Heterochiasmy and the establishment of gsdf as a novel sex determining gene in Atlantic halibut

Abstract

Atlantic Halibut (Hippoglossus hippoglossus) has a X/Y genetic sex determination system, but the sex determining factor is not known. We produced a high-quality genome assembly from a male and identified parts of chromosome 13 as the Y chromosome due to sequence divergence between sexes and segregation of sex genotypes in pedigrees. Linkage analysis revealed that all chromosomes exhibit heterochiasmy, i.e. male-only and female-only meiotic recombination regions (MRR/FRR). We show that FRR/MRR intervals differ in nucleotide diversity and repeat class content and that this is true also for other Pleuronectidae species. We further show that remnants of a Gypsy-like transposable element insertion on chr13 promotes early male specific expression of gonadal somatic cell derived factor (gsdf). Less than 4.5 MYA, this male-determining element evolved on an autosomal FRR segment featuring pre-existing male meiotic recombination barriers, thereby creating a Y chromosome. Our findings indicate that heterochiasmy may facilitate the evolution of genetic sex determination systems relying on linkage of sexually antagonistic loci to a sex-determining factor.

Fig 1. Genetic differentiation between phenotypic males and females highlights chr13 as the putative Y-chromosome.

A Average male vs. female ΔAF for 500 kb windows along each assembled chromosome. The genome-average ΔAF is indicated as a horizontal line together with a line indicating Z-score = 3, i.e. the ΔAF-value corresponding to 3 SD above the genome average). B Average ΔAF for 500 kb windows along chr13 with genome average ΔAF and 3 SD above genome average ΔAF indicated by horizontal lines.

Fig 2. Identification of gsdf as the Atlantic halibut sex determining gene.

A Volcano plot representation of differential expression (DE) between individuals stratified by chr13 genotyping-based sex assignment. Dot colors indicate the chromosome from which genes are expressed. The most highly differentially expressed gene, gsdf, is indicated in the upper right corner. B qPCR result showing the sex specific gsdf expression (relative to gtf3c6) across developmental stages. C Pictures of fish representing the developmental stages sampled for gene expression analyses; 1 hpf to 13 dpf.

Fig 3. Cumulative proportion of all recombination events occurring during male and female meiosis in a two-pedigree intercross.

A chr13 with gsdf location indicated. *Mapping interval of markers most highly associated with sex in a previous study [19]. B Male and female linkage maps differ strongly genome-wide. The outermost track visualizes individual chromosomes. Locations of defined Male-only and Female-only meiotic Recombination Regions (MRRs and FRRs) are shown as alternating black/grey rectangular tracks beneath each chromosome. All 24 chromosomes exhibit clear sex-restricted recombination intervals as shown in the tracks labeled MR and FR, showing the cumulative proportions (range 0–1) of all recombination events on each chromosome for males (MR) and females (FR). The innermost track, labeled ΔAF, shows the male vs. female delta allele frequency for 500kb windows along each chromosome (range 0–0.36).

Fig 4. Nucleotide diversity (π) and repeat content is correlated with sex-restricted recombination throughout the genome.

A π distribution in the Atlantic halibut genome, contrasting regions of male-only vs. female-only meiotic recombination, MRRs and FRRs, respectively. B Distribution of Pacific Halibut π estimates for regions mapped back to the Atlantic halibut genome, classified as either MRRs or FRRs. Chromosomal and MRR/FRR assignments were extracted from the Atlantic halibut genome assembly coordinates. C Genome-wide distributions of estimated π for Atlantic-, Pacific- and Greenland halibut and European plaice in regions corresponding to herein defined Atlantic halibut MRRs and FRRs. Numbers above MRR distributions indicate, within species, how much higher π is in MRR regions compared to FRR regions. * The time since divergence of Greenland halibut from Hippoglossus was estimated from [28]. ** Divergence estimate for European plaice was obtained from TimeTree [29]. D Genome-wide distributions of five repeat types in MRRs and FRRs. LINE and SINE represent Long- and short interspersed nucleotide elements, respectively. LTR indicates Long Terminal Repeats. LCR and SR indicate low complexity- and simple repeats, respectively. The Y-axis shows densities of different repeat types after Z-score transformations in relation to genome averages.

Fig 5. Chr. Y carries a transposon-derived LTR upstream of gsdf.

A Agreement between reference individual Chromium pseudo-haplotype assemblies and chrY-markers from Pool-seq of males and females. Black lines show markers along chr13 where pseudo-haplotype 1 (outer track) and pseudo-haplotype 2 (inner track) are in agreement with the Y-allele defined from pool-seq. In the interval 0-10Mb pseudo-haplotype 1 shows a striking agreement with Y-markers defined from pool-seq data. B Schematic representation of the Atlantic Halibut sex chromosomes as inferred from linkage analysis alone and from comparison of Pool-seq defined Y-markers with Chromium pseudo-haplotypes of the assembly male. X and Y haplotypes differ for a 1.2 kb segment derived from a transposable element 2 kb upstream of the gsdf transcription start site. Males carry the insertion allele which contains a lowly methylated CpG-island inferred to act as a derived promoter of gsdf uniquely in males. The shorter gsdf isoforms were observed in Atlantic halibut adult testis in as well as adult ovary and testis of Pacific halibut. The two gsdf gene models labeled chr13Y-specific indicate isoforms detected only in males from three separate developmental stages. An IGV-plot of this region is presented in S12 Fig.