Comprehensive RAD51C ovarian cancer variant analysis uncouples homologous recombination and replicative functions

Rein, Hayley L.; Rawal, Yashpal; Palovcak-Lightbourn, Anna L.; Ganesan, Gayatri S.; Parker, Phoebe S.; Russell, Reagan; Darrah, Kristie E.; Afsar, Mohammad; Sullivan, Meghan R.; Hengel, Sarah R.; Radke, Marc R.; Opresko, Patricia L.; Yanowitz, Judith L.; Greene, Eric C.; Lee, Jung-Min; Domchek, Susan M.; Swisher, Elizabeth M.; Olsen, Shaun K.; Sung, Patrick; Bernstein, Kara A.

doi:10.1038/s41467-025-61283-2

Download PDF

Article
Open access
Published: 16 July 2025

Comprehensive RAD51C ovarian cancer variant analysis uncouples homologous recombination and replicative functions

Nature Communications volume 16, Article number: 6539 (2025) Cite this article

3655 Accesses
4 Altmetric
Metrics details

Subjects

Abstract

RAD51C is a tumor suppressor gene with over 285 variants of unknown significance (VUS) found in primary ovarian tumors. RAD51C is a paralog of the recombinase RAD51, and it forms complexes with other paralogs to regulate RAD51 activity. We screened 27 ovarian cancer-derived RAD51C VUS to identify those that affect the assembly of functional tetrameric RAD51B-C-D-XRCC2 (BCDX2) complex. With yeast 3-hybrid and biochemical analyses, we identify a mutation cluster of the RAD51C Walker B region affecting protein interactions with other RAD51 paralogs. By further analyzing these variants for homologous recombination (HR), replication fork regression, DNA binding and ATPase activity, and RAD51 filament formation, we identified separation-of-function alleles that uncouple RAD51C distinct enzymatic activities with HR and replication. Thus, our analysis of RAD51C identifies additional VUS with functional defects, which will aid in pathogenicity classification and inform future strategies to treat individuals harboring RAD51C loss-of-function alleles.

Germline RAD51B variants confer susceptibility to breast and ovarian cancers deficient in homologous recombination

Article Open access 11 October 2021

Structure and function of the RAD51B–RAD51C–RAD51D–XRCC2 tumour suppressor

Article 21 June 2023

RAD51-mediated homologous recombination is a pro-tumour driver pathway

Article Open access 24 September 2025

Introduction

Homologous recombination (HR) is a high-fidelity mechanism of DNA double-strand break (DSB) repair with a strong connection to genome stability and cancer avoidance. HR defects have been identified in over 50% of high grade serous ovarian cancers (HGSOC) stemming from genetic mutations, copy number changes, and increased promoter methylation of related genes¹. BRCA1 and BRCA2, the two most frequently mutated HR genes, are routinely screened for in breast and ovarian tumors to guide precision therapy^2,3. Emerging studies show that mutations in several of the other HR genes, including RAD51C, RAD51D, PALB2, BARD1, BRIP1, may also create cancers susceptible to similar targeted therapies developed for BRCA1 and BRCA2 deficient tumors^4,5. However, many of these genes, such as RAD51C and RAD51D, are less well characterized and thus point mutations identified are typically classified as variants of unknown significance (VUS)^3,4,6. Uncertainty regarding possible pathogenicity of these VUS complicates cancer diagnosis and treatment.

