BRCA2 prevents PARPi-mediated PARP1 retention to protect RAD51 filaments

Abstract

The tumour-suppressor protein BRCA2 has a central role in homology-directed DNA repair by enhancing the formation of RAD51 filaments on resected single-stranded DNA generated at double-stranded DNA breaks and stimulating RAD51 activity^1,2. Individuals with BRCA2 mutations are predisposed to cancer; however, BRCA2-deficient tumours are often responsive to targeted therapy with PARP inhibitors (PARPi)^3-6. The mechanism by which BRCA2 deficiency renders cells sensitive to PARPi but with minimal toxicity in cells heterozygous for BRCA2 mutations remains unclear. Here we identify a previously unknown role of BRCA2 that is directly linked to the effect of PARP1 inhibition. Using biochemical and single-molecule approaches, we demonstrate that PARPi-mediated PARP1 retention on a resected DNA substrate interferes with RAD51 filament stability and impairs RAD51-mediated DNA strand exchange. Full-length BRCA2 protects RAD51 filaments and counteracts the instability conferred by PARPi-mediated retention by preventing the binding of PARP1 to DNA. Extending these findings to a cellular context, we use quantitative single-molecule localization microscopy to show that BRCA2 prevents PARPi-induced PARP1 retention at homologous-recombination repair sites. By contrast, BRCA2-deficient cells exhibit increased PARP1 retention at these lesions in response to PARPi. These results provide mechanistic insights into the role of BRCA2 in maintaining RAD51 stability and protecting homologous-recombination repair sites by mitigating PARPi-mediated PARP1 retention.

Extended Data Fig. 1.. Kinetic analysis of BRCA2 modulating distinct states of RAD51 filament.

(a) Autoradiogram of DNA strand exchange reaction with 3′ tailed DNA and increasing concentrations of RAD51 in the presence of ATP. (b) A pseudo-3D plot, Transition Density plot (TDP), generated from HMM analysis of the smFRET trajectories for RAD51 only bound to 3′ tailed DNA highlighting the transition densities corresponding to each state (N, IE, E). (c) Dwell time histograms (bin size = 90 msec) for each of the FRET states as obtained from HMM analysis (Supplementary Table 1) and FRET population histograms for unprocessed and unnormalized raw data from ebFRET output (n=891; r=3). (d-e) TDP generated from HMM analysis of the smFRET trajectories for BRCA2-RAD51 bound to 3′ tailed DNA highlighting the transition densities corresponding to each state (N′, IE′, E′) along with dwell time histograms (bin size = 90 msec) for each of the FRET states as obtained from HMM analysis (Supplementary Table 1) (n=670; r=3). In both cases, individual states were selected from the TDP, and the corresponding dwell-time histograms were fitted using a single exponential function to obtain the apparent rates of transitions out of the states The number of analyzed molecules and replicates are denoted by n and r, respectively. [BRCA2] = 0.04 μM | [RAD51] = 0.4 μM | [ATP] = 2 mM

Extended Data Fig. 2.. Purification of PARP1 From HEK293T cells and characterization of biochemical activity.

(a) Coomassie blue stain and western blot of purified full-length PARP1. Lane 1,4: Marker, lane 2: 2XMBP-PARP1, lane 3: purified PARP1, lane 5: western blot for PARP1. (b) In vitro PARylation assay using recombinant PARP1 in the presence of NAD⁺ and an activator oligonucleotide (0.2 μM) mimicking DNA strand break. Stain free gel shows PARP1 loading in all lanes. (c) In vitro PARylation inhibition assay using recombinant PARP1 in the presence of NAD⁺, activator oligonucleotide and PARPi. (d) PARylated PARP1 dissociation assay using PARP1 and NAD⁺ and 40bp ds-DNA (0.4 nM). (e) EMSA analyses of PARP1 binding to 0.4 nM of dsDNA and 3′ tailed DNA. (f) EMSA analyses of PARP1 binding to 0.4 nM of ss-DNA. (g) Quantification of EMSA results from e and f. Data shown is mean ± s.d (r=3).

