RPA-independent activation of the ATR/CHK1 pathway

Significance

The ATR/CHK1 signaling pathway is a central guardian of genome integrity, canonically thought to depend on RPA-coated single-stranded DNA for activation. Using an acute RPA2-degron system, we uncovered an unexpected RPA-independent route to ATR/CHK1 activation mediated by RAD9 and TOPBP1. This finding reveals a previously unrecognized redundancy in the replication stress response, demonstrating that cells can preserve checkpoint signaling and genome stability even when the canonical RPA-dependent mechanism is disrupted. Our findings expand the conceptual framework of DNA damage signaling, provide insight into the robustness of replication stress responses, and suggest additional targets for therapeutic strategies aimed at exploiting checkpoint vulnerabilities in cancer.

Abstract

The ATR/CHK1 pathway governs a crucial intra-S-phase checkpoint that safeguards genome stability under replication stress by stabilizing stalled replication forks and ensuring high-fidelity DNA replication. Traditionally, activation of this pathway is thought to rely on replication protein A (RPA)-coated single-stranded DNA, which recruits the ATR–ATRIP complex to sites of stalling fork, positioning RPA as essential for ATR signaling. In this study, we report a surprising and previously unrecognized phenomenon: acute depletion of RPA2 triggers robust ATR/CHK1 activation through an RPA-independent mechanism. Using 293A and RPE-1 cells engineered with an inducible RPA2-dTAG degron system, we observed increased phosphorylation of CHK1 at Ser296 and Ser345 in the absence of RPA. Notably, this elevated CHK1 phosphorylation was abolished by ATR inhibition, confirming its dependence on ATR kinase activity. Mechanistic analyses further revealed that this RPA-independent activation requires the checkpoint mediators RAD9 and TOPBP1. These findings uncover dual mechanisms, both RPA-dependent and -independent, of ATR/CHK1 pathway activation, highlighting a robust and flexible replication stress response network that preserves genome integrity even when canonical signaling is disrupted.

The stability of the eukaryotic genome is continually challenged by intrinsic and extrinsic factors that threaten the integrity of replicating DNA. The fidelity of DNA replication is primarily safeguarded by the S-phase checkpoint, particularly the ATR/CHK1 pathway, which is activated in response to replication stress to delay cell cycle progression and stabilize stalled replication forks (1–3). Canonically, ATR, in complex with ATRIP, is recruited to stalled replication forks through direct interaction with replication protein A (RPA), a heterotrimeric complex that binds single-stranded DNA (ssDNA) generated when the MCM helicase becomes uncoupled from DNA polymerases (4, 5). Upon recruitment, ATR is activated by its effector proteins, including TOPBP1 and/or ETAA1, and subsequently phosphorylates CHK1 to propagate the DNA damage response (DDR) (6–11). Despite this well-characterized mechanism, it remains an open question whether RPA-mediated recruitment of the ATR–ATRIP complex is strictly required for ATR/CHK1 activation, or if alternative RPA-independent mechanisms can fulfill this role.

The RPA complex, composed of three tightly associated subunits, RPA1, RPA2, and RPA3, plays an essential role in DNA replication, repair, and recombination. RPA binds ssDNA through six oligonucleotide/oligosaccharide-binding (OB) fold domains (12–15), which not only mediate ssDNA binding but also facilitate the heterotrimeric assembly of the RPA complex (16, 17). By coating ssDNA, RPA protects it from nucleolytic degradation and provides a platform for the recruitment of DDR factors (18). Critically, RPA-coated ssDNA interacts with the ATR–ATRIP complex, while the RAD17–RFC (2 to 5) complex facilitates loading of the RAD9–HUS1–RAD1 (9–1–1) clamp at ssDNA–dsDNA junctions, enabling recruitment of TOPBP1. This spatial coordination promotes interaction between TOPBP1 and ATR, ultimately leading to ATR/CHK1 pathway activation (6, 7, 18–21). However, it remains unclear whether ATR activation via the RAD17/RFC–9–1–1–TOPBP1 axis strictly depends on RPA-mediated recruitment of the ATR–ATRIP complex.

In addition to the RPA complex itself, posttranslational modifications (PTMs) of RPA, particularly phosphorylation of the RPA2 subunit at its N terminus, have been implicated in regulating DNA damage signaling and repair (18). For example, phosphorylation of RPA2 at Ser33 (pRPA2-S33) is commonly used as a marker of DNA end resection during homologous recombination (HR) (22, 23). However, the precise functional role of N-terminal RPA2 phosphorylation remains incompletely understood. In this study, we demonstrate that abolishing N-terminal RPA2 phosphorylation has no detectable impact on cell viability, DDR signaling, RAD51 foci formation, or sensitivity to genotoxic agents. Most notably, we uncover a previously unrecognized RPA-independent mechanism of ATR/CHK1 activation that operates through the RAD9–HUS1–RAD1 complex and TOPBP1. These findings identify an alternative pathway by which cells can respond to replication stress when canonical RPA-dependent signaling is impaired, significantly broadening our understanding of the dynamic and redundant nature of the DDR.

Materials and Methods

Cell Lines.