The central step of HR is the formation of RAD51 filaments. These filaments comprise of the RAD51 recombinase wound around single-stranded DNA derived from the nucleolytic processing of DSB ends. These RAD51 nucleoprotein filaments perform a DNA homology search that leads to the engagement and invasion of a homologous chromosome to execute repair^7,8. Many VUS occur in genes that regulate RAD51 activity such as BRCA2, PALB2, RAD51C, and RAD51D^3,9. RAD51C and RAD51D belong to a family of related proteins called the RAD51 paralogs. The RAD51 paralogs aid in preservation of genome stability through regulation of RAD51 activity during HR, and in newly identified functions during DNA replication¹⁰. The canonical RAD51 paralogs consist of RAD51B, RAD51C, RAD51D, XRCC2 and XRCC3, which form two distinct complexes, namely, the BCDX2 complex (RAD51B, RAD51C, RAD51D, XRCC2) and the CX3 complex (RAD51C, XRCC3). These proteins share up to 30% sequence similarity with RAD51 with the majority of this similarity residing in the Walker A and Walker B ATPase motifs¹¹. Along with ATP binding and hydrolysis, the region in and around the conserved Walker A motif of RAD51C was recently identified as crucial for both protein-protein interactions and HR^12,13,14. A similar variant cluster in the Walker B region has yet to be identified.

RAD51C genetic alterations have been identified in 3% of familial ovarian cancers and ovarian cancer risk for RAD51C variant carriers is on par with BRCA2 variant carriers (Odds ratio RAD51C 8.3 vs BRCA2 9.94 [95% CI 5.43–12.48 and 8.09–12.25, respectively])^3,15. Epidemiology studies have identified RAD51C variants in families with high incidences of breast and ovarian cancers that are otherwise wild-type for BRCA1 and BRCA2¹⁶. Patient samples reveal that RAD51C variants are heterozygous within healthy patient tissue, and biallelic in the tumor itself, suggesting a loss-of-heterozygosity (LOH) event¹⁶. These tumors are also p53 deficient^4,16. Several studies show that a subset of RAD51C missense variants are deleterious to RAD51C function as HR impairment and reduction in DNA damage-induced RAD51 foci are observed^{12,14,16,17,18,19,20,21,22}. Furthermore, the same treatment strategies such as the use of PARP inhibitors, can be equally efficacious against BRCA1/BRCA2 and RAD51C-deficient tumors^12,14,21,22.

Although RAD51C loss is closely associated with increased ovarian cancer risk, less than 50% of the reported ovarian cancer derived RAD51C VUS have been functionally screened for an HR defect (Supplementary Data 1). To close this knowledge gap, we recently created a framework to characterize RAD51C variants for altered protein function. Using a yeast-three-hybrid approach, we showed that RAD51C interaction with RAD51D is highly predictive of impaired HR and chemotherapeutic sensitivity¹². In this study, we performed a comprehensive screen of 27 RAD51C ovarian cancer VUS for altered protein function and HR. Our results identify an ovarian cancer variant mutation cluster in and surrounding the Walker B motif. We show that variants in this region exhibit loss of protein function and identify separation-of-function alleles in the Walker B region that uncouple its enzymatic activities during HR and DNA replication. By performing biochemical analysis of reconstituted BCDX2 complex, we demonstrate that while both RAD51C HR and replication functions are needed for RAD51 filament formation, RAD51C HR function is needed for cisplatin and olaparib resistance whereas RAD51C replicative function require its ATPase activity. Our work thus provides evidence for an important role of the RAD51C Walker B region in RAD51C function and indicates that while variants that impact both RAD51C HR and replication function are found in ovarian cancer patients, those who harbor VUS that impact its HR function would most benefit from targeted therapies tailored against HR deficient tumors.

Results

Selection of 27 ovarian cancer VUS for functional analysis

We identified 50 RAD51C missense variants in ovarian cancer either reported in the literature, in cancer databases (COSMIC, TCGA, cBioPortal), or through our collaborations with clinicians (Supplementary Data 1, Supplementary Table 1). A subset of these 50 RAD51C ovarian cancer variants were functionally analyzed for DNA repair defects by our group and others (n = 23/50) and here we analyzed the remaining 27 variants (Summarized in Supplementary Data 1). The RAD51C ovarian cancer variants are primarily germline and a subset were also found in other cancer types (15 total), including breast (6 total). Furthermore, 24 of the cancer variants were also identified in population studies as single nucleotide polymorphisms (SNPs) through exome sequencing (gnomAD; https://gnomad.broadinstitute.org; Supplementary Data 1).

