The standard reference Caenorhabditis elegans strain, N2, has evolved marked behavioral changes in social feeding behavior since its isolation from the wild. We show that the causal, laboratory-derived mutations in two genes, npr-1 and glb-5, confer large fitness advantages in standard laboratory conditions. Using environmental manipulations that suppress social/solitary behavior differences, we show the fitness advantages of the derived alleles remained unchanged, suggesting selection on these alleles acted through pleiotropic traits. Transcriptomics, developmental timing, and food consumption assays showed that N2 animals mature faster, produce more sperm, and consume more food than a strain containing ancestral alleles of these genes regardless of behavioral strategies. Our data suggest that the pleiotropic effects of glb-5 and npr-1 are a consequence of changes to O2 -sensing neurons that regulate both aerotaxis and energy homeostasis. Our results demonstrate how pleiotropy can lead to profound behavioral changes in a popular laboratory model.
Presented on January 10, 2017 from 8:30 a.m.-9:30 a.m. at the Parker H. Petit Institute for Bioengineering and Bioscience (IBB), room 1128, Georgia Tech.
Over long evolutionary timescales, major changes to the copy number, function, and genomic organization of genes occur, however, our understanding of the individual mutational events responsible for these changes is lacking. In this report, we study the genetic basis of adaptation of two strains of C. elegans to laboratory food sources using competition experiments on a panel of 89 recombinant inbred lines (RIL). Unexpectedly, we identified a single RIL with higher relative fitness than either of the parental strains. This strain also displayed a novel behavioral phenotype, resulting in higher propensity to explore bacterial lawns. Using bulk-segregant analysis and short-read resequencing of this RIL, we mapped the change in exploration behavior to a spontaneous, complex rearrangement of the rcan-1 gene that occurred during construction of the RIL panel. We resolved this rearrangement into five unique tandem inversion/duplications using Oxford Nanopore long-read sequencing. rcan-1 encodes an ortholog to human RCAN1/DSCR1 calcipressin gene, which has been implicated as a causal gene for Down syndrome. The genomic rearrangement in rcan-1 creates two complete and two truncated versions of the rcan-1 coding region, with a variety of modified 5' and 3' non-coding regions. While most copy-number variations (CNVs) are thought to act by increasing expression of duplicated genes, these changes to rcan-1 ultimately result in the reduction of its whole-body expression due to changes in the upstream regions. By backcrossing this rearrangement into a common genetic background to create a near isogenic line (NIL), we demonstrate that both the competitive advantage and exploration behavioral changes are linked to this complex genetic variant. This NIL strain does not phenocopy a strain containing an rcan-1 loss-of-function allele, which suggests that the residual expression of rcan-1 is necessary for its fitness effects. Our results demonstrate how colonization of new environments, such as those encountered in the laboratory, can create evolutionary pressure to modify gene function. This evolutionary mismatch can be resolved by an unexpectedly complex genetic change that simultaneously duplicates and diversifies a gene into two uniquely regulated genes. Our work shows how complex rearrangements can act to modify gene expression in ways besides increased gene dosage.
SUMMARY Foraging strategies that enable animals to locate food efficiently are composed of highly conserved behavioral states with characteristic features. Here, we identify parallel multimodal circuit modules that control an innate foraging state -- local search behavior -- after food removal in the nematode Caenorhabditis elegans . Two parallel groups of chemosensory and mechanosensory glutamatergic neurons that detect food-related cues trigger local search by inhibiting separate integrating neurons through a metabotropic glutamate receptor, MGL-1. The chemosensory and mechanosensory modules are separate and redundant, as glutamate release from either can drive the full behavior. Spontaneous activity in the chemosensory module encodes information about the time since the last food encounter and correlates with the foraging behavior. In addition, the ability of the sensory modules to control local search is gated by the internal nutritional state of the animal. This multimodal circuit configuration provides robust control of an innate adaptive behavior.
Natural isolates of C. elegans differ in their sensitivity to pheromones that inhibit exploratory behavior. Previous studies identified a QTL for pheromone sensitivity that includes alternative alleles of srx-43, a chemoreceptor that inhibits exploration through its activity in ASI sensory neurons. Here we show that the QTL is multigenic and includes alternative alleles of srx-44, a second chemoreceptor gene that modifies pheromone sensitivity. srx-44 either promotes or inhibits exploration depending on its expression in the ASJ or ADL sensory neurons, respectively. Naturally occurring pheromone insensitivity results in part from previously described changes in srx-43 expression levels, and in part from increased srx-44 expression in ASJ, which antagonizes ASI and ADL. Antagonism between the sensory neurons results in cellular epistasis that is reflected in their transcription of insulin genes that regulate exploration. These results and genome-wide evidence suggest that chemoreceptor genes may be preferred sites of adaptive variation in C. elegans.
