28^−^/CD27^+^ intermediate, and CD28^−^/CD27^−^ late) through the course of HIV-1 infection. Abbreviations: H, healthy (*n* = 15); A, acute HIV infection (*n* = 11); C, chronic HIV infection nonprogressor (no ART; *n* = 14); P, chronic HIV infection with signs of disease progression (no ART; *n* = 10). Statistics: \* *p* \< 0.0001 with the ANOVA test and *p* \< 0.005 between each group. \(B) Percentages of CD27^−^ CD8^+^ T-cells that are specific for HLA-B8 HIV (nef) or HLA-A2 CMV in HIV-1-infected individuals at different stages of infection. Statistics: \*\* *p* \< 0.005 with the nonparametric Mann--Whitney test. \(C) Inverse correlation between CD4^+^ T-cell counts and percentage of highly differentiated CD27^−^ cells in the whole CD8^+^ T-cell population of HIV-1-infected donors during chronic infection (untreated nonprogressors and progressors). The *p* value was obtained using the nonparametric Spearman rank correlation test. :::  ::: Increased T-Cell Differentiation with Progression to AIDS {#s2d} --------------------------------------------------------- Persistent and continuous replication is a hallmark of HIV-1 infection, along with chronic activation and constant turnover of T-cells, and these factors are now thought of as playing a critical role in HIV pathogenesis and disease progression ([@pbio-0020020-Giorgi1]; [@pbio-0020020-Hazenberg1]; [@pbio-0020020-Grossman1]; [@pbio-0020020-Hellerstein2]). The detailed distribution of the CD8^+^ T-cell population along the pathway of differentiation during HIV-1 infection was analysed in a cross-sectional study of individuals at different stages of infection. It revealed an increase in the proportion of highly differentiated CD8^+^ T-cells associated with HIV disease progression ([Figure 4](#pbio-0020020-g004){ref-type="fig"}A). Increased proportions of CD28^−^/CD27^+^ CD8^+^ T-cells during acute HIV-1 infection are likely to reflect expansion of HIV-specific CD8^+^ T-cells. The enrichment in highly differentiated CD8^+^ T-cells from acute infection onwards included virus-specific cells, as exemplified by the analysis of populations specific for one HIV epitope or one CMV epitope ([Figure 4](#pbio-0020020-g004){ref-type="fig"}B). The study of individuals during chronic infection (including nonprogressors and donors with evidence of disease progression, both untreated) revealed an inverse correlation between the overall percentage of highly differentiated cells and CD4^+^ T-cell count, as an indicator of disease progression ([Figure 4](#pbio-0020020-g004){ref-type="fig"}C). No significant correlation emerged between the differentiation state of virus-specific CD8^+^ T-cell populations and CD4^+^ T-cell count; a larger number of virus-specific CD8^+^ T-cell populations studied may be required. A problem with the interpretation of increased numbers of highly differentiated T-cells relates to the controversy around the significance of these cells. Some investigators regard these cells as the effector-type population, conferring optimum protective immunity ([@pbio-0020020-van2]; [@pbio-0020020-Zhang1]), but for others, these cells have lost their capacity to proliferate and their incidence may reflect ageing of the lymphocyte population ([@pbio-0020020-Effros2]; [@pbio-0020020-Globerson1]; [@pbio-0020020-Appay2]). Replicative Senescence and Increased T-Cell Differentiation {#s2e} ----------------------------------------------------------- As CD8^+^ T-cells differentiate further, they express increasing levels of CD57 ([Figure 5](#pbio-0020020-g005){ref-type="fig"}A), a marker that has recently been associated with a state of replicative senescence ([@pbio-0020020-Brenchley1]). This is in line with the observation of increased CD57 expression on CD8^+^ T cells following acute HIV infection, including cells specific for HIV, as well as other specificities, such as CMV- and EBV-specific cells ([Figure 5](#pbio-0020020-g005){ref-type="fig"}B). Increased CD57 expression in association with further T-cell differentiation was also seen following priming of T-cells in vitro (see [Figure 2](#pbio-0020020-g002){ref-type="fig"}C), although this remained relatively modest (below 10%), possibly due to the high susceptibility to activation induced cells death of CD57^+^ T-cells ([@pbio-0020020-Brenchley1]; unpublished data) in the interleukin-2 (IL-2)-supplemented assay conditions. In keeping with the finding by [@pbio-0020020-Brenchley1]), we observed that highly differentiated CD27^−^/CD57^+^ CD8^+^ T-cells exhibited a reduced capacity to proliferate despite being activated following stimulation with anti-CD3 antibodies (−/+ addition of IL-2) ([Figure 5](#pbio-0020020-g005){ref-type="fig"}C). In addition, we measured telomere length in CD8^+^ T-cell subsets at different stages of differentiation. The telomere length reflects the mitotic history of cells: in lymphocytes, every cell division shortens the telomeres by approximately 30--60 basespairs ([@pbio-0020020-Rufer1]), until the cells lose their capacity to proliferate any longer. The induction of human telomerase expression (necessary for the maintenance of telomere length) has recently been shown to decrease in T-cells that have expanded in vivo upon antigen encounter ([@pbio-0020020-Roth1]). Shortening of the telomeres appears to occur progressively along T-cell differentiation ([Figure 5](#pbio-0020020-g005){ref-type="fig"}D) so that highly differentiated CD27^−^/CD57^+^ cells display the shortest telomeres, with lengths (4--5 kb) equivalent to those observed in antigen-experienced CD8^+^ T-cells from the elderly ([@pbio-0020020-Rufer2]). All together, these data support the view that T-cells exhibit increasing characteristics of replicative senescence as they differentiate further. The assumption that CD28^−^/CD27^−^ T-cells are protective effector cells is mainly based on the fact that these cells possess strong cytotoxic potential, expressing high levels of perforin, as seen ex vivo ([@pbio-0020020-Hamann1]). However, a recent report suggests that ex vivo Cr51 release assay, and therefore perforin levels, may not be a true reflection of in vivo cytotoxic capacities and, accordingly, this could be misleading in the interpretation of what constitutes a protective "effector cell" ([@pbio-0020020-Barber1]). ::: {#pbio-0020020-g005 .fig} Figure 5 ::: {.caption} ###### CD8^+^ T-Cell Differentiation and Senescence \(A) Expression of the replicative senescence-associated marker CD57 on antigen-experienced CD8^+^ T-cell subsets. The percentage and mean fluorescence intensity for the CD57^+^ cells are shown for one single donor. Data on several donors (HIV-1-infected or healthy) are also shown (*n* = 24). \(B) Expression of CD57 on CD8^+^ T-cells (whole population or antigen-specific) from acute to postacute (on ART) HIV-1 infection. \(C) CD69 expression and CFSE proliferation profile for CD8^+^ T-cell subsets gated on the basis of CD57 and CD27 expression following stimulation with anti-CD3 antibodies. PBMCs were analysed for CD69 expression after 18 h and CFSE labeling after 6 d. Percentages of proliferating cells (with background subtracted) are indicated. Representative results from three experiments (one HIV-infected and two healthy donors) are shown. \(D) Telomere length measurement by flow FISH on naïve and antigen-experienced CD8^+^ T-cell subsets FACS-sorted on the basis of CD57, CD27, CCR7, and CD45RA expression. The average length of telomeres was obtained by substracting the mean fluorescence of the background control (no probe; open histogram) from the mean fluorescence obtained from cells hybridised with the FITC-labeled telomere probe (gray histogram). Representative results from two experiments (on healthy donors) are shown. \(E) CD57 and perforin expression in the CD8^+^ T-cell population dissected into naïve (CD27^+high^, perforin-negative), antigen-experienced CD27^+^ (perforin^low^), and antigen-experienced CD27^−^ perforin^low^ or perforin^high^ subsets. The percentage and mean fluorescence intensity for the CD57^+^ cells are indicated. \(F) Representative staining for perforin and CD57 in CD8^+^ T-cells from a HIV-1-infected or a healthy donor. Percentages of cells present in the top quadrants are shown. \(G) Representative staining for perforin and CD57 in CD4^+^ T-cells from an HIV-1-infected or a healthy donor. Percentages of cells present in the top quadrants are shown. :::  ::: It was previously reported that antigen-specific CD27^−^ CD8^+^ T-cells do proliferate ([@pbio-0020020-van3]). We show here that only a proportion of highly differentiated CD27^−^ CD8^+^ T-cells express CD57, therefore exhibiting reduced proliferative capacities, while the rest of the CD27^−^ CD8^+^ T-cells should indeed be able to expand. Nonetheless, the vast majority of highly differentiated cells with high levels of perforin are CD57^+^ ([Figure 5](#pbio-0020020-g005){ref-type="fig"}E). The association between high levels of perforin and characteristics of replicative senescence is not a particular characteristic of HIV infection, but holds true in both HIV-infected and HIV-noninfected individuals ([Figure 5](#pbio-0020020-g005){ref-type="fig"}F). Increase in the intracellular perforin content seems to be the normal consequence of the process of post-thymic development, and it is also valid in the case of CD4^+^ T-cell differentiation, since cytotoxic CD4^+^ T-cells, whose proportions are increased during HIV-1 infection ([@pbio-0020020-Appay5]), are CD57^+^ ([Figure 5](#pbio-0020020-g005){ref-type="fig"}G). Overall, as HIV-1-infected individuals are progressing, they display increasing proportions of late-differentiated T-cells with characteristics of replicative senescence, with an average of 40% of CD57^+^ CD8^+^ T-cells in progressor/AIDS individuals (data not shown). Overall, the accumulation of highly differentiated CD8^+^ T-cells in HIV infection goes along with reports of reduced proliferative capacities and shorter telomere length characterising the T-cells of the HIV-infected individual ([@pbio-0020020-Wolthers1]; [@pbio-0020020-Bestilny1]; [@pbio-0020020-Effros1]). Discussion {#s3} ========== Here we have studied the interplay between CD8^+^ T-cell activation and differentiation and its implications for HIV pathogenesis. HIV-1 induces a strong immune activation, which is particularly evident within the CD8^+^ T-cell compartment. Our data indicate that HIV-1 infection results in immune activation not only directly, but also indirectly, with the activation of cells specific for non-HIV antigens. In recent years, the role of potential bystander activation has been reevaluated and is now considered less important ([@pbio-0020020-Murali-Krishna1]), suggesting that most of the stimulation observed may be antigen-driven. During acute HIV-1 infection, immunosuppression may develop that favours the replication of host flora like CMV and EBV, as occurs in other immunocompromised individuals ([@pbio-0020020-Yao1]; [@pbio-0020020-Gerna1]). Recently, the help provided by CD4^+^ T-cells to control viral replication has been emphasised in the context of CMV infection ([@pbio-0020020-Gamadia1]). The drop in the CD4^+^ T-cell counts during HIV acute infection may result in suboptimal immune control of CMV and EBV and thus permits the replication of these viruses. Data have indicated that frequent reactivation of CMV likely occurs in the human host, as evidenced by the presence of a large population of CD69^+^ CMV-specific cells, indicative of recent in vivo activation ([@pbio-0020020-Dunn1]). Hence, HIV infection may serve to increase both the frequency and magnitude of CMV reactivation. In addition, inflammatory conditions occurring during HIV acute infection (e.g., release of proinflammatory cytokines) may participate in the reactivation of latent forms of CMV and EBV. We have shown here that T-cell activation and increasing differentiation are closely related. One could speculate that the association between different stages of CD8^+^ T-cell differentiation and viral specificity of these cells, as previously described ([@pbio-0020020-Appay3]; [@pbio-0020020-Tussey1]), may be related to the stimulation intensity received by the cells from priming onwards. CMV may therefore be a particularly potent stimulus for CD8^+^ T-cells, thus promoting a strong differentiation of these cells. Interestingly, a similar phenomenon seems to happen in the context of CD4^+^ T-cells, as CMV-specific CD4^+^ T-cells show further differentiation, in comparison with EBV-specific CD4^+^ T-cells ([@pbio-0020020-Amyes1]). In the context of HIV infection, elevated and chronic immune activation is the most plausible cause for the general shift of the CD8^+^ T-cell population towards the highly differentiated cells that accompanies progression towards AIDS, as we have shown that elevated cellular activation drives further differentiation of CD8^+^ T-cells (including HIV-, CMV-, or EBV-specific cells). Converging evidence suggests that a reduction of replicative potential occurs with extensive T-cell division and differentiation. Differentiation towards late stages (CD28^−^/CD27^−^/CD57^+^) is strongly associated with the display of characteristics of replicative senescence, which may have an impact on viral control. The relevance of perforin^high^ late-differentiated T-cells in conferring protective immunity is controversial. For instance, [@pbio-0020020-van1]) reported a correlation between high numbers of late-differentiated HIV-specific CD8^+^ T-cells and years of AIDS-free survival. However, it remains to be determined whether late-differentiated CD8^+^ T-cells would simply accumulate in these individuals with chronic infection over time, whilst playing no role in delaying disease progression. Overall, there is confusion regarding the ideal functional and phenotypic profile of a "protective effector cell." Protective immunity has recently been associated with the proliferative capacity of virus-specific CD8 T-cells in the mouse model ([@pbio-0020020-Wherry1]). This is supported by [@pbio-0020020-Migueles1]), who showed that HIV-1-infected long-term nonprogressors are characterised by HIV-1-specific CD8^+^ T-cells that maintain a strong proliferative capacity following in vitro stimulation (cells defined mainly as CD45RO^+^/CD28^+^/CD27^+^ early-differentiated cells). In this study, the proliferative potential of these cells was coupled to strong perforin expression, suggesting that early-differentiated cells (which express low perforin levels in a resting state \[[@pbio-0020020-Appay3]\]) are able to express high perforin levels after certain conditions of stimulation. In contrast, the high perforin levels observed in resting late-differentiated T-cells seem to correlate with characteristics of replicative senescence. These findings challenge the view that highly differentiated T-cells are beneficial effector cells that should be the goal of vaccine or immunotherapeutic strategies ([@pbio-0020020-Speiser1]). In keeping with this position, the fraction of perforin^high^ HIV-specific CD8^+^ T-cells has been proposed to be a marker for disease progression ([@pbio-0020020-Heintel1]). One may speculate that this high perforin expression may reflect an alteration of gene expression related to replicative senescence. This may not be dissimilar to the changes in gene expression that occur during replicative senescence in fibroblasts ([@pbio-0020020-Smith1]). More investigations on this matter will be necessary to clarify the cause and consequence of high perforin levels in late-differentiated T-cells. The elevated and chronic stimulation induced by HIV-1 may result in the exhaustion of the capacity to generate new T-cells ([@pbio-0020020-Hazenberg3]), while the pool of antigen-experienced cells is driven to differentiate into aged oligoclonal populations. Interestingly, these characteristics are not unique to HIV infection, but they are also common to other conditions that result in some degree of immunodeficiency, like ataxia telangiectasia ([@pbio-0020020-Giovannetti1]), and normal human ageing ([@pbio-0020020-Nociari1]; [@pbio-0020020-Rufer2]). They may reflect a premature decline of the immune resources necessary for viral control and therefore contribute to the onset of disease progression ([@pbio-0020020-Effros1]; [@pbio-0020020-Hazenberg1]; [@pbio-0020020-Appay2]; [@pbio-0020020-Grossman1]). This hypothesis is also strongly supported by a recent study performed in a mouse model in which persistent immune activation was shown to exhaust the T-cell pool and be sufficient to induce lethal immunodeficiency ([@pbio-0020020-Tesselaar1]). In addition to a direct effect of HIV on the thymus, decreased thymic output and T-cell renewal may originate from thymus involution ([@pbio-0020020-Kalayjian1]) as well as the failure of the bone marrow and the reduction of primitive hemaopoietic stem cell subsets ([@pbio-0020020-Marandin1]; [@pbio-0020020-Moses1]), as observed in HIV-1-infected individuals. Increased proportions of highly differentiated T-cells may relate to the maintenance of homeostasis and "immunological space" in the absence of T-cell renewal. Our study also emphasises the importance of considering the influence of HIV-1 infection on other pathogens as well as the influence of these pathogens on HIV pathogenesis. For instance, CMV is known to drive substantial differentiation of T-cells towards CD57^+^ cells ([@pbio-0020020-Wang1]). CMV may therefore play an important role in the decline of the immune resources, as recently proposed in the HIV-noninfected elderly ([@pbio-0020020-Khan1]; [@pbio-0020020-Wikby1]). CMV infection was recently associated with a higher rate of disease progression in HIV-1-infected infants ([@pbio-0020020-Kovacs1]) and with reduced survival in patients with advanced HIV disease ([@pbio-0020020-Erice1]); it has also been shown to be a cofactor for HIV disease progression and death in some longitudinal studies of HIV-infected haemophiliacs ([@pbio-0020020-Webster1]). The impact of elevated activation and differentiation on immune function appears to have considerable importance in the onset of immunodeficiency and needs to be addressed in the development of current and future anti-HIV strategies. Materials and Methods {#s4} ===================== {#s4a} ### Study subjects. {#s4a1} Samples were taken from HIV-1-infected patients attending clinics in London or Oxford (United Kingdom) and San Diego (United States) who were known to have either acute or chronic HIV-1 infection. The relevant local Institutional Review Boards and Ethics Committees approved the study. Subject ages ranged from 23 to 65 y old. Eleven patients with HIV-1 acute infection were selected from a well-characterised cohort in San Diego on the basis of their having an HLA type (HLA-A\*0201, HLA-B\*0701, or HLA-B\*0801) for which we could detect virus-specific CD8^+^ T-cell populations using tetramers. The donors were diagnosed before or at the time of HIV-1 seroconversion, defined by symptomatic disease, recent high-risk exposure, high-plasma *HIV-1* RNA (ranging from 3 × 10^5^ to 3 × 10^6^ copies/ml \[mean, 8.3 × 10^5^ copies/ml\]), and either a negative HIV-1 ELISA or a negative/indeterminate HIV-1 Western blot. A second sample was analysed at a later timepoint after the start of successful ART (see [Table 1](#pbio-0020020-t001){ref-type="table"}). The study also involved untreated HIV chronically infected individuals: either with indications of viral control (*n* = 14, drug naïve, infected for more than 10 y with a CD4^+^ count above 500 per milliliter and viral load ranging from undetectable to 2 × 10^4^ copies/ml) or with evidence of progressive HIV disease (*n* = 10, with decreasing CD4^+^ count, 500 \< *x* \< 130 per milliliter, and viral load ranging from 5 × 10^3^ to 3 × 10^5^ copies/ml). Blood samples were also obtained from healthy adult volunteers. Peripheral blood mononuclear cells (PBMCs) were separated from heparinised blood and cryopreserved for subsequent studies. HLA typing was carried out by amplification refractory mutation system--polymerase chain reaction (ARMS--PCR) using sequence-specific primers as previously described ([@pbio-0020020-Bunce1]). HLA-typed patients were generally screened first for virus-specific CD8^+^ T-cell responses by means of Elispot assays using known HLA class I-restricted viral epitope peptides. ### Reagents and flow cytometry. {#s4a2} HLA--peptide tetrameric complexes ("tetramers") were produced as previously described ([@pbio-0020020-Altman1]) and included the following specificity: A2 HIV p17-SLYNTVATL and pol-ILKEPVHGV, A2 CMV pp65-NLVPMVATV, A2 EBV BMLF1-GLCTLVAML, A2 influenza matrix-GILGFVFTL, A2 melan-A-ELAGIGILTV, B7 HIV nef-TPGPGVRYPL and gp41-IPRRIRQGL, B7 CMV pp65-TPRVTGGGAM, B8 HIV nef-FLKEKGGL and p24-DIYKRWII, B8 EBV BZLF1-RAKFKQLL, B35 CMV pp65-VFPTKDVAL and B57 HIV p24-KAFSPEVIPMF. Anti-CD8--PerCP (peridinin chlorophyll protein) or APC CY7 (allophycocyanin cyanine 7), anti-CD27--PE (phycoerythrin) or APC, anti-CD28--FITC (fluorescein isothiocyanate), anti-CD38--APC, anti-CD45RA--FITC or ECD (PE--Texas red), anti-CD62L--APC, anti-Ki67--FITC, anti-CD69--FITC, anti-CCR7-purified, anti-granzyme A--FITC, and anti-perforin--PE antibodies were purchased from Becton-Dickinson PharMingen (San Diego, California, United States); anti-CD57--FITC or PE antibodies were from Beckman Coulter (San Diego, California, United States). FACS stainings were performed as previously described ([@pbio-0020020-Appay1]). In brief, titrated tetramers (PE-conjugated) were added to 150 μl of heparinised blood or PBMCs, followed by addition of a panel of titrated antibodies (FITC-, PerCP-, or APC-conjugated). The lymphocytes were then fixed and the red blood cells lysed using FACS^TM^ lysis solution (Becton-Dickinson). Cells were washed, fixed, and permeabilised in FACS^TM^ permeabilisation buffer (Becton-Dickinson). After washing, intracellular perforin staining was performed using titrated antibodies. Cells were then washed and stored in Cell Fix^TM^ buffer (Becton-Dickinson) at 4°C until analysis. Samples were analysed on a Becton-Dickinson FACSCalibur after compensation was checked using freshly stained PBMCs. Carboxyfluorescein diacetate succinimidyl ester (CFSE) labeling was performed by incubating PBMCs with 5 μM CFSE (Molecular Probes, Leiden, The Netherlands) in RPMI 1640 for 10 min at 37°C, before quenching with ice-cold RPMI 1640--10% foetal calf serum (FCS) and washing. The cells were then incubated with immobilised OKT3 (10 μg/ml) for 6 d (with or without 20U/ml of IL-2) before staining. ### Flow fluorescence in situ hybridisation. {#s4a3} Naïve and antigen-experienced CD8^+^ T-cell subsets were sorted ex vivo from freshly isolated PBMCs, on the basis of CD27, CD57, CCR7, and CD45RA expression using a five-color FACS vantage SE (with 98%-- 99% purity). For each subset, 0.5 × 10^5^ to 2 × 10^5^ cells were used to measure the average length of telomere repeats at chromosome ends in individual cells by quantitative flow fluorescence in situ hybridisation (FISH), as previously described ([@pbio-0020020-Rufer1], 1999). FITC-labeled fluorescent calibration beads (Quantum TM-24 Premixed; Bangs Laboratories Inc., Fishers, Indiana, United States) were used to convert telomere fluorescence data to molecules of equivalent soluble fluorescence (MESF) units. The following equation was performed to estimate the telomere length in basepairs from telomere fluorescence in MESF units: basepair = MESF × 0.495 ([@pbio-0020020-Rufer1]). ### In vitro priming of CD8^+^ T-cells with DCs. {#s4a4} DCs were generated as previously described ([@pbio-0020020-Salio1]). Monocytes were purified from healthy donors\' PBMCs (screened for HLA-A2 expression) by positive sorting using anti-CD14-conjugated magnetic microbeads (Miltenyi Biotec, Bergisch-Gladbach, Germany). The recovered cells were greater than 99% CD14^+^. DCs were generated by culturing monocytes in RPMI 1640--10% FCS supplemented with 50 ng/ml GM--CSF (Leucomax, Basel, Switzerland) and 500 U/ml IL-4 (Peprotech, London, United Kingdom) for 5 d. Cells (3 × 10^5^/ml) were stimulated by addition of 1 μg/ml LPS (Sigma, St. Louis, Missouri, United States). Antigen-presenting cells were pulsed for 3 h with various concentrations of melan-A--peptide in serum-free medium before incubation with autologous PBMCs at a 1:5 ratio in RPMI 1640--10% FCS. Human rIL-2 (R&D Systems, Minneapolis, Minnesota, United States) was added from day 4 at 10 U/ml, then at 500 U/ml IL-2 when cells expanded. Melan-A-specific CD8^+^ T-cells were then analysed by flow cytometry over time for up to 50 d. ### Statistics. {#s4a5} Group medians and distributions were compared by the nonparametric Mann--Whitney test. Associations between variables were determined by the nonparametric Spearman rank correlation test. Associations between variables in different patient groups were determined by simple linear regression or ANOVA test. *P* values above 0.05 were considered not significant. We are very grateful to Linda Terry for technical assistance and to the staff and patients of the clinics that provided blood samples, particularly the Caldecot Centre at King\'s College Hospital, London; the clinic of Infectious and Tropical Diseases, L. Sacco Hospital, Milano; and the Veterans Administration San Diego Research Center for AIDS and HIV Infection, the National Institutes of Health (NIH) Acute and Early Infectious Disease Research Program, the University of California, San Diego, Center for AIDS Research (NIH drug-resistance grant AI 29164). This work was supported by the Medical Research Council of the United Kingdom, the Wellcome Trust, the European Union (QLK2-CT-1999--00356), the Elizabeth Glaser Paediatric AIDS Foundation, the Cancer Research UK, and the NIH. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** LP, CAS, MS, NR, PRD, AJM, SLR-J, and VA conceived and designed the experiments. LP, AM, NR, GC, VE, and VA performed the experiments. LP, NR, and VA analysed the data. CAS, AM, MS, NR, SL, TD, AW, PE, DS, VC, PG, CC, and DDR contributed reagents/materials/analysis tools. SLR-J and VA wrote the paper. DOI: [10.1371/journal.pbio.0020020](10.1371/journal.pbio.0020020) Copyright: © 2004 Papagno et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Academic Editor: Philippa Marrack, National Jewish Medical and Research Center \* To whom correspondence should be addressed. E-mail: <vappay@gwmail.jr2.ox.ac.uk> AIDS : acquired immunodeficiency syndrome APC : allophycocyanin ARMS--PCR : amplification refractory mutation system--polymerase chain reaction ART : antiretroviral therapy CFSE : carboxyfluorescein diacetate succinimidyl ester CMV : cytomegalovirus CY7 : cyanine 7 DC : dendritic cell EBV : Epstein--Barr virus FCS : foetal calf serum FISH : fluorescence in situ hybridisation FITC : fluorescein isothiocyanate HIV : human immunodeficiency virus HLA : human leukocyte antigen IFN : interferon IL : interleukin MESF : molecules of equivalent soluble fluorescence PBMC : peripheral blood mononuclear cell PE : phycoerythrin PerCP : peridinin chlorophyll protein SIV : simian immunodeficiency virus

PubMed Central

2024-06-05T03:55:46.057784

2004-2-17

PMC340938

Introduction {#s1} ============ Aberrant retinoblastoma 1 (pRb) pathway activity, resulting from defects in pRb itself, cyclin-dependent kinase inhibitor 2A (p16^INK4a^), cyclin D1 (CCND1), or cyclin-dependent kinase 4 (CDK4), is observed in the majority of human sporadic cancers ([@pbio-0020022-Marshall1]; [@pbio-0020022-Weinberg1]; [@pbio-0020022-Sherr1]; [@pbio-0020022-Ortega1]). This pathway is commonly altered early in cancer development, indicating an ability to predispose cells to tumorigenesis. However, whether the mechanism(s) is similar among cell types is not known. Examination of pRb inactivation in specific cell types in vivo has been technically challenging due to the apparent functional compensation or redundancy among pRb, retinoblastoma-like 1 (p107), and retinoblastoma-like 2 (p130) in many cell types of the mouse ([@pbio-0020022-Luo1]; [@pbio-0020022-Robanus-Maandag1]; [@pbio-0020022-Dannenberg1]; [@pbio-0020022-Sage1]). Thus, genetic inactivation of the *Rb* gene alone, either by conditional deletion ([@pbio-0020022-Marino1]) or by the generation of chimeric mice harboring pRb-deficient cells ([@pbio-0020022-Maandag1]; [@pbio-0020022-Williams1]) yields only medulloblastomas, pituitary, and thyroid tumors. We have begun to systematically examine the role of retinoblastoma protein family (pRb~f~) inactivation in multiple cell types of the mouse by dominant expression of T~121~, a truncation mutant of simian virus 40 (SV40) T antigen that inactivates all three pRb-related proteins ([@pbio-0020022-DeCaprio1]; [@pbio-0020022-Dyson2]; [@pbio-0020022-Ewen1]; [@pbio-0020022-Stubdal1]; [@pbio-0020022-Sullivan1]). In this report we determine the role of pRb inactivation in mammary adenocarcinoma predisposition, establish a role for p53 inactivation in subsequent mammary adenocarcinoma progression, and, together with our previous studies, provide a comprehensive comparison of these mechanisms in distinct epithelial lineages. pRb plays a critical role in eukaryotic cell cycle progression, when cells exit G0 or G1 and enter S phase, thereby acting as a crucial negative regulator of cellular proliferation and neoplasia ([@pbio-0020022-Sherr2]). In quiescent or early G1-phase cells, pRb is hypophosphorylated and associates with specific members of the E2F transcription factor family, converting them to active transcriptional repressors ([@pbio-0020022-Hamel1]; [@pbio-0020022-Weintraub1]). Gene repression is also mediated by pRb and p130 recruitment of histone deacetylase to promote formation of inhibitory nucleosomes ([@pbio-0020022-Brehm1]; [@pbio-0020022-Luo1]; [@pbio-0020022-Magnaghi-Jaulin1]). The many proteins found in association with pRb suggest other regulatory mechanisms may also be involved ([@pbio-0020022-Morris1]), although the biological potential for most of these interactions remains yet unproven. Cell cycle progression from G to S phase occurs when complexes of D-type cyclins/CDK4/CDK6 phosphorylate pRb, thereby derepressing E2Fs to direct transcription of DNA-replication machinery and nucleotide biosynthesis genes ([@pbio-0020022-Dyson1]). Like most human solid tumors, breast cancers harbor frequent alterations in the pRb pathway, including CCND1 overexpression in 45% ([@pbio-0020022-Buckley1]), p16^INK4A^ loss in 49% ([@pbio-0020022-Geradts1]), and pRb loss in 6% of breast tumors ([@pbio-0020022-Geradts1]). In the *Rb*-deficient mouse mammary gland, p107 and/or p130 may play overlapping or compensatory roles, as they do during embryonic development, given that pRb is dispensable for normal mammary development and mammary tumor suppression. pRb-deficient embryonic stem cells participate in normal mammary gland formation in chimeric mice ([@pbio-0020022-Maandag1]), and donor *pRb^−/−^* mammary precursor cells transplanted into wild-type mice can populate a normal mammary gland without evidence of neoplasia, even after multiple pregnancies ([@pbio-0020022-Robinson1]). The interplay between pRb signaling and the tumor protein p53 pathway is also critical to the understanding of breast cancer biology. Since the pRb pathway is defective in a majority of human tumors and the *p53* gene is mutated in about half of them, including approximately a fifth of sporadic breast cancers ([@pbio-0020022-Nigro1]; [@pbio-0020022-Greenblatt1]), these aberrations often coexist. Whether loss of these tumor suppressor pathways collaborate in tumorigenesis is also cell type-specific. In a brain epithelial tumor model, we previously demonstrated that, in the absence of pRb~f~ function, inactivation of p53 significantly decreases apoptosis and accelerates tumor growth in vivo ([@pbio-0020022-Symonds3]). However, in astrocytic brain tumors induced by pRb~f~ inactivation, tumor progression is not accelerated by reduced p53 activity; rather, the phosphatase and tensin homolog (*PTEN*) regulates the apoptosis, and reduction in its function accelerates tumor growth ([@pbio-0020022-Xiao1]). In this report, we extend our analysis of pRb function in vivo and examine the consequence of pRb~f~ loss specifically in mammary epithelium. These studies serve not only to provide insight into the cell specificity of tumor suppression mechanisms, but also to model the stepwise evolution of breast adenocarcinomas that harbor defects in this pathway. Results {#s2} ======= Generation of Mice with Inducible pRb~f~ Deficiency in Mammary Cells {#s2a} -------------------------------------------------------------------- Seven founder mice were generated in which the *T~121~* gene was regulated by the whey acidic protein (WAP) transcriptional signals ([Figure 1](#pbio-0020022-g001){ref-type="fig"}; see [Materials and Methods](#s4){ref-type="sec"}). Of these, two founder animals died spontaneously of unknown causes, while the transgenic progeny of the third line died prematurely, also of unknown cause ([Figure 2](#pbio-0020022-g002){ref-type="fig"}A). The extent to which the transgene contributed to these deaths was not investigated further; however, ectopic transgene expression was detected in several tissues (data not shown). Characterization of female mice of the four remaining lines is the focus of this report. ::: {#pbio-0020022-g001 .fig} Figure 1 ::: {.caption} ###### Diagram of the *WAP-T~121~* Transgene and Protein The fragment consists of the 2.4 kb WAP promoter (hatched) and the mutant SV40 T-antigen coding region (white box) containing two deletions, the 196-bp amino-terminal deletion, which abolishes small t antigen production, and the *dl*1137 deletion, which truncates T antigen. Both the J domain and the LXCXE domain are required for pRb family inactivation (see [Materials and Methods](#s4){ref-type="sec"}). :::  ::: ::: {#pbio-0020022-g002 .fig} Figure 2 ::: {.caption} ###### Expression of T~121~ Protein in *WAP-T~121~* Mice and a Summary of Gross Phenotypes As expected, each line showed mammary-specific expression following lactation induction, while line 4 showed more widespread expression, with protein detected in brain and kidney. Mice from the higher-expressing lines 3 and 4 failed to nurse because of lactation defects. Mammary glands of adult female mice from all four lines showed elevated proliferation and apoptosis. Glands from line 1 and 2 mice were hyperplastic, while glands from lines 3 and 4 were atrophic. Lines 3 and 4 later developed carcinomas and other neoplasms. T~121~ protein was detected by Western blot analysis in lactating mammary glands of animals from all four lines (B), although the lower-expressing lines 1 and 2 required immunoprecipitation with anti-T-antigen antibody prior to Western blot analysis (right panel in \[B\]). Brain tumor extract (see [Materials and Methods](#s4){ref-type="sec"}) was used for a positive control, and nontransgenic mammary tissue extract was used for a negative control. A timecourse analysis of T~121~ expression (C) shows lactation-induced expression peaking at 5 d postpartum. **Abbreviations:** Adeno-Ca, adenocarcinoma; AP, elevated apoptosis in mammary gland; At, atrophy; dpc, postcoital; FTN, failure to nurse; Hyp, hyperplastic acini; MG, mammary gland; MIN, mammary epithelia neoplasia; ND, not determined; nt, nontransgenic; pp, postpartum; Pr, elevated proliferation in mammary gland; pw, post-weaning. **Footnotes:** ^a^Mosaic founder animal.^b^At earlier stages, development defects attributed to atrophy, while MIN and adenocarcinoma were observed at terminal stages.^c^Approximately half of progeny died of unknown cause. :::  ::: T~121~ Is Expressed in Lactating Mammary {#s2b} ---------------------------------------- Western immunoblotting analyses of mammary gland extracts demonstrated that this tissue expresses T~121~ protein at the expected size in all four lines ([Figure 2](#pbio-0020022-g002){ref-type="fig"}B). T~121~ expression in lines 1 and 2 was only revealed following immunoprecipitation using an anti-T-antigen antibody prior to Western blot analysis, indicating lower levels of T~121~ (right panel in [Figure 2](#pbio-0020022-g002){ref-type="fig"}B). A survey of select tissues showed that detectable expression was restricted to the mammary gland in lines 1--3, while expression was more widespread in the higher expressing line 4 (data not shown) and included brain and kidney expression. As expected, T~121~ expression was induced by lactation with highest levels observed 5 d postpartum ([Figure 2](#pbio-0020022-g002){ref-type="fig"}C). Southern blot analyses indicate that mice in line 3, which was used as a representative line for extensive characterization, harbor approximately ten copies of the transgene at a single insertion site (data not shown). Impact of Rb~f~Inactivation in Mammary Epithelium {#s2c} ------------------------------------------------- Representative histological analysis of lactating mammary glands (day 1) from single-pregnancy females of the line 2 founder (F~0~) and a line 3 F~1~ mouse shows that the impact of *Rb* perturbation is severalfold. Compared to an age- and parity-matched control tissue, the normal architecture of the lactating mammary tissue is disturbed. In contrast to normal tissue where acini consist of a single layer of secretory epithelia with milk-filled lumen ([Figure 3](#pbio-0020022-g003){ref-type="fig"}A), transgenic animals have a lower density of acini ([Figure 3](#pbio-0020022-g003){ref-type="fig"}K), consistent with atrophy, and are often atypical ([Figure 3](#pbio-0020022-g003){ref-type="fig"}I). T~121~-positive mammary epithelial cells were associated with abnormalities ([Figure 3](#pbio-0020022-g003){ref-type="fig"}B, [3](#pbio-0020022-g003){ref-type="fig"}F, and [3](#pbio-0020022-g003){ref-type="fig"}J). The line 2 F~0~ animal was mosaic for T~121~ protein expression with distinct regions of expressing and nonexpressing cells ([Figure 3](#pbio-0020022-g003){ref-type="fig"}F), whereas T~121~ expression in the line 3 animal was in secretory epithelium distributed throughout the gland ([Figure 3](#pbio-0020022-g003){ref-type="fig"}J). Increased proliferation, indicated by proliferating cell nuclear antigen (PCNA) staining, was also observed in transgenic mammary glands ([Figure 3](#pbio-0020022-g003){ref-type="fig"}C, [3](#pbio-0020022-g003){ref-type="fig"}G, and [3](#pbio-0020022-g003){ref-type="fig"}K), concomitant with increased levels of apoptosis assayed by TUNEL staining ([Figure 3](#pbio-0020022-g003){ref-type="fig"}D, [3](#pbio-0020022-g003){ref-type="fig"}H, and [3](#pbio-0020022-g003){ref-type="fig"}L). Quantification of T~121~ expression and apoptosis revealed higher protein expression levels (see [Figure 2](#pbio-0020022-g002){ref-type="fig"}B) correlate with higher percentages of apoptotic cells ([Figure 4](#pbio-0020022-g004){ref-type="fig"}A). Consistent with a model for cell-autonomous functioning of T~121~, the pattern of abnormalities of morphology, proliferation, and apoptosis in the mosaic animal mimicked the regionalized T~121~ expression pattern, and conversely, where T~121~ protein was absent, the tissue appeared normal. ::: {#pbio-0020022-g003 .fig} Figure 3 ::: {.caption} ###### Mammary-Specific Inactivation of the pRb Pathway Induces Extensive Abnormalities Histologic comparisons of nontransgenic (A--D), mosaic (F~0~ line 2 \[E--H\]), and transgenic (F~1~, line 3 \[I--L\]) lactating mammary glands reveals that T~121~ expression results in increased proliferation and apoptosis. Hemotoxylin and eosin staining shows acini of the normal lactating gland are composed of a single layer of secretory epithelial cells (A) with milk-filled lumen. Consistent with atrophy, transgenic animals have a lower density of acini demonstrated by the presence of lipid-filled adipocytes (asterisk in \[K\]). Acini composed of T~121~-expressing cells are atypical. Many are collapsed and composed of tall columnar epithelia of large hyperchromatic cells with papillary tufting (arrows in \[I\]). Transgene-expressing cells have large pleomorphic nuclei (open arrows in \[G\]) as compared to nuclei of nonexpressing cells (arrows in \[G\]). Staining for T~121~ expression (blue in \[B\]--\[J\]) indicates the line 2 F~0~ animal is mosaic, showing localized expression (F), whereas the transgene expresses throughout the gland of an F~1~ line 3 animal (J). Increased proliferation assayed by PCNA staining (red) is also localized in the mosaic founder (G), but found throughout the F~1~ transgenic gland (K). Similarly, TUNEL staining (brown) demonstrates increased apoptosis in transgenic animals (H and L); moreover, the regionalized apoptosis in the mosaic gland (H) strongly suggests that transgene expression and not precocious involution is the cause. All samples are from primiparous females on lactation day 1. :::  ::: ::: {#pbio-0020022-g004 .fig} Figure 4 ::: {.caption} ###### Reduced p53 Activity Decreases Apoptosis but Does Not Increase Proliferation Representative apoptosis levels of each mouse line correlate with T~121~ expression as indicated by the percentage of TUNEL positive cells (A). Decreasing levels of p53 activity correlate with lower levels of apoptosis in transgenic mammary glands (B). The mean percentage of apoptotic cells in *p53* wild-type transgenic glands was 21%; in *p53* heterozygous animals, 9%; and in *p53* null animals, 5% (B), indicating that 75% of the apoptosis is p53-dependent. Apoptosis levels are further reduced to 2% in terminal stage tumors (B, Tumors). The percentage of PCNA staining cells remains unchanged in *p53* heterozygous or nullizygous animals (C), indicating that reduction of p53 activity levels had no significant impact on cell proliferation. Samples were derived from primiparous animals on lactation day 1, except as indicated as tumor samples (B). Transgenic animals in (B) and (C) were from line 3. :::  ::: Role of p53 in Apoptosis {#s2d} ------------------------ To investigate the impact of germline loss of p53 on apoptosis levels in Rb~f~-deficient mammary glands, we mated line 3 animals to *p53* null mice to generate transgenic and nontransgenic females of distinct *p53* genotypes (+/+, +/−, −/−). Transgene expression was induced by a single pregnancy, and mammary glands were examined on lactation day 1. As expected, nontransgenic mammary glands showed no appreciable apoptosis regardless of *p53* status ([Figure 4](#pbio-0020022-g004){ref-type="fig"}B). However, in transgenic animals, decreased levels of p53 activity were correlated with lower levels of apoptosis. The mean percentage of apoptotic cells in *p53* wild-type transgenic glands was 21%; in *p53* heterozygous animals, 9%;and in *p53* null animals, 5% ([Figure 4](#pbio-0020022-g004){ref-type="fig"}B), indicating that 75% of the apoptosis is p53-dependent. That we could detect haploinsufficiency of *p53* for apoptosis is remarkable, since in the previously characterized T~121~-expressing choroid plexus epithelium, apoptosis levels were the same in *p53* heterozygous and wild-type backgrounds ([@pbio-0020022-Lu1]). This observation indicates that there is a threshold for p53 levels in eliciting apoptosis and that either the threshold is different between cell types or that the absolute functional p53 level is distinct. Such differences could have significant impact on the requirements for tumorigenesis. Role of p53 in Proliferation {#s2e} ---------------------------- In two other transgenic mouse models of breast cancer, where tumors were initiated by activated Harvey rat sarcoma viral oncogene homolog (v*-Ha-ras*) ([@pbio-0020022-Hundley1]) or *wingless*-related murine mammary tumor virus (MMTV) integration site 1 (*Wnt-1*) ([@pbio-0020022-Donehower1]), inactivation of p53 did not result in a reduction of apoptosis; rather, loss of p53 was associated with increased proliferation of the mammary epithelium. To determine whether p53 inactivation also impacted mammary cell proliferation induced by Rb~f~ inactivation, glands from primiparous lactating (day 1) mice were assessed for the expression of nuclear PCNA. Unlike the tumors initiated by activated *Ras* or *Wnt-1*, *p53* heterozygosity or nullizygosity had no significant impact on the level of cell proliferation ([Figure 4](#pbio-0020022-g004){ref-type="fig"}C). This experiment indicates that p53 can have distinct mechanisms of action depending on the nature of the initiating lesion. pRb Inactivation Predisposes to Tumorigenesis {#s2f} --------------------------------------------- All females from higher-expressing lines (lines 3 and 4) failed to nurse pups because of lactation defects and developed mammary tumors after multiple pregnancies. Because line 4 mice expressed T~121~ in nonmammary tissues, further characterization focused on line 3. For this line, the median time following initial transgene induction until a palpable tumor appeared was 10 mo, and within 16 mo, all mice developed palpable tumors ([Figure 5](#pbio-0020022-g005){ref-type="fig"}A). Interestingly, latency in this line on a BALB/cJ background (see [Materials and Methods](#s4){ref-type="sec"}) was reduced to a median time of 8.5 mo (*p* = 0.0077; [Figure 5](#pbio-0020022-g005){ref-type="fig"}A) indicating the presence of modifier alleles. The condensed timeframe for tumor development in this strain will also be valuable for future preclinical studies using this model. However, all further studies in the current report were carried out on the original B6D2F1 background. ::: {#pbio-0020022-g005 .fig} Figure 5 ::: {.caption} ###### Mammary Tumor Onset and Growth Are Accelerated by p53 Reduction Among line 3 animals, the median time following initial transgene induction until a palpable tumor appeared was 10 mo, and within 16 mo, all mice developed palpable tumors (red line in \[A\]). In *p53^+/−^* transgenic animals (blue line in \[A\]), mammary tumors were detected significantly earlier (*p* \< 0.0003) with a median onset of 6 mo. Among mice with BALB/cJ background (black line in \[A\]), median mammary tumor latency (8.5 mo) was significantly shorter (*p* = 0.0077) compared to mice of the hybrid BDF1 background strain and indistinguishable (*p* = 0.2466) from *WAP-T* *~121~* *;p53^+/−^* mice. Once palpable, *WAP-T* *~121~* *;p53^+/−^* tumors grew faster than the *p53* wild-type counterparts (B). The average growth rates for *p53^+/+^* (black solid) and *p53^+/−^* (dashed) are indicated. :::  ::: The median onset for mammary tumors in line 4 was 14 mo (*n* = 3; data not shown), which indicates that the transgene and not its insertion caused tumorigenesis. With two exceptions, line 3 *WAP-T~121~*mice, regardless of *p53* status, developed a single palpable tumor (87% of *p53^+/+^*, *n* = 15; 78% of *p53^+/−^*, *n* = 9). A single mouse with either two or three palpable tumors was also observed in both *p53* +/+ and +/− backgrounds. At least one additional nonpalpable tumor was visible during necropsy in approximately one-third of all tumor-bearing mice. While the two lower-expressing lines, lines 1 and 2, were able to nurse pups and appeared grossly normal, both had hyperplastic lobular alveoli associated with increased levels of proliferation and apoptosis. However, females from low-expressing lines did not develop adenocarcinomas after at least four pregnancies and 20 mo of age (line 1, *n* = 2; line 2, *n* = 6) (data not shown). Most terminal stage tumors in either wild-type or *p53^+/−^* backgrounds were adenocarcinomas ([Figure 6](#pbio-0020022-g006){ref-type="fig"}A, [6](#pbio-0020022-g006){ref-type="fig"}B, and [6](#pbio-0020022-g006){ref-type="fig"}E); however, we also observed four pilar tumors ([Figure 6](#pbio-0020022-g006){ref-type="fig"}C and [6](#pbio-0020022-g006){ref-type="fig"}E) and one spindle cell carcinoma ([Figure 6](#pbio-0020022-g006){ref-type="fig"}D and [6](#pbio-0020022-g006){ref-type="fig"}E). Terminal-stage mammary adenocarcinomas resembled poorly to moderately differentiated invasive ductal adenocarcinoma in humans. Morphologically, we designate these tumors as mixed solid and glandular carcinomas with necrosis and fibrosis. Poorly differentiated solid tumors ([Figure 6](#pbio-0020022-g006){ref-type="fig"}A) are composed of nests of epithelial cells with large pleomorphic nuclei and delicate chromatin patterns with inverted nuclear:cytoplasmic ratios, while glandular tumors ([Figure 6](#pbio-0020022-g006){ref-type="fig"}B) are composed of irregular glands with varying degrees of differentiation. While most animals had a single tumor mass, the adenocarcinomas were multifocal, with solid tumors consisting of subclones of distinct expansile masses, and with only two exceptions, glandular tumors were coincident with solid tumors. The adenocarcinomas were malignant, infiltrating dense, fibrous connective tissue, and were accompanied by strong peripheral immune response ([Figure 6](#pbio-0020022-g006){ref-type="fig"}A). ::: {#pbio-0020022-g006 .fig} Figure 6 ::: {.caption} ###### Tumor Morphologies Hemotoxylin and eosin staining of *WAP-T~121~* (C and D) and *WAP-T* *~121~* *p53^+/−^* (A and B) (also representative of *WAP-T~121~*) tumor sections shows that terminal stage adenocarcinomas have varied morphologies. Poorly differentiated solid tumors were comprised of nests (A) or cords of epithelial cells (Tu) that infiltrate a fibrous stroma and were accompanied by necrosis (arrow) and strong immune response (arrowheads). Moderately differentiated glandular tumors (B) consisted of irregular, disorganized glands. In animals of wild-type *p53* background, four pilar tumors (C), distinguished by swirls of laminar acellular keratin (arrow), and a single spindle cell carcinoma (D) were also observed. For comparison, a lactating gland from a wild-type animal is shown in [Figure 3](#pbio-0020022-g003){ref-type="fig"}A. The percentage of animals displaying each of the phenotypes is summarized in (G). Since many tumors shared multiple morphologies, the sum exceeds 100%. :::  ::: Mammary Tumor Onset and Growth Are Accelerated by p53 Reduction {#s2g} --------------------------------------------------------------- Since 75% of the apoptosis induced by Rb~f~ inactivation was mediated by *p53* and was indeed reduced even in *p53^+/−^* mice, we investigated the impact of *p53* loss on tumor onset and growth kinetics. Animals harboring either one or two *p53* null alleles were monitored for mammary tumors. As expected, a subset of *p53^+/−^* and *p53^−/−^* mice developed nonmammary tumors (either thymic lymphomas or sarcomas), consistent with published reports ([@pbio-0020022-Jacks1]; [@pbio-0020022-Sandgren1]; [@pbio-0020022-Dannenberg1]). All *p53^−/−^* mice (*n* = 4) succumbed to these tumors by 4 mo of age, prior to developing palpable mammary tumors, so acceleration of this phenotype could not be assessed. In *p53^+/−^* animals, mammary tumors were detected significantly earlier (see [Figure 5](#pbio-0020022-g005){ref-type="fig"}A; *p* = 0.0003) compared with *p53^+/+^* mice. Furthermore, once palpable, *WAP-T~121~;p53^+/−^* tumors grew significantly faster than the *p53* wild-type counterparts (see [Figure 5](#pbio-0020022-g005){ref-type="fig"}B). The observation of four pilar tumors in *p53^+/+^* animals and none in *p53^+/−^* animals is a statistically significant difference (Fisher--Freeman--Halton\'s exact test, *p* = 0.0177) and suggests that the reduction of p53 activity drives tumors to the adenocarcinoma phenotype. Taken together, these studies indicate that *p53* heterozygosity leads to increased tumor growth rates and/or progression and may alter the spectrum of tumor morphologies. Selective Pressure for p53 Inactivation during Adenocarcinoma Development {#s2h} ------------------------------------------------------------------------- Since apoptosis was significantly reduced in *WAP-T~121~;p53^+/−^* mammary tissue compared with that of *WAP-T~121~;p53^+/+^* mice, it was possible that *p53* heterozygosity was sufficient for tumor acceleration. To assess whether this was the case or whether there was selective pressure for p53 inactivation during tumor progression, real-time PCR analysis was employed to determine the status of the wild-type *p53* allele in *WAP-T~121~;p53^+/−^* tumors. Of ten tumors, eight showed loss of the wild-type *p53* allele ([Table 1](#pbio-0020022-t001){ref-type="table"}), indicating that the apoptosis reduction observed in *WAP-T~121~;p53^+/−^*mammary epithelium was not sufficient for tumor progression. Significant selective pressure favored cells that had completely inactivated p53, indicating that tumor progression requires further reduction of apoptotic activity and/or that *p53* loss contributes to tumor progression through additional mechanisms that confer selective advantage. Assessment of apoptosis levels in terminal tumors showed apoptosis levels were indeed reduced in comparison to preneoplastic tissue (see [Figure 4](#pbio-0020022-g004){ref-type="fig"}B). ::: {#pbio-0020022-t001 .table-wrap} Table 1 ::: {.caption} ###### *p53* LOH among the Majority of *p53^+/−^* Tumors :::  Real-time PCR was performed in duplicate to determine the status of the wild-type *p53* alleles in the mammary tumors or tissues as indicated. Analysis of standard samples indicates that copy numbers of 2, 1, and 0 are indicated by 2^-ΔΔCt^ values of greater than or equal to 0.7, 0.2--0.7, and less than 0.2, respectively ([@pbio-0020022-Lu1]). Of ten *WAP-T~121~;p53^+/−^* tumors, eight show LOH of *p53* gene, while all three *WAP-T~121~;p53^+/+^* tumors retained both *p53* alleles. Abbreviation: Tg, transgenic ^a^Tumor samples were derived from line 3 animals, except tumor 1, which was derived from a line 4 animal ^b^ΔΔC~t~ = \[sample C~t~ (*p53*) − sample C~t~ (*β-actin*) \] − \[ *p53^+/+^* control Ct (*p53*) − *p53^+/+^* control C~t~ (*β-actin*) \]. C~t~ = the number of cycles required to reach a threshold value, which is set within the exponential phase of the logarithmic scale amplification plot ::: Comparative Genomic Hybridization Reveals Recurrent Chromosomal Imbalances in Tumors, but Limited Chromosomal Instability {#s2i} ------------------------------------------------------------------------------------------------------------------------- Among the multiple mechanisms of tumor suppression attributed to p53, a common hypothesis is that p53 prevents genetic instability. Indeed, studies using other mouse models indicate loss of p53 function in tumors often correlates with chromosomal instability. These include other breast cancer models such as *Wnt-1p53^+/−^* ([@pbio-0020022-Donehower1]) and *MMTV-ras p53^+/−^* ([@pbio-0020022-Hundley1]) and *p53^+/−^* thymic lymphomas and sarcomas ([@pbio-0020022-Venkatachalam1]). In marked contrast, our study of p53 deficiency in an evolving brain epithelial tumor showed that tumorigenesis progresses without chromosomal instability, indicating p53 loss contributes via alternative mechanisms ([@pbio-0020022-Lu1]). To determine whether this difference was due to cell-type specificity, differences in initiating mechanisms, or differences in experimental approaches, we analyzed the genome of mammary *WAP-T~121~;p53^+/−^* tumors. We employed two methods of comparative genomic hybridization (CGH): chromosome-based CGH (cCGH) (Panel I in [Figure 7](#pbio-0020022-g007){ref-type="fig"}) ([@pbio-0020022-Kallioniemi1]) and microarray CGH (aCGH) (Panel II in [Figure 7](#pbio-0020022-g007){ref-type="fig"}) ([@pbio-0020022-Solinas-Toldo1]; [@pbio-0020022-Pinkel1]). ::: {#pbio-0020022-g007 .fig} Figure 7 ::: {.caption} ###### CGH Analysis Shows Limited Genomic Instability Twelve tumors were analyzed by CGH: ten by cCGH (Panel I, A--J), eight by aCGH (Panel II, B, C, E, and H--L), and six by both procedures (Panels I and II, B, C, E, and H--J). In Panel I, green and red lines adjacent to the ideograms indicate relative gain or loss, respectively. Tumor sample identities are indicated by letters above gain and loss lines. Only a single sample (Panel I, D) shows loss of Chromosome 11. Telomeric sequences of many chromosomes are increased, most frequently Chromosomes 5 and 15. Recurrent losses are seen on Chromosomes 10 and X. For aCGH (Panel II), the genomic map is depicted with chromosomes horizontally aligned centromere to telomere. The relative fluorescence intensities (tumor:normal) are indicated along the vertical axis. Individual BACs are plotted according to their physical map position versus relative fluorescence, with sample identities indicated by a unique symbol for each tumor. To simplify visualization, only BACs with relative intensities greater than 1.25 (gains) or less than 0.75 (losses) are shown. X Chromosome values were halved to account for sex-mismatched samples. Changes spanning the entire length of the chromosome are readily detected on Chromosomes 6, 8, 10, 15, 18, and X. None of the clones showing loss on Chromosome 11 spans the *p53* locus. The original *p53* background of the animal and the *p53* LOH status of each tumor are also indicated in the legend. :::  ::: Twelve mammary tumors were assayed by CGH: ten by cCGH, eight by aCGH, and six by both procedures. Both assays revealed limited genomic imbalances ([Figure 7](#pbio-0020022-g007){ref-type="fig"}), yet only a single tumor showed loss of Chromosome 11 (which harbors *p53*). Among samples tested by both methods, there is strong concordance among large chromosomal changes, encompassing multiple cytological bands to whole chromosome lengths. For example, there is apparent whole chromosomal duplication of Chromosomes 6 and 15 in tumor C and of Chromosomes 8 and 18 in tumor H, monosomy of Chromosome 10 in tumor J, and loss of X Chromosome in tumors E and H (all or partial, respectively). Making comparisons among imbalances spanning shorter chromosome lengths was more difficult, mainly due to the challenge of reconciling cytological and physical maps. Furthermore, technical limitations may account for real differences between the two assays: small imbalances detected by one to several bacterial artificial chromosome (BAC) clones are irresolvable by cCGH; on the other hand, the relatively low density of BAC clones may not adequately sample smaller regions detected by cCGH. Nevertheless, on average, about five imbalances per tumor were detected by cCGH. This number is comparable to the number of changes observed in myelocytomatosis oncogene (c*-myc*)-induced mouse mammary tumors ([@pbio-0020022-Weaver1]) and human tumors ([@pbio-0020022-Ried1]), yet less than the number of changes seen in breast cancer 1 (*Brca1*)-deficient mouse tumors (8.0) and more than v-*erb-b2* erythroblastic leukemia viral oncogene homolog 2 (HER2/*neu*)-induced tumors (2.8) ([@pbio-0020022-Montagna1]; [@pbio-0020022-Weaver2]). Discussion {#s3} ========== Common Mechanisms for Tumor Progression in Epithelial Cells of Distinct Origin {#s3a} ------------------------------------------------------------------------------ Here we report that loss of pRb family function in mammary epithelium predisposes to malignant adenocarcinoma. Using a single transgenic allele, we have thus far inactivated the pRb pathway in several cell types in the mouse: brain choroid plexus epithelium, astrocytes, and mammary epithelium. In each case, despite the marked differences among these divergent cell types, pRb inactivation causes a similar response, initially evoking increased proliferation and apoptosis and, ultimately, predisposing to tumorigenesis ([@pbio-0020022-Chen1]; [@pbio-0020022-Saenz-Robles1]; [@pbio-0020022-Symonds3]; [@pbio-0020022-Xiao1]). Not surprisingly, the long latency of mammary adenocarcinomas indicates that additional events are required for tumor progression. We show that mammary epithelium is similar to brain epithelium ([@pbio-0020022-Symonds3]; [@pbio-0020022-Lu1]) in its requirement for p53 activity in the apoptotic response to aberrant proliferation caused by pRb~f~ inactivation. Previous models using wild-type large T antigen ([@pbio-0020022-Li2]; [@pbio-0020022-Husler1]; [@pbio-0020022-Green1]; [@pbio-0020022-Schulze-Garg1]) are unable to address the relative contribution of pRb and p53, since T antigen also binds and inactivates p53. As in brain epithelium, we show here that when the mammary tumor phenotype is initiated by pRb~f~ inactivation, most of the apoptosis is mediated through p53. Furthermore, as in brain epithelium, heterozygosity for a null *p53* allele significantly shortens tumor latency (discussed further below). Importantly, the Rb~f~ deficiency-induced apoptotic response and inhibition of tumor progression are not universally dependent on p53. In astrocytes, we recently showed that PTEN, and *not* p53, modulates these same responses to Rb~f~ inactivation. In contrast to the p53-dependent apoptosis of mammary epithelial cells in response to pRb~f~ deficiency, apoptosis associated with normal mammary involution subsequent to lactation does not require p53 ([@pbio-0020022-Li1]). Thus, the "wiring" of the apoptotic response within this cell type is not global, but rather depends on the signal. Although loss of p53-dependent apoptosis accounts for the acceleration of mammary tumorigenesis in *WAP-T~121~;p53^+/−^* mice, in models expressing either activated *v-Ha-ras* ([@pbio-0020022-Hundley1]) or *Wnt-1* ([@pbio-0020022-Jones1]), earlier tumor formation in p53 heterozygous and homozygous null mice is accounted for by increased proliferation rather than attenuated apoptosis. An important caveat to this comparison is that the latter studies compared apoptosis in terminal tumors in which loss of apoptosis might have been selected regardless of initial p53 status, leaving open the possibility that tumor growth rates in these models reflect the combined effects of increased proliferation as well as reduced apoptosis. Nevertheless, there is a clear difference in *WAP-T~121~* mammary gland in that, unlike the *Ras* and *Wnt-1* models, proliferation levels do not depend on *p53* status. Taken together, these observations indicate that the specific cellular response to an oncogenic stimulus depends on the nature of the initial insult. Given that the pRb pathway is directly disrupted in T~121~-expressing cells, this could be explained if these other initiating events evoke p53-dependent growth arrest which, in part, functions upstream of pRb. High Selective Pressure for p53 Inactivation in the Transition to Aggressive Mammary Adenocarcinoma {#s3b} --------------------------------------------------------------------------------------------------- Most of the apoptosis induced by pRb~f~ deficiency in both mammary (75%) and brain (85%) epithelia is *p53*-dependent as determined by comparing *p53^+/+^* and *p53^−/−^* tissue. However, while *p53* heterozygosity had no impact on the level of apoptosis in the brain epithelium, in the mammary gland the level was reduced by half in *p53^+/−^* tissue. Given that apoptosis is the basis for selective inactivation of p53 in the brain tumor model ([@pbio-0020022-Lu1]; X. Lu and T. Van Dyke, unpublished data), it was possible that the pressure was relieved or reduced in *WAP-T~121~;p53^+/−^* mice. However, aggressive adenocarcinoma growth was accelerated with 100% penetrance, and 80% of these tumors underwent selective loss of the wild-type *p53* allele, just as in the brain tumor model ([@pbio-0020022-Lu1]). This result indicates that tumor progression requires more than a simple reduction in the level of apoptosis; it follows that *p53* may contribute to tumor suppression by multiple mechanisms. While both mammary and brain carcinomas show high rates of *p53* loss of heterozygosity (LOH), the mechanism of loss may be distinct. Chromosome loss clearly explains *p53* LOH in the brain carcinoma model ([@pbio-0020022-Lu1]) where nearly all tumors (greater than 90%) are monosomic for Chromosome 11, whereas only a single mammary tumor analyzed by CGH showed Chromosome 11 loss. Alternative mechanisms that may explain *p53* LOH in the mammary tumors include somatic recombination or chromosomal reduplication following mitotic nondisjunction. Whether these alternative routes of LOH represent bona fide tissue-specific phenomena or are due to relatively small sample sizes will require further analyses. Interestingly, most mammary tumors derived from *Brca1*-deficient mice lost *p53*; however, regions distal to *p53* were amplified ([@pbio-0020022-Weaver2]). Thus, it is possible that mammary tumor promoting factor(s) is located on distal Chromosome 11, selecting *against* loss. Limited Chromosomal Instability in the Absence of p53 {#s3c} ----------------------------------------------------- Genomic instability is a hallmark of most human solid tumors, and a widely held view is that p53 represses instability to suppress tumorigenesis, although evidence for this activity has been mostly correlative. Contrary to this model, we demonstrated previously that in the absence of p53 activity in brain epithelia, tumors progress without chromosomal instability; except for Chromosome 11 loss, in a *p53^+/−^* background these carcinomas are diploid ([@pbio-0020022-Lu1]). Here we show that mammary tumors similarly harbor limited genome-wide alterations. While the number of aberrations within the mammary tumors is small, it is intriguing that some changes are recurrent, suggesting that their accrual is causal in tumorigenesis. T~121~-induced mammary carcinomas harbor more genomic imbalances than brain tumors (approximately five versus approximately one). One explanation for this observation is that, because the brain is a vital organ, animals succumb to their illness when the brain tumor is at a relatively earlier stage at which fewer changes have accumulated. However, chromosome content of choroid plexus tumors passaged further in xenografts remained stable (X. Lu and T. Van Dyke, unpublished data). The converse experiment, analyses of early mammary tumors subsequent to p53 loss, will be required to determine the kinetics of chromosomal changes in this tissue. Pocket Protein Redundancy {#s3d} ------------------------- Chimera and tissue-grafting experiments with pRb-deficient cells indicate the absence of pRb alone is not sufficient for abnormal mammary development or tumor formation ([@pbio-0020022-Maandag1]; [@pbio-0020022-Robinson1]). Yet mammary-directed overexpression of CCND1, an upstream regulator of pRb~f~, leads to mammary adenocarcinoma ([@pbio-0020022-Wang1]). Given other recent studies indicating the possibility for compensation of pRb function by p107 and/or p130 ([@pbio-0020022-Dannenberg1]; [@pbio-0020022-Sage1]) and the clear redundancy of function in some murine cell types ([@pbio-0020022-Robanus-Maandag1]; [@pbio-0020022-Xiao1]), it is likely that the discrepancy among our results can be explained by overlapping functions of other family members, p107 and/or p130. In our studies, T~121~ abrogates the activities of all *Rb* family members by a dominant interfering mechanism. A subtly distinct alternative explanation is that the acute loss of pRb signaling, rather than a chronic loss as of pRb during mammary development, as in the chimera and grafting models, accounts for the difference. Cell culture experiments that support this hypothesis were recently reported ([@pbio-0020022-Sage2]). In this model, p107 and p130 may be more responsive to pRb regulatory signals during development than in the terminally differentiated tissue; therefore, the developing tissue more easily accommodates for the absence of pRb in the pool of available pocket proteins. In the *WAP-T~121~* model, the gland undergoes normal development and then is subsequently subjected to acute pRb pathway loss. We presume that this scenario more closely mimics the situation of spontaneous somatic loss in adult human breast. The test of this alternative hypothesis awaits analyses of tissue-specific inactivation of pRb and the paralogous pocket proteins using conditional alleles. A Model for Mammary Tumorigenesis Initiated by Targeting the pRb Pathway {#s3e} ------------------------------------------------------------------------ The *WAP-T~121~* model is a significant addition to the current repertoire of preclinical mammary tumor models exploring the role of pRb pathway in tumorigenesis. Despite the prevalence of pRb pathway defects in human sporadic cancers, mice harboring germline mutations of *p16^INK4a^* do not develop mammary cancer ([@pbio-0020022-Krimpenfort1]; [@pbio-0020022-Sharpless1]). In addition, mammary-directed expression of CCND1 is only mildly oncogenic ([@pbio-0020022-Wang1]), and as mentioned above, inactivation of pRb alone is not sufficient for tumorigenesis. Although the WAP promoter was a convenient means of directing mammary-specific expression for an initial assessment this model, it also presents the major shortcoming to this model in that expression of T~121~is linked to lactogenic hormone activity, as in most existing murine mammary tumor models. Future improvements aim to direct expression of T~121~ through hormone-independent methods. Finally, the advantage over wild-type T antigen models is that *WAP-T~121~* uncouples the simultaneous inactivation of pRb and p53 and permits an assessment of the relative contributions of the individual oncogenic pathways. Testing the combinatorial effects of *Rb* loss and other breast cancer mutations (e.g., *BRCA1* and *BRCA2*), along with the further characterization of *WAP-T~121~* tumors, should help provide additional insights into human breast cancer biology. Materials and Methods {#s4} ===================== {#s4a} ### Derivation and characterization of transgenic mice. {#s4a1} The 2.4 kb WAP promoter region was isolated from a WAP-TGFα construct (a gift from David Lee, University of North Carolina at Chapel Hill, United States \[[@pbio-0020022-Sandgren1]\]) and was cloned upstream of a 2.4 kb KpnI--SalI fragment of the *dl*1137′t plasmid ([@pbio-0020022-Chen1]). We targeted T~121~ expression to mammary gland using the WAP promoter, which is induced late in pregnancy and expressed during lactation ([@pbio-0020022-Pittius1]) (see [Figure 1](#pbio-0020022-g001){ref-type="fig"}). T~121~ contains the first 121 amino acids of the SV40 T antigen (see [Figure 1](#pbio-0020022-g001){ref-type="fig"}) that encodes a J domain and a pRb-binding domain, which together are sufficient to cause transformation by inactivating the pRb~f~ proteins ([@pbio-0020022-DeCaprio1]; [@pbio-0020022-Dyson2]; [@pbio-0020022-Ewen1]). Importantly, in contrast to other wild-type T antigen constructs encoding the entire SV40 early region ([@pbio-0020022-Husler1]; [@pbio-0020022-Green1]; [@pbio-0020022-Schulze-Garg1]), small t antigen expression is absent due to a deletion that removes the splice acceptor site. The importance of this is demonstrated by the recent observation that small t antigen alone is sufficient for tumorigenesis in the mammary gland ([@pbio-0020022-Goetz1]). Furthermore, p53 and EP300 (E1A-binding protein p300), which map to the carboxyl half of T antigen, are also abolished, thus permitting assessment of pRb~f~ inactivation without the confounding effects of altering additional suppressor pathways. An EcoRI fragment containing the full transgene (see [Figure 1](#pbio-0020022-g001){ref-type="fig"}) at a concentration of 4 ng/μl was injected in to fertilized eggs harvested from B6D2F1 (Jackson Laboratory, Bar Harbor, Maine, United States) mice as described previously ([@pbio-0020022-Yan1]). Transgenic mice were identified by PCR amplification of a 160 bp fragment using primers 5′-GAATCTTTGCAGCTAATGGACC-3′ and 5′-GCATCCCAGAAGCTCCAAAG-3′ with toe-derived genomic DNA as template. Cycling profile was as follows: 94°C, 2 min; 35 cycles of 94°C, 20 s; 62°C, 45 s; 72°C, 45 s; and final incubation of 72°C, 2 min. *TgWAP-T~121~* mouse lines were maintained by crossing to nontransgenic B6D2F1 mice (Jackson Laboratory) and therefore are designated as B6;D2-Tg(WAP-T~121~) Tvd. To study the effect of background differences, *WAP-T~121~* males were backcrossed to BALB/cJ (Jackson Laboratory) female mice. To increase sample size, tumor onset analysis for BALB/c background mice combined data for N6 (*n* = 6), N7 (*n* = 1), and N9 (*n* = 4) generation mice. For tumor induction, female mice, unless noted as virgin, were housed with male mice to maximize the number of pregnancies, because WAP promoter activity is lactation-dependent ([@pbio-0020022-Pittius1]). To study the effect of *p53* mutation on mammary tumorigenesis in *WAP-T~121~* mice, male *WAP-T~121^+^~* mice were mated to *p53^+/−^* females (*p53^tm1Tyj^*; Jackson Laboratory). *p53* genotypes were determined by PCR using two reactions ([@pbio-0020022-Lowe1]), one that amplifies the neomycin insertion site (neomycin primer: 5′- TCCTCGTGCTTTACGGTATC-3′, *p53* primer: 5′-TATACTCAGAGCCGGCCT-3′; 525 bp product) and a second that amplifies the endogenous *p53* allele (substituting 5′-ACAGCGTGGTGGTACCTTAT-3′ for the *neo* primer, 475 bp product). Cycling parameters were the same as the above *WAP-T~121~* reaction. We performed the cross *WAP-T~121^−^~;p53^+/−^* × *WAP-T~121^+^~;p53^+/−^,* and transgenic female mice that were *p53^+/+^*, *p53^+/−^*, or *p53^−/−^* were used for analyses while nontransgenic littermates served as controls. ### Western immunoblotting analysis. {#s4a2} Protein expression levels were assayed as previously described ([@pbio-0020022-Symonds2]). Fresh or flash-frozen tissue samples were homogenized in lysis buffer (50 mM Tris \[pH 8.0\], 5 mM EDTA, 150 mM NaCl , and 1% NP-40) using a Polytron^®^ homogenizer (Kinematica, Littau-Lucerne, Switzerland). Total protein (10 μg) was electrophoresed through a 15% polyacrylamide denaturing gel and then transferred to nitrocellulose membrane (15 V, 30 min). Alternatively, for low-expressing lines, immunoprecipitation was performed prior to electrophoresis as previously described ([@pbio-0020022-Symonds1]). The filter was preincubated in 3% bovine serum albumin, followed by incubation with primary antibody against SV40 T antigen (PAb419 at a dilution of 1:5,000; [@pbio-0020022-Harlow1]). The filter was then washed, followed by incubation at room temperature with horseradish peroxidase-conjugated goat anti-mouse IgG (Amersham Biosciences, Little Chalfont, United Kingdom). The enhanced chemiluminescence method (Amersham Biosciences) was used for autoradiography. ### Histopathology and immunohistochemistry. {#s4a3} Mammary tissue and tumor samples were dissected from *WAP-T~121~* transgenic or age- and parity-matched B6D2F1 animals. A portion of each tumor was flash-frozen in liquid nitrogen and the remaining tissue was fixed in 10% phosphate buffered formalin, embedded in paraffin, cut to a 5-μm thickness, and stained with hemotoxylin and eosin or immunostained using the Vector ABC system (Vector Laboratories, Burlingame, California, United States) for histopathological examination. Apoptosis levels were evaluated by TUNEL assay ([@pbio-0020022-Gavrieli1]) essentially as described in [@pbio-0020022-Symonds3]). ### Real-time PCR. {#s4a4} Quantitative real-time PCR analysis was performed using a TaqMan approach on DNA derived from terminal tumors to determine the status of the wild-type *p53* allele as previously described ([@pbio-0020022-Lu1]). The primers for the *p53* allele were 5′-ATGGCCATCTACAAGAAGTCACAG-3′ and 5′-ATCGGAGCAGCGCTCATG-3′. The sequence of the *p53* probe was 5′-ACATGACGGAGGTCGTGAGACGCTG-3′. The primers for the internal control *β-actin* gene were 5′-AAGAGCTATGAGCTGCCTGA-3′ and 5′-ACGGATGTCAACGTCACACT-3′. The sequence of the *β-actin* probe was 5′-CACTATTGGCAACGAGCGGTTCCG-3′. Each 25-μl reaction mixture contained 50 ng of DNA template, 18 nM *p53* primers, 80 nM *β-actin* primers, 8 nM probe, and 12.5 μl of TaqMan Universal PCR Master Mix (Applied Biosystems, Foster City, California, United States) containing AmpliTaq Gold polymerase, deoxynucleoside triphosphates, and PCR buffer. The cycling conditions were 50°C for 2 min and 95°C for 10 min for 1 cycle, and 95°C for 15 s and 60°C for 1 min for 40 cycles. The reactions were performed using an ABI 7700 Sequence Detection system (Applied Biosystems), and the data analyzed using Sequence Detector 1.7 (Applied Biosystems) and standard protocols (<http://www.appliedbiosystems.com>). The copy number of each sample was determined by calculating ΔΔCt based on the formula ΔΔC~t~ = \[sample C~t(p53)~ − sample C~t(~ *~β-actin~* ~)~\] − \[p53^+/+^ control Ct~(p53)~ − p53^+/+^ control Ct~(~ *~β-actin~* ~)~\], where C~t~ is the number of cycles required to reach a threshold based on linear amplification. Analyses of standard samples (L. Chin, Harvard University, Cambridge, Massachusetts, United States, personal communication) indicate copy numbers of 2, 1, and 0 are indicated by 2^-ΔCtn^ values of greater than 0.6, 0.15--0.6, and less than 0.15, respectively. Standard samples analyzed along with experimental samples confirmed the accuracy of these assignments. ### Statistical analyses. {#s4a5} Kaplan--Meier survival analysis was used to determine median tumor latencies (StatsDirect, Camcode, Sale, United Kingdom), and the Log-Rank (Peto, StatsDirect) test was performed to evaluate significance. The equivalence of tumor morphology distributions was tested using the Fisher--Freeman--Halton\'s exact test. ### CGH. {#s4a6} Genomic DNA was extracted from end-stage tumors (1 cm in diameter) or tails using a DNeasy genomic tip (Qiagen, Valencia, California, United States) and further purified by proteinase K digestion followed by phenol/chloroform extraction, ethanol precipitation, and resuspension in sterile H~2~O. cCGH was performed as described in [@pbio-0020022-Kallioniemi1]), [@pbio-0020022-Donehower1]), and [@pbio-0020022-Lu1]). aCGH was performed as described in [@pbio-0020022-Snijders1]). For both methods, genomic DNA from tumor and normal tissue was labeled with different fluorochromes and then cohybridized together with *Cot-1* DNA to either normal metaphase chromosomes from cultured cells (cCGH) or microarrayed BAC clones containing mouse genomic DNA (aCGH). Nonequivalent fluorescence intensities indicate relative imbalances of genomic DNA. Aneuploidy and partial chromosome gains and losses are detectable by cCGH with approximately 10 Mb resolution. Graphical output of cCGH data was generated using the National Cancer Institute and National Center for Biotechnology Information Spectral Karyotyping *SKY* and Comparative Genomic Hybridization CGH Database (<http://www.ncbi.nlm.nih.gov/sky/skyweb.cgi>). For aCGH, approximately 1,500 BAC clones span the entire mouse genome with 2--20 Mb spacing. Tumor DNA and normal DNA were sex-mismatched; thus, the X Chromosome served as an internal control, while normal tail DNA was used as a negative control. Gains or losses were scored based on tumor:normal fluorescence ratios that were greater than 1.25 or less than 0.75, respectively. Supporting Information {#s5} ====================== {#s5a} ### Accession Numbers {#s5a1} The accession numbers for the genes and gene products discussed in this paper are *Brca1* (LocusLink ID 12189), CDK2 (LocusLink ID 1017), CDK4 (LocusLink ID 1019), CDK6 (LocusLink ID 1021), c*-myc* (LocusLink ID 17869), cyclin D1 (LocusLink ID 595), cyclin E (LocusLink ID 898), E2F (InterPro ID IPR003316), HER2/*neu* (LocusLink ID 13866), histone deacetylase (LocusLink ID 3065), p16^INK4a^ (LocusLink ID 1029), p53 (LocusLink ID 7157), p107 (LocusLink ID 5933), p130 (LocusLink ID 5934), p300 (LocusLink ID 2033), PCNA (LocusLink ID 18538), pRb (LocusLink ID 5925), *PTEN* (LocusLink ID 5728), v*-Ha-ras* (LocusLink ID 3265), WAP (LocusLink ID 22373), and *Wnt-1* (LocusLink ID 22408). These databases may be found at [www.ncbi.nlm.nih.gov/LocusLink/](www.ncbi.nlm.nih.gov/LocusLink/) (LocusLink), and [www.ebi.ac.uk/InterPro/](www.ebi.ac.uk/InterPro/) (InterPro). The authors thank Lisa Livanos at the University of North Carolina (UNC) chromosome imaging core facility for assistance with cCGH; Joe McCarville for the assembly of the *WAP-T~121~* construct; Drew Fogarty and Anne Wolthusen for expert technical assistance; the UNC histopathology core facility; the UNC Division of Laboratory Animal Medicine; Bing Huey for aCGH; and Karl Sirotkin and Vasuki Gobu at the National Center for Biotechnology Information for their assistance with cCGH data submission. We also thank Xiangdong Lu and Dale Cowley for their critical reading of the manuscript. KS was supported by a National Cancer Institute postdoctoral training grant (T32CA09156) to the Lineberger Comprehensive Cancer Center. This work was also supported by grants from the Susan G. Komen Breast Cancer Foundation (BCTR98), the United States Army (DAMD17--99-1--9332), and two grants from the National Cancer Institute, to TVD (5 R01 CA46283--15) and to DA (CA84118). **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** KS, HW, and TVD conceived and designed the experiments. KS, HW, and LL performed the experiments. KS, HW, RDC, and TVD analyzed the data. DP and DA contributed reagents/materials/analysis tools. KS and TVD wrote the paper. Academic Editor: Chris Marshall, Institute for Cancer Research ¤Current address: Phenomix Corporation, La Jolla, California, United States of America aCGH : microarray comparative genomic hybridization BAC : bacterial artificial chromosome Brca1 : breast cancer 1 cCGH : chromosome comparative genomic hybridization CCND1 : cyclin D1 CCNE1 : cyclin E1 CDK2 : cyclin-dependent kinase 2 CDK4 : cyclin-dependent kinase 4 CDK6 : cyclin-dependent kinase 6 CGH : comparative genomic hybridization c*-myc* : myelocytomatosis oncogene LOH : loss of heterozygosity MMTV : murine mammary tumor virus p16^INK4a^ : cyclin-dependent kinase inhibitor 2A PCNA : proliferating cell nuclear antigen pRb : retinoblastoma 1 pRb~f~ : retinoblastoma protein family PTEN : phosphatase and tensin homolog SV40 : simian virus 40 v-*erb-b2* : erythroblastic leukemia viral oncogene homolog 2 v*-Ha-ras* : Harvey rat sarcoma viral oncogene WAP : whey acidic protein Wnt-1 : *wingless*-related MMTV integration site 1

PubMed Central

2024-06-05T03:55:46.061513

2004-2-17

PMC340939

Nobody questions the importance of good scientific information for sound environmental decision-making. But designing mechanisms to link scientific research to the decision-making process is no easy matter. Research and decision-making often seem to operate in different worlds. Policy-makers\' needs for applied findings and best judgment typically clash with scientists\' pursuit of basic research and statistical significance. Despite this challenge, as needs for scientific input into decision-making arise, a number of institutions have been established to help bridge the science--policy divide. Regulatory agencies like the United States\' Food and Drug Administration have met this need in areas of public health, for example, and Environmental Impact Assessment procedures have helped to introduce better science into project-level decisions. At the scale of global environmental challenges, highly regarded mechanisms have been established such as the Scientific Assessment of Ozone Depletion, which has guided decisions by governments, the private sector, and nongovernmental organizations (NGOs), and the Intergovernmental Panel on Climate Change, which has become the authoritative source of policy-relevant information on climate science. But a significant gap remains in the landscape of institutions designed to link science with policy-making: no mechanism has existed to provide decision-makers with authoritative information on the causes and consequences of changes in the planet\'s ecosystems and on the options for response. Human well-being and progress toward sustainable development are vitally dependent upon improving the management of Earth\'s ecosystems to ensure their conservation and sustainable use. The benefits that human beings extract from nature are the foundation of all economies and the basis of major industries, are sources of knowledge, and are central to many cultures. While demands for ecosystem services such as food and clean water are growing, human actions are at the same time diminishing the capability of many ecosystems to meet these demands. And while many of the changes to ecosystems, such as increased agricultural production, have greatly enhanced human well-being, many others have not. World fisheries are now declining due to overfishing, for instance, and some 40% of agricultural land has been degraded in the past half-century. Other human-induced impacts on ecosystems include alteration of the nitrogen, phosphorous, sulfur, and carbon cycles, causing acid rain, algal blooms, and fish kills in rivers and coastal waters, along with contributions to climate change. > *The benefits that human beings extract from nature are the foundation of all economies and the basis of major industries, are sources of knowledge, and are central to many cultures.* Recognizing the growing scale of these problems, United Nations Secretary General Kofi Annan, in his 2000 Millennium Report to the General Assembly, called for a Millennium Assessment of Global Ecosystems to provide definitive information on the consequences of ecosystem change for human well-being. With further authorization received through three international conventions (on Biological Diversity, Desertification, and Wetlands) and with financial support from the Global Environment Facility, the United Nations Foundation, the David and Lucile Packard Foundation, and the World Bank, the Millennium Ecosystem Assessment (MA) ([www.millenniumassessment.org](www.millenniumassessment.org)) was launched one year later, in 2001. More than 700 authors from 80 countries are now involved in the expert working groups preparing the global assessment; 100 experts will serve on the Editorial Review Board; more than 1,000 experts will be asked to review the materials, and hundreds more are undertaking subglobal assessments as part of the MA. The MA focuses on ecosystem services (the benefits people obtain from ecosystems), how changes in ecosystem services have affected human well-being, how ecosystem changes may affect people in future decades, and response options that might be adopted at local, national, or global scales to improve ecosystem management and thereby contribute to human well-being and poverty alleviation ([Figure 1](#pbio-0020027-g001){ref-type="fig"}). ::: {#pbio-0020027-g001 .fig} Figure 1 ::: {.caption} ###### MA Conceptual Framework The MA examines both indirect and direct drivers (both human-caused and natural) of change in ecosystems, how those changes affect ecosystem services, how those changes, in turn, influence human well-being and poverty reduction, and opportunities for interventions that can ensure ecosystem conservation and enhance human well-being. The assessment must take into consideration the multiple time and spatial scales over which these interactions take place. (Schematic is used by permission from the Millennium Ecosystem Assessment \[2003\] and published under the terms of the Creative Commons Attribution License.) :::  ::: The first report---*Ecosystems and Human Well-Being: A Framework for Assessment*---was published in 2003 and describes the approach and methods used in the MA. The four main assessment volumes---Conditions and Trends, Scenarios, Response Options, and Subglobal Assessments---began the first of two rounds of peer review in January 2004, and the final assessment reports will be published in early 2005. Unlike previous global scientific assessments, the MA is a "multiscale" assessment. Assessments at subglobal scales are needed because ecosystems are highly differentiated in space and time and because sound management requires careful local planning and action. Local assessments alone are insufficient, however, because some processes are global and because local goods, services, matter, and energy are often transferred across regions. The MA subglobal assessments will directly meet needs of decision-makers at the scale at which they are undertaken, strengthen the global findings with on-the-ground reality, and reinforce the local findings with global perspectives, data, and models. In Southern Africa, for example, a series of community-level assessments are being conducted using the MA conceptual framework. The findings from these assessments inform, and are informed by, assessments underway in the Gariep and Zambezi river basins. These local and river basin assessments, in turn, are linked to a regional assessment encompassing the countries in the Southern African Development Community. Other subglobal assessments are now underway in such regions as São Paulo, Brazil; coastal British Columbia, Canada; the Caribbean Sea; western China; Colombia; the Sinai Peninsula, Egypt; several regions within India; Indonesia; the Laguna Lake Basin, the Philippines; Portugal; and Sweden. The ultimate impact of the MA will depend on its credibility, legitimacy, and utility. To ensure its scientific credibility, the assessment involves leading scientists from around the world and has established an independent peer-review process. To ensure the political legitimacy, all of the stakeholders---governments, the private sector, and NGOs---have a role in governing the process, and governments themselves have approved the process through decisions in international conventions. And, to ensure its utility, ongoing interactions with stakeholders are designed to ensure a focus on their questions and issues. The scientific information now available concerning ecosystems and human development holds the promise of significantly improving the choices that the public and decision-makers take concerning the environment. But for that promise to be fulfilled, a bridge needs to be built between the research community holding this information and the decision-makers seeking it. The MA is an attempt to establish that bridge. Walter V. Reid is the director of the Millennium Ecosystem Assessment, located in Penang, Malaysia. E-mail: <reid@millenniumassessment.org> MA : Millennium Ecosystem Assessment NGO : nongovernmental organization

