PBRM1 directs PBAF to pericentromeres and protects centromere integrity

Lane, Karen A.; Harrod, Alison; Wu, Lillian; Roumeliotis, Theodoros I.; Feng, Hugang; Foo, Shane; Begg, Katheryn A. G.; Schiavoni, Federica; Amin, Noa; Zenke, Frank T.; Melcher, Alan A.; Choudhary, Jyoti S.; Downs, Jessica A.

doi:10.1038/s41467-025-57277-9

Download PDF

Article
Open access
Published: 26 February 2025

PBRM1 directs PBAF to pericentromeres and protects centromere integrity

Nature Communications volume 16, Article number: 1980 (2025) Cite this article

4344 Accesses
5 Citations
19 Altmetric
Metrics details

Subjects

Abstract

The specialised structure of the centromere is critical for effective chromosome segregation, but its repetitive nature makes it vulnerable to rearrangements. Centromere fragility can drive tumorigenesis, but protective mechanisms preventing fragility are still not fully understood. The PBAF chromatin remodelling complex is frequently misregulated in cancer, but its role in cancer is incompletely characterized. Here, we identify PBAF as a protector of centromere and pericentromere structure with profound consequences for genome stability. A conserved feature of isogenic cell lines lacking PBRM1, a subunit of PBAF, is compromised centromere and pericentromere integrity. PBAF is present at these regions, and binding patterns of PBAF and H3K9 methylation change when PBRM1 is absent. PBRM1 loss creates a dependence on the spindle assembly checkpoint, which represents a therapeutic vulnerability. Importantly, we find that even in the absence of any perturbations, PBRM1 loss leads to centromere fragility, thus identifying a key player in centromere protection.

Human centromere repositioning activates transcription and opens chromatin fibre structure

Article Open access 24 September 2022

Whole genome sequencing of skull-base chordoma reveals genomic alterations associated with recurrence and chordoma-specific survival

Article Open access 03 February 2021

MUC1-C activates the PBAF chromatin remodeling complex in integrating redox balance with progression of human prostate cancer stem cells

Article Open access 23 June 2021

Introduction

The centromere is a specialised region of the chromosome that is responsible for the attachment of spindle fibres via the kinetochore during mitosis, ensuring that each daughter cell receives an equal and identical set of chromosomes. Kinetochore assembly is a highly regulated process that builds upon the constitutive centromere associated network (CCAN). The CCAN bridges the CENP-A-containing centromeric chromatin with the outer kinetochore components. CENP-B contributes to this assembly by bridging CENP-A and the CCAN subunit CENP-C. Human centromeric DNA is made up of repetitive arrays of alpha satellite (α-Sat) repeats forming higher order repeats (HORs)¹. Many α-Sat repeats contain a motif termed the B-box, which is bound by CENP-B. CENP-B is important for creating three dimensional structures in the centromere that promote appropriate chromosome segregation².

Centromeres are flanked by pericentromeres or transition regions, which are made up of a wider range of repetitive sequences in human cells, including α-Sat and non-α-Sat repeats, transposable elements, and duplications¹. H3K9me3 and other repressive marks are enriched in pericentromeres and the repetitive elements are mostly silenced³. However, pericentromeric regions also contain non-repetitive elements, including protein coding genes, some of which are expressed¹. Therefore, while largely heterochromatic, there are interspersed regions of open chromatin in pericentromeric regions and transition arms.

The repetitive nature of centromeric and pericentromeric sequences facilitates topological organisation, but this comes at a cost, since these repetitive sequences are vulnerable to inappropriate rearrangements^4,5. Moreover, recent work demonstrated that cells use centromere-associated DNA breaks to help specify functional centromeres⁶, which increases vulnerability if these are processed or repaired inappropriately. Pathological centromere fragility contributes to human disease, such as cancer⁴. Cells must therefore balance the use of these specialised and repetitive structures with mechanisms that protect genome integrity. How this is achieved is not yet fully understood.

PBRM1 (or BAF180) is a subunit of the PBAF chromatin remodelling complex, one of three mammalian SWI/SNF remodelling complexes. PBRM1 is frequently mutated in cancer, and evidence supports the idea that PBRM1 can function as a tumour suppressor⁷. Loss of function mutations are particularly prevalent in clear cell renal cell carcinoma (ccRCC), but they are also found across a range of other cancer types⁸. A critical question is what the fundamental functions of PBRM1 are, which, when lost, contribute to the development or evolution of cancer.

Here, we identify PBRM1 as a factor that prevents centromere fragility. We show that cells lacking PBRM1 have lower levels of centromere- and pericentromere-associated proteins and have altered patterns of organisation of these structures in cells. PBRM1 and the SMARCA4 subunit of PBAF are physically present at these regions, and the SMARCA4 binding pattern changes when PBRM1 is absent. Patterns of histone H3K9 methylation in centromeres and pericentromeres also change in cells lacking PBRM1. Furthermore, PBRM1 loss leads to mitotic defects and creates a dependence on the spindle assembly checkpoint, revealing a potential therapeutic vulnerability. Importantly, we find that even without external perturbations, PBRM1 loss leads to significant centromere fragility, highlighting a previously unrecognised role in centromere protection.

Results

Analysis of isogenic PBRM1 knockout (KO) cell lines identifies misregulation of centromere- and pericentromere-associated proteins

To identify core functions of PBRM1, we generated a panel of 17 clonal cell lines with CRISPR-Cas9 engineered loss of function mutations in PBRM1 across five different cell line backgrounds, including both cancer-derived and immortalised non-cancerous parental cell lines (Fig. 1a and ref. ⁹). The growth rate, cell cycle profile, and morphological changes in the knockout (KO) cells were analysed relative to the parental lines (Fig. 1b-f, and Supplementary Fig. 1). We found none of these features were substantially altered in any of the cell lines other than a modestly reduced growth rate when PBRM1 is lost (Fig. 1d and Supplementary Fig. 1).

**Fig. 1: Analysis of isogenic PBRM1 knockout (KO) cell lines identifies centromere associated protein misregulation as a common feature.**

To characterise the molecular profiles of the cells, we performed mass spectrometry on all PBRM1 KO clones and the corresponding parental lines. Pathway analysis showed alterations in chromatin organisation, DNA repair and recombination, and innate immune signalling were apparent across cell lines (Supplementary Fig. 2a and b). While apoptosis was identified as an altered pathway, we found no difference in the percentage of apoptotic cells when PBRM1 is deficient (Supplementary Fig. 2c). Previously, we found that PBRM1 is important for mediating sister chromatid cohesion at centromeres¹⁰, raising the possibility that PBRM1 is important for chromatin structure and organisation at or near centromeres. We therefore interrogated the proteomic datasets for centromere- and pericentromere-associated proteins, including CENP-A interacting proteins, the constitutive centromere-associated network (CCAN) complex, the outer kinetochore, the chromosomal passenger complex (CPC), pericentromeric heterochromatin (PHC) proteins, and other annotated pericentromere associated proteins. Individual protein levels were modestly altered in the PBRM1 KO cells, but looking across the pathways, a trend towards lower protein levels was apparent in the PBRM1 KO cells when compared with their isogenic parental cell lines (Fig. 1g-k, and Supplementary Fig. 2d, e and 3a). In contrast, transcript levels of these genes were not consistently downregulated in the PBRM1 KO cells, indicating that the protein level changes were not primarily being driven by misregulation of gene expression (Fig. 1l, and Supplementary Fig. 3b, c). This is consistent with previous work showing that there is a poor correlation between changes in the transcriptome and the proteome of SWI/SNF-deficient cells¹¹, and suggests that SWI/SNF activity often influences protein stability through mechanisms other than direct regulation of gene expression. Moreover, downregulation was specific to centromere-associated complexes; by contrast, no consistent changes were apparent when centrosome proteins were analysed (Supplementary Fig. 3f, g).

We further investigated available proteomic datasets in the cancer cell line encyclopedia (CCLE) to explore whether downregulated centromere proteins are a general feature of PBRM1 loss. In the absence of isogenic comparisons, we ranked cell lines according to PBRM1 protein levels. Notably, we found a significant correlation between PBRM1 protein levels and centromere- and pericentromere-associated proteins (Supplementary Fig. 3d, e). These data suggest that lower levels of centromere- and pericentromere-associated proteins is a common feature of cells lacking PBRM1 expression.

Loss of PBRM1 results in altered organization of centromeres and pericentromeric heterochromatin

We next set out to understand whether the decreased level of centromere- and pericentromere-associated proteins had any functional consequence on their organisation. We first looked at whether there were any detectable changes in centromere structure using FISH probes against centromere α-Sat repeats in chromosome 2 or 10. We found a modest trend towards increased area in the KO clones compared with the parental RPE1 or 1BR3 cells (Fig. 2a–c and Supplementary Fig. 4). This pattern was previously observed in CENP-A-depleted cells¹² raising the possibility that PBRM1 loss leads to CENP-A deficiency. Since CENP-A was not present in our proteomic dataset, we interrogated CENP-A by immunofluorescence (IF). However, when CENP-A signal area and shape (eccentricity) were measured, we found no clear difference in the PBRM1 KO cells (Supplementary Fig. 5a-c), suggesting that CENP-A deficiency is not driving the changes in α-Sat signal in the PBRM1 KO cells. Since α-Sat repeats are also present outside of CENP-A-containing chromatin, one possible explanation for the slight changes in FISH signals is a failure of these regions to form appropriate folded structures in the KO cells, leading to an increased volume of α-Sat repeat-containing chromatin in the KO cell nuclei.

**Fig. 2: PBRM1 KO cells display increased H3K9me3 intensity around centromeres.**

We used IF to interrogate CENP-B and NDC80 patterns to further understand the impact of PBRM1 loss on centromere-associated structures. While there was no difference in the CENP-B signal area, we found a modest, but reproducible decrease in eccentricity scores in the PBRM1 KO cells (Supplementary Fig. 5d–f), suggesting that the three-dimensional organisation of the region is impacted by the loss of PBRM1. NDC80 patterns were also subtly altered in PBRM1 knockout cells, with an increase in both area and eccentricity (Supplementary Fig. 5g–i). Interestingly, we also noticed an increase in the distance between mitotic sister chromatid-associated NDC80 signals in the PBRM1 KO cells (Supplementary Fig. 5j), consistent with a change in the three-dimensional organisation of pericentromeric regions.

We next examined whether there were any microscopically detectable changes in pericentromeric heterochromatin. We performed immunofluorescence using an antibody against H3K9me3, a marker of heterochromatin, and found that the intensity of this signal is subtly but consistently increased in the absence of PBRM1 (Fig. 2d, e), suggesting altered organisation or prevalence of heterochromatin in these cells. To see whether there were changes specifically in pericentromeric heterochromatin, we used an antibody against CENP-A to identify centromeres and then quantified the surrounding H3K9me3 signal (Fig. 2f). Interestingly, we find that the pattern of H3K9me3 intensity changes in the PBRM1 KO clones (Fig. 2g). The signal was lower in the shell closest to CENP-A, but higher in outer shells in the PBRM1 KO cells, again suggesting that PBRM1 helps to organise chromatin in regions surrounding centromeres.