Abstract Social behaviors are diverse in nature, but it is unclear how conserved genes, brain regions, and cell populations generate this diversity. Here we investigate bower-building, a recently-evolved social behavior in cichlid fishes. We use single nucleus RNA-sequencing in 38 individuals to show signatures of recent behavior in specific neuronal populations, and building-associated rebalancing of neuronal proportions in the putative homolog of the hippocampal formation. Using comparative genomics across 27 species, we trace bower-associated genome evolution to a subpopulation of glia lining the dorsal telencephalon. We show evidence that building-associated neural activity and a departure from quiescence in this glial subpopulation together regulate hippocampal-like neuronal rebalancing. Our work links behavior-associated genomic variation to specific brain cell types and their functions, and suggests a social behavior has evolved through changes in glia.
Abstract Species inhabit a variety of environmental niches, and the adaptation to a particular niche is often controlled by genetic factors, including gene-by-environment interactions. The genes that vary in order to regulate the ability to colonize a niche are often difficult to identify, especially in the context of complex ecological systems and in experimentally uncontrolled natural environments. Quantitative genetic approaches provide an opportunity to investigate correlations between genetic factors and environmental parameters that might define a niche. Previously, we have shown how a collection of 208 whole-genome sequenced wild Caenorhabditis elegans can facilitate association mapping approaches. To correlate climate parameters with the variation found in this collection of wild strains, we used geographic data to exhaustively curate daily weather measurements in short-term (3 month), middle-term (one year), and long-term (three year) durations surrounding the date of strain isolation. These climate parameters were used as quantitative traits in association mapping approaches, where we identified 11 quantitative trait loci (QTL) for three climatic variables: elevation, relative humidity, and average temperature. We then narrowed the genomic interval of interest to identify gene candidates with variants potentially underlying phenotypic differences. Additionally, we performed two-strain competition assays at high and low temperatures to validate a QTL that could underlie adaptation to temperature and found suggestive evidence supporting that hypothesis.
Significance We do not fully understand how behavior evolves. Here we investigate the genomic basis of bower building among Lake Malawi cichlid fishes. Males construct bowers of two major types, pits and castles, to attract females in mating displays. Thousands of genetic variants are strongly associated with divergence in bower behavior. Remarkably, F 1 hybrids of pit-digging and castle-building species perform sequential construction of first pit and then castle bowers. Analysis of brain gene expression in hybrids showed behavior-dependent allele-specific expression with preferential expression of pit-digging alleles during pit digging and castle-building alleles during castle building. Our results suggest that behaviors evolve via complex genetic architectures featuring cis -regulatory differences whose effects on gene expression are specific and context dependent.
Temporally and spatially controlled master regulators drive the Caulobacter cell cycle by regulating the expression of >200 genes. Rapid clearance of the master regulator, CtrA, by the ClpXP protease is a critical event that enables the initiation of chromosome replication at specific times in the cell cycle. We show here that a previously unidentified single domain-response regulator, CpdR, when in the unphosphorylated state, binds to ClpXP and, thereby, causes its localization to the cell pole. We further show that ClpXP localization is required for CtrA proteolysis. When CpdR is phosphorylated, ClpXP is delocalized, and CtrA is not degraded. Both CtrA and CpdR are phosphorylated via the same CckA histidine kinase phospho-signaling pathway, providing a reinforcing mechanism that simultaneously activates CtrA and prevents its degradation by delocalizing the CpdR/ClpXP complex. In swarmer cells, CpdR is in the phosphorylated state, thus preventing ClpXP localization and CtrA degradation. As swarmer cells differentiate into stalked cells (G1/S transition), unphosphorylated CpdR accumulates and is localized to the stalked cell pole, where it enables ClpXP localization and CtrA proteolysis, allowing the initiation of DNA replication. Dynamic protease localization mediated by a phospho-signaling pathway is a novel mechanism to integrate spatial and temporal control of bacterial cell cycle progression.