PubMed Central

2024-06-05T03:55:46.065947

2004-2-17

PMC340940

Introduction {#s1} ============ Because of the abundant availability of mouse and human genome data ([@pbio-0020029-International1]; [@pbio-0020029-Mouse1]), it has come to light that mutation rates vary widely across different regions of the human genome ([@pbio-0020029-Matassi1]; [@pbio-0020029-Mouse1]; [@pbio-0020029-Hardison1]), in agreement with a number of smaller-scale studies ([@pbio-0020029-Wolfe1]; [@pbio-0020029-Casane1]; [@pbio-0020029-Perry1]). Regions of unusually high or low substitution rates have been observed from 4-fold sites and ancestral repeat sequences, two of the best candidates for measuring neutral rates of mutation in mammals ([@pbio-0020029-Sharp2]; [@pbio-0020029-Mouse1]; [@pbio-0020029-Hardison1]). The reasons for such regional variability are unclear, since structural characterizations of the mutation rate are nascent. Whatever the reason for these hot and cold regions, their existence suggests a question that has intriguing consequences for molecular evolution: does the organism take advantage of these hot and cold spots? One way to take advantage of a hot region would be to place genes there for which the hotness is useful---an intuitive example would be receptor proteins, which must respond to a constantly changing ligand set. Similarly, it could be beneficial to place delicate genes in a cold region, to reduce the possibility of deleterious mutations. These potential advantages offer the possibility that regional mutation rates affect the spatial organization of genes. The idea of such organization in mouse and human is bolstered by recent findings of gene organization in yeast. For example, [@pbio-0020029-Pal1]) showed that yeast genes are organized to take advantage of local recombination rates, which is particularly relevant since mutation rate and recombination rate are known to be correlated ([@pbio-0020029-Lercher1]). If the local mutation rate---equivalent to the synonymous (amino acid preserving) substitution rate *K~S~* if synonymous substitutions are neutral---affects gene organization, this would constitute a type of selection complementary to traditional selection on point mutations ([@pbio-0020029-Graur1]). We studied whether local mutation rates affect gene locations by measuring the mutation rates of genes and their organization in the human genome. First, we analyzed the substitution rates of the genes in each of the families defined by the Gene Ontology (GO) Consortium ([@pbio-0020029-Ashburner1]). If the organism is taking advantage of varying *K~S~*, gene families should be biased toward regions of appropriate rate. In fact, we observe that several functional classes of genes preferentially occur in hot or cold regions. Some of the notable hot categories we observe are olfactory genes, cell adhesion genes, and immune response genes, while the cold categories are biased toward regulatory proteins such as those involved in transcription regulation, DNA/RNA binding, and protein modification. Also, to better characterize the hot and cold regions, we measured the length scale over which substitution rates vary. While rough limits on the size of hot and cold regions are known ([@pbio-0020029-Matassi1]; [@pbio-0020029-Hardison1]), this paper presents the first known quantitative calculation of their length scale. Because mutation rates are regional, mutation rates in genes categories could be influenced by events altering the organization of genes in the genome, such as gene relocation or gene duplication. We therefore analyzed mechanisms by which functional categories of genes may have become concentrated in hot or cold regions. A clustering analysis reveals that the hotness of some categories is enhanced by local gene duplications in hot regions. However, there are strong functional similarities among the hot categories---both clustered and unclustered---as well as among the cold categories. These functional similarities imply that the instances of duplicated categories are not random; i.e., selection may have affected which genes have duplicated and persisted. Results {#s2} ======= Mutation Rates Have Regional Biases {#s2a} ----------------------------------- Recently, substitution rates between Mus musculus and Homo sapiens have been measured by several groups on a genome-wide scale ([@pbio-0020029-Kumar1]; [@pbio-0020029-Mouse1]; [@pbio-0020029-Hardison1]). These substitution rates vary significantly across the genome ([@pbio-0020029-Mouse1]; [@pbio-0020029-Hardison1]), suggesting that neutral mutation rates may have regional biases as well. A popular proxy for neutral mutation rates is the substitution rate at 4-fold sites (a recent example is found in [@pbio-0020029-Kumar1]), base positions in coding DNA that do not affect protein sequence and that should hence be under less selective pressure than other sites. The 4-fold sites also offer the advantage of being easily alignable. For these reasons, we estimated the neutral mutation rate from substitution rates at 4-fold sites (which we use interchangeably with the term *K~S~* in this paper). This identification is not without complexities, however, since there are processes that can in principle selectively affect the 4-fold sites. For example, some have argued that exogenous factors such as isochore structure influence the silent sites ([@pbio-0020029-Bernardi1]), and codon usage adaptation has been shown to affect silent sites in bacteria and yeast ([@pbio-0020029-Sharp1]; [@pbio-0020029-Percudani1]). So far, such selective effects have been difficult to detect in mammals ([@pbio-0020029-Smith1]; [@pbio-0020029-Duret1]; [@pbio-0020029-Iida1]; [@pbio-0020029-Kanaya1]). Recently, [@pbio-0020029-Hardison1]) showed that several functionally unrelated measures of mutation rate, including SNP density, substitutions in ancestral repeats, and substitutions in 4-fold sites, are correlated in genome-wide mouse--human comparisons---suggesting that these measures have common neutral aspects. We constructed our own dataset of the 4-fold substitution rates for 14,790 mouse/human orthologous genes, using data from the ENSEMBL consortium. In order to properly account for stochastic finite-size effects, we mapped the observed substitution rates to a normalized value, based on the assumption that all 4-fold sites mutate at the same rate (see [Materials and Methods](#s4){ref-type="sec"}). Under this assumption, it was expected that the normalized substitution rates would follow the normal distribution (a Gaussian distribution with σ = 1). Contrary to these expectations, the distribution of ortholog substitution rates was found to be highly biased toward high or low rates, indicating that 4-fold mutation rates vary substantially by location and on a scale larger than the typical size of a gene. [Figure 1](#pbio-0020029-g001){ref-type="fig"} shows the distribution of substitution rates for all mouse/human orthologs. The observed distribution has excesses of genes at both high and low substitution rates. These results are in agreement with the findings of [@pbio-0020029-Matassi1]), who reported significant mutation rate correlations between neighboring genes. This is not a compositional effect---the distribution remained the same even when corrections for the gene\'s human base composition were made (see [Materials and Methods](#s4){ref-type="sec"}). We further verified that substitution rates of neighboring genes were correlated using an analysis qualitatively similar to [@pbio-0020029-Matassi1])---though with approximately 20 times more orthologs---finding that gene substitution rates are correlated with their neighbors with a *p*-value of 10^−189^ (see [Materials and Methods](#s4){ref-type="sec"}). These results imply that substitution rates have regional biases, acting both within a gene and over longer length scales. ::: {#pbio-0020029-g001 .fig} Figure 1 ::: {.caption} ###### Distribution of Normalized Substitution Rates Histogram of substitution rates based on 14,790 orthologous mouse and human genes (black curve). The rate distribution has significantly more genes at high and low rates than the expected Normal distribution (red curve). This bias toward high and low rates remains even when rates are corrected for human base composition (green curve). :::  ::: Some Gene Categories Are Biased toward Hot or Cold Regions {#s2b} ---------------------------------------------------------- We next considered whether there is a relationship between gene locations and their functions, i.e., whether functional categories of genes have biases for being in regions of particular mutation rate. To test whether such biases exist, we performed an analysis of the GO assignments for each ortholog pair ([@pbio-0020029-Ashburner1]), using data from the ENSEMBL human ENSMART database to assign genes to GO categories. For each GO category, we calculated a *z*-score to measure the overall substitution rate, based on the substitution rates of the genes in the category (see [Materials and Methods](#s4){ref-type="sec"}). The 21 GO categories having statistically significant positive values of *z* are shown in [Table 1](#pbio-0020029-t001){ref-type="table"}. In terms of 4-fold substitution rates, the hot category rate averages were found to range from 0.346 (integral to membrane) to 0.468 (internalization receptor activity), while the genome-wide average was 0.337 (with a genewise standard deviation of 0.08). For a category with several genes, the effective standard deviation is much smaller, equal to 0.08/ , where *N~GO~* where *N~GO~* is the number of genes in the category, so these rate biases are extremely significant. Hot gene categories were focused mainly in receptor-type functions, along with a few other categories such as "proteolysis" and "microtubule motor activity." Some preferences were partially because categories have genes in common; e.g., eight genes are shared among the categories "dynein ATPase activity," "dynein complex," and "microtubule-based movement." However, there were several categories of similar function that were independent; e.g., "membrane" and "olfactory receptor activity" shared no genes, and "cell adhesion" and "immune response" shared only 5% of their genes. Overall, there was a clear bias for the larger hot categories to contain receptor-type proteins: e.g., "receptor activity," "olfactory receptor activity," "G-protein coupled receptor protein signaling pathway," "membrane," and "immune response." For the set of all 1,488 genes where the string "receptor" is part of the GO description, the average 4-fold substitution rate was found to be 0.347. The probability that a random set of 1,488 genes would have an average rate this high is 10^−6^. ::: {#pbio-0020029-t001 .table-wrap} Table 1 ::: {.caption} ###### Statistically Significant Hot GO Categories :::  Listed are the categories with *z* \> 0 having at least five genes and *p~z~* ≤ 10^−3^, sorted by statistical significance (−log~10~ *p~z~*). There is a bias toward proteins involved in extracellular communication. Several of the categories have an unusual number of clustered genes (−log~10~ *p~cluster~* \>3) ::: The 36 statistically significant GO categories with negative *z* scores, are shown in [Table 2](#pbio-0020029-t002){ref-type="table"}. The 4-fold rate averages for the cold categories ranged from 0.220 ("mRNA binding activity") to 0.326 ("protein serine/threonine kinase activity"). The coldest gene categories included "nuclear proteins," "transcription regulation," "DNA and RNA binding," "oncogenesis," "phosphatases," and "kinases," all of which are important to regulatory processes. Many of these genes are also housekeeping genes ([@pbio-0020029-Hsiao1]). For the set of all 1,704 genes where the string "regulat" is part of the GO description, the average 4-fold substitution rate was found to be 0.325. The probability that a random set of 1,704 genes would have an average rate this low is 10^−9^. ::: {#pbio-0020029-t002 .table-wrap} Table 2 ::: {.caption} ###### Statistically Significant Cold GO Categories :::  Listed are the categories with z \< 0 having at least five genes and *p~z~* ≤ 10^−3^, sorted by statistical significance. There is a bias toward proteins involved in DNA, RNA, or protein regulation. None of the cold categories have statistically significant clustering ::: We repeated our *z*-score classifications using several other measures of mutation rate and in each case inferred similar hot and cold categories. For example, under the normalized rate model that accounts for human base composition, the same set of 23 hot categories were found. Of the 37 cold categories, 33 remained classified as cold. The four lost were "regulation of transcription from Pol II promoter," "development," "neurogenesis," and "translation regulator activity." There were six new categories, and these were also largely regulatory: "nucleic acid binding activity," "translation initiation factor activity," "ubiquitin C-terminal hydrolase activity," "collagen," "RNA processing," and "negative regulation of transcription." We also calculated several maximum likelihood (ML) measures of *K~S~* using mutation models in the Phylogenetic Analysis by Maximum Likelihood (PAML) package ([@pbio-0020029-Yang1]), including the [@pbio-0020029-Nei1]) codon-based measure and the TN93 ([@pbio-0020029-Tamura1]) and REV ([@pbio-0020029-Tavere1]) models. We again found qualitatively similar sets of hot and cold categories---receptor genes at high substitution rates and regulatory genes at low substitution rates---though there were changes in the numbers of significant categories. For example, for the TN93 model, we observed ten hot categories---"induction of apoptosis by extracellular signals," "G-protein coupled receptor protein signaling pathway," "olfactory receptor activity," "receptor activity," "apoptosis," "enzyme activity," "chymotrypsin activity," "trypsin activity," "integral to membrane," and "dynein ATPase activity"---and eight cold categories: "calcium-dependent protein serine/threonine phosphatase activity," "ribonucleoprotein complex," "protein serine/threonine kinase activity," "RNA binding activity," "protein amino acid dephosphorylation," "intracellular protein transport," "protein transporter activity," and "nucleus." The categories inferred from our original *z*-score analysis are probably more accurate than those from ML methods, because ML methods tend to produce strong outliers at high substitution rate, skewing calculations of the variance in the *z*-score analysis. Can Gene Duplications Explain the Hot and Cold Categories? {#s2c} ---------------------------------------------------------- Given the existence of hot and cold gene categories, the question then becomes: why do these biases exist? One potentially nonselective factor that could affect category rate biases is local gene duplications. New genes generally arise by duplication, in which a new copy of a gene is generated nearby to the preexisting gene by a recombinatorial event such as unequal crossing-over, followed by evolution to a novel, but often related function ([@pbio-0020029-Graur1]). Such local duplications can cause many genes with similar function to be clustered together. Because there are regional biases in mutation rate (discussed in the section on Block Structure of the Substitution Rate), these functionally related genes will tend to have similar mutation rates. GO categories containing these genes will then be biased toward the mutation rate of the region surrounding the genes. We tested the effect of gene duplications on category rates through a clustering analysis (see [Materials and Methods](#s4){ref-type="sec"}). If gene duplications are not important to category rates, genes in a hot (cold) gene category would be expected to be distributed randomly throughout the many hot (cold) regions around the genome; i.e., clustering of genes would be weak. However, if gene duplications are relevant, we would expect hot (cold) genes of the same category to be tightly clustered since many of these genes would have arisen by local duplications. We therefore studied the location distribution of each of the gene categories and analyzed the significance of its clustering, using the short-range correlation length τ ∼ 10^6^ basepairs (see the section on Block Structure of the Substitution Rate) as a defining length scale. This analysis was similar to that of [@pbio-0020029-Williams1]), who studied clustering of tissue-specific genes, though we analyzed a larger number of more narrowly defined gene families. We found that some of the hot gene categories were indeed clustered, but that none of the cold gene categories were. The results of the clustering for the hot and cold categories are displayed in Tables [1](#pbio-0020029-t001){ref-type="table"} and [2](#pbio-0020029-t002){ref-type="table"}, with the clustering p-values shown via their −log~10~ values. Of the 21 statistically significant hot categories, ten categories had statistically significant clustering (−log~10~ *p~cluster~* \> 3). For example, the "olfactory receptor activity" category has 223 genes, with a randomly expected number of clustered genes equal to 30.6. The actual number of clustered genes was found to be 190, which has a *p*-value of less than 10^−16^. In the set of 37 cold gene GO categories, none had statistically significant clustering. The clustering significance is plotted versus the substitution score *z* for all the GO categories with at least five members in [Figure 2](#pbio-0020029-g002){ref-type="fig"}. There were many categories of hot genes with significant clustering (−log~10~ *p~cluster~* \> 3), but virtually no cold ones. ::: {#pbio-0020029-g002 .fig} Figure 2 ::: {.caption} ###### Clustering versus Substitution Rate for GO Categories Containing at Least Five Members Virtually all clustered gene categories have higher than average substitution rates (*z* \> 0). :::  ::: As an example of clustering in the hot gene categories, we considered the olfactory receptors. It is well-established that olfactory receptors occur in clusters throughout the human genome ([@pbio-0020029-Rouquier1]), and we likewise observed the olfactory receptors to be highly clustered in three regions near the head, middle, and tail of Chromosome 11 ([Figure 3](#pbio-0020029-g003){ref-type="fig"}). The central cluster is displayed in [Figure 4](#pbio-0020029-g004){ref-type="fig"}. This clustering provided evidence that local gene duplications have influenced the high category rate of the olfactory genes. ::: {#pbio-0020029-g003 .fig} Figure 3 ::: {.caption} ###### Clustering of Olfactory Genes on Human Chromosome 11 The olfactory genes are clustered into three regions along the chromosome. The substitution rates of the olfactory genes are almost all hot, while the nonolfactory genes are distributed around *r* = 0. :::  ::: ::: {#pbio-0020029-g004 .fig} Figure 4 ::: {.caption} ###### Olfactory Genes Lie in a Mutational Hot Spot Substitution rates of the olfactory genes in the central region of human Chromosome 11. The substitution rate of ancestral repeat sequences is higher in the region where the olfactory genes lie. :::  ::: We next attempted to determine whether the high olfactory rates are due to a regional bias. The substitution rates of all genes are plotted in [Figure 4](#pbio-0020029-g004){ref-type="fig"}, with the olfactory genes in red. As expected, the olfactory genes exhibited an obvious bias for higher substitution rates than other genes. We next calculated the mutation rate of the region as determined from an independent measure, the substitution rates between ancestral repeat sequences (green curve in [Figure 4](#pbio-0020029-g004){ref-type="fig"}), using data published by [@pbio-0020029-Hardison1]) (see [Materials and Methods](#s4){ref-type="sec"}). The repeat sequence mutation rate was notably higher in the regions where the olfactory genes occur, showing that the hotness of the olfactory genes is a regional property and not specific to the genes. Similar clustering and regional hotness were observed for other hot gene categories. We plot the substitution rates of a cluster of homophilic cell adhesion genes on Chromosome 5 in [Figure 5](#pbio-0020029-g005){ref-type="fig"}, along with the rates of nearby genes and the ancestral repeat sequence substitution rates. The same features observed for the olfactory genes were also present for the cell adhesion genes: clustering, high substitution rates, and an elevated ancestral repeat substitution rate. The repeat substitution rate exhibited a plateau-like behavior over the region defined by the homophilic cell adhesion genes. These factors support the interpretation that significant numbers of hot genes have arisen by duplications in inherently hot regions of the genome. ::: {#pbio-0020029-g005 .fig} Figure 5 ::: {.caption} ###### Homophilic Cell Adhesion Genes Also Lie in a Hot Spot Substitution rates of a cluster of homophilic cell adhesion genes on human Chromosome 5, along with substitution rates of other genes and ancestral repeat sequences. The repeat sequence substitution rate plateaus at a higher level in this region. :::  ::: Block Structure of the Substitution Rate {#s2d} ---------------------------------------- Several explanations have been proposed that could account for the regional biases in mutation rate ([@pbio-0020029-Mouse1]), including recombination-associated mutagenesis ([@pbio-0020029-Perry1]; [@pbio-0020029-Lercher1]), strand asymmetry in mutation rates ([@pbio-0020029-Francino1]), and inhomogeneous timing of DNA replication ([@pbio-0020029-Wolfe1]; [@pbio-0020029-Gu1]). The structure of regional biases could be considered from the perspective of amino acid changing substitutions as well, since linked proteins have been known to have similar substitution rates ([@pbio-0020029-Williams1]). However, the silent sites may be easier to comprehend, since protein sequences are more likely to be complicated by nonneutral pressures. To shed light on the structural properties of the hot and cold mutational regions, we measured the length scale over which substitution rates are correlated. Previously, correlations have been observed in blocks of particular physical (5 Mb) ([@pbio-0020029-Hardison1]) or genetic (1, 2, 5, and 200 cM) ([@pbio-0020029-Matassi1]; [@pbio-0020029-Lercher2]) size. While these studies have focused on whether correlations exist at certain length scales, it is informative to measure the decay of correlations with distance. We therefore measured the length scale of substitution rate correlation, using an analysis of the correlation function ([@pbio-0020029-Huang1]) where *r*(*t*) is the substitution rate of a gene *t* basepairs downstream of a gene with substitution rate *r*(0), and \<...\> indicates an average over the available data (see [Materials and Methods](#s4){ref-type="sec"}). We expect that at small *t*, the correlation function will be positive and then decrease with *t* as rates become decoupled. The length scale over which this decay occurs serves as a measure of the typical size of hot or cold regions. The rate correlation function is plotted in [Figure 6](#pbio-0020029-g006){ref-type="fig"} versus both the human and mouse values for *t*. ::: {#pbio-0020029-g006 .fig} Figure 6 ::: {.caption} ###### Correlation Length Analysis of Substitution Rates Correlation of substitution rates in syntenous blocks as a function of distance between genes measured along the human chromosome (top) and measured along the mouse chromosome (bottom). There are two length scales of correlation decay: a short one of 1 Mb and a long one of 10 Mb. The curve fits are for \<*r*(0)*r*(*t*)\> = *A* ~0~ exp (−*t/τ*) + *A* ~∞~ for the region *t* ∈ \[0, 10000000\] . :::  ::: We observed two notable behaviors: (1) a strong correlation that decays over a region of approximately 1 Mb, and (2) a longer range correlation which plateaus over a region of approximately 10 Mb. At larger distances, correlations are weaker. For example, the human curve first dips below the \<*r*(0)*r*(*t*)\> = 0 threshold at approximately 11 Mb, and the mouse curve first crosses it at approximately 9 Mb. This suggests that there are multiple phenomena that control the mutation rate of regions, both long (10 Mb) and short (1 Mb) length scale. We also measured the characteristic short-range correlation length using an exponential fit. The correlation length τ was determined by fitting the data to the functional form where *A* ~∞~ is the correlation at long distances and (*A* ~0~+ *A* ~∞~) is the correlation at zero distance. Because of the observed plateauing behavior of the data, we performed our curve fit over the region *t* ∈ \[0, 10000000\] . For the human data, we obtained *A* ~0~ = 0.83, τ = 1.21 × 10^6^, *A* ~∞~ = 0.39. For the mouse data, we found values of a similar magnitude (*A* ~0~ = 1.08, τ = 0.73 × 10^6^, *A* ~∞~ = 0.32), suggesting that short-range mutational processes may be alike in mouse and human. The long-range correlation *A* ~∞~ was at least an order of magnitude larger than would be expected by chance at all distances up to 10 Mb (see [Materials and Methods](#s4){ref-type="sec"}). It is unclear what factors are responsible for these two length scales of rate correlation, though some guesses are possible. For the short-range effect, one process that occurs on the appropriate length scale is DNA replication ([@pbio-0020029-Alberts1]). Replication origins in a concerted unit activate under similar timing and similar cell conditions and could have a common regulatory mechanism, making it a reasonable to expect the DNA in such a unit to have similar mutation rates. Long-range correlations have previously been observed at chromosomal-size distances in particular regions of the genome; e.g., it is known that Chromosome 19 is generally hotter than other chromosomes ([@pbio-0020029-Lercher2]; [@pbio-0020029-Castresana1]). However, the 10 Mb correlation was not simply due to selection on chromosomes. We removed the respective chromosomal average from each substitution rate and repeated the correlation analysis, finding that *A* ~∞~ retained a significant value of approximately 0.2. One possible mechanistic explanation for the long-range correlation is suggested by the finding of [@pbio-0020029-Lercher1]) that recombination rate and substitution rate are correlated even in blocks extending to 30 Mb. Therefore, if large regions of similar recombination rate exist, they could be related to the long-range 4-fold correlation effects we observed. Discussion {#s3} ========== Evidence for Selection {#s3a} ---------------------- Recently, there has been evidence for selective factors influencing gene location in yeast ([@pbio-0020029-Pal1]). This suggests the possibility that similar phenomena affect gene locations in mouse/human as well. We therefore considered whether regional mutation rates could have selectively influenced the types of genes occurring in different loci in the genome. Selection due to the local mutation rate would require different mechanisms than that observable through the traditional measure *K~A~/K~S~*, which quantifies selection on point mutations. For example, regional mutation rates could have influenced the fitness of the genome after events that cause gene relocation, such as gene transposition or chromosomal recombination. Or perhaps the duplication of certain genes provided a fitness benefit (a mechanism possibly relevant for the hot clustered categories). Differential duplication rates could force a category to have a mutation rate bias, due to the block structure of the mutation rate and the fact that duplications occur locally. The observed categories of hot and cold genes suggest gene locations have been selectively influenced by regional mutation rates. This is because if mutation rates were unrelated to gene function, then the lists of hot and cold categories would be expected to be random; i.e., the lists shown in [Tables 1](#pbio-0020029-t001){ref-type="table"} and [2](#pbio-0020029-t002){ref-type="table"} would have been evenly sampled from all possible GO categories. However, this was not the case, as the hot and cold categories each had strong internal commonalities. The hot categories were found to be biased toward receptor activities or roles in extracellular communication. Intriguingly, arguments based on protein-level effects appear applicable to the silent-site hotness of these categories. Cellular receptors and those involved in extracellular communication are the proteins that most directly interact with the environment and are therefore the most likely to have experienced a dynamically changing set of selection pressures. This variability of selection pressures would have made it favorable for them to be in hot regions, in order that new mutations be possible to deal with new stimuli. Examples of hot categories with known protein-level diversification pressures include the olfactory receptors ([@pbio-0020029-Lane1]), immune genes ([@pbio-0020029-Papavasiliou1]), and cell adhesion genes ([@pbio-0020029-Uemura1]; [@pbio-0020029-Tasic1]). Arguments normally applied to protein-level selection were found to be appropriate for cold mutation rate categories as well. Cold categories were often related to transcription or other regulatory processes. Regulatory proteins should be tuned to interact with many different nucleic acid or protein targets, in contrast with receptor proteins, which typically interact with only a particular ligand. Mutations to regulatory proteins would therefore be expected to be more deleterious, and hence it would be beneficial for them to have low mutation rates. Strong conservation pressures in the cold categories could also be related to their roles as housekeeping genes ([@pbio-0020029-Zhang1]) or as essential genes. For example, in the dataset of [@pbio-0020029-Winzeler1]), 81 out of 356 essential yeast genes were involved in transcription, whereas only four were involved in signal transduction, the function most similar to extracellular communication for which data were available. The applicability of protein-level arguments to synonymous mutation rates suggests that *K~S~* and *K~A~* are under similar pressures. This is consistent with what would be expected if gene locations have evolved to make use of the block structure of the mutation rate, since relocation to a hot (cold) spot would increase propensities for both high (low) *K~A~* and *K~S~*. More quantitatively, we observed that *K~S~* category biases were similar to category biases caused by selection on amino acid changing point substitutions---i.e., selection observable through the ratio *K~A~/K~S~*. We performed a GO *z*-score analysis on *K~A~/K~S~* (for consistency, the CODEML method in PAML was used to calculate both *K~A~* and *K~S~*). There were eight hot categories common to both the 4-fold and *K~A~/K~S~* classifications ("immune response," "proteolysis receptor activity," "peptidolysis receptor activity," "integral to membrane," "chymotrypsin activity," "cell adhesion," "trypsin activity," "olfactory receptor activity") and 17 common cold categories ("nucleus," "regulation of transcription," "transcription factor activity," "RNA binding activity," "development," "ribonucleoprotein complex," "protein transporter activity," "protein serine/threonine kinase activity," "ubiquitin conjugating enzyme activity," "GTP binding activity," "ubiquitin-dependent protein catabolism," "translation regulator activity," "intracellular protein transport," "neurogenesis," "ubiquitin cycle," "cytoplasm," "regulation of transcription from Pol II promoter"). The strong commonalities between the two types of classification suggest that the selective forces that influenced amino acid changing point mutations also influenced gene locations. The hot and cold categories derived from *K~A~/K~S~* are available as [Dataset S1](#sd001){ref-type="supplementary-material"} and [Dataset S2](#sd002){ref-type="supplementary-material"}. Selection on gene locations would provide an evolutionary explanation for the puzzle of why *K~A~* and *K~S~* are correlated beyond levels expected by neutral evolutionary theory ([@pbio-0020029-Mouchiroud1]; [@pbio-0020029-Ohta1]). Assuming 4-fold sites are neutral, locational selection would have to be realized through the influence of the local mutation rate *K~S~* on the amino acid changing mutation rate *K~A~*. Thus, locational selection and point mutation-based amino acid selection would behave similarly with respect to positive or negative selection on protein sequence, increasing the correlation of *K~A~* and *K~S~*, even if mutations to any individual 4-fold site did not provide a fitness benefit. One caveat is that other, not necessarily exclusive, explanations for the strong correlation of *K~A~* and *K~S~* have been proposed as well---most notably simultaneous substitutions at adjacent sites, so-called tandem substitutions ([@pbio-0020029-Smith2]). Tandem substitutions were not sufficient to explain our hot and cold categories, however. We rederived sets of hot and cold categories after correcting for tandem effects (see [Materials and Methods](#s4){ref-type="sec"}) and once again found similar results. For example, the six hottest categories (of 22 significant) were "dynein ATPase activity," "receptor activity," "homophilic cell adhesion," "olfactory receptor activity," "integral to membrane," and "calcium ion binding activity." The six coldest (of 36) were "nucleus," "regulation of transcription, DNA dependent," "RNA binding activity," "transcription factor activity," "development," and "ribonucleoprotein complex." Mechanisms {#s3b} ---------- For the hot clustered categories, it may be that high mutation rates and high rates of gene duplication are tied to a hidden variable that imposes both phenomena simultaneously. One possibility is the recombination rate along the genome, which [@pbio-0020029-Pal1]) found to have selective effects in yeast. For example, two mechanisms for diversification, gene duplication and mutation, can both be accelerated by recombination ([@pbio-0020029-Graur1]; [@pbio-0020029-Lercher1]). High recombination rates are relevant for a number of the hot gene categories we have studied, as they have been suggested for the protocadherins ([@pbio-0020029-Wu1]), immune response ([@pbio-0020029-Papavasiliou1]), and olfactory families ([@pbio-0020029-Sharon1]). Because both gene duplication and point mutation are useful for diversifying a family, it is difficult to separate the significance of mutation rate and recombination rate. [@pbio-0020029-Pal1]) offered preliminary evidence that in yeast, selection acts on the recombination rate, but not point mutation rates. However, we have observed unusual rate biases for nonclustered gene categories as well, for which recombination would not be expected to play a role. Cold gene categories are not clustered; therefore, the existence of cold categories (as well as nonclustered hot categories) cannot be attributed to duplication events. One alternate phenomenon that could cause cold category biases is gene relocation to cold regions. The concept of relocation brings up a number of questions. First, if cold genes have relocated, this leaves one wondering in what sort of environment cold genes originated. One speculative possibility is that these genes developed in regions of high recombination (the hot regions), which would have allowed for fast duplication and functional diversification, and later dispersed to cooler regions as their functions became fixed. Second, it is unclear whether gene relocations occur frequently enough to account for the observed rate biases. This issue is complicated by the fact that genes have arisen at different times. Many of the cold gene categories occur in diverse sets of tissues and have important regulatory effects, suggesting they should be relatively old. This old age may have allowed them enough time to redistribute through the genome. We verified the correlations of substitution rates along the genome and showed that these correlations lead to an excess of hot and cold genes, confirming studies by [@pbio-0020029-Matassi1]) and [@pbio-0020029-Hardison1]). Our results appear to disagree with those of [@pbio-0020029-Kumar1]), who reported that mutation rates are uniform in the genome. While our rate measurements were qualitatively similar to those of [@pbio-0020029-Kumar1]), one beneficial addition we made was the use of a normalized rate that accounts for the length dependence of rate variance, allowing genes of differing lengths to be treated equally in [Figure 1](#pbio-0020029-g001){ref-type="fig"}. Our correlation length analysis revealed two scales of rate correlation: a short decay length of 1 Mb and a long-range length extending along a syntenous block up to distances of 10 Mb. We have very speculatively proposed that DNA replication units and DNA recombination may be relevant to these length scales. More generally, it is hoped that these scale determinations will be helpful in placing constraints on possible processes that control mutation rate. Some data issues suggest topics for further exploration. First, the resolution of our analysis is dictated by the structure of the GO taxonomy, which currently has 16,000 categories but is evolving. Our category inferences should become more specific as GO gene assignments improve. Second, multispecies data will be invaluable in revealing the mutations that have occurred in each lineage. One promising early result from human--chimpanzee comparisons, based on a set of 96 orthologs derived from HOVERGEN release 44 ([@pbio-0020029-Duret2]), is that olfactory receptors are a hot category. Unfortunately, this is the only statistically significant hot or cold category at present, owing to the lack of data. However, inferences should improve rapidly as more chimpanzee gene identifications become available. Materials and Methods {#s4} ===================== {#s4a} ### Ortholog generation. {#s4a1} We downloaded a list of the available 37,347 human and 27,504 mouse peptides from the ENSEMBL sequence database ([www.ensembl.org](www.ensembl.org)), then used BLAST ([@pbio-0020029-Altschul1]) to find orthologous peptide sequences between the genomes. The peptides studied were the set of all known or predicted peptides in the ENSEMBL 12.31.1 human and 12.3.1 mouse datasets. Sequences were designated as orthologous if the two peptides were each other\'s mutual best hit in the opposing databases, as determined by BLASTALL, and the E-value for the match was 10^−10^ (using the higher score as a worst-case bound) or better. We chose this method of ortholog determination to get a one-to-one relationship between proteins. We found 14,790 ortholog pairs, a coverage rate of approximately 50% in mouse and 40% in human. The observed E-values between orthologs have a median value of 0.0 (\<1e − 180). The aligned peptide orthologs were then used in conjunction with ENSEMBL cDNA data to determine aligned orthologous cDNA. For the chimpanzee--human comparison, human genes from ENSEMBL were compared to chimpanzee genes from HOVERGEN. A mutual best-hit criterion was used to determine the set of 96 orthologs. We manually inspected the mouse--human synteny of the olfactory gene cluster of [Figure 4](#pbio-0020029-g004){ref-type="fig"} to verify that orthologs were assigned correctly. This was to address the concern that orthologs are more difficult to assign in gene categories with many homologous members, since incorrect assignments could distort substitution rates. The synteny structure was found to be almost totally conserved for these genes, as it was for the cell adhesion genes in [Figure 5](#pbio-0020029-g005){ref-type="fig"}. ### Calculation of substitution rates. {#s4a2} We calculated the distribution of substitution rates between the mouse and human genomes using the 4-fold sites of orthologous genes; 4-fold sites are the third bases of codons for which the amino acid is specified by the first two bases. For each of the orthologous gene pairs, we calculated *p*, the fraction of 4-fold sites in which the mouse base differs from the human base. The average value of *p* over all 4-fold sites in all orthologs was \<*p*\> = 0.337. The average 4-fold substitution rate on a genewise basis was 0.338 with a standard deviation of 0.080. These rates were in agreement with substitution rates measured in other studies of 4-fold sites or in ancestral repeats ([@pbio-0020029-Mouse1]; [@pbio-0020029-Hardison1]). Because genes are of finite length, stochastic effects can cause substitution rates to vary from gene to gene, even if all 4-fold sites mutate at the same rate. We defined a normalized substitution rate to correct for these finite-size effects. A gene with *N* 4-fold sites was modeled as having *N*independent events in which substitution can occur with probability \<*p*\>. This formulation can fit both the Jukes--Cantor one-parameter or the Kimura two-parameter model for mutation matrices ([@pbio-0020029-Durbin1]). Although this model is not as sophisticated as other more modern treatments (e.g., see [@pbio-0020029-Tavere1]; [@pbio-0020029-Tamura1]; [@pbio-0020029-Li1]; [@pbio-0020029-Goldman1]), it gives an easily falsifiable prediction that the rate distribution, in the absence of regional correlation, can be approximated by a standard Normal distribution, due to the central limit theorem ([@pbio-0020029-Rice1]). Under this model, at each N the distribution of substitution rates can be described by a binomial distribution with a standard deviation of σ(*N*) = . Therefore, gene substitution rates were normalized by their respective σ(*N*) to get one universal rate distribution, which in the limit of many datapoints should approach the Normal distribution (2π)^−½^ exp (*x* ^2^/2). We defined the normalized substitution rate to be where *p* is the actual 4-fold substitution rate in the gene. The values of *r* for all ortholog pairs were used to calculate the distribution shown in [Figure 1](#pbio-0020029-g001){ref-type="fig"}. The actual rate distribution in genes was found to be skewed toward high or low mutation rates, as shown in [Figure 1](#pbio-0020029-g001){ref-type="fig"}. The observed distribution had a standard deviation of 2.04, significantly higher than the expected σ = 1. Similar excesses of hot and cold genes were found even when corrections were made for base composition. To verify this, we calculated a normalized mutation rate using a four-parameter model in which each site of type A, C, G, or T in the human sequence has its own substitution probability. For each human base (A, C, G, and T), we measured the substitution rate at the corresponding 4-fold locations, yielding 4 rates \<*p~A~*\>, \<*p~C~*\>, \<*p~G~*\>, \<*p~T~*\>. Based on these rates, we then calculated the expected frequency and variance of substitutions for a gene given the gene\'s base composition at the 4-fold sites. This yielded a distribution nearly identical to that in the one-parameter model (see [Figure 1](#pbio-0020029-g001){ref-type="fig"}). We also tested whether neighboring genes have similar substitution rates. The orthologs were ordered by their location along the human genome, after which we calculated the Pearson correlation of a gene\'s substitution rate *r* with that of its following gene. We used only neighboring genes that were in syntenous blocks, as defined by all three conditions of monotonicity (the genes are ordered the same in both species), consistent strand orientation (a block is either in the same strand orientation in both species or completely reversed), and consistent chromosome (no chromosome changes in either species in a block), yielding a dataset of 11,087 neighbor pairs. Under this condition, the Pearson correlation was 0.26, corresponding to a highly significant *p*-value of 10^−189^. ### *z*-score calculation for GO categories. {#s4a3} For each GO category, we calculated a normalized substitution rate (*z*-score) based on the substitution rates of all members of that category. Of the genes in our ortholog set, 9,966 had GO classifications available. The *z*-score was defined to be where \<*r*\>*~GO~* is the average substitution rate *r* for the genes in the GO category, \<*r*\>*~all~* is the average *r* for all of the genes with GO classifications, σ*~all~* is the genewise standard deviation, and *N~GO~* is the number of genes in the category. The *p*-value for *z* was determined from the probability that a Gaussian-distributed variable takes on a value ≥*z*. To reduce the problem of outliers, we limited our analysis to the GO categories containing at least five genes, of which there are 997, and accordingly set a *p*-value cutoff of 1/997 ∼ 10^−3^. We expressed the significance in terms of −log~10~ *p~z~*, which should have a value larger than 3 to be statistically significant. *z*-scores corrected for tandem substitutions were calculated by first removing all possible tandem substitution sites from the dataset. That is, 4-fold sites were only accepted into the dataset if both the preceding and following bases matched in the two species. After culling the dataset, we calculated rates and category *z*-scores as before. ### Clustering analysis. {#s4a4} To measure clustering, for each gene in a GO category we tested whether it had another category member downstream of it within the short-range correlation length of τ = 10^6^ basepairs. In each GO category, we calculated the number of genes satisfying this condition, defining this to be the number of "clustered genes." This "downstream" criterion (rather than a symmetric one) was used to avoid the problem of double counting of genes when several are close together. To test the statistical significance of the number of clustered genes in a category, we used bootstrapping. For each GO category, we performed 5,000 random trials of selecting *N~GO~* random genes from the entire set of orthologs, where *N~GO~* is the number of genes in the GO category. In each trial, we counted the number of clustered genes in this randomly selected group. The average number of clustered genes was used to approximate the random number of clustered genes by a Poisson distribution. These Poisson statistics were then used to calculate the significance of the number of clustered genes for the GO category. A Poisson distribution is appropriate as long as clustering of neighbors is a rare event, i.e., as long as *N~GO~*\<\<*N~allgenes~*, which was generally the case. The random distributions were visually inspected and found to agree with the shape of the Poisson curve. To generate the data for [Tables 1](#pbio-0020029-t001){ref-type="table"} and [2](#pbio-0020029-t002){ref-type="table"}, we also limited ourselves to the 997 categories with at least five genes, implying that −log~10~ *p~cluster~* \> 3 is the cutoff for significance. ### Calculation of repeat sequence mutation rates. {#s4a5} Aligned repeat sequences between mouse and human were obtained from the dataset of [@pbio-0020029-Hardison1]). For each repeat, positions in which a base was defined for both the mouse and human sequence were used to calculate a normalized substitution rate, in analogy with the method used for the 4-fold sites. The genome-wide average value of *p* in these repeat sequences was 0.33, which was very close to the value for 4-fold sites, 0.34. The start position of each repeat sequence was used to define its location in the genome. In order to determine the locations of repeat sequences (based on the June 2002 UCSC genome map) along the physical map used for the gene sequences (based on the ENSEMBL May 2003 map), gene locations according to the two maps were compared. Repeat sequence locations were then corrected using the location differences of nearby genes. For clarity, the ancestral repeat values shown in [Figure 4](#pbio-0020029-g004){ref-type="fig"} and [Figure 5](#pbio-0020029-g005){ref-type="fig"} were smoothed using a moving-window average of 20 repeat sequences. ### Correlation length calculation. {#s4a6} We considered all pairs of genes on continuous orthologous blocks, starting from the first neighbor up to the 35th gene downstream. This allowed us to get hundreds of measurements of *r*(0)*r(t)* for *t* values even as large as several megabases. We binned these data into 100 uniformly spaced groups covering *t* ∈ \[0, 15000000\] and then averaged over each of these bins to determine the correlation function \<*r*(0)*r(t)*\>. The data were plentiful enough for the averaged values shown in [Figure 6](#pbio-0020029-g006){ref-type="fig"} to be statistically significant. It was difficult to extend to larger values of *t* since the amount of data decreases with *t*, a fact manifested in the increasing fluctuations at larger *t* in [Figure 6](#pbio-0020029-g006){ref-type="fig"}. For example, the value of the average correlation \<*r*(0)*r(t)*\> at *t* = 15 Mb in the human data of [Figure 6](#pbio-0020029-g006){ref-type="fig"} was based on only 79 measurements, whereas at *t*= 75,000 it was based on 22,860 measurements. For genes with alternative splicings, only one of the genes was used, in order to avoid spurious effects caused by reuse of DNA. Orthologous block boundaries were defined by genes at which the chromosome changes in either species. Monotonicity and consistent strand orientation were ignored in order to obtain blocks with large values of *t*. Most of the *r*(0)*r(t)* data comes from blocks at least several megabases long. Approximately 5% is in blocks of size less than 10^6^ basepairs, 55% is in blocks of size between 10^6^ and 10^7^ basepairs, and the remaining 40% is in larger blocks. The long-range correlation shown in [Figure 6](#pbio-0020029-g006){ref-type="fig"} was statistically significant. Theoretically, fluctuations in \<*r*(*i*)*r*(*j*)\> should be of the order ∼*O*(1/ , where *N* is the number of data samples in a bin. At a distance of 10 Mb, there were approximately 400 samples, corresponding to an uncertainty of approximately 0.05. This uncertainty was an order of magnitude smaller than the observed value of *A* ~∞~ = 0.4. Supporting Information {#s5} ====================== Dataset S1 ::: {.caption} ###### Hot Gene Categories Based on *K~A~/K~S~* Gene categories with significant positive selection on amino acid changing point mutations. (23 KB XLS). ::: ::: {.caption} ###### Click here for additional data file. ::: Dataset S2 ::: {.caption} ###### Cold Gene Categories Based on *K~A~/K~S~* Gene categories with significant negative selection on amino acid changing point mutations. (21 KB XLS). ::: ::: {.caption} ###### Click here for additional data file. ::: Accession Numbers {#s5a} ----------------- The Gene Ontology (<http://www.geneontology.org/>) ID numbers for the categories discussed in this paper are as follows: brain development (GO:0007420), calcium-dependent cell adhesion molecule activity (GO:0008014), calcium-dependent protein serine/threonine phosphatase activity (GO:0004723), calcium ion binding activity (GO:0005509), carbohydrate metabolism (GO:0005975), cell adhesion (GO:0007155), cell growth and/or maintenance (GO:0008151), chymotrypsin activity (GO:0004263), CTD phosphatase activity (GO:0008420), cytoplasm (GO:0005737), development (GO:0007275), DNA binding activity (GO:0003677), dynein ATPase activity (GO:0008567), dynein complex (GO:0030286), enzyme activity (GO:0003824), G-protein coupled receptor protein signaling pathway (GO:0007186), GTP binding activity (GO:0005525), heterogeneous nuclear ribonucleoprotein (GO:0008436), homophilic cell adhesion (GO:0007156), immune response (GO:0006955), integral to membrane (GO:0016021), internalization receptor activity (GO:0015029), intracellular protein transport (GO:0006886), magnesium-dependent protein serine/threonine phosphatase activity (GO:0004724), membrane (GO:0016020), metabolism (GO:0008152), microtubule-based movement (GO:0007018), microtubule motor activity (GO:0003777), mRNA binding activity (GO:0003729), myosin phosphatase activity (GO:0017018), neurogenesis (GO:0007399), nucleus (GO:0005634), olfactory receptor activity (GO:0004984), oncogenesis (GO:0007048), protein amino acid dephosphorylation (GO:0006470), protein phosphatase type 2A activity (GO:0000158), protein phosphatase type 2B activity (GO:0030357), protein phosphatase type 2C activity (GO:0015071), protein serine/threonine kinase activity (GO:0004674), protein transporter activity (GO:0008565), proteolysis and peptidolysis (GO:0006508), receptor activity (GO:0004872), regulation of transcription, DNA-dependent (GO:0006355), regulation of transcription from Pol II promoter (GO:0006357), regulation of translational initiation (GO:0006446), ribonucleoprotein complex (GO:0030529), RNA binding activity (GO:0003723), RNA polymerase II transcription factor activity (GO:0003702), RNA splicing (GO:0008380), transcription coactivator activity (GO:0003713), transcription factor activity (GO:0003700), transcriptional activator activity (GO:0016563), translation regulator activity (GO:0045182), trypsin activity (GO:0004295), ubiquitin conjugating enzyme activity (GO:0004840), ubiquitin cycle (GO:0006512), and ubiquitin-dependent protein catabolism (GO:0006511). This material is based upon work supported by the National Science Foundation under a grant awarded in 2003. Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the National Science Foundation. This work is also supported in part by a Sandler\'s opportunity grant and a David and Lucile Packard fellowship awarded to HL. JC would like to thank T. Hwa, D. Petrov, and C. S. Chin for comments on the manuscript. HL acknowledges helpful discussions with Pat O\'Farrell and Hiten Madhani. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.**JC and HL conceived and designed the experiments. JC performed the experiments. JC and HL analyzed the data. JC contributed reagents/materials/analysis tools. JC wrote the paper. Academic Editor: Charles H. Langley, University of California, Davis

PubMed Central

2024-06-05T03:55:46.067118

2004-2-17

PMC340941

Introduction {#s1} ============ The analysis of whole-genome sequences has revealed that most signaling systems consist of multiple ligands that converge on a relatively small set of receptors and pathway-specific transcription factors. In the case of human transforming growth factor-β (TGFβ) signaling, 42 TGFβs converge on seven type I receptors, five type II receptors, and two classes of Smad signal transducers (reviewed in [@pbio-0020030-Shi1]). This convergence has raised the question of how ligand diversity and signaling specificity among different signals can be achieved. If different TGFβs activate the same receptors, it is unclear how these ligands can vary in their function (diversity) or how a given signal can have a unique role (specificity). Biochemical studies have suggested that ligand diversity can be attained by differential stability and receptor affinity, leading to differences in signaling strength (reviewed in [@pbio-0020030-Piek1]; [@pbio-0020030-Shi1]). An additional source of ligand variability stems from differential ligand movement through a field of cells, rendering related signals either short- or long-range ([@pbio-0020030-Chen2]). Finally, specificity and diversity can also be determined by ligand-specific cofactors or inhibitors ([@pbio-0020030-Piek1]; [@pbio-0020030-Shi1]). A prominent example involves epidermal growth factor--Cripto/FRL-1/Cryptic (EGF-CFC) coreceptors and the TGFβs Activin, Nodal, and Vg1/GDF1 (growth and differentiation factor-1). In this case, differential dependence on a coreceptor leads to ligand diversity and signaling specificity (reviewed in [@pbio-0020030-Schier1]). Members of the Nodal, Activin, and Vg1/GDF1 subfamilies display similar activities and are potent mesendoderm inducers in vertebrates (reviewed in [@pbio-0020030-Schier2]). Genetic and biochemical studies have shown that EGF-CFC proteins are essential for signaling by Nodal and Vg1/GDF1 ([@pbio-0020030-Gritsman1]; [@pbio-0020030-Reissmann1]; [@pbio-0020030-Yeo1]; [@pbio-0020030-Bianco1]; [@pbio-0020030-Sakuma1]; [@pbio-0020030-Yan2]; [@pbio-0020030-Cheng2]). EGF-CFC proteins are extracellular glycosylphosphatidylinositol (GPI)-linked factors and include One-eyed pinhead (Oep) in zebrafish and mammalian Cripto and Cryptic (reviewed in [@pbio-0020030-Shen2]; [@pbio-0020030-Minchiotti1]; [@pbio-0020030-Schier1]). Genetic studies in zebrafish and mouse have shown that EGF-CFC proteins and Nodal are required for mesoderm and endoderm induction ([@pbio-0020030-Conlon1], 1994; [@pbio-0020030-Zhou1]; [@pbio-0020030-Ding1]; [@pbio-0020030-Feldman1]; [@pbio-0020030-Gritsman1]). For example, zebrafish embryos lacking both the maternal and zygotic contribution of Oep (MZ*oep*) lack all endoderm and most mesoderm, similar to the double mutants for the zebrafish Nodal-related genes *cyclops* and *squint* (*sqt*) ([@pbio-0020030-Feldman1]; [@pbio-0020030-Gritsman1]). Moreover, Nodal and Vg1/GDF1 are inactive in MZ*oep* mutants ([@pbio-0020030-Gritsman1]; [@pbio-0020030-Cheng2]). During later stages of development, Oep, Cryptic, Nodal, and GDF1 are required for proper left--right axis formation ([@pbio-0020030-Gaio1]; [@pbio-0020030-Yan1]; [@pbio-0020030-Bamford1]; [@pbio-0020030-Rankin1]; [@pbio-0020030-Brennan1]; [@pbio-0020030-Long1]). The EGF-CFC protein Cripto is highly overexpressed in human epithelial cancers, such as breast and colon carcinomas (reviewed in [@pbio-0020030-Salomon1]), and has been implicated in tumorigenesis ([@pbio-0020030-Ciardiello1], 1994; [@pbio-0020030-Baldassarre1]; [@pbio-0020030-De1]; [@pbio-0020030-Salomon1]; [@pbio-0020030-Adkins1]). The mechanism by which Cripto mediates tumorigenesis is not well understood. Several possibilities include mediating Nodal/GDF1 signaling ([@pbio-0020030-Gritsman1]; [@pbio-0020030-Reissmann1]; [@pbio-0020030-Yeo1]; [@pbio-0020030-Bianco1]; [@pbio-0020030-Sakuma1]; [@pbio-0020030-Yan2]; [@pbio-0020030-Cheng2]), antagonizing Activin signaling ([@pbio-0020030-Adkins1]; [@pbio-0020030-Gray1]), or activating Akt and mitogen-activated protein kinase (MAPK) pathways independently of the TGFβ signals and Activin receptors ([@pbio-0020030-Ebert1]1999; [@pbio-0020030-Bianco1], 2003). Whatever the molecular mechanism of Cripto activity, inhibition of Cripto by antisense or antibody blockade can inhibit tumor cell proliferation in vitro and in vivo ([@pbio-0020030-Ciardiello2]; [@pbio-0020030-Baldassarre1]; [@pbio-0020030-De1]; [@pbio-0020030-Adkins1]). Biochemically, EGF-CFC proteins can act as coreceptors for Nodal and Vg1/GDF1 to bind and activate the type I Activin receptor Alk4 and the type II Activin receptor ActRIIB ([@pbio-0020030-Reissmann1]; [@pbio-0020030-Yeo1]; [@pbio-0020030-Sakuma1]; [@pbio-0020030-Yan2]; [@pbio-0020030-Bianco1]; [@pbio-0020030-Cheng2]). In the absence of EGF-CFC proteins, these TGFβs cannot form a complex with Activin receptors. Strikingly, Activin utilizes the same receptors as Nodal and Vg1/GDF1, but does not require EGF-CFC coreceptors ([@pbio-0020030-Mathews1]; [@pbio-0020030-Attisano1], 1996; [@pbio-0020030-Hemmati-Brivanlou1]; [@pbio-0020030-Mathews2]; [@pbio-0020030-ten1]; [@pbio-0020030-Chang1]). For instance, Activin can signal in MZ*oep* mutants ([@pbio-0020030-Gritsman1]). This ligand diversity between Activin and Nodal or Vg1/GDF1 raises the question of which sequences confer coreceptor dependence or independence. Activin, Nodal, and Vg1/GDF1 are highly related and are thought to acquire very similar folds. Like other TGFβ ligands, Activin has four major structural features: a β-stranded Finger 1, an α-helical Heel, a β-stranded Finger 2, and three conserved disulfide bonds that form a cysteine knot ([@pbio-0020030-Shi1]; [@pbio-0020030-Thompson1]). Sequence comparisons indicate that the highest divergence among Activin, Nodal, and Vg1/GDF1 is in the N-terminal segment of Finger 1, the central α-helix, and the loop of Finger 2 with approximately 10%, approximately 15%, and approximately 25% sequence identity, respectively. These regions are potential candidates to determine the specificity of receptor--coreceptor--ligand interactions. In addition to coreceptors, antagonists represent another class of extracellular factors that control ligand access to receptors (reviewed in [@pbio-0020030-Piek1]; [@pbio-0020030-Freeman1]; [@pbio-0020030-De2]; [@pbio-0020030-Shi1]). For example, the divergent TGFβ class of Lefty proteins antagonizes Nodal signaling (reviewed in [@pbio-0020030-Hamada1]; [@pbio-0020030-Schier1]). Unlike other TGFβs, Lefty proteins may function as monomers due to the lack of a cysteine residue involved in dimerization ([@pbio-0020030-Meno1]; [@pbio-0020030-Thisse1]; [@pbio-0020030-Sakuma1]). Lefty overexpression in zebrafish induces phenotypes identical to *cyclops;sqt* double mutants and MZ*oep* mutants ([@pbio-0020030-Bisgrove1]; [@pbio-0020030-Meno4]; [@pbio-0020030-Thisse1]). Furthermore, the loss of Lefty activity leads to enhanced Nodal signaling during mesoderm induction and left--right axis determination ([@pbio-0020030-Meno4], 2001; [@pbio-0020030-Agathon1]; [@pbio-0020030-Branford1]; [@pbio-0020030-Chen3]; [@pbio-0020030-Feldman2]). Although it has not been determined whether Lefty directly blocks Vg1/GDF1 signaling ([@pbio-0020030-Branford2]), it has been proposed that Lefty inhibits signaling by Activin. Misexpression of Activin or ActRIIB can overcome the inhibitory effects of Lefty ([@pbio-0020030-Meno4]; [@pbio-0020030-Thisse1]; [@pbio-0020030-Cheng1]; [@pbio-0020030-Tanegashima1]; [@pbio-0020030-Sakuma1]). Hence, some members of the Lefty family have been called Antivins for their anti-Activin properties ([@pbio-0020030-Thisse1]; [@pbio-0020030-Cheng1]; [@pbio-0020030-Ishimaru1]; [@pbio-0020030-Tanegashima1]). However, it has been elusive how Lefty functions at the molecular level. Here we present genetic and biochemical studies in zebrafish and *Xenopus* that indicate that Lefty is an in vivo antagonist of EGF-CFC coreceptors. We find that Lefty can antagonize signaling by the coreceptor-dependent ligands Nodal and Vg1/GDF1, but not Activin. Lefty genetically interacts with Cripto and Oep and competes with Nodal for binding to Cripto, representing a novel mechanism for antagonizing TGFβ signaling. We identify a short region in Finger 2 of Activin, Nodal, and Vg1 that determines EGF-CFC coreceptor-dependent or coreceptor-independent signaling and governs susceptibility to Lefty. These results indicate that subtle sequence variations in TGFβ ligands can dramatically expand signaling diversity by determining interactions with coreceptors and their antagonists. Results {#s2} ======= Lefty Antagonizes Nodal and Vg1 Signaling, but Not Activin Signaling {#s2a} -------------------------------------------------------------------- TGFβ ligands that activate Activin receptors can be categorized into two classes. The Activin class activates Activin receptors in an EGF-CFC coreceptor-independent manner, whereas the Nodal and Vg1/GDF1 classes require EGF-CFC proteins for receptor activation ([@pbio-0020030-Gritsman1]; [@pbio-0020030-Cheng2]). To determine whether these classes are also differentially susceptible to inhibition by the TGFβ antagonist Lefty, we coexpressed zebrafish Lefty1 with *Xenopus* ActivinβB, *Xenopus* ActivinβA, Sqt (a zebrafish Nodal-related protein), or zebrafish Vg1 (a chimeric molecule containing the *Xenopus* ActivinβB prodomain fused to the mature domain of zebrafish Vg1) in zebrafish embryos ([@pbio-0020030-Smith1]; [@pbio-0020030-Thomsen1]; [@pbio-0020030-van1]; [@pbio-0020030-Helde1]; Erter et al. 1998; [@pbio-0020030-Feldman1]; [@pbio-0020030-Thisse1]). As a readout for active signaling, we analyzed the ectopic induction of the Nodal downstream gene *goosecoid* (*gsc*). ActivinβB, ActivinβA, Sqt, and Vg1-induced ectopic *gsc* expression in wild-type embryos ([Figure 1](#pbio-0020030-g001){ref-type="fig"}D, [1](#pbio-0020030-g001){ref-type="fig"}G, [1](#pbio-0020030-g001){ref-type="fig"}J, and [1](#pbio-0020030-g001){ref-type="fig"}M; [@pbio-0020030-Gritsman1]; [@pbio-0020030-Cheng2]). Coexpression of Lefty1 efficiently inhibited *gsc* induction by Sqt ([Figure 1](#pbio-0020030-g001){ref-type="fig"}K and [1](#pbio-0020030-g001){ref-type="fig"}L; [@pbio-0020030-Bisgrove1]; [@pbio-0020030-Meno4]; [@pbio-0020030-Thisse3]) and Vg1 ([Figure 1](#pbio-0020030-g001){ref-type="fig"}N and [1](#pbio-0020030-g001){ref-type="fig"}O), but not ActivinβB or ActivinβA ([Figure 1](#pbio-0020030-g001){ref-type="fig"}E, [1](#pbio-0020030-g001){ref-type="fig"}F, [1](#pbio-0020030-g001){ref-type="fig"}H, and [1](#pbio-0020030-g001){ref-type="fig"}I). To examine whether Lefty1 can antagonize the induction of a gene that responds to very low levels of Activin signaling, we titrated ActivinβB levels so that *no tail* (*ntl*; also known as *brachyury/T*) expression was only weakly induced (see arrowhead in [Figure 1](#pbio-0020030-g001){ref-type="fig"}A). The coexpression of Lefty1 did not inhibit *ntl* induction by ActivinβB ([Figure 1](#pbio-0020030-g001){ref-type="fig"}B and [1](#pbio-0020030-g001){ref-type="fig"}C), but inhibited the dorsal margin expression of *ntl* that is dependent on endogenous Nodal signaling (see asterisks in [Figure 1](#pbio-0020030-g001){ref-type="fig"}B and [1](#pbio-0020030-g001){ref-type="fig"}C; [@pbio-0020030-Feldman1]). In a more quantitative assay, we overexpressed Lefty1, ActivinβB, Sqt, and Vg1 in zebrafish embryos in the presence of the luciferase reporter A3-luc, which contains FoxH1/P-Smad2 response elements ([@pbio-0020030-Chen1]). Consistent with the *gsc* and *ntl* induction assays, Sqt and Vg1 signaling, but not ActivinβB signaling, is inhibited by Lefty1 ([Figure 1](#pbio-0020030-g001){ref-type="fig"}P). These results indicate that Lefty1 efficiently antagonizes Nodal and Vg1/GDF1 signaling, but not Activin signaling. ::: {#pbio-0020030-g001 .fig} Figure 1 ::: {.caption} ###### Lefty Antagonizes Nodal and Vg1 Signaling, but Not Activin Signaling, in Zebrafish *ntl* mRNA expression (A--C) and *gsc* mRNA expression (D--O) in wild-type zebrafish embryos at shield stage, animal pole view. Embryos were injected with low levels (1 pg) of *activin βB* mRNA (A--C), high levels (10 pg) of *activin βB* mRNA (D--F), 200 pg of *activin βA* mRNA (G--I), 75 pg of *sqt* mRNA (J--L), or 200 pg of *Vg1* (M--O). Embryos were further double-injected with either 500 pg of *LacZ* mRNA (A, D, G, J, and M), 100 pg of *lefty1* and 400 pg of *LacZ* mRNAs (B, E, H, K, and N), or 500 pg of *lefty1* mRNA (C, F, I, L, and O). Ectopic *ntl* expression (arrowheads) in *activin βB* mRNA-injected embryos was not inhibited by Lefty1 (B and C) when compared with *LacZ* mRNA-coinjected controls (A). Note the dorsal expression of *ntl* (asterisks)---that is, dependent on endogenous Nodal signaling---is inhibited by Lefty1 in these embryos (B and C). Ectopic *gsc* expression in *activin βB* and *activin βA* mRNA-injected embryos was not inhibited by Lefty1 (E and F and H and I, respectively). In contrast, ectopic *gsc* expression in *sqt* and *Vg1* mRNA-injected embryos was inhibited by both levels of Lefty expression (K and L and N and O, respectively). Wild-type embryos (P) were injected with 10 pg (low) and 20 pg (high) of *activin βB*/HA, 75 pg of *sqt*, or 200 pg of *Vg1* mRNA. Embryos were further double-injected with 500 pg of *LacZ* mRNA, 100 pg of *lefty1*, and 400 pg of *LacZ* mRNAs, or 500 pg of *lefty1* mRNA. Smad2 pathway activation was measured by an Activin response element luciferase reporter, A3-luc. Values are folds over wild-type control injected with 500 pg of *LacZ* mRNA and A3-luc reporter. An asterisk indicates a significant difference from the level of activation with ligand and *LacZ* expression alone (Student\'s *t*-test, *p* \< 0.05). :::  ::: EGF-CFC Proteins Genetically Interact with Lefty {#s2b} ------------------------------------------------ The molecular mechanism of Lefty action has been unresolved. Lefty seems to act upstream of the Activin receptors, as Lefty cannot block signaling from ligand-independent constitutively activated receptors ([@pbio-0020030-Thisse1]). Our finding that Lefty blocks Nodal and Vg1 signaling, but not Activin signaling, suggests that Lefty blocks extracellular components specific to the Nodal and Vg1 pathways. The only such factors identified to date are the EGF-CFC coreceptors. We therefore examined whether the EGF-CFC genes zebrafish *oep* and mouse *Cripto* genetically interact with Lefty1. Overexpression of Lefty1 in wild-type zebrafish resulted in embryos lacking head and trunk mesendoderm due to inhibition of endogenous Nodal signaling ([Figure 2](#pbio-0020030-g002){ref-type="fig"}A1 and [2](#pbio-0020030-g002){ref-type="fig"}A′; [@pbio-0020030-Bisgrove1]; [@pbio-0020030-Meno4]; [@pbio-0020030-Thisse1]). Coexpression of Cripto or Oep partially suppressed Lefty-induced defects ([Figure 2](#pbio-0020030-g002){ref-type="fig"}B1--[2](#pbio-0020030-g002){ref-type="fig"}B3; data not shown), restoring trunk and head mesoderm, including the notochord, and resulting in the separation of the eye field into two eyes. These results indicate an antagonistic relationship between EGF-CFC coreceptors and Leftys. ::: {#pbio-0020030-g002 .fig} Figure 2 ::: {.caption} ###### EGF-CFC Coreceptors Genetically Interact with Lefty Live wild-type zebrafish embryos at 30 h postfertilization (hpf). (A1, B1, and C1) Ventral views of the head. (A1′, B2, B3, and C1′) Lateral views, with anterior to the left, dorsal up. (A1, A1′, B1, B2, and B3) Wild-type embryos were injected with 20 pg of *lefty1* mRNA. Embryos were further double-injected with either 200 pg of *LacZ* mRNA (A1 and A1′) or 200 pg of *Cripto* mRNA (B1, B2, and B3). (C1 and C1′) Wild-type embryos injected with 200 pg of *Cripto* mRNA and 20 pg of *LacZ* mRNA. Misexpression of Lefty1 results in cyclopia and other head and trunk mesoderm defects (\[A1 and A1′\] 32 of 32 embryos had the phenotype shown; arrow shows cyclopia). Coexpression of Cripto with Lefty in embryos leads to rescue of two eyes (\[B1\] four of 50; arrows show two eyes), notochord (\[B2\] 20 of 50; inset shows trunk somites and notochord, red bar delineates notochord), and trunk somites (\[B3\] 50 of 50). Embryos injected with *Cripto* mRNA only show normal wild-type phenotype (\[C1 and C1′\] 30 of 30; arrow in \[C1\] shows two normal eyes, and inset in \[C1′\] shows normal notochord and trunk somites, red bar delineates notochord). :::  ::: Lefty Binds to Cripto, but Not to ActRIIB or Alk4 {#s2c} ------------------------------------------------- Because Lefty and EGF-CFC proteins interact genetically, we examined whether Lefty interacts biochemically with Cripto and/or ActRIIB/Alk4 receptor complexes. We expressed and immunoprecipitated epitope-tagged ligands (zebrafish Lefty1/HA \[hemagglutinin\], zebrafish Lefty1/Glu, or zebrafish Sqt/HA), receptors (mouse ActRIIB\[KR\]/Myc and human Alk4\[KR\]/Flag), and a coreceptor (mouse Cripto/Flag) in *Xenopus* embryos ([@pbio-0020030-Yeo1]; [@pbio-0020030-Cheng2]). Similar to other Nodal-related proteins ([@pbio-0020030-Reissmann1]; [@pbio-0020030-Yeo1]; [@pbio-0020030-Bianco1]; [@pbio-0020030-Sakuma1]; [@pbio-0020030-Yan2]), Sqt formed a complex with the type II receptor ActRIIB, type I receptor Alk4, and Cripto ([Figure 3](#pbio-0020030-g003){ref-type="fig"}A). In contrast, Lefty1 coimmunoprecipitated Cripto, but not ActRIIB or Alk4 ([Figure 3](#pbio-0020030-g003){ref-type="fig"}A). Since Cripto is bound to Alk4 even in the absence of ligand ([@pbio-0020030-Reissmann1]; [@pbio-0020030-Yeo1]; [@pbio-0020030-Bianco1]; [@pbio-0020030-Yan2]), Lefty seemed to disrupt the Cripto--Alk4 interaction. In reverse experiments, Cripto efficiently coimmunoprecipitated Lefty1 ([Figure 3](#pbio-0020030-g003){ref-type="fig"}B). Since Sqt and Lefty1 can both bind to Cripto ([Figure 3](#pbio-0020030-g003){ref-type="fig"}C; [@pbio-0020030-Reissmann1]; [@pbio-0020030-Yeo1]; [@pbio-0020030-Bianco1]; [@pbio-0020030-Sakuma1]; [@pbio-0020030-Yan2]), these two ligands might compete for binding to Cripto. Indeed, the coexpression of Lefty1 led to decreased interactions between Cripto and Sqt ([Figure 3](#pbio-0020030-g003){ref-type="fig"}C). To determine whether Cripto can directly interact with Lefty, we immunoprecipitated purified mouse Lefty1 protein (mLefty1) in the presence of either purified mouse Cripto protein or a control cysteine-rich protein, mouse vascular endothelial growth factor-D (VEGF-D). mLefty1 protein directly interacted with Cripto, but not with VEGF-D. Together, these results suggest that Lefty inhibits Nodal signaling by associating with Cripto and blocking it from interacting with Nodal. ::: {#pbio-0020030-g003 .fig} Figure 3 ::: {.caption} ###### Lefty Binds to Cripto, but Not to the Activin Receptors ActRIIB and Alk4 (A and B) Lefty1 interacts with Cripto. RNAs (1 ng each) encoding ActRIIB(KR)/Myc, Alk4(KR)/Flag, Cripto/Flag, Lefty1/HA, or Sqt/HA were injected into *Xenopus* embryos. After chemical cross-linking, lysates were immunoprecipitated for either Lefty1/HA or Sqt/HA (A) with anti-HA antibody, or ActRIIB(KR)/Myc, Alk4(KR)/HA, Cripto/Flag (B) with, respectively, anti-Myc, anti-HA, or anti-Flag antibodies. Note that Lefty1 specifically interacts with Cripto (A and B), and these Lefty/Cripto complexes do not contain Alk4 (A). Moreover, processed Lefty1 binds much more efficiently to Cripto than full-length Lefty1 precursor (B). In contrast, Sqt can bind to ActRIIB, Alk4, and Cripto (A). The 55 kDa protein marker in (B) is estimated based on molecular weight markers. \(C) Lefty1 competes with Nodal for binding to Cripto. RNAs encoding Sqt/HA (1 ng), Cripto/Flag (100 pg), or Lefty1 (2 ng) were injected and anti-Flag antibody was used to immunopreciptate Cripto/Flag. \(D) mLefty1 binds directly to Cripto. Purified mouse Lefty1 protein (mLefty1; 10 μg/ml) was mixed with either soluble purified Cripto/His protein (5 μg/ml) or purified control VEGF-D/His protein (5 μg/ml). After chemical cross-linking, mLefty1 was immunoprecipitated with anti-mLefty1 antibody. mLefty1 associates with Cripto, but not with control VEGF-D. Proteins in the coimmunoprecipitates and total extracts were probed in Western blot analysis with the indicated antibodies: ActRIIB(KR)/Myc (kinase-defective receptor, approximately 120 kDa; anti-Myc), Alk4(KR)/Flag (kinase-defective receptor, approximately 70 kDa; anti-Flag), Cripto/Flag (approximately 30 kDa; anti-Flag), Lefty1/HA (mature ligand, approximately 36--40 kDa; anti-HA; [@pbio-0020030-Sakuma1]), Sqt/HA (unprocessed precursor, approximately 55 kDa; mature ligand, approximately 22 kDa; anti-HA), Lefty1/Glu (unprocessed precursor, approximately 55 kDa; mature ligand, approximately 38 kDa; anti-Glu; [@pbio-0020030-Sakuma1]), mLefty1 (mature ligand, approximately 36 kDa, anti-mLefty1; [@pbio-0020030-Sakuma1]), Cripto/His (soluble form, approximately 22--25 kDa; anti-His), and VEGF-D/His (mature ligand, approximately 15--20 kDa; anti-His). :::  ::: Activin Loop-β8 Region Confers EGF-CFC Coreceptor Independence to Sqt {#s2d} --------------------------------------------------------------------- The finding that TGFβ ligands that activate Activin receptors can be grouped into a EGF-CFC coreceptor-dependent class that is susceptible to inhibition by Lefty (Nodal and Vg1) and a class that is independent of EGF-CFC proteins and resistant to Lefty (Activin) prompted us to examine which sequences underlie this ligand diversity. We therefore generated chimeras between Activins (EGF-CFC-independent) and Sqt or Vg1 (EGF-CFC-dependent) (Figures [4](#pbio-0020030-g004){ref-type="fig"} and [5](#pbio-0020030-g005){ref-type="fig"}). As a readout for active signaling, we injected mRNAs encoding these chimeric ligands into wild-type and MZo*ep* zebrafish embryos and analyzed the ectopic induction of the downstream genes *ntl* and *gsc*. Sqt, Vg1, and Activins induced these genes in wild-type embryos, allowing us to determine which chimeric ligands were active. Activins, but not Sqt or Vg1, were active in MZ*oep* mutants, allowing us to test which sequences conferred EGF-CFC coreceptor dependence or independence. ::: {#pbio-0020030-g004 .fig} Figure 4 ::: {.caption} ###### Chimera Analysis to Identify TGFβ Sequence Determinants Conferring EGF-CFC Coreceptor Dependence or Independence Schematic depiction of chimeras of mature ligand domains, Finger 1 (F1), Heel (H), and Finger 2 (F2), between *Xenopus* ActivinβB and zebrafish Sqt. HA indicates an hemagglutinin epitope tag. Schematic is not drawn to scale. The letters in these three-lettered (XXX) chimeras represent the Finger 1, Heel, and Finger 2, respectively. *S* denotes Squint; *A* denotes ActivinβB. Synthetic mRNAs (200 pg) encoding chimeras were injected into wild-type and MZ*oep* embryos. *gsc* and *ntl* mRNA expression is shown at shield stage; animal pole views are dorsal to the right. *gsc* is expressed in the dorsal organizer (shield) in wild-type embryos, but is absent in MZ*oep* mutants. *ntl* is expressed around the entire margin in wild-type embryos, but the dorsal margin expression is lost in MZ*oep* mutants. The presence of the ActivinβB prodomain and epitope tag does not alter the specificity or functionality of wild-type ActivinβB (AAA) or Sqt (SSS). AAA can induce ectopic *gsc* and *ntl* expression in both wild-type and MZ*oep* embryos. In contrast, SSS can induce ectopic *gsc* and *ntl* expression in only wild-type embryos. Similar to ActivinβB, chimeras SSA, SAS, ASA, and SAA can induce ectopic *gsc* and *ntl* expression in both wild-type and MZ*oep* embryos. Chimeras ASS and AAS are inactive in both wild-type and MZ*oep* embryos. Western blot analysis indicated that all chimeric constructs produce stable ligands (data not shown). :::  ::: ::: {#pbio-0020030-g005 .fig} Figure 5 ::: {.caption} ###### Sequence Determinants Conferring Independence from EGF-CFC Coreceptors \(A) Sequence alignment of Finger 2 region of EGF-CFC-dependent and EGF-CFC-independent TGFβ ligands. Location of secondary structure elements, β-sheets (β6--β9) and loop, are shown ([@pbio-0020030-Kirsch1]). Residue numbering is from mouse ActivinβA. (B--E) Synthetic mRNAs (200 pg) encoding chimeras of Finger 2 subregions between *Xenopus* ActivinβB or ActivinβA and zebrafish Sqt or Vg1 were injected into wild-type and MZ*oep* embryos. Schematic is not drawn to scale. *gsc* and *ntl* mRNA expression is at shield stage; animal pole views are dorsal to the right. \(B) Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8^ ^β^ ^9\]^ and Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ can induce *gsc* and *ntl* expression in both wild-type and MZ*oep* embryos. \(C) Sqt^Act^ ^β^ ^B\[^ ^β^ ^8\]^ can weakly expand *ntl* expression in MZ*oep* mutants. *ntl* mRNA expression in MZ*oep* mutants is at shield stage; lateral view. \(D) Other TGFβs conform to loop-β8 EGF-CFC-independent determinant. Note that *Xenopus* ActivinβA can induce ectopic *gsc* in both wild-type and MZ*oep* embryos. In contrast, Vg1 can only induce *gsc* in wild-type embryos. Similar to Activins, chimeric Sqt^Act^ ^β^ ^A\[loop^ ^β^ ^8\]^ and Vg1^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ can induce ectopic *gsc* in both wild-type and MZ*oep* embryos. \(E) Wild-type and MZoep embryos were injected with 5 pg of *activin βB*, 100 pg of *sqt*, 100 pg of *Vg1*, 125 pg of *Sqt^ActβB\[loopβ8\]^*, 250 pg of *Sqt^ActβA\[loopβ8\]^*, or 100 pg of *Vg1^ActβB\[loopβ8\]^* mRNA. Smad2 pathway activation was measured by an Activin response element luciferase reporter, A3-luc. Luciferase units are relative to wild-type or MZ*oep* control injected with the A3-luc reporter alone. \(F) Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ can bind to ActRIIB and Alk4 in the absence of EGF-CFC coreceptors. RNAs (1 ng each) encoding ActRIIB(KR)/Myc, Alk4(KR)/Flag, Cripto/Flag, ActivinβB/HA, Sqt/HA, or Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^/HA were injected into *Xenopus* embryos. Proteins in the coimmunoprecipitates and total extracts were probed in Western blot analysis with the indicated antibodies: ActRIIB(KR)/Myc (approximately 120 kDa; anti-Myc), Alk4(KR)/Flag (approximately 70 kDa; anti-Flag), Cripto/Flag (approximately 30 kDa; anti-Flag), ActivinβB/HA (mature ligand, approximately 16 kDa; anti-HA), Sqt/HA (mature ligand, approximately 22 kDa; anti-HA), and Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^/HA (mature ligand, approximately 22 kDa; anti-HA). :::  ::: Initially, swaps of the Finger 1, Heel, or Finger 2 domains of Sqt and ActivinβB were generated. As shown in [Figure 4](#pbio-0020030-g004){ref-type="fig"}, the Finger 2 region of ActivinβB contains sequence determinants that conferred EGF-CFC-independent activity on chimeric ligands. Chimeric SSA, ASA, and SAA that contain the Finger 2 region of ActivinβB were active in both wild-type and MZ*oep* embryos ([Figure 4](#pbio-0020030-g004){ref-type="fig"}). To further delineate this region, we generated additional chimeras ([Figure 5](#pbio-0020030-g005){ref-type="fig"}B). Short stretches of full-length Sqt were replaced by the corresponding region of ActivinβB, including the β6β7, loop, β8, or β9 subregions ([Figure 5](#pbio-0020030-g005){ref-type="fig"}A and [5](#pbio-0020030-g005){ref-type="fig"}B; data not shown). Analysis of these chimeras revealed that the 14 amino acids encoding the loop and β8 region of ActivinβB (Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^; the bracketed region in superscript denotes substituted domains) were sufficient to confer EGF-CFC independence. Further dissection of this region into loop alone (Sqt^Act^ ^β^ ^B\[loop\]^) or β8 alone (Sqt^Act^ ^β^ ^B\[^ ^β^ ^8\]^) yielded no or much weaker activity in MZ*oep* mutants as compared with wild-type embryos ([Figure 5](#pbio-0020030-g005){ref-type="fig"}B and [5](#pbio-0020030-g005){ref-type="fig"}C). These results suggest that a 14 amino acid region in Activin is sufficient to confer EGF-CFC independence when placed into Sqt. Activin Loop-β8 Region Confers EGF-CFC Coreceptor Independence to Vg1 {#s2e} --------------------------------------------------------------------- To determine whether the loop-β8 region has a wider role in conferring coreceptor independence, we generated additional chimeras using ActivinβA (another EGF-CFC-independent ligand) and Vg1. Sqt^Act^ ^β^ ^A\[loop^ ^β^ ^8\]^ (full-length Sqt with an ActivinβA loop-β8 region) and Vg1^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ (Vg1 with an ActivinβB loop-β8 region) both induced *gsc* expression in MZ*oep* mutants with similar efficiencies as in wild-type embryos ([Figure 5](#pbio-0020030-g005){ref-type="fig"}D). These results were also corroborated using the A3-luc reporter assay ([Figure 5](#pbio-0020030-g005){ref-type="fig"}E) and suggest that the loop-β8 region has a general role in conferring EGF-CFC coreceptor independence. Activin Loop-β8 Region Confers Binding to Activin Receptors in the Absence of EGF-CFC Coreceptors {#s2f} ------------------------------------------------------------------------------------------------- Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ can signal in an EGF-CFC-independent manner in vivo, suggesting that this chimeric protein might bind to ActRIIB and Alk4 receptors in the absence of EGF-CFC coreceptors. To test this idea, we coexpressed and immunoprecipitated epitope-tagged ligands (ActivinβB/HA, Sqt/HA, Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^/HA), receptors (ActRIIB\[KR\]/Myc and Alk4\[KR\]/Flag), and a coreceptor (Cripto/Flag) in *Xenopus* embryos ([@pbio-0020030-Yeo1]; [@pbio-0020030-Cheng2]). We found that Sqt binding to the ActRIIB and Alk4 receptor complex required Cripto ([Figure 5](#pbio-0020030-g005){ref-type="fig"}F). In contrast, ActivinβB and Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ can bind to Activin receptors in the absence of Cripto. Moreover, Cripto did not significantly increase Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ ligand--receptor complex formation. These results indicate that the loop-β8 region is a determinant of TGFβ ligand binding to Activin receptors independent of EGF-CFC coreceptors. Multiple Residues in the Loop-β8 Region Contribute to Coreceptor Independence {#s2g} ----------------------------------------------------------------------------- An alignment of the loop-β8 region of EGF-CFC-dependent and EGF-CFC-independent TGFβs ([Figure 5](#pbio-0020030-g005){ref-type="fig"}A) reveals the presence of several residues unique to Activins. These include (i) a Lys102--X--Asp104 motif (numbering from ActivinβA) that forms a significant binding interface with the type II receptor ActRII ([@pbio-0020030-Wuytens1]; [@pbio-0020030-Greenwald1]; [@pbio-0020030-Thompson1]); (ii) Gln/Pro106 and Asn107, which contribute to the dimerization interface responsible for conformational arrangement ([@pbio-0020030-Thompson1]); and (iii) an Asn insertion at position 99. We therefore mutated the corresponding residues in Sqt, individually or in combination, to the ActivinβB sequence and tested them in the *gsc*/*ntl* induction assay ([Figure 6](#pbio-0020030-g006){ref-type="fig"}). All constructs were active in wild-type embryos. The Sqt3 and Sqt5 constructs containing the Lys102--X--Asp104 motif and Asn99 insertion showed weak expansion of *ntl* expression animally and dorsally in MZ*oep* mutants. The incorporation of Pro106--Asn107 (Sqt2, Sqt4, and Sqt5) in Sqt did not enhance activity in MZ*oep* mutants. These results suggest that multiple residues contribute to coreceptor independence, with the type II receptor-binding interface being an essential determinant. ::: {#pbio-0020030-g006 .fig} Figure 6 ::: {.caption} ###### Conserved Residues in Activin Loop-β8 Region Confer Independence from EGF-CFC Coreceptors Synthetic mRNAs (200 pg) encoding Sqt harboring multiple mutations from ActivinβB (shown in red) were injected into wild-type and MZ*oep* embryos. *gsc* and *ntl* mRNA expression is shown at shield stage; animal pole views are dorsal to the right. Schematic is not drawn to scale. Note that the Sqt3 and Sqt5 constructs containing the Lys102--X--Asp104 motif and Asn99 insertion show weak expansion of *ntl* expression animally and dorsally in MZ*oep* mutants. :::  ::: The Loop-β8 Region in Sqt Is Inhibitory {#s2h} --------------------------------------- The results described above identified the loop-β8 region of Activin as a region that confers coreceptor-independent signaling to ligands that are normally EGF-CFC-dependent. In a reverse set of experiments, we asked which regions confer dependence on EGF-CFC coreceptors. To identify domains that confer EGF-CFC dependence, chimeras between ActivinβB and Sqt (see Figures [4](#pbio-0020030-g004){ref-type="fig"} and [7](#pbio-0020030-g007){ref-type="fig"}) were analyzed for their inability to signal in MZ*oep* mutants. Chimeras containing the Sqt Finger 2 domain (AAS and ASS; see [Figure 4](#pbio-0020030-g004){ref-type="fig"}) or only the Sqt loop-β8 region (Act^Sqt\[loop^ ^β^ ^8\]^; [Figure 7](#pbio-0020030-g007){ref-type="fig"}) were inactive in both wild-type and MZ*oep* embryos. Western blot analysis demonstrated that these chimeras generate stable ligands (data not shown). The addition of Finger 1 in SAS or Act^Sqt\[Finger1-loop^ ^β^ ^8\]^ relieved the inhibitory effect of the loop-β8 region of Sqt in wild-type embryos ([Figure 7](#pbio-0020030-g007){ref-type="fig"}). These chimeras were inactive in MZ*oep* mutants. These results indicate that the loop-β8 region in Sqt acts as an inhibitory domain and that Finger 1 relieves this inhibition by conferring dependence on EGF-CFC coreceptors. ::: {#pbio-0020030-g007 .fig} Figure 7 ::: {.caption} ###### Sequence Determinants Conferring EGF-CFC Dependence Synthetic mRNAs (200 pg) encoding ActivinβB with single or double region substitutions from Sqt were injected into wild-type and MZ*oep* embryos. *gsc* and *ntl* mRNA expression is shown at shield stage; animal pole views are dorsal to the right. Schematic is not drawn to scale. HA indicates a hemagglutinin epitope tag. Note that Act^Sqt\[loop^ ^β^ ^8\]^ containing the loop-β8 region of Sqt is inactive in both wild-type and MZ*oep* embryos. In Act^Sqt\[Finger1-loop^ ^β^ ^8\]^, the additional substitution of Sqt Finger 1 region relieves the inhibitory presence of the Sqt loop-β8 region. Similar to Sqt, Act^Sqt\[Finger1-loop^ ^β^ ^8\]^ can induce ectopic *gsc* and *ntl* in wild-type, but not in MZ*oep* embryos. Western blot analysis indicates that these chimeric constructs produce stable ligands (data not shown). :::  ::: Specificity of Antagonism by Lefty Is Determined by EGF-CFC Coreceptor Dependence {#s2i} --------------------------------------------------------------------------------- Our genetic and biochemical studies suggested that Lefty blocks Nodal and Vg1 signaling via EGF-CFC coreceptors. In contrast, the coreceptor-independent signaling by Activins cannot be blocked by Lefty. This finding predicts that the EGF-CFC-independent chimeric ligands Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^, Sqt^Act^ ^β^ ^A\[loop^ ^β^ ^8\]^, and Vg1^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ should also be resistant to Lefty. Conversely, the coreceptor-dependent chimera Act^Sqt\[Finger1-loop^ ^β^ ^8\]^ should be suspectible to inhibition by Lefty. To test this hypothesis, we coexpressed chimeric ligands and Lefty1 and analyzed *gsc* expression and A3-luc reporter induction ([Figure 8](#pbio-0020030-g008){ref-type="fig"}A--[8](#pbio-0020030-g008){ref-type="fig"}M). As predicted, Lefty1 did not inhibit signaling by Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ ([Figure 8](#pbio-0020030-g008){ref-type="fig"}B and [8](#pbio-0020030-g008){ref-type="fig"}C), Sqt^Act^ ^β^ ^A\[loop^ ^β^ ^8\]^ ([Figure 8](#pbio-0020030-g008){ref-type="fig"}E and [8](#pbio-0020030-g008){ref-type="fig"}F), or Vg1^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ ([Figure 8](#pbio-0020030-g008){ref-type="fig"}H and [8](#pbio-0020030-g008){ref-type="fig"}I), but antagonized Act^Sqt\[Finger1-loop^ ^β^ ^8\]^ ([Figure 8](#pbio-0020030-g008){ref-type="fig"}K and [8](#pbio-0020030-g008){ref-type="fig"}L). These results indicate that the incorporation of the Activin loop-β8 region into Nodal and Vg1 can render these ligands EGF-CFC-independent and therefore resistant to Lefty. ::: {#pbio-0020030-g008 .fig} Figure 8 ::: {.caption} ###### EGF-CFC Coreceptor Depen-dence Determines Susceptibility to Antagonism by Lefty (A--L) Embryos were injected with 75 pg of *Sqt^ActβB\[loopβ8\]^* mRNA (A--C), 75 pg of *Sqt^ActβA\[**loop**β**8**\]^* mRNA (D--F), 200 pg of *Vg1^ActβB\[loopβ8\]^* mRNA (G--I), or 200 pg *Act^Sqt\[Finger1-loopβ8\]^* mRNA (J--L). Embryos were further double-injected with either 500 pg of *LacZ* mRNA (A, D, G, and J), 100 pg of *lefty1* and 400 pg *LacZ* mRNAs (B, E, H, and K), or 500 pg of *lefty1* mRNA (C, F, I, and L). *gsc* mRNA expression in wild-type zebrafish embryos is shown at shield stage, animal pole view. Note that both levels of Lefty1 cannot inhibit the ectopic *gsc* expression induced by Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ (B and C), Sqt^Act^ ^β^ ^A\[loop^ ^β^ ^8\]^ (E and F), and Vg1^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ (H and I). In contrast, Lefty1 can inhibit Act^Sqt\[Finger1-loop^ ^β^ ^8\]^ (K and L). \(M) Wild-type embryos were injected with 75 pg of either Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^, Sqt^Act^ ^β^ ^A\[loop^ ^β^ ^8\]^, Vg1^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^, or 200 pg of Act^Sqt\[Finger1-loop^ ^β^ ^8\]^ mRNA. Embryos were further double-injected with 500 pg of *LacZ* mRNA, 100 pg of *lefty1*, and 400 pg of *LacZ* mRNAs, or 500 pg of *lefty1* mRNA. Smad2 pathway activation was measured by an Activin response element luciferase reporter, A3-luc. Values are folds over wild-type control injected with 500 pg of *LacZ* mRNA and A3-luc reporter. An asterisk indicates a significant difference from the level of activation with ligand and *LacZ* expression alone (Student\'s *t*-test, *p* \< 0.05). :::  ::: Discussion {#s3} ========== Lefty Antagonizes EGF-CFC Coreceptors ------------------------------------- Lefty molecules are key regulators of mesendoderm development and left--right axis determination, but the molecular basis of Lefty-mediated antagonism of Activin-like pathways has been elusive ([@pbio-0020030-Hamada1]; [@pbio-0020030-Schier1]). Our genetic and biochemical studies provide three lines of evidence that Lefty blocks EGF-CFC coreceptors. First, Lefty only inhibits EGF-CFC-dependent TGFβ ligands such as Nodal and Vg1, but not EGF-CFC-independent ligands such as Activins. A striking example of this coreceptor-specific interaction is the finding that changing only 14 amino acids in Nodal or Vg1 to the corresponding residues in Activins renders the resulting TGFβ ligands independent of EGF-CFC coreceptors and resistant to Lefty. Second, the EGF-CFC proteins mouse Cripto and zebrafish Oep can partially suppress the effects of Lefty overexpression in zebrafish. Third, Leftys can bind to EGF-CFC coreceptors and block the coreceptors from interacting with Nodal. Furthermore, Lefty/EGF-CFC complexes seem to exclude interactions with type I and type II Activin receptors. Taken together, these results indicate that Lefty blocks a subset of TGFβ signals by the novel mechanism of blocking pathway-specific coreceptors ([Figure 9](#pbio-0020030-g009){ref-type="fig"}A--[9](#pbio-0020030-g009){ref-type="fig"}D). ::: {#pbio-0020030-g009 .fig} Figure 9 ::: {.caption} ###### Model for EGF-CFC, Activin Receptors, Lefty, and TGFβ Interactions \(A) In the absence of ligands, the EGF-CFC coreceptor (solid pink) is constitutively bound to the type I receptor Alk4 (solid green). \(B) Nodal (solid blue) binds to receptor complexes consisting of EGF-CFC/Alk4 and ActRIIB (solid green). \(C) Lefty (solid yellow) sequesters the EGF-CFC coreceptor, thereby preventing Nodal binding to the receptor complexes. \(D) Subtle sequence differences determine the interaction with the EGF-CFC coreceptor and the Lefty inhibitor. Nodal and Vg1/GDF1 (solid blue) require the EGF-CFC coreceptor for signaling through ActRIIB and Alk4, while Activin (solid red) does not. Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ and Vg1^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^ (solid blue with red strip) containing the loop-β8 region of ActivinβB can bind to ActRIIB and Alk4 without the EGF-CFC coreceptor and therefore cannot be blocked by Lefty. Act^Sqt\[Finger1-loop^ ^β^ ^8\]^ (solid red with two blue strips) requires the coreceptor for receptor complex binding and can be inhibited by Lefty. :::  ::: The observation that Lefty does not block signaling by Activin seems in apparent contrast to previous studies that led to naming some Lefty family members Antivins, for their anti-Activin properties ([@pbio-0020030-Thisse1]; [@pbio-0020030-Cheng1]; [@pbio-0020030-Ishimaru1]; [@pbio-0020030-Tanegashima1]). In particular, it has been found that misexpression of Activin can suppress the defects caused by Lefty misexpression in vivo ([@pbio-0020030-Thisse1]). Our results do not undermine this finding, but suggest an alternative explanation. Previous studies have shown that Activin can suppress the loss of EGF-CFC activity in MZ*oep* mutants ([@pbio-0020030-Gritsman1]; [@pbio-0020030-Cheng2]). Analogously, we suggest that the blocking of EGF-CFC activity by Lefty can be bypassed by Activin, because this ligand can activate Activin receptors independent of co-receptors. A similar scenario can also account for the suppression of Lefty gain-of-function phenotypes by misexpression and activation of Activin receptors ([@pbio-0020030-Meno4]; [@pbio-0020030-Thisse1]; [@pbio-0020030-Sakuma1]). Hence, Activin and Activin receptors bypass the loss of EGF-CFC coreceptor function that is caused either by mutations in *oep* or by overexpression of Lefty. Conversely, Lefty cannot block Activin signals and Activin receptors because of their independence from EGF-CFC coreceptors. Is the block of EGF-CFC coreceptors by Lefty a general and conserved mechanism? Although we have only analyzed a representative subset of these protein families (zebrafish and mouse Lefty1; zebrafish Oep; mouse Cripto; zebrafish Sqt), previous studies have suggested that heterologous Nodal, Lefty, and EGF-CFC proteins have similar activities in zebrafish ([@pbio-0020030-Schier1]). For example, mouse Nodal, mouse Lefty2, and mouse Cripto are active in zebrafish, despite less than 30% overall sequence conservation with their zebrafish counterparts ([@pbio-0020030-Toyama1]; [@pbio-0020030-Meno4]; [@pbio-0020030-Gritsman1]). These studies suggest that the molecular mechanisms described here apply to most, if not all, Nodal/Lefty/EGF-CFC interactions. This does not exclude the possibility that Lefty has additional means of blocking TGFβ signaling. First, Leftys might block the processing of Nodals. However, Sqt is processed normally at levels of Lefty that block Nodal signaling (unpublished data). Second, Leftys might bind Nodal signals. This could result in blocking receptor interactions or antagonizing TGFβ dimerization. However, Sqt is not bound by Lefty at Lefty levels that block Nodal signaling and lead to complex formation with Cripto (unpublished data). Moreover, a Sqt protein containing the loop-β8 region of Activin is resistant to Lefty, whereas changing only the dimerization residues in this region does not confer resistance. Third, Leftys might interact with additional extracellular factors. Indeed, the overexpression of the extracellular domain of the type II receptor ActRIIB has been shown to suppress Lefty activity ([@pbio-0020030-Meno4]). Although zebrafish Lefty1 does not appear to bind to ActRIIB, it is conceivable that overexpression of soluble ActRIIB might protect EGF-CFC coreceptors or another yet-to-be identified protein from antagonism by Lefty. In addition, overexpression of EGF-CFC proteins in zebrafish does not induce dominant phenotypes ([@pbio-0020030-Gritsman1]). It is therefore possible that an additional factor would be required to completely block Lefty in these experiments. Alternatively, overexpression levels of EGF-CFC coreceptors might not be high enough to block Lefty at blastula stages. It is also possible that coreceptor overexpression might block Nodal signals, because it has been shown that EGF-CFC proteins and Nodals can directly interact. The complex feedback interactions between Lefty and Nodal might also overcome an initial reduction of Lefty activity by increasing Lefty transcription. These considerations and the data presented here therefore suggest that a major, but perhaps not exclusive, role of Leftys is to block a subset of TGFβ signals by interaction with EGF-CFC coreceptors. Implications for the Role of Lefty during Development {#s3b} ----------------------------------------------------- Our finding that Leftys can block Vg1 signaling also has important implications for the developmental control of TGFβ signaling. Based on previous studies revealing that Lefty proteins inhibit Nodal signaling, the mouse Lefty mutant phenotypes have been interpreted as a consequence of increased or sustained Nodal signaling ([@pbio-0020030-Hamada1]; [@pbio-0020030-Schier1]). For example, loss of mouse Lefty2 has been thought to increase Nodal signaling, resulting in an enlarged primitive streak ([@pbio-0020030-Meno4]). Similarly, the left--right defects observed in mouse *Lefty1* and left-side specific *Lefty2* (*Lefty2^ΔASE^*) mutants have been thought to be caused by inappropriate spread of Nodal signaling ([@pbio-0020030-Meno3], 2001). Our finding that Vg1/GDF1 signaling can also be blocked by Lefty suggests a more complex scenario. In particular, GDF1 (the mouse homologue of Vg1) is required for proper left--right axis determination ([@pbio-0020030-Rankin1]). *GDF1* mutants appear to have the opposite phenotypes as *Lefty1* and *Lefty2^ΔASE^* mutants. While GDF1 promotes the expression of left-side-specific genes such as *Pitx2* on the left, Leftys appear to block *Pitx2* expression on the right ([@pbio-0020030-Meno3], 2001; [@pbio-0020030-Rankin1]). In light of our findings, we suggest that during left--right axis formation, Leftys act as antagonists of not only Nodal, but also GDF1. In this scenario, loss of Lefty1 or Lefty2 would lead to ectopic and sustained GDF1 signaling. This model is particularly attractive when one considers the expression patterns of *Lefty1*, *Lefty2*, the EGF-CFC gene *Cryptic*, *Nodal*, and *GDF1. Lefty 1* and *GDF1* are coexpressed in the developing midline ([@pbio-0020030-Meno1], 1997; [@pbio-0020030-Rankin1]), whereas *Lefty2* and *Nodal* are coexpressed in left-lateral plate mesoderm ([@pbio-0020030-Conlon2]; [@pbio-0020030-Meno2]). *Cryptic* is expressed in both the lateral plate and midline ([@pbio-0020030-Shen3]). It is therefore conceivable that GDF1 signaling is restricted by Lefty1-mediated inhibition of Cryptic in the midline and its progenitors, whereas Nodal signaling is antagonized by Lefty2-mediated block of Cryptic in the lateral plate. Our results might also have implications for the role of Cripto in tumorigenesis. Cripto is highly overexpressed in human epithelial cancers, such as breast and colon carcinomas ([@pbio-0020030-Salomon1]), and has been implicated in tumor formation ([@pbio-0020030-Ciardiello1], 1994; [@pbio-0020030-Baldassarre1]; [@pbio-0020030-De1]; [@pbio-0020030-Salomon1]; [@pbio-0020030-Adkins1]). Although the mechanisms by which Cripto acts in these circumstances are unclear, inhibition of Cripto by antisense or antibody blockade can inhibit tumor cell proliferation ([@pbio-0020030-Ciardiello2]; [@pbio-0020030-Baldassarre1]; [@pbio-0020030-De1]; [@pbio-0020030-Adkins1]; reviewed by [@pbio-0020030-Shen1]). Since Lefty is an in vivo antagonist of EGF-CFC activity, it might also serve as a therapeutic agent to block Cripto. Subtle Sequence Differences Determine the Interaction with Coreceptors and Inhibitors {#s3c} ------------------------------------------------------------------------------------- The finding that the highly related ligands Activin, Nodal, and Vg1/GDF1 activate the same receptors but differ in their dependence on coreceptors allowed us to determine how ligand diversity and signaling specificity can be achieved. We have identified the loop-β8 region as a 14 amino acid domain, a mere 4% of the entire TGFβ signal, that contributes to coreceptor dependence or independence. Sqt and Vg1 incorporating the loop-β8 region of Activin can bind to the Activin receptors in the absence of EGF-CFC proteins. Conversely, Activin incorporating the loop-β8 region of Sqt is inactive, suggesting that the Nodal/Vg1 loop-β8 region might be inhibitory. This inhibition can be relieved by the Finger 1 domain of Sqt, which results in the dependence on EGF-CFC coreceptors ([Figure 9](#pbio-0020030-g009){ref-type="fig"}D). These results indicate that rather subtle sequence variations can lead to striking changes in ligand diversity. Structural considerations suggest that the loop-β8 region determines coreceptor independence or dependence at least in part by its interactions with type II receptors. The conserved Lys102--X--Asp104 motif in the Activin loop-β8 region has been shown to be important for high-affinity binding to the ActRII receptor ([@pbio-0020030-Wuytens1]; [@pbio-0020030-Greenwald1]; [@pbio-0020030-Thompson1]). In the crystal structure of the ActivinβA-ActRIIB complex, Lys102--X--Asp104 forms an intramolecular salt bridge that interacts with a hydrophobic interface on ActRIIB ([@pbio-0020030-Thompson1]). Mutational analysis has shown that substituting Lys102 with a neutral charge (Ala) significantly reduces receptor binding affinity and signaling ([@pbio-0020030-Wuytens1]). In contrast to Activin, EGF-CFC-dependent ligands such as Nodals and Vg1/GDF1 have the differentially charged residues Met/Leu102 and His104 at the corresponding positions (numbering according to ActivinβA). Similarly, in BMP7 the corresponding residues are Leu102 and Lys104. It has been shown that modeling Lys onto the aligned 102 residue in BMP7 positions it within hydrogen-bonding distance to Glu29 of ActRII and may allow for greater hydrophobic packing at the interface ([@pbio-0020030-Greenwald1]). Analogously, we propose that Sqt^Act^ ^β^ ^A\[loop^ ^β^ ^8\]^, Sqt^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^,Vg1^Act^ ^β^ ^B\[loop^ ^β^ ^8\]^, Sqt3, and Sqt5 are coreceptor-independent because of their favorable binding to ActRIIB receptors. Conversely, the corresponding region in Sqt and Vg1 might be inhibitory because of inefficient interaction with ActRIIB receptors. Detailed structural studies should reveal whether EGF-CFC proteins overcome this inhibition by changing the conformation of Nodal and Vg1 or by providing an additional interaction surface that allows the assembly of receptor complexes. In summary, our results lead to two major conclusions. First, Lefty inhibits a subset of TGFβ signals by using the novel mechanism of blocking pathway-specific coreceptors belonging to the EGF-CFC family. Second, subtle sequence changes in TGFβs determine their signaling specificity and dependence on coreceptors. Although *Drosophila* has an Activin signaling pathway, Nodals, Leftys, and EGF-CFC proteins seem to be restricted to chordates ([@pbio-0020030-Brummel1]; [@pbio-0020030-Schier1]). The evolution of Activin-like signaling pathways therefore represents a remarkable example of how a simple signaling pathway consisting of ligand and receptors can be diversified by subtle sequence changes that modulate the interaction with coreceptors and their inhibitors. Materials and Methods {#s4} ===================== {#s4a} ### Strains and embryos. {#s4a1} Adult homozygous fish for *oep^tz57^* were generated as described previously ([@pbio-0020030-Zhang1]; [@pbio-0020030-Gritsman1]). *Xenopus* embryos were obtained as described in [@pbio-0020030-Hemmati-Brivanlou2]). ### Generation of constructs. {#s4a2} Epitope-tagged and chimeric constructs were made using PCR-based methods and confirmed by sequencing. pCS2-zebrafish Lefty1/HA and Lefty1/Glu constructs were generated by inserting three tandem copies of HA-epitope or Glu-epitope, respectively, after Val145. The initial three-lettered (XXX) Sqt/ActivinβB chimeras were generated by subcloning the prodomain of *Xenopus* ActivinβB (codons Met1 to Gly256) fused to an HA-epitope/XhoI fragment (YPYDVPDYALE) and followed by the mature chimeric ligand into pcDNA3 vector. *S* denotes Sqt; *A* denotes ActivinβB. The boundaries for Sqt mature ligand domains are as indicated: Finger 1 (Asn263 to Cys325), Heel (Pro326 to Cys358), and Finger 2 (Val359 to His392). The boundaries for *Xenopus* ActivinβB mature ligand domains are as indicated: Finger 1 (Cys215 to Cys299), Heel (Pro300 to Cys335), and Finger 2 (Ile336 to Ala370). Full-length chimeras were generated by incorporating the indicated regions into Sqt, ActivinβB, or zebrafish Vg1, which were then subcloned into the pT~7~T~S~ vector ([@pbio-0020030-Ekker1]). The boundaries for Sqt Finger 2 structural subregions are as indicated: β6β7 (Val359 to Try370), loop (Tyr371 to Met376), β8 (Val377 to Gly383), and β9 (Met384 to His392). The boundaries for *Xenopus* ActivinβB Finger 2 structural subregions are as indicated: β6β7 (Ile336 to Try347), loop (Phe348 to Ile354), β8 (Val355 to Asn356), and β9 (Met357 to Ala370). The *Xenopus* ActivinβA loop-β8 region sequence is FDRNNNVLKTDIAD (also identical in *Xenopus* ActivinβD). The zebrafish Vg1 loop-β8 region is from Try332 to Asp345. pcDNA3-zebrafish Vg1/HA, pcDNA3-Squint/HA, pCS2-Alk4(KR)/Flag (a kinase-defective mutant of human Alk4 with Lys234 to Arg234 substitution), pCS2-ActRIIB(KR)/Myc (a kinase-defective mutant of mouse ActRIIB with Lys217 to Arg217 substitution), and pCS2-Cripto/Flag have been described elsewhere ([@pbio-0020030-Yeo1]; [@pbio-0020030-Cheng2]). ### Embryo microinjection. {#s4a3} Plasmids were linearized and sense strand-capped mRNA was synthesized using the mMESSAGE mMACHINE system (Ambion, Austin, Texas, United States). Zebrafish embryos were dechorinated by pronase treatment and injected between the one- and four-cell stage. *Xenopus* embryos at the one- to two-cell stage were used for injections into the animal pole. ### Phenotypic analysis. {#s4a4} Zebrafish embryos at 24 h were mounted in 2% methylcellulose and photographed using a Zeiss (Oberkochen, Germany) M2Bio dissecting microscope. In situ hybrization was performed as described previously ([@pbio-0020030-Thisse2]), using RNA probes to *gsc* and *ntl* ([@pbio-0020030-Stachel1]; [@pbio-0020030-Schulte-Merker1]). ### Luciferase reporter assay. {#s4a5} Luciferase assays were performed with three to six samples and five embryos in each sample. Results are representative of three independent experiments. The injection mixtures were equalized with respect to total mRNA amount with *LacZ* mRNA. The A3-luc reporter DNA construct (25 pg) ([@pbio-0020030-Chen1]) was also coinjected. Whole zebrafish embryos were harvested at shield stage. Luciferase activity was analyzed using the Luciferase Reporter Assay system (Promega, Madison, Wisconsin, United States) according to the manufacturer\'s instruction in a Lumat LB9501 (Berthold Technologies, Bad Wildbad, Germany). Owing to the technical aspects of microinjections, in rare circumstances, a single outlier was statistically removed from a population using Grubbs\' test/extreme studentized deviate method. Inclusion of outliers into the populations does not change the statistical significance of the *p* values; that is, *p* remains \<0.05, where indicated. ### Coimmunoprecipitation analysis. {#s4a6} *Xenopus* embryos were harvested at stage 10. For chemical cross-linking of proteins, animal halves were incubated in PBS with 10 mM 3,3′-dithiobis(sulfo-succinimidyl propionate) (DTSSP) (Pierce Biotechnology, Rockford, Illinois, United States) and incubated for 2 h on ice. Coimmunoprecipitation was performed as described previously ([@pbio-0020030-Yeo1]). Purified processed mouse Lefty1, soluble mouse Cripto, and mouse VEGF-D proteins were obtained from R&D Systems (Minneapolis, Minnesota, United States). Activity assays were performed by R&D Systems. The proteins were incubated in PBS with 1 mM DTSSP for 1 h on ice. Coimmunoprecipitation was performed as described previously ([@pbio-0020030-Yeo1]). Samples were treated with 100 mM DTT to cleave DTSSP prior to SDS-PAGE analysis. The following antibodies were used for immunoprecipitation and Western blot analysis: anti-Flag mouse monoclonal antibody (clone M2; Sigma, St. Louis, Missouri, United States), anti-HA mouse monoclonal antibody (clone 16B12; Covance, Princeton, New Jersey, United States), anti-HA rabbit polyclonal antibody (Y-11; Santa Cruz Biotechnology, Santa Cruz, California, United States), anti-c-Myc rabbit polyclonal antibody (A-14; Santa Cruz Biotechnology), anti-c-Myc mouse monoclonal antibody (clone Ab-1; Oncogene Science, Tarrytown, New York, United States), anti-His mouse monoclonal antibody (clone 6-His; Covance), anti-mLefty1 goat polyclonal antibody (R&D Systems), and anti-Glu mouse monoclonal antibody (clone Glu-Glu; Covance). Proteins were visualized using the Super Signal West Pico/Femto Chemiluminescent Substrate system (Pierce). Supporting Information {#s5} ====================== Accession Numbers {#s5a} ----------------- The GenBank (<http://www.ncbi.nlm.nih.gov/entrez/query.fc-gi?db=Nucleotide>) accession numbers for the sequences discussed in this paper are ActivinβA (Q9W6I6), ActivinβB (Q91350), human ActRIIB (P08476), mouse ActRIIB (P27040), rat ActRIIB (P38445), Alk4 (Z22536), BMP7 (P23359), Cripto (P51865), Cryptic (P97766), *cyclops* (P87358), GDF1 (P20863), *goosecoid* (P53544), Lefty1 (Q9W6I6), Lefty2 (P57785), mLefty1 (Q64280), Nodal (P43021), *no tail* (Q07998), *one-eyed pinhead* (O57516), *Pitx2* (P97474), *squint* (O13144), VEGF-D (P97946), and Vg1 (P09534). We thank members of the Schier, Brivanlou, and Yelon laboratories for discussions; Gord Fishell, Kathy Joubin, Alvaro Sagasti, and Will Talbot for comments on the manuscript; Steven Zimmerman, Trisha Bruno, and Nicole Dillon for fish care; Chang-Yeol Yeo and Malcolm Whitman for sharing of plasmids and protocols; and Matthieu Schapira for aiding with structural modeling. SKC was supported in part by National Institutes of Health (NIH) training grant T32HD07520. AHB is supported by the NIH. AFS is a Scholar of the McKnight Endowment Fund for Neuroscience, a Irma T. Hirschl Trust Career Scientist, and an Established Investigator of the American Heart Association and is supported by grants from the NIH. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** SKC, AHB, and AFS conceived and designed the experiments. SKC and FO performed the experiments. SKC, FO, AHB, and AFS analyzed the data. SKC, FO, AHB, and AFS contributed reagents/materials/analysis tools. SKC and AFS wrote the paper. DOI: [10.1371/journal.pbio.0020030](10.1371/journal.pbio.0020030) Copyright: © 2004 Cheng et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Academic Editor: Hiroshi Hamada, Osaka University ActβA : ActivinβA ActβB : ActivinβB ActRIIB : Activin receptor type IIB Alk4 : Activin receptor-like kinase-4 DTSSP : 3,3′-dithiobis(sulfo-succinimidyl propionate) EGF-CFC : epidermal growth factor--Cripto/FRL-1/Cryptic GDF1 : growth and differentiation factor-1 *gsc* : *goosecoid* GPI : glycosylphosphatidylinositol HA : hemagglutinin hpf : hours postfertilization kDa : kilodalton *LacZ* : E. coli β-galactosidease luc : luciferase MAPK : mitogen-activated protein kinase MZ*oep* : maternal- and zygotic-deficient mutant of *oep* *ntl* : *no tail* *oep* : *one-eyed pinhead* *sqt* : *squint* TGFβ : transforming growth factor-β VEGF-D : vascular endothelial growth factor-D