HEK293A (293A), HEK293T (293T), and hTERT RPE-1 (RPE-1) cells (ATCC) were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Corning) supplemented with 10% fetal bovine serum (FBS; Sigma). To generate HEK293A-RPA2cki-RPA2-dTAG (293A-RPA2-dTAG) and RPE1-RPA2cki-dTAG-mCherry (RPE1-RPA2-dTAG) cell lines, HEK293A and hTERT RPE-1 cells were cotransfected with a single-guide RNA (sgRNA) targeting the C terminus of the RPA2 gene and a donor vector. The donor construct contained a dTAG cassette comprising a flexible linker, FKBPF36V-2HA tag, P2A self-cleaving sequence, and either a blasticidin resistance gene or mCherry reporter, all flanked by RPA2 homology arms within a pUC19 backbone, as previously described (24). Forty-eight hours posttransfection, cells were selected with 10 μg/mL blasticidin S (Gibco) for 5 d or enriched based on mCherry fluorescence by fluorescence-activated cell sorting (FACS). Single-cell clones were obtained by limiting dilution into 96-well plates, and positive clones were identified via genomic PCR and validated by western blotting.

To establish wild-type (WT) 293A and 293A-RPA2-dTAG cell lines stably expressing tFucci(SA)5, cells were cotransfected with the Super PiggyBac transposase expression vector (PB210PA-1; System Biosciences) and the tFucci(SA)5 construct (Addgene, Cat. No. 153520) using polyethylenimine (PEI). Following transfection, Fucci(SA)-positive single cells were isolated by FACS based on both green and red fluorescence signals. Sorted cells were subsequently plated into 96-well plates for single-cell cloning (25).

Constructs and Cloning.

The RPA2 sgRNA (GCAGAATAACTGGATCTAAC) was cloned into the pX330 vector (Addgene, Cat. No. 42230) (26) to enable C-terminal dTAG knock-in at the endogenous RPA2 locus. The dTAG DNA fragment used for conditional knock-in (CKI) was PCR-amplified from the pCRIS-PITChv2-dTAG-blasticidin plasmid (Addgene, Cat. No. 91795) (27). To construct the donor vector for RPA2 knock-in, the 5′ homology arm, dTAG cassette, and 3′ homology arm were assembled into a pUC19 backbone using Gibson assembly.

For gene knockdown experiments, RAD9-sgRNA1 (GGTAGAGCTCGTCCCCGATG), RAD9-sgRNA2 (GCGCTGTAAGATCCTGATGA), TOPBP1-sgRNA1 (TCAAAGACAAC GCCACTAAA), and TOPBP1-sgRNA2 (TATATCTGTGACCCTTTTAG) were cloned into the lentiCRISPRv2 vector (Addgene, Cat. No. 52961) to enable CRISPR/Cas9-mediated disruption of RAD9 or TOPBP1.

The RPA2-ST10A fragment was PCR-amplified from pEGFP-hsRPA2-Ala10 (Addgene, Cat. No. 208073) and subcloned into a modified pLEX_307 SFB vector using Gateway recombination cloning. Similarly, RPA2 truncation mutants were PCR-amplified from pEGFP-hsRPA2 (Addgene, Cat. No. 208068) and inserted into the same modified pLEX_307 SFB vector via Gateway cloning.

Antibodies and Inhibitors.

Antibodies against pKAP1 (S824; cat. no. 4127S), pATM (S1981; cat. no. 4526L), pChk2 (T68; cat. no. 2661S), HA (cat. no. 2999S), pChk1 (S345; cat. no. 2348S), pH3 (S10; cat. no. 9701S), and TOPBP1 (cat. no. 14342) were purchased from Cell Signaling Technology. Antibodies against RPA2 purchased from Abcam (cat. no. 2175) and Cell Signaling Technology (cat. no. 2208) were respectively used in 293A and RPE-1 cells. An antibody against vinculin (cat. no. V9264) was purchased from Sigma. An antibody against γH2AX was purchased from BioLegend (cat. no. 613402). An antibody against cyclin B1 was purchased from Santa Cruz Biotechnology (cat. no. sc-245). These antibodies were used for western blot analysis. Antibodies against RAD51 (cat. no. 63801) and γH2AX (cat. no. 05-636) respectively purchased from Abcam and Millipore were used for immunofluorescent staining. Alexa Fluor 488 azide was purchased from Thermo Fisher Scientific (cat. no. A10266). Alexa Fluor 555-conjugated goat anti-rabbit IgG (H + L; cat. no. A-21428) was purchased from Thermo Scientific.

The ATR inhibitor AZD6738 (cat. no. S7693), ATM inhibitor AZD0156 (cat. no. S8375) and DNA-PK inhibitor AZD7648 (cat. no. S8843) were purchased from Selleck Chemicals. The CDK4/6 inhibitor Palbociclib (cat. no.: HY-50767) and CHK1 inhibitor Rabusertib (cat. no. HY-14720) were purchased from MedChemExpress (MCE). dTAG^v-1-NEG (NEG; cat. no. 6915) and dTAG^v-1 (cat. no. 6914) were purchased from Tocris Bioscience.

Western Blot Analysis.

293A-RPA2-dTAG or RPE1-RPA2-dTAG cells were seeded in six-well plates and incubated overnight, followed by treatment with either NEG or dTAG^v-1 for the indicated time periods. For synchronization at early S phase, RPE1-RPA2-dTAG cells were treated with 2 mM thymidine for 16 h, washed twice with PBS, and released into thymidine-free medium for 9 h. Cells were then subjected to a second thymidine block (2 mM) for 14 h. Synchronized cells at the G1/S boundary or early S phase were subsequently released for 2 h before treatment with NEG or dTAG^v-1 as specified.