Extended Data Fig. 3.. Kinetic analysis of BRCA2-RAD51 binding on resected DNA in presence of PARP1 + PARPi.

(a-b) A TDP generated from HMM analysis of the smFRET trajectories for RAD51 and PARP1 + PARPi (a) bound to 3′ tailed DNA highlighting the transition densities corresponding to each state (PR+N, IE, E). From the TDP, individual states were selected, and the corresponding dwell-time histograms (bin size = 90 msec) were fitted using a single exponential function to obtain the apparent rates of transitions out of the states along with the FRET population histograms for unprocessed and unnormalized raw data from ebFRET output (b) (n=850; r=3). (c-d) A TDP generated from HMM analysis of the smFRET trajectories for BRCA2-RAD51 and PARP1 + PARPi (c) bound to 3′ tailed DNA with transition states (N′, IE′, E′) and corresponding dwell-time histograms (bin size = 90 msec) along with the FRET population histograms for unprocessed and unnormalized raw data from ebFRET output (d) (n=652; r=4). Brighter colors represent more frequent transitions, and the frequency scale is shown to the right of the graphs. The number of analyzed molecules and replicates are denoted by n and r, respectively. [BRCA2] = 0.04 μM | [RAD51] = 0.4 μM | [ATP] = 2 mM [PARP1] = 0.4 μM | NAD⁺ = 1 mM | [PARPi] = 0.1 μM talazoparib (BMN673)

Extended Data Fig. 4.. Disruption of RAD51 filament by PARPi-induced PARP1 retention potentiated by PARPi potency.

(a-b) Representative smFRET trajectory and normalized FRET histogram of RAD51 only (first panel), RAD51 with PARP1 + talazoparib (second panel), RAD51 with PARP1 + saruparib (third panel), RAD51 with PARP1 + olaparib (fourth panel) and finally RAD51 with PARP1 + veliparib (fifth panel). All four panels with PARPi show three distinct states (PR+N, IE, E), however, the shift in the population distribution is worth noting where talazoparib and saruparib causes maximum disruption to the RAD51 filament followed by olaparib and then little to no disruption by veliparib. (c) Quantification of the effect of different PARPi to the RAD51 filament stability. Lanes: 1: RAD51 only, 2: talazoparib, 3: saruparib (n=433), 4: olaparib (n=460) and 5: veliparib (n=450). Data shown is mean ± s.d (r=2). (d) Schematic and western blot showing bead-capture experiment to measure binding of RAD51 to 167 nucleotide 3′ tailed DNA pre-incubated with PARP1 ± NAD⁺ ± saruparib/olaparib/veliparib. The number of analyzed molecules and replicates are denoted by n and r, respectively. Cartoons were created using BioRender (https://biorender.com). [RAD51] = 0.4 μM | [ATP] = 2 mM [PARP1]_smFRET = 0.4 μM | NAD⁺ = 1 mM | [PARPi] = talazoparib (0.1 μM), saruparib (0.1 μM), olaparib (10 μM) and veliparib (100 μM)

Extended Data Fig. 5.. BRCA2 protects RAD51 filaments from dismantling by FBH1 and RADX.

Normalized FRET histograms of (a) DNA only (first panel), RAD51 only (second panel), RAD51 with FBH1 (third panel) (n=320), BRCA2 with RAD51(fourth panel) and finally BRCA2-RAD51 with FBH1 (fifth panel) (n=276) (b) DNA only (first panel), RAD51 only (second panel), RAD51 with RADX (third panel) (n=446), BRCA2 with RAD51(fourth panel) and finally BRCA2-RAD51 with RADX (fifth panel) (n=391) (c) DNA only (first panel), RAD51 only (second panel), RAD51 with PARP1 + PARPi (third panel), RAD51 with PARP1 + PARPi + FBH1 (n=240) (fourth panel),RAD51 with PARP1 + PARPi + RADX (n=220) (fifth panel) and BRCA2-RAD51 with PARP1 + PARPi (sixth panel). (d) Effect of different replication fork proteins that governs fork stability such as RADX and FBH1 on RAD51 filament stability, cartoon depiction on the right, defining the specificity that PARPi have on PARP1 mediated changes to RAD51 filaments. Data shown is mean ± s.d (r=2). The number of analyzed molecules and replicates are denoted by n and r, respectively. Data for DNA only, DNA with RAD51 and DNA with BRCA2-RAD51 was replotted here for comparison purposes. Cartoons were created using BioRender (https://biorender.com). [BRCA2] = 0.04 μM | [RAD51] = 0.4 μM | [ATP] = 2 mM [RADX] = 0.04 μM | [FBH1] = 0.04 μM [PARP1] = 0.4 μM | NAD⁺ = 1 mM | [PARPi] = 0.1 μM talazoparib (BMN673)

Extended Data Fig. 6.. Characterization of PARP1 interaction with BRCA2 and RAD51.

(a) Control western blots showing bead-capture experiment to measure binding of RAD51 to ss/dsDNA co-incubated with PARP1 ± NAD⁺ ± PARPi. (b) Coomassie stain and western blot for human XRCC1 protein highlighting the quality of the protein used for the protein pull-down assays. (c) First and last washes from the pulldown assays performed in Fig. 3d. (d) Protocol for Protein-A Magnetic Beads pull-down assays for untagged PARP1 with XRCC1, BRCA2 and RAD51 in absence and presence of PARPi. (e) Protein A Magnetic Beads pull-down assay without PARPi. Lane 1: Marker, lane 2: “Control” no IgG1 against PARP1, lane 3: IgG1 against PARP1 + recombinant PARP1 (0.1 μM), lane 4–5: IgG1 for PARP1 + recombinant PARP1 (0.1 μM) ± NAD⁺ (0.25 μM) with XRCC1/2X MBP-BRCA2/RAD51 (0.1 μM each). (f) Protein A Magnetic Beads pull-down assay with PARPi. Lane 1: Marker, lane 2: “Control” no IgG1 against PARP1, lane 3–5: IgG1 against PARP1 + NAD⁺ + PARPi with XRCC1/2XMBP-BRCA2/RAD51. (g) Control experiment demonstrating addition of NAD⁺ or PARPi does not affect BRCA2-RAD51 mediated strand exchange reaction (Lane5,6). Lanes 1–4 are also shown in Fig. 1c. (h) TDP generated from HMM analysis of the smFRET trajectories for PARP1 + NAD⁺ + PARPi bound to 3′ tailed DNA followed by addition of RAD51 highlighting the transition densities corresponding to each state (Nucleation, N) for RAD51 and (PARP1 Retention, PR) for PARP1 respectively. Brighter colors represent more frequent transitions, and the frequency scale is shown to the right of the graphs. (i) Dwell time histograms (bin size = 90 msec) for each of the FRET states as obtained from HMM analysis along with the FRET population histograms for unprocessed and unnormalized raw data from ebFRET output (Supplementary Table 1) (n=632; r=2). The number of analyzed molecules and replicates are denoted by n and r, respectively. [BRCA2] = 0.04 μM | [RAD51] = 0.4 μM | [ATP] = 2 mM [PARP1] = 0.4 μM | NAD⁺ = 1 mM | [PARPi] = 0.1 μM talazoparib (BMN673)

Extended Data Fig. 7.. Characterization of BRCA2-RAD51 and PARP1-RAD51 interaction levels following DSB induction.

(a) Experimental schematic depicting PLA design to quantify levels of BRCA2-RAD51 5 hrs post DNA damage. (b) Representative immunofluorescence images of BRCA2-RAD51 (magenta) PLA foci and DAPI staining to visualize nuclei (blue) for DLD^par cells at 5 hr post CPT (0.1 μM) or IR (12 Gy) treatment (positive control). Quantification for total number of foci and the average foci per nuclease is shown alongside. Individual data points represent single PLA foci per nucleus (n=180). (c) Single Antibody controls for BRCA2, RAD51 and PARP1 antibodies. (d) Experimental schematic depicting PLA design to quantify levels of RAD51-PARP1 5 hrs post DNA damage. Quantification for total number of foci per nucleus is shown alongside as a violin plot (r=2). Individual data points represent single PLA foci per nucleus. Values on graph indicate P-values of unpaired two-sample t-tests between each sample set. DLD^par (n=349) and DLD^−/− (n=350). The number of cells analyzed, and replicates are denoted by n and r, respectively. Cartoons were created using BioRender (https://biorender.com). *P < 0.05, **P < 0.01, and ***P < 0.001 and ****P < 0.0001.

Extended Data Fig. 8.. Characterization of seDSB damage, DDR and initial resection.

(a) U2OS NT and 0.1 μM CPT treated cells were pulse-labeled with 10 μM EdU for 30 min and analyzed by flow cytometry for DNA synthesis (EdU) and DAPI staining for DNA content (10,000 events per sample) (r=2). (b) Experimental schematic of inducing seDSB repair foci and EdU labeling strategy to visualize nascent DNA (naDNA). (Left) Levels of γH2AX (n=95) and 53BP1 (n=82) localizing at nascent DNA 1 hour after CPT treatment. Individual data points represent result from single cell (r=2). (Right) Representative super-resolution images of a single nucleus, post damage, and stained for naDNA (yellow) and 53BP1 (magenta). Scale bar = 2 μm. (c) Western blot analysis for DNA Damage Response (DDR) pathway of U2OS cells treated with 0.1 μM CPT for 1hr. (d) Experimental schematic depicting spatiotemporal mapping of the arrivals, accumulations, and departures of MRE11(n_1Hr=248; n_3Hrs=234) and RPA (n_1Hr=363; n_3Hrs=238) at individual seDSBs. (e-f) Kinetics of MRE11 and RPA localizing at nascent DNA throughout repair over 3 hr recovery (r=2). (g) Representative super-resolution images of a single nucleus, post damage, and stained for naDNA (yellow), MRE11 (green) and RPA (red). Scale bar = 2 μm. Values on graph indicate P-values of unpaired two-sample t-tests between NT and CPT-treated cells. The number of analyzed regions of interest (ROI) and replicates are denoted by n and r, respectively. *P < 0.05, **P < 0.01, and ***P < 0.001 and ****P < 0.0001. Cartoons were created using BioRender (https://biorender.com).

Extended Data Fig. 9.. Time-Course Analysis of BRCA2, RAD51 and PARP1 recruitment in HR repair.

(a) Representative epifluorescence (inset) and super-resolution SMLM-STORM images of a single nucleus, post damage, and stained for naDNA (yellow), RAD51 (cyan) and BRCA2 (magenta). Scale bar = 2 μm. (b) Spatial and temporal mapping of BRCA2, RAD51 and PARP1 proteins respectively, over 16 hr of recovery post damage. Values on the graph represents the percent change in localization of BRCA2, RAD51 and PARP1 to naDNA compared to (nontreated) NT cells. Data shown is mean ± s.d (r=2). (c-e) Kinetics of BRCA2, RAD51 and PARP1 localizing at nascent DNA throughout repair (over 16hrof recovery for BRCA2-RAD51 and 5hr recovery for PARP1). For all quantification graphs individual data points represent result from single cell. Values on graph indicate P-values of unpaired two-sample t-tests between NT, CPT-treated and CPT followed by PARPi treated cells. Cartoons were created using BioRender (https://biorender.com). *P < 0.05, **P < 0.01, and ***P < 0.001 and ****P < 0.0001. [CPT] = 0.1 μM

Extended Data Fig. 10.. Effects of BRCA2 depletion and BRCA2-RAD51 interruption on PARPi-induced PARP1 retention.

(a) Experimental schematic depicting U2OS cells NT and 0.1 μM CPT treated cells were pulse-labeled with 10 μM EdU for 30 min followed by PARPi treatment at the mature phase of recombinase assembly. (b-d) Quantification of levels of RAD51, PARP1 and BRCA2 at naDNA over 5 hr of recovery in U2OS cells WT, compared to (e,f) U2OS cells lacking BRCA2 expression (siBRAC2), in presence and absence of PARPi. (g) Representative super-resolution images of a single nucleus, post damage, and stained for naDNA (yellow), PARP1 (red) and RAD51 (cyan). Scale bar = 2 μm. Western blots showing siRNA-mediated knockdown of BRCA2 in U2OS cells. Stain free gel used as a loading control. (h) Experimental schematic depicting U2OS cells in media containing CAM833 followed by PARPi treatment at the mature phase of recombinase assembly. Cartoon showing CAM833 disrupting the protein-protein interaction between RAD51 and BRCA2. High-magnification SMLM images of damage induced RAD51 filament (cyan) in U2OS cells (left) and their suppression by CAM833 (right) at naDNA (yellow), under the same experimental conditions. Scale bar = 0.2 μm. (i,j) Levels of RAD51 and PARP1 at naDNA over 5 hr of recovery in U2OS cells with CAM833 in media. CAM833 potentiates toxicity and synergizes with PARPi leading to PARP1 retention. We note that for both RAD51 and PARP1 the CPT only and CPT with PARPi data plotted in this panel is also used in panels e,f for comparison. Quantification is from two independent experiments and values on graph indicate P-values of unpaired two-sample t-tests between each sample set. (k,l) Representative super-resolution images of a single nucleus, post damage, and stained for naDNA (yellow), PARP1 (magenta) and RAD51 (cyan) in presence and absence of PARPi. Scale bar = 2 μm. Cartoons were created using BioRender (https://biorender.com). *P < 0.05, **P < 0.01, and ***P < 0.001 and ****P < 0.0001. CPT = 0.1 μM | [CAM833] = 25 μM

Fig. 1.. BRCA2 modulates distinct states of RAD51 filaments.

(a) Coomassie blue stain and western blot of purified full-length 2XMBP BRCA2 (lanes 2, 5) and RAD51 (lanes 3, 7) highlighting the purity and quality of the proteins used for subsequent experiments. (b) Protein pull-down assay of 2XMBP BRCA2 with RAD51. Lane 1: “Control”, RAD51 with 2XMBP only, lane 2: 2XMBP BRCA2 alone, lane 3: RAD51 alone and lane 4: 2XMBP BRCA2 + RAD51. (c) Schematic and autoradiogram of DNA strand exchange reaction with 3′ tailed DNA and RAD51 in the presence or absence of BRCA2 and ATP demonstrating activity of complexes. (d) Schematic depicting the smFRET assay where the dsDNA with a 3′-(dT)₃₀ ssDNA tail is immobilized on a surface. The Cy3 (donor) FRET probe was placed at the dsDNA-ssDNA junction and the Cy5 (acceptor) probe was placed on the ss-DNA tail 16-nucleotide from the junction (left) such that the changes in FRET efficiency (distance between the probes) is indicative of RAD51 filament formation. (e) Representative smFRET trajectories for DNA only, DNA with RAD51 and DNA with BRCA2-RAD51 complex. Idealized FRET trajectory (black) derived from Hidden Markov Modeling (HMM) is superimposed on the FRET trajectory (blue). (f) Normalized FRET histograms of DNA with RAD51 (E_FRET ~ 0.70, 0.53, 0.34, corresponding to the states N, IE, E, respectively; n=943; r=3) and DNA with BRCA2-RAD51 complex (E_FRET ~ 0.65, 0.46, 0.28, corresponding to the N′, IE′, E′, respectively; n=896; r=4). The number of analyzed molecules and replicates are denoted by n and r, respectively. All the illustrations in this work were created using BioRender (https://biorender.com). [BRCA2] = 0.04 μM | [RAD51] = 0.4 μM | [ATP] = 2 mM

Fig. 2.. PARP1 inhibition destabilizes RAD51 filaments, counteracted by BRCA2.

(a) Schematic depicting the smFRET assay of 3′ tailed DNA with PARP1 + NAD⁺ + PARPi + RAD51 ± BRCA2. (b-c) Representative smFRET trajectories (b) and normalized FRET histograms (c) of DNA with PARP1 + NAD⁺ + PARPi + RAD51 (E_FRET ~ 0.70, 0.54, 0.34, corresponding to the states N+PR, IE, E, respectively; n=951; r=3) and PARP1 + NAD⁺ + PARPi + RAD51 + BRCA2 (E_FRET ~ 0.68, 0.45, 0.28, corresponding to the states N′, IE′, E′, respectively; n=795; r=4). The FRET peak corresponding to 0.70 in the RAD51 without BRCA2 sample is a combined population of nucleation (N) and PARP1 retention (PR) on DNA (Supplementary Figs. 3 and 4). (d) Effect of different PARPi on RAD51 filament stability, measured as the precent change in Extension state (E) population, the magnitude follows their respective retention efficiencies (talazoparib = saruparib > olaparib > veliparib) as shown in Supplementary Fig. 11. Data shown is mean ± s.d (r=2). (e) A comparative illustration of the three-state kinetic models of RAD51-BRCA2 in presence and absence of PARP1 and PARPi. States (circles) are connected by forward and reverse rate constants as specified and the area of the circle correlates to the corresponding population of the states. Cartoons were created using BioRender (https://biorender.com). [BRCA2] = 0.04 μM | [RAD51] = 0.4 μM | [ATP] = 2 mM [PARP1]_smFRET = 0.4 μM | NAD⁺ = 1 mM | [PARPi] = talazoparib (0.1 μM), saruparib (0.1 μM), olaparib (10 μM) and veliparib (100 μM)

Fig. 3.. BRCA2 prevents PARPi-induced PARP1 retention on DNA.

(a) Schematic and western blot showing bead-capture experiment to measure binding of RAD51 to 167 nucleotide 3′ tailed DNA pre-incubated with PARP1 ± NAD⁺ ± PARPi. (b) Schematic and western blot showing bead-capture experiment to measure binding of BRCA2-RAD51 to 167 nucleotide 3′ tailed DNA co-incubated with PARP1 ± NAD⁺ ± PARPi. Controls with ds-DNA and ss-DNA is shown in Extended Data Fig. 6a. (c) Single-molecule colocalization (smCL) approach depicting CF647 labeled Halo-PARP1 retention on DNA (% colocalization) ± BRCA2-RAD51. The magnitude was normalized to the maximum co-localization efficiency of PARP1 + PARPi (100 %). Data shown is mean ± s.d (r=2). (d) Schematic protocol and SYPRO orange-stained gels of BER (XRCC1) and HR (RAD51, 2XMBP BRCA2) proteins pull-down experiment with full-length 2XMBP PARP1. First and last washes are shown in Extended Data Fig. 6c. Cartoons were created using BioRender (https://biorender.com). [BRCA2] = 0.04 μM | [RAD51] = 0.4 μM | [ATP] = 2 mM [PARP1] = 0.4 μM | [CF-Halo-PARP1]_smCL = 0.01 μM | NAD⁺ = 1 mM | [PARPi] = 0.1 μM (talazoparib)

Fig. 4.. PARP1-DNA retention interrupts RAD51 strand exchange activity.

(a-b) Schematic depiction of DNA strand exchange protocol. (c) Quantification of DNA strand exchange reaction in presence (black) and absence (red) of PARPi, clearly depicting that PARP1 retention affects the kinetics of RAD51 mediated DNA strand exchange reaction. Controls are shown in Extended Data Fig. 6g, and in this figure, data shown is mean ± s.d (r=3). (d) Schematic depicting the smFRET assay where dsDNA with a 3′-(dT)₃₀ tail was pre-incubated with PARP1, NAD⁺ and PARPi followed by addition of RAD51 protein. (e) Representative smFRET trajectory and (f) normalized FRET histogram (n=632; r=2) of DNA preincubated with PARP1 + NAD⁺ + PARPi followed by addition of RAD51 (E_FRET ~ 0.74 PR, 0.63 N). Two states (PR, N) were assigned to PARP1-RAD51 bound DNA complex, based on the TDP and the dwell time analysis as shown in Extended Data Fig. 6h, i. An illustration of the PARP1-RAD51 two-state kinetic model is also shown alongside where the states (circles) are connected by forward and reverse rate constants as specified. (g) Strand exchange protocol where 3′ tailed DNA was pre-incubated with PARP1 + NAD⁺ + PARPi followed by addition of BRCA2-RAD51 compared to (h) where all three proteins were co-incubated with 3′ tailed DNA in presence of PARPi, followed by the addition of donor DNA. (i) Quantification of (g) and (h). Data shown is mean ± s.d (r=3). Cartoons were created using BioRender (https://biorender.com). [BRCA2] = 0.04 μM | [RAD51] = 0.4 μM | [ATP] = 2 mM [PARP1] = 0.4 μM | NAD⁺ = 1 mM | [PARPi] = 0.1 μM talazoparib (BMN673)

Fig. 5.. BRCA2 prevents PARPi-induced PARP1 retention at HR repair sites.

(a) Experimental schematic depicting PLA design to quantify levels of RAD51-PARP1 5 hr post IR DNA damage. (b). Representative immunofluorescence images of RAD51-PARP1 (magenta) PLA foci and DAPI staining to visualize nuclei (blue) for DLD^par and DLD^−/− cells at 5 hr post IR (12 Gy) treatment. Quantification of the average number of PLA foci per nucleus is shown for DLD^par (n=349) and DLD^−/− (n=350) (n is the number of cells analyzed). Data shown is mean ± s.d (r=2). Statistics tests are unpaired two-sample t-tests between each sample set. Raw data is shown as violin plots in Extended Data Fig.10d highlighting the distribution of the data points. Scale bar = 5 μm. (c) Experimental schematic of analysis of HR repair at CPT induced seDSB repair foci in U2OS cells and EdU labeling strategy to visualize nascent DNA (naDNA). Graphs represent the percent change in localization of PARP1 ± PARPi to naDNA in (nontreated) WT cells compared to cells with no BRCA2 (siBRCA2) and cells where the interaction between BRCA2 and RAD51 is compromised (CAM833). Representative epifluorescence (inset) and super-resolution SMLM-STORM images of a single nucleus, post damage, and stained for naDNA (yellow), RAD51 (cyan), BRCA2 (magenta) and PARP1 (red). Scale bar = 0.2 μm. Associated microscopy images and raw data are shown Extended Data Figs. 9 and 10, plotted as violin plots highlighting the distribution of the data points. (d) Model depicting proposed mechanism of BRCA2 proficient cells tolerating PARPi treatment and synthetic lethal behavior of PARPi in BRCA2 deficient cells. Data shown is mean ± s.d (r=2). For non-normal distributions the Z-scores were calculated from the mean of each sample set and used to obtain the P-values of unpaired two-sample t-tests between each sample set. *P < 0.05, **P < 0.01, and ***P < 0.001 and ****P < 0.0001. Cartoons were created using BioRender (https://biorender.com). CPT = 0.1 μM | [PARPi] = 0.1 μM talazoparib (BMN673) | [CAM833] = 25 μM