PBRM1 and SMARCA4 are present at pericentromeres, and PBRM1 loss leads to changes in SMARCA4 association

One possible mechanism by which PBRM1 is regulating the structure and organisation of chromatin and associated proteins in centromeres and pericentromeric regions is through direct remodelling activity in these regions. There is some evidence to support this possibility. PBAF was reported to associate with kinetochores of mitotic chromosomes¹³, and SWI/SNF subunits have been identified in protein-interaction studies of centromere-associated proteins, including CENP-C, INCENP, and Bub1^14,15,16. However, in contrast to our understanding of PBAF binding patterns in euchromatic regions of the genome, little is known about the specific binding pattern of PBAF in repetitive regions.

We therefore set out to determine whether PBAF associates with centromeric or pericentromeric chromatin (including HORs, Hsat (Human satellite) repeats, transition (ct) arms; Fig. 3a) and gain a comprehensive view of PBAF localisation patterns. To do this, we performed CUT&RUN sequencing in both low and high salt conditions to ensure the capture of heterochromatic regions¹⁷. We mapped PBRM1 and SMARCA4 (BRG1), one of two catalytic subunits of PBAF, using both IgG and a SMARCA4 KO cell line (Fig. 3a and Supplementary Figs. 6 and 7b) as negative controls, and CENP-B as a positive control. Because of the repetitive nature of the region, we analysed the data in two ways: peak calling of uniquely mapping reads, and a k-mer analysis of reads mapping to these regions, thus allowing analysis of reads mapping to multiple locations (Fig. 3a and Supplementary Fig. 6, described below).

**Fig. 3: SMARCA4 is present at centromeres, and this pattern changes in PBRM1 KOs.**

We first aligned uniquely mapping reads using the reported telomere-to-telomere (T2T) CHM13 genome^1,18. CENP-B localised primarily to HORs, as expected, with some additional sites of enrichment in flanking regions (Fig. 3b and Supplementary Fig. 7a). We found enrichment of SMARCA4 and PBRM1 primarily in pericentromere sequences when compared with negative controls, and most SMARCA4 and PBRM1 peaks were located in the transition (ct) arms (Fig. 3b and Supplementary Fig. 7a). As expected, many of the PBRM1 and SMARCA4 peaks overlapped, but non-overlapping peaks were also identified (Fig. 3c, Supplementary Fig. 7c, e and f). This likely reflects the combinatorial flexibility of the complexes (i.e. PBAF can contain SMARCA2 instead of SMARCA4, and SMARCA4 is found in other SWI/SNF complexes).

When analysing the peaks found in centromeres and pericentromeric regions, we find that the peak distribution profile of PBRM1 has a stronger association with promoters when compared with SMARCA4 (Fig. 3d). This profile is similar to that of the genome-wide distributions of both PBRM1 and SMARCA4 in our dataset (Supplementary Fig. 7d), and is consistent with previous studies showing that, in euchromatin, PBRM1-containing PBAF complexes are enriched at both promoters and enhancers, whereas BAF complexes are more often found at enhancers^19,20. To determine whether pericentromere-association is a conserved feature of PBAF, we interrogated datasets in which PBRM1 and SMARCA4 were mapped^21,22, and strikingly, we find that the association of PBRM1 and SMARCA4 in these regions is apparent in other cell line models (Fig. 3e).

Because of the repetitive nature of these regions, we also analysed centromere- and pericentromere-associated reads that mapped to multiple locations, and were therefore unable to be precisely mapped, by performing a k-mer analysis^1,23. This analysis identifies 51-mer sequences that are significantly enriched in the mapping datasets (see Fig. 3a and Supplementary Fig. 6 for workflow) relative to the IgG control. CENP-B was analysed relative to IgG as a positive control.

In order to look for binding patterns and proximity to other features, enriched centromere- and pericentromere-associated k-mers were mapped back onto the T2T genome. The CENP-B-associated k-mers map primarily to the active higher order repeats (HORs) where CENP-B is known to bind (Fig. 3f, i, and Supplementary Fig. 7g, h), and motif analysis of the CENP-B associated k-mers identified the CENP-B box (Fig. 3h), providing support for the utility of this approach.

Similar to peak enrichment, we found that the majority of PBRM1- and SMARCA4-associated k-mers are located within the transition arms, but a small proportion map to repeats (Fig. 3f, i Supplementary Fig. 7g, h), which may indicate some association across these regions. Again, we find both overlapping as well as PBRM1- and SMARCA4-specific k-mers (Fig. 3g and Supplementary Fig. 7i).

We additionally mapped SMARCA4 in PBRM1 KO cells to interrogate changes to PBAF binding patterns when PBRM1 is absent. A subset of SMARCA4-enriched peaks and SMARCA4-enriched k-mers are lost when PBRM1 is absent (Fig. 4a and Supplementary Fig. 8a), suggesting that PBRM1 is important for targeting SMARCA4 and PBAF to these locations. Interestingly, a considerable number of SMARCA4-enriched peaks and SMARCA4-enriched k-mers are gained when PBRM1 is deficient (Fig. 4a, b and Supplementary Figs. 8 and 9a), suggesting that PBRM1 loss leads to aberrant SMARCA4 binding at sites not normally bound by SMARCA4. Together, these data indicate that PBAF binds to specific sites in chromatin flanking centromeres, and this binding pattern is disrupted when PBRM1 is deficient.

**Fig. 4: SWI/SNF enrichment at a subset of centromeric chromatin marks is altered in a PBRM1-dependent manner.**

Distinct patterns in PBRM1-dependent and -independent SMARCA4 association with pericentromeric regions

To further explore binding patterns, we divided centromere- and pericentromere-specific sites of PBRM1 and SMARCA4 chromatin association (both peaks and k-mers) into three different groups (Fig. 4c, d, and Supplementary Fig. 9b, c). First, we looked at all PBRM1 enriched sites that did not have SMARCA4 enrichment in the PBRM1 KO cells (PBRM1 specific), representing sites that are dependent on PBRM1. Second, we grouped PBRM1 enriched sites present at locations where SMARCA4 is still bound when PBRM1 is deficient (PBRM1 non-specific), which likely represents other SWI/SNF complexes, such as BAF. Third, we grouped all SMARCA4 enriched sites that appear only when PBRM1 is deficient (KO specific), which reflect potentially aberrant or inappropriate binding when PBRM1 is deficient). As expected, a greater proportion of PBRM1-specific sites are located in promoter regions when compared with PBRM1 non-specific sites (Fig. 4e). The majority of enriched k-mers from all three groups were located in ct arms, suggesting there is no substantive shift into other centromeric regions (Supplementary Fig. 9d, e).

We then compared the location of these groups of enriched sites (peaks and k-mers) to mapped PTMs. Again, consistent with genome-wide binding patterns of BAF and PBAF, we find that the PBRM1 specific peaks are associated with both promoter- and enhancer-associated PTMs, such as H3K4me3, and PBRM1 non-specific SMARCA4 peaks show relatively more association with enhancer-associated PTMs, such as H3K27ac (Supplementary Fig. 10a). Interestingly, the SMARCA4 peaks that arise when PBRM1 is deficient (KO specific) display a distinct pattern from either of the other two groups. There is a decrease in association with regions normally enriched in H3K4me3, H3K9ac, and H3K36me1, and an increase in association with H3K36me3-enriched chromatin (Supplementary Fig. 10a). In addition, when enriched k-mers were analysed, we noticed a small but potentially important PBRM1-dependent change in association with H3K9me3 (Supplementary Fig. 10a, b). The number of KO-specific k-mers that intersect with H3K9ac enriched k-mers is lower than for the PBRM1 specific subset, similar to what is seen with H3K4me3, and instead, there is a greater number of KO-specific k-mers that intersect with H3K9me3 enriched k-mers. (Supplementary Fig. 10b). To look at this more quantitatively, we analysed the proximity of these by plotting the relative distance between the mapping locations of subsets of SWI/SNF k-mers and histone PTM k-mers as a cumulative fraction²⁴. We used the promoter-associated H3K4me3 mark as a positive control, which showed, as expected, that PBRM1-specific mapping locations are closer to H3K4me3 than PBRM1 non-specific k-mers (Supplementary Fig. 10c). Interestingly, KO specific k-mer mapping locations tend to be further away from those of H3K4me3 k-mers. In contrast, these are much closer to mapping locations that are normally enriched in H3K9me3 and further away from those enriched in H3K9ac (Supplementary Fig. 10c, d).

Changes in centromere- and pericentromere-associated H3K9 methylation patterns are apparent in the absence of PBRM1

Given the changes evident in H3K9me3 staining (Fig. 2) combined with differences in mapping locations relative to H3K9 marks (Supplementary Fig. 10), we tested the possibility that loss of PBRM1 would lead to altered H3K9 methylation patterns. To do this, we performed CUT&RUN mapping of H3K9me2 and H3K9me3 in parental and PBRM1 KO RPE1 cells.

When uniquely mapping reads were mapped to the centromere and pericentromeric regions, we found a shift in the distribution of both marks. Overall, there was a loss of H3K9me2 peaks and a gain of H3K9me3 peaks in the PBRM1 KO cells (Fig. 5a), and the H3K9me3 signal appears to spread into regions formerly occupied by H3K9me2 (Fig. 5b and Supplementary Fig. 11b). Redistribution of these marks was also apparent genome-wide (Supplementary Fig. 11a).

**Fig. 5: H3K9 methylation patterns in the centromere and pericentromeric regions are altered in the absence of PBRM1.**

Again, we analysed enriched repetitive sequences using k-mer analysis. In doing so, we found a shift in the distribution of H3K9 methylation marks associated with repetitive DNA when PBRM1 is deficient (Fig. 5c), with both losses and gains of k-mers apparent in the PBRM1 KO cells. Interestingly, we found that both H3K9me2- and H3K9me3-enriched k-mers were lost from ct arms and gained in HOR sequences when PBRM1 is deficient (Fig. 5d, e, Supplementary Fig. 11c–e). A loss of H3K9me3-enriched k-mers from divergent HORs and monomeric repeats is also apparent, while H3K9me3 association with Hsat1 sequences increases when PBRM1 is deficient (Supplementary Fig. 11c–h). Together, these data demonstrate that loss of PBRM1 alters the pattern of H3K9 methylation in centromeres and flanking pericentromeric sequences.

