WASp modulates RPA function on single-stranded DNA in response to replication stress and DNA damage

ARTICLE WASp modulates RPA function on single-stranded DNA in response to replication stress and DNA damage Seong-Su Han1,2,6, Kuo-Kuang Wen1,2,6, María L. García-Rubio 3, Marc S. Wold 4, Andrés Aguilera 3, Wojciech Niedzwiedz 5✉& Yatin M. Vyas 1,2✉ Perturbation in the replication-stress response (RSR) and DNA-damage response (DDR) causes genomic instability. Genomic instability occurs in Wiskott-Aldrich syndrome (WAS), a primary immunodeficiency disorder, yet the mechanism remains largely uncharacterized. Replication protein A (RPA), a single-strand DNA (ssDNA) binding protein, has key roles in the RSR and DDR. Here we show that human WAS-protein (WASp) modulates RPA functions at perturbed replication forks (RFs). Following genotoxic insult, WASp accumulates at RFs, associates with RPA, and promotes RPA:ssDNA complexation. WASp deficiency in human lymphocytes destabilizes RPA:ssDNA-complexes, impairs accumulation of RPA, ATR, ETAA1, and TOPBP1 at genotoxin-perturbed RFs, decreases CHK1 activation, and provokes global RF dysfunction. las17 (yeast WAS-homolog)-deficient S. cerevisiae also show decreased ScRPA accumulation at perturbed RFs, impaired DNA recombination, and increased frequency of DNA double-strand break (DSB)-induced single-strand annealing (SSA). Consequently, WASp (or Las17)-deficient cells show increased frequency of DSBs upon genotoxic insult. Our study reveals an evolutionarily conserved, essential role of WASp in the DNA stress-resolution pathway, such that WASp deficiency provokes RPA dysfunction-coupled genomic instability. https://doi.org/10.1038/s41467-022-31415-z OPEN 1Department of Pediatrics, PennState College of Medicine, PennState Health Children’s Hospital, Hershey, PA 17033, USA. 2Division of Pediatric HematologyOncology, University of Iowa Stead Family Children’s Hospital, Iowa City, IA 52242, USA. 3Andalusian Center of Molecular Biology and Regenerative Medicine CABIMER, Department of Genome Biology, University of Seville-CSIC-University Pablo de Olavide, Seville, Spain. 4Department of Biochemistry and Molecular Biology, University of Iowa Carver College of Medicine, Iowa City, IA, USA. 5The Institute of Cancer Research, 237 Fulham Road, London SW3 6JB, UK. 6 These authors contributed equally: Seong-Su Han, Kuo-Kuang Wen. ✉email: [email protected];[email protected] NATURE COMMUNICATIONS | (2022) 13:3743 | https://doi.org/10.1038/s41467-022-31415-z | www.nature.com/naturecommunications 1 1234567890():,; Wiskott-Aldrich syndrome (WAS) is an inborn error of immunity (IEI) caused by deficiency of WASp1. WAS lymphocytes show some combination of actincytoskeletal defect, gene transcription defect, and genome instability2–8, manifesting clinically in immunodeficiency, atopy/ autoimmunity, and lymphoid malignancy9. WASp supports ARP2/3-mediated actin-polymerization in the cytoplasm and RNA Pol II-mediated transcription in the nucleus3,10. Recent evidence links WASp to preserving genome integrity by modulating the cellular load of DSBs in human T and B lymphocytes, cells critical for adaptive immunity. WASp does so by suppressing the ectopic accumulation of pathologic R loops (formed by DNARNA hybrid and displaced ssDNA) that cause DSBs11, and by enabling the early step of transporting DSB-ends for repair by the homology-directed repair (HDR) pathway12. Therefore, WASpdeficient lymphocytes manifest both the afferent arm (increased DSBs) and efferent arm (decreased DSB-repair efficiency) of the genome instability circuit. Furthermore, WASp deficiency undermines the nucleus-to-Golgi signaling elicited by DNA damage that is essential for cell survival after genotoxin-induced damage13. Altogether, these defects render WASp-deficient cells ill-equipped to manage genomic stress, endogenous or exogeneous. Thus, the collective evidence has greatly expanded WASp role from a cytoplasmic regulator of actin-cytoskeleton to a nuclear regulator of genome integrity. A key factor tasked to preserve genome integrity is Replication protein A (RPA), a heterotrimeric protein containing RPA1, RPA2, RPA3 subunits, involved in DNA repair, replication, and recombination14. RPA accumulates at ssDNA sites generated from DNA-processing or stalled RFs. Although RPA is a highaffinity ssDNA-binding protein [Kd: ~10−10 M], it is readily replaced by proteins that bind ssDNA with lower-affinity than RPA, during the temporal progression of the protein “handoff” reactions involved in fork protection and DNA repair15. This indicates that proteins and/or signals that can dynamically enable/disable RPA:ssDNA interaction must exist. As such, Saccharomyces cerevisiae (Sc) RPA employs Rtt105 (Protein-Regulator-of-Ty1-transposition-105) to optimize its binding with ssDNA16,17. Similarly, Xenopus RPA-interacting protein (XRIP) facilitates RPA nuclear localization18. These