Article Figures and data Abstract Introduction Results Discussion Materials and methods References Decision letter Author response Article and author information Metrics Abstract Natural isolates of C. elegans differ in their sensitivity to pheromones that inhibit exploratory behavior. Previous studies identified a QTL for pheromone sensitivity that includes alternative alleles of srx-43, a chemoreceptor that inhibits exploration through its activity in ASI sensory neurons. Here we show that the QTL is multigenic and includes alternative alleles of srx-44, a second chemoreceptor gene that modifies pheromone sensitivity. srx-44 either promotes or inhibits exploration depending on its expression in the ASJ or ADL sensory neurons, respectively. Naturally occurring pheromone insensitivity results in part from previously described changes in srx-43 expression levels, and in part from increased srx-44 expression in ASJ, which antagonizes ASI and ADL. Antagonism between the sensory neurons results in cellular epistasis that is reflected in their transcription of insulin genes that regulate exploration. These results and genome-wide evidence suggest that chemoreceptor genes may be preferred sites of adaptive variation in C. elegans. https://doi.org/10.7554/eLife.21454.001 Introduction Social communication, particularly chemical communication with pheromones, is broadly used throughout nature to organize behavior. Pheromones and the responses that they drive are highly diverse within and between species. The detection of diverse pheromones by large families of finely-tuned G protein-coupled receptors is well-recognized, but it is less clear how these different receptors drive unique, overlapping, and evolving behavioral and physiological responses (Yang and Shah, 2016). The nematode C. elegans communicates through the secretion and detection of pheromones called ascarosides, which reflect population density and vary with an animal's sex, developmental stage, and feeding status (Izrayelit et al., 2012; Jeong et al., 2005; Kaplan et al., 2011). At least four classes of sensory neurons detect ascarosides; these can induce entry into and exit from the developmental dauer stage, aggregation into feeding groups, male attraction to hermaphrodite mating partners, or acute avoidance (Jang et al., 2012; Kim et al., 2009; McGrath et al., 2011; Srinivasan et al., 2008). In addition, a subset of ascarosides potently modifies innate foraging behaviors. On a lawn of bacterial food, C. elegans spontaneously alternates between minute-long foraging states called roaming and dwelling (Ben Arous et al., 2009; Flavell et al., 2013; Fujiwara et al., 2002). Roaming animals move quickly to explore a large area, while dwelling animals move slowly and reverse frequently to exploit local food resources. A number of ascarosides shorten roaming states to suppress exploration at high population density, including the indolated ascaroside icas#9 (also known as IC-asc-C5 or C5)(Butcher et al., 2009; Greene et al., 2016). icas#9 is detected by the G protein-coupled chemoreceptor SRX-43 in ASI sensory neurons, which are components of the distributed neuromodulatory circuit that regulates roaming and dwelling (Flavell et al., 2013; Greene et al., 2016). Some wild-type strains including the German strain MY14 have reduced sensitivity to icas#9 due to variation at roam-1, a 37 kb Quantitative Trait Locus (QTL) that includes the srx-43 gene (Greene et al., 2016). icas#9 insensitivity results in part from reduced srx-43 expression in MY14 compared to N2. Both MY14-like and N2-like roam-1 alleles are distributed across four continents, and population genetic studies suggest that the roam-1 QTL is under balancing selection, such that both alleles are actively maintained in wild populations. Although the natural ecology of C. elegans is incompletely understood, laboratory competition studies indicate that alternative roam-1 alleles can be favored depending on the presence of pheromones and the distribution of food, consistent with balancing selection on this locus. The genetic architecture of individual differences in behavior is a central element of neurogenetics that is only beginning to be understood (Greenspan, 2009). Many studies have suggested that individual mutations with small effect sizes interact to generate natural variation, but counterexamples of large-effect single genes have been described as well (Buchner and Nadeau, 2015; Greenspan, 2009). The contribution of srx-43 to foraging appeared to represent a case in which a single gene has a large effect on a behavioral trait, as the roam-1 QTL explains about 40% of the genetic variance in icas#9 sensitivity. Here we show that the previously identified roam-1 QTL is more complex than suggested by initial studies: it reflects changes both in srx-43 and in an adjacent