In this post, I review the evolution of recombination rate between the sex chromosomes. As considered in previous posts, low rates, and more often absence, of crossover between parts of the X and the Y plays an important role in sex chromosome evolution and is frequently considered a defining characteristic of sex chromosomes. Wide variation in recombination rate between the X and Y is observed across different species: in some species recombination is completely absent in one sex, for example in Drosophila or the amphipod crustacean Gammarus chevreuxi, whereas in other species the non-recombining region of the sex chromosomes is limited to very few loci, for example in fugu pufferfish or asparagus. The full gamut of differences between these extremes in recombination-rate variation exists. As I discuss below, various processes have been invoked to explain this variation, and the empirical support is mixed.
In 1937, Darlington proposed low rates of crossing-over were essential to sex chromosome ontogeny. From his study of sex chromosome cytotypes, Darlington concluded that the evolution of sex chromosomes was inseparable from the evolution of separate sexes in an ancestrally hermaphroditic population, assuming that sex was heritable and followed Mendelian patterns of inheritance. At the very least, Darlington proposed, the complete transition to separate male and female individuals from a hermaphroditic population required two mutations: individuals with a male sterility mutation that exhibit a female phenotype and individuals with a female sterility mutation that exhibit a male phenotype. Such sex-specific sterility alleles could invade a hermaphroditic population when selection acts to avoid the cost of inbreeding depression, to maximize fitness in sexually dimorphic phenotypes, or a combination of both. In the transition to separate sexes, each sex-specific sterility mutation must arise on different haplotypes to avoid the deleterious fitness consequences of double-sterile haplotypes, and therefore the recombination rate between these two haplotypes must be low or evolve to be low. A corollary of the two-sterility model of sex chromosome evolution is that sex chromosomes evolve from regular (autosomal) chromosomes. Indeed, ancient molecular homology between X and Y genes in many species supports the hypothesis that sex chromosomes evolved from a homologous pair. As any chromosome could harbour the pair of sex-specific sterility mutations, the loss of recombination, favoured by selection to avoid the cost of double-sterile individuals, will create sex chromosomes.
Yet recombination loss does not seem to have occurred in a single step. First observed by Lahn and Page in humans, studies of molecular divergence suggest the data are best explained by several distinct divergence times across the sex chromosomes rather than one single loss of recombination event. Whereas some groups of genes which are physically close to each other on the X have similar divergence times from the Y, discrete clusters of genes often have markedly different levels of X-Y divergence. The clusters of divergence have been termed ‘evolutionary strata’ to invoke distinct evolutionary periods across the chromosome, similar to the evolutionary eras captured by geological strata. Strata have been observed on sex chromosomes in a variety of organisms including mammals, Drosophila, the predominantly haploid anther-smut funguses, some birds but not others and some plants like Silene latifolia but not Salix or Rumex. The observation of evolutionary strata suggests that recombination loss across the sex chromosomes can be more complex than a simple transition from hermaphroditism. One way to explain evolutionary strata is that subsequent to the transition to separate males and females, more genomic regions become locked into segregation with the sex-specific sterility mutations. This hypothesis proposes that the non-recombining region of sex chromosomes can evolve and grow in size over evolutionary time.
A primary selective mechanism that has been thought to drive the subsequent spread of recombination suppression on sex chromosomes is sexual antagonism. In 1948, A.J. Bateman revolutionized the view of sex-specific selection in a way that massively broadened the parameter space under which genetic polymorphism could be maintained. In Darwin’s original view, sex-specific selection was a difference in the strength or in the presence/absence of selection between the sexes. In contrast, under Bateman’s proposal, selection could act in opposite directions in each sex because males and females have different reproductive strategies and thus different evolutionary optima. Bateman proposed that phenotypes could be favoured in one sex that were deleterious in the other sex, which can be described as a difference between males and females in their fitness optima. Selection on phenotypes participating in the trade-off between offspring investment and offspring number, for example, should be different between males and females, and this form of selection is known as ‘sexually antagonistic’ selection.
Alleles contributing to sexually antagonistic traits can be maintained for long evolutionary time periods because their selection pressure will switch when inherited in the opposite sex. Whereas one allele would be beneficial to males, the other would be beneficial to females, and under some circumstances neither allele can spread to fixation. In populations evolving under weak natural selection, differential selection between the sexes can cause a departure from mean optimal phenotypes that is characterized as a fitness load. This fitness load can be countered by changes to genome organization that minimize the fitness cost of this load. The stable maintenance of a polymorphism due to sexually antagonistic selection allows for the invasion of genes that lower the rate of recombination between the X and Y. For example, in support of this hypothesis, study of the sex chromosomes of monkeys and guppies suggests species with more sexual dimorphism have lower rates of recombination. As sexually antagonistic alleles evolve to be linked with the sex determination region, the non-recombining region of the sex chromosomes expands. In this way, sex-specific inheritance limits the cost of sexually antagonistic variation.
