Introduction

DNA double-strand break (DSB) is a form of damage, wherein the phosphate backbones of the two complementary DNA strands are simultaneously broken, leading to serious consequences including chromosomal aberrations or cell death (Mehta and Haber 2014). DSBs may be caused by exogenous factors, such as chemical mutagens, chemotherapeutic DNA damaging agents, ionizing radiation, or by endogenous stressors including abnormal DNA replication and reactive oxygen species (Chen and Stubbe 2005, Wyrobek et al. 2005). During replication, prolonged stalling of replication forks results in the generation of extended stretches of single-stranded DNA (ssDNA). Persistent replication stress causes fork collapse, resulting in highly lethal DNA DSBs (Lopes et al. 2001; Aladjem and Redon 2017; Kurat et al. 2017). A cascade of protein modifications involving relocalization and degradation occurs in response to DSBs as follows: the Mre11-rad50-Ns1 (MRN) complex (sensor) recruits Ataxia telangiectasia mutated (ATM) kinase, which is known as the master factor at the DSB sites (Fishman-Lobell, Rudin et al. 1992; Manis, Morales et al. 2004; Ma, Lu et al. 2005; Difilippantonio, Gapud et al. 2008; Bunting, Callen et al. 2012), followed by rapid phosphorylation of H2AX at serine 139 in an ATM-dependent manner (Burma, Chen et al. 2001; Fernandez-Capetillo, Lee et al. 2004). Amplification of this signal can lead to the subsequent recruitment of several proteins, including E3 ubiquitin ligase ring finger protein 8 (RNF8) and RNF168. This ubiquitination-based decision of localization or degradation of the downstream factors facilitates DSB repair by homologous recombination (HR) and nonhomologous end joining (NHEJ) pathways (Huyen, Zgheib et al. 2004; Doil, Mailand et al. 2009; Stewart, Panier et al. 2009; Symington and Gautier 2011; Mallette, Mattiroli et al. 2012).

The AAA + (Associated with diverse cellular Activities) ATPase p97, also known as valosin-containing protein (p97/VCP), is a central component of the ubiquitin–proteasome system (Dantuma and Hoppe 2012; Vaz, Halder et al. 2013; Meyer and Weihl 2014). P97/VCP segregates ubiquitinated substrates involved in DNA replication, transcription, repair, and cell cycle progression using ATP-derived energy for degradation or recycling (Torrecilla, Oehler et al. 2017). DNA repair proteins breast cancer type 1 (BRCA1) and Werner helicase interact with p97/VCP (Zhang, Wang et al. 2000; Partridge, Lopreiato et al. 2003; Indig, Partridge et al. 2004), which physically associates with DSBs and removes ubiquitinated lethal 3 malignant brain tumor 1 (L3MBTL1), Ku, and KRAB-associated protein 1 (KAP1) for the subsequent loading of 53BP1, BRCA1, and Rad51 (Acs, Luijsterburg et al. 2011, Meerang, Ritz et al. 2011, Kuo, Li et al. 2016, van den Boom, Wolf et al. 2016). Inactivation or depletion of the p97/VCP-mediated segregation system compromises HR and NHEJ pathways, demonstrating the essential role of p97/VCP in DNA repair (Acs, Luijsterburg et al. 2011; Meerang, Ritz et al. 2011).

