Investigating the mechanisms and elements responsible for regulatory domain formation during gene regulation

The genome is a highly organised and complex structure; it is separated into distinct chromatin domains whose structures influence transcriptional regulation. Regulatory domains contain the promoters of active genes and their cis-regulatory elements, such as enhancers, and are flanked by boundary elements which bind CTCF. Processes that initiate and maintain this level of genome organisation are not well understood. In this thesis, I have investigated the role of promoters, enhancers, and CTCF-bound elements in generating and defining regulatory domains. To do this, I have exploited the well-studied regulatory domain and elements of the mouse α-globin locus. In mouse erythroblasts, the five active enhancer elements and two active promoters of the α-globin genes interact and form a tissue-specific regulatory domain. These active elements have been well-characterised and are clearly involved in formation of the domain. In addition to the tissue-specific elements, there are numerous CTCF binding sites throughout the α-globin locus which may contribute to the structure of the regulatory domain. I have characterised a series of mouse models that harbour mutations of CTCF binding sites within the regulatory domain (θ1 & θ2) and downstream of the active promoters (HS+44 & HS+48) to understand their role in gene regulation and domain formation. Deletion of the downstream CTCF sites (∆44-48) provides support for the loop extrusion model of domain formation. Deletion of CTCF sites within the active domain (∆θ1θ2 & ∆θ2) suggests that promoters may act as domain boundaries, in addition to CTCF sites. Genetic manipulations of the α-globin locus have helped us understand which elements are necessary for domain formation, but as the mouse α-globin locus is highly conserved and contains numerous elements that could be contributing to domain formation, it is hard to determine which elements are sufficient for domain formation. Therefore, I have identified “neutral” loci on chromosome X, devoid of genes and functional elements, and by genome editing, have generated artificial loci containing the mouse α-globin locus elements in various permutations in male mouse embryonic stem (ES) cells. Introduction of the strongest α-globin enhancer, R2, into one such locus on chromosome X causes local changes in chromatin accessibility in erythroid cells. These changes in accessibility are accompanied by changes in chromatin modifications in which the accessible regions are marked with the active histone modifications H3K27ac and H3K4me1. Additionally, the R2 insertion gains another active histone modification H3K4me3 and produces long, unidirectional non-polyadenylated (~75 kb) and polyadenylated (~4 kb) transcripts. Finally, I found that the activity of the inserted R2 element is sufficient to form a regulatory domain structure similar to that observed at the α-globin locus. These results indicate that strong enhancer elements may be the key drivers in the formation of regulatory domains and provide insights into enhancer evolution and the mechanisms by which they regulate transcription of their target promoters.