Mammalian genomes are spatially organized into compartments, topologically associating domains (TADs), and loops to facilitate gene regulation and other chromosomal functions. How compartments, TADs, and loops are generated is unknown. It has been proposed that cohesin forms TADs and loops by extruding chromatin loops until it encounters CTCF, but direct evidence for this hypothesis is missing. Here, we show that cohesin suppresses compartments but is required for TADs and loops, that CTCF defines their boundaries, and that the cohesin unloading factor WAPL and its PDS5 binding partners control the length of loops. In the absence of WAPL and PDS5 proteins, cohesin forms extended loops, presumably by passing CTCF sites, accumulates in axial chromosomal positions (vermicelli), and condenses chromosomes. Unexpectedly, PDS5 proteins are also required for boundary function. These results show that cohesin has an essential genome-wide function in mediating long-range chromatin interactions and support the hypothesis that cohesin creates these by loop extrusion, until it is delayed by CTCF in a manner dependent on PDS5 proteins, or until it is released from DNA by WAPL.

Nuclear processes like V(D)J recombination are determined by the three-dimensional organization of chromosomes in multiple layers, including the compartments 1 and topologically associated domains (TADs) 2,3 consisting of chromatin loops. TADs are formed by chromatin loop extrusion 5-7 , which depends on the ring-shaped cohesin complex 8-10 with its loop extrusion function 11,12. The cohesin-release factor Wapl 13,14 instead restricts loop extension 10,15. The generation of a diverse antibody repertoire, providing humoral immunity to pathogens, requires the participation of all V genes in V(D)J recombination 16 , which depends on contraction of the 2.8-Mb-long immunoglobulin heavy-chain (Igh) locus by Pax5 17,18. How Pax5 controls Igh contraction in pro-B-cells is, however, unknown. Here, we demonstrate that locus contraction is caused by loop extrusion across the entire Igh locus. Notably, the expression of Wapl is repressed by Pax5 specifically in pro-B and pre-B-cells, which facilitates extended loop extrusion by increasing the residence time of cohesin on chromatin. Pax5 mediates the transcriptional repression of Wapl through a single Pax5-binding site by recruiting the Polycomb repressive complex 2 to induce bivalent chromatin at the Wapl promoter. Reduced Wapl expression causes global alterations in the chromosome architecture, indicating that the potential to recombine all V genes entails structural changes of the entire genome in pro-B-cells.

The cell cycle transition from interphase into mitosis is best characterized by the appearance of condensed chromosomes that become microscopically visible as thread-like structures in nuclei. Biochemically, launching the mitotic program requires the activation of the mitotic cyclin-dependent kinase Cdk1 (cyclin-dependent kinase 1), but whether and how Cdk1 triggers chromosome assembly at mitotic entry are not well understood. Here we report that mitotic chromosome assembly in prophase depends on Cdk1-mediated phosphorylation of the condensin II complex. We identified Thr 1415 of the CAP-D3 subunit as a Cdk1 phosphorylation site, which proved crucial as it was required for the Polo kinase Plk1 (Polo-like kinase 1) to localize to chromosome axes through binding to CAP-D3 and thereby hyperphosphorylate the condensin II complex. Live-cell imaging analysis of cells carrying nonphosphorylatable CAP-D3 mutants in place of endogenous protein suggested that phosphorylation of Thr 1415 is required for timely chromosome condensation during prophase, and that the Plk1-mediated phosphorylation of condensin II facilitates its ability to assemble chromosomes properly. These observations provide an explanation for how Cdk1 induces chromosome assembly in cells entering mitosis, and underscore the significance of the cooperative action of Plk1 with Cdk1.

Eukaryotic genomes are folded into loops. It is thought that these are formed by cohesin complexes via extrusion, either until loop expansion is arrested by CTCF or until cohesin is removed from DNA by WAPL. Although WAPL limits cohesin’s chromatin residence time to minutes, it has been reported that some loops exist for hours. How these loops can persist is unknown. We show that during G1-phase, mammalian cells contain acetylated cohesinSTAG1 which binds chromatin for hours, whereas cohesinSTAG2 binds chromatin for minutes. Our results indicate that CTCF and the acetyltransferase ESCO1 protect a subset of cohesinSTAG1 complexes from WAPL, thereby enable formation of long and presumably long-lived loops, and that ESCO1, like CTCF, contributes to boundary formation in chromatin looping. Our data are consistent with a model of nested loop extrusion, in which acetylated cohesinSTAG1 forms stable loops between CTCF sites, demarcating the boundaries of more transient cohesinSTAG2 extrusion activity.