SMARCA4 (BRG1) and SMARCB1 (INI1) ATPase Motors of SWI/SNF Chromatin Remodeling

SMARCA4 (SWI/SNF-related, matrix-associated, actin-dependent regulator of chromatin, subfamily A, member 4, also designated BRG1, encoded on human chromosome 19p13.2 across 35 exons, ~185 kDa) and its mutually exclusive paralog SMARCA2 (BRM, encoded on chromosome 9p24.3, ~181 kDa) serve as the catalytic adenosine triphosphate-hydrolyzing DNA translocase motors across all three mammalian SWI/SNF (BAF) chromatin remodeling subcomplexes: canonical BAF (cBAF), polybromo-associated BAF (PBAF), and non-canonical BAF (ncBAF). Operating alongside the core structural subunit SMARCB1 (INI1, SNF5, or BAF47, encoded on chromosome 22q11.23 across 9 exons, ~47 kDa), these subunits convert chemical free energy from ATP hydrolysis into mechanical torque that slides, ejects, or reconfigures histone octamers along genomic DNA to regulate promoter and enhancer accessibility.

Somatic and germline mutations in SMARCA4 and SMARCB1 define distinct diagnostic and therapeutic classes of human malignancy. Inactivating mutations in SMARCA4 occur in ~8–10% of non-small-cell lung carcinomas (NSCLC), Thoracic SMARCA4-deficient undifferentiated tumors, and >95% of small-cell carcinomas of the ovary, hypercalcemic type (SCCOHT), establishing a paralog synthetic lethality with residual SMARCA2 (BRM). Conversely, biallelic deletion or truncation of SMARCB1 represents the defining genomic lesion in >95% of pediatric malignant rhabdoid tumors (MRT), atypical teratoid/rhabdoid tumors (AT/RT), and epithelioid sarcomas. Loss of SMARCB1 destabilizes cBAF and PBAF binding to the nucleosome acidic patch, permitting unopposed deposition of repressive trimethylated lysine 27 on histone H3 (H3K27me3) by the Polycomb Repressive Complex 2 (PRC2) methyltransferase EZH2 and creating a clinical vulnerability to EZH2 inhibition.

Snf2-Family RecA Bilobed ATPase Motor Mechanics at SHL2

Both SMARCA4 (1,647 amino acids) and SMARCA2 (1,590 amino acids) share ~75% overall amino acid sequence identity and >86% identity across their central catalytic Snf2-type ATPase domain. From N-terminus to C-terminus, both polypeptides comprise a proline- and glutamine-rich region, a QLQ (Gln-Leu-Gln) protein-interaction motif, a helicase/SANT-associated (HSA) α-helical helix that binds nuclear β-actin (ACTB) and actin-related protein ACTL6A (BAF53a), a bilobed RecA-like ATPase core (DExx Lobe 1 and HELIc Lobe 2 separated by a flexible hinge and braced by Post-HSA and Protrusion-1 helices), an AT-hook DNA-contacting motif, and a C-terminal bromodomain that recognizes acetylated lysine tails on histones H3 and H4 (H3K27ac, H4K16ac).

Cryo-electron microscopy (cryo-EM) reconstructions of nucleosome-engaged human cBAF and PBAF complexes at 2.9–3.7 Å resolution demonstrate that the bilobed SMARCA4 ATPase domain binds internal nucleosomal duplex DNA at superhelical location 2 (SHL2), roughly 20 base pairs from the nucleosomal dyad axis. In the apo and adenosine diphosphate (ADP)-bound open conformations, RecA Lobe 1 and Lobe 2 adopt a twisted orientation that introduces a 1-bp tracking deformation in the minor groove at SHL2. Binding of a single molecule of Mg2+–ATP across the Walker A (P-loop) and Walker B motifs drives cleft closure between Lobe 1 and Lobe 2, pulling 1 bp of linker DNA into the nucleosome core from the entry side. Subsequent ATP hydrolysis and orthophosphate (Pi) release reset the open conformation, propagating a 1-bp DNA bulge across the histone octamer dyad toward the exit site at rates of ~10–30 bp·s⁻¹.

SMARCB1 C-Terminal α-Helix Clamping of the H2A–H2B Acidic Patch

Productive DNA translocation by the SMARCA4 motor requires a stationary mechanical counter-anchor so that ATPase rotation propels DNA across the octamer rather than spinning the remodeler around the DNA duplex. Within cBAF and PBAF complexes (which contain ARID1A/ARID1B or ARID2/PBRM1, respectively), this counter-torque is provided by the base module and SMARCB1 (385 amino acids). SMARCB1 contains an N-terminal Winged-Helix (WH) domain and two tandem imperfectly repeated RPT1 and RPT2 domains that assemble with the SMARCC1/SMARCC2 (BAF155/BAF170) coiled-coil scaffold and SMARCE1.