Following treatment, both adherent and floating cells were harvested and lysed in sodium dodecyl sulfate (SDS) gel-loading buffer. Lysates were boiled and stored for subsequent analysis. Protein samples were resolved by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and analyzed by immunoblotting using the indicated antibodies.

Immunofluorescent Staining.

293A-RPA2-dTAG or RPE1-RPA2-dTAG cells were seeded onto coverslips and incubated overnight. Cells were then treated with either NEG or dTAG^v-1 for the indicated time points, followed by fixation in 3% paraformaldehyde. After fixation, cells were permeabilized with 0.5% Triton X-100 in phosphate-buffered saline (PBS) for 15 min and subsequently blocked with 3% bovine serum albumin (BSA) in PBS for 30 min. Cells were incubated overnight at 4 °C with primary antibodies against γH2AX or RAD51, followed by incubation with the appropriate secondary antibodies the next day. Nuclei were counterstained with DAPI, and fluorescence images were acquired using a Nikon microscope.

Flow Cytometry.

293A-RPA2-dTAG or RPE1-RPA2-dTAG cells were treated with 1 μM NEG or dTAG^v-1 for the indicated time points and pulse-labeled with 10 μM EdU for 30 min. Both adherent and floating cells were collected and fixed in prechilled 70% ethanol. Following fixation, cells were permeabilized with 0.5% Triton X-100 in PBS for 15 min and blocked with 3% BSA in PBS for 30 min. EdU incorporation was detected using the Click-iT cell reaction buffer (Invitrogen) according to the manufacturer’s protocol for 30 min. Cells were then washed with PBS and stained with FxCycle (Thermo Fisher Scientific) for 30 min prior to flow cytometry analysis.

CellTiter-Glo Cell Viability Assay.

For short-term cell survival assays, 293A-RPA2-dTAG cells were seeded in 96-well plates and treated with 1 μM NEG or dTAG^v-1, with or without the indicated chemical agents, in sextuplicate. Cell viability was assessed at the specified time points using the CellTiter-Glo Luminescent Cell Viability Assay (Promega) according to the manufacturer’s instructions.

Colony Formation Assay.

RPE1-RPA2-dTAG cells, 293A-RPA2-dTAG cells, and 293A-RPA2-dTAG cells reconstituted with either SFB-RPA2-WT or SFB-RPA2-ST10A were seeded into six-well plates and treated with 1 μM NEG or dTAG^v-1. After 1 wk of incubation, cells were fixed and stained with crystal violet solution (Sigma-Aldrich) to assess colony formation.

Neutral Comet Assay.

The neutral comet assay was performed using the CometAssay Kit (cat. no. 4250-050-K; R&D Systems) according to the manufacturer’s instructions and as previously described (24). Briefly, 293A-RPA2-dTAG or RPE1-RPA2-dTAG cells treated with 1 μM NEG or dTAG^v-1 for the indicated durations were collected. Comet LMAgarose was melted at 95 °C for 5 min, then cooled and maintained at 37 °C for at least 20 min before use. Collected cells were washed once with PBS and diluted to a concentration of 4 × 10⁵ cells/mL in PBS. A 20 μl aliquot of the cell suspension was mixed with 200 μL of prewarmed LMAgarose and 50 μL of this mixture was immediately pipetted onto a CometSlide. Slides were placed flat at 4 °C for 30 min to solidify the agarose.

Following solidification, slides were immersed in lysis solution and incubated overnight at 4 °C. Slides were then rinsed and equilibrated in 1 × neutral electrophoresis buffer (50 mM Tris base and 150 mM sodium acetate) for 30 min before electrophoresis at 21 V for 45 min. After electrophoresis, slides were washed with distilled water, followed by 70% ethanol, and allowed to air dry overnight. The dried slides were stained with SYBR Gold for 30 min, rinsed with water, and dried again for imaging. Comet images were captured using a Nikon microscope at 10× magnification and analyzed using the OpenComet plugin (28), with olive tail moments measured and reported.

Results

RPA2 Depletion Disrupts the RPA Complex and Impairs DNA Replication.

To assess whether RPA-dependent recruitment of ATR/ATRIP is essential for ATR/CHK1 activation, we employed the degradation tag (dTAG) system to induce acute degradation of the RPA complex, a critical component for DNA replication and cell viability. We generated a homozygous C-terminal dTAG knock-in at the endogenous RPA2 locus in 293A cells and confirmed successful targeting by PCR (Fig. 1A).

Fig. 1.

RPA2 depletion disrupts the RPA complex and impairs DNA replication. (A) Schematic design of the donor template used for generating 293A-RPA2-dTAG cells (Left). Forward and reverse primers outside the left and right homology arms were used for PCR validation of 293A-RPA2-dTAG clones (Right). (B) Western blot validation of 293A-RPA2-dTAG cell line treated with 1 µM NEG or 1 µM dTAG^v-1 for the indicated hours. (C) Flow cytometry analyses of EdU intensity using 293A-RPA2-dTAG cells treated with 1 µM NEG or 1 µM dTAG^v-1 for the indicated hours. More than 10,000 cells were collected for analyses. The X and Y axes respectively represent DNA content and EdU intensity. (D) 293A-RPA2-dTAG Fucci (SA) cells and 293A Fucci (SA) cells were treated with 1 µM NEG or 1 µM dTAG^v-1 for the indicated hours, followed by FACS analyses. More than 10,000 cells were collected for analyses. Bars represent the mean ± SEM (n = 3). (E) 293A-RPA2-dTAG cells were treated with 1 µM NEG or 1 µM dTAG^v-1 and collected at the indicated hours, followed by neutral comet assay. More than 100 cells of each group were counted for analyses. Bars represent the mean ± SEM. Representative images were shown. (Scale bar, 100 μm.)