PBRM1 KO cells are sensitive to mitotic perturbation

We next wanted to test whether the changes in structure at centromeres and pericentromeric regions where PBRM1 is bound lead to sensitivity to mitotic perturbation when PBRM1 is absent. We tested this and found PBRM1 KO cells are sensitive to chronic CDK1 inhibition compared with parental RPE1 cells (Fig. 6a, b). We also treated cells with an acute dose of CDK1 inhibitor (Fig. 6c) and monitored the presence of aberrant mitotic events after release. PBRM1 KO cells show a substantial increase in the number of nuclear abnormalities (Fig. 6d–f, Supplementary Fig. 12a, b).

We also found PBRM1 KO cells are selectively sensitive to depletion of CCNB1 (Fig. 6g, h and Supplementary Fig. 12c). 786-O derived PBRM1 KO clones, however, which do not show downregulation of centromere- or pericentromere-associated proteins (Supplementary Fig. 2e, and Supplementary Fig.14a, b), also do not exhibit sensitivity to CCNB1 depletion (Supplementary Fig. 14c, d). We further tested this using a panel of six renal cancer cell lines, in which three have loss of function mutations in PBRM1, and we found that PBRM1 deficient cells are considerably more sensitive to CCNB1 depletion than the PBRM1 proficient cell lines (Fig. 6i–k, and Supplementary Fig.12d, e), suggesting potential clinical implications. These data suggest that loss of PBRM1 reduces the ability to cope with mitotic perturbations.

PBRM1 KO cells display sensitivity to Mps1 inhibition in vitro and in vivo

Cells with aberrant centromeres rely on the activity of the spindle assembly checkpoint (SAC) if these changes impact on kinetochore-microtubule attachments²⁵. We, therefore, tested whether cells with PBRM1 loss are sensitive to inhibition of the Mps1 kinase that regulates the SAC²⁵. Using three different Mps1 inhibitors, we found that RPE1-derived PBRM1 KO cells are modestly but significantly sensitive when compared with the parental cells (Fig. 7a–c, and Supplementary Fig. 13a–c). In contrast, we found no substantial sensitivity to Mps1 inhibitors in the 786-O derived PBRM1 KO cells (Supplementary Fig. 14e–h), supporting the link between downregulation of centromere-associated proteins and perturbation of mitosis.

To test whether this has clinical potential, we created two PBRM1 KO clones in the mouse melanoma B16-F10 cell line for in vivo studies (Fig. 7d and Supplementary Fig. 13d). We first analysed the proteome of these cells by mass spectrometry, and strikingly, we found downregulation of centromere- and pericentromere-associated proteins (Fig. 7e), suggesting that this is a conserved feature of PBRM1 loss in at least one other species. We next tested the response to Mps1 inhibitors and found the survival of the mouse PBRM1 KO cell lines was significantly lower than the parental B16-F10 cells (Fig. 7f-h, and Supplementary Fig. 13e), consistent with a functional impact of altered centromere-associated proteins.

We used these cell lines to perform in vivo studies with Mps1 inhibitors. Mice were dosed twice weekly with the BOS172722 Mps1 inhibitor over the course of 42 days, at which point the study was terminated (Supplementary Fig. 13f). Mps1 inhibitor treatment had a marginal and statistically insignificant impact on survival of mice bearing B16-F10 parental cell line derived tumours (Fig. 7i and Supplementary Fig. 13g). In contrast, despite the fact the PBRM1-deficient tumours grow more slowly than the parental B16 tumours, survival was further improved following Mps1 inhibitor treatment in mice bearing the PBRM1 KO derived tumours (Fig. 7j, k, and Supplementary Fig. 13g). The disproportionate impact of Mps1 inhibitors on PBRM1-deficient tumours was also evidenced by the fact that tumour growth of PBRM1 KO-derived tumours, but not that of parental B16-F10 derived tumours, was significantly reduced following treatment with Mps1 inhibitors (Fig. 7l-n and Supplementary Fig. 13h). Together, these data suggest that reliance on the SAC in PBRM1 KO cancers represents a therapeutic vulnerability that can be clinically exploited.

Loss of PBRM1 leads to centromere fragility

These results raised the possibility that cells lacking PBRM1 display centromere fragility. To test this, we examined the patterns of sister chromatid exchanges (SCEs) in RPE1 parental and PBRM1 KO cell lines in the absence of any perturbations. We found no significant difference in the total number of SCEs in the PBRM1 KO clones compared with the parental RPE1 cells (Fig. 8a). Notably, however, when we scored the location of the exchanges, we found that they were more likely to localise to centromeres in the PBRM1 KO cells (Fig. 8b–d), leading to whole arm exchanges, consistent with an increased vulnerability at centromeres.

To look more directly at centromere fragility, we made use of an assay in which centromeres are labelled in a strand specific manner to identify sister chromatid exchanges and other centromere-specific rearrangements (termed Cen-CO-FISH¹²; Fig. 8e). As a positive control, we depleted CENP-A, which protects centromere integrity¹². Strikingly, we found that the PBRM1 KO clones showed significantly elevated levels of aberrant centromere signals when compared with the parental RPE1 cells, at a level similar to that seen in CENP-A-depleted cells (Fig. 8f–h). We additionally performed the assay in a second cell line background (1BR3-hTERT) and again find elevated levels of aberrant centromere signals in the PBRM1 KO clones when compared with isogenic controls (Supplementary Fig. 15a, b).

Our mapping data suggests that other SMARCA4-containing complexes are bound at flanking pericentromeric regions (Fig. 3). We were therefore interested to see whether other SWI/SNF complexes are involved in protecting genome stability at centromeres. We performed the Cen-CO-FISH assay in RPE1 cells with CRISPR-mediated KO of the ARID1A subunit of BAF (Supplementary Fig. 15c), which is also commonly mutated in cancer⁸. Notably, in contrast to the effect of PBRM1 loss, we found no detectable difference in aberrant centromere signals in the ARID1A KO cells when compared with parental RPE1 (Supplementary Fig. 15d, e). Collectively, these data indicate that loss of PBRM1 leads to substantially increased genome instability at centromeres, even in the absence of any perturbations.

Discussion

Here, we find that changes to the structure of centromeres and centromere-associated protein levels are conserved features of PBRM1-deficient cells. We find PBRM1 directs SMARCA4 to specific locations in chromatin flanking centromeres. In the absence of PBRM1, cells show mitotic defects and vulnerability to inhibition of the spindle assembly checkpoint, which has the potential to be therapeutically exploited. Most notably, we find centromere fragility in PBRM1-deficient cells, even in the absence of any perturbation.

We find several specific changes when PBRM1 is deficient, including altered H3K9 methylation and changes in both levels and staining patterns of CENP-B. Whether there are additional changes to chromatin composition in these regions remains to be tested. Interestingly in that regard, DNA methylation influences CENP-B binding²⁶ and SWI/SNF can influence DNA methylation²⁷.

Given the PBRM1-dependent changes in H3K9 methylation patterns, the altered immunofluorescence patterns observed (Fig. 2 and Supplementary Figs 4 and 5) likely arise from altered three-dimensional chromatin structure. However, another possibility is that PBRM1-dependent centromere fragility leads to amplification of repetitive elements, which could result in similar patterns. This possibility seems less likely since it should lead to greater signal heterogeneity than what was detected here, but recombination-mediated instability may exacerbate structural changes over longer timeframes in PBRM1-deficient cells.

Together, these data are consistent with a model (Supplementary Fig. 15f) in which PBRM1-containing PBAF complexes work in the transition arms flanking centromeres to help to create a chromatin substrate that promotes the assembly of protein complexes and structures. In the absence of PBRM1, the integrity of these structures is compromised, leading to mitotic vulnerability, dependence on the spindle assembly checkpoint, and inappropriate recombination between repetitive sequences at the centromere.

When mapping SMARCA4 to pericentromeric regions, we find that a subset of binding sites is lost when PBRM1 is deficient, suggesting that PBRM1 is important for directing SMARCA4-containing complexes to these sites. This raises the possibility that there are acetylated substrates at these sites acting as beacons for the PBRM1 bromodomains.

Perhaps more surprisingly, we find that SMARCA4 localises to new sites when PBRM1 is deficient, and many of these sites are normally enriched in H3K9me3. We also find that H3K9me3 patterns are altered when PBRM1 is deficient and that PBRM1 localisation anti-correlates with H3K9me3 enriched regions. These data are consistent with the idea that PBRM1-containing PBAF complexes help to establish appropriate patterns of H3K9me3, which is important for the spatial organization of chromatin²⁸. It was recently reported that DSBs exist in centromeres⁶. We previously showed that PBRM1 influences cohesin dynamics, both at centromeres¹⁰ and at DNA double strand breaks, to help promote repair²⁹. H3K9me3 can also influence centromere function and DSB repair^3,30, and there is an interplay between H3K9me3 and cohesin dynamics^31,32. Therefore, one attractive possibility is that PBRM1 could promote stability at centromeres and pericentromeric regions by modulating the chromatin environment (via H3K9me3 patterns and cohesin dynamics) to prevent inappropriate DSB processing, which would lead to aberrant recombination events between repeats.

Importantly, PBRM1 loss was recently identified as a genetic determinant for a chromosome instability signature that is associated with whole arm or whole chromosome changes³³. This is consistent with the role in preventing centromere fragility we identify here. Notably, ARID1A, which is frequently mutated in cancer, was not identified as a determinant for chromosomal instability, and consistently, we find no impact of ARID1A on centromere fragility in our assays. This suggests that, while BAF may also be bound in regions in or around centromeres, it does not provide the same function as PBAF complexes to prevent centromere fragility. It is significant that we observe PBRM1-dependent centromere fragility in the absence of any exogenous perturbation, suggesting that this fundamental feature of PBRM1 loss is a critical activity that contributes to tumourigenesis and cancer progression.

Methods

All research complied with relevant ethical regulations. The in vivo studies were approved by the Animal Welfare and Ethical Review Body’ at the Institute of Cancer Research (ICR; protocol number BSU_SPF_1467) and carried out under the UK Home Office Project License P5541E04C.

Statistics and reproducibility

Sample sizes were decided based on standard practise from the field, and experiments generally had n = 3 or n = 4 biological replicates (annotated in figure legends). No statistical method was used to predetermine sample size. Biological replicates were used to ensure data was replicable. Three technical replicates were also used where appropriate, i.e. for survival assays, RT-qPCR, proliferation assays. Outliers were removed using Grubbs’ test, and only data with a significance level of 0.05 were excluded. A single proteomics biological replicate in 1BR3-hTERT cells was also excluded as it was identified as an outlier based on principal component analysis (PCA). Ranges of doses of drugs and timepoints were also used, as well as multiple assays to confirm results. Multiple knockout clones were also used to ensure observations were biologically reproducible. Proteomic and transcriptomic data were also used to validate results. Randomisation was not used in this paper; rather results were compared to control samples. Blinding was used for manual analyses of clonogenic survival assays and Cen-CO-FISH quantification.

