Balancing selection shapes density-dependent foraging behaviour

Abstract

The optimal foraging strategy in a given environment depends on the number of competing individuals and their behavioural strategies. Little is known about the genes and neural circuits that integrate social information into foraging decisions. Here we show that ascaroside pheromones, small glycolipids that signal population density, suppress exploratory foraging in Caenorhabditis elegans, and that heritable variation in this behaviour generates alternative foraging strategies. We find that natural C. elegans isolates differ in their sensitivity to the potent ascaroside icas#9 (IC-asc-C5). A quantitative trait locus (QTL) regulating icas#9 sensitivity includes srx-43, a G-protein-coupled icas#9 receptor that acts in the ASI class of sensory neurons to suppress exploration. Two ancient haplotypes associated with this QTL confer competitive growth advantages that depend on ascaroside secretion, its detection by srx-43 and the distribution of food. These results suggest that balancing selection at the srx-43 locus generates alternative density-dependent behaviours, fulfilling a prediction of foraging game theory.

Extended Data Figure 1. Roaming and dwelling states in the presence of ascarosides

a) Roaming and dwelling behaviors scored from video analysis. n=102-214 tracks per data point. b and c) Cumulative distribution of roaming (b) and dwelling (c) state durations for animals in (a). ***P<0.001 by log rank test; ns, not significant. d and e) Scatter plot of average speed and angular speed (a measure of turning rate) in 10 s intervals taken from 1.5 hour long video recordings of wild-type animals in control (d) and icas#9 (e) conditions. Roaming animals move quickly and turn infrequently compared to dwelling animals. Note bimodal distribution defining distinct behavioral states. Control = 161 tracks, icas#9 = 102 tracks. f and g) Speed following a reversal (f) and reversal rate (g) for roaming or dwelling animals. Roaming speed is slightly slower in ascarosides (e,f). Data presented as mean ± SEM. ***P<0.001, *P<0.05 by ANOVA with Dunnett correction; ns, not significant.

Extended Data Figure 2. Roaming and dwelling behavior of MY14

a) Fraction of time MY14 animals spend roaming or dwelling in control, ascr#8, and icas#9 conditions. n=66-109 tracks per data point. Assays were conducted in 8% O₂. Compare Extended Data Figure 1. b and c) Cumulative distribution of roaming (b) and dwelling (c) state durations for MY14 animals scored in (a). Roaming states are significantly shorter in the presence of ascr#8 (t½ = ~150 s, vs ~220 s in controls), but they are not significantly affected by icas#9 (t ½ = ~190 s). Roaming states may also be longer at baseline in MY14 than in N2 (see Extended Data Figure 1). ***P<0.001 by log rank test; ns, not significant.

Extended Data Figure 3. Ascarosides produced by wild-type strains

LC-MS/MS analysis of ascarosides secreted by N2, CX12311, and MY14 strains grown on (a) OP50 or (b) HB101 bacteria. icas#9 is produced at similar levels by icas#9-sensitive and icas#9-resistant strains. n=2 (a) or 3 (b) culture extracts per genotype.

Extended Data Figure 4. Covariate analysis of 94 RILs

Covariate analysis controlling for roam-1 genotype, testing for additive (a) or interactive (b) QTL at other loci. The horizontal line denotes the P<0.05 genome-wide significance threshold. LOD, log likelihood ratio.

Extended Data Figure 5. Signal transduction by SRX-43

a) Expression of Psrx-43::srx-43::SL2::GFP bicistronic reporter transgenes bearing N2 (top) or MY14 (botton) srx-43 sequences. Arrows indicate cell bodies of ASI sensory neurons. Scale bar, 50 μm. b) ASH sensory neurons are insensitive to multiple ascarosides. ASH calcium imaging with GCaMP3 in control animals that do not express srx-43 transgene, isolated as non-transgenic siblings of transgenic animals tested in Figure 3f (n=19). Ascarosides tested at 10 nM. c) SRX-43 from MY14 confers icas#9 sensitivity on ASH neurons. Compare SRX-43 from N2 in Figure 3f. d) icas#9 decreases daf-7::GFP expression in ASI neurons of N2 but not roam-1_MY14 adults. Bars indicate mean fluorescence intensity ± SEM. *P<0.05 by ANOVA with Tukey's multiple comparisons test. n = # of animals, e) icas#9 responses of daf-7(lf) mutants are attenuated in N2 but not in roam-1_MY14 genetic backgrounds. Modified exploration assays were conducted on strains including daf-3(lf) alleles (see methods). *P<0._{05, ns}, not significant by t test. Data represented as mean ± SEM. f) Time course for icas#9 response in exploration assay. Pheromone response expressed as mean ± SEM for 2, 4, 6, 10, and 14 hours following initiation of exploration assay. ***P<0.001, ns, not significant by t test with Bonferonni correction comparing squares entered in control versus 10 nm icas#9 plates. n = 12 for all time points.