data suggest that RPA employs one or more “chaperons”to enable its interaction with ssDNA, and that this requirement is likely evolutionarily conserved. Yet, to date, how RPA-binding to ssDNA is modulated in humans is ill-understood. Here, we show that WASp is a critical factor that directly binds RPA and modulates its ssDNAbinding activity. Furthermore, we provide evidence that WASp:RPA alliance is required for maintaining genome integrity by influencing the RSR and the DDR in human lymphocytes and S. cerevisiae. This study establishes that RPA activity/function in higher and lower eukaryotes depends on WASp (or ScLas17) to efficiently manage DNA stress; and in the absence of WASp, RPA dysregulation triggers replication stress, DNA damage, and cellular/organismal dysfunction. The study, therefore, identifies WASp as an essential part of both normal DNA replication and DNA stress-resolution pathway, and unveils the sources of impaired RSR-linked genome instability in WASp-deficient lymphocytes. Results Genotoxins promote WASp accumulation at stressed RFs and its co-association with the markers of stressed DNA.To investigate a possible role of WASp in the RSR and DDR, we employed hydroxyurea (HU: 1 mM, 2 h; causes reduction of RFvelocity and RF-stalling) and camptothecin (CPT: 2 µM or 5 µM, 2 h; causes ssDNA nick, which after replication would become a single-ended DSB) as genotoxins. Using proximity ligation assay (PLA), we show in human T cells and B cells that these genotoxins induce WASp association in vivo with γH2A.X and RPA2(pSer33) (Fig. 1a), both known to co-accumulate at stressed RFs19,20. In contrast, WASp does not associate spontaneously (i.e., without damage) with these DNA stress markers. We next performed the quantitative in situ analysis of protein interactions at RFs (SIRF)21, a technique that combines principles of iPOND (isolation of proteins at nascent DNA)22 and PLA to visualize WASp enrichment at RFs at the single-cell resolution. We first verified in human T cells that PCNA (proliferating cell nuclear antigen) and GINS-CDC45 (replisome component) are constitutively present at unperturbed RFs, as SIRF positive controls (Fig. 1b). Like PCNA and CDC45, WASp is also constitutively present at unperturbed RFs, albeit at a lower frequency, in T and B cells (Fig. 1b), suggesting a role for WASp in normal DNA replication. However, following HUor CPT-treatment, WASp enrichment at perturbed RFs is increased significantly in the WT T cells and B cells, reported by SIRF (Fig. 1b). Because ssDNA stretches generated at stressed RF could be long, and thus some of the RPA on ssDNA may be far from EdUlabeled DNA, which could potentially underestimate SIRF foci data, we next verified the SIRF data by iPOND/Western blot. This showed that the enrichment of WASp and RPA2 at perturbed RFs is significantly increased over the steady-state levels following HU-induced RF perturbation, even at an early time-point of 30 min post-HU (Fig. 1c). Similarly, by Western blot we show that the relative abundance of WASp expression in the nucleus of human B cells also increases upon HU-stress relative to unstressed control (Supplementary Fig. S1). The collective findings suggest that WASp is part of the molecular apparatus that regulates both normal DNA replication and the RSR. WASp directly binds RPA. Since RPA is a central component of the RSR and DDR apparatus, and because WASp and RPA coaccumulate at perturbed RFs (Fig. 1a-c), we asked if WASp and RPA interact directly. Using ELISA-based protein-protein binding assays, we show that native purified human WASp directly binds native purified heterotrimeric human RPA protein (RPA1, 2, 3; aka, RPA70, RPA32, RPA14) in vitro, in a dose-dependent manner (Fig. 2a). WASp binds in vitro to both the RPA heterotrimer-ssDNA mixture and RPA heterotrimer alone (Fig. 2a), but does not bind to ssDNA, dsDNA, or ssRNA alone (Supplementary Fig. S2a), denoting that WASp and RPA can interact directly. Moreover, purified human WASp does not bind purified human MAX (Myc-associated factor X), Bovine serum albumin (BSA), or Saccharomyces ScRpa protein (Fig. 2a), confirming specificity of WASp:RPA physical interaction. These in vitro findings align with the in-situ proximity of WASp and RPA2 (likely <10 nm distance from each other) in intact T and B cells by PLA, following HUor CPT-induced DNA stress (Fig. 1a). Together, these results demonstrate that WASp is an RPA-interacting protein, in vitro and in vivo. WASp:RPA interaction is mediated by RPA1-binding motif in WASp. We next sought to identify the WASp-domain(s) involved in RPA interaction. Because RPA1-interacting proteins express a consensus motif: D-ϕ-x-ϕ-D-D-ϕ-x-D-D (D: Aspartic acid; ϕ: hydrophobic or charged side-chain amino acid; x: any amino acid)23,24 (Fig. 2b), we inspected the primary sequence of human WASp and found a putative RPA1-binding motif (RBM1) (493D-E-D-E-D-D-E-W-D-D502), located