gene, srx-44. The reduced icas#9 sensitivity in strains bearing the roam-1MY14 QTL arises from reduced srx-43 expression in ASI neurons, and acquisition of srx-44 expression in ASJ sensory neurons. Natural behavioral variation results from the remapping of these two chemoreceptor genes across multiple sensory neurons with antagonistic actions. Results srx-44 increases icas#9 sensitivity in N2 and decreases sensitivity in roam-1MY14 The ascaroside pheromone icas#9 strongly suppresses exploration behavior in the N2 laboratory strain and in most wild strains, as determined by measuring the area a single animal explores on a bacterial lawn in an overnight assay (Figure 1). By contrast, icas#9 does not suppress exploration in animals from the wild MY14 strain or in the Near Isogenic Line (NIL) roam-1MY14, which bears a 182 kb region from MY14 in an otherwise N2 genetic background (Greene et al., 2016) (Figure 2A,B). The srx-43 gene, which is strongly expressed in N2 and weakly expressed in MY14, is an essential element of the roam-1 QTL (Greene et al., 2016). We found that the N2 roam-1 allele was recessive to the MY14 allele: F1 progeny of N2 animals crossed with the roam-1MY14 NIL were insensitive to icas#9 (Figure 2B). However, the F1 progeny of N2 animals crossed with an srx-43 null mutant were sensitive to icas#9, like N2, indicating that an srx-43(lf) mutation is recessive (Figure 2B). The dominance of roam-1MY14 over N2 indicates that the roam-1 QTL is not entirely explained by reduced srx-43 expression in roam-1MY14. Figure 1 Download asset Open asset Exploration assays. (A) Individual animals are allowed to explore a thinly seeded 3.5 cm plate for 16 hr, after which exploration is scored by placing the plate on a grid and counting squares with tracks (Flavell et al., 2013). Diagram from Greene et al., 2016. (B) A pheromone response for each animal on an ascaroside plate was determined with respect to the behavior of control animals that were tested on ascaroside-free plates on the same day. Data are representative values for one day of testing; boxes indicate the mean ± SEM. Below, calculating individual and group pheromone response. https://doi.org/10.7554/eLife.21454.002 Figure 1—source data 1 Individual exploration assays in Figure 1B. https://doi.org/10.7554/eLife.21454.003 Download elife-21454-fig1-data1-v1.xlsx Figure 2 Download asset Open asset srx-43 and srx-44 both influence icas#9 sensitivity. (A) The roam-1 locus. Boxes indicate genomic regions used for srx-43 and srx-44 transgenes. (B) Dominance tests. Pheromone response of parental strains and of the F1 progeny from crosses between N2 and roam-1MY14 or between N2 and N2 srx-43(lf). (C) srx-43 variation is insufficient to explain the roam-1 QTL. (D) Pheromone response of srx-44 loss-of-function mutants. (E) Complementation test between srx-43 and roam-1MY14. Pheromone response of parental strains and of F1 progeny from crosses between roam-1MY14 srx-44(lf) and N2 or N2 srx-43(lf). (F) Reciprocal hemizygosity test for srx-44. Left, pheromone response of the F1 progeny from crosses between N2 and roam-1MY14 srx-44(lf) and between N2 srx-44(lf) and roam-1. Center and right, hemizygosity for srx-44 in the parental strains did not affect behavior. For Figure B–F, boxes indicate the mean pheromone response ± SEM. Box color indicates genotype at the roam-1 locus (Red = N2; Blue = MY14; red and blue = heterozygous). Cartoons of the roam-1 locus show endogenous srx-43 (B–F) and srx-44 (D–F) with the same color code. X indicates null allele. In C, 'MosSCI srx-43' indicates strains with a chromosome II Mos1 Single Copy Inserted srx-43 allele from N2.. ***p<0.001,**p<0.01, *p<0.05; ns, not significant by ANOVA with Dunnett correction (A, D- roam-1MY14, E) or t-test (C, D- N2, F). https://doi.org/10.7554/eLife.21454.004 Figure 2—source data 1 Individual exploration assays in Figure 2B–F. https://doi.org/10.7554/eLife.21454.005 Download elife-21454-fig2-data1-v1.xlsx These results prompted more detailed analysis of the roam-1 QTL. The endogenous srx-43 locus was inactivated by null mutations in N2 and in the roam-1MY14 N2 NIL, resulting in profound icas#9 insensitivity, and rescued with a single-copy N2 srx-43 gene targeted to a different chromosome. The N2 srx-43 single copy transgene restored icas#9 sensitivity to both N2 and roam-1MY14 srx-43 mutants (Figure 2C), but the roam-1MY14 strain was less sensitive to icas#9 than its N2 counterpart (Figure 2C). As both of these strains bear the same N2 srx-43 transgene, these experiments confirm that the difference in srx-43 function cannot fully account for the roam-1 behavioral phenotype. The srx-44 gene was an attractive candidate for a second locus, as srx-43 and srx-44 are adjacent, closely-related paralogs in a large chemoreceptor gene family that might be expected to have related