PubMed Central

2024-06-05T03:55:46.071230

2004-2-17

PMC340942

Introduction {#s1} ============ Polyketides are a large class of structurally and pharmacologically diverse molecules, including many antibiotics and antitumor drugs ([@pbio-0020031-OaHagan1]). They are produced as secondary metabolites primarily by bacteria and fungi ([@pbio-0020031-Hopwood1]). Analogous to fatty acid synthases (FASs), polyketide synthases (PKSs) catalyze the biosynthesis of polyketides through repetitive C--C bond-forming reactions between selected acyl-CoA-derived building blocks ([@pbio-0020031-Cane1]). However, in contrast to fatty acid biosynthesis, the carbon chain backbones of polyketides exhibit greater variety with respect to the choice of acyl-CoA building blocks and the degree of reduction of β-ketone functional groups that result after each round of chain elongation. In complex polyketides, such as the macrolide erythromycin, biosynthetic variability arises from independent control of each round of chain elongation by one module of enzymes within a multimodular PKS ([@pbio-0020031-Cane1]). (The term *module* used in this report refers to a collection of dissociated enzymes. The elongation module consists of enzymes involved in chain extension steps of polyketide biosynthesis, while the initiation module consists of enzymes involved in the nonacetate priming of certain aromatic PKSs.) However, the polyketide backbones of most bacterial aromatic polyketides ([Figure 1](#pbio-0020031-g001){ref-type="fig"}) are synthesized by a single dissociated enzymatic module comprised of a heterodimeric ketosynthase--chain length factor (KS-CLF) complex that catalyzes chain initiation and iterative elongation, an acyl-carrier protein (ACP) that shuttles malonyl extender units to the active site of KS-CLF as malonyl-S-ACP intermediates, and a malonyl-CoA:ACP acyl transferase (MAT), which catalyzes acyl transfer between malonyl-CoA and the ACP and is shared between the PKS and the housekeeping FAS(s) ([@pbio-0020031-Revill1]; [@pbio-0020031-Khosla2]). For example, the minimal PKS from the actinorhodin (*act*) biosynthetic pathway synthesizes an octaketide (C~16~) backbone from eight malonyl-CoA equivalents, whereas the tetracenomycin (*tcm*) minimal PKS synthesizes a decaketide (C~20~) backbone from ten equivalents of malonyl-CoA ([Figure 2](#pbio-0020031-g002){ref-type="fig"}) Although a large number of "unnatural" natural products have been engineered to date by genetic manipulation of bacterial aromatic PKSs ([@pbio-0020031-Khosla1]; [@pbio-0020031-Rawlings1]; [@pbio-0020031-Shen1]), most of this variety has resulted from the combinatorial manipulation of ketoreductases (KRs), aromatases (AROs), and cyclases (CYCs) that ordinarily interact with minimal PKS subunits to channel the exceptionally high reactivity of a poly-β-ketone intermediate ([Figure 2](#pbio-0020031-g002){ref-type="fig"}) into the observed natural product ([@pbio-0020031-McDaniel5]). The development of generally applicable methods for chemo- and regioselective modification of natural and unnatural bacterial aromatic polyketides is an important goal for the medicinal chemist and, more recently, the biosynthetic engineer. ::: {#pbio-0020031-g001 .fig} Figure 1 ::: {.caption} ###### Examples of Aromatic Polyketides Actinorhodin (S. coelicolor) and tetracenomycin (S. glaucescens) are primed by acetate groups through decarboxylation of malonyl-CoA. Actinorhodin and tetracenomycin are synthesized by the *act* and *tcm* PKSs, respectively. Oxytetracycline (S. rimosus), frenolicin (*S. roseofulvus)*, R1128 (*S.* R1128), and doxorubicin (S. peucetius) are primed by nonacetate units as shown. Notably, R1128 family of compounds (a--d) are primed with different alkyl units. Oxytetracycline, frenolicin, R1128, and daunorubicin are synthesized by the *otc*, *frn*, R1128, and *dxr* PKSs, respectively. Actinorhodin and tetracenomycin represent much-studied models of aromatic polyketide biosynthesis. Oxytetracycline is a commonly prescribed antibiotic. Frenolicin is a potent antiparasitic agent. R1128 is an estrogen receptor antagonist that shows minimal agonist activity. Doxorubicin is a widely used anticancer drug in treating late-stage tumors. :::  ::: ::: {#pbio-0020031-g002 .fig} Figure 2 ::: {.caption} ###### Biosynthesis of Acetate-Primed Polyketides \(A) Minimal PKS is necessary and sufficient for the synthesis of a complete polyketide chain. KS-CLF is the condensing enzyme in the minimal PKS, catalyzing each round of condensation between malonyl-ACP and the growing polyketide chain. ACP serves as the carrier for malonyl units, and it is malonylated by the MAT associated with FAS. Chain synthesis initiates with the decarboxylation of malonyl-ACP to acetyl-ACP by the KS-CLF for most aromatic PKSs. The acetyl unit is then transferred to the KS-CLF and primes the enzyme for subsequent condensations. The overall chain length is controlled by the KS-CLF complex. An octaketide synthase (e.g., *act* PKS) uses a total of eight malonyl equivalents (including the primer), while a decaketide synthase (e.g., *tcm* PKS) uses a total of ten malonyl equivalents. \(B) An octaketide can spontaneously form SEK4 and SEK4b in the absence of tailoring enzymes. Members of the *act* KR family can regioselectively reduce the octaketide at C-9, which can then spontaneously form mutactin in the absence of AROs and CYCs. When bifunctional ARO/CYC (e.g., *act*VII) and second-ring CYC (e.g., *act*IV) are present, the reduced octaketide can be transformed into the anthraquinone DMAC. \(C) A decaketide can spontaneously form SEK15 and SEK15b in the absence of tailoring enzymes. KR can regioselectively reduce the C-9 carbonyl. A reduced decaketide can spontaneously form RM20, RM20b, and RM20c. Not shown are the other tailoring enzymes associated with decaketides, which can transform the nascent decaketide into tetracycline or anthracycline structures. :::  ::: The primer unit of a polyketide backbone is an attractive site for introducing unnatural building blocks. For example, genetic and chemobiosynthetic approaches have been devised to replace the natural primer units in the polyketide backbones of erythromycin, avermectin, and rapamycin with a broad range of unnatural functional groups ([@pbio-0020031-Jacobsen1]; [@pbio-0020031-Marsden1]; [@pbio-0020031-Lowden1]; [@pbio-0020031-Long1]). However, most aromatic PKSs initiate polyketide biosynthesis through decarboxylation of malonyl-ACP, resulting in an invariant acetyl primer unit ([Figure 2](#pbio-0020031-g002){ref-type="fig"}A) ([@pbio-0020031-Bisang1]1999; [@pbio-0020031-Dreier1]). Important antitumor antibiotics, such as the anthracycline doxorubicin, are primed with propionate units. Recently, we have investigated the biosynthesis of the estrogen receptor antagonist R1128 ([@pbio-0020031-Hori1]) and the antiparasitic agent frenolicin ([@pbio-0020031-Bibb1]), two aromatic polyketides that are apparently derived from nonacetate primer units (see [Figure 1](#pbio-0020031-g001){ref-type="fig"}). These bimodular PKSs are comprised of a dissociated initiation module consisting of a homodimeric KS (ZhuH, named for R1128 PKS), an acyl transferase (AT) (ZhuC), and an ACP (ZhuG), and an elongation module consisting of a KS-CLF (ZhuA--ZhuB), a second ACP (ZhuN), and the MAT (borrowed from the housekeeping FAS). The proposed biosynthetic mechanism of the R1128 PKS is shown in [Figure 3](#pbio-0020031-g003){ref-type="fig"} ([@pbio-0020031-Marti1]). Biochemical studies have revealed that the KS subunits of the initiation and elongation modules have specific protein--protein interactions with ACPs from the same module ([@pbio-0020031-Tang1]), suggesting that it may be possible to functionally coexpress these initiation modules with heterologous minimal PKSs, so as to regioselectively incorporate nonacetate primer units into aromatic polyketides. Of particular interest is the R1128 initiation module, since it is known to have broad substrate specificity ([@pbio-0020031-Meadows1]) and the X-ray crystal structure of its KS subunit has been solved ([@pbio-0020031-Pan1]). Here we demonstrate the biosynthetic utility of the R1128 initiation module by synthesizing a variety of anthraquinone antibiotics, some with significant biological activities. In the course of these studies, fundamentally novel and unanticipated properties of bacterial aromatic PKSs have been elucidated. ::: {#pbio-0020031-g003 .fig} Figure 3 ::: {.caption} ###### Proposed Priming Mechanisms for R1128 PKS An independent loading module consisting of ZhuG, ZhuH, and ZhuC can generate an alkylacyl-ZhuG intermediate (boxed) from malonyl-CoA and short chain acyl-CoAs such as propionyl-CoA and isobutyryl-CoA. The precursor selectivity is determined by the KSIII analog ZhuH. Ketoreductase, dehydratase, and enoylreductase associated with FAS are presumed to transformed the β-ketoacyl-ZhuG moiety into alkylacyl-ZhuG. The alkylacyl-ZhuG is then able to prime the minimal PKS module (consisting of the ZhuB \[KS\], ZhuA \[CLF\], ACP \[ZhuN\], and MAT) and initiate polyketide synthesis. The mechanism by which the transacylation occurs is not known and is possibly catalyzed by unassigned, but essential, enzyme ZhuC. Homologs of ZhuG, ZhuH, and ZhuC are present in the *frn* PKS (FrnJ, FrnK, and FrnI, respectively) as well. :::  ::: Results {#s2} ======= Using the host/vector system first described by [@pbio-0020031-McDaniel1]), several combinations of initiation modules, minimal PKSs, and auxiliary PKS subunits were coexpressed and analyzed. Streptomyces coelicolor CH999, which contains a chromosomal deletion of the entire *act* gene cluster, was used as the host strain, whereas the shuttle vector pRM5 was used as an expression plasmid. The polyketide product profiles of the recombinant strains are summarized in [Table 1](#pbio-0020031-t001){ref-type="table"}. ::: {#pbio-0020031-t001 .table-wrap} Table 1 ::: {.caption} ###### Plasmids Constructions and Resulting Polyketide Products :::  Each plasmid is a derivative of pRM5. The constructs are transformed into S. coelicolor strain CH999. Produced are analyzed by LC/MS and NMR ^a^LM: Loading module, consists of ZhuC, ZhuH, and ZhuG ^b^The minimal PKS consists of the indicated KS-CLF, ZhuN, and the endogenous MAT ^c^This construct contains the *act* KS-CLF, but lacks the minimal ACP ZhuN ^d^NP: no major polyketide product observed ^e^Number in parenthesis indicates yield of the product as percentage of total polyketide recovered ::: Recombination of an Initiation Module and a Heterologous Minimal PKS {#s2a} -------------------------------------------------------------------- Guided by our recent observation that the KS--ACP pairs of initiation and elongation PKS modules have orthogonal molecular recognition features, we first attempted to coexpress the R1128 initiation module with the *act* minimal PKS. The *zhuC* (AT)*, zhuH* (KS), *and zhuG* (ACP) genes from the R1128 gene cluster were coexpressed with the genes encoding the *act* KS-CLF, ZhuN (ACP), and the *act* KR (on plasmid pYT46). Control plasmids pYT44, lacking the R1128 initiation module, and pYT45, lacking *zhuN*, were also constructed and characterized. All three plasmids were introduced into S. coelicolor CH999 by transformation, and polyketides products were analyzed by liquid chromatography/mass spectrometry (LC/MS) and nuclear magnetic resonance (NMR) spectroscopy. S. coelicolor CH999/pYT44 produced mutactin (and its dehydrated derivative, dehydromutactin), the expected products of the *act* minimal PKS in the presence of *act* KR ([@pbio-0020031-McDaniel3]), whereas CH999/pYT45 did not produce any polyketide, consistent with the requirement for separate ACPs to support turnover of the initiation and elongation modules in a bimodular aromatic PKS ([Figure 4](#pbio-0020031-g004){ref-type="fig"}A) ([@pbio-0020031-Tang1]). Remarkably, CH999/pYT46 produced two new polyketide products in addition to mutactin; these new polyketides were isolated with a combined yield of 40 mg/l, representing 70% of total polyketides produced by this host. The two compounds had molecular masses of 278 and 292. (The 14 mass unit difference is suggestive of one methylene unit difference between the two compounds.) Isotopic labeling studies indicated that the compounds were derived from six acetate equivalents. NMR analyses ([Table 2](#pbio-0020031-t002){ref-type="table"}) revealed that the two compounds, YT46 (**1**) and YT46b (**2**), had structures shown in [Figure 4](#pbio-0020031-g004){ref-type="fig"}. YT46 and YT46b are identical, with the exception of a branched methyl group in YT46b. Each compound contains an α-pyrone moiety, which is commonly observed in aberrantly cyclized aromatic polyketides ([@pbio-0020031-Yu2]). The C-9 carbonyl groups were selectively reduced by the *act* KR. Supplementing the growth medium with^13^C-labeled sodium propionate revealed that the alkyl moiety of **1** was derived from propionyl-CoA. Similarly, supplementing the growth medium with 1 g/l of L-valine resulted in a 2-fold increase in the level of **2**, suggesting the branched alkyl group observed in **2** was derived from isobutyryl-CoA, a primary catabolite of L-valine ([@pbio-0020031-Zhang1]). These findings are consistent with in vitro characterization of the substrate specificity of ZhuH ([@pbio-0020031-Meadows1]). Among substrates tested, ZhuH had an 11-fold higher specificity for both propionyl-CoA and isobutyryl-CoA over the next best substrate, acetyl-CoA ([@pbio-0020031-Meadows1]). YT46 analogs generated by incorporation of isovaleryl and butyryl primer units were also detectable by LC/MS, although these compounds were present at lower levels. Dehydrated analogs of **1**and **2**were also observed in CH999/pYT46. (Purified **1**and **2**slowly dehydrate at room temperature.) ::: {#pbio-0020031-g004 .fig} Figure 4 ::: {.caption} ###### Engineered Biosynthesis of YT46 (1) and YT46b (2) \(A) HPLC trace of extracts from CH999/pYT44-pYT46. A linear gradient of 20%--60% CH~3~CN in water was used. (Left) CH999/pYT44, which only has the minimal PKS module from *act* PKS and KR, produced the expected mutactin (and the dehydrated dehydromutactin). CH999/pYT45, which contains the R1128 loading module and an incomplete minimal PKS, produced no major polyketides. (Right) Upon coexpressing the R1128 loading module with a functional *act* minimal PKS (in the presence of KR), two new major polyketides were produced with the indicated masses. When 1 g/l of valine was added to the growth medium, the yield of the compound with mass of 292 doubled (traces not drawn to the same scale). The two compounds are identified as YT46 and YT46b. \(B) Engineered biosynthesis of YT46 and YT46b. YT46 (1) is derived from propionate, while YT46b (2) is derived from isobutyrate. Alkylacyl-ZhuG supplied by the loading module is able to prime the *act* minimal PKS efficiently. Incorporation of the alkylacyl moiety by the KS-CLF led to a decrease in the number of extender units incorporated in the final chain. The octaketide synthase is now only able to synthesize an alkyl hexaketide. The KR is able to regioselectively reduce the alkyl-hexaketide at the expected C-9 position. The reduced hexaketide spontaneously form the novel bicyclic structure observed in 1 and 2. Dehydrated versions of 1 and 2 are also observed (outside of limits shown in \[A\]). :::  ::: ::: {#pbio-0020031-t002 .table-wrap} Table 2 ::: {.caption} ###### Proton and Carbon NMR Data for YT46 (**1**) and YT46b (**2**) :::  Spectra were obtained at 400 MHz for proton and 100 MHz for carbon and were recorded in acetone-*d* ~6~ ^a^Carbons are labeled as shown in [Figure 4](#pbio-0020031-g004){ref-type="fig"}B ^b^1, 2-^13^C-acetate labeling experiments were performed with 2 and the observed carbon--carbon coupling constants are shown in parenthesis ::: The biosynthesis of **1**and **2** by a recombinant bimodular PKS consisting of the *act* minimal PKS and the R1128 initiation module supported our hypothesis that nonacetate-primed polyketides could be biosynthesized by combinatorial expression of heterologous initiation and elongation PKS modules from bacterial aromatic PKSs. Indeed, the *act* KS-CLF, which is normally primed exclusively by acetate (generated via decarboxylation of a malonyl unit), has a remarkably strong preference for the diketide product of the R1128 initiation module. However, unexpectedly, the incorporation of a longer chain substrate into the catalytic cycle of the *act* minimal PKS results in a reduced number of malonyl units utilized during iterative chain extension ([Figure 4](#pbio-0020031-g004){ref-type="fig"}B). Compound **1**is a hexaketide whose backbone consists of 15 C atoms. Thus, upon introduction of an initiation PKS module into the overall catalytic cycle, the octaketide synthase retains its carbon chain length specificity, rather than executing the normal number of extension cycles. In contrast, the *act* KR retains its selectivity for the C-9 carbonyl in the nascent polyketide backbone, notwithstanding structural differences between a hexaketide and an octaketide. In Vivo Reconstitution of R1128 Biosynthesis Using a Heterologous Bimodular PKS {#s2b} ------------------------------------------------------------------------------- To validate the generality of the above observations, we sought to reconstitute the biosynthesis of R1128 family of compounds in S. coelicolor using a heterologous combination of initiation and elongation PKS modules ([Figure 5](#pbio-0020031-g005){ref-type="fig"}). It follows from the above analysis that to synthesize an alkyl-primed anthraquinone such as **3** and **4**, the following catalytic components are needed: (i) an initiation module; (ii) a decaketide minimal PKS that can extend the five to six carbon products of the initiation module by seven more extender units to yield a 19--20 carbon polyketide; and (iii) appropriate ARO and CYC subunits. ::: {#pbio-0020031-g005 .fig} Figure 5 ::: {.caption} ###### Engineered Biosynthesis of TMAC (7), YT128 (3), and YT128b (4) \(A) ZhuI and ZhuJ are CYCs specific for unreduced octaketides. CH999/pYT105, which coexpressed ZhuI and ZhuJ from the R1128 PKS with the *act* minimal PKS, produced the anthraquinone compound TMAC (7). ZhuI and ZhuJ are thus able to cyclize unreduced octaketides. Previously characterized octaketide CYCs *act* ARO/CYC and *act* CYC are specific for reduced octaketides only. ZhuI and ZhuJ are used for reconstituting R1128 biosynthesis. \(B) Reconstitution of R1128 biosynthesis using the heterologous combination of *tcm* minimal PKS, R1128 loading module, and CYCs ZhuI and ZhuJ. The alkylacyl-ZhuG intermediate synthesized by the loading module is able to prime the decaketide synthase from *tcm* minimal PKS. Owing to backbone size restrictions, the *tcm* KS-CLF primed with the alkylacyl groups are only able to extend the polyketide by seven additional malonyl units, resulting in an alkyl-octaketide. ZhuI and ZhuJ are able to transform the unreduced octaketide into YT128 (3) and YT128b (4). The decarboxylated versions of 3 and 4, which are R1128b (5) and R1128c (6), respectively, are also observed in the extracts of CH999/pYT128. :::  ::: The R1128 family of antibiotics represents a unique set of anthraquinones that contain an unreduced C-9 carbonyl (present as an enolic C-9 hydroxyl in R1128). Since members of the *act* ARO/CYC family are unable to cyclize an unreduced octaketide ([@pbio-0020031-McDaniel4]) and since enzymes from the *tcm* ARO/CYC family have alternative regiospecificity of cyclization ([@pbio-0020031-McDaniel5]), AROs and CYCs were sought from the R1128 biosynthetic pathway. ZhuI and ZhuJ are two putative enzymes present in the R1128 PKS ([@pbio-0020031-Marti1]). ZhuI, which is homologous to the *act* ARO/CYC, was predicted to be a first ring CYC, while ZhuJ was predicted to be a second ring CYC. To test these hypotheses, plasmids pYT105 and pYT92 (see [Table 1](#pbio-0020031-t001){ref-type="table"}) were constructed, coexpressing ZhuI and ZhuJ with the *act* minimal PKS and the *tcm* minimal PKS, respectively. Analysis of compounds produced by CH999/pYT92 revealed the decaketides SEK15 and SEK15b as the major products, suggesting ZhuI and ZhuJ did not recognize an unreduced decaketide. However, the anthraquinone compound ([@pbio-0020031-Bartel1]), 3,6,8-trihydroxy-1-methylanthraquinone-2-carboxylic acid (TMAC) (**7**; also known as laccaic acid D, a well known plant-derived pigment) ([Figure 5](#pbio-0020031-g005){ref-type="fig"}A; [Table 2](#pbio-0020031-t002){ref-type="table"}; atoms are numbered according to order in polyketide backbone), was isolated from CH999/pYT105 at 10 mg/l, in addition to the known products of *act* minimal PKS, SEK4 and SEK4b (20 mg/l). Thus, ZhuI and ZhuJ are able to cyclize an unreduced octaketide into the corresponding anthraquinone. The incomplete transformation of nascent octaketides into TMAC may be due to the fact that an acetate-primed octaketide is not a natural substrate of the two CYCs (see below). The identification of ZhuI and ZhuJ as the appropriate CYCs for the synthesis of R1128-like anthraquinones prompted the design of pYT128, which coexpresses *tcm* KS-CLF, ZhuN, ZhuI, ZhuJ, and the R1128 initiation module. Plasmid pYT92 (which lacks the initiation module) was used as the negative control. In addition to the decaketides SEK15 and SEK15b, two new anthraquinone compounds, YT128 (**3**) and YT128b (**4**), were isolated at comparable levels (7 mg/l each) from S. coelicolor CH999/pYT128. The two compounds account for 50% of total polyketides produced by this recombinant strain. \[^13^C\]Propionate and valine feeding experiments verified the alkyl groups installed at C-16 were indeed derived from either propionyl-CoA or isobutyryl-CoA. NMR and MS analyses confirmed the identities of **3**and **4**as alkyl-primed TMAC analogs ([Table 3](#pbio-0020031-t003){ref-type="table"}). The natural products, R1128b (**5**) and R1128c (**6**) (i.e., decarboxylated derivatives of **3** and **4**), were present at an approximately 20% level to that of **3** and **4**. ::: {#pbio-0020031-t003 .table-wrap} Table 3 ::: {.caption} ###### Proton and Carbon NMR Data for TMAC (**7)**, YT128 (**3**), and YT128b (**4**) :::  Spectra were obtained at 400 MHz for proton and 100 MHz for carbon and were recorded in methanol-*d* ~4~ ^a^Carbons are labeled as shown in [Figure 5](#pbio-0020031-g005){ref-type="fig"} ::: The engineered biosynthesis of compounds **3**--**6** further validated our hypothesis that the initiation module of the R1128 PKS could productively interact with elongation modules from any bacterial aromatic PKS. Moreover, by incorporating a primer unit with at least five carbon atoms, the *tcm* KS-CLF effectively became an octaketide synthase with respect to the rest of the polyketide molecule. This was analogous to the conversion of the *act* KS-CLF into a hexaketide synthase in the presence of the R1128 initiation module. It also suggested that the R1128 KS-CLF was intrinsically a decaketide synthase. Finally, analogous to the *act* KR, ZhuI and ZhuJ were programmed to recognize full-length polyketide chains based upon the number of β-carbonyl groups, rather than the carbon chain length of the backbone. Our findings below suggest this is a general property for all CYCs. It should be noted that the CYCs ZhuI and ZhuJ were apparently more efficient in processing the unreduced alkyl-primed octaketide than an acetate-primed octaketide, since no alkyl-primed analogs of SEK4 and SEK4b were observed in extracts of CH999/pYT128. (In the absence of ZhuI and ZhuJ, alkyl-primed versions of both SEK4 and SEK4b are the major polyketides produced \[data not shown\]). Engineered Biosynthesis of Novel Anthraquinones Using Bimodular PKSs {#s2c} -------------------------------------------------------------------- To demonstrate the utility of hybrid bimodular PKSs for the rational design of new analogs of known polyketides, we targeted the engineered biosynthesis of alkyl-primed 3,8-dihydroxy-methylanthraquinone carboxylic acid (DMAC) analogs **8** and **9** ([Figure 6](#pbio-0020031-g006){ref-type="fig"}). Specifically, we inserted the *act* KR gene into pYT128, along with replacing *zhuI* and *zhuJ* with genes encoding *act* ARO and *act* CYC, to arrive at the plasmid pYT127. We rationalized that the *act* CYCs should be able to recognize the reduced, alkyl-primed octaketide. A control plasmid (pYT90) lacking the R1128 initiation module was also constructed, transformed, and analyzed. ::: {#pbio-0020031-g006 .fig} Figure 6 ::: {.caption} ###### Engineered Biosynthesis of DMAC Analogs YT127 (8) and YT127b (9) When the bimodular PKS containing *tcm* minimal PKS and R1128 loading module are coexpressed with tailoring enzymes KR, *act* ARO/CYC, and *act* CYC, the desired compounds 8 and 9 were produced. The decarboxylated versions of 8 and 9 are 10 and 11, respectively. All three tailoring enzymes are able to process the unnatural alkyl-octaketide. :::  ::: In the absence of the initiation module, the *tcm* minimal PKS outfitted with the *act* KR produced the expected polyketides RM20, RM20b, and RM20c ([@pbio-0020031-McDaniel1]). The targeted anthraquinone carboxylic acids **8** and **9** were isolated at high titers (15 mg/l each, 70% of total polyketide products) in CH999/pYT127. The identities of **8**and **9**were verified by NMR and MS ([Table 4](#pbio-0020031-t004){ref-type="table"}). Decarboxylated analogs of both compounds were also observed; these compounds are alkyl-primed analogs of the natural product aloesaponarin II. These findings confirmed that, analogous to ZhuI and ZhuJ, the *act* KR, *act* ARO, and *act* CYC were able to process the octaketide intermediate possessing unnatural functional groups at C-16. Thus, it appears that the substrate recognition features of all auxiliary PKS subunits have evolved to monitor the number of β-ketone functional groups present in the polyketide chain. ::: {#pbio-0020031-t004 .table-wrap} Table 4 ::: {.caption} ###### Proton and Carbon NMR Data for YT127 (**8**) and YT127b (**9**) :::  Spectra were obtained at 400 MHz for proton and 100 MHz for carbon and were recorded in methanol-*d* ~4~ ^a^Carbons are labeled as shown in [Figure 6](#pbio-0020031-g006){ref-type="fig"} ::: Cytotoxic Properties of Novel Anthraquinones {#s2d} -------------------------------------------- As described above, the biosynthetic engineering methods reported here have yielded practical routes for the production of several new as well as known anthraquinone compounds. Given the track record of this family of natural products as pharmacologically active molecules, compounds **3**,**4**, **9**, and DMAC were assayed for cytotoxic activities against human mammary adenocarcinoma MCF-7 cells. Apoptosis was observed after 24 h of drug treatment, and IC~50~ values were recorded after 5 d of drug addition. The IC~50~ values for reduced compounds DMAC and **9**are 26.9 and 21.7 μg/ml, respectively, while the IC~50~ values for unreduced anthraquinones **3**and **4**are 3.4 and 1.7 μg/ml, respectively. Thus inserting the hydroxyl group at C-9 results in a 10-fold increase in cytotoxic activity. The new compounds also show modest improvement in cytotoxic activity relative to the natural products **5**(R1128b, IC~50~ = 9.5 μg/ml) and **6** (R1128c, IC~50~=6.2 μg/ml) ([@pbio-0020031-Hori1]), suggesting an additive effect of both the C-9 OH and C-2 COOH groups. ### Inhibition of glucose-6-phosphatase. {#s2d1} Recently, the natural product mumbaistatin ([Figure 7](#pbio-0020031-g007){ref-type="fig"}A) was identified as an extremely potent inhibitor (IC~50~= 5 nM) of the glucose-6-phosphate translocase enzyme complex, an attractive target for the treatment of Type II diabetes ([@pbio-0020031-Vertesy1]). The core of mumbaistatin consists of an anthraquinone moiety that is related to several engineered compounds discussed in this report. For example, the carboxylic acid at position C-1 and the reduced C-9 are identical to those present in compounds DMAC, **8**, and **9**. We tested the inhibitory activities of some of these polyketides against glucose-6-phosphate translocase using intact male rat liver microsomes. ::: {#pbio-0020031-g007 .fig} Figure 7 ::: {.caption} ###### Inhibition of Glucose-6-Phosphate Translocase Activity by Anthraquinones \(A) The chemical structure of mumbaistatin. Mumbaistatin is an extremely potent inhibitor of glucose-6-phosphate translocase (IC~50~ = 5 nM). The core of the molecule is a reduced carboxylic acid containing anthraquinone, which is also observed in compounds such as DMAC, YT127, and YT127b. \(B) Inhibition of glucose-6-phosphate translocase activity by novel anthraquinones. The inhibition assay is performed as described in the experimental section. The control contains the rat liver microsome and glucose-6-phosphate only. Chlorogenic acid (IC~50~ = 0.26 mM) was used as a reference. The G6Pase activity of the microsome in the presence of six anthraquinones (DMAC, **3, 4, 7, 8**, and **9**) was measured. For each compound, three different concentrations (50, 25, and 12.5 mM) were used to detect dose-dependent inhibition. No inhibition was observed for DMAC or **7**. Dose-dependent inhibition was observed for compounds **3,** **4**,**8**, and **9**. :::  ::: The integrity of the microsomes was first verified by comparing the glucose-6-phosphatase (G6Pase) activity in the presence of either glucose-6-phosphate or mannose-6-phosphate. The activity of G6Pase was then measured in the presence of chlorogenic acid (IC~50~ = 0.26 mM), DMAC, **3**,**4**,**7**,**8**, or **9** using a colorimetric assay described earlier ([@pbio-0020031-Arion1]) ([Figure 7](#pbio-0020031-g007){ref-type="fig"}B). Dose-dependent inhibition was observed in the 10--50 μM range with the alkylacyl-primed compounds **3**,**4**,**8**, and **9**, but not DMAC or **7**. Our results demonstrate the following: (i) a long substituent at C-16 in an anthraquinone is important for targeting the membrane-bound G6Pase, and (ii) the C-9 position of the anthraquinone can be chemically modified without significantly affecting the enzyme-inhibitor interactions. Discussion {#s3} ========== The engineering of the primer units of macrolide antibiotics is a well-established strategy for generating new natural product analogs with modified chemical and biological properties ([@pbio-0020031-Jacobsen1]; [@pbio-0020031-Marsden1]; [@pbio-0020031-Moore1]). In contrast, manipulation of the ordinarily invariant acetate primer unit of bacterial aromatic polyketides has not been recognized as a general methodology in biosynthetic engineering, presumably owing to the apparently high efficiency with which these PKSs decarboxylate malonyl extender units to generate acetate primers. An exception to this principle has been recently demonstrated in the case of the enterocin PKS, which ordinarily incorporates a benzoic acid primer unit, but can also accept a range of aryl acids to generate substituted enterocins ([@pbio-0020031-Kalaitzis1]). In this report, we have described a general method for modifying the primer unit of any aromatic polyketide by engineering hybrid bimodular PKSs. This method can be used to construct hitherto undiscovered polyfunctional aromatic scaffolds, as illustrated by compounds **1**and **2**; alternatively, regioselective modifications of known polyketides, such as **8**and **9**, can be achieved. Notably, structural analysis of these novel compounds also revealed fundamentally new properties of bacterial aromatic PKSs, as summarized below. The KS-CLF Prefers Nonacetate Priming over Decarboxylative Priming {#s3a} ------------------------------------------------------------------ Most bacterial aromatic PKSs catalyze chain initiation by decarboxylating an ACP-bound malonyl extender unit to yield an acetyl primer unit, a reaction that is catalyzed by the KS-CLF. In order to install nonacetate primer units in an aromatic polyketide backbone, one must bypass this decarboxylative priming mechanism. Genetic and enzymological analysis of the R1128 PKS, which utilizes a range of nonacetate primer units, has revealed the existence of two PKS modules. Each module includes a distinct KS and an ACP. Previous studies have shown that these two KS-ACP pairs have orthogonal molecular recognition features, leading to the speculation that the initiation module may be able to productively interact with other bacterial aromatic PKSs to synthesize hybrid polyketides. However, the ability of the R1128 initiation module to kinetically compete with the intrinsic decarboxylative priming mechanism of the heterologous PKS was unexplored. To address this question, we coexpressed the entire R1128 initiation module with either the *act* or the *tcm* minimal PKS. The efficient biosynthesis of compounds described in this report shows that, although decarboxylation cannot be completely suppressed, both PKSs have an intrinsic preference for nonacetate primers over decarboxylative chain initiation. It should be noted that although acetate-primed products are observed in conjunction with nonacetate-primed compounds (e.g., CH999/pYT46 cosynthesizes mutactin along with compounds **1** and **2**), the former class of products may not be derived via decarboxylative priming in strains carrying bimodular PKSs. Instead, they may arise as a result of premature diketide transfer from the initiation module to the elongation module before the β-carbonyl can be reduced. Future isotope labeling studies on such systems should be useful for quantifying the distribution between polyketide chains derived from bimodular PKSs versus those that arise via decarboxylative priming. Our findings are consistent with the fact that the frenolicin PKS from Streptomyces roseofulvus can synthesize both nanaomycin (an acetate-primed aromatic polyketide) and frenolicin (its butyrate-primed analog) ([@pbio-0020031-Tsuzuki1]). They also explain earlier observations that the doxorubicin (a propionate-primed polyketide) and oxytetracycline (a malonamate-primed polyketide) minimal PKSs yield acetate-primed polyketides, when expressed alone ([@pbio-0020031-Fu1]; [@pbio-0020031-Rajgarhia1]). Thus, notwithstanding its widespread prevalence, decarboxylative priming by the KS-CLF can be regarded as a default mechanism for chain initiation that occurs when alternative primer units are unavailable. The potential for recombining naturally occurring initiation and elongation PKS modules from bacterial aromatic PKSs is enormous. Other than the R1128 biosynthetic pathway, initiation modules with attractive primer unit specificity can also be found in the doxorubicin, frenolicin, enterocin, and (presumably) oxytetracycline biosynthetic pathways. It should be possible to recombine these synthase units with elongation modules from the *act*, *frn*, *tcm*, and *whiE* ([@pbio-0020031-Yu1]) PKSs (which synthesize C~16~--C~24~ backbones) to yield a range of reactive backbones whose subsequent fates can be controlled by previously analyzed auxiliary PKS subunits and tailoring enzymes. Molecular Recognition Features of the KS-CLF and Auxiliary PKS Subunits {#s3b} ----------------------------------------------------------------------- By generating a variety of nascent and highly reactive alkylacyl polyketide intermediates in situ, we have been able to probe the properties of the key aromatic PKS components, including the KS-CLF, the KR, and different subclasses of AROs and CYCs. Such studies have provided insight as to whether carbon chain length of the backbone or the repetitive poly-β-ketone functionality are the primary factors influencing the substrate specificity of these proteins. Our results demonstrate that, since chain length specificity of KS-CLF heterodimers is primarily dictated by the backbone size, incorporation of bulky, nonacetate primer units is compensated for by a reduced number of condensation cycles. Thus, hexaketides and octaketides are synthesized by the *act* and *tcm* KS-CLF, respectively, when these KSs are primed with pentanoyl (or 4-methylpentanoyl) primer units. In contrast to KS-CLF subunits, the regioselectivity of auxiliary PKS enzymes such as KR, ARO, and CYC is unaffected by the incorporation of nonacetate starter units. For example, the *act* KR selectively reduced the C-9 keto group in an acetate-primed octaketide (DMAC), an alkyl-hexaketide (see [Figure 4](#pbio-0020031-g004){ref-type="fig"}B), and an alkyl-octaketide (see [Figure 5](#pbio-0020031-g005){ref-type="fig"}B). This observation confirms an earlier proposal ([@pbio-0020031-McDaniel2]) that such KRs recognize fully synthesized polyketide chains, rather than the β-ketone of a partially elongated intermediate. Similarly, the *act* ARO and CYC process a reduced (but not unreduced) octaketide, regardless of the primer unit, although they are unable to recognize hexaketides or decaketides. In contrast, the R1128 CYCs act upon an unreduced, but not reduced, backbone, regardless of the primer unit. The Initiation Module of a Bimodular Aromatic PKS {#s3c} ------------------------------------------------- Our studies have revealed that each initiation module component, ZhuG (see [Figure 4](#pbio-0020031-g004){ref-type="fig"}A), ZhuH (data not shown), and ZhuC (data not shown) are essential for priming with a nonacetate building block. Deletion of any of these three genes from the constructs shown in [Table 1](#pbio-0020031-t001){ref-type="table"} completely abolishes nonacetate priming by the KS-CLF, but leaves the decarboxylative priming mechanism intact. Although the roles of ZhuG and ZhuH in the initiation pathway have been reported ([@pbio-0020031-Meadows1]; [@pbio-0020031-Tang1]), the role of ZhuC is unclear. ZhuC is homologous to the MAT and was therefore putatively assigned as a second malonyl transferase that malonylates ZhuG. However, subsequently it was shown that MAT can malonylate ZhuG with high efficiency (*k~cat~*, approximately150 s^--1^) ([@pbio-0020031-Tang1]), whereas ZhuC is sluggish in malonylating ACPs (at a rate approximately ten times slower than MAT \[data not shown\]). In light of these observations, we propose that ZhuC catalyzes transacylation between the diketide-ZhuG and ZhuN, leading to an alkylacyl-ZhuN intermediate that can then be transferred onto the KS-CLF. Future biochemical analysis may be able to verify this property of ZhuC. Engineered Biosynthesis of Diverse Aromatic Polyketides via Bimodular PKSs {#s3d} -------------------------------------------------------------------------- To further expand the repertoire of primer units that can be introduced into aromatic polyketides using bimodular PKSs, one could (i) alter the substrate specificity of the R1128 initiation module and/or (ii) engineer in vivo metabolic pathways for new types of primer units. KSIII homologs found in initiation modules serve as gatekeepers in primer unit selection and are therefore attractive targets for protein engineering. The X-ray crystal structure of ZhuH has recently been solved and has led to the identification of a binding pocket for the acyl-CoA moiety ([@pbio-0020031-Pan1]). ZhuH adopts a dimeric thiolase fold and selected residues at the interface between the two subunits control size and flexibility of the binding pocket, preventing acyl groups larger than isovaleryl groups from entering the pocket. The corresponding amino acids in FrnI, the homolog to ZhuH in the *frn* PKS, are occupied by bulkier residues, thus excluding acyl groups larger than acetyl-CoA. Therefore, altering the size and polarity of these gatekeeping residues using rational mutagenesis and directed evolution may enlarge the repertoire of acyl-CoA moieties recognized by KSIII enzymes. Amino acid catabolism in S. coelicolor is the primary source of primer units such as isobutyryl-CoA (valine) and isovaleryl-CoA (leucine). These pathways involve transamination (catalyzed by branched-chain amino acid transaminases) to convert the amino acid into the corresponding α-ketoacid, followed by decarboxylation (catalyzed by acyl-CoA dehydrogenase \[AcdH\]) to yield the corresponding acyl-CoA ([@pbio-0020031-Zhang1]). These catabolic pathways in actinomyces are presumably tolerant of unnatural amino/α-keto acids, as illustrated by the incorporation of a large variety of primer units into the macrolide avermectin through precursor feeding ([@pbio-0020031-Dutton1]). It may therefore be possible to expand the repertoire of primer units in S. coelicolor by feeding the recombinant strains, constructed as above, with unnatural amino acids such as allylglycine, norvaline, norleucine, and fluorinated derivatives thereof. In addition, heterologous expression of enzymes involved in the biosynthesis of novel acyl-CoA moieties, such as cyclohexynoyl-CoA ([@pbio-0020031-Cropp1]) and benzoyl-CoA ([@pbio-0020031-Xiang1]), can be efficient sources of loading module substrates. Successful elaboration of the corresponding acyl-CoA primers into full-length polyketides, through both KSIII-based protein engineering and *Streptomyces* metabolic engineering, can yield additional polyketide variants. Our observation that anthracycline derivatives generated via such engineering approaches can have improved properties over their natural product counterparts provides further motivation to expand the biosynthetic potential of bimodular PKSs. Materials and Methods {#s4} ===================== {#s4a} ### Bacterial strains and general methods for DNA manipulation. {#s4a1} S. coelicolor strain CH999 was used as the host for transformation by shuttle vectors. Protoplast preparation and PEG-assisted transformation were performed as described by [@pbio-0020031-Hopwood2]). All cloning steps were performed in Escherichia coli strain XL-1 Blue. PCR was performed using the *pfuTurbo* polymerase (Strategene, La Jolla, California, United States). PCR products were first cloned into pCRBlunt vector (Invitrogen, Carlsbad, California, United States), followed by DNA sequencing (Stanford PAN Facility, Stanford, California, United States). Unmethylated DNA was obtained by using the methylase-deficient strain GM2163 (New England Biolabs, Beverly, Massachusetts, United States). ### Construction of plasmids. {#s4a2} The following primers were used to amplify the individual genes: *zhuC*: 5′-CC*TCTAGA*TGTACTCGGGTCGAGGAGACCTCCG-3′, 5′-GG*ACTAGT*GCCACGTTCACCGTTCCGCCGCG-3′; *zhuN*: 5′-CATGCGACCCG*TCTAGA*GAAGGAGATTCCG-3′, 5′-CGCGGTTCTGC*ACTAGT*CAGGCCGCGGCC-3′; *zhuG*: 5′-CCTG*TCTAG*AGGGAGGACGAACCC-3′, 5′-TG*CTGCAG*TCAGCCCGCGGTCTCG-3′; *zhuH*: 5′-GA*CTGCAG*CAGAACCGCGAAAGGTGG-3′, 5′-AGTAGTAC*GTTTAAAC*TCAAGCCGGAGTGGACGGC-3′; *act*VII*/act*IV: 5′-GCC*GTTTAAAC*GCTGGCGCCAAGCTTCTC-3′, 5′-CCGGAGACGCGTCACGGCCGAAGC-3′; *zhuI/zhuJ*: 5′-GCC*GTTTAAAC*CGAGGAGCACCCTCATGCGTC-3′, 5′-GGACTAGTCCTCCTCTTCCTGCTCG-3′. The introduced restriction sites are shown in italics. Genes encoding *zhuC*, *zhuH*, *zhuN*, *zhuG*, and *zhuI/zhuJ* were amplified from pHU235 ([@pbio-0020031-Marti1]), and genes encoding *act*VII*/act*IV were amplified from pRM5 ([@pbio-0020031-McDaniel1]). *zhuC*, *zhuN*, and *zhuG* were cloned as a single 2.1 kb XbaI--PstI cassette; *zhuH* was cloned as a 1.3 kb PstI--PmeI cassette; *act*VII*/act*IV (2.5 kb) and *zhuI/zhuJ* (1.3 kb) were each cloned as a PmeI--EcoRI cassette. Different combinations of cassettes, as shown in [Table 1](#pbio-0020031-t001){ref-type="table"}, were introduced into either pRM5 (KR-*act*KS/CLF), pSEK24 (*act*KS/CLF), pSEK33 (*tcm*KS-CLF), or pRM20 (KR-*tcm*KS/CLF) to yield pYT46, pYT105, pYT128, and pYT127, respectively. ### Culture conditions, extraction, and small-scale analysis. {#s4a3} The strains were grown on R5 plates containing 50 mg/l thiostrepton at 30°C for 7--10 d. Acyl-CoA precursors such as sodium propionate and valine were added at 1 g/l when needed. For LC/MS and analytic HPLC analysis, one plate was sufficient. The plate was chopped into fine pieces and extracted with 30 ml of ethyl acetate (EA)/methanol/acetic acid (89:10:1). The solvent was removed in vacuo and the residue was redissolved in 1 ml of methanol. The polyketide products were separated and detected by analytical reversed-phase HPLC using a diode array detector at 280 and 410 nm using an Alltech (Vienna, Virginia, United States) Econosphere C18 column (50 mm × 4.6 mm); linear gradient: 20% acetonitrile (ACN) in water (0.1% TFA) to 60% ACN in water (0.1% TFA) over 30 min; 1 ml/min. HPLC retention times (t~R~, minutes) were as follows: **1**: 13.8; **2**: 15.4; **3**: 20.4; **4**: 21.8; **5**: 24.9; **6**: 26.4; **7**: 17.1; **8**: 22.8; **9**: 24.4; **10**: 27.7;**11**: 29.4. LC/MS was performed at the Vincent Coates Foundation Mass Spectrometry Laboratory at Stanford University using a ThermoFinnigan (San Jose, California, United States) quadrupole ion trap LC/MS system and electrospray ionization (both positive and negative ionization). ### Large-scale production and isolation. {#s4a4} Sufficient number of R5 plates (20--60 plates, depending on the yield of the product) streaked with the desired CH999 strains were grown at 30°C for 1 wk. The plates were chopped into fine pieces and extracted with a minimum of 1 l of EA/methanol/acetic acid (89:10:1). The organic solvents were removed and the residuals were dissolved in 5 ml of methanol. The solution was filtered and injected into a preparative reversed-phase HPLC column (250 × 22.5 mm C-18 column; Alltech Econosil). A 20%--60% ACN in water (0.1 % TFA) gradient (50 min, 5 ml/min) was used to separate the polyketide products. Fractions containing the desired polyketides were combined and concentrated in vacuo. The residuals were redissolved in acetone and applied to a preparative TLC plate (20 cm × 20 cm, 0.25 mm E. Merck \[Readington Township, New Jersey, United States\] silica gel plates \[60F-254\]). TLC plates spotted with **1** (*R* ~f~ = 0.34) and **2** (*R* ~f~ = 0.41) were developed with EA/methanol/acetic acid (97:2:1), while those spotted with **3**(*R* ~f~ = 0.29), **4**(*R* ~f~ = 0.37), **7** (*R* ~f~ = 0.21), **8**(*R* ~f~ = 0.34), and **9** (*R* ~f~ = 0.43) were developed with EA/hexane/acetic acid (90:10:1). The desired bands were excised from the TLC plates and stirred in EA/methanol (10 ml, 9:1) for 2 h. The compounds were eluted from silica using the same solvent and dried in vacuo. ### NMR and MS characterization of novel compounds. {#s4a5} NMR spectra were recorded on Varian (Salt Lake City, Utah, United States) Inova 500 or Mercury 400 instruments and calibrated using residual undeuterated solvent as an internal reference.^1^H and ^13^C NMR spectra data are shown in [Tables 2--4](#pbio-0020031-t002){ref-type="table"}. HRFABMS were collected under negative ionization mode as follows: HRFABMS m/z: **1**: 277.1082 (calcd for C~15~H~17~O~5~: 277.1076); **2**: 291.1247 (calcd for C~16~H~19~O~5~: 291.1232);**3**: 355.0823 (calcd for C~19~H~15~O~7~: 355.0818); **4**: 369.0981 (calcd for C~20~H~17~O~7~: 369.0974); **7**: 313.0354 (calcd for C~16~H~9~O~7~: 313.0348); **8**: 339.0871 (calcd for C~19~H~15~O~6~: 339.0869); **9**: 353.1036 (calcd for C~20~H~17~O~6~: 353.1025). ### Cytotoxicity studies. {#s4a6} The studies were performed as described by Hori et al (1993). The cells were maintained at 37°C and growth was measured with the colorimetric MTT assay after each day. The IC~50~ values were measured after 5 d. ### G6Pase activity assay. {#s4a7} Male rat liver microsome (Sprague Dawley) was purchased from BD Gentest^TM^ (Becton Dickinson, Franklin Lakes, New Jersey, United States). Aliquots (100 μl, 2 mg/ml) in 0.25 M sucrose were stored at --80°C. Compound stock solutions were prepared in 95% ethanol and diluted with DMSO. Glucose-6-phosphate and mannose-6-phosphate were purchased from Sigma (St. Louis, Missouri, United States). The integrity of microsomes and G6Pase activity was measured based on the colorimetric reaction of inorganic phosphate as previously reported ([@pbio-0020031-Arion1]). The enzyme reaction was initiated by adding 3 μl of microsome to the reaction mixture, which contained 51 μl of assay buffer (50 mM HEPES, 100 mM KCl, 2.5 mM EDTA, 2.5 mM MgCl~2~, and 1 mM DTT at pH 7.2), 3 μl of glucose-6-phosphate (final concentration, 1 mM), and 3 μl of inhibitor sample in DMSO. The reaction mixture was incubated at room temperature, and 13 μl of reaction mixture was taken every 10 min and quenched with 117 μl of working solution (6:2:1 mixture of 0.42% ammonium molybdate tetrahydrate in 1N H~2~SO~4~, 10% SDS in water, and 10% ascorbic acid in water). The blue reduced phosphomolybdate complex is formed after incubation at 50°C for 20 min. The absorbance was measured at 820 nm. Supporting Information {#s5} ====================== Accession Numbers {#s5a} ----------------- The SwissProt ([www.ebi.ac.uk/swissprot/](www.ebi.ac.uk/swissprot/)) accession numbers for the proteins and genes discussed in this paper are *act* ARO/CYC (Q02055), *act* KS-CLF (Q02059 and Q02062), ZhuA (Q9F6E1), ZhuB (Q9F6E0), ZhuC (Q9F6D6), ZhuG (Q9F6D5), ZhuH (Q9F6D4), ZhuI (Q9F6D3), ZhuJ (Q9F6D2), and ZhuN (Q9F6C8). This work was supported by a grant from the National Institutes of Health (NIH) (CA 77248 to CK). YT is supported by an NIH postdoctoral fellowship. **Conflicts of interest.** A patent application has been filed by Stanford University for aspects of the work described in this paper. **Author contributions.** YT, TSL, and CK conceived and designed the experiments. YT and TSL performed the experiments. YT and TSL analyzed the data. YT wrote the paper. DOI: [10.1371/journal.pbio.0020031](10.1371/journal.pbio.0020031) Copyright: © 2004 Tang et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Academic Editor: Rowena G. Matthews, University of Michigan AcdH : acyl-CoA dehydrogenase ACN : acetonitrile ACP : acyl carrier protein *act* : actinorhodin ARO : aromatase AT : acyl transferase CLF : chain length factor CYC : cyclase DMAC : 3,8-dihydroxy-methylanthraquinone carboxylic acid EA : ethyl acetate FAS : fatty acid synthase G6Pase : glucose-6-phosphatase KR : ketoreductase KS : ketosynthase LC : liquid chromatography MAT : malonyl-CoA:ACP acyl transferase MS : mass spectrometry NMR : nuclear magnetic resonance PKS : polyketide synthase *tcm* : tetracenomycin TMAC : 3,6,8-trihydroxy-1-methylanthraquinone-2-carboxylic acid

PubMed Central

2024-06-05T03:55:46.075806

2004-2-17

PMC340943

For half a century, natural products from microorganisms have been the main source of medicines for treating infectious disease. The most important chemical class of these antibiotics, apart from the penicillins, is the polyketides. They are made by the stepwise building of long carbon chains, two atoms at a time, by multifunctional enzymes that determine the chain length, oxidation state, and pattern of branching, cyclisation, and stereochemistry of the molecules in a combinatorial fashion to produce an enormous variety of structures. Recent elucidation of the genetic 'programming' of the enzymes has opened a new field of drug discovery based on rationally engineering the enzymes to produce 'unnatural natural products' with novel properties. Following the development of penicillin for the treatment of septicemia in the early 1940s, numerous antibiotics were discovered and introduced into medicine. While a fungus makes penicillin, semisynthetic derivatives of which have been a mainstay of antibacterial therapy for decades, most natural antibacterial antibiotics come from a group of soil-dwelling, filamentous bacteria called the actinomycetes, of which Streptomyces is the best-known genus. These organisms make an amazing array of so-called secondary metabolites that have evolved to give their producers a competitive advantage in the complex soil environment, where they are exposed to stresses of all kinds ([@pbio-0020035-Challis1]). The compounds have many functions, but those with antibiotic activity are the most important from the human perspective. Actinomycete antibiotics include such antibacterial compounds as tetracycline and erythromycin, antifungal agents like candicidin and amphotericin, anticancer drugs such as doxorubicin, and the antiparasitic avermectin ([@pbio-0020035-Walsh1]). While many different chemical classes are represented amongst actinomycete antibiotics, one class accounts for an extraordinary proportion of the important compounds, including all those mentioned above. This chemical family is made up of the polyketides. They are synthesized by multifunctional enzymes called polyketide synthases (PKSs), which are related to the fatty acid synthases that make the lipids essential for the integrity of cell membranes, but they carry out much more complex biosynthetic routines. Repeated rounds of carbon chain building and modification use a series of independently variable reactions selected according to a 'program' characteristic of each PKS ([@pbio-0020035-Reeves1]). Recent research has focused on determining this program so as to be able to modify it in rational ways by genetic engineering and thus generate novel drug candidates. The resulting field of 'combinatorial biosynthesis' of 'unnatural natural products' has been given added urgency by the rise of multidrug-resistant pathogens, of which MRSA (methicillin-resistant Staphylococcus aureus) is simply the most discussed of a series of threats ([@pbio-0020035-Walsh1]). How do PKSs work and how can we make new ones? Molecular Diversity {#s2} =================== The heart of PKS function is the synthesis of long chains of carbon atoms by joining (condensing) together small organic acids, such as acetic and malonic acid, by a so-called ketosynthase function. This uses the building units in the form of activated derivatives, called coenzyme A (CoA) esters, so we speak of acetyl-CoA and malonyl-CoA, for example. The special form of condensation that joins them is driven by loss of carbon dioxide. Thus, when acetyl-CoA, with two carbon atoms, joins with malonyl-CoA, with three carbons, one of the latter is lost and a chain of four carbon atoms results ([Figure 1A](#pbio-0020035-g001){ref-type="fig"}). Further rounds of condensation extend the chain by two carbons at each step. If the chain-extender unit, instead of being malonyl-CoA, is methylmalonyl-CoA, which has four carbon atoms, the linear carbon chain is still extended by two carbons, and the 'extra' carbon forms a methyl side branch. More complex extender units produce more complex side branches. ::: {#pbio-0020035-g001 .fig} Figure 1 ::: {.caption} ###### The Chemistry of Polyketide Chain Assembly \(A) Acetic acid and malonic acid are converted to their coenzyme A (CoA) esters and then attached, by specific acyl transferases, to components of the polyketide synthase (PKS): acetyl-CoA is attached to the active site of the ketosynthase, and malonyl-CoA to a structural component of the PKS called the acyl carrier protein (ACP). Condensation of the two units by the ketosynthase, with loss of one carbon from malonyl-CoA as carbon dioxide, produces a four-carbon chain attached to the ACP. This is transferred back to the ketosynthase, and further rounds of condensation with malonyl-CoA (as shown) or other chain extender units produce a polyketide chain. \(B) The three-step reductive cycle that converts a keto group to a hydroxyl, then to a double bond, and finally to a fully saturated carbon. \(C) A complex polyketide in which keto groups, hydroxyl groups, double bonds, and fully saturated carbons occur at different positions along the chain, depending on the operation of the reductive cycle after each condensation. :::  ::: Choices of the number and type of the building units are variables in determining polyketide structure. Another concerns the keto groups (C=O) that appear at every alternate carbon atom in the growing chain as a result of the condensation process (accounting for the name polyketide). They may remain intact. Alternatively, some may be modified or removed by a series of three steps ([Figure 1B](#pbio-0020035-g001){ref-type="fig"}), any of which may be omitted. This results in keto groups remaining at some points in the chain; hydroxyl groups (--OH), formed by reduction of a keto group, at others; double bonds between some adjacent carbon atoms, resulting from removal of the hydroxyl by loss of water (dehydration); or full saturation with hydrogen atoms elsewhere, arising by 'enoyl' reduction of the double bond ([Figure 1C](#pbio-0020035-g001){ref-type="fig"}). A further variable concerns the stereochemistry of the hydroxyl groups and methyl or other carbon branches, each of which can exist in two possible configurations. Finally, the nascent carbon chain adopts different folding patterns after it leaves the PKS, and 'tailoring' enzymes can then add sugars or other chemical groups to it at many alternative positions, enabled by the pattern of chemical reactivity built into the polyketide by the PKS. The challenge has been to understand the programming of the PKS that accounts for this gamut of structural variation. During the 1990s, the ability to manipulate actinomycete genes, developed over previous decades, mainly using the model species Streptomyces coelicolor ([@pbio-0020035-Hopwood1]), was combined with chemical and biochemical experiments to begin to crack this 'polyketide code'. The first studies were on organisms making antibiotics of the 'aromatic' family, which includes tetracycline and doxorubicin, as well as the model compounds actinorhodin (made by S. coelicolor itself) and tetracenomycin. The main variable in their structure is carbon chain length, with few choices of different building units or keto group modification, so the programming would (in principle) be simple. The DNA sequences responsible for such PKSs revealed sets of genes encoding proteins, including ketosynthases, ketoreductases, and acyl carrier proteins (ACPs) (the unit of the PKS on which the growing carbon chain is tethered; see [Figure 1A](#pbio-0020035-g001){ref-type="fig"}), that would come together to form a multicomponent PKS resembling a typical bacterial fatty acid synthase. In contrast, the DNA sequence of the gene set for the complex polyketide erythromycin, made by a relative of Streptomyces called Saccharopolyspora erythraea, which has more involved programming, revealed multifunctional proteins with the various enzymic functions carried out by active sites on the same polypeptide chain, as in a mammalian fatty acid synthase. The big surprise, though, was the finding of six sets, or modules, of such active sites, corresponding to the six rounds of condensation needed to build the carbon chain ([@pbio-0020035-Cortes1]; [@pbio-0020035-Donadio1]). The modules each contain an acyl transferase (to load the extender unit onto the enzyme), as well as a ketosynthase and an ACP domain, together with exactly those reductive activities needed to generate the required pattern of modification of the chain at each step of elongation. Thus was born an 'assembly line' model in which the program for the PKS is hardwired into the DNA and expressed in a linear array of active sites (domains) along the giant protein. This consists of the six chain-building modules, preceded by a short module for loading the starter unit and ending in a domain for releasing the completed carbon chain from the PKS. The carbon chain of the polyketide would be assembled and modified progressively as the molecule moved along the protein, interacting with each domain in turn, which would select extender units, make carbon--carbon bonds, and modify keto groups as appropriate, depending on the presence or absence of domains for the three steps in the reductive cycle. The model arose from the gene sequence, but was rapidly tested by mutating individual domains or adding or deleting whole modules and by observing predicted changes in the polyketide product. Soon, dozens of engineered compounds had been made, and the field mushroomed with the isolation of more and more clusters of genes for complex polyketides that both proved the generality of the model (with minor variations) and filled the need for spare parts for the engineering of countless new polyketides ([@pbio-0020035-Shen1]). Several biotech companies were founded to exploit the potential for drug discovery. Aromatic PKS Programming {#s3} ======================== Meanwhile, the programming of the aromatic PKSs was harder to understand. They had been found to contain only a single ketosynthase, which has to operate a specific number of times to build a carbon chain of the correct length, so how is this determined? How does a single reductive enzyme know which keto groups to modify? And how is the starter unit for building the carbon chain selected (the extender units are normally all malonyl-CoA, so no choice is involved)? Considerable progress had been made in constructing novel compounds by mixing and matching PKS subunits, but this was largely based on empirical knowledge about which components to put together ([@pbio-0020035-McDaniel2]). A specific subunit of the PKS, named the chain length factor (CLF), was deduced to have a major influence on carbon chain length ([@pbio-0020035-McDaniel1]), but this conclusion was not universally accepted in the absence of experimental evidence on its mode of action. Two recent publications by the Khosla laboratory at Stanford University describe significant advances in understanding aromatic PKS programming and promise to turn the spotlight back onto engineered members of this class of compounds as potential drug candidates by allowing rational manipulation of the two key variables: carbon chain length and choice of starter unit. In the first paper ([@pbio-0020035-Tang1]), the authors explore the hypothesis that the CLF exerts control over carbon chain length by associating closely with the ketosynthase, a protein with which it shares considerable amino acid sequence similarity, giving rise to a channel of a certain size at the interface between the two proteins. By systematically changing amino acids at four key positions in the CLF, the size of the channel was altered. Thus, large amino acid residues in the CLF of a PKS that makes a 16-carbon chain were replaced by less bulky residues found in one that builds a 20-carbon chain, and the chain length of the product increased as expected. The authors propose that the length of the channel is the main factor in controlling the number of chain-extension steps that can take place to fill it. While protein--protein interactions with other PKS subunits may modulate this chain length control, the work represents a major step in understanding and manipulating the chain length of aromatic polyketides. What about the choice of starter unit? Most aromatic polyketides start with acetyl-CoA. An important earlier publication by Leadlay and colleagues ([@pbio-0020035-Bisang1]) had shown that this is not loaded directly onto the PKS, as had been assumed, but is derived by loss of carbon dioxide from a molecule of malonyl-CoA previously loaded onto the enzyme. This decarboxylation is catalysed by the CLF as an activity independent of its role in influencing carbon chain length. There are, however, certain aromatic polyketides, including the anticancer drug doxorubicin, an antiparasitic agent called frenolicin, and the estrogen receptor agonist R1128, that have different starters. What [@pbio-0020035-Tang2] have deduced, as described in this issue of *PloS Biology*, is that the PKSs for these compounds consist of two modules of active sites. The components of each module are not activities carried on the same protein, as in the PKSs for the complex polyketides, but are all separate proteins. They form functional modules nevertheless. The newly recognized modules in the producers of compounds that start with nonacetate units have a dedicated ACP and a special ketosynthase that carries out a first condensation, joining the unusual starter unit to the first malonyl-CoA extender unit. The starter module then hands the resulting 'diketide' on to the second module (first reducing it, if appropriate, using reductive enzymes 'borrowed' from fatty acid biosynthesis) for typical extension by successive condensation with malonyl-CoA units to complete the chain. If the starter module is not present, the second module defaults to its typical habit of decarboxylating malonyl-CoA to acetyl-CoA and starts the chain with that. The excitement of the work for biotechnology is that it offers the prospect of engineering promising drug candidates by making novel combinations of starter and extender modules and perhaps of feeding the starter modules with a whole range of unnatural substrates ([@pbio-0020035-Kalaitzis1]). It is encouraging that already, in the proof-of-principle studies reported by [@pbio-0020035-Tang2], some products with improved in vitro antitumor activity were obtained. David Hopwood is an emeritus fellow at the John Innes Centre in Norwich, United Kingdom. E-mail: <david.hopwood@bbsrc.ac.uk> ACP : acyl carrier protein CLF : chain length factor CoA : coenzyme A MRSA : methicillin-resistant Staphylococcus aureus PKS : polyketide synthase

