Chromatin Remodeling

A nucleosome is not a permanent fixture. It is thermodynamically stable enough to survive cellular life but not so stable that it cannot be moved — and moving it, rather than removing histones altogether by chemistry, turns out to be the cell's preferred strategy for controlling access. Chromatin remodeling refers to the ATP-dependent repositioning of nucleosomes on DNA, carried out by multi-subunit protein complexes built around a catalytic subunit belonging to the Snf2 family of the helicase superfamily 2 (SF2). These ATPases are not helicases in the classical sense — they do not unwind DNA strands — but they use the same fold and a similar translocase mechanism to pull duplex DNA past the surface of the histone octamer. The key distinction to hold onto: histone-modifying enzymes (acetyltransferases, methyltransferases, kinases) change the chemistry of histone tails, altering charge and recruiting reader proteins, while remodeling complexes change the physical position or composition of the nucleosome itself. The two systems talk to each other constantly — bromodomains and chromodomains within many remodeling complexes recognize acetylated or methylated histone tails, targeting the remodeler to the right place — but remodeling is mechanistically its own phenomenon: an energy-consuming, motor-driven translocation reaction. Remodelers accomplish four distinguishable outcomes, and it is useful to name them precisely because different complexes specialize in different ones:

The catalytic core. Every remodeling complex contains an ATPase subunit with two RecA-like lobes forming a split ATPase domain, flanked by regulatory domains (bromodomains, chromodomains, HSA domains, or SANT domains depending on the family) that dock the enzyme onto nucleosomal DNA and onto specific histone marks. The ATPase binds DNA at superhelical location 2 (SHL2), roughly two turns of DNA from the nucleosome's dyad axis — a position determined by cryo-EM structures of remodeler–nucleosome complexes solved over the last decade.

DNA translocation and loop propagation. The accepted mechanism, supported by both structural and single-molecule studies, is a DNA-tracking, loop-propagation model. The two ATPase lobes clamp DNA at SHL2 and, using cycles of ATP binding and hydrolysis, pull DNA from outside the nucleosome inward, past the histone surface — much like an inchworm — creating a small transient loop or bulge of DNA (roughly 1–2 base pairs per ATP cycle in the simplest step, though productive translocation events move larger amounts of DNA cooperatively). This bulge then propagates around the octamer surface by a wave of local histone–DNA contact breakage and reformation, eventually reaching the far edge of the nucleosome and emerging as a net shift of the histone octamer's position by roughly 10 or more base pairs per translocation cycle. Because the octamer's histone contacts are broken only transiently and locally, the octamer never fully dissociates from DNA during a sliding event — this preserves the nucleosome's overall integrity while relocating it.

Directionality and outcome depend on flanking accessory domains. ISWI- and CHD-family remodelers carry domains that sense the length of flanking linker DNA on either side of the nucleosome (an "AutoN" and "NegC" regulatory mechanism in ISWI, and DNA-binding domains in CHD1), biasing the motor to slide nucleosomes toward the center of long linkers. This is precisely why ISWI and CHD1 are the workhorses of nucleosome spacing — generating the regular "beads on a string" arrays with roughly uniform 165–200 bp repeat lengths seen genome-wide. SWI/SNF-family remodelers, by contrast, lack this linker-sensing restraint and instead use their extra bulk (an ATPase lobe plus large accessory subunits like BRG1/BAF and ARID proteins in the mammalian BAF complex) to generate larger DNA loops and apply enough torsional strain to fully evict the histone octamer, especially from destabilized nucleosomes at promoters — this is how SWI/SNF opens up promoter regions to create nucleosome-depleted regions (NDRs).

Histone exchange by INO80/SWR1. The SWR1 complex in yeast (and its mammalian counterpart, the p400/SRCAP-related machinery) does something mechanistically distinct: rather than simply sliding the octamer, it opens the nucleosome enough to remove one H2A–H2B dimer and replace it with an H2A.Z–H2B dimer, using ATP hydrolysis to drive this unidirectional exchange reaction. INO80, its close relative, runs the reverse reaction — removing H2A.Z and restoring canonical H2A — and also plays a major role in evicting nucleosomes ahead of replication forks and RNA polymerase II.

Targeting. No remodeler acts genome-wide indiscriminately. Targeting comes from three converging inputs: sequence-specific transcription factors that recruit remodelers directly (e.g., Gal4 recruiting SWI/SNF at the yeast GAL genes), reader domains recognizing histone marks (bromodomains binding H3K14ac or H4K16ac, chromodomains binding H3K4me3), and intrinsic DNA sequence features (AT-rich or GC-rich stretches that are intrinsically less favorable for histone wrapping, making some nucleosomes easier to evict than others).