functions. In addition, both genes fall within a 20 kb region of the minimal 37 kb roam-1 QTL that has exceptionally high sequence divergence between N2 and MY14. Protein-terminating srx-44(lf) alleles were generated by CRISPR-cas9 mutagenesis in N2 and roam-1MY14 strains. In each case, srx-44(lf) resulted in an icas#9 response that was intermediate between those of N2 and roam-1MY14 (Figure 2D). Since N2 became less icas#9 sensitive in the srx-44 null, and roam-1MY14 became more sensitive, the alternative srx-44 alleles have opposite effects on icas#9 sensitivity. The effects of srx-44(lf) mutations were weaker than those of srx-43(lf) (Figure 2C,D), which suggest that srx-44 acts a modifier gene. We examined this possibility by generating a roam-1MY14 srx-43(lf) srx-44(lf) strain by CRISPR-Cas9 mutagenesis. The resulting double mutant was insensitive to icas#9, like srx-43(lf) (Figure 2D). These results suggest that srx-43 is essential for icas#9 sensitivity in all genetic backgrounds, whereas srx-44 modifies icas#9 response in opposite directions depending on the roam-1 allele. With this additional information, genetic tests were conducted to assess the contributions of srx-43 and srx-44 to the roam-1 QTL. The F1 progeny of a roam-1MY14 srx-44(lf) cross with N2 resembled N2 (Figure 2E), indicating that srx-44 is necessary for the dominance of the roam-1MY14 QTL. The F1 progeny of a cross between roam-1MY14 srx-44(lf) and N2 srx-43(lf) were insensitive to icas#9 (Figure 2E). The failure of the roam-1MY14 srx-43 allele to complement N2 srx-43(lf) confirmed reduced srx-43 function in roam-1MY14. To test for altered srx-44 function, a reciprocal hemizygosity test was conducted: The F1 progeny of N2 srx-44(lf) crossed with roam-1MY14 were compared to the F1 progeny of N2 crossed with roam-1MY14 srx-44(lf). The resulting hemizygotes differed genetically only in whether their single functional srx-44 allele derived from N2 or roam-1MY14 and demonstrated significantly different icas#9 sensitivity (Figure 2F), indicating that srx-44 variation in the roam-1 QTL affects foraging behavior. Sequence variation in the srx-44 promoter alters pheromone response As existing sequence resources underestimated the exceptionally high sequence divergence between N2 and MY14 srx-43 alleles (19.7%), we examined srx-44 by Sanger resequencing. This analysis indicated that MY14 and N2 srx-44 sequences differed by 5.0% across the 2.8 Kb promoter and coding region of the srx-44 gene, approximately 25 times the genome-wide average (Thompson et al., 2015). A phylogeny of srx-44 and related genes confirmed that srx-44 in N2 and MY14 are alleles of one orthologous gene (Figure 3F). Despite this high level of divergence, only five polymorphisms alter amino acids between N2 and MY14 srx-44 alleles, suggesting that the coding region remains under purifying selection (dN/dS = 0.071; five non-synonymous mutations, 20 synonymous mutations). Figure 3 Download asset Open asset The proximal promoter sequence underlies altered srx-44 activity in MY14. (A) Cartoon of srx-44 transgenes (corresponds to grey box in Figure 2A). (B) N2 srx-44 transgenes confer icas#9 sensitivity on roam-1MY14. Similar transgenes bearing a frameshift in the coding region do not, nor do MY14 srx-44 transgenes. (C) srx-44 transgenes with a N2 promoter confer icas#9 sensitivity on roam-1MY14, whereas srx-44 transgenes with a MY14 promoter do not. (D) The proximal promoter element accounts for N2 srx-44 activity in the transgene assay. (E) icas#9 responses of Allele Replacement Lines for the srx-44 proximal promoter element, generated by homologous recombination in the endogenous genomic locus with CRISPR/Cas9, localize activity to the proximal promotor element. (F) Phylogeny constructed for the coding sequence of srx-43, srx-44 and related genes in C. elegans, C. briggsae, and C. remanei. The srx-44 alleles in N2 and MY14 are closely related, confirming they are alleles of a single gene. Genes are color coded by species (green = C. elegans, blue = C. briggsae, orange = C. remanei). Protein sequences and gene names are from Thomas and Robertson (2008). For B-E, boxes indicate the mean pheromone response ± SEM. Box color in the data panels indicates genotype at the genomic roam-1 locus (Red = N2; Blue = MY14). Transgenes and allele replacements are indicated in cartoons with the same color code. ***p<0.001,**p<0.01,* p<0.05, ns = not significant by ANOVA with Dunnett correction (B, C, D) or by t test (E). https://doi.org/10.7554/eLife.21454.006 Figure 3—source data 1 Individual exploration assays in Figure 3B–E. https://doi.org/10.7554/eLife.21454.007 Download elife-21454-fig3-data1-v1.xlsx To localize the biologically relevant sequence changes between N2 and MY14 srx-44 genes, we tested transgenes with N2 and roam-1MY14 sequences for their biological