PubMed Central

2024-06-05T03:55:46.079525

2004-2-17

PMC340944

Introduction {#s1} ============ Most patients who succumb to cancer do so not from primary tumor burden, but from metastatic disease ([@pbio-0020036-Fidler1]). For example, advanced-stage peritoneal carcinomatosis (e.g., from metastatic ovarian and breast cancer) and disseminated peritoneal lymphomas are often resistant to current chemotherapy treatment ([@pbio-0020036-Parsons1]). Posttreatment survival rates for patients presenting with metastatic ovarian peritoneal carcinomatosis or lymphoma are less than 20% and less than 50%, respectively ([@pbio-0020036-Lam1]; [@pbio-0020036-Deppe1]; [@pbio-0020036-Hofstra1]). Consequently, the development of novel therapeutic strategies to reverse these numbers is clearly warranted. A significant effort has been aimed at understanding the function of tumor suppressor gene pathways that are genetically and epigenetically altered during oncogenesis ([@pbio-0020036-Macleod1]). One rationale for the study of tumor suppressor pathways is the hypothesis that reconstitution of these pathways in cancer patients will be therapeutically beneficial ([@pbio-0020036-Macleod1]). The p53 tumor suppressor protein induces growth arrest and apoptosis in response to cellular stress ([@pbio-0020036-Vousden1]). Mutation of genes in the p53 pathway is thought to be nearly universal in human cancer ([@pbio-0020036-Vousden1]). Thus, any strategy designed to restore p53 activity in tumor cells will likely be an effective means of inducing cancer cell death and will be applicable to a large fraction of cancer patients. The inability of large tumor suppressor proteins, all of which are intracellular, to cross the plasma membrane precludes the therapeutic administration of recombinant tumor suppressors in a manner analogous to administration of extracellular biological therapeutics (e.g., insulin or G-CSF). Thus, the development of an efficient methodology for restoring tumor suppressor function to cancer cells in vivo remains a challenge for both basic and clinical researchers. Gene therapy approaches aimed at restoring tumor suppressor function have been extensively investigated. Both viral and nonviral vectors have been employed to express exogenous tumor suppressor genes, such as *p53*, in cancer cells ([@pbio-0020036-McCormick1]). Although gene therapy may be useful under certain conditions, problems associated with immunogenicity and lack of systemic biodistribution to disseminated metastases are likely to curtail its anticancer efficacy ([@pbio-0020036-McCormick1]). Delivery of macromolecules by protein transduction has recently emerged as an alternative methodology for directly introducing tumor suppressor proteins into cancer cells in vivo. Several small cationic peptides, including TAT, Antp, and polyArg (referred to as protein transduction domains \[PTDs\]), are capable of traversing the plasma membrane and entering the cytoplasm of cells by a concentration-dependent, but receptor-independent, macropinocytic mechanism ([@pbio-0020036-Wadia1]). PTDs have recently been used to deliver a wide range of cargo, including biologically active proteins, peptides, nucleic acids, and iron beads, into cells in culture ([@pbio-0020036-Fischer1]; [@pbio-0020036-Lindsay1]). PTDs have also been employed to deliver biologically active cargo into most, if not all, tissues in preclinical models ([@pbio-0020036-Schwarze1]). Because of the presence of either wild-type or mutant p53 protein in most tumors, it has been hypothesized that restoration of endogenous p53 activity in cancer cells will be a therapeutically efficacious alternative to delivery of exogenous p53. However, this hypothesis has been tested in vivo in only a limited number of cases ([@pbio-0020036-Foster1]; [@pbio-0020036-Bykov1]) and has never been tested in preclinical models of terminal human malignancy. We therefore focused on a strategy to activate endogenous p53 in cancer cells by PTD-mediated delivery. The C-terminus of p53 is a lysine-rich domain that is subjected to a variety of posttranslational modifications ([@pbio-0020036-Apella1]). A peptide derived from the C-terminus was previously shown by D. Lane\'s group (University of Dundee, United Kingdom) to activate specific DNA binding by p53 in vitro by an unknown mechanism ([@pbio-0020036-Hupp1]). In cancer cells, p53C′ peptide can induce apoptosis by activating wild-type p53 protein and by restoring function to several p53 DNA contact mutants. Importantly, the p53C′ peptide also restores specific DNA binding to some p53 DNA contact mutants in vitro and induces apoptosis in cancer cells expressing p53 DNA contact mutants ([@pbio-0020036-Selivanova1], 1998; [@pbio-0020036-Kim1]). However, the peptide fails to induce apoptosis in p53-deficient tumor cells or in tumor cells containing *p53* structural mutations. In contrast, primary cells are resistant to p53C′ peptide action ([@pbio-0020036-Selivanova1]; [@pbio-0020036-Kim1]). This resistance is likely a result of the extremely low levels of endogenous p53 present in normal cells and the absence of continual DNA damage often associated with tumor cells ([@pbio-0020036-Selivanova1]; [@pbio-0020036-Kim1]). Here we report a proof-of-concept that in vivo delivery of a transducible, proteolytically stable p53C′ peptide (termed RI-TATp53C′) is a therapeutically effective means of activating the p53 tumor suppressor pathway in preclinical models of terminal metastatic cancer. Results {#s2} ======= Activation of p53 by Transducible Retro-Inverso D-Isomer p53C′ Peptide {#s2a} ---------------------------------------------------------------------- Although PTDs solve one major obstacle to the use of intracellular peptides as therapeutics, the susceptibility of peptides to degradation in vivo remains problematic. To circumvent the problem of proteolytic degradation, we synthesized a retro-inverso version of the parental p53C′ peptide by inverting the peptide sequence, using D-isomer residues, and adding the TAT PTD to obtain a transducible RI-TATp53C′ peptide ([Figure 1](#pbio-0020036-g001){ref-type="fig"}A). This double inversion of peptide structure often leaves the surface topology of the sidechains intact and has been used extensively to stabilize biologically active peptides for in vivo applications ([@pbio-0020036-Chorev1]). Because of their greater stability, retro-inverso peptides often display increased potency. ::: {#pbio-0020036-g001 .fig} Figure 1 ::: {.caption} ###### RI-TATp53C′ Induces the Hallmarks of p53 Activity in TA3/St Mammary Carcinoma Cells \(A) Sequence of p53C′TAT peptide (L-amino acids) and its retro-inverso analogue (D-amino acids). To generate a negative control peptide, three essential lysine residues ([@pbio-0020036-Selivanova1]) were mutated while leaving the remaining peptide sequence intact. \(B) Induction of G1 arrest in TA3/St cells by wild-type RI-TATp53C′, but not mutant peptide, 24 h after peptide addition. \(C) Dose-dependent induction of G1 arrest by RI-TATp53C′ (open square) (D-amino acids) and the less potent p53C′TAT (open circle) (L-amino acids) but not mutant (open triangle) peptide at 24 h (left) and 48 h (right) after single treatment. \(D) Induction of a permanent growth arrest in TA3/St cells by RI-TATp53C′. Cells were treated with RI-TATp53C′ peptide or vehicle for 2 d, replated, and allowed to proliferate in the presence of serum for 10 d. Colonies were then stained with methylene blue. \(E) Induction of a senescence-like phenotype in TA3/St cells by RI-TATp53C′. Cells were treated with RI-TATp53C′ peptide and stained for acidic β-galactosidase activity. :::  ::: To determine whether the RI-TATp53C′ peptide retained functionality, we compared the transducible parental L-isomer p53C′TAT and D-isomer RI-TATp53C′ peptides for the ability to induce a cell cycle arrest ([Figure 1](#pbio-0020036-g001){ref-type="fig"}B). Treatment of murine TA3/St mammary carcinoma cells (which express wild-type p53) with either the L-isomer p53C′TAT or D-isomer RI-TATp53C′ peptides resulted in a concentration-dependent G1 cell cycle arrest ([Figure 1](#pbio-0020036-g001){ref-type="fig"}B and [1](#pbio-0020036-g001){ref-type="fig"}C). The control D-isomer mutant peptide had little to no effect ([Figure 1](#pbio-0020036-g001){ref-type="fig"}B and [1](#pbio-0020036-g001){ref-type="fig"}C). Compared to the L-isomer p53C′TAT, the D-isomer RI-TATp53C′ peptide induced a stronger cell cycle arrest at substantially lower concentrations ([Figure 1](#pbio-0020036-g001){ref-type="fig"}C). A single administration of the L-isomer p53C′TAT peptide partially arrested cells for 24 h, but by 48 h cells had reentered the cell cycle ([Figure 1](#pbio-0020036-g001){ref-type="fig"}C). In contrast, a single dose of the D-isomer RI-TATp53C′ peptide was sufficient to sustain a G1 arrest for greater than 7 d ([Figure 1](#pbio-0020036-g001){ref-type="fig"}C; data not shown). To ascertain whether sustained arrest required the continuous presence of RI-TATp53C′ peptide, TA3/St cells were treated with peptide or vehicle for 2 d and then replated under mitogenic conditions in the absence of peptide. Peptide-treated tumor cells formed less than 1% as many colonies as vehicle-treated cells ([Figure 1](#pbio-0020036-g001){ref-type="fig"}D). This observation suggested that RI-TATp53C′ peptide induced a permanent growth arrest in TA3/St cells. We therefore assayed RI-TATp53C′ peptide-treated cells for induction of senescence, a state of terminal arrest that can be induced by p53 activation ([@pbio-0020036-Roninson1]). By 6 d after peptide addition, greater than 80% of viable TA3/St cells were positive for acidic β-galactosidase activity ([Figure 1](#pbio-0020036-g001){ref-type="fig"}E), the standard marker of senescence ([@pbio-0020036-Roninson1]). The treated cells also displayed other features of senescence ([@pbio-0020036-Roninson1]), including increased size, increased granularity, and a flattened morphology ([Figure 1](#pbio-0020036-g001){ref-type="fig"}E; data not shown). These observations suggest that treatment of mammary carcinoma cells with the RI-TATp53C′ peptide induces hallmarks of p53 activation, namely a G1 cell cycle arrest followed by induction of senescence. We next investigated the ability of the RI-TATp53C′ peptide to transcriptionally activate p53-responsive genes. We transiently transfected p53 null human H1299 lung carcinoma cells with p53-dependent luciferase reporter plasmid (PG13-Luc) and either wild-type p53 expression plasmid or empty vector. The use of p53 null cells allows for negative controls that are not possible in cells expressing endogenous p53. As expected, we observed a p53-dependent induction of luciferase activity in cells transfected with p53 expression vector ([Figure 2](#pbio-0020036-g002){ref-type="fig"}A). However, RI-TATp53C′ peptide treatment of cells transfected with p53 expression vector resulted in a significant increase in p53-dependent luciferase activity ([Figure 2](#pbio-0020036-g002){ref-type="fig"}A, left). Consistent with observations in TA3/St cells (see [Figure 1](#pbio-0020036-g001){ref-type="fig"}B and [1](#pbio-0020036-g001){ref-type="fig"}C), mutant peptide and L-isomer p53C′TAT displayed substantially reduced potency in this assay when compared to RI-TATp53C′ peptide (data not shown). Importantly, RI-TATp53C′ peptide treatment of cells transfected with empty vector and luciferase plasmids caused no increase in p53 target promoter activity ([Figure 2](#pbio-0020036-g002){ref-type="fig"}A). In addition, RI-TATp53C′ peptide activated p53-dependent transcription in SW480 colon carcinoma cells expressing a p53 DNA contact mutant (R273H) and in H1299 p53 null colon carcinoma cells transfected with a p53 DNA contact mutant (R248Q and R273H) ([Figure 2](#pbio-0020036-g002){ref-type="fig"}A, right), though to a lesser extent than in the presence of wild-type p53 ([Figure 2](#pbio-0020036-g002){ref-type="fig"}A, left). These observations show that RI-TATp53C′ peptide retains the ability to specifically activate p53-dependent gene transcription. ::: {#pbio-0020036-g002 .fig} Figure 2 ::: {.caption} ###### RI-TATp53C′ Peptide Activates p53-Dependent Transcription and Inhibits Tumor Cells Expressing p53 (A, left) Induction of transcription from a p53-dependent promoter by RI-TATp53C′ only when p53 protein is expressed. H1299 cells (*p53^−/−^*) were cotransfected with p53-responsive reporter (PG13-Luc) and either empty vector or p53 expression vector. Depicted are mean and standard deviation of triplicate results that are representative of multiple experiments. (A, right) RI-TATp53C′ peptide activates p53-dependent transcription in cells expressing DNA contact mutant p53. SW480 cells containing a DNA contact mutant (R273H) p53 were transfected with p53-dependent reporter (PG13-Luc). H1299 cells (*p53* ^−/−^) were co-transfected with PG13-Luc and either R248Q or R273H mutant p53 expression vector. RI-TATp53C′ was added to cells, and promoter activity was assessed 24 h later. \(B) Inhibition of tumor cell proliferation in a p53-dependent manner by RI-TATp53C′. Increasing concentrations of peptide were added to HCT 116 cells (*p53^+/+^*) and their *p53^−/−^* isogenic derivatives. After 2 d, the number of viable cells was assessed by Trypan blue exclusion and normalized to the number of viable untreated cells. Mean and standard deviation of multiple experiments are depicted. \(C) Inhibition of the proliferation of tumor cells expressing wild-type or mutant p53, but not *p53^−/−^* tumor cells or nontransformed human fibroblasts. Cell viability was assessed as in (B). Mean and standard deviation of multiple experiments are depicted. :::  ::: To confirm that the RI-TATp53C′ peptide inhibited tumor cell proliferation in a p53-dependent fashion, we compared parental *p53^+/+^* HCT116 colorectal carcinoma cells to HCT116 cells that were rendered p53-deficient at both loci by targeted genetic recombination ([@pbio-0020036-Bunz1]). Treatment of wild-type *p53* HCT116 cells with RI-TATp53C′ peptide inhibited cell proliferation in a dose-dependent manner ([Figure 2](#pbio-0020036-g002){ref-type="fig"}B). In contrast, RI-TATp53C′ peptide treatment of p53-deficient HCT116 cells did not significantly alter the number of viable cells. p53-deficient human H1299 lung adenocarcinoma cells also failed to respond to the RI-TATp53C′ peptide ([Figure 2](#pbio-0020036-g002){ref-type="fig"}C), further confirming the specificity of peptide action. RI-TATp53C′ peptide inhibited proliferation of TA3/St cells (*p53^+/+^*) and human Namalwa lymphoma cells that express a p53 hotspot DNA contact mutant (R248Q) ([Figure 2](#pbio-0020036-g002){ref-type="fig"}C). In contrast, RI-TATp53C′ peptide did not alter the proliferation of normal human foreskin fibroblasts containing wild-type p53 ([Figure 2](#pbio-0020036-g002){ref-type="fig"}C). These results are consistent with previous observations that certain p53 contact mutations are susceptible to p53C′ peptide activation and that the p53C′ peptide induces apoptosis in tumor cells, but not normal cells ([@pbio-0020036-Selivanova1], 1998; [@pbio-0020036-Kim1]). Taken together, these observations demonstrate both the p53 and tumor dependency of the RI-TATp53C′ peptide. Systemic Delivery of RI-TATp53C′ Peptide Inhibits Solid Tumor Growth {#s2b} -------------------------------------------------------------------- We ([@pbio-0020036-Schwarze1]) and others ([@pbio-0020036-Datta1]; [@pbio-0020036-Harada1]) have previously shown that intraperitoneal (IP) administration of TAT--fusion peptides and proteins results in systemic delivery in animal models. Consistent with these observations, IP injection of a biotinylated RI-TATp53C′ peptide into mice harboring subcutaneous tumors resulted in distribution of the peptide throughout the tumor ([Figure 3](#pbio-0020036-g003){ref-type="fig"}A). ::: {#pbio-0020036-g003 .fig} Figure 3 ::: {.caption} ###### Solid Tumor Growth Is Inhibited by Systemic RI-TATp53C′ Peptide Administration \(A) Delivery of RI-TATp53C′-biotin to subcutaneous TA3/St tumors after IP administration to immune competent A/J mice. \(B) Reduction of solid TA3/St tumor growth in immune competent mice as a result of systemic administration of RI-TATp53C′. TA3/St cells were injected subcutaneously into A/J mice and allowed to grow to an average size of approximately 100 mm^3^. Mice were then sorted into treatment groups that received eight daily injections of vehicle (open circle) (*n* = 17), 650 μg of mutant peptide (open diamond) (*n* = 7), or 650 μg of wild-type RI-TATp53C′ peptide (open triangle) (*n* = 11). Final mean tumor volumes were 573 mm^3^ for vehicle-treated mice, 550 mm^3^ for mice treated with mutant peptide, and 268 mm^3^ for the wild-type RI-TATp53C′ peptide group. :::  ::: We next tested the ability of IP administration of RI-TATp53C′ peptide to inhibit the growth of distant solid tumors in immune competent mice. Subcutaneous tumors in mice receiving either vehicle or mutant peptide grew rapidly, reaching an average volume of nearly 600 mm^3^ by the end of treatment ([Figure 3](#pbio-0020036-g003){ref-type="fig"}B). In contrast, tumors in mice treated with wild-type RI-TATp53C′ peptide were significantly retarded in growth and reached a final mean volume less than 50% that of tumors in the control-treated mice (*p* = 0.01) ([Figure 3](#pbio-0020036-g003){ref-type="fig"}B). These observations demonstrate that systemic delivery of RI-TATp53C′ peptide in immune competent mice can significantly inhibit the growth of an aggressively proliferating solid tumor at a distant site. RI-TATp53C′ Peptide Treatment of Terminal Peritoneal Carcinomatosis {#s2c} ------------------------------------------------------------------- Because of their encapsulation and ectopic site of growth, subcutaneous tumors fail to replicate many of the features of terminal human cancer. We therefore tested the efficacy of RI-TATp53C′ peptide in a terminal peritoneal carcinomatosis mouse model that more closely resembles metastatic human disease. TA3/St carcinoma cells inoculated into the peritoneum of immune competent, syngeneic A/Jax (A/J) mice proliferated in a rapid logarithmic fashion, doubling in 24 h and increasing their numbers 100-fold 5 d postinoculation ([@pbio-0020036-Nagy1]). This aggressive, terminal peritoneal carcinomatosis model of human disease has been used extensively to study the pathophysiology of peritoneal tumor growth ([@pbio-0020036-Nagy1], [@pbio-0020036-Nagy2]). We assayed the ability of the RI-TATp53C′ peptide to alter the tumor burden and increase the longevity of mice harboring TA3/St peritoneal carcinomatosis. Vehicle-treated mice rapidly succumbed to peritoneal tumor burden with a mean survival time of 11 d ([Figure 4](#pbio-0020036-g004){ref-type="fig"}A). Mice treated with control mutant peptide succumbed to their tumor burden with similar kinetics and a mean survival time of 10 d ([Figure 4](#pbio-0020036-g004){ref-type="fig"}A). In contrast, peritoneal tumor-bearing mice treated with wild-type RI-TATp53C′ peptide lived on average more than 70 d after tumor inoculation ([Figure 4](#pbio-0020036-g004){ref-type="fig"}A), a greater than 6-fold increase in lifespan over mutant peptide- or vehicle-treated mice (*p* \< 10^−6^). These observations demonstrate the ability of transducible peptides to significantly extend survival in a mouse model of terminal peritoneal carcinomatosis. ::: {#pbio-0020036-g004 .fig} Figure 4 ::: {.caption} ###### RI-TATp53C′ Treatment Extends Survival of Mice Harboring Terminal Peritoneal Carcinomatosis \(A) A 6-fold increase in survival of A/J immune-competent mice harboring lethal TA3/St mammary peritoneal carcinomatosis burden after RI-TATp53C′ peptide treatment. A/J mice were given IP injections of TA3/St cells, and cells were allowed to double in number (approximately 24 h). Peritoneal tumor-bearing mice were then treated once a day for 12 consecutive days with vehicle (*n* = 15), 600 μg of wild-type RI-TATp53C′ (*n* = 10), or 600 μg of mutant peptide (*n* = 10). Mean survival duration was 11 d for vehicle-treated mice, 10 d for mice receiving mutant peptide, and greater than 70 d for the group receiving wild-type RI-TATp53C′ peptide. \(B) Reduction of tumor cell number in vivo by RI-TATp53C′ treatment. Mice were injected with TA3/St tumor cells and treated with wild-type peptide as in (A). Three days after tumor cell injection, cells were flushed from the peritoneal cavity and serially diluted in 6-well plates. Growth of colonies was then assessed by methylene blue staining and used to measure the number of viable tumor cells present in the peritoneum after treatment with vehicle or wild-type peptide. :::  ::: We next investigated the biological consequences of peptide treatment to tumor cells in vivo. Peritoneal-TA3/St tumor-bearing mice were given daily injections of wild-type RI-TATp53C′ peptide or vehicle control. Mice were sacrificed 3 d after tumor cell inoculation for assessment of tumor burden. Vehicle-treated mice contained a significant tumor burden of recoverable dividing TA3/St tumor cells ([Figure 4](#pbio-0020036-g004){ref-type="fig"}B). In contrast, mice treated with wild-type RI-TATp53C′ peptide showed a dramatic reduction in tumor cell number, suggesting that RI-TATp53C′ peptide treatment extended survival by directly inhibiting overall tumor proliferation. Consistent with cell culture studies, cell cycle analysis of tumor cells from peptide-treated mice showed an increase in G1 phase of the cell cycle (data not shown). RI-TATp53C′ Peptide Treatment of Terminal Peritoneal Lymphoma {#s2d} ------------------------------------------------------------- To broaden these results, we also tested the efficacy of the RI-TATp53C′ peptide in a mouse model of aggressive, disseminated peritoneal lymphoma. Wild-type RI-TATp53C′ peptide, but not mutant peptide, induced G1 phase accumulation and substantial apoptosis in Namalwa human lymphoma cells ([Figure 5](#pbio-0020036-g005){ref-type="fig"}A). When injected IP into SCID (severe combined immune deficiency) mice, Namalwa cells proliferate in the peritoneum and disseminate to other locations (e.g., spleen, lymph nodes, and blood \[[@pbio-0020036-de1]\]), modeling human B-cell lymphoma ([@pbio-0020036-Bertolini1]). Mice harboring peritoneal lymphoma succumbed to tumor burden with similar kinetics when treated with either vehicle or mutant peptide, with a mean survival time of 35 d and 33 d, respectively ([Figure 5](#pbio-0020036-g005){ref-type="fig"}B). In contrast, wild-type RI-TATp53C′ peptide treatment resulted in 50% long-term survival (*p* \< 0.0007) ([Figure 5](#pbio-0020036-g005){ref-type="fig"}B), with six of 12 treated mice still healthy at more than 200 d after tumor cell injection. Taken together, these observations demonstrate that in models of terminal metastatic human disease, transducible p53-activating peptides can modulate tumor biology in vivo, resulting in significantly decreased tumor burden, increased lifespan, and long-term disease-free survival. ::: {#pbio-0020036-g005 .fig} Figure 5 ::: {.caption} ###### RI-TATp53C′ Treatment Leads to 50% Long-Term Survival of Mice Bearing Terminal Peritoneal Lymphoma \(A) Treatment of human Namalwa B-cell lymphoma cells with RI-TATp53C′ peptide induces apoptosis. Cells were treated with wild-type or mutant peptide, and DNA content was analyzed by flow cytometry 24 h after peptide addition. \(B) Long-term survival of SCID mice harboring lethal peritoneal Namalwa lymphoma tumor burden after RI-TATp53C′ peptide treatment. Namalwa lymphoma cells were IP injected into SCID mice and allowed to proliferate for 48 h. Mice were then injected 16 times over 20 d with vehicle control (*n* = 16), 900 μg of wild-type RI-TATp53C′ peptide (*n* = 12), or 900 μg of mutant peptide (*n* = 6). Mean survival duration was 35 d for vehicle-treated mice and 33 d for mice receiving mutant peptide, whereas 50% of mice treated with wild-type RI-TATp53C′ peptide remained healthy at 150 d after tumor cell injection. :::  ::: The subset of RI-TATp53C′ peptide-treated animals that succumbed to peritoneal carcinomatosis could have failed treatment either because of the emergence of peptide-resistant tumor cells or because of insufficient treatment duration. To distinguish between these two possibilities, we isolated TA3/St and Namalwa cells from animals that failed treatment. In both cases, the cells readily proliferated in culture under the same conditions as the parental cell population (data not shown). RI-TATp53C′ peptide treatment of reconstituted TA3/St cells induced a G1 arrest similar in extent to that of the parental cell line ([Figure 6](#pbio-0020036-g006){ref-type="fig"}A). RI-TATp53C′ peptide treatment of reconstituted Namalwa cells also inhibited the viability of both parental and reconstituted cells to the same degree ([Figure 6](#pbio-0020036-g006){ref-type="fig"}B). These observations demonstrate that treatment failure is not due to acquisition of RI-TATp53C′ peptide resistance and suggests that an extended treatment protocol (greater than 12 d) may lead to a further enhancement of survival in these preclinical cancer models. ::: {#pbio-0020036-g006 .fig} Figure 6 ::: {.caption} ###### Tumor-Reconstituted Cells from Treated Mice Remain Sensitive to RI-TATp53C′ Peptide-Induced G1 Arrest or Apoptosis in Culture \(A) TA3/St cells were recovered from an A/J mouse treated with RI-TATp53C′ peptide and grown in DMEM/10% FBS. Recovered cells were treated with increasing concentrations of RI-TATp53C′ peptide and then analyzed for DNA content by flow cytometry 24 h later. \(B) Namalwa cells were recovered from a SCID mouse treated with RI-TATp53C′ peptide and grown in RPMI plus 10% FBS. Recovered cells were treated with increasing concentrations of RI-TATp53C′ peptide. After 2 d, the number of viable cells was assessed by Trypan blue exclusion and normalized to the number of viable untreated cells. Mean and standard deviation of multiple experiments are depicted. :::  ::: Discussion {#s3} ========== Advanced-stage peritoneal carcinomatosis and disseminated peritoneal lymphomas are often resistant to current chemotherapy treatment ([@pbio-0020036-Parsons1]), and new strategies for treating these diseases are clearly needed. The need to develop different therapeutic modalities to restore tumor suppressor function is acutely illustrated by the current limitations of viral/DNA-based strategies for delivering tumor suppressor genes to cancer cells in patients ([@pbio-0020036-McCormick1]). Here we show that macromolecular biological cargo can be delivered via TAT-mediated transduction in order to modulate tumor biology in vivo. Specifically, we find that delivery of a transducible p53-activating peptide in sensitive tumor cells inhibits solid tumor growth in vivo (see [Figure 3](#pbio-0020036-g003){ref-type="fig"}) and dramatically extends survival (greater than 6-fold), yielding disease-free animals in terminal peritoneal cancer models of human metastatic disease (see Figures [4](#pbio-0020036-g004){ref-type="fig"} and [5](#pbio-0020036-g005){ref-type="fig"}). The vast majority of tumors express either wild-type p53 protein or a full-length p53 point mutant ([@pbio-0020036-Vousden1]). This observation has led to the hypothesis that reactivation of endogenous p53 protein will be a useful means of treating cancer. The data presented here provide evidence for this hypothesis by showing that TAT-mediated delivery of a p53-activating peptide in vivo is an effective treatment for multiple preclinical cancer models. This macromolecular approach to p53 reactivation has certain advantages over the limited number of small molecule-based strategies reported to reactivate mutant p53 in vivo ([@pbio-0020036-Foster1]; [@pbio-0020036-Bykov1]). First, the RI-TATp53C′ peptide can activate wild-type p53 in addition to several p53 contact mutants. Second, small molecules may suffer from a lack of specificity ([@pbio-0020036-Rippin1]) in comparison to larger, more information-rich macromolecules. Finally, recent investigations into the mechanism of TAT-mediated transduction ([@pbio-0020036-Richard1]; [@pbio-0020036-Fittipaldi1]) suggest that, unlike small molecules, TAT-linked cargo is taken up by macropinocytosis (Wadia et al. 2004) and is therefore not susceptible to the multidrug resistance phenotype. Theoretically, tumors could avoid the RI-TATp53C′ peptide action by mutating or deleting p53; however, we did not observe peptide resistance here (see [Figure 6](#pbio-0020036-g006){ref-type="fig"}). Therefore, we conclude that linking PTDs to p53C′ and to other p53-activating peptides may be an effective therapeutic strategy applicable to a significant fraction of human cancers. The work presented here provides several broad lines of evidence for the general feasibility of applying in vivo TAT-mediated transduction to cancer therapy. First, we find that inversion of the p53C′TAT peptide sequence and synthesis with D-amino acids results in a highly stable peptide (RI-TATp53C′) that retains both biological activity and the ability to transduce into cells. Given the rapid degradation of L-residue-containing peptides in vivo ([@pbio-0020036-Chorev1]), use of retro-inverso transformations with D-isomer residues and/or other stabilizing procedures will likely be essential for the pharmacological use of transducible peptides. Second, given the history of virus-mediated gene delivery, the necessity of validating new therapeutic approaches to systemic disease in the context of an intact immune system cannot be underestimated. Consequently, here we demonstrate that TAT-mediated systemic delivery inhibits tumor growth in immune competent animals. Finally, most studies on anticancer transduction peptides have relied primarily on the use of solid, subcutaneous tumor growth as a measure of efficacy ([@pbio-0020036-Datta1]; [@pbio-0020036-Harada1]). Although informative, such studies are inherently limited by the minimal impact that subcutaneous tumors have on the biology of the host and by the failure of this type of tumor to closely mimic human disease. In contrast, the more rigorous peritoneal carcinomatosis and peritoneal lymphoma models used here require that therapeutic agents be able to suppress tumors to such an extent that the deleterious effects of the tumor on host physiology are substantially ameliorated. This is a particularly salient point because cancer patients generally do not succumb to the primary tumor burden but to complications from metastatic disease ([@pbio-0020036-Fidler1]). Indeed, anticancer therapeutics are defined as clinically successful by their ability to alleviate pathology and extend survival and not simply by their ability to reduce tumor volume. Our work here, combined with that of [@pbio-0020036-Fulda1]), demonstrates that transducible agents can effectively treat rigorous models of terminal cancer. Current clinical use of macromolecular biological therapies is limited to agents that have an extracellular mode of action. The preclinical data presented here demonstrate a proof-of-concept that intracellular delivery of biologically active macromolecular cargo by TAT-mediated transduction can modify specific pathways in vivo and that this approach potentially serves as a foundation for the generation of new classes of intracellular biological therapeutics. Materials and Methods {#s4} ===================== {#s4a} ### Cell culture and flow cytometry. {#s4a1} TA3/St (gift of W. G. Kaelin), H1299 (gift of R. K. Brachmann), and human foreskin fibroblast (M. Haas) cells were maintained in DMEM plus 10% fetal bovine serum (FBS) and penicillin/streptomycin (P/S). Namalwa cells (American Type Culture Collection, Manassas, Virginia, United States) were maintained in RPMI plus 10% FBS, P/S. HCT116 cells (gift of B. Vogelstein) were grown in McCoy\'s medium plus 10% FBS, P/S. All cells were maintained at 37°C in 5% CO~2~. Short-term cell viability was assessed by counting Trypan blue-excluding cells on a hemocytometer. Long-term cell viability was assessed by colony formation assay. After serial dilution and 10 d of culture, colonies were washed in PBS and stained with 1% methylene blue. Cellular senescence was assessed by X-Gal staining as previously described ([@pbio-0020036-Schwarze1]), except for the use of PBS (pH 6.0). For cell cycle analysis, TA3/St cells were treated with 0.25--10 μM peptide and Namalwa cells with 40 μM peptide. DNA was stained 24 h later with 10 μg/ml propidium iodide in 0.5% NP-40 (TA3/St cells) or Draq5 (Namalwa cells) per the manufacturer\'s instructions (Qbiogene, Carlsbad, California, United States). DNA profiles were analyzed using a FACScan and CellQuest software (Becton Dickinson, Palo Alto, California, United States). ### Peptide synthesis. {#s4a2} Peptides were synthesized by standard Fmoc chemistry on an ABI 433A Peptide Synthesizer (Applied Biosystems, Foster City, California, United States). Crude peptides were purified by reverse-phase HPLC over a C18 preparatory column (Varian, Palo Alto, California, United States). The identity of all peptides was confirmed by mass spectrometry. ### Promoter activity assays. {#s4a3} In a 96-well dish, 4 × 10^4^ cells were plated per well. The next day, H1299 cells were transfected with 15 ng of TK-Renilla (Promega, Madison, Wisconsin, United States), 200 ng of PG13-Luc, and one of the following: 0.3 ng of empty vector, 0.3 ng of p53 expression vector, or 1 ng mutant of p53 expression vector (gift of R. K Brachmann). SW480 cells were transfected with 25 ng of TK-Renilla and 250 ng of PG13-Luc reporter plasmid. Cells were all transfected using Lipofectamine 2000 per the manufacturer\'s protocol (Invitrogen, Carlsbad, California, United States). After 5 h, the transfection medium was removed and peptides were added to cells. Luciferase activity was measured 24 h later with the Dual Luciferase Reporter Assay System per the manufacturer\'s instructions (Promega). ### Animal tumor models. {#s4a4} For TA3/St tumor models, 4- to 8-wk-old immune competent A/J female mice were obtained from Jackson Laboratory (Bar Harbor, Maine, United States). Solid TA3/St tumors were generated by subcutaneous injection of 1.5 × 10^6^ TA3/St cells in 200 μl of Hanks\' balanced salt solution (HBSS). Tumor volume was estimated by *V* = (*a* ^2^ × *b*)/2, where *a* is the short axis and *b* is the long axis of the tumor. IP TA3/St tumors were generated by injection of 2 × 10^6^ TA3/St cells IP in 400 μl of HBSS. For the Namalwa lymphoma tumor model, 6- to 8-wk-old CB17 SCID female mice were obtained from Charles River Laboratory (Wilmington, Massachusetts, United States). Then, 5 × 10^5^ Namalwa lymphoma cells were injected IP in 400 μl of HBSS. Peptide was dissolved in water, brought to 600 μl in PBS, and injected IP. All animal studies were approved by the University of California, San Diego, Institutional Animal Care and Use Committee. ### Histology. {#s4a5} Mice harboring solid TA3/St tumors were injected with 650 μg of biotinylated RI-TATp53C′ peptide and sacrificed 1 h postinjection. Sections from frozen tumors were stained with Vectastain Elite ABC Kit and DAB substrate per the manufacturer\'s instructions (Vector Laboratories, Burlingame, California, United States). ### Statistical analysis. {#s4a6} Student\'s *t*-test was used to determine statistical significance (*p* \< 0.05) in all experiments except animal survival experiments, in which the Wilcoxon Rank-Sum Test was performed. We thank R. K. Brachmann, B. Vogelstein, W. G. Kaelin, and M. Haas for reagents and cell lines. We thank P. Gent for technical assistance with mice and L. Gross for assistance with mass spectrometry. This work was supported by the Howard Hughes Medical Institute and the National Institutes of Health (CA96098). **Conflicts of interest.** SFD is the scientific founder of Ansata Therapeutics, a biotech company utilizing protein transduction technology to treat human disease. However, he owns less than 5% of the stock, does not receive any research funds from Ansata, does not have a seat on the board of directors, and has no say in the governance of the company. **Author contributions.** ELS and SFD conceived and designed the experiments. ELS, BRM, and CCS performed the experiments. CCS and SFD analyzed the data. BRM and CCS contributed reagents/materials/analysis tools. ELS and SFD wrote the paper. DOI: [10.1371/journal.pbio.0020036](10.1371/journal.pbio.0020036) Copyright: © 2004 Snyder et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Academic Editor: Nicholas Hastie, Western General Hospital A/J : A/Jax strain FBS : fetal bovine serum HBSS : Hanks\' balanced salt solution IP : intraperitoneal P/S : penicillin/streptomycin PTD : protein transduction domain SCID : severe combined immune deficiency

PubMed Central

2024-06-05T03:55:46.081370

2004-2-17

PMC340945

For people who received their introduction to cancer genetics in college in the first half of the 1990s, everything looked simple and straightforward. It was the stuff you could explain to sincerely interested relatives who wanted to know what you were spending your time on. There were oncogenes and there were tumour suppressor genes. Oncogenes were overactive genes and proteins that somehow caused cancer because they were overactive; therefore, they were dominant. Tumour suppressor genes were genes that would normally prevent a tumour from happening and that needed to be inactivated for a tumour to start to form; both copies of a tumour suppressor gene had to be inactivated, and the mutation was recessive. If inactivation of these genes is a random process, it was understandable that people who inherit an inactivated copy of a tumour suppressor gene had a higher risk of developing the associated form(s) of cancer than people born with two normal copies, as postulated in [@pbio-0020040-Knudson1] two-hit model. And, indeed, it was shown that in the tumours in these predisposed patients, the remaining wild-type copy of the tumour suppressor gene was lost, a process referred to as loss of heterozygosity. For me, in 1998 things started to change. [@pbio-0020040-Venkatachalam1] published a paper in the *EMBO Journal* describing a detailed study of tumours in mice lacking one copy of the p53 tumour suppressor gene (*Trp53*). This gene is known to be the most mutated gene in human cancer and its function to be central to many processes that are involved in the cellular prevention of cancer. Mice lacking both copies of this gene are for the most part viable, but succumb to cancer (mainly thymic lymphomas) at three to five months of age ([@pbio-0020040-Donehower1]). Mice born with one copy of the *Trp53* gene start to develop cancer at around nine months, and incidence increases with age. In their study, Venkatachalam and colleagues analysed an impressive group of 217 *Trp53* ^+/−^ mice of controlled genetic background and followed the fate of the *Trp53* wild-type allele in the tumours. According to the two-hit model, it was expected that in these tumours this copy would have been lost or inactivated. However, this turned out not to be the case. Half of the tumours from mice younger than 18 months were found to have retained the wild-type copy of *Trp53*, a number that increased to 85% in mice older than 18 months. In two tumours, the researchers sequenced the complete coding region of the remaining wild-type allele and showed it was structurally intact. To exclude the possibility of downregulation or inactivation at the level of protein expression, they irradiated tumour-bearing mice prior to sacrifice, a treatment known to increase p53 protein levels via posttranslational mechanisms. Their data showed the retained wild-type allele in these tumours was expressed normally and suggested it had a normal wild-type conformation. Next, the authors did a rigorous test of different functions of the p53 protein. They first tested whether the tumours showed amplification of Mdm2. This protein, whose expression is regulated by p53, stimulates breakdown of p53, thereby forming a negative feedback mechanism that keeps p53 levels low. Some tumours therefore amplify the *Mdm2* gene as a means of inactivating p53. However, this was not found in the tumours from the *Trp53* ^+/−^ mice. Subsequently, the researchers tested to what extent the retained *Trp53* copy behaved normally. Irradiation of many tissues leads to p53-dependent apoptosis, and, indeed, in tumours that had retained the wild-type allele, irradiation did lead to an increase in apoptosis, whereas in tumours that had lost the wild-type allele, it did not. The p53 protein is known to function as a transcriptional regulator by either up- or down regulating target genes in response to different forms of cellular stress, including irradiation-induced DNA damage. The authors studied the expression of two p53-upregulated genes (*Cdkn1a*, which encodes p21, and *Mdm2*) and one downregulated gene (*Pcna*) in p53-positive tumours after irradiation and showed responses indicative of normal p53 function. Furthermore, it was shown that the p53 protein from the tumours was able to bind to a p53-binding DNA sequence in an in vitro setting. Finally, since it is known that p53 absence in tumours is correlated with chromosomal instability, [@pbio-0020040-Venkatachalam1] used comparative genome hybridisation to compare this feature between p53-negative and p53-positive tumours and found a 5-fold greater stability in the latter. In short, this paper clearly showed that, at least in mice, in many *Trp53* ^+/−^ tumours the wild-type allele of *Trp53* is not only retained, but also appears to function normally. This strongly suggested that a decrease of dosage in p53 is already sufficient for tumourigenesis, a phenomenon referred to as haploinsufficiency. Shortly before, the group of Moshe Oren ([@pbio-0020040-Gottlieb1]) had shown that a *Trp53* ^+/−^ background leads to a greater than 50% reduction in p53 activity using a p53-responsive *lacZ* reporter gene in transgenic mice. Venkatachalam and colleagues suggested the strong concentration dependence of p53 function could be explained by the fact that p53 functions as a tetramer. A 50% decrease in p53 monomers can easily be imagined to result in a greater than 50% decrease in functional tetramers, which in turn increases the chances of these cells becoming cancerous. This paper by [@pbio-0020040-Venkatachalam1] made me realise how important it is to remain critical, even of long-established theories and models. Since then, haploinsufficient mechanisms have been described in more tumour suppressor genes in humans and mice (reviewed in [@pbio-0020040-Fodde1]). For instance, in a recent paper in *PLoS Biology*, [@pbio-0020040-Trotman1] used mouse models to describe how the dosage of the *Pten* tumour suppressor gene influences the occurrence of prostate cancer. Further genes have been described with other unexpected roles in the tumourigenic process. There is a long-standing debate in the literature about the number and role of mutations in a tumour, and without going into the details, it is clear that haploinsufficient mechanisms for tumour suppressor genes will greatly influence the statistics on which these discussions are based. At a time when microarray analysis has become a standard experiment and the many thousands of changes in tumour cells are analysed across the whole genome, it is important to keep in mind that the correct interpretation of this wealth of information might be more complicated than the widely accepted models would have us believe. ::: {#pbio-0020040-g001 .fig} Figure 1 ::: {.caption} ###### Initiating Genetic Aberrations in Tumourigenesis \(A) According to the two-hit model, the first hit at the rate-limiting tumour suppressor gene provides no selective advantage for the cell. Only after the loss of the second allele of this gene is tumour formation initiated. Extra genetic changes are needed for complete transformation of the cell. \(B) In a haploinsufficient mechanism, the first hit on the rate-limiting tumour suppressor gene already provides the cell with sufficient selective advantage to initiate tumour formation. Further events are necessary for complete transformation. These events might or might not include the loss of the wild-type allele of the rate-limiting tumour suppressor gene. :::  ::: Peter Hohenstein is a European Union Marie Curie postdoctoral research fellow in the laboratory of Nick Hastie at the Department of Comparative and Developmental Genetics of the Medical Research Council Human Genetics Unit at Western General Hospital in Edinburgh, United Kingdom. E-mail: <peter.hohenstein@hgu.mrc.ac.uk>

PubMed Central

2024-06-05T03:55:46.084361

2004-2-17

PMC340946

Of the 300 or so viruses that cause disease in humans, HIV may have the greatest adaptive advantage. Like most persistent viruses---including the herpesviruses Epstein--Barr and cytomegalovirus (CMV)---HIV employs various strategies to counteract its host\'s response to infection. But HIV possesses a unique ability to sustain a progressive attack on the immune system---infecting the very cells that coordinate the immune response---leaving the body susceptible even to normally harmless microorganisms. It is these so-called opportunistic infections, rather than the human immunodeficiency virus itself, that makes HIV so deadly. The specific mechanisms that engineer this ongoing systemic attack have been the subject of intense research. HIV targets white blood cells with protein surface receptors called CD4. These CD4, or helper, T-cells normally orchestrate the body\'s immune response by signaling killer T-cells (which are also called CD8 T-cells, after their CD8 surface receptors) and other immune cells to multiply and differentiate---that is, become specially equipped to recognize a particular pathogen, or antigen. At the onset of infection, the immune system appears to respond normally, with a strong attack led by HIV-specific CD8 T-cells that initially contain the virus. But as the infection progresses, CD4 counts drop and the body\'s ability to renew T-cells decreases while its proportion of "antigen-experienced" CD8 T-cells increases. While the biological effect of this hyperactivity is unclear, it is apparent that patients with elevated immune activity face a poor prognosis. Investigating the interaction among immune activation, CD8 T-cell differentiation, and HIV prognosis, Victor Appay and colleagues report that a close connection between elevated immune activation and elevated levels of highly differentiated T-cells may bring further insights into how HIV exhausts the immune system. To examine the effect of elevated immune activation on T-cells, the researchers analyzed T-cells from a group of HIV-infected individuals collected at two distinct points in time: at the onset of acute infection---which is characterized by vigorous HIV replication---and after treatment, when viral replication is suppressed. To explore the connection between T-cell differentiation and clinical status, the researchers analyzed the T-cells from a group of untreated infected individuals divided into three subsets based on stage of infection: acute infection, chronic infection without progression, and chronic infection with signs of progression. During acute HIV infection, the vast majority (80%--90%) of the CD8 T-cell population was activated---not just the HIV-specific CD8 T-cells. Surprisingly, CD8 T-cells specific to the Epstein--Barr and CMV viruses showed significant activation levels during acute infection, suggesting that HIV may indirectly promote the replication of these viruses. When the researchers investigated the effects of this activation on T-cell differentiation, they found a correlation between increasing antigen concentrations and increasing CD8 T-cell activation and proliferation. And when Laura Papagno et al. analyzed the differentiation state of CD8 T-cells in individuals at different stages of infection, they found a progression in the proportion of highly differentiated CD8 T-cells associated with HIV disease progression. These results, the researchers conclude, show that chronic overactivation of the immune system during HIV infection produces the large pool of highly differentiated T-cells observed in HIV infection. T-cells go through various stages toward late differentiation, and it may be that the early-differentiated CD8 T-cells, which maintain the ability to proliferate, offer protective immunity. But highly differentiated cells, they propose, exhibit characteristics associated with "replicative senescence"---they are in effect old, worn-out cells that can no longer proliferate. Though replicative senescence is a natural process for most cells, in the context of HIV---in which infected individuals also lose the ability to replenish T-cells---it creates an aging population of T-cells that are less effective at fighting infection. ::: {#pbio-0020041-g001 .fig} ::: {.caption} ###### Two T-cells, one of which recognizes a target cell :::  :::

PubMed Central

2024-06-05T03:55:46.085239

2004-2-17

PMC340947

Introduction {#s1} ============ It is well established that learning can have a strong impact on neural responses to visual stimuli in high-level association cortices such as inferior temporal (IT) or prefrontal (PF) cortex, where the activity of single neurons reflects learning in pair association, object identification, or categorization tasks ([@pbio-0020044-Sakai1]; [@pbio-0020044-Logothetis1]; [@pbio-0020044-Booth1]; [@pbio-0020044-Kobatake1]; [@pbio-0020044-Erickson1]; [@pbio-0020044-Rainer1]; [@pbio-0020044-Freedman1]; [@pbio-0020044-Sigala1]). In these studies, learning is thought to modify neural activity to represent task-relevant attributes, such as trained views of three dimensional objects ([@pbio-0020044-Logothetis1]) or associations between paired visual stimuli ([@pbio-0020044-Sakai1]; [@pbio-0020044-Erickson1]). The learned representations often exhibit invariance for stimulus features such as size ([@pbio-0020044-Logothetis1]), rotation ([@pbio-0020044-Booth1]), or stimulus degradation ([@pbio-0020044-Rainer1]). Similar neural activity to within-category stimuli during categorization ([@pbio-0020044-Freedman1]) can also be thought of as a learning-dependent form of invariance. Several lines of evidence suggest that these learning effects involve synaptic plasticity and thus represent long-lasting modifications to visual association cortices. Recent evidence suggests that neurons in early visual sensory areas can also modify their response properties with learning. In particular, several studies have revealed learning-related changes in primary visual cortex (V1) ([@pbio-0020044-Crist1]; [@pbio-0020044-Schoups1]; [@pbio-0020044-Ghose1]), although the extent and functional significance of these learning effects remains somewhat controversial ([@pbio-0020044-Schoups1]; [@pbio-0020044-Ghose1]). Available evidence suggests that classical V1 response properties such as receptive field size or orientation tuning parameters are affected relatively little by learning, while learning does appear to cause general reduction in activity for trained stimuli as well as a task-dependent increase in the influence of nonclassical surround stimulation on the neuron\'s response. Learning thus appears to affect both low and high level areas of the ventral visual stream. The results obtained by studies in these two areas are, however, difficult to compare directly, owing to substantial differences in experimental design. In studies of IT or PF cortex, studies typically employ 'complex\' visual stimuli such as Fourier descriptors ([@pbio-0020044-Sakai1]), computer-rendered animals ([@pbio-0020044-Freedman1]), or colored photographs and artwork ([@pbio-0020044-Erickson1]). These stimuli are generally presented at the center of gaze and can be from 1° up to 10° of visual angle in size. Many studies also include a selection process that determines which of the neurons encountered in a given penetration are chosen for further quantitative study. By contrast, available learning studies in early visual areas follow well-established rules for investigation of primary and extrastriate visual areas. These studies employ 'simple\' visual stimuli such as oriented bars ([@pbio-0020044-Crist1]) or gratings ([@pbio-0020044-Schoups1]; [@pbio-0020044-Ghose1]). These stimuli are generally presented at eccentric locations, with stimulation parameters adjusted to the receptive field and orientation selectivity of the single neuron currently under investigation. Thus, both stimulus type and experimental procedure generally differ substantially, depending on whether a study investigates low-level sensory or high-level associative visual cortex. For a comprehensive account of how learning affects visual processing, the same stimuli and experimental procedure must be used to study different levels of the visual processing hierarchy. What kind of stimuli might be suitable to study visual areas as different as early sensory visual and PF cortex? We decided to use natural images for several reasons: The primate visual system evolved in the natural environment under conditions of 'natural\' stimulation; much is known about their statistical properties and they can therefore be well-controlled; they contain structure at all spatial scales and thus can be expected to activate a large fraction of visually responsive neurons. We avoid subjectively biasing our sample of recorded neurons by always recording from the first neurons whose waveforms we are able to reliably isolate. This ensures that our population of recorded neurons represents an unbiased sample in each brain region under study, and this in turn allows us to compare data obtained from different brain regions. We obtain a sensitive measure of behavioral performance and associated neural activity by employing a stimulus degradation procedure that makes stimuli harder to discriminate by adding various amounts of noise (see [Figure 1](#pbio-0020044-g001){ref-type="fig"}A). With degradation, stimuli become increasingly indeterminate because all stimuli in a given session are combined with the same noise pattern. Noise is newly generated for every session so that monkeys cannot rely on the specific individual characteristics of a particular noise pattern. Instead, they need to extract task-relevant information from degraded displays, whose particular details vary from day to day. Similarly, outside the laboratory we are rarely presented with familiar stimuli in canonical views and conditions of standard lighting, but instead need to extract this information from complex scenes in which it is embedded. Previously these kind of stimuli were used to study neural activity in the PF cortex ([@pbio-0020044-Rainer1]), where learning made neural activity more robust to stimulus degradation. After learning, PF neurons tended to fire in a similar manner to undegraded and moderately degraded versions of the same stimulus. Learning thus resulted in a form of neural response invariance, because degradation no longer had an impact on PF neural activity. ::: {#pbio-0020044-g001 .fig} Figure 1 ::: {.caption} ###### Stimuli and Behavioral Task \(A) An example natural image is shown at three coherence levels, corresponding to 100% (undegraded), 45% (degraded), and 0% (pure visual noise). \(B) The sequence of trial events for the DMS task used in this study. After a fixation period, a sample stimulus (S) is briefly presented, followed by a delay period and the presentation of a probe stimulus (P). While sample stimuli were presented at different coherence levels, probe stimuli were always presented in undegraded form (100% coherence). The monkeys were required to release a lever if the probe matched the sample. :::  ::: Here our aim is to use similar stimuli and behavioral procedures to characterize how learning modifies neural activity in extrastriate visual cortical area V4. Area V4 was chosen because it is considered to be a sensory visual area at an intermediary processing stage in the ventral stream and because it is directly connected to parts of the PF cortex ([@pbio-0020044-Petrides1]). Our task was a modified version of delayed-matching-to-sample (DMS) (see [Figure 1](#pbio-0020044-g001){ref-type="fig"}B). After grasping a metal lever and subsequently attaining central fixation, monkeys viewed a sample stimulus presented at one of six coherence levels ranging from undegraded (100% coherence) to fully degraded (0% coherence). After a brief delay, monkeys were presented with a probe stimulus (always at 100% coherence) and had to release a lever if the probe matched the sample (i.e., if the sample was identical to or was a degraded version of the probe stimulus). During each session, we employed four highly familiar stimuli and four 'novel\' stimuli that monkeys had not seen previously. Great care was taken to ensure that novel and familiar images differed only in terms of their familiarity to the animal (see [Materials and Methods](#s4){ref-type="sec"}). Using novel and familiar stimuli allowed us to ask whether learning had any effect on monkeys\' ability to identify degraded and undegraded versions of natural images. Intermixing novel and familiar images in the same session had the additional advantage of allowing us to estimate for each single neuron in our population, whether there were any learning-related changes in the amount of stimulus-specific information these neurons communicated. Results {#s2} ======= We found that learning resulted in significant and robust improvements in monkeys\' ability to identify degraded stimuli. Behavioral performance varied systematically with coherence ([Figure 2](#pbio-0020044-g002){ref-type="fig"}A). Monkeys performed at chance level (50% correct) when stimuli were presented at 0% coherence and thus contained no task-relevant information. For degraded stimuli (35%--65% correct), monkeys performed significantly better with familiar than with novel stimuli (*t*-test, *p* \< 0.01). For undegraded stimuli at 100% coherence, the monkeys\' performance was near ceiling for both novel and familiar stimuli (92% and 95% respectively; t-test, *p* = 0.12). Learning-dependent performance improvements for degraded stimuli were highly consistent across stimuli and monkeys. There were in fact no significant differences in the monkeys\' performance to each of the familiar stimuli across sessions at all coherence levels (one-way ANOVAs, *p* \> 0.1), and this was also true for novel stimuli. In addition, performance for novel and familiar stimuli did not differ significantly between the two monkeys at any coherence level (t-tests, *p* \> 0.1). Note that the monkeys\' excellent perfor-mance with undegraded novel objects reflects the fact that they have acquired the rule of the DMS task and are thus able to perform it near ceiling with novel stimuli. The timecourse of this learning-dependent difference in performance is shown in [Figure 2](#pbio-0020044-g002){ref-type="fig"}B. Session 1 represents a session in which a set of four initially novel stimuli is arbitrarily chosen and kept constant in subsequent sessions, thus becoming more and more familiar. Comparing performance for these stimuli with performance of novel stimuli that are randomly chosen in each session reveals that it takes several sessions for the learning effect to appear. Performance averaged across the first five session was similar for novel and familiar stimuli (*t*-test, *p* = 0.43). Furthermore, the learning-dependent difference in performance appeared to asymptote after around ten sessions. In summary, learning led to robust improvements in the monkeys\' ability to identify degraded natural images while the monkeys performed near ceiling for novel and familiar undegraded images. ::: {#pbio-0020044-g002 .fig} Figure 2 ::: {.caption} ###### Learning Improved Monkeys\' Ability to Identify Degraded Stimuli \(A) Behavioral performance for the sessions during which neural data was collected (*n* = 11) is shown as a function of the coherence of the sample stimulus for novel and familiar stimuli. Asterisks denote significant differences in performance for novel and familiar stimuli. \(B) The performance at 45% coherence (%Correct~45~) is shown for a set of novel stimuli that is introduced in the first session and then used during all subsequent sessions and thus becomes more and more familiar during subsequent sessions (circles). For comparison, performance with novel stimuli that are new and unique to each session is shown (diamonds). Sessions 1--20 represent purely behavioral training sessions (TRAIN), and sessions 21--26 represent combined behavioral and single unit recording sessions (REC). :::  ::: We now examine whether there were any learning-depen-dent changes in the activity of V4 neurons. Results described in this report are based on a population of 83 V4 neurons. We first asked whether there was any general difference in mean activity elicited by novel and familiar undegraded stimuli. We found that the response of V4 neurons to novel (〈FR~nov~〉 = 36.7 ± 2.8 Hz) and familiar stimuli (〈FR~fam~〉 = 34.2 ± 2.7 Hz) was similar (*t*-test, *p* = 0.14; see also [Table 1](#pbio-0020044-t001){ref-type="table"}). Out of the 14 neurons that individually showed a significant difference in activity between novel and familiar stimuli (*t*-test, *p* \< 0.05), similar fractions preferred familiar or novel stimuli (6/14 or 43% and 8/14 or 57% respectively; χ^2^ test, *p* = 0.45). We thus found that learning did not lead to a change in the average activity of V4 neurons for undegraded stimuli. Next, we examined whether learning resulted in any change in the amount of stimulus-specific information that V4 neurons communicated. To do this, we computed the mutual information between the set of four familiar or novel stimuli and the associated neural responses ([see Materials and Methods](#s4){ref-type="sec"}). We found that V4 neurons on average communicated similar amounts of information about novel and familiar undegraded stimuli ([Figure 3](#pbio-0020044-g003){ref-type="fig"}A). The average information communicated by each neuron in the entire population of 83 V4 neurons was similar for novel stimuli 〈I~nov~〉 = 0.48 bits and for familiar stimuli 〈I~fam~〉 = 0.45 bits (*t*-test, *p* = 0.16). We selected 25% of the population (21 out of 83 neurons), which communicated most information about novel or familiar objects ([see Materials and Methods](#s4){ref-type="sec"}). For this population of most informative neurons (white circles in [Figure 3](#pbio-0020044-g003){ref-type="fig"}A), we also found no difference between novel and familiar stimuli (〈I~nov~〉 = 0.67 bits, 〈I~fam~〉 = 0.65 bits; *t*-test, *p* = 0.48). Thus, for natural images (undegraded stimuli) we saw no significant learning-dependent difference in performance and also no learning-dependent changes in the average activity or in the amount of stimulus-specific information communicated by V4 neurons. ::: {#pbio-0020044-g003 .fig} Figure 3 ::: {.caption} ###### Learning Led to an Increase in V4 Neural Information about Degraded but Not Undegraded Stimuli Here we summarize how much information V4 neurons communicated about novel (I~nov~) and familiar (I~fam~) stimuli for undegraded (A) and degraded (B) stimuli. Each symbol in the scatter plot represents a single neuron and shows how much information this neuron communicated about familiar (x-axis) and novel (y-axis) stimuli. In each scatter plot, white-shaded symbols represent the 25% most informative neurons, i.e., the one-quarter of the population communicating most information about either familiar or novel stimuli. The remaining three-quarters of the population are shown in gray shading. The single neuron example in [Figure 5](#pbio-0020044-g005){ref-type="fig"} is represented by the star. The black 'x\' represents the population mean for the 25% most informative neurons. :::  ::: ::: {#pbio-0020044-t001 .table-wrap} Table 1 ::: {.caption} ###### Mean Sample-Evoked Activity and Information Values :::  This table summarizes the mean sample-evoked activity and information values for the entire population and for the set of informative V4 neurons (open circles in [Figure 3](#pbio-0020044-g003){ref-type="fig"}B) as a function of degradation and learning. Significance of paired *t*-tests comparing estimated values for familiar versus novel stimuli are reported to the right of each pair of values NS, not significant; \* *p* \< 0.05; \*\* *p* \< 0.001 ::: At intermediate coherence levels, the monkeys\' ability to correctly identify degraded stimuli was improved by learning, and we asked whether this behavioral improvement was associated with any changes in the activity of V4 neurons. We found that V4 neurons indeed communicated significantly more information about degraded familiar than about degraded novel stimuli ([Figure 3](#pbio-0020044-g003){ref-type="fig"}B). Considering the entire population, learning led to a significant increase in information about degraded stimuli from 〈I~nov~〉 = 0.34 bits to 〈I~fam~〉 = 0.40 bits (*t*-test, *p* \< 0.05). For the 25% most informative neurons (white circles in [Figure 3](#pbio-0020044-g003){ref-type="fig"}B), we observed an even larger change from 〈I~nov~〉 = 0.47 bits to 〈I~fam~〉 = 0.67 bits (*t*-test, *p* \< 0.001), corresponding to a 40% increase in information with learning. We further characterized this effect by examining how degradation affected the amount of information separately for novel and familiar stimuli. For both novel ([Figure 4](#pbio-0020044-g004){ref-type="fig"}A) and familiar ([Figure 4](#pbio-0020044-g004){ref-type="fig"}B) stimuli, V4 neurons communicated on average more information about undegraded (I~100~) than degraded (I~degrad~) stimuli (paired *t*-tests, *p* \< 0.001), reflecting the fact that behavioral perfor-mance was better for undegraded than degraded stimuli (see [Figure 2](#pbio-0020044-g002){ref-type="fig"}A). The ΔI distributions (I~100~− I~degrad~) for familiar and novel stimuli shown in the insets ([Figure 4](#pbio-0020044-g004){ref-type="fig"}A and [4](#pbio-0020044-g004){ref-type="fig"}B), however, differed significantly (paired *t*-test, *p* \< 0.001), and learning was associated with a rightward shift in this distribution (〈ΔI~fam~〉 = 0.06, 〈ΔI~nov~〉 = 0.13). Interestingly, the kurtosis or skewness of the ΔI distribution changed by an order of magnitude from 0.13 for novel stimuli to 5.5 for familiar stimuli, similar to experience-dependent effects that have been observed on hippocampal place cell activity ([@pbio-0020044-Mehta2]; [@pbio-0020044-Mehta1]). As a consequence of these learning-dependent changes, many V4 neurons actually communicated more information about degraded than undegraded familiar stimuli (25/83 or 30%), whereas only a small minority did so for novel stimuli (6/83 or 7%). This difference in proportions was significant (χ^2^ test, *p* \< 0.001). Taken together, learning accordingly resulted in an increase in the amount of information communicated by V4 neurons about degraded stimuli and many neurons actually communicated more information about degraded than undegraded familiar stimuli. ::: {#pbio-0020044-g004 .fig} Figure 4 ::: {.caption} ###### Many Neurons Communicated More Information about Degraded than Undegraded Familiar Stimuli Here we replot the data from [Figure 3](#pbio-0020044-g003){ref-type="fig"} to illustrate how much information V4 neurons communicated about degraded (I~degrad~) and undegraded (I~100~) stimuli separately for novel (A) and familiar (B) stimuli. Each symbol in the scatter plot represents a single neuron. The insets show how degradation affected the information communicated by V4 neurons, by plotting histograms of the ΔI distributions (I~100~− I~degrad~) for novel and familiar stimuli. While 25 neurons (30% of the population) communicated more information about degraded than undegraded familiar stimuli, only six neurons (7% of the population) did so for novel stimuli. :::  ::: How did single V4 neurons mediate this learning-depen-dent increase in information about degraded stimuli? The activity of an example neuron is shown in [Figure 5](#pbio-0020044-g005){ref-type="fig"} in histogram and raster format for its preferred and nonpreferred familiar stimulus. This neuron showed little or no response to pure visual noise (0% coherence) or to its nonpreferred stimulus at any coherence level ([Figure 5](#pbio-0020044-g005){ref-type="fig"}B). It was activated to a peak firing rate of about 20Hz by its preferred stimulus (red curve in [Figure 5](#pbio-0020044-g005){ref-type="fig"}A). Degradation of the preferred stimulus resulted in brisk activity of this neuron, and activity was greater to the preferred stimulus at all intermediate coherence levels (35%--65%) than to the undegraded preferred stimulus (paired *t*-tests, *p* \< 0.01). For this neuron (see star in [Figure 3](#pbio-0020044-g003){ref-type="fig"}), degradation resulted in a large increase in information about familiar stimuli from I~100~ = 0.18 bits to I~degrad~ = 0.74 bits. This example neuron thus displayed a nonmonotonic, inverted U-shaped response as a function of degradation. The responses of this neuron for the preferred and nonpreferred familiar stimuli and also for the corresponding novel stimuli are summarized in [Figure 5](#pbio-0020044-g005){ref-type="fig"}C. While the preferred novel undegraded stimulus also activated the neuron, degradation of this stimulus was not associated with significant response enhancement. To examine whether the inverted U-shaped response was in fact characteristic of the V4 neurons that communicated most information about degraded stimuli, we plotted the activity of the neurons which were highly selective for degraded stimuli (see white circles in [Figure 3](#pbio-0020044-g003){ref-type="fig"}B), as a function of coherence, using the preferred stimulus for each neuron ([Figure 6](#pbio-0020044-g006){ref-type="fig"}). We found that across this population, neural activity was indeed significantly enhanced for familiar stimuli at intermediate coherence levels of 55% and 65% relative to activity to undegraded familiar stimuli (paired *t*-tests: *p* \< 0.05). By contrast, activity to novel stimuli systematically decreased with degradation and was significantly below activity to undegraded stimuli at coherence levels of 35% and 45% (paired *t*-tests, *p* \< 0.05). As expected, V4 neurons generally showed greater activity to novel and familiar stimuli than to pure noise at 0% coherence (paired *t*-tests, *p* \< 0.05). As detailed in [Table 1](#pbio-0020044-t001){ref-type="table"}, mean activity was similar for undegraded familiar and novel stimuli, but significantly greater for degraded familiar than degraded novel stimuli (paired *t*-test, *p* \< 0.05). Taken together, learning resulted in an increase in information communicated by V4 neurons about degraded or indeterminate stimuli. This increase in information was mediated by neurons that showed an enhancement in neural activity to degraded compared to undegraded familiar stimuli. ::: {#pbio-0020044-g005 .fig} Figure 5 ::: {.caption} ###### Learning-Dependent Enhancement for Degraded Stimuli---Single Neuron Example (A and B) The activity for an example neuron for its preferred (A) and nonpreferred (B) familiar stimulus is shown in peri-stimulus-time-histogram (PSTH) and raster format. \(C) The average firing rate during stimulus presentation as a function of coherence is summarized for this neuron for its preferred (+) and nonpreferred (−) familiar (fam) and novel (nov) stimuli. :::  ::: ::: {#pbio-0020044-g006 .fig} Figure 6 ::: {.caption} ###### Learning-Dependent Enhancement for Degraded Stimuli---Population Activity These panels show the activity of neurons that communicated most information about degraded stimuli (i.e., white-shaded symbols in [Figure 3](#pbio-0020044-g003){ref-type="fig"}B) as a function of degradation for familiar (A and B) and novel (C and D) stimuli. The preferred stimulus was used for each neuron. The left column shows activity in PSTH format and the right column shows the mean stimulus-evoked activity at each coherence level; asterisks denote significant differences between activity at each coherence level and activity to undegraded stimuli at 100% coherence (paired *t*-tests, *p* \< 0.05). :::  ::: We performed additional behavioral experiments to assess whether learning led to any changes in fixational eye movements, because such changes might shed light on what mediates monkeys\' behavioral advantages for familiar degraded stimuli. In these studies, we allowed the monkeys to freely view sample stimuli during task performance and then estimated a fixation probability map (FPM) for each familiar and novel stimulus presented at 45% and 100% coherence ([see Materials and Methods](#s4){ref-type="sec"}. We applied a threshold to this map to identify regions where monkeys tended to fixate with high probability. The thresholded FPMs for 45% and 100% coherence versions of an example familiar and novel stimulus, along with the overlap between these regions, are shown in [Figure 7](#pbio-0020044-g007){ref-type="fig"}. As can be seen, there was substantially more overlap between the regions of focused eye position at 45% and at 100% after learning. This effect was significant across sessions and stimuli: On average, the overlap region increased by a factor of 2.8 from 0.54 ± 0.14 dva^2^ (degrees of visual angle squared) for novel stimuli to 1.47 ± 0.16 dva^2^ for familiar stimuli (unpaired *t*-test, *p* \< 0.0001). There were also significant learning-dependent increases in the high-probability FPM areas at 45% and 100% coherence (at 45% from 1.04 ± 0.25 dva^2^ to 1.88 ± 0.19 dva^2^, unpaired *t*-test, *p* \< 0.01; at 100% from 0.84 ± 0.21 dva^2^ to 1.74 ± 0.21 dva^2^, unpaired *t*-test, *p* \< 0.01). This learning-dependent increase in the high probability FPM regions and their overlap was highly consistent across sessions and monkeys, and we observed it during all six sessions in both monkeys. Note that the lower FPM values for novel stimuli indicate that eye position was less focused and therefore more distributed before learning, whereas for familiar stimuli robust regions of focused eye position developed. ::: {#pbio-0020044-g007 .fig} Figure 7 ::: {.caption} ###### Eye Movement Analysis during Free Viewing Regions of high fixation probability during free viewing of an example familiar and novel stimulus are shown. Monkeys viewed stimuli at 100% coherence (red-shaded regions) and at 45% coherence (yellow-shaded regions). The green-shaded regions represent regions with high fixation probability at both 45% and 100% coherence. :::  ::: Discussion {#s3} ========== V4 neurons are generally conceptualized as detectors of visual features of intermediate complexity, such as non-Cartesian gratings ([@pbio-0020044-Gallant1]) or contour features ([@pbio-0020044-Pasupathy1]). We have found that learning does not affect how V4 neurons respond to undegraded natural images, both in terms of mean firing rate and information communicated about these stimuli. This absence of learning-dependent differences suggests that this V4 selectivity for features of intermediate complexity is not modified by learning, at least during the several weeks of training in the adult monkey during our task. Basic response properties of V4 neurons thus appear not be altered by learning, similar to findings in V1 that have found that parameters such as receptive field size or orientation tuning width remain unchanged even after extensive training ([@pbio-0020044-Crist1]). Learning does however lead to robust changes in how V4 neurons respond in the presence of degradation. For novel stimuli, V4 neurons tend to act as simple passive feature detectors for which the addition of increasing amounts of noise to the display results in successive reduction in neural activity. Consistent with this finding, we observed a systematic decrease of blood-oxygen level-dependent (BOLD) levels with decreasing stimulus coherence in area V4 of anesthetized monkeys using novel stimuli ([@pbio-0020044-Rainer2]) . After learning, many V4 neurons showed increased activity with the degradation of familiar stimuli, suggesting that they were specifically recruited for difficult discriminations involving the processing of these indeterminate visual inputs. The extraction and amplification of task-relevant elements from visual scenes is a key problem of intermediate-level vision. Our results suggest that V4 neurons play a crucial part in resolving indeterminate visual stimuli and signaling the presence of salient stimulus features. Consistent with this interpretation, studies have found that deactivation or ablation of V4 in monkeys has little impact on basic visual functions, but severely affects shape discrimination ([@pbio-0020044-Girard1]), the identification of images that are occluded or have incomplete contour information ([@pbio-0020044-Schiller1]) or the visual selection in the presence of salient distracters ([@pbio-0020044-De1]). A recent study found severe deficits after V4 ablation in tasks that required making judgments about oriented line segments embedded in distracter arrays ([@pbio-0020044-Merigan1]), a task that has many similarities to the extraction of task-relevant features from degraded displays in our study. We suggest that lesion-induced deficits are a result of disrupting the operation of V4 neurons which are engaged in selective amplification of task-relevant elements of the visual scene. This idea is consistent with our analysis of eye movements, because monkeys focused more reliably on particular stimulus features for familiar than for novel stimuli. This raises the possibility that allocation of focused attention during task performance under central fixation might have contributed to our results, since attention can greatly enhance the response of V4 neurons to visual stimulation ([@pbio-0020044-Moran1]; [@pbio-0020044-Connor1]). Indeed, we suggest that the enhancement in activity and information about degraded familiar stimuli can be conceptualized as a learning-dependent form of attention. Our findings in V4 are in stark contrast to data obtained in the PF cortex using similar task and stimuli ([@pbio-0020044-Rainer1]). In the PF cortex, learning resulted in qualitatively different changes in neural activity. Learning resulted in a robust reduction in average neural activity to undegraded stimuli in PF cortex, whereas we found no general differences in activity in V4. This implies that while PF cortex may play a particularly important role in processing novel stimuli ([@pbio-0020044-Ranganath1]), extrastriate visual areas communicate feature-specific information largely in the absence of learning-related changes for easy-to-discriminate stimuli. Learning led to neural response invariance across degradation in the PF cortex: neurons that responded differentially to two stimuli maintained this response difference for degraded stimuli after learning, whereas the difference in neural response collapsed with degradation for novel stimuli. Response invariance across degradation implies that the PF cortex does not differentiate between degraded and undegraded versions of a stimulus. Learning thus builds response invariance in the PF cortex. In V4, we found that learning led to a selective enhancement of activity for degraded stimuli over and above the response for undegraded stimuli. While PF neurons showed invariant activity, V4 neurons showed inverted U-shaped noise tuning and were thus most active during difficult discriminations, showing responses consistent with selective amplification of feature-specific activity. Our results suggest that the enhancement observed in V4 may be instrumental in establishing invariance in PF cortex and that interaction between these areas may be required to maintain it. Further experiments using simultaneous recordings from both regions are needed to directly test such a hypothesis. Several studies have identified learning-dependent increases in BOLD signals in extrastriate and temporal visual areas ([@pbio-0020044-Dolan1]; [@pbio-0020044-Grill-Spector1]). Because BOLD measures aggregate activation across many neurons, these studies cannot dissociate whether learning-dependent increases are due to building of invariance or selective enhancement of a subpopulation of neurons. This kind of question is certainly important for characterizing functional properties of brain regions and can be answered definitively only by detailed comparison of neural population activity with simultaneously acquired BOLD signal ([@pbio-0020044-Logothetis2], [@pbio-0020044-Logothetis3]). The task dependence of learning effects in V1 ([@pbio-0020044-Gilbert1]; [@pbio-0020044-Gilbert2]) has been taken as evidence that top-down modulation plays an important role in the learning-dependent modifications seen in V1 neurons and that, accordingly, these changes are reflections of plasticity in higher areas of the visual system. Our findings are certainly consistent with this view and suggest that vision is an active process involving recurrent interaction of different brain regions rather than a purely feed-forward process ([@pbio-0020044-Thorpe1]), although our data are consistent with largely feed-forward processing for familiar undegraded stimuli. A possible biophysical mechanism for this interaction was identified by a recent study, which demonstrated that subthreshold activation of the distal apical dendrite of layer V pyramidal neurons can greatly enhance their response to more proximal inputs ([@pbio-0020044-Larkum1]). Because feedback projections from higher cortical areas tend to arrive in upper cortical layers, this represents a mechanism by which feedback could exert control over activity in sensory cortices ([@pbio-0020044-Siegel1]) and thus contribute to the inverted U-shaped responses observed in the present study. Several computational models have investigated how brain regions might interact during stimulus identification. A key feature of such models is the interaction between bottom-up and top-down processing ([@pbio-0020044-Carpenter1]; [@pbio-0020044-Ullman1]). Consider a neuron in an intermediate visual area such as V4, receiving bottom-up feature-tuned visual input from visual areas lower in the hierarchy and top-down feedback from higher areas representing possible interpretations of the stimulus. It has been hypothesized that a match between top-down and bottom-up inputs could result in elevated activity or nonlinear response enhancement. We have observed such enhancement for familiar but not for novel stimuli, indicating that learning plays a critical role in facilitating interaction between top-down and bottom-up processing streams. Another type of model has suggested that top-down feedback may represent a predictive code, where top-down signals effectively cancel predictable responses in the bottom-up signal ([@pbio-0020044-Mumford1]; [@pbio-0020044-Rao1]). In this scheme, activity would be reduced for undegraded stimuli because it can be accurately 'predicted away\' by higher level areas. Degraded stimuli containing noise might not be accurately predicted, leaving more remaining activity compared to undegraded images. However, based on this model, one would predict lower activity for familiar than for novel degraded stimuli, because more of the familiar stimuli can be predicted away---exactly the opposite of what we have found. Thus, our results are more consistent with theories that conceptualize top-down feedback as high-level stimulus interpretations rather than as an error signals. Materials and Methods {#s4} ===================== {#s4a} ### Behavioral and electrophysiological methods. {#s4a1} Two adult male rhesus monkeys (Macaca mulatta) participated in these experiments. All studies were approved by local authorities and were in full compliance with applicable guidelines (EUVD 86/609/EEC) for the care and use of laboratory animals. Stimuli were 10° × 10° in size, 24-bit color depth, and presented at the center of gaze on a γ-corrected 21-inch monitor with linear display characteristics placed at a distance of 97 cm from the monkeys. Stimuli were generated using Fourier techniques that have been described in detail elsewhere ([@pbio-0020044-Rainer2]). In brief, a large set of natural images was first normalized to have identical Fourier amplitude spectra. Degraded versions of natural images were generated mixing the Fourier phase spectra of natural images with a random phase spectrum corresponding to visual noise, independently for each of the RGB color channels. A different random phase spectrum was used during each session, and it was mixed with all images used during that session. Each trial began when the monkey grasped a lever and then acquired fixation on a central fixation point. After 1000 ms, a sample stimulus was presented for 320 ms, which could be any one of eight different images at six coherence levels (0%, 35%, 45%, 55%, 65%, and 100%). After a delay of 1000 ms, a probe stimulus was presented for 600 ms, which could be any one of the eight undegraded images (100% coherence). The monkeys were required to release the lever if the probe matched the sample (i.e., if the sample had been identical to or a degraded version of the probe). In case of a nonmatch, a second brief delay (200 ms) followed the probe, and this delay was always terminated by the presentation of the correct matching stimulus, ensuring that monkeys had to make a behavioral response on every trial. The monkeys were rewarded with apple juice for making correct responses and were rewarded randomly at 0% coherence where the sample contained no task-relevant information. During each session, the monkeys performed the task with a set of four familiar stimuli, with which they had many weeks of practice, as well as with a set of four novel stimuli that they had never seen before. Matches occurred on 50% of trials; the other 50% were non-matches selected randomly from the remaining stimuli. Owing to the normalization procedure, familiar and novel stimuli did not differ in terms of low-level characteristics of spatial frequency content and image intensity. Familiar stimuli from four categories were used (faces, flowers, birds, and landscapes), and one of the four novel stimuli also came from each of these categories. Fixation was monitored with a scleral search coil and sampled at 200Hz (CNC Engineering, Enfield, Connecticut, United States), and the monkeys were required to maintain fixation within a ±1.25° window at all times during the trial. The monkeys completed at least ten trials per condition during each session. Recordings were made from V4 using standard electrophysiological techniques. We employed a grid system (CRIST, Damascus, Maryland, United States) with eight tungsten microelectrodes (FHC Inc., Bowdoinham, Maine, United States). Preoperative magnetic resonance imaging (MRI) was used to identify the stereotaxic coordinates of V4, which was then covered by a recording chamber. To ensure an unbiased estimate of neural activity, we made no attempt to select neurons based on task selectivity. Instead, we advanced each electrode until the activity of one or more neurons was well isolated and then began collecting data. Comparison of the monkeys\' performance during the last six training sessions to performance during recording sessions revealed that performance was unchanged for novel objects (*t*-test, *p* = 0.87), but significantly lower during recording sessions for familiar stimuli (*t*-test, *p* \< 0.01), likely owing to nonspecific factors such as additional wait periods during these sessions. ### Eye movement analysis. {#s4a2} To determine whether there were any learning-related changes in the monkeys\' fixational eye movements, we performed separate behavioral experiments in which we allowed the monkeys to freely view the sample stimulus for a period of about 2 s. As before, we presented four familiar and four novel stimuli during each session, but we only used two coherence levels of 100% and 45% to allow us to assess whether learning led to changes both for degraded and undegraded stimuli. Monkeys performed around 20 trials for each stimulus at each degradation levels during each session, and we report here the results from a total of six sessions. We identified periods of fixation during free-viewing as periods as periods of at least 100 ms duration during which eye position did not change by more than 0.3°. We then marked off a region of 0.3° × 0.3° around this position and superimposed these regions for all fixations during all relevant trials. By normalizing the volume under this function to a value of 1, we created an FPM for each stimulus. We then applied a single threshold to the FPM for all stimuli and degradation levels. The threshold η was chosen to be an order of magnitude greater than the FPM value corresponding to randomly distributed eye position, i.e., to a value of η = 10 × 1/256^2^, and these areas were converted to degrees squared of visual angle (dva^2^). The thresholded FPMs shown in [Figure 7](#pbio-0020044-g007){ref-type="fig"} depict the regions of the FPM that passed threshold for each of the two stimuli during an example session and thus represent the foci of eye position or regions of high fixation probability for that stimulus. Because FPMs are all normalized, a small or absent thresholded FPM region indicates that eye position was distributed on the stimulus without a clear focus. Note that for familiar stimuli, thresholded FPMs were highly consistent across sessions confirming the robustness of this measure. ### Data analysis. {#s4a3} Neural activity was assessed during a fixed period of 310 ms duration, beginning 50 ms after the onset of the visual stimulus to take response latency into account. Such a period roughly corresponds to the time between saccades during natural viewing conditions. Out of a total population of 116 neurons, 83 task-related neurons were identified as showing significant differences in activity between any of the eight stimuli at any coherence level using a Bonferroni-corrected *t*-test evaluated at *p* \< 0.05. Mean firing rates, reported in [Table 1](#pbio-0020044-t001){ref-type="table"}, were computed using the preferred stimulus for each neuron. To assess whether learning had any systematic effect on the amount of stimulus-specific information communicated by V4 neurons, we quantified how much information was contained in the pattern of neural firing rates about novel and familiar stimuli separately. This quantity is given by the mutual information between the set of four familiar or novel stimuli and the set of associated firing rates ([@pbio-0020044-Shannon1]). We thus computed the mutual information (I) among the set of stimuli (*s*) and the neural responses (*r*): where *P*(*s*) is the probability of showing stimulus *s*, *P*(*r*\|*s*) is the probability of observing a response *r* when stimulus *s* is presented, and *P*(*r*) is the probability of observing response *r*. Because calculation of information requires many trials, we computed information for two conditions: degraded and undegraded stimuli. For degraded stimuli, we pooled the coherence levels from 35% to 65%. For undegraded stimuli, we estimated the mutual information for 100% coherence stimuli during the sample period as well as during the probe period on nonmatch trials (to exclude possible movement-related activity). We report here estimates during the probe period because they are based more trials, but results were similar for the sample period. This ensured that information measures for degraded and undegraded stimuli were based in a similar number of trials. For each neuron we estimated four different information values, describing how much stimulus-specific information was contained in its firing rate distributions about undegraded, as well as degraded, familiar (I~fam,100~, I~fam,degrad~) and novel (I~nov,100~, I~nov,degrad~) stimuli. Note that although across all sessions we employed many more novel than familiar stimuli, each individual neuron from which we recorded during a given session 'saw\' exactly the same number of four familiar and four novel stimuli. We identified highly selective neurons in each population by selecting the 25% neurons that communicated most stimulus information about either novel or familiar stimuli (*n* = 21 out of 83 neurons total); i.e., we chose the top 25% of the distribution max(I~fam~,I~nov~). We did this because, owing to our unbiased procedure, our sample contains neurons that did not communicate large amounts of information, and we thus wanted to establish that our conclusions also applied to the neurons that communicated most information. These neurons are shown as white filled circles in [Figure 3](#pbio-0020044-g003){ref-type="fig"}A and [3](#pbio-0020044-g003){ref-type="fig"}C, whereas the remaining 75% of neurons (*n* = 62) are shown as gray filled circles. There was significant overlap (13/21, 62%) between the populations of informative neurons for degraded and undegraded stimuli (χ^2^ test, *p* \< 0.05), indicating that the majority of neurons that were informative for undegraded stimuli were also informative for degraded stimuli. There were no significant differences between informative neurons and the entire population in terms of mean firing rate. Unless otherwise noted, we used paired *t*-tests to compare information measures obtained for novel and familiar stimuli. We have benefited from discussions with M. Bar, L. Chelazzi, K. Cheng, P. Dayan, C. Ranganath, and A. Tolias. We thank K. Nielsen for assistance with the eye movement analysis. This work was supported by the Austrian Academy of Sciences (APART 669) and by the Max Planck Society. GR is a DFG Heisenberg Investigator (RA 1025/1--1). **Conflicts of interest.** The authors have declared that no conflicts of interest exist. DOI: [10.1371/journal.pbio.0020044](10.1371/journal.pbio.0020044) Copyright: © 2004 Rainer et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Academic Editor: Robert Desimone, National Institute of Mental Health BOLD : blood-oxygen level dependent DMS : delayed matching to sample dva : degrees of visual angle FPM : fixation probability map IT : inferior temporal MRI : magnetic resonance imaging PF : prefrontal PSTH : peri stimulus time histogram V1 : primary visual cortex