The replication origin-binding protein RepID, also known as Pleckstrin homology domain-interacting protein (PHIP) and DDB1/CUL4-associated factor 14 (DCAF14), is a multifunctional protein that facilitate proliferation of cancer cells (Farhang-Fallah, Randhawa et al. 2002; Podcheko, Northcott et al. 2007; Tokuda, Fujita et al. 2007; Li, Francisco et al. 2010; De Semir, Nosrati et al. 2012; de Semir, Bezrookove et al. 2018). Chromatin-localized RepID selectively associates with a subgroup of replication origin proteins to initiate DNA replication, possibly by regulating the formation of chromatin loops (Zhang, Huang et al. 2016). RepID-depleted cells show reduced initiation frequency as well as slow elongation and frequent stalling at replication forks (Zhang, Huang et al. 2016). The Cullin-RING ubiquitin ligase complex 4 (CRL4) is recruited to the chromatin by RepID (Jang, Redon et al. 2018a, b; Jang, Zhang et al. 2018a, b). In the absence of RepID, CRL4 cannot ubiquitinate substrates localized on the chromatin, resulting in re-replication via accumulation of the DNA replication origin licensing factor CDT1 or delay in the transition from metaphase to anaphase during mitosis owing to the failure to degrade BUB3, a subunit of the spindle assembly checkpoint proteins blocking the anaphase-promoting complex/cyclosome ubiquitin E3 ligase (APC/C)(Jang, Nathans et al. 2020a, b; Jang, Redon et al. 2020a, b). These abnormal replication and chromosomal segregation patterns in the absence of RepID suggest that RepID plays a role in the prevention of DNA damage during cell proliferation.

Here, we report that RepID is a safeguard protein that prevents excessive endogenous DNA damage. We observed that RepID-depleted cells showed increased levels of DSBs. This event correlated with higher recruitment of p97/VCP on the chromatin in the absence of RepID, leading to hypersensitivity of cancer cells to the p97/VCP-targeting inhibitor CB5083. Our findings suggest that RepID levels could be investigated as possible effectors of cancer therapy based on the DNA damage-p97/VCP axis.

Materials and methods

Cell culture and chemicals

Human U2OS osteosarcoma cells with and without RepID were incubated in Dulbecco’s modified Eagle’s medium (Invitrogen, 10,569–010) supplemented with 10% heat-inactivated fetal bovine serum in a 37 °C/5% CO2 humidified incubator. U2OS original cancer cell lines were obtained from The American Type Culture Collection (ATCC; www.atcc.org). All cell lines were tested negative for mycoplasmas (Lonza, LT07-418). P97/VCP inhibitor CB5083 (Selleckchem, S8101) or CPT (Selleckchem, S1288) was added to the medium at indicated concentrations.

RepID-depleted cell lines

RepID was depleted in U2OS cells using the CRISPR/CAS9 tool. A 20 base-pair guide sequence targeting the fifth exon of RepID (5ʹ-CTGCAAATATGTCATCGACTAGG-3ʹ) and the eighth exon of RepID (5ʹ-GTGATAAAATGATCCGAGTCTGG-3ʹ) in U2OS cells was selected from a published database of predicted high-specificity protospacer-PAM target sites in the human exome. Cells were cultured in six-well dishes to 70%-80% confluency for co-transfection with 2 μg RepID single-guide RNA (sgRNA) plasmid, 2 μg linearized pCR2.1 vector harboring a puromycin-resistance gene expression cassette, and 10 μL Lipofectamine 2000 (Life Technologies). Cloning, selection, and verification were performed using polymerase chain reaction (PCR).

Fluorescence-activated cell sorting (FACS) analysis

Cells were labeled with 10 μM EdU for 30 min before harvesting using the Click-iT EdU kit (Invitrogen, C10424). Cell staining was performed according to the manufacturer’s protocol. We used 4′,6-diamidino-2-phenylindole (DAPI) for DNA counterstaining. An LSR Fortessa cell analyzer (BD Biosciences) with FlowJo 10.5.2 software was used for cell cycle analyses. All experiments reported representative results of at least three independent repetitions.

Clonogenic survival assay

U2OS cells were plated in six-well plates (500 cells/well) in triplicates and treated with CB5083 for 10 days. Colonies were fixed with 1% paraformaldehyde (PFA) and stained with crystal violet. Well intensity was measured using ImageJ software. All experiments reported representative results of at least three independent repetitions.

