High resolution interaction surface mapping by PRISMA reveals novel ARID1A interactions

High resolution interaction surface mapping by PRISMA reveals novel ARID1A interactions | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article High resolution interaction surface mapping by PRISMA reveals novel ARID1A interactions Mercedes Pardo, Chiara Marcozzi, Karen Lane, Fernando Sialana, and 7 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9415881/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 8 You are reading this latest preprint version Abstract Background The cBAF SWI/SNF chromatin remodelling complex controls cell proliferation and fate determination by regulating chromatin accessibility and programs of gene expression, DNA damage response, replication, splicing, translation, and cell plasticity. This functional diversity stems from varied protein interactions. ARID1A is the most frequently mutated SWI/SNF subunit in cancer and scaffold of cBAF through the globular C-terminal domain. The rest of ARID1A mostly contains intrinsically disordered regions, key interaction surfaces that are notoriously difficult to study due to highly dynamic conformations. The molecular basis of ARID1A interactions remains obscure, and few have been explored functionally or mapped at an interface level. We perform a PRotein Interaction Screen on peptide MAtrix (PRISMA) to identify novel ARID1A interactions and map short linear motifs that mediate interactions at high sequence resolution. Results Our PRISMA assay recapitulates binding sites of cBAF subunits to ARID1A and detects previously described SLiM-mediated interactions. We reveal binding sites for direct interaction with transcriptional repressor SIN3A, interaction and confirm its role in SIN3A chromatin recruitment. We identify TOX4, CDK2 and CCNA2 as novel ARID1A interactors. Mutation of a cell cycle-regulated CDK2 phosphorylation site in ARID1A leads to altered gene expression of microtubule factors and defective cell division. Conclusions Our study provides novel mechanistic insight into ARID1A functions and highlights a role for CDK2-dependent phosphorylation of ARID1A in the regulation of cell division, with important ramifications for cancer. Our data represents a valuable resource to further study ARID1A and SWI/SNF biology, and more broadly, to investigate sequence and structural features governing protein interactions. ARID1A SWI/SNF PRISMA protein interactions SLiM peptide interactions mitosis phosphorylation Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Background ARID1A is a defining subunit of the multi-protein cBAF complex, one of several types of SWI/SNF chromatin remodelling complexes that use energy derived from ATP hydrolysis to alter the position of nucleosomes and modify DNA accessibility for RNA polymerase II and other transcriptional regulators, thereby regulating many DNA-based processes, such as transcription, splicing, DNA replication and DNA damage repair ( 1 – 7 ). ARID1A directly binds to many of the subunits of the BAF complex and in fact acts as a scaffold for its assembly ( 8 , 9 ). Loss of ARID1A leads to complex dissociation, reduced levels of other BAF subunits and BAF complex deficiency ( 10 – 12 ). ARID1A has also shown to interact with ribosomes and be important for translation ( 13 , 14 ) and reported to have a role in cell plasticity, through a mechano-regulated interaction with transcription factors in the Hippo signalling pathway ( 15 , 16 ). Subsequently, ARID1A and BAF function impact many key aspects of cell and organismal physiology ( 17 , 18 ). Mutations in ARID1A lead to neurodegenerative diseases and account for ~ 6% of all cancer mutations ( 19 – 22 ). Better understanding of regulatory mechanisms of ARID1A function could help develop novel therapeutic strategies. ARID1A is a large protein containing two conserved domains of globular structure, an ARID domain located centrally that mediates binding to DNA and a C-terminal domain that mediates interactions with other BAF subunits and is required for BAF complex formation. In contrast, the N-terminal half of the protein is mostly made up of intrinsically disordered regions (IDRs) ( 23 ). Not much is known about the function of this region of ARID1A. Despite their lack of structure, IDRs play crucial roles in the control of signalling pathways and can also work as platforms for the assembly of large proteinaceous structures. IDRs frequently contain SLiMs or short linear motifs that mediate interactions with globular domains ( 24 ). They also often carry post translational modifications (PTMs) that function as regulatory switches ( 25 ). Many disordered regions are capable of phase separation, and this is indeed a general property of most SWI/SNF subunits ( 26 ). The N-terminus of ARID1A can partition into phase-separated condensates, possibly through interaction with FUS, a nucleic acid binding protein that forms oncogenic fusion proteins ( 26 , 27 ). Recently, the IDRs of ARID1A were also shown to be required for genomic targeting of the BAF complex to enhancers and interaction with transcription factors and other transcriptional machinery ( 28 , 29 ). The majority of ARID1A interacting proteins that have been mapped to date bind to the C-terminal half of ARID1A (Fig. 1 A). This includes, in addition to other BAF subunits, proteins involved in DNA damage response and repair such as ATR and MSH2, cell cycle regulators such as p53 and RB, and other transcriptional regulators such as Polycomb protein EZH2 and the glucocorticoid receptor, transcription factor HIC2, topoisomerase enzyme TOP2A, and E3 ubiquitin ligase bTrCP. Characterised binding partners on the disordered N-terminal half of ARID1A include FUS, and YAP1/TAZ, transcription factors that regulate cell plasticity ( 15 ). In addition to these, many other ARID1A-interacting proteins have been reported in the literature. However, the molecular basis for these interactions is unknown. Recent reports suggest that many functional ARID1A interactions occur through the disordered regions. IDR-mediated interactions are inherently challenging to study due to their reversible and weak nature, and traditional approaches are not best suited to investigate them ( 30 , 31 ). Here, we take advantage of a peptide interactomics screen on a tiled peptide array to systematically identify amino acid residues, domains and motifs that mediate ARID1A interactions with other proteins at high sequence resolution. Our PRISMA data recapitulates interaction surfaces of ARID1A with other BAF complex subunits, and SLiM-mediated interactions. We identify and validate several novel ARID1A-interacting proteins with roles in transcriptional regulation and cell cycle control, and map their binding sites. Finally, we functionally characterise and uncover a role for ARID1A phosphorylation in the transcriptional regulation of microtubule organisation and cell division, providing the first insights into ARID1A regulation through phosphorylation. Results A sequence-resolution biophysical interaction map of ARID1A To systematically map the landscape of ARID1A protein-protein interactions at high sequence resolution and delineate interaction surfaces, domains and motifs in ARID1A, we took advantage of the recently developed Protein Interaction Screen on peptide MAtrices (PRISMA) technology ( 32 ). The PRISMA assay consists of high-throughput affinity purifications using hundreds of overlapping peptides spanning the entire sequence of a target protein immobilised on a nitrocellulose membrane, combined with shotgun mass spectrometry analysis, for identification and quantitation of peptide-bound proteins. We designed our PRISMA assay with 20-mer peptides with a 10-aa (amino acid) overlap covering the whole length of human ARID1A (2285 aas), resulting in a high-density tiled array of 228 peptide spots (Fig. 1 B, Table S1 ). Binding proteins were screened using nuclear extract from RPE1 cells in the presence of nuclease to eliminate DNA-mediated interactions. A total of 2224 proteins identified by at least 2 peptides were quantified (Fig. 1 C, D). Around a quarter of the proteins bound ubiquitously across the peptide array at low intensities, while the rest displayed much higher signals in discrete regions of the array. The frequency of binding was calculated for each spot, revealing that the number of interacting proteins per spot ranged from 1 to 235 (Fig. 1 D). Approximately 700 proteins (30%) were quantified in more than half of the peptide array (100 peptide spots), whilst around 100 were found to bind to over 200 peptide spots (Fig. 1 D). This so called “promiscuous” binding has also previously been observed in a PRISMA assay for C/EBPβ ( 32 ). To reduce the noise from background binding and highlight the strongest signals across the peptide array, binding signals below the 75th percentile intensity were removed. Furthermore, to discriminate robust interactions, a second filter based on the presence of consecutive signal intensities across a minimum of two contiguous peptides was applied. The consecutive binding is driven by the overlap in the tiled peptide sequences (Fig. 1 B, C). This strict filter criterion, applied after the 75th percentile outlier filtering, yielded an ARID1A core PRISMA interaction set of 1454 proteins. It is worth pointing out that the consecutive binding filter when applied to the ARID1A PRISMA dataset is much more restrictive that the one applied in the original PRISMA paper, where the matrix peptides were shorter (11 aas) and also had a smaller overlap (4 aas), which means that generally any sequence stretch was covered by three peptides, whilst in the ARID1A PRISMA assay described here any sequence is only covered by two peptides. The distribution of peptide interactions was not attributed to physicochemical parameters of the peptides, as evidenced by comparison of the number of binding proteins with peptide properties such as hydrophobicity, charge, and isoelectric point (Figure S1 ). These results suggest that the PRISMA peptides on the matrix retained their specific protein-binding properties. The global binding profiles from the PRISMA dataset were represented as a hierarchically clustered heatmap of normalised intensities (Fig. 1 D, Figure S1 , Table S2 ). This revealed there were large differences in the binding profiles of the matrix peptides indicating that binding was specific, as background binding would be expected to result in a homogeneous distribution of interacting proteins. Peptides encompassing the conserved A RID s caffolding d omains ASD1 and ASD2, and particularly those containing the LXXLL motifs #2 (aas 1709–1713) and #4 (aas 2085–2089), displayed the highest number of interactions, followed by regions flanking the ARID domain (Figure S1 ). The ARID DNA-binding domain and the NLS appeared devoid of interactions. The region immediately upstream of the ARID domain that sustains a high number of interactions is an unstructured stretch that is predicted to fold upon binding to a protein partner by the ANCHOR2 algorithm ( 33 ). The N-terminal region also displayed significant numbers of interactions, seemingly organised in locally enriched hotspots, some of which overlapped with predicted disordered sequences or disordered binding regions. GO term enrichment analysis of the PRISMA core interaction dataset revealed significant over-representation of proteins annotated with GO terms related to known ARID1A functions (Figure S2 A, Table S3 ), with enriched terms including translation, mRNA splicing, ATP-dependent chromatin remodelling and DNA damage response. Enrichment of protein domains revealed that nucleotide binding domain, RNA recognition domain, Armadillo, DNA/RNA helicase domain and AAA ATPase domain were some of the most frequently represented domains and sites (data not shown). To further assess the biological relevance of the PRISMA-derived data, we compared it to the ARID1A physical interactome retrieved from the BioGRID repository of protein interactions. Interactors identified by yeast two-hybrid or RNA-purification, viral proteins and duplicates were removed prior to comparison, leaving 211 ARID1A physical interactors. Forty-nine of these were detected by PRISMA and 31 of them were included in the core dataset (Figure S2 B). These results suggest that the PRISMA-derived interaction data shows a high level of overlap and coverage of the native ARID1A interactome and represents a useful resource to explore sequence-to-function relationships at high sequence resolution. PRISMA recapitulates binding of BAF subunits To assess the ability of the ARID1A PRISMA to detect localised interactions, we investigated the binding of components of the BAF complex. Core module subunits SMARCB1, SMARCC1/2, SMARCD1/3 and SMARCE1, ATPse module subunits SMARCA4 and ACTL6A, and accessory subunit DPF2 displayed correlated interaction footprints on the conserved ASD1 and ASD2 domains, immediately upstream of the ARID domain, and on the central part of the disordered N-terminal domain, with the latter displaying the strongest signals (Fig. 2 A). Consecutive binding signals of SMARCA4, SMARCB1 and SMARCE1 were observed in the ASD2 domain aligned with the LXXLL motif located at position 2085–2089, which has been shown to be essential for binding of BAF subunits ( 34 ). We then compared binding sites of BAF subunits with cryogenic electron microscopy (cryoEM) structural model 6LTJ data ( 9 ). The PRISMA interaction footprint was in good agreement with cryoEM maps, demonstrating multiple contacts of the ASD1 and ASD2 conserved domains of ARID1A with SMARCA4, SMARCB1, SMARCC1/2, SMARCD1, SMARCE1 and DPF2 (Fig. 2 B). Furthermore, binding signals of SMARCA4, SMARCB1 and SMARCE1 showed remarkable overlap with protein contact sites from the cryoEM structural models (Fig. 2 B). In this cryoEM study only a stretch of the ARID1A C-terminus (aas 1639–2285) was modelled, precluding the comparison of contact sites at the protein N-terminal half. Therefore, we also compared the PRISMA data to BAF complex crosslinking-mass spectrometry (XL-MS) data ( 8 ) that encompasses the whole ARID1A sequence. The agreement between both was particularly noteworthy (Fig. 2 B): in addition to the well-documented binding of several BAF subunits on the C-terminal region of ARID1A, both PRISMA and XL-MS showed SMARCD, SMARCC and ACTL6A strongly binding to ARID1A N-terminal region. The overlap was observed in regions, not in exact positions, as the different approaches provide different resolution and measure different properties, but converged on a consistent structure. Hence, the PRISMA data mirrored binding data from other experimental approaches and provided information on contact sites between associated proteins. Our PRISMA data suggests that while the BAF250_C region is essential for scaffolding the assembly of the BAF complex ( 34 ), the disordered N-terminus also mediates contacts with different BAF subunits, in agreement with XL-MS ( 8 ). Most SWI/SNF subunits have disordered prion-like regions that are involved in multivalent protein interactions ( 23 ). Interactions engaged by these disordered regions may serve a regulatory function. Next, the PRISMA data was compared with published sites of interaction in ARID1A with other proteins (Fig. 3 A). Out of 16 mapped ARID1A direct binders found in the PRISMA data, three displayed PRISMA binding patterns that agreed with their mapped site of interaction with ARID1A. Transcriptional coactivator YAP1 displayed consecutive binding at 3 peptides spanning ARID1A aas 131–170 (Fig. 3 B). These peptides contain the proline-rich PPXY sequence, a SLiM that mediates binding of ARID1A to the YAP1 WW domain, and mutation of ARID1A Tyr148 to Ala has been shown to eliminate YAP1 binding ( 15 ). Therefore, our PRISMA data successfully pinpointed the site of YAP1 interaction. Other WW domain proteins such as ITCH and NEDD4L were also detected binding to these PPXY containing peptides in our data, indicating that PRISMA can truly recapitulate SLiM-mediated direct associations. In contrast, transcriptional coactivator TAZ, which belongs to the same family as YAP, displayed consecutive binding at peptides from the ARID1A C-terminus and semi-consecutive binding between aas 2061–2100. TAZ only contains one WW domain in contrast to YAP1 which contains 2, so this might explain the lack of detection of TAZ at the 144–148 PPXY motif. The retinoblastoma protein (RB1) displayed weak consecutive binding signals on the two most C-terminal peptides (position 2261–2285), in agreement with published data ( 35 ). ARID1A has been shown to interact with Rb and suppress its phosphorylation, with two truncation fragments, ARID1A:1110–1976 and ARID1A:1977–2285, being able to mediate the interaction. The binding footprint of paraspeckle protein FUS showed strongest signals in the disordered N-terminus of ARID1A, and some residual binding in a disordered region upstream of the ASD1 domain. This is also consistent with published data showing that ARID1A interacts and partitions into biocondensates with FUS through its N-terminal prion-like domain ( 26 , 36 ). Other mapped ARID1A binders displayed only scattered single signals that could not be confidently attributed to specific binding (TOP2A, MSH2, ATR) or showed promiscuous binding (TGM2, ERH and PRKDC). In summary, our results demonstrate that the ARID1A PRISMA assay can recapitulate FUS, RB1 and YAP1 in vivo binding. These have not previously been identified in affinity purification-mass spectrometry (AP-MS) interactomes, suggesting that the in vitro set-up of PRISMA may be able to reveal complementary interactions that might be too labile or of too low abundance for traditional AP-MS approaches. PRISMA reveals binding of SIN3A to a predicted SIN3A SLiM Seeing as PRISMA can recapitulate SLiM-mediated interactions, we next investigated whether our assay could identify binding to other ARID1A SLiMs. We used the Eukaryotic Linear Motif database (ELMdb) ( 37 ) prediction tool to obtain a list of potential SLiMs in human ARID1A and checked whether peptides containing these motifs displayed significant specific binding of proteins containing the matching globular domain. We observed highest binding signal of E3 ubiquitin ligases on peptides containing degrons, and SH3 domain proteins on peptides containing an SH3-binding SLiM (Table S1 ). Detailed inspection of the data revealed strong consecutive binding of the transcriptional repressor SIN3A to peptides encompassing a predicted Sin3-interacting domain (SID) of low conservation (Fig. 4 A), suggesting this motif could mediate binding between SIN3A and ARID1A. The PRISMA data also showed SIN3A binding to the two C-terminal peptides (aas 2261–2285). ARID1A and SIN3A have been shown to physically interact to repress the expression of cell cycle genes in mouse osteoblast precursor cells ( 38 ), but the direct binding determinants have never been mapped. SIN3A has also been shown to interact with other subunits of BAF complexes ( 39 – 42 ). Therefore, our results suggest that ARID1A and SIN3A may interact directly through a Sin3 SLiM and other supporting nearby residues. ARID1A is required for the recruitment of SIN3A to chromatin in ovarian cancer cells and quiescent mouse pre-osteoblastic MC3T3-E1 cells ( 38 , 43 , 44 ). To assess if ARID1A is necessary for chromatin recruitment of SIN3A in RPE1 cells, we isolated chromatin from ARID1A KO RPE1 cells and their wild type counterparts and performed quantitative profiling of chromatin-binding proteins by mass spectrometry. Inspection of the chromatome revealed that levels of SIN3A on chromatin were significantly lower in ARID1A KO cells compared to WT cells (Fig. 4 B), confirming that ARID1A is required for the recruitment of SIN3A to chromatin in RPE1 cells. To assess if the association between ARID1A and SIN3A requires the ARID1A N-terminus, we expressed an ARID1A C-terminal fragment encompassing aas 1018–2285 tagged with Emerald-GFP (Em-ARID1A_C) in ARID1A KO RPE1 cells and performed GFP-trap immunoprecipitation followed by MS analysis. All subunits of the SIN3 complex were enriched in ARID1A_C immunoprecipitates compared to control GFP-only (Fig. 4 C), confirming that ARID1A C-terminus is sufficient for association with the SIN3 complex. The SID SLiMs of MAD1 and other proteins specifically interact with the PAH domain of SIN3 ( 45 ). To confirm if ARID1A and SIN3A interact directly, we individually expressed full length ARID1A fused to the V5 tag ( 46 ) and a SIN3A fragment encompassing aas 449–1086 which includes the PAH3 domain (SIN3A_PAH3) tagged with StrepII in insect cells and performed pull-down experiments (Fig. 4 B, C). Full length KDM5A-StrepII was included as a positive control, since its paralog KDM3A has previously been shown to bind to ARID1A ( 47 ). Bead-bound ARID1A captured KDM5A and SIN3A_PAH3 (Fig. 4 C). Although SIN3A_PAH3 was also detected in the negative control IgG immunoprecipitated, it was consistently much more abundant in V5 immunoprecipitates. Therefore, our results indicate that ARID1A interacts directly with the PAH domain of SIN3A. PRISMA identifies novel ARID1A interactor TOX4 Comparison of PRISMA data to ELM database predictions revealed potential novel candidate interactor TOX4 binding to two ARID1A peptides spanning aas 611–630 and 1691–1710 (Fig. 5 A). TOX4 is a regulatory subunit of the protein phosphatase 1 (PP1) complex that is almost exclusively associated with chromatin and has been shown to mediate dephosphorylation of the C-terminal domain (CTD) of RNA polymerase II, thus regulating transcriptional elongation ( 48 ). Despite being distal from each other, both TOX4-binding peptides encompassed phosphorylation consensus sites for several kinases (polo-like kinase 1 (PLK1), Casein kinase I (CK1) and glycogen synthase kinase 3 (GSK3)), suggesting that these were specific binding events. Phosphorylation of ARID1A serine 631 (Ser631) located in the kinase consensus site included in peptide 611–630 has been previously detected (Figure S3 , ( 49 )). We used the AlphaFold protein–protein interaction prediction algorithm ( 50 ) to predict the structure of the ARID1A peptides complexed with TOX4. For both peptides we obtained low confidence output models with low interface-predicted template modelling (0.13 and 0.41 respectively) and high PAE scores, not unexpectedly, given the highly disordered nature of TOX4 and one of the ARID1A peptides (Fig. 5 B). To validate the interaction, we immunoprecipitated TOX4 from RPE1 and HEK293T cells and probed the immunoprecipitates with an anti-ARID1A antibody (Fig. 5 C). We reproducibly detected anti-ARID1A reactive bands at 270 kDa band and above in TOX4 immunoprecipitates, confirming that TOX4 and ARID1A interact in vivo . Intriguingly, the ladder pattern of ARID1A bands was reminiscent of ubiquitination or sumoylation. We therefore probed TOX4 immunoprecipitates with an anti-ubiquitin antibody and observed that the ladder-like bands above 270 kDa were indeed ubiquitin-containing bands. These results confirm that TOX4 interacts with (poly)ubiquitinated ARID1A and suggest an interplay between (de)phosphorylation and ubiquitination of ARID1A. PRISMA uncovers ARID1A binding to CDK2/cyclin A2 The detection of CDK2 and cyclin A2 as overlapping consecutive binders in at least two different regions on ARID1A was of particular interest (Fig. 6 A). The highest binding signal for both proteins was detected on peptides spanning aas 1391–1430, where CDK2 and cyclin A2 binding overlapped on 2 out of 3 peptides. CDK2 also bound strongly between aas 531–600, with cyclin A2 displaying residual binding on one peptide in that span. Both binding sites overlapped with sequences of intrinsic disorder that were predicted to fold upon binding to a protein partner by the ANCHOR2 algorithm ( 33 ). CDK2 alone also displayed strong binding on peptides spanning aas 1451–1480, and residual binding on peptides spanning aas 331–370, which encompass a CDK2 consensus site. Phosphorylation of Serine 363 within this sequence is one of the most frequently detected ARID1A post-translational modifications (Figure S2 ). Exploration of a published CDK2-dependent phosphopeptidome study revealed that phosphorylation of Ser363 is dependent on CDK2 ( 51 ). Therefore, our results endorse the direct interaction between ARID1A with CDK2 and/or cyclin A via small regions of intrinsic disorder. To validate this, we synthesised a peptide subset array containing ARID1A peptides that had shown binding to CDK2 and/or CCNA2, an ARID1A peptide containing an RxL motif as positive control (RxL is the best characterised docking motif for cyclin A ( 52 )), and the corresponding scrambled sequences counterparts as negative controls (Table S4 ). The peptide array was then probed with a mix of recombinant full length CDK2 and a CCNA3 fragment spanning aas 174–432 that contains the hydrophobic patch that docks on the RxL motif ( 52 ) (Fig. 6 B). Western blot analysis revealed binding of CDK2 to the RxL peptide and not to its scrambled counterpart. We also detected binding of CDK2 to the ARID1A peptide spanning aas 1451–1470, but not to its scrambled counterpart. We could not perform cyclin A detection in this experiment because the fragment used does not contain the epitope of our anti-cyclin A antibody. To establish which of CDK2 and cyclin A is the direct binder to ARID1A, we probed a second copy of the array with recombinant full length CCNA2 (Fig. 6 C). Detection with an antibody against CCNA2 showed that CCNA2 alone binds to ARID1A peptide spanning aas 1451–1470, suggesting that CCNA2 mediates the binding of CDK2 to this sequence. Interestingly, CDK2 or cyclin A2 have never been identified as ARID1A-associated proteins in our AP-MS experiments or any published ARID1A AP-MS studies ( 4 , 53 , 54 ). Thus, our data highlight the utility of the PRISMA strategy in revealing weak, transient and/or low abundance ARID1A interactions, possibly mediated by SLiMs, that are not detectable by traditional affinity purification strategies. Exploring the role of ARID1A Ser363 phosphorylation A survey of our recent cell cycle phosphoproteomics study ( 55 ) revealed that ARID1A Ser363 phosphorylation is modulated along the cell cycle, being higher in G1 and S phases, and decreasing at the end of S phase and throughout G2 and mitosis (Fig. 6 D). To investigate the role of Ser363 phosphorylation, we generated RPE1 cells lines expressing inducible EmeraldGFP (emGFP)-tagged wild type (WT) ARID1A, phospho-mimetic Ser 363 to Asp mutated ARID1A (ARID1A_S363D) or phospho-dead Ser 363 to Ala ARID1A (ARID1A_S363A) in an ARID1A KO background. ARID1A_S363A was observed as a ~ 250 kDa band in agreement with its predicted size, but the S363D mutant could not be detected with anti-ARID1A antibodies, and in fact was only expressed as a small protein fragment containing emGFP (Figure S4 A, B). We confirmed this through