CTCF primes repressive chromatin for rapid BRCA1/BARD1 recruitment to drive homologous recombination repair of DNA double-strand breaks

Soon Young Hwang; Daeki Choi; Myoung Joo Choi; Hye-Ram Jo; Jae-Hoon Jeong; Mi Ae Kang; Jong-Soo Lee

doi:10.1093/nar/gkag715

. 2026 Jul 20;54(14):gkag715. doi: 10.1093/nar/gkag715

CTCF primes repressive chromatin for rapid BRCA1/BARD1 recruitment to drive homologous recombination repair of DNA double-strand breaks

Soon Young Hwang ^1,^2,^c, Daeki Choi ^3,^4,^c, Myoung Joo Choi ⁵, Hye-Ram Jo ⁶, Jae-Hoon Jeong ⁷, Mi Ae Kang ^8,^9,^✉, Jong-Soo Lee ^10,^11,^✉

¹ Department of Life Sciences, College of Natural Sciences, Ajou University, 16499,Suwon, Republic of Korea

² Research Institute of Basic Sciences, Ajou University, 16499 Suwon, Republic of Korea

³ Department of Life Sciences, College of Natural Sciences, Ajou University, 16499,Suwon, Republic of Korea

⁴ Research Institute of Basic Sciences, Ajou University, 16499 Suwon, Republic of Korea

⁵ Department of Life Sciences, College of Natural Sciences, Ajou University, 16499,Suwon, Republic of Korea

⁶ Division of Radiation Biomedical Research, Korea Institute of Radiological and Medical Sciences, 01812 Seoul, Republic of Korea

⁷ Division of Radiation Biomedical Research, Korea Institute of Radiological and Medical Sciences, 01812 Seoul, Republic of Korea

⁸ Department of Life Sciences, College of Natural Sciences, Ajou University, 16499,Suwon, Republic of Korea

⁹ Research Institute of Basic Sciences, Ajou University, 16499 Suwon, Republic of Korea

¹⁰ Department of Life Sciences, College of Natural Sciences, Ajou University, 16499,Suwon, Republic of Korea

¹¹ Research Institute of Basic Sciences, Ajou University, 16499 Suwon, Republic of Korea

^✉

To whom correspondence should be addressed. Email: jsjlee@ajou.ac.kr

^✉

Correspondence may also be addressed to Mi Ae Kang. Email: makang@ajou.ac.kr

The first two authors should be regarded as Joint First Authors.

Roles

Soon Young Hwang: Data curation, Formal analysis, Investigation, Methodology, Validation, Writing - original draft

Daeki Choi: Data curation, Formal analysis, Investigation, Methodology, Validation, Writing - original draft

Myoung Joo Choi: Data curation, Formal analysis, Investigation, Methodology, Validation

Hye-Ram Jo: Data curation, Investigation

Jae-Hoon Jeong: Funding acquisition, Investigation, Writing - review & editing

Mi Ae Kang: Funding acquisition, Investigation, Validation, Writing - review & editing

Jong-Soo Lee: Conceptualization, Funding acquisition, Investigation, Supervision, Validation, Writing - review & editing

PMCID: PMC13384258 PMID: 42475683

Abstract

Multifaceted CTCF facilitates double-strand break (DSB) repair via homologous recombination (HR), and BRCA1/BARD1 play key roles in HR. However, whether and how CTCF cooperates with BRCA1/BARD1 for HR remain poorly understood. Here, we show CTCF recruits repressive chromatin proteins, HP1γ, KAP1, SUV39H1, SETDB1, and KDM5A, thereby enabling rapid accumulation of BRCA1/BARD1 at DSBs. CTCF depletion abrogates the enrichment of these repressive chromatin proteins, accompanied by loss of H3K9 methylation/H3K4 demethylation at DSBs, resulting in defective rapid BRCA1/BARD1 recruitment. Deprivation of one component of the repressive chromatin proteins or deregulating H3K9 methylation/H3K4 demethylation compromises the immediate accumulation of the other repressive chromatin proteins as well as BRCA1/BARD1 on DSBs. Consequently, recruitment of the exonucleases EXO1/DNA2 is abrogated, which is causally linked to defective DNA end resection and dysfunctional HR. Collectively, our findings suggest that CTCF establishes a hierarchical recruitment cascade in HR, by coordinating the repressive chromatin assembly and BRCA1/BARD1 recruitment at DSBs, facilitating the exonucleases EXO1/DNA2-mediated extensive DNA resection. Finally, we identify a critical role of CTCF in coupling BRCA1/BARD1 with the HR licensing point, the DNA end resection step, for HR.

Graphical Abstract

For image description, please refer to the figure legend and surrounding text.

Introduction

Mutations in BRCA1 (breast cancer susceptibility gene 1) and BARD1 (BRCA1-associated really interesting new gene [Ring] domain 1) can compromise genome fidelity, causally predisposing carriers to familial breast and ovarian cancers as well as sporadic cancers of diverse origins [1–5]. The closely linked tumor suppressors BRCA1 and BARD1 form an obligate heterodimer complex (BRCA1/ BARD1), which acts as a key mediator and regulator of homologous recombination (HR)-directed repair of double-stranded DNA breaks (DSBs). This complex engages in nucleolytic resection of the DNA ends of DSBs, loading of the recombinase RAD51 onto resected DNA and its stimulation, as well as homologous DNA pairing [6–9].

Extensive studies have unveiled multiple molecular mechanisms that govern the recruitment and function of BRCA1/BARD1 on damaged chromatin. Upon DNA damage, BRCA1/BARD1 accumulates at DSBs through the BRCA1- and/or BARD1-driven pathways via interactions and modifications, where it exerts its principal functions in HR. In the well-characterized BRCA1-driven pathway, the BRCA1-A complex recruits BRCA1/BARD1 to DNA DSBs in an ubiquitin-dependent manner. Upon DNA damage, RNF8 mediates K63-linked polyubiquitination of histone H1 in an ATM-dependent manner [10–13]. Subsequently, the ubiquitin-interacting motif of RAP80 recognizes this ubiquitin chain and recruits the Abraxas-BRCA1 complex via a phosphorylation-binding pocket within the tandem BRCA1 C-terminal (BRCT) domain of BRCA1, thereby promoting BRCA1/BARD1 accumulation at the DSB site. This complex restricts DNA end resection, thereby limiting HR repair, but also it prevents excessive resection, ensuring accurate and efficient HR [14, 15].

Notably, the BARD1-driven mechanism for BRCA1/BARD1 recruitment at DSBs implicate BARD1 C-terminal ankyrin and/or BRCT domains, by which the complex can ultimately facilitate HR of DSBs. One such mechanism involves BARD1 utilizing the DSB signal poly (ADP-ribose) (PAR), where its BRCT domain binds to PAR to recruit BRCA1/BARD1 at the DSBs [16]. Additionally, RNF168-mediated histone H2AK15 ubiquitination (H2AK15Ub), a common DSB mark, binds both the HR-mediator BARD1 and non-homologous end joining (NHEJ)-mediator 53BP1 through the BARD1 BRCT domain [17] and 53BP1 ubiquitination-dependent recruitment motif, respectively [18]. Additional repair pathway-specific marks or steps may shift DSB repair toward either HR or NHEJ. One such specific HR mark is unmodified H4K20 (H4K20me0) incorporated into the chromatin during DNA replication. It is a hallmark of availability of a newly replicative chromatin and sister chromatid for HR that is specifically recognized by the ankyrin domain in BARD1. Thus, it serves to recruit BARD1 tethered to BRCA1 [19], in contrast with the DSB engagement of 53BP1 requiring specific binding to H4K20me2 to facilitate NHEJ [20].

Another mechanism for the BRCA1/BARD1 recruitment at DSBs involves HP1γ binding to H3K9me2/3-accumulated DSB chromatin. This occurs through its chromodomain (CD) recognition of H3K9 methylation at DSBs. Thus, HP1γ anchors BRCA1/BARD1 to DSB chromatin through the interaction of its chromoshadow domain (CSD) with the HP1-binding motif (PxVxL/I) in the BRCT domain of BARD1 [21]. The CD of HP1γ binds H3K9me2/3, while its CSD mediates dimerization and interactions with various proteins, including BARD1 [21, 22]. Besides H3K9me2/3 and HP1γ, histone marks (H3K27me3 and macroH2A) and proteins (KAP1, SUV39H1, and the Polycomb complex) associated with repressive chromatin are enriched around the DSB chromatin [23–30]. KAP1, which serves as a scaffold protein for repressive chromatin components, including H3K9 methyltransferases such as SUV39H1/2 and SETDB1, as well as HP1γ, is localized at DSB sites [23, 24]. However, the upstream mechanisms responsible for such accumulation of these repressive chromatin-related components and molecules and the precise function of the DNA damage-induced repressive chromatin formed around DSBs remain poorly understood.

CCCTC-binding factor (CTCF) is a pleiotropic protein in genome regulation and expression [31]. It was initially identified as a transcriptional repressor and named for its tendency to bind CCCTC repeats. Its role in DNA repair as a genome guardian extends beyond insulation and organization via long-range chromatin looping [31]. Evidence supports that CTCF rapidly accumulates at DSBs across the genome regardless of its original DNA-binding module sequences [31–34]. Moreover, it plays an essential role in HR by interacting with critical proteins, including CtIP, BRCA2, and RAD51, and consecutively recruiting them at DSBs [32, 34]. Additionally, CTCF may facilitate the initial resection of broken DNA ends by collaborating with the endonuclease MRE11 and cofactor CtIP, generating 3′-overhang single-stranded DNA (ssDNA) to promote HR [34]. BRCA1/BARD1 also mediates the nucleolytic resection of DNA lesions and regulation of the recombinase RAD51 [6]. Further, it ubiquitylates H2A K127 (H2AK127Ub) at DSBs, which recruits the chromatin-remodeling factor SMARCAD1. Thus, SMARCAD1 repositions nucleosomes at DSB chromatin to promote long-range DNA end resection [35–37]. Despite the shared role of CTCF and BRCA1/BARD1 in DNA end resection, their functional interplay in HR and its underlying mechanisms remain unclear.

In this study, we demonstrated that CTCF physically and functionally interacts with chromatin-binding proteins (HP1γ and KAP1) and histone-modifying enzymes (SUV39H1/SETDB1 and KDM5A) related with repressive chromatin, and recruits them to DNA lesions, which is required for prompt recruitment of BRCA1/BARD1 at DSBs. Consequently, cells lacking CTCF or one of these proteins exhibit defective BRCA1/BARD1 recruitment at DSBs and resultant impairment of the BRCA1/BARD1-dependent recruitment of SMARCAD1 and exonucleases EXO1/DNA2. This hampers DNA end resection and the dysfunctional HR-mediated repair of DSBs. Additionally, compromising SUV39H1/SETDB1-dependent H3K9 methylation or KDM5A-dependent H3K4 demethylation has similar impact. Taken together, our data reveal a crucial function of CTCF in HR-mediated DNA repair by facilitating the sequential recruitment of pro-HR factors, including repressive chromatin factors and BRCA1/BARD1. This accelerates DNA end resection and establishes CTCF as the mediator of hierarchical relationships in this pro-HR recruitment cascade.

Materials and methods

Cells and reagents

HEK-293T and U2OS cells were provided by American Type Culture Collection and were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Welgene Inc., Gyungsan, Republic of Korea) supplemented with 10% fetal bovine serum (Hyclone, GE Healthcare, Chicago, IL, USA) and 1% penicillin/streptomycin (Welgene Inc.). U2OS-based DSB reporter cell lines were cultured in DMEM with puromycin and stably expressed the following inducible recombinant proteins: AsiSI-Estrogen receptor (ER) (AsiSI-ER-U2OS, gifted by Dr Gaëlle Legube, Université de Toulouse) [38]; and ER-mCherry-LacI-FokI-Destabilization domain (DD) (FokI-U2OS) [39] provided by Dr Roger Greenberg (University of Pennsylvania). Additionally, U2OS-based DR-GFP and EJ5-GFP cells, contributed by Dr Jeremy Stark (Beckman Research Institute of the City of Hope) [40]. All U2OS-based reporter cells were cultured in DMEM without sodium pyruvate (Hyclone) supplemented with 10% fetal bovine serum, 1% penicillin/streptomycin, and puromycin. All cells were cultured in a humidified incubator at 37°C and 5% CO₂ (BB15, Thermo Fisher Scientific). The reagents used in the study were 4′-hydroxytamoxifin (4-OHT), Etoposide, Ku-55933, and Bix 01 294 (Sigma–Aldrich, St. Louis, MO, USA), UNC0638, Chaetocin, and CPI-455 (Selleckchem, Houston, TX, USA), and Phleomycin (InvivoGen, San Diego, CA, USA).

Plasmid construction and transfection

Plasmids contatining full-length (FL) and truncated EGFP-tagged and FLAG-tagged BARD1 were constructed using pEGFP-N1 (Addgene #6085-1) and pcDNA3-FLAG, respectively. The recombinant glutathione S-transferase (GST)-tagged truncated CTCF constructs (N, 1–267; ZF, 246–588; C, 578–727) were generated using pGEX-5X-1 (Addgene #27-4584-01). Constructs of HA-tagged HP1 subtypes and HP1γ point mutants were generated using pcDNA3-HA. Point mutantions were introduced using QuikChange II (Agilent Technologies, Santa Clara, CA, USA). The primer oligonucleotide sequences utilized for cloning and small interfering RNA (siRNA) sequences utilized for knockdown are presented in Supplementary Tables S1 and S2, respectively. Lipofectamine^TM 3000 or Lipofectamine^TM RNAiMAX Transfection Reagent (Invitrogen, Waltham, MA, USA) was used to transfect mammalian cells for expression or knockdown.

Immunofluorescence

After the indicated treatments, the cells were fixed with 4% paraformaldehyde for 10 min at room temperature, permeabilized with 0.5% Triton X-100, and blocked using 1% bovine serum albumin (BSA). The cells were then incubated with indicated primary and secondary antibodies diluted in 1% BSA. The nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; Sigma–Aldrich). Coverslips were mounted using fluorescence mounting medium (Dako, Glostrup, Denmark), and images were captured using confocal (LSM 800, Carl Zeiss, Oberkochen, Germany) or super-resolution confocal microscope (Nikon A1R HD25 N-SIM S, Nikon Instruments Inc., Melville, NY, USA).

Immunoblot assay

Immunoblotting (IB) experiments were performed as previously described [34]. Representative results from at least three independent experiments are shown in the figures.

Immunoprecipitation

After the indicated treatments, the HEK-293T cells were lysed with Benzo lysis buffer (40 mM Tris–HCl, pH 7.4, 150 mM NaCl, 2 mM MgCl₂, 0.2% NP-40, 0.4% Triton X-100) containing protease inhibitor cocktail on ice for 30 min and then incubated with 150 U/ml Benzonase^® (Sigma–Aldrich) for additional 10 min at room temperature. The whole cell lysates were incubated with indicated antibodies for overnight at 4°C and then incubated with protein A beads for 1 h at 4°C. The beads were washed thrice with immunoprecipitation (IP) wash buffer (20 mM Tris–HCl, pH 7.4, 150 mM NaCl, and 0.2% Triton X-100). Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) was performed to separate the immunoprecipitated proteins, which were then immunoblotted using appropriate antibodies.

Irradiation and laser micro-irradiation

Using a dual-source ¹³⁷Cs unit, γ-radiation was delivered at a dose rate of 3.2 Gy/min with a GC-3000 Elan irradiator (MDS Nordion, Ottawa, Canada). For micro-irradiation, U2OS cells were seeded on glass-bottomed cell culture dishes (SPL Life Sciences Co., Ltd., Pocheon, Republic of Korea) and transfected with plasmids or siRNAs as indicated. The cells were pretreated with 20 μM of 5-bromo-2-deoxyuridine (BrdU, Sigma–Aldrich) for 24–30 h and then subjected to laser micro-irradiation for 1 s (32 lines/s) using a Nikon A1R HD25 N-SIM S.

Chromatin immunoprecipitation assay

AsiSI-ER-U2OS cells were treated with 300 nM of 4-OHT for 4 h to trigger the movement of AsiSI-ER into the nucleus and initiate DSB formation through AsiSI nuclease activity. FokI-U2OS reporter cells were modified to express an inducible ER-mCherry-LacI-FokI-DD protein, were treated with 300 nM of 4-OHT and 1 μM Shield-I (Clontech Laboratories, Mountain View, CA, USA) for 4 h, leading to the induction of nuclear expression and stabilization of ER-mCherry-LacI-FokI-DD. This process ultimately generated clusterd DSBs through FokI nuclease activity at specific Lac operator sequences within the transgene. Afterwards, the chromatin immunoprecipitation (ChIP) procedure was performed using the ChIP Assay Kit (Millipore, Burlington, MA, USA), following the guidelines provided by the manufacturer. The immunoprecipitated and input (control) DNA were subjected to quantitative polymerase chain reaction (qPCR) analysis using the Rotor-Gene SYBR® Green PCR kit on a Rotor-Gene Q system (Qiagen, Carlsbad, CA, USA). The PCR protocol consisted of an initial pre-incubation step at 95°C for 5 min, followed by 45 cycles of amplification at 95°C for 5 s and 60°C for 30 s. After the last amplification cycle, a melting curve analysis was performed to verify the specificity of the PCR amplification. Relative IP values were determined by calculating the threshold cycle (Ct) value using the 2^-△△Ct method [41]. The specific primer sequences used for qPCR are listed in Supplementary Table S3.

Protein binding assay

GST fusion CTCF fragment proteins (N, 1–267; ZF, 246–588; C, 578–727) were expressed in Escherichia coli BL21 (DE3) cells (Agilent Technologies, Inc. Santa Clara, CA, USA, 200 131) and purified with Glutathione Sepharose^TM 4B resin (Citiva). To analyze interaction between GST-CTCF and FLAG-BARD1 fragment proteins, protein binding assays were performed by using FLAG-BARD1 fragments (1–424 harboring RING domain; 424R harboring ankyrin and BRCT domains; 570R harboring BRCT domains) immobilized on beads as baits. The pull-downed GST-CTCF fragment proteins were resolved using SDS–PAGE, and binding was analyzed via IB using anti-GST. In addition, GST pulldown assays were performed using whole cellular extracts (WCEs) from KAP1-depleted or control 293T cells complemented with HA-tagged HP1γ to analyze the interaction between GST-CTCF fragment and HP1γ. GST-CTCF fragment proteins (N, ZF, and C) were immobilized on GST-Sepharose beads and incubated with the WCE of KAP1-depleted or control 293T expressing HA-HP1γ proteins. The pulled-down HA-HP1γ proteins were probed via IB using anti-HA antibody.

DNA end resection assay

Genomic DNA was isolated from specific siRNA transfected AsiSI-ER-U2OS cells using a DNeasy blood & Tissue Kit (Qiagen) following the manufacturer’s protocol. Subsequently, the DNA was treated with BamHI to measure resection at chromosome position 1:89 458 296 (CCBL2 gene). To evaluate the level of ssDNA generated by DNA end resection at the targeted AsiSI sites (DSBs), qPCR was conducted using the specific primer sets described in Supplementary Table S4. The percentage of ssDNA (ssDNA%) was determined as previously described [42, 43]. Briefly, for each sample, a ∆Ct value was calculated by subtracting the Ct value of the mock-digested sample from that of the digested sample. The ssDNA% was calculated using the following equation: ssDNA% = 1/(2(∆Ct − 1) + 0.5) × 100 [44].

