Polycomb repression of Hox genes involves spatial feedback but not domain compaction or demixing

Abstract

Polycomb group (PcG) proteins modulate higher-order genome folding and play a critical role in silencing transcription during development. It is commonly proposed that PcG dependent changes in genome folding, which compact chromatin, contribute directly to repression by blocking binding of activating complexes and demixing repressed targets from non-repressed chromatin. To test this model we utilized Optical Reconstruction of Chromatin Architecture (ORCA) to trace the 3-dimensional folding of the Hoxa gene cluster, a canonical Polycomb target, allowing us to analyze thousands of DNA traces in single cells. In cell types ranging from embryonic stem cells to brain tissue, we find that PcG-bound chromatin frequently explores decompact states and partial mixing with neighboring chromatin, while remaining uniformly repressed, challenging the repression-by-compaction model. Using polymer physics simulations, we show that the flexible ensembles we observe can be explained by dynamic contacts mediated by multivalent interactions that are too weak to induce phase separation. Instead, these transient contacts contribute to accurate propagation of the epigenetic state without ectopic spreading or gradual erosion. We propose that the distinctive 3D organization of Polycomb chromatin, reflects a mechanism of 'spatial feedback' required for stable repression.

Single-cell chromatin state transitions during epigenetic memory formation

“…This heterogeneity suggests that even with H3K9me3 enrichment across more than 10 kb, chromatin does not undergo an all-or-none transition from fully open to fully compact, but instead it is dynamic in both the active and silenced states. Similar single-cell results were recently reported for Polycomb repression in mouse ES cells 43 . It is possible that the change in average compaction is a result of different transition rates between open and compact conformations in active versus silent loci.…”

“…Epigenetic modifications spread across the genome 8,21,38 by either chromatin looping, reader-writer feedback, or diffusion of locally concentrated histone modifiers 8,39,40 . Increasing compaction, and hence chromatin contact frequency, would therefore help propagation or maintenance of epigenetic modifications [41][42][43] .…”

Design principles of 3D epigenetic memory systems

“…Semimarked domains also fold into irregular structures (fig. S6B)—as recently observed by super-resolution microscopy targeting Polycomb-repressed Hox genes ( 49 )—instead of spheres, as would fully marked domains. Additionally, semimarking explains the counterintuitive findings of Alabert et al .…”

4D epigenomics: deciphering the coupling between genome folding and epigenomic regulation with biophysical modeling

“…All this will allow a finer description of the crosstalk between the structural and functional epigenomic regulatory pathways in normal and pathological conditions. Specific engineered/synthetic systems of CRs [1,97] may provide a template to study the aforementioned components of epigenome regulation, while development of single cell experimental techniques might facilitate deciphering the CR recruitment mode (thereby the mechanism of spreading) in vivo [93]. Importantly, there is also a need to improve our current description of the epigenetic inheritance during and after replication (distribution of the maternal marks among the sister chromatids, dynamics of reestablishment during G2, etc.)…”

“…Such insulation, coupled to enough genomic bookmarking (sequence-dependent recruitment of HMEs) [86,92] may significantly enhance stability over many cell generations [83]. Limitation of the number of HMEs coupled to 3D spreading and 3D compaction/segregation, as observed for the Polycomb regulation during fly embryogenesis [93], is also compatible with stable confined memory [85,87]. In this case, an optimal HME concentration may be a prerequisite for robust maintenance of an epigenomic landscape: at too low concentrations (≪ 1𝜇𝑀), the spreading strength is not strong enough to overcome the perturbations occurring at replication while, at too high concentrations (≫ 1𝜇𝑀), it may lead to the formation of spurious epigenomic domains and to unconfined memory [85].…”