RAPID is uniquely suited for studying cell cycle-linked processes, by defining when and where stable ‘marks’ are set in chromatin. RAPID will define fundamental rules for inheritance of histone and other chromatin-associated proteins. In a first application, we have profiled the dynamics of histone variant H3.3 elucidating its intriguing role in dynamic heterochromatin maintenance in mouse embryonic stem cells. Nucleosome turnover concomitant with incorporation of the replication-independent histone variant H3.3 is a hallmark of regulatory regions in the animal genome. In our current understanding, nucleosome turnover is universally linked to DNA accessibility and histone acetylation. However, we have shown that H3.3 contributes to a unique heterochromatin state at endogenous retroviral elements in mouse embryonic stem cells (mESC). We recently further demonstrated that fast histone turnover and H3.3 incorporation is compatible with stereotypical hallmarks of heterochromatin. We find that Smarcad1 ATPase triggers nucleosome eviction and histone H3.3 is required to maintain minimal DNA accessibility in this surprisingly dynamic heterochromatin state. Loss of H3.3 in mouse embryonic stem cells elicits a highly specific opening of interstitial heterochromatin with minimal effects on other silent or active regions of the genome. Following up on these results, we used RAPID to perform time-lapse interactomics using mass spectrometry and quantitative ChIP to understand the mechanism of H3.3 deposition and turnover (Katsori et. al., manuscript in preparation). We were able to assign distinct specific remodeling activities to different functional outcomes, e.g. heterochromatin maintenance versus the opening of enhancers.
We also apply our technology to understanding the function of Polycomb group (PcG) proteins, evolutionary conserved master regulators of cell identity and differentiation that have long been a paradigm for epigenetic gene control. Two major PcG complexes, PRC1 and PRC2, cooperate in the repression of target genes catalyzing histone H2A K119 monoubiquitination (H2Aub) and H3K27 trimethylation (H3K27me3), respectively. Applying our quantitative MINUTE-ChIP in mESCs lead us to a number of interesting findings: 1) Catalytic activity of both Polycomb complexes is not confined to so-called Polycomb targets (e.g. bivalent domains) but establishes a considerable ‘background’ across most of the genome both in mouse and human pluripotent cells. We did not find a dependence of PRC1 activity on presence of PRC2 or H3K27me3 as was also described by others. Instead, loss of H3K27me3 lead to a global increase in H2Aub (unpublished data), potentially consistent with a previously suggested competitive/compensatory interplay of PRC1 and PRC2. This and much emerging evidence in the field suggests that our understanding of PRC1 and PRC2 function needs to be fundamentally revisited. For example, PRC1 has recently been suggested to mediate promoter-enhancer contacts. In one implementation of RAPID, we developed a synthetic strategy to quickly deplete H2Aub, without disturbing PRC1 and PRC2 complexes itself. Surprisingly, we found that neither PRC1 nor PRC2 dissociate from the H2Aub-depleted chromatin immediately. A high degree of multivalent binding, or, as already documented in principle for PRC1, a phase separation phenomenon would explain such ‘resilient’ binding of PRC1.