In situ proximity ligation assay

The proximity ligation assay (PLA) was conducted using specified antibodies, following the manufacturer’s instructions (Sigma–Aldrich) with slight adjustments [45]. Cells were fixed and permeabilized using the same methods as described for immunofluorescence (IF). After co-incubation with primary antibodies, the cells were incubated with proximal probes (anti-mouse PLUS and anti-rabbit MINUS) (Sigma–Aldrich) for 1 h at 37°C. Ligation was performed for 30 min at 37°C, followed by polymerization for 2 h at 37°C with Duolink In Situ Detection reagents Green (Sigma–Aldrich). The cells were counterstained with DAPI, and fluorescence images were captured with a confocal microscope (LSM800).

HR and NHEJ assays

U2OS DR-GFP and EJ5-GFP cells were transfected with siRNA targeting specific genes, and 24 h later, the I-SceI construct was transfected into each cell line. After 48 h, the GFP-positive cell population was analyzed by flow cytometry using a FACSAria III (BD Biosciences, San Jose, CA, USA), and HR and NHEJ efficiencies were evaluated using BD FACSDiva software (Ver. 8.0.3) and Flowing Software 2 (Ver. 2.5.1).

Antibodies

The CTCF antibodies were obtained from Abcam (ab128873, 1:2000 dilution for IB and 1:2000 for IF), Cell Signaling Technology (#2899S, 1:1000 for IP), and Millipore (07-729, 1 μg for each ChIP sample). The other antibodies used for IB, IF, IP, and ChIP analyses were as follows: FLAG (F3165 from Sigma–Aldrich, 1:3000 for IB), HA (ab9110 from Abcam, 1:2000 for IB and IP), GFP (ab290 from Abcam, 1:2000), H2AX (05-636 from Millipore, 1:2000 for IF and IB; ab2893 from Abcam, 1 μg for each ChIP sample), Tubulin (05-829 from Millipore, 1:3000 for IB), BRCA1 (sc-6954 from Santa Cruz, 1:500 for IB and IF; A300-000A from Bethyl laboratories, 1:2000 for IF), BARD1 (GTX132094 from Genetex, 1:2000 for IF and IB; A300-263A from Bethyl laboratories, 1:1000 for IP), HP1α (05-689 from Millipore, 1:2000 for IB), HP1β (MAB3448 from Millipore, 1:2000 for IB), HP1γ (05-690 from Millipore, 1:1000 for IP and 1 μg for each ChIP sample; ab217999 from Abcam, 1:2000 for IF and IB), KAP1 (ab22553 from Abcam, 1:2000 for IF, IB, and IP; 1 μg for each ChIP sample), SUV39H1 (#39785 from Active motif, 1:1000 for IF and 1 μg for each ChIP sample), SETDB1 (ab107225 from Abcam, 1:2000 for IB and IF, and 1 μg for each ChIP sample), KDM5A (ab70892 from Abcam, 1:2000 for IF and IP, and 1 μg for each ChIP sample), KDM5B (H00010765-MO2 from Abnova, 1:2000 for IB and IF), KDM5C (ab34718 from Abcam, 1:2000 for IB and IF), H3K9me2 (07-441 from Millipore, 1 μg for each ChIP sample), H3K9me3 (ab8898 from Abcam, 1:2000 for IB and 1 μg for each ChIP sample; 05-1250 from Millipore, 1:2000 for IF), H3K4me0 (05-1341 from Millipore, 1 μg for each ChIP sample) H3K4me1 (ab8895 from Abcam, 1 μg for each ChIP sample), H3K4me2 (07-303 from Millipore, 1 μg for each ChIP sample), H3K4me3 (ab8580 from Abcam, 1:2000 for IB and IF, and 1 μg for each ChIP sample), pan-H3 (07-690 from Millipore, 1 μg for each ChIP sample), CtIP (61 141 from Active Motif, 1:1000 for IF), RNF168 (ab58063 from Abcam, 1:2000 for IF), SMARCAD1 (A301-593A from Bethyl laboratories, 1:2000 for IF), EXO1 (ab95012 from Abcam, 1:2000 for IF), DNA2 (ab96488 from Abcam, 1:2000 for IF), RAD51 (ABE257 from Millipore, 1:2000 for IF), and BRCA2 (05-666 from Millipore, 1:2000 for IF).

Statistical analysis

The data in this study are presented as means ± standard deviation (SD) and are based on a minimum of three separate replicates. Differences between datasets were analyzed using the Student’s t-test, and statistical significance was set as *P ≤ .05, **P ≤ .01, and ***P ≤ .001.

Results

Rapid enrichment of BRCA1/BARD1 at DSB sites depends on CTCF

CTCF facilitates the recruitment of multiple HR-directed repair proteins, including CtIP, RAD51, and BRCA2, at DNA lesions [32–34]. To clarify the functional relation between CTCF and the crucial HR mediator BRCA1/BARD1 in HR, we first investigated potential reciprocal recruitment to DNA damage sites. Using the previously reported FokI-U2OS reporter system in which the mCherry-LacI-FokI endonuclease fusion protein introduces clustered DSBs in the genome [39], we observed that BRCA1/BARD1 localization at the FokI-induced DSB site was abolished upon CTCF knockdown (Fig. 1A and B). To rule out the possibility of off-target effects of siRNAs, we used three additional independent siRNAs targeting different regions of CTCF messenger RNA (mRNA) and observed comparable phenotype in cells depleted of endogenous CTCF by each of the four siRNAs (Fig. 1A–D and Supplementary Fig. S1A). Rapid recruitment of BRCA1/BARD1 to sites of DNA damage was similarly reduced. This supports the notion that the phenomenon observed in this study is not due to off-target effects of specific siRNAs, but rather to CTCF deficiency itself. For subsequent complementation studies by reintroducing exogenous FL or fragmented CTCF, we used siRNA targeting the 3′UTR of CTCF mRNA (siCTCF) to selectively target endogenous CTCF but not exogenous CTCF, thereby enabling functional analysis of reintroduced siRNA-resistant CTCF through specific depletion of endogenous CTCF within the cells. Similarly, ChIP assay revealed that BRCA1/BARD1 accumulation at the FokI-induced DSB site was dramatically reduced in FokI-U2OS reporter cells lacking CTCF, compared with that in control cells (Supplementary Fig. S1B).

To further validate the requirement of CTCF for localization of BRCA1/BARD1 at multiple arbitrary DSBs, we assessed the foci formation of BRCA1/BARD1 at sites of DNA lesions induced by phleomycin treatment that randomly introduces DNA damage to the whole genome. Consistent with the results from using the U2OS-DSB-reporter system with FokI (Fig. 1A and B; Supplementary Fig. S1B), CTCF depletion remarkably reduced the number of BRCA1/BARD1 foci in the treated cells compared to that in the control U2OS cells (Fig. 1C and D). This confirmed our observation that enrichment of BRCA1/BARD1 at multiple DNA damage sites depends on CTCF. These random foci results raised the possibility that CTCF may efficiently recruit BRCA1/BARD1 onto DNA lesions, irrespective of transcriptional activity or chromatin status prior to DNA damage. This aligns with the previous findings that CTCF robustly localizes CtIP at DSBs in both transcriptionally active (intragenic region at chromosome 1:89 458 296) and inactive (the intergenic region at chromosome 1:110 319 090) chromatin [34, 46]. Nonetheless, accumulating evidence indicates that the occurrence of DSBs at actively transcribed regions, where HR proteins are readily accumulated, causally relates to its more predominant repair by HR than by NHEJ [47–50]. Therefore, to test these two possibilities, we employed an additional AsiSI-ER-U2OS reporter system [38], in which the AsiSI nuclease cuts multiple AsiSI sites throughout the genome, and evaluated the recruitment of BRCA1/BARD1 at the AsiSI-induced DSBs located at transcriptionally active (intragenic) and inactive (intergenic) chromatin regions. First, to confirm whether the intragenic region corresponds to transcriptionally active chromatin and the intergenic region corresponds to transcriptionally inactive chromatin, we evaluated the levels of H3K36me3 deposition and S2 phosphorylation of RNA polymerase II, two highly correlated marker of transcriptionally active chromatin using a ChIP. Two transcriptionally active chromatin markers were more highly accumulated in the intragenic region, compared to the intergenic region (Supplementary Fig. S1C). BRCA1/BARD1 ChIP results revealed that CTCF depletion consistently abolished the accumulation of BRCA1/BARD1 at both regions (Fig. 1E), while the recruitment of BRCA1/BARD1 remained comparable at both regions in the presence of CTCF (Fig. 1E). To generalize the CTCF-dependent BRCA1/BARD1 enrichment at DSBs across intragenic and intergenic chromatin regions, we further examined BRCA1/BARD1 recruitment to DSBs at additional locations (intergenic DSB IV at chromosome 17:77 844 122 and intragenic DSB 3 at chromosome 6:90 404 901) in the presence or absence of CTCF depletion. Consistently, BRCA1/BARD1 recruitment to DSBs occurring in other intragenic and intergenic regions was significantly reduced in CTCF-deficient cells compared to control cells without CTCF depletion (Supplementary Fig. S1D). Collectely, CTCF is essential for the recruitment of BRCA1/BARD1 to sites of DNA lesions regardless of chromatin status or position.

To further ascertain the CTCF-reliant BRCA1/BARD1 accumulation at sites of DNA damage, a laser micro-irradiation system was used. We found that depletion of CTCF abrogated rapid BRCA1/BARD1 recruitment on laser strips at 5 min, but not late recruitment at 20 min (Fig. 1F and G) and 60 min (Supplementary Fig. S1E) after micro-irradiation. As in a functional study on BRCA1 [16], BRCA1/BARD1 accumulation within the first 5 min after micro-irradiation was considered early recruitment, whereas accumulation at later time points was considered late recruitment. Intriguingly, CTCF knockdown had a moderate effect on the late accumulation of BRCA1/BARD1 on the laser strips, contrary to the profound abolishment of the early BRCA1/BARD1 accumulation by CTCF depletion. Consistent with the BRCA1 localization at the laser strips after micro-irradiation (Fig. 1F and G), CTCF depletion significantly reduced endogenous BRCA1 foci after 1 h (early) neocarzinostatin (NCS) treatment, but had little effects on BRCA1 foci after 3 h (late) NCS treatment (Supplementary Fig. S1F). This indicates that CTCF is primarily required for early BRCA1/BARD1 recruitment at sites of DNA damage, whereas other factors/pathways may be responsible for late recruitment. These differences may be due to the biphasic mode of BRCA1/BARD1 recruitment to DNA damage sites: with a rapid γ-H2AX-independent recruitment stage followed by a late γ-H2AX-dependent stage [51], explaining the engagement of CTCF only in the rapid recruitment. In contrast, depletion of either BRCA1 or BARD1 did not affect later CTCF accumulation on the laser strips (Supplementary Fig. S1G), while their depletion severely impaired their heterodimer partner’s recruitment to the FokI/AsiSI-induced and genome-wide DSB sites (Fig. 1A–D). This indicates that CTCF recruitment at DSBs does not depend on BRCA1/BARD1, whereas BRCA1 and BARD1 recruitment is interdependent. Collectively, these observations indicate that rapid accumulation of BRCA 1/BARD1 to sites of DNA damage requires CTCF, but not vice versa.

BRCA1/BARD1 is a component of multiple complexes, including BRCA1-A complex that negatively regulates HR by limiting end resection of DSBs [14], and BRCA1-B or -C complexes that promote it. To determine whether CTCF, which promotes HR by accelerating DNA end resection via recruiting CtIP to DSBs [34], could recruit BRCA1-A complex, we investigated whether CTCF is involved in translocation of RAP80, a BRCA1-A complex member, or BRCA1-A complex recruitment pathway factors MDC1 and RNF8 to DNA damage sites. We analyzed their recruitment to DSBs in CTCF-deficient cells and found that MDC1, RNF8, and RAP80 were recruited to DSBs regardless of the presence of CTCF (Supplementary Fig. S1H). This indicates that CTCF is not implicated in recruitment of MDC1, RNF8, and RAP80 to DNA damage sites, and therefore CTCF may not take a part in translocation of the BRCA1/BARD1 member of BRCA1-A complex to DNA damage sites, which relies on MDC1, RNF8, and RAP80. Collectively, these results (Fig. 1 and Supplementary Fig. S1) suggest that CTCF, which interacts with MRE11-RAD50-NBS1 and CtIP, components of BRCA1-C complex [34], may facilitate rapid accumulation of the BRCA1/BARD1 member of BRCA1-C complex, but not BRCA1-A complex, thereby boosting HR.

CTCF assists rapid BRCA1/BARD1 recruitment on DNA lesions through CTCF-BRCA1/BARD1 interaction

To next address how CTCF contributes to rapid BRCA1/BARD1 enrichment at DNA lesions, we examined whether CTCF might interact with BRCA1 and/or BARD1 using the co-immunoprecipitation (co-IP) assay. We found that CTCF associated with BARD1 but rarely BRCA1 in the absence of DNA damage (Fig. 2A). Etoposide treament augmented the CTCF-BARD1 interaction and also enabled CTCF to interact with BRCA1, although this interaction was much weaker than that with BARD1 (Fig. 2A), indicating that the strength of CTCF-BRCA1/BARD1 interaction can change over time in response to DNA damage. Accordingly, we conducted a time course analysis of the CTCF-BARD1 interaction after γ-irradiation. The interaction was strong at 5 min but declined gradually over time (20, 40, and 60 min) (Fig. 2B). These observations were confirmed by reciprocal co-IP (Supplementary Fig. S2A). Based on our findings that BRCA1/BARD1 depends on CTCF for its early enrichment at DNA lesions (Fig. 1F and G; Supplementary Fig. S1F) and that the CTCF-BARD1 interaction peaks at 5 min post-DNA damage (Fig. 2B and Supplementary Fig. S2A), we speculated that this rapid, CTCF-dependent recruitment of BRCA1/BARD1 at DSBs is driven by CTCF-BRCA1/BARD1 interaction (Fig. 2A and B; Supplementary Fig. S2A). This is evidenced by the correlation between their interaction strength and recruitment timing. Given the DNA damage-induced rapid increment and subsequent reduction in the CTCF-BARD1 interaction (Fig. 2B and Supplementary Fig. S2A), we posited that the CTCF-BARD1 interaction kinetics may be linked to the prompt colocalization of CTCF-BRCA1/BARD1 at DNA lesions and the ensuing separation. Also, we postulated that CTCF may affect the timing of BRCA1/BARD1 recruitment rather than retention at DNA lesions. To test this hypothesis, we performed time-lapse imaging analyses of CTCF-BARD1 recruitment on laser strips. The CTCF recruitment to laser lines was detected within 10 s after micro-irradiation and increased up to 60 s (Fig. 2C), aligning with a previous study [34]. BARD1 showed comparably kinetics, with recruitment detected as fast as 10 s and augmented by 60 s. Despite the coincident recruitment timings of CTCF and BARD1, they became spatially separated at 60 s after micro-irradiation. Initially, CTCF was colocalized with BARD1 on the irradiated lines (up to 60 s after micro-irradiation) (Fig. 2C). However, at 120 s, CTCF from the DSB strips became redistributed to the periphery of the DSB strips, while BARD1 remained at the DSB strips for up to 180 s (Fig. 2C). Consistent with the BARD1 results, BRCA1 was also remained in the laser strips at 180 s (Supplementary Fig. S2B). These spatiotemporal imaging analyses of CTCF-BARD1 in parallel with their interaction kinetics suggested that, initially, CTCF and BARD1 interact tightly so that they are recruited and co-located on the DSB strips immediately after DNA damage; however, CTCF subsequently decouples from BARD1, possibly through the loosened CTCF-BARD1 interaction over time.

Therefore, to investigate whether the CTCF-BARD1 interaction is indeed implicated in the CTCF-dependent rapid BRCA1/BARD1 accumulation at DNA damage sites, the regions responsible for the interaction were identified using truncated fragments of CTCF and BARD1 (Fig. 2D–F). Endogenous BARD1 pulled down FL CTCF and fragments harboring either the N-terminus (residues 1–267) or ZF domains (residues 246–588) (Fig. 2D). Considering the protein expression levels of the N-terminal and ZF fragments, the N-terminus by itself was likely to be associated less with BARD1 than ZF. In contrast, the C-terminus alone could not interact with BARD1. This indicates that the N-terminal and ZF domains are marginally and robustly responsible, respectively, for the CTCF-BARD1 interaction, while the C-terminus is dispensable. Using four fragments of BARD1, its CTCF-interacting region was mapped to the N-terminal and the internal region while C-terminal BRCT domain repeats of BARD1 did not participate in its CTCF interaction (Fig. 2E). [17, 19, 52, 53].

To determine whether the CTCF-BARD1 interaction is direct or indirect and which regions interact directly, we performed pull-down assays using two CTCF-interactable (1–424 and 424R) and one CTCF-non-interatable (570R) FLAG-tagged BARD1 protein fragments immobilized on anti-FLAG M2-conjugated agarose beads, and two BARD1-interactable GST-fused recombinant CTCF protein fragments (two BARD1-interactble N-terminal and ZF, and one BARD1-non-interactable C-terminal) (Fig. 2F). GST-N terminal of CTCF bound to FLAG-BARD1 1–424 fragment, but neither GST-ZF and GST-C-terminal of CTCF bound to any FLAG-BARD1 fragments. This result suggests that CTCF directly interacts with the 1–424 BARD1 fragment via its N-terminal region and indirectly interacts with the internal region of BARD1 (amino acid residues 424–570). The CTCF-interactable 1–424 BARD1 fragment harbors its ubiquitin ligase activity as well as the BRCA1-interactable RING finger domain [52]. Our findings that CTCF interacts more strongly with BARD1 than with BRCA1, and directly binds to the RING finger domain of BARD1, which is hooked to the BRCA1 RING finger domain, arguably indicate that CTCF facilitates the rapid enrichment of BRCA1/BARD1 to sites of DNA damage through the CTCF-BARD1 interaction, which may be a potential mechanism by which BRCA1/BARD1 is rapidly recruited to sites of DNA damage in a CTCF-dependent manner.

