Role of 3D-organization of the X-inactivation centre in imprinted X-chromosome inactivation

Abstract

The connection between 3D-genome organization and genome regulation is one of the fundamental questions in modern biology. In a nucleus, the genome is organized through different layers of 3D-organization such as compartmentalization, formation of topologically associated domains (TADs), chromatin looping etc. TADs are one of the central structural units of 3D-genome organization. However, the role of TADs in genome regulation remains controversial. In this study, through studying TADs of the mouse X-inactivation center (XIC), we have provided significant insight into the role of TADs in genome regulation. X-chromosome inactivation serves as a paradigm of 3D-genome organization and gene regulation. In mouse, there are two forms of X-chromosome inactivation: in early embryos, paternal X-gets inactivated, which is known as imprinted X-inactivation and later, in post- implantation epiblast, it switches to random X-inactivation, where either paternal or maternal X gets inactivated. Upon initiation of random X-inactivation in differentiating mouse embryonic stem cells (ESCs), TADs are largely lost from the inactive-X, and the inactive-X gets bipartitely reorganized into two large megadomains. Interestingly, the XIC harbors two TADs – at the locus of long non-coding RNA Xist (Xist-TAD) and Tsix (Tsix-TAD). Xist is the master regulator of X-inactivation, which coats the inactive-X chromosome and facilitates heterochromatinization. The role of Xist and Tsix TAD in the orchestration of X-chromosome inactivation remains poorly understood. Here, we show that mouse female extra-embryonic endoderm stem cells (XEN), which undergo imprinted X-inactivation, also have Xist-TAD at their XIC, like differentiating ESCs with random X-inactivation. To explore the role of Xist-TAD in the maintenance of imprinted X- inactivation, we deleted Xist upstream sequences (~6 kb) near the Xist TAD boundary in XEN cells. This Xist upstream deletion at the inactive-X chromosome in XEN cells led to the major rearrangement of Xist intra-TAD contacts and impairment of interactions of the Xist loci across X-chromosome and autosomes. Notably, impaired topological interactions were accompanied by loss or gain of binding of architectural proteins CTCF/Rad21 at many loci of the inactive- X, including Xist-TAD. Furthermore, CTCF/Rad21 binding was significantly altered at Firre and x75 loci of the inactive-X, which are critical for the maintenance of inactive-X 11 conformation. Together, our analysis suggested that Xist upstream sequences near the Xist-TAD boundary are crucial for maintaining proper inactive-X topology in XEN cells. Moreover, we found that there was upregulation of Xist expression, dispersal of Xist coating and loss of enrichment of repressive marks at the inactive-X upon deletion of Xist upstream sequences. Surprisingly, there was no effect on X-linked gene silencing at the inactive-X, however, autosomal genes were dysregulated in these cells. Collectively, we conclude that Xist upstream sequences are necessary for the maintenance of proper topological contacts of Xist locus, Xist coating/expression and autosomal gene expression. Altogether, our study provides significant insight into the role of 3D-genome organization in genome regulation.