Intergenomic conflict revealed by patterns of sex-biased gene expression

Intergenomic conflict can affect the distribution of genes across eukaryotic genomes. Because the phenotypic optima of males and females often differ, the fitness consequences of newly arisen alleles might not be concordant between the sexes and can be sexually antagonistic – genetic variants favored in one sex are deleterious in the other. In this article, we demonstrate that previously unexplained patterns of sex-biased gene expression in Drosophila melanogaster might have evolved by sexual antagonism, and that the majority of sex-biased expression is due to adaptive changes in males, implying that males experience stronger selection than females.

Introduction

The ultimate fate of new mutations is inextricably linked to the evolutionary forces acting on them. Selective pressures commonly differ between males and females, and might actually operate in opposing directions [1]. The evolution of sex-limited gene-expression is expected as a result of this ‘sexually antagonistic’ genomic tug of war, leaving a characteristic genetic footprint that depends on the degree of dominance on an allele and which of the two sexes benefits 2, 3, 4. The X chromosome is predicted to be a haven for sexually antagonistic variation because dominant alleles that benefit females are under positive selection during two-thirds of their evolution, and rare recessive male-benefiting alleles are masked in females [3].

Because changes in gene regulation, as opposed to coding sequence, are important generators of phenotypic diversity [5], and coding sequence and regulatory evolution are often decoupled 6, 7, microarrays (i.e. gene expression profiles) offer a window into the evolutionary processes shaping the genome. Studies have revealed that an amazing proportion of the genome is sex-biased in expression (i.e. expression level is sexually dimorphic) 8, 9, 10, 11, 12, 13 and that a significant number of genes show newly derived sex-limitation between recently diverged species [11]. Male- and female-biased genes also exhibit idiosyncratic X-linkage patterns: male-biased genes are strongly deficient, whereas female-biased genes are slightly overabundant 10, 13. However, inferences about the processes driving patterns of sex-biased expression have been constrained by the lack of a conceptual framework in which to (i) identify sex-specific adaptive change; and (ii) translate gene expression data into the classic population genetic constructs of recessive and dominant mutations. Both are fundamental to testing whether patterns of genomic divergence reflect sexually-antagonistic selection.

In this article, we consider published cDNA microarray data [11] under alternative models of adaptive change, with attention to whether shifts in expression correspond to recessive or dominant substitutions. Within this framework, we analyzed thousands of sex-biased genes, including >800 genes that appear to have evolved recent sex-limited expression in either Drosophila melanogaster or its sister taxon D. simulans. Our results demonstrate that previous assumptions about sex-biased expression are too simplistic – sexually dimorphic profiles often reflect downregulation in one sex but more commonly arise from upregulation in the other sex. We also show that the majority of shifts to sex-biased expression are due to adaptive change in males, including cases of increased expression (producing a male-biased gene) and decreased expression (producing a female-biased gene), implying that males experience stronger selection than females. Finally, analysis of the genomic distribution of loci reveals that previously unexplained genomic patterns of sex-biased expression might have evolved by sexual antagonism – a process that is often cited 14, 15, 16, 17, 18, 19, 20, 21 but rarely directly tested.