To test this possibility, we assessed the recruitment of endogenous BRCA1 and BARD1 to the laser strips in the endogenous CTCF knockdown cells, complemented with or without HA-tagged FL CTCF or its fragments. As shown in Supplementary Fig. S2C and Fig. 2G, BRCA1/BARD1 failed to rapidly translocate onto the laser lines in CTCF-depleted cells; however, the reintroduction of FL CTCF rescued the CTCF-depleted cells from impaired early BRCA1/BARD1 recruitment. The N-terminus and ZF fragment of CTCF, which are capable of interacting with BARD1 (Fig. 2D), restored early BRCA1/BARD1 recruitment to a lesser extent than that in the presence of FL CTCF (Supplementary Fig. S2C and Fig. 2G). The N-terminal showed a modest BRCA1/BARD1 recruitment recovery, while the ZF fragment conferred a robust recovery. The C-terminal, which is unable to interact with BARD1 (Fig. 2D), failed to restore the early BRCA1/BARD1 recruitment to the laser lines. The early BRCA1/BARD1 recruitment recovery conferred by the CTCF fragments correlated with the ability of each fragment to interact with BARD1. The sum of the BRCA1/BARD1 recruitment recovery efficiencies of the N-terminal and ZF fragment was roughly similar to that of FL CTCF. To further verify the importance of CTCF-BARD1 interaction for CTCF-facilitated BRCA1/BARD1 localization at DSBs, we evaluated the translocation of endogenous BRCA1/BARD1 into repair foci after DSB induction using phleomycin treatment in CTCF-knockdown cells. These cells were complemented with or without EGFP-tagged FL CTCF or its fragments (Supplementary Fig. S2D and Fig. 2H). Indeed, we observed a significant reduction in endogenous BRCA1 and BARD1 foci in CTCF-depleted cells after DSB induction (Supplementary Fig. S2D and Fig. 2H, respectively). However, reintroduction of EGFP-tagged FL CTCF recovered both. CTCF-deficient cells containing fragments harboring N-terminus or ZF domains showed less recovery of the BRCA1 and BARD1 foci compared to cells with FL CTCF. Thus the combined recovery efficiencies of the N-terminal and ZF fragments were approximately equivalent to that of FL CTCF. In contrast, complementation of the C-terminus of CTCF failed to rescue the reduced foci of BRCA1 and BARD1 in CTCF-deficient cells. These observations further support that the CTCF-BARD1 interaction is necessary for rapid BRCA1 and BARD1 accumulation at the repair foci formed around DNA lesions, and this is attributable to the correlation between the BRCA1/BARD1 recruitment abilities and BARD1 interaction strengths of the N-terminal and ZF domains of CTCF. Collectively, the complementation analyses reveal that CTCF facilitates BRCA1/BARD1 recruitment at DNA lesions in a CTCF-BARD1 interaction-dependent manner. Through this mechanism, the interactions of N-terminal and ZF regions in CTCF with BARD1 may additively promote BRCA1/BARD1 accumulation at sites of DNA damage. However, these results do not exclude the possibility that additional components co-binding to CTCF and BARD1 act as a bridge to enable CTCF to rapidly recruit BRCA1/BARD1 at the lesions.

CTCF couples with HP1γ to rapidly recruit BRCA1/BARD1 at sites of DNA DSBs

Thus, we postulated that CTCF may act in concert and/or have hierarchical relationships with other BARD1-driven BRCA1/BARD1 recruitment mechanisms, as CTCF is indeed required for early BRCA1/BARD1 recruitment at DSBs and it more strongly interacts with BARD1 than with BRCA1 (Figs 1 and 2; Supplementary Figs S1 and S2). Therefore, to gain further insight into how CTCF operates the BRCA1/BARD1 recruitment pathway, we assessed two known BARD1-driven mechanisms mediated by either RNF168 or HP1γ to link BRCA1/BARD1 with the DSB chromatin [13, 17, 21]. We first examined whether RNF168 may collaborate with CTCF in the early recruitment of BRCA1/BARD1 at DSBs by testing CTCF recruitment onto DSBs in RNF168-depleted cells and vice versa. RNF168 depletion did not affect the CTCF accumulation at DSBs (Supplementary Fig. S3A) or its recruitment kinetics (Supplementary Fig. S3B). Moreover, it exerted no influence on the enrichment of the critical HR factor CtIP at the DSB sites (Supplementary Fig. S3C), which relies on CTCF [34]. CTCF depletion also had no impact on the localization of RNF168 localization and RNF168-mediated H2A K15 ubiquitination at the DSBs (Supplementary Fig. S3D). However, consistent with a previous report [17], RNF168 depletion compromised the early accumulation of BRCA1/BARD1 at the DSBs (Supplementary Fig. S3E), similar to CTCF depletion (Figs 1 and 2). Moreover, co-depletion of CTCF and RNF168 resulted in a more severe defect in BARD1 accumulation at DSBs than individual depletion of either protein alone (Supplementary Fig. S3F), suggesting an additive-like effect. These results indicate that CTCF and RNF168 operate complementary pathways for prompt BRCA1/BARD1 recruitment at DNA lesions.

We also examined whether the CTCF-accelerated BRCA1/BARD1 recruitment at DSBs may be linked to the HP1γ-mediated BARD1 recruitment pathway. In this pathway, bivalent interactions of HP1γ with DNA damage-induced H3K9 methylation [22, 24] and the HP1-binding motif PxVxL (amino acid residues 569–573) overlapping the start of two BRCT repeats (amino acid residues 570–788) of BARD1 [21] provide the molecular bases for BRCA1/BARD1 recruitment at DSBs. We thus investigated the interaction of CTCF with the HP1 family and found that among the three HP1 subtypes, CTCF interacts preferentially with HP1γ, modestly with HP1α, and weakly with HP1β (Fig. 3A and B), suggesting that CTCF primarily interacts with HP1γ. As shown in Fig. 3A and B, the endogenous CTCF-HP1γ interaction was modestly enhanced, similarly to the DNA damage-augmented CTCF-BARD1 interaction (Fig. 2A). This verifies the DNA damage-responsive interactions of CTCF with BARD1 and HP1γ. The discrepancy in the interaction strengths of CTCF with endogenous HP1 subtypes and its relatively higher affinity for HP1γ might be due to the differences in antibody sensitivities. To exclude this possibility, we assessed the interaction strengths of CTCF with the HP1 subtypes through IP, using cells coexpressing exogenous FLAG-tagged HP1 subtypes and HA-tagged CTCF with HA antibody. Despite showing the lowest ectopic HP1γ expression level among all HP1 subtypes, HP1γ interacted robustly with CTCF (Supplementary Fig. S3G). In comparison, CTCF interacted to a much lesser extent with HP1α and barely with HP1β, supporting its preference for HP1γ. To next identify the region(s) of CTCF responsible for the CTCF-HP1γ interaction, we performed forward and reciprocal coimmunoprecipitation (co-IP) experiments using truncated fragments of CTCF (Fig. 3C) or HP1γ (Fig. 3D). Endogenous HP1γ pulled down FL CTCF and the ZF fragment (Fig. 3C). On the contrary, the N-terminus or C-terminus could not interact with HP1γ. The CTCF ZF domain participated in binding to HP1γ (Fig. 3C) in addition to BARD1 (Fig. 2D), while the N-terminus interacted solely with BARD1 (Fig. 2D). According to the reciprocal co-IP results, the HP1γ CD was less likely to be associated with CTCF than FL HP1γ that was proficiently pulled-down with endogenous CTCF (Fig. 3D). This weak interaction between the HP1γ CD and CTCF may be attributed to the absence of the HP1 CSD. This domain is required for HP1 oligomerization that promotes protein–protein interaction and function. The HP1γ CD, mapped to interact with CTCF (Fig. 3D), also interacts with the trimethylation at K9 of H3 (H3K9me3), while its CSD domain interacts with the HP1 binding motif (PxVxL) of BARD1 [21]. Together, these CTCF-BARD1 (Fig. 2D and E) and CTCF-HP1γ (Fig. 3C and D) interactions suggest that the CTCF ZF domain interacts with HP1γ (Fig. 3C) and BARD1 (Fig. 2D), while HP1γ interacts with CTCF and BARD1 via its CD (Fig. 3D) and CSD [21], respectively. Thus, HP1γ may act as a molecular bridge between CTCF and BARD1, enabling CTCF to rapidly recruit BARD1 at DNA lesions (Fig. 3E). This possibility is further supported by the observation that the N terminus of CTCF binds directly to BARD1, whereas the ZF that interacts with HP1γ is unable to bind directly to BARD1 (Fig. 2F).

To test this possibility, we investigated whether CTCF cooperates with HP1γ to recruit BARD1 at sites of DNA damage. We first assessed the HP1γ recruitment at DSBs with and without CTCF knockdown and vice versa. At the FokI-induced clustered DSBs, CTCF depletion abolished HP1γ enrichment, in contrast to its robust accumulation in control cells (Fig. 3F). In a similar manner, HP1γ ChIP results using AsiSI-ER-U2OS reporter systems revealed that HP1γ was recruited to the AsiSI-induced DSBs at both intragenic and intergenic regions, but significantly less in CTCF knockdown cells than in control cells (Supplementary Fig. S3H). Consistent with these results, CTCF knockdown dramatically abated the early HP1γ accumulation on laser-induced DSB strips at 5 min following micro-irradiation, while HP1γ was rapidly enriched in the presence of CTCF (Fig. 3G). Remarkably, HP1γ was barely localized at the DSB strips at 20 min (Fig. 3G) and 60 min (Supplementary Fig. S3I) regardless of CTCF depletion. These results indicate that HP1γ is rapidly and transiently localized at DSBs through a CTCF-dependent mechanism, and shortly displaced from damaged sites. However, depletion of HP1γ had little effect on the recruitment of CTCF to the laser strips (Supplementary Fig. S3J). Thus, our results have established that CTCF interacts with the HP1γ CD through its ZF domain and is required for the rapid and transient recruitment of HP1γ at DNA lesions but not vice versa. Next, we speculated whether the CTCF-HP1γ interaction was required for the CTCF-mediated HP1γ recruitment to sites of DNA damage. Accordingly, we assessed the recruitment of endogenous HP1γ to laser strips in cells cotransfected with siRNA targeting the 3′-UTR of endogenous CTCF and with FL or truncated CTCF. HP1γ was found to be recruited to laser strips in endogenous CTCF-depleted cells expressing exogenous, FL CTCF or the ZF domain (Fig. 3H), which is capable of associating with HP1γ. In contrast, the N- and C-terminal fragments did not restore the impaired HP1γ recruitment to laser strips in CTCF-depleted cells. These results suggest that the ZF domain of CTCF is sufficient for HP1γ recruitment to DNA lesions and CTCF-HP1γ interaction is necessary for HP1γ recruitment to DSBs.

To determine whether and how the CTCF-dependent rapid and transient HP1γ recruitment at DNA lesions acts during early BARD1 recruitment, we analyzed the genetic interactions between CTCF and HP1γ. Thus, we examined early BARD1 recruitment to DSBs in CTCF/HP1γ single- or double-depleted cells with or without reintroduction of their FL or fragments. As expected, single HP1γ and HP1γ-CTCF-combined double depletions abolished seriously early but modestly late BARD1 recruitment at the laser strips (Fig. 4A and Supplementary Fig. S4A), similar to the phenomenon observed in CTCF single-depleted cells (Fig. 1G). This indicates that HP1γ, like CTCF, is required for early BARD1 enrichment on DSBs. Similarly, early BRCA1 localization on laser strips was severely affected by HP1γ single depletion or its co-depletion with CTCF, but late BRCA1 enrichment was less affected (Supplementary Fig. S4B). This suggests that both BRCA1 and BARD1 are recruited on DSBs in an HP1γ-dependent manner. Deprivation of BARD1 did not affect HP1γ localization on the laser strips (Supplementary Fig. S4C), verifying that HP1γ is essential for early BARD1 recruitment into DNA lesions but not vice versa. We next found that only FL HP1γ but not its fragments (CD and CSD) restored BARD1 recruitment onto laser strips (Fig. 4B) and γH2AX foci as a marker for genome-wide DSBs (Supplementary Fig. S4D) in HP1γ-depleted cells, indicating that early BARD1 enrichment at DNA lesions requires both CD and CSD of HP1γ, which interacts with CTCF and BARD1, respectively. However, neither FL nor fragments of HP1γ could robustly rescue HP1γ and CTCF co-depleted cells from impaired early BARD1 recruitment to laser strips (Supplementary Fig. S4E). This suggests that CTCF acts upstream of HP1γ for early BARD1 accumulation at DSBs, where it first accelerates HP1γ recruitment, which in turn leads to early BARD1 enrichment. This is consistent with the theory that CTCF uses HP1γ as a molecular bridge that interacts with both CTCF and BARD1 (Fig. 3D and E; [21]) to rapidly recruit BARD1 at DNA lesions (Fig. 2C).

If CTCF acted in concert with HP1γ for early BRCA1/BARD1 recruitment at DNA lesions, ZF and other CTCF fragments may fail to rescue BRCA1/BARD1 enrichment in CTCF- and HP1γ co-depleted cells. The ZF domain is capable of binding to HP1γ and could, by itself, recover its BRCA1/BARD1 recruitment ability in CTCF single-depleted cells. However, we found that the CTCF fragments harboring the N-terminal region (N or N-ZF) modestly rescued impaired early BARD1 localization on laser lines in the co-depleted cells (Fig. 4C). Meanwhile, both the N-terminal and ZF fragments rescued the localization in CTCF single-depleted cells (Fig. 2G and H). Importantly, the observation that the N-terminal fragment of CTCF by itself exhibited a modest BARD1 enrichment recovery in the co-deprived cells aligns with the idea that CTCF promotes early BARD1 recruitment both with and without HP1γ, via an additional parallel pathway. The differential recovery efficiencies delivered by FL CTCF and its fragments raise indicate that the two CTCF-mediated pathways (with and without the assistance of HP1γ) additively support early BARD1 recruitment at DNA lesions. The CTCF-HP1γ pathway involving the ZF domain has greater responsibility than the HP1γ-independent CTCF pathway involving the N-terminus. Consistent with this idea, we found that a BARD1 mutant (PEELI) lacking the intact HP1-binding motif was much less and faintly recruited at laser lines, compared with WT BARD1 (Fig. 4D and E). In this mutant, the center valine residue required for HP1 binding, PxVxI (x indicates any amino acid), is changed to glutamate [21]. This supports the idea that the CTCF pathway uncoupled to HP1γ may modestly act in the early BARD1 recruitment at DNA lesions. Further, we reason that this lower BARD1 recruitment level is attributable to the weak interaction of CTCF with the HP1-binding defective BARD1 mutant (PEELI) because CTCF marginally interacted with the BARD1 mutant (Fig. 4F) through its N-terminus (Fig. 2D–F). In contrast, the efficient WT BARD1 recruitment coupled to HP1γ was achieved by the strong interaction of CTCF with BARD1 through the ZF domains of CTCF and the CD of HP1γ (Fig. 3C and D). In this interaction, HP1γ was found to be sandwiched between the two regions and potentially acted as a bridge between them (Fig. 3C–E; [21]) in concert with the weak interaction of the CTCF N-terminus with the BARD1 N-terminus (Fig. 2D–F). This idea was also supported by the reduced interaction between CTCF and BARD1 in HP1γ-depleted cells (Fig. 4G). The reduced interaction might causally link to the reduced BARD1 recruitment at DSBs during HP1γ deficiency. This indicates that CTCF and HP1γ promote early BARD1 recruitment at DSBs through their interdependent interactions with BARD1.

CTCF supports transient enrichment of H3K9 methylation to promptly recruit BRCA1/BARD1 around DSBs

We found that the methylated H3K9 (H3K9me) reader HP1γ is rapidly and transiently enriched around DSBs in both intragenic and intergenic regions [34, 46] through an CTCF-dependent mechanism (Fig. 3G and Supplementary Fig. S3H). This is required to rapidly recruit BARD1 (Fig. 4), raising the possibility that CTCF may promote the local induction of H3K9me proximal to DSBs to facilitate HP1γ-mediated BARD1 recruitment. We therefore tested whether the surrogate repressive histone marker H3K9me3, which is recognized and bound by HP1γ [21], is induced around DSBs and relies on CTCF. To assess H3K9 methylation around DSB chromatin, we performed IF microscopy of FokI-U2OS reporter cells. Increased H3K9me3 were detected in proximity to and at the FokI-broken DSB foci; however, CTCF depletion disrupted its DSB colocalization (Fig. 5A). This suggests that H3K9me3 is enhanced in the vicinity of and at the FokI-induced break foci in a CTCF-dependent manner. Rapid and transient accumulation of the H3K9me3 reader HP1γ around DSB sites prompted us to examine whether H3K9me3 accumulation at DSBs occur transiently. Consistent with this idea, laser micro-irradiation microscopy revealed that H3K9me3 was rapidly and transiently elevated at DSB strips at 5 min post-DNA damage, but at 20 min, the accumulation was undetectable (Fig. 5B). However, CTCF deprivation perturbed the early H3K9 trimethylation at DSB sites (Fig. 5B), suggesting CTCF-dependent rapid and transient H3K9me3 establishment around the DSBs. To verify CTCF-dependent DSB induction of H3K9 methylation, we assessed the H3K9me2/3 levels around DSBs in intragenic and intergenic regions generated by the AsiSI restriction enzyme. DSBs increased H3K9me2 and H3K9me3 levels by >10-fold compared to those in the control, while the total H3 amount remained unchanged (Fig. 5C and Supplementary Fig. S5A). However, similar to the defective HP1γ recruitment at DSBs during CTCF deprivation (Fig. 3F and G), H3K9me2/3 levels did not elevate around DSBs in the absence of CTCF (Fig. 5C and Supplementary Fig. S5A). This indicates that CTCF is required to rapidly increase H3K9me2/3 levels (Fig. 5A–C and Supplementary Fig. S5A) as well as transient and rapid HP1γ at DSBs (Fig. 3G and Supplementary Fig. S3H). This spatiotemporal correlation between the accumulation of H3K9me3 and HP1γ at DSBs implies that the induction of repressive chromatin, such as H3K9me3, causally relates to the accumulation of its reader, HP1γ, which in turn facilitates BARD1 recruitment at DSBs. This idea—relationship between the accumulation of H3K9me3 and HP1γ at DSBs—is further supported by the findings that, reintroduction of FL CTCF or its ZF fragment, which is capable to recover HP1γ recruitment at DSBs in endogenous CTCF-depleted cells, restored defective H3K9me3 accumulation onto laser strips (Supplementary Fig. S5B). In contrast, the N- or C-terminal fragment, which can not rescue poor recruitment of HP1γ at DSBs, did not restore the impaired trimethylation at H3K9 to laser strips in CTCF-depleted cells (Supplementary Fig. S5B).

Given that in the CTCF-dependent mechanism, H3K9me2/3 was transiently enriched around DSB chromatin (Fig. 5A–C), we considered whether histone-modifying enzymes related to H3K9 methylation might be locally recruited upon DSB damage through this mechanism. In accordance with the transient H3K9me2/3 accumulation at DSBs, we found that H3K9 methyl transferases—SET domain bifurcated 1 (SETDB1) and suppressor of variegation 3–9 homologue 1 (SUV39H1)—were accumulated at the laser strips during the initial response (5 min) to laser micro-irradiation and were released thereafter (20 min) (Fig. 5D and Supplementary Fig. S5C). Also, we examined whether KRAB-associated protein 1 (KAP1), which forms a repressive chromatin-associated protein complex with SETDB1, SUV39H1, and HP1, and works in transcriptional repression and DNA repair [54], might engage in DSB-induced transient H3K9 methylation in the CTCF-reliant way. Similar to SETDB1 and SUV39H1, KAP1 was also recruited to DSB strips early (5 min) after micro-irradiation and then released later (20 min) (Fig. 5D and Supplementary Fig. S5C). In contrast, we did not detect their recruitment at DSBs in CTCF-depleted cells (Fig. 5D), suggesting that SETDB1, SUV39H1, and KAP1 require CTCF for their rapid and transient recruitment. This is consistent with the CTCF-dependent rapid and transient accumulation of HP1γ (Fig. 3G) and H3K9me3 (Fig. 5B) at DSBs.