Cell culture

All cell lines were obtained from ATCC unless otherwise specified. hTERT-RPE1 (catalogue number CRL-4000) cells were cultured in Dulbecco Modified Minimal Essential Medium (DMEM)/F-12 (Sigma) supplemented with 10% FBS (Gibco), 200 µM glutamax (Gibco), 0.26% sodium bicarbonate (Gibco), and 1% penicillin/streptomycin (P/S)(Sigma). 1BR3-hTERT (Gift from Penny Jeggo, Sussex University), HEK293TN (Systems Biosciences LV900A-1), U2OS (catalogue number HTB-96), RCC4-VO (European Collection of Authenticated Cell Cultures, ECACC, catalogue number 03112702), and B16-F10 (catalogue number CRL-6475) cells were cultured in DMEM supplemented with 10% FBS and 1% P/S. 786-O (catalogue number CRL-1932), 769-P (catalogue number CRL-1933), and RCC-FG2 (Cell Lines Service, CLS GmbH, catalogue number 300249) cells were cultured in RPMI 1640 medium (Sigma) supplemented with 10% FBS and 1% P/S. Caki-1 (catalogue number HTB-46) and Caki-2 (catalogue number HTB-47) cells were cultured in McCoy’s 5 A modified medium (Gibco) supplemented with 10% FBS and 1% P/S. All cells were maintained at 37^oC in a humified incubator with 5% CO2 and were regularly tested for mycoplasma contamination.

Drug treatments

For clonogenic survival and SRB assays, cells were treated with a range of concentrations of RO-3306 (Merck), reversine (Sigma), AZ3146 (Sigma), or BOS172722 (MedChemExpress). For immunofluorescence experiments, cells were treated with 3 µM RO-3306. At this dose, partial inhibition of CDK1 is seen, perturbing mitotic timing and progression, but not causing the G2 arrest seen at higher doses of RO-3306³⁴. For in vivo experiments, BOS172722 ditosylate salt was dissolved in vehicle (10% DMSO and 1% Tween-20 in 89% saline).

CRISPR/Cas9-mediated gene knockout

For PBRM1 knockout in a range of cell lines and ARID1A knockout in RPE1, sgRNAs (Sigma) targeting PBRM1 or ARID1A Exon 1 were cloned into the pSpCas9(BB)-2A-GFP (PX458) construct. For SMARCA4 knockout in RPE1 cells, two sgRNAs targeting Exon 2 of SMARCA4 were cloned into the CRISPR-Cas9 nickase pSpCas9n(BB)-2A-GFP (PX461) construct. gRNAs are listed in Supplementary Table 1. PX458 and PX461 plasmids were a gift from Professor Feng Zhang (Addgene plasmids #48138 and #48140 respectively). Plasmids were transfected into cells using Lipofectamine 3000 (Invitrogen), according to the manufacturer’s instructions. 48 h after transfection, GFP- (PX458) or mRuby-positive (PX461) cells were single-cell sorted into 96-well plates using a FACSAria™ III sorter (BD). Resulting clones were screened for loss of PBRM1 or SMARCA4 using western blotting, followed by Sanger sequencing (Eurofins) of the targeted genomic region, as well as immunofluorescence microscopy and proteomic profiling. The PBRM1 knockouts in RPE1 and 1BR3 were described in ref. ⁹.

siRNA mediated gene depletion/RNA interference

100 µM of the specified siRNAs (Dharmacon) were reverse transfected into cells using Lipofectamine RNAiMAX (Invitrogen), according to the manufacturer’s instructions. Multiple single siRNAs or pools containing a mix of 4 independent siRNAs were used as indicated in Figure legends. Corresponding single or pooled non-targeting siRNAs (Dharmacon) were used as controls. Cells were assayed 24-72 h after transfection.

Clonogenic survival assay

Cells in culture were diluted to the appropriate density (300–1000 cells per dish, depending on the cell line) and seeded onto 6 cm dishes in triplicate. For survival after drug treatments, drugs were added 24 h after seeding. For clonogenic survival after RNAi, cells were reverse transfected as described with the appropriate siRNA 24 h before seeding. Cells were incubated in culture for 10-21 days (depending on cell line), after which media was aspirated and colonies were fixed and stained with 1% methylene blue in 70% methanol for 1 h at room temperature with gentle rocking. Dishes were washed extensively with water and allowed to dry overnight. Colony number was counted using a Digital Colony Counter (Stuart), and survival was defined as the % of colonies in treated conditions versus control conditions (vehicle only or control siRNA for drug treatments and RNAi respectively).

Proliferation assays

For cell growth assays, 1×10⁵ cells were seeded to 6 well plates. Every 24 h, triplicate wells were trypsinised and counted, up to a total of 96 h, to quantify the speed of proliferation. For hTERT-RPE1 cells, cells were seeded in triplicate into 96-well plates, and the rate of proliferation was quantified on the Incucyte SX5 (Sartorius) using phase contrast images of cells taken every 4 h, up to a total of 116 h.

Detection of dead cells

Cells were seeded in triplicate to clear-bottom black 96 well plates (Greiner Bio-One) and incubated in culture for 48 to 96 h. 0.5 µg/mL Hoechst and 1 µg/mL propidium iodide was added to cells 30 minutes before detection of total and dead cells using a Celigo S image cytometer (Nexcelom Bioscience).

Flow cytometry

For flow cytometric analysis of cell cycle distribution, growth medium was collected, and cells were trypsinised and combined with cells suspended in growth medium. Cells were pelleted by centrifugation and washed once with PBS. After removal of most supernatant, cells were gently resuspended in remaining PBS and fixed by addition of 70% ethanol dropwise with gentle vortexing and incubated at -20^oC overnight. Cell suspensions were centrifuged for 5 minutes at 300 x g and the ethanol removed. Pellets were washed twice in cold PBS and then resuspended in the appropriate volume of PBS containing 5 µg/mL propidium iodide (Sigma) and 0.1 mg/mL RNase A. Cells were incubated at 37^oC for 30 minutes, protected from light. DNA content of at least 10,000 cells per condition was detected on a BD LSR II or BD FACSymphony A5 (BD Biosciences). Cell debris and doublets were removed by gating (Supplementary Fig. 2f) and cell cycle phases were quantified using FlowJo software (v10.8.1).

Immunofluorescence microscopy

Cells for immunofluorescence (IF) imaging were seeded onto glass coverslips in 6 well plates and assayed. Cells were fixed with 100% ice-cold methanol for 15 minutes at -20^oC, followed by rehydration with four 5-minute washes in PBS. Samples were blocked with 1% BSA in PBS for 1 h at room temperature and incubated overnight at 4^oC in primary antibody diluted in blocking solution. Following washes in PBS, cells were incubated with the appropriate secondary antibody and DNA stain Hoechst 33342 (Sigma) for 2 h at room temperature. Cells were then washed in PBS, mounted onto frosted glass microscope slides with ProLong Gold (Thermo Fisher Scientific), and cured overnight. Antibodies used for immunofluorescence imaging are indicated in Supplementary Table 2. Cells were imaged with a CSU-W1 Yokogawa Advanced Spinning Disk or CSU-W1 SoRa Yokogawa Super Resolution Spinning Disk Confocal microscope with SlideBook imaging software (3i). 1.02 µm Z-stacks were imaged using a 40x or 63x oil objective, or 0.27 µm Z-stacks when using the SoRa spinning disk, and exported for analysis as maximum intensity projections.

Fluorescence in situ hybridisation (FISH)

Cells were trypsinised and washed once in PBS. Washed cells were then fixed by adding ice-cold fixative solution (3:1 methanol:acetic acid) dropwise with gentle vortexing. Cells were incubated for 15 minutes at room temperature, centrifuged, and subjected to a second round of fixation. Cells were then resuspended in 30 µL fixative solution, dropped onto a glass microscope slide, and allowed to dry overnight. After drying, cells were rehydrated in 2x sodium chloride and sodium citrate (SSC) buffer for 2 minutes followed by dehydration in an ethanol series (70, 80, and 95%) for 2 minutes each. The chromosome 2 and chromosome 10 α-satellite FISH probes (Cytocell) were denatured and hybridised with slides according to the manufacturer’s protocol. After hybridisation, cells were washed in 0.25x SSC for 2 minutes at 72^oC and 2x SSC with 0.05% Tween-20 for 30 seconds at room temperature. Slides were drained and mounted with Vectashield Antifade mounting medium with DAPI (Vector labs).

Metaphase spreads

Cells in culture were treated with 0.1 µg/mL colcemid (Gibco) for 3-5 h to accumulate cells in mitosis. Cells were trypsinised, saving floating cells and PBS washes in a falcon tube, and the supernatant and resuspended cells were pelleted, washed with PBS, and slowly resuspended in 10 mL of pre-warmed hypotonic solution (75 mM potassium chloride) and incubated for 30 minutes at 37^oC with intermittent gentle inversion. 200 µL fresh ice-cold fixative solution (3:1 methanol:acetic acid) was added to solution directly before centrifugation at 175 x g for 5 minutes. The majority of the supernatant was removed, leaving less than 0.3 mL, and the pellet was gently resuspended in the remaining supernatant by tapping. Cells were fixed by adding 10 mL fixative solution dropwise with gentle vortexing and incubated at room temperature for 15 minutes. Cells were washed once with fixative solution and resuspended in the appropriate amount of fixative solution to result in a cell concentration of approximately 1×10⁶ cells/mL. Drops of approximately 20 µL were dropped onto humid glass slides from a height and air-dried overnight.

Sister chromatid exchange (SCE) assay

SCE assay was performed as described previously^35,36. Briefly, cells were labelled for two cell cycles, calculated based on doubling time, with 20 µM BrdU. Metaphases were fixed and dropped onto slides, as described above. After drying, slides were immersed in 10 µg/mL Hoechst-33258 (Sigma) for 20 minutes at room temperature. After rinsing in ddH₂O, slides were exposed to 365 nm UV light for 30 minutes, immersed in 2x SSC buffer. Slides were then immersed in preheated 2x SSC at 50^oC for 1 hr, before being incubated in 10% Karyo-MAX^TM Giemsa stain solution (Thermo Fisher) in Sorenson buffer (pH=6.8) for 30 minutes at room temperature. After incubation, slides were rinsed with ddH₂O and allowed to dry, followed by mounting with DPX medium (Sigma). Metaphases were imaged on an RGB colour camera at 100x using 3i imaging software.

