Evidence for Male-Driven Evolution in Drosophila

Abstract

Haldane (1935) proposed that in humans the mutation rate in males would be higher than in females because the male germ line goes through many more cell divisions per generation than the female. Thus, if mutations arise by replication error, most mutations would occur in males and DNA sequence evolution would be male driven. The male-to-female mutation rate (α) can be estimated by contrasting DNA divergence data at X-linked, Y-linked, and autosomal loci as these chromosomes reside differently in males and females. In particular, the Y chromosome is restricted to males, autosomes spend half their times in males and females, whereas X chromosomes are transmitted only 1/3 of the time through males. Thus, under male-driven evolution, Y chromosomes should have the highest mutation rate and X chromosomes the lowest. Assume the mutation rates for a X-linked, Y-linked, and autosomal sequence are X, Y, and A, respectively. Miyata et al. (1987) showed that Y/X = 3α/(2 + α), Y/A = 2α/(1 + α), and X/A = (2/3)(2 + α)/(1 + α). From these formulas, one can estimate α if the Y/X, Y/A, or X/A ratio is known. To estimate these ratios, divergence at a pair of homologous nonfunctional sequences on the different chromosomes can be used because in a nonfunctional sequence the rate of nucleotide substitution is equal to the rate of mutation (Kimura 1983).

In several vertebrate taxa, including fish, birds, and mammals, researchers found evidence for male-driven evolution based on comparisons of homologous X and Y or autosomal sequences (Ellegren and Fridolfsson 1997, 2003; Makova and Li 2002; Sandstedt and Tucker 2005). In different organisms, estimates of α were found to vary roughly between 2 and 10 (see Li et al. [2002] for a recent review). To date, male-driven evolution has not been found outside vertebrates. Previous studies in Drosophila have contrasted X and autosomal divergence and found no evidence for male-driven evolution (Bauer and Aquadro 1997; Betancourt et al. 2002). However, there are potential problems when comparing levels of divergence at nonhomologous loci. In particular, neutral mutation rates may differ between loci due to differences in base composition or chromosomal effects. Also, the number of loci compared in previous studies was relatively small, and the power to detect male-driven evolution is maximized in contrasts involving X- versus Y-linked data (Miyata et al. 1987). However, the lack of homologous sequences between the X and Y chromosome has prevented such comparisons in the Drosophila melanogaster species group.

Here, I test for male-driven evolution in Drosophila taking advantage of the recently formed neo-sex chromosomes of Drosophila miranda. In this species, a pair of autosomes became sex linked only about 1 MYA, creating so-called neo-sex chromosomes (Bachtrog and Charlesworth 2002). The neo-sex chromosomes of D. miranda show high levels of sequence similarity over wide stretches (Bachtrog 2005), which allow me to compare rates of molecular evolution at homologous neo-X- and neo-Y-linked loci. To test for male-driven evolution, a total of 205 homologous gene pairs located on the neo-sex chromosomes were analyzed (Bachtrog et al. 2008). Changes along the neo-X and neo-Y chromosome of D. miranda are polarized using Drosophila pseudoobscura as an outgroup species, assuming parsimony. Substitution rates along the neo-X and neo-Y chromosome were calculated using Jukes–Cantor. Given the low amount of sequence divergence between the neo-sex chromosomes (∼3%, see table 1), parsimony reconstruction to infer the ancestral state is accurate (Kimura 1983).

To estimate neutral mutation rates of the neo-sex chromosomes, I consider 2 potentially unconstrained classes of neutral DNA: synonymous sites and short introns. In Drosophila, synonymous codons are not used randomly but instead some codons are used preferentially over a different codon encoding for the same amino acid; for example, preferred (P) and unpreferred (U) codons (Bulmer 1991; Akashi and Schaeffer 1997). Thus, synonymous sites are not necessarily free of selective constraints. Nevertheless, selection for codon bias is strongly reduced in D. miranda, and synonymous sites are evolving probably close to neutral in this species (Bartolomé and Charlesworth 2006; Bachtrog 2007). Synonymous changes accumulate on the neo-Y chromosome at a higher rate than on the neo-X (d_S neo-Y = 1.88% vs. d_S neo-X = 1.33%, table 1), corresponding to a ratio of Y/X divergence of roughly 1.4. A similar excess of neo-Y-linked mutations is estimated using the absolute number of mutations accumulated on the neo-X or neo-Y branch (table 2).

