The canonical model of sex-chromosome evolution predicts that, as recombination is suppressed along sex chromosomes, gametologs will progressively differentiate, eventually becoming heteromorphic. However, there are numerous examples of homomorphic sex chromosomes across the tree of life. This homomorphy has been suggested to result from frequent sex-chromosome turnovers, yet we know little about which forces drive them. Here, we describe an extremely fast rate of turnover among 28 species of Ranidae. Transitions are not random, but converge on several chromosomes, potentially due to genes they harbour. Transitions also preserve the ancestral pattern of male heterogamety, in line with the ‘hot-potato’ model of sex-chromosome transitions, suggesting a key role for mutation-load accumulation in non-recombining genomic regions. The importance of mutation-load selection in frogs might result from the extreme heterochiasmy they exhibit, making frog sex chromosomes differentiate immediately from emergence and across their entire length.

Comparative genomic studies are revealing that, in sharp contrast with the strong stability found in birds and mammals, sex determination mechanisms are surprisingly labile in cold-blooded vertebrates, with frequent transitions between different pairs of sex chromosomes. It was recently suggested that, in context of this high turnover, some chromosome pairs might be more likely than others to be co-opted as sex chromosomes. Empirical support, however, is still very limited. Here we show that sex-linked markers from three highly divergent groups of anurans map to Xenopus tropicalis scaffold 1, a large part of which is homologous to the avian sex chromosome. Accordingly, the bird sex determination gene DMRT1, known to play a key role in sex differentiation across many animal lineages, is sex linked in all three groups. Our data provide strong support for the idea that some chromosome pairs are more likely than others to be co-opted as sex chromosomes because they harbor key genes from the sex determination pathway. K E Y W O R D S :Amphibian, Bufo siculus, convergent evolution, conserved synteny, DMRT1, Hyla arborea, Rana temporaria, sex chromosome turnover.

According to the canonical model of sex-chromosome evolution, the degeneration of Y or W chromosomes (as observed in mammals and birds, respectively) results from an arrest of recombination in the heterogametic sex, driven by the fixation of sexually antagonistic mutations. However, sex chromosomes have remained homomorphic in many lineages of fishes, amphibians, and nonavian reptiles. According to the "fountain-of-youth" model, this homomorphy results from occasional events of sex reversal. If recombination arrest in males is controlled by maleness per se (and not by genotype), then Y chromosomes are expected to recombine in XY females, preventing their long-term degeneration. Here, we provide field support for the fountain-of-youth, by showing that sex-chromosome recombination in Rana temporaria only depends on phenotypic sex: naturally occurring XX males show the same restriction of recombination as XY males (average map length ∼2 cM), while XY females recombine as much as XX females (average map length ∼150 cM). Our results challenge several common assumptions regarding the evolution of sex chromosomes, including the role of sexually antagonistic genes as drivers of recombination arrest, and that of chromosomal inversions as underlying mechanisms.

Sex chromosome differentiation in Rana temporaria varies strikingly among populations or families: whereas some males display well-differentiated Y haplotypes at microsatellite markers on linkage group 2 (LG2), others are genetically undistinguishable from females. We analysed with RADseq markers one family from a Swiss lowland population with no differentiated sex chromosomes, and where sibship analyses had failed to detect any association between the phenotypic sex of progeny and parental haplotypes. Offspring were reared in a common tank in outdoor conditions and sexed at the froglet stage. We could map a total of 2177 SNPs (1123 in the mother, 1054 in the father), recovering in both adults 13 linkage groups (= chromosome pairs) that were strongly syntenic to Xenopus tropicalis despite > 200 My divergence. Sexes differed strikingly in the localization of crossovers, which were uniformly distributed in the female but limited to chromosome ends in the male. None of the 2177 markers showed significant association with offspring sex. Considering the very high power of our analysis, we conclude that sex determination was not genetic in this family; which factors determined sex remain to be investigated.

In sharp contrast with birds and mammals, sex-determination systems in ectothermic vertebrates are often highly dynamic and sometimes multifactorial. Both environmental and genetic effects have been documented in common frogs (Rana temporaria). One genetic linkage group, mapping to the largest pair of chromosomes and harbouring the candidate sex-determining gene Dmrt1, associates with sex in several populations throughout Europe, but association varies both within and among populations. Here, we show that sex association at this linkage group differs among populations along a 1500-km transect across Sweden. Genetic differentiation between sexes is strongest (FST  = 0.152) in a northern-boreal population, where male-specific alleles and heterozygote excesses (FIS  = -0.418 in males, +0.025 in females) testify to a male-heterogametic system and lack of X-Y recombination. In the southernmost population (nemoral climate), in contrast, sexes share the same alleles at the same frequencies (FST  = 0.007 between sexes), suggesting unrestricted recombination. Other populations show intermediate levels of sex differentiation, with males falling in two categories: some cluster with females, while others display male-specific Y haplotypes. This polymorphism may result from differences between populations in the patterns of X-Y recombination, co-option of an alternative sex-chromosome pair, or a mixed sex-determination system where maleness is controlled either by genes or by environment depending on populations or families. We propose approaches to test among these alternative models, to disentangle the effects of climate and phylogeography on the latitudinal trend, and to sort out how this polymorphism relates to the 'sexual races' described in common frogs in the 1930s.