in the 3’-end acidic (A)-region of WASp’s VCA-domain, which we show is evolutionarily-conserved down to yeast, and resembles the RBM1-consensus in ATRIP, ETAA1, and RAD9A, other known ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31415-z 2NATURE COMMUNICATIONS | (2022) 13:3743 | https://doi.org/10.1038/s41467-022-31415-z | www.nature.com/naturecommunications RPA1-binding proteins (Fig. 2b). To test the functionality of RBM1, we generated WASp-mutant lacking aa:493-502 (ΔRBM1*WASp) (Supplementary Fig. S3a). We purified WT*WASp and ΔRBM1*WASp proteins, and by ELISA method show that the in vitro binding of ΔRBM1*WASp to purified heterotrimeric-RPA is dramatically reduced relative to WT*WASp (Fig. 2c). Notably, ΔRBM1*WASp mutant can still bind purified Arp2 protein (Fig. 2c), the latter previously shown to bind VCA-domain with higher affinity than Arp325, thus providing subdomain-delimited separation-of-function for actinpolymerization versus RPA-activity. Next, we expressed GFPtagged ΔRBM1*WASp-mutant or WT*WASp into WASpFig. 1 WASp is a DNA stress-response protein. a Proximity ligation assay (PLA). Confocal immunofluorescence (IF) microscopy showing a representative set of collapsed composite IF images from a z-stack of ~30-40 images acquired per optical field. Shown are the PLA signals between the indicated interacting proteins induced by the indicated genotoxins or no damage control in human T and B cells, along with their corresponding PLA statistics in boxand-whisker plots (whiskers @10–90%, horizontal bar denotes median, “+”denotes mean). In both panels, unpaired, two-tailed, Mann-Whitney nonparametric pvalues ****<0.0001, n=150 cells/condition. Scale bar: 3 μm. bIn situ analysis of protein interactions at replication forks (SIRF). Right, Representative z-stack collapsed composite confocal IF images showing SIRF signals between 5-ethynyl-2’-deoxyuridine (EdU)-labeled nascent DNA and the indicated protein in steady-state (no damage) or after HUor CPT-induced DNA perturbation, along with their corresponding statistics (n=100 cells). Left, representative SIRF images of positive and negative controls (ctrl), under replicative stress or no stress/damage. Statistical details as per a. Scale bar: 2μm. cIsolation of proteins on nascent DNA (iPOND). Left, schematic of nascent DNA labeling, in which red line denotes DNA labelled with EdU, followed by black line denoting chase into media containing either hydroxyurea (replication stressor) or thymidine (control for true replication proteins that will not enrich in this sample due to EdU displacement). Middle, shows Western blots of input and iPOND purified proteins under indicated conditions, including no EdU sample (no-click negative control). Right, bar graphs depict gel densitometric quantitation of iPOND/Western blot signals normalized to their respective inputs. n=3, +SEM. Mann-Whitney unpaired two-tailed nonparametric p-value *<0.01. Source data are provided as a Source Data file. NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31415-z ARTICLE NATURE COMMUNICATIONS | (2022) 13:3743 | https://doi.org/10.1038/s41467-022-31415-z | www.nature.com/naturecommunications 3 deficient (WKO) T cells (Supplementary Fig. S3b) and show by the PLA method in the FACS-enriched GFP-expressing T cells that the in vivo binding of transfected ΔRBM1*WASp with endogenous RPA1 is significantly lower relative to that of transfected WT*WASp after 2 h of CPT treatment (Fig. 2d). In contrast, transfected ΔRBM1*WASp can still bind endogenous Arp2 in the cytoplasm and nucleus (PLA signals captured in both subcompartments; cytoplasmic PLA signals appear to be concentrated in the region typically occupied by Golgi and/or microtubule-organizing center) (Fig. 2d). Together, these results indicate that RBM1 primarily mediates WASp interaction with RPA, in vitro and in vivo. WASp enhances the ssDNA-binding activity of RPA. To elucidate a functional relevance of the WASp:RPA association, we investigated how WASp modulates RPA’s innate function of binding ssDNA. In EMSA-binding assays, we show that adding ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31415-z 4NATURE COMMUNICATIONS | (2022) 13:3743 | https://doi.org/10.1038/s41467-022-31415-z | www.nature.com/naturecommunications increasing concentrations of purified WASp to a fixed mixture of ssDNA oligo-probe (61-nt) plus either purified-RPA1 or heterotrimeric-RPA protein, results in an augmented RPA:ssDNA complexation (Fig. 2e). Maximum RPA1:ssDNA complexation is observed with equal molar concentrations of WASp, i.e., 25 nM (Fig. 2e). In contrast, addition of purified FANCA, XRCC2, DNMT1, MYC, NF-κB1, STAT1, or BSA, does not change the binding affinity of RPA1 to ssDNA (Supplementary Fig. S2b), denoting unique effect of WASp on RPA:ssDNA interaction. These results suggest that a direct interaction between WASp and RPA is modulating RPA interaction with