Sister chromatid cohesion is mediated by cohesin and is essential for accurate chromosome segregation. The cohesin subunits SMC1, SMC3, and Rad21 form a tripartite ring within which sister chromatids are thought to be entrapped. This event requires the acetylation of SMC3 and the association of sororin with cohesin by the acetyltransferases Esco1 and Esco2 in humans, but the functional mechanisms of these acetyltransferases remain elusive. Here, we showed that Esco1 requires Pds5, a cohesin regulatory subunit bound to Rad21, to form cohesion via SMC3 acetylation and the stabilization of the chromatin association of sororin, whereas Esco2 function was not affected by Pds5 depletion. Consistent with the functional link between Esco1 and Pds5, Pds5 interacted exclusively with Esco1, and this interaction was dependent on a unique and conserved Esco1 domain. Crucially, this interaction was essential for SMC3 acetylation and sister chromatid cohesion. Esco1 localized to cohesin localization sites on chromosomes throughout interphase in a manner that required the Esco1-Pds5 interaction, and it could acetylate SMC3 before and after DNA replication. These results indicate that Esco1 acetylates SMC3 via a mechanism different from that of Esco2. We propose that, by interacting with a unique domain of Esco1, Pds5 recruits Esco1 to chromatin-bound cohesin complexes to form cohesion. Furthermore, Esco1 acetylates SMC3 independently of DNA replication.

Cornelia de Lange syndrome (CdLS) is a dominantly inherited congenital malformation disorder, caused by mutations in the cohesin-loading protein NIPBL for nearly 60% of individuals with classical CdLS, and by mutations in the core cohesin components SMC1A (~5%) and SMC3 (<1%) for a smaller fraction of probands. In humans, the multisubunit complex cohesin is made up of SMC1, SMC3, RAD21 and a STAG protein. These form a ring structure that is proposed to encircle sister chromatids to mediate sister chromatid cohesion and also has key roles in gene regulation. SMC3 is acetylated during S-phase to establish cohesiveness of chromatin-loaded cohesin, and in yeast, the class I histone deacetylase Hos1 deacetylates SMC3 during anaphase. Here we identify HDAC8 as the vertebrate SMC3 deacetylase, as well as loss-of-function HDAC8 mutations in six CdLS probands. Loss of HDAC8 activity results in increased SMC3 acetylation and inefficient dissolution of the ‘used’ cohesin complex released from chromatin in both prophase and anaphase. SMC3 with retained acetylation is loaded onto chromatin, and chromatin immunoprecipitation sequencing analysis demonstrates decreased occupancy of cohesin localization sites that results in a consistent pattern of altered transcription seen in CdLS cell lines with either NIPBL or HDAC8 mutations.

The ring-shaped chromosomal cohesin complex holds sister chromatids together by topological embrace, a prerequisite for accurate chromosome segregation. Cohesin plays additional roles in genome organization, transcriptional regulation and DNA repair. The cohesin ring includes an ABC family ATPase, but the molecular mechanism by which the ATPase contributes to cohesin function is not yet understood. Here we have purified budding yeast cohesin, as well as its Scc2-Scc4 cohesin loader complex, and biochemically reconstituted ATP-dependent topological cohesin loading onto DNA. Our results reproduce previous observations obtained using fission yeast cohesin, thereby establishing conserved aspects of cohesin behavior. Unexpectedly, we find that non-hydrolyzable ATP ground state mimetics ADP·BeF2, ADP·BeF3 - and ADP·AlFx, but not a hydrolysis transition state analog ADP·VO4 3-, support cohesin loading. The energy from nucleotide binding is sufficient to drive the DNA entry reaction into the cohesin ring. ATP hydrolysis, thought to be essential for in vivo cohesin loading, must serve a subsequent reaction step. These results provide molecular insight into cohesin function and open new experimental opportunities that the budding yeast model affords.

Structural maintenance of chromosome (SMC) protein complexes are able to extrude DNA loops. While loop extrusion constitutes a fundamental building block of chromosomes, other factors may be equally important. Here, we show that yeast cohesin exhibits pronounced clustering on DNA, with all the hallmarks of biomolecular condensation. DNA-cohesin clusters exhibit liquid-like behavior, showing fusion of clusters, rapid fluorescence recovery after photobleaching and exchange of cohesin with the environment. Strikingly, the in vitro clustering is DNA length dependent, as cohesin forms clusters only on DNA exceeding 3 kilo-base pairs. We discuss how bridging-induced phase separation, a previously unobserved type of biological condensation, can explain the DNA-cohesin clustering through DNA-cohesin-DNA bridges. We confirm that, in yeast cells in vivo, a fraction of cohesin associates with chromatin in a manner consistent with bridging-induced phase separation. Biomolecular condensation by SMC proteins constitutes a new basic principle by which SMC complexes direct genome organization.