Although perhaps an intuitive explanation for the extension of recombination loss across sex chromosomes, these models require very strong sexually antagonistic selection. The requirements for the strength of sexually antagonistic selection are less stringent when some recombination suppression, and thus linkage, already exists. The invasion of a recombination modifier is much less restricted when the locus under selection and the Sex-Determining Region (SDR) are already on the same chromosome and even less constrained when recombination rate is already low in that region. For example, the sex chromosomes of papaya seem to have arisen in a region with ancestrally low rates of recombination. This observation also predicts that recombination loss is more likely to spread across the sex chromosomes if a SDR is established than it is to recruit other chromosomes to fuse with the sex chromosomes, known as a neo sex chromosomes.
Gene expression may also evolve to be sex-specific or sex-biased in cases where alleles with opposing fitness effects between the sexes are maintained by selection. To avoid the fitness costs of sexually antagonistic genetic load, one can expect the invasion of gene expression modifiers that cause sex-specific expression of genes underlying sexually antagonistic traits. Sex-specific gene expression can evolve to increase male and/or female fitness through cis regulatory changes. The existence of the evolution of sex-specific gene expression as a means to alleviate load from sexually antagonistic variation is currently being actively investigated in animals. Some metrics suggest differentiation in allele frequencies between the sexes at loci with sex-biased gene expression may represent this type of process. Studies of sex-specific gene expression suggest genes with sex-biased expression are more likely to be sexually antagonistic in humans, Drosophila, and fishes, but recent simulation studies suggest the association between sexual antagonism and sex-biased gene expression can also arise from random processes. Although hinting at an association between gene expression and sexual antagonism, these studies struggle to associate sex-specific gene expression to fitness.
Sexually dimorphic gene expression with sex-specific fitness effects is expected to cause sex-specific dominance reversal for genetic variation in fitness. Sex-specific dominance reversals have been observed for sexually antagonistic major effect loci such as RXFP2 in soay sheep and VGLL3 in salmonids, and sex-specific dominance reversals have significant and broad fitness effects in the seed beetle Callosobruchus maculatus. Quantitative trait locus (QTL) analysis suggests sex-specific trait expression may avoid load in plants and in haplodiploid systems, such as Nylanderia ants, but not for the mandibles of flour beetles. These studies suggest selection may act to avoid sexually antagonistic load, but the underlying genetics remains unclear.
The dominance of alleles under selection will factor into the expectations of change in recombination rate, and therefore the evolution of recombination rate between sex chromosomes may be affected by the evolution of sexually dimorphic gene expression. As the intensity of selection will be mediated by dominance, in cases where dominance reversals relieve the pressure of sexual antagonism, recombination modifiers are less likely to invade. For example, birds with homomorphic sex chromosomes have more sexually dimorphic gene expression and changes to sex-specific gene expression can account for recombination rate on the sex chromosomes of guppies. Similarly, haploid gene expression can have important effects on recombination rate evolution on sex chromosomes. A recent extension to the Charlesworth and Charlesworth (1980) model of invasion of recombination modifiers between loci under sexually antagonistic selection considers haploid selection. As selection is expected to be more efficient on the haploid stage of the lifecycle, increased efficacy of selection in the haploid stage can indirectly increase selection on a recombination modifier beyond, or even counter to, the expectation for selection in the diploid phase. These studies suggest gene expression can have a significant impact on changes in recombination rate between the sex chromosomes.
In summary, the evolution of reduced recombination between X and Y is a dynamic process. Recombination rates can continue to evolve beyond the initial invasion of sex chromosomes, whereas in some systems the non-recombining region remains small and stable for long time periods. Sexual antagonism is also likely to play an important role in the evolution of recombination rate. Because both sex-specific inheritance and sex-specific gene expression can participate in alleviating the load of sexually antagonistic genetic variation, both processes are intertwined. While sex-specific gene expression in the diploid phase may limit the loss of recombination between X-Y, recent work suggests haploid-specific expression of sexually antagonistic variation could extend the region of recombination loss.