Investigating catalysis-independent roles of PRC2

Abstract

The catalytic subunit EZH2 of Polycomb Repressive complex 2 (PRC2) is responsible for placing the histone H3 lysine 27 trimethylation (H3K27me3) histone modification at nucleosomes, a mark associated with silenced genes. This histone modification can be found at domains that also carry marks usually associated with a transcriptionally active state, such as H3K4me3 and H3K4me1.

Genes that bear this overlap of repressive and active marks around their promoters and/or enhancers are known as bivalent or poised genes, thought to be in a state of readiness for transcriptional silencing or activation upon stimulation by additional factors during differentiation. However, it is not known how much of Polycomb repression relies on H3K27me3 alone and whether the binding of PRC2 itself plays a part that is independent of this histone modification. It has further been suggested that PRC2 may interact with chromatin organisers and be involved in long-range chromatin interactions. Recently, it has also been noted that poised enhancers are already looped to contact their respective promoters, in a PRC2 dependent manner, before gene activation occurs. Again, it is not yet known whether these contacts depend on H3K27me3 or whether they are mediated by the protein complex itself. To address these questions, I used CRISPR/Cas9 gene editing to generate both a knock out and catalytically inactive form of EZH2 in mouse embryonic stem cells, both in the presence and absence of its less active paralogue, EZH1. By performing gene expression, chromatin conformation, and differentiation assays, I aimed to identify roles of PRC2 that may be independent of the histone mark that it catalyses and that may mediate the role of PRC2 in genome organisation, as well as seeking to further elucidate the role of PRC2 throughout development. The findings of this project showed that some catalytic activity of PRC2 was required for the maintenance of Polycomb gene silencing in mESCs. Interestingly, the levels of H3K27me3 that required to maintain repression were much lower than those that are found at these genes in wild-type conditions. I found that not only was the activity of EZH1 sufficient to maintain the silencing of Polycomb bound genes in mESCs, but also for the majority of RING1b recruitment to these sites, and the maintenance of the PRC2-dependent enhancer-promoter contact at the Lhx5 locus. Furthermore, I found that throughout differentiation EZH1 was able to maintain the repression of genes that were silenced in undifferentiated mESCs, but EZH2 and robust H3K27me3 methylation was required to allow for the correct upregulation of certain developmental genes required for transition to the neural lineage. This failure to upregulate such a large number of genes was somewhat unexpected as PRC2 is known as a repressive complex and this may suggest it also possesses some gene activating functions. Based on these findings I concluded that while the low catalytic activity of EZH1 is sufficient to maintain a Polycomb repressive state at target genes in a steady state environment such as undifferentiated mESCs, the activity of EZH2 is required in the dynamic conditions of cell differentiation to establish robust levels of H3K27me3 deposition and Polycomb binding at novel target sites.