Recurrent events of positive selection in independent Drosophila lineages at the spermatogenesis gene roughex

Abstract

Our understanding of the role of positive selection in the evolution of genes with male-biased expression can be hindered by two observations. First, male-biased genes tend to be overrepresented among lineage-specific genes. Second, novel genes are prone to experience bursts of adaptive evolution shortly after their formation. A thorough study of the forces acting on male-biased genes therefore would benefit from phylogenywide analyses that could distinguish evolutionary trends associated with gene formation and later events, while at the same time tackling the interesting question of whether adaptive evolution is indeed idiosyncratic. Here we investigate the roughex (rux) gene, a dose-dependent regulator of Drosophila spermatogenesis with a C-terminal domain responsible for nuclear localization that shows a distinct amino acid sequence in the melanogaster subgroup. We collected polymorphism and divergence data in eight populations of six Drosophila species, for a total of 99 rux sequences, to study rates and patterns of evolution at this male-biased gene. Our results from two phylogeny-based methods (PAML and HyPhy) as well as from population genetics analyses (McDonald-Kreitman-based tests) indicate that amino acid replacements have contributed disproportionately to divergence, consistent with adaptive evolution at the Rux protein. Analyses based on extant variation show also the signature of recent selective sweeps in several of the populations surveyed. Most important, we detect the significant and consistent signature of positive selection in several independent Drosophila lineages, which evidences recurrent and concurrent events of adaptive evolution after rux formation.

Figure 1.—

Phylogenetic relationship among the nine species of the D. melanogaster subgroup inferred from rux gene genealogy. The proposed phylogeny is in complete agreement with that obtained from the 12 Drosophila Genomes Project (Clark et al. 2007).

Figure 2.—

Distribution of positively selected sites along exon 2 of the rux gene identified by PAML (BEB) (diamonds, posterior probabilities >0.50), FEL (circles, posterior probabilities >0.50), and REL (triangles, P < 0.20) in species of (A) the melanogaster subgroup, (B) the D. yakuba–D. erecta clade, and (C) the D. simulans–D. melanogaster clade.

Figure 3.—

Results of lineage-specific MK tests in the rux locus [goodness-of-fit (G) values and probabilities (P)]. Branch lengths do not represent real genetic distances.