- Top of page
- Materials and Methods
- Results and Discussion
- Conflict of Interest
- Appendix 1: FASTA alignment of Acp62F from five Drosophila species
- Appendix 2: Genotype counts and inferred allele frequencies at generations 6 and 8
In a broad range of taxa including vertebrates, invertebrates, fungi, and plants, some of the most rapidly evolving proteins encoded in the genome are those contributing to reproductive success (for reviews, see Clark et al. 2006; Chapman 2008; Wong 2011). In Drosophila, for example, seminal fluid proteins (SFPs) that are transferred from males to females during copulation diverge quickly at the sequence level, and the proportion of SFP-encoding genes subject to positive selection is unusually high compared with the rest of the genome (e.g., Civetta and Singh 1995; Haerty et al. 2007). Moreover, the complement of SFPs produced by different species can vary drastically, with apparently high rates of gene loss and recruitment of new genes (e.g., Holloway and Begun 2004; Mueller et al. 2005; Wagstaff and Begun 2005a, 2007; Haerty et al. 2007; Findlay et al. 2008; Kelleher et al. 2009).
Several hypotheses have been proposed to explain the rapid evolution of reproductive proteins. In internally fertilizing animals, a leading proposal is that postcopulatory sexual selection is responsible, for example, via sperm competition or sexual conflict. Sperm competition arises when sperm from multiple males are present at the same time in the reproductive tract of a female. If an SFP variant leads to an increased paternity share, then that variant should be favored by sexual selection. Consistent with an effect of sperm competition on SFP evolution, several rapidly evolving Drosophila melanogaster SFP genes have known effects on sperm storage and/or sperm competition (Acp36DE: Neubaum and Wolfner 1999; Chapman et al. 2000; Acp29AB: Wong et al. 2008; CG9997: Ram and Wolfner 2007).
Similarly, a number of Drosophila SFPs are known to have effects on processes potentially involved in interlocus sexual conflict. Interlocus sexual conflict arises when the optimal outcome of an interaction between the sexes is different for males and females. For example, females of some species may gain by mating with multiple males (Simmons 2005) owing to sperm depletion (e.g., Thornhill and Alcock 1983), sperm quality (e.g., Keller and Reeve 1995), and other indirect benefits, and/or from direct benefits such as nuptial gifts (e.g., Simmons et al. 1999). By contrast, it is typically in a male's interest that his partner does not mate with other males. As such, the optimal remating rate may be higher for females than it is for males (e.g., Gavrilets and Hayashi 2006). In D. melanogaster, the “sex-peptide (SP) network” of proteins modulates remating rate, as well as several other postmating responses (Chapman et al. 2003; Liu and Kubli 2003; Ram and Wolfner 2009; LaFlamme et al. 2012), and at least one member of the network has been subject to positive selection (CG9997; Wong et al. 2012). The SP network may in fact be involved in sexual conflict in several different ways; in addition to its effects on female remating, a functional SP network is necessary for the mating-induced reduction in female life span observed in this species (Wigby and Chapman 2005).
Consistent with the hypothesis that sexual selection is a primary cause of rapid reproductive protein evolution, comparative studies indicate a correlation between mating system and the rates of evolution of some reproductive proteins, including SFPs. That is, for some individual proteins, rates of evolution are higher in polyandrous lineages than in monandrous lineages, where postcopulatory sexual selection is less likely to occur (e.g., Ramm et al. 2008; Finn and Civetta 2010; Prothmann et al. 2012). Moreover, average rates of reproductive protein evolution are higher in polyandrous lineages in comparison with monandrous lineages in primates and in Drosophila (Wagstaff and Begun 2005b, 2007; Kelleher et al. 2007; Almeida and Desalle 2009), again suggesting that sexual conflict is important in the evolution of reproductive proteins. Additional sequence-based evidence for an effect of sexual selection on rates of reproductive protein evolution comes from Clark et al. (2009), who showed strong linkage disequilibrium between the abalone sperm protein lysin and its egg-receptor vitteline envelope receptor for lysin, as predicted under models of sexual selection.
In this study, we investigated the evolution of the D. melanogaster SFP Acp62F. Acp62F has been extensively studied using genetic and biochemical methods, but its function nevertheless remains unclear. The Acp62F protein is a protease inhibitor (Lung et al. 2002), a biochemical characteristic that it shares with numerous other SFPs (Mueller et al. 2004; Laflamme and Wolfner 2012). Overexpression of Acp62F in larvae or adults is toxic (Lung et al. 2002; Mueller et al. 2007), and a QTL containing the Acp62F locus is associated with postmating female mortality in interspecific introgression lines (Civetta et al. 2005), suggesting that this protein may be involved in seminal fluid toxicity. If so, then one would expect this protein to have other fitness-enhancing effects that outweigh its costs to females (unless toxicity is itself beneficial to males – see Johnstone and Keller 2000). Such a benefit has not been identified; indeed, Mueller et al. (2008) found that Acp62F knockout males performed better in sperm competition than did wild-type (wt) males. Mueller et al. (2008) also showed that Acp62F is required for normal proteolytic cleavage of the egg-laying hormone Ovulin, but the functional consequences of this phenotype are unclear as mates of Acp62F knockout males did not show any obvious defect in egg laying.
Here, we use computational and manipulative approaches to investigate the nature of selection acting on Acp62F. Using comparative sequence analyses, we test for a signature of positive selection on Acp62F. Furthermore, in a two-way factorial evolution experiment, we independently manipulate the opportunities for selection arising from variation in sexual and nonsexual fitness in replicate experimental populations polymorphic for Acp62F wt and null alleles, and then determine the consequences of this manipulation by tracking changes across generations in allele frequencies. This approach allows the net effects of Acp62F on sexual and nonsexual fitness to be integrated by the evolutionary process itself.