Cen-CO-FISH

Cen-CO-FISH was performed as described³⁷. Briefly, cells were labelled for a single cell cycle, calculated based on doubling time, with 7.5 mM 5’-Bromodeoxyuridine (BrdU)(Sigma) and 2.5 mM 5’-Bromodeoxycytidine (BrdC)(MP Biomedicals). Cells were trypsinised and metaphase spreads were prepared as described above, ensuring that metaphase spreads were protected from light when drying. The following day, slides were rehydrated in PBS, and treated with 0.5 mg/mL RNase A (Sigma) for 10 minutes at 37^oC. Slides were then stained with 1 µg/mL Hoechst-33258 (Sigma) in 2x SSC buffer. Slides covered in 2x SSC buffer were exposed to 365 nm UV irradiation for 30 minutes using a UV lamp (Analytik Jena). Following UV nicking, DNA was digested using Exonuclease III (Promega) dissolved in the appropriate buffer according to the manufacturer’s instructions, twice for 10 minutes each. Slides were washed, dehydrated using an ethanol series (70, 90, and 100%), and dried overnight. The following day, slides were hybridised with peptide nucleic acid (PNA) probes against the CENP-B box (PNABio) diluted to a 50 nM concentration in hybridisation solution (10 mM Tris-HCl (pH=7.4) and 0.5% blocking reagent (Roche) in 70% formamide) and heated to 60^oC for 10 minutes directly before use. Slides were first hybridised in a dark hybridisation chamber for 2 h at room temperature with the forward CENP-B probe (PNA Bio, #3004), washed for 30 seconds in Wash Buffer #1 (10 mM Tris-HCl (pH=7.4) and 0.1% BSA in 70% formamide), and incubated for 2 h with the reverse CENP-B probe (PNA Bio, #F3009). Slides were then washed twice in Wash Buffer #1 for 15 minutes each with gentle rocking, followed by three washes in Wash Buffer #2 (0.1 M Tris-HCl (pH=7.4), 0.15 M NaCl, and 0.1% Tween-20), including 1 µg/mL DAPI (Sigma) in the second wash. Slides were dehydrated in an ethanol series (70, 90, and 100%), air dried, and mounted using ProLong Gold. Slides were imaged as before on an Advanced Spinning Disk confocal microscope but instead imaging 0.2 µm Z-stacks and were exported as maximum intensity projections.

Whole protein extraction

Cell pellets were resuspended in the appropriate volume of lysis buffer (10% glycerol, 50 mM Tris-HCl (pH=7.4), 0.5% NP-40, 150 mM NaCl) containing 0.25U/µL benzonase nuclease (Sigma), 1x cOmplete™ EDTA-free protease inhibitor cocktail (Roche), 1 mM MgCl₂, 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^oC. The resulting supernatant containing whole cell protein extracts was collected. For the SMARCA4 western blot in Supplementary Fig. 7, whole cell extracts were obtained using urea buffer as described previously³⁸. Protein concentration was estimated using a Bradford assay (Bio-Rad) according to the manufacturer’s protocol.

Western blotting

For each western blot, approximately 30 µg of protein extract was combined with LSB containing 5% β-mercaptoethanol and boiled for 10 minutes at 95^oC to denature proteins. Proteins were separated via SDS-polyacrylamide gel electrophoresis and transferred to 0.2 µm nitrocellulose membranes (Fisher Scientific). Successful protein transfer was confirmed by incubating membranes in Ponceau S solution (Sigma) for 5 minutes with rocking and membranes were blocked in 5% milk in TBS buffer containing 0.1% Tween-20 for 1 h with gentle rocking. Membranes were incubated in the appropriate antibody diluted in blocking buffer at 4^oC overnight, washed 3 times with TBS buffer containing 0.1% Tween-20, and incubated in the appropriate horseradish peroxidase (HRP)-conjugated secondary antibody diluted in blocking buffer for 1 h at room temperature (except for β-actin antibody, which is already HRP-conjugated). Proteins were visualised on an iBright CL750 imager (Thermo Fisher Scientific) using Immobilon Forte HRP substrate (Merck Millipore) for chemiluminescence. Details of the primary and secondary antibodies used for western blotting in this study are detailed in Supplementary Table 2.

RT-qPCR

RNA was extracted from cells using an RNeasy Mini Kit (Qiagen) according to the manufacturer’s protocol. 0.5 µg of RNA was then reverse transcribed to cDNA with the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems) according to the manufacturer’s protocol. 4 ng of cDNA was used for each qPCR reaction, along with the indicated forward and reverse primers (Supplementary Table 1) at 200 nM concentration, and Power SYBR green PCR master mix (Applied Biosystems). Samples were run in triplicate in MicroAmp Fast Optical 96-well plates (Applied Biosystems) on a StepOne Plus Real-Time PCR system (Applied Biosystems), according to the manufacturer’s protocol. cDNA levels were compared using the Ct (comparative cycle) method, and GAPDH and PPIA were used as housekeeping genes to normalise data.

Proteomics analysis – sample preparation

Cell pellets were lysed in 150 μL buffer containing 1% sodium deoxycholate (SDC), 100 mM triethylammonium bicarbonate (TEAB), 10% isopropanol, 50 mM NaCl and Halt protease and phosphatase inhibitor cocktail (100x) (ThermoFisher Scientific) on ice, assisted with probe sonication, followed by heating at 90 °C for 5 min and re-sonication for 5 sec. Protein concentration was measured with the Quick Start Bradford protein assay (Bio-Rad) according to the manufacturer’s instructions. Protein aliquots of 60 μg or 100 μg (for TMTpro and TMT11plex respectively) were reduced with 5 mM tris-2-carboxyethyl phosphine (TCEP) for 1 h at 60 °C and alkylated with 10 mM iodoacetamide (IAA) for 30 min in the dark, followed by overnight digestion with trypsin at a final concentration of 75 ng/μL (Pierce). Peptides were labelled with the TMT-11plex or TMTpro reagents (ThermoFisher Scientific) according to the manufacturer’s instructions. The TMT labelled peptide pool was acidified at 1% formic acid, the precipitated SDC was removed by centrifugation, and the supernatant was SpeedVac dried. Peptides were fractionated with high-pH Reversed-Phase (RP) chromatography with the XBridge C18 column (2.1 x 150 mm, 3.5 μm) (Waters) on a Dionex UltiMate 3000 HPLC system. Mobile phase A was 0.1% (v/v) ammonium hydroxide and mobile phase B was acetonitrile, 0.1% (v/v) ammonium hydroxide. Peptides were fractionated at a flow rate of 0.2 mL/min using the following gradient: 5 min at 5% B, for 35 min gradient to 35% B, gradient to 80% B in 5 min, isocratic for 5 minutes and re-equilibration to 5% B. Fractions were collected every 42 sec, combined in 28 fractions and SpeedVac dried.

Proteomics analysis – LC-MS and protein quantification

LC-MS analysis was performed on a Dionex UltiMate 3000 UHPLC system coupled to the Orbitrap Lumos Mass Spectrometer (ThermoFisher Scientific). Peptides were loaded onto the Acclaim PepMap 100, 100μm × 2 cm C18, 5μm, trapping column at 10 μL/min flow rate. Peptides were analysed with the EASY-Spray C18 capillary column (75μm × 50 cm, 2μm). Mobile phase A was 0.1% formic acid and mobile phase B was 80% acetonitrile, 0.1% formic acid. Peptides were separated over a 90 min gradient 5%-38% B at a flow rate of 300 nL/min. Survey scans were acquired in the range of 375-1,500 m/z with a mass resolution of 120 K. Precursors were selected in the top speed mode in cycles of 3 sec and isolated for CID fragmentation with quadrupole isolation width 0.7 Th. Collision energy was 35% with AGC 1e4 and max IT 50 ms. Quantification was obtained at the MS3 level with HCD fragmentation of the top 5 most abundant CID fragments isolated with Synchronous Precursor Selection (SPS). Quadrupole isolation width was 0.7 Th and collision energy was 65%. The HCD MS3 spectra were acquired for the mass range 100-500 with 50 K resolution. Targeted precursors were dynamically excluded from further fragmentation for 45 seconds with 7 ppm mass tolerance. The mass spectra were analysed in Proteome Discoverer 2.2 or 2.4 (ThermoFisher Scientific) with the SequestHT search engine for protein identification and quantification. The precursor and fragment ion mass tolerances were set at 20 ppm and 0.5 Da respectively. Spectra were searched for fully tryptic peptides with maximum 2 missed cleavages and minimum length of 6 amino acids. TMT6plex or TMTpro at N-terminus/K and Carbamidomethyl at C were selected as static modifications. Oxidation of M and Deamidation of N/Q were selected as dynamic modifications. Peptide confidence was estimated with the Percolator node and proteins were identified with at least one peptide with q-value < 0.01 based on target-decoy database search. All spectra were searched against reviewed UniProt Homo sapiens protein entries. The reporter ion quantifier node included a TMT11plex or TMTpro quantification method with an integration window tolerance of 15 ppm at the MS3 level. Only unique peptides were used for quantification, considering protein groups for peptide uniqueness. Only peptides with average reporter signal-to-noise>3 were used for protein quantification.

Statistical analysis of proteomics data

Log2 ratios against the parental (PBRM1 WT) cells were computed for each PBRM1 knockout (KO) clone, per cell line, using the sum-normalised abundances exported from Proteome Discoverer. For visualising effects on each cell line (Fig. 1g-i and Supplementary Fig. 2d,e), the median log2(KO/WT) was calculated from replicate clones or biological repeats and was further normalised by subtracting column median; p-values were obtained with two-sided unpaired one-sample t-test in the Perseus platform v1.6.2.2³⁹. The following number of replicates were used: RPE1 × 3 clones vs 2 × Parental, × 2 replicates each; 1BR3 × 6 clones vs 1 × Parental; HEK293TN × 3 clones vs 1 × Parental; U2OS × 2 clones vs 2 × Parental, × 2 replicates each; 786 O × 4 clones vs 2 × Parental, × 2 replicates each. Proteins detected in least one parental and PBRM1 knockout cell line were used for GSEA analysis (v4.3.2)⁴⁰. The log2 ratio of classes was used as the metric for ranking genes, and normalised abundances from proteomic data from all 5 cell lines was used as input. Hallmarks and Gene Ontology: Biological Processes gene set collections were used to detect enriched gene sets. Hallmark gene sets enriched in at least 3 out of 5 cell lines (in the same direction, negatively or positively) and with an FDR < 0.25 were included, and the top 20 negatively enriched Biological Processes were plotted. Cancer cell line proteomic data was downloaded from the DepMap portal and analyses were performed on the normalised protein quantitation data of 375 cell lines performed as described previously⁴¹. Proteins which were not detected in at least 50% of the profiled cell lines were removed from the analysis, and cell lines were ordered based on PBRM1 Log2FC. Pearson correlation was used to determine the correlation between the abundance of PBRM1 and the listed proteins. Heatmaps were generated in R using ComplexHeatmap (v2.14.0)⁴² and clustering was performed using default Euclidean distance methods.

