The impact of toggling 3D genome structure on Polycomb-mediated repression, interchromosomal interactions, and X-inactivation

Abstract

Microscopy and chromosome conformation capture (3C) techniques revealed that mammalian genomes fold into chromosome territories, compartmental structures such as active/inactive (A/B) compartments, and dynamic chromatin loops, mainly formed by the ring-shaped complex cohesin and stabilized by the zinc-finger protein CTCF. Interestingly, previous CTCF and cohesin loss-of-function studies disrupted 3D genome structure but had relatively small impact on gene expression. In this thesis, I revisit this structure-function relationship by first examining the role of a post-translational modification of CTCF, Serine 224 phosphorylation, on gene expression and domain formation. I then examine the role of 3D structure in X chromosome inactivation (XCI), the process which balances dosage of X-linked genes between male and female mammals. XCI is mediated by Xist RNA, which spreads across the inactive X chromosome (Xi), triggering recruitment of repressive factors, such as the Polycomb Repressive Complexes. Simultaneously, the Xi folds into a unique 3D structure not seen elsewhere in the genome. Therefore, XCI provided an ideal system in which to examine the role of 3D structure in epigenetic spreading and gene regulation. I generated degron lines in which architectural factors CTCF, RAD21, an essential component of the cohesin complex, or WAPL, a cohesin release factor, could be rapidly and efficiently degraded. Surprisingly, CTCF degradation had little impact on 3D genome structure or XCI. Toggling cohesin, in contrast, had striking results. Genome wide, a proper balance of cohesin was needed to maintain appropriate interchromosomal interactions as well as Polycomb-mediated repression. On the Xi, cohesin degradation ablated megadomain superstructures. However, the Dxz4-Firre superloop persisted upon cohesin degradation. In contrast, WAPL degradation did not affect the megadomain border but weakened long-range intra-megadomain interactions. Intriguingly, cohesin imbalance also created Xi-like superstructures on the active X. Finally, I examined how these structural changes impacted XCI. WAPL degradation led to both decreased Xist and Polycomb spreading and disrupted gene silencing, especially to the ends of the chromosome where the most contacts with the Xist transcription locus were lost. Thus, a balance of cohesin is needed for proper Polycomb-mediated repression, interchromosomal interactions, and XCI.