PubMed Central

2024-06-05T03:55:46.086081

2004-2-17

PMC340948

Introduction {#s1} ============ Practice can induce behavioral improvement that is often specific to the situation experienced during the practice sessions (or "training"). Such findings suggest that changes occur in neurons with fine selectivity (or "tuning") for the stimuli experienced or the movements made during training. In the visual system, for example, behavioral improvement is specific to the trained stimulus, such as the orientation of a light bar ([@pbio-0020045-Fiorentini1]; [@pbio-0020045-Crist1]), and is paralleled by specific changes in neurons that are tuned to the orientation of a light bar ([@pbio-0020045-Schoups1]) or, in other experiments, the direction of visual motion ([@pbio-0020045-Zohary1]). In the auditory system, changes in response properties of single neurons and cochleotopic maps are specific to the parameters characterizing the sound ([@pbio-0020045-Suga1]). In the motor system, skill acquisition induces expansion in the cortical representation of the used forelimb ([@pbio-0020045-Nudo1]) and enhance synaptic connections in the trained contralateral hemisphere ([@pbio-0020045-Rioult-Pedotti1]). A line of studies found that when monkeys perform reaching movements and adapt to directional errors induced by force fields, primary motor cortex (M1) cells shift their preferred direction (PD) in about the same way as for the muscle activity needed to perform the task ([@pbio-0020045-Gandolfo2]; [@pbio-0020045-Li1]; [@pbio-0020045-Padoa-Schioppa1]). We have recently shown that learning a local rotational visuomotor task can induce an elevation in the activity of single neurons in M1 ([@pbio-0020045-Paz1]) and that these changes are observed only in a specific subpopulation of neurons, those with a PD close to the movement direction used during the learning. Whereas many studies indicate that learning can induce specific changes in brain activity, this finding does not necessarily imply that newly learned skills are "better" represented in the brain. The crucial question is this: Do neurons encode task parameters, such as movement direction, any better after learning? In the motor system, such improved encoding ([@pbio-0020045-Chen2]) can be used for decoding by downstream areas and as an efference copy for further computation ([@pbio-0020045-Wolpert1]; [@pbio-0020045-Sommer1]). It can also be used by an external observer to allow for more accurate prediction of behavior ([@pbio-0020045-Laubach1]). In this paper, we examine two questions. First, do learning-induced changes in firing rates provide more information on the task? And, second, what aspect of the cells\' activity contributes mostly to this improvement? To address the first question, we employed an information-theory analysis ([@pbio-0020045-Cover1]; [@pbio-0020045-Rieke1]) to calculate the mutual information (MuI) (see [Figure 2](#pbio-0020045-g002){ref-type="fig"}) between cells\' activity and direction of movement. Informational measures have two relevant advantages. First, they use the full distribution (estimated from the data) of neuronal activity and do not assume any specific shape of the tuning curve or noise distribution. This allows for a more fine-tuned examination of learning-related changes. Second, they provide a measure as to how well different directions can be differentiated, based on neuronal activity. To address the second question, we examined two features of the neuronal response that could contribute to the increase in information: response variability and the slope of the tuning curve. Finally, to demonstrate that the observed increase in information can be extracted, we use the neuronal activity to decode the actual movement direction. ::: {#pbio-0020045-g002 .fig} Figure 2 ::: {.caption} ###### MuI between Neuronal Activity and Direction of Movement The example shows a simulation of the activity of one cell during 64 movements to evenly spaced eight directions, presented in a random order (eight trials per direction). Each dot in the raster plots a and b describes the spike count of the cell in a specific trial. Without prior knowledge about the direction of movement (A), a large uncertainty exists about the responses of the neuron. However, ordering the trials according to the movement direction (B) reveals a large reduction in the uncertainty about the cell responses. The probability p(*r,d*) of observing a trial with direction *d* and spike count *r* is shown in (C); along with a specific conditional distribution *p*(*r*\|*d* ~0\ =\ 90)~ . The entropy is a measure of the uncertainty about movement direction: H(D) = log(8) = 3 bits, in the case that all eight directions have equal probability to occur. The conditional entropy is defined as and describes the mean uncertainty about direction given the cell response. The MuI *I*(*R*;*D*) = *H*(*D*) − *H*(*D*\|*R*) measures the reduction in uncertainty about movement direction given the response of the cell. The MuI is symmetric, in the sense that it also measures the reduction in uncertainty about cell response given the direction of movement *I*(*R*;*D*) = *H*(*D*) − *H*(*D*\|*R*) . This relation is graphically depicted in (D). :::  ::: Results {#s2} ======= Monkeys adapted to visuomotor rotations on a daily basis by altering the relationship between the visual feedback (cursor) and the hand movement ([Figure 1](#pbio-0020045-g001){ref-type="fig"}). Learning was confined to only one target in space, i.e., learning that is local in direction. We tested neuronal sensitivity to direction by comparing the information content conveyed in the firing rate of single cells during the pre- and post-learning epochs (identical task of standard movements to eight directions spanning the two-dimensional working surface, only differentiated by a learning epoch). We specifically looked for a change in representation that was related selectively to the learned direction, i.e., the hand direction that was used to bring the cursor to the target during the transformation. ::: {#pbio-0020045-g001 .fig} Figure 1 ::: {.caption} ###### Behavioral Paradigm and Movement Kinematics \(A) Session flow (left to right). Every session (day) consisted of pre-learning, learning, post-learning, and relearning epochs. Pre- and post-learning epochs were standard eight-target tasks with a default (one-to-one) mapping between cursor movement and the movement of the hand. In the learning epoch, only one target (upwards) appeared, and a visuomotor rotational transformation was imposed on the relationship between movement of the hand and cursor movement. The example shown is for a transform of −90° (see[Materials and Methods](#s4){ref-type="sec"} for a full description). (B--D) Similar kinematics pre- and post-learning. (B) Example of 1-day trajectories from the two epochs; the transform in this session was of --45°. (C) Velocity profiles. Peak velocity was slightly lower in the post-learning epoch (*t*-test, *p* = 0.05), but no difference was found between the learned direction and other directions (*t*-test, *p* = 0.3). (D) Improvement in directional deviation was calculated as the deviation of the instantaneous hand direction from the required target direction, calculated every 10 ms starting from the go-signal. All four movement types (learned and nonlearned, pre- and post-learning) exhibited the same temporal pattern. Here and for analysis of neuronal activity, we excluded the first trials in the post-learning epoch---those exhibiting significant aftereffects due to learning. :::  ::: Activity was measured from the hold period that immediately follows the target appearance, but before the go-signal, and was therefore termed preparatory activity (PA). There were three reasons for this choice. First, such PA has been reported in many motor cortices and is thought to participate in movement planning and in computing visuomotor transformations ([@pbio-0020045-Kurata1]; [@pbio-0020045-Alexander1]; [@pbio-0020045-Kalaska1]; [@pbio-0020045-Shen1]; [@pbio-0020045-Zhang1]; [@pbio-0020045-Kakei1]). Second, as previously found in this experimental paradigm, learning-related changes have only been reported for this period ([@pbio-0020045-Paz1]). Third, as a means of eliminating any kinematic-related changes ([@pbio-0020045-Wise1]), we further verified that movements shared similar kinematics before and after learning (see [Materials and Methods](#s4){ref-type="sec"}; [Figure 1](#pbio-0020045-g001){ref-type="fig"}). Mutual Information {#s2a} ------------------ The MuI between one-cell activity and direction of movement is exemplified in [Figure 2](#pbio-0020045-g002){ref-type="fig"}. We compared the MuI between pre- and post-learning ([Figure 3](#pbio-0020045-g003){ref-type="fig"}A). The figure depicts the distributions of MuI between direction and spike count for all cells ([Figure 3](#pbio-0020045-g003){ref-type="fig"}A, corrected for bias) for pre-learning (dashed line) and for post-learning (solid line). No difference was found between the MuI on the population level, either by comparing the distributions (Kolmogorov--Smirnoff, *p* = 0.3) or by comparing their means (paired *t*-test, *p* = 0.53). We further tested the average information about direction conveyed by each spike by normalizing each cell\'s information by its firing rate and again found no significant difference (inset in [Figure 3](#pbio-0020045-g003){ref-type="fig"}A; Kolmogorov--Smirnoff, *p* = 0.25, paired *t*-test, *p* = 0.7). ::: {#pbio-0020045-g003 .fig} Figure 3 ::: {.caption} ###### Comparing MuI of Single Cells Pre- and Post-Learning \(A) Distributions of single-cell information about direction of movement pre-learning (dashed) and post-learning (solid). No significant difference was found between the distributions (Kolmogorov--Smirnoff, *p* = 0.3). The inset shows the MuI per spike, calculated by dividing the information per cell by the cell\'s firing rate (Kolmogorov--Smirnoff, *p* = 0.25). \(B) Improvement in information of individual cells. Histogram of *p*-values for all cells; a significant (*p* \< 0.01, χ^2^) number of cells (*n* = 37) had a *p*-value greater than 0.95, representing cells that significantly increase their information content about direction after learning; 18 cells had a *p*-value lower than 0.05, representing cells that decreased their information content, but this was found to be only marginally significant (*p* = 0.06, χ^2^). \(C) Histograms of difference in information, post- minus pre-learning, for all cells (upper) and only for cells that increase (*p* \> 0.95) or decrease (*p* \< 0.05) their information content significantly (lower). \(D) Circular histogram for PD of cells that significantly increased their information. The cells\' PDs were normalized to the learned direction in each cell recording session, revealing a unimodal distribution (Rayleigh test, *p* \< 0.05). The upper inset shows the circular histogram for all cells and lower inset shows the circular histogram for cells that decreased their information; in both cases, the distributions seem homogenous (Rayleigh test, *p* \> 0.1). :::  ::: Although the population as a whole did not change significantly, single neurons could still increase or decrease their information about direction. To explore this, we probed each neuron individually for changes in MuI. Using a bootstrap method, we shuffled trials from pre- and post-learning and randomly reselected two different groups of trials, we then calculated the MuI for each group and the difference between the two MuIs. The procedure was repeated 1,000 times to produce a distribution of MuI differences. The actual MuI difference (between the pre- and post-learning) was compared to this distribution to obtain a *p*-value. A high *p*-value means that the MuI in the post-learning epoch was significantly higher than the MuI in the pre-learning epoch. [Figure 3](#pbio-0020045-g003){ref-type="fig"}B plots the histogram of the *p*-values for all cells. A significant number of cells showed an increase in MuI with a *p*-value larger than 0.95 (black in [Figure 3](#pbio-0020045-g003){ref-type="fig"}B; *n* = 37 out of 177, *p* \< 0.01, χ^2^), a nearly significant number of cells showed a decrease in MuI with a *p*-value lower than 0.05 (white/transparent in [Figure 3](#pbio-0020045-g003){ref-type="fig"}B; *n* = 18, *p* = 0.06), while all the rest did not (gray in [Figure 3](#pbio-0020045-g003){ref-type="fig"}B). We also examined the actual change in information content for all cells ([Figure 3](#pbio-0020045-g003){ref-type="fig"}C, upper) and specifically for the cells that had a significant change ([Figure 3](#pbio-0020045-g003){ref-type="fig"}C, lower). Following the rationale explained in the Introduction, the association between the learned parameter (direction) in local rotational transformations and the sensitivity of many cells to direction, we probed for a relation between cells\' PD and the learned direction. [Figure 3](#pbio-0020045-g003){ref-type="fig"}D plots a circular histogram of PDs of cells that exhibited a significant increase in their MuI. For the plot, we normalized each cell\'s PD to the learned direction in its recording session, and this revealed a unimodal distribution (Rayleigh test, *p* \< 0.05) with its center on the learned direction. In contrast, the PD distributions of the whole population ([Figure 3](#pbio-0020045-g003){ref-type="fig"}D, upper inset) and of cells that significantly decrease their information content ([Figure 3](#pbio-0020045-g003){ref-type="fig"}D, lower inset) did not exhibit this trend and seemed homogenous. To test that this change in information is indeed owing to the learning of visuomotor transformations and not owing to the mere repetition of a single movement during the learning epoch, we conducted the same analysis for control, repetition sessions. Only a nonsignificant (*p* \> 0.1, χ^2^) number of cells (eight out of 126) had a *p*-value greater than 0.95 ([Figure 4](#pbio-0020045-g004){ref-type="fig"}A). Further, this population did not exhibit any specific distribution of PDs ([Figure 4](#pbio-0020045-g004){ref-type="fig"}B; Rayleigh test, *p* \> 0.1). ::: {#pbio-0020045-g004 .fig} Figure 4 ::: {.caption} ###### Changes Were Not Observed after Mere Repetition of Movement to One Direction Same as in Figure 3B and 3D, but for control sessions that included the mere repetition of standard, nontransformed movement to one target during the learning epoch. The number of cells that exhibited an increase in their information content was not significant (\[A\] right bar, eight out of 126), and their distribution was homogenous and showed no specific relation to the direction of the repeated movement (B). :::  ::: Individual Information per Direction {#s2b} ------------------------------------ The MuI represents the information that a cell\'s spike count conveys about all the eight tested directions. We further investigated how much information a cell conveys about one direction in particular, which is termed the individual information per direction (DI) ([@pbio-0020045-Rolls1]; [@pbio-0020045-Buracas1]) and is measured as the reduction in uncertainty about the spike counts, given a specific direction. We calculated the DI of each cell for each of the eight possible directions, pre- and post-learning. The distribution of the differences between the post-learning DI and pre-learning DI for the learned direction was significantly above zero ([Figure 5](#pbio-0020045-g005){ref-type="fig"}A, "'Learned"\'). This indicates that after learning, cells\' firing rates conveyed more information about the learned direction. [Figure 5](#pbio-0020045-g005){ref-type="fig"}A also shows that information about other nonlearned directions did not change. As with the MuI, to probe for the directional tuning of the cells, we plotted the circular histogram of PDs of cells that increased their information about the learned direction (a positive post-learning minus pre-learning). Again, a unimodal distribution (Rayleigh test, *p* = 0.01) was found with its peak on the learned direction ([Figure 5](#pbio-0020045-g005){ref-type="fig"}B). ::: {#pbio-0020045-g005 .fig} Figure 5 ::: {.caption} ###### Comparing Individual DI \(A) Mean (with 95% confidence intervals, by fitting a Gaussian distribution) of post-learning information minus pre-learning information for one direction. Abscissa represents the distance from the learned-movement direction; all directions were normalized according to the learned direction in the cell\'s session. An increase is evident only for the learned-movement direction, with mean at 0.1 and 95% confidence intervals at 0.036 and 0.164. \(B) Circular histogram of PDs for cells with a positive difference of post-learning minus pre-learning information about the learned direction (Rayleigh test, *p* = 0.01). :::  ::: Possible Origins for Improvement in Information {#s2c} ----------------------------------------------- Information theory makes use of the complete (estimated from data) stimulus--response distribution and thus does not tell us what feature in cell activity primarily contributed to the increase in information content. However, we found that the increase in information is specific to a single-learned direction and that cells with a PD close to the learned direction mainly contributed to this increase. We have previously reported that cells with PD close to the learned direction increased their firing rate after learning when movement was to the learned direction ([@pbio-0020045-Paz1]). We therefore explored more closely this elevation in firing rates and its relationship to the increase in information content. [Figure 6](#pbio-0020045-g006){ref-type="fig"}A histograms the net changes in activity (post- minus pre-learning) at the cells\' PDs for the whole population. [Figure 6](#pbio-0020045-g006){ref-type="fig"}B shows the same net changes for cells that significantly increased their information about direction, where a significant positive trend was found (by fitting a normal distribution; see legend to [Figure 6](#pbio-0020045-g006){ref-type="fig"}B). We further aligned each cell tuning curve on the cell\'s PD and calculated the average tuning curve. This revealed that this group of cells indeed elevated their activity mainly around their PD ([Figure 6](#pbio-0020045-g006){ref-type="fig"}C). ::: {#pbio-0020045-g006 .fig} Figure 6 ::: {.caption} ###### Learning-Induced Elevation of Information and Activity (A and B) Histograms of changes in firing rate in the PD (post-learning minus pre-learning) for all the cells (A) and for cells that significantly increased their information (B). The horizontal line below the histogram represents its mean and the 95% confidence intervals, by fitting a Gaussian distribution. \(C) Average tuning curves (baseline subtracted, ± SEM) of cells that significantly increased their information (*n* = 37). Comparing pre-learning (gray) and post-learning (black). Cell tuning curves were first aligned to each cell\'s PD. :::  ::: Two natural features of a cell\'s tuning curve can contribute to the improvement in information content. First, a cell can increase the slope of the tuning curve just near the learned direction, and thus small changes in direction can lead to a larger difference in the cell\'s response, providing a better differentiation of direction based on the neuronal activity (illustrated in [Figure 7](#pbio-0020045-g007){ref-type="fig"}A). Second, cells can reduce the variability of their response near the learned direction. This is also termed "reliability," because when variability is low, each single report made by the cell is more reliable (illustrated in [Figure 7](#pbio-0020045-g007){ref-type="fig"}B). A standard method for characterizing this is the Fano factor ([@pbio-0020045-Berry1]), calculated as the variance of the response divided by its mean. We correlated the net change in information content (post-learning minus pre-learning) to these two factors: change in slope near the learned direction ([Figure 7C1](#pbio-0020045-g001){ref-type="fig"}--[7C3](#pbio-0020045-g007){ref-type="fig"}) and change in the Fano factor ([Figure 7D1](#pbio-0020045-g001){ref-type="fig"}--[7D3](#pbio-0020045-g007){ref-type="fig"}). [Figure 7](#pbio-0020045-g007){ref-type="fig"} shows that whereas no systematic change in the corresponding factor was found for the whole population ([Figure 7](#pbio-0020045-g007){ref-type="fig"}C1 for slope and [Figure 7](#pbio-0020045-g007){ref-type="fig"}D1 for FF), a significant positive trend was observed for the population of neurons that significantly increased their information after learning. This trend was obvious for the slope factor ([Figure 7](#pbio-0020045-g007){ref-type="fig"}C2) and also, but to a much lesser extent, for the Fano factor ([Figure 7](#pbio-0020045-g007){ref-type="fig"}D2). [Figure 7C3](#pbio-0020045-g003){ref-type="fig"} and [7D3](#pbio-0020045-g007){ref-type="fig"} extends this relation and shows the correlation between the corresponding factor and the increase in information. A significant positive correlation was only found for the slope factor and only for cells that significantly increased their information ([Figure 7](#pbio-0020045-g007){ref-type="fig"}C3, black asterisks and line). No correlation was observed between the change in slope and the change in information for the rest of the cells ([Figure 7](#pbio-0020045-g007){ref-type="fig"}C3, gray dots) or between the change in Fano factor and the change in information, either for the whole population ([Figure 7](#pbio-0020045-g007){ref-type="fig"}D3, gray dots) or for those that significantly increased their information ([Figure 7](#pbio-0020045-g007){ref-type="fig"}D3, black asterisks). Further, the increase in the slope of the tuning curve near the learned direction was specific to this direction only and to cells that significantly increased their information content ([Figure 8](#pbio-0020045-g008){ref-type="fig"}). ::: {#pbio-0020045-g007 .fig} Figure 7 ::: {.caption} ###### Increased Slope of Tuning Curve Is Correlated with the Increase in Information Possible mechanisms for improving the information content of single cells. (A and C) The slope of the tuning curve at the learned direction indicates the magnitude of change in activity in response to small changes in direction. The higher slope suggests that nearby directions can be discriminated better. (B and D) Reliability of coding. The variability at each direction indicates how well different directions can be differentiated based on single trials. (C1 and D1) Histograms of the difference between pre- and post-learning for the corresponding mechanism for the whole population of cells. The horizontal line below the histogram represents its mean and the 95% confidence intervals. (C2 and D2) Histograms for cells that significantly increased their information about direction. (C3 and D3) Correlation between the difference in information (post- minus pre-learning) and the corresponding mechanism. Gray dots are all the cells, and black asterisks are cells that significantly increased their information content. The black line represents the linear regression fit. The corresponding Pearson correlation (C) and its significance (*p*-value) are designated. The histogram in (C2) is shifted to the right, indicating that cells that increased their information content also increased the slope of the tuning curve in the learned direction. In these cells only, a significant (*p* = 0.002) correlation coefficient (*c* = 0.492) was found. :::  ::: ::: {#pbio-0020045-g008 .fig} Figure 8 ::: {.caption} ###### Slope Increase Is Specific to the Learned Direction Mean change (± SEM) in the slope of the tuning curve surrounding each direction, for cells that significantly increased their information content (black) and for the rest of the cells (gray). :::  ::: These results suggest that cells increased the slope of their tuning curve near the learned direction and improve the information content in their activity. Cells can use several strategies to do so and we considered three possibilities: first, by shifting their tuning curve and positioning the learned direction at a better "slope-wise" location on the tuning curve (illustrated in [Figure 9](#pbio-0020045-g009){ref-type="fig"}A); second, by narrowing the tuning curve ([Figure 9](#pbio-0020045-g009){ref-type="fig"}B); and, third, by local changes increasing or decreasing specific points near the desired (learned) location ([Figure 9](#pbio-0020045-g009){ref-type="fig"}C). Although the three possibilities are not mutually exclusive and might be interrelated, we attempted to distinguish among them by correlating the change in information to each one. [Figure 9](#pbio-0020045-g009){ref-type="fig"}D--[9](#pbio-0020045-g009){ref-type="fig"}F shows that the increase in information was correlated with the increased firing rate at the learned direction ([Figure 9](#pbio-0020045-g001){ref-type="fig"}F1--[9](#pbio-0020045-g009){ref-type="fig"}F3), but not with shifts in PD ([Figure 9](#pbio-0020045-g001){ref-type="fig"}D1--[9](#pbio-0020045-g009){ref-type="fig"}D3) or with the narrowing of tuning curves ([Figure 9](#pbio-0020045-g001){ref-type="fig"}E1--[9](#pbio-0020045-g009){ref-type="fig"}E3). We therefore suggest that cells locally increased their firing rate to increase the slope of their tuning curve at the learned direction. ::: {#pbio-0020045-g009 .fig} Figure 9 ::: {.caption} ###### Increased Information after Learning Is Correlated with Elevation of Firing Rate in the Learned Direction Possible mechanisms for increased slope of the tuning curve in the learned direction. (A and D) Shift of PD, i.e., shifting the whole tuning curve, may position the learned direction at a higher slope location. (B and E) Narrowing of the tuning curve, as measured by the width at half-height. (C and F) Local changes (Increase) in activity in the learned direction can increase the slope. This is similar to the observed learning-induced changes in our data (see Figure 6C). In (A)--(C), an illustration of the measured difference is indicated. (D1--D3, E1--E3, and F1--F3) Same format as in Figure 4 for the three possible mechanisms. The histogram in (F2) is shifted to the right, indicating that cells that increased their information content also elevated their firing rate in the learned direction. In these cells only a significant (*p* \< 0.001) correlation coefficient (*c* = 0.566) was found (F3, asterisks and line). :::  ::: Decoding Movement Direction {#s2d} --------------------------- We hypothesized that the improved information regarding movements in the learned-movement direction would lead to an improved ability to reconstruct movements from population activity. To test this assumption, we applied two reconstruction methods: the population vector (PV) approach, a widely used decoding scheme for M1 activity ([@pbio-0020045-Georgopoulos1]; [@pbio-0020045-Moran1]), and a maximum a posteriori (MAP) estimator ([@pbio-0020045-Sanger2]). For the PV analysis, we selected 129 of the 177 cells, only including cells that exhibited directional tuning by a cosine fit. Neurons were pooled according to the learned-movement direction in their recording session, and we computed the PV from the pre-learning and post-learning activity. [Figure 10](#pbio-0020045-g010){ref-type="fig"}A shows the deviation of the PV direction, i.e., the difference between the PV prediction and the actual movement direction for the four possible learned-movement directions. A marked and statistically significant improvement was observed in the predicted direction (*p* \< 0.05 for all four learned directions, bootstrap and *t*-test). We verified that this improvement was due to learning in two ways: first, by shuffling trials from the pre-learning and the post-learning epochs, and second, by shuffling cells from days with different transformations. In both cases, the mean of the distribution of improvements was not significantly different from zero. Furthermore, the improvement in the PV prediction was specific to the learned-movement direction. [Figure 10](#pbio-0020045-g010){ref-type="fig"}B shows the signal-to-noise ratio (mean/SD) of improvements in PV accuracy (the difference between the accuracy of the pre-learning prediction and the post-learning prediction). We normalized each session directions to the learned direction in the session. A statistically significant improvement was found only for the learned-movement direction (χ^2^, *p* \< 0.01). This improvement in the PV prediction can be accounted for by the enhanced firing of cells with a PD near the learned-movement direction, as shown above (see [Figure 6](#pbio-0020045-g006){ref-type="fig"}). Cells with their PD close to the learned-movement direction made a larger contribution to the PV, but mostly when the movement was in that direction. Because each cell contributes a weighted vector in the direction of its own PD, only the learned-movement directions showed improvement in PV accuracy. This improvement in prediction due to altered directional tuning is reminiscence of studies that examined learning of visuomotor associations in frontal eye fields ([@pbio-0020045-Chen1], [@pbio-0020045-Chen2]) and of studies showing evolvement of directional tuning in M1 when monkeys received real-time visual feedback of brain-controlled trajectories ([@pbio-0020045-Taylor1]). ::: {#pbio-0020045-g010 .fig} Figure 10 ::: {.caption} ###### Improved Decoding of Movement Direction Only for the Learned Direction (A and B) Using PV. (A) PV errors given as the distance in degrees between the predicted and the actual direction for the four learned-movement directions (± SEM, bootstrap test). (B) Signal-to-noise ratio (mean/SD) of PV improvement (pre-learning deviation minus post-learning deviation) for all directions (four learned directions are pooled together and all other directions are normalized to them). A significant improvement was observed only for the learned direction (*p* \< 0.005, Bonferoni correction for multiple tests, i.e., the eight directions). (C and D) Using a MAP estimator, we predicted 100 times the actual hand direction using neuronal activity. Shown is the fraction of correct predictions for pre-learning (C) and post-learning (D). A significant increase was observed only for the learned direction (*p* \< 0.005, Bonferoni correction for multiple tests, i.e., the eight directions). The dashed line is the chance level (0.125). :::  ::: The PV method includes several assumptions about the coding and the decoding of the M1 population activity and is not guaranteed to be optimal ([@pbio-0020045-Sanger1]; [@pbio-0020045-Snippe1]; [@pbio-0020045-Pouget1]; [@pbio-0020045-Scott2]). Therefore, we also tested the performance of a probabilistic approach. Using a MAP estimator, we predicted the movement direction for all possible directions, including the learned-movement direction pooled and normalized from all sessions. [Figure 10](#pbio-0020045-g010){ref-type="fig"}C depicts the success rate for 100 repetitions (by cross-validation) for each direction. [Figure 10](#pbio-0020045-g010){ref-type="fig"}D shows the same, but in the post-learning epoch. A higher success rate of correctly predicting the movement direction can be observed for learned direction only in the post-learning epoch (χ^2^, *p* \< 0.01, chance level is at 0.125;dashed line in [Figure 10](#pbio-0020045-g010){ref-type="fig"}D). This indicates that after learning and by using this decoding method, we could better predict the actual movement direction from neuronal activity. Discussion {#s3} ========== This report describes improved encoding and decoding of specific directions by neurons in M1 of monkeys after learning a visuomotor skill that requires learning only for one direction in space. Our results suggest a close link between properties of neurons, such as directional tuning of cells, and learning a skill that is local in the same parameter, in this case direction, a finding that is concordant with ideas and findings in sensory systems ([@pbio-0020045-Zohary1]; [@pbio-0020045-Suga1]; [@pbio-0020045-Sharma1]). The fact that the increased information we found was associated with an increased slope of the tuning curve, as also reported in a recent visual study ([@pbio-0020045-Schoups1]), further suggests that similar mechanisms may govern neuronal interactions and learning throughout the central nervous system. The fact that improved information in neuronal activity was evident mainly for the learned direction is in accordance with studies showing confined generalization of learning a sensorimotor skill, one that requires adaptation to directional errors. The width of the behavioral generalization function (i.e., the angular distance from the learned direction where aftereffects could still be observed) was similar for our monkeys ([@pbio-0020045-Paz1]) and in human studies, ranging from 45° ([@pbio-0020045-Gandolfo1]; [@pbio-0020045-Krakauer1]) to 90° ([@pbio-0020045-Imamizu1]; [@pbio-0020045-Thoroughman1]). The neuronal changes we previously observed occurred mainly for cells with PD within 30° of the learned direction, and the change in slope observed in this study was sharply focused and not seen for directions 45° away from the learned direction (see [Figure 8](#pbio-0020045-g008){ref-type="fig"}; note, however, that changes in firing rate were wider \[see [Figure 6](#pbio-0020045-g006){ref-type="fig"}\]). While narrower primitives reasonably lead to narrower generalization function ([@pbio-0020045-Donchin1]), the exact generalization width depends not only on the primitives\' width, but also on the connectivity and the specific model used. These are still largely unknown. An intriguing result in this study is that learning-related changes were observed and persistent in the post-learning epoch, when performing a standard task that required no transformation. Further, measured kinematics was the same as in the pre-learning epoch. If the improved information can be used, why isn\'t it? First, our monkeys were trained on a task that did not require improved performance in the standard task after learning, but did encourage them to reserve learning for future use of the same visuomotor task. This is in agreement with our previous report, showing that the monkeys retained the task until the performance of the relearning epoch (i.e., they exhibited immediate recall rather than learning in this second learning epoch), and suggests that the neuronal change should persist but somehow gated. Indeed, everyday behavior shows that we can learn new tasks without interfering with the performance of existing ones. An alternative possibility is that we did not measure the appropriate kinematic variable that was altered and improved due to the neuronal changes. For example, a task that would demand finer directional sensitivity (i.e., angular distance of less than 45°) might show a change in performance after learning. It is also worth noting that our experiment was performed in a local region in space and was not constrained to a specific posture ([@pbio-0020045-Scott1]) or joint or muscle combination ([@pbio-0020045-Scott2]). Therefore, we cannot conclude that locality and specificity of change in information content are related to external direction of movement. Our results may be consistent with other reference frames as well ([@pbio-0020045-Mussa-Ivaldi1]; [@pbio-0020045-Todorov1]). One important question is what kind of learning can induce such an increase in information content. Although psychophysics studies have shown that adapting to new kinematics and/or dynamics environments results in the formation of internal representations in the brain (reviewed by [@pbio-0020045-Kawato1]; [@pbio-0020045-Wolpert1]), changes were also observed after extensive training and mere repetition ([@pbio-0020045-Nudo1]). Moreover, many sensory systems exhibit stimulus-related adaptations ([@pbio-0020045-Dragoi1]; [@pbio-0020045-Suga1]), where repeated presentations of a stimulus induce a change in activity of neurons. To control for this possibility, we conducted sessions with a repetition condition, which entailed a one-target task without angular transformations. Cells recorded in these sessions did not exhibit a change in their information content, and PV analysis produced similar results before and after repetition. An alternative explanation could be attention-related modulations ([@pbio-0020045-Spitzer1]). We discuss elsewhere why this is an unlikely source for the changes we observed ([@pbio-0020045-Paz1]), yet we cannot rule out the possibility that increased attention might lead to similar improvement in information. MuI measures are more often used in sensory research, describing the information that neurons convey about a presented stimulus, and only few papers have applied such measures to the motor system (e.g., [@pbio-0020045-Hatsopoulos1]). We believe this stems from the fact that in sensory systems, neurons respond to the stimulus, whereas in the motor system, neurons "cause" the movement. In this study, we treated direction of movement as a stimulus to which the neuron responds. This can be justified because MuI is a symmetric measure and the point of view can be reversed; e.g., we can interpret the results as neuronal activity → movement, rather than movement → neuronal activity. More importantly, frontal motor fields, M1 included, are only part of the brain\'s learning system and project to many brain areas that take part in processing an upcoming movement, such as the basal ganglia and cerebellum ([@pbio-0020045-Middleton1]). Therefore, M1 activity may be decoded by those areas involved in coplanning of the movement. Moreover, an efference copy of the planned motor command is probably used for error estimation and correction ([@pbio-0020045-Wolpert1]; [@pbio-0020045-Sommer1]). Indeed, we are aware of our movements before they have actually started ([@pbio-0020045-Haggard1]). This suggests that when learning new sensorimotor tasks, activity in M1 should not only produce the correct behavior, but also change in a way that enables other brain areas a better readout of the motor command. This will allow more efficient computation and better control of the forthcoming movement. Although higher information content implies better encoding by neurons, it does not entail better decoding; this is highly dependent on the algorithm used and on the error function introduced. Since our task involved manipulation of movement direction and since real-time prediction of movement trajectory has taken on major interest in recent years ([@pbio-0020045-Wessberg1]; [@pbio-0020045-Taylor1]), we used the discrepancy between the actual movement direction and the predicted one from neuronal activity as the error signal (either categorical, for the MAP, or continuous, for the PV). The MAP method ([@pbio-0020045-Sanger2]; [@pbio-0020045-Zhang2]) is theoretically optimal ([@pbio-0020045-Seung1]) and requires fewer assumptions on the tuning-curve shapes and distribution of PDs ([@pbio-0020045-Sanger1]), but requires larger amounts of data to estimate the true distributions ([@pbio-0020045-Pouget1]). The PV method has been shown to be robust in many scenarios ([@pbio-0020045-Georgopoulos1]; [@pbio-0020045-Moran1]) and very useful in predicting hand movement in real time ([@pbio-0020045-Taylor1]). In our experiment, both methods yielded a better prediction of the learned-movement direction during its planning stages and long before its initiation (see also [@pbio-0020045-Laubach1]). Although we cannot determine whether neurons further downstream use this improved information to decode a previous layer of neurons, we believe it is possible. Further, our findings could lead to improved strategies for recovering trajectory information from populations of M1 cells ([@pbio-0020045-Wessberg1]; [@pbio-0020045-Serruya1]; [@pbio-0020045-Taylor1]). The specificity of the learning is of high importance here. The large number of degrees of freedom, the complexity of movements, and the technical difficulties of recording many neurons simultaneously are only starting to be addressed, and a plausible strategy might require learning and practicing specific and essential movements. Our results suggest that this would modify brain activity in a way that would enable earlier and better readout of brain activity from fewer neurons. Materials and Methods {#s4} ===================== {#s4a} The experimental setup and data acquisition procedures are described in detail in [@pbio-0020045-Paz1]). The sampled cells were taken from the same database. ### Physiological procedures. {#s4a1} Two female rhesus (Macaca mulatta) monkeys (approximately 4.5 kg) were implanted with recording chambers (27 × 27 mm) above both the right and left hemispheres. Animal care and surgical procedures complied with the National Institutes of Health *Guide for the Care and Use of Laboratory Animals* (rev. 1996) and with the Hebrew University guidelines supervised by the Institutional Committee for Animal Care and Use. The monkeys were seated in a dark chamber, and eight microelectrodes were introduced into each hemisphere. The electrode signals were amplified, filtered, and sorted (MCP-PLUS, Alpha-Omega, Nazareth, Israel), and all spike shapes were sampled at 24 KHz. We used a template-based method for real-time isolation of spike shapes (MSD, Alpha-Omega). Penetration locations were verified by MRI (Biospec Bruker 4.7 Tesla, Bruker BioScences, Billerica, Massachusetts, United States) before recordings. At the end of each session, we examined the activity of neurons evoked by passive manipulation of the limbs and applied intracortical microstimulation (50 ms of 200-μs cathodal pulses at 300 Hz) to evoke movements. Only penetration sites that evoked single-joint shoulder or elbow movement at thresholds of lesser than or equal to 40 μA were used in this study. In one monkey, we also made anatomical observations, to verify the accurate penetration sites relative to the central sulcus. ### Behavioral paradigm. {#s4a2} Monkeys moved a manipulandum to control the movement of a cursor on a video screen located 50 cm from their torso and eyes with the goal of moving the cursor from a starting point at the center of the screen (origin) to a visual target in a delayed go-signal paradigm; this required the monkey to hold (as verified by hand velocity and EMG) the cursor in the origin circle for a random 750--1,500 ms after the target onset. The disappearance of the origin indicated the go-signal. In each session (day), four consecutive epochs were introduced: (1) pre-learning epoch (more than 100 trials), a standard, eight-target task in which the target direction was randomly chosen from eight possible directions uniformly distributed over the circle; (2) learning epoch (more than 30 trials), a transformed, one-target task in which only one target (upwards, 90°) was presented and a rotational transformation was introduced between the cursor on the screen and the manipulandum; (3) post-learning epoch (more than 100 trials), where the default eight-target task was presented again; and (4) relearning epoch, same as the learning epoch. Rotations were 90°, 45°, --45°, or --90° and were chosen randomly for each session, but fixed for the duration of the learning epoch in a session. Note that learning here is local in direction since only one target direction was introduced during the learning epoch. The term *learned-movement direction* refers to the direction of hand movement needed to bring the cursor to the target for these rotations (thus, there were four possible learned-movement directions in this study: 0°, 45°, 135°, and 180°, associated with the --90°, --45°, +45°, and +90° transforms, respectively). Monkeys were trained for several months with the standard eight-target, task but did not see the transformations before the recordings. To achieve learning on a daily basis during the whole recording period (rather than switching between pre-learned behaviors), a different rotational transformation was randomly chosen for each day from the set of four possible transformations. To observe systematic change in the activity of neurons, the same transformation was repeated (greater than or equal to four repetitions for each transformation and each monkey, on different days). Note that in this paradigm, the monkeys learn the visuomotor rotation by repeated performance of a single movement (to the learned direction). To test whether the repetition could be responsible for the neuronal changes observed, we conducted control sessions. These sessions (termed "'repetition"\' sessions) consisted of a one-target task without any visuomotor transformation (namely, a standard task to one direction only). We performed 16 such sessions (ten with repeated movements to 90° and six with movements to 180°). ### Data analysis. {#s4a3} Psychophysics studies have shown that immediately after learning, humans exhibit aftereffects, which is evidence for the formation of an internal representation of the newly acquired skill ([@pbio-0020045-Lackner1]; [@pbio-0020045-Shadmehr1]; [@pbio-0020045-Kawato1]). This has been observed in monkeys as well ([@pbio-0020045-Paz1]). To compare neuronal activities for movements with same kinematics, we excluded the first trials (three to five) in the post-learning epoch that exhibited significant aftereffects (measured as the directional deviation at peak velocity from a straight movement and compared to the distribution of deviations before learning). For the remaining trials, we compared velocity profiles, initial direction as a function of time, and actual trajectories to verify that there were similar to the trajectories in the pre-learning epoch (see [Figure 1](#pbio-0020045-g001){ref-type="fig"}B--1D). We also compared reaction times and perpendicular deviations at peak velocity and endpoint locations. No significant difference was found between the pre- and post-learning in all three groups (*t*-test, *p* \> 0.1). We also verified that learning was the same during the whole recording period. We divided the recording period into two to three consecutive segments and compared (1) learning rates in the learning epochs and (2) aftereffect magnitudes and washout rates in the post-learning epoch ([@pbio-0020045-Paz1]). To further avoid changes in activity that result from any kinematic or dynamic differences, and since learning-related changes were only observed in activity taken from preparation for movement (before the go-signal), here we only report neuronal activity from this period, i.e., activity during the 600 ms following the target appearance but before the go-signal. We isolated 177 cells (113 from monkey W and 64 from monkey X) based on (1) the lack of significant change in activity during the first-hold period (during which no information was available about the upcoming trial) for the pre-learning epoch versus the post-learning epoch (by Mann-Whitney *U*-test); (2) the results of a one-way ANOVA showing a significant effect for direction; (3) more than five trials per direction both pre- and post-learning. We calculated spike counts in the 600-ms range following the target onset, referred to as the PA. Examining the neurons for changes in PD did not reveal any systematic or significant changes (bootstrap test, three of 177 showed a significant change) and PDs were uniformly distributed (Rayleigh test). MuI between the direction of the movement and each cell response was calculated by standard methods ([@pbio-0020045-Cover1]) using the formula where *d* is the direction of movement and *r* is the number of spikes (see [Figure 2](#pbio-0020045-g002){ref-type="fig"}). We used either the direct method for calculating *P(r)* or by assuming a Poisson distribution with the mean taken from all trials. We compensated for the limited number of trials (bias correction) by applying either analytical correction ([@pbio-0020045-Panzeri1]) or by shuffling trials between directions to obtain mean baseline and confidence intervals for the MuI; since both methods produced similar qualitative results, we report here the direct method, corrected analytically. For calculating the individual DI that each neuron conveys, we used the following formula ([@pbio-0020045-Rolls1]; [@pbio-0020045-Buracas1]) (for alternative definitions, see [@pbio-0020045-DeWeese1]): calculated separately for each direction *d*. To predict the direction of hand movement based on neuronal activity, we used two standard decoding methods: \(1) MAP estimator. This was carried out by assuming Poisson distribution of rates and independency between neurons. We sampled (with repetition) 100 cells and then cross-validated by selecting randomly one trial from each cell and direction, calculating the cell\'s mean firing rate from the rest of the trials and used the following formula ([@pbio-0020045-Sanger2]) to obtain the most likely direction: where σ*~i~*(*d*) denotes the mean firing rate of cell *i* in direction *d* and *r~i~* is the rate in the randomly drawn trial. For ease of computation, we took the log of the probability and did not calculate *N*, the normalization factor. The process was repeated 100 times and performed separately for the pre-learning and post-learning. \(2) PV analysis ([@pbio-0020045-Georgopoulos1]; [@pbio-0020045-Schwartz1]). One hundred twenty-nine cells (91 from monkey X and 38 from monkey W) were characterized as directionally tuned by fitting a cosine function (*r^2^* \> 0.5). The cells\' PDs were homogenously distributed both pre- and post-learning (Rao test, pre-learning, *p* = 0.4, post-learning, *p* = 0.5). We performed two bootstrap tests for significance. First, we shuffled trials from pre- and post-learning and calculated the difference between the deviations of the PV prediction pre-learning to that of the post-learning. The process was repeated 1,000 times to obtain confidence intervals. Second, we shuffled cells from days in which different transformations were learned and again obtained confidence limits. This process tests whether the improvement in prediction was indeed related to the specific direction learned in the session. We thank Chen Nathan and Thomas Boraud for participating in the recordings, Gal Chechick and Amir Globerson for helpful discussions of Information theory, and Hagai Bergman and Steven P. Wise for many fruitful discussions. This study was supported in part by a Center for Excellence grant (8006/00) administered by the Israel Science Foundation (ISF), by the Bundesministerium für Bildung und Forschung--Deutsch-Israelische Projektkooperation(BMBF--DIP), by grant 2001073 administrated by the Binational Science Foundation (BSF), and by a special contribution of the Golden Charitable Trust. RP was supported by a Constantiner fellowship. **Conflicts of interest.**The authors have declared that no conflicts of interest exist. **Author contributions.** RP and EV conceived and designed the experiments. RP and EV performed the experiments. RP analyzed the data. RP and EV wrote the paper. RP and EV performed the surgeries. DOI: [10.1371/journal.pbio.0020045](10.1371/journal.pbio.0020045) Copyright: © 2004 Paz and Vaadia. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Academic Editor: James Ashe, University of Minnesota DI : information per direction M1 : primary motor cortex MAP : maximum a posteriori MuI : mutual information PA : preparatory activity PD : preferred direction PV : population vector

PubMed Central

2024-06-05T03:55:46.089838

2004-2-17

PMC340949

Stuttering, with its characteristic disruption in verbal fluency, has been known for centuries; earliest descriptions probably date back to the Biblical Moses\' "slowness of speech and tongue" and his related avoidance behavior (Exodus 4, 10--13). Stuttering occurs in all cultures and ethnic groups ([@pbio-0020046-Andrews1]; [@pbio-0020046-Zimmermann1]), although prevalence might differ. Insofar as many of the steps in how we produce language normally are still a mystery, disorders like stuttering are even more poorly understood. However, genetic and neurobiological approaches are now giving us clues to causes and better treatments. What Is Stuttering? {#s2} =================== Stuttering is a disruption in the fluency of verbal expression characterized by involuntary, audible or silent, repetitions or prolongations of sounds or syllables ([Figure 1](#pbio-0020046-g001){ref-type="fig"}). These are not readily controllable and may be accompanied by other movements and by emotions of negative nature such as fear, embarrassment, or irritation ([@pbio-0020046-Wingate1]). Strictly speaking, stuttering is a symptom, not a disease, but the term *stuttering* usually refers to both the disorder and symptom. ::: {#pbio-0020046-g001 .fig} Figure 1 ::: {.caption} ###### Speech Waveforms and Sound Spectrograms of a Male Speaker Saying "PLoS Biology" The left column shows speech waveforms (amplitude as a function of time); the right column shows a time--frequency plot using a wavelet decomposition of these data. In the top row, speech is fluent; in the bottom row, stuttering typical repetitions occur at the "B" in "Biology." Four repetitions can be clearly identified (arrows) in the spectrogram (lower right). :::  ::: Developmental stuttering evolves before puberty, usually between two and five years of age, without apparent brain damage or other known cause ("idiopathic"). It is important to distinguish between this persistent developmental stuttering (PDS), which we focus on here, and acquired stuttering. Neurogenic or acquired stuttering occurs after a definable brain damage, e.g., stroke, intracerebral hemorrhage, or head trauma. It is a rare phenomenon that has been observed after lesions in a variety of brain areas ([@pbio-0020046-Grant1]; [@pbio-0020046-Ciabarra1]). The clinical presentation of developmental stuttering differs from acquired stuttering in that it is particularly prominent at the beginning of a word or a phrase, in long or meaningful words, or syntactically complex utterances ([@pbio-0020046-Karniol1]; [@pbio-0020046-Natke1]), and the associated anxiety and secondary symptoms are more pronounced ([@pbio-0020046-Ringo1]). Moreover, at repeated readings, stuttering frequency tends to decline (adaptation) and to occur at the same syllables as before (consistency). Nonetheless, the distinction between both types of stuttering is not strict. In children with perinatal or other brain damage, stuttering is more frequent than in age-matched controls, and both types of stuttering may overlap ([@pbio-0020046-Andrews1]). Who Is Affected? {#s3} ================ PDS is a very frequent disorder, with approximately 1% of the population suffering from this condition. An estimated 3 million people in the United States and 55 million people worldwide stutter. Prevalence is similar in all social classes. In many cases, stuttering severely impairs communication, with devastating socioeconomic consequences. However, there are also many stutterers who, despite their disorder, have become famous. For instance, Winston Churchill had to rehearse all his public speeches to perfection and even practiced answers to possible questions and criticisms to avoid stuttering. Charles Darwin also stuttered; interestingly, his grandfather Erasmus Darwin suffered from the same condition, highlighting the fact that stuttering runs in families and is likely to have a genetic basis. The incidence of PDS is about 5%, and its recovery rate is up to about 80%, resulting in a prevalence of PDS in about 1% of the adult population. As recovery is considerably more frequent in girls than in boys, the male-to-female ratio increases during childhood and adolescence to reach three or four males to every one female in adulthood. It is not clear to what extent this recovery is spontaneous or induced by early speech therapy. Also, there is no good way of predicting whether an affected child will recover ([@pbio-0020046-Yairi1]). The presence of affected family members suggests a hereditary component. The concordance rate is about 70% for monozygotic twins ([@pbio-0020046-Andrews1]; [@pbio-0020046-Felsenfeld1]), about 30% for dizygotic twins ([@pbio-0020046-Andrews1]; [@pbio-0020046-Felsenfeld1]), and 18% for siblings of the same sex ([@pbio-0020046-Andrews1]). Given the high recovery rate, it may well be that the group abnormalities observed in adults reflects impaired recovery rather than the causes of stuttering ([@pbio-0020046-Andrews1]). Changing Theories {#s4} ================= Over the centuries, a variety of theories about the origin of stuttering and corresponding treatment approaches have been proposed. In ancient Greece, theories referred to dryness of the tongue. In the 19^th^ century, abnormalities of the speech apparatus were thought to cause stuttering. Thus, treatment was based on extensive "plastic" surgery, often leading to mutilations and additional disabilities. Other treatment options were tongue-weights or mouth prostheses ([@pbio-0020046-Katz1]) ([Figure 2](#pbio-0020046-g002){ref-type="fig"}). In the 20th century, stuttering was primarily thought to be a psychogenic disorder. Consequently, psychoanalytical approaches and behavioral therapy were applied to solve possible neurotic conflicts ([@pbio-0020046-Plankers1]). However, studies of personality traits and child--parent interactions did not detect psychological patterns consistently associated with stuttering ([@pbio-0020046-Andrews1]). ::: {#pbio-0020046-g002 .fig} Figure 2 ::: {.caption} ###### Two Different Apparatuses to Prevent Stuttering On the left is a device by Gardner from 1899 to artificially add weight to the tongue (United States patent number 625,879). On the right is a more complex speech apparatus by Peate from 1912 (United States patent number 1,030,964). :::  ::: Other theories regard stuttering as a learned behavior resulting from disadvantageous external, usually parental, reactions to normal childhood dysfluencies ([@pbio-0020046-Johnson1]). While this model has failed to explain the core symptoms of stuttering ([@pbio-0020046-Zimmermann1]), it may well explain secondary symptoms ([@pbio-0020046-Andrews1]), and guided early parental intervention may prevent persistence into adulthood ([@pbio-0020046-Onslow1]). The severity of PDS is clearly modulated by arousal, nervousness, and other factors ([@pbio-0020046-Andrews1]). This has led to a two-factor model of PDS. The first factor is believed to cause the disorder and is most likely a structural or functional central nervous system (CNS) abnormality, whereas the second factor reinforces the first one, especially through avoidance learning. However, one should be careful to call the latter factor "psychogenic" or "psychological," because neuroscience has shown that learning is not simply "psychogenic" but leads to measurable changes in the brain ([@pbio-0020046-Kandel1]). In some cases, arousal actually improves stuttering instead of making it worse. Consequently, some famous stutterers have "treated" their stuttering by putting themselves on the spot. Anecdotally, the American actor Bruce Willis, who began stuttering at the age of eight, joined a drama club in high school and his stuttering vanished in front of an audience. Is Stuttering a Sensory, Motor, or Cognitive Disorder? {#s5} ====================================================== Stuttering subjects as a group differ from fluent control groups by showing, on average, slightly lower intelligence scores on both verbal and nonverbal tasks and by delays in speech development ([@pbio-0020046-Andrews1]; [@pbio-0020046-Paden1]). However, decreased intelligence scores need to be interpreted carefully, as stutterers show a schooling disadvantage of several months ([@pbio-0020046-Andrews1]). Associated symptoms comprise delays in tasks requiring a vocal response ([@pbio-0020046-Peters1]) and in complex bimanual timed tasks such as inserting a string in the eye of a needle ([@pbio-0020046-Vaughn1]), whereas many other studies on sensory--motor reaction times yielded inconsistent results ([@pbio-0020046-Andrews1]). Alterations of auditory feedback (e.g., delayed auditory feedback, frequency-altered feedback), various forms of other auditory stimulation (e.g., chorus reading), and alteration of speech rhythm (e.g., syllable-timed speech) yield a prompt and marked reduction of stuttering frequency, which has raised suspicions of impaired auditory processing or rhythmic pacemaking in stuttering subjects ([@pbio-0020046-Lee1]; [@pbio-0020046-Brady2]; [@pbio-0020046-Hall1]; [@pbio-0020046-Salmelin1]). Other groups have also reported discoordinated and delayed onset of complex articulation patterns in stuttering subjects ([@pbio-0020046-Caruso1]; [@pbio-0020046-vanLieshout1]). The assumption that stuttering might be a form of dystonia---involuntary muscle contractions produced by the CNS---specific to language production ([@pbio-0020046-Kiziltan1]) was not supported by a study on motor cortex excitability ([@pbio-0020046-Sommer2]). Neurochemistry, however, may link stuttering with disorders of a network of structures involved in the control of movement, the basal ganglia. An increase of the neurotransmitter dopamine has been associated with movement disorders such as Tourette syndrome ([@pbio-0020046-Comings1]; [@pbio-0020046-Abwender1]), which is a neurological disorder characterized by repeated and involuntary body movements and vocal sounds (motor and vocal tics). Accordingly, like Tourette syndrome, stuttering improves with antidopaminergic medication, e.g., neuroleptics such as haloperidol, risperidone, and olanzapine ([@pbio-0020046-Brady1]; [@pbio-0020046-Lavid1]; [@pbio-0020046-Maguire1]), and anecdotal reports suggest that it is accentuated or appears under treatment with dopaminergic medication ([@pbio-0020046-Koller1]; [@pbio-0020046-Anderson1]; [@pbio-0020046-Shahed1]). Hence, a hyperactivity of the dopaminergic neurotransmitter system has been hypothesized to contribute to stuttering ([@pbio-0020046-Wu1]). Although dopamine antagonists have a positive effect on stuttering, they all have side effects that have prevented them from being a first line treatment of stuttering. Lessons from Imaging the Brain {#s6} ============================== Given reports on acquired stuttering after brain trauma ([@pbio-0020046-Grant1]; [@pbio-0020046-Ciabarra1]), one might think that a lesion analysis (i.e., asking the question where do all lesions that lead to stuttering overlap) could help to find the location of an abnormality linked to stuttering. Unfortunately, lesions leading to stuttering are widespread and do not seem to follow an overlapping pattern. Even the contrary has been observed, a thalamic stroke after which stuttering was "cured" in a patient ([@pbio-0020046-Muroi1]). In fluent speakers, the left language-dominant brain hemisphere is most active during speech and language tasks. However, early studies on EEG lateralization already strongly suggested abnormal hemispheric dominance ([@pbio-0020046-Moore1]) in stutterers. With the advent of other noninvasive brain imaging techniques like positron emission tomography (PET) and functional magnetic resonance imaging (fMRI), it became possible to visualize brain activity of stutterers and compare these patterns to fluent controls. Following prominent theories that linked stuttering with an imbalance of hemispherical asymmetry ([@pbio-0020046-Travis1]; [@pbio-0020046-Moore1]), an important PET study ([@pbio-0020046-Fox1]) reported increased activation in the right hemisphere in a language task in developmental stutterers. Another PET study ([@pbio-0020046-Braun1]) confirmed this result, but added an important detail to the previous study: Braun and colleagues found that activity in the left hemisphere was more active during the production of stuttered speech, whereas activation of the right hemisphere was more correlated with fluent speech. Thus, the authors concluded that the primary dysfunction is located in the left hemisphere and that the hyperactivation of the right hemisphere might not be the cause of stuttering, but rather a compensatory process. A similar compensatory process has been observed after stroke and aphasia, where an intact right hemisphere can at least partially compensate for a loss of function ([@pbio-0020046-Weiller1]). Right hemisphere hyperactivation during fluent speech has been more recently confirmed with fMRI ([@pbio-0020046-Neumann1]). PET and fMRI have high spatial resolution, but because they only indirectly index brain activity through blood flow, their temporal resolution is rather limited. Magnetoencephalography (MEG) is the method of choice to investigate fine-grained temporal sequence of brain activity. Consequently, MEG was used to investigate stutterers and fluent controls reading single words ([@pbio-0020046-Salmelin2]). Importantly, stutterers were reported to have read most single words fluently. Nevertheless, the data showed a clear-cut difference between stutterers and controls. Whereas fluent controls activated left frontal brain areas involved in language planning before central areas involved in speech execution, this pattern was absent, even reversed, in stutterers. This was the first study to directly show a neuronal correlate of a hypothesized speech timing disorder in stutterers ([@pbio-0020046-VanRiper1]). Thus, functional neuroimaging studies have revealed two important facts: (i) in stutterers, the right hemisphere seems to be hyperactive, and (ii) a timing problem seems to exist between the left frontal and the left central cortex. The latter observation also fits various observations that have shown that stutterers have slight abnormalities in complex coordination tasks, suggesting that the underlying problem is located around motor and associated premotor brain areas. Are there structural abnormalities that parallel the functional abnormalities? The first anatomical study to investigate this question used high-resolution MR scans and found abnormalities of speech--language areas (Broca\'s and Wernicke\'s area) ([@pbio-0020046-Foundas1]). In addition, these researchers reported abnormalities in the gyrification pattern. Gyrification is a complex developmental procedure, and abnormalities in this process are an indicator of a developmental disorder. Another recent study investigated the hypothesis that impaired cortical connectivity might underlie timing disturbances between frontal and central brain regions observed in MEG studies ([Figure 3](#pbio-0020046-g003){ref-type="fig"}). Using a new MRI technique, diffusion tensor imaging (DTI), that allows the assessment of white matter ultrastructure, investigators saw an area of decreased white matter tract coherence in the Rolandic operculum ([@pbio-0020046-Sommer1]). This structure is adjacent to the primary motor representation of tongue, larynx, and pharynx ([@pbio-0020046-Martin1]) and the inferior arcuate fascicle linking temporal and frontal language areas, which both form a temporofrontal language system involved in word perception and production ([@pbio-0020046-Price1]). It is thus conceivable that disturbed signal transmission through fibers passing the left Rolandic operculum impairs the fast sensorimotor integration necessary for fluent speech production. This theory also explains why the normal temporal pattern of activation between premotor and motor cortex is disturbed ([@pbio-0020046-Salmelin2]) and why, as a consequence, the right hemisphere language areas try to compensate for this deficit ([@pbio-0020046-Fox1]). ::: {#pbio-0020046-g003 .fig} Figure 3 ::: {.caption} ###### Decreased Fiber Coherence Decreased fiber coherences, as observed with DTI, in persistent developmental stutterers compared with a fluent control group. A red dot indicates the peak difference in a coronal (top left), axial (top right), and a sagittal (bottom) slice. :::  ::: These new data also provide a theory to explain the mechanism of common fluency-inducing maneuvers like chorus reading, singing, and metronome reading that reduce stuttering instantaneously. All these procedures involve an external signal (i.e., other readers in chorus reading, the music in singing, and the metronome itself). All these external signals feed into the "speech production system" through the auditory cortex. It is thus possible that this external trigger signal reaches speech-producing central brain areas by circumventing the frontocentral disconnection and is able to resynchronize frontocentral decorrelated activity. In simple terms, these external cues can be seen as an external "pacemaker." Future Directions in Research {#s7} ============================= There are numerous outstanding issues in stuttering. If structural changes in the brain cause PDS, the key question is when this lesion appears. Although symptoms are somewhat different, it would be interesting to find out to what extent transient stuttering (which occurs in 3%--5% in childhood) is linked to PDS. It is possible that all children who show signs of stuttering develop a structural abnormality during development, but this is transient in those who become fluent speakers. If this is the case, it is even more important that therapy starts as early as possible if it is to have most impact. This question can now be answered with current methodology, i.e., noninvasive brain imaging using MRI. Given that boys are about four times less likely to recover from stuttering than girls, it is tempting to speculate that all stutterers have a slight abnormality, but only those that can use the right hemisphere for language can develop into fluent speakers. Language lateralization is less pronounced in women ([@pbio-0020046-McGlone1]) and might therefore be related to the fact that women show an overall lower incidence in PDS. Again, a developmental study comparing children who stutter with fluent controls and, most importantly, longitudinal studies on these children should be able to answer these questions. It is unlikely that stuttering is inherited in a simple fashion. Currently, a multifactorial model for genetic transmission is most likely. Moreover, it is unclear whether a certain genotype leads to stuttering or only represents a risk factor and that other environmental factors are necessary to develop PDS. Again, this question might be answered in the near future, as the National Institutes of Health has recently completed the data collection phase of a large stuttering sample for genetic linkage analysis. We thank Tobias Sommer and Andreas Starke for fruitful discussions and the Volkswagen Foundation as well as the Deutsche Forschungsgemeinschaft for funding. Christian Büchel is at NeuroImage Nord in the Department of Neurology at the University of Hamburg in Hamburg, Germany. Martin Sommer is at the Department of Clinical Neurophysiology at the University of Göttingen in Göttingen, Germany. E-mail: <buechel@uke.uni-hamburg.de> (CB) CNS : central nervous system DTI : diffusion tensor imaging fMRI : functional magnetic resonance imaging MEG : magnetoencephalography MRI : magnetic resonance imaging PDS : persistent developmental stuttering PET : positron emission tomography

PubMed Central

2024-06-05T03:55:46.094421

2004-2-17

PMC340950

Creative human beings are the torch-bearers of civilization. How does their creativity arise? What causes some minds/brains to achieve awe-inspiring artistic or scientific achievements? We cannot help but be fascinated by the fact that Shakespeare---a merchant\'s son with "small Latin and less Greek"---could emerge from the "nowhere" of rural Stratford to create the richest literary treasure in the English language. We wonder how Michelangelo---a stonecutter\'s son who also came from a rural nowhere---found within himself the vision to see the shape of David in a block of discarded marble or the apolcalyptic fresco of *The Last Judgment* on the wall of the Sistine Chapel. What genetic influences shaped their brains to create---and to create these very specific wondrous things? How did their environments promote or impede them? Would Michelangelo have been great without the patronage of the Medicis or the competitive edge induced by Leonardo? Great art and great science are indeed often forged in the smithy of pain---with the fire fueled by self-doubt, obsessive preoccupation, sorrow, depression, competition, or economic needs. *The Midnight Disease: The Drive to Write, Writer\'s Block, and the Creative Brain* by Alice Weaver Flaherty unites two intrinsically fascinating domains of knowledge---the workings of the brain and the nature of creativity. Its author, a neurologist who has also become a writer by virtue of having published her first nonacademic book, draws on her knowledge of neuroscience, her medical career as a clinician, and her experiences as a patient. Early in the book, she describes her own hospitalization for manic-depressive illness, a disclosure that implicitly places her in the pantheon of other artists who have suffered from serious mental illness and provides her with lustre-by-association. The result of all these juxtapositions is, however, a somewhat disconcerting blend of pop-science and pop-confessional genres. The author frequently talks to us in the first person, but one is not quite sure which person (the neuroscientist, the doctor, or the patient) is actually speaking. In other words, this book has a jarring lack of a strong single voice, despite a knack for often finding a fine turn-of-phrase or a clever word choice. Given that the book topic is promising and that the author can often write very well, it is dismaying that this book is not better than it is. It is written for the intelligent lay public, many of whom avidly collect and read "brain books" to expand their minds. Most painful is the fact that this book is filled with factual errors, glib and misleading generalizations, and careless misstatements. Perhaps most shocking and most erroneous, we are told (by a neurologist!) that "The tips of the temporal lobe can be lopped off without much changing a person\'s behavior." HM, the most famous patient to receive bilateral temporal lobectomy, remains frozen in a past linked to a never-changing present because he lost the capacity to retain new memories. Temporal lobe syndromes are discussed more accurately later in the book, but that is a weak excuse for this early error. We are also told that "manic depression is a genetically transmitted syndrome" (when, in fact, no replicable genetic loci have yet been identified), that "a very high proportion of manic depressives become writers" (the lifetime prevalence rate of bipolar disorder is approximately 1%, and only a tiny proportion of that 1% are writers), and that "electrophysiology, because it is dangerous, is rarely performed" (electrophysiology tools---e.g., the study of evoked potentials or electroencephalograms---are noninvasive and frequently used; recordings of the activity of individual neurons with electrodes placed in the gray matter are indeed rare, but nothing from the context suggests that this particular type of electrophysiology is being discussed). There are many more such careless misstatements. The intelligent lay reader deserves better than this. The book raises and addresses a variety of interesting questions that have intrigued many thoughtful people for more than two millennia. What is the nature of creativity? What is the difference between skill and creativity? What is the relation between mental illness and creativity? Is creativity inhibited when mental illnesses are treated? What is the relation between mind and brain? The book also addresses some unique and interesting twists on these questions. Its focus is the domain of writing, drawing from the author\'s own experience of a compulsion to write, or hypergraphia, following a pyschic break. What is the relationship between hypergraphia and the brain? Between writer\'s block and the brain? Are these problems always pathological, or do they sometimes enhance creativity? Does that college student who can\'t finish a term paper have a "disease"? Can "mind-expanding" drugs that affect the brain enhance creativity? In short, *The Midnight Disease* raises many important questions, but fails to address them completely and accurately. There is much more to learn, and much more to say, about the nature of creativity, its origins in the mind/brain and in the human genome, and its boundaries with health and disease. Nancy C. Andreasen is the Andrew H. Woods Chair of Psychiatry at the University of Iowa Hospitals and Clinics in Iowa City, Iowa, and director of The MIND Institute in Albuquerque, New Mexico, United States of America. E-mail: <nancy-andreasen@uiowa.edu> **Book Reviewed** Flaherty AW (2004) The midnight disease: The drive to write, writer\'s block, and the creative brain. New York: Houghton Mifflin. 272 pp. ISBN (hardcover) 0-618-23065-3. US\$24.00.

PubMed Central

2024-06-05T03:55:46.096287

2004-2-17

PMC340951

"*"I look forward to such an organization of the literary records of medicine that a puzzled worker in any part of the civilized world shall in an hour be able to gain the knowledge pertaining to a subject of the experience of every other man in the world."*---George Gould, first president of the Association of Medical Librarians (now the Medical Library Association), May 1898" For over 100 years, the Medical Library Association (MLA) has upheld the belief that quality information is essential for improved health and has worked to ensure that health sciences librarians have the skills, knowledge, and leadership necessary for the delivery of information in a biomedical setting. The association has also promoted the concept of unrestricted, affordable, and permanent access of health information worldwide. For example, MLA\'s peer-reviewed *Journal of the Medical Library Association* *(JMLA)*, formerly the *Bulletin of the Medical Library Association (BMLA)*, has been made available since January 2000 on PubMed Central (PMC), a digital archive of the life sciences journal literature developed and managed by the National Center for Biotechnology Information (NCBI) and the United States National Library of Medicine (NLM). Access to PMC is free and unrestricted. Recently, NLM, working with MLA headquarters, made the full-text archives of *BMLA* from 1911 onward available online through PMC. This is an excellent resource for the study of health information sciences and the management of knowledge-based information, putting into practice MLA\'s belief in open access. The association has supported open access to information in several other ways, including memberships in the Scholarly Publishing and Academic Resources Coalition (SPARC) and the Information Access Alliance (IAA). MLA\'s statement on open access, found at <http://www.mlanet.org/government/info_access/openaccess_statement.html>, defines the association\'s position on this important topic. However, open access increases the need for more sophisticated information management tools and systems, such as quality filtering and customization of clinical and research information at the point of need and decision-making. MLA is pursuing a number of initiatives that address the specific information needs of clinicians, healthcare students, biomedical researchers, and institutional leaders. Our members are excited to be in a unique position to develop tools, resources, and advice on how to find relevant information on the Internet. For example, MLA members have developed a *User\'s Guide to Finding and Evaluating Health Information on the Web* for the Pew Internet and American Life Project. The guide provides access and evaluation guidelines and MLA\'s top ten most useful Web sites, as well as lists of top Web sites for cancer, diabetes, and heart disease. MLA is furthering the concept of evidence-based medicine through its exploration and definition of expert searching techniques (see <http://www.mlanet.org/resources/expert_search/>) and the provision of continuing education opportunities in this area (see <http://www.mlanet.org/education/telecon/ebhc/index.html>). These techniques identify best practices and cutting-edge clinical and research knowledge and cull through a sometimes overwhelming amount of medical literature that continues to grow exponentially. MLA\'s work in the area of expert searching was prompted by the increased emphasis on evidence-based practice by the Institute of Medicine. This, along with the publicity following the unfortunate death of a healthy research volunteer at Johns Hopkins about the need for more vigilance in maintaining the quality of literature searching, has created a renewed interest in the knowledge base and skill set required for expert literature searching and expert consultation. The use of evidence- or knowledge-based information retrieved through the expert searching process can help insure the clinical, administrative, educational, and research success and positive performance of the individual healthcare provider as well as the hospital or academic health center. In addition to retrieving the best evidence, it is also important to deliver knowledge and services within the specialized context to patient care, research, and learning. MLA\'s exploration, along with NLM, of the informationist concept, i.e., specialist librarians who blend the knowledge and skills of both the clinical and information sciences, is defining new roles for librarians for providing filtered and customized clinical/research information at the point of need and decision-making (for more information, see <http://www.mlanet.org/research/informationist/>). Librarians are being recruited to join clinical and research teams as clinical medical librarians and information specialists in context and to provide expert consultation on issues ranging from informatics literacy to evidence-based medicine classes. Besides health-care providers, millions of consumers search for health information on the Web every year. Recognizing the documented difficulties and frustrations health professionals and consumers face in coping with the barrage of available information in a way that results in informed healthcare decisions, MLA has established its health information literacy program (see <http://www.mlanet.org/resources/healthlit/index.html>) to stress the importance of "information" in health literacy. The association defines health information literacy as the set of abilities needed to recognize a health information need; to identify likely information sources and use them to retrieve relevant information; to assess the quality of the information and its applicability to a specific situation; and to analyze, understand, and use the information to make good health decisions. MLA has also developed a resources Web site for health consumers at <http://www.mlanet.org/resources/consumr_index.html> and <http://caphis.mlanet.org/consumer/index.html> that helps them find quality health information on the Web. These tools are publicly available to anyone in the world at any time. MLA recognizes that this is a time of rapid change in our society in which the availability of digital resources in a networked environment provides unprecedented opportunities for more open access to the scientific and medical literature. As health sciences librarians, we are excited about the potential to serve a much wider group of international consumers, ranging from medical researchers to patients and their relatives. We will continue to work to convert access to information into access to knowledge. ::: {#pbio-0020048-g001 .fig} ::: {.caption} ###### [www.mlanet.org](www.mlanet.org) :::  ::: Patricia L. Thibodeau is the president and Carla J. Funk is the executive director of the Medical Library Association, located in Chicago, Illinois, United States of America. E-mail: <funk@mlahq.org> *BMLA* : *Bulletin of the Medical Library Association* IAA : Information Access Alliance *JMLA* : *Journal of the Medical Library Association* MLA : Medical Library Association NCBI : National Center for Biotechnology Information NLM : National Library of Medicine PMC : PubMed Central SPARC : Scholarly Publishing and Academic Resources Coalition