At the extreme C-terminus of SMARCB1 (residues 357–378), a positively charged arginine-rich α-helix (Arg366, Arg370, Arg373, Arg374, and Arg377) docks directly into the negatively charged H2A–H2B acidic patch (Glu56, Glu61, Glu64, Asp90, Glu91, Glu92 on H2A and Glu105, Glu113 on H2B) on the bottom face of the nucleosome disk, opposite the SMARCA4 ATPase motor clamping the upper gyre at SHL2. This bilateral pincer grip sandwiches the histone octamer between SMARCA4 and SMARCB1, coupling ATP hydrolysis to high-efficiency nucleosome sliding and octamer eviction. Notably, the non-canonical ncBAF (GBAF) subcomplex lacks both ARID subunits and SMARCB1, utilizing BRD9 and GLTSCR1/GLTSCR1L to remodel promoter-proximal and CTCF-bound insulator sites without acidic-patch clamping.

Mutational Spectrum of SMARCA4 in NSCLC and SCCOHT

Across human malignancies, SMARCA4 exhibits two distinct mutational regimes depending on tissue lineage. In small-cell carcinoma of the ovary, hypercalcemic type (SCCOHT)—an aggressive tumor affecting young women with a median age of ~24 years—biallelic truncating nonsense, frameshift, or splice-site mutations in SMARCA4 occur in >95% of cases within an otherwise quiet diploid genome, accompanied by epigenetic silencing of the paralogous SMARCA2 locus. In non-small-cell lung adenocarcinoma (NSCLC), SMARCA4 mutations occur in ~8–10% of tumors and frequently co-occur with STK11, KEAP1, and KRAS alterations. Roughly half of SMARCA4-mutant NSCLC cases harbor Class 1 truncating loss-of-function (LoF) variants that abolish BRG1 protein expression, while the remaining half carry Class 2 missense mutations (Thr910Met, Arg1157Trp, Gly1232Ser) clustered inside the ATPase cleft that act as dominant-negative motors able to assemble into BAF complexes but unable to hydrolyze ATP.

Targeted SMARCA2 PROTAC Degradation in SMARCA4-Deficient Tumors

Because every functional cBAF, PBAF, and ncBAF complex requires exactly one catalytic ATPase subunit (SMARCA4 or SMARCA2), tumor cells that lose SMARCA4 through Class 1 truncating mutations become dependent on residual SMARCA2 (BRM) to maintain basal enhancer and promoter chromatin accessibility. In the Cancer Dependency Map (DepMap), SMARCA2 ranks as the top selective genetic dependency in SMARCA4-null cancer cell lines (Chronos score < −1.5).

Because the ATPase active sites of SMARCA4 and SMARCA2 are >93% identical, dual catalytic ATPase inhibitors (BRM014, FHD-286) block both enzymes simultaneously and cause dose-limiting gastrointestinal and hematopoietic toxicity in diploid host tissues that rely on SMARCA4. Conversely, constitutional Smarca2 knockout mice are viable and fertile because normal tissues express abundant SMARCA4. To achieve paralog selectivity, structure-based drug discovery campaigns targeted the non-catalytic C-terminal bromodomain of SMARCA2/SMARCA4—leveraging minor conformational and lysine-positioning differences to engineer VHL-recruiting proteolysis-targeting chimeras (PROTACs) such as ACBI1, PRT3789, and oral PRT7732. Although the warhead binds both bromodomains with comparable micromolar affinity, ternary complex formation (SMARCA2–PROTAC–VHL) is cooperative for SMARCA2 (>20-to-30-fold degradation selectivity over SMARCA4), driving rapid ubiquitylation and 26S proteasomal clearance of SMARCA2 and regression of SMARCA4-null NSCLC xenografts while sparing SMARCA4-wild-type cells.

Acidic Patch Loss and Unopposed H3K27me3 Deposition

From Drosophila genetics to human developmental biology, the SWI/SNF (Trithorax group) chromatin remodeling complexes and the Polycomb Repressive Complex 2 (PRC2) operate in a dynamic tug-of-war across lineage-specifying developmental enhancers and bivalent (H3K4me3/H3K27me3) promoters. In normal cells, SMARCB1-containing cBAF and PBAF complexes clamp the nucleosome H2A–H2B acidic patch and rapidly evict PRC2 (composed of EZH2, EED, SUZ12, and RBBP4), preventing EZH2 from trimethylating histone H3 lysine 27 (H3K27me3) at tumor-suppressor and differentiation loci such as CDKN2A (p16INK4a).

When biallelic loss of SMARCB1 occurs on chromosome 22q11.23 in malignant rhabdoid tumors (MRT) or epithelioid sarcomas—or when the oncogenic SS18–SSX fusion protein in synovial sarcoma displaces SMARCB1 from cBAF complexes—the mutilated cBAF and PBAF assemblies lose their C-terminal acidic-patch anchor. Deprived of SMARCB1-mediated mechanical eviction, PRC2 accumulates across CDKN2A and lineage differentiation enhancers, depositing dense H3K27me3 domains that lock the cell in an undifferentiated, proliferating state.