To next address whether SETDB1, SUV39H1, and KAP1 functionally relate to the accumulation of H3K9me3 and HP1γ to rapidly recruit BRCA1/BARD1 and at DSBs, we assessed the early localization of H3K9me3 and BRCA1/BARD1 at laser strips in SUV39H1-, SETDB1-, KAP1, or HP1γ-depleted cells. The specific depletion of these proteins from the KAP1 complex abrogated immediate H3K9me3 accumulation (Fig. 5E) following laser micro-irradiation, suggesting that SUV39H1, SETDB1, KAP1, and HP1γ are required for H3K9me3 accumulation at DSBs. However, CTCF recruitment at DSBs was not detectably affected (Supplementary Figs S3J and S5D). This indicates that SUV39H1 and SETDB1 mediate the prompt and transient H3K9 methylation around the chromatin proximal to DSBs, and HP1γ may subsequently bind and stabilize the DSB-induced H3K9me3. Deprivation of HP1γ, SUV39H1, SETDB1, or KAP1 consequently disturbed early BRCA1/BARD1 recruitment to the laser strips (Fig. 5F), further indicating that BRCA1/BARD1 rapidly accumulates at DSBs through its interaction with HP1γ bound to H3K9me3. Notably, both of the histone methyltransferases, SETDB1 and SUV39H1, are required for the rapid recruitment of BRCA1/BARD1 at DSBs, and either one is insufficient on its own. These results support the idea that CTCF facilitates the recruitment of the KAP1 complex components, SUV39H1 and SETDB1, and their scaffold KAP1, to establish H3K9me3 around the DSB chromatin. This in turn stimulates HP1γ assembly to stabilize H3K9me3 and allow rapid BRCA1/BARD1 recruitment through HP1γ-BARD1 interaction (Fig. 4; [21]). Consistently, inhibitors blocking H3K9 methylation, including UNC-0638, chaetocin, and Bix-01294, disturbed the immediate accumulation of H3K9me3, HP1γ, and BRCA1/BARD1 but not CTCF at the DSBs (Supplementary Fig. S5E–G). To further support the upstream role of CTCF-dependent repressive chromatin proteins-mediated H3K9me3 accumulation in early BRCA1/BARD1 recruitment at DSBs (Fig. 5E and F; Supplementary Fig. S5E and F), we confirmed that deprivation of BRCA1 or BARD1 had little effects on H3K9me3 enrichment at DSBs (Supplementary Fig. S5H). These findings corroborate our cellular observations regarding the collaboration of HP1γ and SUV39H1/ SETDB1 in H3K9me3 accumulation at DNA lesion for this BRCA1/BARD1 recruitment cascade directed by CTCF.

We next investigated how CTCF assists the recruitment of the KAP1 complex components—HP1γ, SETDB1, SUV39H1, and KAP1—at DSBs. CTCF reportedly acts as a scaffold for several DNA repair factors to promote their recruitment at sites of DNA damage [31, 34]. Consistent with this idea, SETDB1 and KAP1 have been identified as putative CTCF interaction partners in proteomic studies [34, 55]. We therefore examined their interactions and, as expected, detected interaction of CTCF with HP1γ, SETDB1, SUV39H1, and KAP1 by co-immunoprecipitating CTCF-containing protein complexes (Supplementary Fig. S5I and J). Also, their interaction with CTCF augmented in response to DNA damage, indication that CTCF interactions with the KAP1 complex components may causally contribute to their early recruitment at DSBs but also their later release from DSBs. To test this hypothesis, we performed a time-course analysis of the CTCF interactions with HP1γ, SUV39H1, or KAP1 after γ-irradiation. Their interactions with CTCF were strong at 5 and 20 min but declined to basal levels after 60 min (Supplementary Fig. S5I). These results support that the enhanced interaction of CTCF and KAP1 complex components during the early response (5 min) to γ-irradiation is associated with the rapid recruitment (5 min) of these components to the DSB strips. Also, the strong interaction of CTCF with the KAP1 complex during the late response (20 min) is likely mechanistically linked to its later release from the DSB (20 min). When CTCF is recruited to DNA damage sites (within seconds after microirradiation, Fig. 2C), CTCF can rapidly bring KAP1 complex components to the DNA damage site through the strong DNA damage-induced interaction. Meanwhile, when CTCF is released from DNA lesions (20 min after microirradiation, Supplementary Fig. S5K), CTCF may help remove these components from the DNA damage site through the strong interaction between CTCF and KAP1 complex components.

Given the similarity in the CTCF-dependent transient localization of the KAP1 complex components at sites of DNA damage (Figs 3G and 5D; Supplementary Fig. S5C) and in the dynamics of CTCF interaction with them in response to DNA damage (Supplementary Fig. S5I), it is likely that KAP1 complex components interact with CTCF as a single complex and respond to DNA damage as a single complex.

To test this hypothesis, we examined whether the depletion of KAP1 affects the interactions of CTCF with the histone methyltransferase and HP1γ, based on the report that KAP1 forms a complex with HP1 and SETDB1 by acting as a scaffold [56]. KAP1 deprivation caused significant reduction of the CTCF interactions with SUV39H1/SETDB1 and HP1γ in both the presence and absence of DNA damage (Supplementary Fig. S5J), suggesting that the KAP1 scaffold is necessary for the strong interaction of CTCF with SUV39H1/SETDB1 and HP1γ. To further determine whether KAP1 complex components interact with CTCF as a single complex, we examined the effect of KAP1 on the interaction between CTCF and HP1γ using a pull-down assay (Supplementary Fig. S5L). KAP1 depletion reduced the interaction between HP1γ and ZF fragment of CTCF, which can bind HP1γ (Fig. 3C and E; Supplementary Fig. S5L). These results indicate that KAP1, as a scaffold, forms a complex with SUV39H1/SETDB1 and HP1γ, which as a complex, strongly interacts with CTCF, thereby promoting the CTCF-directed rapid translocation of SUV39H1/SETDB1 and HP1γ to DSB chromatin in response to DNA damage. This leads to the accumulation of H3K9me3 and HP1γ, and eventually BRCA1/BARD1, on the lesions (Fig. 5E and F). The interaction of CTCF with the components of KAP1 complex raised a hypothesis that CTCF may recruit the components to DNA lesions as a single complex. Thus, we investigated whether the specific deprivation of HP1γ, KAP1, SUV39H1, or SETDB1 in the KAP1 complex affects the recruitment of other components of the complex at DSBs. We found that deprivation of even one component of the complex perturbed the recruitment of the other components (Supplementary Fig. S5M). This indicates that the recruitment of each component is interdependent with that of the others, and the KAP1 complex containing at least HP1γ, KAP1, SUV39H1, and SETDB1 is localized at the DSB chromatin as a functional unit. Collectively, CTCF acts upstream of the KAP1 complex harboring HP1γ, SUV39H1, and SETDB1 to accumulate at the DSB chromatin in an interdependent manner and to establish the H3K9me3 platform. This promotes the rapid BRCA1/BARD1 recruitment to DSBs in response to DNA damage.

CTCF assists the transient disappearance of H3K4me3 to promote rapid BRCA1/BARD1 localization at sites of DNA damage

Lysine demethylase 5A (KDM5A) is recruited to damaged DNA, causing the removal of methyl groups from a marker of active chromatin, methylated histone H3 at lysine 4 (H3K4me3) [57, 58], and strengthening HR repair [59]. Therefore, we speculated whether H3K4me3 might be relieved (i.e. demethylated) from DSB chromatin sites, in line with the CTCF-dependent H3K9me2/3 increment at these sites, to establish repressive chromatin in vicinity of the DSBs (Fig. 5C and Supplementary Fig. S5A). We detected the disappearance of H3K4me3 from the FokI-induced DNA break foci in FokI-U2OS reporter cells. However, H3K4me3 might be irregularly and randomly dispersed across the DSB foci in CTCF-depleted cells (Fig. 6A and Supplementary Fig. S6A), suggesting that H3K4me3 is erased from the DSB sites in a CTCF-dependent manner. To validate the DSB-induced H3K4me3 reduction, we examined whether H3K4me3 is erased from DSB strips upon laser micro-irradiation. Consistent with a previous study [57], H3K4me3 was removed from the DSB strips at 5 min after micro-irradiation, while H3K4me3 was found randomly and irregularly dispersed across the DSB chromatin at 20 min (Fig. 6B and Supplementary Fig. S6B). This indicates the transient disappearance of H3K4me3 upon DNA damage. In accordance with the CTCF-dependent transient H3K9me3 accumulation at the DSB strips (Fig. 5B), no changes in H3K4me3 were detected in CTCF-depleted cells, suggesting that CTCF ensures the transient removal of H3K4me3 from DSB chromatin following DNA damage. Next, we evaluated H3K4me3, H3K4me2, H3K4me1, and unmethylated H3K4me0 levels at DSBs through ChIP. We found that the level of H3K4me3 dropped as it did at the FokI- and micro-irradiation-induced DSBs (Fig. 6A and B), and consequently, the levels of all three demethylated H3K4me0/1/2 from H3K4me3 increased at the FokI-induced DSB chromatin in a CTCF-reliant manner (Fig. 6C and Supplementary Fig. S6C). Furthermore, the depletion of CTCF impaired DSB-induced H3K4me3 reduction along with the increase in H3K4me0/1/2 (Fig. 6C and Supplementary Fig. S6C), indicating that the two processes occur in a CTCF-dependent manner. The repressive chromatin marks, H3K9me3, HP1, and H3K4me0 are functionally interdependent in the formation and maintenance of damage-induced repressive chromatin, where one modification promotes the other, or HP1 binding to H3K9me3 promotes stabilization or further deposition of H3K9me3 [22, 60]. This raises the possibility that accumulation of H3K9me3 and HP1γ at the sites of DNA damage can promote H3K4 demethylation. To test this possibility, we investigated whether the ZF fragment of CTCF, which can bind HP1γ (Fig. 3C) and rescues impaired accumulation of HP1γ and H3K9me3 in CTCF-depleted cells (Fig. 3H and Supplementary Fig. S5B), could rescue H3K4 demethylation. Similar to the restoration of HP1γ and H3K9me3 accumulation at DSB strips by reintroduction of FL CTCF or its ZF fragment (Fig. 3H and Supplementary Fig. S5B), H3K4me3 demethylation at DSB strips was restored in endogenous CTCF-depleted cells expressing exogenous FL or ZF domains (Supplementary Fig. S6D). These results suggest that HP1γ, H3K9me3, and H3K4me0 interdepedently function in the CTCF-reliant transient damage-induced repressive chromatization at DNA damage sites.

To identify the KDM5 (lysine demethylase 5) subtypes (KDM5A–C) primarily responsible for DSB-induced H3K4me3 demethylation, we investigated the ones recruited at the DSB chromatin. PLA [61] allowed the in situ detection of protein interactions and revealed that KDM5A was colocalized with the canonical DNA damage maker γH2AX upon laser micro-irradiation, while CTCF depletion abrogated KDM5A collocation with γH2AX (Fig. 6D). This indicates that KDM5A is located at DSB strips during the immediate response to DNA damage in a CTCF-dependent manner, following which it may demethylate H3K4me3 at the DSB chromatin (Supplementary Fig. S6E). On the other hand, KDM5B was recruited at DSB strips following laser micro-irradiation regardless of CTCF depletion, while KDM5C was not recruited at all (Supplementary Fig. S6E). However, the CTCF-independent KDM5B localization at DSB did not exclude the possibility that KDM5B as well as KDM5A, may demethylate H3K4me3 at the DSB chromatin. We thus investigated whether deprivation of KDM5A/B affects the removal of H3K4me3 from DSB chromatin. We found that KDM5A deprivation failed to erase H3K4me3 from DSB chromatin, whereas H3K4me3 around the DSB chromatin was not visibly affected in KDM5B-depleted cells (Fig. 6E). This demonstrates the functional relationship between KDM5A recruitment and the subsequent H3K4me3 removal from DSB chromatin, which is supported by the CTCF-dependency of both processes. These results suggest that KDM5A could indeed demethylate H3K4me3 at DSB chromatin to erase H3K4me3 in a CTCF-dependent manner (Fig. 6A and B). Moreover, reflecting the CTCF-dependent localization of KDM5A at DSBs (Fig. 6C and D; Supplementary Fig. S6E) and CTCF-independent localization of KDM5B (Supplementary Fig. S6E), CTCF interacted with KDM5A but not with KDM5B/C (Fig. 6F). These findings on the functional and physical association of CTCF and KDM5A were validated by the restoration of the impaired KDM5A localization at DSBs in CTCF-depleted cells with the reintroduction of FL CTCF or its ZF domain (Supplementary Fig. S6F).

To investigate the functional significance of KDM5A-mediated demethylation of H3K4me3 at DSB chromatin, we assessed early BARD1 recruitment at DSBs in KDM5A-depleted cells. KDM5A deprivation perturbed the recruitment of BARD1 and its partner BRCA1 to the DSB sites, while KDM5B deprivation had little effect (Fig. 6G and Supplementary Fig. S6G). This suggests that KDM5A mediates H3K4me3 displacement from DSBs, contributing to the early recruitment of BRCA1/BARD1 to DSBs. To verify the relevance of H3K4me3 demethylation at the DSB chromatin to BRCA1/BARD1 recruitment, we examined the effects of a KDM5 inhibitor, CPI455, on BRCA1/BARD1 recruitment at DSBs. Early localization of BARD1 as well as BRCA1 at DSBs was impaired by the KDM5 inhibitor (Supplementary Fig. S6H), further indicating that KDM5A-mediated H3K4me3 demethylation is essential for this process.

Based on the findings that the two histone modifications associated with repressive chromatin, H3K9me3 accumulation (Fig. 5A–C) and H3K4me3 dissolution (Fig. 6A–C), simultaneously and transiently occur around the DSB chromatin, we determined whether DNA damage might induce them as interdependent co-adjuvants. To test this hypothesis, we investigated whether KDM5A depletion influences H3K9me3 accumulation at DSB chromatin, and whether SUV39H1 or SETDB1 depletion influences H3K4me3 dissolution from the chromatin. We found that depletion of KDM5A or the KDM5A inhibitor abrogated the accumulation of H3K9me3 at FokI-induced DSBs (Supplementary Fig. S6I). Similarly, depletion of SETDB1 or SUV39H1 and H3K9 methyltransferase inhibitors impaired H3K4me3 disappearance from FokI-induced DSBs (Supplementary Fig. S6J). This interdependency between H3K9me3 accumulation and H3K4me3 removal indicates that the two repressive chromatin markers may collaborate in BRCA1/BARD1 recruitment at DSBs. This is supported by the further decrease in BRCA1/BARD1 recruitment to laser strips in KDM5A and SUV39H1-codepleted cells, compared with that in either KDM5A or SUV39H1 single-depleted cells (Supplementary Fig. S6K). Cotreatment with inhibitors of H3K9 methylase and H3K4 demethylase disturbed BRCA1/BARD1 recruitment to DSB strips more severely than single-inhibitor treatment (Supplementary Fig. S6K).

CTCF-reliant rapid recruitment of BRCA1/BARD1 drives EXO1/DNA2-mediated extensive resection of DSB ends

BRCA1/BARD1-mediated ubiquitination of H2A or H2AX functions as a recognition site to recruit a chromatin remodeling factor, SWI/SNF-related matrix-associated actin-dependent regulator of chromatin (SMARCAD1). This factor facilitates the recruitment of exonucleases EXO1/DNA2 through sliding or eviction of nucleosomes to support long-range resection [62]. Therefore, we next examined whether and how CTCF assists in establishing the BRCA1/BARD1–SMARCAD1–EXO1/DNA2 pathway for extensive resection at DSBs via early BRCA1/BARD1 recruitment. First, we examined the recruitment of SMARCAD1 and EXO1/DNA2 at the DSBs in CTCF-depleted cells. The localization of SMARCAD1 and EXO1/DNA2 was ablated by CTCF deprivation (Fig. 7A), indicating that their recruitment at DSBs relies on CTCF and that CTCF acts upstream of their BRCA1/BARD1-mediated recruitment. Notably, we found the difference in recruitment timings between BRCA1/BARD1 (Fig. 1G) and SMARCAD1–EXO1/DNA2 (Supplementary Fig. S7A). The enrichment of SMARCAD1 and EXO1/DNA2 at DSBs occurred later (at ~5 and 10 min, respectively) than that of BRCA1/BARD1, HP1γ, SUV39H1, SETDB1, KDM5A, and KAP1 (within 1 min) immediately after micro-irradiation (Figs 1–6 and Supplementary Fig. S7A). This may correspond with steps in the HR process where these components function. Accordingly, the difference in the CTCF-dependent localization time-points, from rapid HP1γ–SUV39H1–SETDB1–KDM5A–KAP1 and then BRCA1/BARD1 to SMARCAD1 and finally EXO1/DNA2 may be attributable to steps in the HR process in which they participate. The repressive chromatin-associated proteins (HP1γ, SUV39H1, SETDB1, KDM5A, and KAP1), which were instantly recruited at the DSBs (within 1 min), are likely to establish the repressive chromatin context around DSBs to concurrently promote rapid BRCA1/BARD1 recruitment at the 1-min point (Figs 1–4). Moreover, BRCA1/BARD1 may in turn recruit SMARCAD1 and EXO1/DNA2 within ~5 min (Fig. 7A and Supplementary Fig. S7A) to extensively resect the DSB ends via chromatin remodeling. These results suggest that CTCF may contribute to extended resection by facilitating EXO1/DNA2 recruitment to DSBs, in addition to the initial resection via CtIP recruitment [34].

To next ascertain whether CTCF-dependent rapid BRCA1/BARD1 recruitment may be causally linked to the recruitment of SMARCAD1 and subsequent EXO1/DNA2 at DSBs, we assessed the requirement of BRCA1 or BARD1 for the localization of SMARCAD1 or EXO1/DNA2 at DSBs. Depletion of BRCA1 or BARD1 reduced the enrichment of SMARCAD1 and EXO1/DNA2 (Fig. 7A) at DSBs.

This suggests the possibility that CTCF facilitates rapid BRCA1/BARD1 recruitment, whereby EXO1 and DNA2 may be readily localized at DSBs in a sequential CTCF- and BRCA1/BARD1-co-dependent manner. Further, this theory was also supported by the localization of SMARCAD1 and EXO1/DNA2 at DSBs, which were aborted in cells lacking HP1γ, SUV39H1, SETDB1, KAP1, or KDM5A where rapid BRCA1/BARD1 recruitment at DSBs was equivalently impaired (Fig. 7A and Supplementary Fig. S7B). Similarly, in cells treated with inhibitors of the histone-modifying enzymes, SUV39H1, SETDB1, and KDM5A, rapid BRCA1/BARD1 recruitment was consistently ablated (Supplementary Figs S5F and S6H) and SMARCAD1 and EXO1/DNA2 failed to locate at the DSBs (Supplementary Fig. S7C). This indicates that enrichment of SMARCAD1 and EXO1/DNA2 at DSBs depends on the CTCF-facilitated rapid BRCA1/BARD1 localization at DSBs (Fig. 7A).