Western blot analysis showed that treatment with dTAG^v-1 led to rapid and near-complete degradation of RPA2 within 1 h compared to NEG-treated controls (Fig. 1B). Notably, RPA2 depletion also significantly reduced RPA1 levels (Fig. 1B), indicating destabilization of the entire RPA complex.

To evaluate the functional impact on DNA replication, we performed EdU pulse labeling following dTAG^v-1 treatment. RPA2-depleted cells exhibited markedly reduced DNA synthesis relative to NEG controls (Fig. 1C). Extended treatment for 24 h resulted in an accumulation of cells with sub-2 N DNA content, consistent with elevated cell death (Fig. 1C).

Since RPA stabilizes ssDNA during both replication initiation and elongation, its loss is expected to disrupt fork stability and progression, potentially leading to widespread replication stress and genomic instability. We first utilized the Fucci (SA) fluorescent biosensor to monitor cell cycle progression. Flow cytometry of 293A cells stably expressing Fucci (SA) revealed no significant changes upon dTAGv-1 treatment (Fig. 1 D, Lower panel). However, dTAGv-1-treated 293A-RPA2-dTAG Fucci (SA) cells accumulated at the late S/G2/M phases (Fig. 1 D, Upper panel), suggesting that RPA2 depletion may not affect replication initiation, but arrest cells probably in S phase.

Given the known role of RPA in protecting ssDNA, we hypothesized that RPA2 loss may lead to formation of replication-associated double-strand breaks (DSBs). Neutral comet assays revealed a progressive increase in comet tail moments in RPA2-dTAG cells treated with dTAG^v-1 for 1 to 24 h (Fig. 1E), indicative of sustained DSB accumulation.

Collectively, these results demonstrate that acute RPA2 depletion destabilizes the RPA complex, disrupts DNA replication, and leads to unresolved replication stress and DNA damage. Thus, 293A-RPA2-dTAG cells provide a robust system for delineating RPA-dependent and RPA-independent mechanisms of ATR/CHK1 pathway activation.

Robust ATR/CHK1 Activation Observed in RPA-Depleted Cells Mediated by RAD9/TOPBP1.

We next examined how RPA depletion influences ATR/CHK1 activation and other DNA damage signaling in 293A-RPA2-dTAG cells. In line with the comet assay results, we observed significant activation of ATM (pATM-S1981) as early as 1 h after dTAG^v-1 treatment, persisting through 2, 4, and 8 h. Activation of DNA-PKcs (pDNA-PKcs-S2056) became evident at later time points (8 to 24 h) (Fig. 2A), consistent with the accumulation of DSBs in RPA2-depleted cells.

Fig. 2.

ATR/CHK1 activation persists in RPA-depleted cells via a RAD9/TOPBP1-dependent, RPA-independent mechanism. (A) Immunoblots of the indicated proteins prepared from 293A-RPA2-dTAG cells treated with 1 µM NEG or 1 µM dTAG^v-1 for the indicated hours. Both floated and adhesive cells were collected for analysis. (B) Immunoblots of the indicated proteins prepared from 293A-RPA2-dTAG cells treated with 1 µM ATM inhibitor (AZD0156), 10 µM ATR inhibitor (AZD6738), or 10 µM DNA-PK inhibitor (AZD7648) for 2 h, followed by treatment with 1 µM NEG or 1 µM dTAG^v-1 for 6 h. (C) Immunoblots of the indicated proteins prepared from 293A-RPA2-dTAG cells treated with DMSO or 10 µM ATR inhibitor (AZD6738) for 2 h, followed by treatment with 1 µM NEG or 1 µM dTAG^v-1 for the indicated hours. (D) Immunoblots of the indicated proteins prepared from 293A-RPA2-dTAG cells treated with 10 µM ATR inhibitor (AZD6738) for 2 h, followed by a combined 2-h treatment with dTAG^v-1, treatment with dTAG^v-1 alone, or treatment with 2 mM HU for 2 or 4 h. Both floated and adhesive cells were collected for analysis. (E) 293A-RPA2-dTAG cells were infected with empty virus (vector) or virus expressing sgRNA1/2 targeting RAD9. Twenty-four hours after infection, cells were treated with 1 µM NEG or 1 µM dTAG^v-1 for 2 or 4 h and then lysed for western bot analyses. (F) 293A-RPA2-dTAG cells were infected with empty virus (vector) or virus expressing sgRNA1/2 targeting TOPBP1. Twenty-four hours after infection, cells were treated with 1 µM NEG or 1 µM dTAG^v-1 for 2 or 4 h and then lysed for western bot analyses.

Surprisingly, we also detected robust phosphorylation of CHK1 at Ser296 and Ser345 in RPA2-depleted cells, which was similar but reduced when compared to their levels induced by hydroxyurea (HU), a positive control for ATR/CHK1 activation (Fig. 2A). Since phosphorylation at Ser317 or Ser345 induces CHK1 conformational changes that promote autophosphorylation of CHK1 at Ser296, we tested whether CHK1 activity was responsible for the observed pCHK1-S296 increase. Indeed, treatment with a CHK1 inhibitor abolished pCHK1-S296 but not pCHK1-S345 (SI Appendix, Fig. S1A), confirming that CHK1 remains active in RPA-depleted cells. These results suggest the existence of an RPA-independent pathway for CHK1 activation.