Clinical EZH2 Inhibition (Tazemetostat) and Collateral ncBAF (BRD9) Dependency

This epigenetic antagonism creates two clinically validated synthetic-lethal dependencies in SMARCB1-deficient tumors:

1. EZH2 Methyltransferase Inhibition (Tazemetostat): Genetic depletion or S-adenosyl-L-methionine (SAM)-competitive pharmacological inhibition of EZH2 with tazemetostat (EPZ-6438) erases aberrant H3K27me3 marks at CDKN2A and differentiation loci in SMARCB1-null cells, restoring p16INK4a transcription and inducing cell-cycle arrest and terminal differentiation. In 2020, tazemetostat received FDA approval for metastatic or locally advanced INI1 (SMARCB1)-negative epithelioid sarcoma.
2. Non-Canonical ncBAF (BRD9) Dependency: Because SMARCB1 loss destabilizes cBAF and PBAF while leaving the SMARCB1-independent ncBAF complex intact, SMARCB1-deficient rhabdoid and synovial sarcoma cells rely on ncBAF to maintain RNA polymerase II transcription at super-enhancer and CTCF sites. Targeted PROTAC degraders of the ncBAF-specific bromodomain subunit BRD9 (dBRD9, CFT8634, FHD-609) selectively collapse residual SWI/SNF activity in SMARCB1-null tumors.

Why do Coffin–Siris and Nicolaides–Baraitser syndromes arise from germline missense mutations in SMARCA4 and SMARCA2 rather than truncations?

Whereas somatic carcinomas frequently acquire biallelic truncating mutations in SMARCA4, germline heterozygous variants in SMARCA4 (causing Coffin–Siris syndrome 4) and SMARCA2 (causing Nicolaides–Baraitser syndrome) are almost exclusively non-truncating missense substitutions (>90%) clustered inside the RecA ATPase lobes and HSA cleft. These mutant polypeptides assemble normally into cBAF/PBAF/ncBAF holoenzymes, poisoning 50% of cellular complexes with catalytically dead motors (dominant-negative mechanism) rather than allowing the wild-type allele or paralog to compensate.

Why is SCCOHT (>95% SMARCA4-null) exceptionally sensitive to SMARCA2 degraders even without a SMARCA2 PROTAC?

In small-cell carcinoma of the ovary, hypercalcemic type (SCCOHT), SMARCA4 is inactivated by biallelic somatic or germline mutations while the paralogous SMARCA2 locus is transcriptionally silenced via promoter histone deacetylation rather than genetic mutation. Consequently, SCCOHT cells contain near-zero levels of both SMARCA4 and SMARCA2 and survive on trace residual SMARCA2—making them sensitive both to SMARCA2 PROTACs (which eliminate the trace pool) and paradoxically to HDAC or EZH2 inhibitors (tazemetostat), which re-express SMARCA2 to trigger growth-suppressive differentiation.

How do SMARCB1 C-terminal missense mutations in Schwannomatosis differ from truncations in Malignant Rhabdoid Tumor?

Complete biallelic loss of SMARCB1 via nonsense or frameshift deletions on Chr22q11.23 abolishes SMARCB1 protein expression and drives aggressive pediatric malignant rhabdoid tumors (MRT) and AT/RT. By contrast, germline hypomorphic missense or exon-skipping splice variants that preserve partial SMARCB1 protein expression and partial acidic-patch binding predispose adults to benign schwannomatosis (frequently requiring a somatic second hit alongside co-deletion of neighboring NF2 on chromosome 22q).

1. Wilson, B. G., Wang, X., Shen, X., et al. (2010). "Epigenetic antagonism between polycomb and SWI/SNF complexes during oncogenic transformation." Cancer Cell, 18(4), 316–328. https://doi.org/10.1016/j.ccr.2010.09.006
2. Valencia, A. M., Collings, C. K., Dao, H. T., et al. (2019). "Recurrent SMARCB1 mutations reveal a nucleosome acidic patch interaction site that potentiates mSWI/SNF complex chromatin remodeling." Cell, 179(6), 1342–1356. https://doi.org/10.1016/j.cell.2019.10.044
3. Farnaby, W., Koegl, M., Roy, M. J., et al. (2019). "BAF complex vulnerabilities in cancer demonstrated via structure-based PROTAC design." Nature Chemical Biology, 15(7), 672–680. https://doi.org/10.1038/s41589-019-0294-6
4. Michel, B. C., D'Avino, A. R., Cassel, S. H., et al. (2018). "A non-canonical SWI/SNF complex is a synthetic lethal target in cancers driven by BAF complex perturbation." Nature Cell Biology, 20(12), 1410–1420. https://doi.org/10.1038/s41556-018-0221-1