Effective purifying selection in ancient asexual oribatid mites

Abstract

Sex is beneficial in the long term because it can prevent mutational meltdown through increased effectiveness of selection. This idea is supported by empirical evidence of deleterious mutation accumulation in species with a recent transition to asexuality. Here, we study the effectiveness of purifying selection in oribatid mites which have lost sex millions of years ago and diversified into different families and species while reproducing asexually. We compare the accumulation of deleterious nonsynonymous and synonymous mutations between three asexual and three sexual lineages using transcriptome data. Contrasting studies of young asexual lineages, we find evidence for strong purifying selection that is more effective in asexual as compared to sexual oribatid mite lineages. Our results suggest that large populations likely sustain effective purifying selection and facilitate the escape of mutational meltdown in the absence of sex. Thus, sex per se is not a prerequisite for the long-term persistence of animal lineages.Asexual reproduction is thought to be an evolutionary dead end in eukaryotes because deleterious mutations will not be purged effectively. Here, Brandt and colleagues show that anciently asexual oribatid mites in fact have reduced accumulation of deleterious mutations compared to their sexual relatives.

Fig. 1

Phylogenetic tree of 30 oribatid mite species analysed in this study. The tree is based on partial sequences of 18S rDNA, ef1α and hsp82 (see ‘Methods’). The taxon sampling covers asexual (blue) and sexual (red) species from four out of six major groups of oribatid mites (brackets). Species used for transcriptome-wide analyses are indicated with full name, remaining species names are listed from top to bottom in Supplementary Table 1

Fig. 2

Per-gene differences in dN/dS ratios between reproductive modes. The histogram shows the distribution of per-gene differences in dN/dS ratios between asexual and sexual terminal branches in a tree comprising three sexual and three asexual oribatid mite species for 3539 orthologous loci under purifying selection (dN/dS < 1). To improve data presentation, the histogram range is limited to −0.1 to 0.1, representing 3539 out of 3545 orthologs. The distribution mean (−0.008; dotted line) is shifted to the left indicating overall lower dN/dS ratios at asexual as compared to sexual branches (Wilcoxon signed-rank test V = 4.84*10⁶, P < 0.001). Frequencies of genes with significantly lower dN/dS ratios at asexual as compared to sexual branches or sexual as compared to asexual branches are coloured blue and red, respectively

Fig. 3

Hydrophobicity scores at asexual and sexual terminal branches. The boxplots show hydrophobicity scores (HS) at three asexual (blue) and three sexual (red) terminal branches for (a) 3545 nuclear orthologous genes (involving 92,351 and 133,606 HS at asexual and sexual terminal branches, respectively) and (b) ten mitochondrial orthologous genes (involving 1494 and 2468 HS at asexual and sexual terminal branches, respectively). HS measures the strength in changes of hydrophobicity from ancestral to replacement amino acids and indicates the ‘deleteriousness’ of a nonsynonymous mutation. The lower the HS the stronger is the change in hydrophobicity and ‘deleteriousness’ of a nonsynonymous mutation. For nuclear orthologs, amino acids shifted to more dissimilar hydrophobicity at sexual branches (generalised linear mixed model z = 2.4, P = 0.017), indicating stronger ‘deleteriousness’ of nonsynonymous changes in sexual as compared to asexual oribatid mites. For mitochondrial orthologs, there was no difference (generalised linear mixed model z = 0.81, P = 0.421). Significant differences in HS are marked with an asterisk (*P < 0.05). Whiskers correspond to 1.5 times the interquartile range

Fig. 4

Comparison of CDC values between asexual and sexual oribatid mite species. Boxplots show CDC values of three asexual (blue) and three sexual (red) oribatid mite species for (a) 3545 nuclear orthologous genes (involving 10,635 CDC values for both asexual and sexual species) and (b) ten mitochondrial orthologous genes (involving 30 CDC values for both asexual and sexual species). CDC measures the deviation of observed from expected codon usage and allows for inference of the effectiveness of purifying selection acting on synonymous sites. A lower CDC value corresponds to more ‘relaxed’ selection on codon usage bias (see ‘Methods’). For nuclear orthologs, per-gene CDC was slightly but significantly higher in asexual species (gene effect P < 0.001, reproductive mode effect P = 0.008, interaction P = 0.616; permutation ANOVA). For mitochondrial orthologs, per-gene CDC differed more strongly between reproductive modes (gene effect P = 0.002, reproductive mode effect P < 0.001, interaction P = 0.960; permutation ANOVA). Significant differences in CDC are marked with asterisks (***P < 0.001 and **P < 0.01). Whiskers correspond to 1.5 times the interquartile range