An easy solution?
Conflict resolution is an active topic for biologists studying IASC. This owes to the uncertainty of whether traits experiencing IASC will eventually reach resolution, or whether they will remain in this state indefinitely. In order to address this question, we need to consider the possible mechanisms of resolution. This will help us to dissect traits on an individual basis to predict their evolutionary fates and the consequences for whole-organism fitness.
An abundance of theoretical work suggests that conflict could be resolved via a number of mechanisms, which together or in isolation, would relieve the gender load that arises when the sexes are displaced from their fitness optima. Sexual dimorphism is suspected to represent conflict resolution and is thought to be caused by underlying changes in the genetic architecture of a particular trait, which then permits males and females to evolve along their own independent trajectories. This occurs when underlying genetic changes cause the intersexual genetic correlation (rmf) to deviate from 0. In fact, a negative correlation between rmf and sexual dimorphism was identified across most traits in the fly, Prochyliza xanthostoma (Bonduriansky and Rowe 2005a).
To test whether sexual dimorphism represents a robust resolution to IASC, Tigreros and Lewis (2011) applied artificial selection to a dimorphic trait (body size) in opposing directions to each sex. They were able to demonstrate that once dimorphism evolves, it can be irreversible under short-term selection; thus signifying a resistant resolution to sexual conflict. It may then be reasonable to assume that as the evolution of sexual dimorphism is biologically widespread (Darwin 1871), then perhaps conflict resolution is too.
This might hold true to some extent; however, sexual antagonism has in fact been found to affect even highly dimorphic traits (Pischedda and Chippindale 2006; Long and Rice 2007; and see Bedhomme et al. 2008). Furthermore, a review of selection estimates for 89 traits taken from 34 species reinforces these findings (Cox and Calsbeek 2009). As Cox and Calsbeek (2009) state, if the extent of dimorphism does not match up to the fitness peaks of the two sexes, then sexual dimorphism will not be an indication of permanent conflict resolution. More support is provided by Innocenti and Morrow (2010), who identify existing conflict over traits with sexually dimorphic gene expression. In their study, almost 92% of the genes identified were found to be sex biased in expression (Innocenti and Morrow 2010), and only 8% of these were actually shown to be sexually antagonistic. As conflict may be absent for many of these dimorphic transcripts, this could be an indication of widespread conflict resolution. To predict whether these patterns have evolved under positive selection in response to IASC between the sexes, it is necessary to assess the fitness consequences of sex-specific expression levels. Indeed, a look at genome-wide transcription profiles reveals that a considerable amount of sex-biased gene expression is related to sex-specific functions with positive fitness effects (Connallon and Clark 2011a). Therefore, we could envisage that the dimorphic gene expression patterns shown by Innocenti and Morrow (2010) might have evolved as a mechanism to resolve conflict; however, for some genes identified in this study, sex-specific transcription did not always predict sex-specific functions or fitness consequences. This highlights how some transcriptional differences between the sexes may not have evolved directly in response to IASC. These studies provide consistent evidence that although sexual dimorphism could theoretically permit resolution, its use as a signature of resolved conflict should be avoided.
Nevertheless, there are many theoretical examples of how sex-specific gene expression (which may lead to sexual dimorphism) could resolve IASC. One way to achieve this is via sex-specific hormonal cascades or modifiers (Rice 1984). For example, secondary sexual trait expression is determined by testosterone levels in vertebrates (Mougeot et al. 2004; Blas et al. 2006) and titers of juvenile hormone in insects (Emlen et al. 2006), both of which differ between the sexes. These hormone levels will subsequently affect the induction of intracellular signaling that leads to changes in gene transcription. Concentrations of regulatory proteins that target specific genes can also affect the level of gene transcription. These regulatory proteins play an important role in D. melanogaster and Caenorhabditis elegans, for example, by initiating sex-specific developmental pathways (Yi and Zarkower 1999; Yi et al. 2000). There are outstanding questions regarding the birth of such gene expression patterns, as dimorphism could either result from the repression or gain of gene expression in one sex relative to the other (Williams et al. 2008). Nevertheless, there are a handful of studies addressing this question, where the authors have been able to identify genes involved in regulating sexual dimorphism, and predict an ancestral state of monomorphic expression for some traits (Emlen et al. 2007; Williams et al. 2008; Moczek and Rose 2009; Williams and Carroll 2009; Khila et al. 2012). A phylogenetic analysis of wing pattern evolution in butterflies also found evidence that for some traits, sex-limited gene expression occurred simultaneously as the trait arose in a lineage; whereas, for other traits there was an ancestral state of dimorphic expression, followed by the subsequent loss of expression in one sex (Oliver and Monteiro 2011).
