Uncovering a New Step in Sliding Nucleosomes

Nucleosome sliding is achieved by a highly conserved ATPase motor that has the same architecture as bona fide helicases, with two core domains that open and close in a nucleotide-dependent fashion [1]. For monomeric helicases, repeated cycles of opening and closing, powered by ATP binding, hydrolysis, and release, allow the bilobed motor to ratchet itself along single-stranded DNA or RNA in an inchworm fashion [2]. A longstanding question has been how a helicase-like motor can overcome the formidable challenge of shifting nearly 150 bp of duplex DNA wrapped around a histone core.

One popular model for nucleosome sliding suggests that the ATPase motor spools out loops of DNA on the nucleosome that then whip around the histone core like a hula hoop. An alternative model proposed long before the first nucleosome crystal structure, called twist diffusion, posited that nucleosomes could be repositioned if short DNA segments between histone contacts could transiently accommodate an additional base pair, which would alter the twist of that segment [3]. The transfer of the twist (called a twist defect) from one DNA segment to the next would reflect each segment taking on and then giving up an additional base pair as the defect traveled around the nucleosome. Remarkably, the first nucleosome crystal structure showed a natural propensity to accommodate twist defects at an internal DNA location called superhelix location 2 (SHL2) [4], which corresponds to the binding site for many chromatin remodelers.

In contrast to features required for loop-spooling models, observations for the Chd1 remodeler recently supported a direct connection between remodeler action and twist defects. Unlike the classical notion of helicases shifting their nucleic acid substrates in a single power stroke, Chd1 was found to shift nucleosomal DNA in both the open and closed states, which can be explained by differences in DNA twist [5]. In the open state (apo and ADP bound), the ATPase motor effectively undertwists DNA at SHL2, creating a bulge that pulls one to two base pairs of entry-side DNA toward itself (Figure 1A). The closed state of the ATPase, promoted by transition-state analogs (ADP·BeF₃⁻, ADP·AlF_X, and ADP·MgF_X), disfavors the bulged and undertwisted duplex created by the open ATPase. Instead, the closed ATPase stabilizes a more canonical DNA duplex at SHL2, forcing the twist defect to propagate toward the dyad. A missing piece of this puzzle was understanding how the remodeler ATPase grips and distorts DNA at the SHL2-binding site to initiate and stabilize the twist defect.

Now, two recent cryo-electron microscopy (EM) studies by Chen and coworkers [6,7] have captured remodeler-induced bulges at SHL2, revealing a striking intermediate in the nucleosome sliding cycle. These studies focused on two different remodeler types, ISWI and SWI/SNF, and trapped each in open and closed states on the nucleosome. The closed state was achieved by loading remodelers with the ATP mimic/transition state analog ADP·BeF₃⁻, and these structures closely resembled previously visualized Chd1-nucleosome complexes also stabilized with ADP·BeF₃⁻ [8,9]. For these complexes, the DNA gripped by the remodeler ATPases followed largely the same path as in the free nucleosome, consistent with the closed ATPase motor re-establishing the canonical twist of nucleosomal DNA.

In contrast to the closed state, the open state observed for ISWI and SWI/SNF remodelers revealed an unanticipated change in the structure of nucleosomal DNA. In the open state, the DNA duplex was expected to bulge at the SHL2-binding site, which would allow entry DNA to ratchet forward by one to two base pairs. However, in these structures, a DNA bulge was only observed for one of the two DNA strands, called the tracking strand. Lifted away from its previous location on the nucleosome by 4–5 Å, the bulge at SHL2 ratcheted the tracking strand along the histone octamer by one nucleotide. Since the other strand, called the guide strand, did not move, the shift of the tracking strand caused a tilting of base pairs (Figure 1B). This remarkable base pair tilt extended all the way from SHL2 through the entry side of the nucleosome. The authors supported this unexpected finding in several ways. Not only was the same tilt in base pairs found for two different remodeler–nucleosome complexes, but a shift of the tracking but not guide strand was also apparent for a cryo-EM structure of the SWI/SNF ATPase bound to a nucleosome with a different sequence (MMTV rather than the Widom 601) [6]. Moreover, using single-molecule Förster resonance energy transfer (FRET), they showed that SWI/SNF shifted only the tracking strand in the open state, with the guide strand catching up when the ATPase transitioned to the closed state.

Since the guide strand remains unshifted in the open state, the job of the remodeler to move the DNA duplex is only half done. Yet, the extensive tilting of base pairs caused from translocation of just the tracking strand likely represents strain in the DNA duplex that could help ratchet the guide strand forward. Potentially, such a strain-assisted shift may help explain why nucleosome sliding rates are influenced by DNA sequence. This intriguing half-shifted DNA of the open state implies that the remodeler should subsequently promote a distinctly bulged DNA structure at SHL2, corresponding to a one nucleotide shift of both tracking and guide strands from the entry side of the nucleosome. A possible example of such a bulged structure is the ADP·BeF₃⁻-trapped complex of SWR1 with the nucleosome [10]. However, SWR1 specializes in histone exchange and does not reposition nucleosomes; therefore, more structural and biochemical evidence will be required to determine whether the SHL2 bulge created by SWR1 is analogous to an intermediate in the nucleosome sliding process.

Built on nearly two decades of work, the elusive mechanism of nucleosome sliding appears at long last to be coming into view. Given the rapid advancements fueled by cryo-EM, the field will no doubt soon solidify a mechanistic understanding of how remodelers work.