PubMed Central

2024-06-05T03:55:46.096937

2004-2-17

PMC340952

Introduction {#s1} ============ Chemotaxis is the process by which motile bacteria sense changes in their chemical environment and move to more favorable conditions ([@pbio-0020049-Bren1]). In peritrichously flagellated bacteria such as Escherichia coli and Bacillus subtilis, swimming alternates between smooth runs and reorientating tumbles. Smooth runs require that the flagellar motors spin counterclockwise, whereas tumbles result from clockwise spins. Bacteria follow a random walk that is biased in the presence of gradients of attractants and repellents by alternating the frequency of runs and tumble. Owing to their small size, most bacteria are unable to sense chemical gradients across the length of their body. Rather, they respond only to temporal changes. In particular, their stimulated response always returns to prestimulus levels despite the sustained presence of attractants or repellents. Sensory adaptation involves a rudimentary form of memory that allows bacteria to compare their current and past environments. Bacteria regulate chemotaxis using a network of interacting proteins. The basic mechanism in flagellated bacteria involves receptor-mediated phosphorylation of a cytoplasmic protein (CheY) that binds to the flagellar motor and changes the spin direction ([@pbio-0020049-Falke1]). This pathway is characterized best in the γ-proteobacteria---E. coli and Salmonella enterica serovar *typhimurium*. Even though less is known about chemotaxis in other species of bacteria, the evidence so far suggests that the pathways are mechanistically different despite the homology of the individual genes to their γ-proteobacteria counterparts. B. subtilis, *Helicobacter pylori, Myxococcus xanthus, Rhodobacter sphaeriodes*, and Sinorhizobium meliloti, for example, all use similar yet distinct set of pathway components to regulate chemotaxis ([@pbio-0020049-Armitage2]; [@pbio-0020049-Ward1]; [@pbio-0020049-Pittman1]; [@pbio-0020049-Sonenshein1]). E. coli and B. subtilis bias their motion towards favorable conditions with nearly identical behavior by adjusting the frequency of straight runs and reorienting tumbles. Both pathways (summarized in [Figure 1](#pbio-0020049-g001){ref-type="fig"} and [Table](#pbio-0020049-t001){ref-type="table"} 1) share five orthologous proteins with apparently identical biochemistry. How these individual orthologs contribute to the overall function, however, is different, as illustrated when synonymous orthologs are deleted in each organism. Deletion of the CheY response regulator causes E. coli to run exclusively and B. subtilis to tumble exclusively ([@pbio-0020049-Bischoff1]). When the CheR methyltransferase is deleted in E. coli, the cells are incapable of tumbles and only run. Likewise, when the CheB methylesterase is deleted, E. coli cells are incapable of runs and only tumble. In B. subtilis, cells still run and tumble when either CheB or CheR is deleted, though they no longer precisely adapt ([@pbio-0020049-Kirsch2], [@pbio-0020049-Kirsch1]). Remarkably, both genes complement in the heterologous host. Deletion of the CheW adaptor protein in E. coli results in a run-only phenotype, whereas there is no change in phenotype for the synonymous deletion in B. subtilis. When the genes involved in regulating methylation are deleted (*cheBR* in E. coli and *cheBCDR* in B. subtilis), E. coli does not adapt ([@pbio-0020049-Segall1]), whereas B. subtilis either oscillates or partially adapts when exposed to attractants ([@pbio-0020049-Kirby3]). These differences demonstrate that the pathways are different even though they involve homologous proteins. ::: {#pbio-0020049-g001 .fig} Figure 1 ::: {.caption} ###### The Chemotaxis Pathways in E. coli and B. subtilis \(A) E. coli. (B) B. subtilis. Both organisms respond to extracellular signals by regulating the activity of the CheA histidine kinase. CheA is coupled to transmembrane receptors (MCP) by an adaptor protein CheW. Chemoattractants, by binding the receptor, inhibit CheA in E. coli (red line) ([@pbio-0020049-Borkovich1]) and stimulate CheA in B. subtilis (green line) ([@pbio-0020049-Garrity1]). CheA phosphorylates CheY. Phosphorylated CheY binds to the flagellar motor and increases the frequency of tumbles in E. coli ([@pbio-0020049-Cluzel1]) and runs in B. subtilis ([@pbio-0020049-Bischoff1]). Phosphorylated CheY is also predicted to inhibit the receptor complex in B. subtilis (dashed line). Both organisms tune the sensitivity of CheA to ligands by reversibly methylating the receptors using the CheR methytransferase and CheB methylesterase ([@pbio-0020049-Zimmer1]; [@pbio-0020049-Sourjik2]). Phosphorylation of CheB by CheA increases its methylesterase activity nearly 100-fold ([@pbio-0020049-Anand1]). CheA activity is proportional to the degree of receptor methylation in E. coli. In B. subtilis, CheA activity depends on which residue is methylated, akin to a binary switch. E. coli possesses a phosphatase, CheZ, not present in B. subtilis, that enhances the rate of CheY dephosphorylation. B. subtilis possesses three chemotaxis proteins not found in E. coli: CheC, CheD, and CheV. CheC is a negative regulator of receptor methylation and homologous to the CheY-binding domain (P2) in CheA ([@pbio-0020049-Rosario3]; [@pbio-0020049-Rosario1]). CheD is a positive regulator of receptor methylation and also deamidates specific residues on the receptor ([@pbio-0020049-Kristich1]). CheV is a CheW-response regulator fusion. CheV is functionally redundant to CheW and is predicted to negatively regulate receptor activity (dashed line) ([@pbio-0020049-Rosario2]; [@pbio-0020049-Karatan1]). :::  ::: ::: {#pbio-0020049-t001 .table-wrap} Table 1 ::: {.caption} ###### Summary of Differences between E. coli and B. subtilis Chemotaxis :::  ::: To analyze and compare the two networks, we constructed mathematical models of both pathways. Numerous mathematical models exist for the chemotaxis pathway in E. coli ([@pbio-0020049-Goldbeter1]; [@pbio-0020049-Asakura1]; [@pbio-0020049-Knox1]; [@pbio-0020049-Bray2]; [@pbio-0020049-Bray1]; [@pbio-0020049-Hauri1]; [@pbio-0020049-Barkai1]; [@pbio-0020049-Spiro1]; [@pbio-0020049-Morton-Firth1]), and we combined the models proposed by [@pbio-0020049-Barkai1] and [@pbio-0020049-Sourjik1]). For B. subtilis, we constructed a mathematical model that proposes an alternative mechanism for sensory excitation and adaptation. We validated the model against published data for B. subtilis chemotaxis. As there are fewer data concerning chemotaxis in B. subtilis, the model makes predictions regarding the function of the chemotaxis proteins CheC, CheD, and CheV not present in *E. coli.* Both models demonstrate how two divergent species mediate the same task using orthologous genes with different circuitry. Despite the differences, both pathways involve the same control strategy. The mathematical details of both models are described in [Materials and Methods](#s4){ref-type="sec"}. Model Assumptions and Justification {#s1a} ----------------------------------- Both E. coli and B. subtilis regulate motility by controlling the phosphorylation of the CheY response regulator using the CheA histidine kinase. Phosphorylated CheY binds to the flagellar motor and increases the likelihood of reorientating tumbles in E. coli and straight runs in B. subtilis ([@pbio-0020049-Bischoff1]). CheY is dephosphorylated by the CheZ phosphatase in *E. coli. B. subtilis* does not possess a homolog to the CheZ phosphatase. Instead, the motor switch protein FliY is the phosphatase for CheY in B. subtilis. CheA forms a complex with transmembrane receptors and CheW. When chemoattractants bind to the receptors, CheA is inhibited in E. coli and activated in B. subtilis. The net result is the same in both organisms: chemoattractants increase the likelihood of straight runs. Building on the success of the E. coli models ([@pbio-0020049-Barkai1]; [@pbio-0020049-Morton-Firth1]), we employed a variant of the two-state model for receptor activation in B. subtilis. The two-state model treats the chemotaxis receptors, CheW, and CheA as a single entity and assumes the receptor complex adopts either an active or inactive comformation. Implicit in the two-state model is the assumption that the receptor complex is stable. The model assumes that the rate of CheA autophosphorylation is proportional to the average number of active receptor complexes in the cell. CheA, in turn, controls the rate of the phosphorylation for CheB, CheV, and CheY, as it is the phosphodonor. As the phosphorylation kinetics in B. subtilis have not been extensively investigated, the model uses the mechanism and parameters for phosphorylation cascade in E. coli proposed by [@pbio-0020049-Sourjik1]). Both organisms respond and adapt to chemoattractants at comparable speeds ([@pbio-0020049-Kirby3]; [@pbio-0020049-Sourjik2]), so it is reasonable to assume that the phosphorylation rates are similar. The model assumes that the mechanism for CheV phosphorylation is the same as CheY and CheB. In E. coli, CheW regulates CheA activity in a biphasic manner ([@pbio-0020049-Gegner1]). Ternary signaling complexes form when CheW joins receptor dimers with CheA dimers. The actual stoichiometry of the signaling complex is unknown, though it is known to form higher-order structures ([@pbio-0020049-Stock1]). At low concentrations, the number of signaling complexes is proportional to the concentration of CheW. At higher concentrations, CheW inhibits the formation of ternary signaling complexes. Instead of ternary (active) complexes, partial (inactive) complexes of receptor--CheW and CheW--CheA form. Only at intermediate, stoichiometric concentrations of CheW do the majority of free receptors and CheA form active ternary complexes. In addition to CheW, chemotaxis in B. subtilis involves CheV, a CheW--response regulator fusion. CheV is functionally redundant to CheW: deletion of either gene has no visible effect on chemotaxis ([@pbio-0020049-Rosario2]). Unlike CheW, the additional response regulator domain on CheV is necessary for proper function ([@pbio-0020049-Karatan1]). We propose that CheV forms an additional layer of regulation in B. subtilis, where phosphorylation of the response regulator domain activates CheV. By regulating the number of active CheV molecules, B. subtilis could dynamically regulates the number of functional signaling complexes using a biphasic mechanism similar to CheW. The model simplifies this proposed mechanism for parsimony and assumes unphosphorylated CheV disrupts the receptor complex and inhibits the activation of CheA. This feedback mechanism proposes a role for CheV in addition to its functional redundancy to CheW. We note that H. pylori precisely adapts using a methylation-independent process involving three CheV paralogs ([@pbio-0020049-Pittman1]), suggesting that perhaps it involves the same proposed CheV feedback mechanism for adaptation. B. subtilis also employs a methylation-independent chemotaxis mechanism; unlike E. coli, it still partially adapts to chemoattractants even when receptor methylation is disabled ([@pbio-0020049-Kirsch1], [@pbio-0020049-Kirsch2]; [@pbio-0020049-Rosario3]; [@pbio-0020049-Rosario1]). The model assumes that phosphorylated CheY forms a negative feedback loop, where it inactivates CheA by binding to receptors. No such loop exists in E. coli. Experimental data for B. subtilis (discussed later) indicate that CheY interacts with the receptors. This model provides one possible feedback mechanism for methylation-independent chemotaxis. The other possibility is CheV. While either CheY or CheV is sufficient for methylation-independent chemotaxis, the model predicts that both feedback loops are necessary to generate the oscillations that are observed in the *cheBCDR* strains ([@pbio-0020049-Kirby3]). The phosphorylation cascade is summarized in [Figure 2](#pbio-0020049-g002){ref-type="fig"}. ::: {#pbio-0020049-g002 .fig} Figure 2 ::: {.caption} ###### Model for the Phosphorylation Cascade in B. subtilis The model assumes that the receptor complex (receptor, CheA, CheC, CheD, and CheW) exists either in an active (*T^A^*) or inactive (*T^I^*) state. Active receptors stimulate CheA. CheA phosphorylates CheB, CheV, and CheY. Phosphorylated CheY (*Y~p~*) binds the receptor and increases the likelihood a receptor adopts an inactive conformation (thick red line). Phosphorylated CheY also binds the flagellar motor (*M*). The motor switch enhances the rate of CheY dephosphorylation ([@pbio-0020049-Szurmant1]). The model assumes that unphosphorylated CheV inhibits CheA by disrupting the receptor complex (thick blue line). :::  ::: In E. coli, CheA activity is roughly proportional to the number of methylated residues on the receptor ([@pbio-0020049-Bornhorst1]). E. coli adapts by altering the level of receptor methylation ([@pbio-0020049-Goy1]). In B. subtilis, CheA activity depends on the specific residue methylated. In the model, we propose that methylation of residue E630 increases activity, whereas methylation of residue E637 decreases activity. The model is supported by the following experiments ([@pbio-0020049-Zimmer1]). The amino acid substitution E630D, which renders the site permanently demethylated, decreases the activity of CheA, as inferred by analyzing the spin of the flagellar motor. Likewise, the substitution E637D increases the activity of CheA. In addition to residues E630 and E637, residue Q371 is also reversibly methylated. However, the substitution Q371D does not alter the activity or interfere with adaptation. As a result, we ignored it in the model. The model predicts that B. subtilis adapts to the addition of attractants by demethylating residue E630 and methylating residue E637. The reverse process is used to adapt to the loss of attractants. When B. subtilis is stimulated either by the addition or removal of attractants, the chemotaxis receptors are rapidly demethylated and then slowly remethylated ([@pbio-0020049-Kirby2]). Cast in terms of the model, one residue is demethylated and then the other is methylated. As a comparison, the receptors in E. coli are methylated when the cells are exposed to attractants and demethylated when the attractants are removed. When the *cheY* gene is deleted in B. subtilis, a methylation pattern similar to E. coli is observed: the receptors are demethylated when the cells are exposed to attractants and methylated when the attractants are removed ([@pbio-0020049-Kirby3]). These results demonstrate that CheY is necessary for normal patterns of methylation in B. subtilis. Similar behavior is observed when mutations are made to the active site of CheY ([@pbio-0020049-Kirby3]) or when missense mutations are made to a small region on the C-terminus of the McpB receptor (C. J. Kristich, unpublished data). These results suggest that phosphorylated CheY interacts with the receptor to coordinate selective methylation. In the model ([Figure 3](#pbio-0020049-g003){ref-type="fig"}), we propose that CheY forms a switch for selective methylation. Residue E637 is preferentially methylated when phosphorylated CheY binds to the receptor. Otherwise, residue E630 is methylated. This proposed mechanism explains the mutant behavior: when the interaction between phosphorylated CheY and the receptor is disrupted, only residue E630 is methylated. As methylation of this residue increases the activity of the CheA kinase, we expect that residue E630 is demethylated when cells are exposed to attractants and methylated when the attractants are removed (as observed in *cheY* mutants). However in the mutant, there are no complementary changes at residue E637, as it cannot be methylated. ::: {#pbio-0020049-g003 .fig} Figure 3 ::: {.caption} ###### Model for Selective Methylation in B. subtilis The model assumes that the receptor dimers exist in six different methylation states. The different methylation states are denoted by the variable *T~ij~*, where the index *i* denotes the methylation state of residue 630 and *j* denotes the state of residue 637. For example, *T* ~20~ denotes the concentration of dimers with both residues methylated at position 630 and none at position 637. For simplicity, the model assumes that at most two residues are methylated as additional states are superfluous. When receptors are methylated at residue 630, the signaling complex preferentially adopts an active conformation. When residue 637 is methylated, the signaling complex preferentially adopts an inactive conformation. When the dimers are partially methylated, the strength of activation or inhibition is attenuated. Selective methylation is coordinated by phosphorylated CheY (*Y~p~*). CheR methylates residue 637 when phosphorylated CheY is bound to the receptor and methylates residue 630 otherwise. :::  ::: As discussed previously, the model also predicts that the proposed interaction between phosphorylated CheY and the receptor forms a negative feedback loop that inhibits the CheA kinase in addition to its role in methylation. These two mechanisms form the following regulatory feedback loop. When there is an excess of phosphorylated CheY, CheA is inhibited and residue E637 is preferentially methylated (inhibiting residue). Likewise, when the majority of CheY is unphosphorylated, CheA is not repressed and residue E630 is preferentially methylated (activating residue). This feedback loop provides a regulatory mechanism for adaptation otherwise absent in B. subtilis. While in E. coli CheB phopshorylation is not necessary for adaptation ([@pbio-0020049-Alon1]), it forms a negative feedback loop as the rate of demethylation---catalyzed by CheB---is proportional to the activity of CheA ([@pbio-0020049-Anand1]). This feedback loop likely controls the basal activity and the speed of response ([@pbio-0020049-Hauri1]). However, in B. subtilis, the receptors are demethylated in response to both positive and negative stimuli. It is implausible that CheB phosphorylation provides a regulatory mechanism for selective methylation and, based on the available data, CheY provides the logical alternative. *cheC* and *cheD*, chemotaxis genes present in B. subtilis and missing in E. coli, are not treated explicitly in the model. Mutations to either gene are modeled implicitly by perturbing the kinetic parameters governing CheA activation and selective methylation. CheC is homologous to the P2 domain of CheA and the N-terminal domain of FliM ([@pbio-0020049-Kirby5]). Both domains bind CheY in E. coli. When CheC is deleted, the steady-state level of receptor methylation is roughly twice wild-type levels ([@pbio-0020049-Rosario1]). When CheD is deleted, the receptors are unmethylated ([@pbio-0020049-Rosario3]). Yeast two-hybrid experiments suggest that CheC and CheD interact with one another ([@pbio-0020049-Rosario1]). Collectively, these results suggest that CheC and CheD coordinate CheY-dependent selective methylation by protecting one residue and exposing the other using phosphorylated CheY as the cue. In addition to its role in methylation, CheD deaminates glutamine residues on the receptors ([@pbio-0020049-Kristich1]). As *cheD* mutants respond weakly to the addition of chemoattractants ([@pbio-0020049-Kirby5]), we hypothesize that deamidation strengthens the coupling between the receptor and CheA kinase. Simple loss of methylation is insufficient to explain the phenomena, since unmethylated *cheR* mutants still respond strongly to chemoattractants ([@pbio-0020049-Kirsch2]). We model deletions to CheD by decreasing the transition rate between active and inactive receptor complexes. Our justification, based on the model, is that the period of oscillations of flagellar rotation in the *cheBCDR* mutant is 100 s ([@pbio-0020049-Kirby3]), far slower than the response in wild-type (less than 1 s). Our biological justification is that the CheD modifications strengthen the coupling between the receptors and CheA. [@pbio-0020049-Barkai1]) demonstrated that activity-dependent methylation is necessary for robust adaptation in E. coli chemotaxis. They propose that CheB demethylates only active receptors. Subsequent models, involving more detail, require that CheR methylates only inactive receptor ([@pbio-0020049-Morton-Firth1]; [@pbio-0020049-Barkai2]; [@pbio-0020049-Mello1]). Adaptation results by balancing the rates of methylation and demethylation at steady state. In the B. subtilis model, activity-dependent methylation is also necessary for robust adaptation, albeit in a different form. With selective methylation, one option is that CheB demethylates residue 630 when the receptor is active and residue 637 when it is inactive. No equivalent assumption is necessary for CheR. Other alternatives are possible, though this one was the simplest considered. How CheB distinguishes between active and inactive receptors is unknown. Phosphorylation is not sufficient: receptors are also demethylated when CheA is inhibited ([@pbio-0020049-Kirby2]). The cue likely involves the same feedback loop regulating selective methylation: CheB binds residue 630 when phosphorylated CheY is bound to the receptor and binds residue 637 otherwise. In the present two-state model, however, this mechanism is not sufficient for robust adaptation. It is necessary to assume that CheB explicitly distinguishes between active and inactive receptors (as is the case with the E. coli models). Few kinetic measurements have been made for B. subtilis. On the one hand, we expect that the rates and concentrations are comparable to their E. coli counterparts, given that many B. subtilis chemotaxis proteins complement in E. coli. On the other hand, the additional feedback loops involving CheV and CheY could mask differences in the rates and concentrations between the two species. Unlike E. coli, many properties of the B. subtilis model, such as the steady-state bias and adaptation time, are insensitive to the kinetic parameters, suggesting that perhaps chemotaxis is more robust in B. subtilis than in E. coli. For lack of a better alternative, we used E. coli parameters for the B. subtilis model when available, as they produce results in the B. subtilis model consistent with experimental measurements. Many regulatory interactions proposed in B. subtilis model were inferred from mutants and lack explicit experimental confirmation. There are a number of experiments that could test the predictions made by the model, and we describe just a few. One experiment is to correlate receptor methylation with CheA activity in vitro using purified components ([@pbio-0020049-Ninfa1]; [@pbio-0020049-Borkovich2]). This in vitro setup could also be used to test CheD; the model predicts that CheD enhances CheA activity by post-translationally modifying the receptors. Another experimental option for correlating receptor methylation with CheA is to fuse fluorescent proteins to FliY and CheY and use fluorescence resonance energy transfer to measure the relative concentration of phosphorylated CheY for different engineered methylation states in vivo ([@pbio-0020049-Sourjik2]). The in vitro setup using purified components could test the proposed regulatory interactions between CheY and the receptor. We could also test the predicted regulatory interactions involving CheV by measuring the stability of the ternary receptor complex (receptor, CheV, and CheA) for different concentrations of phosphorylated CheA or CheV. Another option is to compare the response to ligand for different *cheV* mutants (e.g., *cheBCDR* versus *cheBCDRV*). Results {#s2} ======= Alternate Mechanisms for Adaptation {#s2a} ----------------------------------- Timecourse simulations of the models illustrate the process of adaptation in E. coli ([Figure 4](#pbio-0020049-g004){ref-type="fig"}A) and B. subtilis ([Figure 4](#pbio-0020049-g004){ref-type="fig"}B). Both models accurately reproduce the observed adaptation kinetics ([@pbio-0020049-Segall1]; [@pbio-0020049-Kirby3]). Upon the addition of attractant, the CheA kinase is inhibited in E. coli and activated in B. subtilis. This change correlates with a rapid decrease in the concentration of phosphorylated CheY in E. coli ([@pbio-0020049-Borkovich1]) and a rapid increase in B. subtilis ([@pbio-0020049-Garrity1]). Both species adapt by changing the methylation state of their receptors. Whereas adaptation to attractants in E. coli is commensurate with an increase in receptor methylation, adaptation in B. subtilis is commensurate with the change in the relative state of receptor methylation. The average number of residues methylated at position 630 decreases and the average number at position 637 increases. The relative change in methylation in B. subtilis correlates with the absolute change in methylation in E. coli. Both organisms adapt to the loss of attractants by reversing the process. ::: {#pbio-0020049-g004 .fig} Figure 4 ::: {.caption} ###### Simulation of Adaptation in E. coli and B. subtilis Attractant (10 μM) is added at 500 s and removed at 1,000 s. \(A) Timecourse simulation of phosphorylated CheY (left) and receptor methylation (right) in E. coli. \(B) Timecourse simulation of phosphorylated CheY (left) and receptor methylation (right) in B. subtilis. In both species, adaptation correlates with changes in receptor methylation. :::  ::: The concentration of phosphorylated CheB is proportional to the concentration of active receptors in E. coli and B. subtilis. This mechanism makes sense for E. coli, where CheB phosphorylation forms a negative feedback loop by de-methylating active receptors. However, it makes little sense in B. subtilis, where both active and inactive receptors are demethylated. Remarkably, however, experiments and simulation demonstrate that inactive receptors are demethylated just as efficiently as active receptors in B. subtilis, despite the fact that phosphorylation is necessary for CheB activity. What role phosphorylation of CheB plays in B. subtilis is unknown. We note that the homolog to CheB in Campylobacter jejuni lacks a response regulator domain. The B. subtilis model predicts that differential changes in methylation are symmetric. The increase in methylation at position 637 is matched by an equal decrease in methylation at position 630. These results predict that the average number of residues methylated is constant at all times. Experiments, however, paint a different picture ([@pbio-0020049-Kirby3]). While the total level of methylation is constant at steady state, dynamic changes in differential methylation are not symmetric. Upon the addition or removal of attractants, there is a rapid decrease in receptor methylation proportional to the amount of attractant added or removed. This rapid decrease is followed by slow increase in receptor methylation. Despite considerable effort, we were unable to develop a robust model that captures this asymmetric behavior. Likely, there are additional mechanisms involved. The logical suspects are CheC and CheD. One hypothesis is that CheC and CheD form a switch, where CheC protects one residue and CheD exposes the other. In such a model, the rate of demethylation needs to be much faster than that predicted by the E. coli kinetic parameters. While conceptually appealing, we are currently unable to propose such a mechanism that robustly adapts. Further elucidation of CheC and CheD is necessary. The model in this case clearly points out deficiencies in our knowledge. Adaptation Involves Similar Regulatory Strategy {#s2b} ----------------------------------------------- The two-state model for chemotaxis in E. coli assumes that CheR (*R*) binds only inactive receptors (*T^I^*) and that phosphorylated CheB (*B~P~*) binds active receptors (*T^A^*). In a simplified version of the model ([@pbio-0020049-Barkai1]), receptor methylation *m* is described by the differential equation where *k~B~* and *k~R~* are the rate constants and *K~B~* and *K~R~* are the Michaelis constants for receptor demethylation and methylation, respectively. We assume that the concentration of phosphorylated CheB is proportional to the concentration of active receptors. As argued previously by [@pbio-0020049-Barkai1]), the rates of receptor methylation and demethylation are, respectively, monotonically decreasing and increasing functions of receptor activity. As they are monotonic, the two rates intersect only once ([Figure 5](#pbio-0020049-g005){ref-type="fig"}A). Therefore, [Equation (1)](#pbio-0020049-e001){ref-type="disp-formula"} admits a single steady-state activity. As the rates are functions of receptor activity and not ligand concentration, the model precisely adapts to all ligand concentrations. The model is also robust; the rates are monotonic for all choices of kinetic parameters. However, where they intersect depends on the choice of kinetic parameters. Adaptation is robust, but other properties of the network are not. Similar arguments extend to the full model ([@pbio-0020049-Yi1]; [@pbio-0020049-Mello1]). ::: {#pbio-0020049-g005 .fig} Figure 5 ::: {.caption} ###### Graphical Illustration of Mechanism for Robust Adaptation \(A) Qualitative relationship among receptor activity, methylation, and demethylation in E. coli. The rate of demethylation is proportional to the number of active receptors, and the rate of methylation is inversely proportional to the number of active receptors. The system reaches steady state only when the two solid lines cross. As the rate of methylation decreases monotonically with receptor activity and the rate of demethylation increases monotonically with receptor activity, only one steady state is possible (*A\**) if the rates depend solely on receptor activity. The kinetic parameters change the slope of the curves, but not their monotonicity. Hence, adaptation is robust with respect to changes in the kinetic parameters. However, the point where they intersect does change with the parameters. \(B) Qualitative relationship between receptor activity and the differential rate of methylation in B. subtilis. The net rate of methylation at residue 630 decreases monotonically with receptor activity, and the net rate of methylation at residue 637 increases monotonically with receptor activity. By the same arguments, only one steady state (*A\**) is possible and, hence, adaptation is robust in B. subtilis. :::  ::: The B. subtilis model assumes that methylation is coordinated by phosphorylated CheY (*Y~p~*) and that CheB demethylates active receptors (*T^A^*) at residue 630 and inactive receptors (*T^I^*) at residue 637. If we simplify the model, the concentrations of receptors with residues methylated at 630 (*m* ~630~)and 637 (*m* ~637~) are described by the following two differential equations: where *K~Y~* is the Michaelis constant for phosphorylated CheY and the receptor. Subtracting [Equation (3)](#pbio-0020049-e003){ref-type="disp-formula"} from [Equation (2)](#pbio-0020049-e002){ref-type="disp-formula"}, we obtain the differential equation where Δ*m* = *m* ~630~ -- *m* ~637~. We assume the concentration *Y~P~* is proportional to the concentration of active receptors. The relative rate of methyation at residue 630 in [Equation (2)](#pbio-0020049-e002){ref-type="disp-formula"} is a monotonically decreasing function of receptor activity, and the relative rate of methylation at residue 637 in [Equation (3)](#pbio-0020049-e003){ref-type="disp-formula"} is an monotonically increasing function of receptor activity. By the same arguments used for the E. coli model, [Equation (4)](#pbio-0020049-e004){ref-type="disp-formula"} admits a single steady state ([Figure 5](#pbio-0020049-g005){ref-type="fig"}B) and the system robustly adapts to all concentrations. The difference between the two species is how receptor methylation forms memory. E. coli forms memory using the absolute level of receptor methylation *m*, and B. subtilis forms memory using the differential level of receptor methylation Δ*m*. The structure of Equations [(1)](#pbio-0020049-e001){ref-type="disp-formula"} and [(4)](#pbio-0020049-e004){ref-type="disp-formula"} are identical. One rate---proportional to the number of inactive receptors---increases the memory term, while the other rate---proportional to the number of active receptors---decreases the memory term. Both processes reach steady state only when the memory matches the current state. The structural similarities imply that both species employ the same core control strategy. The decision process is the same; the difference is in how the process is instantiated. The analogy is to running the same program on two different kinds of computers: same software, different machine code. However, as the next section demonstrates, how susceptible these pathways are to perturbation is different, suggesting a distinct evolutionary advantage for each underlying design. Both mechanisms are robust; adaptation does not depend on the values of the kinetic parameters. Robust adaptation requires feedback with integral memory ([@pbio-0020049-Yi1]). The same strategy is used in many engineering designs and, in fact, is a necessary component for robustness ([@pbio-0020049-Wonham1]). By including a memory term, a feedback controller is able to determine whether regulation is improving or degrading with time and dynamically compensate for changes in control. This similarity between biological and artificial controls suggests that engineering concepts such as integral feedback can be used to predict the regulatory structure of intracellular pathways as they direct model development and help exclude alternate models. As we have argued, the difference between the two organisms is how memory is stored using receptor methylation. From an engineering perspective, both designs---*m* and Δ*m*---are equivalent. Chemotaxis Is Robust {#s2c} -------------------- Adaptation is robust in E. coli chemotaxis; changes in the relative level of CheR expression did not alter the ability of E. coli to adapt to attractants ([@pbio-0020049-Alon1]). It has previously been argued that robustness is necessary for complex networks ([@pbio-0020049-Gerhart1]; [@pbio-0020049-Hartwell1]). The model predicts that adaptation is also robust in B. subtilis---not surprisingly, as we explicitly considered robustness in model development. While adaptation is robust in E. coli, other network properties, such as the steady-state levels of phosphorylated CheY and adaptation time, are not. As these properties also affect the ability of bacteria to respond effectively to their environment and find food sources, we hypothesize that the two additional feedback loops present in B. subtilis chemotaxis (see the blue and red thick lines in [Figure 2](#pbio-0020049-g002){ref-type="fig"}) buffer against mutation and stochastic fluctuations in protein expression. As a comparison, we plotted the steady-state levels of CheY phosphorylation and adaptation time as a function of CheB and CheR concentrations ([Figure 6](#pbio-0020049-g006){ref-type="fig"}). [Figure 6](#pbio-0020049-g006){ref-type="fig"} demonstrates that both properties in E. coli are sensitive to the concentrations of CheB and CheR. These predictions are consistent with experimental results ([@pbio-0020049-Alon1]). The B. subtilis model, on the other hand, predicts that the steady-state level of CheY phosphorylation is insensitive to the concentrations of CheB and CheR and that the adaptation time is insensitive to the concentration of CheR. These results are also consistent with experimental data, as deletions to either CheB or CheR do not change the network behavior in B. subtilis as strongly as they do in E. coli ([@pbio-0020049-Kirsch1], 1993b). ::: {#pbio-0020049-g006 .fig} Figure 6 ::: {.caption} ###### Sensitivity to Parameters in E. coli and B. subtilis \(A) E. coli. (B) B. subtilis. The top figures are plots of the steady-state concentration of phosphorylated CheY as a function of CheB and CheR concentrations. The bottom figures are plots of the adaptation time as a function of CheB and CheR concentrations. Adaptation time is defined as the length of time from the peak concentration in phosphorylated CheY (*Y~p~*) to within 5% of the steady-state concentration after the addition of attractant (10 μM). For all the concentrations considered, both models precisely adapt. :::  ::: While adaptation is a necessary component of chemotaxis, there are other design requirements of equal importance. One is positioning the concentration of phosphorylated CheY in a narrow functional range. The flagellar motor is exquisitely sensitive to changes in the concentration of phosphorylated CheY ([@pbio-0020049-Cluzel1]). Simulations of the models suggest that the steady-state concentration of phosphorylated CheY in B. subtilis, unlike E. coli, is robust to changes in the relative level of CheR expression ([Figure 6](#pbio-0020049-g006){ref-type="fig"}). As the B. subtilis pathway is more complex than that of E. coli, the robust positioning of phosphorylated CheY provides one possible benefit to offset the evolutionary cost associated with the additional complexity. Obviously, both organisms inhabit different ecological niches (colon and gut versus soil) and, as a result, are subject to different selective pressures, so it is difficult to explain their differences without further investigating the role of their environment. There is also the issue of sensitivity; E. coli is able to sense gradients in concentrations spanning five orders of magnitude. As formulated, both models fail to capture this observed behavior. Other mechanisms, such as receptor clustering ([@pbio-0020049-Maddock1]; [@pbio-0020049-Bray3]) and interactions between heterogeneous receptors ([@pbio-0020049-Mello2]), are needed to explain this sensitivity in E. coli. Experimental data suggest that the same mechanisms are involved in B. subtilis ([@pbio-0020049-Kirby4]; [@pbio-0020049-Zimmer2]). Methylation-Independent Chemotaxis {#s2d} ---------------------------------- In the absence of CheR and CheB, computer simulations, consistent with experiments ([@pbio-0020049-Kirsch1], [@pbio-0020049-Kirsch2]), demonstrate that B. subtilis partially adapts in response to the addition of chemoattractants (data not shown). The results are similar when either gene is deleted. A subpopulation (60%) of *B. subtilis cheBCDR* cells oscillates when stimulated with chemoattractants ([@pbio-0020049-Kirby3]). To model this behavior, we reduced the rate of transition between active and inactive receptor complexes by a factor of 500. This change produced a relaxation oscillator with a period of roughly 100 s that is observed experimentally ([Figure 7](#pbio-0020049-g007){ref-type="fig"}A). Wild-type cells respond in less than 1 s to attractants, thereby suggesting that the rate of signaling is slower in the mutant. We needed therefore to adjust the model to account for the relatively long period in the mutants. *cheD* mutants weakly respond to chemoattractants, suggesting that the coupling between the receptor and kinase is attenuated. These results suggest that CheD, which deaminates glutamine residues on the receptors ([@pbio-0020049-Kristich1]), enhances the coupling in the signaling complex. ::: {#pbio-0020049-g007 .fig} Figure 7 ::: {.caption} ###### Oscillations and Methylation-Independent Chemotaxis \(A) Timecourse simulation of *cheBCDR* strain in B. subtilis subject to the addition of attractants (100 μM) at 200 s and the removal at 500 s. Concentration of CheV was set at 8 nM. \(B) Timecourse simulation of the *cheBCDR* strain in B. subtilis subject to the addition of attractants (100 μM) at 200 s and the removal at 500 s, where the concentration of CheV is halved (4 nM). :::  ::: Oscillations are very sensitive to the choice of kinetic parameters. Experiments indicate that only a fraction of the *cheBCDR* mutants oscillate (60%). The remaining cells partially adapt to the addition of attractants ([@pbio-0020049-Kirby3]). We propose that the differences arise from stochastic variations in protein concentrations. In our simulations, we transition between the two phenotypes by adjusting the concentration of CheV by a factor of 2 ([Figure 7](#pbio-0020049-g007){ref-type="fig"}B). A similar change has no effect in simulated wild-type stains, consistent with the fact that experimental deletions of CheV do not produce a detectable phenotype. Chemical oscillations typically arise from the interplay of positive and negative feedback loops ([@pbio-0020049-Ferrell1]; [@pbio-0020049-Tyson1]). The model proposes that CheV and CheY form these feedback loops. There is no evidence to suggest that other feedback loops exist, as the remaining regulatory proteins are not present in the oscillating strain. The model predicts that CheV inhibition produces a positive feedback loop. Unphosphorylated CheV inhibits CheA activation (see the blue thick line in [Figure 2](#pbio-0020049-g002){ref-type="fig"}). As the concentration of phosphorylated CheV increases, the inhibition of CheA decreases, as there is less unphosphorylated CheV. Less inhibition leads to more phosphorylated CheV, and the cycle repeats itself. The net result is a positive feedback loop. This positive feedback loop forms a hysteresis: the kinase still remains active after the attractant is removed. Hysteresis is a common cause of oscillations in signal transduction cascades, as it results in two stable steady states: one where the concentration of phopshorylated CheY is high and the other where the concentration is low ([@pbio-0020049-Ferrell1]). When this hysteresis is coupled with negative feedback by CheY, the pathway oscillates as the negative feedback loop drives the pathway away from the high steady state towards the low steady state and then the low towards the high. The hysteresis disappears when the model accounts for CheD owing to the associated change in the kinetics. Even in the model for wild-type B. subtilis, the CheV positive feedback loop increases the sensitivity of the signaling response to chemoattractants. These predictions assign another possible function to CheV distinct from CheW. CheV is present in many motile species of bacteria, including coliform bacteria such as *S. typhimurium.* CheY Feedback Is a Relic of Vestigial Chemotaxis Pathway {#s2e} -------------------------------------------------------- We speculate that CheY feedback is a relic of a primitive chemotaxis pathway. It is unlikely that bacteria started with all of the necessary chemotaxis genes from the outset, but rather evolved or acquired methylation later ([@pbio-0020049-Boyd1]). The core pathway involving chemoreceptors, CheW, CheA, and CheY is present in all known species of motile bacteria. Homologs to the remaining chemotaxis genes are present in some species and absent in others, suggesting that they were subsequent innovations to the core pathway ([Table 2](#pbio-0020049-t002){ref-type="table"}). If the core pathway was present before these additional genes were acquired, there would need to be some sort of stopgap regulation. As many of these additional genes are involved in methylation, we suspect that early pathways were regulated by a methylation-independent process. CheY feedback is the logical first step towards a functioning chemotaxis pathway, as it provides a mechanism for precise adaptation involving the core pathway without the need for additional genes ([Figure 8](#pbio-0020049-g008){ref-type="fig"}). The mechanism is not robust; the model is sensitive to the choice of parameters. If robustness is important for survival and environmental adaptation, perhaps then the methylation genes were acquired (CheB, CheC, CheD, and CheR) to address this flaw. Additional factors also favor the acquisition of methylation: methylation broadens the range of concentrations over which the bacteria are able to detect gradients and further implicates methylation as an evolutionary upgrade to primitive CheY feedback. ::: {#pbio-0020049-g008 .fig} Figure 8 ::: {.caption} ###### CheY Feedback Is Sufficient for Precise Adaptation Timecourse simulation of model subject to the addition of attractants (10 μM) at 200 s and removal at 500 s. The model is described in [Materials and Methods](#s4){ref-type="sec"}. :::  ::: ::: {#pbio-0020049-t002 .table-wrap} Table 2 ::: {.caption} ###### Distribution of Chemotaxis-Like Genes and Number of Paralogs for a Representative Set of Microbial Organisms :::  Genes were determined either by annotation or simple BLAST searches. R. sphaeroides genes were taken from [@pbio-0020049-Porter1]). Some chemotaxis-like genes are not directly involved in motility, but are involved in other process, such as development ([@pbio-0020049-Kirby1]). Other genes, in particular paralogs to CheY, may be false positives. For further information, including FASTA files and alignments, refer to <http://genomics.lbl.gov/chris/chemotaxis/genes.html>. This table updates a similar table presented by [@pbio-0020049-Armitage1]) ^a^There are additional orthologs to CheA, CheB, and CheR on the plasmids pSymA and pSymB ^b^There is also a CheABR fusion ^c^CheA is fused to CheY in C. jejuni and H. pylori ^d^CheB lacks a response regulator domain in C. jejuni ::: Discussion {#s3} ========== That the two pathways are different is not surprising, as E. coli and B. subtilis likely diverged over 1 billion years ago ([@pbio-0020049-Kunst1]). That both organisms use homologous genes is also not surprising. Divergent species of bacteria likely tinker with a limited set of genes, as mutations that change regulatory interactions between genes are far more frequent than mutations that confer novel function ([@pbio-0020049-Jacob1]; [@pbio-0020049-Carroll1]). The genes may be similar, but how they interact with one another is different. In fact, other species of bacteria, each with their own idiosyncrasies, also have evolved novel chemotaxis pathways by tinkering with a small set of conserved genes and protein domains (see [Table 2](#pbio-0020049-t002){ref-type="table"}). The question then is whether other properties of the network, in addition to the genes, are conserved. The chemotaxis models for E. coli and B. subtilis indicate that the decision-making process is identical. The biochemistry is different, but the regulatory strategy is the same. Does this mean that regulation is conserved? Selective pressures likely constrain the evolution of most networks to ensure they function robustly despite intrinsic noise due to molecular fluctuations, stochastic gene expression, and mutation ([@pbio-0020049-Hartwell1]; [@pbio-0020049-von1]). Consequently, regulation becomes an indirect object of selection. As diverse physiological processes have equivalent regulatory needs such as homeostasis and adaptation, the underlying pathways, based on this hypothesis, involve identical control strategies. Bacteria constantly prune their genome, removing redundant and nonessential genes ([@pbio-0020049-Mira1]). As the chemotaxis pathways in E. coli and B. subtilis are functionally equivalent, it is not evident why chemotaxis is more complex in B. subtilis than in E. coli. One hypothesis is that the additional genes and feedback loops buffer against genetic mutation, though why B. subtilis is more robust is not clear. As both organisms inhabit different environments, the alternate designs and associated tradeoffs likely reflect niche adaptation. A similar hypothesis regarding the evolution of regulatory networks was proposed by [@pbio-0020049-Savageau1]) in his demand theory for metabolism. As evident from bacterial chemotaxis, we cannot necessarily predict the structure and behavior of a network based on protein homology alone, as subtle differences in the proteins affect how they function in the network and with whom they interact. As these differences result from alternate regulatory interactions, comparing and analyzing these loops in divergent organisms provide insight regarding the properties and design of intracellular networks. By studying bacteria in different environments, we can learn how network structures evolve. By constructing a model of B. subtilis chemotaxis and comparing it to models of E. coli chemotaxis, we were able to explore two mechanisms for sensory adaptation involving homologous genes. These models enabled us to interpret a large class of data involving many different experimental conditions and mutants. The conclusion from this theoretical study is that both networks involve the same core control process, though the physical interactions and feedback loops that form this process are different. The implication is that we need to study the systematic properties of the homologous pathway in divergent organisms, rather than focusing exclusively on the individual genes. The hope is to understand the relative advantage and significance of each design and not exhaustively study each special case. Materials and Methods {#s4} ===================== {#s4a} All simulations were performed in Matlab (Mathworks, Natick, Massachusetts, United States). Matlab m-files are available from <http://genomics.lbl.gov/~chris/chemotaxis>. ### *E. coli* chemotaxis model. {#s4a1} The chemotaxis model combines the two-state model proposed for adaptation by [@pbio-0020049-Barkai1]), with the model for the phosphorylation cascade proposed by [@pbio-0020049-Sourjik1]). The two-state model assumes that receptor complexes *T* exist in either an active (*T^A^*) or inactive (*T^I^*) state. Let *T~i~* denote the concentration of receptor complexes with *i* residues methylated and α*~i~*(*L*) denote the probability that the receptor complex *T~i~* is active when the concentration of chemoattractant is *L*. The concentration of active receptors is and the concentration of inactive receptors is For simplicity, we assumed that ligand binding is fast and employed the quasi-steady-state assumption. The probabilities α*~i~*(*L*) are given by the expression with these parameters: α~0~ = 0; α~1~ = 0.1; α~2~ = 0.5; α~3~ = 0.75; α~4~ = 1; α^L^ ~0~ = 0; α^L^ ~1~ = 0; α^L^ ~2~ = 0.1; α^L^ ~3~ = 0.5; α^L^ ~4~ = 1; *K~L~* = 10 μM ([@pbio-0020049-Barkai1]). We modeled the phosphorylation cascade using the mechanism and parameters proposed by [@pbio-0020049-Sourjik1]). We extended their model to include CheB phosphorylation. The parameters for CheB phosphorylation were inferred from the wild-type adaptation kinetics ([@pbio-0020049-Sourjik2]): The terms *A* and *A~p~* denote the concentrations of CheA and phosphorylated CheA, *Y* and *Y~p~* denote the concentrations of CheY and phosphorylated CheY, *B* and *B~p~* denote the concentrations of CheB and phosphorylated CheB, *M* denotes the concentration of FliM, and \[*MY~p~*\] denotes the concentration of phosphorylated CheY bound to FliM. We modeled receptor methylation using the mechanism proposed by [@pbio-0020049-Barkai1]), with the extensions proposed by [@pbio-0020049-Morton-Firth1]). For simplicity, we assume that the methylation reactions follow Michaelis--Menten kinetics. Similar results were obtained using mass action kinetics. In the Morton-Firth model, CheR binds only inactive receptors and phosphorylated CheB binds only active receptors. For the receptor *T~i~*, the rate of demethylation is *r~Bα~* *~i~*(*L*)*T~i~* and the rate of methylation is *r~B~*(1 -- α*~i~*(*L*))*T~i~*, where and Note that the rate of methylation is proportional to concentration of inactive receptors (1 -- α*~i~*(*L*))*T~i~* and the rate of demethylation is proportional to the concentration of active receptors α*~i~*(*L*)*T~i~*. A simple mass balance yields the following set of differential equations for the receptors: The parameters for the model are: *k~r~* = 0.255 s^--1^; *K~R~* = 0.251 nM; *k~B~* = 0.5 s^--1^; *K~B~* = 5.5 nM; *A* + *A~p~* = 5 nM; *B* + *B~p~* = 2 nM; *Y* + *Y~p~* + \[*MY~p~*\] = 17.9 nM; *M* + \[*MY~p~*\] = 5.8 nM; and *T* ~0~ + *T* ~1~ + *T* ~2~ + *T* ~3~ + *T* ~4~ = 5 nM ([@pbio-0020049-Sourjik1]). We note that the estimated concentrations for FliM and CheY were for fluorescent fusion proteins expressed from a plasmid and may be different from the wild-type concentrations. ### *B. subtilis* chemotaxis model. {#s4a2} The B. subtilis model employs a variation of the two-state model proposed for E. coli. The model assumes that the receptor complex adopts either an active or inactive conformation. However, receptors can adopt one of four signaling states: either active, inactive, weakly active, or weakly inactive. In this regard, the model distinguishes between the signaling state of receptor complex and receptor, and it can be considered a heterogeneous two-state model ([@pbio-0020049-Bornhorst2]). Let *T~ij~*denote the concentration of receptor dimers with *i* residues at position 630 methylated and *j* residues at position 637 methylated. We assume that at most two residues on a dimer are methylated. Additional methylation states are superfluous. The concentration of (strongly) active receptors is given by the expression and the concentration of (strongly) inactive receptors is given by the expression where *i~ij~* is the probability that the receptor complex *T~ij~* adopts an inactive conformation. The concentration of weakly active receptors is given by the expression where β(*L*) is the probability that a weakly active receptor adopts an active conformation. The concentration of weakly inactive receptors is given by the expression The physical picture is the following. Receptors can either activate or inactivate the CheA kinase. Receptor methylation increases the magnitude of activation or inactivation, likely by stabilizing the conformational change and the coupling between the receptor and kinase. When receptors are methylated (either at residue 630 or 637), the probability that they adopt a strong conformation increases. Unmethylated receptors always adopt a weak (active or inactive) conformation. These assumptions were necessary to construct a robust model. In the E. coli model, there are two boundary conditions: fully methylated receptors and unmethylated receptors. Furthermore, methylated receptors are active (α = 1) and unmethylated receptors are inactive (α = 0). In the B. subtilis model, there are four boundary conditions: *T* ~20~, *T* ~02~, *T* ~11~, and *T* ~00~. Furthermore, the methylated receptors *T* ~11~and unmethylated receptors *T* ~00~are partially active. We needed, therefore, to distinguish additional states to construct a robust model involving activity-dependent methylation. In a similar manner to the E. coli model, we assume that the kinetics for ligand binding are fast and employ the quasi-steady-state assumption for simplicity. The probabilities α*~ij~*(*L*) and *i~ij~*(*L*) are given by the expressions with these parameters: α~20~ = 1; α~10~ = 0.4, α~11~ = 0.2; α~00~ = α~01~ = α~02~ = 0; α^0^ ~20~ = 1; α^0^ ~10~ = 0.99, α^0^ ~11~ = 0.8; α^0^ ~00~ = α^0^ ~01~ = α^0^ ~02~ = 0; *i* ~02~ = 1; *i* ~01~ = 0.99, *i* ~11~ = 0.8; *i* ~00~ = *i* ~10~ = *i* ~20~ = 0; *i* ^0^ ~02~ = 1; *i* ^0^ ~01~ = 0.4, *i* ^0^ ~11~ = 0.2; *i* ^0^ ~00~ = *i* ^0^ ~10~ = *i* ^0^ ~20~ = 0; β = 0.2; β^0^ = 0.8; *K~L~* = 10 μM. The parameters were inferred from tethering experiments, where the attractant asparagine is added and then removed in a flowcell containing wild-type cells and the rotation of the flagellar motor is observed ([@pbio-0020049-Kirby3]) The model assumes that CheY negatively regulates CheA activity. The model assumes that only phosphorylated CheY (*Y~p~*) binds receptors. We model receptor binding with the following two differential equations: where \[*T*\] and \[*TY~p~*\] denote, respectively, the concentration of unbound and *Y~p~*-bound receptors. We assume that the fraction of active receptor complexes *C^A^* satisfies the following differential equation: where *k~A~* = 0.5 \[*T*\](1 + 10*T^A^* + 0.1*T^WA^*) and *k~I~* = 0.5 \[*TY~p~*\](2 + 10*T^I^* + 0.1*T^WI^*). The term *C* denotes the concentration of inactive receptor complexes. Evident from the expressions for *k~A~*and *k~I~*, weakly active and inactive receptors contribute less to the state of the receptor complex. The model for the phosphorylation cascade in B. subtilis is an extension of the model proposed for E. coli. The key differences are the addition of CheV and the loss of CheZ. We used a Michaelis--Menten-type expression to model inhibition of the CheA kinase by unphosphorylated CheV (*V*). There is no dedicated phosphatase for CheY in B. subtilis. However, the motor switch appears to enhance the CheY dephosphorylation when phosphorylated CheY is bound to the motor ([@pbio-0020049-Szurmant1]). We assume the rate of CheY dephosphorylation increases when phosphorylated CheY is bound to the motor: As we lack kinetic parameters for B. subtilis, we used the parameters from the E. coli model when available. The parameters for CheV and CheY dephosphorylation were chosen so that the dynamics of the model were similar to those observed in tethering experiments involving wild-type bacteria and *cheBCDR* mutants ([@pbio-0020049-Kirby3]). For simplicity, we used Michaelis--Menten kinetics to model the methylation reactions. Similar results were obtained using mass action kinetics. For the receptor *T~ij~* ~,~ the rate of demethylation for residue 630 is *r~Bα~* *~ij~*(*L*)*T~ij~* and the rate of demethylation for residue 637 is *r~B\ i~ij~~*(*L*)*T~ij~*, where The model assumes that only (strongly) active and inactive receptors are demethylated. The rate of demethylation for residue 630 is proportional to the concentration of (strongly) active receptors, and the rate for residue 637 is proportional to the concentration of (strongly) inactive receptors. The rate of methylation for residue 630 is *r^1^~R\ T~ij~~*and the rate for residue 637 is *r^2^~R\ T~ij~~*, where and Note that the rate of methylation for residue 637 is simply the rate of methylation times the probability that the receptor is bound with *Y~p~* and vice versa. A simple mass balance yields the following differential equation for the receptors: The parameters are the same as the E. coli model: *k~r~* = 0.255 s^--1^; *K~R~* = 0.251 nM; *k~B~* = 0.5 s^--1^; *K~B~* = 5.5 nM; *A* + *A~p~* = 5 nM; *B* + *B~p~* = 2 nM; *Y* + *Y~p~* + \[*MY~p~*\] + \[*TY~p~*\] = 17.9 nM; *M* + \[*MY~p~*\] = 5.8 nM; *T* ~20~ + *T* ~10~ + *T* ~00~ + *T* ~01~ + *T* ~02~ + *T* ~11~ = 5 nM; \[*T*\] + \[*TY~p~*\] = 5 nM. The model assumes that the concentration of CheV is 8 nM: *V* + *V~p~* = 8 nM. To model oscillations for the *cheBCDR* strain described in [Figure 7](#pbio-0020049-g007){ref-type="fig"}, we used the following differential equation to describe the fraction of active receptor complexes *C^A^* where *k~A~* = 0.001*T* (1 + 0.1*T^WA^*) and *k~I~* = 0.001\[*TY~p~*\] (2 + 0.1*T^WI^*) with the initial condition *T* ~00~ = 5 nM. The concentrations of CheB and CheR were set to 0 to account for their deletion. The subpopulation that partially adapts was modeled by setting the concentration of CheV = 4 nM. In this formulation, receptors adopt either a weakly active or weakly inactive conformation. We also induced a timescale separation necessary for a relaxation oscillator by decreasing the transition rate between active and inactive receptor complexes by a factor of 500. This change produced oscillations with a period of 100 s. To model precise adaptation with simple negative feedback by CheY as described in [Figure 8](#pbio-0020049-g008){ref-type="fig"}, we used the following differential equation to describe the fraction of active receptor complexes *C^A^*: where *k~A~* = 0.1\[*T*\]*T^WA^* and *k~I~* = 0.1\[*TY~p~*\]*T^WI^* with the initial condition *T* ~00~ = 5 nM. The concentrations of CheB and CheR were set to 0. We also needed to change the model for receptor binding: where *k~A~* = 0.01/(10 + *L*) + 0.036*L*/(10 + *L*). Supporting Information {#s5} ====================== Accession Numbers {#s5a} ----------------- The GenBank (<http://www.ncbi.nlm.nih.gov/Genbank/>) accession numbers for the genes and gene products discussed in this paper are E. coli CheA (AAC74958) and B. subtilis CheA (CAB13516), E. coli CheB (AAC74953) and B. subtilis CheB (CAB13506), B. subtilis CheC (CAB13518), B. subtilis CheD (CAB13519), E. coli CheR (AAC74954) and B. subtilis CheR (CAB14188), B. subtilis CheV (CAB13274), E. coli CheW (AAC74957) and B. subtilis CheW (CAB13517), E. coli CheY (AAC74952) and B. subtilis CheY (CAB13506), E. coli CheZ (AAC74951), and B. subtilis FliY (CAB13505). We thank Dennis Bray, Thomas Shimizu, and Matthew Levin for criticism of the manuscript and suggestions regarding the models. We also thank Denise Wolf, Keith Erickson, Hanna Walukiewicz, Brian Aufderheide, and Tim Grammer for their comments and criticisms. This work was supported by the Howard Hughes Medical Institute and the Defense Advanced Research Project Agency. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** CVR, JRK, and APA conceived and designed the experiments. CVR performed the experiments. CVR and JRK analyzed the data. CVR and APA contributed reagents/materials/analysis tools. CVR wrote the paper. Academic Editor: Marc W. Kirschner, Harvard University

PubMed Central

2024-06-05T03:55:46.097789

2004-2-17

PMC340953

The growing library of sequenced genomes is challenging scientists to extract new biological meaning from DNA sequences. Comparative analysis of the mouse and human genome, for example, has already revealed that mutation rates in the 3 billion base pairs of the human genome vary considerably. What accounts for this regional disparity, however, is unclear. Mutations---substitutions in the nucleotide bases of DNA---produce variation in the genome. In classical evolutionary theory, natural selection drives evolutionary change by determining which of these mutations live on in the next generation or die with the organism. Mutations can be neutral, harmful, or beneficial, though the neutral theory of molecular evolution predicts that most mutations are "nearly" neutral or only slightly deleterious, while beneficial mutations---which confer a survival advantage on an organism and, if it reproduces, on its progeny---are quite rare. As a whole, mutations occur at the rate of approximately five substitutions per billion nucleotide sites per year. There are many types of neutral mutations---that is, mutations that have no effect on function. DNA base substitutions that lie outside of gene-coding regions or occur within introns (regions that are excised before being translated into a protein sequence) can fall into this category. Neutral mutations can also occur within gene-coding regions. For example, there are many instances where more than one codon---say, CUU, CUC, CUA, CUG---specify the same amino acid---in this case, leucine. Since these mutations can be used to gauge the neutral mutation rate of a region in the genome, they can be used to analyze the relationship between local mutation rates and gene location. Correlating gene mutation rates with their location in the genome, Jeffrey Chuang and Hao Li not only confirm that regional mutation rates indeed exist, but also calculate the size of these regions. Strikingly, certain classes of genes tend to congregate in mutational "hot spots"---regions with high mutation rates---while other types of genes gravitate toward "cold spots"---regions with relatively low mutation rates. Chuang and Li first determined whether mutation rates have regional biases---that is, whether the frequency and distribution of mutations follow a distinct pattern along the genome. The researchers calculated the substitution rates of neutral mutations in nearly 15,000 orthologous mouse and human genes---orthologous genes are genes that have evolved from a common ancestor without diverging in biological function---and found that mutation rates were in fact skewed toward either high or low rates. Mutation rate analysis of the orthologs\' neighbors revealed rates similarly skewed toward high or low substitutions, indicating that the region itself, rather than a particular gene, is prone to these differential rates. These regions, Chuang and Li report, were either one megabase or ten megabases long, affecting up to roughly 100 genes. But the question remained: Does the organism take advantage of these mutational hot and cold spots? If there is an adaptive advantage, gene families should occur in an appropriate mutational zone. In mutational hot spots, for example, one would expect to find genes that would benefit from high rates of mutation, which would in turn facilitate flexible responses to constantly changing environmental stimuli. Likewise, one would expect genes in cold regions to need protection from potentially deleterious mutations. And that\'s just what Chuang and Li found. Overall, genes in hot regions code for proteins involved in cell signaling, such as olfactory receptors, G-protein coupled receptors, membrane proteins, and immune response proteins---being in an area subject to high mutation rates means these genes can evolve quickly enough to adapt to constantly changing stimuli. Cold-region genes code for "housekeeping" proteins involved in core cellular processes, like transcription regulation and protein modification---these genes tend to be highly conserved, changing very little since they first evolved. Thus, it appears that natural selection may also operate at the level of gene location, relegating genes to different mutational genomic niches according to their function. While Chuang and Li explore possible mechanisms to account for these genomic niches---such as gene duplication and gene transposition---they argue that the selective pressures that influence gene location are the same that influence mutations in genes. By calculating the sizes of these mutational hot and cold regions, the researchers lay the groundwork for investigating genetic mechanisms that operate on these scales. And by showing that location matters, they have revealed a new force in genome evolution. ::: {#pbio-0020051-g001 .fig} ::: {.caption} ###### Olfactory genes lie in a mutational "hot spot" :::  :::

PubMed Central

2024-06-05T03:55:46.103302

2004-2-17

PMC340954

The AIDS crisis has brought to public notice what has always been generally true---that the existing business model for drug development leads to high prices and unequal access. There is now widespread dissatisfaction with drug prices in both the developed ([@pbio-0020052-FUSA1]) and developing world ([@pbio-0020052-Correa1]). Governments and health insurers are finding ways to deny access to the newest and priciest products. In the United States and other countries without a universal public health system, the uninsured simply cannot afford the newest medicines. In developing countries, life-saving medicines are priced beyond the reach of most people, a morally offensive outcome ([@pbio-0020052-TV1]). Huge publicity surrounds negotiated price reductions for specific drugs in specific developing countries, yet the effect on the overall access problem is tiny. Today\'s high drug prices are a direct consequence of a business model that uses a single payment to cover both the cost of manufacture of a drug and the cost of the research and development (R&D) carried out by manufacturers to discover it. A 20-year patent-based marketing monopoly is then granted to the drug\'s developers to prevent their prices being undercut by 'generic' copies produced by manufacturers who do not have R&D costs to recover. Preventing such 'free riding' on R&D has become a global trade issue at the World Trade Organisation (WTO) ([@pbio-0020052-Drahos1]). The implementation of the TRIPS (Trade-Related Aspects of Intellectual Property Rights) agreement and a growing number of regional and bilateral agreements on intellectual property require most countries to implement tough patent systems that discourage or eliminate competition from manufacturers of generic medicines ([Box 1](#box1){ref-type="boxed-text"}). Unfortunately, monopoly-based business models have unpleasant side effects. Since the primary responsibility of any company is to maximise return on investment, it is unsurprising that there is pressure on pharmaceutical companies to set drug prices to whatever level gives the highest return, excluding those individuals who cannot afford to pay, rather than maximising the number of patients treated. There is also pressure to misuse the power given by patents, using them as anticompetitive weapons to block innovation and extend marketing monopolies. And there are growing fears that the huge growth in the use of patents is in itself starting to inhibit research ([@pbio-0020052-CIPR1]; [@pbio-0020052-Anonymous1]; [@pbio-0020052-RS1]). Something that is less well recognised is that this system is an enormously inefficient way of purchasing R&D. There is a considerable lack of transparency in pharmaceutical R&D investment, but the available data indicate that only about 10% of drug sales go towards R&D on new products. Only about one-quarter of new drug approvals are rated by the United States Food and Drug Administration (FDA) to have therapeutic benefit over existing treatments ([@pbio-0020052-NIHCM1]; see [Figure 1](#pbio-0020052-g001){ref-type="fig"}). Measured by investment, only about one-fifth of the 10% is invested in innovative products ([@pbio-0020052-Love1]). There is also very little research for diseases that primarily afflict the poor ([@pbio-0020052-Trouiller1]; [@pbio-0020052-WHO1]). ::: {#pbio-0020052-g001 .fig} Figure 1 ::: {.caption} ###### A Breakdown of the 1,035 New Drugs Approved by the FDA between 1989 and 2000 More than three-fourths are classed as having no therapeutic benefit over existing products, so-called 'me too' drugs ([@pbio-0020052-NIHCM1]). Less than 1% address diseases that primarily afflict the poor, for which new treatments would have the greatest effect on world healthcare ([@pbio-0020052-WHO1]). Industry trade associations, reports to investors, and data from income tax returns suggest somewhere between 10% and 15% of the \$430 billion revenues (reported in 2002) are spent on R&D, but data from regulatory bodies imply that only approximately 2%--3% is actually spent on R&D that leads to new medicines with therapeutic benefits over existing ones, and even this is inflated by research primarily designed to achieve marketing outcomes ([@pbio-0020052-Love1]; [@pbio-0020052-WHO1]). :::  ::: Propping up the present structure for financing R&D ([Figure 2A](#pbio-0020052-g002){ref-type="fig"}) is the widely held belief that the private sector plays a key role in the development of new medicines and that it is necessary to grant patents to incentivise private-sector financing. If this were true, it would make sense to tolerate all sorts of bad outcomes, because the fruits of R&D eventually benefit everyone. But granting a 20-year marketing monopoly on a patented invention is only one way to finance R&D, and the shortcomings of the present system are increasingly hard to ignore. Suggestions for alternatives are beginning to come from many quarters ([@pbio-0020052-Baker1]; [@pbio-0020052-CGSD1]; [@pbio-0020052-Hubbard1]; [@pbio-0020052-Weisbrod1]). In this essay, we present practical proposals to modify trade rules based solely on intellectual property so that alternative policy instruments can be used to encourage innovation. ::: {#pbio-0020052-g002 .fig} Figure 2 ::: {.caption} ###### Funding Healthcare R&D \(A) A schematic of the way the public currently funds healthcare R&D. Academic research funded by government research agencies is paid for via taxes. This is a mixture of pure research into fundamentals and directed research, including clinical trials. Despite this, there is a dogma that academic R&D cannot produce drugs since it does not have the required commercial pressures to turn ideas into products. Patents ensure the public pays for commercial R&D via their purchase of medicines at high prices, compared to those of generic copies. The distortion of research priorities (too much spent on 'me too' drugs and too little on neglected diseases) has been recognised by governments for some time, and a variety of push-and-pull mechanisms have been introduced (or are being considered) to encourage research that more closely reflects public priorities. Examples of push incentives are tax breaks for R&D and other incentives such as special marketing monopolies for products as a reward for investing in research on orphan drugs or testing with pediatric patients. Pull incentives currently being discussed are advance-purchase commitments, with which governments guarantee to buy a certain amount of a drug if one is developed, or prize models. Some of these schemes are thought to be inefficient, particularly those that are indiscriminate and provide expensive subsidies relative to the amount of new R&D they 'encourage'. \(B) A schematic of the way funding of healthcare R&D could work if separate competitive markets for sales and R&D were created. A crucial difference is the absence of monopolies on final products, enabling competition between generic producers and greatly reduced prices. Incentives to develop new drugs would be provided by a new virtual market in R&D. 'Nationally directed R&D funds' could represent anything from rewards for innovation using market based mechanisms such as prize models (see text) to centralised funding agencies, similar to the NIH model, or multiple R&D investment funding bodies that compete for new resources. Contributions to R&D could be via taxation or as a legal obligation when paying for private healthcare plans (see text). The ability to design what would be rewarded in the virtual market would allow governments to set R&D priorities and build up local capacity within their own countries. Countries could choose weaker patent protection and create an environment in which all research groups could build on each other\'s work. :::  ::: A New Trade Framework {#s2} ===================== Analysis of worldwide drug expenditure shows that spending varies, but is close to 1% of the gross domestic product (GDP) in most developed and developing countries ([@pbio-0020052-Love2]). Assuming that about a tenth of the revenue from the sale of drugs is ploughed back into R&D on new products, that means that countries already indirectly contribute about 0.1% GDP to support this. This contribution is enforced by trade agreements, which require the granting of patents to prevent 'free riding' via the purchase of generic drugs (see [Box 1](#box1){ref-type="boxed-text"}). Suppose the World Health Organisation (WHO) developed an R&D contribution 'norm' based upon this or a more appropriate figure and that there was international agreement that countries evaluated as meeting this norm would no longer be regarded as 'free riding'. Trade rules could then be modified to allow countries to meet this norm *by any means*, not just by the implementation of strict TRIPS intellectual property rules, as at present. Countries that met the norm would then be free to decide whether they wanted to follow a strictly patent-based system as at present, with high drug prices for 20 years, or experiment with new models based on the creation of separate competitive markets for sales and R&D ([Figure 2B](#pbio-0020052-g002){ref-type="fig"}). Countries adopting the latter system would remove patents on final drug compounds, placing them in the public domain. This would allow them to become a freely traded commodity, creating a competitive manufacture and sales market with low generic prices. At the same time, in order to meet the required R&D contribution norm, they would have to create an efficient R&D virtual market alongside. However, the costs of this would be more than offset by the reduction in drugs prices, making substantial savings for that country overall. Business Models for an Effective Virtual R&D Market {#s3} =================================================== The existing system ([Figure 2A](#pbio-0020052-g002){ref-type="fig"}), despite its failings, does lead to the development of new drugs. The challenge in creating a virtual R&D market is to find viable business models for successful drug development in the absence of marketing monopoly incentives. One obvious approach is direct funding of drug development. For example, the National Institutes of Health (NIH), the national agency in the United States, already spends \$27 billion per year on research, a substantial amount of which is directed towards drug development, including clinical trials. The NIH already has a track record in developing important drugs for severe illnesses, such as cancer or AIDS, showing that this is a viable model. It is also widely recognised that much of the research carried out across the world by similar agencies underpins the existing commercial research that leads to new drugs. Governments could expand direct funding for drug development, either through the existing structures in academia or through funding R&D arms of existing companies to carry out specific drug R&D. Such directed drug development funding could be similar to existing nonprofit development projects, such as those currently resourced to address treatments for neglected diseases like malaria and tuberculosis (TB). Examples of such projects are the Medicines for Malaria Venture ([www.mmv.org](www.mmv.org)), the Global Alliance for TB Drug Development ([www.tballiance.org](www.tballiance.org)), the International AIDS Vaccine Initiative ([www.iavi.org](www.iavi.org)), the Drugs for Neglected Diseases Initiative ([@pbio-0020052-Butler2]) ([www.dndi.org](www.dndi.org)), and the Institute for One World Health ([www.oneworldhealth.org](www.oneworldhealth.org)). Many are doubtful that increased direct funding would generate sufficient incentives or be managed efficiently enough. An alternative market-based approach is one in which R&D organisations compete for rewards for specific R&D output, referred to by economists as a prize model ([@pbio-0020052-Wright1]; [@pbio-0020052-Kremer1]; [@pbio-0020052-Shavell1]). In a simple formulation, governments would place large sums into a fund that would be allocated every year to firms that bring new products to market. This could work with or without patents. If products were protected by patents or other intellectual property claims, the government could grant compulsory licenses (a procedure allowed by trade agreements to override monopoly rights on a patent, in return for compensation to rights owners; see [Box 1](#box1){ref-type="boxed-text"}) and permit rapid introduction of generic competition. The reward system could be a lump-sum payment, eliminating any incentive to continue to market the product, or a long-term payout structure, which would depend upon evidence of both usage and efficacy. Prize systems could be designed to be fairly similar to the current system, with big payoffs for successful entrepreneurs, but even with this approach, there would be huge opportunities to improve welfare. The reward system could be more rational than the existing system, allocating greater rewards for innovative products and less for 'me too' products that do not work better than existing products. Premiums could be given for therapies that address treatment gaps or for inventions that pave the way to new classes of drugs. Organisations competing for prizes might be expected to behave secretly to ensure that they are the ones to obtain 'credit' for the fruits of their work. However, progress in research is also driven by free exchange of information. It may be possible to design models that both reward R&D outputs and at the same time encourage complete and continuous openness with intermediate research outputs. There are now a number of examples of open collaborative public goods models ([@pbio-0020052-Cukier1]), such as those used for the Human Genome Project. The proponents of such models point to the success of GNU/Linux in the software field as evidence that major projects can be undertaken with radically different business models. One of the benefits of complete openness is that it allows independent and open evaluation of R&D outputs, which helps in the allocation of 'credit' whether in the form or prizes or new research grants. The open-access publishing movement ([@pbio-0020052-Brown1]) has the potential to help in this process by allowing independent analysis of published science, which will help research funding agencies measure research outputs. Competitive Intermediators {#s4} ========================== An R&D contribution norm, established by treaty, would ensure that the amount of money being spent on R&D is maintained. However, new mechanisms would be needed to collect the money to finance the R&D, as it would no longer come via drug sales. This could be via general taxation, although in countries with a private health insurance system this may be anathema. Many will also worry that a centralised national drug development agency taking decisions on R&D priorities and allocation of funds (via prizes or grants as discussed above) could easily become bureaucratic and inefficient. As a possible alternative, we propose a competitive financing scheme that would work through R&D investment intermediators. These R&D funds would be licensed and regulated (like pension funds). Their role would be to manage R&D assets on behalf of consumers. Individuals (or employers) would be required to make minimum contributions into R&D funds, much as there are mandatory contributions to social security or health insurance or to pension funds. Government would set the required contribution, but the individual (or employer) would be free to choose the particular intermediator that received their contributions. Intermediators would compete to attract funds to invest in R&D on the basis of their prowess for drug development and upon their priorities. Different business models for financing R&D could be tested in such a market, with intermediators experimenting with prize systems, direct investments in profit or nonprofit entities, open collaborative public good models, or other approaches. A Change for the Common Good {#s5} ============================ We believe the economics of a change in the paradigm for funding R&D are highly favourable. Taken together, the two core steps of changing the trade framework and moving away from marketing monopolies can change the world in a positive way. We can raise global R&D levels as a matter of policy and ensure that resources flow into the areas of the greatest need, and we can do so knowing that the poor and the rich will have access to new inventions at marginal cost. Policy-makers will be weaned from their current unhealthy addiction to ever-higher levels of intellectual property rights as the only instrument to raise R&D levels, a path that has increasingly reached diminishing returns or become counterproductive. With new instruments to address the overall levels of R&D investment, policy-makers can more constructively address the well-known inefficiencies in the patent system without the fear that global R&D levels will suffer and explore alternative models ([@pbio-0020052-Butler1]). At the same time, the system of prescribing medicines will be transformed by a substantial reduction in the distorting influences of the current multibillion-dollar industry of marketing medicines to doctors and (increasingly) directly to the public. Similarly, without marketing monopolies to protect, there will be far less spent to influence the governments that set the rules that regulate such monopolies. If implemented worldwide, one of our most vexing ethical dilemmas can be resolved in a manner that actually promotes the Doha Declaration on TRIPS and Public Health mandate to encourage access to medicine for all. ::: {.caption} ###### Box 1. Trade Agreements on Intellectual Property ::: The most important is the World Trade Organisation (WTO) agreement on Trade-Related Aspects of Intellectual Property (TRIPS), which requires member countries to issue 20-year patents on all fields of technology. All but the least-developed countries must comply by 2005. Going much further than the TRIPS are a plethora of regional and bilateral 'TRIPS-Plus' trade agreements, pushed in particular by the United States, which require even higher levels of intellectual property protection, such as limitations on the use of compulsory licensing, a tool used by governments to override the strong exclusive rights of a patent in return for compensation to patent owners. In 2001 the WTO adopted the Doha Declaration on TRIPS and Public Health, which said that 'the Agreement can and should be interpreted and implemented in a manner supportive of WTO Members' right to protect public health and, in particular, to promote access to medicines for all\'. In order to promote 'access to medicines for all', countries have to find new ways of financing R&D. These ideas have been developed in collaboration with attendees of a series of workshops hosted by Aventis, the TransAtlantic Consumer Dialogue, the Rockefeller Foundation, Médicins sans Frontières, Oxfam, Health Action International, and others. The views expressed in this article are those of the authors and do not necessarily reflect those of the Sanger Institute or the Wellcome Trust. Tim Hubbard is Head of Human Genome Analysis at the Wellcome Trust Sanger Institute in Hinxton, United Kingdom. James Love is Director of the Consumer Project on Technology (CPTech) in Washington, District of Columbia, the United States of America. E-mail: <james.love@cptech.org> FDA : Food and Drug Administration GDP : gross domestic product NIH : National Institutes of Health R&D : research and development TB : tuberculosis TRIPS : Trade-Related Aspects of Intellectual Property WHO : World Health Organisation WTO : World Trade Organisation

PubMed Central

2024-06-05T03:55:46.104257

2004-2-17

PMC340955

By next summer, more than 40% of Streptococcus pneumoniae strains in the United States will resist both penicillin and erythromycin, according to a recent prediction from the Harvard School of Public Health. The forecast, based on mathematical modeling, was published in the spring of 2003. It\'s too early to tell whether that prediction is precisely on track, according to the senior author on that paper, Marc Lipsitch. But no one doubts that multidrug resistance in this common bug---responsible for diseases that range from sinus trouble and ear infections to meningitis and pneumonia---is speeding up. It is the certain fate of all antibacterials to be fought off eventually by the pathogens they target. The fact that the process is accelerating has been alarming public health officials for some time, especially in the United States. We need new ways to defeat disease, and we will need them forever. Tried and True---and Tired? {#s2} =========================== Antibiotics have traditionally been plucked from nature\'s battleground. For billions of years, tiny organisms have engaged in an arms race, hurling toxic molecules at each other in the struggle to prosper. Nearly all of today\'s antibiotics are versions of weapons long wielded by microbes and fungi. Chemical synthesis of entirely human-created antibiotics has so far yielded only fluoroquinolones, a group of broad-spectrum antibiotics that includes Cipro, which became famously scarce during the 2001 anthrax scare, and linezolid (trade-named Zyvox), which is effective against some resistant strains of Staphylococcus, Streptococcus, and Enterococcus. The usual way to find a new antibiotic has been laborious screening of immense libraries of compounds, natural and otherwise. Some argue that screening chemical libraries is approaching a deadend. There may be diminishing returns from screening, but it\'s not quite dead yet: in October, researchers at the University of Wisconsin at Madison reported a new class of bacterial RNA polymerase inhibitors with antibiotic potential. They were found by screening for molecules that prevent Escherichia coli from transcribing RNA. Christopher T. Walsh of Harvard Medical School says screening\'s problem may be simply that libraries aren\'t good enough. Marine organisms have not been studied well, he points out, and 90% of organisms in the biosphere can\'t be cultured in standard ways. He says, "We\'re missing 90% of them every time we go and look in nature." Walsh is doing his bit to create new libraries. He and his colleagues have recently employed combinatorial biosynthesis to learn how to use part of the machinery for assembling cyclic peptide antibiotics to control their architecture. The result was a small library of natural product analogs, some of which have improved antibiotic activity against common bacterial pathogens. "There are dozens of such enzymatic domains that in principle one could clone, express, and test with other substrates. I view that as the kind of thing we should do," he says. For example, Walsh suggests, it is a reasonable approach to second-generation improvement of daptomycin, the antibiotic most recently approved for sale in the United States. Improving on Nature {#s3} =================== Walsh collaborates with Chaitan Khosla of Stanford University on finding ways to make existing antibiotics better. They are studying biosynthesis of rifamycin, an antibiotic that is increasingly less effective against its prime target, tuberculosis (TB) (see [Figure 1](#pbio-0020053-g001){ref-type="fig"}). "In the course of learning about that pathway, we\'ve learned a few interesting things lately about how that molecule is initiated, and we\'re trying to apply it in other contexts, especially in the context of erythromycin biosynthesis," Khosla says. The idea would be to make a molecule that might be more effective against bacteria that are becoming resistant to rifamycin---and are already naturally resistant to molecules like erythromycin. ::: {#pbio-0020053-g001 .fig} Figure 1 ::: {.caption} ###### TB Drug Resistance This 40-year-old Estonian truck driver\'s TB is resistant to drugs and his right lung was removed three days before this picture was taken. (Photograph by WHO/STB/Colors Magazine/J. Langvad.) :::  ::: "Basically, what we do is to try and figure out new ways to hijack the biosynthesis of antibiotics in nature so as to modify their structures with the goal of improving them," Khosla explains. He works with an important class of natural antibiotics called polyketides that have generated dozens of drugs, including erythromycin. Polyketides are secondary metabolites (which give their producers a competitive advantage in their environment) produced mostly by bacteria and fungi and made by a complex and structurally diverse family of enzymes called polyketide synthases (see the primer by David Hopwood in this issue of *PLoS Biology*). Among them are the anthracyclines, a group of anticancer drugs and antibacterials that includes tetracycline. In this issue of *PloS Biology*, Khosla and his colleagues report that they can make selective positional modifications in existing anthracycline antibiotics by starting in a different way with a different starting molecule. The molecule came from a natural anthracycline antibiotic, an estrogen receptor antagonist called R1128. R1128 is made via two modules of enzymes that work sequentially; the first module starts the process, and the second completes it. This division of labor permitted the researchers to tack the first R1128 module onto two other enzyme systems, thus engineering completely new anthracyclines. Some were more active in two types of assays than the natural parent molecule. "One setting was an assay on an estrogen-sensitive cancer cell line. Another setting was an assay to probe activity of an enzyme that\'s of particular interest in Type 2 diabetes, called glucose-6-phosphate translocase." The work also revealed fundamental mechanistic features of the polyketide synthases, Khosla says. The researchers didn\'t study the new anthracyclines\' effects on bacteria, but Khosla notes that the general principle should apply to other classes of compounds, although the details of how it\'s implemented will vary from system to system. He says, "The upshot of this paper is that it is now possible to modify a particular methyl group in just about any anthracycline antibiotic." Finding New Targets {#s4} =================== Instead of searching for new antibiotics by modifying existing ones, some researchers are trying something completely different---first finding the most vulnerable targets in a bacterium and then designing something that hits one or more of them hard. "You have to understand a helluva lot more about how these little cells work. In fact, we think we understand a lot, but I think we can understand almost everything now that we have all the genomes," says Lucy Shapiro of Stanford University School of Medicine. While having full genome sequences---more than 100 microbe sequences have been completed---is essential, Shapiro believes that knocking outs genes galore to find out which ones are necessary and going after them all is not a sensible strategy. She observes, "People have been doing that for a while with absolutely no success. That\'s really going after the problem with a Howitzer instead of with an intelligent approach." So instead of screening libraries of existing compounds, Shapiro prefers using structural information about drug targets or their natural ligands to create new drugs, an approach known as rational drug design. And instead of looking at all essential genes in a bacterium and choosing one to target, she and her colleagues look at genetic circuitry that controls the cell cycle, the pathway that coordinates cell growth and differentiation. They have identified key control points, or nodes, in the circuitry for their favorite study subject, Caulobacter crescentus. Thus, they have found critical genes encoding proteins that control several critical functions in the cell. Their first candidate was an essential enzyme, a methyltransferase called CcrM, that prevents a particular piece of DNA from being expressed in a cell by tagging it with a methyl group. Antibiotic discovery is all chemistry, Shapiro says, which is why she joined with biochemist Stephen J. Benkovic of Pennsylvania State University. They didn\'t know the structure of CcrM, Benkovic explains, but the literature about other methyltransferases suggested that the adenine molecule, which is the substrate for CcrM within DNA, binds to a specific region of the enzyme. The researchers designed adenine-like molecules that would bind to CcrM and then developed inhibitors. Benkovic says, "We already knew what kind of structure we wanted, and we simply fine-tuned it." They worked their way through 1,000 inhibitor candidates, ending up with a small subset---no more than about 20---that not only inhibited CcrM, but also killed Caulobacter very quickly. And not only inoffensive Caulobacter. The compounds knock out other gram-negative bacteria, such as the pathogens Brucella abortus and Francisella tularensis. Some even killed off anthrax, a big surprise because it is gram-positive and so has much thicker cell walls than gram-negative bacteria. The researchers undertook an exhaustive series of experiments to identify which gram-positive bacteria would be affected by which compounds. The list of sensitive pathogens now includes multidrug-resistant Streptococcus, Staphylococcus, and Mycobacterium tuberculosis. More recently, Shapiro reports, they have demonstrated efficacy against rats infected with anthrax or multidrug-resistant Staph, although the compounds save only about 60% of the rats at present. She notes, "So we have a long way to go. But this has proven that if you go after something using some rational approach instead of hit-and-miss, you\'ll probably have more success than by the other method." Benkovic points out that theirs is an entirely new class of compounds, small molecular weight compounds that can be made in a few steps. He says, "They don\'t look like the normal antibiotic, so that\'s why I think they\'re fairly unique." The basic research was done under a grant from the Defense Advanced Research Projects Agency (DARPA), the United States Department of Defense\'s (DOD) central research and development organization, and once the researchers realized they wanted to develop drugs against three agents that have been considered bioterrorism threats --- Brucella, tularensis, and anthrax --- they established a separate operation, Anacor Pharmaceuticals, which is developing them with DOD funding and without Shapiro. In her Stanford lab, she continues her fundamental research to define the complete genetic circuitry of Caulobacter, hoping to identify additional nodes in the circuit. She says, "I am not doing it to develop antibiotics; that\'s what comes out of the work. My goal is to understand how the cell works. I think a lot of studies in pathogenesis should not be just to understand pathogenic organisms, but to understand the complete network of regulatory mechanisms that controls the bacterial cell." Phage Therapy {#s5} ============= The most radical approach to new antibiotics may be the resurrection of an old idea: bacteriophage therapy (see [Figure 2](#pbio-0020053-g002){ref-type="fig"}). Late in the 19th century, a researcher noticed that water from some of India\'s sacred rivers combated cholera. Some years later, the active agents were identified as viruses that infected bacteria. Such viruses are called bacteriophage, or phage for short. There were reports of phage success against dysentery, typhoid, and plague, and bacteriophage therapy had a brief heyday, especially in the 1920s. Results on other diseases were mixed, and with the appearance of antibiotics, phage therapy became unfashionable in the United States, although it has continued in Russia and Eastern Europe. ::: {#pbio-0020053-g002 .fig} Figure 2 ::: {.caption} ###### Phage Negative stain electron micrograph of the gamma phage from which the PlyG lytic enzyme was cloned for use to control B. anthracis. (Photograph courtesy of Vincent Fischetti and Raymond Schuch, The Rockefeller University.) :::  ::: Phage were the model organisms of choice for genetics research in the 1930s and 1940s, but became less fashionable as research tools when investigators moved on to eukaryotes. A few held on, like Ry Young of Texas A&M University, who has made phage-induced cell lysis his life\'s work. "The cell is basically genetically dead as soon as the phage goes in there, but it will keep living as sort of an infected zombie for as long as the phage wants it to, with virus particles accumulating inside the cell," he explains. "Only when the phage is ready and has decided that it\'s the right time will it pull the trigger. And the cell blows up." The freed phage then spew forth to infect new cells. Antibiotic resistance has led to new interest in phage therapy by several small biotech companies. Young continues basic research at Texas A&M, but has also joined one of them, GangaGen, providing bacteriophage expertise to its labs. Phage do kill pathogenic bacteria effectively, and they do it without penetrating human cells, which they can\'t even recognize. So what is keeping phage therapy out of the clinic? Problems that some doubt can be overcome. Because bacteria develop resistance to phage rapidly, phage therapy companies will need to direct cocktails against a single pathogen, according to Vincent Fischetti at The Rockefeller University. Phage are also antigenic, and the antibodies they stimulate will neutralize their effects during subsequent treatment, he says. But the chief problem appears to be regulatory---regulatory in the political, rather than the genetic, sense. When bacteriophage package their DNA, they occasionally include varying amounts of their hosts\' DNA, too. This miscellany, Fischetti points out, is likely to make the Food and Drug Administration unhappy. "Phage normally are very fragile, their tails break, so lot-to-lot homogeneity could be a problem too," he adds. "So even though it will work, I think they\'ll have an uphill battle." Phage may well enter agricultural or veterinary use, he predicts, but are probably not going to be available to patients in the United States any time soon. Fischetti chose a different approach to phage therapy. It does not rely on phage themselves, but on enzymes that phage produce to smash their way out of their host bacteria so they can infect new hosts. He and his colleagues employ these enzymes externally to kill bacteria. He reports, "We now have enzymes that will kill Strep pyogenes, pneumococci, Strep pneumoniae, Bacillus anthracis, Enterococcus faecalis, and group B Strep. The beauty of these enzymes is that they are targeted killing. You only kill the organism you intend to kill, without destroying or affecting the surrounding organisms that are necessary for health." The enzymes can be loaded into a nasal spray that wipes out pathogens such as Pneumococcus, Staphylococcus, and group A Strep on contact with mucous membranes. The strategy might prevent bacterial infections from spreading in close quarters like hospitals, nursing homes, and daycare centers. Fischetti says, "Clinical trials would tell us how often we had to treat, but more important, we\'d have a reagent that could treat people who walk out the door of the hospital to eliminate or reduce the transmission of resistant organisms into the community. We don\'t have that capability right now." Fischetti and his colleagues have moved on to using the enzymes systemically to wipe out Bacillus anthracis spores, preventing them from germinating and seething through the bloodstream, producing deadly toxins. An IV drip would be started after exposure to the spores. The method, Fischetti reports, is already successful in mice; clinical trials will determine how long treatment must be continued, perhaps a week or so. They have also eliminated septicemia from pneumocci with the same intravenous method. Up to now the enzymes must make contact with bacteria to kill, but Fischetti is hoping that a new generation of engineered enzymes will be able to kill pathogens inside cells too. A second disadvantage is that they are effective only against gram-positive bacteria, although that group includes many vicious pathogens. But phage enzymes seem to offer one very big advantage: resistance to them has yet to develop. Fischetti says, "We\'ve tried very hard to identify resistant bacteria, but so far we haven\'t found resistant organisms in all three of the enzymes we\'re working with. It appears to be a very rare event, much rarer than resistance to antibiotics." Fischetti cautions against expecting that gladsome state to last forever, but he points out that even if widespread resistance takes the same 40 or 50 years that antibiotics required to become significantly resistant, phage enzymes could buy researchers decades for inventing other approaches. Antibiotics in the 21st Century {#s6} =============================== There is no shortage of ideas for unearthing new antibiotic candidates. Why are they so slow to enter medical practice? The bottleneck, researchers agree, lies in the development process of turning them into effective therapies. Several researchers blame the big pharmaceutical companies that got so big by leading the way to new drugs for battling infectious disease, but in recent years have dropped out. Fischetti complains, "These are the big companies that have the money to develop antiinfectives, but they leave it to small biotech companies, and it\'s not going to happen as rapidly as it should. I think it\'s really unconscionable for these big companies to drop the ball because it\'s not going to be a billion-dollar market for them and that\'s what they\'re looking for." Half a billion at least, says Francis Tally, a big pharmaceuticals veteran who is now chief scientific officer at Cubist Pharmaceuticals, a biotech company located in Lexington, Massachusetts. According to Tally, Cubist produced daptomycin, approved in September 2003, by licensing it from Eli Lilly, which shelved the new compound after concluding its potential market was only \$250 million. But, Tally argues, the size of the market is not the only barrier to new antibiotics. Combinatorial chemistry and the genomics revolution have simply not delivered on their early promise. "The pipeline is very dry," he says. "There\'s been a real lag at the basic research level." "Antibiotic discovery is hard," Shapiro says. "It\'s a huge long process to get a decent antibiotic." Walsh agrees. "It\'s easier to find inhibitors of particular enzymes for particular processes---and a very long road to convert that into something for development." In the meantime, there is a rising clamor to slow down the rate at which bacteria develop resistance. Doctors are exhorted to cut back on prescribing antibiotics and decline to prescribe for viral diseases, which antibiotics can\'t combat, even when their patients badger them. But even if antibiotic consumption slowed, we will still need new antibiotics. "I always say it\'s not a matter of if, it\'s only a matter of when," says Walsh. "There will always be a need for new antibiotics because the clock starts ticking on the useful lifetime of any antibiotic once you start to use it. That cannot be argued." Tabitha M. Powledge is a science writer based near Washington, DC. E-mail: <tam@nasw.org> DARPA : Defense Advanced Research Projects Agency DOD : Department of Defense TB : tuberculosis