ssDNA, either by increasing affinity or promoting the stability of the RPA:ssDNA-complex. To further refine this finding, we tested WASp modulation of RPA1-binding to multiple other DNA conformations. Because RPA1 binds ssDNA with a preference for 3’-protrusions (3’- ended tail, 3’-ended flap, 3’-splayed arm) over 5’-protrusions26,27, we first verified these RPA1-binding preferences by EMSA. We show that RPA1 binding to the 3’-ended ssDNA protrusions of 30-nt (considered both optimum and stable ssDNA-binding length for hRPA)27 is also increased by WASp in a dosedependent manner (Fig. 2f). Notably, WASp increases binding of RPA1 to DNA-structures containing ssDNA-protrusion even at a low concentration (25 nM RPA1), which otherwise show a modest binding on its own, i.e., without added WASp (Fig. 2f). These data suggest that WASp improves the efficiency of RPA1 binding to ssDNA. In contrast, WASp does not ectopically enable RPA binding to nucleic-acid structures that normally do not efficiently bind RPA, e.g., ssRNA or DNA Holliday junction26,27(Supplementary Fig. S2c). Together, the data demonstrate specificity of WASp effect on the intrinsic activity of RPA to bind ssDNA. Finally, since RPA-binding to ssDNA is dynamic, involving cycles of association-dissociation-reassociation of RPA-subunits to ssDNA28–30, we tested if the effect of WASp on RPA involves modulating the DNA binding-activity of individual RPAsubunits. To this end, we used purified RPA-subunits containing DNA-binding domain (DBD) DBD-F/A/B (RPA1), DBD-A/B (RPA1), and DBD-D/wh/E (RPA2/RPA3), (Supplementary Fig. S2d), which we had previously generated and characterized29–31. In EMSA-binding assays we show that purified WASp increases the ssDNA-binding activity of RPA1subunits DBD-F/A/B and DBD-A/B (Supplementary Fig. S2e), considered high-affinity DNA-binding domains (DBDs)29–31.In contrast, WASp does not ectopically induce DNA-binding of RPA2/RPA3-subunits DBD-D/wh/E (Supplementary Fig. S2e), considered the trimerization core that intrinsically has lowaffinity for binding ssDNA. This data suggests that WASp likely participates in optimizing the interaction of RPA’s high-affinity DBDs with the available binding sites on ssDNA, i.e., by modulating RPA’s conformational state that favors more OBfold domains to bind ssDNA32. Together, our results indicate that WASp directly enhances the binding and/or stabilization of RPA to multiple ssDNA conformations. WASp-deficiency disrupts genotoxin-induced RPA:ssDNAcomplex formation. To investigate the role of WASp in RPAdependent RSR and DDR, we tested how WASp-deficiency resulting from patient-derived mutation in B cells (WAS03) or CRISPR/Cas9-mediated WASp depletion in T cells (ND1-WKO) (Supplementary Fig. S2f, WASp-expression profile) influences RPA activity at perturbed RFs. By SIRF, we show that the enrichment of endogenous RPA2(pSer33) at HUand CPTperturbed RFs (4 h post-treatment) is significantly increased relative to unperturbed RFs in WT T cells (Fig. 3a), which is consistent with an essential role of RPA in the RSR and DDR signaling. In contrast, the enrichment of RPA2(pSer33) is significantly decreased in WKO T cells relative to WT control, this despite a significant increase in γH2A.X enrichment at perturbed RFs relative to unperturbed RFs in both WT and WKO T cells (Fig. 3a). Notably, expression of total cellular RPA1 protein as well as the nuclear localization of RPA1 in WKO T cells is comparable with WT T cells in the steady-state in vitro culture condition (Supplementary Figs. S2g, S2h), which rules out the possibility that the decreased occupancy of RPA on stressed RFs is due to an overall reduction in the amount of RPA present in WASp-deficient cells. Similarly, a significant reduction in the SIRF-enrichment of RPA at perturbed RFs is observed also in WAS03 B cells relative to normal B cells (Fig. 3a). We next verified the SIRF data by iPOND/Western blot. This showed that the enrichment of endogenous RPA2 at HUperturbed RFs is significantly lower in WKO T cells relative to WT (Fig. 3b). Furthermore, in EMSA-binding assays we show Fig. 2 WASp directly binds RPA and enables RPA binding to ssDNA. a, ELISA. In vitro protein-protein interaction monitored by ELISA for the indicated purified proteins at the indicated concentrations. 