activity. Although roam-1MY14 was dominant to N2 in an F1 cross, high-copy transgenes that expressed the N2 srx-44 gene were able to confer pheromone sensitivity on the roam-1MY14 strain (Figure 3A and B). This result suggests that the N2 and MY14 srx-44 alleles have antagonistic activities. Transgenes containing an N2 srx-44 gene with an early nonsense mutation or a MY14 srx-44 gene did not confer pheromone sensitivity (Figure 3B). Exchanging the promoters and coding regions of srx-44 showed that transgenes with the N2 srx-44 promoter conferred icas#9 sensitivity to roam-1MY14 regardless of whether the coding region was from N2 or MY14, whereas transgenes with the MY14 promoter did not enhance pheromone sensitivity (Figure 3C). Therefore, promoter sequences account for differential activity of N2 and MY14 srx-44 genes in this assay. N2 and MY14 differ at 35 of the 515 bases between the start codon of srx-44 and the end of the adjacent srx-43 gene. Nine of these changes cluster in a region 34–72 bp upstream of the start codon of srx-44. Exchanging just this proximal promoter sequence in N2 and MY14 srx-44 transgenes switched their activity (Figure 3D). To confirm the biological importance of this potential regulatory site, we precisely exchanged the sequences 34–72 bp upstream of the srx-44 start codon at the endogenous genomic loci of N2 and roam-1MY14 using oligonucleotide-templated homologous recombination with CRISPR/Cas9. Introducing the MY14 srx-44 proximal promoter element into the N2 genome reduced response to icas#9 (Figure 3E). Conversely, introducing the N2 proximal promoter element for srx-44 strongly enhanced the icas#9 sensitivity of roam-1MY14 (Figure 3E). These result match the genetic predictions for an exchange of N2 and roam-1MY14 srx-44 alleles, and localize altered icas#9 sensitivity to the proximal promoter sequences of srx-44. srx-44 acts in different neurons to promote or inhibit icas#9 sensitivity To understand how altered srx-44 activity was conferred by the promoter sequence, we examined the expression of srx-44 genomic sequences linked to the green fluorescent protein (GFP) in bicistronic transcripts. Transgenes bearing the N2 promoter sequence drove GFP expression selectively in the two ADL sensory neurons, while transgenes bearing the MY14 promoter sequence drove GFP in the ADL and ASJ sensory neurons (Figure 4A). The cell type specificity of srx-44 expression was determined by the proximal promoter element that functionally differentiated N2 and MY14 srx-44 genes: exchanging this promoter element between N2 and MY14 transgenes resulted in GFP reporter gene expression that matched the proximal element (ADL for N2, ADL and ASJ for MY14) (Figure 4A). These experiments localize both biological activity and expression to a proximal region 34–72 bp upstream of srx-44. Figure 4 Download asset Open asset srx-44 site of expression determines whether it potentiates (ADL) or suppresses (ASJ) behavioral response to icas#9. (A) Expression of GFP from srx-44 reporter genes. Images show overlaid fluorescence and differential interference contrast images of the anterior ganglion, with anterior at the left and dorsal up. Arrowheads indicate ADL, arrows indicate ASJ. Cartoons show transgenes tested in each image. Top left = Psrx-44(N2)::GFP; Top right = Psrx-44(MY14)::GFP. Bottom left = Psrx-44(N2 distal promoter/MY14proximal promoter)::GFP; Bottom right = Psrx-44(MY14 distal promoter/N2 proximal promoter)::GFP. (B) Transgenes expressing srx-44 under ASJ- or ADL-specific promoters oppositely affect pheromone responses. (C) Tetanus toxin light chain (TeTx) inhibition of neurotransmitter and neuropeptide release from ASJ or ADL affects icas#9 sensitivity in an srx-44 dependent manner. (D) TeTx inhibition of ASI, but not ADL or ASJ, reduces exploration on pheromone-free control plates, expressed as mean squares entered ± SEM. Boxes indicate the mean ± SEM. Box color indicates genotype at the roam-1 locus (Red = N2; Blue = MY14). All data presented as mean ± SEM. ***p<0.001,**p<0.01, ns = not significant by t test (B, C- N2 and roam-1MY14, D) or by ANOVA with Dunnett correction (C- N2 srx-44(lf)). https://doi.org/10.7554/eLife.21454.008 Figure 4—source data 1 Individual exploration assays in Figure 4B–D. https://doi.org/10.7554/eLife.21454.009 Download elife-21454-fig4-data1-v1.xlsx Since N2 and MY14 srx-44 alleles have opposite effects on foraging behavior, both sites of expression are likely to be functionally important: srx-44 might decrease icas#9 sensitivity when expressed in ASJ and increase icas#9 sensitivity when expressed in ADL. We tested this hypothesis by expressing srx-44 selectively in either ASJ or ADL neurons under the control of cell-specific promoters. Expression of srx-44 under an ASJ-specific srh-11 promoter reduced icas#9 responses in N2 (Figure 4B), supporting