PubMed Central

2024-06-05T03:55:46.106200

2004-2-17

PMC340956

From Bob Clark\'s snug office in Boise, Idaho, where he manages the United States government\'s Joint Fire Science Program (JFSP), he figures his computer provides fingertip reach to just about everybody who\'s anybody in wildfire research. This points to a primary need of nations worldwide in combating the scourge of recurrent wildfires: tools and technology suited to the job. It\'s no small order in places as economically, socially, and ecologically varied as, say, Brazil, South Africa, Australia, Indonesia, and the United States, which are among the countries where wildfire creates the greatest havoc. More than 750,000 acres (303,500 hectares) were burned in southern California alone during last year\'s wildfires. The 2000 season was one of the country\'s worst on record, destroying 8.4 million acres (3.4 million hectares), more than double the decade\'s 10-year average. Australia\'s summer months around the turn of 2002--2003 brought perhaps the worst drought in a century to the populous southeast and the biggest fire season for two decades. Mountain forests were extensively burned and more than 500 houses were lost. In 2002, Brazil suffered 217,000 wildfires, a number that is almost certainly too low because remote imaging cannot detect many fires under the forest canopy. In Indonesia, wildfires that burned for months during 1997--1998 were later estimated to have released the equivalent to 13%--40% of annual global carbon emissions from fossil fuels, inflicting smoke-related ailments on thousands. Where wildfire is concerned, the many differences between such countries can perhaps be pinned down to two essentials. The first is whether a blaze occurs in temperate or tropical forest, and the second is whether the nation is developed or developing. "The science can be rock solid, but it can only go so far before social, economic, and political pressures take over," Clark says. "That\'s what a forest service manager\'s job is, picking the best option based on all those considerations." Unfortunately, having science-based options that are applicable to local conditions is largely a luxury for developed countries. Managers there can choose to let a fire burn under hopefully contained conditions, a policy known in the United States as "wildland fire use." They can set experimental crown fires to study their effects, as was done recently in\"journal\" Canada ([Figure 1](#pbio-0020054-g001){ref-type="fig"}). And they can take preemptive measures, such as reducing fuel in the forest to lower fire hazard. ::: {#pbio-0020054-g001 .fig} Figure 1 ::: {.caption} ###### Northwest Crown Fire Experiment (Photograph used by permission of the USDA Forest Service.) :::  ::: The two main fuel-reduction methods are mechanical removal of combustible materials and controlled or "prescribed" burning ([Figure 2](#pbio-0020054-g002){ref-type="fig"}). During Bill Clinton\'s administration, prescribed burns were encouraged in protected areas, but thinning was allowed only for trees with trunks of nine inches (22.8 cm) in diameter or less. Under George W. Bush, prescribed burning remains a choice, but the United States Department of Agriculture\'s (USDA) Forest Service policy is much more focused on mechanical means. The argument runs that there\'s been too much concern about removing trees, when what counts most is the enhanced fire-resistance of the thinned habitat. ::: {#pbio-0020054-g002 .fig} Figure 2 ::: {.caption} ###### Prescribed Burns in the Intermountain Region of the United States (Photograph used by permission of the USDA Forest Service.) :::  ::: Fire hazard reduction methods must be tailored to an understanding of fuel characteristics in a given area, says David Peterson of the Forest Service\'s Pacific Wildland Fire Sciences Laboratory in Seattle, Washington. "There\'s no uniform way of doing it, partly because, as scientists, we haven\'t given the management folks any quantitative guidelines." Working with other ecologists, social scientists, and economists, he\'s currently producing just such guidelines for the dry interior forests of the Pacific Northwest. "One thing we don\'t want to do is take choices out of the hands of field managers working at the local level." Forecasting Tools: Models and Simulations {#s2} ========================================= For those choices to be meaningful, managers need reliable information on the risk of wildfire outbreaks and on the future behavior of existing fires. This requires models and simulations that incorporate climatic conditions, particularly wind ([Figure 3](#pbio-0020054-g003){ref-type="fig"}). At the Forest Service\'s Fire Sciences Laboratory in Missoula, Montana, researchers have created a "gridded wind" tool based on the engineering discipline of computational fluid dynamics. The program maps wind speed and direction using a digital elevation model, which is a grid of elevation points every 30--100 feet (9--30.5 meters) over a terrain 10--40 square miles (25.9--103.6 square kilometers) in size. This map forms the floor of a box extending up to five miles (eight kilometers) high, which is subdivided into a million or more cubes. Wind flow from either real observations or estimates can be entered into the software, and the layer of cubes nearest the grid floor is used to create surface wind maps at resolutions of every 100 meters (109 yards) or less. In contrast, the usual resolution of weather forecasts is 12 kilometers (7.5 miles), down to 4 kilometers (2.5 miles) in some urban areas. ::: {#pbio-0020054-g003 .fig} Figure 3 ::: {.caption} ###### Shaded Surface Images of Areas in Northwestern Montana That Suffered Wildfires during 2003 The winds are from the west at 20 mph/32 kph. The white lines represent fire perimeters. (Image used by permission of the USDA Forest Service, Fire Sciences Laboratory.) :::  ::: "Two or three years ago, we couldn\'t have done this simulation on a single-processor laptop," says one of its developers, physical engineer Bret Butler. "It would have taken two or three days. Now we can do it in a matter of hours." The ability of these maps to show varying wind flow in valleys, at midslope and on ridgetops, is just the beginning. The next step is to feed these data into models that predict wildfire spread. Butler and colleagues have coupled their gridded wind technology to a fire growth model and tested it against the actual spread of several wildfires, including in Southern California last summer. Maps of actual and predicted surface winds showed strong similarities, encouraging Butler to foresee an ideal scenario in which fire fighting teams enter wind flow data online or by telephone to a central base where gridded wind maps and fire growth simulations are generated within hours, before operational decisions are made. Yet he admits that challenges remain, including the current inability of fire behavior simulation to account for diurnal winds in addition to cold front-driven flow. In mountainous terrain, for example, winds often move up-canyon in the morning and down-canyon in the evening. Moreover, the effect of vegetation on wind is not yet included in such models. Those issues and others are being addressed by researchers working on improvements to the regional weather forecasts of so-called mesoscale models. At the Forest Service\'s Rocky Mountain Research Center in Fort Collins, Colorado, meteorologist Karl Zeller and colleagues are contributing calculations of biological processes to mesoscale weather models. Their algorithms not only can account for diurnal winds but can predict the effects on local weather when vegetation takes in carbon dioxide and releases water vapor. This process can produce different fluxes of carbon dioxide drawn into the canopy and water vapor coming out, depending largely on the type of vegetation and its canopy density. Zeller\'s group has analyzed current mesoscale forecasts in the Rocky Mountains and found that in the daytime, they often are too hot in the high country and too cold in the plains. Water vapor estimates are too low in the mountains and too high in the plains, which Zeller thinks is because the models feed off soil moisture estimates, not off vegetation. In coupling his team\'s new biophysical interface to gridded wind and mesoscale forecast models, Zeller says "point forecasts" are being developed that can focus on a prescribed burn area or even a single house. Wildfire and Species Diversity {#s3} ============================== Fitting the appropriate mix of strategies to a given situation is an issue that has also received close attention in Australia. After the bushfires of 2002--2003, media commentators called for increased "hazard reduction burning" in national parks, prompting ecologists around the country to distribute a joint statement declaring that such a strategy would not further reduce bushfire risk, but would actually threaten biodiversity. Australian species are often well-adapted to fire, and researchers have learned that different fire regimes---meaning the type of fire, its intensity, severity, extent, season, and frequency---favor different species ([Box 1](#box1){ref-type="boxed-text"}). In the southeast of Australia, prescribed burns of high frequency and low intensity can alter the habitat in ways that therefore threaten survival of numerous plant and animal species. "A generic problem or conundrum seems to be that species which do not prosper under relatively frequent fires can be found in most fire-prone environments," notes Ross Bradstock, principal research scientist in the New South Wales Department of Environment and Conservation. He says it\'s very difficult to determine how human interventions in various habitats can foster the coexistence of species that have different fire regime requirements. Fire Suppression and Tropical Forests {#s4} ===================================== As tough as such questions are to answer in developed countries, they pale compared to the problems of tropical forest wildfire researchers and managers in developing countries. In these countries, a destructive cycle of human behavior begins with land-clearing and burning for farming, logging, mining, road-building, and other uses that open gaps in the rainforest\'s canopy cover. This lets in sunlight and air, reducing the forest\'s ability to smother fire by trapping moisture, and it encourages the growth of smaller, more fire-prone plants. The first wildfires that occur are bad, but successive ones can eventually transform tropical forest to scrub savanna ([Figure 4](#pbio-0020054-g004){ref-type="fig"}). Of course, the remaining forest is thereby broken into fragments that continue to suffer incursions at their edges, as the cycle continues. ::: {#pbio-0020054-g004 .fig} Figure 4 ::: {.caption} ###### Forest Regeneration \(A) Dense understory regeneration three years after a low-intensity fire. \(B) The almost total loss of live, above-ground biomass six months after a forest burnt for the third time in 15 years. (Photographs by Jos Barlow; used by permission.) :::  ::: In a recent paper in Science, Michigan State University Amazon expert Mark Cochrane pointed out that prescribed burning is ineffective in tropical forests, because the collateral damage outweighs any benefits. Indeed, tools and technologies employed in temperate conditions can seldom be applied usefully to tropical forests without significant alterations. "One of the main issues in fire science is that the U.S. has no capacity to develop new tools," charges Ernesto Alvarado, a research scientist at the University of Washington in Seattle. He\'s been working for several years with United States Forest Service and Brazilian scientists on field studies in Mata Grosso, the southernmost state of the Brazilian Amazon. He says that fire prediction simulations developed decades ago have not yet been replaced by ones that account for tropical wildfire extremes, including either large-scale crown fires or surface fires, which often reach only 10 centimeters (3.9 inches) in height and move slowly but can burn for weeks and kill many trees. Fire behavior models don\'t work for tropical surface fires because the physics are different from those in temperate forests, he explains. A slow wind generated from the unburned forest blows toward the fire, forcing the small flames to advance against, rather than with, the wind. Another difference is that the fuel is mostly leaf litter, not conifer needles or sticks. Alvarado and colleagues light experimental fires in clear-cuts to determine factors limiting ignition and spread. Such experimental work is rare in tropical forests, where observation and description still predominate. But the team also monitors surface wildfires, measuring fire length, spread, and heat release. "We\'re trying to find applications that people can use to control fires or to explain implications of fire policy," he says. Most wildfires originate from deliberately set burns. For example, many farmers still clear land by the ancient method of slash-and-burn, in which forest is chopped, left to dry, and then burned. These farmers are now banned by Brazilian federal law from burning at the height of the dry season, mid-July to mid-September. They cut in May, but if the rains come early in September, they can\'t burn after the ban ends and must wait until the next season, with nowhere to grow their crops in the meantime. Alvarado thinks a more flexible burning schedule is a solution. The challenge is to pass on technological understanding to decision-makers. For example, even ranchers in Mata Grosso\'s economic elite usually haven\'t heard of fire management techniques, says Amazonian ecologist Carlos Peres at the University of East Anglia in the United Kingdom. Educational projects from nongovernmental organizations have helped to turn some farmers away from heavy reliance on slash-and-burn techniques, but fire suppression information remains to be distributed on the frontiers. "What we really need are very large areas of primary forest that effectively serve as fire breaks," he says. Conservation plans have been made by the federal government in collaboration with international agencies, but implementation remains a question, particularly given the high level of economic pressure from multinational resource developers eager to enter the Amazon. Major roads through the jungle are also on the drawing board. "Different categories of conservation units can be gazetted on paper, but in practice they\'re a long way from working. Someone draws lines on a map high in an office in Brasilia, but when you go out to that place in the forest, no one knows it\'s a conservation zone." Fire Prevention: Developing the Technology {#s5} ========================================== Information transfer faces similar barriers in much of Southeast Asia, as Canadian forestry researchers discovered during a five-year project (now winding up) to create a computerized early warning tool for wildfire outbreaks. The program was instigated after the 1997--1998 fires created a regional haze hazard, largely because of peat deposits up to 20 meters (21.8 yards) thick that had become susceptible to burning in swampy forests drained and cleared for development. Michael Brady, who managed the Canadian project in Jakarta, points out that headmen in remote communities are still likely to believe that wildfires start spontaneously, by grasses rubbing together or even by magic. A fire scientist whose doctorate is in tropical forest peat dynamics, Brady sees the project as a medium to strengthen regional fire ecology in general. "In some ways, that\'s more important to me than the tool itself." The tool is a variation of the Fire Danger Rating System used in Canada and, with various permutations, in many other countries. The Canadian system has two components, one for indexing fire weather and another predicting fire behavior. The weather component models moisture input and output in fuels generically classed as fine, moderate, and heavy. Brady and Indonesian university scientists grouped grasslands in the fine fuel category, fallen leaves and litter as medium, and peat and woody materials as heavy. They spent three years calibrating these fuels to local weather conditions, examining moisture dynamics and performing ignition tests. In developed countries, fuels are further specified in numerous classes for fire behavior prediction, but that requires decades of field work. Brady\'s team concentrated instead on helping key agencies in seven Southeast Asian countries, especially Indonesia and Malaysia, to obtain and use the appropriate computing tools. Brady doesn\'t expect immediate results in terms of reducing acreage burned. "Canada and the U.S. still have huge fire problems after working on it for a century." But he does hope for a change of thinking, away from a current fascination in the region with satellite imaging of "hot spots" where fires are likely to be occurring. Fire danger rating concentrates on where fires are most likely to begin. "It allows you to add prevention into your management program." Beyond Prevention {#s6} ================= In South Africa, "retention" is a conservation buzzword referring to strategies that, in a sense, go beyond prevention of problems. What ecologists hope to retain is biodiversity in the midst of changes that can\'t be stopped, and their methods are producing major repercussions throughout government. The work is centered on the Cape Floristic Region of Africa\'s southwestern tip. Almost 90,000 square kilometers (34,750 square miles) in area, it\'s the world\'s smallest floral kingdom. A conservation plan was launched in 1998 that has drawn cooperation from tourism, mining, water use, agricultural, and land use planning groups. The project has the ambitious goal of protecting not only the usual biodiversity patterns of conservation areas but also the spatial components of evolutionary processes that enable species to adapt to potentially harmful changes. This entails a complex effort to determine which parts of developed and undeveloped lands are most necessary to such processes, including rivers, sand movement corridors, gradients from uplands to coastal lowlands, and major wilderness areas. University of Port Elizabeth botanist Richard Cowling, one of the scheme\'s principal architects, estimates that it might require 60%--70% of the region\'s landscape. As in Australia, fires are important to the Cape\'s biodiversity, but too-frequent burns are a problem. Cowling thinks that by consolidating mountainous megawilderness under the project\'s plan and protecting spatial transitions between fire-prone areas and those that resist fire, managers could move toward allowance of natural fire regimes. The current problem, he says, is that protected areas usually stop short of the transition to semidesert areas that are privately owned. When fire spreads from public to private land, the government often gets sued. Under the evolving Cape plan, landowners will sit on governing boards, and property that they contract for conservation will be tax-exempt. The Cape plan has attracted millions of dollars in support from the World Bank and other international sources, but Cowling regards that achievement as much less important than the progress made in gaining support from various interest groups. "The key issue is the extent to which you can get biodiversity concerns mainstreamed to other sectors," he says. Threats to habitat retention, not least of which is wildfire, endanger every species. "It\'s about making people realize that biodiversity is the basis upon which all other things will succeed." ::: {.caption} ###### Box 1. Fire-Adapted Species ::: Plants and animals of many countries evolved for millennia with wildfire as a natural occurrence, but when human interventions increase the frequency of fire, species suffer. African fire lilies and Australian "grass trees" are among plants that are stimulated to flower by smoke constituents such as ethylene. Plant seeds in fire-prone landscapes of Australia and South Africa often require fire to stimulate their germination, but it can take more than a decade for new seed banks to mature in some species. If a second fire arrives before then, the species could die out. Animals can be similarly affected. For example, a threatened marsupial called the potoroo is capable of surviving a high-intensity wildfire, but cannot tolerate the habitat changes caused by frequent, low-intensity fires. Likewise, some species of Australian honeyeaters are threatened with extinction because too-frequent fires have changed the proportion of mature and immature nectar plants. On the other hand, ecosystems can also be transformed by fire suppression. In southern Africa, decades of such activity have encouraged forests to replace grasslands. Yet the lovely marsh rose almost disappeared from the Cape before land managers realized that fire suppression was preventing its seeds from germinating. In such ways, biodiversity must find its place among the goals and tradeoffs of human intervention. Steve Bunk is a freelance writer based in Boise, Idaho, United States of America. E-mail: <stevebunk@sbcglobal.net> JFSP : Joint Fire Science Program USDA : United States Department of Agriculture

PubMed Central

2024-06-05T03:55:46.108202

2004-2-17

PMC340957

Faced with new and ongoing threats to public health, researchers are becoming increasingly resourceful in their quest to discover new drugs. Drug researchers have long looked to living organisms for inspiration, either mimicking or extracting chemical formulas from naturally occurring compounds. Bacteria and fungi, for example, produce a wide range of compounds---some of which give them a selective advantage in their own environments---that provide important pharmaceutical activities. One class of these natural compounds are the polyketides, which make up a large portion of the antibiotics (including erythromycin and tetracycline) and antitumor drugs (such as doxorubicin and epothilone) that have been isolated from various microorganisms. Polyketides are synthesized by bacteria and fungi by the appropriately named polyketide synthases (PKSs). PKSs can be thought of as large molecular factories containing a series of enzymes working on an assembly line: each enzyme in the line adds molecules to a primer, or starter, unit---which is usually an acetate molecule---and then hands off the growing chain to the next enzyme. The specific enzymes set all the characteristics of the polyketide, including the chain length, the building blocks used, and the branching pattern of the molecules. Although microorganisms generate polyketides with a variety of characteristics, one goal of drug discovery research is to increase this diversity even further---a larger pool of polyketides promises more drugs with enhanced pharmaceutical applications. Early attempts at creating artificial polyketides focused on altering the functional characteristics of naturally occurring polyketides---the length of the chain, the building blocks, and the patterns of the branches. Chaitan Khosla and colleagues have taken this approach one very large step further. Rather than changing the machinery to modify the growing structure of a polyketide, they engineered bacteria to use an alternative, nonacetate primer molecule. This has important practical implications because some medicinally significant compounds do not use the usual acetate primer unit. By dissecting out the specificities of the "starter" and longer, multiunit "elongation" PKS enzymes and by mixing and matching modules, they have produced novel polyketide analogs (in this case, of anthraquinone) with more effective medically relevant properties. One of the compounds they engineered shows enhanced efficacy in blocking the growth of breast cancer cells that depend on the activity of the estrogen receptor, while a second polyketide inhibits an enzyme linked to adult-onset diabetes, demonstrating just two possible new therapeutic applications for synthesized polyketides. But, as the authors propose, this method promises to reveal new pharmaceutical agents that haven\'t even been discovered yet. ::: {#pbio-0020056-g001 .fig} ::: {.caption} ###### A synthetic polyketide :::  :::

PubMed Central

2024-06-05T03:55:46.110682

2004-2-17

PMC340958

Late-stage cancers are notoriously unresponsive to treatment, making certain hard-to-detect cancers particularly insidious. Ovarian cancer, for example, most often escapes diagnosis until the tumor has already metastasized. At this stage, ovarian cancer is classified as peritoneal carcinomatosis, a terminal condition characterized by widespread tumor growth throughout the peritoneum, the large serous membrane that lines the abdominal cavity, pelvis, and associated organs. Advanced cases of peritoneal carcinomatosis are largely resistant to chemotherapy and account for the bleak 15%--20% survival rate of ovarian cancer. Biologists often view cancer as an evolutionary process in which cells that would normally cooperate with their neighbors begin to compete with them. Selective advantage for cancer cells often begins with mutations that inhibit tumor suppressor pathways. p53, like other tumor suppressor genes, arrests cell growth and induces apoptosis (programmed cell death) in response to cellular stress, such as chromosomal damage. Cells with p53 mutations often escape these constraints, leading to the uncontrolled growth characteristic of "immortal" cancer cells. Nearly all types of tumors have mutations in the p53 pathway. Treatments focused on restoring p53 function---which is likely to be defective only in cancer cells---should prove more effective than chemotherapies, which indiscriminately kill all dividing cells, healthy or cancerous. With the goal of developing targeted therapeutic strategies, Steven Dowdy and colleagues show that restoring p53 protein function in tumor cells not only dramatically increases lifespan in mice but also eliminates disease. While past efforts to restore tumor suppressor function in cancer cells have focused on gene therapy, Dowdy and colleagues introduced modified p53 peptides, or protein fragments, into cancer cells. p53 works as a "transcriptional" activator that binds to specific sequences of DNA and triggers apoptosis in response to DNA damage. Its biological function flows from this binding ability. One region of this large protein, called the C-terminal domain, facilitates effective binding. In cancer cells, synthesized peptides (called p53C′) derived from this region can induce apoptosis by activating p53---which is normally present in low levels in a biologically inactive form---and restoring function to p53 proteins with DNA-binding mutations. To get p53C′ peptides into cancer cells, Eric Snyder et al. used a technique pioneered by Dowdy that delivers large proteins into the cell interior. Since the cell membrane normally limits passage to only small molecules (larger molecules generally enter through surface receptors), this is no small feat. The technique exploits the ability of a small peptide region from the HIV TAT protein to smuggle macromolecules through cell membranes that normally prohibit entry to such large molecules. After synthesizing a structurally modified form of p53C′ less prone to degradation, the researchers first confirmed that the peptide was functional and then that it activates p53-specific genes in tumor cells, but not in normal cells. Testing the effectiveness of the peptide therapy on mouse strains used to model human metastatic disease, they found that mice treated with the TATp53C′ peptide showed a significant reduction in tumor growth and lived six times longer than both mice treated with a control peptide and untreated mice, with some mice remaining disease-free more than 200 days after treatment. This macromolecular delivery approach, Snyder et al. argue, works with greater specificity and avoids the tumor-generated neutralizing effects observed in small molecule strategies. Because a mutation in the p53 gene is one of the most common events in the development of cancer, these results could have implications for a wide variety of cancers. And by working with mouse models that approximate the physiological burdens metastatic cancer imposes on humans, Dowdy\'s team demonstrates the promise of developing targeted "intracellular biologic" therapeutics that treat the systemic pathology of cancer---inhibiting tumor growth as well as alleviating the lethal complications associated with the disease. ::: {#pbio-0020058-g001 .fig} ::: {.caption} ###### TATp53C′ treatment extends survival of mice :::  :::

PubMed Central

2024-06-05T03:55:46.111404

2004-2-17

PMC340959

Thanks to the tools of molecular biology, our understanding of the 100-plus diseases known collectively as cancer has increased dramatically over the past decades. While each of these cancers exhibits unique characteristics reflecting the particular cell or tissue it springs from, the disease follows a similar arc in nearly all its forms. Cancer is a multistep disease that begins when genetic damage---initiated by a multitude of agents---unleashes a single cell from the normal constraints on cellular proliferation. This single transformed cell generates a colony of similarly abnormal progeny that can take decades to develop into malignancies. While events that stimulate uncontrolled cell division can promote cancer, mutations in tumor suppressor genes figure prominently in tumor progression. Disruptions in the pRb (retinoblastoma 1) tumor suppressor, for example, are often seen early in cancer development, sensitizing cells to tumorigenesis. pRb, along with other "pocket proteins"---so-called because they share an amino acid domain called the Rb pocket---regulate cell cycle progression, apoptosis (programmed cell death), and cellular differentiation. Some tumor suppressors, such as p53, can trigger apoptosis, ultimately sacrificing cells that have sustained DNA damage or other types of cellular stress. Mutations in both the pRb and p53 tumor suppressor pathways are commonly seen in human cancers, though their interactions appear to vary depending on the tissue. In mouse brain epithelial cells, for example, loss of p53 function coupled with loss of pRb results in reduced apoptosis and increased tumor growth, while p53 loss in mouse brain astrocytes (cells that support neurons) does not affect tumor growth. Building on this work, Terry Van Dyke and colleagues report that loss of the pRb tumor suppressor in mammary tissue has the same effect---predisposition to tumor formation---seen in these other cell types. Despite the different environment inherent in each cell type, the initial events following loss of the pRb pathway were the same: increased proliferation and apoptosis, followed by tumorigenesis. But, surprisingly, pRb and p53 interactions varied in different cell types. Like most cancers, mammary gland cancer has a long latency period, prompting the researchers to ask what events engineer tumor progression. To investigate the relative contribution of pRb and p53 in tumorigenesis, the researchers generated a novel mouse model with a dysfunctional pRb pathway and various levels of p53 function in several cell types. This is a significant achievement in itself, as many agents that inactivate the pRb pathway also disrupt the p53 pathway. pRb inactivation, they show, causes abnormalities in mammary cell proliferation, apoptosis, and tissue morphology. In these mammary-specific pRb-deficient mice, p53 was responsible for most of the apoptotic response---decreased levels of p53 resulted in reduced apoptosis and accelerated tumorigenesis, but had no effect on proliferation. Interestingly, in other mouse models where aberrant proliferation is caused by disabling other pathways, loss of p53 was associated with increased proliferation---rather than reduced apoptosis---and early tumor formation. And while p53 is the main effector of apoptosis in brain and mammary epithelial cells, this is not the case in all tissues: in astrocytes, for example, the tumor suppressor Pten regulates apoptosis in response to pRb inactivation. Together these results indicate that specific cellular responses to a cancer-causing stimulus vary depending on the nature of the initial genetic injury and the cell type and that pRb and p53 interact in different ways in different tissues. And p53, it appears, contributes to tumor suppression---and thus progression---through multiple mechanisms. By creating a mouse model that disentangles the pRb and p53 pathways, Van Dyke and colleagues have added a valuable resource for studying breast cancer. This model, they propose, will facilitate further investigations into the relative contributions of these overlapping pathways to cancer progression. What\'s more, the model offers a vehicle for examining how pRb interacts with other breast cancer mutations, like the inherited mutations in the human *BRCA1* and *BRCA2* genes, to shed light on the complex series of events that ultimately cause breast cancer. ::: {#pbio-0020059-g001 .fig} ::: {.caption} ###### Transgene expression is associated with increased cell proliferation and cell death (apoptosis) :::  :::

PubMed Central

2024-06-05T03:55:46.112086

2004-2-17

PMC340960

If you\'ve ever hit a patch of ice on the road that sent your car swerving left while you resolutely---and futilely---steered right to get back in your lane, you\'ve experienced what neuroscientists call a "visuomotor rotation task." On a dry road, your response would have been appropriate. But under icy conditions, the same sensory cue produces a decidedly negative result: a car fishtailing out of control. While you\'re figuring out what movements will straighten out the car, the neurons in your primary motor cortex---the region of the brain responsible for movement---are taking notes. Chances are, your next icy encounter was less dramatic. But how does your brain learn to produce a different movement in response to the same visual cue? Neuroscientists investigate such questions by recording and analyzing the electrical activity of neurons during learning and performance of new sensory-motor transformations. Such studies, for example, show that populations of neurons in different brain areas map sensory cues and desired arm motion by creating an internal representation of the corresponding sensory and motor coordinates in a way that allows flexible responses to changing conditions. In previous studies, Rony Paz and Eilon Vaadia, of The Hebrew University in Israel, found that neurons in the primary motor cortex that fire before monkeys move their arm in a particular direction have higher firing rates after the monkey learns to dissociate the arm direction from the cursor direction (an indicator of visual feedback). Interestingly, changes in activity preferentially occurred in a subset of neurons that were already tuned (that is, maximally activated during movement) to the direction experienced while learning. While many studies indicate that learning new tasks can generate specific changes in brain activity, it had not been clear how or if such changes improve the internal representation inside the brain. Specifically, is the neuronal code any "better" after learning? Now Paz and Vaadia show that while these neurons are firing at higher rates they are also transmitting more information about specific task parameters. Paz and Vaadia trained two rhesus monkeys to learn various visual-motor tasks---which involved operating a joystick to move a cursor on a screen---and then changed the relationship between the visual feedback (the cursor) and hand movement. Using information-theory analysis---which measures the amount of information that single neurons can tell about the movement---they were able to correlate neuron activity with direction of movement and, conversely, distinguish differences between directions based on neuron activity. Their analysis revealed that the neurons transmit more information about the direction of movement after the monkeys learn a task. To figure out what aspect of neuron activity conveys this improved information, Paz and Vaadia examined two features of neuron signaling---response variability and directional sensitivity---which they reasoned might plausibly accomplish this. Increased information content after learning a task, they found, corresponded to sensitivity to a single direction, and neurons attuned to that direction contributed to the increase. These findings suggest that subsets of directionally sensitive neurons increase their firing rates to more finely tune their sensitivity to that direction. By successfully reconstructing the movement direction from the neuron signals captured after learning a task, Paz and Vaadia also demonstrate that the observed learning improvement can be extracted to predict behavior. The authors argue that their results suggest a close association between properties of neurons---such as directional tuning of cells---and learning a skill that is focused on the same parameter---in this case, direction. Together with results from visual and auditory areas, they propose that similar mechanisms may control the interplay between neurons and learning throughout the central nervous system. ::: {#pbio-0020060-g001 .fig} ::: {.caption} ###### Mutual information between neuronal activity and direction of movement :::  :::

PubMed Central

2024-06-05T03:55:46.112694

2004-2-17

PMC340961

Most animals (including humans) show a high level of bilateral symmetry: on the surface, the right side of our body resembles the left. A closer and deeper look, however, reveals an underlying asymmetry. The heart, for example, is on the left side in most humans, and the liver on the right. This left-right asymmetry develops early on in the embryo, and research in the past few years has revealed some of the molecular and cellular mechanisms that establish the left-right axis, which conveys positional information to cells in the growing embryo. We know that the formation of the axis relies on "crosstalk" between cells, which involves long-range signaling molecules (or ligands) and cell-surface receptors on cells that receive the signal. The molecules involved in the formation of the left-right axis during embryogenesis, along with their functions, are conserved among vertebrates. They include members of the Transforming Growth Factor beta (TGF-ß) family---such as the agonists (or ligands) Nodal, Vg1/GDF, and activin, and the antagonist (a molecule that interferes with agonist/ligands) Lefty---on the signaling side and members of the EGF-CFC family---such as the activin receptor and its coreceptors---on the receiving side. The EGF-CFC proteins play important roles in early vertebrate embryogenesis; mutations in these genes in the zebrafish (and mouse) result in a range of developmental defects, including problems in left-right axis specification. While ligand-stimulation of the activin receptor by Nodal and Vg1/GDF requires the EGF-CFC coreceptors, activin can activate the activin pathway without a coreceptor. Lefty---being an antagonist---can block activation of the activin receptor, though it is not clear how. Through genetic and biochemical studies in zebrafish and frog embryos, Simon Cheng, Alex Schier, and colleagues have now clarified a piece of this very complex signaling puzzle by demonstrating that Lefty inhibits a subset of TGF-ß signals---Nodal and Vg1/GDF but not activin---by blocking EGF-CFC coreceptors. They went on to show that a short, specific region of the signal molecules---accounting for less than 4% of the entire protein---determines whether the signals activate the activin receptor in an EGF-CFC coreceptor-dependent or -independent fashion and therefore governs susceptibility to Lefty. These findings suggest that subtle sequence differences between related signals can dramatically influence their function. Gene families are thought to arise from gene duplications, and the studies described here illustrate how members of the same gene families can gain diverse roles by specific interactions with coreceptors and antagonists. Additional studies will be necessary to reveal the structural basis for the observed diversity. ::: {#pbio-0020061-g001 .fig} ::: {.caption} ###### Zebrafish embryo heart loops correctly in wild-type, incorrrectly in mutant :::  :::

PubMed Central

2024-06-05T03:55:46.113179

2004-2-17

PMC340962

The primate brain processes a remarkably diverse array of visual cues to recognize objects in dynamic settings crammed with unfamiliar objects. Not surprisingly, repeated viewing aids recognition, but how the brain orchestrates this experience-driven improvement is unclear. Visual input to the brain travels from the eye to the primary visual cortex (V1), at the back of the brain. From there, signals are sent to nearby extrastriate cortical areas, which process "early" visual cues. Both the "lower level" extrastriate cortex and "higher level" inferior temporal (IT) cortex are important for object recognition in primates. In monkeys and humans, lesions in the IT cortex severely affect the ability to recognize objects. In these higher-level cortical regions, neurons carry more information about an object after subjects learn to recognize that object. This modified neural activity is thought to reflect internal representations of specific aspects of the learned task---such as learned recognition of three-dimensional objects---and these representations often remain stable even though certain features of the visual stimulus---such as size or image degradation---change. With recent evidence suggesting that lower level brain regions like the primary visual cortex are also capable of learning-related modifications, it appears that both early and higher brain areas of the "ventral visual stream" benefit from learning. It is not clear, however, how learning modifies these discrete brain regions to coordinate this processing. By training monkeys to recognize degraded images, Gregor Rainer, Han Lee, and Nikos Logothetis of the Max Planck Institute for Biological Cybernetics in Germany have identified a subset of neurons that compensate for indistinct visual inputs by coordinating disparate regions in the brain. The monkeys\' improved performance, they propose, stems from the informational enrichment of a subset of lower level neurons. Along with an increase in learning-induced firing activity, V4 neurons---extrastriate cortical neurons associated with detecting visual input of intermediate complexity---encode more information about relevant details to resolve indeterminate visual cues. V4 neurons likely interact with higher cortical levels to help the monkeys interpret the degraded indeterminate images as something recognizable. The researchers presented the monkeys with different "natural" images, including pictures of birds and humans, then subjected the images to different levels of "stimulus degradation"---making them harder to read by adding varying amounts of visual noise. Using this approach, the researchers could record the activity of the V4 neurons as the monkeys were presented with the different images. The monkeys viewed a sample image and then signaled whether a second image, presented after a brief delay, was a match or not. When Rainer et al. analyzed the activity of the V4 neurons associated with the different images, they found there was no significant change in the activity or information conveyed by V4 neurons associated with novel or undegraded familiar images. On the other hand, learning not only significantly improved the monkeys\' ability to recognize degraded stimuli but also increased both the activity and informational encoding of the V4 neurons. But how did individual V4 neurons facilitate this enhanced ability to recognize degraded stimuli? After identifying a subset of neurons that showed enriched neural activity in response to degraded or indeterminate stimuli, the researchers studied the monkeys\' eye movements to determine any behaviors that might explain why monkeys performed better with familiar degraded stimuli. They mapped the monkeys\' eye movements while allowing them to freely view the different familiar and novel images---but this time with just two coherence levels (undegraded and 45% coherent). There was substantially more overlap, in terms of where the monkeys looked for the 45% and 100% coherent images after learning. This suggests that monkeys learned to focus their attention on particular salient features, and were thus better able to identify degraded versions of these images. Neurons in the V4 area appear to be recruited to distinguish the relevant visual signal from the visual noise, and thus play a critical role in resolving indeterminate stimuli when salient features are present. These results, together with previous studies showing the sensitivity of prefrontal cortex neurons to novel stimuli, indicate that the prefrontal cortex processes novel stimuli while the V4-rich extrastriate visual areas convey details about hard to decipher images. It may be that as the V4 neurons refine their competence through learning, they also support the ability of the prefrontal cortex to process different but similar visual cues. Vision is a dynamic process, Rainer et al. conclude, characterized by ongoing interactions between stimulus-driven brain regions and feedback from higher-order cognitive regions. ::: {#pbio-0020062-g001 .fig} ::: {.caption} ###### Monkeys can learn to recognize degraded images :::  :::

PubMed Central

2024-06-05T03:55:46.113609

2004-2-17

PMC340963

Open access is gaining momentum. Authors are submitting papers in ever-increasing numbers to open-access journals. Several prominent research sponsors, including the Wellcome Trust, the Max Planck Society, the Centre National de la Recherche Scientifique (CNRS), and the Institut National de la Santé et de la Recherche (INSERM), have recently pronounced that open access is the best way for the researchers they support to publish their work. Several established commercial and not-for-profit publishers have announced plans to experiment with open-access models for some or all of their journals. Delighted and encouraged, we gear up for the launch of *PLoS Medicine* this autumn---the next step in our efforts to bring the benefits of open access to the entire scientific and medical community. We aim to make *PLoS Medicine* a premier journal, providing open access to the best medical research to researchers, to physicians and other caregivers, and to the public. The case for open access to medical research is even stronger than it is for basic research in biology. There are more interested parties: patients and their advocates; biotechnology and pharmaceutical companies that develop drugs and medical devices; doctors, nurses, and other healthcare providers; and health policy-makers at the national and international levels. The goal of the medical research enterprise is---or should be---scientifically, ethically, and socially responsible medicine, which means research that will benefit patients worldwide. The reality looks somewhat different. Large investments into basic research have not yet lived up to their full potential to save lives and improve their quality. Doctors, patients, and their advocates do not have ready access to the combined peer-reviewed evidence from medical research. The prices for the latest drugs often put them out of reach of patients in poor countries or poor patients in countries without universal healthcare systems. Moreover, research focuses disproportionately on the potentially lucrative treatments for diseases of wealthy societies, shortchanging the poorer countries, which bear the greatest burden of disease. A medical journal by itself cannot change this reality. But with the help of researchers and practicing physicians around the world who recognize the need and opportunity for change, we seek to create a journal that promotes medical research and practice that is both scientifically rigorous and compassionate. Open access to this literature will strengthen the medical research community by giving *all* stakeholders free and immediate access to the latest medical research, along with new and more powerful search tools and links between the literature and other relevant information. *PLoS Medicine* will be an international, modern, general medical journal, covering all areas in the medical sciences, from basic studies to large clinical trials and cost-effectiveness analyses. We will concentrate on human studies that enhance our understanding of disease epidemiology, etiology, and physiology; the development of prognostic and diagnostic technologies; and trials that test the efficacy of specific interventions and those that compare different treatments. We will publish original research and commentary that promotes translation both of basic research into clinical investigation and of clinical evidence into practice. A truly broad medical journal is an ambitious project, but we want *PLoS Medicine* to promote an integrated understanding of the patient---to make it easy for people to read outside their specialty area. "Doctors are systems biologists," as one medical researcher put it, and inspiration can often be found in unfamiliar territory. Articles published in *PLoS Medicine* will be rigorously peer-reviewed. Academic and professional editors, supported by expert peer-reviewers, will select those studies that drive research forward---in this case, toward medical applications and benefits for patients. This issue of *PLoS Biology* contains two "human" studies that met our criteria for excellence and originality, a paper by Howard Chang and colleagues (found at DOI: [10.1371/journal.pbio.0020007](10.1371/journal.pbio.0020007)) on the microarray analysis of tumors and one by Sarah Rowland-Jones and coworkers (found at DOI: [10.1371/journal.pbio.0020020](10.1371/journal.pbio.0020020)) that examines how HIV exhausts the capacity of the immune system. Similar papers submitted in the future will be published in *PLoS Medicine*, alongside studies that have more direct implications for clinical practice. This issue also contains several articles describing more basic advances with medical implications: a study by Terry van Dyke and colleagues (found at DOI: [10.1371/journal.pbio.0020022](10.1371/journal.pbio.0020022)) describing a new mouse model for breast cancer, a report on a novel approach to drug synthesis by Chaitan Khosla and coworkers (found at DOI: [10.1371/journal.pbio.0020031](10.1371/journal.pbio.0020031)), and an article by Stephen Dowdy et al. (found at DOI: [10.1371/journal.pbio.0020036](10.1371/journal.pbio.0020036)) on targeted modulation of p53 activity. *PLoS Biology* will continue to solicit and publish such articles, but we will bring them---and similar ones published elsewhere---to the attention of the readers of *PLoS Medicine*. Like *PLoS Biology*, *PLoS Medicine* must be a community journal to achieve its goals. If you are a researcher or an individual anywhere in the world who has a stake in medical research and if the goals of *PLoS Medicine* outlined here resonate with you, please contact us. *PLoS Medicine* will accept submissions beginning in April 2004, and we are looking for advocates who will help to spread the word about open access and *PLoS Medicine* in the global medical community and for investigators who will submit excellent research, review submitted articles, and contribute editorials and commentaries. *PLoS Medicine* is and will stay a work in progress, and we want to consult with as many people as possible, both before the launch and as *PLoS Medicine* evolves. Sign up to join the *PLoS Medicine* community at <http://www.plos.org/medicine> and help us to make the best of medical research and practice accessible to a global audience.

PubMed Central

2024-06-05T03:55:46.114392

2004-2-17

PMC340964

When we think of foraging for food, we usually imagine animals wandering in the woods, poking behind bushes and trees, trying to find something tasty. Amazingly, even single-cell bacteria display a simplified version of this behavior. Many species of bacteria can respond to chemical or nutritional cues (chemoattractants) in the environment by moving toward locations with more favorable conditions, a process known as chemotaxis. The bacteria adjust their movements by rotating threadlike projections called flagella either clockwise or counterclockwise; these adjustments are made by a network of proteins in response to chemoattractants. Chemotaxis has been identified in many bacterial species, but two of the best-studied examples are in Escherichia coli and Bacillus subtilis. Computer modeling of chemotaxis in these species now reveals some important differences in the network architecture that controls this complex behavior. Most of the proteins involved in chemotaxis in E. coli and B. subtilis have been identified and well studied, but much remains to be learned about this biological process. As scientists have begun to understand how the proteins work together, they\'re discovering a network of interactions that operates a bit like an electronic circuit. Researchers have found that using the circuit as a model for protein networks has helped them to understand how complex system properties arise from seemingly simple interactions between proteins. These properties can be explored with the aid of computer simulations, whereby researchers can rapidly test a given system under many different situations and can tweak the properties of the proteins and their connections. The team, led by Adam Arkin of the University of California at Berkeley, has compared the system level properties of chemotaxis in the two bacterial species E. coli and B. subtilis. Not surprisingly, the proteins involved in the signaling pathway are conserved---that is, they have changed very little since they first evolved---even though these species are evolutionarily very distant. In many cases, a gene from one species can even substitute for the ortholog (a conserved gene that retains the same function even though two species have diverged) in the other. Despite these similarities, however, disrupting the function of orthologous genes in these two species often has different, even opposite, effects. This is surprising, especially given that the chemotactic behaviors of E. coli and B. subtilis are almost identical. In order to understand this puzzling observation, the researchers constructed a network model of the chemotaxis system in B. subtilis and used simulations to understand how the network properties differ from those of existing models of E. coli chemotaxis. The group found that despite the similarities in proteins and the nearly identical behavior between the two species, the mechanisms underlying the behavior are quite distinctive. When comparing the system properties of these two bacterial systems, the researchers also made an unusual observation. Though the two "circuits" have different wiring, the system properties underlying the behavior, called the control strategy, are very similar. The two species of bacteria therefore achieve the same chemotaxis behavior by using similar proteins, but in different ways. Arkin and colleagues draw two important conclusions from these results. First, these two systems have conserved proteins, but the proteins are wired together differently. This means that wiring of signaling networks cannot be inferred simply by identifying the conserved proteins in the network. Second, in these systems, conserved proteins use different mechanisms to accomplish the same overall control strategy. This raises the question of how such systems evolve. The authors suggest that the control strategy itself may be an evolutionarily conserved property. These conclusions will be important to keep in mind as researchers examine these systems in more detail and begin to examine more complex systems as well. ::: {#pbio-0020066-g001 .fig} ::: {.caption} ###### Modeling chemotaxis :::  :::

PubMed Central

2024-06-05T03:55:46.115121

2004-2-17

PMC350664

Introduction {#s1} ============ The endoribonuclease Dicer produces two types of small regulatory RNAs that regulate gene expression: small interfering RNAs (siRNAs) and microRNAs (miRNAs) ([@pbio-0020098-Bernstein1]; [@pbio-0020098-Grishok3]; [@pbio-0020098-Hutvagner2]; [@pbio-0020098-Ketting3]; [@pbio-0020098-Knight1]). In animals, siRNAs direct target mRNA cleavage ([@pbio-0020098-Elbashir2]b, [@pbio-0020098-Elbashir2]c), whereas miRNAs block target mRNA translation ([@pbio-0020098-Lee2]; [@pbio-0020098-Reinhart1]; [@pbio-0020098-Brennecke2]; [@pbio-0020098-Xu1]). Recent data suggest that both siRNAs and miRNAs incorporate into similar, perhaps even identical, protein complexes and that a critical determinant of mRNA destruction versus translation regulation is the degree of sequence complementary between the small RNA and its mRNA target ([@pbio-0020098-Hutvagner1]; [@pbio-0020098-Mourelatos1]; [@pbio-0020098-Zeng2]; [@pbio-0020098-Doench1]; [@pbio-0020098-Saxena1]; [@pbio-0020098-Zeng3]). Target RNA cleavage directed by siRNA is called RNA interference (RNAi). RNAi is a powerful method for the study of gene function in animals and plants and has even been proposed as a therapy for treating genetic disorders and viral infections. Biochemical studies in *Drosophila* S2 cells ([@pbio-0020098-Bernstein1]; [@pbio-0020098-Hammond1]a; [@pbio-0020098-Caudy1]; [@pbio-0020098-Liu1]) and affinity purification ([@pbio-0020098-Martinez1]) or immunoprecipitation ([@pbio-0020098-Hutvagner1]) from cultured human HeLa cells have identified protein components of the RNAi effector complex, the RNA-induced silencing complex (RISC). Genetic mutations that disrupt RNAi in Caenorhabditis elegans, *Drosophila*, green algae, fungi, and plants have likewise identified proteins required for RNAi ([@pbio-0020098-Cogoni1], [@pbio-0020098-Cogoni2]a, [@pbio-0020098-Cogoni2]b; [@pbio-0020098-Ketting2]; [@pbio-0020098-Tabara1], [@pbio-0020098-Tabara2]; [@pbio-0020098-Catalanotto1], [@pbio-0020098-Catalanotto2]; [@pbio-0020098-Dalmay1], [@pbio-0020098-Dalmay1]; [@pbio-0020098-Fagard1]; [@pbio-0020098-Grishok2]; [@pbio-0020098-Ketting1]; [@pbio-0020098-Mourrain1]; [@pbio-0020098-Wu-Scharf1]; [@pbio-0020098-Grishok1]; [@pbio-0020098-Tijsterman1]a, [@pbio-0020098-Tijsterman1]b). Key steps in the RNAi pathway have also emerged from studies of RNAi reconstituted in cell-free extracts ([@pbio-0020098-Tuschl1]; [@pbio-0020098-Zamore1]; [@pbio-0020098-Hammond2]b; [@pbio-0020098-Nykanen1]; [@pbio-0020098-Martinez1]; [@pbio-0020098-Schwarz1]; [@pbio-0020098-Tang1]). Recently, hundreds of miRNAs have been identified in animals and plants ([@pbio-0020098-Lagos-Quintana1], [@pbio-0020098-Lagos-Quintana2]; [@pbio-0020098-Lau1]; [@pbio-0020098-Lee1]; [@pbio-0020098-Reinhart2]; [@pbio-0020098-Ambros1]; [@pbio-0020098-Aravin1]; [@pbio-0020098-Brennecke1]; [@pbio-0020098-Lim1]). Of these, the biological functions of only four animal miRNAs are known. In C. elegans, the miRNAs *lin-4* ([@pbio-0020098-Lee2]; [@pbio-0020098-Olsen1]) and *let-7* ([@pbio-0020098-Reinhart1]) regulate developmental timing, whereas the *Drosophila* miRNAs *bantam* and *miR-14* control cell survival by repressing translation of proapoptotic genes ([@pbio-0020098-Brennecke2]; [@pbio-0020098-Xu1]). Computational approaches promise to identify the mRNA targets of other miRNAs ([@pbio-0020098-Enright1]; [@pbio-0020098-Lewis1]; [@pbio-0020098-Stark1]), but these proposed miRNA/target mRNA pairs will require experimental confirmation. Despite the widespread use of RNAi to "knock down" gene function, the RNAi pathway itself remains poorly understood. Similarly, despite intensive efforts to identify all miRNAs in vertebrates, worms, and flies, the mechanisms underlying miRNA function remain mysterious and no biological function has been ascribed to the vast majority of miRNAs. Here we show that 2′-*O*-methyl oligonucleotides are potent and irreversible inhibitors of small RNA-directed RNA silencing in vivo and in vitro. Our experiments using 2′-*O*-methyl oligonucleotides also demonstrate that the acquisition of a target RNA by an siRNA-programmed RISC is far more efficient than the binding of an antisense oligonucleotide to the same region of the target. To demonstrate the utility of 2′-*O*-methyl oligonucleotides in probing RNA silencing pathways, we show that 2′*-O-*methyl oligonucleotides efficiently block siRNA-directed RISC activity in cell extracts and in cultured human HeLa cells. When injected into C. elegans larvae, a *let-7*-complementary 2′-*O*-methyl oligonucleotide can efficiently suppress *lin-41* translational repression by the *let-7* miRNA. Finally, we use a tethered 2′-*O*-methyl oligonucleotide to demonstrate association of the C. elegans Argonaute proteins ALG-1 and ALG-2 with *let-7*. Results {#s2} ======= Inhibition of RNAi by 2′-*O*-Methyl Oligonucleotides {#s2a} ---------------------------------------------------- Although RNAi has proved a straightforward and cost-effective method to assess the function of protein-coding mRNAs ([@pbio-0020098-Fire1]; [@pbio-0020098-Caplen1], [@pbio-0020098-Caplen2]; [@pbio-0020098-Carthew1]; [@pbio-0020098-Elbashir1]a) and even some noncoding RNAs ([@pbio-0020098-Liang1]), no comparable method allows the sequence-specific inactivation of the siRNA or miRNA components of the RISC. The ideal inhibitor of RISC function would be a nucleic acid that would be recognized by the RISC by nucleotide complementarity, but be refractory to RISC-directed endonucleolytic cleavage or translational control. In theory, such a molecule would titrate out RISCs containing a complementary siRNA or miRNA, but not affect the function of RISCs containing guide RNAs unrelated in sequence. Such a RISC inhibitor should also resist degradation by cellular ribonucleases so that it persists long enough to bind RISC and block its function. Finally, the ideal inhibitor of small RNA function would act at concentrations unlikely to elicit nonspecific responses to the inhibitor itself, i.e., in the low nanomolar range, the same concentration at which siRNAs themselves are effective. At micromolar concentration, DNA antisense oligonucleotides may block miRNA function in *Drosophila* embryos ([@pbio-0020098-Boutla1]), but the poor stability of DNA oligonucleotides in vivo may limit their utility. Phosphorothioate-substituted DNA oligonucleotides, which show good in vivo stability, do not inhibit RISC function in vitro (data not shown). 2′-*O*-methyl oligonucleotides are also highly resistant to cellular ribonucleases ([@pbio-0020098-Inoue1]). To test whether 2′-*O*-methyl oligonucleotides can act as RISC inhibitors, we asked whether a 2′-*O*-methyl oligonucleotide, tethered to streptavidin paramagnetic beads via a 5′ biotin linkage, could be used to deplete siRNA-programmed RISC from the reaction. *Drosophila* embryo lysate was programmed with a synthetic siRNA duplex directed against a firefly (Photinus pyralis) luciferase (*Pp*-luc) mRNA target ([Figure 1](#pbio-0020098-g001){ref-type="fig"}A). Then, a tethered 31-nt 2′-*O*-methyl oligonucleotide complementary to the 21-nt siRNA antisense strand was added. Finally, the beads were removed from the solution using a magnet, and the supernatant was tested for siRNA-programmed RISC activity. Under these conditions, the 2′-*O*-methyl oligonucleotide completely depleted the reaction of the RISC programmed with the antisense strand of the siRNA, but not of RISC programmed with the sense strand ([Figure 1](#pbio-0020098-g001){ref-type="fig"}B). Thus, depletion occurred only when the siRNA strand contained within RISC was complementary to the tethered oligonucleotide. ::: {#pbio-0020098-g001 .fig} Figure 1 ::: {.caption} ###### A 2′*-O-*Methyl RNA Oligonucleotide Inhibits RNAi In Vitro in *Drosophila* Embryo Lysate \(A) Sequences of the sense and antisense *Pp*-luc target RNAs (black), the siRNA (red, antisense strand; black, sense strand), and the sense and antisense 2′*-O-*methyl oligonucleotides (blue) used. \(B) Sequence-specific depletion of RNAi activity by immobilized 2′-*O*-methyl oligonucleotides from *Drosophila* embryo lysate programmed with siRNA. siRNA was incubated with lysate to assemble RISC; then, immobilized 2′-*O*-methyl oligonucleotide was added. Finally, the beads were removed from the supernatant, and either sense or antisense ^32^P-radiolabeled target RNA was added to the supernatant to measure RISC activity for each siRNA strand. Symbols and abbreviations: Ø, target RNA before incubation with siRNA-programmed lysate; T, total reaction before depletion; unbound, the supernatant after incubation with the immobilized antisense (AS) or sense (S) 2′-*O*-methyl oligonucleotides shown in (A). The absence of 5′ cleavage product demonstrates that the sense oligonucleotide depleted RISC containing antisense siRNA, but not sense siRNA, and the antisense oligonucleotide depleted the sense RISC, but not that containing antisense siRNA. Bi, 5′ biotin attached via a six-carbon linker. :::  ::: We extended this method to measure the amount of RISC formed in the in vitro reaction at different concentrations of the siRNA duplex. An siRNA duplex in which the antisense strand was 5′-^32^P-radiolabeled was incubated in the reaction; then, the tethered 2′-*O*-methyl oligonucleotide was added to deplete the reaction of antisense siRNA-programmed RISC. The beads were then washed and the fraction of ^32^P-siRNA bound to the beads determined. Depletion was verified by testing the supernatant for RISC activity. Formally, the amount of ^32^P-siRNA retained on the beads for a given concentration of siRNA duplex places an upper limit on the concentration of RISC formed. However, our results using this assay are, within error, identical to the amount of RISC measured by two independent methods: the accumulation of single-stranded siRNA from functionally asymmetric siRNA duplexes ([@pbio-0020098-Schwarz2]), and the magnitude of the burst of target cleavage measured by pre-steady-state kinetics (B. Haley and P. D. Zamore, unpublished data). The simplest explanation for our results is that this assay directly measures siRNA incorporated into RISC. [Figure 2](#pbio-0020098-g002){ref-type="fig"}A shows the results of this assay for six different concentrations of siRNA duplex (5 nM, 15 nM, 25 nM, 50 nM, 100 nM, and 200 nM siRNA). First, the data show that RISC assembly in vitro is inefficient; the majority of siRNA duplexes do not contribute to RISC production. Second, RISC assembly is saturable, suggesting that some component of RISC itself is limiting. ::: {#pbio-0020098-g002 .fig} Figure 2 ::: {.caption} ###### 2′-*O*-Methyl Oligonucleotides Act as Stoichiometric, Irreversible Inhibitors of RISC Function \(A) The immobilized sense 2′-*O*-methyl oligonucleotide was used to determine the concentration of ^32^P-radiolabeled antisense siRNA assembled into RISC in *Drosophila* embryo. The 2′-*O*-methyl oligonucleotide and siRNA duplex are shown in [Figure 1](#pbio-0020098-g001){ref-type="fig"}A. (B--G) Inhibition of RNAi was measured using free 2′-*O*-methyl oligonucleotide and 1.3 nM (B), 4.6 nM (C), 9.3 nM (D), 14.5 nM (E), 18 nM (F), and 23.5 nM (G) RISC. The concentration of 2′-*O*-methyl oligonucleotide required for half-maximal inhibition (IC~50~) was calculated by fitting each dataset to a sigmoidal curve using a Hill coefficient of 1. \(H) A plot of IC~50~ versus RISC concentration suggests that each 2′-*O*-methyl oligonucleotide binds a single RISC. The data suggest that binding is essentially irreversible. :::  ::: To understand better the mechanism by which the 2′-*O*-methyl oligonucleotide interacted with RISC, we measured the concentration of free 2′-*O*-methyl oligonucleotide required for half-maximal inhibition of RISC activity (IC~50~; [Figure 2](#pbio-0020098-g002){ref-type="fig"}B--[2](#pbio-0020098-g002){ref-type="fig"}G) at the six different RISC concentrations determined in [Figure 2](#pbio-0020098-g002){ref-type="fig"}A. The IC~50~ for inhibition by free 2′-*O*-methyl oligonucleotide is shown for each RISC concentration in [Figure 2](#pbio-0020098-g002){ref-type="fig"}H. The IC~50~ for the 2′-*O*-methyl oligonucleotide was remarkably close to half the RISC concentration, suggesting that a single 31-nt 2′-*O*-methyl oligonucleotide binds each RISC and blocks its function. Consistent with this apparent 1:1 stoichiometry, the data for the 2′-*O*-methyl oligonucleotide titrations fit well to sigmoidal curves, with a Hill coefficient of 1 ([Figure 2](#pbio-0020098-g002){ref-type="fig"}B--[2](#pbio-0020098-g002){ref-type="fig"}G). The sequence specificity of 2′-*O*-methyl oligonucleotide inhibition of RISC function clearly shows that inhibition reflects binding of the oligo to the RISC. Our data are most easily explained if the concentration of the 2′-*O*-methyl oligonucleotide required for inhibition in these experiments is much greater than the K~D~ for binding; i.e., the experiments were conducted in a stoichiometric binding regime. Under a stoichiometric binding regime, inhibition by the 2′-*O*-methyl oligonucleotides would be essentially irreversible. In theory, the 2′*-O*-methyl oligonucleotide might act by displacing the passenger (sense) strand of the siRNA duplex, thereby blocking incorporation of the guide (antisense) strand into RISC ([@pbio-0020098-Elbashir3]c). We can exclude this possibility because a 5′ tethered 31-nt 2′-*O*-methyl oligonucleotide complementary to the passenger strand of the siRNA did not deplete the guide-strand RISC activity (see [Figure 1](#pbio-0020098-g001){ref-type="fig"}B). Similarly, an antisense sequence 2′-*O*-methyl oligonucleotide cannot pair with antisense RISC, but can pair with a sense target mRNA. We anticipated that this antisense 2′-*O*-methyl oligonucleotide would pair with the sense target mRNA and occlude the antisense RISC from the target. Surprisingly, the antisense 2′-*O*-methyl oligonucleotide was a poor inhibitor of antisense RISC function when it was used to bind the target site, requiring 300 nM for half-maximal inhibition in a reaction containing 14.5 nM RISC and 3 nM sense target RNA ([Figure 3](#pbio-0020098-g003){ref-type="fig"}A). By contrast, the same antisense 2′-*O*-methyl oligonucleotide was highly effective in blocking the activity of the sense RISC, to which it is complementary, acting with an IC~50~ of 8.2 nM in a reaction containing 16.8 nM sense-strand RISC and 3 nM antisense target RNA ([Figure 3](#pbio-0020098-g003){ref-type="fig"}B). (In this experiment, sense-strand RISC was generated by changing the first nucleotide of the sense strand from C to U, thereby reversing its functional asymmetry \[[@pbio-0020098-Schwarz2]\].) ::: {#pbio-0020098-g003 .fig} Figure 3 ::: {.caption} ###### RISC Does Not Act through an Antisense Mechanism \(A) Inhibition of sense target cleavage by an antisense 2′*-O-*methyl oligonucleotide requires an approximately 40-fold higher concentration than by a sense oligonucleotide. The antisense oligonucleotide can pair completely with the sense target RNA, but not with the antisense siRNA-programmed RISC. The IC~50~ value and the RISC concentration are indicated. Also shown are the sequences of the sense *Pp*-luc RNA target (black), the siRNA (red, antisense strand; black, sense strand), and the 2′*-O-*methyl oligonucleotide (blue). \(B) The same antisense 2′-*O*-methyl oligonucleotide is an effective competitor of antisense target cleavage. In this experiment, inhibition occurs via binding of the antisense oligonucleotide to the sense siRNA-programmed RISC, not the target RNA. The IC~50~ value and the RISC concentration are indicated. Also shown are the sequences of the *Pp*-luc antisense RNA target (black), the siRNA (red, antisense strand; black, sense strand), and the 2′*-O-*methyl oligonucleotide (blue). The G:U wobble in the siRNA duplex in (B) acts to direct the sense strand into RISC and improving its efficacy in target cleavage. :::  ::: Thus, the interaction of 2′-*O*-methyl oligonucleotide with RISC is dramatically different from the interaction of 2′-*O*-methyl oligonucleotide with target RNA; RISC has a more than 40-fold greater affinity for the 2′-*O*-methyl oligonucleotide than the oligonucleotide has for an RNA target (compare Figures [2](#pbio-0020098-g002){ref-type="fig"}E and [3](#pbio-0020098-g003){ref-type="fig"}A). These data imply that the interaction of RISC with target is more complex than simple nucleic acid hybridization. Inhibition of the siRNA-programmed RISC by a 2′-*O*-methyl oligonucleotide with the sequence of the target RNA is more effective than inhibition mediated by binding of an oligonucleotide to the target RNA itself. Thus, the RISC is more adept at finding or remaining bound (or both) to the target RNA than a 2′-*O*-methyl oligonucleotide. These data suggest that specific proteins in the RISC facilitate either target finding, target binding, or both. Consistent with this idea, inhibition of RISC function was incomplete using 21-nt 2′-*O*-methyl oligonucleotides (data not shown). Thus, target sequence flanking the site of complementarity to the siRNA guide strand may play a role in target RISC binding. Perhaps an active mechanism that involves target sequences flanking the siRNA facilitates the search for the target sequence; future studies will clearly be needed to test this idea. Inhibition of RNAi in Cultured Human Cells {#s2b} ------------------------------------------ Our data show that 2′*-O-*methyl oligonucleotides are stoichiometric, irreversible, sequence-specific inhibitors of siRNA function in RNAi reactions using *Drosophila* embryo lysate. Can 2′-*O*-methyl oligonucleotides block siRNA function in vivo? To address this question, we carried out sequential transfection experiments using 1 nM, 5 nM, 10 nM, or 25 nM siRNA duplex. siRNA was transfected on the first day; then, reporter and control plasmids were cotransfected together with various amounts of 2′-*O*-methyl oligonucleotide on the second day. Silencing of *Pp*-luc, relative to the sea pansy (Renilla reniformis) luciferase (*Rr*-luc) control, was measured on the third day. For each siRNA concentration, we determined the concentration of 2′-*O*-methyl required for half-maximal inhibition of RNAi ([Figure 4](#pbio-0020098-g004){ref-type="fig"}A--[4](#pbio-0020098-g004){ref-type="fig"}D). Increasing amounts of the 2′-*O*-methyl oligonucleotide gradually extinguished the ability of the siRNA to silence *Pp*-luc in all four experiments. The inhibition of silencing in the cultured cells cannot be a consequence of the 2′-*O*-methyl oligonucleotide displacing the sense strand of the siRNA duplex, because assembly of siRNA into RISC occurred a full day before the oligonucleotide was introduced. When 10 nM siRNA was used in the transfection, approximately 1 nM 2′-*O*-methyl RNA was required for half-maximal inhibition of RNAi ([Figure 4](#pbio-0020098-g004){ref-type="fig"}C and [4](#pbio-0020098-g004){ref-type="fig"}E). At 25 nM siRNA, approximately 1.1 nM 2′-*O*-methyl RNA was required to inhibit half the RNAi activity ([Figure 4](#pbio-0020098-g004){ref-type="fig"}D and [4](#pbio-0020098-g004){ref-type="fig"}E). In [Figure 4](#pbio-0020098-g004){ref-type="fig"}E, we plot the siRNA concentration versus the amount of 2′-*O*-methyl oligonucleotide required for half-maximal inhibition of silencing (IC~50~). The data fit well to a sigmoidal curve, consistent with the idea that, at these concentrations, increasing amounts of siRNA do not produce a corresponding increase in RISC activity. Higher concentrations of siRNA were not examined because they produce sequence-independent changes in gene expression ([@pbio-0020098-Persengiev1]; [@pbio-0020098-Semizarov1]). We conclude that both cells and extracts have a limited capacity to assemble RISC on exogenous siRNA. Our data suggest that the use of siRNA concentrations greater than that required to produce the maximum amount of RISC will lead to the accumulation of double-stranded siRNA in vivo and may thus contribute to the undesirable, sequence-independent responses sometimes observed in cultured mammalian cells ([@pbio-0020098-Persengiev1]; [@pbio-0020098-Semizarov1]; [@pbio-0020098-Sledz1]). ::: {#pbio-0020098-g004 .fig} Figure 4 ::: {.caption} ###### A 2′-*O*-Methyl Oligonucleotide Is a Potent Inhibitor of RNAi in Human Cultured HeLa Cells (A--D) HeLa cells were transfected with 1 nM (A), 5 nM (B), 10 nM (C), or 25 nM (D) siRNA-targeting *Pp-*luc mRNA. The next day the cells were cotransfected with *Rr-*luc*-*and *Pp*-luc-expressing plasmids together with various amounts of a 31-nt 2′*-O-*methyl oligonucleotide complementary to the antisense strand of the siRNA. The half-maximal concentration of 2′-*O*-methyl oligonucleotide required to inhibit (IC~50~) was determined by fitting the data to a sigmoidal curve using a Hill coefficient of 1. \(E) IC~50~ plotted as a function of the concentration of transfected siRNA. :::  ::: Inhibition of miRNA Function In Vitro and In Vivo {#s2c} ------------------------------------------------- In animal cells, miRNAs are thought predominantly to function as translational regulators. Nonetheless, a growing body of evidence suggests that they function through a similar, if not identical, RISC as siRNAs ([@pbio-0020098-Hutvagner1]; [@pbio-0020098-Zeng2], [@pbio-0020098-Zeng2]; [@pbio-0020098-Doench1]; [@pbio-0020098-Khvorova1]; [@pbio-0020098-Schwarz2]). Because 2′-*O*-methyl oligonucleotides block siRNA function in vitro and cultured human cells, we asked whether these oligonucleotides might likewise disrupt the function of a specific miRNA in vitro and in vivo. An ideal candidate for such an miRNA is *let-7*. Classical genetic mutations in *C. elegans let-7* produce well-characterized, readily scored phenotypes. Furthermore, human HeLa cells express multiple *let-7* family members, and endogenous *let-7* is present naturally in RISC ([@pbio-0020098-Hutvagner1]; [@pbio-0020098-Zeng1]). We tested whether a 31-nt 2′*-O-*methyl oligonucleotide complementary to *let-7* could block target cleavage guided by the endogenous *let-7*-programmed RISC present in HeLa S100 extract ([Figure 5](#pbio-0020098-g005){ref-type="fig"}A). (Our assay detects the target-cleaving activity of *let-7*; we have not examined endogenous human mRNA targets whose translation may be repressed by *let-7*.) As a control, we also tested whether the oligonucleotide could block the activity of a *let-7*-containing RISC assembled in vitro in *Drosophila* embryo lysate. Addition of this 2′-*O*-methyl oligonucleotide efficiently blocked target RNA cleavage directed by the endogenous *let-7*-programmed RISC in the HeLa S100 extract and by the RISC programmed with exogenous *let-7* siRNA duplex in *Drosophila* embryo lysate ([Figure 5](#pbio-0020098-g005){ref-type="fig"}C). ::: {#pbio-0020098-g005 .fig} Figure 5 ::: {.caption} ###### A Complementary 2′*-O-*Methyl Oligonucleotide Blocks Endogenous *let-7*-Containing RISC Function \(A) Sequence of the *let-7*-complementary site in the target RNA (black), of the siRNA (red, antisense strand; black, sense strand), and of the *let-7*-complementary 2′*-O-*methyl oligonucleotide (blue). \(B) Schematic representation of the target RNA, which contained both *Pp*-luc and antisense *let-7* sequences. \(C) *Drosophila* embryo lysate (left) was programmed with *let-7* siRNA; then, the target RNA and the 2′*-O-*methyl oligonucleotide were added together. Target RNA and 2′-*O*-methyl oligonucleotide (right) were added to HeLa S100 extract, which contains endogenous human *let-7*-programmed RISC. \(D) An RNA target containing both *Pp*-luc and antisense *let-7* sequence can be simultaneously targeted by *Pp*-luc siRNA and endogenous *let-7* in HeLa S100 lysate. The *let-7*-complementary 2′-*O*-methyl oligonucleotide blocks *let-7*-programmed, but not *Pp*-luc siRNA-programmed, RISC function. The bottom panel shows the same samples analyzed separately to better resolve the *let-7* 5′ cleavage product. \(E) *Drosophila* embryo lysate was programmed with *let-7* siRNA and then incubated with biotinylated 2′-*O*-methyl oligonucleotide tethered to paramagnetic streptavidin beads. The beads were removed and the supernatant tested for RNAi activity. Symbols and abbreviations: Ø, target RNA before incubation with siRNA-programmed lysate; T, total reaction before depletion; unbound, the supernatant after incubation with the paramagnetic beads. "Mock" indicates that no oligonucleotide was used on the beads; "*let-7*" indicates that the beads contained the *let-7*-complementary oligonucleotide shown in (A). :::  ::: In addition to containing endogenous *let-7*-programmed RISC, HeLa S100 can be programmed with exogenous siRNA duplexes ([@pbio-0020098-Martinez1]; [@pbio-0020098-Schwarz1]). The target RNA used in [Figure 5](#pbio-0020098-g005){ref-type="fig"}B also contains sequence from the *Pp*-luc mRNA and can therefore be targeted by a *Pp*-luc-specific siRNA duplex (see Figures [1](#pbio-0020098-g001){ref-type="fig"}A and [5](#pbio-0020098-g005){ref-type="fig"}C). We incubated the *Pp*-luc siRNA duplex with the human HeLa S100 to form *Pp*-luc-directed RISC, then added the *let-7*-complementary 2′-*O*-methyl oligonucleotide and the target RNA. The oligonucleotide blocked cleavage by the endogenous *let-7*-programmed RISC, but had no effect on cleavage directed by the exogenous *Pp*-luc siRNA in the same reaction ([Figure 5](#pbio-0020098-g005){ref-type="fig"}D). When tethered to a paramagnetic bead, this oligonucleotide could also quantitatively deplete the *let-7*-programmed RISC from the *Drosophila* embryo lysate ([Figure 5](#pbio-0020098-g005){ref-type="fig"}E), demonstrating that, again, the interaction between the 2′-*O*-methyl oligonucleotide and the RISC was apparently irreversible. The 2′-*O*-methyl oligonucleotide was a specific and potent inhibitor of target cleavage directed by a naturally occurring, miRNA-programmed RISC. Furthermore, our data demonstrate that individual RISCs act independently even when they target the same RNA. Next we asked whether 2′-*O*-methyl oligonucleotides can inhibit miRNA function in vivo. Translational repression directed by miRNAs occurs in C. elegans, where both the *lin-4* and *let-7* miRNAs have been shown to block translation of their target mRNAs without altering mRNA stability ([@pbio-0020098-Wightman1]; [@pbio-0020098-Ha1]; [@pbio-0020098-Moss1]; [@pbio-0020098-Olsen1]; [@pbio-0020098-Reinhart1]; [@pbio-0020098-Seggerson1]). The genetics of *lin-4* and *let-7* function are well-characterized in worms, where they are required during larval development to control the timing and pattern of cell division in the hypodermis ([@pbio-0020098-Lee2]; [@pbio-0020098-Reinhart1]). First, we tested whether injection into the germline of wild-type adult hermaphrodites of 2′*-O-*methyl oligonucleotides complementary to either *lin-4* or *let-7* could block *lin-4* or *let-7* function during the larval development of the resulting progeny. Although the 2′*-O-*methyl oligonucleotides were not toxic and when coinjected with an unrelated DNA transformation reporter did not prevent the uptake and expression of the coinjected DNA, we did not observe inhibition of *lin-4* or *let-7* activity (data not shown). This finding suggests that single-stranded 2′*-O-*methyl oligonucleotides are not efficiently transmitted to the progeny of injected animals. To circumvent this problem, we next injected 2′*-O-*methyl oligonucleotides directly into larvae and examined the phenotypes of the injected animals. The *lin-4* miRNA functions in L1/L2 larvae, and we have found that, in our hands, L1 larvae do not survive microinjection (data not shown); thus, it was not possible to assay for inhibition of *lin-4* function by direct injection. In contrast, *let-7* functions during the L4 stage, and we found that L2 and L3 larvae survive the microinjection procedure ([see Materials and Methods](#s4){ref-type="sec"}). Loss of *let-7* function causes worms to reiterate the L4 larval molt and inappropriately produce larval cuticle at the adult stage. Loss-of-function *let-7* phenotypes include weak cuticles prone to bursting at the vulva, defects in egg-laying, and loss of adult-specific cuticular structures that run the length of the animal\'s body, the alae ([@pbio-0020098-Reinhart1]). After larvae were injected with the *let-7*-specific 2′*-O-*methyl oligonucleotide, 80% of the adult worms lacked alae; 77% lacked alae and also exhibited bursting vulvae ([Figure 6](#pbio-0020098-g006){ref-type="fig"}A). In contrast, animals injected with an unrelated control 2′*-O-*methyl oligonucleotide displayed no abnormal phenotypes ([Figure 6](#pbio-0020098-g006){ref-type="fig"}A). ::: {#pbio-0020098-g006 .fig} Figure 6 ::: {.caption} ###### Injection of a 2′-*O*-Methyl Oligonucleotide Complementary to *let-7* miRNA Can Phenocopy the Loss of *let-7* Function in C. elegans \(A) Wild-type and *lin-41(ma104)* L2-stage C. elegans larvae were injected with either a 2′*-O-*methyl oligonucleotide complementary to *let-7* miRNA ([Figure 5](#pbio-0020098-g005){ref-type="fig"}A) or an unrelated *Pp*-luc 2′-*O*-methyl oligonucleotide. Absence of alae and presence of bursting vulvae were scored when the injected animals reached adulthood. \(B) Isolation of *let-7*-associated proteins with a tethered 2′-*O*-methyl oligonucleotide. Northern blot analysis of *let-7* miRNA remaining in the supernatant of the worm lysate after incubation with the *let-7*-complementary (*let-7*) or *Pp*-luc (unrelated) oligonucleotide. Input represents the equivalent of 50% of the total extract incubated with tethered oligonucleotide. \(C) Western blot analysis of the GFP-tagged ALG-1 and ALG-2 proteins associated with *let-7*. The upper band corresponds to GFP-tagged ALG-1 and the lower to GFP-tagged ALG-2. Extracts from a transgenic strain expressing the tagged proteins was incubated with the indicated tethered 2′-*O*-methyl oligonucleotide; then, the beads were washed and bound proteins were fractionated on an 8% SDS-polyacrylamide gel. Western blots were probed using anti-GFP monoclonal or anti-RDE-4 polyclonal antibody. The RDE-4-specific band is marked with an asterisk ([@pbio-0020098-Tabara2]). (D and E) Analysis of *let-7* miRNA in ALG-1/ALG-2 complexes (D). Extracts prepared from mixed-stage wild-type worms (N2) or from GFP::ALG-1/ALG-2 transgenic worms were immunoprecipitated using anti-GFP monoclonal antibodies. The unbound and immunoprecipitated RNAs were analyzed by Northern blot hybridization for *let-7* (D), and 5% of the immunoprecipitated protein was analyzed by Western blotting for GFP to confirm recovery of the GFP-tagged ALG-1/ALG-2 proteins (E). :::  ::: All of the phenotypes associated with injection of the *let-7*-complementary 2′*-O-*methyl oligonucleotide are consistent with a loss of *let-7* activity. *let-7* represses translation of *lin-41* mRNA by binding to a partially complementary site in the *lin-41* 3′-untranslated region ([@pbio-0020098-Reinhart1]; [@pbio-0020098-Slack1]; [@pbio-0020098-Vella1]). Consequently, many of the phenotypes associated with the loss of *let-7* reflect overexpression of LIN-41 protein; *let-7* mutants are partially suppressed by mutations in *lin-41*. We reasoned that if the phenotypes observed in the injected larvae reflect a loss of *let-7* activity, then they should likely be partially suppressed by a *lin-41* mutation ([@pbio-0020098-Reinhart1]; [@pbio-0020098-Slack1]). To test this possibility, we injected the *let-7*-specific 2′-*O*-methyl oligonucleotide into the *lin-41(ma104)* strain and compared the penetrance of the phenotypes with those observed for injection into wild-type*.* Consistent with the idea that the injected oligonucleotide specifically inactivates *let-7*, the absence of alae- and vulval-bursting phenotypes were both suppressed in the *lin-41(ma104)* mutant strain ([Figure 6](#pbio-0020098-g006){ref-type="fig"}A). The number of worms lacking alae was reduced from 80% to 16%, and worms with bursting vulvae were dramatically reduced (74% in wild-type compared to 3.8% in the *lin-41(ma104)* strain). The observed suppression (64%) was nearly identical to that reported for a *let-7, lin-41* genetic double mutant (70%; [@pbio-0020098-Reinhart1]; [@pbio-0020098-Slack1]). Together, our data support the idea that 2′*-O-*methyl oligonucleotides are potent inhibitors of miRNA function that can be used to probe the function of specific miRNAs in C. elegans. Isolation of Protein--miRNA Complex Using a Tethered 2′-O-Methyl Oligonucleotide {#s2d} -------------------------------------------------------------------------------- Our in vitro experiments suggest that both siRNA- and miRNA-containing RISCs are stably bound by 2′*-O-*methyl oligonucleotides. In theory, tethered 2′-*O*-methyl oligonucleotides could be used to isolate cellular factors associated with specific miRNAs. In human cells, miRNAs such as *let-7* are in a protein complex that contains Argonaute proteins ([@pbio-0020098-Hutvagner1]; [@pbio-0020098-Mourelatos1]; [@pbio-0020098-Dostie1]). In C. elegans, the Argonaute protein-encoding genes *alg-1* and *alg-2* are required for the biogenesis or function (or both) of the miRNAs *lin-4* and *let-7* ([@pbio-0020098-Grishok3]), but it has not been shown whether ALG-1 and ALG-2 proteins are directly associated with *let-7*. We prepared extracts from wild-type adult worms carrying a transgene expressing GFP-tagged ALG-1 and ALG-2 proteins. The extracts were then incubated with the *let-7*-complementary 2′*-O-*methyl oligonucleotide tethered by a 5′ biotin to streptavidin-conjugated paramagnetic beads. As a control, the experiment was performed in parallel using an oligonucleotide not complementary to *let-7*. The *let-7*-complementary, but not the control, oligonucleotide depleted nearly all the *let-7* miRNA from the extract ([Figure 6](#pbio-0020098-g006){ref-type="fig"}B). Western blotting using an anti-GFP antibody revealed that both GFP-tagged ALG-1 and ALG-2 protein copurified with the *let-7*-complementary oligonucleotide, but not the control oligonucleotide ([Figure 6](#pbio-0020098-g006){ref-type="fig"}C). In contrast, the RNA-binding protein RDE-4, which is required for RNAi but not for miRNA function in C. elegans, did not copurified with the *let-7*-complementarity oligonucleotide, providing further support for the specificity of the *let-7*:ALG-1/ALG-2 interaction ([Figure 6](#pbio-0020098-g006){ref-type="fig"}C). Finally, we used a coimmunoprecipitation assay to examine the interaction between *let-7* and ALG-1/ALG-2. In this assay, a monoclonal anti-GFP antibody was used to coimmunoprecipitate ALG-1/ALG-2 and small RNAs from the GFP::ALG-1/GFP::ALG-2 strain, which expresses GFP::ALG-1/ALG-2 fusion proteins. Northern blot analysis of the immune complex showed that it contained mature 22-nt *let-7* miRNA ([Figure 6](#pbio-0020098-g006){ref-type="fig"}D). No detectable *let-7* was recovered with the anti-GFP antibody from the N2 wild-type strain. By comparing the fraction of *let-7* associated with GFP::ALG-1/ALG-2 with the unbound fraction of *let-7* miRNA, we estimate that approximately 30% of the 22-nt *let-7* RNAs coimmunoprecipitate with GFP::ALG-1 and GFP::ALG-2. These data support a model in which that ALG-1 and ALG-2 form a complex, in vivo*,* that contains a substantial fraction of the mature *let-7* miRNA. Discussion {#s3} ========== Our studies indicate that 2′-*O*-methyl oligonucleotides bind efficiently and essentially irreversibly to RISC by basepairing with the small guide RNA. These findings provide a rapid and reliable method to measure programmed RISC concentration in vitro and to identify the in vivo functions of small RNA and the identities of their associated proteins. The ability to measure RISC concentration should enable detailed kinetic studies of the enzymatic activity of RISC, an essential step in understanding RISC function. In fact, this method was recently put to use in analyzing the molecular basis of asymmetry in siRNA function ([@pbio-0020098-Schwarz2]). In this study, we have used a tethered 2′-*O*-methyl oligonucleotide to demonstrated the association of ALG-1/ALG-2, two C. elegans Argonaute proteins, with the endogenous worm miRNA *let-7*. Our in vitro and in vivo studies using 2′-*O*-methyl oligonucleotides demonstrate that cells and extracts have a limited capacity to assemble RISC on exogenous siRNA. Our in vitro experiments suggest that inhibition of RISC by 2′-*O*-methyl oligonucleotides is stoichiometric and essentially irreversible. Using a sequential transfection protocol in cultured cells, we find that the half-maximal amount of 2′-*O*-methyl oligonucleotide required to inhibit silencing (IC~50~) is less than the amount of siRNA transfected. These data suggest that only a fraction of the transfected siRNA forms RISC. Furthermore, the data are consistent with stoichiometric and irreversible binding of the 2′-*O*-methyl oligonucleotide to RISC in vivo. Our data hint that recognition of the 2′-*O*-methyl oligonucleotide by RISC and, by inference, recognition of target RNA by RISC are qualitatively different from the simple binding of two complementary nucleic acids by basepairing. We observed that RISC function was far more readily inhibited by binding a 2′-*O*-methyl oligonucleotide to RISC than by binding the same 2′-*O*-methyl oligonucleotide to the site of RISC recognition on a target RNA. A clear implication of this finding is that RISC does not acquire its RNA target by a passive basepairing mechanism that zippers together 21 nt of complementary RNA. Thus, RNAi is not merely a form of antisense inhibition in which the antisense strand is stabilized in a duplex. Rather, an active mechanism---perhaps involving target sequences flanking the region of complementarity---underlies the specificity and efficiency of RISC targeting. Finally, we have shown the utility of 2′*-O-*methyl oligonucleotides to probe miRNA function in vivo. Injection of a 2′-*O*-methyl oligonucleotide complementary to the *let-7* miRNA into C. elegans larvae phenocopied a *let-7* loss-of-function mutation, demonstrating that 2′*-O-*methyl oligonucleotides can disrupt the function of a single miRNA in vivo. These data, combined with our studies in vitro and in cultured cells, show the promise of 2′-*O*-methyl oligonucleotides as a tool for dissecting the function of the numerous miRNAs found in a wide range of organisms. In this regard, 2′-*O*-methyl oligonucleotides provide a tool similar in practice, but mechanistically distinct from, RNAi itself and thus may facilitate the study of small RNA function in cases in which classical genetic mutations in miRNA genes are unavailable. Materials and Methods {#s4} ===================== {#s4a} ### General methods {#s4a1} *Drosophila* embryo lysate preparation, in vitro RNAi reactions, and cap-labeling of target RNAs were as described elsewhere ([@pbio-0020098-Haley1]). Target RNAs were used at approximately 3 nM concentration. Cleavage products of RNAi reactions were analyzed by electrophoresis on 5% or 8% denaturing polyacrylamide gels. Gels were dried, exposed to image plates, and then scanned with a FLA-5000 phosphorimager (Fuji Photo Film Company, Tokyo, Japan). Images were analyzed using Image Reader FLA-5000 version 1.0 (Fuji) and Image Gauge version 3.45 (Fuji). Data analysis was performed using Excel (Microsoft, Redmond, Washington, United States) and IgorPro 5.0 (Wavemetrics, Lake Oswego, Oregon, United States). ### siRNA and 2′-*O*-methyl oligonucleotides {#s4a2} Synthetic siRNA (Dharmacon, Lafayette, Colorado, United States) was deprotected according to the manufacturer, annealed ([@pbio-0020098-Elbashir2]b, [@pbio-0020098-Elbashir3]c), and used at 50 nM final concentration unless otherwise noted. 2′-*O*-methyl oligonucleotides (either from IDT, Santa Clara, California, United States, or from Dharmacon) were 5′-CAU CAC GUA CGC GGA AUA CUU CGA AAU GUC C-3′ and 5′-Bio-CAU CAC GUA CGC GGA AUA CUU CGA AAU GUC C-3′ (complementary to the *Pp*-luc siRNA sense strand); 5′-GGA CAU UUC GAA GUA UUC CGC GUA CGU GAU G-3′ and 5′-Bio-A CAU UUC GAA GUA UUC CGC GUA CGU GAU GUU-3′ (complementary to the *Pp*-luc antisense strand); and 5′-Bio-UCU UCA CUA UAC AAC CUA CUA CCU CAA CCU U-3′ (complementary to *let-7*); 5′ biotin was attached via a six-carbon spacer arm. ### Immobilized 2′-*O*-methyl oligonucleotide capture of RISC {#s4a3} Biotinylated 2′-*O*-methyl oligonucleotide (10 pmol) was incubated for 1 h on ice in lysis buffer containing 2 mM DTT with 50 μl of Dynabeads M280 (as a suspension as provided by the manufacturer; Dynal, Oslo, Norway) to immobilize the oligonucleotide on the beads. To ensure that the tethered oligonucleotide remained in excess when more than 50 nM siRNA was used, 20 pmol of biotinylated 2′-*O*-methyl oligonucleotide was immobilized. For RISC capture assays, siRNA was preincubated in a standard 50 μl in vitro RNAi reaction for 15 min at 25°C. Then, the immobilized 2′-*O*-methyl oligonucleotide was added to the reaction and incubation continued for 1 h at 25°C. After incubation, beads were collected using a magnetic stand (Dynal). The unbound supernatant was recovered and an aliquot assayed for RISC activity as previously described ([@pbio-0020098-Elbashir2]b; [@pbio-0020098-Nykanen1]) to confirm that RISC depletion was complete. The beads were then washed three times with ice-cold lysis buffer containing 0.1% (w/v) NP-40 and 2 mM DTT, followed by a wash without NP-40. To determine the amount of RISC formed, input and bound radioactivity was determined by scintillation counting (Beckman Instruments, Fullerton, California, United States). To isolate *let-7*-containing complexes from C. elegans adults, we incubated 20 pmol of immobilized 2′-*O*-methyl oligonucleotide with 1 mg of total protein. ### Sequential transfection {#s4a4} HeLa S3 cells were transfected in a 24-well plate (200 mm^2^ per well) using Lipofectamine 2000 (GIBCO, San Diego, California, United States) according to the manufacturer\'s protocol, first with various concentrations of siRNA targeting *Pp-*luc mRNA. After 6 h, the cells were washed with PBS and the media replaced. On the next day, the cells were cotransfected with *Rr*-luc-expressing (0.1 μg/well) and *Pp*-luc-expressing (0.25 μg/well) plasmids and 2′-*O*-methyl oligonucleotides using Lipofectamine 2000 (GIBCO) according to the manufacturer\'s protocol. The luciferase activity was measured 24 h later with the Dual Luciferase assay kit (Promega, Madison, Wisconsin, United States) using a Mediators Diagnostika (Vienna, Austria) PhL luminometer. ### Worm injection {#s4a5} For in vivo inhibition of *let-7* function, 1 mg/ml *let-7*-complementary 2′-*O*-methyl oligonucleotide in water (100 μM) was injected into either wild-type (N2) or *lin-41(ma104)* L2 larvae. Injection of L2 larvae was essentially as described elsewhere ([@pbio-0020098-Conte1]). The 2′-*O*-methyl oligonucleotide solution was injected into the body cavity of the larvae using the low flow and pressure setting to prevent animals from dying. Despite these precautions, approximately 60% of the animals do not survive injection, irrespective of the oligonucleotide injected. *let-7* phenotypes were also observed at 10 μM oligonucleotide, but were less penetrant. Phenotypes were scored after the injected animals survived to adulthood. ### Other methods {#s4a6} Synchronized transgenic animals carrying GFP::ALG-1, GFP::ALG-2 were harvested at adulthood and homog-enized in ice-cold buffer (25 mM HEPES--NaOH \[pH 7.4\], 150 mM NaCl, 1 mM EDTA, 1 mM DTT, 10% \[v/v\] glycerol, 0.5% \[v/v\] Triton X-100, 2% \[v/v\] SUPERaseIn \[Ambion, Austin, Texas, United States\]) and Mini Complete Protease Inhibitor cocktail (1 tablet/10 ml solution) (Roche, Basel, Switzerland) using a stainless-steel Dounce homogenizer (Wheaton Incorporated, Millville, New Jersey, United States). The homogenized extract was clarified by a centrifugation at 13,817 × *g* for 10 min at 4°C. To recover the proteins associated with the *let-7* miRNA, the beads were boiled for 10 min in 20 μl of SDS loading buffer (10 mM Tris--HCl \[pH 6.8\], 2% \[w/v\] SDS, 100 mM DTT, and 10% \[v/v\] glycerol). Proteins were resolved by SDS-PAGE on an 8% gel and transferred to Hybond-C membrane (Amersham Biosciences, Little Chalfont, United Kingdom). To detect GFP-tagged ALG-1, ALG-2, and RDE-4 proteins, the membrane was incubated overnight at 4°C with either monoclonal anti-GFP (Roche) or an affinity-purified polyclonal anti-RDE-4 antibody ([@pbio-0020098-Tabara2]) diluted 1:1000 into TBST-milk solution (100 mM Tris--HCl \[pH 7.5\], 150 mM NaCl, 0.1% \[v/v\] Tween-20, and 5% \[w/v\] dried milk), incubated 1 h at room temperature with either anti-mouse (GFP-tagged ALG-1/ALG-2) or anti-rabbit (RDE-4) HRP-conjugated secondary antibody (Jackson Laboratory, Bar Harbor, Maine, United States) diluted 1:5,000 in TBST and then visualized by enhanced chemulinescence (New England Nuclear, Boston, Massachusetts, United States). Immunoprecipitation of GFP-tagged ALG-1/ALG-2 complexes was performed by preclearing worm extract with 50 μl of protein G--agarose beads (Roche) per 5 mg of total protein for 1 h at 4°C. The cleared extract was then incubated with 10 μg of monoclonal antibody anti-AFP 3E6 (Qbiogene, Montreal, Quebec, Canada) for 1 h at 4°C followed by 50 μl of protein G--agarose. The agarose beads were then washed three times with ice-cold homogenization buffer. Depletion of *let-7* miRNA was monitored by Northern blotting. RNA was eluted from the immobilized 2′-*O*-methyl oligonucleotide by digestion with 1 mg/ml proteinase K in 200 mM Tris--HCl (pH 7.5), 25 mM EDTA, 300 mM NaCl, 2% (w/v) SDS at 50°C for 30 min, followed by extraction with phenol--chloroform, and recovered by precipitation with ethanol. Recovered RNA was resuspended in 10 μl of formamide-loading buffer (98% \[v/v\] deionized formamide, 10 mM EDTA, 0.025% \[w/v\] xylene cyanol, 0.025 % \[w/v\] bromophenol blue), heated to 100°C for 2 min. RNA was resolved on a 15% denaturing polyacrylamide gel, transferred to Hybond-N membrane (Amersham Biosciences), and detected by Northern blot analysis using a 5′-^32^P-radiolabeled antisense *let-7* RNA probe (UAU ACA ACC UAC UAC CUC AUU) as described elsewhere ([@pbio-0020098-Hutvagner1]). Supporting Information {#s5} ====================== Accession Numbers {#s5a} ----------------- The GenBank (<http://www.ncbi.nlm.nih.gov/Genbank/>) accession number for Photinus pyralis is X65324 and for Renilla reniformis is AF025846. Rfam (<http://www.sanger.ac.uk/Software/Rfam/index.shtml>) accession numbers for the *let-7* family members are MI0000060--MI0000068, MI0000433, and MI0000434. The LocusLink (<http://www.ncbi.nlm.nih.gov/LocusLink/>) ID numbers for the genes discussed in this paper are *alg-1* (181504), *alg-2* (173468), *bantam* (117376), *let-7* (266954), *lin-4* (266860), *lin-41* (172760), *miR-14* (170868), and *rde-4* (176438). We thank Darryl Conte, Jr., for the *gfp*::*alg-1* and *gfp*::*alg-2* plasmids; Erbay Yigit for generating the transgenic line expressing GFP::ALG-1, GFP::ALG-2; and members of the Zamore and Mello laboratories for helpful discussion and encouragement. Research support was provided by National Institutes of Health to PDZ (GM62862-01 and GM65236-01) and CCM (GM58800). GH is a Charles A. King Trust fellow of the Medical Foundation; MJS is a Canadian Institutes of Health Research postdoctoral fellow; CCM is a Howard Hughes Medical Institute assistant investigator; and PDZ is a Pew Scholar in the Biomedical Sciences and a W. M. Keck Foundation Young Scholar in Medical Research. **Conflicts of interest.** The authors have declared that no conflicts of interest exist. Academic Editor: Gerald Joyce, Scripps Research Institute \*To whom correspondence should be addressed. E-mail: <phillip.zamore@umassmed.edu> miRNA : microRNA *Pp*-luc : Photinus pyralis (firefly) luciferase RISC : RNA-induced silencing complex RNAi : RNA interference *Rr*-luc : Renilla reniformis (sea pansy) luciferase siRNA : small interfering RNA