Sister chromatid cohesion, mediated by cohesin, is required for accurate chromosome segregation [1, 2]. This process requires acetylation of cohesin subunit SMC3 by evolutionarily conserved cohesin acetyltransferases: Eco1 in budding yeast; XEco1 and XEco2 in Xenopus; and ESCO1 and ESCO2 in human [3-10]. Eco1 is recruited to chromatin through physical interaction with PCNA [11] and is degraded by the Skp1/Cul1/F-box protein complex after DNA replication to prevent ectopic cohesion formation [12]. In contrast, XEco2 recruitment to chromatin requires prereplication complex formation [13] and is degraded by the anaphase-promoting complex (APC) [14]. In human, whereas ESCO1 is expressed throughout the cell cycle, ESCO2 is detectable in S phase and is degraded after DNA replication [6, 15]. Although PDS5, a cohesin regulator, preferentially promotes ESCO1-dependent SMC3 acetylation [16], little is known about the molecular basis of the temporal regulation of ESCO2. Here, we show that ESCO2 is recruited to chromatin before PCNA accumulation. Whereas no interaction between PCNA and ESCO proteins is observed, ESCO2, but not ESCO1, interacts with the MCM complex through a unique ESCO2 domain. Interestingly, the interaction is required to protect ESCO2 from proteasomal degradation and is attenuated in late S phase. We also found that ESCO2 physically interacts with the CUL4-DDB1-VPRBP E3 ubiquitin ligase complex in late S phase and that post-replicative ESCO2 degradation requires the complex as well as APC. Thus, we propose that the MCM complex couples ESCO2 with DNA replication and that the CUL4-DDB1-VPRBP complex promotes post-replicative ESCO2 degradation, presumably to suppress cohesion formation during mitosis.

Cohesin is a conserved, ring-shaped protein complex that topologically embraces DNA. Its central role in genome organization includes functions in sister chromatid cohesion, DNA repair, and transcriptional regulation. Cohesin loading onto chromosomes requires the Scc2-Scc4 cohesin loader, whose presence on chromatin in budding yeast depends on the RSC chromatin remodeling complex. Here we reveal a dual role of RSC in cohesin loading. RSC acts as a chromatin receptor that recruits Scc2-Scc4 by a direct protein interaction independent of chromatin remodeling. In addition, chromatin remodeling is required to generate a nucleosome-free region that is the substrate for cohesin loading. An engineered cohesin loading module can be created by fusing the Scc2 C terminus to RSC or to other chromatin remodelers, but not to unrelated DNA binding proteins. These observations demonstrate the importance of nucleosome-free DNA for cohesin loading and provide insight into how cohesin accesses DNA during its varied chromosomal activities.

Sequences that form DNA secondary structures, such as G-quadruplexes (G4s) and intercalated-Motifs (iMs), are abundant in the human genome and play various physiological roles. However, they can also interfere with replication and threaten genome stability. Multiple lines of evidence suggest G4s inhibit replication, but the underlying mechanism remains unclear. Moreover, evidence of how iMs affect the replisome is lacking. Here, we reconstitute replication of physiologically derived structure-forming sequences to find that a single G4 or iM arrest DNA replication. Direct single-molecule structure detection within solid-state nanopores reveals structures form as a consequence of replication. Combined genetic and biophysical characterisation establishes that structure stability and probability of structure formation are key determinants of replisome arrest. Mechanistically, replication arrest is caused by impaired synthesis, resulting in helicase-polymerase uncoupling. Significantly, iMs also induce breakage of nascent DNA. Finally, stalled forks are only rescued by a specialised helicase, Pif1, but not Rrm3, Sgs1, Chl1 or Hrq1. Altogether, we provide a mechanism for quadruplex structure formation and resolution during replication and highlight G4s and iMs as endogenous sources of replication stress.

The chromosomal cohesin complex establishes sister chromatid cohesion during S phase, which forms the basis for faithful segregation of DNA replication products during cell divisions. Cohesion establishment is defective in the absence of either of three non-essential Saccharomyces cerevisiae replication fork components Tof1-Csm3 and Mrc1. Here, we investigate how these conserved factors contribute to cohesion establishment. Tof1-Csm3 and Mrc1 serve known roles during DNA replication, including replication checkpoint signaling, securing replication fork speed, as well as recruiting topoisomerase I and the histone chaperone FACT. By modulating each of these functions independently, we rule out that one of these known replication roles explains the contribution of Tof1-Csm3 and Mrc1 to cohesion establishment. Instead, using purified components, we reveal direct and multipronged protein interactions of Tof1-Csm3 and Mrc1 with the cohesin complex. Our findings open the possibility that a series of physical interactions between replication fork components and cohesin facilitate successful establishment of sister chromatid cohesion during DNA replication.

Concomitant with DNA replication, the chromosomal cohesin complex establishes cohesion between newly replicated sister chromatids. Cohesion establishment requires acetylation of conserved cohesin lysine residues by Eco1 acetyltransferase. Here, we explore how cohesin acetylation is linked to DNA replication. Biochemical reconstitution of replication-coupled cohesin acetylation reveals that transient DNA structures, which form during DNA replication, control the acetylation reaction. As polymerases complete lagging strand replication, strand displacement synthesis produces DNA flaps that are trimmed to result in nicked double-stranded DNA. Both flaps and nicks stimulate cohesin acetylation, while subsequent nick ligation to complete Okazaki fragment maturation terminates the acetylation reaction. A flapped or nicked DNA substrate constitutes a transient molecular clue that directs cohesin acetylation to a window behind the replication fork, next to where cohesin likely entraps both sister chromatids. Our results provide an explanation for how DNA replication is linked to sister chromatid cohesion establishment.