Next, we investigated whether rapid localization of BRCA1/BARD1 at DSBs in a CTCF-dependent manner (Fig. 1) is functionally related to the extensive resection of DSB ends by EXO1/DNA2. For this, we conducted a DNA end resection assay by which resection activity was measured in an AsiSI reporter system [ 34, 43] with and without BRCA1- or BARD1-knockdown. DNA end-resection activity in BRCA1- or BARD1-deficient AsiSI reporter cells was remarkably reduced, to a similar extent as that with CTCF depletion (Fig. 7B). This suggests that early BRCA1/BARD1 is required for the proficient resection of DSB ends in HR. Consistent with the reduced end resection in BRCA1/BARD1-deficient cells, depletion of HP1γ, SUV39H1, SETDB1, KAP1, or KDM5A (Fig. 7B) and their inhibitors (Fig. 7C) disrupted the end resection at DSBs. This demonstrates the co-dependence of end resection on HP1γ, SUV39H1, SETDB1, KAP1, and KDM5A—all factors required for rapid BRCA1/BARD1 assembly and subsequent BRCA1/BARD1-dependent EXO1/DNA2 recruitment at DSBs (Figs 3–6), suggesting that EXO1/DNA2-mediated DNA end resection at DSBs depend on CTCF, the repressive chromatin factors (HP1γ, SUV39H1, SETDB1, KAP1, and KDM5A), and BRCA1/BARD1.

Further, we determined whether DNA end-resection activity is governed by the preexisting chromatin state or by DSB-induced transient repressive chromatin. We performed a resection assay in which inhibitors of H3K9 methyl transferases or H3K4 demethylases were pretreated before DNA damage for either a short period (2 h before DSB induction; may inhibit enzyme activity close to or at the time of DNA damage) or a long period (12 h before DSB induction; may establish a preexisting, relatively less repressive chromatin state). Regardless of the pretreatment duration times, all inhibitors impaired resection activity (Fig. 7C). This suggests that the recruitment of BRCA1/BARD1 is facilitated by the DSB-induced transient repressive chromatin rather than the preexisting chromatin state, which might be crucial for promoting DNA end resection.

We also noted that recruitment of HP1γ, SUV39H1, SETDB1, KAP1, KDM5A, BRCA1/BARD1, and SMARCAD1, which are essential for EXO1/DNA2 recruitment at DSBs, was CTCF-dependent (Figs 1–6 and 7A; Supplementary Fig. S7B). CTCF has been implicated in the initiation of DNA end resection by recruiting CtIP to DNA damage sites [34]. Therefore, we examined the possibility that, the CTCF-dependent repressive chromatin factors facilitate CtIP recruitment at DSBs, and this might in turn affect end resection. Unlike CTCF deficiency, depletion of any of these factors had no effect on CtIP recruitment (Supplementary Figs S3C and S7D). This suggests that early BRCA1/BARD1 and its recruitment regulators, the repressive chromatin factors, may promote extensive resection mediated by EXO1/DNA2 rather than the initial resection mediated by CtIP, whereas CTCF engages in both initial and extensive resections.

Based on the defective recruitment of SMARCAD1 and EXO1/DNA2 to DSBs observed in cells lacking early BRCA1 at DSBs, we sought to determine whether early BRCA1/BARD1 actually plays a critical role in recruiting these factors. To this end, we analyzed the recruitment of SMARCAD1 and EXO1/DNA2 to DSBs and their colocalization with BRCA1 under three distinct conditions: (i) where only late BRCA1/BARD1 was present at DSBs via the depletion of CTCF, HP1γ, or other individual repressive chromatin factors, (ii) where only early BRCA1/BARD1 was present at DSBs via ATM inhibitor treatment; and (iii) where both early and late BRCA1/BARD1 were absent at DSBs via either BRCA1 depletion or ATM inhibitor treatment in combination with the depletion of CTCF or HP1γ. In control cells, BRCA1 quickly accumulated on laser strips within 1 min after microirradiation and persisted thereafter (Figs 1F and 2C). Concurrently, SMARCAD1 and EXO1 readily accumulated on laser strips within 5 min and subsequently disappeared at 20 min after microirradiation (Supplementary Figs S7A and S8). In marked contrast, BRCA1 depletion abolished the accumulation of SMARCAD1 and EXO1 at DSBs (Fig. 7A), confirming that BRCA1 is essential for the recruitment of SMARCAD1 and EXO1 to DSBs. Meanwhile, CTCF depletion resulted in impaired early BRCA1 recruitment to laser strips but allowed late BRCA1 accumulation on laser strips at 15 min (Supplementary Fig. S8A and B). Based the times of recruitment and clearance of these HR factors at DSBs, we optimized the time points for observing their localization and colocalization. Assuming that late BRCA1 retains the ability to recruit SMARCAD1 and EXO1 similarly to early BRCA1, and considering that in control cells they are recruited ~4 to 5 min after BRCA1 is recruited to DSBs, we estimated that they would be recruited around 5 min after the appearance of late BRCA1 (15 min). Therefore, 5 min after microirradiation were selected as time points to evaluate the colocalization of early BRCA1 with SMARCAD1/EXO1, whereas 15, 20, and 25 min were selected to analyze that of late BRCA1 with SMARCAD1/EXO1.

First, we examined the recruitment of SMARCAD1 and EXO1 at various time points after micro-irradiation under conditions in which early BRCA1/BARD1 recruitment was impaired, whereas late BRCA1/BARD1 was preserved at DSBs. In cells depleted of CTCF or individual repressive chromatin-associated factors, including HP1γ, SUV39H1, SETDB1, KAP1, or KDM5A, the recruitment of SMARCAD1 and EXO1 to laser strips was impaired across all tested time points. This defect persisted even that the later points (15, 20, and 25 min) when late BRCA1 was robustly localized to the damage sites (Supplementary Fig. S8A–D). This result indicates that early BRCA1/BARD1 is essential for the recruitment of SMARCAD1 and EXO1 to DSBs, whereas late BRCA1/BARD1 is dispensable. Next, to disrupt late BRCA1 while preserving early BRCA1 at DSBs, we utilized the ATM inhibitor Ku-55933. ATM inhibition is known to disrupt both the γH2AX-dependent recruitment and retention of BRCA1 at DNA damage sites [11, 16, 63], both of which depend on ATM-mediated signaling. As expected, ATM inhibition resulted in severely reduced late BRCA1 at DSBs (15 min) (Supplementary Fig. S8E). Concurrently, in Ku-55933-treated cells, SMARCAD1 and EXO1 accumulated and were colocalized with early BRCA1 at 5 min, and this accumulation persisted up to 15 min even in the absence of late BRCA1 (Supplementary Fig. S8F), further supporting the role of early BRCA1 in recruiting SMARCAD1 and EXO1. Importantly, the accumulation of SMARCAD1 and EXO1/DNA2 at DSBs was completely abolished when both early and late BRCA1 were absent, which achieved by treating CTCF- or HP1γ-deficient cells with Ku-55933 (Supplementary Fig. S8G). This stood in sharp contrast to cells treated with Ku-55933 alone, where early BRCA1 remained present despite the absence of late BRCA1 (Supplementary Fig. S8F). These findings reinforce the notion that early BRCA1, rather than late BRCA1, is critical for recruiting SMARCAD1 and EXO1 to DNA damage sites. If this is the case, ATM inhibition alone should not affect DNA end resection. Consistently, ATM inhibition had no effect on end resection compared to control cells (Supplementary Fig. S8H), yet it markedly reduced HR efficiency (Supplementary Fig. S8I). Notably, concurrent ATM inhibition and CTCF or HP1γ depletion, a condition to abrogate both early and late BRCA1 recruitment, failed to further aggravate impaired end resection or reduced HR efficiency beyond those reduced by CTCF or HP1γ depletion alone (Supplementary Fig. S8H and I). These non-additive relationships indicate that late BRCA1/BARD1 is dispensable for DNA end resection per se, but instead participates in a post-resection setp of the HR process. Collectively, these results demonstrate that the recruitment of SMARCAD1 and EXO1/DNA2 to DSBs relies on CTCF-dependent early BRCA1/BARD1, rather than CTCF-independent late BRCA1/BARD1.

Given the established role of CTCF in HR, we posited that the CTCF downstream heterochromatin factors required for quick BRCA1/BARD1 recruitment at DSBs and accelerated extensive DNA end resection would also act as pro-HR factors. Thus, we examined whether deprivation of these factors reduces HR repair activity and found that HR activities were significantly reduced by deprivation of HP1γ, SUV39H1, SETDB1, KAP1, or KDM5A. This is consistent with results from the depletion of known pro-HR factors (CTCF, BRCA1, BARD1, DNA2/EXO1, or CtIP), indicating that all tested repressive chromatin-associated factors are indeed required for RAD51 recruitment (Fig. 7D) and vital for efficient HR (Supplementary Fig. S9A). Further, we re-evaluated whether the repressive chromatin factors acted downstream of the CTCF-operated HR pathway by investigating the repercussions of their individual depletion when combined with CTCF deficiency. Expectedly, depletion of HP1γ, SUV39H1, SETDB1, KAP1, or KDM5A had little effect on the HR activity of CTCF-deficient cells (Supplementary Fig. S9B). Similarly, depletion of other CTCF downstream factors—BRCA1/BARD1, EXO1/DNA2, and CtIP—had no impact on the HR efficiency in CTCF-depleted cells. These findings align with the idea that CTCF acts as an upstream pro-HR factor to control the BRCA1/BARD1–EXO1/DNA2 pathway by involving the heterochromatin factors in HR. Furthermore, the CTCF-directed HR pathway involves both CtIP-dependent short-range resection and the subsequent EXO1/DNA2-dependent long-range resection.

Next, we explored whether the downstream HR factors in the CTCF-governed cascade participate in NHEJ, given that CTCF itself remains uninvolved while BRCA1 antagonizes NHEJ. Consistent with previous findings [34, 64], the deprivation of BRCA1 enhanced NHEJ activity, while that of CTCF had little effect (Supplementary Fig. S9C). Additionally, NHEJ activities remained unchanged in cells lacking HP1γ, KAP1, KDM5A, or DNA2/EXO1. However, deprivation of SUV39H1 or SETDB1 suppressed NHEJ to a similar extent as that with the depletion of 53BP1, a critical NHEJ mediator, suggesting that they are required for robust NHEJ activity. Accordingly, SUV39H1 and SETDB1 appear to be associated with both NHEJ and HR, unlike CTCF and its other cascade factors, which are associated with HR but not NHEJ. As such, CTCF and its downstream factors participate in the DNA damage repair through various mechanisms and function differentially, although they share a role in rapid BRCA1/BARD1 localization at DSBs during HR. Collectively, our findings suggest a hierarchical recruitment cascade model for how CTCF quickly recruits BRCA1/BARD1 to DSBs. Whithin this framework, the CTCF-dependent and DNA damage-induced repressive chromatin at DSBs provides a platform for early BRCA1/BARD1 recruitment; subsequently, the newly localized BRCA1/BARD1 acts as a recruiter for SMARCAD1 and EXO1/DNA2 to facilitate extensive DNA end resection at DSBs and promote HR (Fig. 7E).

Discussion

The newly identified role of CTCF in HR, seemingly distinct from its classical roles in genome organization and expression that maintain genome fidelity, has recently gained recognition [31, 34]. CTCF has been shown to be localized at DSBs through its interaction with the MRE11–RAD50–NBS1 complex. This in turn, recruits the resection factor CtIP, which is required to initiate DNA end resection in HR, in conjunction with the MRE11 nuclease [34]. These reports extend our understanding of genome fidelity mediated by CTCF and also offer testable hypotheses for how CTCF directs its versatile activities. In this study, we found that CTCF engages multiple repressive chromatin-associated factors to rapidly recruit BRCA1/BARD1 at DSB sites, and this early BRCA1/BARD1 subsequently accelerates extensive resection for HR-directed repair.

CTCF prompts BRCA1/BARD1 localization at DNA damage sites

Substantial progress has been made in elucidating how CTCF plays a definite role in HR by promoting the initial step, in which DNA end resection regulates the choice between the HR and NHEJ pathways [34]. However, its functional interplay with the crucial HR mediator BRCA1/BARD1 has hitherto not been investigated.

Here, our discovery of the role of CTCF in the rapid recruitment of BRCA1/BARD1 to damaged chromatin highlights CTCF as a key player in HR repair pathway. Following DNA damage, CTCF prompted BRCA1/BARD1 accumulation at DSBs by involving repressive chromatin factors—HP1γ, SUV39H1/SETDB1, KDM5A, and KAP1—in an intricate, interdependent manner (Fig. 7E). Depletion of CTCF hindered the rapid recruitment of BRCA1/BARD1 by disturbing the localization of the repressive chromatin factors, SUV39H1/SETDB1-mediated H3K9 methylation, and KDM5A-mediated H3K4me3 demethylation at the damaged chromatin, but not vice versa. Moreover, deficiency of one of the factors that assemble at DSBs in a CTCF-dependent manner or the adopted inhibitors of SUV39H1/SETDB1 and KDM5A interrupted the localization of BRCA1/BARD1 and other repressive chromatin factors at DSBs (Figs 5F and 6G; Supplementary Figs S5F and S6G–I). These data indicate the epistatic function of CTCF in rapid BRCA1/BARD1 recruitment at damaged chromatin. They also demonstrate a functional scenario in which CTCF acts upstream in the HP1γ/SUV39H1/SETDB1/KDM5A/KAP-mediated transient repressive chromatinization around DSBs for rapid BRCA1/BARD1 recruitment, forming a defined CTCF–HP1γ/SUV39H1/SETDB1/KDM5A/KAP1–BRCA1/BARD1 axis (Fig. 7E). Furthermore, our results proposed two hypotheses: (i) CTCF interacts with these proteins prior to and in response to DNA damage, thereby contributing to the localization of the repressive chromatin factors and subsequent BRCA1/BARD1 at DSBs; (ii) the repressive chromatin proteins are interdependently recruited to damaged DNA sites and functionally interplay thereafter, such that HP1γ, which binds to SUV39H1/SETDB1-catalyzed H3K9me3 around the damaged chromatin, interacts with BARD1 and translocates BRCA1/BARD1 to DSBs. In support of the first hypothesis, our data suggest that CTCF interacts with the repressive chromatin factors as well as BARD1 in the absence of DNA damage, but these interactions are enhanced in the presence of DNA damage, and, notably, it can also interact with BRCA1 (Figs 2A, 3A, and 6F; Supplementary Fig. S5I). Additionally, earlier works have demonstrated that CTCF interacts with HP1 while KAP1 forms a complex with HP1/SUV39H1/SETDB1/KDM5A as a scaffold [54]. Further, it is worth noting that HP1γ is capable of recruiting BRCA1/BARD1 via its interaction with BARD1 and bonding to H3K9me3 at DSBs [21]. Accordingly, we speculated that in concert with SUV39H1/SETDB1/KDM5A-mediated establishment of H3K9me3 and H3K4me0 around the DSB chromatin, the multilateral interactions between CTCF-HP1γ, CTCF-BARD1, and HP1γ-BARD1 contribute to the CTCF-governed BRCA1/BARD1 localization at DSBs. The identification of CTCF as a substrate for PARP1-mediated PARylation [65] and ATM-mediated phosphorylation of KAP1 [26] suggests that they may act as a damage-regulatable protein to guide HP1γ/SUV39H1/SETDB1/KDM5A to the damaged chromatin and catalyze H3K9 methylation and H3K4 demethylation to promptly recruit BRCA1/BARD1. We therefore propose that CTCF localizes BRCA1/BARD1 to DSBs via its multivalent cooperative interactions with HP1γ/SUV39H1/SETDB1/KDM5A in addition to BARD1. Moreover, rapid BRCA1/BARD1 recruitment at DSBs requires methylated H3K9 and demethylated H3K4 at the chromatin surrounding DSBs, all of which is guided by CTCF.

Our analysis provides a mechanistic basis for CTCF-directed HP1γ-mediated BARD1 recruitment at DSBs via the tripartite interaction-mode (Figs 2–4). In this mechanism, the N-terminal of CTCF interacts with the BARD1 RING domain, while HP1γ is sandwiched between the ZF domain of CTCF and BRCT domain of BARD1, forming a triangular coordination network (Fig. 3E). To better understand the molecular mechanism underlying the CTCF–BARD1 interaction, we employed AlphaFold-based interaction prediction to explore potential amino acid residues in the N-terminal region of CTCF that are likely involved in BARD1 binding (data not shown). While this analysis does not definitively identify the specific amino acid residues involved in the interaction, future additional experiments using the predicted putative amino acid residues will allow us to determine the specific amino acid residues. Our findings further reinforce this model established by complementation studies in CTCF single- or CTCF/HP1γ co-depleted cells. Ectopic expression of fragments harboring the N-terminal or ZF region of CTCF fully recover early BARD1 recruitment at DSBs in CTCF-depleted cells (Fig. 2G and H). In contrast, the ZF region of CTCF failed to recover BARD1 recruitment at DSBs in CTCF/HP1γ co-depleted cells (Fig. 4C), despite rescue of BARD1 recruitment of as well as that of HP1γ by the ZF fragment of CTCF in cells lacking CTCF to a similar extent as FL CTCF (Figs 2G and 3H). These findings indicate that the ZF domain requires HP1γ for its BARD1 recruitment at DSBs and further underscore the importance of interactions (i) between the ZF region of CTCF and HP1γ and (ii) between HP1γ and the BRCT domain of BARD1. On the other hand, FL or fragments harboring N-terminal region of CTCF—capable to interact with RING domain of BARD1—partially rescued BARD1 recruitment at DSBs in CTCF/HP1γ co-depleted cells (Fig. 4C). Thus, these data uncover two ways to promptly recruit BARD1 at DSBs with CTCF, whereby (i) partially, CTCF facilitates early BARD1 localization at DSBs by its N-terminal alone without involving HP1γ, and (ii) largely with ZF region by using HP1γ as a bridge. Accordingly, the combined multivalent interactions of CTCF N-terminal and ZF region with BARD1, without and with HP1γ involvement respectively, contribute to prompt assembly of BRCA1/BARD1 at DSBs, promoting HR.

In this study, we demonstrated that CTCF regulates BRCA1/BARD1 recruitment in an upstream manner. By contrast, a previous study [32] concluded that CTCF functions downstream of BRCA1, as BRCA1 recruitment appeared unaffected by CTCF knockdown. This discrepancy can be attributed to the timing of analysis. Hilmi and colleagues assessed BRCA1 foci formation 3 h after NCS treatment, a late time point at which DNA damage is already resolved, whereas radiomimetic NCS is known to induce maximal DNA damage within 30–60 min [66]. To directly address this, we examined BRCA1 recruitment in CTCF-depleted U2OS cells at both 1 and 3 h after NCS treatment. At 1 h, BRCA1 recruitment was markedly reduced, whereas at 3 h, BRCA1 foci formation was comparable between control and CTCF-depleted cells. (Supplementary Fig. S1F) These data clearly indicate that CTCF is critical for the rapid recruitment of BRCA1 but not for its late recruitment. Taken together, our findings reconcile the apparent inconsistency between the two studies. While our data highlight the requirement of CTCF for early BRCA1 recruitment, Hilmi et al. reported the delayed BRCA1 recruitment at late time points. Therefore, the two studies should not be viewed as contradictory but rather as complementary, collectively delineating the temporal dynamics of BRCA1 recruitment in response to DNA damage.