PubMed Central

2024-06-05T03:55:46.115859

2004-2-24

PMC350667

Introduction {#s1} ============ Eukaryotic small RNAs of approximately 21--24 nucleotides function as guide molecules in a remarkably wide range of biological processes, including developmental timing and patterning, formation of heterochromatin, genome rearrangement, and antiviral defense ([@pbio-0020104-Carrington1]; [@pbio-0020104-Finnegan1]; [@pbio-0020104-Lai1]). They belong to at least two general classes, microRNA (miRNA) and short interfering RNA (siRNA). miRNAs (approximately 21--22 nucleotides) are found in plants and animals and are often phylogenically conserved within their respective kingdoms. They arise from non-protein-coding genes through formation of a precursor transcript followed by one or more nucleolytic processing steps ([@pbio-0020104-Lai1]). Part of the precursor adopts a fold-back structure that interacts with a multidomain RNaseIII-like enzyme termed DICER or DICER-LIKE (DCL1 in *Arabidopsis*), which catalyzes accurate excision of the mature miRNA ([@pbio-0020104-Denli1]). The miRNAs then associate with ribonucleoprotein complexes that function to negatively regulate target genes controlling a range of developmental events, such as timing of cell fate decisions, stem cell maintenance, apoptosis, organ morphogenesis and identity, and polarity ([@pbio-0020104-Ambros1]; [@pbio-0020104-Carrington1]). siRNAs are chemically similar to miRNAs, although in plants they typically range in size between 21 and 24 nucleotides ([@pbio-0020104-Hamilton2]; [@pbio-0020104-Llave1]a; [@pbio-0020104-Tang1]). They are associated with both post-transcriptional forms of RNA interference (RNAi) and transcriptional silencing involving chromatin modification ([@pbio-0020104-Finnegan1]). siRNAs are processed from precursors containing extensive or exclusive double-stranded RNA (dsRNA) structure, such as transcripts containing inverted repeats or intermediates formed during RNA virus replication ([@pbio-0020104-Hannon1]). siRNA precursors can also be formed by the activity of one or more cellular RNA-dependent RNA polymerases (RdRp), as was shown genetically in several screens for RNA silencing-defective mutants ([@pbio-0020104-Cogoni1]; [@pbio-0020104-Dalmay1]; [@pbio-0020104-Mourrain1]; [@pbio-0020104-Smardon1]; [@pbio-0020104-Volpe1]). *Arabidopsis* plants contain at least three active RdRp genes, termed *RDR1*, *RDR2*, and *RDR6* (also known as *SDE1/SGS2*) ([@pbio-0020104-Dalmay1]; [@pbio-0020104-Mourrain1]; [@pbio-0020104-Yu1]). *RDR6* is necessary for sense transgene-mediated RNAi, but not for silencing of constructs that encode transcripts with hairpins containing extensive dsRNA structure ([@pbio-0020104-Dalmay1]; [@pbio-0020104-Mourrain1]; [@pbio-0020104-Beclin1]). In many animals, both miRNAs and siRNAs are formed by the activity of the same DICER enzyme ([@pbio-0020104-Grishok1]; [@pbio-0020104-Hutvagner2]; [@pbio-0020104-Ketting1]; [@pbio-0020104-Knight1]; [@pbio-0020104-Provost1]; [@pbio-0020104-Zhang1]; [@pbio-0020104-Myers1]), although in plants they are formed by distinct DCL activities ([@pbio-0020104-Finnegan2]). *Arabidopsis* contains four *DCL* genes (*DCL1* to *DCL4*), only one of which (*DCL1*) has been assigned a definitive function in small RNA biogenesis ([@pbio-0020104-Park1]; [@pbio-0020104-Reinhart3]; [@pbio-0020104-Schauer1]). Biochemical data indicate, however, that multiple DCL activities or pathways catalyze formation of siRNAs of small-sized (approximately 21 nucleotides) and large-sized (approximately 24 nucleotides) classes ([@pbio-0020104-Tang1]). Endogenous siRNAs in plants arise from many types of retroelements and transposons, other highly repeated sequences, pseudogenes, intergenic regions (IGRs), and a few expressed genes ([@pbio-0020104-Hamilton2]; [@pbio-0020104-Llave1]a; [@pbio-0020104-Mette2]). Exogenous siRNAs can arise from both sense and hairpin transcript-forming transgenes and by viruses ([@pbio-0020104-Hamilton1]; [@pbio-0020104-Mette1]). Both siRNAs and miRNAs function post-transcriptionally to suppress or inactivate target RNAs. siRNAs guide sequence-specific nucleolytic activity of the RNA-induced silencing complex (RISC) to complementary target sequences ([@pbio-0020104-Hannon1]). Among other proteins, RISCs contain ARGONAUTE (AGO) family members that likely bind siRNAs or target sequences ([@pbio-0020104-Carmell1]). In plants and insects, post-transcriptional RNAi serves as an adaptive antiviral defense response ([@pbio-0020104-Waterhouse1]; [@pbio-0020104-Li1]). miRNAs are fully competent to guide nucleolytic function of RISC, provided that a target sequence with sufficient complementarity is available ([@pbio-0020104-Hutvagner1]; [@pbio-0020104-Doench1]; [@pbio-0020104-Tang1]). Many plant miRNAs function as negative regulators through this cleavage-type mechanism ([@pbio-0020104-Llave2]b; [@pbio-0020104-Rhoades1]; [@pbio-0020104-Emery1]; [@pbio-0020104-Kasschau1]; [@pbio-0020104-Palatnik1]; [@pbio-0020104-Tang1]; [@pbio-0020104-Xie1]). In animals, the level of complementarity between target and miRNA sequences is generally low, which inhibits nucleolytic activity. Animal miRNAs suppress translation of target mRNAs ([@pbio-0020104-Olsen1]; [@pbio-0020104-Reinhart2]). Some plant miRNAs may also function as translational suppressors ([@pbio-0020104-Aukerman1]; [@pbio-0020104-Chen1]). siRNAs also guide chromatin-based events that result in transcriptional silencing. Two lines of evidence support this view. First, in Schizosaccharomyces pombe and *Arabidopsis*, endogenous siRNAs from repeated sequences corresponding to centromeres, transposons, and retroelements are relatively abundant ([@pbio-0020104-Llave1]a; [@pbio-0020104-Mette2]; [@pbio-0020104-Reinhart1]). RNAi-related factors (DICER, RdRp, and AGO proteins) are required to maintain S. pombe centromeric repeats and nearby sequences in a transcriptionally inactive, heterochromatic state ([@pbio-0020104-Hall1]; [@pbio-0020104-Volpe1]). Mutants that lose RNAi component activities lose heterochromatic marks, such as histone H3 methylation at the K9 position (H3K9), as well as centromere function ([@pbio-0020104-Hall1]; [@pbio-0020104-Volpe1], [@pbio-0020104-Volpe2]). In plants, AGO4 is necessary to maintain transcriptionally silent epialleles of *SUPERMAN*. The *ago4* mutants lose both cytosine methylation, particularly at non-CpG positions, and H3K9 methylation at *SUPERMAN* and other constitutive heterochromatic sites (the Arabidopsis thaliana short interspersed element 1 \[*AtSN1*\] locus) ([@pbio-0020104-Zilberman1]). And, second, heterochromatin formation of nuclear DNA can be triggered, in a sequence-specific manner, by post-transcriptional silencing of cytoplasmic RNAs ([@pbio-0020104-Jones1]; [@pbio-0020104-Aufsatz1]; [@pbio-0020104-Schramke1]). The RNA-directed DNA methylation (RdDM) signal transmitted from the cytoplasm to the nucleus is most likely siRNA. The prevailing view states that chromatin-based silencing guided by siRNAs serves, among other purposes, as a genome defense system to suppress mobile genetic elements or invasive DNA ([@pbio-0020104-Dawe1]; [@pbio-0020104-Schramke1]). Using a genetic approach, we show here the existence of three small RNA-generating pathways with unique requirements in *Arabidopsis.* Plants with point mutations or insertions in several members of the *DCL* and *RDR* gene families were examined. The data indicate that plants genetically diversified several factors involved in formation of functionally distinct small RNAs. Results {#s2} ======= Genetic Requirements for miRNA Formation {#s2a} ---------------------------------------- At least two factors, DCL1 and HEN1 (HUA ENHANCER1), are involved in *Arabidopsis* miRNA formation. As shown for miR-171, miR-159 ([Figure 1](#pbio-0020104-g001){ref-type="fig"}A), and several other miRNAs ([@pbio-0020104-Park1]; [@pbio-0020104-Reinhart3]), mutants with *dcl1* loss-of-function alleles lose most of their miRNA populations ([Figure 1](#pbio-0020104-g001){ref-type="fig"}B). Plants with mutant *hen1* alleles either lose miRNAs or the apparent size of miRNAs is increased by one or more nucleotides ([@pbio-0020104-Park1]; [@pbio-0020104-Boutet1]) ([Figure 1](#pbio-0020104-g001){ref-type="fig"}B). miRNA function to suppress target mRNAs is diminished in both *dcl1* and *hen1* mutants ([@pbio-0020104-Boutet1]; [@pbio-0020104-Kasschau1]; [@pbio-0020104-Xie1]). To determine whether other DCL or RDR proteins are required for miRNA formation in *Arabidopsis*, miR-171 and miR-159 were analyzed in four new mutants. The *dcl2-1* and *dcl3-1* mutants contained T-DNA insertions in *DCL2* (At3g03300) and *DCL3* (At3g43920) genes, respectively ([Figure S1](#sg001){ref-type="supplementary-material"}). In wild-type plants, *DCL2* and *DCL3* transcripts accumulated to detectable levels in inflorescence tissues, but not in leaves. The mutant *dcl2-1* and *dcl3-1* transcripts were not detected in either tissue type ([Figure S1](#sg001){ref-type="supplementary-material"}). The *rdr1-1* and *rdr2-1* mutants contained T-DNA insertions in *RDR1* (At1g14790) and *RDR2* (At4g11130), respectively ([Figure S1](#sg001){ref-type="supplementary-material"}). *RDR1* and *RDR2* transcripts accumulated in inflorescence tissue, but not leaves, of untreated wild-type plants ([Figure S1](#sg001){ref-type="supplementary-material"}). The *RDR1* transcript levels were elevated in salicylic acid (SA)-treated leaves, as shown previously ([@pbio-0020104-Yu1]), but *RDR2* transcript levels were not affected by SA ([Figure S1](#sg001){ref-type="supplementary-material"}). Both *rdr1-1* and *rdr2-1* transcripts were below the detection limit in the corresponding mutant plants. In addition, a mutant containing an insertion in the *RDR6* gene (also known as *SDE1/SGS2*; At3g49500) was analyzed in parallel with the *rdr1* and *rdr2* mutants. This *rdr6-1* mutant displayed a weak virus-susceptibility phenotype that was consistent with previously reported *sde1* and *sgs2* mutants ([@pbio-0020104-Mourrain1]; [@pbio-0020104-Dalmay2]). However, no differences in *RDR6* transcript levels were detected between wild-type and *rdr6-1* mutant plants (data not shown). ::: {#pbio-0020104-g001 .fig} Figure 1 ::: {.caption} ###### Genetic Requirements for miRNA and Endogenous siRNA Generation \(A) miRNA genes and selected loci corresponding to three siRNAs or siRNA populations. Cloned small RNA sequenc-es are shown as green (sense orientation relative to the genome) or red (antisense orientation) bars. Protein-coding and miRNA genes are indicated by blue arrowheads. From top to bottom: miR-171 and miR-159a loci; siRNA02 loci, with each siRNA02 sequence indicated by an asterisk and the inverted duplication shown by the gray arrows; cluster2 siRNA locus; a segment of chromosome III showing 10 5S rDNA repeats (blue indicates 5S rRNA, gray indicates spacer) containing the siRNA1003 sequence. \(B) Small RNA blot assays for miR-171, miR-159, and endogenous siRNAs. Ethid-ium bromide-stained gels (prior to transfer) in the zone corresponding to tRNA and 5S RNA are shown at the bottom. Each mutant is presented in a panel with the corresponding wild-type control (Col-0 or La-*er*). :::  ::: Accumulation of miR-171 and miR-159 was unaffected in the *dcl2* and *dcl3* mutants (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}B). This was in contrast to the low level or shifted mobility of miR-171 and miR-159 in *dcl1-7* and *hen1-1*, respectively (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}B). Similarly, accumulation of miR-171 and miR-159 was unaffected in *rdr1* and *rdr2* mutants. Composition of Endogenous siRNA Populations {#s2b} ------------------------------------------- A library of cloned small RNAs from inflorescence tissues of Col-0 ecotype plants was partially sequenced and analyzed. Initial characterization of 125 of these sequences revealed that most of the clones corresponded to siRNA-like sequences ([@pbio-0020104-Llave1]a). A total of 1,368 distinct small RNAs, ranging in size between 20 and 26 nucleotides, were provisionally categorized here as siRNAs, with 24 nucleotides representing the most common size ([Figure 2](#pbio-0020104-g002){ref-type="fig"}A; all sequences are available to view or download at <http://cgrb.orst.edu/smallRNA/db/>). The siRNA sequences were identified at 5,299 genomic loci ([Table S1](#sd001){ref-type="supplementary-material"}). Approximately 27% of endogenous siRNAs derived from transposon or retroelement sequences in the sense or antisense polarity ([Figure 2](#pbio-0020104-g002){ref-type="fig"}B). Centromeric and pericentromeric siRNAs were common, which was partly due to the prevalence of transposons and retroelements at these sites. Forty-five small RNAs of sense and antisense polarity arose from highly repeated 5S, 18S, and 25S rDNA. While it is likely that some rDNA-derived sequences resulted from nonspecific breakdown of highly abundant rRNAs, some had specific genetic requirements and properties that were consistent with functional siRNAs (see below). Thirty-one siRNAs came from sequences annotated as psuedogenes and 147 from hypothetical or predicted genes ([Figure 2](#pbio-0020104-g002){ref-type="fig"}B). Only 28 were identified as originating from genes that are known to be expressed ([Figure 2](#pbio-0020104-g002){ref-type="fig"}B). The remaining 816 sequences mapped to loci that were collectively labeled as an IGR sequence. The IGR-derived siRNAs arose from unique sequences adjacent to known genes, inverted duplications, satellites, and other repeated sequences, although many of these may actually correspond to transposon or retroelement sequences that were not recognized by the search programs. ::: {#pbio-0020104-g002 .fig} Figure 2 ::: {.caption} ###### Endogenous siRNAs in *Arabidopsis* \(A) Size distribution of endogenous siRNAs. \(B) Distribution of distinct siRNAs in different sequence categories. \(C) Density of siRNAs from highly repeated (mainly transposons and retroelements; the asterisk shows repeat sequences identified using RepeatMasker), 5S rDNA, and unique genomic sequence. :::  ::: The frequency of unique siRNAs arising from highly repeated sequences (mainly transposons and retroelements), 5S rDNA repeats, and nonrepetitive sequence was calculated ([Figure 2](#pbio-0020104-g002){ref-type="fig"}C). siRNAs in the library occurred at a frequency of 2.42 per 100 kb repetitive DNA, which was approximately 2.4-fold higher that the frequency of siRNAs from nonrepetitive sequence (1.02 per 100 kb). Based on the number of repeats in the most current version of the *Arabidopsis* genome sequence, unique siRNAs corresponding to 5S rDNA were identified at a frequency of 7.55 per 100 kb. These data indicate that siRNAs arise more frequently from highly repeat genome sequences. Genetic Requirements for Endogenous siRNA Formation {#s2c} --------------------------------------------------- A set of four siRNAs or siRNA populations, representing the major categories identified in the library, were selected for genetic analysis. Twenty-six siRNAs corresponded to SINE retroelements, one of which (*AtSN1*) was selected for detailed analysis. *AtSN1*-derived siRNA formation requires *AGO4* ([@pbio-0020104-Zilberman1]) and *SDE4* ([@pbio-0020104-Hamilton2]). One siRNA (siRNA1003) originating from 5S rDNA was selected. The 5S rRNA genes occur in tandem arrays in chromosomes III, IV, and V, with the typical repeat unit (approximately 500 nucleotides) being composed of transcribed sequence (120 nucleotides) and flanking spacer sequences ([@pbio-0020104-Cloix1]; [@pbio-0020104-Mathieu1]). The siRNA1003 sequence was identified in the sense orientation within the spacer sequence in 202 repeats in chromosome III and four repeats in chromosome V (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}A). The cluster2 siRNA population from a 125-nucleotide IGR segment in chromosome I was represented by seven unique siRNAs in the library (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}A). Finally, the siRNA02 sequence corresponded to two loci separated by approximately 2.1 kb in chromosome V. One locus occurred in an IGR sequence, and the other within a hypothetical gene (At5g56070) of unknown function. The two siRNA02 loci occur in sequences that correspond to arms of an inverted duplication (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}A) ([@pbio-0020104-Llave1]a). The *AtSN1*, cluster2, and siRNA02 probes detected populations that accumulated as 24-nucleotide RNAs, while the siRNA1003 probe detected a population containing 21- to 24-nucleotide species (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}B). The abundance of each siRNA population was decreased in the *dcl3-1* mutant, but not in the *dcl1-7* or *dcl2-1* mutants (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}B). This was in strict contrast to miR-171, miR-159 (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}B), and several other miRNAs tested (data not shown), which depended specifically on *DCL1*. Interestingly, weak signals corresponding to siRNA02, *AtSN1* siRNAs, and cluster2 siRNAs were detected in faster-migrating positions in the *dcl3-1* mutant (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}B). This may have resulted from exposure of siRNA precursors to alternate DCL activities in the absence of DCL3. Notably, both small and large siRNAs detected by the 5S rDNA-derived siRNA1003 probe were diminished in *dcl3-1* plants. Each siRNA population was eliminated in the *rdr2-1* mutant, but not in the *rdr1-1* mutant (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}B). In preliminary experiments, each siRNA population was unaffected by the *rdr6-1* mutation, although these data should be interpreted cautiously because of the possibility that the *rdr6-1* allele is weak (data not shown). The endogenous siRNA requirement for *RDR2* contrasted with the miRNAs, which exhibited complete insensitivity to each of the *rdr* mutations tested (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}B). These data genetically identify DCL3 and RDR2 as components of an endogenous siRNA generating system that differs functionally from the miRNA-generating apparatus. The HEN1 protein was implicated in post-transcriptional silencing of sense-, but not hairpin-forming, transgenes ([@pbio-0020104-Boutet1]). We tested the requirement of HEN1 for endogenous siRNA formation using the *hen1-1* mutant. Two of the siRNA populations, siRNA1003 and the *AtSN1*-siRNAs, were reduced to undetectable levels in *hen1-1* plants (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}B). The siRNA02 and cluster2 siRNAs, on the other hand, reproducibly accumulated to higher levels in *hen1-1* plants compared to wild-type La-*er* plants. Thus, each type of endogenous siRNA tested requires DCL3 and RDR2, but only the highly repeated 5S rDNA and retroelement-derived siRNAs require HEN1. In fact, the requirement for, or independence from, HEN1 was precisely the same as AGO4 at each of these loci (D. Zilberman and S. Jacobsen, unpublished data). Function of the Endogenous siRNA-Generating System {#s2d} -------------------------------------------------- Two previous studies showed that SDE4 and AGO4 are required for *AtSN1* siRNA accumulation and methylation of cytosine positions at the *AtSN1* locus ([@pbio-0020104-Hamilton2]; [@pbio-0020104-Zilberman1]). In an *ago4* mutant, loss of *AtSN1* siRNA is associated with decreased histone H3K9 methylation ([@pbio-0020104-Zilberman1]). Cytosine methylation and increased histone H3K9 methylation are hallmarks of transcriptionally silent and heterochromatic DNA in plants and other organisms, and siRNAs may recruit chromatin modification complexes to specific loci ([@pbio-0020104-Grewal1]). To determine whether DCL3 and RDR2 catalyze formation of siRNAs that functionally interact with chromatin, cytosine methylation at *AtSN1* and 5S rDNA loci and methylation of H3K9 and H3K4 positions in *AtSN1* were examined in wild-type, *dcl3-1*, and *rdr2-1* plants. We also analyzed *AtSN1*-derived transcript levels to determine whether the mutations affected expression of the locus. Consistent with previous reports ([@pbio-0020104-Hamilton2]; [@pbio-0020104-Zilberman1]), bisulfite sequencing of *AtSN1* genomic DNA revealed extensive CpG (72.0%), CpNpG (43.1%), and asymmetric CpHpH (16.3%) methylation in Col-0 wild-type plants ([Figure 3](#pbio-0020104-g003){ref-type="fig"}A; [Table S2](#sd002){ref-type="supplementary-material"}). In the *rdr2-1* mutant, CpNpG and CpHpH methylation was reduced to 24.6% and 4.5%, respectively. Only a slight reduction in CpG methylation was detected in *rdr2-1* plants ([Figure 3](#pbio-0020104-g003){ref-type="fig"}A). This methylation pattern was similar to that detected in mutants lacking CHROMOMETHYLASE3 (*cmt3-7*; [Figure 3](#pbio-0020104-g003){ref-type="fig"}A), which is necessary for efficient methylation of *AtSN1* at non-CpG sites, and in a mutant lacking AGO4 ([@pbio-0020104-Zilberman1]). In the *dcl3-1* mutant, however, cytosine methylation was decreased only at asymmetric sites, while CpG and CpNpG methylation was similar to that of wild-type plants ([Figure 3](#pbio-0020104-g003){ref-type="fig"}A). ::: {#pbio-0020104-g003 .fig} Figure 3 ::: {.caption} ###### Effects of Mutations on *AtSN1* and 5S rDNA Chromatin Structure and Gene Expression \(A) Analysis of CpG (left), CpNpG (center), and CpHpH (right) methylation in *AtSN1* by bisulfite sequencing of genomic DNA. \(B) Blot analysis of 5S rDNA digested with methylation-sensitive restriction enzymes HpaII (left) and MspI (right). HpaII is sensitive to CpG and CpNpG methylation, whereas MspI is sensitive to only CpNpG methylation. Methylation is indicated by the ascending ladder, which corresponds to 5S rDNA multimers (monomer = approximately 0.5 kb). Duplicate samples from each plant were analyzed. \(C) ChIP assays using antibodies against dimethyl-histone H3K9 and dimethyl-histone H3K4. Genomic DNA associated with immunoprecipitated chromatin was analyzed by semiquantitative PCR with primer pairs specific for *AtSN1*, retrotransposon reverse transcriptase (At4g03800) (internal control for H3K9 methylation), and PFK (At4g04040) (internal control for H3K4 methylation). The PCR products were quantitated and compared against the respective internal controls, and the relative H3K4 and H3K9 methylation levels were expressed relative to that in Col-0 (arbitrarily set to 1.00). \(D) Detection of *AtSN1*-specific transcripts by semiquantitative RT-PCR. Primers specific for PFK transcripts were used as the internal control. A parallel set of reactions without addition of reverse transcriptase (RT) was run as a quality control for genomic DNA contamination. The PCR products were normalized relative to PFK, and the expression levels were calculated relative to that in Col-0 (arbitrarily set to 1.00). :::  ::: Because of the number of 5S rDNA repeats, analysis of cytosine methylation was done using restriction enzymes HpaII or MspI and DNA blot assays. Sensitivity to HpaII indicates lack of methylation at CpG or CpNpG sites (or both), whereas sensitivity to MspI indicates lack of methylation at only CpNpG sites. In wild-type Col-0 and La-*er* plants, 5S rDNA loci were heavily methylated at CpG+CpNpG sites, as shown by detection of only high molecular weight forms using HpaII, and partially methylated at CpNpG as shown using MspI ([Figure 3](#pbio-0020104-g003){ref-type="fig"}B). In *rdr2-1* plants, methylation was partially lost at CpNpG sites (increased MspI sensitivity; [Figure 3](#pbio-0020104-g003){ref-type="fig"}B, lanes 15--16), although to a lesser degree than in *cmt3-7* plants ([Figure 3](#pbio-0020104-g003){ref-type="fig"}B, lanes 21--22). Methylation detected by HpaII sensitivity was partially lost in the *rdr2-1* mutant ([Figure 3](#pbio-0020104-g003){ref-type="fig"}B, lanes 3--4), which was most likely due to loss of CpG methylation. Loss of only CpNpG methylation in *rdr2-1* plants would not account for the increased sensitivity to HpaII, as HpaII sensitivity in *cmt3-7* plants (lacking nearly all CpNpG methylation) was unaffected ([Figure 3](#pbio-0020104-g003){ref-type="fig"}B, lanes 9--10). Sensitivity of 5S rDNA sites to HpaII and MspI in *dcl3-1* plants was only slightly increased ([Figure 3](#pbio-0020104-g003){ref-type="fig"}B, lanes 5--6 and 17--18). In the *ago4-1* mutant, CpG methylation was partially lost as revealed by increased sensitivity to HpaII ([Figure 3](#pbio-0020104-g003){ref-type="fig"}B, lanes 11--12). Chromatin immunoprecipitation (ChIP) assays were used to detect changes in H3K4 and H3K9 methylation at *AtSN1* in *rdr2-1* and *dcl3-1* mutant lines. Loci containing genes encoding a retrotransposon reverse transcriptase and phosphofructokinase β subunit (PFK) were used as positive controls for sequences associated primarily with K9- and K4-methylated histone H3, respectively ([@pbio-0020104-Gendrel1]). At *AtSN1*, decreased levels of histone H3K9 methylation were detected in both *rdr2-1* and *dcl3-1* mutants (see [Figure 3](#pbio-0020104-g003){ref-type="fig"}C). This was accompanied by a slight increase in H3K4 methylation (see [Figure 3](#pbio-0020104-g003){ref-type="fig"}C). The extent to which H3 methylation changed was greater in *rdr2-1* relative to *dcl3-1* plants. Little or no change in H3K4 and H3K9 methylation was detected at the control loci. In addition, no changes in H3K4 or H3K9 methylation were detected at *AtSN1* in *cmt3-7* plants (data not shown). The changes in H3 methylation shown here are similar to those at several heterochromatic or silenced loci in *ago4* mutant plants ([@pbio-0020104-Zilberman1]). The level of *AtSN1*-derived transcripts was measured in *rdr2-1* and *dcl3-1* mutant plants and compared against the level of PFK transcript using semiquantitative RT-PCR. As shown in [Figure 3](#pbio-0020104-g003){ref-type="fig"}D, relatively low levels of *AtSN1* transcripts were detected in wild-type Col-0 plants. However, the normalized level of *AtSN1* transcripts was over 8- and 3-fold higher in *rdr2-1* and *dcl3-1* mutant plants, respectively, compared to wild-type plants. Therefore, loss of siRNA-forming capability correlated with loss of heterochromatic marks and elevated transcript levels at an endogenous locus that is normally silenced at the chromatin level. Given that RDR2, DCL3, and AGO4 are involved in chromatin-associated events and that HEN1 is required for accumulation of certain endogenous siRNAs associated with chromatin modification, it was hypothesized that each of these proteins accumulates in the nucleus. The presence of nuclear transport signals in each protein was tested by transient expression and analysis of green fluorescent protein (GFP) fusions in a heterologous plant, Nicotiana benthamiana, using an *Agrobacterium* infiltration assay. Subcellular accumulation sites for these proteins were compared to those of β-glucurodinase (GUS)--GFP (cytosolic control) and nuclear inclusion a protein (NIa)--GFP (nuclear control). The DCL3--GFP, HEN1--GFP, and GFP--AGO4 fusion proteins were detected exclusively in the nucleus ([Figure 4](#pbio-0020104-g004){ref-type="fig"}; [Figure S2](#sg002){ref-type="supplementary-material"}), indicating that DCL3, HEN1, and AGO4 possess independent nuclear transport capability. Subcellular localization experiments with RDR2--GFP and GFP--RDR2 fusion proteins, however, were inconclusive due to low expression levels and protein instability (data not shown). ::: {#pbio-0020104-g004 .fig} Figure 4 ::: {.caption} ###### Subcellular Localization of GFP Fusion Proteins Pairwise presentation of confocal microscopic images showing GFP fluorescence (top) and DAPI fluorescence (bottom) in N. benthamiana expressing the indicated GFP fusion proteins. Arrowheads indicate the location of nuclei. Note that the GUS--GFP control protein accumulates in cytoplasm at the cell periphery and immediately surrounding nuclei, while the NIa--GFP control protein accumulates in nuclei. Scale bar = 25μm. :::  ::: Genetic Requirements for Virus-Derived siRNA Formation {#s2e} ------------------------------------------------------ The involvement of DCL1, DCL2, and DCL3 in siRNA formation in response to infection by three dissimilar RNA viruses was tested using the *dcl* mutant series. Two of the viruses, a GFP-tagged version of turnip mosaic virus (TuMV--GFP) and turnip crinkle virus (TCV), infect *Arabidopsis* systemically and cause moderate to severe disease symptoms. The third virus, cucumber mosaic virus strain Y (CMV-Y), infects plants systemically, but causes only mild symptoms. Wild-type (Col-0 and La-*er*) and mutant plants were inoculated on rosette leaves, and upper, noninoculated tissue (cauline leaves and inflorescences) was analyzed for virus-specific siRNAs at 7 and 14 d post-inoculation (dpi). Viral siRNAs were detected in systemic tissues from wild-type plants at both timepoints ([Figure 5](#pbio-0020104-g005){ref-type="fig"}A--[5](#pbio-0020104-g005){ref-type="fig"}C, lanes 3, 5, 10, and 13), with siRNA levels generally higher at 14 dpi. In TuMV- and CMV-infected *dcl1-7*, *dcl2-1*, and *dcl3-1* mutant plants, siRNAs accumulated to levels that were similar to those in infected wild-type plants at 7 and 14 dpi (Figures [5](#pbio-0020104-g005){ref-type="fig"}A and [5](#pbio-0020104-g005){ref-type="fig"}B). TuMV and CMV titers and symptom phenotypes in the three mutants were indistinguishable from those in their respective parents (data not shown). Similarly, in TCV-infected *dcl1-7* and *dcl3-1* plants, viral siRNA levels, virus titer, and symptom severity were essentially the same as in wild-type plants ([Figure 5](#pbio-0020104-g005){ref-type="fig"}C; [Figure 6](#pbio-0020104-g006){ref-type="fig"}A and [6](#pbio-0020104-g006){ref-type="fig"}B; data not shown). ::: {#pbio-0020104-g005 .fig} Figure 5 ::: {.caption} ###### Genetic Requirements for DCLs in Viral siRNA Generation Blot analysis of viral siRNA. Systemic tissue samples were analyzed at the indicated time points from parental and mutant lines that were infected with TuMV--GFP (A), CMV-Y (B), and TCV (C). RNA blots were analyzed using virus-specific probes to detect siRNAs. Ethidium bromide-stained gels in the zone corresponding to tRNA and 5S RNA are shown. Relative accumulation (RA) of siRNAs is indicated at the bottom of each panel, with the level measured in infected control plants (Col-0 or La-*er*, depending on the mutant) at 7 dpi arbitrarily set to 1.0. :::  ::: ::: {#pbio-0020104-g006 .fig} Figure 6 ::: {.caption} ###### Altered Susceptibility to TCV Infection in *dcl2-1* Mutant Plants \(A) Noninfected control (left) and TCV-infected (right) Col-0, *dcl2-1*, and *dcl3-1* plants at 14 dpi. \(B) TCV accumulation, as measured by ELISA, in the systemic tissues of infected wild-type and mutant plants at 7 dpi (open bars) and 14 dpi (filled bars). \(C) Plant height (left), number of flowers/plant (center), and fresh weight of bolt tissue (right) were measured at 14 dpi in noninfected (open bars) and infected (filled bars) plants (*n* = 9). :::  ::: In contrast, TCV-derived siRNAs accumulated to levels that were 5-fold lower in *dcl2-1* plants compared to wild-type plants at 7 dpi (see [Figure 5](#pbio-0020104-g005){ref-type="fig"}C, lanes 10--11). This was a transient deficit, as TCV siRNA levels rebounded to near wild-type levels by 14 dpi (see [Figure 5](#pbio-0020104-g005){ref-type="fig"}C, lanes 13--14). The slow accumulation of siRNAs was not due to lack of TCV replication or movement in the tissues analyzed, as TCV titer in the *dcl2-1* mutant was similar to (7 dpi) or significantly higher than (*p* \< 0.05, 14 dpi) the titers in wild-type plants ([Figure 6](#pbio-0020104-g006){ref-type="fig"}B). Additionally, TCV-induced disease was more severe in *dcl2-1* plants, as plant height, fresh weight of bolts, and number of flowers in infected *dcl2-1* plants were each significantly (*p* \< 0.01 for plant height and flower number; *p* \< 0.05 for weight of bolts) lower compared to infected wild-type plants ([Figure 6](#pbio-0020104-g006){ref-type="fig"}A and [6](#pbio-0020104-g006){ref-type="fig"}C). Therefore, DCL2 functions as a component of the antiviral silencing response in TCV-infected plants. The DCL2--GFP fusion protein accumulated predominantly in the nucleus of N. benthamiana cells in the transient assay system, although some cytosolic localization was also detected (see [Figure 4](#pbio-0020104-g004){ref-type="fig"}). Thus, DCL1 ([@pbio-0020104-Papp1]), DCL2, and DCL3 each have nuclear transport activity. Discussion {#s3} ========== Genetic Diversification of Small RNA-Generating Systems in Plants {#s3a} ----------------------------------------------------------------- We show here that *Arabidopsis* has at least three systems to generate distinct classes of endogenous or virus-induced small RNAs and that these are associated with specialized regulatory or defensive functions. First, the miRNA-generating system requires DCL1, as shown previously ([@pbio-0020104-Park1]; [@pbio-0020104-Reinhart3]), but none of the RDR proteins tested. In principle, there should be no requirement for an RDR activity during miRNA biogenesis, as the DCL1 substrate is formed directly as a result of DNA-based transcription. DCL1 likely functions in the nucleus ([@pbio-0020104-Papp1]). It also functions, either directly or indirectly, with HEN1, which may confer substrate specificity, processing accuracy, or catalytic function. The second system requires DCL3 and RDR2 and generates endogenous siRNAs primarily of the large-sized (approximately 24 nucleotides) class. While DCL3 undoubtedly functions as the ribonuclease to process dsRNA precursors, RDR2 presumably functions as a polymerase to form dsRNA molecules de novo using templates resulting from transcription of DNA. At some loci, however, RDR2 may be unnecessary as a catalytic subunit, but rather contribute to the formation or stability of a complex that contains active DCL3. This could be the case at some sites, such as the siRNA02 locus, that contain inverted duplications and that may form transcripts with extensive dsRNA structure. Interestingly, accumulation of siRNAs specific to a hairpin construct was shown to be RdRp dependent in fission yeast ([@pbio-0020104-Schramke1]). At some loci, this system appears to interface with AGO4, HEN1, and SDE4. The third system functions in antiviral defense and involves DCL2. Loss of this system was specifically detected in TCV-infected *dcl2-1* plants, which exhibited delayed viral siRNA accumulation and increased susceptibility and sensitivity. However, there are several reasons to suspect that multiple antiviral, siRNA-generating systems exist. siRNAs triggered by TCV were not eliminated in *dcl2-1* plants, but rather siRNA accumulation was delayed. Although this could be due to incomplete loss of DCL2 function in the mutant, it could also reflect the existence of secondary or redundant DCL activities. Among the three viruses tested, two were unaffected by the *dcl2-1* mutation. This strongly implies the existence of one or more other siRNA-generating activities with unique or redundant antiviral specificity. Further, the DCL2-dependent system may have functions in addition to those associated with antiviral defense. The DCL2--GFP fusion protein was detected primarily in the nucleus, whereas TCV replicates and accumulates outside of the nucleus. Experiments to determine the genetic requirements for RDR1 and RDR2 during antiviral silencing against the three viruses were inconclusive, again possibly the result of functional redundancies or the presence of confounding viral RdRp activities ([@pbio-0020104-Ahlquist1]). [@pbio-0020104-Mourrain1]), on the other hand, showed that *rdr6* (*sde1/sgs2*) mutants were deficient in CMV-induced silencing. Additionally, [@pbio-0020104-Yu1]) showed that *RDR1* contributed to defense against tobamoviruses. [@pbio-0020104-Tang1]) identified two siRNA-generating DCL activities in wheat-germ extracts. These were detected using dsRNA as a substrate. Although monocots contain a *DCL* gene family, the members do not correlate one-for-one with those in *Arabidopsis* (Z. Xie and J. Carrington, unpublished data). Further study is required to correlate the DCL activities from wheat germ with those in *Arabidopsis.* The degree of genetic diversification of the DCL family in plants is in contrast to the situation in animals. Caenorhabditis elegans and human, for example, contain only one DICER ([@pbio-0020104-Grishok1]; [@pbio-0020104-Ketting1]; [@pbio-0020104-Knight1]; [@pbio-0020104-Provost1]; [@pbio-0020104-Zhang1]), even though both possess miRNA and siRNA functions. Thus, whereas plants diversified and functionally specialized DCL family members during evolution, animals evolved functionally distinct small RNA systems around one or relatively few DICER activities. Animals, however, evolved relatively large AGO-related families ([@pbio-0020104-Carmell1]), and these may provide modules for functional specialization. Roles of Endogenous siRNA-Generating Systems in Plants {#s3b} ------------------------------------------------------ Both DCL3 and RDR2 cooperate with AGO4, and possibly also with SDE4 and HEN1, at the *AtSN1* locus to initiate or maintain a heterochromatic state ([@pbio-0020104-Hamilton2]; [@pbio-0020104-Zilberman1]). Loss of DCL3, RDR2, and AGO4 factors correlates with loss of DNA methylation and histone H3K9 methylation. Interestingly, these factors are also necessary for silencing triggered de novo during the transformation process using transgenic *FWA* ([@pbio-0020104-Chan1]). Silencing of *FWA* is due to cytosine methylation of a region in the promoter that contains direct repeats ([@pbio-0020104-Soppe1]). The effect of the *rdr2-1* mutation on chromatin structure and gene silencing of *AtSN1* and *FWA* was generally stronger than the effect of the *dcl3-1* mutation. This may be explained by the presence of residual siRNAs formed by another DCL activity in the *dcl3* mutant (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}B). The picture that emerges from these and other results shows that DCL3 and RDR2 function as components of an endogenous siRNA-generating system and that the resulting siRNAs may guide chromatin modification events through effector complexes containing AGO4. Given that AGO proteins are components of RISCs that catalyze sequence-specific RNA degradation ([@pbio-0020104-Carmell1]) and that different AGO proteins have DNA- or RNA-binding activities ([@pbio-0020104-Lingel1]; [@pbio-0020104-Song1]; [@pbio-0020104-Yan1]), it seems reasonable to speculate that AGO4 engages a chromatin-associated RISC-like complex and interacts with nuclear siRNAs or target sequences. But unlike RNAi events in the cytoplasm, chromatin-associated complexes likely interact with DNA methyltransferase and histone methyltransferase systems. RdDM can occur at CpG and non-CpG sites, but maintenance of non-CpG methylation after DNA replication may generally require the continued activity of the siRNA-guided complex ([@pbio-0020104-Luff1]; [@pbio-0020104-Jones2]; [@pbio-0020104-Aufsatz1]). Methylation at CpG sites, in contrast, can be maintained by template-driven methylation on hemimethylated products of DNA replication, which explains why CpG methylation frequently persists in subsequent generations after one or more silencing factors or trigger loci are lost. Accumulation of siRNA from endogenous loci and transgenes does not necessarily require AGO4 (D. Zilberman and S. Jacobsen, unpublished data), suggesting that AGO4 acts downstream of siRNA formation to direct DNA methylation. Losses of AGO4 and HEN1 have nearly identical effects on all siRNAs tested, possibly because HEN1 and AGO4 affect a similar point in the pathway. If AGO4 and HEN1 function downstream of siRNA formation, why do siRNAs derived from some sites (*AtSN1* and 5S rDNA) accumulate to such low levels in *ago4* and *hen1* mutants? One possibility is that heterochromatic marks (DNA and H3K9 methylation) and associated factors serve to recruit RDR2, DCL3, or both to specific sites on chromatin, thus establishing a reinforcement loop. Loss of heterochromatin in an *ago4* mutant, for example, would result in failure to recruit the siRNA-generating enzymes to transcripts originating from a target locus and, therefore, the absence of siRNAs. This hypothesis, however, does not hold for some other siRNA-generating sites, such as those that yield cluster2 siRNAs and siRNA02. Accumulation of siRNAs from these sites is unaffected or even enhanced in *ago4* and *hen1* mutants. In wild-type plants, these loci are both hypomethylated at CpG and non-CpG sites and are associated with histone H3 that largely lacks K9 methylation (data not shown). The siRNAs formed from these loci clearly require RDR2 and DCL3, but they appear not to affect chromatin structure. These siRNAs may be sequestered elsewhere in the cell and unable to interact with chromatin or chromatin-associated factors. The spectrum of naturally occurring siRNAs in *Arabidopsis* is informative about the roles of these molecules in genome maintenance, genome expression, and defense. The fact that siRNAs from highly repeated sequences, largely retroelements and transposons, are overrepresented compared to unique genome sequences suggests that sequence duplication events are sensed and dealt with through RNA-guided formation of heterochromatin. This is frequently discussed within the context of genome defense, whereby suppression of mobile DNA promotes genome stability ([@pbio-0020104-Plasterk1]; [@pbio-0020104-Dawe1]). Indeed, loss of heterochromatin is often associated with increased activity of transposons and retroelements ([@pbio-0020104-Hirochika1]; [@pbio-0020104-Miura1]; [@pbio-0020104-Singer1]; [@pbio-0020104-Gendrel1]). However, it should be appreciated that these and other repeated sequences might also serve as *cis*-active, epigenetic regulatory modules if positioned near or within functional genes ([@pbio-0020104-Kinoshita1]). The rapidly expanding number of examples, such as vernalization ([@pbio-0020104-Bastow1]), of cellular memory conditioned by epigenetic events hint that siRNA-directed processes may be embedded broadly as a regulatory mechanism during growth and development ([@pbio-0020104-Goodrich1]). Materials and Methods {#s4} ===================== {#s4a} ### Plant materials {#s4a1} All plants were grown under standard greenhouse conditions. The *dcl1-*7, *hen1-1*, *cmt3-7*, and *ago4-1* mutant lines were described previously ([@pbio-0020104-Cao1]; [@pbio-0020104-Golden1]; [@pbio-0020104-Park1]; [@pbio-0020104-Zilberman1]). Other mutant lines were obtained from the Salk Institute Genome Analysis Laboratory (SIGnAL, La Jolla, California, United States) and Torrey Mesa Research Institute (now a subsidiary of Syngenta, Basel, Switzerland). *dcl2-1* has a T-DNA insertion within predicted intron 9 (after nucleotide 2,842 from ATG of the genomic DNA) of *DCL2* (At3g03300). *dcl3-1* has a T-DNA insertion within predicted exon 7 of *DCL3* (At3g43920) at a point 2,136 nucleotides beyond the ATG in genomic DNA. This introduces four codons after the serine 288 codon, followed by a premature stop codon. *rdr1-1* has a T-DNA insertion within predicted exon 1 after nucleotide 2,366 beyond the ATG of *RDR1* (At1g14790). *rdr2-1* has a T-DNA insertion within predicted exon 1 (in front of nucleotide 316 from the ATG) of *RDR2* (At4g11130). *rdr6-1* has a T-DNA insertion within predicted exon 2 (in front of nucleotide 3,977 from ATG of the genomic DNA) of *RDR6* (also known as *SDE1/SGS2*; At3g49500). Each insertion line was backcrossed twice to Col-0 and brought to homozygosity. Additional information about the insertion lines are provided in the supplemental online materials. For analysis of each insertion mutant, Col-0 was the wild-type control plant. For *dcl1-7*, *hen1-1*, *ago4-1*, and *cmt3-7* mutants, La-*er* was the wild-type control. ### RNA blot analysis {#s4a2} Extraction of low- and high-molecular weight RNAs and blot assays were done as described previously ([@pbio-0020104-Llave1]a). Low-molecular weight RNA (20 μg) from *Arabidopsis* inflo-rescence tissue was used for miRNA and endogenous siRNA analysis. Probes for miR-171 and *AtSN1*-siRNA analysis were described previously ([@pbio-0020104-Llave2]b; [@pbio-0020104-Zilberman1]). miR-159 was detected using an end-labeled DNA oligonucleotide AS-159 (5′-TAGAGCTCCCTTCAATCCAAA-3′). siRNA02 and siRNA1003 were detected using the end-labeled DNA oligonucleotides AS-02 (5′-GTTGACCAGTCCGCCAGCCGAT-3′) and AS-1003 (5′-ATGCCAAGTTTGGCCTCACGGTCT-3′), respectively. The probe for cluster2 siRNAs was a random primer-labeled fragment spanning a 235-nucleotide IGR of chromosome I (nucleotides 4,506,544--4,506,778) (see [Figure 1](#pbio-0020104-g001){ref-type="fig"}A) and was amplified from genomic DNA using primers AS-285 (5′-TTGCTGATTTGTATTTTATGCAT-3′) and S-786 (5′-CTTTTTCAAACCATAAACCAGAAA-3′). ### Analysis of DNA and histone methylation {#s4a3} Cytosine methylation was analyzed by bisulfite sequencing of genomic DNA or by DNA blot assay following digestion with methylation-sensitive restriction endonucleases, as described elsewhere ([@pbio-0020104-Jacobsen1]; [@pbio-0020104-Zilberman1]). The region of *AtSN1* analyzed (chromosome III, nucleotides 15,805,617--15,805,773) was treated with sodium bisulfite and amplified using primers *AtSN1*-BS1 (5′-GTTGTATAAGTTTAGTTTTAATTTTAYGGATYAGTATTAATTT-3′) and *AtSN1*-BS2 (5′-CAATATACRATCCAAAAAACARTTATTAAAATAATATCTTAA-3′). At least 18 independent clones were sequenced for each genotype. ChIP assays were done using antibodies specific for dimethyl-histone H3K4 (Upstate Biotechnology, Lake Placid, New York, United States) or dimethyl-histone H3K9 (kindly provided by T. Jenuwein, Research Institute of Molecular Pathology, Vienna, Austria) as described elsewhere ([@pbio-0020104-Gendrel1]). Methylation of H3K4 and H3K9 at *AtSN1* in wild-type Col-0 and *rdr2-1* and *dcl3-1* mutants was measured relative to that at internal control loci, At4g04040 and At4g03800. The data were then normalized against the values measured in Col-0. ### Analysis of GFP fusion proteins {#s4a4} The *35S:DCL3--GFP* construct contained the *DCL3* coding region fused to GFP coding sequence, flanked by the cauliflower mosaic virus (CaMV) 35S promoter and terminator sequences. The expression cassette was cloned in pSLJ755I5. All other GFP fusion constructs were made by cloning the coding sequence into pGWB5 (for C-terminal GFP) or pGWB6 (for N-terminal GFP), a set of gateway-compatible binary vectors designed for 35S promoter-driven expression of GFP fusion proteins (kindly provided by T. Nakagawa, Shimane University, Izumo, Japan). Cloning using gateway vectors was done using reagents and protocols from Invitrogen (Carlsbad, California, United States). Constructs were introduced into Agrobacterium tumefaciens strain GV2260 and expressed in N. benthamiana leaves as described previously ([@pbio-0020104-Johansen1]). Fusion proteins were detected by confocal microscopy and immunoblot assay using a monoclonal antibody against GFP (Roche, Basel, Switzerland). ### Virus infection assays {#s4a5} Wild-type and mutant *Arabidopsis* plants (approximately 4 wk old, prior to bolting) were infected with TuMV--GFP, CMV-Y, and TCV as described previously ([@pbio-0020104-Whitham1]; [@pbio-0020104-Lellis1]). At 7 and 14 dpi, systemic tissues consisting of inflorescences and cauline leaves were harvested for ELISA and RNA blot assays. Antibodies used for TuMV and TCV ELISAs were as described previously ([@pbio-0020104-Lellis1]). ### Computational methods {#s4a6} Computational identification of repeat sequences, including transposons and retroelements, in the *Arabidopsis* genome was done using RepeatMasker (http://ftp.genome.washington.edu/RM/RepeatMasker.html) and Repbase (<http://www.girinst.org/index.html>). Further information about *Arabidopsis* siRNAs and miRNAs, including those that were analyzed in this work, can be found in the *Arabidopsis* Small RNA Project database (<http://cgrb.orst.edu/smallRNA/db/>). Supporting Information {#s5} ====================== Figure S1 ::: {.caption} ###### *DCL* and *RDR* Mutant Lines \(A) Exon (bars)/intron (lines) organization of the *Arabidopsis DCL* and *RDR* genes and location of T-DNA insertion sites in mutant lines. \(B) RNA blot analysis (20 μg of total RNA) for *DCL2* and *DCL3* mRNA in Col-0 and the respective mutants. DNA fragments corresponding to nucleotides 2,652--3,292 of the *DCL2* open reading frame and nucleotides 2,805--3,571 of the *DCL3* open reading frame were used as hybridization probes. As a control, the blots were stripped and hybridized with a β-tubulin-specific probe ([@pbio-0020104-Kasschau1]). \(C) RNA blot analysis (10 μg of total RNA) for *RDR1* and *RDR2* mRNA in Col-0 and the respective mutants. DNA fragments corresponding to nucleotides 2,900--3,300 of the *RDR1* open reading frame and nucleotides 10--271 of the *RDR2* open reading frame were used as gene-specific probes. RNA samples from SA-treated leaf tissues were also included in the analysis. (5.9 MB EPS). ::: ::: {.caption} ###### Click here for additional data file. ::: Figure S2 ::: {.caption} ###### Immunoblot Analysis of GFP Fusion Proteins The 35S promoter-driven GFP fusion constructs were transiently expressed in N. benthamiana using an *Agrobacterium*-injection procedure. Leaf tissue from injected zones was excised at 2 dpi for immunoblot assay using a monoclonal antibody against GFP and confocal microscopy (see [Figure 4](#pbio-0020104-g004){ref-type="fig"}). An arrow indicates the position of predicted full-sized fusion protein. (10.8 MB EPS). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S1 ::: {.caption} ###### Cloned siRNA Loci in the *Arabidopsis* Genome (25 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Table S2 ::: {.caption} ###### Cytosine Methylation of *Arabidopsis AtSN1* (24 KB DOC). ::: ::: {.caption} ###### Click here for additional data file. ::: Accession Numbers {#s5a} ----------------- The GenBank (<http://www.ncbi.nlm.nih.gov/Genbank/>) accession numbers for the entities discussed in this paper are At1g14790 (NM\_101348), At3g03300 (NM\_111200), At3g43920 (NM\_114260), At3g49500 (NM\_114810), At4g11130 (NM\_117183), chromosome I (NC\_003070.3), chromosome III (NC\_003074.4), and siRNA02 (AF501743). The SAIL (formerly Garlic) (<http://signal.salk.edu/cgi-bin/tdnaexpress>) accession numbers for the T-DNA insertion lines discussed in this paper are *rdr1-1* (SAIL\_672F11), *rdr2-1* (SAIL\_1277H08), and *rdr6-1* (SAIL\_388H03). The SIGnAL database (<http://signal.salk.edu/>) accession numbers for the T-DNA insertion lines discussed in the paper are *dcl2-1* (SALK\_064627) and *dcl3-1* (SALK\_005512). We thank Scott Givan and Chris Sullivan for invaluable assistance and advice with computational resources; Thomas Jenuwein (Research Institute of Molecular Pathology, Vienna, Austria) for providing dimethyl-histone H3K9-specific antibodies; Xuemei Chen (Rutgers University, Camden, New Jersey, United States) for providing *hen1-1* seeds; Tsuyoshi Nakagawa (Shimane University, Izumo, Japan) for providing the pGWB vector series; Bridget Timony for assistance with virus infection experiments; and Zachary Lippman and Robert A. Martienssen for sharing their ChIP protocol and control locus information. We also thank Syngenta and the Salk Institute Genome Analysis Laboratory for access to their T-DNA insertion lines. This work was supported by grants from the National Science Foundation (MCB-0209836) and National Institutes of Health (AI43288, F32A1051097, and GM60398). **Conflicts of interest.** The authors have declared that no conflicts of interest exist. **Author contributions.** ZX, LKJ, KDK, and JCC conceived and designed the experiments. ZX, LKJ, KDK, and ADL did most of the experimental work. AMG, KDK, and JCC conceived and designed the small RNA database. AMG constructed the database. DZ and SEJ generated the *ago4-1* and *cmt3-7* mutants and developed chromatin analysis protocols. ZX, LKJ, KDK, JCC, and SEJ analyzed the data. ZX and JCC wrote the paper. Academic Editor: Detlef Weigel, Max Planck Institute for Developmental Biology AGO : ARGONAUTE *AtSN1* : Arabidopsis thaliana short interspersed element 1 CaMV : cauliflower mosaic virus ChIP : chromatin immunoprecipitation CMT3 : CHROMOMETHYLASE3 CMV : cucumber mosaic virus*;* DCL dpi : days post-inoculation dsRNA : double-stranded RNA GFP : green fluorescent protein GUS : β-glucurodinase H3K4 : histone H3 at the lysine 4 H3K9 : histone H3 at the lysine 9 HEN1 : HUA ENHANCER 1 IGR : intergenic region miRNA : microRNA NIa : nuclear inclusion a protein PFK : phosphofructokinase β subunit RdDM : RNA-directed DNA methylation *RDR* : RNA-dependent RNA polymerase gene RdRp : RNA-dependent RNA polymerases RISC : RNA-induced silencing complex RNAi : RNA interference RT : reverse transcriptase SA : salicylic acid siRNA : short interfering RNA TCV : turnip crinkle virus TuMV : turnip mosaic virus

PubMed Central

2024-06-05T03:55:46.119475

2004-2-24

PMC350672

Since small RNA molecules were discovered just over ten years ago, it\'s become clear that these once overlooked bits of genetic material play a decidedly large role in controlling gene expression. Though typically just 21 to 24 nucleotides long, small RNAs regulate a diverse array of cellular processes, from developmental patterning and genome rearrangement to antiviral defense. They typically accomplish these tasks by targeting specific nucleotide sequences to shut down gene expression.[](#pbio-0020107-g001){ref-type="fig"} ::: {#pbio-0020107-g001 .fig} ::: {.caption} ###### A. thaliana at the rosette stage :::  ::: Found in both plants and animals, small RNAs come mainly in two classes---microRNA (miRNA) and short interfering RNA (siRNA). miRNAs arise from nonprotein-coding transcripts that adopt extended "fold-back" structures, which are then cleaved by enzymes called Dicer or Dicer-like (DCL). siRNAs, on the other hand, arise from perfectly base-paired double-stranded RNA, which are also cleaved by Dicer. Some siRNAs require enzymes called RNA-dependent RNA polymerases (RdRp). miRNAs and many types of siRNAs function post-transcriptionally---that is, they affect genes that have been expressed, or transcribed, into RNA---to guide cleavage or prevent translation into protein. In plants and some animals, this post-transcriptional RNA interference (RNAi) acts as an adaptive antiviral response, among other things. siRNAs can also "silence" gene expression by altering chromatin---the DNA-protein complex into which chromosomes assemble---and preventing transcription. It is thought that chromatin silencing acts as a genome defense mechanism, guarding against potential damage from mobile genetic elements or invasive DNA (say, from a virus) by keeping genes in the tightly coiled, and thus inaccessible, "heterochromatic" state. While much remains to be learned about the mechanisms and pathways that govern small RNAs, it\'s becoming clear that they add an important layer of regulation and flexibility to gene expression. Now a team led by James Carrington at Oregon State University and Steve Jacobsen at the University of California at Los Angeles demonstrates that plants have evolved multiple systems to produce distinct classes of small RNAs with specialized regulatory and defensive functions. The first generates miRNAs; the second produces siRNAs that regulate chromatin structure; and the third generates siRNAs in response to viral infections. Each system requires a unique spectrum of functions of three different DCL proteins; the siRNA systems each function in coordination with one of several RdRp proteins. The researchers propose that the expansion and subsequent diversification of these proteins, which occurred in plants but not in many animals, has contributed to the diversification of specialized small RNA-directed pathways. Working in Arabidopsis thaliana, a favorite model organism for plant biologists, Zhixin Xie et al. analyzed a series of mutants with nonfunctional *dcl* and *rdr* genes, as well as a few other mutants of interest, to determine how the small RNAs responded to loss of these proteins. Two mutations (one in a *dcl* gene and one in another gene) affected the miRNAs, either impairing their function or reducing their populations. None of the RdRp proteins had any effect on miRNAs. The researchers performed the same type of genetic analyses on siRNAs and found that a different DCL mutant caused a reduction in one class of siRNAs and that an RdRp mutant nearly eliminated these populations of siRNAs. The diversity of siRNAs produced by the Arabidopsis genome reveals an important role in genome maintenance, expression, and defense, the authors conclude. Given that large numbers of siRNAs arise from highly repeated sequences---such as those introduced by viruses or mobile genetic elements---it may be that the cell senses such "invasive" sequence duplication events and enlists siRNAs to run interference by silencing these potentially damaging sequences. In this way, chromatin-associated siRNAs may offer an additional line of defense against invasive sequences, on top of that offered by post-transcriptional RNAi---a dual adaptive advantage since a fast-spreading virus or over-proliferating transposon (also known as a jumping gene) could wreak havoc on a plant population. Whatever other roles small RNAs may play in genome regulation---they have also been implicated in regulating growth and development---their primary responsibility appears to be blocking gene expression. Whether they accomplish that by controlling chromosome activity to prevent gene transcription or by inhibiting or degrading RNA transcripts to block translation into protein, small RNAs appear to make wide-ranging contributions to the overall gene expression program of the cell.

PubMed Central

2024-06-05T03:55:46.123658

2004-2-24

PMC350674

Regulation of gene expression---deciding how much of what proteins are produced in the cell---is controlled by a myriad of different molecules. One type of naturally occurring regulatory molecule is small interfering RNA (siRNA), which selectively disrupts the production of a protein it is programmed to recognize, a process called RNA interference. These short stretches of nucleotides combine with other cellular proteins to form an RNA-induced silencing complex, called RISC, which locates and destroys a targeted messenger RNA---the molecule that carries a protein recipe from the nucleus to the site of production in the cytoplasm. While RNA interference has been widely exploited by researchers as a tool to knock out gene expression and infer the function of missing proteins, very little is known about the mechanisms behind this regulatory process. Recently, biologists have discovered hundreds of other short pieces of regulatory RNA, called microRNAs, in both plants and animals. Like siRNA, they also affect gene expression, through similar, possibly even identical RISC molecules. Animal microRNAs, however, target messenger RNA at a different stage in protein production. Though researchers have determined the sequences of these microRNAs, uncovering their function---that is, which protein they interrupt and, in turn, what the interrupted protein does---has progressed slowly and sporadically without any decisive tool to study them. Only four animal microRNAs have known biological functions, despite the intense level of work going on in this field. In this issue, György Hutvágner and colleagues report a rapid and reliable method for knocking out both siRNAs and microRNAs and thereby exploring their functions. The authors found that a short stretch of nucleotides, called a 2′-*O*-methyl oligonucleotide, whose sequence mirrors the targeted siRNA or microRNA, could bind and inhibit their function, allowing researchers an unprecedented glimpse at the regulatory roles and mechanisms behind RNA interference. The authors first tested their oligonucleotide design against an siRNA known to interfere with production of the firefly protein, luciferase---this luminescent protein is often used as a "reporter," lighting up when cells successfully produce the protein. Any interference means the glow is gone. Using extracts from fruitfly embryos as the test-bed, the researchers mixed in the luciferase-associated siRNA and the sequence-specific oligonucleotide. What holds these two molecules together is complementary base-pairing, the same force that holds two molecules of DNA together. As predicted, the oligonucleotide inhibited RISC activity---it could no longer silence the production of luciferase. Because the authors could easily control the concentration of both the siRNA and the oligonucleotide inhibitor in these fly extract experiments, they were able to answer several questions about how these two molecules interact. They found that adding greater and greater concentrations of siRNA molecules did not result in equally great numbers of RISC; the process became saturated, indicating that a protein in the RISC assembly pathway limits production. Furthermore, the authors saw a marked 1:1 relationship between the concentration of the oligonucleotide and the concentration of RISC, indicating that each inhibitor binds to one RISC molecule in order to inactivate, a binding that appears to be irreversible. The results also showed that, though RISC molecules bind to the inhibitor through complementary base-pairing, a very different and more complex interaction is used by RISC molecules to find and bind their natural interference targets. The authors then went on to use the luciferase siRNA to test the function of their oligonucleotide inhibitor in cultured human cells, which had been engineered to contain the luciferase gene. This in vivo experiment, using living and metabolizing cells, showed results similar to those with fruitfly extracts. But the real test for these inhibitors was to use them in a whole animal against a previously identified microRNA where the outcome of its inactivation was already known. Hutvágner and colleagues constructed an oligonucleotide inhibitor based on the sequence of a microRNA called *let-7*, which blocks the production of the protein Lin-41 and is important for proper developmental timing in roundworm larvae. Larvae injected with the oligonucleotide had the exact features of a *let-7* deficient worm, showing that the inhibitor did indeed block this microRNA\'s function. The authors also used the oligonucleotides to provide evidence that two proteins, previously suggested to be involved with *let-7*, were directly associated with its interfering activity. Using the technique described here, scientists could make rapid headway toward uncovering the biological functions of hundreds of microRNAs, their accessory RISC proteins, and even the proteins and genes they are programmed to interrupt. Furthermore, finding that RISC production is saturable could have significant implications for genetic studies that use RNA interference to uncover the function of sequenced, but unknown, genes; knowing the minimum required concentration of siRNA, researchers can avoid a buildup and any unwanted cell activity that goes along with it.

PubMed Central

2024-06-05T03:55:46.124237

2004-2-24

PMC359389

We are two of the scientist members of the President\'s Council on Bioethics. In late 2001, we were invited by the President of the United States to serve on this Council. The Bioethics Council was appointed by the President to "monitor stem-cell research, to recommend appropriate guidelines and regulations, and to consider all of the medical and ethical ramifications of biomedical innovation.... This council will keep us apprised of new developments and give our nation a forum to continue to discuss and evaluate these important issues." This was a difficult invitation to accept. On the one hand, the President\'s views on the use of human embryonic stem cell research and somatic cell nuclear transfer techniques were well-known and in conflict with our own beliefs about the costs and benefits of the use of progressive technologies to advance biomedical research. On the other hand, we were grateful that the President, despite his views in opposition to these therapies, was willing to invite serious biomedical scientists to help formulate advice to him---and ultimately to contribute to the development of national policy---on these critically important advances. We knew that on this originally 18-member (but for most of the past two years a 17-member) Council, as scientists we would be in the minority in our belief of the good to be gained through these and other areas of biomedical research. We were also aware that some others on the Council had strong opposing views. Thus, it was only with the assurances of the Council chairman, Leon Kass of the University of Chicago, and of the President of the United States himself that we were persuaded that our voices would be heard and integrated into the statements of the Council. Furthermore, we felt, and continue to feel, that bioethical issues are important not only to all biologists, but also to society at large, and thus especially worthy of engaging debate and discussion. Two recently issued reports of the Council, "Beyond Therapy: Biotechnology and the Pursuit of Happiness" (<http://bioethics.gov/reports/beyondtherapy/index.html>) and "Monitoring Stem Cell Research" (<http://bioethics.gov/reports/stemcell/index.html>), are therefore of deep concern to us. We discuss them in turn below. Concerns about the "Beyond Therapy" Report {#s2} ========================================== The "Beyond Therapy" report deals with issues of direct concern for every thoughtful person. However, in the interests of setting straight the record of our views, as Council members and scientists, on the content of this report and for a proper assessment of the scientific content of the "Beyond Therapy" report, we feel it is important to point out aspects of the report for which we had requested revisions and for which those requests were declined. In the discussions of preimplantation genetic diagnosis, the specter of designer babies is raised by implying that selecting embryos for intelligence and other traits, such as temperament is a possibility. Scientifically, this simply is highly unlikely and indeed may not even be feasible. While such scientific unlikelihood is mentioned in passing in the report, it is easy to take away from the report the feeling that such genetic manipulation will happen and is even imminent. The report also claims that "the underlying impulse driving age-retardation research is, at least implicitly, limitless, the equivalent of a desire for immortality." Furthermore, the title of Chapter 4 of the report, "Ageless Bodies," implies that immortality is the goal of this research, despite all reliable scientific evidence to the contrary. Such a title is not consistent with the knowledge, stated in that chapter, that there is no scientific basis for immortality and implies that, by seeking to maintain and extend "youth," research into aging, including stem cell research, is predominantly to serve vanity. Also, without presenting scientific or reliable evidence, the report presents the opinion that research into prolonging healthy life may result in a lifetime obsession with immortality. Hence, this chapter in the report falls short of explaining the serious challenge of preventing and curing age-related disease to extend health---very different from attempting immortality. The same chapter offers a sensational quote from a researcher that "the real goal \[of aging research\] is to keep people alive forever." The request that quotes from researchers more representative of the biomedical research community also be included was declined. This leads to a misleading misrepresentation of the motivation of reputable researchers in the field of aging. In suggesting that slowing biological aging may increase the disjunction between "social aging" (the age at which children are exposed to "adult" images and concepts) and "biological aging" (expected lifespan), only one view, a conservative one, of the supposed "best" way to raise children is presented. The report also suggests, with no clear reasoning behind it, that longer lives will somehow undermine human determination to contribute as much as one can during a lifetime. Despite requests for inclusion of material that would allow for a balanced treatment of these topics, the report minimized discussion of potential positive aspects of slowing biological aging, such as prolonged good health. Finally, the report repeatedly emphasizes a "profound and mysterious" link between longevity and fertility, thereby leaving the reader with the distinct but erroneous impression that anything done to extend healthy life will be traded for decreased fertility, despite the fact that current scientific literature, which was made available for inclusion in the report, shows a *lack* of any necessary *mechanistic* linkage of the two. Concerns about the "Monitoring Stem Cell Research" Report {#s3} ========================================================= With respect to the "Monitoring Stem Cell Research" report, we feel that some facts that would help the public and scientists better assess the content of the report were not brought out clearly or were omitted entirely. First, from the published scientific literature in peer-reviewed journals on stem cells, a major message can be distilled: namely, the vast difference that currently exists in our understanding of, and the potential utility of, embryonic versus adult stem cells as sources of material for research and clinical purposes. In brief, human stem cells have been isolated from a variety of embryonic, fetal, and adult tissue sources. However, enormous differences exist in purity, properties, data reproducibility, and understanding of cells from these different sources. Much of our ignorance is related to the relative paucity of funding for research using embryonic stem cells. Years of rigorous and careful research in animal models have documented that embryonic stem cells have great utility for scientific studies. This work has also rigorously and reproducibly established the great plasticity of these cells and supports the opinion that human embryonic stem cells possess the greatest broadest potential and promise for clinical applications. As well as therapeutic uses, important potential applications include studies of embryonic stem cells bearing complex genotypes susceptible to poorly understood common human diseases and testing and screening drug efficacy. The report does not make clear that the best-characterized adult stem cells are hematopoietic stem cells. Currently, major difficulties and inadequate understanding exist with most other types of adult stem cells reported to date. In addition, many experiments suggesting that adult stem cells have broad plasticity may be incorrectly interpreted owing to an error caused by an experimental artifact of cell fusion present in some unknown proportion of the experiments. Research on some of the reported adult stem cell preparations may conceivably in the future demonstrate that they, too, like hematopoietic stem cells, can also be prospectively identified, "single cell cloned," expanded considerably by growth in vitro with retention of normal chromosome structure and number, and preserved by freezing and storage at low temperatures. But it should be strongly cautioned that this has not been done for most adult stem cell preparations, and, even if possible, it is not clear that any of the just-mentioned procedures will be accomplished in the near future, owing to the technically very demanding nature of such experiments. We feel it is important to emphasize a point that the report mentioned, that the reported isolation and properties of multipotent adult progenitor cells (MAPCs) must be reproduced in additional laboratories for any reliable interpretation of the results reported with these cells. After considerable effort, this has still not been achieved. Thus, in the reported results, the possible significance of the reported isolation and properties of human MAPCs is left unclear, as is their potential as a source of stem cells for clinical purposes. Hence, a strong overall caution is that many of the reports on the properties of cells differentiated from adult stem cell preparations are to date preliminary and incomplete. If results with any isolated and characterized adult stem cells are validated, it will then be very important to compare their properties---and those of any more differentiated cells that can be derived from them---with other stem cell sources, such as the well-characterized hematopoietic stem cells, and with human embryonic stem cell preparations. Two major considerations argue strongly for non-commercial, federal, peer-reviewed funding to be made available for this work. The first is the sustained effort this work will require. The second is the importance of reliable and unbiased design of experiments and of open, public availability of the complete findings. Reasons for Our Concern {#s4} ======================= In being concerned about the content of these reports, neither of which makes any recommendations for legislative or policy actions, are we worrying too much? We think not. Indeed, already, sadly as a result of the way the sections on aging research in the report were written, the myth that longevity has an inevitable tradeoff of diminished fertility is now gaining a further foothold: witness the January 26, 2004, issue of the *The New Republic*. In it, an article about this report of the Council falls right into the trap: it states, "But changes come with longer life. Worms and mice that are altered for extended lifespans become sterile, or barely reproduce." The public is done a disservice when science is presented incompletely; myths are then perpetuated. This is but one example of the dangers that three of the Council members who are scientists (the two of us along with Michael Gazzaniga of Dartmouth College) pointed out, in a Commentary within the edition of the "Beyond Therapy" report published by the Dana Foundation in November 2003. In that Commentary, we stated that "Our concern ... is that, moving forward, the debate carry on with all of the scientific evidence---or as much as such a widespread public discussion can include---and take care not to leave an erroneous impression as to the nature of the potential problems at hand." We ended the Commentary by saying "We urge both good reading and *critical* reading!" (our italics). These reports had as their premise the aim of neutrality in the scientific analysis of the issues addressed. But our concern is that some of their contents, as in the few examples outlined above, may have ended up distorting the potential of biomedical research and the motivation of some of its researchers. Continuing discussions will form the basis for future decisions on these topics; keeping such discussion open and balanced is of paramount importance. ::: {.caption} ###### Box 1. President\'s Council on Bioethics ::: The President\'s Council on Bioethics was created on November 28, 2001. Its mission includes: to "advise the President on bioethical issues that may emerge as a consequence of advances in biomedical science and technology. In connection with its advisory role, the mission of the Council includes the following functions: to undertake fundamental inquiry into the human and moral significance of developments in biomedical and behavioral science and technology;to explore specific ethical and policy questions related to these developments;to provide a forum for a national discussion of bioethical issues;to facilitate a greater understanding of bioethical issues; andto explore possibilities for useful international collaboration on bioethical issues." From Executive Order 13237 George W. Bush The White House, November 28, 2001 Federal Register date: November 30, 2001 Federal Register page: 66 FR 59851 The members of the President\'s Council on Bioethics at the time these reports were written included Leon R. Kass, M.D., Ph.D. (Chair), American Enterprise Institute; Elizabeth H. Blackburn, Ph.D.\*, University of California, San Francisco; Rebecca S. Dresser, J.D., M.S., Washington University School of Law; Daniel W. Foster, M.D., University of Texas, Southwestern Medical School; Francis Fukuyama, Ph.D., Johns Hopkins University; Michael S. Gazzaniga, Ph.D., Dartmouth College; Robert P. George, J.D., D.Phil., Princeton University; Mary Ann Glendon, J.D., M.Comp.L., Harvard University; Alfonso Gómez-Lobo, Dr. Phil., Georgetown University; William B. Hurlbut, M.D., Stanford University; Charles Krauthammer, M.D., syndicated columnist; William F. May, Ph.D.\*, Southern Methodist University; Paul McHugh, M.D., Johns Hopkins Hospital; Gilbert C. Meilaender, Ph.D., Valparaiso University; Janet D. Rowley, M.D., University of Chicago; Michael J. Sandel, D.Phil., Harvard University; and James Q. Wilson, Ph.D., University of California, Los Angeles. \* These members had their Council terms terminated by White House directive on February 27, 2004. Elizabeth Blackburn is a recent member of the President\'s Council on Bioethics and is in the Department of Biochemistry and Biophysics at the University of California, San Francisco, in the United States. Janet Rowley is a current member of the Council and is in the Section of Hematology and Oncology at the University of Chicago in Chicago, Illinois, in the United States. E-mail: <telomer@itsa.ucsf.edu> (EB) MAPC : multipotent adult progenitor cell

PubMed Central

2024-06-05T03:55:46.124918

2004-3-5

PMC368155

During the last few years, we have seen enormous strides in our abilities to sequence genomes, and the information that has poured out of these sequences is quite astonishing. With more than 150 complete genome sequences now available and many laboratories rushing into microarray analysis, proteomic initiatives, and even systems biology, it seems an appropriate time to consider not just the opportunities those sequences present, but also their shortcomings. By far the most serious problem is the quality and degree of completeness of the annotation of those genomes. Most troublesome are the large numbers of open reading frames that have been identified by computer programs, but remain labeled as a "conserved hypothetical protein" when they occur in more than one genome or simply a "hypothetical protein" when they appear unique to the genome in question. Between them, these two categories of annotated open reading frames often represent more than half of the potential protein-coding regions of a genome. These annotations highlight just one portion of our ignorance about the information content of genomes and our lack of fundamental knowledge about the function of so many of the building blocks of cells. Unless we rectify this situation, it is likely to undermine many of the other "-omic" efforts currently underway. Here I advocate a rather straightforward approach to address this problem---focused initially on the bacterial genomes. In contrast to the numerous proposals for big science initiatives to understand the fundamental workings of biological organisms, I propose a small science, relatively low-tech approach that could have a dramatic pay off. A relatively small investment could yield a massive amount of information that would greatly enhance our current efforts to use genomic approaches to study life. Initial Proposal {#s2} ================ The initial proposal is directed at deciphering the role of the "hypothetical proteins" encoded in the microbial genomes and would involve a community-wide approach to determine the function of these hypotheticals based on solid, old-fashioned biochemistry. The essence of the idea is to undertake an interdisciplinary effort that couples our current bioinformatics capabilities to predict protein function with a directed exploration by experimental laboratories to test those predictions. I would encourage a consortium of bioinformaticians to produce a list of all of the conserved hypothetical proteins that are found in multiple genomes, to carry out the best possible bioinformatics analysis, and then to offer those proteins to the biochemical community as potential targets for research into their function. To energize laboratories with appropriate expertise to participate in this community-wide effort, I suggest that a special program be set up by one or more of the funding agencies so that laboratories undertaking the investigation of any particular protein receive a small grant upfront as a supplement to an existing grant. Upon completion of the project and the identification of the function, they would receive a further supplement to that grant as a reward. In this way, one might hope to rally some of the best biochemical talent and apply it to this problem of determining function for a wide range of new proteins. The cost of such an operation could be quite minimal, and the bureaucracy and review process could be equally simple. Here is a case where a modest infusion of funds could greatly enhance our ability to annotate both existing and new genome sequences and ensure that our current investments in genomic sequences yield the richest biological harvest possible. There are two key steps in the proposed plan. Key Steps {#s3} ========= The first step is to encourage some bioinformaticians with appropriate expertise in the functional annotation of genomes to form a consortium and undertake the assembly of a list of prime targets for which an experimental demonstration of function would be most valuable. Three general classes of such genes come to mind: (1) The conserved hypothetical genes. These belong to the set of genes that have orthologs in many other genomes, but for which no function has been experimentally determined in any case. A recent success among such genes is illustrated in [Box 1](#box1){ref-type="boxed-text"}. (2) The hypothetical genes. These form the set of genes that are predicted to be protein coding, but that lack similar genes in any other organism in GenBank. They, too, have no assigned function. (3) The misannotated genes. These genes are ones for which a function has been assigned, but for which there is a good reason to believe the annotation is incorrect. These sets of targets would be combined and arranged into a prioritized list in which each was accompanied by the best assessment of potential function. The priorities would be based on which genes were most likely to prove broadly informative. For instance, a conserved hypothetical gene that occurred in most genomes would be of higher priority than one that had only two orthologs. The list would be on a public Web site where these targets and the predicted functions could be examined and modified by alternative or additional predictions from other groups to guide future experimentation. As function was derived, that information could be presented and the target removed from the main list. The second step would be to invite experimentalists to peruse the list and find those potential genes whose protein products might lie within their realm of expertise so that they could use their experimental knowledge and reagents to quickly test for function. Initially, I would advocate allowing laboratory teams to pick and choose among the list and sign up to study just one of these open reading frames. I would recommend allowing one laboratory per open reading frame in the initial stages. A laboratory wishing to sign up would generate a short document highlighting why its expertise might be suitable for a particular protein. A one-page proposal should suffice, with no experimental plan demanded. At this point, a small panel could choose among competing efforts and the laboratory chosen would be given a small grant and up to six months to carry out its analysis. If it was successful in delineating the function of their target protein, a paper would be written and submitted for peer review. If the paper was accepted for publication, then an additional sum would be allocated as a supplement to the laboratory\'s existing grant. If, after six months, a laboratory had not managed to delineate the function, it would submit a short report describing the approaches that have been tried, with the results of its analyses. This would be posted on the public Web site and that target would then become open for analysis by other laboratories, under the same conditions as before. While the initial list of target genes should probably be based on a well-studied and experimentally tractable organism such as Escherichia coli, I would not demand that the biochemical experiments be done on the E. coli gene. Any of the orthologs would do, so long as the similarity was sufficiently strong to give high expectations that function would be conserved. In fact, for a laboratory that happened to be already working on one of the homologs, this program might provide an added bonus and greatly speed its work. I would also encourage both biochemical and genetic approaches, since one can never be certain when one method might be better than another. The list would, of course, also include conserved genes not found in E. coli, but commonly distributed in other genomes. In particular, I would make a pitch for including all genes in Mycoplasma genitalium, which, as the free-living organism with the fewest genes, might be the most suitable as a model system for in-depth understanding of its biology. The Importance of Community {#s4} =========================== This proposal for experimental attack on hypothetical genes is really a very traditional approach that becomes large-scale simply because of the parallel nature of the implementation. It resembles the successful approach used by the Europeans to achieve the complete sequence of the Saccharomyces cerevisiae genome ([@pbio-0020042-Goffeau1]). The results would significantly increase our functional knowledge of the genes within the microbial genomes thus far sequenced. Such annotation would be immediately applicable across orthologs and could dramatically improve the value of the sequenced genomes. This, in turn, would facilitate our ability to annotate new genomes as they appear. The proposal also reinforces the notion that the overwhelming value of bioinformatics is to generate hypotheses that can be tested experimentally. By enabling the community to join in this effort, we would also demonstrate that science really is the collaborative enterprise that requires all of our contributions, not just a select few. Finally, if this initiative succeeds, it would serve as a suitable model from which to begin the more daunting task of trying to annotate the functions of the complex eukaryotic genomes, such as the human genome. ::: {.caption} ###### Box 1. HemK, a Very Highly Conserved Protein Methyltransferase ::: During studies of the genetics of heme biosynthesis in E. coli, one gene, *hemK*, was found that had no immediately known protein product ([@pbio-0020042-Nakayashiki1]). Since one of the missing biosynthetic activities required for heme biosynthesis is protoporphyrinogen oxidase, the original authors suggested that the *hemK* gene might encode this enzyme. Subsequent *hemK* homologs were annotated as putative protoporphyrinogen oxidases and soon the "putative" was dropped. Then one group noticed that the gene product contained protein sequence motifs typical of DNA adenine methyltransferases. From then on, the annotations in GenBank alternated between these two assignments. This *hemK* gene is of ubiquitous occurrence from humans to Chlamydia, and yet until 2002 its true biochemical function was unknown. At that point, two groups ([@pbio-0020042-Heurgue-Hamard1]; [@pbio-0020042-Nakahigashi1]) demonstrated that neither previous assignment was correct. Instead, they found that in E. coli the HemK gene product was an N5 glutamine methyltransferase that transferred a methyl group from S-adenosylmethionine to the amide nitrogen of a specific glutamine residue in the protein chain release factors *prfA* and *prfB*. Particularly noteworthy was the observation that the *hemK* gene is positioned immediately adjacent to the *prfA* gene in many microbial genomes! Here is a case where bioinformatics suggested strongly that the *hemK* gene encoded a methyltransferase, but an experiment was needed to identify the substrate. Since the adjacent gene in the genome encoded the substrate, it might have been possible to make that prediction too. There are additional paralogs of *hemK* in several genomes, but their biochemical activity and substrates remain to be identified. Richard J. Roberts is a Nobel Laureate and a research director at New England Biolabs, located in Beverly, Massachusetts, United States of America. E-mail: <roberts@neb.com>

PubMed Central

2024-06-05T03:55:46.126462

2004-3-16