Extended Data Figure 6. Alternative roam-1 allele have high sequence variability

a) The roam-1 QTL region (top). roam-1 SNPs = SNPs with respect to the N2 reference genome that are shared by JU360, MY2, MY14, ED3021, JU1171, MY16, and MY6 and not by any other strains, according to the Million Mutation Project, defining the roam-1_MY14 haplotype. Other SNPs = all other SNPs with respect to the N2 reference genome found in any of the 40 wild isolates in the Million Mutation Project. b) Polymorphisms revealed by Sanger sequencing of srx-43 promoter and coding region. Despite the high rate of polymorphism, there are only 4 nonsynonymous mutations in the MY14 coding sequence. We confirmed that the MY14 and N2 sequences are alleles of the same gene by examining sequence reads of the MY14-like strain MY23 in the CeNDR dataset (www.elegansvariation.org) and aligning each read to N2 and MY14 sequence for the srx-43 region as determined by Sanger sequencing. 7272 of the MY23(MY14) reads better matched the MY14 Sanger sequence and 4 of the reads better matched the N2 reference sequence, as expected if MY14 and N2 each bear one alternative allele of the gene. c) Phylogeny constructed for srx-43 and related genes in C. elegans, C. briggsae, and C. remanei demonstrates that the srx-43 alleles in N2 and MY14 are closely related 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.

Extended Data Figure 7. Significant recombination between roam-1 and surrounding regions

Top, phylogenies constructed with 152 diverse wild isolates revealing differences for the region surrounding srx-43 and the regions immediately to the left and right of the 30 kb haplotype. Bottom, graph showing the number of variants and Tajima's D score calculated for 5 kb bins across a 250 kb region. The bin containing srx-43 has 250 polymorphisms and a Tajima's D of 1.01, which is high both at the genomic level (<3.4% of bins had a higher value) and for the chromosomal location of srx-43 (<3.6% of bins had a higher value).

Extended Data Figure 8. Recombination between srx-43 and glc-1 in natural isolates

a) The glc-1 gene has previously been shown to be under balancing selection, and is near srx-43. The blue line shows SNPs/kb for N2 and MY14 averaged over 5 kb intervals for the region spanning srx-43 and glc-1. The large region of low heterozygosity between srx-43 and glc-1 indicates that balancing selection on glc-1 is unlikely to account for the high heterozygosity near srx-43. b) Dendrogram for the glc-1 region for strains shown in Figure 4a. roam-1_MY14 clades and glc-1 clades are not identical.

Extended Data Figure 9. Standard curve for dPCR experiments

Best fit line of dPCR results for known ratios of N2 to roam-1_MY14 DNA created by mixing different ratios of genomic DNA extracted from independent N2 or roam-1_MY14 populations.

Figure 1. Ascaroside pheromones suppress exploratory foraging behavior

A) Exploration assay. B) Wild-type N2 response to crude pheromone extract, showing exploration scores and a pheromone response index, presented as mean ± SEM. C) Structures and names of selected ascarosides. D) N2 response to individual ascarosides, presented as mean ± SEM. ***P<0.001, **P<0.01 by ANOVA with Dunnett correction; ns, not significant.

Figure 2. Natural genetic variation in pheromone sensitivity

A) Response of wild-type C. elegans strains to synthetic ascarosides. ***P<0.001 by ANOVA with Dunnett correction. B) icas#9 response of 94 CX12311-MY14 recombinant inbred lines (RILs) and parental strains. C) QTL analysis of RILs shown in B. Horizontal line denotes P<0.05 genome-wide significance threshold. LOD, log likelihood ratio. D) Mean icas#9 response of near-isogenic lines (NILs) used to map the roam-1 locus, in MY14 and CX12311 (i) or N2 (ii,iii) backgrounds. ***P<0.001, **P<0.01 by ANOVA with Dunnett correction; ns, not significant. All data are presented as mean ± SEM

Figure 3. The roam-1 QTL includes the icas#9 receptor SRX-43

A) The roam-1 locus. B-D) icas#9 sensitivity as affected by (B) high-copy srx-43 transgenes (C) srx-43(lf) mutations (D) single-copy integrated srx-43 transgenes. E) srx-43::GFP translational reporter (top) and Nomarski image (bottom). Arrowhead, ASI sensory cilium. Scale bar, 10 μm. F) ASH calcium responses in ASH::srx-43 transgenic strain. G) icas#9 sensitivity after exchanging promoters in single-copy srx-43 transgenes. H) Endogenous srx-43 mRNA levels. All data are presented as mean ± SEM. ***P<0.001, **P<0.01, *P<0.05 by ANOVA with Dunnett correction or t-test; ns, not significant. Color indicates roam-1 genotype.

Figure 4. Population genetics of the roam-1 locus and icas#9 sensitivity

A) A dendrogram across 41 natural isolates representing a 20 kb region surrounding srx-43. The two major roam-1 haplotypes from N2 and MY14 strains are indicated in red and blue, respectively. B) Whole-genome dendrogram for the same strains as A, showing relationships among strains. Red and blue colors follow roam-1 haplotypes in A. C) icas#9 responses of natural isolates with the MY14 haplotype (blue) are consistently lower than those with the N2 haplotype (red). Pheromone response shown as mean ± SEM.

Figure 5. Bidirectional competitive selection at the roam-1 locus

A) Diagram of “boom-bust” competition experiments, with food depletion followed by 48 hours of starvation before transfer. B-D) Competition on “simple lawn” showing allele ratio of DNA harvested at transfers 1 and 3. (B) N2 versus roam-1_MY14 NIL (C) N2 srx-43(lf) versus roam-1_MY14 srx-43(lf) (D) N2 daf-22(lf) versus roam-1_MY14 daf-22(lf), without or with exogenous icas#9 (10 nM). E) “Patchy lawn” competition between N2 and roam-1_MY14 NIL. Grey points = individual competition experiments, red line = mean, ***P<0.001, **P<0.01, *P<0.05 compared to expected value of 0.5 by t-test with Bonferroni correction; ns, not significant.