primary sequence analysis by mass spectrometry of GFP-immunoprecipitates (Figure S4 C). ARID1A peptides spanning the entire ARID1A sequence were detected in ARID1A_S363A immunoprecipitates. In contrast, in ARID1A_S363D immunoprecipitates we could only detect GFP peptides and ARID1A peptides upstream of S363D, indicating that the ARID1A_S363D protein detected by anti-GFP western blotting is a truncated form of ARID1A. This suggests that the ARID1A_S363D mutant protein is unstable and gets degraded to a truncated version in a rapid manner or was strongly selected against and a genomic rearrangement led to the truncated form. We could not reverse this degradation with proteasome inhibition (Figure S4 B). Mass spectrometry analysis of GFP-trap immunoprecipitates also detected most BAF complex subunits in ARID1A_S363A immunoprecipitates (Figure S4 D), suggesting mutation of Ser363 into Ala does not impact the ability of ARID1A to underpin the assembly of the complex. As expected, no other BAF subunits were detected in ARID1A_S363D immunoprecipitates, since the complex integrity and assembly relies on the ARID1A C-terminal domain ( 56 ). Whilst EmGFP-ARID1A_S363A fusion protein localised mostly to the cell nucleus (Figure S4 E) in a similar manner to endogenous ARID1A, the S363D mutant localised to the cytoplasm, again as expected due to the loss of the NLS motif at position 1368–1387. To test if S363 phosphorylation is important for cell growth, we induced the expression of emGFP, wild type and S363A mutant emGFP-ARID1A proteins in ARID1A KO RPE1 cells and monitored their growth by measuring cell confluency over 120 h (Fig. 7 A). We observed that in these conditions, cells expressing em-GFP-ARID1A_S363A had a mild proliferation defect compared to cells expressing wild type emGFP-ARID1A, suggesting that ARID1A S363 phosphorylation may be important for cell proliferation. To explore the effects of the S363A mutation on ARID1A transcriptional regulation activity, we performed transcriptomics analysis using RNA sequencing (RNA-Seq) of ARID1A KO cells expressing emGFP-tagged wild type (WT) ARID1A, ARID1A_S363A or emGFP only, the untransfected ARID1A KO and the parental wild type cells after induction with doxycycline for 24 h. We then compared parental to ARID1A KO cells, and KO cells expressing wild type ARID1A to those expressing ARID1A_S363A. Differences between wild type and ARID1A-defficient cells were modest, as has previously been reported, and so were those between cells expressing induced wild type or mutant ARID1A. Gene set enrichment analysis (GSEA) revealed that ARID1A-deficient cells showed reduced expression of genes involved in chromatin organisation, replication, DNA repair, translation and peptide presentation compared to the parental (Fig. 7 B, Table S5 ), as has been reported previously ( 11 ). Notably, cells expressing ARID1A_S363A did not display these molecular phenotypes when compared to cells expressing WT ARID1A. They did however show increased expression of genes involved in oxidative phosphorylation, TNF signalling and interferon response, all of which are hallmarks of loss of ARID1A function ( 57 – 60 ). Strikingly, they also showed reduced expression of genes involved in microtubule-related processes, such as microtubule binding, mitotic spindle organisation, microtubule motor activity and kinesin complex (Fig. 7 B, C; Table S5 ). To assess whether these effects translate on the protein level, we performed proteome profiling using quantitative MS. As expected, ARID1A KO cells displayed lower abundance of other BAF complex subunits at the protein level compared to the WT parental lines (Fig. 7 D) as previously shown ( 12 ), but this was not the case for EmGFP-ARID1A_S363A cells compared to cells expressing emGFP-ARID1A_WT, confirming that the S363A mutation does not affect the BAF complex assembly or stability. GSEA of proteomics data recapitulated the microtubule-based defects observed in the transcriptomics analysis (Fig. 7 C, D, Table S6 ). This is striking given that transcriptomic and proteomics changes elicited by loss of key SWI/SNF subunits are generally not well correlated ( 11 , 12 , 61 ). Cells expressing mutated ARID1A showed reduced abundance of proteins involved in microtubule organisation and mitotic events such as mitotic spindle assembly, chromosome alignment and nuclear membrane reassembly after mitosis. There was considerable overlap of downregulated genes/proteins annotated with “microtubule cytoskeleton” (50%) or “mitotic spindle organization” (75%) between RNA-seq and proteomics datasets. These results show that phosphorylation of S363 is important for specific aspects of ARID1A function related to the microtubule cytoskeleton. The reduced abundance of proteins involved in microtubule dynamics in ARID1A_S363A mutant cells suggested that they may have defects in mitosis. To explore this, we induced expression of wild type or mutant ARID1A in ARID1A KO cells and monitored cell division using live-cell imaging over a period of 72 h (Fig. 7 E). We observed that cells expressing ARID1A_S363A rarely underwent mitosis over the 72 h, whilst cells expressing WT ARID1A divided at similar rates to the untransfected KO cells (Fig. 7 F). In summary, our results suggest that ARID1A Ser363 phosphorylation is important for normal mitosis and cell division. Discussion Here we have used a peptide array interactomics strategy (PRISMA) to uncover novel interaction partners and binding sites of the SWI/SNF scaffold subunit ARID1A. This relatively new approach detects biophysical protein binding and has proven very useful to identify interactions that are mediated by SLiMs ( 32 , 62 – 64 ), but has only been applied to a small number of proteins to date. Our study significantly contributes to the understanding of molecular mechanisms of ARID1A regulation and function, by exposing novel ARID1A interacting partners and interaction sites, and providing a novel resource that can be used by the SWI/SNF community to further explore ARID1A biology. Integration of SLiM annotations on the target protein with the PRISMA screen data can significantly contribute to the detection of confident interaction candidates for validation and follow-up experiments, enhancing the confidence in potential true interactions and providing insight into the regulation, molecular mechanism and function of protein interactions. Matching of ARID1A SLiMs to proteins containing the partnering globular domains highlighted SIN3A as a direct interactor through its PAH domain(s). Although ARID1A has been shown to be required for the recruitment of SIN3A to chromatin to regulate the expression of cell cycle and cytokine genes and TERT ( 38 , 43 , 65 ), and interactions of SIN3A with the BAF complex have been reported ( 41 , 42 , 66 ), only one study had previously demonstrated an interaction between SIN3A and ARID1A ( 38 ). Our data shows that the interaction between ARID1A and SIN3A is direct and mediated through the Sin3 SLiM. The high concentration of peptide on the PRISMA arrays facilitates the detection of weak and transient interactions that are challenging to capture in traditional affinity purification approaches employing endogenous proteins. This may have been instrumental in the identification of novel interactor TOX4. The detection of the ubiquitin-containing ARID1A laddered pattern in the TOX4 immunoprecipitate suggests that TOX4 preferentially binds to (poly)ubiquitinated ARID1A in vivo , and this may explain why this interaction is not readily detected in traditional ARID1A IP-MS interactomes. Interestingly, the TOX4-binding ARID1A peptide encompassing aas 1691–1710 contains a predicted ABBA motif, a SLiM that has been shown to act as a degradation signal for key mitotic regulators and a signal for kinetochore recruitment ( 67 ), raising the question of whether the interaction with TOX4 plays a role in ARID1A degradation. PP1 regulatory subunit TOX4 has been shown to have a role in cell fate reprogramming and differentiation ( 68 , 69 ), and was reported to bind to DNA adducts generated by cisplatin ( 70 ). Seeing as it plays a role in fundamental processes including genome stability ( 48 , 71 ), it is no surprise that it has been shown to promote tumourigenesis and so has interesting therapeutic potential. Both PLK1 and GSK3 form reciprocal regulatory loops with PP1 ( 72 – 74 ), and our data suggests that the interplay of PP1 and either or both kinases might regulate some aspect of ARID1A function. Indeed, PLK1 inhibition was recently shown to be synthetic lethal with ARID1A deficiency ( 75 ). Similarly, this seems to be the case with CDK2 and CCNA2, which have not been previously detected in any ARID1A interactome study. We detect weak binding of the kinase on a peptide containing a CDK2 consensus site that has previously been shown to be phosphorylated in a CDK2-dependent manner, and stronger binding, together with cyclin A, on peptides within the region of intrinsic disorder in ARID1A between the ARID and the C-terminal globular domains. The ARID1A peptide with the strongest CDK2/CCNA2 binding signal does not contain a canonical cyclin A docking motif (RxL). It is, however, reminiscent of a yeast cyclin binding motif PxxPxF ( 76 ). Intriguingly, there is an RxL motif in the C-terminal of ARID1A that overlaps with the SIN3 binding motif, that indeed binds cyclin A in vitro (Fig. 6 C). Our exploration of published cell cycle phosphorylation data ( 55 ) suggests that ARID1A Ser363 is regulated in a cell cycle dependent manner and represents the first report of ARID1A regulation through phosphorylation. The mutation experiments prompted by findings from the PRISMA data indicate that the dynamic behaviour of Ser363 phosphorylation is specifically required for the expression of genes involved in microtubule-based processes, many of which are cell cycle genes. We find that expression of ARID1A_S363A mutant, where S363 cannot be phosphorylated, leads to decreased expression of genes and proteins involved in microtubule organisation and movement. Previous work has shown that ARID1A loss results in reduced expression of cell cycle and mitotic spindle genes at RNA and protein level ( 38 , 77 ), and our data suggests that ARID1A S363 dynamic phosphorylation specifically regulates this transcriptional program. Whilst SWI/SNF controls gene expression, many of the transcriptional changes observed when specific subunits are lost do not often correlate well with changes at the protein level ( 11 , 12 , 61 ). Importantly, our study shows that the regulation of MT-based genes by ARID1A S363 is cascaded down to the protein level and therefore results in a functional phenotypic effect on the cells. With ARID1A (and SWI/SNF) being master regulators of a huge number of genes and many different transcriptional programs, it is not surprising that regulation of these different sets of genes must happen through precise and specific mechanisms. This is not unprecedented for SWI/SNF. For instance, during mitosis SMARCE1 and SMARCB1 remain bound specifically to genes required for the completion of mitotic cell division and lineage-identity genes. By contrast, catalytic subunits SMARCA2 and SMARCA4 have been shown to localise to the nucleus through interphase but are displaced from condensed chromatin during mitosis ( 78 , 79 ). ARID1A has also been shown to be excluded from chromatin during mitosis ( 79 ). The genes annotated to microtubule processes that are affected by the S363A mutation are enriched in SIN3A targets (ENCODE ChIP data from EnrichR), and their expression along the cell cycle is positively correlated to that of CCNA2 and inversely to the phosphorylation of ARID1A S363 (data not shown). This raises the possibility that S363 (de)phosphorylation may regulate SIN3A chromatin recruitment and/or maintenance. Notably, ARID1A S363 phosphorylation is upregulated in breast cancer ( 80 ). Incidentally, this residue is in a stretch of sequence that is missing in ARID1B, an ARID1A paralog with an antagonistic role in cell cycle control ( 38 ). Some well documented ARID1A interactions were not identified by our PRISMA assay. There are many reasons why this can happen, such as interactions being dependent on the presence of PTMs, cell line or context-dependent, or simply of too low abundance to be detected by PRISMA. It is also possible that missing interactions are reliant on globular structures on ARID1A that are not recapitulated by the short peptides used in our array. We also noted that PRISMA binding profiles of known ARID1A interacting proteins that are involved in DNA damage did not agree with published binding regions. This could reflect a dependency on DNA damage signalling, including PTMs on either ARID1A or interacting proteins. Overall, our study has uncovered new biology on ARID1A, provides new avenues for further mechanistic exploration of ARID1A and SWI/SNF function and could inform therapeutic strategies impacting ARID1A function. Furthermore, our work represents a useful resource more broadly to investigate sequence or structural features governing protein interactions. Conclusions We demonstrate direct binding of transcriptional repressor SIN3A to ARID1A through its Sin3 SLiM and recruitment to chromatin. We identify TOX4, CDK2 and CCNA2 as novel ARID1A interacting proteins and map their binding sites. We reveal a previously uncharacterised role for CDK2-dependent phosphorylation of ARID1A Ser 363 in the regulation of expression of microtubule factors and cell division. Our study provides novel mechanistic insight into ARID1A functions and highlights important ramifications for cancer. Our data represents a valuable resource to further study ARID1A and SWI/SNF biology, and more broadly, to investigate sequence and structural features governing protein interactions. Declarations Ethics approval and consent to participate Not applicable Consent for publication Not applicable Availability of data and materials RNA-sequencing datasets generated in this study are available in the GEO repository with accession number GSE313270. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (81) partner repository with dataset identifiers PXD076037 (PRISMA), PXD074557 (chromatin) and PXD074744 (ARID1A_S363A proteome profiling). Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Mercedes Pardo ( [email protected] ). All unique/stable reagents generated in this study are available from the lead contact with a completed materials transfer agreement. Competing interests The authors declare that they have no competing interests. Funding This work was supported by the Wellcome Trust (223745/Z/21/Z and 302297/Z/23/Z), BBSRC (BB/Y004477/1), Cancer Research UK (DRCRPG-June24/100004) and ICR core funding. Author’s contributions MP and JC conceived the project. MP, CM, KAL, FS, LS and ZK performed experiments. MP, CM, and KAL performed data analysis. MW and FY contributed reagents. MP, CA, JAD and JC obtained funding and provided input and supervision. MP prepared the original draft. MP, CM, KAL, CA, JAD and JC contributed to review and editing. All authors read and approved the final manuscript. Acknowledgements We thank the Institute of Cancer Research Flow Cytometry Facility, Insect Cell Culture Facility and Scientific Computing Team for support. Resource availability Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Mercedes Pardo ( [email protected] ). All unique/stable reagents generated in this study are available from the lead contact with a completed materials transfer agreement. RNA-sequencing datasets generated in this study are available in the GEO repository with accession number GSE313270. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (81) partner repository with dataset identifiers PXD076037 (PRISMA), PXD074557 (chromatin) and PXD074744 (ARID1A_S363A proteome profiling). STAR Methods Cell culture hTERT-RPE1 (RPE1, ATCC) were cultured in Dulbecco Modified Minimal Essential Medium (DMEM)/F-12 (Merck) supplemented with 10% FBS (Gibco), 200µM glutamax (Gibco), 0.26% sodium bicarbonate (Gibco), and 1% penicillin/streptomycin (P/S)(Merck). HEK293TN (Systems Biosciences) were cultured in DMEM supplemented with 10% FBS and 1% P/S. All cells were maintained at 37°C in a humified incubator with 5% CO 2 and were regularly tested for mycoplasma contamination. Plasmids Emerald-GFP (emGFP) and emGFP-tagged ARID1A, ARID1A S363A and ARID1A C-terminus (aa. 1009-2285) were amplified by PCR with Q5® High-Fidelity DNA Polymerase (M0491S, NEB) using pLenti-puro-ARID1A plasmid (Addgene #39478) as template and cloned into pCW57-MCS1-2A-MCS2 plasmid (Addgene #71782) through NEBuilder® HiFi DNA Assembly (NEB, E2621S). Plasmids were prepared using the Qiagen Plasmid Plus kit as per the manufacturer’s protocol and stored at -20 o C. Correct nucleotide sequence was confirmed by whole plasmid sequencing (Plasmidsaurus). The lentiviral packaging and envelope plasmids psPAX2 and pMD2.G were gifts from Professor Didier Trono (Addgene plasmids #12259 and #12260). Oligos used are listed in Table S8. Cell line generation For lentivirus production, 10 µg of the target plasmid was transfected into HEK293TN cells along with the lentiviral packaging and envelope plasmids, psPAX2 and pMD.2G, using Lipofectamine 3000 (Thermo) according to the manufacturer’s protocol. Viral particle-containing medium was harvested at 48 hr and 72 hr post transfection, combined, centrifuged for 5 min at 300 x g to remove debris, and filtered through 0.45 µm cellulose acetate filters (Sartorius). The resulting supernatant was concentrated using the Lenti-X Concentrator (Takara Bio) according to the manufacturer’s protocol. For transduction, concentrated viral supernatant was added to RPE1 cells in 6-well plates along with 6 µg/mL polybrene (Merck). Cells were expanded to 10 cm 2 dishes, and 1 µg/mL DOX was added to cells to induce protein expression. After 24 hr, the pool of GFP-positive cells was sorted into 6-well plates by the BD FACSAria III sorter (BD Biosciences). Protein expression in the resulting cell lines was confirmed by Western blot analysis and immunofluorescence microscopy. Compounds Doxycycline hyclate (DOX, Merck) was dissolved in RNase-free H 2 O to a stock concentration of 1 mg/mL. 5-ethynyl-2'-deoxyuridine (EdU, Invitrogen) was dissolved in DMSO (Merck) to a stock concentration of 10 mM. Peptide arrays Cellulose acetate membranes containing immobilised PRISMA peptides spanning the whole length of ARID1A were purchased from JPT Peptide Technologies (Berlin, Germany). Peptide sequences are listed in Table S1. Peptide arrays for the CDK2/CCNA2 validation experiments (Table S2) were prepared in-house. Peptides were synthesised on derivatised cellulose membranes using a Multipep1 instrument (CEM Microwave Technology Ltd.) by cycles of coupling of N(a)-Fmoc-protected aminoacids (Sigma-Aldrich) through carboxylic group activation with diisopropylcarbodiimide (Sigma-Aldrich) and Oxyma pure (CEM), followed by deprotection of alpha-amino groups by piperidine (Sigma-Aldrich), followed by acetylation of N-terminal residues with acetic anhydride (Sigma-Aldrich). All peptides were N-terminally acetylated. After peptide synthesis, side chain deprotection was performed as previously described (82). Full length ARID1A PRISMA The PRISMA screen was carried out a single biological replicate with technical replicate MS measurements. Nuclear isolation was performed by lysing wild type RPE1 cells in an isotonic buffer (10 mM Tris-HCl, 10 mM NaCl, 1.5 mM MgCl2, 0.34 M sucrose containing 1 mM dithiothreitol (DTT) and Halt Protease inhibitors (Thermo) and 0.05% NP-40 as described previously (83). Nuclei were lysed by incubation for 10 mins on ice with PRISMA high salt lysis buffer (50 mM HEPES pH 8.0, 450 mM NaCl, 1 mM EDTA, 10% glycerol, 0.5% NP-40, 1 mM DTT, Halt protease inhibitors) followed by Dounce homogenisation. The nuclear lysate was incubated with universal nuclease (Pierce) for 1 hour on ice and then centrifuged at 13000 rpm for 15 min at 4°C. The cleared nuclear lysate was diluted 1:3 with PRISMA dilution buffer (50 mM HEPES pH 8.0, 1 mM EDTA, 10% glycerol, 0.5% NP-40, 0.3 mM DTT, Halt Protease inhibitors (Thermo). The PRISMA membrane was wetted for 15 mins in PRISMA detergent-free wash buffer (50 mM HEPES pH 8.0, 150 mM NaCl, 1 mM EDTA, 10% glycerol), followed by blocking with yeast tRNA at 1 mg/ml (Life Technologies) in the same buffer. The membrane was then washed with PRISMA low salt lysis buffer (50 mM HEPES pH 8.0, 150 mM NaCl, 1 mM EDTA, 10% glycerol, 0.5% NP-40) 5 times for 5 mins. The membrane was then incubated with 40 mg of RPE1 nuclear extract at ~1.4 mg/ml in the diluted lysis buffer for 1h at 4°C. After removal of the nuclear lysate, the membrane was first washed with PRISMA low salt lysis buffer 3 times for 5 mins, followed by PRISMA detergent-free wash buffer 3 times for 5 mins, and then air dried for 1h. The peptide spots were excised and placed one to a well in a 96-well HTS filter plate (Millipore) containing 200 μl of 5 mM TCEP (Sigma-Aldrich) in 50 mM ammonium bicarbonate. Excised membrane spots were incubated for 15 mins at 37°C with shaking to allow reduction, followed by alkylation with 10 mM iodoacetamide for 30 mins shaking at room temperature in the dark. TCEP and iodoacetamide were removed by spinning the plate at 1000 rpm for 1 min, and spots were washed 3 times with 50 mM ammonium bicarbonate. Samples were digested in the filter plate with 0.5 μg of trypsin (sequencing grade, Roche) and 0.25 μg of LysC (Wako Chemicals) per spot/well for 2 hours at 37°C, followed by overnight incubation at room temperature. Peptides were collected by centrifugation onto a fresh plate, and the filter plate stack was flushed twice with 50 μl of 100% HPLC methanol and the flowthrough pooled with the first peptide collection. The peptides were dried and then resuspended in 0.5% formic acid ready for MS analysis. PRISMA mass spectrometry analysis Peptides were analysed with online nanoLC-MS/MS on an Orbitrap Fusion Tribrid mass spectrometer coupled with an Ultimate 3000 RSLCnano System. Samples were first loaded and desalted on a nanotrap (100 µm id x 2 cm) (PepMap C18, 5 µm, 100A) at 10 µl/min with 0.1% formic acid for 10 min and then separated on an analytical column (75 µm id x 50 cm) (PepMap C18, 5 µm, 100A) over a 60 min linear gradient of 4-32% acetonitrile (ACN):0.1% formic acid at 300 nL/min. This was followed by a 5 min column wash with 85% ACN:0.1% formic acid on the analytical column and a 10 min trap-column wash of a cleaning mixture containing 25% ACN:25% Isopropanol:25% methanol: 25% water. The total cycle time was 90 min. The Orbitrap Fusion was operated in standard data-dependent acquisition. For peptide analysis, the precursors were fragmented in HCD (higher collision dissociation) cell at 32% collision energy with a quadrupole isolation width of 0.7 Th (Thomson unit), detected in the ion trap at a rapid scan rate, AGC of 4 × 10e5, and maximum injection time of 50 ms. Targeted precursors were dynamically excluded for further isolation and activation for 45 secs with 7 ppm mass tolerance. Peptide fragments were detected in the ion trap with a rapid scan rate setting. Technical replicate MS measurements were acquired for 50% of peptide spots. PRISMA-MS data analysis The PRISMA raw MS data were analysed with MaxQuant software v. 2.0.3.0 in one batch and searched against the human Uniprot reference database (v. October 2022). A maximum of 2 missed cleavages were allowed, variable modifications were set to methionine oxidation, pyro-glutamic acid conversion of asparagine and glutamine, and N-terminal protein acetylation, and cysteine carbamidomethylation was set as fixed modification. The label free quantification (LFQ) and match between runs options were activated. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (81) partner repository with the dataset identifier PXD076037. Further data processing was carried out using Perseus and Excel. Technical replicate measurements were integrated by averaging the LFQ intensity where two values were available, and the single value was used if only one was measured. Only proteins that were identified by two or more peptides were taken forward. Intensity values for the 228 peptides in each of the data rows were filtered according to the outlier filtering criterion, where the values were replaced by 0 if they were smaller than the 75 th percentile of the intensity value distribution of the protein (32). For visualisation and clustering, each data row was normalised between 0 and 1. Proteins that showed a consecutive binding pattern across a minimum of two matrix peptides were marked as making up the “core interaction dataset”. Peptide properties were calculated using the Peptides R package v.2.4.4 (84). Annotation of GO terms was carried out with Perseus v.2.0.7.0 using an annotation file generated from Uniprot in October 2022. GO term and protein domain enrichment was carried out using the DAVID functional annotation tool (85, 86). Venn diagrams were created with Venny 2.1 (bioinfogp.cnb.csic.es/tools/venny/). Preparation of cell lysates For western blotting, cell pellets were resuspended in the appropriate volume of lysis buffer (10% glycerol, 50mM Tris-HCl (pH 7.4), 0.5% NP-40, 150mM NaCl) containing 0.25U/µL benzonase nuclease (Merck), 1x cOmplete™ EDTA-free protease inhibitor cocktail (Roche), and 1x PhosSTOP™ phosphatase inhibitor cocktail (Roche). Cells were lysed on ice for 45 minutes followed by centrifugation for 15 minutes at 16,000 x g at 4 o C. The resulting supernatant containing whole cell protein extracts was collected. Insect cell pellets were lysed in 50 mM Tris-HCl pH 8, 150 mM NaCl, 0.1% NP40, 1mM EDTA, 1x Halt protease and phosphatase inhibitor (Pierce) and 1 mM DTT on ice for 10 min, homogenised using a Dounce homogeniser and centrifugation for 15 minutes at 16,000 x g at 4 o C. For TOX4 immunoprecipitation, cell lysates were prepared as previously described (87). Immunoprecipitation Relevant antibodies and non-relevant IgG control (normal)antibodies (1–10 μg) were coupled to 25-50 µl of Protein G Dynabeads (Life Technologies) following manufacturer’s protocols. Antibodies are listed in Table S7. Antibody-coupled beads were incubated with whole cell or nuclear lysates for 1–2 h at 4 °C, lysates were removed, and beads were washed with IPP150 buffer (150 mM NaCl, 10 mM Tris-HCl, 0.1% NP-40). Immunoprecipitated proteins were eluted by incubating the beads in 1x LDS sample loading buffer (Life Technologies) containing 50 mM DTT at 70 °C for 10 min. The cell lysate (input) and eluted samples were analysed by western blotting. Western blotting and Far Western on peptide arrays Protein samples were separated in NuPAGE 4-12% BisTris or 3-8% Tris Acetate gels using MOPS or Tris Acetate running buffer (Life Technologies) respectively and transferred to nitrocellulose membranes (Cytiva). Blocking was performed with 5% milk in Tris Buffered Saline-0.1% Tween (TTBS) for 1 h or 10% milk in TTBS for 10 mins. Membranes were incubated in the appropriate primary antibody diluted in 5% milk blocking buffer at 4°C overnight, washed 3 times with TTBS buffer, and incubated in the appropriate horseradish peroxidase (HRP)-conjugated secondary antibody diluted in blocking buffer for 1 hour at room temperature. Blots were developed with ECL Prime, ECL Ultra (Cytiva) or Immobilon Forte (Merck) and chemiluminescence signal acquired on an iBright CL750 imager (Life Technologies). For far western experiments, membranes containing peptide spots were first incubated in TTBS for 15 min, then blocked with 5% milk in TTBS for 1 h. Cyclin A (Bio-techne) at 5 ug/ml or CDK2/CCNA3[174-432] at 10 ug/ml in blocking solution were added to the membrane and incubated for 4 h at room temperature. Primary and secondary antibody incubations were performed at the corresponding dilutions in 5% milk in TTBS. Blots were developed with ECL Prime (Cytiva) and chemiluminescence signal acquired on an iBright used CL750 imager (Life Technologies). Antibodies are listed in Table S7. Expression of proteins in insect cells Codon-optimised human SIN3A (residues 449-1086) fused to StrepII tag and ARID1A-V5 (from Addgene plasmid #39311) cDNAs for insect cell expression were synthesised and subcloned into the pFastBac1 vector (GenScript). The SIN3A and ARID1A constructs were transformed into DH10MultiBac cells for bacmid generation, and bacmid DNAs were isolated by isopropanol precipitation. Bacmid DNAs were transfected into Sf9 ( Spodoptera frugiperda ) cells using CellFectinII (Thermo), and baculovirus stocks were amplified through two consecutive rounds of infection in Sf9 cells. For protein expression, Sf9 cells at a density of 1.5 x 10e6 cells/mL in Sf-900 III medium supplemented with penicillin/streptomycin were infected with the amplified virus at 1:30 (v/v). After 72 h incubation at 27°C with shaking at 130 RPM, cells were harvested and flash-frozen in liquid nitrogen for long term storage at -80°C until processing. Purification of CDK2 and cyclin A3 Both CDK2 (full length) and cyclin A3 (residues 174–432) were induced and expressed individually with 1 mM IPTG in BL21(DE3) Star cells at 18 °C overnight. Pellets containing GST-tagged CDK2 and His-tagged cyclin A3 were combined (two litres of Cdk2 cell culture and 2 litres of cyclin A3 cell culture) and resuspended in lysis buffer (50 mM Hepes pH8.0, 150 mM NaCl, 2 mM MgCl 2 , 5 % glycerol, 0.5 mM TCEP) supplemented with 5 units per ml of benzonase and cOmplete™ EDTA-free protease inhibitors (Roche). After sonication of the cell suspension, the cell lysate was centrifuged at 21,000 rpm. for 1 h at 4 °C and the supernatant was loaded onto pre-equilibrated 5 ml GSTrap™ HP column (Merck). The column was washed with lysis buffer and then the proteins were eluted out with elution buffer (20 mM reduced glutathione added into the lysis buffer). The eluted proteins were buffer exchanged into lysis buffer to remove reduced glutathione and GST-tag of CDK2 was then cleaved off with 3C PreScission protease overnight at 4 °C. After 3C protease cleavage, the sample was loaded onto 5 ml GSTrap™ HP column again, the flow-through was collected, concentrated and loaded onto pre-equilibrated size exclusion column (Superdex75 16/60 column, GE Healthcare). The protein complex was eluted out with gel filtration buffer (20 mM Hepes pH 8.0, 150 mM NaCl, 0.5 mM TCEP) and concentrated to final concentration 1.4 mg/ml. The purified protein was flash frozen and stored in -80 °C. Chromatin extraction Cell pellets were resuspended in isotonic nuclear extraction buffer for chromatin extraction (15 mM Tris-HCL pH 7.5, 60 mM KCl ,15 mM NaCl, 5 mM MgCl2, 1 mM CaCl2, 250 mM Sucrose, 1 mM DTT, Halt Protease inhibitors) and 0.3% NP-40 was added dropwise to final concertation of 0.03% (Sigma-Aldrich), mixed gently and incubated on ice for 5 mins. Nuclei were pelleted at 400-600 rcf for 5 mins at 4°C. Nuclei were washed again with nuclear isolation buffer for chromatin extractions. Small fraction was verified to contain nuclei by light microscopy. Isolated nuclei were resuspended in hypotonic chromatin isolation buffer (3 mM EDTA, 0.2 mM ethylene glycol tetraacetic acid, 1 mM DTT, Halt Protease inhibitors (Thermo) and incubated on ice for 1 h, vortexing every 5-10 mins. Insoluble chromatin fraction was collected by centrifugation for 4 min at 1,700 rcf. Fraction was washed 2 times in hypotonic chromatin isolation buffer. Chromatin fractions were frozen on dry ice and stored at -20°C. Samples were re-suspended in the chromatin preparation buffer (100 mM Triethylamonium bicarbonate (TEAB) (Sigma-Aldrich), 1% sodium deoxycholate (SDC), 10% isopropanol, 50 mM NaCl, Halt Protease and Phosphatase inhibitor (Thermo) supplemented with and Benzonaze (Pierce). The chromatin pellet was disrupted with brief probe sonication and the samples were incubated on ice for 30 min. 30 μg of each sample was labelled TMTpro™ 16plex Label Reagent Set (Thermo) according to manufactures instructions. Chromatin profiling mass spectrometry Offline peptide fractionation was based on high pH reverse-phase (RP) chromatography using the Waters XBridge C18 column (2.1 mm × 150 mm, 3.5 μm) on a Dionex UltiMate 3000 HPLC system at a flow rate of 0.2 ml/min. Mobile phase A was 0.1% (v/v) ammonium hydroxide, and mobile phase B was acetonitrile, 0.1% (v/v) ammonium hydroxide. Pooled TMT-peptides were resuspended in 200 μL of buffer A, centrifuged at 14 000 rpm for 5 min, and the supernatant was injected for fractionation with the following gradient: isocratic for 5 min at 5% phase B, gradient for 40 min to 35% phase B, gradient to 80% phase B in 5 min, isocratic for 5 min, and re-equilibrated to 5% phase B. The peptides were then resuspended with 0.1% formic acid and finally pooled into four fractions by combining the first and the last fractions. LC-MS analysis was performed on the Dionex UltiMate 3000 UHPLC system coupled with the LTQ Orbitrap Lumos mass spectrometer (Thermo). Samples were analysed with the EASY-Spray C18 capillary column (75 μm × 50 cm, 2 μm) at 50 °C. Mobile phase A was 0.1% formic acid and mobile phase B was 80% acetonitrile, 0.1% formic acid. The gradient separation method was as follows: 150 min gradient up to 38% B, for 10 min up to 95% B, for 5 min isocratic at 95% B, re-equilibration to 5% B in 10 min, for 10 min isocratic at 5% B. MS1 scans were acquired at a resolution of 120,000 over an m/z range of 375–1500 in the Orbitrap, with an AGC target of 4 × 10⁵ and a maximum injection time of 50 ms. Precursors were selected using Top Speed mode with a 3 s cycle time and isolated with a quadrupole isolation width of 0.7 Th. Collision energy for fragmentation via collision induced dissociation (CID) was set at 35%, with AGC target of 1 × 10⁴ and maximum injection time of 50 ms. MS3 spectra for TMT-based quantitation were acquired using synchronous precursor selection (SPS) of the top seven CID fragment ions. SPS-selected ions were fragmented by HCD with a collision energy of 35%. MS3 spectra were acquired in the Orbitrap over an m/z range of 100–500 at a resolution of 45,000, with an AGC target of 1 × 10⁵ and maximum injection time of 105 ms. Dynamic exclusion was activated for 45 s with a mass tolerance of 7 ppm. Chromatin profiling data analysis MS raw data from IPs was analysed using Proteome Discoverer 2.4 (Thermo). Sequest search engine was used for protein identification and quantification followed by Percolator. Peptides were filtered at q-value <1% based on a decoy database search. The searches were performed against the full UniProt Human database (reference entries, January 2022) together with the cRAP contaminant database (ftp://ftp.thegpm.org/fasta/cRAP). Precursor mass tolerance was set to 20 ppm. Fragment mass tolerance was set to 0.5 Da. TMTpro (K and peptide N-terminus) and carbamidomethylation (C) were set as static modifications, while oxidation (M) and deamidation (NQ) were set as dynamic modifications. Protein abundances were normalised by multiplying each value by the maximum median of abundances across all channels and then dividing by the median of abundances of the corresponding channel. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (81) partner repository with the dataset identifier PXD074557. Incucyte proliferation assay and cell fate profiling Cells were seeded in triplicate into 96-well plates at a density of 1,000 cells/well, and allowed to attach and grow for 24 hours. When indicated, 1 µg/mL DOX was then added to cells, and plates were transferred to the Incucycte SX5 (Sartorius) maintained at 37°C in a humidified 5% CO2 atmosphere. For tracking cell growth, phase contrast and fluorescence images (4 images per well) were collected every 4 hours up to a total of 120 hours. The average of three technical replicates was used, and % confluence (% of area covered by cells) was used as a readout of proliferation. For cell fate profiling, phase contrast and fluorescence images (4 images per well) were collected every 10 minutes up to a total of 72 hours. Image sequences were exported, and individual cells were analysed manually to generate cell fate profiles. Sulforhodamine B (SRB) proliferation assay Cells were seeded in triplicate into 96-well plates at a density of 1,000 cells/well, and allowed to attach and grow for 24 hours. When indicated, 1 µg/mL DOX was then added to cells and cells were left to incubate for 6 days in total. After this, cells were fixed by the addition of 100 µl 10% TCA (Merck) and incubated at 4°C for at least 24 hours. Plates were then washed five times with water and allowed to completely dry before staining with 0.57% SRB powder (Merck) dissolved in 1% acetic acid and incubated for 1 hour room temperature. Unbound dye was removed by washing with 1% acetic acid, and the plate was allowed to dry completely. Protein-bound dye was solubilised in 10 mM Tris-HCl pH 9.5 for 30 minutes at room temperature with gentle rocking. Absorbance at 565 nm was quantified using the SpectraMax Microplate Absorbance Reader (PerkinElmer). The average of three technical replicates was used, and proliferation was defined as a percentage of the absorbance measured in treated wells compared with control wells. RNA-sequencing For RNA-sequencing, cells were treated with 1 µg/mL DOX for 24 hours before harvesting. Pellets were harvested by scraping cells in ice-cold PBS following by centrifugation at 7,500 x g for 10 minutes. Three independent biological replicates were harvested for all cell lines. Pellets were resuspended in 500 µL TRIzol reagent and total RNA was extracted using a Direct-zol RNA miniprep kit (Zymo Research) according to the manufacturer’s protocol. RNA concentration and quality was confirmed with the High Sensitivity RNA ScreenTape (Agilent), using the TapeStation 4150 System (Agilent). Library preparation and sequencing was performed by Novogene Corporation Ltd. Novogene NGS Stranded RNA Library Prep Set was used to generate 250-300 bp insert strand specific libraries, and ribosomal RNA was removed using TruSeq Stranded Total RNA Library Prep. 50 million 150 bp paired-end reads were sequenced on an Illumina NovaSeq 6000. RNA-sequencing analysis Fastq reads were checked using FastQC (v0.11.9) (88) and trimmed using TrimGalore (v0.6.6) (89). Residual ribosomal RNA reads were removed using Ribodetector with -e norRNA setting (v0.2.7) and strandedness was detected using RSeQC (v4.0.0). Reads were aligned to the human hg38.14 genome (GenBank GCA_000001405.29) using STAR alignment software (v2.7.6a) and reads mapping to genes were quantified using HTSeqCount (v0.12.4). Differential analysis of gene expression was calculated in R using DESeq2 (v1.38.3). Whole cell proteomic profiling Cell pellets were lysed in buffer containing 1% (w/v) sodium deoxycholate (SDC), 100 mM triethylammonium bicarbonate (TEAB), 10% (v/v) isopropanol, and 50 mM NaCl, freshly supplemented with 5 mM TCEP (Thermo, Bond-Breaker), 10 mM iodoacetamide (IAA), universal nuclease (1:2000, v/v; Pierce, #88700), and Halt protease and phosphatase inhibitor cocktail (1×; Thermo, #78442). Lysis was assisted by 5 min of bath sonication. Protein concentration was determined using the Quick Start Bradford protein assay (Bio-Rad). Aliquots containing 30 µg of total protein were digested overnight at room temperature with trypsin (Pierce) at a 1:20 (enzyme:substrate) ratio. Peptides were labelled with TMTpro reagents (Thermo) by adding 5 µL of reagent to 15 µL of peptide solution. Following labelling, samples were acidified to a final concentration of 2% formic acid, and precipitated SDC was removed by centrifugation. The combined TMT-labelled peptide mixture was fractionated by high-pH reversed-phase chromatography using an XBridge C18 column (2.1 × 150 mm, 3.5 µm; Waters) on an UltiMate 3000 HPLC system over a 35-min gradient at 1% slope. Mobile phase A consisted of 0.1% (v/v) ammonium hydroxide, and mobile phase B consisted of 0.1% (v/v) ammonium hydroxide in acetonitrile. LC–MS analysis was performed on a Vanquish Neo UHPLC system coupled to an Orbitrap Ascend mass spectrometer (Thermo). Peptides were separated on a 25 cm capillary column (nanoEase MZ PST BEH130 C18, 1.7 µm, 75 µm × 250 mm; Waters) using a 110-min gradient from 5–35% mobile phase B (80% acetonitrile, 0.1% formic acid). Samples were first loaded onto a PepMap 100 C18 trapping column (5 µm, 0.3 × 5 mm, 1500 bar) before separation on the analytical column, which was coupled to a stainless-steel emitter via a Nanospray Flex ion source. MS1 spectra were acquired in the Orbitrap at a resolution of 120,000. Precursors were selected for HCD fragmentation in top-speed mode (3 s cycle time) with a normalized collision energy of 32% and detected in the ion trap using turbo scan rate. MS3 spectra were triggered by real-time search (RTS) against a FASTA database containing reviewed UniProt Homo sapiens canonical and isoform sequences, using multi-notch isolation (10 notches) and HCD fragmentation at 65% normalized collision energy, with detection in the Orbitrap at 90,000 resolution. Dynamic exclusion was enabled for 45 s with a mass tolerance of 10 ppm, RTS close-out was set to a maximum of four peptides per protein, and targeted precursors were excluded from further activation. Static modifications used for RTS included TMTpro16plex on peptide N-termini and lysine residues and carbamidomethylation of cysteine. Variable modifications included oxidation of methionine and deamidation of asparagine and glutamine, with a maximum of one missed cleavage and up to two variable modifications per peptide. Raw data were searched using the Sequest HT and Comet nodes implemented in Proteome Discoverer version 3.0 (Thermo) against a reviewed UniProt Homo sapiens FASTA database. Precursor mass tolerance was set to 20 ppm, and fragment ion mass tolerance was set to 0.5 Da for Sequest HT and 1 Da for Comet, allowing up to two missed tryptic cleavages. TMTpro modifications on peptide N-termini and lysine residues, as well as carbamidomethylation of cysteine, were specified as static modifications, while oxidation of methionine and deamidation of asparagine/glutamine were specified as dynamic modifications. Peptide-spectrum matches were filtered using Percolator with a peptide-level false discovery rate (FDR) of 1% based on a target–decoy strategy. Only unique peptides were used for protein quantification, considering protein groups for peptide uniqueness, and peptides with an average reporter ion signal-to-noise ratio greater than 3 were included in quantitative analysis. Data are available via ProteomeXchange with identifier PXD074744. Bioinformatics analysis of transcriptomics and proteomics data Peptide properties were calculated using the Peptides R package v.2.4.4 (84). GO term and protein domain enrichment analysis of the ARID1A PRISMA core dataset was carried out using the DAVID functional annotation tool (86). Gene set enrichment analyses for whole cell proteomic profiling and RNA-seq data were performed using the 1D annotation enrichment algorithm (90) within the Perseus software platform (91). GSEA plots were generated in R using the fgsea package (92). References Harrod A, Lane KA, Downs JA. The role of the SWI/SNF chromatin remodelling complex in the response to DNA double strand breaks. DNA Repair (Amst). 2020;93:102919. Schick S, Grosche S, Kohl KE, Drpic D, Jaeger MG, Marella NC, et al. Acute BAF perturbation causes immediate changes in chromatin accessibility. Nat Genet. 2021;53:269–78. Centore RC, Sandoval GJ, Soares LMM, Kadoch C, Chan HM. Mammalian SWI/SNF Chromatin Remodeling Complexes: Emerging Mechanisms and Therapeutic Strategies. Trends Genet. 2020;36:936–50. Blumli S, Wiechens N, Wu MY, Singh V, Gierlinski M, Schweikert G, et al. Acute depletion of the ARID1A subunit of SWI/SNF complexes reveals distinct pathways for activation and repression of transcription. Cell Rep. 2021;37:109943. Dykhuizen EC, Hargreaves DC, Miller EL, Cui K, Korshunov A, Kool M, et al. BAF complexes facilitate decatenation of DNA by topoisomerase IIalpha. Nature. 2013;497:624–7. Simões-Sousa S, Amin N, Lane KA, Harrod A, Pedersen M, Pardo M, et al. ARID1A stabilizes non-homologous end joining factors at DNA breaks induced by the G4 ligand pyridostatin. Cell Rep. 2025;44:116277. Varga J, Kube M, Luck K, Schick S. The BAF chromatin remodeling complexes: structure, function, and synthetic lethalities. Biochem Soc Trans. 2021;49:1489–503. Mashtalir N, D'Avino AR, Michel BC, Luo J, Pan J, Otto JE, et al. Modular Organization and Assembly of SWI/SNF Family Chromatin Remodeling Complexes. Cell. 2018;175:1272-88 e20. He S, Wu Z, Tian Y, Yu Z, Yu J, Wang X, et al. Structure of nucleosome-bound human BAF complex. Science. 2020;367:875–81. Schick S, Rendeiro AF, Runggatscher K, Ringler A, Boidol B, Hinkel M, et al. Systematic characterization of BAF mutations provides insights into intracomplex synthetic lethalities in human cancers. Nat Genet. 2019;51:1399–410. Santamarina JB, Astier C, Garrido M, Daage LC, Roumeliotis TI, Anthony E, et al. Integrated multiomic profiling reveals SWI/SNF subunit-specific pathway alterations and targetable vulnerabilities. bioRxiv 2024; doi: 10.1101/2024.07.16.603530 . Roumeliotis TI, Williams SP, Goncalves E, Alsinet C, Del Castillo Velasco-Herrera M, Aben N, et al. Genomic Determinants of Protein Abundance Variation in Colorectal Cancer Cells. Cell Rep. 2017;20:2201–14. Kanno S-i, Kobayashi T, Watanabe R, Kurimasa A, Tanaka K, Yasui A, et al. Armadillo domain of ARID1A directly interacts with DNA-PKcs to couple chromatin remodeling with nonhomologous end joining (NHEJ) pathway. Nucleic Acids Research. 2025;53:gkaf150. Ulicna L, Kimmey SC, Weber CM, Allard GM, Wang A, Bui NQ, et al. The Interaction of SWI/SNF with the Ribosome Regulates Translation and Confers Sensitivity to Translation Pathway Inhibitors in Cancers with Complex Perturbations. Cancer Res. 2022;82:2829–37. Chang L, Azzolin L, Di Biagio D, Zanconato F, Battilana G, Lucon Xiccato R, et al. The SWI/SNF complex is a mechanoregulated inhibitor of YAP and TAZ. Nature. 2018;563:265–9. Wang Y, Chen X, Qiao X, Xie Y, Guo D, Li B, et al. Chromatin Remodelling Molecule ARID1A Determines Metastatic Heterogeneity in Triple-Negative Breast Cancer by Competitively Binding to YAP. Cancers (Basel). 2023;15:2447. Fontana B, Gallerani G, Salamon I, Pace I, Roncarati R, Ferracin M. ARID1A in cancer: Friend or foe? Frontiers in Oncology. 2023;13:1136248. Hodges C, Kirkland JG, Crabtree GR. The Many Roles of BAF (mSWI/SNF) and PBAF Complexes in Cancer. Cold Spring Harb Perspect Med. 2016;6:a026930. Bögershausen N, Wollnik B. Mutational Landscapes and Phenotypic Spectrum of SWI/SNF-Related Intellectual Disability Disorders. Frontiers in Molecular Neuroscience. 2018;11. Kadoch C, Hargreaves DC, Hodges C, Elias L, Ho L, Ranish J, et al. Proteomic and bioinformatic analysis of mammalian SWI/SNF complexes identifies extensive roles in human malignancy. Nat Genet. 2013;45:592–601. Li J, Wang W, Zhang Y, Cieslik M, Guo J, Tan M, et al. Epigenetic driver mutations in ARID1A shape cancer immune phenotype and immunotherapy. J Clin Invest. 2020;130:2712–26. Alfert A, Moreno N, Kerl K. The BAF complex in development and disease. Epigenetics Chromatin. 2019;12:19. El Hadidy N, Uversky VN. Intrinsic Disorder of the BAF Complex: Roles in Chromatin Remodeling and Disease Development. Int J Mol Sci. 2019;20:5260. Tompa P, Davey NE, Gibson TJ, Babu MM. A million peptide motifs for the molecular biologist. Mol Cell. 2014;55:161–9. Bah A, Forman-Kay JD. Modulation of Intrinsically Disordered Protein Function by Post-translational Modifications. J Biol Chem. 2016;291:6696–705. Davis RB, Kaur T, Moosa MM, Banerjee PR. FUS oncofusion protein condensates recruit mSWI/SNF chromatin remodeler via heterotypic interactions between prion-like domains. Protein Sci. 2021;30:1454–66. Kim YR, Joo J, Lee HJ, Kim C, Park JC, Yu YS, et al. Prion-like domain mediated phase separation of ARID1A promotes oncogenic potential of Ewing's sarcoma. Nat Commun. 2024;15:6569. Patil A, Strom AR, Paulo JA, Collings CK, Ruff KM, Shinn MK, et al. A disordered region controls cBAF activity via condensation and partner recruitment. Cell. 2023;186:4936-55.e26. Chan YS, Gao Q, Robinson SA, Wang W, Filandrova R, Weinhold LM, et al. β-catenin functions as a molecular adapter for disordered cBAF interactions. Mol Cell. 2025;85:3041-56.e9. Van Roey K, Uyar B, Weatheritt RJ, Dinkel H, Seiler M, Budd A, et al. Short linear motifs: ubiquitous and functionally diverse protein interaction modules directing cell regulation. Chem Rev. 2014;114:6733–78. Meyer K, Selbach M. Peptide-based Interaction Proteomics. Mol Cell Proteomics. 2020;19:1070–5. Dittmar G, Hernandez DP, Kowenz-Leutz E, Kirchner M, Kahlert G, Wesolowski R, et al. PRISMA: Protein Interaction Screen on Peptide Matrix Reveals Interaction Footprints and Modifications- Dependent Interactome of Intrinsically Disordered C/EBPbeta. iScience. 2019;13:351–70. Mészáros B, Erdos G, Dosztányi Z. IUPred2A: context-dependent prediction of protein disorder as a function of redox state and protein binding. Nucleic Acids Res. 2018;46:W329-w337. Wang Z, Chen K, Jia Y, Chuang JC, Sun X, Lin YH, et al. Dual ARID1A/ARID1B loss leads to rapid carcinogenesis and disruptive redistribution of BAF complexes. Nat Cancer. 2020;1:909–22. Luo Q, Wu X, Chang W, Zhao P, Nan Y, Zhu X, et al. ARID1A prevents squamous cell carcinoma initiation and chemoresistance by antagonizing pRb/E2F1/c-Myc-mediated cancer stemness. Cell Death Differ. 2020;27:1981–97. Linden M, Thomsen C, Grundevik P, Jonasson E, Andersson D, Runnberg R, et al. FET family fusion oncoproteins target the SWI/SNF chromatin remodeling complex. EMBO Rep. 2019;20:e45766. Kumar M, Michael S, Alvarado-Valverde J, Mészáros B, Sámano-Sánchez H, Zeke A, et al. The Eukaryotic Linear Motif resource: 2022 release. Nucleic Acids Res. 2022;50:D497-d508. Nagl NG, Jr., Wang X, Patsialou A, Van Scoy M, Moran E. Distinct mammalian SWI/SNF chromatin remodeling complexes with opposing roles in cell-cycle control. EMBO J. 2007;26:752–63. Kuzmichev A, Zhang Y, Erdjument-Bromage H, Tempst P, Reinberg D. Role of the Sin3-histone deacetylase complex in growth regulation by the candidate tumor suppressor p33(ING1). Mol Cell Biol. 2002;22:835–48. Underhill C, Qutob MS, Yee SP, Torchia J. A novel nuclear receptor corepressor complex, N-CoR, contains components of the mammalian SWI/SNF complex and the corepressor KAP-1. J Biol Chem. 2000;275:40463–70. Sif S, Saurin AJ, Imbalzano AN, Kingston RE. Purification and characterization of mSin3A-containing Brg1 and hBrm chromatin remodeling complexes. Genes Dev. 2001;15:603–18. Pal S, Yun R, Datta A, Lacomis L, Erdjument-Bromage H, Kumar J, et al. mSin3A/histone deacetylase 2- and PRMT5-containing Brg1 complex is involved in transcriptional repression of the Myc target gene cad. Mol Cell Biol. 2003;23:7475–87. Kim M, Lu F, Zhang Y. Loss of HDAC-Mediated Repression and Gain of NF-κB Activation Underlie Cytokine Induction in ARID1A- and PIK3CA-Mutation-Driven Ovarian Cancer. Cell Rep. 2016;17:275–88. Suryo Rahmanto Y, Jung JG, Wu RC, Kobayashi Y, Heaphy CM, Meeker AK, et al. Inactivating ARID1A Tumor Suppressor Enhances TERT Transcription and Maintains Telomere Length in Cancer Cells. J Biol Chem. 2016;291:9690–9. Sahu SC, Swanson KA, Kang RS, Huang K, Brubaker K, Ratcliff K, et al. Conserved Themes in Target Recognition by the PAH1 and PAH2 Domains of the Sin3 Transcriptional Corepressor. Journal of Molecular Biology. 2008;375:1444–56. Guan B, Wang TL, Shih Ie M. ARID1A, a factor that promotes formation of SWI/SNF-mediated chromatin remodeling, is a tumor suppressor in gynecologic cancers. Cancer Res. 2011;71:6718–27. Abe Y, Rozqie R, Matsumura Y, Kawamura T, Nakaki R, Tsurutani Y, et al. JMJD1A is a signal-sensing scaffold that regulates acute chromatin dynamics via SWI/SNF association for thermogenesis. Nat Commun. 2015;6:7052. Liu Z, Wu A, Wu Z, Wang T, Pan Y, Li B, et al. TOX4 facilitates promoter-proximal pausing and C-terminal domain dephosphorylation of RNA polymerase II in human cells. Commun Biol. 2022;5:300. Hornbeck PV, Zhang B, Murray B, Kornhauser JM, Latham V, Skrzypek E. PhosphoSitePlus, 2014: mutations, PTMs and recalibrations. Nucleic Acids Res. 2015;43:D512-20. Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596:583–9. Chi Y, Welcker M, Hizli AA, Posakony JJ, Aebersold R, Clurman BE. Identification of CDK2 substrates in human cell lysates. Genome Biol. 2008;9:R149. Brown NR, Noble MEM, Endicott JA, Johnson LN. The structural basis for specificity of substrate and recruitment peptides for cyclin-dependent kinases. Nature Cell Biology. 1999;1:438–43. Nagarajan S, Rao SV, Sutton J, Cheeseman D, Dunn S, Papachristou EK, et al. ARID1A influences HDAC1/BRD4 activity, intrinsic proliferative capacity and breast cancer treatment response. Nat Genet. 2020;52:187–97. Shen J, Peng Y, Wei L, Zhang W, Yang L, Lan L, et al. ARID1A Deficiency Impairs the DNA Damage Checkpoint and Sensitizes Cells to PARP Inhibitors. Cancer Discov. 2015;5:752–67. Rega C, Tsitsa I, Roumeliotis TI, Krystkowiak I, Portillo M, Yu L, et al. High resolution profiling of cell cycle-dependent protein and phosphorylation abundance changes in non-transformed cells. Nat Commun. 2025;16:2579. Sandhya S, Maulik A, Giri M, Singh M. Domain architecture of BAF250a reveals the ARID and ARM-repeat domains with implication in function and assembly of the BAF remodeling complex. PLoS One. 2018;13:e0205267. Mehrvarz Sarshekeh A, Alshenaifi J, Roszik J, Manyam GC, Advani SM, Katkhuda R, et al. ARID1A Mutation May Define an Immunologically Active Subgroup in Patients with Microsatellite Stable Colorectal Cancer. Clinical Cancer Research. 