PDF, English Restricted access: Repository staff only until 17 October 2024. Login+Download (9MB) | Terms of use

Abstract

Histone post-translational modifications (PTMs) are generally associated with defined transcriptional states. For instance, histone H3 lysine acetylation (e.g., H3K27ac, H3K9ac) conventionally marks cell-type specific active regulatory elements, whereas H3 lysine 9 tri-methylation (H3K9me3) and lysine 27 tri-methylation (H3K27me3) are enriched at repressed repeat elements and promoters, respectively. Despite strong correlations, it is still unclear whether histone modifications play an instructive role in transcriptional regulation. Histone residue mutagenesis represents an elegant approach to address this question. Nevertheless, applying histone mutagenesis is challenging in the vast majority of metazoans because canonical histone genes are found in large clusters on different chromosomes. The histone variant H3.3 is instead encoded by two genes outside the clusters, which makes mutagenesis more feasible in mammalian systems. Here, I applied H3.3 mutagenesis to study the relevance of H3.3 K9 and K27 residues in gene expression regulation, with the primary goal of characterizing the contribution of these histone marks in repressive chromatin environments. Using CRISPR-Cas9, I generated mouse embryonic stem cells (mESCs) bearing lysine-to-alanine substitutions and applied multi-omics to profile changes in transcriptome and epigenome. H3.3 K9A and K27A mutants resulted in severe gene expression changes whereby a large proportion of deregulated genes overlapped in both mutants. Developmentally-regulated genes harbouring bivalent (i.e., H3K4me3-H3K27me3) promoters were de-repressed in both mutants, and the in-vitro differentiation potential of the mESCs was disrupted. The H3.3K27A mutation affected local deposition and spreading of the H3K27me3 mark, which particularly impacts a subset of bivalent genes with a higher basal level of H3.3 on their promoters. On the other hand, the H3.3K9A mutation resulted in the global loss of H3K9me3 at endogenous retroviruses (ERVs) and distal intergenic regions. These regions gained histone H3 acetylation and exhibited increased levels of nascent transcription. Constructing a gene regulatory network (GRN) revealed that the de-repression of these territories leads to a profound rewiring of transcriptional regulatory networks in mESCs. This rewiring resulted in the upregulation of bivalent genes and genes related to immune system processes in H3.3K9A and also culminated in the expression of proteins generally restricted to highly specialized immune cell types. At canonical enhancers, H3.3K27A displayed decreased H3K27ac and induced a collateral loss of acetylation at other H3 lysines. Nascent transcription levels were reduced, and the genes connected to these regions through the GRN were downregulated, indicating that comprehensive—rather than residue-specific—H3 acetylation might play a crucial role in maintaining enhancers in an active state. In H3.3K9A mESCs, H3K27ac was primarily affected at enhancers, and the nascent transcription was reduced at regulatory elements particularly active in mESCs. I propose that the reduced activity of canonical enhancers in H3.3K9A mESCs is a consequence of the alien activation of ERV-derived regulatory elements, which potentially compete with canonical enhancers resulting in rerouting of transcription factors to binding sites that are not available in the basal state. Collectively, this study demonstrates that H3.3 K9 and K27 residues are critical for maintaining repressive chromatin states at promoters and distal genomic regions and are, therefore, essential for transcriptional homeostasis in mESCs.