CTCF governs rapid BRCA1/BARD1 recruitment pathway via its interaction with BARD1 in two ways with or without involving HP1γ, which is a mechanistically different manner from RNF168 where BARD1 binds to RNF168-mediated mono-ubiquitination of H2A at K13/15, thereby being located around DSB chromatin. In agreement with this, we found that CTCF knockdown has little effects on recruitment of RNF168 and its H2A ubiquitination at DSBs and vice versa—RNF168 knockdown does not affect the localization of CTCF at DSBs (Supplementary Fig. S3A). However, CTCF-knockdown cells phenocopy the compromised rapid BRCA1/BARD1 recruitment at DSBs in RNF168-knockdown cells (Fig. 1 and Supplementary Fig. S3E). We further demonstrate that combined ablation of CTCF and RNF168 results in an additively severe defects of BRCA1/BARD1 recruitment at DSBs (Supplementary Fig. S3F), indicating that these two pathways operated by CTCF or RNF168 act in parallel. Therefore, it may be that the common phenotypes in cells lacking one of CTCF and RNF168 and severe defects observed in double depleted cells result from disparate mechanisms. These findings that mechanisms or routes for localization of BRCA1/BARD1 at DSBs are likely multiple and multi-factorial. Characterizing the precise nature and timing of rapid BRCA1/BARD1 recruitment at DSBs directed by CTCF remains an important avenue for further investigation.

CTCF coordinates transient repressive histone marks and heterochromatin proteins to accelerate localization of BRCA1/BARD1 at DSBs

Our findings further establish that the SUV39H1/SETDB1-mediated H3K9me3 and KDM5A-mediated H3K4me0 are damage-inducible and are required for rapid BRCA1/BARD1 accumulation at DSBs generated by the FokI and AsiSI nucleases or microirradiation (Figs 5A–C and 6A–C). They regulate the DSB repair pathway directed by CTCF. Additionally, CTCF is required for the accumulation of interdependent heterochromatin proteins (HP1γ, KAP1, KDM5A, SUV39H1, and SETDB1) at DSBs, which supports rapid BRCA1/BARD1 recruitment at DSBs (Figs 3F and G, 5D, and 6D; Supplementary Figs S5M and S6E). We also demonstrated that the heterochromatin proteins localize at DSBs through an intricate multilateral interdependent mechanism. The depletion of a single protein abrogates the localization of all the other proteins at the DSBs.

CTCF epistatically directs the recruitment of HP1γ, KAP1, SUV39H1/SETDB1, and KDM5A at sites of DNA damage, which helps to assemble H3K9me2/3, H3K4me0, and subsequently BRCA1/BARD1 at the DSB sites (Fig. 7E). Several studies provide evidence that repressive histone marks, including H3K9me2/3 [24], and proteins related to heterochromatinization (HP1, SUV39H1, KAP1 and KDM5A) [24–26, 57, 58], accumulate at DSBs, consistent with the CTCF-orchestrated accumulation of repressive histone marks (H3K9me2/3 and H3K4me0) and heterochromatin proteins (HP1γ, SUV39H1, SETDB1, KDM5A, and KAP1) (Figs 3F and G, 5A–D, and 6A–D; Supplementary Figs S5A and S6E). By contrast, a genome-wide ChIP-sequencing study reports no changes in H3K9me2/3 [46]. This discrepancy suggests that some of the events related to repressive chromatin-associated proteins and histone marks may take place within a specific context of DSBs or are too transient to be detected by ChIP-sequencing.

Further, our data imply that pre-existing levels of H3K9me3 and H3K4me0 before DNA damage, which are inversely correlated with transcriptional activities, may have little effect on the CTCF-directed rapid BRCA1/BARD1 recruitment at DSBs. This is because CTCF can induce H3K9 methylation and H3K4me3 demethylation around both intergenic and intragenic DSB chromatin regions in a similar fashion (Fig. 5C). This is seemingly contradictory to previous evidence that DSBs occurring within or near transcriptionally active genes are preferentially repaired by HR [47–49], possibly by recruiting HR factors such as BRCA1/BARD1. In contrast, other studies provide evidence that H3K9me3 recruits HP1, which is capable of associating with BRCA1/BARD1, to promote HR [21, 67]. Consistent with our finding that the CTCF-directed methylation of H3K9 and demethylation of H3K4me3 around DSBs favors HR (Fig. 7E and Supplementary Fig. S9A), the critical NHEJ factor RIF1 favors NHEJ by associating with H3K4me3 for end protection [20]. H3K9 methylation and H3K4 demethylation at DSBs may occur independent of the preexisting chromatin state to help recruit BRCA1/BARD1. This suggests that although the DNA damage-induced actual levels of H3K9me2/3 and H3K4me0 at DSBs occurring in transcriptionally active regions may be lower than those in inactive regions, they may be sufficient for BRCA1/BARD1 recruitment and local transcriptional silencing. These modifications may either act as a signal to direct HR or they may work in combination with other diverse histone modifications at DSBs.

A CTCF-driven cascade to accelerate DNA end resection at DSBs for homologous recombination repair

High-fidelity DSB repair depends on DNA end resection—a critical process that both enables HR and governs repair pathway selection. End resection is initiated by MRE11-catalyzed nucleolytic digestion of the DSB ends in cooperation with CtIP. Further long-range extensive resection mediated by EXO1/DNA2 generates 3′ single-strand overhang at DSB ends for subsequent strand invasion into a homologous DNA. CTCF facilitates initial DNA end resection at DSBs by recruiting CtIP, the cofactor crucial for MRE11 nuclease activity [34]. Herein, we assign an additional role to CTCF in long-range resection via rapid assembly of BRCA1/BARD1 at DSBs to engage EXO1/DNA2 with DSB ends, thereby expediting wire-to-wire resection of DSB ends during HR (Fig. 7E). Importantly, we have shown that the assembly of heterochromatin proteins and histone marks by virtue of CTCF is vital for quick BRCA1/BARD1 accumulation at DSBs, thereby facilitating EXO1/DNA2-mediated extensive resection and subsequent HR-dependent repair of DSBs. In line with these findings, we propose a model for HR, which requires CTCF-driven CtIP recruitment for the initial resection of DSB ends. This model necessitates the subsequent CTCF-accelerated BRCA1/BARD1 localization for the EXO1/DNA2-mediated expansion of resected single-strand DNA. This expansion is achieved through the concerted actions of the CTCF-directed heterochromatin factors and repressive histone marks assembled at the DSBs (Fig. 7E). Notably, to override NHEJ and promote HR, CTCF completes the resection of DSB ends critical for HR. This process is initiated by recruiting CtIP and expanded by the rapid recruitment of BRCA1/BARD1, a cascade orchestrated by CTCF to promote HR for repairing DSB damage. CTCF quickly engages BRCA1/BARD1 in the initial step of HR, thereby stimulating extensive end resection through the recruitment of the nucleases CtIP and EXO1/DNA2 at DSBs. Furthermore, CTCF epistatically recruits repressive chromatin proteins and histone marks for rapid BRCA1/BARD1 recruitment, which provides an elegant mechanism to ensure HR in response to DSBs.

CTCF as well as the CTCF-governed heterochromatin proteins and repressive histone modifications are critical for the rapid accumulation of BRCA1/BARD1 and subsequent recruitment of EXO1/DNA2 at DSBs. This allows extensive end resection at DSBs and HR. These findings reinforce a hierarchical cascade model, in which CTCF recruits HP1γ, KAP1, SUV39H1/SETDB1, and KDM5A to quickly localize BRCA1/BARD1 at DSBs, promoting EXO1/DNA2 recruitment to DSBs for extensive resection (Fig. 7E). In line with this model, CTCF and all its downstream factors share a role in early BRCA1/BARD1 recruitment at DSBs and HR-mediated DNA repair. Notably, all the heterochromatin-associated proteins as well as the two histone marks directed by CTCF are essential HR factors. A lack of any single factor leads to disturbed early BRCA1/BARD1 recruitment, impaired end resection, and compromised HR as sequentially causal consequences. Moreover, given that the essential repressive chromatin proteins are recruited at DSBs in an interdependent manner, it is conceivable that they are localized at DSBs as one assembly in the form of a single complex or set. Based on our results that (i) CTCF physically interacts with its downstream heterochromatin factors prior to DNA damage (Fig. 6F and Supplementary Fig. S5I) and (ii) it is required for their assembly at DSBs upon DNA damage (Figs 5D and 6C and D; Supplementary Fig. S6E), CTCF may act as a scaffold for these factors. Thus, it sets a precedent for the essential factors to participate in rapid BRCA1/BARD1 recruitment at DSBs during HR.

Our CTCF-driven cascade model implies that the repressive chromatin-related proteins and histone modifications in HR-mediated DNA repair, far from being a physical chromatin barrier, do a great deal of assistance for the access and localization of BRCA1/BARD1 onto DSBs. Here we show that blocking repressive chromatin establishment at DSBs via knockdown of HP1γ, SUV39H1/SETDB1, KAP1, or KDM5A, or inhibiting H3K9 methylation or H3K4 demethylation hinders rapid BRCA1/BARD1 recruitment to DSBs and the subsequent localization of SMARCAD1 and EXO1/DNA2. Thus, the downstream CTCF factors are localized at DSBs in a hierarchical relationship, further supporting the elaborate HR cascade model (Fig. 7E). Notably, in our model, the repressive chromatin-associated factors and histone markers are localized on DSBs at exactly the time when they are needed and only during the necessary period, after which they are released. They are transiently accumulated at DSBs and promote the rapid recruitment of BRCA1/BARD1, but they have little effect on its late recruitment. Therefore, the proposed CTCF-directed HR cascade pathway entails a chain of timely assembly/disassembly of various pro-HR factors and chromatin modifications and coordinates multistep HR procedures to restore the original DNA sequences of DSBs. The transient assembly of heterochromatin-linked proteins and histone modifications at DSBs during early HR may serve multiple regulatory functions. These include: (i) fine-tuning end resection to efficiently load RAD51 onto the resected single-strand DNA coated with RPA; (ii) blocking excessive or persistent resection; and (iii) pausing/interfering with transcription when DSBs occur near transcribed regions.

Deprivation of CTCF or BRCA1 results in similar severe defects in DNA end resection and HR (Fig. 7B and Supplementary Fig. S9A). This shared role in DNA end resection and HR between CTCF and BRCA1 might be attributed to early BRCA1/BARD1 at DSBs, given that deprivation of CTCF impairs recruitment of early but not late BRCA1/BARD1 at DSBs (Fig.1 and Supplementary Fig. S1F). In particular, early BRCA1/BARD1 was found to promote extensive DNA end resection through the recruitment of SMARCAD1 and EXO1/DNA2, whereas late BRCA1/BARD1 was dispensable for the recruitment of them and DNA end resection itself (Fig. 7B and Supplementary Fig. S8H). However, these findings raise an important question regarding the role of CTCF-independent late BRCA1/BARD1 in HR, given that absence of late BRCA1/BARD1 via ATM inhibition had no effect on DNA end resection but notably impaired HR (Supplementary Fig. S8H and I). Considering that HR is a multi-step sequential process, the defective phenotype observed under BRCA1 deficiency suggest that late BRCA1 may contribute to a post-resection step. In addition to regulating DNA end resection, BRCA1 is known to function in a post-resection step of HR; specifically, it facilitates RAD51 loading onto resected ssDNA by linking BRCA2–RAD51 to DSBs, using PALB2 as a bridge to form BRCA1–PALB2–BRCA2–RAD51 complex [68, 69]. Therefore, we hypothesized that late BRCA1/BARD1 promotes BRCA2 retention at DSBs, thereby supporting subsequent RAD51 loading. To test this hypothesis, we analyzed BRCA2 localization at DSBs under conditions where only late BRCA1/BARD1 is present, where only early BRCA1/BARD1 was present, or where neither was present, following an experimental strategy similar to our previous analyses of SMARCAD1 and EXO1 recruitment. In line with previous findings demonstrating that CTCF promotes BRCA2 recruitment to DSBs [32], BRCA2 localization at DSBs was markedly impaired in CTCF-deficient cells (Supplementary Fig. S10A). In contrast to CTCF-deficient cells, BRCA2 accumulated abundantly at DSBs in cells depleted of HP1γ, SUV39H1, SETDB1, KAP1, or KDM5A, despite the absence of early BRCA1/BARD1 at DSBs (Supplementary Fig. S10A). This suggests that while BRCA2 recruitment to DSBs depends on CTCF, repressive chromatin factors-mediated early BRCA1/BARD1 is dispensable for both the initial recruitment and stable residence of BRCA2 at DSBs. Conversely, when late BRCA1/BARD1 was reduced by ATM inhibition, BRCA2 was initially localized to DSBs at 5 min but failed to persist thereafter (Supplementary Fig. S10B). This indicates that late BRCA1/BARD1 is required for the stabilization and retention of BRCA2 at DSBs. Similarly, in BRCA1- or PALB2-deficient cells, BRCA2 initially accumulated at DSBs but failed to maintain its localization, supporting the model where PALB2 serves as a molecular bridge linking BRCA1 and BRCA2 at DSBs. To further corroborate a role of late BRCA1/BARD1 in BRCA2 retention at DSBs, we treated CTCF- or HP1γ-deficient cells with Ku-55933 and monitored BRCA2 localization at DSBs. In CTCF-depleted cells treated with Ku-55933, BRCA2 accumulation at DSBs was completely abolished (Supplementary Fig. S10C), which was consistent with the defect observed in CTCF-deficient cells without Ku-55933 (Supplementary Fig. S10B). These results reinforce the idea that CTCF is critical for initial recruitment of BRCA2 to DSBs, regardless of the presence of late BRCA1/BARD1 at DSBs. Furthermore, unlike HP1γ-depleted cells, which lacked early BRCA1/BARD1 but retained abundant late BRCA1/BARD1, HP1γ depletion in combination with ATM inhibition (lacking both early and late BRCA1/BARD1) allowed initial BRCA2 accumulation at 5 min but markedly reduced its residence at DSBs over time (Supplementary Fig. S10C). Collectively, these findings support the conclusion that late BRCA1/BARD1 contributes to stabilization and retention of BRCA2 at DSBs.

Based on the CTCF-dependent but mutually independent recruitment of CtIP, BRCA1, and BRCA2 to DSBs (Supplementary Figs S7D and S10A, D, and E), CTCF is thought to facilitate recruitment of CtIP, repressive chromatin factor-mediated early BRCA1/BARD1, and BRCA2 to DSBs through distinct parallel routes, rather than a single epistatic linear pathway. However, from a functional perspective, these factors are sequentially linked within the HR pathway, where CtIP participates in initial resection, early BRCA1/BARD1 promotes SMARCAD1- and EXO1/DNA2-dependent extensive resection, and BRCA2 mediates RAD51 loading. Consistent with this scenario, double or triple co-depletion of CtIP, HP1γ, BRCA1, or BRCA2 did not further aggravate DNA end resection or HR efficiency compared with the single depletion of each individual factor (Supplementary Fig. S10F and G). Taken together, these findings support a model in which CTCF coordinates the recruitment of CtIP, early BRCA1/BARD1, and BRCA2 to DSBs through multiple parallel routes, enabling these factors to sequentially participate in a single HR pathway, spanning from CtIP-mediated initial resection to early BRCA1/BARD1-dependent extensive resection, and subsequent BRCA2-mediated RAD51 loading (Supplementary Fig. S10H).

Collectively, these studies reveal a decisive role for CTCF in the midst of HR, whereby CTCF operates an essential and functional HR cascade pathway to facilitate long-range resection via rapid BRCA1/BARD1 recruitment at DSBs. We demonstrate that CTCF employs heterochromatin proteins (HP1γ, KAP1, SUV39H1/SETDB1, and KDM5A) and repressive histone marks (H3K9me1/2/3 and H3K4me0) for prompt BRCA1/BARD1 localization at DSBs. This is followed by extensive resection by downstream factors SMARCAD1 and EXO1/DNA2, thereby facilitating HR-mediated repair of DSBs. Such knowledge will help understand the molecular mechanisms underlying the diverse functions of CTCF in ensuring genome fidelity. Also, understanding the functional crosstalks and collaborations between CTCF and its downstream HR cascade factors will illuminate the molecular basis of various human diseases that arise from dysfunctional repair and range from developmental disorders to cancer, thereby warranting further studies.

Supplementary Material

gkag715_Supplemental_File

gkag715_supplemental_file.pdf^{(4.5MB, pdf)}

Acknowledgements

We are greatful to Dr Roger A. Greenberg for the ER-mCherry-LacI-FokI-DD-U2OS reporter (FokI-U2OS), Dr Gaëlle Legube for the AsiSI-ER-U2OS reporter, and Dr Jeremy M. Stark for the GFP-reporter cell lines used for the HR and NHEJ repair assay. We thank members of the Lee laboratory for comments on the manuscript.

Author contributions: Soon Young Hwang (Data curation [equal], Formal analysis [equal], Investigation [equal], Methodology [equal], Validation [equal], Writing – original draft [lead]), Daeki Choi (Data curation [equal], Formal analysis [equal], Investigation [equal], Methodology [equal], Validation [equal], Writing – original draft [supporting]), Myoung Joo Choi (Data curation [supporting], Formal analysis [supporting], Investigation [supporting], Methodology [supporting], Validation [supporting]), Hye-Ram Jo (Data curation [supporting], Investigation [supporting]), Jae-Hoon Jeong (Funding acquisition [supporting], Investigation [supporting], Writing – review & editing [supporting]), Mi Ae Kang (Funding acquisition [equal], Investigation [equal], Validation [equal], Writing – review & editing [supporting]), and Jong-Soo Lee (Conceptualization [lead], Funding acquisition [equal], Investigation [equal], Supervision [lead], Validation [equal], Writing – review & editing [lead])

Contributor Information

Soon Young Hwang, Department of Life Sciences, College of Natural Sciences, Ajou University, 16499,Suwon, Republic of Korea; Research Institute of Basic Sciences, Ajou University, 16499 Suwon, Republic of Korea.

Daeki Choi, Department of Life Sciences, College of Natural Sciences, Ajou University, 16499,Suwon, Republic of Korea; Research Institute of Basic Sciences, Ajou University, 16499 Suwon, Republic of Korea.

Myoung Joo Choi, Department of Life Sciences, College of Natural Sciences, Ajou University, 16499,Suwon, Republic of Korea.

Hye-Ram Jo, Division of Radiation Biomedical Research, Korea Institute of Radiological and Medical Sciences, 01812 Seoul, Republic of Korea.

Jae-Hoon Jeong, Division of Radiation Biomedical Research, Korea Institute of Radiological and Medical Sciences, 01812 Seoul, Republic of Korea.

Mi Ae Kang, Department of Life Sciences, College of Natural Sciences, Ajou University, 16499,Suwon, Republic of Korea; Research Institute of Basic Sciences, Ajou University, 16499 Suwon, Republic of Korea.

Jong-Soo Lee, Department of Life Sciences, College of Natural Sciences, Ajou University, 16499,Suwon, Republic of Korea; Research Institute of Basic Sciences, Ajou University, 16499 Suwon, Republic of Korea.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (RS-2020-NR048960, RS-2021-NR060141, RS-2022-NR075491) and a grant of the Korea Institute of Radiological and Medical Sciences (KIRAMS), funded by the Ministry of Science and ICT (MSIT), Republic of Korea (No. 50531-2025). Funding to pay the Open Access publication charges for this article was provided by the Korea Institute of Radiological and Medical Sciences (KIRAMS), funded by the Ministry of Science and ICT (MSIT), Republic of Korea (No.50531-2025).