2021;27:1663–70. Perez JM, Ryu J, Khan H, Shetty M, Parker E, D'Arcy P, et al. Proteomic Analysis of ARID1A-Deficient Ovarian Clear Cell Carcinoma Cells Reveals Differential Mitochondria ETC Subunit Abundances and Targetable Mitochondrial Pathways. Int J Mol Sci. 2025;26:5466. Lissanu Deribe Y, Sun Y, Terranova C, Khan F, Martinez-Ledesma J, Gay J, et al. Mutations in the SWI/SNF complex induce a targetable dependence on oxidative phosphorylation in lung cancer. Nat Med. 2018;24:1047–57. Bakr A, Corte GD, Veselinov O, Kelekçi S, Chen MM, Lin YY, et al. ARID1A regulates DNA repair through chromatin organization and its deficiency triggers DNA damage-mediated anti-tumor immune response. Nucleic Acids Res. 2024;52:5698–719. Lane KA, Harrod A, Wu L, Roumeliotis TI, Feng H, Foo S, et al. PBRM1 directs PBAF to pericentromeres and protects centromere integrity. Nat Commun. 2025;16:1980. Ramberger E, Sapozhnikova V, Kowenz-Leutz E, Zimmermann K, Nicot N, Nazarov PV, et al. PRISMA and BioID disclose a motifs-based interactome of the intrinsically disordered transcription factor C/EBPalpha. iScience. 2021;24:102686. Sandmann C-L, Schulz JF, Ruiz-Orera J, Kirchner M, Ziehm M, Adami E, et al. Evolutionary origins and interactomes of human, young microproteins and small peptides translated from short open reading frames. Molecular Cell. 2023;83:994–1011. e18. Suarez-Artiles L, Breiderhoff T, Girardello R, Gonschior H, Rodius S, Lesur A, et al. Pan-claudin family interactome analysis reveals shared and specific interactions. Cell Rep. 2022;41:111588. Bui C-B, Le HK, Vu DM, Truong K-DD, Nguyen NM, Ho MAN, et al. ARID1A-SIN3A drives retinoic acid-induced neuroblastoma differentiation by transcriptional repression of TERT. Molecular Carcinogenesis. 2019;58:1998–2007. Kemper JK, Kim H, Miao J, Bhalla S, Bae Y. Role of an mSin3A-Swi/Snf chromatin remodeling complex in the feedback repression of bile acid biosynthesis by SHP. Mol Cell Biol. 2004;24:7707–19. Di Fiore B, Davey NE, Hagting A, Izawa D, Mansfeld J, Gibson TJ, et al. The ABBA motif binds APC/C activators and is shared by APC/C substrates and regulators. Dev Cell. 2015;32:358–72. Vanheer L, Song J, De Geest N, Janiszewski A, Talon I, Provenzano C, et al. Tox4 modulates cell fate reprogramming. Journal of Cell Science. 2019;132: jcs232223. Niu H, Wang H. TOX regulates T lymphocytes differentiation and its function in tumor. Frontiers in Immunology. 2023;14:990419. Puch CBMd, Barbier E, Kraut A, Couté Y, Fuchs J, Buhot A, et al. TOX4 and its binding partners recognize DNA adducts generated by platinum anticancer drugs. Archives of Biochemistry and Biophysics. 2011;507:296–303. Lobbardi R, Pinder J, Martinez-Pastor B, Theodorou M, Blackburn JS, Abraham BJ, et al. TOX Regulates Growth, DNA Repair, and Genomic Instability in T-cell Acute Lymphoblastic Leukemia. Cancer Discovery. 2017;7:1336–53. Cordeiro MH, Smith RJ, Saurin AT. Kinetochore phosphatases suppress autonomous Polo-like kinase 1 activity to control the mitotic checkpoint. J Cell Biol. 2020;219:e202002020. Mardin BR, Agircan FG, Lange C, Schiebel E. Plk1 controls the Nek2A-PP1gamma antagonism in centrosome disjunction. Curr Biol. 2011;21:1145–51. Szatmari E, Habas A, Yang P, Zheng J-J, Hagg T, Hetman M. A Positive Feedback Loop between Glycogen Synthase Kinase 3β and Protein Phosphatase 1 after Stimulation of NR2B NMDA Receptors in Forebrain Neurons*. Journal of Biological Chemistry. 2005;280:37526–35. Srinivas US, Tay NSC, Jaynes P, Anbuselvan A, Ramachandran GK, Wardyn JD, et al. PLK1 inhibition selectively induces apoptosis in ARID1A deficient cells through uncoupling of oxygen consumption from ATP production. Oncogene. 2022;41:1986–2002. Ord M, Puss KK, Kivi R, Moll K, Ojala T, Borovko I, et al. Proline-Rich Motifs Control G2-CDK Target Phosphorylation and Priming an Anchoring Protein for Polo Kinase Localization. Cell Rep. 2020;31:107757. Schlösser RM, Krumbach F, Corrales E, Andrieux G, Preisinger C, Liss F, et al. Multidimensional OMICs reveal ARID1A orchestrated control of DNA damage, splicing, and cell cycle in normal-like and malignant urothelial cells. Mol Oncol. 2025;19:3784–805. Muchardt C, Reyes JC, Bourachot B, Leguoy E, Yaniv M. The hbrm and BRG-1 proteins, components of the human SNF/SWI complex, are phosphorylated and excluded from the condensed chromosomes during mitosis. Embo j. 1996;15:3394–402. Zhu Z, Chen X, Guo A, Manzano T, Walsh PJ, Wills KM, et al. Mitotic bookmarking by SWI/SNF subunits. Nature. 2023;618:180–7. Kumar A, Dcunha L, Mahin A, Shivamurthy PB, Subair S, Gopalakrishnan AP, et al. Phosphoproteome landscape of ARID1A and its implications in DNA damage response and breast cancer pathogenesis. Discov Oncol. 2026;17:312. Perez-Riverol Y, Bandla C, Kundu DJ, Kamatchinathan S, Bai J, Hewapathirana S, et al. The PRIDE database at 20 years: 2025 update. Nucleic Acids Res. 2025;53:D543-D53. Roeselová A, Maslen SL, Shivakumaraswamy S, Pellowe GA, Howell S, Joshi D, et al. Mechanism of chaperone coordination during cotranslational protein folding in bacteria. Molecular Cell. 2024;84:2455-71.e8. Shcherbakova L, Pardo M, Roumeliotis T, Choudhary J. Identifying and characterising Thrap3, Bclaf1 and Erh interactions using cross-linking mass spectrometry. Wellcome Open Res. 2021;6:260. Osorio D, Rondón-Villarreal P, Torres R. Peptides: A Package for Data Mining of Antimicrobial Peptides. The R Journal. 2015;7:4–14. Huang da W, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc. 2009;4:44–57. Sherman BT, Hao M, Qiu J, Jiao X, Baseler MW, Lane HC, et al. DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update). Nucleic Acids Res. 2022;50:W216-w21. Pardo M, Lang B, Yu L, Prosser H, Bradley A, Babu MM, et al. An expanded Oct4 interaction network: implications for stem cell biology, development, and disease. Cell Stem Cell. 2010;6:382–95. Andrews S. FastQC: a quality control tool for high throughput sequence data. 2010. https://www.bioinformatics.babraham.ac.uk/projects/fastqc/ Kruegel F. Trim Galore. 2023. https://github.com/FelixKrueger/TrimGalore Cox J, Mann M. 1D and 2D annotation enrichment: a statistical method integrating quantitative proteomics with complementary high-throughput data. BMC Bioinformatics. 2012;13 Suppl 16:S12. Tyanova S, Temu T, Sinitcyn P, Carlson A, Hein MY, Geiger T, et al. The Perseus computational platform for comprehensive analysis of (prote)omics data. Nat Methods. 2016;13:731–40. Korotkevich G, Sukhov V, Budin N, Shpak B, Artyomov MN, Sergushichev A. Fast gene set enrichment analysis. bioRxiv. 2021:060012. Additional Declarations No competing interests reported. Supplementary Files FigS1.pdf FigS2.pdf FigS3.pdf FigS4.pdf TableS1.xlsx TableS3.xlsx TableS4.xlsx TableS5.xlsx TableS6.xlsx TableS8.xlsx TableS7.xlsx TableS2.xlsx Supplementalfiles.docx Cite Share Download PDF Status: Under Review Version 1 posted Reviews received at journal 13 May, 2026 Reviews received at journal 11 May, 2026 Reviewers agreed at journal 06 May, 2026 Reviewers agreed at journal 29 Apr, 2026 Reviewers invited by journal 28 Apr, 2026 Editor assigned by journal 15 Apr, 2026 Submission checks completed at journal 15 Apr, 2026 First submitted to journal 14 Apr, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9415881","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":635770353,"identity":"20e07ac4-ea22-41ca-879c-59d79cb09e1f","order_by":0,"name":"Mercedes Pardo","email":"data:image/png;base64,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","orcid":"","institution":"Institute of Cancer Research","correspondingAuthor":true,"prefix":"","firstName":"Mercedes","middleName":"","lastName":"Pardo","suffix":""},{"id":635770354,"identity":"b71ed1a5-0dd8-404c-af6a-7405949b6525","order_by":1,"name":"Chiara Marcozzi","email":"","orcid":"","institution":"Institute of Cancer Research","correspondingAuthor":false,"prefix":"","firstName":"Chiara","middleName":"","lastName":"Marcozzi","suffix":""},{"id":635770355,"identity":"021e3bc0-ecad-4bb5-9ea8-34b693f7141a","order_by":2,"name":"Karen Lane","email":"","orcid":"","institution":"Institute of Cancer Research","correspondingAuthor":false,"prefix":"","firstName":"Karen","middleName":"","lastName":"Lane","suffix":""},{"id":635770356,"identity":"8272a1a4-0633-4138-b464-7d962190f865","order_by":3,"name":"Fernando Sialana","email":"","orcid":"","institution":"Institute of Cancer Research","correspondingAuthor":false,"prefix":"","firstName":"Fernando","middleName":"","lastName":"Sialana","suffix":""},{"id":635770357,"identity":"cce5b118-bc6e-4be3-bf20-9f3d48cf2936","order_by":4,"name":"Liudmila Shcherbakova","email":"","orcid":"","institution":"Institute of Cancer Research","correspondingAuthor":false,"prefix":"","firstName":"Liudmila","middleName":"","lastName":"Shcherbakova","suffix":""},{"id":635770358,"identity":"ad0c5c45-b623-48ee-bbd9-056160b3316e","order_by":5,"name":"Zuza Kozik","email":"","orcid":"","institution":"Institute of Cancer Research","correspondingAuthor":false,"prefix":"","firstName":"Zuza","middleName":"","lastName":"Kozik","suffix":""},{"id":635770359,"identity":"25b00bca-9812-4728-98f9-5713eb3bc07c","order_by":6,"name":"Mandy Wan","email":"","orcid":"","institution":"Institute of Cancer Research","correspondingAuthor":false,"prefix":"","firstName":"Mandy","middleName":"","lastName":"Wan","suffix":""},{"id":635770361,"identity":"d975cb06-7962-49d8-8274-cf336fec2b5e","order_by":7,"name":"Fuzhou Ye","email":"","orcid":"","institution":"Institute of Cancer Research","correspondingAuthor":false,"prefix":"","firstName":"Fuzhou","middleName":"","lastName":"Ye","suffix":""},{"id":635770363,"identity":"7486c97f-026f-4cf5-814a-cb8b83dff23b","order_by":8,"name":"Claudio Alfieri","email":"","orcid":"","institution":"Institute of Cancer Research","correspondingAuthor":false,"prefix":"","firstName":"Claudio","middleName":"","lastName":"Alfieri","suffix":""},{"id":635770364,"identity":"7cd944da-68f1-40cb-8337-f05f4f52b009","order_by":9,"name":"Jessica Downs","email":"","orcid":"","institution":"Institute of Cancer Research","correspondingAuthor":false,"prefix":"","firstName":"Jessica","middleName":"","lastName":"Downs","suffix":""},{"id":635770365,"identity":"072476b4-0a8d-4d4c-a8fc-d070b5c4a7af","order_by":10,"name":"Jyoti Choudhary","email":"","orcid":"","institution":"Institute of Cancer Research","correspondingAuthor":false,"prefix":"","firstName":"Jyoti","middleName":"","lastName":"Choudhary","suffix":""}],"badges":[],"createdAt":"2026-04-14 13:09:31","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9415881/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9415881/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":108805550,"identity":"fa125d69-a529-4e79-8b43-3b7c15b7aae6","added_by":"auto","created_at":"2026-05-08 15:26:14","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":736494,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eARID1A PRISMA interactome. A, \u003c/strong\u003eCurated ARID1A physical interaction map. Coloured boxes in ARID1A diagram represent significant domains, regions of motifs. The top graph represents intrinsic disorder regions (IDR). Intrinsic disorder scores were calculated with IUPred2A (https://iupred2a.elte.hu/). The blue line represents the IUPred2 prediction (disordered protein regions) and the orange line represents the ANCHOR2 prediction (disordered binding regions). Scores above 0.5 represent intrinsic disorder. \u003cstrong\u003eB, \u003c/strong\u003ePRISMA\u003cstrong\u003e \u003c/strong\u003eexperimental workflow\u003cstrong\u003e. \u003c/strong\u003eNuclear lysates from RPE1 cells were incubated with an ARID1A tiled peptide array. After membrane washes the individual peptide spots containing binding proteins were processed and analysed by mass spectrometry. \u003cstrong\u003eC, \u003c/strong\u003ePRISMA\u003cstrong\u003e \u003c/strong\u003edata analysis workflow\u003cstrong\u003e. D, \u003c/strong\u003eClustered heat-map of PRISMA binding profiles aligned with the ARID1A protein diagram and IUPred2A predictions, representing data from a single biological replicate with technical replicate MS measurements for 50% of peptide spots. Only proteins that were identified by at least two peptides were used, and the two replicate datasets were integrated by calculating the average LFQ intensity (hereafter intensity) for each protein in each spot where two measurements were available or using the single available value. The signal intensity for each protein across the 228 bait peptides was then normalised between 0 and 1. See also Tables S1 and S2 and Figures S1 and S2.\u003c/p\u003e","description":"","filename":"fig1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/83cc194693cc090438c2129d.jpg"},{"id":108687614,"identity":"504f5d79-5c0f-4789-b30e-8f7b9e516564","added_by":"auto","created_at":"2026-05-07 10:20:20","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1050941,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePRISMA captures ARID1A binding interfaces to BAF subunits. A, \u003c/strong\u003eARID1A\u003cstrong\u003e \u003c/strong\u003ePRISMA profiles of BAF subunits\u003cstrong\u003e. B, \u003c/strong\u003eDetailed comparison of PRISMA data on BAF subunits to amino acid contacts in cryoEM model structure 6LTH (9) and XL-MS data of the SWI/SNF complex (8). No contacts between ARID1A and ACTL6A were observed in cryoEM structure 6LTH. Contacts by different paralog subunits are represented where available.\u003c/p\u003e","description":"","filename":"fig2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/c45332b32c3fd4d871c2f87b.jpg"},{"id":108687617,"identity":"bfa9af90-b31a-42e9-8641-e73c1bd4af27","added_by":"auto","created_at":"2026-05-07 10:20:21","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":315725,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePRISMA refines interaction interface maps of known ARID1A interactions. A, \u003c/strong\u003ePRISMA profiles of ARID1A mapped interacting partners\u003cstrong\u003e. B, \u003c/strong\u003ePRISMA detects YAP1 binding to a PPXY SLiM in ARID1A\u003cstrong\u003e. \u003c/strong\u003eELMdb represents Eukaryotic Linear Motifs database annotation of PPxY motifs (LIG_WW_1) on ARID1A. Intensity scale as in A.\u003c/p\u003e","description":"","filename":"fig3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/17ea37283aa399142006725f.jpg"},{"id":108806202,"identity":"a0c3336d-215f-404d-ae67-e92d51fb0777","added_by":"auto","created_at":"2026-05-08 15:27:59","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":414300,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSIN3A binds to ARID1A through a poorly conserved Sin3 SLiM. A, \u003c/strong\u003eAlignment of the SIN3A PRISMA binding profile to ARID1A tiled peptide array and the Sin3 SLiM ELM prediction on ARID1A\u003cstrong\u003e. B, \u003c/strong\u003eBar graph representing levels of SIN3A on chromatin in ARID1A KO and parental RPE1 cells (n=3, mean ± SD, * p\u0026lt; 0.05, *** p\u0026lt; 0.00005). \u003cstrong\u003eC, \u003c/strong\u003eCo-immunoprecipitation of endogenous SIN3 complex (circled) by ARID1A C-terminal fragment GFP pulldown. \u003cstrong\u003eD, \u003c/strong\u003eSchematic of the \u003cem\u003ein vitro\u003c/em\u003e binding experiment between recombinant ARID1A-V5 and SIN3A fragment (449-1086) containing the PAH3 domain tagged with StrepII tag. \u003cstrong\u003eE,\u003c/strong\u003e Co-immunoprecipitation of recombinant ARID1A-V5 and SIN3A_PAH3-StrepII. ARID1A-V5 was immunoprecipitated from insect cell lysates. IgG was used as a negative control. Beads were then incubated with cell lysate from insect cells expressing SIN3A_PAH3-StrepII. Co-immunoprecipitation was assessed with anti-V5 and anti-StrepII antibodies to detect ARID1A and SIN3A_PAH3 respectively. KDM5A-StrepII was used as a positive control. Lanes labelled with X are irrelevant for this work.\u003c/p\u003e","description":"","filename":"fig4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/f3d87710424561fce62d5280.jpg"},{"id":108687620,"identity":"ddd9e932-8d6c-46dc-ba94-f832dadcae40","added_by":"auto","created_at":"2026-05-07 10:20:21","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":456759,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eA novel interaction between ARID1A and protein phosphatase 1 regulatory protein TOX4. A, \u003c/strong\u003eTOX4 PRISMA binding profile to ARID1A tiled peptide array.\u003cstrong\u003e \u003c/strong\u003ePredicted PLK1 consensus sites in TOX4-binding ARID1A peptides are shown\u003cstrong\u003e. B, \u003c/strong\u003eAlphaFold structural models of TOX4 with the two ARID1A peptides where TOX4 binding signal was detected.\u003cstrong\u003e C, \u003c/strong\u003eCo-immunoprecipitation of TOX4 and ARID1A. TOX4 was immunoprecipitated from RPE1 and HEK293T cells. IgG was used as a negative control. TOX4 immunoprecipitates were probed for the presence of ARID1A and ubiquitin. Short and long exposures are shown for ARID1A detection.\u003c/p\u003e","description":"","filename":"fig5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/f4b28b6092d9fa26492880b2.jpg"},{"id":108805573,"identity":"a841682d-4262-4d66-84e3-d06942a800b9","added_by":"auto","created_at":"2026-05-08 15:26:19","extension":"jpg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":342247,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCDK2/Cyclin A bind to ARID1A. A, \u003c/strong\u003eCDK2 and cyclin A2 PRISMA binding profiles with zoomed in sections of CDK2/CCNA2 binding overlap and CDK2 phosphorylation site on ARID1A. \u003cstrong\u003eB,\u003c/strong\u003eValidation of CDK2/CCNA2 binding to a subset of ARID1A peptides by far-western blotting using a recombinant mix of CDK2/CCNA2. WT, wild type peptide sequence; SCR, scrambled sequence control. \u003cstrong\u003eC,\u003c/strong\u003e Validation of CCNA2 direct binding to a subset of ARID1A peptides by far-western blotting using recombinant CCNA2. Peptides as in B. \u003cstrong\u003eD,\u003c/strong\u003e Heatmap showing levels of phosphorylated S363 (normalised to ARID1A abundance) along the cell cycle. See also Table S4.\u003c/p\u003e","description":"","filename":"fig6.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/b3cc033190a283793e224ea3.jpg"},{"id":108687625,"identity":"a5687557-72e6-4ec5-a9f2-d55242acf989","added_by":"auto","created_at":"2026-05-07 10:20:21","extension":"jpg","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":698565,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMutation of ARID1A S363 to Alanine impacts ARID1A function. A, \u003c/strong\u003eBar\u003cstrong\u003e \u003c/strong\u003egraph and curve\u003cstrong\u003e \u003c/strong\u003eshowing growth (as % confluence) of ARID1A KO cells alone or expressing emGFP only, wild type emGFP-ARID1A WT or emGFP-ARID1A_S363A. Points correspond to independent biological replicates, n = 3, mean ± SEM. \u003cstrong\u003eB,\u003c/strong\u003e Heatmap of enriched GO and KEGG terms from Gene Set Enrichment Analysis of RNASeq data from wild type ARID1A, ARID1A KO alone, and ARID1A KO expressing ARID1A wild type or S363A mutant form. \u003cstrong\u003eC,\u003c/strong\u003eEnrichment plot of genes (RNA-seq) and proteins (Proteomics) with the GO term Microtubule cytoskeleton annotation, showing lower levels of genes and proteins annotated with this term are associated with the ARID1A S363A mutation. \u003cstrong\u003eD,\u003c/strong\u003eHeatmap of enriched GO and KEGG terms from Gene Set Enrichment Analysis of proteomics data from wild type ARID1A, ARID1A KO alone, and ARID1A KO expressing ARID1A WT or S363A mutant form. \u003cstrong\u003eE,\u003c/strong\u003e Cell fate profiles comparing ARID1A KO, ARID1A KO expressing either emGFP-ARID1A WT or emGFP-ARID1A_S363A. Each track represents an individual cell, with distinct colours indicating specific cell cycle events. \u003cstrong\u003eF,\u003c/strong\u003e Bar graph showing mitotic events recorded over a 72-hour imaging period for ARID1A KO, ARID1A KO expressing either EmGFP-ARID1A WT or EmGFP-ARID1A S363A. See also Tables S5 and S6.\u003c/p\u003e","description":"","filename":"fig7.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/f6e51426431b5fee585f0d6f.jpg"},{"id":108809783,"identity":"47de2c64-9c8b-4daa-9c5d-fb309f058055","added_by":"auto","created_at":"2026-05-08 15:55:26","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4462277,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/c973813e-7a4e-4dc3-8209-51a5d4fc3031.pdf"},{"id":108806224,"identity":"366bd8aa-f2e4-4418-bf74-30962989d3fb","added_by":"auto","created_at":"2026-05-08 15:28:03","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":122122,"visible":true,"origin":"","legend":"","description":"","filename":"FigS1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/fe469752b28894add6003655.pdf"},{"id":108806661,"identity":"6fea689f-05e6-4e22-8e05-1927515e938e","added_by":"auto","created_at":"2026-05-08 15:29:12","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":119621,"visible":true,"origin":"","legend":"","description":"","filename":"FigS2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/ff102060f94c3952eda24967.pdf"},{"id":108806562,"identity":"a699e9d4-21b1-4153-bbb8-625920539e98","added_by":"auto","created_at":"2026-05-08 15:28:55","extension":"pdf","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":287958,"visible":true,"origin":"","legend":"","description":"","filename":"FigS3.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/6708592836fe812a36a36c85.pdf"},{"id":108805547,"identity":"6e6b4457-fe37-4a06-9a1c-3c522712cac5","added_by":"auto","created_at":"2026-05-08 15:26:13","extension":"pdf","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":1692941,"visible":true,"origin":"","legend":"","description":"","filename":"FigS4.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/5cd59e1b603a3fb4a24320e5.pdf"},{"id":108687622,"identity":"06ebc7d3-9cb7-4469-8cac-077c67db453c","added_by":"auto","created_at":"2026-05-07 10:20:21","extension":"xlsx","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":18957,"visible":true,"origin":"","legend":"","description":"","filename":"TableS1.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/5b89caf69fa0028b5e9c465d.xlsx"},{"id":108806329,"identity":"e6c94d0a-3c22-419f-a341-89daebf9f326","added_by":"auto","created_at":"2026-05-08 15:28:16","extension":"xlsx","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":92872,"visible":true,"origin":"","legend":"","description":"","filename":"TableS3.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/06749bfb97df5801a61e6e60.xlsx"},{"id":108806690,"identity":"8ad6d04e-258b-46d9-9b41-051fed64193a","added_by":"auto","created_at":"2026-05-08 15:29:15","extension":"xlsx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":10119,"visible":true,"origin":"","legend":"","description":"","filename":"TableS4.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/39c6c3fa5bd7d05a17ea51da.xlsx"},{"id":108806965,"identity":"8cd99af0-01fd-4a91-9c95-aa098b826a20","added_by":"auto","created_at":"2026-05-08 15:29:47","extension":"xlsx","order_by":8,"title":"","display":"","copyAsset":false,"role":"supplement","size":124807,"visible":true,"origin":"","legend":"","description":"","filename":"TableS5.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/8e9fc85b798f4518f82ebcb9.xlsx"},{"id":108805986,"identity":"25848de1-5f5b-497f-883a-c3b4efd01317","added_by":"auto","created_at":"2026-05-08 15:27:23","extension":"xlsx","order_by":9,"title":"","display":"","copyAsset":false,"role":"supplement","size":31337,"visible":true,"origin":"","legend":"","description":"","filename":"TableS6.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/5ec1c021ae1dc74ca308c19c.xlsx"},{"id":108805620,"identity":"8ba1c043-3c57-4d39-83e7-3f2d71508e2e","added_by":"auto","created_at":"2026-05-08 15:26:27","extension":"xlsx","order_by":10,"title":"","display":"","copyAsset":false,"role":"supplement","size":10529,"visible":true,"origin":"","legend":"","description":"","filename":"TableS8.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/360d8b0ab268a57724ba33b7.xlsx"},{"id":108806045,"identity":"01b7d3f0-e8c6-40ff-8ce7-27f3e06ac06f","added_by":"auto","created_at":"2026-05-08 15:27:34","extension":"xlsx","order_by":11,"title":"","display":"","copyAsset":false,"role":"supplement","size":10171,"visible":true,"origin":"","legend":"","description":"","filename":"TableS7.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/9519fbf4c914c594af00ea91.xlsx"},{"id":108687629,"identity":"a0cf56f6-474d-4f0f-ba17-90090dc461b4","added_by":"auto","created_at":"2026-05-07 10:20:21","extension":"xlsx","order_by":12,"title":"","display":"","copyAsset":false,"role":"supplement","size":3502824,"visible":true,"origin":"","legend":"","description":"","filename":"TableS2.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/e56bf26a0ee83494f91885e6.xlsx"},{"id":108687630,"identity":"2852ceec-ea5c-48ef-bb88-ab46c41dc660","added_by":"auto","created_at":"2026-05-07 10:20:21","extension":"docx","order_by":13,"title":"","display":"","copyAsset":false,"role":"supplement","size":15370,"visible":true,"origin":"","legend":"","description":"","filename":"Supplementalfiles.docx","url":"https://assets-eu.researchsquare.com/files/rs-9415881/v1/6972f020503eaf5c58635fef.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"High resolution interaction surface mapping by PRISMA reveals novel ARID1A interactions","fulltext":[{"header":"Background","content":"\u003cp\u003eARID1A is a defining subunit of the multi-protein cBAF complex, one of several types of SWI/SNF chromatin remodelling complexes that use energy derived from ATP hydrolysis to alter the position of nucleosomes and modify DNA accessibility for RNA polymerase II and other transcriptional regulators, thereby regulating many DNA-based processes, such as transcription, splicing, DNA replication and DNA damage repair (\u003cspan additionalcitationids=\"CR2 CR3 CR4 CR5 CR6\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e). ARID1A directly binds to many of the subunits of the BAF complex and in fact acts as a scaffold for its assembly (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). Loss of ARID1A leads to complex dissociation, reduced levels of other BAF subunits and BAF complex deficiency (\u003cspan additionalcitationids=\"CR11\" citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e). ARID1A has also shown to interact with ribosomes and be important for translation (\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e, \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e) and reported to have a role in cell plasticity, through a mechano-regulated interaction with transcription factors in the Hippo signalling pathway (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e, \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e). Subsequently, ARID1A and BAF function impact many key aspects of cell and organismal physiology (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e). Mutations in ARID1A lead to neurodegenerative diseases and account for ~\u0026thinsp;6% of all cancer mutations (\u003cspan additionalcitationids=\"CR20 CR21\" citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e). Better understanding of regulatory mechanisms of ARID1A function could help develop novel therapeutic strategies.