Discussion
Using two newly assembled genomes, reduced-representation sequencing (DArTseq) and whole-genome resequencing (WGS) for panels of individuals of known sex, and amplicon sequencing, we aimed to identify sex-linked genomic regions and develop and test molecular sexing markers in twoMacquaria perch sister-species along with two confamilial species. One of these species, golden perch, is inferred by karyology to have XY sex-determination and homomorphic sex chromosomes, and the confamilial Murray cod to have XY and heteromorphic chromosomes (Shams et al., 2019). We did not uncover DArTseq SNPs consistent with Y-linkage or XY-gametology in either golden perch or Macquarie perch, but found loci polymorphic only in one sex (sometimes males, sometimes females), scattered through different scaffolds. This suggests that regions with old strata (with long-term lack of recombination), if present on sex-chromosomes of these species, are short. Finding an order of magnitude more sex-linked DArT loci in golden perch than in Macquarie perch, none of which were on the golden perch scaffold VMKM01003747 homologous to Macquarie perch scaffold 633, indicates that these species have independently-evolved sex-determination systems, with the Macquarie perch one being more recent. Alternatively, Macquarie perch could have lost sex-linked variation during the Pleistocene bottleneck involved in colonization of the Murray Darling Basin from the coastal range or during recent range contraction and population isolation (Pavlova, Gan, et al., 2017). DArTseq has been used to find sex-linked loci and infer sex-determination systems, including in catfish Clarias gariepinus with homomorphic sex-chromosomes, where both XY and ZW sex-determination were inferred based on the presence of moderately male- and moderately female-linked loci (Nguyen et al., 2021).
Exploring male and female population pools of WGS data for Macquarie perch with CMH tests guided us to focus on scaffold 633 region 90152-94017, enriched with sex-linked loci, including XY-gametologous region 93182-94017, encompassing five SNPs and two indels in the 146-bp sexing region 93182-93327, and an additional SNP at 94017. Female pools had a single allele and male pools had two approximately equally represented alleles for these six SNPs. Individual WGS genotypes showed all females to be homozygous and 70% of males heterozygous at these loci. Rare past or ongoing recombination within this XY-gametologous region was suggested by 30% of males being homozygous for X- or Y- alleles for up to seven of these eight loci. Comparisons of male and female read depth for sex-linked loci and separately for sex-linked alleles (Figs. 1, 2) suggested that some females and males might be haploid for part of this XY-gametologous region (i.e. through a deletion on one of the X-variants). Whereas similar read depth for males and females for the first and last loci in this region (93182 and 94017) implies diploidy of these SNP loci in both sexes, for intervening loci 93229, 93299, 93315 and 93327 most females appeared haploid (X0) for this region: females had consistently lower depth than males (Fig. 1), and similar depth for X-alleles as each of the X- and Y- alleles in males (Fig. 2). Lower average X-allele depth in males than the Y-allele suggests that some males could be haploid for Y (i.e. Y0).
Upstream of the XY-gametologous region 93182-94017 was the region 90152-92914 containing 13 SNPs bearing recent-Y-specific polymorphism, which might be a more recently emerging region of sex-difference. Patterns of appearance of these recent-Y-specific polymorphisms in males (Fig. C1) suggest that multiple occasionally recombining Y-haplotypes are segregating in Macquarie perch. The XY-gametologous region itself included another recent-Y-specific SNP 93879, suggesting occasional recombination between X and Y haplotypes (Fig. C1). The lower read depth for 93879 in males compared to females suggested Y-specific deletions at this locus (Figs. 1 and 2). Manta analysis did not detect any structural variants that would support X- or Y- deletion polymorphism consistent with these patterns (Fig. D1), however our read depth coverage (~10-20x) was much lower than the ~50x used in original tests of Manta (Chen et al., 2016), potentially underpowering our analysis.
Overall, in the panel of 22 loci with sex-biased patterns of heterozygosity in scaffold 633 region 90152-94017 (8 XY-gametologs and 14 loci bearing recent-Y-specific polymorphism; Fig. C1), screened in 100 individuals from Dartmouth and Yarra, all females were homozygous for X-alleles, and Y-alleles were present in 100% of males for 94017, 98% for 93229, 93299 and 93327, 96% of for 93182, 93184, 93278 and 93315, and in 14%-76% for the other loci. This suggests a working hypothesis of sex-determination in which only individuals homozygous at the X-alleles at all 22 of these loci develop a female phenotype, whereas presence of Y-alleles at some of the loci leads to a male phenotype. However, presence of the Y-allele at 93229 in 36% of King Parrot Creek and Abercrombie females is not consistent with this mechanism based on the Dartmouth and Yarra populations (Fig. B4). This single deviation from the suggested model call for other mechanisms of producing a female phenotype.