The efficacy of natural selection is reduced on the neo-Y chromosome, due to its smaller effective population size and interference selection, which could result in a faster accumulation of unpreferred substitutions on the neo-Y (Kliman and Hey 1993; Bachtrog and Charlesworth 2002). To distinguish whether synonymous substitutions accumulate on the neo-Y either due to a higher mutation rate (i.e., male-driven evolution) or due to reduced selective constraints against unpreferred synonymous codons, I classified synonymous changes into deleterious (P → U), beneficial (U → P), and presumably neutral changes (i.e., U → U and P → P), using preferred codons identified in D. pseudoobscura (Bachtrog 2007). If increased mutation rates on the neo-Y account for the elevated rate in synonymous substitutions, both the neutral and the deleterious class should show a faster rate of accumulation on the neo-Y. If elevated divergence on the neo-Y solely results from a reduced efficacy of natural selection against unpreferred synonymous changes, only the deleterious class should show an increased number of changes on the neo-Y. I find that both deleterious and neutral changes accumulate similarly and significantly faster on the neo-Y (table 2). This suggests an elevated mutation rate on the neo-Y chromosome as the cause for the increased rate of fixations of synonymous substitutions. Only preferred changes show no significant acceleration in their rate of accumulation on the neo-Y (table 2). This is consistent with a larger mutation rate of the neo-Y and weak positive selection for preferred codons operating on the neo-X (Bartolomé and Charlesworth 2006; Bachtrog 2007).

Another class of putatively unconstrained sequence in Drosophila is short introns. In comparisons between D. melanogaster and Drosophila simulans or between D. miranda and D. pseudoobscura, short introns are found to evolve significantly faster than long introns at rates comparable with synonymous sites (Haddrill et al. 2005; Bachtrog and Andolfatto 2006; Halligan and Keightley 2006). Thus, I also classified mutations at short introns (<100 bp) to accumulate either on the neo-X or on the neo-Y branch. Like synonymous sites, mutations at short introns accumulate faster on the neo-Y compared with the neo-X; the ratio of Y/X divergence is roughly 1.2–1.4 (tables 1 and 2), similar to the ratio inferred using synonymous site. Thus, both synonymous sites and short introns suggest a higher mutation rate for the neo-Y chromosome compared with the neo-X.

Elevated mutation rates on the neo-Y chromosome indicate a higher mutation rate in males relative to females. One can use the amount of nucleotide divergence on the neo-X and neo-Y chromosome to calculate α, the male-to-female ratio of mutation rate. Confidence intervals (CIs) were estimated following Makova and Li (2002). I infer that α = 1.8 (95% CI 1.5–2.1) using synonymous changes and α = 1.7 (95% CI 1.2–2.3) using short introns (see tables 1 and 2), that is, the mutation rate in male Drosophila might almost be twice as high as the female mutation rate. Previous studies in Drosophila, which compared sequence divergence at X-linked and autosomal loci found no evidence for male-driven evolution (Bauer and Aquadro 1997; Betancourt et al. 2002). This probably is due to a combination of the relatively small number of loci considered, lower power in X–autosome than X–Y comparisons (Miyata et al. 1987), and the use of nonhomologous loci. A recent study comparing X and autosomal divergence of the genome sequences of D. melanogaster, D. simulans, and Drosophila yakuba found that X-linked sequences—including intron, intergenic, and synonymous sites—generally evolve faster than autosomal sequences (Begun et al. 2007). However, no distinction was made between potentially functional and nonfunctional sites (i.e., neutral synonymous changes or short introns), preventing inferences about male-driven evolution in this species group. It is also possible that either the number of germ line cell divisions in males and females differ between the D. melanogaster and D. pseudoobscura species group or the age of reproduction, such that only the latter shows male-biased evolution (Bauer and Aquadro 1997). In particular, in D. melanogaster the ratio of germ line cell divisions appears to change over time from indicating a weak female bias to a male bias as the age of reproduction increases (Bauer and Aquadro 1997). Thus, differences in the average age of reproduction among Drosophila species could account for varying levels of male-driven evolution in this genus, but little is known about differences in life history strategies among Drosophila species in nature.

It should be noted that α = 1.8 is likely an underestimate of the extent of the male mutation bias, due to different amounts of segregating polymorphism on the neo-sex chromosomes (Bartolome et al. 2005). Sampling of a single allele from the neo-X and neo-Y chromosome, as done in this study, does not allow fixed and segregating mutations to be distinguished. That is, a small fraction of the mutations assigned as a fixed difference occurring on the neo-X chromosome might in fact be segregating in the population. The severe reduction in levels of variability on the neo-Y chromosome implies that nearly all mutations on the neo-Y are fixed within the population (Bachtrog 2004; Bartolomé and Charlesworth 2006). However, a few of the variants that are assigned as a fixed difference occurring on the neo-X might be polymorphic in the population, causing us to underestimate the amount of male-driven evolution. Also note that the neo-Y chromosome of D. miranda lacks recombination, whereas the neo-X recombines in females. Thus, if recombination were mutagenic this would also result in underestimating α. There is, however, no evidence for mutagenic effects of recombination in Drosophila (Begun and Aquadro 1992).

Mutation rates have been found to be higher in males in several vertebrate taxa studied to date. My study is the first to report male-biased mutation rates outside vertebrates. Taking advantage of the recently formed neo-sex chromosomes of D. miranda, I estimate that there are approximately twice as many mutations in male Drosophila compared with females. This provides evidence that DNA sequence evolution is male driven in a wide variety of taxa, even outside vertebrates.

This work was funded by a National Institute of Health Grant (GM076007) to D.B.

References

Author notes