Impact of 3-dimensional genome organization, guided by cohesin and CTCF looping, on sex-biased chromatin interactions and gene expression in mouse liver

Abstract

More than 1,000 genes are differentially expressed between male and female mouse liver. However, ~90% of 4,000 sex-differential open chromatin regions harboring regulatory elements implicated in liver sex bias are distal enhancers. To elucidate the role of 3D nuclear organization and address the challenge of linking distal regulatory regions to sex-biased gene targets, we performed ChIP-seq for CTCF and cohesin, known mediators of 3D nuclear organization, and identified ~1,000 binding sites showing significant differential occupancy by these factors in male and female mouse liver. Distal genomic regions showing sex-differential CTCF and/or cohesin binding were evaluated as potential factors in sex-dependent looping controlling sex-biased proximal effects on gene expression. Results were integrated with a study of cohesin depletion in mouse liver to evaluate the impact of cohesin loss on the expression of putative gene targets. Further, circularized chromosome conformation capture with sequencing (4C-seq) was used to discover striking sex-differences in 3D looping, linking distal sex-biased enhancers to proximal sex-biased gene expression for both male-biased (C9, Nudt7, Nox4) and female-biased genes (A1bg, Sult3a2). Our findings also illustrate the importance of a nested intra-TAD loop for insulation of sex-specific enhancer-promoter contacts for Nox4. These studies reveal how chromatin interactions and 3-dimensional genome organization guided both directly and indirectly by CTCF and cohesin looping contribute to liver sex bias. Matthews and Waxman -page 2Here we took a multi-pronged approach to characterize the role of architectural proteins and 3D genome organization in the maintenance of mouse liver sex differences. First, we performed ChIP-seq for CTCF and cohesin using chromatin from male and female mouse livers and show that the majority of sex-biased binding of these factors occurs at distal intergenic sites. Only a minority of these binding sites are cis to sex-biased genes, which may cis-proximal (within 20 kb) or more cis-distal (> 20 kb, but within the same TAD) to a sex biased gene. Specific examples of cis proximal and distal CTCF/cohesin binding are shown for male-biased genes (proximal: Cml5, Nat8; distal, C8a, C8b) and female-biased genes (proximal: Fam84b; distal: Slc22a29). Using a publicly available cohesin depletion dataset for male liver [21], we find that distally-regulated malebiased genes are considerably more sensitive to cohesin loss than those with proximal sex bias. To directly measure differences in chromatin interactions, we use 4C-seq for six viewpoints at or nearby five genes showing liver sex bias in expression, ranging from 2-fold to 700-fold (female-biased: A1bg and Sult3a1; malebiased: C9, Nudt7, and Nox4). Finally, we show the importance of a nested intra-TAD loop for insulation of enhancer-promoter contacts for Nox4. Overall, we describe how 3-dimensional genome organization contributes to liver sex bias in both direct and indirect ways. RESULTS Global differences in CTCF/Coh...