In vivo studies

2 ×10⁵ B16-F10 cells (parental or PBRM1 KOs) were resuspended in 100 µL PBS and injected subcutaneously into the right flank of 6-8 week old female C57BL/6 J mice (Charles River Laboratories). After 3 days growth, mice were treated with vehicle or 50 mg/kg BOS172722 via oral gavage. Mice were treated twice weekly (Monday and Thursday) until the experiment endpoint (Day 42) or were culled once the tumours reached 15 mm size in any direction (maximum permitted size, which was never exceeded). Housing conditions were 21°C + /- 2°C at 45-65% humidity with 12 h/12 h light/dark cycles.

CUT&RUN sequencing

CUT&RUN (Cleavage Under Targets & Release Using Nuclease) was performed according to the CUT&RUN Assay Kit protocol (Cell Signalling Technology) with the following modifications. Cells were detached with Accutase (Sigma). 2 × 10⁵ cells were collected per experiment and pelleted by centrifugation for 5 minutes at 600 x g. Beads were incubated in the indicated primary antibody (Supplementary Table 2) on a nutator overnight at 4 °C. MNase digestion was carried out at 0 °C in an ice water bath for 30 minutes. Salt fractionation and DNA purification were carried out according to the CUT&RUN.salt protocol¹⁷. Briefly, after STOP buffer addition, samples were incubated at 4 °C for 1 h, and the supernatant containing the low salt fraction was collected. Beads were resuspended in low salt buffer (175 mM NaCl, 10 mM EDTA, 2 mM EGTA, 0.1% TritonX-100, 20 μg/mL glycogen). High salt buffer (825 mM NaCl, 10 mM EDTA, 2 mM EGTA, 0.1% TritonX-100, 20 μg/mL glycogen) was added to beads dropwise with gentle vortexing. Samples were again rocked at 4 °C for 1 h, centrifuged for 5 minutes at 16,000 x g, and the supernatant containing the high salt fraction was collected. Low salt fractions were adjusted to 500 mM NaCl. RNAse A (Thermo Fisher Scientific) was added to all fractions and samples were incubated at 37 °C for 20 minutes. SDS (sodium dodecyl sulphate) and Proteinase K (Cell Signaling Technology) were added and samples were incubated at 50 °C for 1 h. DNA was extracted using phenol/chloroform and precipitated in 100% ethanol as described¹⁷ before library preparation. Libraries were prepared using the NEBNext Ultra II DNA library prep kit for Illumina (New England Biolabs), profiled using the Agilent TapeStation D1000 high sensitivity ScreenTape on the Agilent 4150 TapeStation System, and sequenced on the Illumina Novaseq 6000 with 150×150 bp reads by Novogene (Novogene Corporation Inc. UK). For PBRM1 knockout cells, SMARCA4 replicates 1 and 2 were performed with KO1 and SMARCA4 replicate 3 was performed with KO2. H3K9me2/3 replicates 1 and 3 were performed with KO1 and replicate 2 with KO2.

CUT&RUN analysis

Fastq reads were trimmed using TrimGalore (v0.6.6) (Krueger F, Trimgalore (2023), GitHub repository, https://github.com/FelixKrueger/TrimGalore) using the options --trim-n --paired. The human CHM13-T2Tv1.1 (GenBank GCA_009914755.3) and S. cerevisiae S288C (GenBank GCA_000146045.2) assemblies were combined into one FASTA file, then trimmed reads were mapped using Bowtie2 (v2.4.2)⁴³ using the parameters --local --very-sensitive-local --no-unal --no-mixed --no-discordant --dovetail --soft-clipped-unmapped-tlen --non-deterministic --phred33 -I 50 -X 1500. Reads with more than 3 mismatches were removed with Sambamba (v0.5.0)⁴⁴, and their corresponding mates were removed with Picard tools (v2.23.8) (“Picard Toolkit.” 2019. Broad Institute, GitHub Repository. https://broadinstitute.github.io/picard/; Broad Institute). Sam files were converted to bam with SAMtools (v1.11)⁴⁵. Reads mapped to the CHM13-T2Tv1.1 assembly were extracted using BAMtools (v2.5.1)⁴⁶ split. Duplicate reads were removed using Picard if deemed necessary. Low and high salt BAM files from each sample were merged, sorted, and indexed using SAMtools.

Uniquely mapped reads analysis

Uniquely mapping reads were extracted with Sambamba⁴⁴, using the parameters: view -h -t 6 -f bam -F “[XS] == null and not unmapped and not duplicate” Bigwig files were made using Deeptools (bamCoverage --extendReads --maxFragmentLength 1500 --binSize 10 –outFileFormat bigwig, v3.1.3)⁴⁷ and normalised by total number of mapped reads, including non-uniquely mapping ones (using --scaleFactor parameter). Peak calling was performed on uniquely mapping read bam files using MACS2 (v2.2.7.1)⁴⁸ callpeak using IgG as control, with options -g 3054832041 -f BAMPE --keep-dup all -q 0.01 --broad --broad-cutoff 0.01. Bedtools (v 2.29.2)⁴⁹ multiinter and merge functions were used to merge peaks that were called in at least 2 reps. Screenshots of bigwig files and peaks were taken from IGV (v2.11.9)⁵⁰. The ChIPseeker R package (v1.40.0)⁵¹ was used to analyse the genomic distribution of peaks across features. The intervene package (v0.6.5)⁵² was used to calculate overlapping peak intersections and resulting venn diagrams were re-drawn using the ggVennDiagram R package (v1.5.2)⁵³. The EnrichedHeatmap R package (v1.34.0)⁵⁴ was used to generate heatmaps of CUT&RUN and ChIP-seq signal across regions extended +/-5kb around target peaks, with an averaged signal plot above.

Analysis of downloaded publicly available ChIP-seq datasets

Histone mark ChIP-seq data for RPE1 cells across the cell cycle, and associated input data was downloaded from GEO accession GSE175752⁵⁵. Fastq files across the cell cycle (G1, ES, LS, G2) in RPE.Ctrl cells were merged for downstream analysis. ChIP-seq data of PBRM1 (SRR12036678, SRR23588908), SMARCA4 (SRR12036684, SRR23588912) and input (SRR12036698) in parental 786-O cells, as well as PBRM1 (SRR12036715, SRR23588903) and input (SRR12036716) from A498, and PBRM1 (SRR12036717-SRR12036722) and input (SRR12036723- SRR12036728) from HK2 were downloaded from the GEO accession GSE152681²¹. Fastq files, where multiple SRA files were available for the same sample, were merged for downstream analysis. ChIP-seq data of SMARCA4 (SRR2133615) and input (SRR2133624) from parental HCT116 cells were downloaded from the GEO accession GSE71510²². These ChIP-seq datasets were analysed as above for the CUT&RUN with a few exceptions, as they have single-end reads rather than paired-end reads, the Picard tools step to remove corresponding mates following removing reads with 3 mismatches was no longer required. Bowtie2 parameters were changed to -I 300 -X 700 for analysing GSE152681. Following uniquely mapping read filtering, bigwig files were generated with updated bamCoverage parameters to not include –maxFragmentLength, but to update to --extendReads 250 (for GSE175752), --extendReads 500 (for GSE152681) and --extendReads 180 (for GSE71510).

Read filtering steps for k-mer analysis

The seqkit (v2.5.1)⁵⁶ fx2tab function was used to identify read IDs of CUT&RUN reads that mapped to the spike-in, S. cerevisiae S288C (converted bam file to fastq using bedtools bamtofastq (v 2.29.2)⁴⁹. The BBmap (v38.84) (BBMap – Bushnell B. – sourceforge.net/projects/bbmap/) filterbyname.sh tool was used to filter out and exclude reads mapping to the S. cerevisiae S288C genome from the original fastq files. CUT&RUN reads mapped to the CHM13-T2Tv1.1 assembly, following extraction of S. cerevisiae S288C reads, were used to calculate the total base count for each replicate for normalisation purposes, following deduplication, trimming and interleaving of paired reads (described below). Reads which intersected with the centromere and pericentromere were extracted for full k-mer analysis. Briefly, sorted bam files were converted to paired end bed files using bedtools (v 2.29.2)⁴⁹ bamtobed -bedpe function, and then reads that mapped to the centromere and pericentromere were extracted using bedtools intersect function with the CHM13-T2Tv1.1 cenSatAnnotation.bed file. Intersecting read names were printed and used with the BBmap (BBMap – Bushnell B. – sourceforge.net/projects/bbmap/) filterbyname.sh tool to filter to only include centromere and pericentromere mapping reads from the original fastq files. These reads were then used for the full k-mer analysis.

k-mer analysis pipeline

The (reference-free) k-mer analysis pipeline was adapted from previously described methods^1,23. First, the clumpify.sh tool from BBmap was used to remove duplicate paired reads that could be PCR duplicates. Then, trimmomatic (v0.39)⁵⁷ was used to trim adapters from paired reads, using the following parameters PE -phred33 ILLUMINACLIP: TruSeq3-PE.fa:2:30:12 SLIDINGWINDOW:10:10 MINLEN:71. The reformat.sh tool from BBmap was used to interleave paired reads for subsequent analysis. Then the kmc package (v3.2.1)⁵⁸ and kmc tools command were used to count and dump 51 bp k-mers that occur at least twice in each replicate, using the parameters -k51 -sm -ci2 -cs100000000. The R package dplyr (v1.1.0)(Wickham H, François R, Henry L, Müller K, Vaughan D (2023). _dplyr: A Grammar of Data Manipulation_) was used to ‘full_join’ the k-mer counts of all replicates and any missing values were replaced with a value of 1 (as missing values could have a count of 0 or 1, so to underestimate enrichment, and to allow fold change calculation). K-mer counts were divided by the total base counts for each replicate (from the total genome reads) to calculate the normalised k-mer counts. Low and high salt fastq files were analysed separately and their normalised k-mer counts were added together for downstream analysis. K-mers were filtered to only include those with a normalised k-mer count of >5e-09 for each replicate individually, and then fold changes were calculated, comparing one sample replicate against the averaged IgG normalised k-mer count as a control (e.g. SMARCA4 vs average IgG). Note that 9 IgG replicates were included in the average from each experiment (n = 6 from parental RPE1 cells, and n = 3 from PBRM1 knockout cells). Tidyverse (v2.0.0) (doi:10.21105/joss.01686) and tibble (v3.2.0) (Müller K, Wickham H (2022). _tibble: Simple Data Frames_) packages were also used above for data formatting purposes, using r-base v4.2.3.