An additional mechanism for controlling sex-specific gene expression is through alternative splicing (McIntyre et al. 2006). Here, sex will determine the final protein form that is produced from a shared coding region in the genome. This is a posttranscriptional process, where the RNA produced from a single gene is spliced in alternate ways through the joining of different exon combinations. McIntyre et al. (2006) conducted a genome-wide analysis of alternative splicing in D. melanogaster, discovering that at least 12% of all genes are spliced in this sex-specific manner. Although empirical data are lacking, it is possible that the patterns of sex-biased alternative splicing described here may have evolved to resolve IASC.
The translocation of genes to sex chromosomes could also facilitate sex-limited gene expression (Charlesworth and Charlesworth 1980; Rice 1984; Bachtrog 2006). It is thought that some male-benefit, female-detriment genes have been translocated from autosomes to the Y chromosome, for example (Bachtrog 2006), consequently enabling males to evolve independently of females (in species where females are the homogametic sex). In order for this to resolve conflict, however, females must not require the translocated gene for functional purposes. As this is unlikely for most genes, a favorable alternative hypothesis is that genes are duplicated, translocated, and then undergo sex-limited gene expression (Ellegren and Parsch 2007; Baur et al. 2008; Connallon and Clark 2011b; Gallach and Betrán 2011). In this scenario, duplications initially produce additional gene copies with identical function, but they can later be released from the ancestral function by evolving freely through mutation and selection (Wyman et al. 2012). When followed by sex-limited expression, this can subsequently allow the sexes to diverge in their trait values. The duplications produced can also be translocated to nonsex-chromosome locations (Mank 2009), with sex-specific modifiers evolving to control their expression on autosomes too. A recent analysis of gene expression by Wyman et al. (2012) found that these duplicate pairs are typically male biased in expression. This is suggested to be a product of sperm competition, as this can create strong sexual selection pressures on male traits, such as those related to ejaculate function (and may also have implications for IRSC; see Box 1).
Genomic imprinting presents another possible mechanism to alleviate IASC through sexual dimorphism (Day and Bonduriansky 2004; Patten and Haig 2008). Imprinting relies on changes to DNA methylation patterns that occur during gametogenesis in parents and affect the expression of genes in offspring. The best-known examples are igf and igf2, growth factors that are known to be silenced when inherited paternally (Barlow 1995; Ferguson-Smith and Surani 2001). Simulations indicate that this pattern could arise due to IASC, as long as the benefits of imprinting an antagonistic allele in one sex outweigh the costs of doing so in the other (Day and Bonduriansky 2004). The possibility for an imprinting modifier allele to invade a population in this way is also heavily dependent on dominance, as shown in simulations by Cleve and Feldman (2007). Their study built upon a previous model by Day and Bonduriansky (2004), where only additive variation for fitness was considered. Despite these findings, for imprinting to fully resolve conflict it would be necessary for parents to imprint genes in a gamete karyotype-specific manner. For example, males should imprint genes so that male-benefit sexually antagonistic alleles are switched off in X-bearing sperm only. This would enable males to increase the fitness of sons, without detrimentally affecting the fitness of daughters. This mechanism would include imprints on autosomes that were dependent on whether they are found in X or Y sperm. Despite this theoretical requirement for resolution, the occurrence of imprinting in this manner is yet to be proven empirically. So far, 80 genes are recognized as being imprinted in mammals (Morison et al. 2005), although others propose that this figure could actually exceed 600 (Luedi et al. 2005). Imprinting therefore presents another potential mechanism with capabilities of resolving conflict on a genome-wide scale, but one that lacks empirical support.