To determine whether this activation was still dependent on ATR, we examined pCHK1-S296/S345 levels following treatment with inhibitors of ATR, ATM, or DNA-PKcs. ATR inhibition completely suppressed CHK1 phosphorylation, while ATM or DNA-PKcs inhibition had modest or no effect (Fig. 2B). Time-course analysis further confirmed that ATR inhibition reduced pCHK1-S345 levels in dTAG^v-1–treated cells, whereas ATM or DNA-PKcs inhibitors had minimal impact (Fig. 2C and SI Appendix, Fig. S1 B and C).

Given that RPA2 degradation reached undetectable levels within 1 h (SI Appendix, Fig. S2A), in which period the residual amount of RPA was still able to support some replication activity (Fig. 1C), we considered whether transient ATR recruitment might occur before RPA loss. To address this possibility, we pretreated cells with ATR inhibitor for 2 h, followed by the combined dTAG^v-1 treatment for another 2 h to ensure RPA2 degradation. Then the ATR inhibitor and dTAG^v-1 were washed out before NEG or dTAG^v-1 treatment for 1, 2, 4, and 14 h. In the timepoint immediately after the ATR inhibitor and dTAG^v-1 were washed out (0 h), CHK1 phosphorylation at S296/S317/S345 was absent, confirming successful ATR inhibition and RPA2 degradation (SI Appendix, Fig. S2B, lane 4). When these pretreated cells were subsequently exposed to dTAG^v-1 or NEG, CHK1 phosphorylation levels remained equivalent to those in cells treated with dTAG^v-1 alone (SI Appendix, Fig. S2B, lanes 3 vs. lane 5 to 12), indicating that ATR/CHK1 activation does not require prior RPA–ATRIP–ATR complex formation. These findings rule out early ATR recruitment as a confounding factor, reinforcing that ATR/CHK1 activation can occur independently of RPA.

To compare the relative contribution of RPA-dependent vs. RPA-independent ATR/CHK1 activation, we used HU-treated cells as a benchmark. We found that the RPA-independent pathway accounted for a fraction of CHK1 activation compared to HU-induced signaling (Fig. 2D, lane 4 vs. lane 5). Furthermore, pretreatment with ATRi+dTAG^v-1 yielded CHK1 phosphorylation levels nearly identical to dTAG^v-1 alone (Fig. 2D, lane 3 vs. lane 4). Moreover, HU treatment induced robust pCHK1-S296 and pCHK1-S345 levels, which were significantly attenuated by RPA2 depletion (SI Appendix, Fig. S2C). These results suggest that under normal circumstance RPA plays a critical role in ATR/CHK1 activation in response to replication stress.

Finally, we investigated whether the RAD9–HUS1–RAD1–TOPBP1 axis mediates this RPA-independent ATR/CHK1 activation. Knockdown of either RAD9 or TOPBP1 markedly impaired pCHK1-S345 induction in RPA2-depleted cells (Fig. 2 E and F). Together, these findings establish the existence of an RPA-independent ATR/CHK1 activation mechanism that operates through RAD9 and TOPBP1, offering insights into how cells respond to replication stress in the absence of canonical RPA signaling.

RPA-independent Activation of CHK1 also Occurs in RPE-1 Cells.

To further support our findings, we generated a homozygous C-terminal dTAG knock-in at the RPA2 locus in RPE-1 cells, a nontransformed human cell line with a distinct genetic background. Consistent with the results observed in 293A-RPA2-dTAG cells, treatment with dTAG^v-1 led to a rapid and near-complete depletion of RPA2 within 1 h, accompanied by a substantial reduction in RPA1 levels (Fig. 3A), indicating disruption of the RPA complex. EdU pulse labeling confirmed that DNA replication was severely impaired in dTAG^v-1-treated RPE1-RPA2-dTAG cells within 1 h (SI Appendix, Fig. S3A). These cells also exhibited pronounced growth defects relative to NEG-treated controls (Fig. 3B).

Fig. 3.

RPA-independent activation of CHK1 also occurs in RPE-1 cells (A) Western blot validation of RPE1-RPA2 cell line treated with 1 µM NEG or 1 µM dTAG^v-1 for the indicated hours. (B) Clonogenic survival of RPE1-RPA2-dTAG cells was determined in the presence of 1 µM NEG or 1 µM dTAG^v-1. (C) RPE1-RPA2-dTAG cells were treated with 1 µM NEG or 1 µM dTAG^v-1 and fixed at the indicated hours, followed by γH2AX staining. (D) RPE1-RPA2-dTAG cells were treated with 1 µM NEG or 1 µM dTAG^v-1 and collected at the indicated hours, followed by neutral comet assay. Representative images were shown. (Scale bar, 100 μm.) (E) More than 100 cells of each group were counted for analyses using the same cells in (D). Bars represent the mean ± SEM. (F) Immunoblots of the indicated proteins prepared from RPE1-RPA2-dTAG cells treated with 1 µM NEG or 1 µM dTAG^v-1 for the indicated hours.

