The CCCTC-binding factor (CTCF) works together with the cohesin complex to drives the formation of chromatin loops and topologically associating domains, but its role in gene regulation has not been fully defined. Here, we investigated the effects of acute CTCF loss on chromatin architecture and transcriptional programs in mouse embryonic stem cells undergoing differentiation to neural precursor cells. We identified CTCF-dependent enhancer-promoter contacts genome-wide and found that they disproportionally affect genes that are bound by CTCF at promoter and dependent on long-distance enhancers. Disruption of promoter-proximal CTCF binding reduced both long-range enhancer-promoter contacts and transcription, which are restored by artificial tethering of CTCF to the promoter. Promoter-proximal CTCF binding is correlated with transcription of over 2,000 genes, across a diverse set of adult tissues. Taken together, our study shows that CTCF binding to promoters may promote long-distance-enhancer dependent transcription at specific genes in diverse cell types.

The spatial organization of chromosomes has key functional roles, yet how chromosomes fold remains poorly understood at the single-molecule level. Here, we employ models of polymer physics to investigate DNA loci in human HCT116 and IMR90 wild-type and cohesin depleted cells. Model predictions on single-molecule structures are validated against singlecell imaging data, providing evidence that chromosomal architecture is controlled by a thermodynamics mechanism of polymer phase separation whereby chromatin self-assembles in segregated globules by combinatorial interactions of chromatin factors that include CTCF and cohesin. The thermodynamics degeneracy of single-molecule conformations results in broad structural and temporal variability of TAD-like contact patterns. Globules establish stable environments where specific contacts are highly favored over stochastic encounters. Cohesin depletion reverses phase separation into randomly folded states, erasing average interaction patterns. Overall, globule phase separation appears to be a robust yet reversible mechanism of chromatin organization where stochasticity and specificity coexist.

No abstract

Loop-extrusion and phase-separation have been proposed as mechanisms that shape chromosome spatial organization. It is unclear, however, how they perform relative to each other in explaining chromatin architecture data and whether they compete or co-exist at the single-molecule level. Here, we compare models of polymer physics based on loop-extrusion and phase-separation, as well as models where both mechanisms act simultaneously in a single molecule, against multiplexed FISH data available in human loci in IMR90 and HCT116 cells. We find that the different models recapitulate bulk Hi-C and average multiplexed microscopy data. Single-molecule chromatin conformations are also well captured, especially by phase-separation based models that better reflect the experimentally reported segregation in globules of the considered genomic loci and their cell-to-cell structural variability. Such a variability is consistent with two main concurrent causes: single-cell epigenetic heterogeneity and an intrinsic thermodynamic conformational degeneracy of folding. Overall, the model combining loop-extrusion and polymer phase-separation provides a very good description of the data, particularly higher-order contacts, showing that the two mechanisms can co-exist in shaping chromatin architecture in single cells.

Graphical AbstractHighlights d The structure of the mouse a-globin locus is investigated with polymer physics models d The conformation is highly intermingled when the globin genes are inactive in ESCs d The locus is folded in a hairpin-shaped conformation when active in erythroid cells d The model accurately predicts three-way contacts, as confirmed by TriC experimental data Correspondence chiariello@na.infn.it (A.M.C.), mario.nicodemi@na.infn.it (M.N.) In BriefChiariello et al. use polymer physics models to infer the 3D conformations of the murine a-globin locus. In the transition from ESCs to erythroid cells, the locus rearranges from an inactive highly intermingled conformation to an active, hairpin-shaped structure, marked by cell-specific three-way contacts accurately predicted by the models. SUMMARYWe investigate the three-dimensional (3D) conformations of the a-globin locus at the single-allele level in murine embryonic stem cells (ESCs) and erythroid cells, combining polymer physics models and high-resolution Capture-C data. Model predictions are validated against independent fluorescence in situ hybridization (FISH) data measuring pairwise distances, and Tri-C data identifying three-way contacts. The architecture is rearranged during the transition from ESCs to erythroid cells, associated with the activation of the globin genes. We find that in ESCs, the spatial organization conforms to a highly intermingled 3D structure involving non-specific contacts, whereas in erythroid cells the a-globin genes and their enhancers form a self-contained domain, arranged in a folded hairpin conformation, separated from intermingling flanking regions by a thermodynamic mechanism of micro-phase separation. The flanking regions are rich in convergent CTCF sites, which only marginally participate in the erythroid-specific geneenhancer contacts, suggesting that beyond the interaction of CTCF sites, multiple molecular mechanisms cooperate to form an interacting domain.