Tests of our molecular sexing protocol, that targeted the Y-allele of 93229 in the 146-bp sexing region of Macquarie perch scaffold 633, in other populations of Macquarie perch, in golden perch, and in more distant relatives from the family Percichthyidae (Murray cod and trout cod) showed that this sexing region is Macquarie perch-specific, and could even be limited to a subset of the species. This Y-allele was present in all but one male out of 50 tested, hence the test should correctly identify Dartmouth and Yarra individuals as males (based on two or three bands on a gel) or females (based on one band), except ~2% of males may be missexed as females. The same test is not expected to work well in other populations of Macquarie perch, because 20% of putative females from King Parrot Creek and 60% from Abercrombie had the male-specific allele. Unless these individuals are feminized (egg-producing) males, female homozygosity for the reference allele at 93299 might be limited to some populations. Unique variation at these loci in amplicon sequencing for two coastal individuals of unknown sex further suggests that sex-specific variation in Macquarie perch could be specific to some populations. Intraspecific variation in chromosomal locations of sex-determining loci occurs in other vertebrates (Kijas et al., 2018a; Lubieniecki et al., 2015; Rodrigues, Merilä, Patrelle, & Perrin, 2014). In the spotted snow skinkNiveoscincus ocellatus , highland and lowland populations, which differ in their degree of influence of environmental temperature on sex, shared- and population-specific sex-linked loci were detected, suggesting a different degree of sex-chromosome differentiation between populations (Hill, Burridge, Ezaz, & Wapstra, 2018). Different sex-determination systems can evolve rapidly in populations of the same species in response to loss of sex-chromosomes by drift. For example, natural populations of zebrafish (Danio rerio ) have a monogenic ZW sex-determination system, whereas polygenic sex-determination systems of different laboratory strains have been inferred (Anderson et al., 2012; Bradley et al., 2011; Liew et al., 2012) which evolved de-novo after loss of W-alleles (Wilson et al., 2014).
Understanding the nature of homo/hemizygosity at the XY-gametologous region of scaffold 633 was obscured here by the drop in read depth for the sex-linked region compared to the genome-average, including zero coverage around position 93400 for 59% of individuals regardless of their overall sequence depth (Fig. 1). This lower depth might be linked to difficulty of sequencing through a low-complexity region, including the (GT)n microsatellite that ends around 93193 (Fig. B1). Accumulation of repetitive sequences was previously reported for golden perch and Murray cod sex-chromosomes (Shams et al., 2019). Given its high read-depth coverage, amplicon sequencing may provide a more reliable way of genotyping genomic regions that contain low-complexity fragments. Despite the locally reduced read depth in WGS, WGS-based genotypes were identical to amplicon-based genotypes for the four individuals analysed with both techniques. Apparently real Y-allele homozygotes were observed even for males with reasonable WGS depth, including a male for which amplicon sequencing was done. Low depth also did not preclude detecting sex-specific and population-specific patterns of depth coverage (Fig. 2). Thus, an X0 or Y0 genotypes for some XY-homologous loci in some individuals likely represents a biological phenomenon. It is unlikely that the drop in depth for the scaffold 633 sexing region was due to failed mapping due to repetitive regions: BLAST search of the 146-bp sexing region did not find highly similar regions in other Macquarie perch scaffolds.
The scaffold 633 sex-linked region 90152-94017 was not annotated to a known gene, but was located ~39 Kb upstream of gene ofSOX transcription factor family, and 41 Kb downstream of a protein with Dbl homology (DH) domain superfamily, associated with guanyl-nucleotide exchange factor activity and protein binding (interacting selectively with any protein or protein complex).SOX genes regulate fate of stem and progenitor cells during development (Sarkar & Hochedlinger, 2013). Three genes in particular,SOX9 , SOX8 (Koopman, 2005) and SOX3 , a progenitor of the SRY gene in mammals (Graves, 2013), have a role in male sex determination in different species, with their expression inducing testes development. Thus, it is possible that SOX gene g10286, modulated by the variation within the sex-linked region 90152-94017, controls gonadal development in Macquarie perch.
Alternatively, polygenic sex-determination in Macquarie perch is possible, given the many loci detected as putatively female- and male-linked by the CMH tests (at p<1e-5; Supplementary Material S3). In sea bass Dicentrarchus labrax at least four loci are associated with sex determination (Palaiokostas et al., 2015). Also, both XY and ZW sex-determining systems are present on different chromosomes in some Metriaclima cichlids, with epistasis between genotypes at two loci controlling sex (Ser, Roberts, & Kocher, 2010). Three alleles (X, Y, and W) segregate at the sex determination locus and interact in platyfish Xiphophorus maculatus (Schultheis, Böhne, Schartl, Volff, & Galiana-Arnoux, 2009). Future studies are needed to demonstrate whether sequence variation at the sex-linked region on Macquarie perch scaffold 633 has mechanistic links with development of female or male phenotypes. Acquiring WGS data for known-sex samples from other populations will be crucial for improving our understanding of sex-determination in this species. Sex-determination in Macquariaappears to be complex and dynamic, potentially involving multiple sites or genomic regions and different mechanisms in quite closely-related lineages, including within species.