Selection shapes turnover and magnitude of sex-biased expression in Drosophila gonads

Background: Sex-biased gene expression is thought to drive the phenotypic differences in males and females in metazoans. Drosophila has served as a primary model for studying male-female differences in gene expression, and its effects on protein sequence divergence. However, the forces shaping evolution of sex-biased expression remain largely unresolved, including the roles of selection and pleiotropy. Research on sex organs in Drosophila, employing original approaches and multiple-species contrasts, provides a means to gain insights into factors shaping the turnover and magnitude (fold-bias) of sex-biased expression. Results: Here, using recent RNA-seq data, we studied sex-biased gonadal expression in 10,740 protein coding sequences in four species of Drosophila, D. melanogaster, D. simulans, D. yakuba and D. ananassae (5 to 44 My divergence). Using an approach wherein we identified genes with lineage-specific transitions (LSTs) in sex-biased status (amongst testis-biased, ovary-biased and unbiased; thus, six transition types) standardized to the number of genes with the ancestral state (S-LSTs), and those with clade-wide expression bias status, we reveal several key findings. First, the six categorical types of S-LSTs in sex-bias showed disparate rates of turnover, consistent with differential selection pressures. Second, the turnover in sex-biased status was largely unrelated to cross-tissue expression breadth, suggesting pleiotropy does not restrict evolution of sex-biased expression. Third, the fold-sex-biased expression, for both testis-biased and ovary-biased genes, evolved directionally over time toward higher values, a crucial finding that could be interpreted as a selective advantage of greater sex-bias, and sexual antagonism. Fourth, in terms of protein divergence, genes with LSTs to testis-biased expression exhibited weak signals of elevated rates of evolution (than ovary-biased) in as little as 5 My, which strengthened over time. Moreover, genes with clade-wide testis-specific expression (44 My), a status not observed for any ovary-biased genes, exhibited striking acceleration of protein divergence, which was linked to low pleiotropy. Conclusions: By studying LSTs and clade-wide sex-biased gonadal expression in a multi-species clade of Drosophila, we describe evidence that interspecies turnover and magnitude of sex-biased expression have been influenced by selection. Further, whilst pleiotropy was not connected to turnover in sex-biased gonadal expression, it likely explains protein sequence divergence.