1st protein is coated onto the plate at a fixed concentration; 2nd protein is added at the indicated increasing concentrations. hRPA1-3, human heterotrimeric complex of RPA1, 2, 3; hMAX, human Myc-associated factor X; BSA, bovine serum albumin; ScRpa, Saccharomyces Rpa; WT*WASp, wild-type WASp. The displayed data are mean+SEM, n=5 independent assays. The intrinsic property of WASp molecules to spontaneously oligomerize (via VCA:VCA-domain interaction) served as +ve control, and is shown for the highest concentration (1 μg) of WT*WASp protein. bSchematic of the multi-modular domain structure of human WASp showing WH1-domain, Basic-domain containing the NLS (B), GTPase-binding domain (GBD), Polyproline-domain (Pro), VCA-domain (WH2, C, A subdomains), in which the location of evolutionarily conserved RPA1binding motif (RBM1) within the VCA-domain is shown. WH2 (aka, V region) binds monomeric G-actin, C and A regions bind Arp2/3-complex. Aminoacid alignments of human WASp’s RBM1 with those of other proteins and species are shown, in which residues that are conserved (in evolution) and common (with other RPA1-binding proteins) are highlighted in blue. The prototypical RBM1-consensus is shown at the top, D: Aspartic acid; ϕ: hydrophobic or charged side-chain amino acid; x: any amino acid. cELISA. Shown is the binding efficiency of purified heterotrimeric RPA1-3 protein (on left) or purified Arp proteins (on right) with either purified Wt*WASp or RBM1-deleted WASp mutant (ΔRBM1*WASp) at the indicated concentrations. Physical interactions of WT*WASp:WT*WASp and mutant ΔRBM1*WASp:ΔRBM1*WASp (from spontaneous oligomerization of their respective VCA-domains) served as +ve controls. n =3 independent assays, mean+SEM. p-value: ****<0.0001. ns, nonsignificant by Mann–Whitney unpaired two-tailed nonparametric. dProximity ligation assay. Z-stack collapsed composite confocal IF images showing PLA signals (dots) between the indicated protein pairs after transfecting WASp-deficient human T cells with GFP-tagged Wt*WASp or ΔRBM1*WASp mutant after CPT-induced damage. The images/data are for the GFP+cells enriched by FACS-sorting. Box plots are from n=150 cells, Mann-Whitney unpaired two-tailed nonparametric p-value: ****<0.0001; *=0.01; ns, nonsignificant. DAPI (in blue) demarcates the nucleus. Scale bar: 4 μm. EMSA. Assays performed using the indicated purified proteins at the indicated concentrations (nM) against fixed concentration of ssDNA (panel e) or other 3’-ended DNA structures (panel f). Data is representative of at least 3 replicates per condition. Red arrows indicate antibody super-shifted RPA band; other arrows indicate location of non-shifted RPA bands. Other related EMSA results are shown in Supplementary Fig. S2. NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31415-z ARTICLE NATURE COMMUNICATIONS | (2022) 13:3743 | https://doi.org/10.1038/s41467-022-31415-z | www.nature.com/naturecommunications 5 that the ssDNA-binding activity of endogenous heterotrimericRPA protein in the nuclear lysates of CPT-treated (5 µM; 2 h) WKO T-cells and patient-derived WAS03 B-cells is decreased relative to WT cells (Fig. 3c). Finally, we show by PLA that the coassociation of RPA2(pSer33) with γH2A.X in situ is also impaired in HU-treated WASp-deficient T-cells compared to WT T-cells (Fig. 3d). Together, the data suggest that the targeting and/or binding of endogenous RPA to the sites of DNA-damage or replication-stress is compromised in the absence of WASp. WASp-deficiency impairs activation of ATR/CHK1-signaling by disrupting TOPBP1 and ETAA1 enrichment at perturbed RFs. Since ATR recruitment to RPA-coated ssDNA is required ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31415-z 6NATURE COMMUNICATIONS | (2022) 13:3743 | https://doi.org/10.1038/s41467-022-31415-z | www.nature.com/naturecommunications for the upregulation of ATR-kinase activity and phosphorylation of its substrate CHK133,34, we tested if the observed RPA:ssDNA association defect is sufficient to subvert ATR signaling at perturbed RFs in WASp-deficient cells. By SIRF, we show that ATR enrichment is increased at genotoxin-perturbed RFs compared to unperturbed RFs, in both WT (normal) T cells and B cells (Fig. 4a). In contrast, ATR enrichment at perturbed RFs is significantly decreased in WASp-deficient T and B cells (Fig. 4a). Moreover, enrichment of both ETAA1 and TOPBP1, the 2 canonical ATR-kinase activators, at perturbed RFs is also significantly reduced in WASp-deficient T cells relative to WT (Fig. 4a). To directly test how WASp influences ATR:ETAA1 and ATR:TOPBP1 co-associations under replication-stress or DNAdamage, we performed PLA experiments. These show that HU (low or high dose)-induced co-association (reported by in situ proximity) of ATR with both ETAA1 and TOPBP1 was significantly increased in WT T cells relative to WASp-deficient T cells. Under CPT-condition, however, ATR:TOPBP1 coassociation was somewhat increased in WKO T cells relative to WT T cells, whereas that of ATR:ETAA1 was comparable (Fig. 4b). Finally, since dimerization of the TOPBP1 and ETAA1 oligomeric-complexes is required for optimal ATR function35,36, we next tested how RF-resident defects of both TOPBP1 and ETAA1 influences ATR function of activating CHK1. Western blot analyses show that although genotoxins increase CHK1(pSer345) levels in both WT and WKO T cells, the magnitude of this increase normalized to total CHK1 is significantly lower in WKO T cells relative to WT T cells, this despite a comparable (or even higher) degree