Immunofluorescence analysis

U2OS cells were incubated in phosphate-buffered saline (PBS)-T (0.2% Triton X-100 in 1 × PBS, protease inhibitor cocktail [Sigma, P8340], phenylmethylsulfonyl fluoride [PMSF], and phosphatase inhibitor cocktail [Roche, P4906845001]) for 5 min on ice, and then fixed with 2% PFA. Primary antibody staining was performed as follows: anti-phosphorylated ATM (Thermo Fisher Scientific, MA5-15,185, 1:200), anti-53BP1 (Novus, NB100-305A, 1:200), anti-p97/VCP (Abcam, ab11433, 1:100), and anti-γH2AX (Millipore, 05–636, 1:1000) for 3 h at room temperature. Secondary antibody staining was performed as follows: Alexa 488-conjugated anti-mouse IgG and Alexa 555-conjugated anti-rabbit IgG (1:500, Thermo Fisher Scientific, A11029 and A21428). EdU was detected using the Click-iT assay kit according to the manufacturer’s protocol. A Zeiss LSM710 confocal microscope and a Visitech VT-ISIM were used, and coefficients were generated using ImageJ software and plugins.

Chromatin fractionation and immunoblotting

Harvested U2OS cells were incubated in a cytosol extraction buffer containing NP-40 (20 mM Tris–HCl pH 7.4, 10 mM sodium chloride [NaCl], 3 mM magnesium chloride [MgCl2], 0.5% NP-40, PMSF, protease inhibitor cocktail, and phosphatase inhibitor cocktail). Cells were harvested by centrifugation at 3000×g for 5 min at 4 °C, washed, and resuspended in a nuclear extraction buffer (10 mM Tris–HCl pH 7.4, 100 mM NaCl, 1% Triton X-100, 1 mM ethylenediaminetetraacetic acid [EDTA] pH 8, 1 mM ethylene glycol tetraacetic acid [EGTA], 0.1% sodium dodecyl sulfate [SDS], 10% glycerol, 0.5% sodium deoxycholate, protease inhibitor cocktail, and phosphatase inhibitor cocktail). The suspension was vortexed, incubated on ice, and then centrifuged at 5000 × g for 5 min at 4 °C. The pellet was resuspended in nuclear extraction buffer containing 5 mM calcium chloride (CaCl2) and micrococcal nuclease (New England Biolabs, Cat. M0247S), vortexed, and incubated at 37 °C for 5 min. Chromatin-bound fractions were collected after centrifugation at 18,000×g for 5 min at 4 °C. Total cell lysates and chromatin-bound proteins were detected by SDS–polyacrylamide gel electrophoresis. The following primary antibodies were used: anti-RepID (NCI186, 1:1,000), anti-p97/VCP (Abcam, ab11433, 1:2,000), anti-γH2AX (Millipore, 05–636, 1:2,000), anti-α-tubulin (Sigma, T9026, 1:2,000), and anti-histone H3 (Millipore, 07–690, 1:20,000). For secondary antibodies, horseradish peroxidase (HRP)-linked anti-mouse IgG (Cell Signaling, 7076) and HRP-linked anti-rabbit IgG (Cell Signaling, 7074) were used following the manufacturer’s protocols.

Results

RepID prevents excessive DNA damage

To determine whether RepID plays a role in regulating the generation of DSBs, we first measured the levels of phosphorylated ATM, the master regulator of DNA damage response in U2OS human osteosarcoma cell lines, which were deficient of RepID (RepID knock-out [KO]). RepID depletion was performed using the clustered regularly interspaced short palindromic repeats (CRISPR)/caspase 9 (CAS9) gene editing tool and single clone was selected followed by verification (Jang, Zhang et al. 2018a, b). Confocal microscopy showed that phosphorylated ATM foci were more frequent in RepID-deficient cells than in RepID-proficient cells; this phenotype was consistent during all cell cycles, including EdU-positive S-phase or EdU-negative G1/G2 phase (Fig. 1a, c, dimethyl sulfoxide [DMSO] panel). Treatment with camptothecin (CPT), a DSB-inducing agent that inhibits topoisomerase I (Wall and Wani 1995), significantly induced phosphorylated ATM signals that were stronger in RepID-deficient cells than in RepID-proficient cells (Fig. 1a, c, CPT panel). Consistent with the observed behavior of ATM foci, foci with the DSB marker γH2AX and the NHEJ player 53BP1 were notably detected after RepID depletion (Fig. 1b, d, e). The colocalization of 53BP1 and γH2AX increased in CPT-treated RepID KO cells (Fig. 1f). These results suggest that RepID prevents excessive DNA damage.

