DNA passes through cohesin’s hinge as well as its Smc3-kleisin interface

The ring model (Haering et al. 2002) proposes that sister chromatid cohesion is mediated by co-entrapment of sister DNAs inside a tripartite cohesin ring created by a pair of rod-shaped proteins (Smc1 and Smc3) whose two ends are connected through dimerization of their hinges at one end and by association of their ATPase domains at the other end with the N- and C-terminal domains of a kleisin subunit (Scc1). The model explains how Scc1 cleavage triggers anaphase (Uhlmann, Lottspeich, and Nasmyth 1999) but has hitherto only been rigorously tested using small circular mini-chromosomes in yeast, where crosslinking the ring’s three interfaces, creating a covalent circular molecule, induces catenation of individual sister DNAs (Haering et al. 2008; Srinivasan et al. 2018). If the model applies to real chromatids, then the ring must have a DNA entry gate essential for mitosis. Whether this is situated at the Smc3/Scc1 (Murayama and Uhlmann 2015; Murayama et al. 2018) or Smc1/Smc3 hinge (Gruber et al. 2006) interface is an open question. Using an in vitro system (Collier et al. 2020), we show that cohesin in fact possesses two DNA gates, one at the Smc3/Scc1 interface and a second at the Smc1/3 hinge. Unlike the Smc3/Scc1 interface, passage of DNAs through SMC hinges depends on both Scc2 and Scc3, a pair of regulatory subunits necessary for entrapment in vivo (Srinivasan et al. 2018). This property together with the lethality caused by locking this interface but not that between Smc3 and Scc1 in vivo (Gruber et al. 2006) suggests that passage of DNAs through the hinge is essential for building sister chromatid cohesion. Passage of DNAs through the Smc3/Scc1 interface is necessary for cohesin’s separase-independent release from chromosomes (Chan et al. 2012) and may therefore largely serve as an exit gate.