\u003c/p\u003e \u003cp\u003eARID1A is a large protein containing two conserved domains of globular structure, an ARID domain located centrally that mediates binding to DNA and a C-terminal domain that mediates interactions with other BAF subunits and is required for BAF complex formation. In contrast, the N-terminal half of the protein is mostly made up of intrinsically disordered regions (IDRs) (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e). Not much is known about the function of this region of ARID1A. Despite their lack of structure, IDRs play crucial roles in the control of signalling pathways and can also work as platforms for the assembly of large proteinaceous structures. IDRs frequently contain SLiMs or short linear motifs that mediate interactions with globular domains (\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e). They also often carry post translational modifications (PTMs) that function as regulatory switches (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e). Many disordered regions are capable of phase separation, and this is indeed a general property of most SWI/SNF subunits (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e). The N-terminus of ARID1A can partition into phase-separated condensates, possibly through interaction with FUS, a nucleic acid binding protein that forms oncogenic fusion proteins (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e). Recently, the IDRs of ARID1A were also shown to be required for genomic targeting of the BAF complex to enhancers and interaction with transcription factors and other transcriptional machinery (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eThe majority of ARID1A interacting proteins that have been mapped to date bind to the C-terminal half of ARID1A (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA). This includes, in addition to other BAF subunits, proteins involved in DNA damage response and repair such as ATR and MSH2, cell cycle regulators such as p53 and RB, and other transcriptional regulators such as Polycomb protein EZH2 and the glucocorticoid receptor, transcription factor HIC2, topoisomerase enzyme TOP2A, and E3 ubiquitin ligase bTrCP. Characterised binding partners on the disordered N-terminal half of ARID1A include FUS, and YAP1/TAZ, transcription factors that regulate cell plasticity (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn addition to these, many other ARID1A-interacting proteins have been reported in the literature. However, the molecular basis for these interactions is unknown. Recent reports suggest that many functional ARID1A interactions occur through the disordered regions. IDR-mediated interactions are inherently challenging to study due to their reversible and weak nature, and traditional approaches are not best suited to investigate them (\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e). Here, we take advantage of a peptide interactomics screen on a tiled peptide array to systematically identify amino acid residues, domains and motifs that mediate ARID1A interactions with other proteins at high sequence resolution. Our PRISMA data recapitulates interaction surfaces of ARID1A with other BAF complex subunits, and SLiM-mediated interactions. We identify and validate several novel ARID1A-interacting proteins with roles in transcriptional regulation and cell cycle control, and map their binding sites. Finally, we functionally characterise and uncover a role for ARID1A phosphorylation in the transcriptional regulation of microtubule organisation and cell division, providing the first insights into ARID1A regulation through phosphorylation.\u003c/p\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eA sequence-resolution biophysical interaction map of ARID1A\u003c/h2\u003e \u003cp\u003eTo systematically map the landscape of ARID1A protein-protein interactions at high sequence resolution and delineate interaction surfaces, domains and motifs in ARID1A, we took advantage of the recently developed \u003cem\u003eProtein Interaction Screen on peptide MAtrices\u003c/em\u003e (PRISMA) technology (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e). The PRISMA assay consists of high-throughput affinity purifications using hundreds of overlapping peptides spanning the entire sequence of a target protein immobilised on a nitrocellulose membrane, combined with shotgun mass spectrometry analysis, for identification and quantitation of peptide-bound proteins. We designed our PRISMA assay with 20-mer peptides with a 10-aa (amino acid) overlap covering the whole length of human ARID1A (2285 aas), resulting in a high-density tiled array of 228 peptide spots (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB, Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). Binding proteins were screened using nuclear extract from RPE1 cells in the presence of nuclease to eliminate DNA-mediated interactions. A total of 2224 proteins identified by at least 2 peptides were quantified (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eC, D).\u003c/p\u003e \u003cp\u003eAround a quarter of the proteins bound ubiquitously across the peptide array at low intensities, while the rest displayed much higher signals in discrete regions of the array. The frequency of binding was calculated for each spot, revealing that the number of interacting proteins per spot ranged from 1 to 235 (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD). Approximately 700 proteins (30%) were quantified in more than half of the peptide array (100 peptide spots), whilst around 100 were found to bind to over 200 peptide spots (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD). This so called \u0026ldquo;promiscuous\u0026rdquo; binding has also previously been observed in a PRISMA assay for C/EBPβ (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e). To reduce the noise from background binding and highlight the strongest signals across the peptide array, binding signals below the 75th percentile intensity were removed. Furthermore, to discriminate robust interactions, a second filter based on the presence of consecutive signal intensities across a minimum of two contiguous peptides was applied. The consecutive binding is driven by the overlap in the tiled peptide sequences (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eB, C). This strict filter criterion, applied after the 75th percentile outlier filtering, yielded an ARID1A core PRISMA interaction set of 1454 proteins. It is worth pointing out that the consecutive binding filter when applied to the ARID1A PRISMA dataset is much more restrictive that the one applied in the original PRISMA paper, where the matrix peptides were shorter (11 aas) and also had a smaller overlap (4 aas), which means that generally any sequence stretch was covered by three peptides, whilst in the ARID1A PRISMA assay described here any sequence is only covered by two peptides. The distribution of peptide interactions was not attributed to physicochemical parameters of the peptides, as evidenced by comparison of the number of binding proteins with peptide properties such as hydrophobicity, charge, and isoelectric point (Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). These results suggest that the PRISMA peptides on the matrix retained their specific protein-binding properties.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe global binding profiles from the PRISMA dataset were represented as a hierarchically clustered heatmap of normalised intensities (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eD, Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e, Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). This revealed there were large differences in the binding profiles of the matrix peptides indicating that binding was specific, as background binding would be expected to result in a homogeneous distribution of interacting proteins. Peptides encompassing the conserved \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003eA\u003c/span\u003eRID \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003es\u003c/span\u003ecaffolding \u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003ed\u003c/span\u003eomains ASD1 and ASD2, and particularly those containing the LXXLL motifs #2 (aas 1709\u0026ndash;1713) and #4 (aas 2085\u0026ndash;2089), displayed the highest number of interactions, followed by regions flanking the ARID domain (Figure \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). The ARID DNA-binding domain and the NLS appeared devoid of interactions. The region immediately upstream of the ARID domain that sustains a high number of interactions is an unstructured stretch that is predicted to fold upon binding to a protein partner by the ANCHOR2 algorithm (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e). The N-terminal region also displayed significant numbers of interactions, seemingly organised in locally enriched hotspots, some of which overlapped with predicted disordered sequences or disordered binding regions.\u003c/p\u003e \u003cp\u003eGO term enrichment analysis of the PRISMA core interaction dataset revealed significant over-representation of proteins annotated with GO terms related to known ARID1A functions (Figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003eA, Table \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e), with enriched terms including translation, mRNA splicing, ATP-dependent chromatin remodelling and DNA damage response. Enrichment of protein domains revealed that nucleotide binding domain, RNA recognition domain, Armadillo, DNA/RNA helicase domain and AAA ATPase domain were some of the most frequently represented domains and sites (data not shown). To further assess the biological relevance of the PRISMA-derived data, we compared it to the ARID1A physical interactome retrieved from the BioGRID repository of protein interactions. Interactors identified by yeast two-hybrid or RNA-purification, viral proteins and duplicates were removed prior to comparison, leaving 211 ARID1A physical interactors. Forty-nine of these were detected by PRISMA and 31 of them were included in the core dataset (Figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003eB). These results suggest that the PRISMA-derived interaction data shows a high level of overlap and coverage of the native ARID1A interactome and represents a useful resource to explore sequence-to-function relationships at high sequence resolution.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003ePRISMA recapitulates binding of BAF subunits\u003c/h3\u003e\n\u003cp\u003eTo assess the ability of the ARID1A PRISMA to detect localised interactions, we investigated the binding of components of the BAF complex. Core module subunits SMARCB1, SMARCC1/2, SMARCD1/3 and SMARCE1, ATPse module subunits SMARCA4 and ACTL6A, and accessory subunit DPF2 displayed correlated interaction footprints on the conserved ASD1 and ASD2 domains, immediately upstream of the ARID domain, and on the central part of the disordered N-terminal domain, with the latter displaying the strongest signals (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e2\u003c/span\u003eA). Consecutive binding signals of SMARCA4, SMARCB1 and SMARCE1 were observed in the ASD2 domain aligned with the LXXLL motif located at position 2085\u0026ndash;2089, which has been shown to be essential for binding of BAF subunits (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e). We then compared binding sites of BAF subunits with cryogenic electron microscopy (cryoEM) structural model 6LTJ data (\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e). The PRISMA interaction footprint was in good agreement with cryoEM maps, demonstrating multiple contacts of the ASD1 and ASD2 conserved domains of ARID1A with SMARCA4, SMARCB1, SMARCC1/2, SMARCD1, SMARCE1 and DPF2 (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). Furthermore, binding signals of SMARCA4, SMARCB1 and SMARCE1 showed remarkable overlap with protein contact sites from the cryoEM structural models (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e2\u003c/span\u003eB). In this cryoEM study only a stretch of the ARID1A C-terminus (aas 1639\u0026ndash;2285) was modelled, precluding the comparison of contact sites at the protein N-terminal half. Therefore, we also compared the PRISMA data to BAF complex crosslinking-mass spectrometry (XL-MS) data (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e) that encompasses the whole ARID1A sequence. The agreement between both was particularly noteworthy (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e2\u003c/span\u003eB): in addition to the well-documented binding of several BAF subunits on the C-terminal region of ARID1A, both PRISMA and XL-MS showed SMARCD, SMARCC and ACTL6A strongly binding to ARID1A N-terminal region. The overlap was observed in regions, not in exact positions, as the different approaches provide different resolution and measure different properties, but converged on a consistent structure. Hence, the PRISMA data mirrored binding data from other experimental approaches and provided information on contact sites between associated proteins. Our PRISMA data suggests that while the BAF250_C region is essential for scaffolding the assembly of the BAF complex (\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e), the disordered N-terminus also mediates contacts with different BAF subunits, in agreement with XL-MS (\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e). Most SWI/SNF subunits have disordered prion-like regions that are involved in multivalent protein interactions (\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e). Interactions engaged by these disordered regions may serve a regulatory function.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eNext, the PRISMA data was compared with published sites of interaction in ARID1A with other proteins (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e3\u003c/span\u003eA). Out of 16 mapped ARID1A direct binders found in the PRISMA data, three displayed PRISMA binding patterns that agreed with their mapped site of interaction with ARID1A. Transcriptional coactivator YAP1 displayed consecutive binding at 3 peptides spanning ARID1A aas 131\u0026ndash;170 (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e3\u003c/span\u003eB). These peptides contain the proline-rich PPXY sequence, a SLiM that mediates binding of ARID1A to the YAP1 WW domain, and mutation of ARID1A Tyr148 to Ala has been shown to eliminate YAP1 binding (\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e). Therefore, our PRISMA data successfully pinpointed the site of YAP1 interaction. Other WW domain proteins such as ITCH and NEDD4L were also detected binding to these PPXY containing peptides in our data, indicating that PRISMA can truly recapitulate SLiM-mediated direct associations. In contrast, transcriptional coactivator TAZ, which belongs to the same family as YAP, displayed consecutive binding at peptides from the ARID1A C-terminus and semi-consecutive binding between aas 2061\u0026ndash;2100. TAZ only contains one WW domain in contrast to YAP1 which contains 2, so this might explain the lack of detection of TAZ at the 144\u0026ndash;148 PPXY motif. The retinoblastoma protein (RB1) displayed weak consecutive binding signals on the two most C-terminal peptides (position 2261\u0026ndash;2285), in agreement with published data (\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e). ARID1A has been shown to interact with Rb and suppress its phosphorylation, with two truncation fragments, ARID1A:1110\u0026ndash;1976 and ARID1A:1977\u0026ndash;2285, being able to mediate the interaction. The binding footprint of paraspeckle protein FUS showed strongest signals in the disordered N-terminus of ARID1A, and some residual binding in a disordered region upstream of the ASD1 domain. This is also consistent with published data showing that ARID1A interacts and partitions into biocondensates with FUS through its N-terminal prion-like domain (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e). Other mapped ARID1A binders displayed only scattered single signals that could not be confidently attributed to specific binding (TOP2A, MSH2, ATR) or showed promiscuous binding (TGM2, ERH and PRKDC).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn summary, our results demonstrate that the ARID1A PRISMA assay can recapitulate FUS, RB1 and YAP1 in vivo binding. These have not previously been identified in affinity purification-mass spectrometry (AP-MS) interactomes, suggesting that the \u003cem\u003ein vitro\u003c/em\u003e set-up of PRISMA may be able to reveal complementary interactions that might be too labile or of too low abundance for traditional AP-MS approaches.\u003c/p\u003e\n\u003ch3\u003ePRISMA reveals binding of SIN3A to a predicted SIN3A SLiM\u003c/h3\u003e\n\u003cp\u003eSeeing as PRISMA can recapitulate SLiM-mediated interactions, we next investigated whether our assay could identify binding to other ARID1A SLiMs. We used the Eukaryotic Linear Motif database (ELMdb) (\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e) prediction tool to obtain a list of potential SLiMs in human ARID1A and checked whether peptides containing these motifs displayed significant specific binding of proteins containing the matching globular domain. We observed highest binding signal of E3 ubiquitin ligases on peptides containing degrons, and SH3 domain proteins on peptides containing an SH3-binding SLiM (Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). Detailed inspection of the data revealed strong consecutive binding of the transcriptional repressor SIN3A to peptides encompassing a predicted Sin3-interacting domain (SID) of low conservation (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e4\u003c/span\u003eA), suggesting this motif could mediate binding between SIN3A and ARID1A. The PRISMA data also showed SIN3A binding to the two C-terminal peptides (aas 2261\u0026ndash;2285). ARID1A and SIN3A have been shown to physically interact to repress the expression of cell cycle genes in mouse osteoblast precursor cells (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e), but the direct binding determinants have never been mapped. SIN3A has also been shown to interact with other subunits of BAF complexes (\u003cspan additionalcitationids=\"CR40 CR41\" citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e). Therefore, our results suggest that ARID1A and SIN3A may interact directly through a Sin3 SLiM and other supporting nearby residues.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eARID1A is required for the recruitment of SIN3A to chromatin in ovarian cancer cells and quiescent mouse pre-osteoblastic MC3T3-E1 cells (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e). To assess if ARID1A is necessary for chromatin recruitment of SIN3A in RPE1 cells, we isolated chromatin from ARID1A KO RPE1 cells and their wild type counterparts and performed quantitative profiling of chromatin-binding proteins by mass spectrometry. Inspection of the chromatome revealed that levels of SIN3A on chromatin were significantly lower in ARID1A KO cells compared to WT cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e4\u003c/span\u003eB), confirming that ARID1A is required for the recruitment of SIN3A to chromatin in RPE1 cells.\u003c/p\u003e \u003cp\u003eTo assess if the association between ARID1A and SIN3A requires the ARID1A N-terminus, we expressed an ARID1A C-terminal fragment encompassing aas 1018\u0026ndash;2285 tagged with Emerald-GFP (Em-ARID1A_C) in ARID1A KO RPE1 cells and performed GFP-trap immunoprecipitation followed by MS analysis. All subunits of the SIN3 complex were enriched in ARID1A_C immunoprecipitates compared to control GFP-only (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e4\u003c/span\u003eC), confirming that ARID1A C-terminus is sufficient for association with the SIN3 complex.\u003c/p\u003e \u003cp\u003eThe SID SLiMs of MAD1 and other proteins specifically interact with the PAH domain of SIN3 (\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e). To confirm if ARID1A and SIN3A interact directly, we individually expressed full length ARID1A fused to the V5 tag (\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e) and a SIN3A fragment encompassing aas 449\u0026ndash;1086 which includes the PAH3 domain (SIN3A_PAH3) tagged with StrepII in insect cells and performed pull-down experiments (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e4\u003c/span\u003eB, C). Full length KDM5A-StrepII was included as a positive control, since its paralog KDM3A has previously been shown to bind to ARID1A (\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e). Bead-bound ARID1A captured KDM5A and SIN3A_PAH3 (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). Although SIN3A_PAH3 was also detected in the negative control IgG immunoprecipitated, it was consistently much more abundant in V5 immunoprecipitates. Therefore, our results indicate that ARID1A interacts directly with the PAH domain of SIN3A.\u003c/p\u003e\n\u003ch3\u003ePRISMA identifies novel ARID1A interactor TOX4\u003c/h3\u003e\n\u003cp\u003eComparison of PRISMA data to ELM database predictions revealed potential novel candidate interactor TOX4 binding to two ARID1A peptides spanning aas 611\u0026ndash;630 and 1691\u0026ndash;1710 (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e5\u003c/span\u003eA). TOX4 is a regulatory subunit of the protein phosphatase 1 (PP1) complex that is almost exclusively associated with chromatin and has been shown to mediate dephosphorylation of the C-terminal domain (CTD) of RNA polymerase II, thus regulating transcriptional elongation (\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e). Despite being distal from each other, both TOX4-binding peptides encompassed phosphorylation consensus sites for several kinases (polo-like kinase 1 (PLK1), Casein kinase I (CK1) and glycogen synthase kinase 3 (GSK3)), suggesting that these were specific binding events. Phosphorylation of ARID1A serine 631 (Ser631) located in the kinase consensus site included in peptide 611\u0026ndash;630 has been previously detected (Figure \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e, (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e)).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWe used the AlphaFold protein\u0026ndash;protein interaction prediction algorithm (\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e) to predict the structure of the ARID1A peptides complexed with TOX4. For both peptides we obtained low confidence output models with low interface-predicted template modelling (0.13 and 0.41 respectively) and high PAE scores, not unexpectedly, given the highly disordered nature of TOX4 and one of the ARID1A peptides (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e5\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003eTo validate the interaction, we immunoprecipitated TOX4 from RPE1 and HEK293T cells and probed the immunoprecipitates with an anti-ARID1A antibody (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e5\u003c/span\u003eC). We reproducibly detected anti-ARID1A reactive bands at 270 kDa band and above in TOX4 immunoprecipitates, confirming that TOX4 and ARID1A interact \u003cem\u003ein vivo\u003c/em\u003e. Intriguingly, the ladder pattern of ARID1A bands was reminiscent of ubiquitination or sumoylation. We therefore probed TOX4 immunoprecipitates with an anti-ubiquitin antibody and observed that the ladder-like bands above 270 kDa were indeed ubiquitin-containing bands. These results confirm that TOX4 interacts with (poly)ubiquitinated ARID1A and suggest an interplay between (de)phosphorylation and ubiquitination of ARID1A.\u003c/p\u003e\n\u003ch3\u003ePRISMA uncovers ARID1A binding to CDK2/cyclin A2\u003c/h3\u003e\n\u003cp\u003eThe detection of CDK2 and cyclin A2 as overlapping consecutive binders in at least two different regions on ARID1A was of particular interest (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e6\u003c/span\u003eA). The highest binding signal for both proteins was detected on peptides spanning aas 1391\u0026ndash;1430, where CDK2 and cyclin A2 binding overlapped on 2 out of 3 peptides. CDK2 also bound strongly between aas 531\u0026ndash;600, with cyclin A2 displaying residual binding on one peptide in that span. Both binding sites overlapped with sequences of intrinsic disorder that were predicted to fold upon binding to a protein partner by the ANCHOR2 algorithm (\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e). CDK2 alone also displayed strong binding on peptides spanning aas 1451\u0026ndash;1480, and residual binding on peptides spanning aas 331\u0026ndash;370, which encompass a CDK2 consensus site. Phosphorylation of Serine 363 within this sequence is one of the most frequently detected ARID1A post-translational modifications (Figure \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e). Exploration of a published CDK2-dependent phosphopeptidome study revealed that phosphorylation of Ser363 is dependent on CDK2 (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e). Therefore, our results endorse the direct interaction between ARID1A with CDK2 and/or cyclin A via small regions of intrinsic disorder.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo validate this, we synthesised a peptide subset array containing ARID1A peptides that had shown binding to CDK2 and/or CCNA2, an ARID1A peptide containing an RxL motif as positive control (RxL is the best characterised docking motif for cyclin A (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e)), and the corresponding scrambled sequences counterparts as negative controls (Table \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e). The peptide array was then probed with a mix of recombinant full length CDK2 and a CCNA3 fragment spanning aas 174\u0026ndash;432 that contains the hydrophobic patch that docks on the RxL motif (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e6\u003c/span\u003eB). Western blot analysis revealed binding of CDK2 to the RxL peptide and not to its scrambled counterpart. We also detected binding of CDK2 to the ARID1A peptide spanning aas 1451\u0026ndash;1470, but not to its scrambled counterpart. We could not perform cyclin A detection in this experiment because the fragment used does not contain the epitope of our anti-cyclin A antibody. To establish which of CDK2 and cyclin A is the direct binder to ARID1A, we probed a second copy of the array with recombinant full length CCNA2 (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e6\u003c/span\u003eC). Detection with an antibody against CCNA2 showed that CCNA2 alone binds to ARID1A peptide spanning aas 1451\u0026ndash;1470, suggesting that CCNA2 mediates the binding of CDK2 to this sequence.