Enriched k-mer selection and downstream analysis

Enriched k-mers were selected if they had a fold change of >2 in the normalised k-mer count of at least 2 replicates vs the averaged IgG normalised k-mer count. Enriched k-mer output files were converted to FASTA format, using sed ‘1 d’ and awk -F ‘[]’ ‘BEGIN{{OFS = “\n”}}{{n=NR; x = “>“n; print x, $1}}’. Bowtie2 (v2.4.2)⁴³ was used to map these k-mers back to the CHM13-T2Tv1.1 genome, using the specific parameters -f -k 5000 --score-min C,0,0 to allow only exact sequences to be mapped back up to 5000 times. Bedtools (v 2.29.2)⁴⁹ intersect was again used to extract k-mer mapping sites that intersect with the centromere annotation, and awk -F [“_”,”\t”] ‘OFS = ”\t” ‘¹’ was used to split the centromere annotation in the subsequent bed file from e.g. ct_1_1(p_arm) to ct. The GNU datamash package (v1.1.0) (Free Software Foundation, I. (2014). GNU Datamash. Retrieved from https://www.gnu.org/software/datamash/) was then used to create a pivot table summarising how many times each k-mer mapped to each part of the centromere annotation (following the split described above), using the following script: sort -k 1 | datamash -s crosstab 1,2 | sed ‘s ~ N/A ~ 0~g’. Heatmaps representing centromere distribution of k-mers were generated in R using ComplexHeatmap (v2.20.00). Bigwig files were generated from enriched k-mer bam files using deepTools bamCoverage tool using the following parameters --smoothLength 1 --binSize 1 --scaleFactor 1 and screenshots were taken from IGV⁵⁰. Motifs associated with enriched k-mers were discovered using the MEME-suite SEA k-mer enrichment package⁵⁹. Venn diagrams were created with the ggVennDiagram R package⁵³. Bedtools (v 2.29.2) reldist function⁴⁹ (based on²⁴) was used to calculate the relative distance between a query set (e.g. H3K9me3 k-mer mapping locations) and a reference set (e.g. PBRM1-specific k-mer mapping locations). This was calculated using the following formula: min(d1,d2)/r – where d1 and d2 are the distances between each location of the query set and the two closest locations of the reference set, and r is the total distance between the two closest locations of the reference set, i.e. (d1 + d2). The cumulative fractions of query k-mer mapping locations at each relative distance up to a maximum possible of 0.5 were calculated and plotted.

RNA-sequencing

For RNA-sequencing, pellets were harvested by scraping cells in ice-cold PBS following by centrifugation at 7500 x g for 10 minutes. Three independent biological replicates were used for parental cells (n = 3), and two independent biological replicates of two individual PBRM1 knockout clones were used for PBRM1 knockouts, which were combined for n = 4 in total. 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 Agilent High Sensitivity RNA ScreenTape, using the Agilent Tapestation as above. 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. Fastq reads were checked using FastQC (v0.11.9) (Andrews, S. (2010). FastQC: A Quality Control Tool for High Throughput Sequence Data [Online]. http://www.bioinformatics.babraham.ac.uk/projects/fastqc/) and trimmed using TrimGalore (v0.6.6) (Krueger F, Trimgalore (2023), GitHub repository, https://github.com/FelixKrueger/TrimGalore). Residual ribosomal RNA reads were removed using Ribodetector with -e norRNA setting (v0.2.7)⁶⁰ and strandedness was detected using RSeQC (v4.0.0)⁶¹. Reads were aligned to the T2T-CHM13v2 genome using STAR alignment software (v2.7.6a)⁶² and reads mapping to genes were quantified using HTSeqCount (v0.12.4)⁶³. Differential analysis of gene expression was calculated in R using DESeq2 (v1.38.3)⁶⁴.

Software & statistical analyses

The number of biological repeats for each experiment and the statistical analyses used are indicated in the figure legend for each experiment. Graphs were generated and statistical analyses were performed using GraphPad Prism (v9.5.1). Outliers were removed using Grubbs’ test using a significance level of 0.05. Microscopy images were analysed using ImageJ (v1.5.3 or 1.5.4) or CellProfiler (v4.0.7) software, and visualized using GraphPad Prism or ggplot2 (v3.4.4)(Valero-Mora (2010). ggplot2: Elegant Graphics for Data Analysis. https://doi.org/10.18637/jss.v035.b01). For statistical analyses of microscopy data, the median of each biological repeat was used for the appropriate statistical test, described in the corresponding Figure legend). Omics data were visualised using the indicated packages in R (v4.2.1, 4.2.2, or 4.4.0 except where noted) and R Studio (v2021.09.0, 2023.03.1, 2024.04.2 except where noted). Figures were generated using Adobe Illustrator (v27.5). Graphics in Figs. 1a, j, 2f, and Supplementary Figure 15f were generated using Adobe Illustrator. The schematic in Fig. 8e was generated using BioRender (biorender.com). More details of script and packages used are available at https://github.com/Downs-Lab/PBRM1_centromere.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

The RNA-seq datasets generated in this study are available in the GEO repository (accession number GSE235342; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE235342). The CUT&RUN datasets generated in this study are available in the GEO repository (accession number GSE235294; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE235294). The proteomics datasets generated in this study are available through the ProteomeXchange Consortium via the PRIDE repository (accession number PXD043209; https://www.ebi.ac.uk/pride/archive/projects/PXD043209). All cell lines used in this study have been made available at CancerTools.org. All data generated in this study are provided in the Supplementary Information figures and tables. Source data are provided with this paper.

Code availability

The scripts and packages used are described in the methods and available at https://github.com/Downs-Lab/PBRM1_centromere and https://zenodo.org/records/14604006⁶⁵.

