SMC complexes can traverse physical roadblocks bigger than their ring size

Abstract

The ring-shaped structural-maintenance-of-chromosomes (SMC) complexes condensin and cohesin extrude loops of DNA as a key motif in chromosome organization. It remains, however, unclear how these SMC motor proteins can extrude DNA loops in chromatin that is bound with proteins. Here, using in vitro single-molecule visualization, we show that nucleosomes, RNA polymerase, and dCas9 pose virtually no barrier to DNA loop extrusion by yeast condensin. Strikingly, we find that even DNA-bound nanoparticles as large as 200 nm, much bigger than the SMC ring size, can be translocated into DNA loops during condensin-driven extrusion. Similarly, human cohesin can pass 200 nm particles during loop extrusion, which even occurs for a single-chain version of cohesin in which the ring-forming subunits are covalently linked and cannot open up to entrap DNA. These findings disqualify all common loop-extrusion models where DNA passes through the SMC rings (pseudo)topologically, and instead point to a nontopological mechanism for DNA loop extrusion.

Cohesin mediates DNA loop extrusion by a “swing and clamp” mechanism

“…The ability of NIPBL to interact in a mutually exclusive manner with the hinge and the SMC3 head, and by doing so to “jump ship” from one DNA binding site to another, may therefore be a key element of the DNA translocation process. Importantly, this model could also explain how cohesin might be able to move DNA in the presence of nucleosomes or other DNA bound proteins ( Pradhan et al., 2021 ). Because chromatin fibers would be translocated beyond the ATPase heads when these are disengaged and could be 30–50 nm apart from each other, there would be space for DNA bound proteins to pass by the ATPase heads.…”

Three-Dimensional Loop Extrusion

“…where condensin/cohesin do not topologically embrace the DNA substrate [50,51]. Intriguingly, recent in vitro evidence provided direct evidence that loop extrusion is non-topological [38] and that some SMCs can form long-range bridges [7]. In light of this, inter-strand grabbing cannot be excluded based on structural or topological considerations and may be particularly likely to occur in dense DNA conditions.…”

“…non-trivial positioning of SMCs yielding crossing looping topologies (see Fig. 1B), or the bypassing of roadblocks several times the size of condensin [38]. At the same time, loop-capture mechanisms [34,36] cannot explain the observed processive extrusion of condensin on tethered DNA [11].…”

The Interplay of Cohesin and the Replisome at Processive and Stressed DNA Replication Forks

“…However, while cohesin can move past nucleosomes (~10 nm) in vitro, larger replisome-sized DNA-bound proteins of ~20 nm form an impenetrable barrier [ 10 , 11 ], and sub-20 nm sized DNA translocases, including RNA polymerases, translocate cohesin instead of passing through the cohesin lumen in vitro [ 10 , 11 ]. Note that this is different during loop extrusion, which may be non-topological and is largely unimpeded even by 200 nm barriers [ 62 ]. Nevertheless, in S. cerevisiae , cohesion establishment can take place in the absence of Scc2-dependent cohesin loading during S phase [ 63 , 64 ], and in U2OS cells with inactivated WAPL-mediated cohesin turnover, pre-loaded cohesin can remain associated with chromatin during replication fork passage [ 65 ].…”