of DNA damage/ replicative-stress (reported by γH2A.X) (Fig. 4c). CHK1(pSer345) activation defect was also observed in WAS patient-derived B cells (WAS03) subjected to HU-mediated replication stress (Fig. 4d). To further substantiate the necessity of RPA1:WASp interaction in the activation of ATR-mediated checkpoint, we show that WAS-null T cells re-expressing the WASp-mutant selectively lacking the RPA1-binding motif (ΔRBM1*WASp) also shows reduced CHK1(pSer345) activation relative to cells reexpressing WT*WASp control (Fig. 4e). We conclude that disruption of RPA:ssDNA complexation in WASp-deficient cells impairs ATR-ETAA1-TOPBP1 signaling and global checkpoint activation during the RSR and DDR. WASp-deficiency causes global RF dysfunction during replicative stress. To understand how WASp deficiency-linked RPA-ATR-CHK1 signaling defect sensitizes T and B cells to replication stress, we conducted DNA fiber assays by labelling cells with sequential pulses of iododeoxyuridine (IdU) and chlorodeoxyuridine (CldU) in the presence or absence of HU to determine RF events (Fig. 5a-d). First, unperturbed (no added genotoxins) WT and WASpdeficient T cells showed similar RF progression speed, with an average RF-velocity of ~0.6 kb/min for both WT and WASpdeficient (WKO) T cells (Fig. 5a). In contrast, HU-perturbed WKO T cells showed a significantly lower mean RF-velocity of ~0.02 kb/min compared to WT T cells, which showed mean RFvelocity of ~0.06 kb/min (Fig. 5b). WASp-deficient B cells (WAS03) also showed increased RF stalling relative to WT B cells following HU treatment (Supplementary Fig. S4). Second, WASp-deficiency also affects replication resumption after release from HU, as monitored by RF restart (Fig. 5c). In WT T cells, ~85% of RFs have restarted at 30 mins post-HU removal (i.e., 2nd CIdU-track display at least 50% length relative to 1st IdU-track) with only 15% failing to restart (i.e., absent 2nd CIdU-track) (Fig. 5c). By contrast, in WKO T cells, ~60% of RFs fail to restart (i.e., 1st IdU-tracks only, or IdU-track followed by a dot-like 2nd CIdU signal) (Fig. 5c). Impaired replication resumption was seen also in WAS03 B cells after lifting HU stress (Supplementary Fig. S4). Third, since impaired RF restart could occur from degradation of stalled RFs and ATR defect is known to provoke RF degradation37,38, we tested if WASp is involved in protecting nascent DNA from degradation. To this end, we sequentiallylabeled DNA with IdU followed by CIdU for 25 min each, and then induced replication stress with 4 mM HU for 4 h. We show that in those RFs that were dual-labelled (sequentially, but not overlapping, green and red tracks), the ratio of CIdU/IdU track lengths was ~1.5 for WT T cells, while in WKO T cells, the ratio was ~0.5, indicating that CIdU track-length was significantly shorter after HU-treatment, consistent with an increased degradation of HU-stalled RFs (Fig. 5d). Together, these results implicate WASp in influencing global replication program under replicative stress and propose that increased degradation of stalled RFs contribute in part to genomic instability in WASp-deficient cells. WASp-deficiency causes heightened genome-instability in genotoxin-treated human T and B cells. Next, we analyzed the effect of fork dysfunction on the cellular DSB load as a measure of genome instability. Employing neutral comet assays (monitors DSBs), we show that the frequency of DSBs, as inferred from comet-tail moments, is significantly higher in WASp-deficient T cells relative to WT, under genotoxic stress (Fig. 5e). As such, we observe a higher frequency of comet tail moments in WKO T cells relative to WT, even in the unstressed cells (Fig. 5e), suggesting spontaneous genome instability likely arising from the accumulation of pathological R-loops and/or impaired HDR in Fig. 3 WASp-deficiency impairs RPA occupancy at perturbed RFs. a, SIRF. Shown are the representative z-stack collapsed composite confocal IF images and their box-and-whisker plots quantifying the enrichment of the indicated proteins at RFs, unperturbed or perturbed by the indicated genotoxins (post-4h treatment) in human T and B cells. n=150 cells, Mann-Whitney unpaired two-tailed nonparametric p-value: **** <0.0001. WKO T cells and WAS03 B cells are WASp-deficient (see Supplementary Fig. S2f for WASp expression profiles). a.u. denotes arbitrary units. Scale bar: 2μm. b, iPOND. Left panel, schematic of nascent DNA labeling, description as per the legend for Fig. 1c. Middle panel, shows Western blots of input and iPOND purified proteins under indicated conditions for wild-type (WT) and WAS knock-out (WKO) isogenic pair of human T cells. iPOND signal for WASp in WKO T cells served as a negative control. Right panel, bar graphs depict gel densitometric quantitation of iPOND/Western blot band signals normalized to their respective inputs. n=3, +SEM. Mann-Whitney unpaired two-tailed nonparametric *<0.01. Source data are provided as a Source Data file. c, EMSA/super-shift assays. Experiments performed using purified nuclear extracts from the indicated cell types, treated (+) or not (-) with CPT at indicated dose/duration. The location of endogenous heterotrimeric RPA bands, as verified by anti-RPA2 antibody mediated super-shifted band (left panel, red arrow denotes supershifted band). Red hatched box denotes the general location of the endogenous RPA1-3 bands, verified by supershift. p, probe only lane. Data is representative of n=3 independent assays. d, PLA. Representative z-stack collapsed composite confocal IF images and their quantification showing RPA2 localization at HU-mediated replication-stress sites (monitored by γH2A.X) post-2h in T cells, wild-type (WT) and WASp-deficient (WKO). Box-andwhisker plots, n=150 cells, Mann-Whitney unpaired two-tailed nonparametric p-value: **** <0.0001. Scale bar: 3μm. NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31415-z ARTICLE NATURE COMMUNICATIONS | (2022) 13:3743 | https://doi.org/10.1038/s41467-022-31415-z | www.nature.com/naturecommunications 7 WASp-deficient cells in the in vitro culture conditions11,12. Furthermore, employing confocal immunofluorescence imaging of DAPI-labelled cells, we show increased micronuclei formation, likely due to sequestered damaged DNA caused by unresolved stressed RFs, in WASp-deficient T and B cells compared to WT controls treated with CPT (Fig. 5f). Together, the data imply increased DNA damage in WASp-deficient cells. Because cell-cycle phase influences DSB repair pathway choice (SSA vs. HDR vs. NHEJ), dependent partly on whether endresection is activated (S/G2 phase) or not (G1 phase), we next analyzed the effect of fork dysfunction on cell-cycle distribution profiles. Flow-cytometry analyses revealed that, like ETAA1deficiency in human HCT116 and HeLa cells24, WASp-deficiency also does not markedly alter the overall cell-cycle distribution in human T and B cells (Supplementary Fig. S5). Under low-dose HU, both WT and WASp-deficient T and B cells show intra-S checkpoint activation. Under high-dose HU and CPT, WT and WASp-deficient T and B cells show a combination of G1 arrest, intra-S, and/or G2/M checkpoint activation with some notable differences in WT and WKO T-cells. Specifically, under CPTFig. 4 WASp deficiency disrupts ATR signaling at perturbed RFs and impairs global CHK1 activation. a, SIRF. Shown are the representative z-stack collapsed composite confocal IF images and their box-and-whisker plots quantifying the enrichment of the indicated proteins at RFs, unperturbed or perturbed by the indicated genotoxins and doses (post-4h treatment) in human T and B cells, WT or WASp-deficient (WKO T cells; WAS03 B cells). In box-and-whisker plots, whiskers @10-90%, horizontal bar denotes median, “+”denotes mean. The box-and-whisker plots are from n=150 cells analyzed, Mann-Whitney unpaired two-tailed nonparametric p-value: **** <0.0001; *** <0.001; ns, nonsignificant. a.u. denotes arbitrary units. Arrows show WASpdeficient T and B cells with micronuclei formation (See Fig. 5f for additional data on micronuclei). Scale bar: 2μm. b, PLA. Representative z-stack collapsed composite confocal IF images and their quantification plots for the indicated protein:protein interactions in T cells, wild-type (WT) and WASp-deficient (WKO), treated with the indicated genotoxin (post-4h treatment), or no treatment. In box-and-whisker plots, whiskers @10-90%, horizontal bar denotes median, “+”denotes mean. The box-and-whisker plots are from n=150 cells analyzed, Mann-Whitney unpaired two-tailed nonparametric p-value: **** <0.0001; ** <0.001; *<0.01; ns, nonsignificant. Scale bar: 4μm. c–e, Western blot. Representative images of the indicated proteins expressed in total cell extracts of human T cells, WT and WKO (panel c), B cell lines (Normal donor and WAS03 patient) (panel d), and WASp-deficient (WKO) human T cells stably-transfected to re-express GFP-tagged WT*WASp or RPA1-binding domain-deleted mutant of WASp (ΔRBP1*WASp) (panel e) treated with the indicated genotoxin for 4 h or untreated (control, ctrl), along with their gel densitometric analyses. In box-and-whisker plots, whiskers @10-90%, horizontal bar denotes median, “+”denotes mean. n=4 independent assays. In box plots shown in panels d and e, data is for mean±SEM. n=3 independent assays. Mann-Whitney unpaired two-tailed p-value: *<0.01, **<0.001. In panel c, the p-values for HU (0.2 low-dose) and untreated (no damage) comparing WT and WKO T cells were not significant. Source data are provided as a Source Data file. ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31415-z 8NATURE COMMUNICATIONS | (2022) 13:3743 | https://doi.org/10.1038/s41467-022-31415-z | www.nature.com/naturecommunications induced damage, WKO T cells showed a higher percentage of G1arrested cells (~70% at 24 h; ~44% at 48 h) compared to WT T cells (~37% at 24 h; ~20% at 48 h). These cell-cycle profiles suggest that WASp likely influences DSB repair pathway choices in a context and cell-type dependent manner, as further evidence from our yeast studies described later. Together, our results indicate that WASp is important for maintaining genome stability in human lymphocytes. Yeast Las17-inactivation manifests RPA and recombinational DNA repair defects. Since WASp has an ortholog in Saccharomyces cerevisiae, Las17, we did a genetic analysis to test whether results in human cells can be extrapolated to yeast to establish conservation of function in eukaryotes. Because in S. cerevisiae, homologous recombination is the primary mechanism for repairing DSBs, we used Las17-deficient Saccharomyces mutants carrying the las17-14 allele (see Methods section) to directly test WASp (Las17) role in HDR. In addition, we generated a las17 auxin-inducible degron strain (las17-aid) (see Methods section) to be able to conditionally deplete cells of Las17 expression after auxin (IAA) addition. By Western blot, we show that the degron allele effectively depleted the Las17 protein expression after IAA addition (Supplementary Fig. S6), thus validating the auxin-based degron system. Notably, hypersensitivity to DNA-damaging agents HU and Methyl methanesulfonate (MMS) is seen in both the las17-14 and the las17-aid mutants, the latter in the presence of IAA that activates the degron (Fig. 6a). This is accompanied by an increase in Rad52 foci (Fig. 6b), denoting an accumulation of recombinogenic DSBs, and thus confirming a defect in HDR under replication-stress and DNA damageinducing conditions. Since las17-14 and las17-aid (+IAA) mutants exhibited similar phenotypes for the DNA-damage readouts of our interest, we employed las17-14 mutant for all subsequent studies. Next, since accumulating evidence indicate that a natural source of replication stress and DNA damage is the formation of ectopic DNA-RNA hybrids (R loops)39,40, and we have previously shown that WASp deficiency leads to accumulation of such structures11 and that RPA facilitates ribonuclease H1 (RNH1) action in suppressing ectopic R loops41, we tested if Las17 has a role in preventing R loop accumulation in yeast. Indeed, the las17-14 mutant accumulates significantly more R loops at two previously validated genes, as determined by DNA-RNA immunoprecipitation (DRIP) using the antibody S9.6. The specificity of DRIP-signal was further verified by in vitro treatment with RNH1, which eliminated the DRIP-signals (Fig. 6c). Importantly, accumulation of DNA breaks as determined by Rad52 foci were also suppressed by RNH1 overexpression in vivo (Fig. 6d), confirming that R loops in these mutants are a natural source of replication stress and DNA damage. This implies that Las17, like human WASp11, is required to maintain a healthy R loop balance and prevent R-loopmediated DSBs in yeast. Fig. 5 WASp-deficiency undermines RF integrity and causes genome instability. a–d, DNA fiber assays showing 4 different labeling protocols, their representative replication track images, and their corresponding RF statistics in human T cells (WT, WKO isogenic pair) under unperturbed (panel a) or HU-perturbed (panels b–d) conditions. For assays in which the 1st and 2nd labeling times were similar, RF velocity was calculated (panels a, b), whereas, when labeling times were dissimilar, the individual RF track length was calculated (panel c). For quantifying RF degradation (panel d), the ratio of 2nd:1st labeled track lengths was calculated only in those tracks that showed sequential, but not overlapping, dual-labeling. The box-and-whisker plots (whiskers @10-90%, horizontal bar denotes median, “+”denotes mean) are from n =150 tracks analyzed from 3 independent experiments. Mann-Whitney unpaired two-tailed nonparametric p-value: **** <0.0001; ns, nonsignificant. See Supplementary Fig. S4 for additional RF data in human B cells. e, Representative confocal IF images of the neutral comet assay reporting on the frequency of DSBs in human T cells (WT, WKO), untreated or treated with indicated genotoxins and doses for 4 h. The box-and-whisker plots (whiskers @10-90%, horizontal bar denotes median, “+”denotes mean) reporting on the tailmoment are from n =150 cells analyzed, Mann-Whitney unpaired two-tailed nonparametric p-value: **** <0.0001. a.u. denotes arbitrary units. Scale bar: 14μm. f,Left, Representative confocal IF images of micronuclei formation (indicated by arrows) in DAPI-stained WASp-deficient T and B cells are shown for the CPT condition at 4 h. Right, statistical bar graph data, mean +SEM, n=150 cells for all genotoxin conditions from n =3 independent assays. Scale bar: 10 μm. NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-022-31415-z ARTICLE NATURE COMMUNICATIONS | (2022) 13:3743 | https://doi.org/10.1038/s41467-022-31415-z | www.nature.com/naturecommunications 9