Genetics first: SWI and SNF as separate screens that converged. The founding evidence for remodeling did not begin as biochemistry — it began as two independent genetic screens in budding yeast in the early-to-mid 1980s. One screen, led by Ira Herskowitz's and later Rando/Winston-lineage groups, searched for mutants defective in mating-type SWItching (hence "SWI"), because switching requires the HO endonuclease gene to be transcribed, and HO transcription requires opening a chromatin barrier. A separate screen by Marian Carlson's lab searched for mutants unable to grow on sucrose because they could not de-repress SUC2 (invertase) — a phenotype named Sucrose Non-Fermenting, "SNF". When the genes identified by both screens were cloned and sequenced in the late 1980s, several turned out to be identical: SWI2 and SNF2 were the same gene. This was the first hint that a single protein complex, acting through chromatin rather than through a dedicated transcription factor pathway, was required for activating multiple, seemingly unrelated genes.

Biochemical purification and the ATP requirement. Through the early 1990s, labs including those of Craig Peterson, Roger Kornberg, and Robert Kingston purified the SWI/SNF complex from yeast and showed, using purified components — recombinant nucleosomes reconstituted on defined DNA templates, purified SWI/SNF complex, and an ATP-regenerating system — that SWI/SNF could increase the accessibility of nucleosomal DNA to restriction enzymes and to transcription factors, but only in the presence of ATP. Omitting ATP, or substituting the non-hydrolyzable analog ATP-γ-S, abolished the effect. This was the direct biochemical proof that remodeling is an energy-consuming enzymatic reaction, not a passive structural change — establishing the entire conceptual category of "ATP-dependent remodeling complex" as distinct from histone-modifying enzymes (which were being characterized in the same years, notably by C. David Allis's identification of a GCN5-related histone acetyltransferase in 1996).

Direct visualization of sliding. The cleanest mechanistic proof that remodelers actually move the histone octamer along DNA — rather than simply destabilizing it in place — came from restriction-enzyme accessibility assays and, more directly, from mapping experiments using mononucleosomes reconstituted on end-labeled DNA fragments. Peterson and colleagues, and independently Wolffe, Owen-Hughes, and Workman, showed that after incubation with SWI/SNF or ISWI-family complexes and ATP, the histone octamer's position on the DNA fragment — mapped by nuclease protection footprinting — shifted measurably along the same molecule, changing which restriction sites became newly cleavable. Because the total mass of the nucleosome and its DNA-wrapping stoichiometry were unchanged, this ruled out histone loss as the explanation and established cis-sliding as a real, ATP-dependent, physical translocation event.

Single-molecule confirmation of loop propagation. More recent single-molecule FRET and optical/magnetic tweezer studies (notably from the Bhattacharya, Zhuang, and Blosser/Bustamante-associated groups in the 2000s and 2010s) tracked individual RSC (a SWI/SNF-family complex) and ISWI molecules acting on single nucleosomes in real time. These experiments observed exactly the predicted intermediate: transient DNA bulges appearing at a fixed internal position and propagating directionally around the octamer in discrete steps, with step sizes and kinetics matching the biochemical loop-propagation model and directly confirming that the octamer itself does not dissociate mid-reaction.

Cryo-EM structures close the loop (literally). Beginning around 2017, cryo-EM structures of Chd1, Snf2, and SWI/SNF-family ATPases bound to nucleosomes (from the Cairns, Chen, and He/Panne laboratories among others) visualized the ATPase clamped at SHL2 with the two RecA lobes in different nucleotide states, providing atomic-resolution confirmation of exactly where and how the motor grips DNA, replacing what had been an inferred mechanism with a directly observed structural one.

Example 1 — Predicting the effect of a SWI/SNF-family loss-of-function mutation on a specific promoter. Suppose a yeast strain carries a null mutation in SNF2 (the ATPase subunit of SWI/SNF) and you measure SUC2 mRNA under glucose-starvation conditions (the inducing condition) versus glucose-rich conditions (the repressing condition). In wild-type cells, SUC2 mRNA increases roughly 200-fold upon glucose starvation because SWI/SNF is recruited to the SUC2 promoter and evicts the nucleosome occluding the TATA box, allowing RNA Pol II and general transcription factors to load. In the snf2Δ strain, you would predict induction is largely abolished — not reduced to, say, half, but reduced close to baseline, because SUC2 activation in this system is essentially SWI/SNF-dependent rather than partially redundant. Actual data from Carlson-lab and Peterson-lab strains bear this out: snf2Δ cells show only a few-fold induction, roughly a 95–98% loss of the induction response, while a handful of other glucose-repressed genes (which do not depend on nucleosome eviction at their promoters, or which use redundant remodelers like RSC) remain largely unaffected. The lesson: SWI/SNF dependence is promoter-specific, determined by whether that particular promoter's activation is rate-limited by a positioned, inhibitory nucleosome.