Fig. 1
figure 1

RepID prevents excessive DNA damage during cell cycle progression. a, b RepID proficient (Wild Type—WT) or RepID-depleted (Knock Out—KO) U2OS cells were incubated with/without 0.5 M CPT for 1 h. EdU-labeled, pre-extracted, and phosphorylated ATM (green) (a), γH2AX (green), and 53BP1 (red) (b), DNA content (DAPI) and EdU (magenta) were detected. Cells in replicating S-phase (EdU-positive) or non-replicating G1/G2 phase (EdU-negative) were identified by EdU staining. Scale bar indicates 10 m. (c-e) Relative intensity. Phosphorylated ATM (c), 53BP1 (d), and γH2AX (E). (f) Extent of colocalization between 53BP1 and γH2AX. Pearson’s correlation coefficients (n = 20). p-values were calculated using a two-tailed t-test, and error bars represent standard deviation from three independent experiments (***p-value < 0.001)

Full size image

Recruitment of P97/VCP to DSB sites increases RepID-deficient cancer cells

P97/VCP segregase is an essential factor regulating DSB repair by orchestrating the segregation of ubiquitinated DSB repair proteins (Torrecilla, Oehler et al. 2017). We investigated p97/VCP localization in the RepID-deficient cells which generate excessive DSBs. Nuclei were pre-extracted to remove soluble nuclear proteins and incubated with p97/VCP and γH2AX antibodies. Most p97/VCP was detected in the nucleoli of RepID-proficient and -deficient cells (Fig. 2a, DMSO panel). However, in RepID-deficient cells, some p97/VCP was detected outside the nucleoli and colocalized with γH2AX foci (Fig. 2a, b, DMSO panel). RepID-proficient cells also showed some colocalization of p97/VCP and γH2AX after CPT treatment along with an increase in γH2AX foci together with diffused patterns of p97/VCP (Fig. 2a, b, CPT panel). However, intensity profile analyses revealed that the colocalization of p97/VCP and γH2AX was more evident in CPT-treated RepID-deficient cells (Fig. 2a, b, CPT panel) than in RepID-proficient cells.

Fig. 2
figure 2

Recruitment of p97/VCP to DSB sites is increased in RepID-deficient cancer cells. a U2OS cells were treated with/without 1.0 M CPT for 1 h, pre-extracted, and incubated with p97/VCP (green) and γH2AX (red) antibodies together with DAPI (blue). Scale bar indicates 10 m. b Colocalization of p97/VCP and γH2AX was analyzed by intensity profiling as indicated in (a). Gray backgrounds indicate colocalized region. c Levels of p97/VCP from RepID WT or RepID KO U2OS cancer cell lines after exposure to MG132 proteasome inhibitor. Histone H3 and α-tubulin were used as loading controls. Total, total cell lysates; chromatin, chromatin-bound proteins. d Quantification of chromatin-bound p97/VCP normalized to the levels in untreated RepID WT cells. Error bars represent standard deviations from three independent experiments. p-values were calculated using a two-tailed t-test, and error bars represent standard deviation from three independent experiments (***p-value < 0.001)

Full size image

To evaluate the extent of p97/VCP recruitment to DSBs, we used MG132 that accumulates ubiquitinated proteins, including DSB repair factors, by preventing proteasomal degradation. In the absence of RepID or MG132, we did not observe any effect on the expression of p97/VCP (Fig. 2c). However, the chromatin-bound fractions of p97/VCP significantly increased in RepID-deficient cells (2.43-fold). The recruitment of p97/VCP on the chromatin was notably clearer after MG132 treatment in RepID-depleted cells (6.21-fold in RepID wild-type [WT] vs 13.7-fold in RepID KO cells) (Fig. 2c, d). These results show that p97/VCP recruitment to the chromatin is enhanced along with excessive DNA damage after RepID depletion.