\u003c/p\u003e \u003cp\u003eInterestingly, CDK2 or cyclin A2 have never been identified as ARID1A-associated proteins in our AP-MS experiments or any published ARID1A AP-MS studies (\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e, \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e). Thus, our data highlight the utility of the PRISMA strategy in revealing weak, transient and/or low abundance ARID1A interactions, possibly mediated by SLiMs, that are not detectable by traditional affinity purification strategies.\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eExploring the role of ARID1A Ser363 phosphorylation\u003c/h2\u003e \u003cp\u003eA survey of our recent cell cycle phosphoproteomics study (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e) revealed that ARID1A Ser363 phosphorylation is modulated along the cell cycle, being higher in G1 and S phases, and decreasing at the end of S phase and throughout G2 and mitosis (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e6\u003c/span\u003eD). To investigate the role of Ser363 phosphorylation, we generated RPE1 cells lines expressing inducible EmeraldGFP (emGFP)-tagged wild type (WT) ARID1A, phospho-mimetic Ser 363 to Asp mutated ARID1A (ARID1A_S363D) or phospho-dead Ser 363 to Ala ARID1A (ARID1A_S363A) in an ARID1A KO background. ARID1A_S363A was observed as a\u0026thinsp;~\u0026thinsp;250 kDa band in agreement with its predicted size, but the S363D mutant could not be detected with anti-ARID1A antibodies, and in fact was only expressed as a small protein fragment containing emGFP (Figure \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003eA, B). We confirmed this through primary sequence analysis by mass spectrometry of GFP-immunoprecipitates (Figure \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003eC). ARID1A peptides spanning the entire ARID1A sequence were detected in ARID1A_S363A immunoprecipitates. In contrast, in ARID1A_S363D immunoprecipitates we could only detect GFP peptides and ARID1A peptides upstream of S363D, indicating that the ARID1A_S363D protein detected by anti-GFP western blotting is a truncated form of ARID1A. This suggests that the ARID1A_S363D mutant protein is unstable and gets degraded to a truncated version in a rapid manner or was strongly selected against and a genomic rearrangement led to the truncated form. We could not reverse this degradation with proteasome inhibition (Figure \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003eB). Mass spectrometry analysis of GFP-trap immunoprecipitates also detected most BAF complex subunits in ARID1A_S363A immunoprecipitates (Figure \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003eD), suggesting mutation of Ser363 into Ala does not impact the ability of ARID1A to underpin the assembly of the complex. As expected, no other BAF subunits were detected in ARID1A_S363D immunoprecipitates, since the complex integrity and assembly relies on the ARID1A C-terminal domain (\u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e56\u003c/span\u003e). Whilst EmGFP-ARID1A_S363A fusion protein localised mostly to the cell nucleus (Figure \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003eE) in a similar manner to endogenous ARID1A, the S363D mutant localised to the cytoplasm, again as expected due to the loss of the NLS motif at position 1368\u0026ndash;1387.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo test if S363 phosphorylation is important for cell growth, we induced the expression of emGFP, wild type and S363A mutant emGFP-ARID1A proteins in ARID1A KO RPE1 cells and monitored their growth by measuring cell confluency over 120 h (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e7\u003c/span\u003eA). We observed that in these conditions, cells expressing em-GFP-ARID1A_S363A had a mild proliferation defect compared to cells expressing wild type emGFP-ARID1A, suggesting that ARID1A S363 phosphorylation may be important for cell proliferation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eTo explore the effects of the S363A mutation on ARID1A transcriptional regulation activity, we performed transcriptomics analysis using RNA sequencing (RNA-Seq) of ARID1A KO cells expressing emGFP-tagged wild type (WT) ARID1A, ARID1A_S363A or emGFP only, the untransfected ARID1A KO and the parental wild type cells after induction with doxycycline for 24 h. We then compared parental to ARID1A KO cells, and KO cells expressing wild type ARID1A to those expressing ARID1A_S363A. Differences between wild type and ARID1A-defficient cells were modest, as has previously been reported, and so were those between cells expressing induced wild type or mutant ARID1A. Gene set enrichment analysis (GSEA) revealed that ARID1A-deficient cells showed reduced expression of genes involved in chromatin organisation, replication, DNA repair, translation and peptide presentation compared to the parental (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e7\u003c/span\u003eB, Table \u003cspan refid=\"MOESM5\" class=\"InternalRef\"\u003eS5\u003c/span\u003e), as has been reported previously (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e). Notably, cells expressing ARID1A_S363A did not display these molecular phenotypes when compared to cells expressing WT ARID1A. They did however show increased expression of genes involved in oxidative phosphorylation, TNF signalling and interferon response, all of which are hallmarks of loss of ARID1A function (\u003cspan additionalcitationids=\"CR58 CR59\" citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e). Strikingly, they also showed reduced expression of genes involved in microtubule-related processes, such as microtubule binding, mitotic spindle organisation, microtubule motor activity and kinesin complex (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e7\u003c/span\u003eB, C; Table \u003cspan refid=\"MOESM5\" class=\"InternalRef\"\u003eS5\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTo assess whether these effects translate on the protein level, we performed proteome profiling using quantitative MS. As expected, ARID1A KO cells displayed lower abundance of other BAF complex subunits at the protein level compared to the WT parental lines (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e7\u003c/span\u003eD) as previously shown (\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e), but this was not the case for EmGFP-ARID1A_S363A cells compared to cells expressing emGFP-ARID1A_WT, confirming that the S363A mutation does not affect the BAF complex assembly or stability. GSEA of proteomics data recapitulated the microtubule-based defects observed in the transcriptomics analysis (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e7\u003c/span\u003eC, D, Table \u003cspan refid=\"MOESM6\" class=\"InternalRef\"\u003eS6\u003c/span\u003e). This is striking given that transcriptomic and proteomics changes elicited by loss of key SWI/SNF subunits are generally not well correlated (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e). Cells expressing mutated ARID1A showed reduced abundance of proteins involved in microtubule organisation and mitotic events such as mitotic spindle assembly, chromosome alignment and nuclear membrane reassembly after mitosis. There was considerable overlap of downregulated genes/proteins annotated with \u0026ldquo;microtubule cytoskeleton\u0026rdquo; (50%) or \u0026ldquo;mitotic spindle organization\u0026rdquo; (75%) between RNA-seq and proteomics datasets. These results show that phosphorylation of S363 is important for specific aspects of ARID1A function related to the microtubule cytoskeleton.\u003c/p\u003e \u003cp\u003eThe reduced abundance of proteins involved in microtubule dynamics in ARID1A_S363A mutant cells suggested that they may have defects in mitosis. To explore this, we induced expression of wild type or mutant ARID1A in ARID1A KO cells and monitored cell division using live-cell imaging over a period of 72 h (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e7\u003c/span\u003eE). We observed that cells expressing ARID1A_S363A rarely underwent mitosis over the 72 h, whilst cells expressing WT ARID1A divided at similar rates to the untransfected KO cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig13\" class=\"InternalRef\"\u003e7\u003c/span\u003eF). In summary, our results suggest that ARID1A Ser363 phosphorylation is important for normal mitosis and cell division.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eHere we have used a peptide array interactomics strategy (PRISMA) to uncover novel interaction partners and binding sites of the SWI/SNF scaffold subunit ARID1A. This relatively new approach detects biophysical protein binding and has proven very useful to identify interactions that are mediated by SLiMs (\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e, \u003cspan additionalcitationids=\"CR63\" citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e), but has only been applied to a small number of proteins to date. Our study significantly contributes to the understanding of molecular mechanisms of ARID1A regulation and function, by exposing novel ARID1A interacting partners and interaction sites, and providing a novel resource that can be used by the SWI/SNF community to further explore ARID1A biology.\u003c/p\u003e \u003cp\u003eIntegration of SLiM annotations on the target protein with the PRISMA screen data can significantly contribute to the detection of confident interaction candidates for validation and follow-up experiments, enhancing the confidence in potential true interactions and providing insight into the regulation, molecular mechanism and function of protein interactions. Matching of ARID1A SLiMs to proteins containing the partnering globular domains highlighted SIN3A as a direct interactor through its PAH domain(s). Although ARID1A has been shown to be required for the recruitment of SIN3A to chromatin to regulate the expression of cell cycle and cytokine genes and \u003cem\u003eTERT\u003c/em\u003e (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e, \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e), and interactions of SIN3A with the BAF complex have been reported (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e, \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e, \u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e), only one study had previously demonstrated an interaction between SIN3A and ARID1A (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e). Our data shows that the interaction between ARID1A and SIN3A is direct and mediated through the Sin3 SLiM.\u003c/p\u003e \u003cp\u003eThe high concentration of peptide on the PRISMA arrays facilitates the detection of weak and transient interactions that are challenging to capture in traditional affinity purification approaches employing endogenous proteins. This may have been instrumental in the identification of novel interactor TOX4. The detection of the ubiquitin-containing ARID1A laddered pattern in the TOX4 immunoprecipitate suggests that TOX4 preferentially binds to (poly)ubiquitinated ARID1A \u003cem\u003ein vivo\u003c/em\u003e, and this may explain why this interaction is not readily detected in traditional ARID1A IP-MS interactomes. Interestingly, the TOX4-binding ARID1A peptide encompassing aas 1691\u0026ndash;1710 contains a predicted ABBA motif, a SLiM that has been shown to act as a degradation signal for key mitotic regulators and a signal for kinetochore recruitment (\u003cspan citationid=\"CR67\" class=\"CitationRef\"\u003e67\u003c/span\u003e), raising the question of whether the interaction with TOX4 plays a role in ARID1A degradation. PP1 regulatory subunit TOX4 has been shown to have a role in cell fate reprogramming and differentiation (\u003cspan citationid=\"CR68\" class=\"CitationRef\"\u003e68\u003c/span\u003e, \u003cspan citationid=\"CR69\" class=\"CitationRef\"\u003e69\u003c/span\u003e), and was reported to bind to DNA adducts generated by cisplatin (\u003cspan citationid=\"CR70\" class=\"CitationRef\"\u003e70\u003c/span\u003e). Seeing as it plays a role in fundamental processes including genome stability (\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e, \u003cspan citationid=\"CR71\" class=\"CitationRef\"\u003e71\u003c/span\u003e), it is no surprise that it has been shown to promote tumourigenesis and so has interesting therapeutic potential. Both PLK1 and GSK3 form reciprocal regulatory loops with PP1 (\u003cspan additionalcitationids=\"CR73\" citationid=\"CR72\" class=\"CitationRef\"\u003e72\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR74\" class=\"CitationRef\"\u003e74\u003c/span\u003e), and our data suggests that the interplay of PP1 and either or both kinases might regulate some aspect of ARID1A function. Indeed, PLK1 inhibition was recently shown to be synthetic lethal with ARID1A deficiency (\u003cspan citationid=\"CR75\" class=\"CitationRef\"\u003e75\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSimilarly, this seems to be the case with CDK2 and CCNA2, which have not been previously detected in any ARID1A interactome study. We detect weak binding of the kinase on a peptide containing a CDK2 consensus site that has previously been shown to be phosphorylated in a CDK2-dependent manner, and stronger binding, together with cyclin A, on peptides within the region of intrinsic disorder in ARID1A between the ARID and the C-terminal globular domains. The ARID1A peptide with the strongest CDK2/CCNA2 binding signal does not contain a canonical cyclin A docking motif (RxL). It is, however, reminiscent of a yeast cyclin binding motif PxxPxF (\u003cspan citationid=\"CR76\" class=\"CitationRef\"\u003e76\u003c/span\u003e). Intriguingly, there is an RxL motif in the C-terminal of ARID1A that overlaps with the SIN3 binding motif, that indeed binds cyclin A \u003cem\u003ein vitro\u003c/em\u003e (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e6\u003c/span\u003eC).\u003c/p\u003e \u003cp\u003eOur exploration of published cell cycle phosphorylation data (\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e) suggests that ARID1A Ser363 is regulated in a cell cycle dependent manner and represents the first report of ARID1A regulation through phosphorylation. The mutation experiments prompted by findings from the PRISMA data indicate that the dynamic behaviour of Ser363 phosphorylation is specifically required for the expression of genes involved in microtubule-based processes, many of which are cell cycle genes. We find that expression of ARID1A_S363A mutant, where S363 cannot be phosphorylated, leads to decreased expression of genes and proteins involved in microtubule organisation and movement. Previous work has shown that ARID1A loss results in reduced expression of cell cycle and mitotic spindle genes at RNA and protein level (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e, \u003cspan citationid=\"CR77\" class=\"CitationRef\"\u003e77\u003c/span\u003e), and our data suggests that ARID1A S363 dynamic phosphorylation specifically regulates this transcriptional program. Whilst SWI/SNF controls gene expression, many of the transcriptional changes observed when specific subunits are lost do not often correlate well with changes at the protein level (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e, \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e, \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e). Importantly, our study shows that the regulation of MT-based genes by ARID1A S363 is cascaded down to the protein level and therefore results in a functional phenotypic effect on the cells. With ARID1A (and SWI/SNF) being master regulators of a huge number of genes and many different transcriptional programs, it is not surprising that regulation of these different sets of genes must happen through precise and specific mechanisms. This is not unprecedented for SWI/SNF. For instance, during mitosis SMARCE1 and SMARCB1 remain bound specifically to genes required for the completion of mitotic cell division and lineage-identity genes. By contrast, catalytic subunits SMARCA2 and SMARCA4 have been shown to localise to the nucleus through interphase but are displaced from condensed chromatin during mitosis (\u003cspan citationid=\"CR78\" class=\"CitationRef\"\u003e78\u003c/span\u003e, \u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e79\u003c/span\u003e). ARID1A has also been shown to be excluded from chromatin during mitosis (\u003cspan citationid=\"CR79\" class=\"CitationRef\"\u003e79\u003c/span\u003e). The genes annotated to microtubule processes that are affected by the S363A mutation are enriched in SIN3A targets (ENCODE ChIP data from EnrichR), and their expression along the cell cycle is positively correlated to that of CCNA2 and inversely to the phosphorylation of ARID1A S363 (data not shown). This raises the possibility that S363 (de)phosphorylation may regulate SIN3A chromatin recruitment and/or maintenance. Notably, ARID1A S363 phosphorylation is upregulated in breast cancer (\u003cspan citationid=\"CR80\" class=\"CitationRef\"\u003e80\u003c/span\u003e). Incidentally, this residue is in a stretch of sequence that is missing in ARID1B, an ARID1A paralog with an antagonistic role in cell cycle control (\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eSome well documented ARID1A interactions were not identified by our PRISMA assay. There are many reasons why this can happen, such as interactions being dependent on the presence of PTMs, cell line or context-dependent, or simply of too low abundance to be detected by PRISMA. It is also possible that missing interactions are reliant on globular structures on ARID1A that are not recapitulated by the short peptides used in our array. We also noted that PRISMA binding profiles of known ARID1A interacting proteins that are involved in DNA damage did not agree with published binding regions. This could reflect a dependency on DNA damage signalling, including PTMs on either ARID1A or interacting proteins.\u003c/p\u003e \u003cp\u003eOverall, our study has uncovered new biology on ARID1A, provides new avenues for further mechanistic exploration of ARID1A and SWI/SNF function and could inform therapeutic strategies impacting ARID1A function. Furthermore, our work represents a useful resource more broadly to investigate sequence or structural features governing protein interactions.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eWe demonstrate direct binding of transcriptional repressor SIN3A to ARID1A through its Sin3 SLiM and recruitment to chromatin. We identify TOX4, CDK2 and CCNA2 as novel ARID1A interacting proteins and map their binding sites. We reveal a previously uncharacterised role for CDK2-dependent phosphorylation of ARID1A Ser 363 in the regulation of expression of microtubule factors and cell division. Our study provides novel mechanistic insight into ARID1A functions and highlights important ramifications for cancer. Our data represents a valuable resource to further study ARID1A and SWI/SNF biology, and more broadly, to investigate sequence and structural features governing protein interactions.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent for publication\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAvailability of data and materials\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRNA-sequencing datasets generated in this study are available in the GEO repository with accession number GSE313270. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (81) partner repository with dataset identifiers PXD076037 (PRISMA),\u0026nbsp;PXD074557 (chromatin) and PXD074744 (ARID1A_S363A proteome profiling). Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Mercedes Pardo ([email protected]). All unique/stable reagents generated in this study are available from the lead contact with a completed materials transfer agreement.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis work was supported by the Wellcome Trust (223745/Z/21/Z and 302297/Z/23/Z), BBSRC (BB/Y004477/1), Cancer Research UK (DRCRPG-June24/100004) and ICR core funding.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor\u0026rsquo;s contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMP and JC conceived the project. MP, CM, KAL, FS, LS and ZK performed experiments. MP, CM, and KAL performed data analysis. MW and FY contributed reagents. MP, CA, JAD and JC obtained funding and provided input and supervision. MP prepared the original draft. MP, CM, KAL, CA, JAD and JC contributed to review and editing. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe thank the Institute of Cancer Research Flow Cytometry Facility, Insect Cell Culture Facility and Scientific Computing Team for support.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResource availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRequests for further information and resources should be directed to and will be fulfilled by the lead contact, Mercedes Pardo ([email protected]). All unique/stable reagents generated in this study are available from the lead contact with a completed materials transfer agreement. RNA-sequencing datasets generated in this study are available in the GEO repository with accession number GSE313270. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (81) partner repository with dataset identifiers PXD076037 (PRISMA), PXD074557 (chromatin) and PXD074744 (ARID1A_S363A proteome profiling).\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e"},{"header":"STAR Methods","content":"\u003cp\u003e\u003cstrong\u003eCell culture\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ehTERT-RPE1 (RPE1, ATCC) were cultured in Dulbecco Modified Minimal Essential Medium (DMEM)/F-12 (Merck) supplemented with 10% FBS (Gibco), 200\u0026micro;M glutamax (Gibco), 0.26% sodium bicarbonate (Gibco), and 1% penicillin/streptomycin (P/S)(Merck). HEK293TN (Systems Biosciences) were cultured in DMEM supplemented with 10% FBS and 1% P/S. All cells were maintained at 37\u0026deg;C in a humified incubator with 5% CO\u003csub\u003e2\u003c/sub\u003e and were regularly tested for mycoplasma contamination.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePlasmids\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eEmerald-GFP (emGFP) and emGFP-tagged ARID1A, ARID1A S363A and ARID1A C-terminus (aa. 1009-2285) were amplified by PCR with Q5\u0026reg; High-Fidelity DNA Polymerase (M0491S, NEB) using pLenti-puro-ARID1A plasmid (Addgene #39478) as template and cloned into pCW57-MCS1-2A-MCS2 plasmid (Addgene #71782) through NEBuilder\u0026reg; HiFi DNA Assembly (NEB, E2621S). Plasmids were prepared using the Qiagen Plasmid Plus kit as per the manufacturer\u0026rsquo;s protocol and stored at -20\u003csup\u003eo\u003c/sup\u003eC. Correct nucleotide sequence was confirmed by whole plasmid sequencing (Plasmidsaurus). The lentiviral packaging and envelope plasmids psPAX2 and pMD2.G were gifts from Professor Didier Trono (Addgene plasmids #12259 and #12260). Oligos used are listed in Table S8.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCell line generation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor lentivirus production, 10 \u0026micro;g of the target plasmid was transfected into HEK293TN cells along with the lentiviral packaging and envelope plasmids, psPAX2 and pMD.2G, using Lipofectamine 3000 (Thermo) according to the manufacturer\u0026rsquo;s protocol. Viral particle-containing medium was harvested at 48 hr and 72 hr post transfection, combined, centrifuged for 5 min at 300 \u003cem\u003ex g\u003c/em\u003e to remove debris, and filtered through 0.45 \u0026micro;m cellulose acetate filters (Sartorius). The resulting supernatant was concentrated using the Lenti-X Concentrator (Takara Bio) according to the manufacturer\u0026rsquo;s protocol. For transduction, concentrated viral supernatant was added to RPE1 cells in 6-well plates along with 6 \u0026micro;g/mL polybrene (Merck). Cells were expanded to 10 cm\u003csup\u003e2\u003c/sup\u003e dishes, and 1 \u0026micro;g/mL DOX was added to cells to induce protein expression. After 24 hr, the pool of GFP-positive cells was sorted into 6-well plates by the BD FACSAria III sorter (BD Biosciences). Protein expression in the resulting cell lines was confirmed by Western blot analysis and immunofluorescence microscopy.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompounds\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDoxycycline hyclate (DOX, Merck) was dissolved in RNase-free H\u003csub\u003e2\u003c/sub\u003eO to a stock concentration of 1 mg/mL. 5-ethynyl-2\u0026apos;-deoxyuridine (EdU, Invitrogen) was dissolved in DMSO (Merck) to a stock concentration of 10 mM.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePeptide arrays\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCellulose acetate membranes containing immobilised PRISMA peptides spanning the whole length of ARID1A were purchased from JPT Peptide Technologies (Berlin, Germany). Peptide sequences are listed in Table S1.\u003c/p\u003e\n\u003cp\u003ePeptide arrays for the CDK2/CCNA2 validation experiments (Table S2) were prepared in-house. Peptides were synthesised on derivatised cellulose membranes using a Multipep1 instrument (CEM Microwave Technology Ltd.) by cycles of coupling of N(a)-Fmoc-protected aminoacids (Sigma-Aldrich) through carboxylic group activation with diisopropylcarbodiimide (Sigma-Aldrich) and Oxyma pure (CEM), followed by deprotection of alpha-amino groups by piperidine (Sigma-Aldrich), followed by acetylation of N-terminal residues with acetic anhydride (Sigma-Aldrich). All peptides were N-terminally acetylated. After peptide synthesis, side chain deprotection was performed as previously described (82).