a suppressive activity in ASJ. Overexpression of srx-44 under ADL-specific sre-1 or srh-220 promoters increased icas#9 responses in roam-1MY14 (Figure 4B), supporting a positive role in ADL even in the presence of antagonistic ASJ expression in roam-1MY14. This effect of ADL overexpression resembles the effect of high-copy N2 srx-44 transgenes, which enhance icas#9 sensitivity in roam-1MY14 (Figure 3B). Neither the ADL neurons nor the ASJ neurons have previously been implicated in foraging behavior. As an independent way to interrogate their functions, we expressed the tetanus toxin light chain, which cleaves synaptobrevin to reduce neurotransmitter and neuropeptide secretion. ADL::tetanus toxin reduced the response to icas#9 in N2 (Figure 4C), indicating that vesicle release from ADL enhances N2 pheromone sensitivity. ASJ::tetanus toxin increased icas#9 responses in roam-1MY14 (Figure 4C), indicating that vesicle release from ASJ antagonizes pheromone sensitivity. In srx-44(lf) mutants, neither ADL::tetanus toxin nor ASJ::tetanus toxin affected behavioral responses to icas#9, indicating that icas#9 acts through the SRX-44 chemoreceptor in ADL and ASJ. Notably, ASI::tetanus toxin reduced exploration in the absence of pheromones, whereas ADL::tetanus toxin or ASJ::tetanus toxin had no significant effect (Figure 4D). These results suggest that ASI is necessary for maintaining basal levels of exploration, in agreement with previous results (Flavell et al., 2013), whereas ADL and ASJ activity are relevant only in the presence of icas#9. These results also show that while srx-43 activity in ASI promotes pheromone response and srx-44 activity in ASJ inhibits pheromone response, both ASI and ASJ induce roaming through vesicle release (Figure 5H). Icas#9 influences behavior through TGF-β and insulin signaling pathways Early in development, the ASI and ASJ neurons regulate the decision to enter the dauer larva stage by secreting insulin-related peptides whose transcription is regulated by ascarosides (Li et al., 2003). Insulin signaling also promotes roaming (Ben Arous et al., 2009), suggesting that insulins could be effectors of ASI and ASJ functions. To ask whether insulins are regulated by icas#9, we examined the expression of the insulin gene daf-28 with integrated daf-28::GFP reporters that are expressed in both ASI and ASJ neurons. Animals were treated with icas#9 in the same protocol used in behavioral assays, and then examined using quantitative microscopy. In N2 animals, icas#9 decreased daf-28::GFP by 37% in ASI and by 28% in ASJ (Figure 5A). In roam-1MY14 animals, icas#9 did not suppress daf-28::GFP expression, which also appeared elevated at baseline in the absence of pheromones (Figure 5B). Thus the expression of daf-28::GFP can be regulated by icas#9, with differential sensitivity in the N2 and MY14 alleles of the roam-1 QTL. Figure 5 with 1 supplement see all Download asset Open asset srx-43 and srx-44 regulate insulin and TGF-β endocrine signaling pathways. (A–D) Effect of icas#9 on ASI and ASJ daf-28::GFP expression in N2, roam-1MY14, N2 srx-43(lf), and roam-1MY14 srx-44(lf) animals. Bars indicate mean fluorescence intensity ± SEM. (E) Exploration in the absence of pheromones in insulin pathway mutants, expressed as mean squares entered ± SEM. ***p<0.001, ns = not significant by ANOVA with Dunnett correction. (F) Pheromone response of insulin pathway mutants expressed as mean ± SEM. (G) daf-12, a convergence point of daf-7 (TGF-beta) and daf-28 (insulin) signaling in development, is necessary for icas#9 modulation of foraging behavior. Bars indicate icas#9 response expressed as mean ± SEM. (H) Schematic of the proposed relationships between icas#9, the chemoreceptors srx-43 or srx-44, the sensory neurons ASI, ASJ or ADL, alterations in insulin and TGF-beta gene expression, vesicle release, and roaming behavior. SRX-43 and SRX-44 confer sensitivity to icas#9, which leads to changes in gene expression both within the sensing cell and within other sensory neurons. Activation of SRX-43 in ASI reduces the expression of daf-7 (TGF-beta) and daf-28 (insulin) genes as shown in Figure 5A and C and in Greene et al. (2016). The activity of SRX-44 can antagonize this effect as shown in Figure 5B and D. Vesicle release from the sensory neurons, potentially releasing DAF-7, DAF-28, or other neurotransmitters or neuropeptides, can stimulate roaming at baseline (ASI; shown in Figure 4D), stimulate roaming in the presence of icas#9 (ASJ; Figure 4C), or inhibit roaming in the presence of icas#9 (ADL; Figure 4C). Boxes indicate the mean ± SEM, color indicates genotype at roam-1 locus (red = N2, blue = MY14). ***p<0.001,**p<0.01, ns = not significant by t test (A, D) or by ANOVA with Dunnett correction (B, C). https://doi.org/10.7554/eLife.21454.010 Figure 5—source data 1 GFP quantification in