Gonadal supporting cells acquire sex-specific chromatin landscapes during mammalian sex determination

Abstract

Cis-regulatory elements are critical for the precise spatiotemporal regulation of genes during development. However, identifying functional regulatory sites that drive cell differentiation in vivo has been complicated by the high numbers of cells required for whole-genome epigenetic assays. Here, we identified putative regulatory elements during sex determination by performing ATAC-seq and ChIP-seq for H3K27ac in purified XX and XY gonadal supporting cells before and after sex determination in mice. We show that XX and XY supporting cells initiate sex determination with similar chromatin landscapes and acquire sex-specific regulatory elements as they commit to the male or female fate. To validate our approach, we identified a functional gonad-specific enhancer downstream of Bmp2, an ovary-promoting gene. This work increases our understanding of the complex regulatory network underlying mammalian sex determination and provides a powerful resource for identifying non-coding regulatory elements that could harbor mutations that lead to Disorders of Sexual Development.

Keywords: Cell differentiation; Chromatin landscapes; Enhancers; Gonad development; Regulatory elements; Sex determination.

Figure 1.. Chromatin architecture is remodeled during mammalian sex determination.

A) Overview of sex determination and work flow. Briefly, supporting progenitor cells (yellow) are bipotential and indistinguishable between XX and XY gonads at E10.5. Expression of the Y-linked Sry gene (at E11.5) directs Sertoli cell differentiation in the testis (XY, blue). Absence of Sry directs differentiation of granulosa cells in the ovary (XX, pink). XX and XY progenitor cells (E10.5), Sertoli cells (E13.5), and granulosa cells (E13.5) were FACS-purified and used for ATAC-seq and ChIP-seq for H3K27ac. Further analysis made use of microarray expression data from purified supporting cells (Jameson et al., 2012b). B) Percent (and number) of H3K27ac-negative (grey) and H3K27ac-postive (green) NDRs in XX and XY cells at E10.5 (left) and E13.5 (right). C) Venn diagrams of all NDRs in XX (pink) and XY (blue) supporting cells at E10.5 (left) and E13.5 (right). D) Percent of NDRs that are shared between XX and XY cells at E10.5 (purple) and at E13.5 (orange), or specific to either XX or XY cells at E10.5 (black) or at E13.5 (grey). The purple and black bars at E13.5 represent NDRs that were retained from E10.5, while the orange and grey represent newly acquired NDRs.

Figure 2.. H3K27ac+ gonad-specific NDRs are distal enhancers that neighbor granulosa-and Sertoli-promoting genes.

A-C) Graphs showing the average number of NDRs associated to the nearest TSS as determined by HOMER. Based on their cell-type specificity, NDRs were classified as ubiquitous (A), gonad-specific (B) or H3K27ac+ gonad-specific (C). NDRs were associated to the nearest TSS of genes that become granulosa cell-specific (pink, 823 total genes) and Sertoli cell-specific (blue, 1026 total genes). Granulosa-specific and Sertoli-specific genes were obtained from (Jameson et al., 2012b). Error bars represent the standard error of the mean (SEM) and ** represents p<0.001 as determined by a student’s t test. D) Ubiquitous, gonad-specific and gonad-specific H3K27ac-positive NDRs were subcategorized based on their genomic location (promoter, grey; exonic, orange; intronic, green; intergenic, purple). E) Ubiquitous, gonad-specific and gonad-specific H3K27ac-positive NDRs were subcategorized based on their 5’ (left) and 3’ (right) distance to the associated TSS (TSS, black middle line; 0–1kb, light grey; 1–3kb, purple; 3–5kb, blue; 5–10kb, green; 10–100kb, medium grey; >100kb, dark grey).

Figure 3.. Gonad-specific H3K27ac-negative NDRs are potential silencers of the alternate fate.

Analysis of gonad-specific (GS) H3K27ac-negative NDRs in granulosa cells (A&B), and GS H3K27ac-negative NDRs in Sertoli cells (C&D). A&C) Graphs showing the average number of NDRs associated to the nearest TSS as determined by HOMER. NDRs were associated to the nearest TSS of genes that become granulosa cell-specific (pink) and Sertoli cell-specific (blue). Error bars represent the standard error of the mean (SEM) and *** represents p<0.0001 as determined by a student’s t test. B&D) Top five motifs identified by TF binding motif analysis of protein families with members expressed in the gonad ranked in order of significance (pWnt4 that is H3K27ac-positive in granulosa cells and H3K27ac-negative in Sertoli cells (G) and an NDR near the granulosa-promoting gene Bmp2 that is H3K27ac-negative in Sertoli cells and is not shared in granulosa cells (H) are shown.

Figure 4.. Gonad-specific NDRs are either retained from the progenitor state or arise de novo.

A) Genome Browser tracks of ATAC-seq in XX E10.5 (light pink), XY E10.5 (light blue), granulosa (dark pink) and Sertoli (dark blue) purified cells of the genomic location surrounding Axin2. Bold black lines above tracks represent significant enrichment compared to flanking regions as determined by HOMER. A resolved granulosa (RG) and a de novo granulosa (DNG) NDR (boxed) lie downstream of Axin2. B) Axin2 gene expression profile in purified XX (pink) and XY (blue) supporting cells throughout sex determination (E11.5-E13.5) (Jameson et al., 2012b). C) Number of resolved and de novo NDRs in granulosa cells. D and E) Top motifs of families with members expressed in the gonad identified by TF binding motif analysis in resolved and de novo NDRs in granulosa cells ranked in order of significance (additional motifs in Table S1). F) Genome Browser tracks of ATAC-seq in XX E10.5 (light pink), XY E10.5 (light blue), granulosa (dark pink) and Sertoli (dark blue) purified cells of the genomic location surrounding Sox8. Bold black lines above tracks represent significant enrichment compared to flanking regions as determined by HOMER. A resolved Sertoli (RS) and a de novo Sertoli (DNS) NDR (boxed) lie downstream of Sox8. G) Sox8 gene expression profile in purified XX (pink) and XY (blue) supporting cells throughout sex determination (E11.5-E13.5) (Jameson et al., 2012b). H) Number of resolved and de novo NDRs in Sertoli cells. I and J) Top motif families (with members expressed in the gonad) identified by TF binding motif analysis in resolved and de novo NDRs in Sertoli cells ranked in order of significance (additional motifs in Table S1).

Figure 5.. Transient transgenic analysis of a granulosa-specific NDR downstream of Bmp2.

A) Genome Browser tracks of ATAC-seq in XX E10.5 (light pink), XY E10.5 (light blue), granulosa (dark pink) and Sertoli (dark blue) purified cells and corresponding H3K27ac ChIP-seq profiles (green) at the genomic location surrounding Bmp2. Bold black lines above tracks represent significant enrichment compared to flanking regions as determined by HOMER. Grey boxes represent ubiquitous NDRs (top). Sequence conservation between rat, human, cow, chicken, lizard, and zebrafish are shown below. A de novo gonad-specific H3K27ac+ granulosa NDR downstream of Bmp2 (boxed) was selected for transient transgenic analysis and cloned into an hsp68-LacZ reporter cassette. B) Gene expression profile in purified XX (pink) and XY (blue) supporting cells throughout sex determination (E10.5-E13.5) showing that Bmp2 becomes granulosa-specific (Nef et al., 2005). C) An E13.5 XX gonad from a transgenic embryo shows β-galactosidase expression (TgBmp2_enh, dark blue) in the gonad (g), but not the mesonephros (m). TgBmp2_enh expression is stronger in the anterior pole of the gonad (left).