Sex-biased transcriptome divergence along a latitudinal gradient

Although not all sex-dependent gene expression is adaptive, it is likely an important genomic mechanism that allows each sex to independently adapt to environmental changes. Among Drosophila species, sex-biased genes display remarkably consistent evolutionary patterns; male-biased genes evolve faster than unbiased genes in both coding sequence and expression level, suggesting sex-differences in selection through time. However, comparatively little is known of the evolutionary process shaping sex-biased expression within species. Latitudinal clines offer an opportunity to examine how changes in key ecological parameters also influence sex-specific selection and the evolution of sex-biased gene expression. We assayed male and female gene expression in Drosophila serrata along a latitudinal gradient in eastern Australia spanning most of its endemic distribution. Analysis of 11,631 genes across eight populations revealed strong sex differences in the frequency, mode, and strength of divergence. Divergence was far stronger in males than females and while latitudinal clines were evident in both sexes, male divergence was often population-specific, suggesting responses to localized selection pressures that do not covary predictably with latitude. While divergence was enriched for male-biased genes, there was no overrepresentation of X-linked genes in males. By contrast, X-linked divergence was elevated in females, especially for female biased genes. Many genes that diverged in D. serrata have homologs also showing latitudinal divergence in D. simulans and D. melanogaster on other continents, likely indicating parallel adaptation in these distantly related species. Our results suggest that sex differences in selection play an important role in shaping the evolution of gene expression over macro- and micro-ecological spatial scales. The goal of our study was to estimate “common garden” mean expression level for genes in each sex, rather than to estimate within population genetic variation. Flies were sampled from eight populations along the east-coast of Australia, covering a straight-line distance of approximately 2,300 km, which spans much of the species natural range (Figure 1). To preserve the natural genetic differences among populations and minimize adaptation to the lab, flies for each population were maintained as isofemale lines (David et al. 2005) (n=12 for all populations with the exceptions of Airlie Beach, n=10 and Cooktown, n=6) until the gene expression assay. At this point, to ensure gene expression was measured on outbred flies, we crossed the isofemale lines within populations following a double round robin mating design that included reciprocal crosses by sex (Stich 2009; Verhoeven et al. 2006). For example, isofemale line 1 × isofemale line 2, isofemale line 1 × isofemale line 3, isofemale line 2 × isofemale line 3, isofemale line 2 × isofemale line 4, and so on. Owing to a smaller number of available lines, a triple round-robin mating design was used for Airlie and all possible pair-wise crosses were performed for Cooktown. A total of 18 F1 crosses were randomly selected for RNA processing from each population with 6 crosses assigned to each of three biological replicates. Five flies were randomly selected from each cross in order to produce pools of 30 adult virgin flies (3 days old) per biological replicate. The samples were snap frozen in liquid nitrogen without the use of CO2 anesthesia. Freezing began at 10:22 am and was completed by 1:40 pm. All flies were frozen in a random order with respect to sex and population. All flies were reared in 50ml vials containing standard yeast medium at 25°C with a 12 hour day/night cycle and adult flies were held in vials for three days in same sex groups of 5 before being frozen.