ADL and ASJ neurons for Figure 5A–D. https://doi.org/10.7554/eLife.21454.011 Download elife-21454-fig5-data1-v1.xlsx Figure 5—source data 2 Individual exploration assay results for Figure 5E–G. https://doi.org/10.7554/eLife.21454.012 Download elife-21454-fig5-data2-v1.xlsx With this molecular readout of icas#9 signaling, we examined the effects of the srx-43 and srx-44 genes on daf-28 regulation. In N2 srx-43(lf) animals, icas#9 did not decrease daf-28::GFP in either ASI or ASJ, unlike N2 (Figure 5C). In roam-1MY14 srx-44(lf) animals, icas#9 reduced daf-28::GFP expression in ASI neurons, although it did not have this effect in roam-1 MY14 (Figure 5D). Together these experiments show that srx-43 and srx-44 influence icas#9 regulation of daf-28::GFP, and follow the pattern that the N2 roam-1 allele depends upon srx-43, and the MY14 roam-1 allele depends upon srx-44. However, they are inconsistent with simple cell-autonomous action of each receptor: Comparing Figure 5A and C shows that srx-43, which is expressed only in ASI, affects daf-28::GFP expression in both ASI and ASJ. Comparing Figure 5B to 5D shows that srx-44 affects daf-28::GFP expression in ASI, where srx-44 is not expressed. These results show that signaling between ASI and ASJ neurons affects their gene expression patterns (Figure 5H). The influence of insulins on foraging behavior was further examined in insulin signaling mutants. Animals lacking daf-2, the receptor for all insulins, explored significantly less than N2 controls at baseline, precluding an analysis of their pheromone sensitivity (Figure 5E). Mutants bearing either the dominant interfering daf-28(sa191) insulin mutation or the recessive daf-28(tm2308) null allele resembled N2 in their exploration of control plates (Figure 5E), but responded only weakly to ascarosides compared to N2 (Figure 5F). Animals null for daf-16, which encodes a transcription factor that is inhibited by the insulin pathway, also responded weakly to icas#9 compared to N2 (Figure 5F). These results suggest that unidentified insulins and the DAF-2 insulin receptor regulate basal exploration, while daf-28 acts in the ascaroside regulation of exploration. In pheromone-regulated dauer larva development, insulin signaling converges with a parallel TGF-β signaling pathway to regulate a final common target, the nuclear hormone receptor DAF-12. Previous studies demonstrated that srx-43 regulates the TGF-beta protein encoded by daf-7, and that the behavioral effects of icas#9 are partly suppressed in TGF-β signaling mutants, as they are in insulin signaling mutants (Greene et al., 2016). daf-12(lf) animals were profoundly insensitive to icas#9, with stronger defects than eliminating either insulin (daf-16) or TGF-beta (daf-3) signaling (Figure 5G). These genetic results suggest that icas#9 foraging behavior, like dauer development, may involve convergence of insulin and TGF-β signaling onto the nuclear hormone receptor DAF-12. Full-length SRX-44 translational fusions were localized to the sensory cilia of ADL (N2 and MY14 transgenes) and ASJ (MY14 transgenes) (Figure 5—figure supplement 1A), consistent with a sensory function for SRX-44. However, we could not detect acute icas#9 responses in ADL neurons using genetically-encoded calcium indicators (n = 7). This result is consistent with previous studies of the SRBC-64, SRX-43, SRG-36, and SRG-37 chemoreceptors, which detect ascaroside pheromones but do not elicit calcium transients in the ASK or ASI neurons in which they are normally expressed (Kim et al., 2009; McGrath et al., 2011). Ectopic expression of SRX-43, SRG-36, and SRG-37 in the ASH sensory neurons is sufficient to drive calcium transients in response to pheromones, but we were unable to elicit a response upon SRX-44 expression in ASH (Figure 5—figure supplement 1B), leaving its exact sensory properties unclear. Chemoreceptor genes are associated with balancing selection and positive selection in the C. elegans genome Several studies have linked C. elegans chemoreceptor genes to natural or artificial selection. srg-36 and srg-37, which encode chemoreceptors for the ascaroside ascr#5, have repeatedly mutated to inactivity in laboratory strains maintained under high density growth conditions (McGrath et al., 2011). srbc-64, which encodes a chemoreceptor for the ascaroside ascr#2, falls within a mating incompatibility locus that is under balancing selection (Kim et al., 2009; Seidel et al., 2008). These repeated associations raise the possibility that chemoreceptors are preferred substrates of C. elegans adaptation and evolution. In agreement with this hypothesis, chemoreceptor genes are enriched in regions that are hypervariable between the well-characterized N2 and CB4856 wild-type strains (Thompson et al., 2015). To ask if this observation extends across populations at a genomic level