RepID-deficient cancer cells are sensitive to p97/VCP inhibitor

The above observations suggest that the function of p97/VCP might be crucial for DNA repair in RepID-deficient cancer cells. To determine whether p97/VCP-mediated segregation of ubiquitinated proteins during DNA damage response and repair is critical for the proliferation of RepID-deficient cells, both RepID WT and RepID KO cells were exposed to the p97/VCP inhibitor (p97i) CB5083 (Anderson, Le Moigne et al. 2015). As shown in Fig. 3a and b, RepID-deficient cells were more sensitive to p97/VCP inhibition than RepID proficient cells (no surviving colonies in RepID deficient cells treated with 0.6 M p97i). On the other hand, the cells expressing intact RepID were resistant to p97i treatment. A fraction of both RepID-expressing and RepID-depleted cells (SubG1 fraction) died following exposure to CB5083 for up to 4 days. The number of cells in SubG1 was higher in RepID-depleted group than in RepID-expressing group (24 h: 1.36% and 3.61% in WT and KO, respectively; 48 h: 2.28% and 7.06% in WT and KO, respectively; 96 h: 1.03% and 10.3% in WT and KO, respectively) (Fig. 3c, d). Levels of p97/VCP on the chromatin were clearly higher in RepID-depleted cells than in RepID WT cells, consistent with the induced expression of γH2AX following CB5083 treatment (Fig. 3e). Taken together, our observations suggest that RepID expression may modulate the sensitivity of cancer cells to p97/VCP-targeting drugs.

Fig. 3
figure 3

RepID-deficient cancer cells are sensitive to p97/VCP inhibitor. a, b Colony formation assay with U2OS RepID WT and KO cells upon CB5083 treatment (a). Bar charts indicate relative growing intensity by measuring covered area (b). c CB5083-treated RepID WT and KO U2OS cells were labeled with EdU for 30 min and analyzed by flow cytometry. Percentages of cells in each cell cycle phase are indicated in the flow cytometry plots. d Fold changes in the SubG1 phase were based on the values from non-treated RepID WT cells. P-values were calculated using a two-tailed t-test and error bars in all results represent standard deviation from three independent experiments (**p-value < 0.01, ***p < 0.001). e Immunoblot analysis showing elevated accumulation of p97/VCP on the chromatin of RepID-deficient cells in response to CB5083 treatment. The numbers under the panels represent the intensity ratios for p97/VCP normalized by the intensity of the signal in untreated RepID WT or the intensity ratios for γH2AX normalized by each histone H3 signals from three independent experiments

Full size image

Discussion

The data presented here demonstrate the crucial role of RepID protein in preventing excessive DNA damage at endogenous levels. Fewer replication initiation events and stalled replication forks with asymmetric progression in RepID-depleting cells may cause single-strand DNA breaks (SSBs) by generating excessive ssDNA through the continuous DNA unwinding in a minichromosome maintenance (MCM) helicase-dependent manner (Smith, Fu et al. 2009; Van, Yan et al. 2010). These replication stresses are susceptible to fork collapse, subsequently leading to DSB formation (Pfeiffer, Goedecke et al. 2000). DSB can also occur in the absence of CRL4 recruitment on the chromatin in RepID-deficient cells. Failure to degrade CDT1, known as an origin-licensing factor and a substrate of CRL4, precedes re-replication mediated by CDT1-based re-licensing in RepID KO cells (Machida, Hamlin et al. 2005; Jang, Zhang et al. 2018a, b). Accumulation of other CRL4 substrates such as the histone methyltransferase SET8 or the CDK inhibitor p21, which is required for proper S-phase progression, prevention of spontaneous DNA damage and activation of the G2/M checkpoint, can also generate DSB (Jorgensen, Elvers et al. 2007; Abbas, Sivaprasad et al. 2008; Huen, Sy et al. 2008; Kim, Starostina et al. 2008; Liu, Lee et al. 2009; Tardat, Brustel et al. 2010). The BUB3-mediated delay in the metaphase-anaphase transition, followed by abnormal chromosomal segregation in RepID-deficient cells are also possibly associated with the occurrence of DNA damage foci (Janssen, van der Burg et al. 2011, Jang, Nathans et al. 2020a, b).

