The Spatial Organization of the X-chromosome and its Impact on X-inactivation

Abstract

Female mammals have two X-chromosomes, males have one X-chromosome. X-chromosome inactivation (XCI) is a process where most genes on one female X-chromosome is transcriptionally silenced in order to balance gene dosage on the X-chromosome between males and females. It has been known since the discovery of the Barr body in the 1940’s that the active (Xa) and inactive (Xi) X-chromosomes fold differently. However, the precise structures adopted by the two Xs, the mechanisms that cause the Xa and Xi to fold differently, and the role of the unique structure of the Xi in silencing were unknown. In this thesis, I describe our work defining the structures of the Xa and the Xi, the role of Xist RNA and cohesin proteins in shaping the organization of the Xi, and testing whether the “super-structure” of the Xi is necessary for gene silencing. I performed allele-specific Hi-C to map the structures of the Xa and Xi at high resolution and found that while the Xa is organized into megabase-sized topologically associated domains (TADs) like all the other chromosomes, the Xi is largely devoid of TADs. Allele-specific ChIP-seq showed that cohesin proteins, which are necessary for folding chromosomes into TADs, are depleted from the Xi. Deletion of Xist allows partial restoration of both cohesin binding and TADs across large regions of the Xi, suggesting that Xist RNA removes cohesin from chromatin and breaks down TADs on the Xi. Although the Xi lacks TADs, it folds into two very large “megadomains”, each tens of megabases in size. There is an unusual tandem repeat locus called Dxz4 at the border between the two megadomains. Dxz4 also forms a ~25 Mb looping interaction with another tandem repeat called Firre specifically on the Xi. I generated single and double deletions of Dxz4 and Firre in mouse embryonic stem cells. Deletion of Dxz4 prevented megadomain formation and deletion of Firre partially disrupted interactions within megadomains. However, deletion of Dxz4, Firre or both loci had no impact on the ability of XCI to initiate, suggesting that the large-scale organization of the Xi is not needed for gene silencing.