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFull length ARID1A PRISMA\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe PRISMA screen was carried out a single biological replicate with technical replicate MS measurements. Nuclear isolation was performed by lysing wild type RPE1 cells in an isotonic buffer (10 mM Tris-HCl, 10 mM NaCl, 1.5 mM MgCl2, 0.34 M sucrose containing 1 mM dithiothreitol (DTT) and Halt Protease inhibitors (Thermo) and 0.05% NP-40 as described previously (83). Nuclei were lysed by incubation for 10 mins on ice with PRISMA high salt lysis buffer (50 mM HEPES pH 8.0, 450 mM NaCl, 1 mM EDTA, 10% glycerol, 0.5% NP-40, 1 mM DTT, Halt protease inhibitors) followed by Dounce homogenisation. The nuclear lysate was incubated with universal nuclease (Pierce) for 1 hour on ice and then centrifuged at 13000 rpm for 15 min at 4\u0026deg;C. The cleared nuclear lysate was diluted 1:3 with PRISMA dilution buffer (50 mM HEPES pH 8.0, 1 mM EDTA, 10% glycerol, 0.5% NP-40, 0.3 mM DTT, Halt Protease inhibitors (Thermo). The PRISMA membrane was wetted for 15 mins in PRISMA detergent-free wash buffer (50 mM HEPES pH 8.0, 150 mM NaCl, 1 mM EDTA, 10% glycerol), followed by blocking with yeast tRNA at 1 mg/ml (Life Technologies) in the same buffer. The membrane was then washed with PRISMA low salt lysis buffer (50 mM HEPES pH 8.0, 150 mM NaCl, 1 mM EDTA, 10% glycerol, 0.5% NP-40) 5 times for 5 mins. The membrane was then incubated with 40 mg of RPE1 nuclear extract at ~1.4 mg/ml in the diluted lysis buffer for 1h at 4\u0026deg;C. After removal of the nuclear lysate, the membrane was first washed with PRISMA low salt lysis buffer 3 times for 5 mins, followed by PRISMA detergent-free wash buffer 3 times for 5 mins, and then air dried for 1h. The peptide spots were excised and placed one to a well in a 96-well HTS filter plate (Millipore) containing 200 \u0026mu;l of 5 mM TCEP (Sigma-Aldrich) in 50 mM ammonium bicarbonate. Excised membrane spots were incubated for 15 mins at 37\u0026deg;C with shaking to allow reduction, followed by alkylation with 10 mM iodoacetamide for 30 mins shaking at room temperature in the dark. TCEP and iodoacetamide were removed by spinning the plate at 1000 rpm for 1 min, and spots were washed 3 times with 50 mM ammonium bicarbonate. Samples were digested in the filter plate with 0.5\u0026nbsp;\u0026mu;g of trypsin (sequencing grade, Roche) and 0.25\u0026nbsp;\u0026mu;g of LysC (Wako Chemicals) per spot/well for 2 hours at 37\u0026deg;C, followed by overnight incubation at room temperature. Peptides were collected by centrifugation onto a fresh plate, and the filter plate stack was flushed twice with 50\u0026nbsp;\u0026mu;l of 100% HPLC methanol and the flowthrough pooled with the first peptide collection. The peptides were dried and then resuspended in 0.5% formic acid ready for MS analysis.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePRISMA mass spectrometry analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePeptides were analysed with online nanoLC-MS/MS on an Orbitrap Fusion Tribrid mass spectrometer coupled with an Ultimate 3000 RSLCnano System. Samples were first loaded and desalted on a nanotrap (100 \u0026micro;m id x 2 cm) (PepMap C18, 5 \u0026micro;m, 100A) at 10 \u0026micro;l/min with 0.1% formic acid for 10 min and then separated on an analytical column (75 \u0026micro;m id x 50 cm) (PepMap C18, 5 \u0026micro;m, 100A) over a 60 min linear gradient of 4-32% acetonitrile (ACN):0.1% formic acid at 300 nL/min. This was followed by a 5 min column wash with 85% ACN:0.1% formic acid on the analytical column and a 10 min trap-column wash of a cleaning mixture containing 25% ACN:25% Isopropanol:25% methanol: 25% water. The total cycle time was 90 min.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe Orbitrap Fusion was operated in standard data-dependent acquisition. For peptide analysis, the precursors were fragmented in HCD (higher collision dissociation) cell at 32% collision energy with a quadrupole isolation width of 0.7 Th (Thomson unit), detected in the ion trap at a rapid scan rate, AGC of 4\u0026thinsp;\u0026times;\u0026thinsp;10e5, and maximum injection time of 50\u0026nbsp;ms. Targeted precursors were dynamically excluded for further isolation and activation for 45 secs with 7 ppm mass tolerance. Peptide fragments were detected in the ion trap with a rapid scan rate setting. Technical replicate MS measurements were acquired for 50% of peptide spots.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePRISMA-MS data analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe PRISMA raw MS data were analysed with MaxQuant software v. 2.0.3.0 in one batch and searched against the human Uniprot reference database (v. October 2022). A maximum of 2 missed cleavages were allowed, variable modifications were set to methionine oxidation, pyro-glutamic acid conversion of asparagine and glutamine, and N-terminal protein acetylation, and cysteine carbamidomethylation was set as fixed modification. The label free quantification (LFQ) and match between runs options were activated. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (81) partner repository with the dataset identifier PXD076037.\u003c/p\u003e\n\u003cp\u003eFurther data processing was carried out using Perseus and Excel. Technical replicate measurements were integrated by averaging the LFQ intensity where two values were available, and the single value was used if only one was measured. Only proteins that were identified by two or more peptides were taken forward. Intensity values for the 228 peptides in each of the data rows were filtered according to the outlier filtering criterion, where the values were replaced by 0 if they were smaller than the 75\u003csup\u003eth\u003c/sup\u003e percentile of the intensity value distribution of the protein (32). For visualisation and clustering, each data row was normalised between 0 and 1. Proteins that showed a consecutive binding pattern across a minimum of two matrix peptides were marked as making up the \u0026ldquo;core interaction dataset\u0026rdquo;.\u003c/p\u003e\n\u003cp\u003ePeptide properties were calculated using the Peptides R package v.2.4.4 (84). Annotation of GO terms was carried out with Perseus v.2.0.7.0 using an annotation file generated from Uniprot in October 2022. GO term and protein domain enrichment was carried out using the DAVID functional annotation tool (85, 86). Venn diagrams were created with Venny 2.1 (bioinfogp.cnb.csic.es/tools/venny/).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePreparation of cell lysates\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor western blotting, cell pellets were resuspended in the appropriate volume of lysis buffer (10% glycerol, 50mM Tris-HCl (pH 7.4), 0.5% NP-40, 150mM NaCl) containing 0.25U/\u0026micro;L benzonase nuclease (Merck), 1x cOmplete\u0026trade; EDTA-free protease inhibitor cocktail (Roche), and 1x PhosSTOP\u0026trade; phosphatase inhibitor cocktail (Roche). Cells were lysed on ice for 45 minutes followed by centrifugation for 15 minutes at 16,000 \u003cem\u003ex g\u003c/em\u003e at 4\u003csup\u003eo\u003c/sup\u003eC. The resulting supernatant containing whole cell protein extracts was collected.\u003c/p\u003e\n\u003cp\u003eInsect cell pellets were lysed in 50 mM Tris-HCl pH 8, 150 mM NaCl, 0.1% NP40, 1mM EDTA, 1x Halt protease and phosphatase inhibitor (Pierce) and 1 mM DTT on ice for 10 min, homogenised using a Dounce homogeniser and centrifugation for 15 minutes at 16,000 \u003cem\u003ex g\u003c/em\u003e at 4\u003csup\u003eo\u003c/sup\u003eC.\u003c/p\u003e\n\u003cp\u003eFor TOX4 immunoprecipitation, cell lysates were prepared as previously described (87).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eImmunoprecipitation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRelevant antibodies and non-relevant IgG control (normal)antibodies (1\u0026ndash;10\u0026thinsp;\u0026mu;g) were coupled to 25-50 \u0026micro;l of Protein G Dynabeads (Life Technologies) following manufacturer\u0026rsquo;s protocols. Antibodies are listed in Table S7. Antibody-coupled beads were incubated with whole cell or nuclear lysates for 1\u0026ndash;2\u0026thinsp;h at 4\u0026thinsp;\u0026deg;C, lysates were removed, and beads were washed with IPP150 buffer (150 mM NaCl, 10 mM Tris-HCl, 0.1% NP-40). Immunoprecipitated proteins were eluted by incubating the beads in 1x LDS sample loading buffer (Life Technologies) containing 50 mM DTT at 70\u0026thinsp;\u0026deg;C for 10\u0026thinsp;min. The cell lysate (input) and eluted samples were analysed by western blotting.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eWestern blotting and Far Western on peptide arrays\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eProtein samples were separated in NuPAGE 4-12% BisTris or 3-8% Tris Acetate gels using MOPS or Tris Acetate running buffer (Life Technologies) respectively and transferred to nitrocellulose membranes (Cytiva). Blocking was performed with 5% milk in Tris Buffered Saline-0.1% Tween (TTBS) for 1 h or 10% milk in TTBS for 10 mins. Membranes were incubated in the appropriate primary antibody diluted in 5% milk blocking buffer at 4\u0026deg;C overnight, washed 3 times with TTBS buffer, and incubated in the appropriate horseradish peroxidase (HRP)-conjugated secondary antibody diluted in blocking buffer for 1 hour at room temperature. Blots were developed with ECL Prime, ECL Ultra (Cytiva) or Immobilon Forte (Merck) and chemiluminescence signal acquired on an iBright CL750 imager (Life Technologies).\u003c/p\u003e\n\u003cp\u003eFor far western experiments, membranes containing peptide spots were first incubated in TTBS for 15 min, then blocked with 5% milk in TTBS for 1 h. Cyclin A (Bio-techne) at 5 ug/ml or CDK2/CCNA3[174-432] at 10 ug/ml in blocking solution were added to the membrane and incubated for 4 h at room temperature. Primary and secondary antibody incubations were performed at the corresponding dilutions in 5% milk in TTBS. Blots were developed with ECL Prime (Cytiva) and chemiluminescence signal acquired on an iBright used CL750 imager (Life Technologies). Antibodies are listed in Table S7.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eExpression of proteins in insect cells\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCodon-optimised human SIN3A (residues 449-1086) fused to StrepII tag and ARID1A-V5 (from Addgene plasmid #39311) cDNAs for insect cell expression were synthesised and subcloned into the pFastBac1 vector (GenScript). The SIN3A and ARID1A constructs were transformed into DH10MultiBac cells for bacmid generation, and bacmid DNAs were isolated by isopropanol precipitation. Bacmid DNAs were transfected into Sf9 (\u003cem\u003eSpodoptera frugiperda\u003c/em\u003e) cells using CellFectinII (Thermo), and baculovirus stocks were amplified through two consecutive rounds of infection in Sf9 cells. For protein expression, Sf9 cells at a density of 1.5 x 10e6 cells/mL in Sf-900 III medium supplemented with penicillin/streptomycin were infected with the amplified virus at 1:30 (v/v). After 72 h incubation at 27\u0026deg;C with shaking at 130 RPM, cells were harvested and flash-frozen in liquid nitrogen for long term storage at -80\u0026deg;C until processing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePurification of CDK2 and cyclin A3\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBoth CDK2 (full length) and cyclin A3 (residues 174\u0026ndash;432) were induced and expressed individually with 1 mM IPTG in BL21(DE3) Star cells at 18\u0026thinsp;\u0026deg;C overnight. Pellets containing GST-tagged CDK2 and His-tagged cyclin A3 were combined (two litres of Cdk2 cell culture and 2 litres of cyclin A3 cell culture) and resuspended in lysis buffer (50 mM Hepes pH8.0, 150 mM NaCl, 2 mM MgCl\u003csub\u003e2\u003c/sub\u003e, 5 % glycerol, 0.5 mM TCEP) supplemented with 5 units per ml of benzonase and cOmplete\u0026trade; EDTA-free protease inhibitors (Roche). After sonication of the cell suspension, the cell lysate was centrifuged at 21,000\u0026thinsp;rpm. for 1\u0026thinsp;h at 4\u0026thinsp;\u0026deg;C and the supernatant was loaded onto pre-equilibrated 5 ml GSTrap\u0026trade; HP column (Merck). The column was washed with lysis buffer and then the proteins were eluted out with elution buffer (20 mM reduced glutathione added into the lysis buffer). The eluted proteins were buffer exchanged into lysis buffer to remove reduced glutathione and GST-tag of CDK2 was then cleaved off with 3C PreScission protease overnight at 4\u0026thinsp;\u0026deg;C. After 3C protease cleavage, the sample was loaded onto 5 ml GSTrap\u0026trade; HP column again, the flow-through was collected, concentrated and loaded onto pre-equilibrated size exclusion column (Superdex75 16/60 column, GE Healthcare). The protein complex was eluted out with gel filtration buffer (20 mM Hepes pH 8.0, 150 mM NaCl, 0.5 mM TCEP) and concentrated to final concentration 1.4 mg/ml. The purified protein was flash frozen and stored in -80 \u0026deg;C.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eChromatin extraction\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCell pellets were resuspended in isotonic nuclear extraction buffer for chromatin extraction (15 mM Tris-HCL pH 7.5, 60 mM KCl ,15 mM NaCl, 5 mM MgCl2, 1 mM CaCl2, 250 mM Sucrose, 1 mM DTT, Halt Protease inhibitors) and 0.3% NP-40 was added dropwise to final concertation of 0.03% (Sigma-Aldrich), mixed gently and incubated on ice for 5 mins. Nuclei were pelleted at 400-600 rcf for 5 mins at 4\u0026deg;C. Nuclei were washed again with nuclear isolation buffer for chromatin extractions. Small fraction was verified to contain nuclei by light microscopy. Isolated nuclei were resuspended in hypotonic chromatin isolation buffer (3 mM EDTA, 0.2 mM ethylene glycol tetraacetic acid, 1 mM DTT, Halt Protease inhibitors (Thermo) and incubated\u0026nbsp;on ice for 1 h, vortexing every 5-10 mins. Insoluble chromatin fraction was collected by centrifugation for 4 min at 1,700 rcf. Fraction was washed 2 times in hypotonic chromatin isolation buffer. Chromatin fractions were frozen on dry ice and stored at -20\u0026deg;C. Samples were re-suspended in the chromatin preparation buffer (100 mM Triethylamonium bicarbonate (TEAB) (Sigma-Aldrich), 1% sodium deoxycholate (SDC), 10% isopropanol, 50 mM NaCl, Halt Protease and Phosphatase inhibitor (Thermo) supplemented with and Benzonaze (Pierce). The chromatin pellet was disrupted with brief probe sonication and the samples were incubated on ice for 30 min. 30 \u0026mu;g of each sample was labelled TMTpro\u0026trade; 16plex Label Reagent Set (Thermo) according to manufactures instructions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eChromatin profiling mass spectrometry\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eOffline peptide fractionation was based on high pH reverse-phase (RP) chromatography using the Waters XBridge C18 column (2.1 mm \u0026times; 150 mm, 3.5 \u0026mu;m) on a Dionex UltiMate 3000 HPLC system at a flow rate of 0.2 ml/min. Mobile phase A was 0.1% (v/v) ammonium hydroxide, and mobile phase B was acetonitrile, 0.1% (v/v) ammonium hydroxide. Pooled TMT-peptides were resuspended in 200 \u0026mu;L of buffer A, centrifuged at 14 000 rpm for 5 min, and the supernatant was injected for fractionation with the following gradient: isocratic for\u0026nbsp;5 min at 5% phase B, gradient for 40 min to 35% phase B, gradient to 80% phase B in 5 min, isocratic for 5 min, and re-equilibrated to 5% phase B. The peptides were then resuspended with 0.1% formic acid and finally pooled into four fractions by combining the first and the last fractions.\u003c/p\u003e\n\u003cp\u003eLC-MS analysis was performed on the Dionex UltiMate 3000 UHPLC system coupled with the LTQ Orbitrap Lumos mass spectrometer (Thermo). Samples were analysed with the EASY-Spray C18 capillary column (75 \u0026mu;m \u0026times; 50 cm, 2 \u0026mu;m) at 50 \u0026deg;C. Mobile phase A was 0.1% formic acid and mobile phase B was 80% acetonitrile, 0.1% formic acid. The gradient separation method was as follows: 150 min gradient up to 38% B, for 10 min up to 95% B, for 5 min isocratic at 95% B, re-equilibration to 5% B in 10 min, for 10 min isocratic at 5% B.\u003c/p\u003e\n\u003cp\u003eMS1 scans were acquired at a resolution of 120,000 over an m/z range of 375\u0026ndash;1500 in the Orbitrap, with an AGC target of 4 \u0026times; 10⁵ and a maximum injection time of 50 ms. Precursors were selected using Top Speed mode with a 3 s cycle time and isolated with a quadrupole isolation width of 0.7 Th. Collision energy for fragmentation via collision induced dissociation (CID) was set at 35%, with AGC target of 1 \u0026times; 10⁴ and maximum injection time of 50 ms. MS3 spectra for TMT-based quantitation were acquired using synchronous precursor selection (SPS) of the top seven CID fragment ions. SPS-selected ions were fragmented by HCD with a collision energy of 35%. MS3 spectra were acquired in the Orbitrap over an m/z range of 100\u0026ndash;500 at a resolution of 45,000, with an AGC target of 1 \u0026times; 10⁵ and maximum injection time of 105 ms. Dynamic exclusion was activated for 45 s with a mass tolerance of 7 ppm.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eChromatin profiling data analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMS raw data from IPs was analysed using Proteome Discoverer 2.4 (Thermo). Sequest search engine was used for protein identification and quantification followed by Percolator. Peptides were filtered at q-value \u0026lt;1% based on a decoy database search. The searches were performed against the full UniProt Human database (reference entries, January 2022) together with the cRAP contaminant database (ftp://ftp.thegpm.org/fasta/cRAP). Precursor mass tolerance was set to 20 ppm. Fragment mass tolerance was set to 0.5 Da. TMTpro (K and peptide N-terminus) and carbamidomethylation (C) were set as static modifications, while oxidation (M) and deamidation (NQ) were set as dynamic modifications.\u0026nbsp;Protein\u0026nbsp;abundances were normalised by multiplying each value by the maximum median of abundances across all channels and then dividing by the median of abundances of the corresponding channel. The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (81) partner repository with the dataset identifier PXD074557.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eIncucyte proliferation assay and cell fate profiling\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCells were seeded in triplicate into 96-well plates at a density of 1,000 cells/well, and allowed to attach and grow for 24 hours. When indicated, 1 \u0026micro;g/mL DOX was then added to cells, and plates were transferred to the Incucycte SX5 (Sartorius) maintained at 37\u0026deg;C in a humidified 5% CO2 atmosphere. For tracking cell growth, phase contrast and fluorescence images (4 images per well) were collected every 4 hours up to a total of 120 hours. The average of three technical replicates was used, and % confluence (% of area covered by cells) was used as a readout of proliferation. For cell fate profiling, phase contrast and fluorescence images (4 images per well) were collected every 10 minutes up to a total of 72 hours. Image sequences were exported, and individual cells were analysed manually to generate cell fate profiles.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSulforhodamine B (SRB) proliferation assay\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCells were seeded in triplicate into 96-well plates at a density of 1,000 cells/well, and allowed to attach and grow for 24 hours. When indicated, 1 \u0026micro;g/mL DOX was then added to cells and cells were left to incubate for 6 days in total. After this, cells were fixed by the addition of 100 \u0026micro;l 10% TCA (Merck) and incubated at 4\u0026deg;C for at least 24 hours. Plates were then washed five times with water and allowed to completely dry before staining with 0.57% SRB powder (Merck) dissolved in 1% acetic acid and incubated for 1 hour room temperature. Unbound dye was removed by washing with 1% acetic acid, and the plate was allowed to dry completely. Protein-bound dye was solubilised in 10 mM Tris-HCl pH 9.5 for 30 minutes at room temperature with gentle rocking. Absorbance at 565 nm was quantified using the SpectraMax Microplate Absorbance Reader (PerkinElmer). The average of three technical replicates was used, and proliferation was defined as a percentage of the absorbance measured in treated wells compared with control wells.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRNA-sequencing\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFor RNA-sequencing, cells were treated with 1 \u0026micro;g/mL DOX for 24 hours before harvesting. Pellets were harvested by scraping cells in ice-cold PBS following by centrifugation at 7,500 \u003cem\u003ex g\u003c/em\u003e for 10 minutes. Three independent biological replicates were harvested for all cell lines. Pellets were resuspended in 500 \u0026micro;L TRIzol reagent and total RNA was extracted using a Direct-zol RNA miniprep kit (Zymo Research) according to the manufacturer\u0026rsquo;s protocol. RNA concentration and quality was confirmed with the High Sensitivity RNA ScreenTape (Agilent), using the TapeStation 4150 System (Agilent). Library preparation and sequencing was performed by Novogene Corporation Ltd. Novogene NGS Stranded RNA Library Prep Set was used to generate 250-300 bp insert strand specific libraries, and ribosomal RNA was removed using TruSeq Stranded Total RNA Library Prep. 50 million 150 bp paired-end reads were sequenced on an Illumina NovaSeq 6000.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRNA-sequencing analysis\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eFastq reads were checked using FastQC (v0.11.9) (88) and trimmed using TrimGalore (v0.6.6) (89). Residual ribosomal RNA reads were removed using Ribodetector with -e norRNA setting (v0.2.7) and strandedness was detected using RSeQC (v4.0.0). Reads were aligned to the human hg38.14 genome (GenBank GCA_000001405.29) using STAR alignment software (v2.7.6a) and reads mapping to genes were quantified using HTSeqCount (v0.12.4). Differential analysis of gene expression was calculated in R using \u003cem\u003eDESeq2\u003c/em\u003e (v1.38.3).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eWhole cell proteomic profiling\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCell pellets were lysed in buffer containing 1% (w/v) sodium deoxycholate (SDC), 100 mM triethylammonium bicarbonate (TEAB), 10% (v/v) isopropanol, and 50 mM NaCl, freshly supplemented with 5 mM TCEP (Thermo, Bond-Breaker), 10 mM iodoacetamide (IAA), universal nuclease (1:2000, v/v; Pierce, #88700), and Halt protease and phosphatase inhibitor cocktail (1\u0026times;; Thermo, #78442). Lysis was assisted by 5 min of bath sonication. Protein concentration was determined using the Quick Start Bradford protein assay (Bio-Rad). Aliquots containing 30 \u0026micro;g of total protein were digested overnight at room temperature with trypsin (Pierce) at a 1:20 (enzyme:substrate) ratio. Peptides were labelled with TMTpro reagents (Thermo) by adding 5 \u0026micro;L of reagent to 15 \u0026micro;L of peptide solution. Following labelling, samples were acidified to a final concentration of 2% formic acid, and precipitated SDC was removed by centrifugation.\u003c/p\u003e\n\u003cp\u003eThe combined TMT-labelled peptide mixture was fractionated by high-pH reversed-phase chromatography using an XBridge C18 column (2.1 \u0026times; 150 mm, 3.5 \u0026micro;m; Waters) on an UltiMate 3000 HPLC system over a 35-min gradient at 1% slope. Mobile phase A consisted of 0.1% (v/v) ammonium hydroxide, and mobile phase B consisted of 0.1% (v/v) ammonium hydroxide in acetonitrile. LC\u0026ndash;MS analysis was performed on a Vanquish Neo UHPLC system coupled to an Orbitrap Ascend mass spectrometer (Thermo). Peptides were separated on a 25 cm capillary column (nanoEase MZ PST BEH130 C18, 1.7 \u0026micro;m, 75 \u0026micro;m \u0026times; 250 mm; Waters) using a 110-min gradient from 5\u0026ndash;35% mobile phase B (80% acetonitrile, 0.1% formic acid). Samples were first loaded onto a PepMap 100 C18 trapping column (5 \u0026micro;m, 0.3 \u0026times; 5 mm, 1500 bar) before separation on the analytical column, which was coupled to a stainless-steel emitter via a Nanospray Flex ion source.\u003c/p\u003e\n\u003cp\u003eMS1 spectra were acquired in the Orbitrap at a resolution of 120,000. Precursors were selected for HCD fragmentation in top-speed mode (3 s cycle time) with a normalized collision energy of 32% and detected in the ion trap using turbo scan rate. MS3 spectra were triggered by real-time search (RTS) against a FASTA database containing reviewed UniProt \u003cem\u003eHomo sapiens\u003c/em\u003e canonical and isoform sequences, using multi-notch isolation (10 notches) and HCD fragmentation at 65% normalized collision energy, with detection in the Orbitrap at 90,000 resolution. Dynamic exclusion was enabled for 45 s with a mass tolerance of 10 ppm, RTS close-out was set to a maximum of four peptides per protein, and targeted precursors were excluded from further activation. Static modifications used for RTS included TMTpro16plex on peptide N-termini and lysine residues and carbamidomethylation of cysteine. Variable modifications included oxidation of methionine and deamidation of asparagine and glutamine, with a maximum of one missed cleavage and up to two variable modifications per peptide.\u003c/p\u003e\n\u003cp\u003eRaw data were searched using the Sequest HT and Comet nodes implemented in Proteome Discoverer version 3.0 (Thermo) against a reviewed UniProt Homo sapiens FASTA database. Precursor mass tolerance was set to 20 ppm, and fragment ion mass tolerance was set to 0.5 Da for Sequest HT and 1 Da for Comet, allowing up to two missed tryptic cleavages. TMTpro modifications on peptide N-termini and lysine residues, as well as carbamidomethylation of cysteine, were specified as static modifications, while oxidation of methionine and deamidation of asparagine/glutamine were specified as dynamic modifications. Peptide-spectrum matches were filtered using Percolator with a peptide-level false discovery rate (FDR) of 1% based on a target\u0026ndash;decoy strategy. Only unique peptides were used for protein quantification, considering protein groups for peptide uniqueness, and peptides with an average reporter ion signal-to-noise ratio greater than 3 were included in quantitative analysis. Data are available via ProteomeXchange with identifier PXD074744.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eBioinformatics analysis of transcriptomics and proteomics data\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePeptide properties were calculated using the Peptides R package v.2.4.4 (84). GO term and protein domain enrichment analysis of the ARID1A PRISMA core dataset was carried out using the DAVID functional annotation tool (86). Gene set enrichment analyses for whole cell proteomic profiling and RNA-seq data were performed using the 1D annotation enrichment algorithm (90) within the Perseus software platform (91). GSEA plots were generated in R using the \u003cem\u003efgsea\u003c/em\u003e package (92).\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eHarrod A, Lane KA, Downs JA. The role of the SWI/SNF chromatin remodelling complex in the response to DNA double strand breaks. DNA Repair (Amst). 2020;93:102919.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchick S, Grosche S, Kohl KE, Drpic D, Jaeger MG, Marella NC, et al. Acute BAF perturbation causes immediate changes in chromatin accessibility. Nat Genet. 2021;53:269\u0026ndash;78.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCentore RC, Sandoval GJ, Soares LMM, Kadoch C, Chan HM. Mammalian SWI/SNF Chromatin Remodeling Complexes: Emerging Mechanisms and Therapeutic Strategies. Trends Genet. 2020;36:936\u0026ndash;50.