Immunofluorescence analysis showed that γH2AX foci accumulated progressively in RPE1-RPA2-dTAG cells treated with dTAG^v-1, indicating the build-up of DNA damage over time (Fig. 3C). This was further supported by neutral comet assay results, which revealed increased tail moments after 1 or 2 h of dTAG^v-1 treatment, with damage persisting at 8 and 24 h (Fig. 3 D and E). Moreover, when RPE1-RPA2-dTAG cells were treated with dTAG^v-1 for 4 h followed by ionizing radiation (IR), no RAD51 foci were observed (SI Appendix, Fig. S3B), suggesting that RPA2 depletion impairs HR-mediated DNA repair. Furthermore, we observed significant cell death in RPE1-RPA2-dTAG cells when RPA2 was depleted for 24 or 48 h (SI Appendix, Fig. S3C). These results suggest that RPA2 depletion results in the accumulation of DNA damage, which could not be repaired efficiently and eventually lead to cell death.

We then determined whether RPA2 depletion would affect replication initiation in synchronized cells. The RPE1-RPA2-dTAG cells were synchronized in G1 phase with CDK4/6 inhibitor treatment for 22 h followed by the combined dTAG^v-1 treatment for another 2 h to ensure RPA2 degradation. Cells were then released from G1 arrest and treated with NEG or dTAG^v-1 for 1, 2, 4, 8, or 24 h and labeled with EdU for 30 min prior to flow cytometry analyses (SI Appendix, Fig. S4 A and B). As shown in SI Appendix, Fig. S4B, about 88.7% of cells were synchronized in G1. After release, NEG-treated control cells gradually entered S phase as shown by EdU incorporation, whereas dTAG^v-1-treated cells exhibited deficient EdU incorporation and increased DNA damage signals, including γH2AX, pKAP1, pCHK1-S296, and pCHK1-S345 (SI Appendix, Fig. S4 B and C). These results indicate that cells with RPA2 depletion in G1 could not synthesize DNA. Moreover, depletion of RPA2 renders ssDNA unprotected, destabilizing the replication bubble and increasing nuclease sensitivity, as evidenced by heightened DNA damage signals (SI Appendix, Revised Fig. S4C). Consequently, DNA synthesis could not continue, which is reflected by deficient EdU incorporation (SI Appendix, Revised Fig. S4B). Collectively, cells with RPA depletion in G1 showed abolished EdU incorporation and increased DNA damage signals, which indicate that RPA may not be essential for DNA unwinding and formation of initial replication bubbles, it is likely required and/or essential for subsequent DNA synthesis.

To determine whether the RPA-independent ATR/CHK1 activation observed in 293A cells also occurs in RPE-1 cells, we assessed CHK1 phosphorylation following RPA2 depletion. As shown in Fig. 3F, pCHK1-S296 and pCHK1-S345 levels were upregulated at 1, 4, 8, and 24 h after dTAG^v-1 treatment (Fig. 3F). Interestingly, a transient drop in CHK1 phosphorylation was observed at 2 h (Fig. 3F), a pattern distinct from that seen in 293A-RPA2-dTAG cells.

Collectively, these results confirm that CHK1 can be activated via an RPA-independent pathway in both 293A and RPE-1 human cell lines, further highlighting the robustness and adaptability of the ATR/CHK1 signaling pathway under conditions of replication stress.

RPA-independent ATR/CHK1 Activation Exhibits a Distinct Biphasic Pattern in RPE-1 Cells.

We were puzzled by the biphasic pattern of ATR/CHK1 activation in RPE-1 cells. To investigate whether both phases of CHK1 activation in RPE-1 cells are dependent on ATR, we treated RPE1-RPA2-dTAG cells with an ATR inhibitor prior to dTAG^v-1 administration. As shown in Fig. 4A, ATR inhibition significantly impaired pCHK1-S345 induction compared to DMSO-treated controls, indicating that both waves of CHK1 activation are primarily ATR-dependent. However, a residual level of pCHK1-S345 persisted during the second activation phase, and pCHK1-S296 remained consistently elevated, suggesting potential involvement of additional kinases in CHK1 phosphorylation under ATR-inhibited conditions.

Fig. 4.

RPA-independent ATR/CHK1 activation exhibits a distinct biphasic pattern in RPE-1 cells. (A) Immunoblots of the indicated proteins prepared from RPE1-RPA2-dTAG cells treated with DMSO or 10 µM ATR inhibitor (AZD6738) for 2 h, followed by treatment with 1 µM NEG or 1 µM dTAG^v-1 for the indicated hours. (B) Immunoblots of the indicated proteins prepared from RPE1-RPA2-dTAG cells treated with DMSO or 1 µM ATM inhibitor (AZD0156) for 2 h, followed by treatment with 1 µM NEG or 1 µM dTAG^v-1 for the indicated hours. (C) Immunoblots of the indicated proteins prepared from RPE1-RPA2-dTAG cells treated with DMSO or 10 µM DNA-PK inhibitor (AZD7648) for 2 h, followed by treatment with 1 µM NEG or 1 µM dTAG^v-1 for the indicated hours. (D) RPE1-RPA2-dTAG cells were infected with empty virus (vector) or virus expressing sgRNA1/2 targeting RAD9. Twenty-four hours after infection, cells were treated with 1 µM NEG or 1 µM dTAG^v-1 for 1, 8, or 24 h and then lysed for western bot analyses. (E) RPE1-RPA2-dTAG cells were infected with empty virus (vector) or virus expressing sgRNA1/2 targeting TOPBP1. Twenty-four hours after infection, cells were treated with 1 µM NEG or 1 µM dTAG^v-1 for 1, 8, or 24 h and then lysed for western bot analyses. (F) Immunoblots of the indicated proteins prepared from RPE1-RPA2-dTAG cells treated with 1 µM NEG or 1 µM dTAG^v-1 plus 100 ng/µL nocodazole for the indicated hours. (G) Immunoblots of the indicated proteins prepared from double-thymidine synchronized RPE1-RPA2-dTAG cells. The synchronized cells were released for 2 h before treatment with NEG or dTAG^v-1 for the indicated hours.

