Chromatin landscape associated with sexual differentiation in a UV sex determination system

Abstract

In many eukaryotes, such as dioicous mosses and many algae, sex is determined by UV sex chromosomes and is expressed during the haploid phase of the life cycle. In these species, the male and female developmental programs are initiated by the presence of the U- or V-specific regions of the sex chromosomes but, as in XY and ZW systems, sexual differentiation is largely driven by autosomal sex-biased gene expression. The mechanisms underlying the regulation of sex-biased expression of genes during sexual differentiation remain elusive. Here, we investigated the extent and nature of epigenomic changes associated with UV sexual differentiation in the brown alga Ectocarpus, a model UV system. Six histone modifications were quantified in near-isogenic lines, leading to the identification of 16 chromatin signatures across the genome. Chromatin signatures correlated with levels of gene expression and histone PTMs changes in males versus females occurred preferentially at genes involved in sex-specific pathways. Despite the absence of chromosome scale dosage compensation and the fact that UV sex chromosomes recombine across most of their length, the chromatin landscape of these chromosomes was remarkably different to that of autosomes. Hotspots of evolutionary young genes in the pseudoautosomal regions appear to drive the exceptional chromatin features of UV sex chromosomes.

Figure 1.

Histone PTMs and chromatin signatures of female and male Ectocarpus. (A) A model of prevalent chromatin emission states found in Ectocarpus using ChromHMM. Permissive 1 and Permissive 2, activation-associated states; Mixed, States that mix activation-associated and repression-associated chromatin PTMs; Silent, repression-associated chromatin states; Null, absence of assayed histone PTMs. (B) Representative region of the chromosome 19 showing profiles of mapped ChIP-seq reads for the six histone PTMs in females and males. Coverage is represented as the ratio of IP DNA relative to H3 for H3K36me3, H4K20me3 and H3K79me2 and input for TSS marks (H3K4me3, H3K9ac, H3K27ac). (C) Chromatin signatures assigned to genes based on ChromHMM states (see methods). Percentages of the total gene set associated with each chromatin signature in males (M) and females (F) are shown to the right. (D) Proportions of transcribed (TPM≥1TPM), silent (TPM<1TPM), housekeeping (tau<0.25) and narrowly expressed genes (tau>0.75) associated with each chromatin signature in males and females.

Figure 2.

Gene expression and chromatin states. (A) Transcript abundances for genes associated with different chromatin signatures in males and females. The colour code is the same as that used in Figure 1A,B. Transcript abundances for genes exhibiting either activation-associated (S1-S5) or repression-associated (S13 or S16) chromatin signatures in females (dark pink) and males (gray). (C) GO term enrichment for genes marked with activation-associated or repression-associated chromatin signatures in males and females. (D) Venn diagrams representing the proportion of genes marked with activation-associated or repression-associated chromatin signatures in males and females.

Figure 3.

Histone PTM patterns at sex-biased genes in Ectocarpus males and females. (A) Proportions of the 16 chromatin signatures for female-biased, male-biased and unbiased genes in females (left) and males (right). The number of genes in each category are inside brackets. Chromatin signature color codes correspond to Figure 1C. (B) Proportions of genes associated with each of the 16 chromatin signatures for female-biased (FBG) and male-biased (MBG) genes in females (left) and males (right). The intensity of the gray squares is proportional to the number of genes corresponding to each signature. Colored squares represent the different chromatin signatures (see Figure 1A). (C) Chord diagrams comparing chromatin signatures associated with female-biased (left) and male-biased (right) genes in females and males. The color code for the chromatin states is the same as that used in Figure 1C. Each chord represents a sex-biased gene and illustrates whether a gene changes from one signature to another in the opposing sex. (D) Representative chromatin profiles for a male-biased gene on chromosome 12 in females and males. (E) Representative chromatin profiles for a female-biased gene in chromosome 5 in females and males.

Figure 4.

Chromatin landscape of the U and V sex chromosomes compared with the autosomes. (A) Chromatin signature distribution for each autosome and for the SDR and PAR regions of the sex chromosome in females (left panel) and in males (right panel). (B) Proportions of genes associated with each of the 16 chromatin signatures for all autosomes and for the PAR and SDR regions of the sex chromosome in females and in males. The intensity of the grey is proportional to the number of genes in each signature. The color code for the chromatin states is the same as that used in Figure 1A. (C) Transcript abundances, measured as log2(TPM+1), for autosomal and for PAR genes in males and females. Significant differences were assessed using pairwise Wilcoxon rank sum test. (D) Transcript abundances for autosomal and PAR genes associated with different chromatin signatures. Permissive, activation-associated signatures (S1-S5); Silent, repression-associated signatures (S13-S16). Significant differences were assessed using pairwise Wilcoxon rank sum test. (E) Chromatin state distribution of evolutionary young genes compared with autosomal conserved genes with similar expression patterns. See also Supplementary Table S15. (F) Transcript abundances, measured as log2(TPM+1), for individual genes located in the female and male sex determining regions (SDRs). Colored plots represent chromatin signatures corresponding to the color code indicated in Figure 1C (see also Supplementary Table S11). (G) Transcript abundances of genes located within the female and male sex-specific regions (SDRs). Significant differences were assessed using pairwise Wilcoxon rank sum test.