Example 2 — Reasoning through a clinical case: SMARCB1 loss and rhabdoid tumor. A two-year-old presents with an aggressive renal or CNS tumor; genomic sequencing reveals biallelic loss of SMARCB1</strong> (also called INI1/*hSNF5), a core, non-catalytic subunit shared by essentially all mammalian SWI/SNF (BAF) complexes. Reasoning through the mechanism: SMARCB1 loss does not simply abolish BAF complex activity uniformly across the genome; instead it selectively cripples BAF's ability to be retained at a specific subset of enhancers, notably those regulating <strong>polycomb target genes and differentiation programs</strong>, while BAF-independent (or PBAF-buffered) loci are less affected. The consequence is a specific derepression of EZH2-driven Polycomb Repressive Complex 2 (PRC2) target genes and a block in normal differentiation, producing highly undifferentiated, aggressive tumors in infancy — malignant rhabdoid tumor — despite the tumor genome otherwise carrying very few additional mutations (rhabdoid tumors are notable for having one of the lowest mutation burdens of any pediatric cancer, consistent with a single driver event, SMARCB1* loss, being sufficient). This is why EZH2 inhibitors (e.g., tazemetostat) have been tested and approved in this specific molecular subtype — the therapeutic logic follows directly from the mechanistic prediction that BAF loss creates dependence on unopposed PRC2 activity.

Craig L. Peterson, working at the University of Massachusetts Medical School, became one of the central figures in converting SWI/SNF from a genetically defined "activator complex" into a biochemically and structurally understood machine. Building on the yeast genetics that first identified SWI2/SNF2, Peterson's laboratory reconstituted the yeast SWI/SNF complex on defined nucleosomal templates in vitro and demonstrated, in a series of papers through the 1990s, that the complex used ATP hydrolysis to alter nucleosome structure and mobility directly — bridging genetics and mechanistic biochemistry at a time when the very idea of an "ATP-dependent remodeling complex" as a distinct enzymatic category was still being established. His work, alongside that of Robert Kingston, Bradley Cairns, Carl Wu, and Tom Owen-Hughes, helped define the loop-propagation/DNA-translocation model that later structural biology would confirm directly.

Cancer. Subunits of the mammalian SWI/SNF (BAF) complex are mutated in an estimated 20% of all human cancers, making BAF-complex genes collectively among the most frequently altered gene sets in oncology — comparable in scale to TP53. ARID1A mutations are especially common in ovarian clear-cell carcinoma (roughly 50% of cases) and endometrial cancers; SMARCB1 loss defines malignant rhabdoid tumors; PBRM1 mutations occur in a large fraction of clear-cell renal cell carcinomas. This has motivated synthetic-lethality drug strategies — for example, ARID1A-deficient tumors show heightened sensitivity to EZH2 or ATR inhibition, because loss of one chromatin-regulatory arm creates dependence on a compensatory pathway.

Developmental biology and stem cell reprogramming. The BAF complex (in its embryonic-stem-cell-specific configuration, "esBAF") is required to maintain pluripotency gene expression and is functionally linked to core reprogramming factors; remodeler activity is now a standard consideration in optimizing induced pluripotent stem cell (iPSC) protocols.

Rare disease. Coffin–Siris syndrome and related neurodevelopmental disorders arise from heterozygous mutations in BAF-complex subunits (ARID1B, SMARCA4, SMARCB1), directly linking remodeler dosage to normal neural crest and cortical development.

Agricultural and model-organism research. ISWI- and CHD-family remodelers are conserved from yeast to plants and are studied in crop species to understand vernalization and stress-response chromatin dynamics, connecting basic remodeling mechanism to breeding strategies for climate resilience.

Drug discovery. Small-molecule inhibitors of the SWI/SNF ATPase subunits (targeting SMARCA2/SMARCA4) are in clinical development specifically for SMARCA4-mutant or SMARCB1-deficient tumors, directly exploiting the synthetic-lethal relationships that fall out of the mechanistic biology described above. Understanding remodeling also sets up the next level of genome organization: how repeatedly remodeled, variably compacted chromatin fibers are further folded, looped, and condensed into the structures examined in Chromosome Structure and the Mitotic Chromosome.

Ask the tutor: How do bromodomains and chromodomains within remodeling complexes physically couple histone modification state to remodeler recruitment and activity?