To explore this further, we treated RPE1-RPA2-dTAG cells with inhibitors targeting ATM or DNA-PKcs prior to RPA2 degradation. ATM inhibition attenuated pCHK1-S345 levels (Fig. 4B), while DNA-PKcs inhibition markedly suppressed the second wave of pCHK1-S345 phosphorylation (Fig. 4C). In these conditions, pCHK1-S296 levels remained high and followed different time courses, similar to those seen with ATR inhibition alone, suggesting that ATR, ATM, and DNA-PKcs all contribute to CHK1 activation, particularly in the context of RPA depletion.

Additionally, we examined whether the RAD9–HUS1–RAD1–TOPBP1 axis also contributes to the RPA-independent ATR/CHK1 activation in RPE-1 cells. Consistent with the results observed in 293A-RPA2-dTAG cells, knockdown of either RAD9 or TOPBP1 markedly impaired pCHK1-S296/S345 induction in RPA2-depleted RPE-1 cells (Fig. 4 D and E).

The pattern of RPA-independent CHK1 activation differed between 293A and RPE-1 cells. To investigate this difference, we considered the potential influence of cell cycle phase, given that RPA depletion is expected to primarily affect cells in S phase. We hypothesized that the biphasic CHK1 activation observed in RPE-1 cells might reflect asynchronous entry into S phase, resulting in a staggered activation of the ATR/CHK1 pathway across the population.

To test this, we cotreated RPE1-RPA2-dTAG cells with dTAG^v-1 and nocodazole to prevent cells escape mitosis. This treatment did not alter the biphasic pattern of pCHK1-S296/S345 compared to dTAG^v-1 alone (Fig. 4F). Next, we synchronized cells at early mitosis using a thymidine block followed by release and nocodazole treatment. Cells were then released into cell cycle with the addition of dTAG^v-1. CHK1 phosphorylation only became obvious after cells reentered S phase (SI Appendix, Fig. S5), indicating that CHK1 phosphorylation mainly occurs in S phase cells. We then synchronized cells at early S phase using a double thymidine block followed by a 2-h release. Again, biphasic CHK1 activation was observed in dTAG^v-1 treated cells (Fig. 4G), similar to the pattern in unsynchronized cells (Fig. 3F).

These findings suggest that both waves of ATR/CHK1 activation occur within S phase and are modulated by the replication status of the cell population. Moreover, CHK1 activation under RPA-depleted conditions can involve multiple kinases, including ATR, ATM, and DNA-PKcs, highlighting the robustness and redundancy of the DDR in response to replication stress.

Abrogation of RPA2 N-terminal Phosphorylation Does Not Affect Cell Survival, RAD51 Foci Formation, or Sensitivity to DNA Damage.

Phosphorylation of the RPA2 N-terminal region is one of the most extensively studied PTMs of the RPA complex. However, its functional relevance in ATR/CHK1 signaling and HR-mediated DNA repair remains controversial, particularly in mammalian systems. A key challenge has been the inability to fully eliminate endogenous WT RPA2. The RPA2-dTAG system provides a powerful approach to overcome this limitation.

To assess the functional role of RPA2 N-terminal phosphorylation, we used 293A-RPA2-dTAG cells reconstituted with either WT RPA2 (SFB-RPA2-WT) or a mutant form in which 10 serine/threonine residues in the N terminus were mutated to alanine (ST10A) (Fig. 5A). As shown in Fig. 5B, abrogation of RPA2 N-terminal phosphorylation (via SFB-RPA2-ST10A expression and dTAG^v-1 treatment) had no detectable effect on cell proliferation. Growth curves were comparable to those of NEG-treated controls and WT-reconstituted cells treated with dTAG^v-1. Similarly, deletion of the entire N-terminal domain (RPA2-Δ1) had no effect on survival in 293A- and RPE1-RPA2-dTAG cells, while deletion of the OB-fold domain, critical for trimerization and ssDNA binding, abrogated cell viability (SI Appendix, Fig. S6 A–C). Deletion of the winged helix (WH) domain (RPA2-Δ4) had a modest effect (SI Appendix, Fig. S6 A–C).

Fig. 5.

Abrogation of RPA2 N-terminal phosphorylation does not affect cell survival, RAD51 foci formation, or sensitivity to DNA damage. (A) Schematic of RPA2 and its potential N-terminal phosphorylation sites (serine/threonine, S/T). The ST10A mutant contains S/T-to-alanine (A) to block phosphorylation. (B) Cell proliferation was measured by CellTiter-Glo luminescence assay (Top) and clonogenic assay (Bottom) in 293A-RPA2-dTAG cells with or without reconstitution of SFB-RPA2 or SFB-RPA2-ST10A. (C) Immunoblots of the indicated proteins prepared from 293A-RPA2-dTAG cells with or without reconstitution of SFB-RPA2 or SFB-RPA2-ST10A, following treatment with 1 µM NEG or 1 µM dTAG^v-1 for the indicated hours. (D) 293A-RPA2-dTAG cells with or without reconstitution of SFB-RPA2 or SFB-RPA2-ST10A, were treated with 1 µM NEG or 1 µM dTAG^v-1 for 4 h, followed by exposure to 10 Gy IR and recovery for 16 h. Cells were then fixed and stained for RAD51. Representative images were shown. (E) Cell sensitivity to DNA damage was assessed in 293A-RPA2-dTAG cells, with or without reconstitution of SFB-RPA2 or SFB-RPA2-ST10A, using the CellTiter-Glo luminescence assay.

