Genetically encoded multivalent sensors to detect bivalent epigenetic modifications in living stem cells

Eukaryotic DNA is organized in the form of chromatin whose basic unit is the nucleosome. The four core histones forming the nucleosome, H2A, H2B, H3 and H4 can be highly post-translationally modified, especially on their N-terminal tail protruding from the nucleosome particle. Histone post-translational modifications (PTMs) work combinatorially to establish chromatin states defined by specific gene expression status. Found in embryonic stem cells (ESCs) at promoters of key developmental genes, bivalent chromatin is the combination of the active chromatin mark, trimethylation of lysine 4 on histone H3 (H3K4me3) and the repressive mark, H3K27me3. Established and maintained by Polycomb (Pc) and Trithorax (Trx) proteins, bivalency is proposed to keep gene transcription repressed but poised for activation. How bivalent domains are organized within the nucleus and how it is installed by Pc and Trx is still unkown. In this work, we aim to answer these questions by designing probes that allow live cell imaging of bivalent domains and by studying the establishment and removal of H3K4 methylation on nucleosomes. The current lack of live cell imaging methods for PTM patterns prompt us to engineer genetically encoded sensors which bind to bivalent marks in a multivalent fashion. These sensors contain a fluorescent protein and two reader domains, joined by flexible linkers. Their selectivity for bivalent nucleosomes was tested in a pulldown assay with a library of differently modified reconstituted nucleosomes. For this, we obtained site-specifically modified histones via expressed protein ligation. The best probe was then applied in live ESCs to visualize bivalent domains. Subsequent imaging by confocal microscopy revealed the organization of bivalent chromatin into discrete and local clusters. Furthermore, this probe was employed to monitor loss of bivalency upon treatment with a small molecule epigenetic inhibitor. Then, we studied the histone-lysine N-methyltransferase Set1B and lysine-specific histone demethylase 1A (LSD1), enzymes which deposit and remove methyl groups of H3K4 respectively. We measured a mid-micromolar affinity for LSD1 to reconstituted nucleosomes using microscale thermophoresis. This interaction might have an impact on the recruitment of LSD1 to its target genes. We then measured the activity of Set1B complex on symmetrically modified H3K27me3 nucleosomes and on asymmetrically modified H3K4me3 nucleosomes. We showed that H3K27me3 does not influence the activity of Set1B complex whereas H3K4me3 activates Set1B-mediated deposition of methyl groups at K4 on the opposite H3 tail. This findings might have important implications concerning the establishment of bivalent chromatin. Together these results gave insights about multivalent binding of tandem reader domains, subnuclear bivalent chromatin organization and establishment of bivalent marks. In the future, we envisage to develop multivalent sensors for other PTM patterns of biological interest.