Data availability

The data underlying this article are available in the article and in its online supplementary material.

References

1. Welcsh PL, King MC. BRCA1 and BRCA2 and the genetics of breast and ovarian cancer. Hum Mol Genet. 2001;10:705–13. 10.1093/hmg/10.7.705 [DOI] [PubMed] [Google Scholar]
2. Burgess M, Puhalla S. BRCA 1/2-mutation-related and sporadic breast and ovarian cancers: more alike than different. Front Oncol. 2014;4:19. 10.3389/fonc.2014.00019 [DOI] [PMC free article] [PubMed] [Google Scholar]
3. Thai TH, Du F, Tsan JT et al. Mutations in the BRCA1-associated RING domain (BARD1) gene in primary breast, ovarian and uterine cancers. Hum Mol Genet. 1998;7:195–202. 10.1093/hmg/7.2.195 [DOI] [PubMed] [Google Scholar]
4. Suszynska M, Kozlowski P. Summary of BARD1 mutations and precise estimation of breast and ovarian cancer risks associated with the mutations. Genes. 2020;11:798. 10.3390/genes11070798 [DOI] [PMC free article] [PubMed] [Google Scholar]
5. Hawsawi YM, Shams A, Theyab A et al. BARD1 mystery: tumor suppressors are cancer susceptibility genes. BMC Cancer. 2022;22:599. 10.1186/s12885-022-09567-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
6. Cruz-García A, López-Saavedra A, Huertas P. BRCA1 accelerates CtIP-mediated DNA-end resection. Cell Rep. 2014;9:451–9. 10.1016/j.celrep.2014.08.076 [DOI] [PubMed] [Google Scholar]
7. Zhao W, Steinfeld JB, Liang F et al. BRCA1–BARD1 promotes RAD51-mediated homologous DNA pairing. Nature. 2017;550:360–5. 10.1038/nature24060 [DOI] [PMC free article] [PubMed] [Google Scholar]
8. Tarsounas M, Sung P. The antitumorigenic roles of BRCA1-BARD1 in DNA repair and replication. Nat Rev Mol Cell Biol. 2020;21:284–99. 10.1038/s41580-020-0218-z [DOI] [PMC free article] [PubMed] [Google Scholar]
9. Wang M, Li W, Tomimatsu N et al. Crucial roles of the BRCA1–BARD1 E3 ubiquitin ligase activity in homology-directed DNA repair. Mol Cell. 2023;83:3679–91.e8. 10.1016/j.molcel.2023.09.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
10. Glover JN, Williams RS, Lee MS. Interactions between BRCT repeats and phosphoproteins: tangled up in two. Trends Biochem Sci. 2004;29:579–85. 10.1016/j.tibs.2004.09.010 [DOI] [PubMed] [Google Scholar]
11. Her J, Soo Lee N, Kim Y et al. Factors forming the BRCA1-A complex orchestrate BRCA1 recruitment to the sites of DNA damage. Acta Biochim Biophys Sin (Shanghai). 2016;48:658–64. 10.1093/abbs/gmw047 [DOI] [PubMed] [Google Scholar]
12. Lee BL, Singh A, Mark Glover JN et al. Molecular basis for K63-linked ubiquitination processes in double-strand DNA break repair: a focus on kinetics and dynamics. J Mol Biol. 2017;429:3409–29. 10.1016/j.jmb.2017.05.029 [DOI] [PMC free article] [PubMed] [Google Scholar]
13. Krais JJ, Wang Y, Patel P et al. RNF168-mediated localization of BARD1 recruits the BRCA1–PALB2 complex to DNA damage. Nat Commun. 2021;12:5016. 10.1038/s41467-021-25346-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
14. Coleman KA, Greenberg RA. The BRCA1–RAP80 complex regulates DNA repair mechanism utilization by restricting end resection. J Biol Chem. 2011;286:13669–80. 10.1074/jbc.M110.213728 [DOI] [PMC free article] [PubMed] [Google Scholar]
15. Hu Y, Scully R, Sobhian B et al. RAP80-directed tuning of BRCA1 homologous recombination function at ionizing radiation-induced nuclear foci. Genes Dev. 2011;25:685–700. 10.1101/gad.2011011 [DOI] [PMC free article] [PubMed] [Google Scholar]
16. Li M, Yu X. Function of BRCA1 in the DNA damage response is mediated by ADP-ribosylation. Cancer Cell. 2013;23:693–704. 10.1016/j.ccr.2013.03.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
17. Becker JR, Clifford G, Bonnet C et al. BARD1 reads H2A lysine 15 ubiquitination to direct homologous recombination. Nature. 2021;596:433–7. 10.1038/s41586-021-03776-w [DOI] [PMC free article] [PubMed] [Google Scholar]
18. Fradet-Turcotte A, Canny MD, Escribano-Díaz C et al. 53BP1 is a reader of the DNA-damage-induced H2A Lys 15 ubiquitin mark. Nature. 2013;499:50–4. 10.1038/nature12318 [DOI] [PMC free article] [PubMed] [Google Scholar]
19. Nakamura K, Saredi G, Becker JR et al. H4K20me0 recognition by BRCA1–BARD1 directs homologous recombination to sister chromatids. Nat Cell Biol. 2019;21:311–8. 10.1038/s41556-019-0282-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
20. Bayley R, Borel V, Moss RJ et al. H3K4 methylation by SETD1A/BOD1L facilitates RIF1-dependent NHEJ. Mol Cell. 2022;82:1924–39.e10. 10.1016/j.molcel.2022.03.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
21. Wu W, Nishikawa H, Fukuda T et al. Interaction of BARD1 and HP1 is required for BRCA1 retention at sites of DNA damage. Cancer Res. 2015;75:1311–21. 10.1158/0008-5472.CAN-14-2796 [DOI] [PMC free article] [PubMed] [Google Scholar]
22. Bannister AJ, Zegerman P, Partridge JF et al. Selective recognition of methylated lysine 9 on histone H3 by the HP1 chromo domain. Nature. 2001;410:120–4. 10.1038/35065138 [DOI] [PubMed] [Google Scholar]
23. Schultz DC, Ayyanathan K, Negorev D et al. SETDB1: a novel KAP-1-associated histone H3 lysine 9-specific methyltransferase that contributes to HP1-mediated silencing of euchromatic genes by KRAB zinc-finger proteins. Genes Dev. 2002;16:919–32. 10.1101/gad.973302 [DOI] [PMC free article] [PubMed] [Google Scholar]
24. Ayrapetov MK, Gursoy-Yuzugullu O, Xu C et al. DNA double-strand breaks promote methylation of histone H3 on lysine 9 and transient formation of repressive chromatin. Proc Natl Acad Sci USA. 2014;111:9169–74. 10.1073/pnas.1403565111 [DOI] [PMC free article] [PubMed] [Google Scholar]
25. Alagoz M, Katsuki Y, Ogiwara H et al. SETDB1, HP1 and SUV39 promote repositioning of 53BP1 to extend resection during homologous recombination in G2 cells. Nucleic Acids Res. 2015;43:7931–44. 10.1093/nar/gkv722 [DOI] [PMC free article] [PubMed] [Google Scholar]
26. White D, Rafalska-Metcalf IU, Ivanov AV et al. The ATM substrate KAP1 controls DNA repair in heterochromatin: regulation by HP1 proteins and serine 473/824 phosphorylation. Mol Cancer Res. 2012;10:401–14. 10.1158/1541-7786.MCR-11-0134 [DOI] [PMC free article] [PubMed] [Google Scholar]
27. Campbell S, Ismail IH, Young LC et al. Polycomb repressive complex 2 contributes to DNA double-strand break repair. Cell Cycle. 2013;12:2675–83. 10.4161/cc.25795 [DOI] [PMC free article] [PubMed] [Google Scholar]
28. Fitieh A, Locke AJ, Mashayekhi F et al. BMI-1 regulates DNA end resection and homologous recombination repair. Cell Rep. 2022;38:110536. 10.1016/j.celrep.2022.110536 [DOI] [PubMed] [Google Scholar]
29. Kumbhar R, Sanchez A, Perren J et al. Poly(ADP-ribose) binding and macroH2A mediate recruitment and functions of KDM5A at DNA lesions. J Cell Biol. 2021;220:e202006149. 10.1083/jcb.202006149 [DOI] [PMC free article] [PubMed] [Google Scholar]
30. Kendek A, Wensveen MR, Janssen A. The sound of silence: how silenced chromatin orchestrates the repair of double-strand breaks. Genes. 2021;12:1415. 10.3390/genes12091415 [DOI] [PMC free article] [PubMed] [Google Scholar]
31. Kang MA, Lee JS. A newly assigned role of CTCF in cellular response to broken DNAs. Biomolecules. 2021;11:363. 10.3390/biom11030363 [DOI] [PMC free article] [PubMed] [Google Scholar]
32. Hilmi K, Jangal M, Marques M et al. CTCF facilitates DNA double-strand break repair by enhancing homologous recombination repair. Sci Adv. 2017;3:e1601898. 10.1126/sciadv.1601898 [DOI] [PMC free article] [PubMed] [Google Scholar]
33. Lang F, Li X, Zheng W et al. CTCF prevents genomic instability by promoting homologous recombination-directed DNA double-strand break repair. Proc Natl Acad Sci USA. 2017;114:10912–7. 10.1073/pnas.1704076114 [DOI] [PMC free article] [PubMed] [Google Scholar]
34. Hwang SY, Kang MA, Baik CJ et al. CTCF cooperates with CtIP to drive homologous recombination repair of double-strand breaks. Nucleic Acids Res. 2019;47:9160–79. 10.1093/nar/gkz639 [DOI] [PMC free article] [PubMed] [Google Scholar]
35. Densham RM, Garvin AJ, Stone HR et al. Human BRCA1–BARD1 ubiquitin ligase activity counteracts chromatin barriers to DNA resection. Nat Struct Mol Biol. 2016;23:647–55. 10.1038/nsmb.3236 [DOI] [PMC free article] [PubMed] [Google Scholar]
36. Densham RM, Morris JR. The BRCA1 ubiquitin ligase function sets a new trend for remodelling in DNA repair. Nucleus. 2017;8:116–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
37. Hu Q, Botuyan MV, Zhao D et al. Mechanisms of BRCA1–BARD1 nucleosome recognition and ubiquitylation. Nature. 2021;596:438–43. 10.1038/s41586-021-03716-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
38. Iacovoni JS, Caron P, Lassadi I et al. High-resolution profiling of gammaH2AX around DNA double strand breaks in the mammalian genome. EMBO J. 2010;29:1446–57. 10.1038/emboj.2010.38 [DOI] [PMC free article] [PubMed] [Google Scholar]
39. Tang J, Cho NW, Cui G et al. Acetylation limits 53BP1 association with damaged chromatin to promote homologous recombination. Nat Struct Mol Biol. 2013;20:317–25. 10.1038/nsmb.2499 [DOI] [PMC free article] [PubMed] [Google Scholar]
40. Gunn A, Stark JM. I-SceI-based assays to examine distinct repair outcomes of mammalian chromosomal double strand breaks. Methods Mol Biol. 2012;920:379–91. [DOI] [PubMed] [Google Scholar]
41. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods. 2001;25:402–8. 10.1006/meth.2001.1262 [DOI] [PubMed] [Google Scholar]
42. Nicolette ML, Lee K, Guo Z et al. MRE11-Rad50-Xrs2 and Sae2 promote 5 strand resection of DNA double-strand breaks. Nat Struct Mol Biol. 2010;17:1478–85. 10.1038/nsmb.1957 [DOI] [PMC free article] [PubMed] [Google Scholar]
43. Zhou Y, Caron P, Leqube G et al. Quantitation of DNA double-strand break resection intermediates in human cells. Nucleic Acids Res. 2014;42:e19. 10.1093/nar/gkt1309 [DOI] [PMC free article] [PubMed] [Google Scholar]
44. Zierhut C, Diffley JF. Break dosage, cell cycle stage and DNA replication influence DNA double-strand break response. EMBO J. 2008;27:1875–85. 10.1038/emboj.2008.111 [DOI] [PMC free article] [PubMed] [Google Scholar]
45. Ikura M, Furuya K, Fukuto A et al. Coordinated regulation of TIP60 and poly(ADP-Ribose) polymerase 1 in damaged-chromatin dynamics. Mol Cell Biol. 2016;36:1595–607. 10.1128/MCB.01085-15 [DOI] [PMC free article] [PubMed] [Google Scholar]
46. Clouaire T, Rocher V, Lashgari A et al. Comprehensive mapping of histone modifications at DNA double-strand breaks deciphers repair pathway chromatin signatures. Mol Cell. 2018;72:250–62.e6. 10.1016/j.molcel.2018.08.020e6 [DOI] [PMC free article] [PubMed] [Google Scholar]
47. Aymard F, Bugler B, Schmidt CK et al. Transcriptionally active chromatin recruits homologous recombination at DNA double-strand breaks. Nat Struct Mol Biol. 2014;21:366–74. 10.1038/nsmb.2796 [DOI] [PMC free article] [PubMed] [Google Scholar]
48. Yasuhara T, Kato R, Hagiwara Y et al. Human Rad52 promotes XPG-mediated R-loop processing to initiate transcription-associated homologous recombination repair. Cell. 2018;175:558–70.e11. 10.1016/j.cell.2018.08.056 [DOI] [PubMed] [Google Scholar]
49. Ouyang J, Yadav T, Zhang JM et al. RNA transcripts stimulate homologous recombination by forming DR-loops. Nature. 2021;594:283–8. 10.1038/s41586-021-03538-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
50. Marnef A, Cohen S, Legube G. Transcription-coupled DNA double-strand break repair: active genes need special care. J Mol Biol. 2017;429:1277–88. 10.1016/j.jmb.2017.03.024 [DOI] [PubMed] [Google Scholar]
51. Savage KI, Harkin DP. BRCA1, a ‘complex’ protein involved in the maintenance of genomic stability. FEBS J. 2015;282:630–46. 10.1111/febs.13150 [DOI] [PubMed] [Google Scholar]
52. Hashizume R, Fukuda M, Maeda I et al. The RING heterodimer BRCA1-BARD1 is a ubiquitin ligase inactivated by a breast cancer-derived mutation. J Biol Chem. 2001;276:14537–40. 10.1074/jbc.C000881200 [DOI] [PubMed] [Google Scholar]
53. Burdett H, Foglizzo M, Musgrove LJ et al. BRCA1-BARD1 combines multiple chromatin recognition modules to bridge nascent nucleosomes. Nucleic Acids Res. 2023;51:11080–103. 10.1093/nar/gkad793 [DOI] [PMC free article] [PubMed] [Google Scholar]
54. Groner AC, Meylan S, Ciuffi A et al. KRAB-zinc finger proteins and KAP1 can mediate long-range transcriptional repression through heterochromatin spreading. PLoS Genet. 2010;6:e1000869. 10.1371/journal.pgen.1000869 [DOI] [PMC free article] [PubMed] [Google Scholar]
55. Aitken SJ, Ibarra-Soria X, Kentepozidou E et al. CTCF maintains regulatory homeostasis of cancer pathways. Genome Biol. 2018;19:106. 10.1186/s13059-018-1484-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
56. Randolph K, Hyder U, D’Orso I. KAP1/TRIM28: transcriptional activator and/or repressor of viral and cellular programs?. Front Cell Infect Microbiol. 2022;12:834636. 10.3389/fcimb.2022.834636 [DOI] [PMC free article] [PubMed] [Google Scholar]
57. Gong F, Clouaire T, Aguirrebengoa M et al. Histone demethylase KDM5A regulates the ZMYND8-NuRD chromatin remodeler to promote DNA repair. J Cell Biol. 2017;216:1959–74. 10.1083/jcb.201611135 [DOI] [PMC free article] [PubMed] [Google Scholar]
58. Kumbhar R, Sanchez A, Perren J et al. Poly(ADP-ribose) binding and macroH2A mediate recruitment and functions of KDM5A at DNA lesions. J Cell Biol. 2021;220:e202006149. 10.1083/jcb.202006149 [DOI] [PMC free article] [PubMed] [Google Scholar]
59. Feng S, Ma S, Li K et al. RIF1-ASF1-mediated high-order chromatin structure safeguards genome integrity. Nat Commun. 2022;13:957. 10.1038/s41467-022-28588-y [DOI] [PMC free article] [PubMed] [Google Scholar]
60. Lachner M, O’Carroll D, Rea S et al. Methylation of histone H3 lysine 9 creates a binding site for HP1 proteins. Nature. 2001;410:116–20. 10.1038/35065132 [DOI] [PubMed] [Google Scholar]
61. Rassoolzadeh H, Coucoravas C, Farnebo M. The proximity ligation assay reveals that at DNA double-strand breaks WRAP53β associates with γH2AX and controls interactions between RNF8 and MDC1. Nucleus. 2015;6:417–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
62. Densham RM, Garvin AJ, Stone HR et al. Human BRCA1–BARD1 ubiquitin ligase activity counteracts chromatin barriers to DNA resection. Nat Struct Mol Biol. 2016;23:647–55. 10.1038/nsmb.3236 [DOI] [PMC free article] [PubMed] [Google Scholar]
63. Ciccia A, Elledge SJ. The DNA damage response: making it safe to play with knives. Mol Cell. 2010;40:179–204. 10.1016/j.molcel.2010.09.019 [DOI] [PMC free article] [PubMed] [Google Scholar]
64. Swift ML, Beishline K, Flashner S et al. DSB repair pathway choice is regulated by recruitment of 53BP1 through cell cycle-dependent regulation of Sp1. Cell Rep. 2021;34:108840. 10.1016/j.celrep.2021.108840 [DOI] [PubMed] [Google Scholar]
65. Yu W, Ginjala V, Pant V et al. Poly(ADP-ribosyl)ation regulates CTCF-dependent chromatin insulation. Nat Genet. 2004;36:1105–10. 10.1038/ng1426 [DOI] [PubMed] [Google Scholar]
66. Lee J-H, Cheong H-M, Kang M-Y et al. Ser1778 of 53BP1 plays a role in DNA double-strand break repairs. Korean J Physiol Pharmacol. 2009;13:343–8. 10.4196/kjpp.2009.13.5.343 [DOI] [PMC free article] [PubMed] [Google Scholar]
67. Choi JD, Park MA, Lee JS. Suppression and recovery of BRCA1-mediated transcription by HP1γ via modulation of promoter occupancy. Nucleic Acids Res. 2012;40:11321–38. 10.1093/nar/gks947 [DOI] [PMC free article] [PubMed] [Google Scholar]
68. Zhang F, Ma J, Wu J et al. PALB2 links BRCA1 and BRCA2 in the DNA-damage response. Curr Biol. 2009;19:524–9. 10.1016/j.cub.2009.02.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
69. Sy SMH, Huen MSY, Chen J. PALB2 is an integral component of the BRCA complex required for homologous recombination repair. Proc Natl Acad Sci USA. 2009;106:7155–60. 10.1073/pnas.0811159106 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

gkag715_Supplemental_File

gkag715_supplemental_file.pdf^{(4.5MB, pdf)}

Data Availability Statement

The data underlying this article are available in the article and in its online supplementary material.