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBlumli S, Wiechens N, Wu MY, Singh V, Gierlinski M, Schweikert G, et al. Acute depletion of the ARID1A subunit of SWI/SNF complexes reveals distinct pathways for activation and repression of transcription. Cell Rep. 2021;37:109943.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDykhuizen EC, Hargreaves DC, Miller EL, Cui K, Korshunov A, Kool M, et al. BAF complexes facilitate decatenation of DNA by topoisomerase IIalpha. Nature. 2013;497:624\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSim\u0026otilde;es-Sousa S, Amin N, Lane KA, Harrod A, Pedersen M, Pardo M, et al. ARID1A stabilizes non-homologous end joining factors at DNA breaks induced by the G4 ligand pyridostatin. Cell Rep. 2025;44:116277.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVarga J, Kube M, Luck K, Schick S. The BAF chromatin remodeling complexes: structure, function, and synthetic lethalities. Biochem Soc Trans. 2021;49:1489\u0026ndash;503.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMashtalir N, D'Avino AR, Michel BC, Luo J, Pan J, Otto JE, et al. Modular Organization and Assembly of SWI/SNF Family Chromatin Remodeling Complexes. Cell. 2018;175:1272-88 e20.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHe S, Wu Z, Tian Y, Yu Z, Yu J, Wang X, et al. Structure of nucleosome-bound human BAF complex. Science. 2020;367:875\u0026ndash;81.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchick S, Rendeiro AF, Runggatscher K, Ringler A, Boidol B, Hinkel M, et al. Systematic characterization of BAF mutations provides insights into intracomplex synthetic lethalities in human cancers. Nat Genet. 2019;51:1399\u0026ndash;410.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSantamarina JB, Astier C, Garrido M, Daage LC, Roumeliotis TI, Anthony E, et al. Integrated multiomic profiling reveals SWI/SNF subunit-specific pathway alterations and targetable vulnerabilities. bioRxiv 2024; doi:\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1101/2024.07.16.603530\u003c/span\u003e\u003cspan address=\"10.1101/2024.07.16.603530\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoumeliotis TI, Williams SP, Goncalves E, Alsinet C, Del Castillo Velasco-Herrera M, Aben N, et al. Genomic Determinants of Protein Abundance Variation in Colorectal Cancer Cells. Cell Rep. 2017;20:2201\u0026ndash;14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKanno S-i, Kobayashi T, Watanabe R, Kurimasa A, Tanaka K, Yasui A, et al. Armadillo domain of ARID1A directly interacts with DNA-PKcs to couple chromatin remodeling with nonhomologous end joining (NHEJ) pathway. Nucleic Acids Research. 2025;53:gkaf150.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUlicna L, Kimmey SC, Weber CM, Allard GM, Wang A, Bui NQ, et al. The Interaction of SWI/SNF with the Ribosome Regulates Translation and Confers Sensitivity to Translation Pathway Inhibitors in Cancers with Complex Perturbations. Cancer Res. 2022;82:2829\u0026ndash;37.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChang L, Azzolin L, Di Biagio D, Zanconato F, Battilana G, Lucon Xiccato R, et al. The SWI/SNF complex is a mechanoregulated inhibitor of YAP and TAZ. Nature. 2018;563:265\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang Y, Chen X, Qiao X, Xie Y, Guo D, Li B, et al. Chromatin Remodelling Molecule ARID1A Determines Metastatic Heterogeneity in Triple-Negative Breast Cancer by Competitively Binding to YAP. Cancers (Basel). 2023;15:2447.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFontana B, Gallerani G, Salamon I, Pace I, Roncarati R, Ferracin M. ARID1A in cancer: Friend or foe? Frontiers in Oncology. 2023;13:1136248.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHodges C, Kirkland JG, Crabtree GR. The Many Roles of BAF (mSWI/SNF) and PBAF Complexes in Cancer. Cold Spring Harb Perspect Med. 2016;6:a026930.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eB\u0026ouml;gershausen N, Wollnik B. Mutational Landscapes and Phenotypic Spectrum of SWI/SNF-Related Intellectual Disability Disorders. Frontiers in Molecular Neuroscience. 2018;11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKadoch C, Hargreaves DC, Hodges C, Elias L, Ho L, Ranish J, et al. Proteomic and bioinformatic analysis of mammalian SWI/SNF complexes identifies extensive roles in human malignancy. Nat Genet. 2013;45:592\u0026ndash;601.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi J, Wang W, Zhang Y, Cieslik M, Guo J, Tan M, et al. Epigenetic driver mutations in ARID1A shape cancer immune phenotype and immunotherapy. J Clin Invest. 2020;130:2712\u0026ndash;26.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAlfert A, Moreno N, Kerl K. The BAF complex in development and disease. Epigenetics Chromatin. 2019;12:19.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEl Hadidy N, Uversky VN. Intrinsic Disorder of the BAF Complex: Roles in Chromatin Remodeling and Disease Development. Int J Mol Sci. 2019;20:5260.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTompa P, Davey NE, Gibson TJ, Babu MM. A million peptide motifs for the molecular biologist. Mol Cell. 2014;55:161\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBah A, Forman-Kay JD. Modulation of Intrinsically Disordered Protein Function by Post-translational Modifications. J Biol Chem. 2016;291:6696\u0026ndash;705.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDavis RB, Kaur T, Moosa MM, Banerjee PR. FUS oncofusion protein condensates recruit mSWI/SNF chromatin remodeler via heterotypic interactions between prion-like domains. Protein Sci. 2021;30:1454\u0026ndash;66.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim YR, Joo J, Lee HJ, Kim C, Park JC, Yu YS, et al. Prion-like domain mediated phase separation of ARID1A promotes oncogenic potential of Ewing's sarcoma. Nat Commun. 2024;15:6569.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePatil A, Strom AR, Paulo JA, Collings CK, Ruff KM, Shinn MK, et al. A disordered region controls cBAF activity via condensation and partner recruitment. Cell. 2023;186:4936-55.e26.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChan YS, Gao Q, Robinson SA, Wang W, Filandrova R, Weinhold LM, et al. β-catenin functions as a molecular adapter for disordered cBAF interactions. Mol Cell. 2025;85:3041-56.e9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVan Roey K, Uyar B, Weatheritt RJ, Dinkel H, Seiler M, Budd A, et al. Short linear motifs: ubiquitous and functionally diverse protein interaction modules directing cell regulation. Chem Rev. 2014;114:6733\u0026ndash;78.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMeyer K, Selbach M. Peptide-based Interaction Proteomics. Mol Cell Proteomics. 2020;19:1070\u0026ndash;5.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDittmar G, Hernandez DP, Kowenz-Leutz E, Kirchner M, Kahlert G, Wesolowski R, et al. PRISMA: Protein Interaction Screen on Peptide Matrix Reveals Interaction Footprints and Modifications- Dependent Interactome of Intrinsically Disordered C/EBPbeta. iScience. 2019;13:351\u0026ndash;70.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eM\u0026eacute;sz\u0026aacute;ros B, Erdos G, Doszt\u0026aacute;nyi Z. IUPred2A: context-dependent prediction of protein disorder as a function of redox state and protein binding. Nucleic Acids Res. 2018;46:W329-w337.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang Z, Chen K, Jia Y, Chuang JC, Sun X, Lin YH, et al. Dual ARID1A/ARID1B loss leads to rapid carcinogenesis and disruptive redistribution of BAF complexes. Nat Cancer. 2020;1:909\u0026ndash;22.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLuo Q, Wu X, Chang W, Zhao P, Nan Y, Zhu X, et al. ARID1A prevents squamous cell carcinoma initiation and chemoresistance by antagonizing pRb/E2F1/c-Myc-mediated cancer stemness. Cell Death Differ. 2020;27:1981\u0026ndash;97.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLinden M, Thomsen C, Grundevik P, Jonasson E, Andersson D, Runnberg R, et al. FET family fusion oncoproteins target the SWI/SNF chromatin remodeling complex. EMBO Rep. 2019;20:e45766.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKumar M, Michael S, Alvarado-Valverde J, M\u0026eacute;sz\u0026aacute;ros B, S\u0026aacute;mano-S\u0026aacute;nchez H, Zeke A, et al. The Eukaryotic Linear Motif resource: 2022 release. Nucleic Acids Res. 2022;50:D497-d508.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNagl NG, Jr., Wang X, Patsialou A, Van Scoy M, Moran E. Distinct mammalian SWI/SNF chromatin remodeling complexes with opposing roles in cell-cycle control. EMBO J. 2007;26:752\u0026ndash;63.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKuzmichev A, Zhang Y, Erdjument-Bromage H, Tempst P, Reinberg D. Role of the Sin3-histone deacetylase complex in growth regulation by the candidate tumor suppressor p33(ING1). Mol Cell Biol. 2002;22:835\u0026ndash;48.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUnderhill C, Qutob MS, Yee SP, Torchia J. A novel nuclear receptor corepressor complex, N-CoR, contains components of the mammalian SWI/SNF complex and the corepressor KAP-1. J Biol Chem. 2000;275:40463\u0026ndash;70.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSif S, Saurin AJ, Imbalzano AN, Kingston RE. Purification and characterization of mSin3A-containing Brg1 and hBrm chromatin remodeling complexes. Genes Dev. 2001;15:603\u0026ndash;18.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePal S, Yun R, Datta A, Lacomis L, Erdjument-Bromage H, Kumar J, et al. mSin3A/histone deacetylase 2- and PRMT5-containing Brg1 complex is involved in transcriptional repression of the Myc target gene cad. Mol Cell Biol. 2003;23:7475\u0026ndash;87.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKim M, Lu F, Zhang Y. Loss of HDAC-Mediated Repression and Gain of NF-κB Activation Underlie Cytokine Induction in ARID1A- and PIK3CA-Mutation-Driven Ovarian Cancer. Cell Rep. 2016;17:275\u0026ndash;88.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSuryo Rahmanto Y, Jung JG, Wu RC, Kobayashi Y, Heaphy CM, Meeker AK, et al. Inactivating ARID1A Tumor Suppressor Enhances TERT Transcription and Maintains Telomere Length in Cancer Cells. J Biol Chem. 2016;291:9690\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSahu SC, Swanson KA, Kang RS, Huang K, Brubaker K, Ratcliff K, et al. Conserved Themes in Target Recognition by the PAH1 and PAH2 Domains of the Sin3 Transcriptional Corepressor. Journal of Molecular Biology. 2008;375:1444\u0026ndash;56.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuan B, Wang TL, Shih Ie M. ARID1A, a factor that promotes formation of SWI/SNF-mediated chromatin remodeling, is a tumor suppressor in gynecologic cancers. Cancer Res. 2011;71:6718\u0026ndash;27.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbe Y, Rozqie R, Matsumura Y, Kawamura T, Nakaki R, Tsurutani Y, et al. JMJD1A is a signal-sensing scaffold that regulates acute chromatin dynamics via SWI/SNF association for thermogenesis. Nat Commun. 2015;6:7052.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu Z, Wu A, Wu Z, Wang T, Pan Y, Li B, et al. TOX4 facilitates promoter-proximal pausing and C-terminal domain dephosphorylation of RNA polymerase II in human cells. Commun Biol. 2022;5:300.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHornbeck PV, Zhang B, Murray B, Kornhauser JM, Latham V, Skrzypek E. PhosphoSitePlus, 2014: mutations, PTMs and recalibrations. Nucleic Acids Res. 2015;43:D512-20.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596:583\u0026ndash;9.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChi Y, Welcker M, Hizli AA, Posakony JJ, Aebersold R, Clurman BE. Identification of CDK2 substrates in human cell lysates. Genome Biol. 2008;9:R149.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrown NR, Noble MEM, Endicott JA, Johnson LN. The structural basis for specificity of substrate and recruitment peptides for cyclin-dependent kinases. Nature Cell Biology. 1999;1:438\u0026ndash;43.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNagarajan S, Rao SV, Sutton J, Cheeseman D, Dunn S, Papachristou EK, et al. ARID1A influences HDAC1/BRD4 activity, intrinsic proliferative capacity and breast cancer treatment response. Nat Genet. 2020;52:187\u0026ndash;97.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShen J, Peng Y, Wei L, Zhang W, Yang L, Lan L, et al. ARID1A Deficiency Impairs the DNA Damage Checkpoint and Sensitizes Cells to PARP Inhibitors. Cancer Discov. 2015;5:752\u0026ndash;67.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRega C, Tsitsa I, Roumeliotis TI, Krystkowiak I, Portillo M, Yu L, et al. High resolution profiling of cell cycle-dependent protein and phosphorylation abundance changes in non-transformed cells. Nat Commun. 2025;16:2579.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSandhya S, Maulik A, Giri M, Singh M. Domain architecture of BAF250a reveals the ARID and ARM-repeat domains with implication in function and assembly of the BAF remodeling complex. PLoS One. 2018;13:e0205267.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMehrvarz Sarshekeh A, Alshenaifi J, Roszik J, Manyam GC, Advani SM, Katkhuda R, et al. ARID1A Mutation May Define an Immunologically Active Subgroup in Patients with Microsatellite Stable Colorectal Cancer. Clinical Cancer Research. 2021;27:1663\u0026ndash;70.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePerez JM, Ryu J, Khan H, Shetty M, Parker E, D'Arcy P, et al. Proteomic Analysis of ARID1A-Deficient Ovarian Clear Cell Carcinoma Cells Reveals Differential Mitochondria ETC Subunit Abundances and Targetable Mitochondrial Pathways. Int J Mol Sci. 2025;26:5466.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLissanu Deribe Y, Sun Y, Terranova C, Khan F, Martinez-Ledesma J, Gay J, et al. Mutations in the SWI/SNF complex induce a targetable dependence on oxidative phosphorylation in lung cancer. Nat Med. 2018;24:1047\u0026ndash;57.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBakr A, Corte GD, Veselinov O, Kelek\u0026ccedil;i S, Chen MM, Lin YY, et al. ARID1A regulates DNA repair through chromatin organization and its deficiency triggers DNA damage-mediated anti-tumor immune response. Nucleic Acids Res. 2024;52:5698\u0026ndash;719.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLane KA, Harrod A, Wu L, Roumeliotis TI, Feng H, Foo S, et al. PBRM1 directs PBAF to pericentromeres and protects centromere integrity. Nat Commun. 2025;16:1980.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRamberger E, Sapozhnikova V, Kowenz-Leutz E, Zimmermann K, Nicot N, Nazarov PV, et al. PRISMA and BioID disclose a motifs-based interactome of the intrinsically disordered transcription factor C/EBPalpha. iScience. 2021;24:102686.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSandmann C-L, Schulz JF, Ruiz-Orera J, Kirchner M, Ziehm M, Adami E, et al. Evolutionary origins and interactomes of human, young microproteins and small peptides translated from short open reading frames. Molecular Cell. 2023;83:994\u0026ndash;1011. e18.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSuarez-Artiles L, Breiderhoff T, Girardello R, Gonschior H, Rodius S, Lesur A, et al. Pan-claudin family interactome analysis reveals shared and specific interactions. Cell Rep. 2022;41:111588.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBui C-B, Le HK, Vu DM, Truong K-DD, Nguyen NM, Ho MAN, et al. ARID1A-SIN3A drives retinoic acid-induced neuroblastoma differentiation by transcriptional repression of TERT. Molecular Carcinogenesis. 2019;58:1998\u0026ndash;2007.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKemper JK, Kim H, Miao J, Bhalla S, Bae Y. Role of an mSin3A-Swi/Snf chromatin remodeling complex in the feedback repression of bile acid biosynthesis by SHP. Mol Cell Biol. 2004;24:7707\u0026ndash;19.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDi Fiore B, Davey NE, Hagting A, Izawa D, Mansfeld J, Gibson TJ, et al. The ABBA motif binds APC/C activators and is shared by APC/C substrates and regulators. Dev Cell. 2015;32:358\u0026ndash;72.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVanheer L, Song J, De Geest N, Janiszewski A, Talon I, Provenzano C, et al. Tox4 modulates cell fate reprogramming. Journal of Cell Science. 2019;132: jcs232223.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNiu H, Wang H. TOX regulates T lymphocytes differentiation and its function in tumor. Frontiers in Immunology. 2023;14:990419.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePuch CBMd, Barbier E, Kraut A, Cout\u0026eacute; Y, Fuchs J, Buhot A, et al. TOX4 and its binding partners recognize DNA adducts generated by platinum anticancer drugs. Archives of Biochemistry and Biophysics. 2011;507:296\u0026ndash;303.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLobbardi R, Pinder J, Martinez-Pastor B, Theodorou M, Blackburn JS, Abraham BJ, et al. TOX Regulates Growth, DNA Repair, and Genomic Instability in T-cell Acute Lymphoblastic Leukemia. Cancer Discovery. 2017;7:1336\u0026ndash;53.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCordeiro MH, Smith RJ, Saurin AT. Kinetochore phosphatases suppress autonomous Polo-like kinase 1 activity to control the mitotic checkpoint. J Cell Biol. 2020;219:e202002020.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMardin BR, Agircan FG, Lange C, Schiebel E. Plk1 controls the Nek2A-PP1gamma antagonism in centrosome disjunction. Curr Biol. 2011;21:1145\u0026ndash;51.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSzatmari E, Habas A, Yang P, Zheng J-J, Hagg T, Hetman M. A Positive Feedback Loop between Glycogen Synthase Kinase 3β and Protein Phosphatase 1 after Stimulation of NR2B NMDA Receptors in Forebrain Neurons*. Journal of Biological Chemistry. 2005;280:37526\u0026ndash;35.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSrinivas US, Tay NSC, Jaynes P, Anbuselvan A, Ramachandran GK, Wardyn JD, et al. PLK1 inhibition selectively induces apoptosis in ARID1A deficient cells through uncoupling of oxygen consumption from ATP production. Oncogene. 2022;41:1986\u0026ndash;2002.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOrd M, Puss KK, Kivi R, Moll K, Ojala T, Borovko I, et al. Proline-Rich Motifs Control G2-CDK Target Phosphorylation and Priming an Anchoring Protein for Polo Kinase Localization. Cell Rep. 2020;31:107757.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchl\u0026ouml;sser RM, Krumbach F, Corrales E, Andrieux G, Preisinger C, Liss F, et al. Multidimensional OMICs reveal ARID1A orchestrated control of DNA damage, splicing, and cell cycle in normal-like and malignant urothelial cells. Mol Oncol. 2025;19:3784\u0026ndash;805.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMuchardt C, Reyes JC, Bourachot B, Leguoy E, Yaniv M. The hbrm and BRG-1 proteins, components of the human SNF/SWI complex, are phosphorylated and excluded from the condensed chromosomes during mitosis. Embo j. 1996;15:3394\u0026ndash;402.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu Z, Chen X, Guo A, Manzano T, Walsh PJ, Wills KM, et al. Mitotic bookmarking by SWI/SNF subunits. Nature. 2023;618:180\u0026ndash;7.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKumar A, Dcunha L, Mahin A, Shivamurthy PB, Subair S, Gopalakrishnan AP, et al. Phosphoproteome landscape of ARID1A and its implications in DNA damage response and breast cancer pathogenesis. Discov Oncol. 2026;17:312.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePerez-Riverol Y, Bandla C, Kundu DJ, Kamatchinathan S, Bai J, Hewapathirana S, et al. The PRIDE database at 20 years: 2025 update. Nucleic Acids Res. 2025;53:D543-D53.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRoeselov\u0026aacute; A, Maslen SL, Shivakumaraswamy S, Pellowe GA, Howell S, Joshi D, et al. Mechanism of chaperone coordination during cotranslational protein folding in bacteria. Molecular Cell. 2024;84:2455-71.e8.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShcherbakova L, Pardo M, Roumeliotis T, Choudhary J. Identifying and characterising Thrap3, Bclaf1 and Erh interactions using cross-linking mass spectrometry. Wellcome Open Res. 2021;6:260.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOsorio D, Rond\u0026oacute;n-Villarreal P, Torres R. Peptides: A Package for Data Mining of Antimicrobial Peptides. The R Journal. 2015;7:4\u0026ndash;14.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang da W, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc. 2009;4:44\u0026ndash;57.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSherman BT, Hao M, Qiu J, Jiao X, Baseler MW, Lane HC, et al. DAVID: a web server for functional enrichment analysis and functional annotation of gene lists (2021 update). Nucleic Acids Res. 2022;50:W216-w21.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePardo M, Lang B, Yu L, Prosser H, Bradley A, Babu MM, et al. An expanded Oct4 interaction network: implications for stem cell biology, development, and disease. Cell Stem Cell. 2010;6:382\u0026ndash;95.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAndrews S. FastQC: a quality control tool for high throughput sequence data. 2010. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.bioinformatics.babraham.ac.uk/projects/fastqc/\u003c/span\u003e\u003cspan address=\"https://www.bioinformatics.babraham.ac.uk/projects/fastqc/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKruegel F. Trim Galore. 2023. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/FelixKrueger/TrimGalore\u003c/span\u003e\u003cspan address=\"https://github.com/FelixKrueger/TrimGalore\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCox J, Mann M. 1D and 2D annotation enrichment: a statistical method integrating quantitative proteomics with complementary high-throughput data. BMC Bioinformatics. 2012;13 Suppl 16:S12.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTyanova S, Temu T, Sinitcyn P, Carlson A, Hein MY, Geiger T, et al. The Perseus computational platform for comprehensive analysis of (prote)omics data. Nat Methods. 2016;13:731\u0026ndash;40.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKorotkevich G, Sukhov V, Budin N, Shpak B, Artyomov MN, Sergushichev A. Fast gene set enrichment analysis. bioRxiv. 2021:060012.\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"bmc-biology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [BMC Biology](https://bmcbiol.biomedcentral.com/)","snPcode":"12915","submissionUrl":"https://submission.springernature.com/new-submission/12915/3","title":"BMC Biology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"ARID1A, SWI/SNF, PRISMA, protein interactions, SLiM, peptide interactions, mitosis, phosphorylation","lastPublishedDoi":"10.21203/rs.3.rs-9415881/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9415881/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eThe cBAF SWI/SNF chromatin remodelling complex controls cell proliferation and fate determination by regulating chromatin accessibility and programs of gene expression, DNA damage response, replication, splicing, translation, and cell plasticity. This functional diversity stems from varied protein interactions. ARID1A is the most frequently mutated SWI/SNF subunit in cancer and scaffold of cBAF through the globular C-terminal domain. The rest of ARID1A mostly contains intrinsically disordered regions, key interaction surfaces that are notoriously difficult to study due to highly dynamic conformations. The molecular basis of ARID1A interactions remains obscure, and few have been explored functionally or mapped at an interface level. We perform a PRotein Interaction Screen on peptide MAtrix (PRISMA) to identify novel ARID1A interactions and map short linear motifs that mediate interactions at high sequence resolution.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eOur PRISMA assay recapitulates binding sites of cBAF subunits to ARID1A and detects previously described SLiM-mediated interactions. We reveal binding sites for direct interaction with transcriptional repressor SIN3A, interaction and confirm its role in SIN3A chromatin recruitment. We identify TOX4, CDK2 and CCNA2 as novel ARID1A interactors. Mutation of a cell cycle-regulated CDK2 phosphorylation site in ARID1A leads to altered gene expression of microtubule factors and defective cell division.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e \u003cp\u003eOur study provides novel mechanistic insight into ARID1A functions and highlights a role for CDK2-dependent phosphorylation of ARID1A in the regulation of cell division, with important ramifications for cancer. Our data represents a valuable resource to further study ARID1A and SWI/SNF biology, and more broadly, to investigate sequence and structural features governing protein interactions.\u003c/p\u003e","manuscriptTitle":"High resolution interaction surface mapping by PRISMA reveals novel ARID1A interactions","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-05-07 10:20:10","doi":"10.21203/rs.3.rs-9415881/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"editorInvitedReview","content":"","date":"2026-05-13T16:55:11+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-11T23:28:35+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"147868028198827238155235188336887440747","date":"2026-05-06T21:57:36+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"48558657539589569449203935896196600477","date":"2026-04-29T07:43:42+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-04-28T16:32:52+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-04-15T19:01:01+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-04-15T08:52:32+00:00","index":"","fulltext":""},{"type":"submitted","content":"BMC Biology","date":"2026-04-14T12:56:17+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"bmc-biology","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"Learn more about [BMC Biology](https://bmcbiol.biomedcentral.com/)","snPcode":"12915","submissionUrl":"https://submission.springernature.com/new-submission/12915/3","title":"BMC Biology","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"BMC Series","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d3d2b185-ae94-4876-843c-857bd789550d","owner":[],"postedDate":"May 7th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"","date":"2026-05-13T16:55:11+00:00","index":16,"fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-05-11T23:28:35+00:00","index":15,"fulltext":""},{"type":"reviewerAgreed","content":"147868028198827238155235188336887440747","date":"2026-05-06T21:57:36+00:00","index":14,"fulltext":""},{"type":"reviewerAgreed","content":"48558657539589569449203935896196600477","date":"2026-04-29T07:43:42+00:00","index":10,"fulltext":""}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-05-07T10:20:11+00:00","versionOfRecord":[],"versionCreatedAt":"2026-05-07 10:20:10","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9415881","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9415881","identity":"rs-9415881","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}