Balancing selection maintains hyper-divergent haplotypes in Caenorhabditis elegans

Abstract

Across diverse taxa, selfing species have evolved independently from outcrossing species thousands of times. The transition from outcrossing to selfing significantly decreases the effective population size, effective recombination rate, and heterozygosity within a species. These changes lead to a reduction in genetic diversity, and therefore adaptive potential, by intensifying the effects of random genetic drift and linked selection. Within the nematode genus Caenorhabditis , selfing has evolved at least three times and all three species, including in the model organism Caenorhabditis elegans , show substantially reduced genetic diversity relative to outcrossing species. Selfing and outcrossing Caenorhabditis species are often found in the same niches, but we still do not know how selfing species with limited genetic diversity can adapt to these environments. Here, we examine the whole-genome sequences from 609 wild C. elegans strains isolated worldwide and show that genetic variation is concentrated in punctuated hyper-divergent regions that cover 20% of the C. elegans reference genome. These regions are enriched in environmental response genes that mediate sensory perception, pathogen response, and xenobiotic stress response. Population genomic evidence suggests that genetic diversity in these regions has been maintained by long-term balancing selection. Using long- read genome assemblies for 15 wild strains, we show that hyper-divergent haplotypes contain unique sets of genes and show levels of divergence comparable to levels found between Caenorhabditis species that diverged millions of years ago. These results provide an example for how species can avoid the evolutionary “dead end” associated with selfing.

Diversification of small RNA pathways underlies germline RNAi incompetence in wild C. elegans strains

“…Estimates of nucleotide diversity (p) per gene were generally high (Table 1) relative to genome-553 wide averages (Lee et al, 2021), with p>0.001 for 14/62 genes (13/55 genes not characterized as 554 pseudogenes). Elevated variation was often, but not always, associated with high-impact 555 mutations in one or more strains (Figure S5).…”

Nematode nuclear receptors as integrators of sensory information

“…In addition, the correlated expression of nhrs and gpcrs is not restricted only to genes in the hyperdivergent regions of the C. elegans genome (Figures S3A-S3G). 19 However, we find that the sensory neuron-enriched expression of gpcr genes more strongly resembles the expression pattern of C. elegans-expanded nhrs compared to that of evolutionarily conserved nhrs (Figures S3H-S3M). 3 Though correlated expression of genes does not necessarily imply an underlying functional relationship, it is consistent with our hypothesis that individual nuclear receptors potentially modulate the expression of GPCR-encoding genes to dictate an animal's sensory experience in a particular environment.…”

“…3 Identifying NHR and GPCR genes in hyperdivergent genomic regions The genomic coordinates for all members of the NHR and GPCR gene families were compared to that of the previously identified hyperdivergent regions of the C. elegans genome. 19 A gene was identified to be in a hyperdivergent region if at least 10% of the gene's length resides within hyperdivergent regions of the C. elegans genome, which cover $20% of the C. elegans reference genome. 19 Unlike for NHR (55%) and GPCR (52%) genes, only 17.5% of all insulin genes reside in hyperdivergent regions and hence they were excluded from this analysis.…”

“…19 A gene was identified to be in a hyperdivergent region if at least 10% of the gene's length resides within hyperdivergent regions of the C. elegans genome, which cover $20% of the C. elegans reference genome. 19 Unlike for NHR (55%) and GPCR (52%) genes, only 17.5% of all insulin genes reside in hyperdivergent regions and hence they were excluded from this analysis. Expression analyses for NHR and GPCR genes were performed using data from the CeNGEN scRNA atlas, as described above.…”

Natural variation in theirldgene family affects starvation resistance inC. elegans

“…irld-52 also contains a high-impact variant associated with starvation resistance (Fig 3F). irld-11 and irld-57 overlap with 'hyper-divergent' regions containing numerous variants in some starvation-resistant strains (LEE et al 2021), and hyper-divergence at these loci is associated with starvation resistance . Because irld-39 and irld-52 contain singular, putative high-impact variants associated with starvation resistance, we used genome editing to generate these specific variants in the starvation-sensitive MY2147 and the laboratoryadapted N2 backgrounds (Fig 3I).…”