Sexual dimorphism has also been shown to increase for some traits as a result of condition dependence, by weakening rmf (Bonduriansky and Rowe 2005b). Condition dependence is expected to evolve for traits that are under strong sexual selection, which as a result become exaggerated to a point where they are costly to produce – hence the expression of these traits comes to reflect condition (Rowe and Houle 1996). If the level of condition dependence of a trait becomes unequal between the sexes, then this may permit the elaboration of a trait in only one sex, consequently exaggerating the degree of sexual dimorphism; however, Bonduriansky and Rowe (2005b) do not quantify the fitness consequences of sexual dimorphism through condition dependence; therefore, its ability to resolve conflict was not clear. It is also necessary to explore the genetic mechanisms facilitating this as it is also unclear how trait rmf affects the potential for condition dependence (Bonduriansky and Rowe 2005b).
Rather than confronting the genetic basis of IASC, some species appear to have evolved an alternative way to mask the effects of sexually antagonistic genes – sex ratio adjustment (SRA). It is conceivable that this strategy presents a means of partially resolving IASC when it is not possible to achieve sex-limited gene expression via changes to trait genetic architecture. A study conducted in the wild and follow-up laboratory investigation revealed how side blotched lizards, Uta stansburiana, can choose sperm depending on the phenotype of their mate (Calsbeek and Sinervo 2004). This enables females to select the sex of their offspring as a remarkable way to diffuse IASC. For instance, females mated to larger males produce more sons because size is positively correlated to male fitness, but negatively correlated to female fitness. In accordance, a small sire results in increased production of daughters. Both sexes benefit from this since it presents an opportunity to maximize the fitness of their progeny in the face of antagonistic alleles. There are parallel findings in brown anoles, Anolis sagrei (Calsbeek and Bonneaud 2008; Cox and Calsbeek 2010); fruit flies, D. melanogaster (Connallon and Jakubowski 2009), and barn owls, Tyto alba (Roulin et al. 2010).
Katsuki et al. (2012) focused on SRA in broad-horned flour beetles. Interestingly, the sex of offspring produced by a female depended on her own fitness, rather than that of her mate. A low fitness female produced opposite-sex offspring, whereas higher fitness females increased the production of daughters. Why the fitness of their mate had no effect on offspring ratio was not clear, but by basing offspring ratio on recognition of their own fitness, females could increase their inclusive fitness and that of their mates. Although lacking any pertinent evidence, they suggest females could alter their hormone levels to determine offspring sex.
A simple model was also developed by Blackburn et al. (2010) to explore the circumstances under which SRA could evolve. Providing that sufficient genetic variation exists at SRA loci, then SRA is expected to evolve rapidly. They note that while they only looked at a single gene, SRA is equally likely to evolve in the presence of many sexually antagonistic genes if it results in a net increase in fitness.
As well as allowing us to understand the selection pressures leading to sex ratio adjustment, these studies reinforce the argument that IASC can in fact have evolutionarily important outcomes. Nevertheless, to obtain a more complete picture, the proximate mechanisms leading to SRA require much greater empirical attention.
It would seem that IASC could be eliminated through both genetic and strategic innovations; however, this is not to say that sexual antagonism for every trait may be so easily resolved. In particular, there is much to learn about the genetic mechanisms behind the evolution of sexual dimorphism and how these work to alleviate IASC (Rhen 2000; Rice and Chippindale 2001, 2002; Day and Bonduriansky 2004; Bonduriansky and Rowe 2005b); especially in the face of strong intersexual genetic correlations (Lande 1980) or when pleiotropic genes are involved (Badyaev 2002; Ellegren and Parsch 2007; Van Doorn 2009). Moreover, despite expectations that sex-biased gene expression could rapidly evolve to diminish sexual conflict (Reeve and Fairbairn 1996, 2001; Van Doorn 2009), others describe this conclusion as premature (Stewart et al. 2010). This is supported by evidence that low levels of sexual antagonism can exist for traits that appear to be sexually dimorphic (Harano et al. 2010). Perhaps, IASC in some traits can only ever be partially resolved, with a simmering level of sexual antagonism always maintaining fitness levels below optima for the sexes. To understand, this requires a look at the potential barriers to conflict resolution, for which there is some convincing evidence.