[B1] 1. Welcsh PL, King MC. BRCA1 and BRCA2 and the genetics of breast and ovarian cancer. Hum Mol Genet. 2001;10:705–13. 10.1093/hmg/10.7.705 [DOI] [PubMed] [Google Scholar]

[B2] 2. Burgess M, Puhalla S. BRCA 1/2-mutation-related and sporadic breast and ovarian cancers: more alike than different. Front Oncol. 2014;4:19. 10.3389/fonc.2014.00019 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B3] 3. Thai TH, Du F, Tsan JT et al. Mutations in the BRCA1-associated RING domain (BARD1) gene in primary breast, ovarian and uterine cancers. Hum Mol Genet. 1998;7:195–202. 10.1093/hmg/7.2.195 [DOI] [PubMed] [Google Scholar]

[B4] 4. Suszynska M, Kozlowski P. Summary of BARD1 mutations and precise estimation of breast and ovarian cancer risks associated with the mutations. Genes. 2020;11:798. 10.3390/genes11070798 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B5] 5. Hawsawi YM, Shams A, Theyab A et al. BARD1 mystery: tumor suppressors are cancer susceptibility genes. BMC Cancer. 2022;22:599. 10.1186/s12885-022-09567-4 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B6] 6. Cruz-García A, López-Saavedra A, Huertas P. BRCA1 accelerates CtIP-mediated DNA-end resection. Cell Rep. 2014;9:451–9. 10.1016/j.celrep.2014.08.076 [DOI] [PubMed] [Google Scholar]

[B7] 7. Zhao W, Steinfeld JB, Liang F et al. BRCA1–BARD1 promotes RAD51-mediated homologous DNA pairing. Nature. 2017;550:360–5. 10.1038/nature24060 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B8] 8. Tarsounas M, Sung P. The antitumorigenic roles of BRCA1-BARD1 in DNA repair and replication. Nat Rev Mol Cell Biol. 2020;21:284–99. 10.1038/s41580-020-0218-z [DOI] [PMC free article] [PubMed] [Google Scholar]

[B9] 9. Wang M, Li W, Tomimatsu N et al. Crucial roles of the BRCA1–BARD1 E3 ubiquitin ligase activity in homology-directed DNA repair. Mol Cell. 2023;83:3679–91.e8. 10.1016/j.molcel.2023.09.015 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B10] 10. Glover JN, Williams RS, Lee MS. Interactions between BRCT repeats and phosphoproteins: tangled up in two. Trends Biochem Sci. 2004;29:579–85. 10.1016/j.tibs.2004.09.010 [DOI] [PubMed] [Google Scholar]

[B11] 11. Her J, Soo Lee N, Kim Y et al. Factors forming the BRCA1-A complex orchestrate BRCA1 recruitment to the sites of DNA damage. Acta Biochim Biophys Sin (Shanghai). 2016;48:658–64. 10.1093/abbs/gmw047 [DOI] [PubMed] [Google Scholar]

[B12] 12. Lee BL, Singh A, Mark Glover JN et al. Molecular basis for K63-linked ubiquitination processes in double-strand DNA break repair: a focus on kinetics and dynamics. J Mol Biol. 2017;429:3409–29. 10.1016/j.jmb.2017.05.029 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B13] 13. Krais JJ, Wang Y, Patel P et al. RNF168-mediated localization of BARD1 recruits the BRCA1–PALB2 complex to DNA damage. Nat Commun. 2021;12:5016. 10.1038/s41467-021-25346-4 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B14] 14. Coleman KA, Greenberg RA. The BRCA1–RAP80 complex regulates DNA repair mechanism utilization by restricting end resection. J Biol Chem. 2011;286:13669–80. 10.1074/jbc.M110.213728 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B15] 15. Hu Y, Scully R, Sobhian B et al. RAP80-directed tuning of BRCA1 homologous recombination function at ionizing radiation-induced nuclear foci. Genes Dev. 2011;25:685–700. 10.1101/gad.2011011 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B16] 16. Li M, Yu X. Function of BRCA1 in the DNA damage response is mediated by ADP-ribosylation. Cancer Cell. 2013;23:693–704. 10.1016/j.ccr.2013.03.025 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B17] 17. Becker JR, Clifford G, Bonnet C et al. BARD1 reads H2A lysine 15 ubiquitination to direct homologous recombination. Nature. 2021;596:433–7. 10.1038/s41586-021-03776-w [DOI] [PMC free article] [PubMed] [Google Scholar]

[B18] 18. Fradet-Turcotte A, Canny MD, Escribano-Díaz C et al. 53BP1 is a reader of the DNA-damage-induced H2A Lys 15 ubiquitin mark. Nature. 2013;499:50–4. 10.1038/nature12318 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B19] 19. Nakamura K, Saredi G, Becker JR et al. H4K20me0 recognition by BRCA1–BARD1 directs homologous recombination to sister chromatids. Nat Cell Biol. 2019;21:311–8. 10.1038/s41556-019-0282-9 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B20] 20. Bayley R, Borel V, Moss RJ et al. H3K4 methylation by SETD1A/BOD1L facilitates RIF1-dependent NHEJ. Mol Cell. 2022;82:1924–39.e10. 10.1016/j.molcel.2022.03.030 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B21] 21. Wu W, Nishikawa H, Fukuda T et al. Interaction of BARD1 and HP1 is required for BRCA1 retention at sites of DNA damage. Cancer Res. 2015;75:1311–21. 10.1158/0008-5472.CAN-14-2796 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B22] 22. Bannister AJ, Zegerman P, Partridge JF et al. Selective recognition of methylated lysine 9 on histone H3 by the HP1 chromo domain. Nature. 2001;410:120–4. 10.1038/35065138 [DOI] [PubMed] [Google Scholar]

[B23] 23. Schultz DC, Ayyanathan K, Negorev D et al. SETDB1: a novel KAP-1-associated histone H3 lysine 9-specific methyltransferase that contributes to HP1-mediated silencing of euchromatic genes by KRAB zinc-finger proteins. Genes Dev. 2002;16:919–32. 10.1101/gad.973302 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B24] 24. Ayrapetov MK, Gursoy-Yuzugullu O, Xu C et al. DNA double-strand breaks promote methylation of histone H3 on lysine 9 and transient formation of repressive chromatin. Proc Natl Acad Sci USA. 2014;111:9169–74. 10.1073/pnas.1403565111 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B25] 25. Alagoz M, Katsuki Y, Ogiwara H et al. SETDB1, HP1 and SUV39 promote repositioning of 53BP1 to extend resection during homologous recombination in G2 cells. Nucleic Acids Res. 2015;43:7931–44. 10.1093/nar/gkv722 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B26] 26. White D, Rafalska-Metcalf IU, Ivanov AV et al. The ATM substrate KAP1 controls DNA repair in heterochromatin: regulation by HP1 proteins and serine 473/824 phosphorylation. Mol Cancer Res. 2012;10:401–14. 10.1158/1541-7786.MCR-11-0134 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B27] 27. Campbell S, Ismail IH, Young LC et al. Polycomb repressive complex 2 contributes to DNA double-strand break repair. Cell Cycle. 2013;12:2675–83. 10.4161/cc.25795 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B28] 28. Fitieh A, Locke AJ, Mashayekhi F et al. BMI-1 regulates DNA end resection and homologous recombination repair. Cell Rep. 2022;38:110536. 10.1016/j.celrep.2022.110536 [DOI] [PubMed] [Google Scholar]

[B29] 29. Kumbhar R, Sanchez A, Perren J et al. Poly(ADP-ribose) binding and macroH2A mediate recruitment and functions of KDM5A at DNA lesions. J Cell Biol. 2021;220:e202006149. 10.1083/jcb.202006149 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B30] 30. Kendek A, Wensveen MR, Janssen A. The sound of silence: how silenced chromatin orchestrates the repair of double-strand breaks. Genes. 2021;12:1415. 10.3390/genes12091415 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B31] 31. Kang MA, Lee JS. A newly assigned role of CTCF in cellular response to broken DNAs. Biomolecules. 2021;11:363. 10.3390/biom11030363 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B32] 32. Hilmi K, Jangal M, Marques M et al. CTCF facilitates DNA double-strand break repair by enhancing homologous recombination repair. Sci Adv. 2017;3:e1601898. 10.1126/sciadv.1601898 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B33] 33. Lang F, Li X, Zheng W et al. CTCF prevents genomic instability by promoting homologous recombination-directed DNA double-strand break repair. Proc Natl Acad Sci USA. 2017;114:10912–7. 10.1073/pnas.1704076114 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B34] 34. Hwang SY, Kang MA, Baik CJ et al. CTCF cooperates with CtIP to drive homologous recombination repair of double-strand breaks. Nucleic Acids Res. 2019;47:9160–79. 10.1093/nar/gkz639 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B35] 35. Densham RM, Garvin AJ, Stone HR et al. Human BRCA1–BARD1 ubiquitin ligase activity counteracts chromatin barriers to DNA resection. Nat Struct Mol Biol. 2016;23:647–55. 10.1038/nsmb.3236 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B36] 36. Densham RM, Morris JR. The BRCA1 ubiquitin ligase function sets a new trend for remodelling in DNA repair. Nucleus. 2017;8:116–25. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B37] 37. Hu Q, Botuyan MV, Zhao D et al. Mechanisms of BRCA1–BARD1 nucleosome recognition and ubiquitylation. Nature. 2021;596:438–43. 10.1038/s41586-021-03716-8 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B38] 38. Iacovoni JS, Caron P, Lassadi I et al. High-resolution profiling of gammaH2AX around DNA double strand breaks in the mammalian genome. EMBO J. 2010;29:1446–57. 10.1038/emboj.2010.38 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B39] 39. Tang J, Cho NW, Cui G et al. Acetylation limits 53BP1 association with damaged chromatin to promote homologous recombination. Nat Struct Mol Biol. 2013;20:317–25. 10.1038/nsmb.2499 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B40] 40. Gunn A, Stark JM. I-SceI-based assays to examine distinct repair outcomes of mammalian chromosomal double strand breaks. Methods Mol Biol. 2012;920:379–91. [DOI] [PubMed] [Google Scholar]

[B41] 41. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods. 2001;25:402–8. 10.1006/meth.2001.1262 [DOI] [PubMed] [Google Scholar]

[B42] 42. Nicolette ML, Lee K, Guo Z et al. MRE11-Rad50-Xrs2 and Sae2 promote 5 strand resection of DNA double-strand breaks. Nat Struct Mol Biol. 2010;17:1478–85. 10.1038/nsmb.1957 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B43] 43. Zhou Y, Caron P, Leqube G et al. Quantitation of DNA double-strand break resection intermediates in human cells. Nucleic Acids Res. 2014;42:e19. 10.1093/nar/gkt1309 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B44] 44. Zierhut C, Diffley JF. Break dosage, cell cycle stage and DNA replication influence DNA double-strand break response. EMBO J. 2008;27:1875–85. 10.1038/emboj.2008.111 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B45] 45. Ikura M, Furuya K, Fukuto A et al. Coordinated regulation of TIP60 and poly(ADP-Ribose) polymerase 1 in damaged-chromatin dynamics. Mol Cell Biol. 2016;36:1595–607. 10.1128/MCB.01085-15 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B46] 46. Clouaire T, Rocher V, Lashgari A et al. Comprehensive mapping of histone modifications at DNA double-strand breaks deciphers repair pathway chromatin signatures. Mol Cell. 2018;72:250–62.e6. 10.1016/j.molcel.2018.08.020e6 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B47] 47. Aymard F, Bugler B, Schmidt CK et al. Transcriptionally active chromatin recruits homologous recombination at DNA double-strand breaks. Nat Struct Mol Biol. 2014;21:366–74. 10.1038/nsmb.2796 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B48] 48. Yasuhara T, Kato R, Hagiwara Y et al. Human Rad52 promotes XPG-mediated R-loop processing to initiate transcription-associated homologous recombination repair. Cell. 2018;175:558–70.e11. 10.1016/j.cell.2018.08.056 [DOI] [PubMed] [Google Scholar]

[B49] 49. Ouyang J, Yadav T, Zhang JM et al. RNA transcripts stimulate homologous recombination by forming DR-loops. Nature. 2021;594:283–8. 10.1038/s41586-021-03538-8 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B50] 50. Marnef A, Cohen S, Legube G. Transcription-coupled DNA double-strand break repair: active genes need special care. J Mol Biol. 2017;429:1277–88. 10.1016/j.jmb.2017.03.024 [DOI] [PubMed] [Google Scholar]

[B51] 51. Savage KI, Harkin DP. BRCA1, a ‘complex’ protein involved in the maintenance of genomic stability. FEBS J. 2015;282:630–46. 10.1111/febs.13150 [DOI] [PubMed] [Google Scholar]

[B52] 52. Hashizume R, Fukuda M, Maeda I et al. The RING heterodimer BRCA1-BARD1 is a ubiquitin ligase inactivated by a breast cancer-derived mutation. J Biol Chem. 2001;276:14537–40. 10.1074/jbc.C000881200 [DOI] [PubMed] [Google Scholar]

[B53] 53. Burdett H, Foglizzo M, Musgrove LJ et al. BRCA1-BARD1 combines multiple chromatin recognition modules to bridge nascent nucleosomes. Nucleic Acids Res. 2023;51:11080–103. 10.1093/nar/gkad793 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B54] 54. Groner AC, Meylan S, Ciuffi A et al. KRAB-zinc finger proteins and KAP1 can mediate long-range transcriptional repression through heterochromatin spreading. PLoS Genet. 2010;6:e1000869. 10.1371/journal.pgen.1000869 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B55] 55. Aitken SJ, Ibarra-Soria X, Kentepozidou E et al. CTCF maintains regulatory homeostasis of cancer pathways. Genome Biol. 2018;19:106. 10.1186/s13059-018-1484-3 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B56] 56. Randolph K, Hyder U, D’Orso I. KAP1/TRIM28: transcriptional activator and/or repressor of viral and cellular programs?. Front Cell Infect Microbiol. 2022;12:834636. 10.3389/fcimb.2022.834636 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B57] 57. Gong F, Clouaire T, Aguirrebengoa M et al. Histone demethylase KDM5A regulates the ZMYND8-NuRD chromatin remodeler to promote DNA repair. J Cell Biol. 2017;216:1959–74. 10.1083/jcb.201611135 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B58] 58. Kumbhar R, Sanchez A, Perren J et al. Poly(ADP-ribose) binding and macroH2A mediate recruitment and functions of KDM5A at DNA lesions. J Cell Biol. 2021;220:e202006149. 10.1083/jcb.202006149 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B59] 59. Feng S, Ma S, Li K et al. RIF1-ASF1-mediated high-order chromatin structure safeguards genome integrity. Nat Commun. 2022;13:957. 10.1038/s41467-022-28588-y [DOI] [PMC free article] [PubMed] [Google Scholar]

[B60] 60. Lachner M, O’Carroll D, Rea S et al. Methylation of histone H3 lysine 9 creates a binding site for HP1 proteins. Nature. 2001;410:116–20. 10.1038/35065132 [DOI] [PubMed] [Google Scholar]

[B61] 61. Rassoolzadeh H, Coucoravas C, Farnebo M. The proximity ligation assay reveals that at DNA double-strand breaks WRAP53β associates with γH2AX and controls interactions between RNF8 and MDC1. Nucleus. 2015;6:417–24. [DOI] [PMC free article] [PubMed] [Google Scholar]

[B62] 62. Densham RM, Garvin AJ, Stone HR et al. Human BRCA1–BARD1 ubiquitin ligase activity counteracts chromatin barriers to DNA resection. Nat Struct Mol Biol. 2016;23:647–55. 10.1038/nsmb.3236 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B63] 63. Ciccia A, Elledge SJ. The DNA damage response: making it safe to play with knives. Mol Cell. 2010;40:179–204. 10.1016/j.molcel.2010.09.019 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B64] 64. Swift ML, Beishline K, Flashner S et al. DSB repair pathway choice is regulated by recruitment of 53BP1 through cell cycle-dependent regulation of Sp1. Cell Rep. 2021;34:108840. 10.1016/j.celrep.2021.108840 [DOI] [PubMed] [Google Scholar]

[B65] 65. Yu W, Ginjala V, Pant V et al. Poly(ADP-ribosyl)ation regulates CTCF-dependent chromatin insulation. Nat Genet. 2004;36:1105–10. 10.1038/ng1426 [DOI] [PubMed] [Google Scholar]

[B66] 66. Lee J-H, Cheong H-M, Kang M-Y et al. Ser1778 of 53BP1 plays a role in DNA double-strand break repairs. Korean J Physiol Pharmacol. 2009;13:343–8. 10.4196/kjpp.2009.13.5.343 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B67] 67. Choi JD, Park MA, Lee JS. Suppression and recovery of BRCA1-mediated transcription by HP1γ via modulation of promoter occupancy. Nucleic Acids Res. 2012;40:11321–38. 10.1093/nar/gks947 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B68] 68. Zhang F, Ma J, Wu J et al. PALB2 links BRCA1 and BRCA2 in the DNA-damage response. Curr Biol. 2009;19:524–9. 10.1016/j.cub.2009.02.018 [DOI] [PMC free article] [PubMed] [Google Scholar]

[B69] 69. Sy SMH, Huen MSY, Chen J. PALB2 is an integral component of the BRCA complex required for homologous recombination repair. Proc Natl Acad Sci USA. 2009;106:7155–60. 10.1073/pnas.0811159106 [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

CTCF primes repressive chromatin for rapid BRCA1/BARD1 recruitment to drive homologous recombination repair of DNA double-strand breaks

Soon Young Hwang

Daeki Choi

Myoung Joo Choi

Hye-Ram Jo

Jae-Hoon Jeong

Mi Ae Kang

Jong-Soo Lee

Roles

Abstract

Graphical Abstract

Graphical Abstract.

Introduction

Materials and methods

Cells and reagents

Plasmid construction and transfection

Immunofluorescence

Immunoblot assay

Immunoprecipitation

Irradiation and laser micro-irradiation

Chromatin immunoprecipitation assay

Protein binding assay

DNA end resection assay

In situ proximity ligation assay

HR and NHEJ assays

Antibodies

Statistical analysis

Results

Rapid enrichment of BRCA1/BARD1 at DSB sites depends on CTCF

Figure 1.

CTCF assists rapid BRCA1/BARD1 recruitment on DNA lesions through CTCF-BRCA1/BARD1 interaction

Figure 2.

CTCF couples with HP1γ to rapidly recruit BRCA1/BARD1 at sites of DNA DSBs

Figure 3.

Figure 4.

CTCF supports transient enrichment of H3K9 methylation to promptly recruit BRCA1/BARD1 around DSBs

Figure 5.

CTCF assists the transient disappearance of H3K4me3 to promote rapid BRCA1/BARD1 localization at sites of DNA damage

Figure 6.

CTCF-reliant rapid recruitment of BRCA1/BARD1 drives EXO1/DNA2-mediated extensive resection of DSB ends

Figure 7.

Discussion

CTCF prompts BRCA1/BARD1 localization at DNA damage sites

CTCF coordinates transient repressive histone marks and heterochromatin proteins to accelerate localization of BRCA1/BARD1 at DSBs

A CTCF-driven cascade to accelerate DNA end resection at DSBs for homologous recombination repair

Supplementary Material

Acknowledgements

Contributor Information

Supplementary data

Conflict of interest

Funding

Data availability

References

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases