BRCA1/BARD1 ubiquitinates PCNA in unperturbed conditions to promote continuous DNA synthesis

Article https://doi.org/10.1038/s41467-024-48427-6 BRCA1/BARD1 ubiquitinates PCNA in unperturbed conditions to promote continuous DNA synthesis Daniel Salas-Lloret 1,7 , Néstor García-Rodríguez 2,3,7 , Emily Soto-Hidalgo 3 , Lourdes González-Vinceiro 3 ,CarmenEspejo-Serrano 3 , Lisanne Giebel 1 , María Luisa Mateos-Martín 4 ,ArnoudH.deRu 5 , Peter A. van Veelen 5 , Pablo Huertas 2,3 ,AlfredC.O.Vertegaal 1 & Román González-Prieto 1,3,6 Deficiencies in the BRCA1 tumor suppressor gene are the main cause of hereditary breast and ovarian cancer. BRCA1 is involved in the Homologous Recombination DNA repair pathway and, together with BARD1, forms a heterodimer with ubiquitin E3 activity. The relevance of the BRCA1/BARD1 ubiquitin E3 activity for tumor suppression and DNA repair remains controversial. Here, we observe that the BRCA1/BARD1 ubiquitin E3 activity is not required for Homologous Recombination or resistance to Olaparib. Using TULIP2 methodology, which enables the direct identification of E3-specific ubiquitination substrates, we identify substrates for BRCA1/BARD1. We find that PCNA is ubiquitinated by BRCA1/BARD1 in unperturbed conditions independently of RAD18. PCNA ubiquitination by BRCA1/BARD1 avoids the formation of ssDNA gaps during DNA replication and promotes continuous DNA synthesis. These results provide additional insight about the importance of BRCA1/BARD1 E3 activity in Homologous Recombination. Breast cancer susceptibility type 1 (BRCA1) binds its partner BRCA1- Associated RING Domain 1 (BARD1) to form an obligated and multifunctional heterodimer with ubiquitin E3 ligase activity1–3.TheBRCA1/ BARD1 heterodimer acts as tumor suppressor and maintains genome stability, generally, by repairing deleterious double-strand DNA breaks (DSBs) via error-free homologous recombination (HR)4,5.DSBscan originate either from endogenous agents, such as reactive oxygens, replication fork progression and single-stranded DNA (ssDNA), or from environmental exposure to chemicals, ionizing radiation and ultraviolet light (UV)6. Germline mutations in BRCA1 and BARD1 are the main cause of hereditary breast and ovarian cancers, conferring a life-time probability of up to 90% for developing breast cancer and 50% for ovarian cancer7–9. Mice studies suggested that the E3 ligase activity is not essential to prevent tumor development10, as the effects of patientderived BRCA1 RING domain mutations such as C61G or C64G, which are required for tumor suppression and HR, are likely due to the disruption of the interaction of BRCA1 with BARD111. Nevertheless, mutations in BARD1 that impair Histone 2A (H2A) ubiquitination have been identified in familial breast cancer12. BRCA1/BARD1 has a very well-established histone H2A ubiquitin ligase activity on lysines K127/K12913–16. This ubiquitination has been related to the maintenance of heterochromatin integrity, genetic stability, and senescence2,13,16. Received: 3 August 2023 Accepted: 30 April 2024 Check for updates 1 Department of Cell and Chemical Biology, Leiden University Medical Centre, Leiden, The Netherlands. 2 Departamento de Genética, Facultad de Biología, Universidad de Sevilla, Sevilla, Spain. 3 Andalusian Centre for Regenerative Medicine and Molecular Biology (CABIMER), Universidad de Sevilla-CSIC- Universidad Pablo de Olavide-Junta de Andalucía, Sevilla, Spain. 4 Institute of Biomedicine of Seville, IBiS/Hospital Universitario Virgen del Rocío/CSIC/ Universidad de Sevilla, Proteomics Facility, Sevilla, Spain. 5 Center for Proteomics and Metabolomics, Leiden University Medical Center, Leiden, The Netherlands. 6 Departamento de Biología Celular, Facultad de Biología, Universidad de Sevilla, Sevilla, Spain. 7 These authors contributed equally: Daniel Salas- Lloret, Néstor García-Rodríguez. e-mail: rom[email protected] Nature Communications | (2024) 15:4292 1 1234567890():,; 1234567890():,; Up to date, the relevance of the BRCA1/BARD1 E3 activity for DSBs repair, tumor suppression, and resistance to PARP inhibitors and platinum-based compounds is still controversial. Histone H2A K127/ 129 ubiquitination is required for DNA DSB repair by HR and RAD51 focus formation2.However,deficiencies in histone H2A ubiquitination by BRCA1/BARD1 can be rescued by other ubiquitin E3s such as RNF16817,18. While a study from the Morris group2using siRNA-based knockdowns in HeLa cells showed that the BRCA1/BARD1 E3 activity was required for resistance to agents such as the PARP inhibitor Olaparib and cisplatin, a degrON-based strategy on HCT116 cells showed that the BRCA1/BARD1 E3 activity was dispensable for resistance to Olaparib19, consistent with most of the published literature10,11,17,19,20. Moreover, many research groups have addressed the challenge of identifying BRCA1/BARD1 ubiquitination substrates. However, all these studies have relied on indirect evidence to identify putative BRCA1/ BARD1 ubiquitination substrates, by either overexpressing or depleting BRCA1 and identification of changes in the ubiquitin proteome by mass spectrometry-based proteomics16,21–25.However,sinceBRCA1/ BARD1 plays an important role in several signaling pathways including cell cycle regulation, replication fork protection, gene transcription regulation, and DNA damage repair5,26, these approaches could indirectly affect the ubiquitination state of proteins. Recently, we developed the TULIP and TULIP2 methodologies27,28 which enable the unambiguous identification of direct ubiquitination substrates for an E3 of interest. Here, we investigate the role of the BRCA1/BARD1 heterodimer E3 activity in the non-tumoral cell line RPE1 and apply the TULIP2 methodology to identify novel substrates for ubiquitination by BRCA1/ BARD1. Combined, our data show that BRCA1/BARD1 promotes PCNA ubiquitination in unperturbed conditions to facilitate continuous DNA synthesis and provides insights about the ubiquitin E3 function of BRCA1/BARD1 during HR. Results BRCA1/BARD1 ubiquitin E3 activity is dispensable for DSB repair and resistance to treatment with Olaparib The relevance of the BRCA1/BARD1 ubiquitin E3 activityfor HR in terms of DNA DSB repair and treatment resistance remains contradictory in current literature, due to the use of different approaches involving diverse cancer cell lines and models2,10,11,19,20.Toaddressthiscontroversy, we employed RPE1 TP53−/−(Parental) and RPE1 TP53−/−BRCA1−/− (BRCA1-KO) cells3rescued or not with GFP-tagged BRCA1 wild type or I26A mutant at similar expression levels to the Parental cells BRCA1 levels (Supplementary Fig. 1a). The BRCA1 I26A mutant abrogates the BRCA1/BARD1 E3 activity without disrupting the formation of the BRCA1/BARD1 heterodimer (Supplementary Fig. 1b)29,30. We decided to test the proficiency in HR of the ubiquitin E3 activity BRCA1 mutant I26A in non-tumoral cells both in terms of resistance to Olaparib (Fig. 1a) and RAD51 focus formation in response to Ionizing Radiation (IR) (Fig. 1b, c). As expected, BRCA1-KO cells were sensitive to Olaparib and deficient in RAD51 focus formation in response to IR, and BRCA1- GFP rescued cells could restore both phenotypes to the Parental levels. Importantly,BRCA1-I26A-GFP rescued cells could also restore Olaparib resistance and RAD51 focus to the Parental levels, which is consistent with most of the published literature about the relevance of BRCA1 ubiquitinE3activityinHRinDSBrepairandresistancetoOlapariband Platinum-based compounds10,11,17,19,20. BRCA1/BARD1 ubiquitinates PCNA in a constitutive manner Previous studies aiming to identify BRCA1/BARD1 ubiquitination substrates are based on indirect evidence14,16,31. Moreover, the ubiquitination of histone H2A-K127/K129, which is the best studied BRCA1/BARD1 ubiquitination substrate14,16, can be rescued by RNF168, another E3 enzyme, upon BRCA1 deficiencies18. Thus, we hypothesized that BRCA1/BARD1 ubiquitination substrates other than histone H2A isoforms exist. We recently developed the TULIP2 methodology for the systematic application of Ubiquitin Activated Interaction Traps27,28,32, which enables the identification of E3-specific targets in a direct manner. The rationale behind this approach is that if we have a linear fusion between an E3 and ubiquitin, this E3 could use its fused ubiquitin to modify a substrate protein. This strategy enables the co-purification and subsequent identification by mass spectrometry-based proteomics of the E3 together with ubiquitin and the covalently modified substrate. Therefore, we decided to apply TULIP2 methodology to identify BRCA1/BARD1-specific ubiquitination substrates. Three different TULIP2 constructs were generated to identify BRCA1 direct ubiquitination targets: BRCA1-WT-TULIP2, BRCA1-I26A-TULIP2, and BRCA1- WT-TULIP2ΔGG (Supplementary Fig. 1c; Fig. 1d). Both BRCA1-I26A and BRCA1-WT-TULIP2ΔGG served as negative controls. While BRCA1-I26A is a catalytic dead mutant without E3 activity, BRCA1-WT-TULIP2ΔGG fused ubiquitin cannot be conjugated to target proteins due to lack of the C-terminal diGly motif (Supplementary Fig. 1c). First, we confirmed the functionality of the BRCA1-TULIP2 constructs for HR by introducing them in BRCA1-KO cells in a stable inducible manner and testing their resistance to Olaparib treatment (Supplementary Fig. 1e), and RAD51 focus formation in response to IR (Supplementary Fig. 1f, g). As previously seen for the GFP-tagged constructs, BRCA1-KO cells were hypersensitive to Olaparib and failed to form RAD51focus in response to IR. However, BRCA1-KO cells stably expressing either BRCA1-WT-TULIP2 or BRCA1-I26A-TULIP2 completely rescued both Olaparib sensitivity and deficiency in RAD51 focus formation (Supplementary Fig. 1e–g). Besides, BRCA1-TULIP2 constructs co-localized with RAD51 similarly to endogenous BRCA1 (Supplementary Fig. 1f). These results corroborate the functionality of BRCA1-TULIP2 constructs and indicate a minor role of BRCA1/BARD1 E3 activity on HR pathway for DNA DSB repair and Olaparib resistance. Next, we performed the TULIP2 assay to identify BRCA1 ubiquitin E3 activity-specific substrates using our BRCA1-TULIP2 cell lines (Supplementary Fig. 1c) and performed mass spectrometry-based proteomics analysis (Supplementary Data 1; Supplementary Fig. 1h-j). Analysis by immunoblotting failed to show a clear smear up from the wild type and I26A corresponding to BRCA1-TULIP2 conjugates as expected from functional TULIPs (Supplementary Fig. 1d)27,28,33. Moreover, proteomics analysis was unable to identify neither histone H2A nor macro-H2A as BRCA1-TULIP2 substrates, which would have served as positive controls, and the top hit was RAB43, which is not a nuclear protein. We concluded that the BRCA1-TULIP2 constructs were functional for BRCA1 activity but not regarding the TULIP2 assay, likely due to steric hindrance. Nevertheless, a recent cryo-EM study has shown that it is BARD1, and not BRCA1, the heterodimer partner which positions the E2 enzyme and directs the ubiquitination of histone H2A15. Therefore, we made stable inducible BARD1-TULIP2 constructs and introduced them in Parental cells, including the wild type and ΔGG TULIP2 constructs. As an additional negative control, in this case, we introduced wild type BARD1-TULIP2 constructs in BRCA1-KO cells (Fig. 1d). Analysis by immunoblotting of the BARD1-TULIP2- expressing cells (Fig. 1e) revealed that: (1) The endogenous BARD1 levels were similar in Parental and BRCA1-KO cells; (2) the BARD1- TULIP2 construct levels were not detectable compared to endogenous BARD1 levels in the input samples, which is important to avoid overexpression artifacts; and (3) The BARD1-TULIP2 constructs were very efficiently enriched in the His-pulldown following TULIP2 methodology and a smear up from the wild type BARD1- TULIP2 construct corresponding to BARD1-TULIP2 conjugates, which was absent in the ΔGG control, could be detected. Mass spectrometry analysis of BARD1-TULIP2 samples (Supplementary Data 2; Fig. 1f–h) identified both histones H2A and macro-H2A as BARD1 ubiquitination substrates, serving as internal positive controls. Interestingly, we could identify another top hit, PCNA, as a Article https://doi.org/10.1038/s41467-024-48427-6 Nature Communications | (2024) 15:4292 2 BRCA1/BARD1-specific ubiquitination substrate using TULIP2 methodology (Fig. 1h). BRCA1/BARD1 ubiquitinates PCNA to prevent ssDNA accumulation in unperturbed conditions The proteomics spectra obtained in our BARD1-TULIP2 experiments did not enable the identification of the ubiquitin acceptor lysine in PCNA for BRCA1/BARD1. K164 is the main ubiquitin acceptor site in PCNA, and this ubiquitination promotes the recruitment of Trans- Lesion Synthesis (TLS) DNA polymerases34. However, other acceptor lysines for ubiquitin have been identified in PCNA, which physiological relevance are unknown35. Thus, there was the possibility that the BRCA1/BARD1 heterodimer was ubiquitinating PCNA on a lysine other than K164. Therefore, we performed the BARD1-TULIP2 assay on RPE1 a 0510 0.0 0.5 1.0 Olaparib (PM) Survival Parental - WT I26A BRCA1-KO - 0 20 40 60 80 BRCA1-KO 0.0069 0.0032 0.0085 %Cells ≥10 rad51 foci Parental WT I26A b BRCA1: BARD1 TULIP2 BARD1 TULIP2 'GG BARD1 TULIP2 ++- BARD1 TULIP2 BARD1 TULIP2 'GG BARD1 TULIP2 ++- BARD1 Ponceau Input PD 191 97 64 kDa BARD1 BARD1-TULIP2 Conjugates 97 d e fg BRCA1-KO Parental WT I26A DAPI EGFP RAD51 MERGE - h c Fold Change (Log2) HIST1H2AJ HIST1H2AC 0 2 4 6 0246 PCNA H2AFY BRCA1-independent BARD1-TULIP2 conjugates (WT-BRCA1-KO vs ∆GG) - Log p Fold Change (Log2) H2AFY PCNA HIST1H2AJ HIST1H2AC 0 2 4 6 0246 BRCA1/BARD1-TULIP2 targets (WT vs Controls) - Log p Fold Change (Log2) H2AFY PCNA HIST1H2AJ HIST1H2AC 0 2 4 6 0246 BARD1-TULIP2 conjugates (WT vs ∆GG) - Log p Interactor BRCA1 E2 Target BARD1 HIS Ub LC-MS/MS-Analisys Target HIS BARD1 HIS Ub Ub Ub Target HIS BARD1 HIS Ub Ub Ub Target HIS BARD1 HIS Ub Ub Ub Interactor BRCA1 E2 Target BARD1 HIS Ub UbUb b Ub Ub U U U U U U 'GG HIS BARD1 HIS Ub'GG Interactor E2 Target BARD1 HIS UbUb Ub b Ub U Ub + x BRCA1-KO Parental BRCA1 BRCA1 Article https://doi.org/10.1038/s41467-024-48427-6 Nature Communications | (2024) 15:4292 3 TP53−/−PCNAK164R/K164R (PCNA-K164R) cells36 (Supplementary Data 3; Fig. 2a). In this assay, we could identify histones H2A and macroH2A as BARD1 ubiquitination substrates. However, PCNA was lost as a BARD1- specific substrate, which indicates that BRCA1/BARD1 ubiquitinates PCNA on lysine K164. PCNA K164 ubiquitination has recently been shown to participate in replication fork protection36 and in preventing the accumulation of ssDNA gaps during DNA replication in unperturbed conditions37,two functions that have been previously also attributed to BRCA138–41.Different groups have shown that BRCA1-deficient cancer cells Fig. 1 | BRCA1 E3 activity is not required for homologous recombination. aClonogenic survival assay of Parental and BRCA1-KO cells complemented or not with either BRCA1-WT-GFP or BRCA1-I26A-GFP in response to different concentrations of Olaparib. Average and standard deviation of three independent experiments with three technical repeats are depicted (N=3).b,cImmunofluorescence analysis Parental and BRCA1-KO cells complemented or nor with either BRCA1-WT-GFP or BRCA1- I26A-GFP after exposure to 10 Gy IR. Representative images (b)andquantification (c) of cells with RAD51 focus is shown. Size bars in fluorescence microscopy images represent 10 µm. Data corresponds to three independent experiments which value is shown by an orange, green or blue circle respectively. Average and Standard Deviation is shown. Unpaired two-sided t-tests were performed with pvaluesshowninthe figure (d) Cartoon depicting BARD1-TULIP2 rationale. eAnalysis by immunoblotting of BARD1-TULIP2 samples. This analysis was performed for two out of four of the samples prior to LC-MS/MS analysis. f–h. Volcano plots depicting differences between each of the specified BARD1-TULIP2 constructs. Pvalue corresponds to unpaired twosided t-tests. Histones H2A and macro-H2A and PCNA are labeled. Each dot represents a protein from Supplementary Data 2. Source data are provided as a Source Data file. a Fold Change (Log2) -Log p HIST1H2AC HIST1H2AJ HIST1H1C H2AFY PCNA 02468 0 2 4 6 8 Fold Change (Log2) BARD1-TULIP2 conjugates (PCNA-K164R) BARD1-TULIP2 conjugates (Parental) - Log p HIST1H2AC HIST1H2AJ HIST1H1C H2AFY PCNA 0 2 4 6 8 02468 b cd BARD1 BRCA1 BARD1 BRCA1 BRCA1-I26A BARD1 BRCA1 PCNA Ub Ub Ub Unperturbed Ub BARD1 BRCA1 PCNA Ub Ub Ub BRCA1-WT x ssDNA ssDNA x x 0 20 40 60 80 CldU tract lenght (μm) S1: ---- Parental BRCA1KO WT I26A - <0.0001 0.159 0.157 <0.0001 n: 392 477 368 424 375 543 516 610 0 20 40 60 80 CldU tract lenght (μm) 0.268 <0.0001 S1: -- Parental PCNA K164R n: 305 355 410 636 + S1 10 μm Parental BRCA1KO + S1 WT I26A BRCA1KO IdU CldU 20’ 60’ +/- S1 30’ Fig. 2 | BRCA1-BARD1 ubiquitinates PCNA on K164 for ssDNA gap suppression. aVolcano plots depicting BARD1-TULIP2 conjugates in the Parental and PCNA- K164R cell lines. PCNA is represented in purple and the Histones H2A and macro- H2A arein green. Each dot represents a protein from Supplementary Data 3. Pvalue corresponds to unpaired two-sided t-tests. bCartoon depicting schematic representation of ssDNA accumulation in BRCA1-BARD1 E3 catalytic dead mutants. Formation of ssDNA gaps in unperturbed conditions in Parental, BRCA1-KO and BRCA1-WT and BRCA1-I26A rescued cells (c) and in Parental and PCNA-K164R mutant cells (d). Top left: Scheme of the IdU/CldUpulse-labelingprotocol,followed by S1 nuclease treatment. Bottom left: Representative images of DNA fibers. Graphs represent CldU tract lengths in the indicated cell lines with and without S1 nuclease treatment. Each dot represents one fiber and the green bar represents the median. n: indicates number of fibers measured from three biological independent experiments. pvalues were calculated using the two-sided Mann-Whitney test. Coloreddots indicate the median of each independent experiment. Source data are provided as a Source Data file. Article https://doi.org/10.1038/s41467-024-48427-6 Nature Communications | (2024) 15:4292 4 accumulate ssDNA gaps and thus can be therapeutically exploited with PARP or REV1 inhibitors39,42. We hypothesized that BRCA1/BARD1 mediates PCNA ubiquitination to protect replication forks and avoid the accumulation of ssDNA gaps in unperturbed conditions. Therefore, mutations on either PCNA or BARD1-BRCA1 that compromise its E3 activity could lead to ssDNA accumulation and genetic instability (Fig. 2b). To test BRCA1/BARD1 E3 activity on ssDNA gaps formation, we cultured different RPE1 cell lines with 10 μM CldU for 48 h and analyzed by immunofluorescence in non-denaturing conditions using an anti-BrdU antibody that will only recognize CldU present in ssDNA gaps38 (Supplementary Fig. 2a). In line with previous research, we observed an increased number of ssDNA gaps in BRCA1-KO and PCNA K164R mutant RPE1 cells under normal growth conditions36,39. BRCA1 cells reconstituted with BRCA1-GFP suppressed the formation of ssDNA gaps to the Parental levels. However, complementation with the BRCA1-I26A-GFP mutant construct could not suppress ssDNA gap formation and showed similar levels of ssDNA foci to BRCA1-KO cells. However, this assay cannot distinguish replication-induced ssDNA gaps from ssDNA originating from other sources, such as DSB resection or other DNA repair intermediates. Alternatively, to corroborate that the E3 activity of the BRCA1/BARD1 heterodimer prevents the formationofssDNAgaps,weperformedtheS1-nucleaseassayonDNA fibers (Fig. 2c)38,42. In this assay, DNA fibers containing ssDNA gaps would be shortened after treatment with S1 nuclease, as it is the case in BRCA1-KO cells. Importantly, DNA fiber shortening was suppressed after complementation with wild-type BRCA1-GFP but not by BRCA1- I26A-GFP. Consistently, PCNA-K164R mutant cells also accumulated gaps behind the fork (Fig. 2d). Subsequently, we performed the S1-nuclease assay on DNA fibers after mild replication stress (0.5 mM hydroxyurea (HU) for 2 h) (Supplementary Fig. 2b). Mild replication stress exacerbated the accumulation of ssDNA gaps in both BRCA1-KO and PCNA-K164R cells. As previously observed in unchallenged conditions (Fig. 2c, d), wild-type BRCA1-GFP, but not by BRCA1-I26A-GFP, could suppress the accumulation of ssDNA gaps, although the accumulation of ssDNA gaps was ameliorated in BRCA1-I26A-GFP cells compared to BRCA1-KO cells. Additionally, the increased replication fork speed observed in BRCA1- KO and BRCA1-I26A cells in response to mild HU treatment was also evident in PCNA-K164R cells when compared to Parental cells (Supplementary Fig. 2b, c). ssDNA gap-suppression function of PCNA Ubiquitination and BRCA1/BARD1 are in the same pathway Next, we decided to investigate if the gap suppression functions of PCNA ubiquitination and BRCA1/BARD1 operate in the same pathway or through independent mechanisms. Thus, we performed the S1 nuclease fiber assay in unperturbed conditions in Parental or PCNA- K164R cells after treatment with siRNA control or siBRCA1 (Fig. 3a; Supplementary Fig. 3a). Both BRCA1-depleted Parental cells38 and PCNA-K164R mutantcells showed increased replication fork speed and ssDNA gap accumulation in unperturbed conditions. However, BRCA1 depletion in PCNA-K164R cells did not cause an additional replication fork speed increase nor an additional accumulation of ssDNA gaps, thus indicating that the PCNA ubiquitination and BRCA1/BARD1 act in the same pathway regarding ssDNA gap suppression. In contrast, inhibition of PolΘin PCNA-K164R cells with ART558 resulted in an additional formation of ssDNA gaps (Fig. 3b), which provide further evidence that non-epistatic interactions can be detected with this assay as previously observed for BRCA1-KO cells43. As lack of BRCA1 E3 activity causes the accumulation of replication associated ssDNA gaps, BRCA1-I26A mutant should be sensitive to replicative stress. Then, we decided to test the sensitivity to HU of the Parental and BRCA1-KO cells complemented or not with GFP-tagged constructs of either wild type or I26A mutant BRCA1. PCNA-K164R mutant was also included (Fig. 3c). In contrast to what we observed for Olaparib (Fig. 1a), where sensitivity of BRCA1-deficient cells could be rescued by complementation with the BRCA1-I26A-GFP mutant, BRCA- I26A-GFP complemented cells were more sensitive to HU treatment than their wild-type counterpart. Notably, PCNA-K164R mutant cells were more sensitive than the I26A mutant, mirroring the sensitivity observed in BRCA1-deficient cells, indicative of the relevance of BRCA1 and PCNA ubiquitination in additional DNA damage repair or tolerance pathways. Accordingly, treating PCNA-K164R cells with an shRNA targeting BRCA1 caused additional sensitivity to replication stress (Fig. 3d, Supplementary Fig. 3b), at least at 2 mM HU, as previously described36. In contrast when introducing shRNA-resistant BRCA1-WT- GFP or BRCA1-I26A-GFP constructs in the PCNA-K164R cells (Fig. 3e), no additional sensitivity of the BRCA1 shRNA treatment was observed. Which indicates that the additional sensitivity to HU of BRCA1 knockdown in PCNA-K164R cells is not related to BRCA1 E3 activity. We found intriguing that BRCA1-I26A cells were resistant to Olaparib (Fig. 1a), which induces ssDNA gaps38,39, but not to gaps induced by HU (Fig. 3c, Supplementary Fig. 2c). Therefore, we decided to investigate how these treatments affected PCNA ubiquitination levels (Supplementary Figs. 3c, d). Analysis by immunoblotting of PCNA ubiquitination levels after treatment with Olaparib (Supplementary Fig. 3c), showed no effect. However, treatment with HR (Supplementary Fig. 3d) caused an accumulation of PCNA ubiquitination in Parental and BRCA1-WT-GFP rescued BRCA1-KO cells. This response to HU did not occur in BRCA1-I26A-GFP rescued cells. Interestingly, BRCA1- KO cells were able to respond to hydroxyurea treatment with an increase in relative PCNA ubiquitination levels, similar to the parental cells, which indicates that sensitivity of BRCA1-KO cells to HU is not related to PCNA ubiquitination. Noteworthy, in unperturbed conditions (Supplementary Fig. 3e) PCNA-Ub level was reduced 60% in BRCA1-KO cells, and this decrease was partially rescued by BRCA1-WT. However, PCNA-Ub level was equally rescued by the BRCA1-I26A mutant, suggesting that the contribution of BRCA1 E3 activity to the total levels of ubiquitinated PCNA is negligible. BRCA1/BARD1 and RAD18 ubiquitinate PCNA-K164 in distinct scenarios BRCA1 has been shown to promote the recruitment of RAD18 to chromatin in response to replication-blocking lesions44. Since RAD18 is considered the canonical ubiquitin ligase for PCNA in response to replication barriers45–47 and protein-protein interactions between BRCA1 and RAD18 have been described44, there was the possibility that BRCA1 absence was negatively affecting the E3 activity of RAD18 towards PCNA. To test this hypothesis, first, we performed iPOND experiments48 to investigate if the absence of BRCA1 or its E3 activity was negatively affecting the recruitment of BRCA1 or RAD18 to replication forks (Fig. 4a, b; Supplementary Data 4). Neither the recruitment of RAD18 to replication forks was affected by the absence of BRCA1 nor the recruitment of BRCA1 was affected by the lack of E3 activity. Noteworthy, MDC1 was enriched in BRCA1-KO cells compared to Parental cells, indicative of DNA damage at the forks. Next, we irradiated with UV light Parental and BRCA1-KO cells and followed the accumulation of ubiquitinated PCNA in a time course manner (Supplementary Fig. 4a). No substantial differences in ubiquitinated PCNA accumulation kinetics could be observed after UV irradiation between Parental and BRCA1-KO cells. However, depleting the cells from RAD18, completely suppressed the accumulation of ubiquitinated PCNA after UV irradiation (Supplementary Fig. 4b), indicating that the absence of BRCA1 was not negatively affecting the E3 activity of RAD18 towards PCNA. Nevertheless, we decided to monitor the ubiquitin E3 activity of RAD18 towards PCNA in a direct manner. Therefore, we made RAD18-TULIP2 constructs, introduced them in Parental and BRCA1-KO cells and performed the TULIP2 assay with Article https://doi.org/10.1038/s41467-024-48427-6 Nature Communications | (2024) 15:4292 5 and without UV light irradiation (Supplementary Data 5; Fig. 4c). As expected, immunoblotting analysis (Supplementary Fig. 4c) showed that the wild type RAD18-TULIP2 samples formed a smear up from the RAD18-TULIP2 construct, indicative of its functionality, that was suppressed in the ΔGG negative controls. Mass spectrometry-based proteomics analysis of the non-irradiated samples (Fig. 4d) revealed that among the identified RAD18 substrates, PCNA was not present. This favors the hypothesis that BRCA1/BARD1 is, compared to RAD18, a more active ubiquitin E3 enzyme for PCNA under unperturbed conditions (Fig. 1f). However, there was still the possibility thatthe absence of PCNA among the RAD18-specific substrates, regardless of the formation of the smear, was due to steric hindrance, similarly as previously observed for BRCA1 (Supplementary Fig. 1). To serve as positive control, we irradiated with UV light RAD18- TULIP2 expressing Parental cells, performed the TULIP2 methodology and identified the substrates after UV irradiation (Supplementary a c 024 6 8 10 mM HU Parental BRCA1KO WT I26A PCNA K164R - 02468 0.0 0.2 0.4 0.6 0.8 1.0 0.0778 0.0961 0.0029 0.0938 HU (mM) Survival Parental - BRCA1-WT BRCA1-I26A PCNA K164R BRCA1-KO CldU tract lenght (μm) 0 20 40 60 80 <0.0001 0.919 0.583 <0.0001 <0.0001 0.004 S1 Nuclease Parental PCNA K164R ---- ART558 ---- IdU CldU 20’ 60’ +/- S1 30’ 25 μM ART558 for 24h b S1 Nuclease Parental PCNA K164R ---- siNT siBRCA1 - -- -- -- - 0 20 40 60 80 CldU tract lenght (μm) 0.327 0.886 0.744 0.001 <0.0001 IdU CldU 20’ 60’ +/- S1 30’48h siRNA <0.0001 <0.0001 <0.0001 n: 317 368 447 457 399 413 387 508 80246 0.0 0.5 1.0 Parental vs PCNA-K164R Parental shControl Parental shBRCA1 KR shControl KR shBRCA1 Survival [HU] mM de PCNA-K164R-BRCA1-GFP vs PCNA-K164R-BRCA1-I26A KR-BRCA1-WT shControl KR-BRCA1-WT shBRCA1 KR-BRCA1-I26A shControl KR-BRCA1-I26A shBRCA1 02468 0.0 0.5 1.0 Survival [HU] mM n: 268 198 300 298 196 228 213 270 Article https://doi.org/10.1038/s41467-024-48427-6 Nature Communications | (2024) 15:4292 6 Data 5; Fig. 4e). PCNA became now a substrate while it was not in unperturbed conditions, confirming the functionality of the RAD18- TULIP2 construct. PCNA ubiquitination levels were lower in BRCA1 knockout cells compared to Parental cells in unperturbed conditions (Supplementary Fig. 3e). However, after UV irradiation, ubiquitination kinetics of PCNA were not affected by BRCA1 knockout (Supplementary Fig. 4a). Additionally, we confirmed by RAD18 knockdown that PCNA ubiquitination after UV exposure was RAD18-dependent and BRCA1-independent (Supplementary Fig. 4b). Finally, we performed the RAD18-TULIP2 assay in BRCA1-KO cells in unperturbed conditions and after UV irradiation (Fig. 4f, g). Mass spectrometry analysis revealed a substantial enrichment of PCNA as a RAD18-specificsub- strate compared to Parental cells in both unperturbed and UV conditions. The conjugation of PCNA to the RAD18-TULIP2 construct could be confirmed by immunoblotting (Fig. 4h), since PCNA antibody staining produced a band between 100 and 130 kDa, corresponding the size of a branched polypeptide consisting of RAD18-TULIP2 covalently bound to PCNA. Next, we investigated how PCNA ubiquitination levels were affected by the absence of RAD18 and how these levels responded to HU treatment (Supplementary Fig. 4d). RAD18-KO cells had a big reduction on PCNA-Ub levels in unperturbed conditions, yet they were still able to respond to HU treatment. Importantly, the remaining PCNA ubiquitination levels in RAD18-KO cells sufficed to suppress the accumulation of ssDNA gaps in unperturbed conditions (Fig. 4i). Taken together, our data indicate that BRCA1/BARD1, regarding the ssDNA gap-suppression function of PCNA ubiquitination, is the main E3 for PCNA in unperturbed conditions. However, it is likely that other E3s ubiquitinate PCNA in different scenarios also present in unperturbed conditions which are responsible of most of the ubiquitinated PCNA fraction. In the absence of BRCA1, RAD18 might compensate for PCNA ubiquitination levels in a scenario downstream the ssDNA gap formation. Collectively, these data suggest that BRCA1/BARD1 and RAD18 share PCNA as substrate but act in different scenarios. While RAD18 would ubiquitinate PCNA when the replication fork encounters an obstacle on the DNA template, BRCA1/BARD1 would perform PCNA ubiquitination at low levels to facilitate continuous DNA synthesis during unperturbed conditions or upon induction of mild replicative stress that do not imply replication fork stalling, such as after mild HU treatment (Fig. 4j). Discussion Here, we provide novel insight in the role of the BRCA1/BARD1 ubiquitin E3 activity in promoting continuous DNA synthesis during DNA replication. Our results (Fig. 1a–c) are in accordance with most studies and show that the E3 activity of BRCA1/BARD1 is not relevant for HR and Olaparib resistance10,19,20. However, we noticed that cells deficient in the BRCA1/BARD1 ubiquitin E3 activity were sensitive to HU (Fig. 3c) and accumulated more ssDNA gaps behind the replication forks (Fig. 2c, d; Supplementary Fig. 2), which indicates that deficiencies in the E3 activity of BRCA1/BARD1 do not affect the DSB repair function of HR but do affect the DNA replication function of HR49. Using TULIP2 methodology in unperturbed conditions, we identified PCNA-K164 as a ubiquitination substrate for BRCA1/BARD1 but not for RAD18 (Fig. 1d–h; Fig. 2a; Fig. 4b). Interestingly, previous studies show synthetic lethality in BRCA1 deficient cells with either RAD18 loss or the use of REV1 inhibitors42,50. PCNA is known to be monoubiquitinated at low levels in mammalian cells in unperturbed conditions36,51,52, and PCNA ubiquitination levels in unperturbed conditions were reduced in BRCA1-KO cells (Supplementary Fig. 3e), although still some PCNA ubiquitination could be detected. Ectopic BRCA1-WT expression could partially rescue basal PCNA ubiquitination levels. However, BRCA1-I26A could also do it at the same extent, which suggests that the reduction of basal PCNA ubiquitination levels is due to replication fork collapse in the absence of BRCA153,54,as reflected in the accumulation of MDC1 at replication forks in BRCA1- KO cells compared to Parental cells (Fig. 4a; Supplementary Data 4). RAD18-KO also caused a big reduction in basal PCNA Ubiquitination levels (Supplementary Fig. 4d). Absence of RAD18 also causes transcription-dependent fork collapse55. Thus, we cannot exclude the possibility that the decrease in PCNA ubiquitination in RAD18-KO cells is also due to replication fork collapse. Moreover, BRCA1-KO also promotes the accumulation of R-loops56, which can explain why activity of RAD18 towards PCNA increases by several orders of magnitude (Fig. 4f–g) but this ubiquitination is still unable to suppress the formation of replication-dependent ssDNA gaps. BRCA1 and RAD18 would be working in different scenarios arising in response to endogenous stresses but the absence of one would be promoting the accumulation of the toxic structures for which the other E3 is required. Thus, determining which fraction of the basal PCNA ubiquitination levels correspond to which E3 by quantifying this modification after the knockout of different E3s is misleading as can be due to an indirect effect (i.e., fork collapses occurring upon knockout of these E3s or rewiring of the ubiquitination machinery). Noteworthy, our TULIP2 approach (Figs. 1d–h, 2a, 3c–h) enabled us to access these hurdles by studying the activity of an E3 towards its substrates in a scenario where all the other E3s are present and with expression below the endogenous levels, avoiding the appearance of overexpression artifacts. With this approach, our data indicates that in unperturbed conditions BRCA1/BARD1 is the predominant E3 for PCNA, compared to RAD18, responsible of the replication-associated ssDNA gapsuppressionfunction of PCNA ubiquitination.Nevertheless, it is likely that other E3s, yet to be characterized, also modify PCNA in other situations arising in unperturbed conditions, given the big proportion of E3s in the human proteome which physiological relevance has not been studied yet. E3s for PCNA other than RAD18 and BRCA1 have been described, including the Cullin 4 RING ligase, CRL457, or RNF21458.However,while CRL4 works synergistically with RAD18 and CRL4 knockdowns affect the accumulation of mono-ubiquitinated PCNA in response to UV irradiation57, no differences were observed in PCNA ubiquitination in response to UV light in BRCA1-deficient cells (Supplementary Fig. 4a, b). This also supports the model that BRCA1/BARD1 and RAD18 modify PCNA in alternative pathways (Fig. 4g), without excluding the possibility that a crosstalk between BRCA1 and RAD18 exists as previously proposed44. Fig. 3 | PCNA-K164 ubiquitination and BRCA1 ssDNA gap suppression activities are in the same pathway. Epistasis analysis of ssDNA gap formation. Cells were transfected with siRNA control (siNT) or siBRCA1 48 h before IdU and CldU labeling (a), or 24 h with the PolΘinhibitor ART558 (b) followed by S1 nuclease treatment. Graphs represent CldU tract lengths in the indicated cell lines with and without S1 nuclease treatment. Each dot represents one fiber and the green bar represents the median. n: indicates number of fibers measured from three biological independent experiments. pvalues were calculated using two-sided Mann-Whitney test. Colored dots indicate the median of each independent experiment. cSurvival assay of Parental, PCNA-K164R mutant and BRCA1-KO cell lines rescued with either BRCA1- WT or BRCA1-I26A against different HU concentrations. Three independent experiments with two technical repeats were performed (N=3).Pvalues of unpaired two-sided t-tests between BRCA1-WT and BRCA1-I26A cells is indicated. dSurvival assay after treatment with HU of Parental and PCNA-K164R mutant cells treated either with a control shRNA of BRCA1-targetting shRNA. eSurvival assay after treatment with HU of PCNA-K164R mutant cells ectopically expressing an shRNA-resistant construct of BRCA1-WT or BRCA1-I26A treated either with a control shRNA of BRCA1-targetting shRNA. c–eAverage and Standard Deviation of 3 independent experiments is depicted. Source data are provided as a Source Data file. Article https://doi.org/10.1038/s41467-024-48427-6 Nature Communications | (2024) 15:4292 7 Recently, it has been proposed that the synthetic lethality between BRCA1-KO and Olaparib treatments comes from an increased replication fork speed in BRCA1 deficient cells and subsequent hyper accumulation on ssDNA gaps behind replication forks38,39.Theaccumulation of ssDNA gaps behind the fork have been also detected in PCNA-K164R mutants36, as well as when treating BRCA1-deficient cells with mild replicative stress like HU concentrations not sufficient to produce a complete nucleotide depletion38,42. In this study, we show that the E3 activity BRCA1-I26A mutant show similar levels of ssDNA gaps formation as PCNA-K164R mutant and BRCA1-KO cells (Fig. 2c, d), which seems, in principle, contradictory with the E3-deficient mutant being resistant to Olaparib treatment (Fig. 1a10,19,20). However, Olaparib treatment does not induce PCNA ubiquitination (Supplementary Fig. 3b), thus suggestingthat PARP inhibition-associated ssDNA gaps in ssDNA PCNA UV Unperturbed j 5x 5x RAD18-TULIP2 WT 'GG LC-MS/MS-Analisys TULIP2 methodology RAD18-TULIP2 BRCA1-KO +Doxycycline o/n +/- 40 J/m2 UV 5h Recovery WT 'GG c hi WT ∆GG WT ∆GG WT ∆GG WT ∆GG --++--++UV: Parental BRCA1-KO RAD18-TULIP2: IB:RAD18 IB:PCNA -70 kDa -130 -100 -130 -100 0 20 40 60 80 CldU tract lenght (μm) S1: -- Parental RAD18KO 0.803 0.716 IdU CldU 20’ 60’ +/- S1 30’ n: 249 307 231 271 H2AFY PCNA HIST2H2AC Fold Change (Log2) 0 2 4 6 8 -Log p 02468 RAD18-TULIP2 conjugates +UV BRCA1KO (WT vs 'GG) Fold Change (Log2) PCNA H2AFY 0 2 4 6 0246 HIST1H2AC RAD18-TULIP2 conjugates +UV (WT vs 'GG) -Log p PCNA 02468 Fold Change (Log2) 0 2 4 6 8 -Log p RAD18-TULIP2 conjugates BRCA1KO (WT vs 'GG) Fold Change (Log2) PCNA H2AFY HIST1H2AC 0 2 4 6 0246 RAD18-TULIP2 conjugates (WT vs 'GG) -Log p df eg ● ● ●●● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ●● ● ● ● ●● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ●● ●● ● ● ● ● ● ●● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ●● ● ● ● ●●●● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ●● ● ● ● ●● ●● ●●● ●● ●●● ●●● ● ●● ● ● ●● ● ●● ●● ● ● ● ● ● ●● ● ● ●● ● ●● ● ●● ● ● ● ● ●● ●● ● ●● ● ●● ● ●● ● ● ●● ● ●● ● ● ● ●● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ●● ●●● ● ● ●● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ●●● ● ●● ● ● ●● ● ● ● ●●● ● ●● ●● ● ●●● ●● ●● ● ● ● ● ● ●●● ●● ●● ● ● ●● ● ● ● ●● ● ● ● ●● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●●● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ●● ●● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ●● ● ● ● ● ● ●● ● ●● ● ● ● ● ● ●● ● ●● ●● ● ● ● ●● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ●● ● ●● ●● ●● ● ● ● ● ●● ●●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ●● ● ●● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ●● ●●● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ●● ●●● ●● ●● ● ● ● ● ● ●● ● ● ●● ●● ● ●● ● ● ●● ● ● ● ●● ● ● ● ●● ● ●● ●● ● ● ● ●● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ●●● ●● ●● ● ● ●● ● ● ● ● ●● ●● ● ●● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ●● ● ● ●● ●● ● ● ● ● ● ● ●● ● ● ●● ● ● ●● ● ● ●● ● ●● ● ● ● ●● ●● ●● ● ● ● ●● ● ● ●● ● ●● ● ● ●● ●● ●● ● ● ●● ●● ●● ● ● ● ●● ● ● ● ● ● ● ●● ● ●● ● ● ● ●●● ●●● ● ● ● ● ● ● ● ● ● ● ●● ● ●●● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ●●● ● ●● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ●● ●● ● ●● ● ●● ● ● ● ●● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ●●●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ●● ●●● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ●● ● ●● ● ● ●● ● ●● ● ●● ● ● ● ● ● ● ● ●● ● ●● ● ●● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ●● ● ●● ● ● ● ● ●● ● ●●● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ●●● ●● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ●●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ●● ● ●● ●● ● ● ●● ● ●●● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ●●● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ●● ●● ● ● ● ● ●● ● ● ●● ● ● ● ●● ● ●● ● ●● ● ●●● ●● ● ● ● ● ● ● ● ●●● ●● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ●●●● ● ● ● ●● ● ● ●● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ●● ● ●● ● ● ●● ● ●● ● ● ● ●● ● ●● ● ●● ●● ● ●● ●● ● ● ● ● ● ● ●● ● ● ●● ●● ● ●● ●● ● ● ● ● ● ● ● ●● ● ● ●● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ●●● ●● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ●● ● ● ● ●●● ●● ● ● ● ● ● ●● ● ●● ● ●● ● ●● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ●● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ●● ● ● ●● ● ● ●● ● ● ● ● ● ●● ● ● ●●● ● ● ● ● ● ●● ● ● ● ● ● ●● ●● ●●● ● ● ● ● ● ● ●● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ●● ● ● ● ● ● ● ●● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ●●●● ●● ● ●● ●● ● ●● ● ● ●● ● ●● ● ●●● ● ●● ● ● ● ● ● ●● ●● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ●● ● ●● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ●● ●● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ●●● ● ● ● ●● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ●● ●● ● ● ●● ● ●● ● ● ●● ●● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ●● ● ●● ● ● ●● ● ● ●● ● ● ● ● ●● ● ●● ●● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ●● ●● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ●●● ● ● ●● ● ● ●● ● ● ● ● ●● ● ● ● ●● ● ● ● ●●● ●● ● ● ● ● ●● ●● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ●● ● ● ●● ● ● ●●● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ●●● ● ● ● ●● ● ● ● ● ●● ●● ●● ● ● ●●● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●●● ●● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ●●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ● ●● ●● ● ● ● ●● ●● ● ●● ● ●● ● ●● ●●● ● ● ● ●●● ● ● ● ●● ● ●● ● ●● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ●● ● ●● ●● ●● ●● ● ● ●● ● ● ●● ●● ● ● ●● ● ● ● ●● ● ● ● ●● ● ●● ●● ● ●● ● ●● ● ● ● ● ● ●●●● ● ● ● ● ● ●● ● ● ●● ● ● ● ● PCNA BRCA1 RAD18 MDC1 0 1 2 3 4 5 05 Fold Change (log2) - Log p IPOND Parental vs BRCA1-KO ab ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ●● ●● ● ●● ● ●● ● ● ● ● ●● ● ● ● ●●● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ●● ● ● ● ● ● ● ●●● ● ●● ●● ● ● ●● ● ●● ●● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ●● ● ●●● ●● ● ● ●● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ●● ● ● ●● ● ● ●● ●● ● ●● ●● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ●●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ●●● ●● ● ● ●● ● ● ●● ●● ●●● ● ●● ●●● ● ● ● ● ● ●●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ●●● ● ●● ● ●● ●● ● ● ●● ●● ● ● ● ●● ● ●●● ●● ● ●● ● ●● ● ● ● ● ● ●● ● ●● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ●● ● ● ●● ●● ● ● ● ● ● ●● ●● ●● ●● ● ● ● ● ● ●● ● ●●● ● ● ●● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ●● ● ● ●● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ●● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ●● ●● ● ●● ●● ●● ● ● ● ● ●● ● ● ● ●●●● ● ● ● ● ●● ● ● ● ● ●● ● ● ●● ●● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ●●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ●●● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ●● ●● ● ● ●● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ●● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ●● ● ●● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●●● ● ●● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ●●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ●●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ●● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ●● ●● ● ● ●● ● ● ●● ● ● ●●● ●● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ●● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ●● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ●● ●● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ●● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ●● ● ● ● ● ● ●● ● ●●● ●● ●● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ●● ● ●● ● ● ●● ●● ● ● ● ●● ● ●● ● ● ● ● ● ●● ● ● ●● ●● ● ● ● ●● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●●● ● ● ● ● ● ● ● ● ●●● ● ● ●● ● ● ● ●● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ●●● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ●● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ●● ● ●● ● ● ●● ● ● ● ● ● ● ●● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ●● ● ● ● ● ● ●● ● ● ● ●● ●● ● ● ● ●● ●● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ●● ● ●● ●●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ●● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ●● ● ● ● ● ● ● ● ●● ●● ● ● ● ●● ●● ● ●● ● ● ● ●●● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ●● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ●● ● ●● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ●● ● ● ● ● ● ●● ●● ● ● ● ● ● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ●● ● ● ● ● ● ● ● ● ●● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● PCNA BRCA1 RAD18 MDC1 0 1 2 3 4 −2 0 2 4 Fold Change (log2) IPOND BRCA1-WT vs BRCA1-I26A - Log p Article https://doi.org/10.1038/s41467-024-48427-6 Nature Communications | (2024) 15:4292 8 BRCA1-deficient cells and BRCA1 E3 activity-deficiency associated gaps are formed through independent processes. We then propose a model (Fig. 5) in which upon replication stress the BRCA1/BARD1 E3 activity would promote replication fork stability and continuous, ssDNA-free, DNA synthesis. Deficiencies in the E3 activity would eventually promote the appearance of DNA DSBs. While in BRCA1-deficient cells these DSBs would lead to cell death, in the E3-deficient mutants could still be repaired through HR-mediated DSB repair. Consistently, BRCA1-KO and PCNA-K164R cells are more sensitive to HU than the BRCA1-I26A-GFP cells (Fig. 3c), as they also lack alternative resistance mechanisms such as HR-mediated DSB repair and RAD18-mediated PCNA ubiquitination pathways, respectively. Finally, BRCA1 tumors carrying premature stop codon mutations or disruption of the RING domain are sensitive to PARP inhibitor therapies. Unfortunately, these tumors can revert by restoring the expression of a BRCA1 protein lacking the RING domain functionality but expressing the rest of the domains11. Alternatively, treating these patients with other chemo therapeutic agents such as HU, for which BRCA1/BARD1 E3-deficient cellsare still sensitive (Fig. 3b) might open a therapeutic opportunity for cancer treatment. Methods Generation of TULIP2 and GFP-tagged constructs BRCA1 gene was amplified and STOP codon removed from pDEST-FRT/ T0-GFP-BRCA1 (Addgene #71116) by PCR using BP-tailed primers and cloned into pDNOR207 by Gateway® cloning BP reaction (Thermo Fisher Scientific). BRCA1-I26A mutation was introduced by site-directed mutagenesis. RAD18-pDNOR223 plasmid was obtained from Open Biosystems (Clone 1782) and the STOP codon was removed by site-directed mutagenesis. TULIP2 constructs were generated by Gateway® cloning LR reaction between donor plasmids containing BARD128,BRCA1or RAD18 and acceptor TULIP2 plasmids27. For the GFP-tagged constructs, the LR reactions were performed with pLVU-GFP, gift from Lars Ittner (Addgene #24177)59. Primers are listed in Supplementary Table 1. Cell culture 293 T HEK and RPE1 cells were cultured in Dulbecco’smodified Eagle’s medium (DMEM) supplemented with 10 % Fetal Bovine Serum (FBS) and 100 U/mL penicillin/100 µg/mL streptomycin at 37 °C and 5% CO2 unless specifically specified. All the Parental RPE1 cells in this work were TP53−/− . Cells were regularly tested for mycoplasma contamination. TULIP2 Lentivirus production 293 T HEK cells were seeded at 30% confluency in a T175 flask containing 17 mL of DMEM + 10% FBS. After 24 h, the transfection mixture was prepared by combining lentiviral packaging plasmids 7.5 µg pMD2.G (#12259, Addgene), 11.4 µg pMDLg-RRE (#12251, Addgene), 5.4 µg pRSV-REV (#12253, Addgene) and 13.7 µgTULIP2plasmidwith 114 µL of 1 mg/mL Polyethylenimine (PEI) in 2 mL 150 mM NaCl. The mixture was vortexed and incubated 15 min at room temperature before adding to the HEK cells. Next day, culture medium was refreshed with DMEM 10% FBS. 72 h post-transfection, lentiviral suspension was harvested by filtering through a 0.45 µmsyringefilter (PN4184, Pall Corporation) and kept at −20 °C for further use. Lentiviral particle concentration was determined using the HIV Type 1 p24 antigen ELISA Kit (ZeptoMetrix Corporation). Generation of cell lines RPE1-hTERT TP53−/−and RPE1-hTERT TP53−/−BRCA1−/−cells were kindly provided by Dr. Sylvie Noordermeer60 and were seeded in 15 cm Fig. 4 | RAD18 and BARD1 ubiquitinates PCNA in different pathways. Volcano Plots depicting statistical differences in iPOND experiments comparing (a) Parental and BRCA1-KO cells or (b) BRCA1-WT or BRCA1-I26A rescued cells. Pvalue corresponds to unpaired two-sided t-tests. Each dot represents a protein from Supplementary Data 4. cSchematic representation of RAD18-TULIP2 set up in Parental and BRCA1-KO cell lines. TULIP2 constructs were induced with 1 μg/μL doxycycline overnight. Cells were treated or not with 40J/m2of UV light and allowed to recover for 5 h prior to cell lysis, protein purification and identification by LC-MS/MS of the RAD18-TULIP2 conjugates. RAD18-WT constructs were compared with their correspondent ΔGGconstructs for conjugates identification after MS analysis. Volcano plots of RAD18-TULIP2 conjugates without UV damage (d) and upon UV treatment (e) in Parental cell lines. Volcano plots depicting RAD18-TULIP2 conjugates without UV damage (f) and after UV damage (g) in BRCA1-KO cell lines. Histones are represented in green and PCNA is shown in purple. Each dot represents a protein from Supplementary Data 5 and proteins enriched over 0.8 fold (log2) and p>0.05 are displayed in red. Pvalue corresponds to unpaired two-sided t-tests. hImmunoblotting of RAD18-TULIP2 pull down samples with and without UV damage in Parental and BRCA1-KO cells. Antibodies used are indicated. This experiment was repeated three times with similar results. iFormation of ssDNA gaps in unperturbed conditions in Parental, and RAD18-KO cells. Top: Scheme of the IdU/CldU pulse-labeling protocol, followed by S1 nuclease treatment. Bottom: CldU tract lengths inthe indicated cell lines with andwithout S1 nuclease treatment. Each dot represents one fiber and the green bar represents the median. n: indicates number of fibers measured from three biological independent experiments. p values were calculated using two-sided Mann-Whitney tests. Colored dots indicate the median of each independent experiment. jProposed model for PCNA ubiquitination in two different scenarios. During unperturbed conditions BRCA1/BARD1 ubiquitinates PCNA at low levels to promote continuous DNA synthesis. Upon replication fork barrier-inducing lesions, such as UV irradiation, RAD18 ubiquitinates PCNA at higher levels. Source data are provided as a Source Data file. cIt has been adapted and modified from Yalçin Z, Koot D, Bezstarosti K, Salas-Lloret D, Bleijerveld OB, Boersma V, Falcone M, González-Prieto R, Altelaar M, Demmers JAA, Jacobs JJL. Ubiquitinome Profiling Reveals in Vivo UBE2D3 Targets and Implicates UBE2D3 in Protein Quality Control. Mol Cell Proteomics. 2023 Jun;22(6):100548. https://doi.org/10.1016/j.mcpro.2023.100548. which was published under CC-BY license [https://creativecommons.org/licenses/by/4.0/]. BRCA1-KO ssDNA Cell death BRCA1-I26A RPA ssDNA RPA DSB DSB Cell Survival BRCA1-I26A No repair Repair BRCA1-KO Fig. 5 | Proposed model. Model representing the relevance of BRCA1/BARD1 E3 activity in preventing ssDNA gap accumulation and promoting replication fork stability by PCNA ubiquitination. In the absence of BRCA1/BARD1 E3 activity, ssDNA gaps accumulate, which can eventually lead to the formation of DSBs. While in the absence of BRCA1, these DSBs would lead to cell death, in E3-deficient mutants, they can still be repaired by Homologous Recombination-mediated DSB repair. Article https://doi.org/10.1038/s41467-024-48427-6 Nature Communications | (2024) 15:4292 9