RepID’s functional domains mediate its interaction with chromatin and might underlie its role in regulating the DNA damage response. RepID directly interacts with chromatin using its cryptic Tudor domain and bromodomain (Ji, Dadon et al. 2015; Morgan, Rickels et al. 2017). Many bromodomain-containing proteins have been recently identified as crucial factors promoting genome stability (Kim, Lee et al. 2019). In line with previous reports, we observed elevated DSB markers such as phosphorylated ATM, γH2AX, and 53BP1 foci in RepID-deficient cells regardless of whether those cells were actively engaged in DNA synthesis. These observations suggest that RepID is required to protect cells from excessive DNA damage during all phases of the cell cycle. Future studies are warranted to identify proteins that serve as direct targets of CRL4 and how RepID prevents the generation of endogenous DNA damage.

We observed that a p97/VCP-targeting anticancer drug was highly potent in RepID-deficient cancer cells. This finding is in line with the association of p97/VCP function in DSB and repair (Torrecilla, Oehler et al. 2017). Notably, the chromatin association of p97/VCP increased in RepID-deficient cells, correlated with an elevated prevalence of DNA damage. Taken together with the redistribution and mobilization of nucleolar p97/VCP, along with other repair proteins (e.g. WRN), into the nucleoplasm after exposure to CPT, these observations are consistent with the proposed role of the nucleolus as a depository of repair proteins that are rapidly deployed upon exposure to potentially genotoxic perturbations (Yang, Maiguel et al. 2002; Nalabothula, Indig et al. 2010; Indig, Rybanska et al. 2012; Torrecilla et al. 2017).

We propose the following model for the RepID-p97/VCP axis in DNA damage response and repair (Fig. 4). Depletion of RepID results in endogenous DNA damage caused by abnormal replication, attenuation of CRL4 chromatin recruitment and low HR efficiency. This spontaneous DNA damage facilitates p97/VCP chromatin binding to catalyze the segregation of ubiquitinated proteins during the repair process. As RepID-deficient cells exhibit excessive DNA damage, such cells may depend on p97/VCP activity to a higher extent than RepID-proficient cells. This is consistent with the increased potency of p97/VCP inhibitor in RepID-KO cells.

Fig. 4
figure 4

Model depicting the RepID-p97/VCP axis during DNA damage response and repair. In RepID WT cells, DNA replication in S-phase (bidirectional replication as indicated in green and red arrows with same length) or chromosomal segregation during mitosis progresses normally. RepID possibly contributes to the HR DNA repair pathway. In RepID-depleted (RepID KO) cells, excessive DSBs can be generated at the endogenous levels by following reasons: (1) Abnormal DNA replication that includes fewer replication initiation and asymmetric progression (different length of the green and red arrows) resulting in stalled replication forks; (2) accumulation of CRL4 substrates crucial for the prevention of re-replication, spontaneous DNA damage, or mis-segregation of the chromosome owing to compromised CRL4 recruitment on the chromatin; (3) attenuation of HR activity. Thus, p97/VCP recruitment to DSBs increased in RepID KO cells, leading to increased sensitivity to the p97/VCP inhibitor

Full size image

Our study has potentially significant clinical implications. The 97/VCP network is now regarded as an attractive anticancer target, and CB5083 is currently being tested in clinical trials (Anderson et al. 2015). Our findings provide an important connection that underlies the DNA damage and repair effects of RepID and p97/VCP in cancer and lay the foundation for new strategies to target p97/VCP based on RepID interactions in cancer.