RPA plays a key role in sensing replication stress and initiating DDR signaling. To determine whether N-terminal phosphorylation of RPA2 contributes to this function, we assessed DDR markers. RPA2 depletion led to robust induction of pCHK1-S296/S345, γH2AX, and pKAP1, effects that were fully rescued by reexpression of either SFB-RPA2-WT or SFB-RPA2-ST10A (Fig. 5C).

Similarly, reconstitution with either form restored IR-induced RAD51 foci formation and RPA complex integrity (Fig. 5 C and D). Furthermore, cells reconstituted with RPA2-WT or ST10A exhibited comparable sensitivity to DNA-damaging agents (Fig. 5E).

Collectively, these results demonstrate that phosphorylation of the RPA2 N terminus is dispensable for cell survival, DDR signaling, RAD51 foci formation, and sensitivity to DNA damage, at least under the experimental conditions tested.

Discussion

This study provides compelling evidence for a previously unrecognized, RPA-independent pathway that contributes to ATR/CHK1 activation. Canonical models have established that recruitment of the ATR–ATRIP complex to sites of DNA damage relies on its physical interaction with RPA1 within RPA-coated ssDNA, a mechanism we define as the RPA-dependent pathway (18). Based on our findings, we propose that ATR/CHK1 activation can occur via both RPA-dependent and RPA-independent mechanisms. Importantly, the parallel activation of ATR/CHK1 ensures robust response to replication stress, since in RPA-depleted cells, the ATR/CHK1 activation is significantly reduced when compared to normal circumstance, i.e. in the presence of RPA. Notably, these pathways are not mutually exclusive but may act synergistically to ensure an efficient response to replication stress and safeguard genome stability (SI Appendix, Fig. S7).

Studying the RPA-independent pathway has historically been challenging due to the essential role of the RPA complex in DNA replication and cell viability. By using the dTAG system to acutely degrade RPA2, we overcame this limitation and revealed that ATR/CHK1 activation can proceed in the absence of RPA, via a mechanism dependent on the RAD9–HUS1–RAD1 complex and TOPBP1. Notably, this mechanism was observed in both 293A and RPE-1 cells, supporting its generalizability. Nonetheless, important questions remain regarding how the RAD17–RFC clamp loader and the RAD9–HUS1–RAD1 complex are recruited to ssDNA–dsDNA junctions at stalled replication forks, in the absence of RPA.

One possibility is that RPA degradation results in replisome collapse, which then leads to the mislocalized and uncontrolled ATR/CHK1 activation by TOPBP1, since ATR and TOPBP1 associate with replisome. Previous studies have shown that ATR activity at the replisome plays a physiological role in limiting origin firing, which is localized and distinct from the global DDR (29, 30), indicating that such localized ATR activity may not require RPA and contribute to RPA-independent ATR/CHK1 activation. It remains unclear how replication collapse may lead to ATR/CHK1 activation, which will be explored in future studies.

Another possibility is that the dTAG system, albeit powerful, would not be able to eliminate 100% of RPA protein in cells. Therefore, we cannot exclude that there may still be some residual RPA, which is sufficient for ATR recruitment and promotes ATR/CHK1 activation.

Although RPA2-depleted cells accumulate unprotected ssDNA, blunt DSBs, as indicated by pDNA-PKcs-S2056, were not detected until approximately 8 h postdepletion, much later than theoretically expected. This delayed onset of DSB formation parallels findings in HU-induced replication stress, where fork collapse and DSB formation, mediated by SLX4 and MUS81, occur several hours after stalling (31–35). One plausible explanation is that stalled replication forks are initially stabilized by protective structures or proteins that shield ssDNA regions. When repair fails, these forks eventually collapse, producing DSBs. Identifying the molecular components that mediate fork protection and delay DSB formation will be a valuable avenue for future research.

The rapid degradation enabled by the dTAG system also allowed us to assess the functional domains of RPA2 and the role of its N-terminal phosphorylation. As expected, the OB fold domain, critical for trimerization and ssDNA binding, was essential for cell viability (36). Interestingly, deletion of the winged-helix (WH) domain, known to mediate protein–protein interactions (37), had only a modest effect on cell survival in both 293A and RPE-1 cells. Determining which RPA2-interacting proteins are crucial for the basal proliferative functions of the RPA complex will provide further mechanistic insight.

Unexpectedly, we found that mutation of all N-terminal serine/threonine residues (ST10A) in RPA2 had no effect on cell viability, RAD51 foci formation, or sensitivity to DNA-damaging agents. These results challenge the presumed importance of RPA2 N-terminal phosphorylation and raise questions about its physiological relevance. However, the reconstituted SFB-RPA2 WT and ST10A mutant proteins were expressed at higher levels than the endogenously tagged RPA2, raising the possibility that high levels of ST10A may override subtle defects of this mutant. Nevertheless, our dTAG-based system offers a powerful platform for exploring patient-derived hypomorphic RPA2 mutations, which may further illuminate regulatory mechanisms of the RPA complex in human disease.

Altogether, this study not only broadens our understanding of ATR/CHK1 activation and RPA complex function, but also highlights the utility of acute degradation systems in dissecting essential cellular processes.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All data are included in the article and/or SI Appendix.