Mutation Load in Sunflower Inversions Is Negatively Correlated with Inversion Heterozygosity

Abstract

Recombination is critical both for accelerating adaptation and purging deleterious mutations. Chromosomal inversions can act as recombination modifiers that suppress local recombination in heterozygotes and thus, under some conditions, are predicted to accumulate such mutations. In this study, we investigated patterns of recombination, transposable element abundance and coding sequence evolution across the genomes of 1,445 individuals from three sunflower species, as well as within nine inversions segregating within species. We also analyzed the effects of inversion genotypes on 87 phenotypic traits to test for overdominance. We found significant negative correlations of long terminal repeat retrotransposon abundance and deleterious mutations with recombination rates across the genome in all three species. However, we failed to detect an increase in these features in the inversions, except for a modest increase in the proportion of stop codon mutations in several very large or rare inversions. Consistent with this finding, there was little evidence of overdominance of inversions in phenotypes that may relate to fitness. On the other hand, significantly greater load was observed for inversions in populations polymorphic for a given inversion compared to populations monomorphic for one of the arrangements, suggesting that the local state of inversion polymorphism affects deleterious load. These seemingly contradictory results can be explained by the low frequency of inversion heterozygotes in wild sunflower populations, apparently due to divergent selection and associated geographic structure. Inversions contributing to local adaptation represent ideal recombination modifiers, acting to facilitate adaptive divergence with gene flow, while largely escaping the accumulation of deleterious mutations.

Chromosomal inversion polymorphisms shape the genomic landscape of deer mice

“…With reduced efficacy of purifying selection in the absence of recombination, the expectation is that inversions will accumulate a deleterious mutational load (when inversion heterozygotes are common) 32 , which will limit their spread 46 . In deer mouse inversions, however, we did not find evidence for the accumulation of mutational load based on nonsynonymous mutations (although these inversions may harbour an excess of other types of deleterious variants such as transposable elements, which future work will further resolve), consistent with a recent study in sunflowers 47 . In both deer mice and sunflowers, inversion homozygotes are common 47 ; as recombination proceeds uninterrupted in inversion homozygotes, deleterious mutations can efficiently be removed once an inversion reaches substantial allele frequency 32 , especially if effective population sizes (N e ) are high, as in many populations of deer mice (for example, N e ≈ 4 × 10 6 in a single population 20 ).…”

“…In both deer mice and sunflowers, inversion homozygotes are common 47 ; as recombination proceeds uninterrupted in inversion homozygotes, deleterious mutations can efficiently be removed once an inversion reaches substantial allele frequency 32 , especially if effective population sizes (N e ) are high, as in many populations of deer mice (for example, N e ≈ 4 × 10 6 in a single population 20 ). As in sunflowers 47 , we hypothesize that these inversions, which act as large-scale modifiers of recombination when heterozygous, largely evaded deleterious costs associated with suppressed recombination by quickly spreading to high frequencies in deer mice, whose large population sizes could facilitate effective purifying selection in inversion homozygotes 32 (noting that gene conversion between inversion and standard haplotypes may also have a role in reducing deleterious mutational load 32 ).…”

The genomics of linkage drag in sunflower

“…The statistic was negatively correlated with recombination rate (SI Appendix, Fig. S19), in accord with previous understanding of the role of recombination in eliminating deleterious mutations (Huang et al 2022). P nonsense / P nonsyn of polymorphic primary introgressions was lower in null recombination rate regions than that of non-introgressed regions and comparable to non-introgressed regions in regions of reduced and high recombination rate (SI Appendix, Fig.…”

“…Plant breeders typically monitor the size and location of introgressions with molecular markers and/or restrict pre-breeding efforts to fully compatible wild relatives (i.e., members of the primary gene pool; Harlan and de Wet 1971) to reduce the impact of the linkage drag (Young and Tanksley 1989; Tanksley and McCouch 1997; Frary et al 2004). However, in large plant genomes, regions of low recombination are widespread, making it difficult to reduce the sizes of some introgressions (Rodgers-Melnick et al 2015; Brazier and Glémin 2022; Huang et al 2022). Also, key traits may be found outside of the primary gene pool, making it necessary to tap less compatible wild relatives (e.g., Duriez et al 2019).…”

The rate of inversion fixation in plant genomes is highly variable

“…Although the genomes we included achieved chromosome-scale scaffolding, centromeric regions are especially dynamic and repetitive which means they are highly challenging to assemble correctly. Inversions can overlap with centromeres, and if breakpoints fell within centromeric regions we may not accurately locate them (Kirubakaran et al 2020; Harringmeyer and Hoekstra 2022; Huang et al 2022).…”