References

Altemose, N. et al. Complete genomic and epigenetic maps of human centromeres. Science 376, eabl4178 (2022).
Article CAS PubMed PubMed Central Google Scholar
Chardon, F. et al. CENP-B-mediated DNA loops regulate activity and stability of human centromeres. Mol. Cell 82, 1751–1767 e1758 (2022).
Article CAS PubMed MATH Google Scholar
Muller, S. & Almouzni, G. Chromatin dynamics during the cell cycle at centromeres. Nat. Rev. Genet 18, 192–208 (2017).
Article CAS PubMed MATH Google Scholar
Barra, V. & Fachinetti, D. The dark side of centromeres: types, causes and consequences of structural abnormalities implicating centromeric DNA. Nat. Commun. 9, 4340 (2018).
Article ADS CAS PubMed PubMed Central Google Scholar
Balzano, E. & Giunta, S. Centromeres under pressure: evolutionary innovation in conflict with conserved function. Genes 11, 10.3390/genes11080912 (2020).
Saayman X., Graham E., Nathan W. J., Nussenzweig A., Esashi F. Centromeres as universal hotspots of DNA breakage, driving RAD51-mediated recombination during quiescence. Mol Cell, 10.1016/j.molcel.2023.01.004 (2023).
Nargund, A. M. et al. The SWI/SNF protein PBRM1 restrains VHL-loss-driven clear cell renal cell carcinoma. Cell Rep. 18, 2893–2906 (2017).
Article CAS PubMed PubMed Central MATH Google Scholar
Harrod, A., Lane, K. A. & Downs, J. A. The role of the SWI/SNF chromatin remodelling complex in the response to DNA double strand breaks. DNA Repair (Amst.) 93, 102919 (2020).
Article CAS PubMed Google Scholar
Feng, H. et al. PBAF loss leads to DNA damage-induced inflammatory signaling through defective G2/M checkpoint maintenance. Genes Dev. 36, 790–806 (2022).
Article CAS PubMed PubMed Central MATH Google Scholar
Brownlee, P. M., Chambers, A. L., Cloney, R., Bianchi, A. & Downs, J. A. BAF180 promotes cohesion and prevents genome instability and aneuploidy. Cell Rep. 6, 973–981 (2014).
Article CAS PubMed PubMed Central Google Scholar
Roumeliotis, T. I. et al. Genomic Determinants of Protein Abundance Variation in Colorectal Cancer Cells. Cell Rep. 20, 2201–2214 (2017).
Article CAS PubMed PubMed Central MATH Google Scholar
Giunta, S. & Funabiki, H. Integrity of the human centromere DNA repeats is protected by CENP-A, CENP-C, and CENP-T. Proc. Natl Acad. Sci. USA 114, 1928–1933 (2017).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Xue, Y. et al. The human SWI/SNF-B chromatin-remodeling complex is related to yeast rsc and localizes at kinetochores of mitotic chromosomes. Proc. Natl Acad. Sci. USA 97, 13015–13020 (2000).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Santos-Barriopedro, I., van Mierlo, G. & Vermeulen, M. Off-the-shelf proximity biotinylation using ProtA-TurboID. Nat Protoc 1, 36–37 (2022).
Garcia, Y. A. et al. Mapping Proximity Associations of Core Spindle Assembly Checkpoint Proteins. J. Proteome Res 20, 3414–3427 (2021).
Article CAS PubMed PubMed Central MATH Google Scholar
Samejima, I. et al. Whole-proteome genetic analysis of dependencies in assembly of a vertebrate kinetochore. J. Cell Biol. 211, 1141–1156 (2015).
Article CAS PubMed PubMed Central MATH Google Scholar
Thakur, J. & Henikoff, S. Unexpected conformational variations of the human centromeric chromatin complex. Genes Dev. 32, 20–25 (2018).
Article CAS PubMed PubMed Central MATH Google Scholar
Nurk, S. et al. The complete sequence of a human genome. Science 376, 44–53 (2022).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Carcamo, S. et al. Altered BAF occupancy and transcription factor dynamics in PBAF-deficient melanoma. Cell Rep. 39, 110637 (2022).
Article CAS PubMed PubMed Central MATH Google Scholar
Park, Y. K. et al. Interplay of BAF and MLL4 promotes cell type-specific enhancer activation. Nat. Commun. 12, 1630 (2021).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Yao, X. et al. PBRM1-deficient PBAF complexes target aberrant genomic loci to activate the NF-kappaB pathway in clear cell renal cell carcinoma. Nat. Cell Biol. 25, 765–777 (2023).
Article CAS PubMed MATH Google Scholar
Mathur, R. et al. ARID1A loss impairs enhancer-mediated gene regulation and drives colon cancer in mice. Nat. Genet 49, 296–302 (2017).
Article CAS PubMed MATH Google Scholar
Smith, O. K. et al. Identification and characterization of centromeric sequences in Xenopus laevis. Genome Res 31, 958–967 (2021).
Article CAS PubMed PubMed Central MATH Google Scholar
Favorov, A. et al. Exploring massive, genome scale datasets with the GenometriCorr package. PLoS Comput Biol. 8, e1002529 (2012).
Article CAS PubMed PubMed Central MATH Google Scholar
Lara-Gonzalez, P., Pines, J. & Desai, A. Spindle assembly checkpoint activation and silencing at kinetochores. Semin Cell Dev. Biol. 117, 86–98 (2021).
Article CAS PubMed PubMed Central Google Scholar
Tanaka, Y., Kurumizaka, H. & Yokoyama, S. CpG methylation of the CENP-B box reduces human CENP-B binding. FEBS J. 272, 282–289 (2005).
Article CAS PubMed MATH Google Scholar
Banine, F. et al. SWI/SNF chromatin-remodeling factors induce changes in DNA methylation to promote transcriptional activation. Cancer Res 65, 3542–3547 (2005).
Article CAS PubMed MATH Google Scholar
Padeken, J., Methot, S. P. & Gasser, S. M. Establishment of H3K9-methylated heterochromatin and its functions in tissue differentiation and maintenance. Nat. Rev. Mol. Cell Biol. 23, 623–640 (2022).
Article CAS PubMed PubMed Central MATH Google Scholar
Meisenberg, C. et al. Repression of transcription at DNA breaks requires cohesin throughout interphase and prevents genome instability. Mol. Cell 73, 212–223.e217 (2019).
Article CAS PubMed PubMed Central Google Scholar
Xu, Y., Xu, C. & Price, B. D. Mechanistic links between ATM and histone methylation codes during DNA repair. Prog. Mol. Biol. Transl. Sci. 110, 263–288 (2012).
Article CAS PubMed MATH Google Scholar
Haarhuis, J. H. I. et al. A Mediator-cohesin axis controls heterochromatin domain formation. Nat. Commun. 13, 754 (2022).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Spracklin, G. et al. Diverse silent chromatin states modulate genome compartmentalization and loop extrusion barriers. Nat. Struct. Mol. Biol. 30, 38–51 (2023).
Article CAS PubMed Google Scholar
Drews, R. M. et al. A pan-cancer compendium of chromosomal instability. Nature 606, 976–983 (2022).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
McCloy, R. A. et al. Partial inhibition of Cdk1 in G2phase overrides the SAC and decouples mitotic events. Cell Cycle 13, 1400–1412 (2014).
Article CAS PubMed PubMed Central MATH Google Scholar
Stults, D. M., Killen, M. W., Marco-Casanova, P. & Pierce, A. J. The sister chromatid exchange (SCE) assay. Methods Mol. Biol. 2102, 441–457 (2020).
Article CAS PubMed Google Scholar
Tumini, E. & Aguilera, A. The Sister-Chromatid Exchange Assay in Human Cells. Methods Mol. Biol. 2153, 383–393 (2021).
Article CAS PubMed Google Scholar
Giunta, S. Centromere chromosome orientation fluorescent in situ hybridization (Cen-CO-FISH) detects sister chromatid exchange at the centromere in human cells. Bio Protoc. 8, e2792 (2018).
Article PubMed PubMed Central Google Scholar
Schiavoni, F. et al. Aneuploidy tolerance caused by BRG1 loss allows chromosome gains and recovery of fitness. Nat. Commun. 13, 1731 (2022).
Article ADS CAS PubMed PubMed Central MATH Google Scholar
Tyanova, S. et al. The Perseus computational platform for comprehensive analysis of (prote)omics data. Nat. Methods 13, 731–740 (2016).
Article CAS PubMed MATH Google Scholar
Subramanian, A. et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl Acad. Sci. USA 102, 15545–15550 (2005).
Article ADS CAS PubMed PubMed Central Google Scholar
Nusinow, D. P. et al. Quantitative proteomics of the cancer cell line encyclopedia. Cell 180, 387–402 e316 (2020).
Article CAS PubMed PubMed Central MATH Google Scholar
Gu, Z., Eils, R. & Schlesner, M. Complex heatmaps reveal patterns and correlations in multidimensional genomic data. Bioinformatics 32, 2847–2849 (2016).
Article CAS PubMed MATH Google Scholar
Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–359 (2012).
Article CAS PubMed PubMed Central MATH Google Scholar
Tarasov, A., Vilella, A. J., Cuppen, E., Nijman, I. J. & Prins, P. Sambamba: fast processing of NGS alignment formats. Bioinformatics 31, 2032–2034 (2015).
Article CAS PubMed PubMed Central Google Scholar
Danecek, P. et al. Twelve years of SAMtools and BCFtools. Gigascience 10, giab008 (2021).
Barnett, D. W., Garrison, E. K., Quinlan, A. R., Stromberg, M. P. & Marth, G. T. BamTools: a C++ API and toolkit for analyzing and managing BAM files. Bioinformatics 27, 1691–1692 (2011).
Article CAS PubMed PubMed Central Google Scholar
Ramirez, F., Dundar, F., Diehl, S., Gruning, B. A. & Manke, T. deepTools: a flexible platform for exploring deep-sequencing data. Nucleic Acids Res 42, W187–W191 (2014).
Article CAS PubMed PubMed Central Google Scholar
Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol. 9, R137 (2008).
Article PubMed PubMed Central MATH Google Scholar
Quinlan, A. R. & Hall, I. M. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26, 841–842 (2010).
Article CAS PubMed PubMed Central MATH Google Scholar
Robinson, J. T. et al. Integrative genomics viewer. Nat. Biotechnol. 29, 24–26 (2011).
Article CAS PubMed PubMed Central Google Scholar
Wang, Q. et al. Exploring Epigenomic Datasets by ChIPseeker. Curr. Protoc. 2, e585 (2022).
Article CAS PubMed Google Scholar
Khan, A. & Mathelier, A. Intervene: a tool for intersection and visualization of multiple gene or genomic region sets. BMC Bioinforma. 18, 287 (2017).
Article MATH Google Scholar
Gao, C. H., Yu, G. & Cai, P. ggVennDiagram: an intuitive, easy-to-use, and highly customizable r package to generate venn Diagram. Front Genet 12, 706907 (2021).
Article PubMed PubMed Central Google Scholar
Gu, Z., Eils, R., Schlesner, M. & Ishaque, N. EnrichedHeatmap: an R/Bioconductor package for comprehensive visualization of genomic signal associations. BMC Genomics 19, 234 (2018).
Article PubMed PubMed Central Google Scholar
Van Rechem, C. et al. Collective regulation of chromatin modifications predicts replication timing during cell cycle. Cell Rep. 37, 109799 (2021).
Article PubMed PubMed Central MATH Google Scholar
Shen, W., Le, S., Li, Y. & Hu, F. SeqKit: a cross-platform and ultrafast toolkit for FASTA/Q file manipulation. PLoS One 11, e0163962 (2016).
Article PubMed PubMed Central Google Scholar
Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114–2120 (2014).
Article CAS PubMed PubMed Central Google Scholar
Kokot, M., Dlugosz, M. & Deorowicz, S. KMC 3: counting and manipulating k-mer statistics. Bioinformatics 33, 2759–2761 (2017).
Article CAS PubMed MATH Google Scholar
Bailey, T. L., Johnson, J., Grant, C. E. & Noble, W. S. The MEME Suite. Nucleic Acids Res 43, W39–W49 (2015).
Article CAS PubMed PubMed Central MATH Google Scholar
Deng, Z. L., Munch, P. C., Mreches, R. & McHardy, A. C. Rapid and accurate identification of ribosomal RNA sequences via deep learning. Nucleic Acids Res 50, e60 (2022).
Article CAS PubMed PubMed Central MATH Google Scholar
Wang, L., Wang, S. & Li, W. RSeQC: quality control of RNA-seq experiments. Bioinformatics 28, 2184–2185 (2012).
Article CAS PubMed MATH Google Scholar
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Article CAS PubMed Google Scholar
Anders, S., Pyl, P. T. & Huber, W. HTSeq–a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166–169 (2015).
Article CAS PubMed MATH Google Scholar
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
Article PubMed PubMed Central MATH Google Scholar
Harrod A., Wu L., Lane K. A., Downs J. A. PBRM1 directs PBAF to pericentromeres and protects centromere integrity (2025).
Lane, K. https://BioRender.com/g59e957 (2025).

Download references

Acknowledgements

We thank the Institute of Cancer Research Flow Cytometry, Light Microscopy and the Biological Services Unit for support, and we thank the ICR Scientific Computing Team for HPC services. We thank all members of the Downs lab, Jon Pines, and Bill Earnshaw for helpful discussions, Swen Hoelder and Florence Raynaud for reagents and advice, and former lab members Peter Brownlee and Cornelia Meisenberg for early observations. This work was supported by Cancer Research UK C7905/A25715 (JAD), C309/A25144 (AAM), and DRCRPGTD-Nov21100001 (JCS) and the Medical Research Council MR/W001276/1 (JAD, AAM, and JCS).

Author information

Hugang Feng
Present address: The Francis Crick Institute; London, London, UK
These authors contributed equally: Karen A. Lane, Alison Harrod.

Authors and Affiliations

Division of Cell and Molecular Biology, The Institute of Cancer Research; London, London, UK
Karen A. Lane, Alison Harrod, Lillian Wu, Theodoros I. Roumeliotis, Hugang Feng, Katheryn A. G. Begg, Federica Schiavoni, Noa Amin, Jyoti S. Choudhary & Jessica A. Downs
Division of Radiotherapy and Imaging, The Institute of Cancer Research; London, London, UK
Shane Foo & Alan A. Melcher
Merck KGaA, Biopharma R&D, Translational Innovation Platform Oncology, Darmstadt, Germany
Frank T. Zenke

Authors

Karen A. Lane
View author publications
Search author on:PubMed Google Scholar
Alison Harrod
View author publications
Search author on:PubMed Google Scholar
Lillian Wu
View author publications
Search author on:PubMed Google Scholar
Theodoros I. Roumeliotis
View author publications
Search author on:PubMed Google Scholar
Hugang Feng
View author publications
Search author on:PubMed Google Scholar
Shane Foo
View author publications
Search author on:PubMed Google Scholar
Katheryn A. G. Begg
View author publications
Search author on:PubMed Google Scholar
Federica Schiavoni
View author publications
Search author on:PubMed Google Scholar
Noa Amin
View author publications
Search author on:PubMed Google Scholar
Frank T. Zenke
View author publications
Search author on:PubMed Google Scholar
Alan A. Melcher
View author publications
Search author on:PubMed Google Scholar
Jyoti S. Choudhary
View author publications
Search author on:PubMed Google Scholar
Jessica A. Downs
View author publications
Search author on:PubMed Google Scholar

Contributions

K.A.L., A.H., and J.A.D. conceived the project. K.A.L., A.H., L.W., T.I.R., H.F., S.F., K.B., N.A. and F.S. designed and performed experiments. A.H., L.W. and T.I.R. performed bioinformatics analyses. F.T.Z., A.A.M., J.C.S., and J.A.D. obtained funding and provided input and supervision. K.A.L., A.H., and J.A.D. prepared the original draft, and all authors contributed to review and editing.

Corresponding author

Correspondence to Jessica A. Downs.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review file

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Lane, K.A., Harrod, A., Wu, L. et al. PBRM1 directs PBAF to pericentromeres and protects centromere integrity. Nat Commun 16, 1980 (2025). https://doi.org/10.1038/s41467-025-57277-9

Download citation

Received: 29 June 2024
Accepted: 16 February 2025
Published: 26 February 2025
DOI: https://doi.org/10.1038/s41467-025-57277-9

This article is cited by

The centromere: the punching bag of the chromosome
- Annapaola Angrisani
- Daniele Fachinetti
Chromosome Research (2025)