Support for the Dominance Theory in Drosophila Transcriptomes

Abstract

Interactions among divergent elements of transcriptional networks from different species can lead to misexpression in hybrids through regulatory incompatibilities, some with the potential to generate sterility. While the possible contribution of faster-male evolution to this misexpression has been explored, the role of the hemizygous X chromosome (i.e., the dominance theory for transcriptomes) remains yet to be determined. Here, we study genome-wide patterns of gene expression in females and males of Drosophila yakuba, Drosophila santomea and their hybrids. We used attached-X stocks to specifically test the dominance theory, and we uncovered a significant contribution of recessive alleles on the X chromosome to hybrid misexpression. Our analyses also suggest a contribution of weakly deleterious regulatory mutations to gene expression divergence in genes with sex-biased expression, but only in the sex toward which the expression is biased (e.g, genes with female-biased expression when analyzed in females). In the opposite sex, we found stronger selective constraints on gene expression divergence. Although genes with a high degree of male-biased expression show a clear signal of faster-X evolution of gene expression, we also detected slower-X evolution in other gene classes (e.g., female-biased genes). This slower-X effect is mediated by significant decreases in cis- and trans-regulatory divergence. The distinct behavior of X-linked genes with a high degree of male-biased expression is consistent with these genes experiencing a higher incidence of positively selected regulatory mutations than their autosomal counterparts.

Keywords: X chromosome; faster-Autosome evolution; faster-X evolution; sexually dimorphic gene expression; speciation.

Figure 1

Diagram of the karyotypes analyzed in this study. Horizontal short bars represent sex chromosomes (the Y chromosome is shown with a hook) while long bars represent autosomes. Gray, D. yakuba chromosomes; white, D. santomea chromosomes. Att-X, attached-X females; T × S, hybrids of first generation from the cross between D. yakuba Taï18E2 females and D. santomea STO.4 males; S × T, hybrids of first generation from the cross between D. santomea STO.4 females and D. yakuba Taï18E2 males; Cy × Cs, hybrids of first generation from the cross between D. yakuba C(1)RM females and D. santomea C(1)RM males; Cs × Cy, hybrids of first generation from the cross between D. santomea C(1)RM females and D. yakuba C(1)RM males. RNAseq detailed information for each of the 12 karyotypes/genotypes is shown in Table S1.

Figure 2

Distribution of genes with different degrees of sex-biased expression across the genome. The number of genes in each class is shown on the right column. See Table S2 for the number of genes in each category.

Figure 3

Gene expression divergence as a sexually dimorphic trait in the female and male transcriptomes. The horizontal line inside each box indicates the median. The length of the box and the whiskers represent 50% and 90% CI, respectively. The numbers of genes analyzed in each sex-biased category are shown under the female and male symbols. Note that only genes expressed in both sexes were included in the analysis. Probabilities are based on Mann-Whitney tests, where NS indicates not significant, * P < 0.05, ** P < 0.001, *** P < 1 × 10⁻⁴. Gene expression divergence data are based on the comparison of D. yakuba Taï18E2 and D. santomea STO.4 stocks.

Figure 4

Tests of faster-X evolution for gene expression divergence in the female transcriptome (A) and the male transcriptome (B). Gene expression divergence data are based on the comparison of D. yakuba Taï18E2 and D. santomea STO.4 stocks. X, X chromosome; A, Autosomes. Probabilities are based on Mann-Whitney tests, where NS indicates not significant, * P < 0.05, ** P < 0.001, *** P < 1 × 10⁻⁴. Black and red asterisks indicate slower-X and faster-X, respectively (see Figure 3 legend for boxplot explanation).

Figure 5

Misexpression of autosomal genes with different degrees of sex-biased expression in standard hybrid females, attached-X (att-X) hybrid females, and hybrid males. The numbers of genes analyzed in each sex-biased category are shown on top of the panels (chromosomes 2 and 3). Note that only genes expressed in both sexes were included in the analysis. Misexpression data are based on first generation hybrids from the crosses between D. yakuba females [Taï18E2 or C(1)RM] and D. santomea males [STO.4 or C(1)RM]. Probabilities are based on Mann-Whitney tests, where NS indicates not significant, * P < 0.05, ** P < 0.001, *** P < 1 × 10⁻⁴. (See Figure 3 legend for boxplot explanation.).

Figure 6

Misexpression of autosomal genes with sex-specific and nonsex-biased expression in standard hybrid females, attached-X (att-X) hybrid females, and hybrid males. The numbers of genes analyzed in each category are shown on top of the panels (chromosomes 2 and 3). FSGs, genes with female-specific expression; MSGs, genes with male-specific expression. Misexpression data based on first generation hybrids from the crosses between D. yakuba females [Taï18E2 or C(1)RM] and D. santomea males [STO.4 or C(1)RM]. Probabilities are based on Mann-Whitney tests, where NS indicates not significant, * P < 0.05, ** P < 0.001, *** P < 1 × 10⁻⁴ (see Figure 3 legend for boxplot explanation).