We clinically identified three ovarian cancer patients with RAD51C variants (F164L [c.492 T > G], L238R [c.713 T > G], V239G [c.716 T > A]) that had additional information about their treatment and disease progression. These patients were diagnosed with Stage III (or stage unknown) ovarian cancer and in each case, the RAD51C variants uncovered were germline, or likely germline. Two of these individuals (one with RAD51C-F164L and the other with RAD51C-V239G) showed pedigrees with familial breast and ovarian cancer (Supplementary Table 1, Supplementary Fig. 1). One of these patient families (RAD51C-F164L) also harbored a pathogenic BRCA1 variant, while the other two were BRCA wild-type. Each patient was treated with platinum-based chemotherapy, with the RAD51C-V239G ovarian cancer patient being in remission for 13+ years. In patients with RAD51C-L238R or F164L, the cancer recurred after platinum treatment and therefore continued to receive additional platinum-based therapy along with other targeted therapies including PARP inhibitors, VEGF inhibitors and/or CHK1 inhibitor. The patient with RAD51C-L238R ovarian cancer responded well to CHK1 inhibitor therapy but withdrew from the clinical trial during the COVID-19 pandemic and her status is currently unknown. The individual with RAD51C-F164L died from disease three years post diagnosis. We sought to analyze these patient derived variants as well as the other ovarian cancer variants that were publicly reported for functional changes (Fig. 1b; 27 total).

**Fig. 1: RAD51C ovarian cancer variants are conserved amongst vertebrate species and a subset exhibit reduced protein interactions with its binding partner RAD51D.**

A subset of RAD51C ovarian cancer variants are conserved with other vertebrate species and RAD51

The RAD51C protein is highly conserved amongst vertebrate species, especially within the regions concerned with ATP binding and hydrolysis, including the Walker A and Walker B motifs. Of the 27 ovarian cancer VUS, nine of the residues are highly conserved (identical) in monkey, mouse, zebrafish, and frogs, six are moderately conserved and four somewhat conserved (Fig. 1a). These RAD51C vertebrate-conserved residues are concentrated in the ATPase core domain. RAD51C also shares sequence similarity to RAD51 itself. When comparing RAD51C to RAD51, five of the RAD51C residues are highly conserved in RAD51 and five are somewhat or moderately conserved (Supplementary Fig. 2a).

A subset of RAD51C ovarian cancer variants exhibit reduced protein interaction with RAD51D

RAD51C is an integral member of the BCDX2 complex and directly interacts with RAD51B and RAD51D²³. The BCDX2 complex protein interactions can be readily assessed by yeast-2-hybrid and yeast-3-hybrid^6,24. We previously found that RAD51C yeast-three-hybrid interaction with RAD51D is a significant predictor of HR proficiency (r = 0.85)¹². Therefore, we analyzed RAD51C VUS for interaction with RAD51D by yeast-three-hybrid. Using this approach, we identified five RAD51C variants with 50% or more reduced interaction with RAD51D when compared to wild-type (WT) RAD51C (Fig. 1c, d; RAD51C-T120A [c.358 A > G], -G162A [c.485 G > C], -F164L, -V239G, -A279P [c.835 G > A]). To determine if RAD51C variant interaction loss is due to altered protein expression, we performed western blot analysis using protein extracted from the yeast-three-hybrid experiments. While all the RAD51C interaction deficient variants are expressed, a subset of these variants have reduced protein expression compared to wild-type RAD51C and this may explain the yeast-three-hybrid results (Fig. 1e; Supplementary Fig. 2b).

Testing RAD51C ovarian cancer variant expressing cells for HR

To examine the function of RAD51C ovarian cancer variants of interest, we complemented a CRISPR/Cas9 RAD51C knockout U2OS cell line (KO) with either wild-type RAD51C or the RAD51C variant of interest. We confirmed expression of the RAD51C variants in our polyclonal stable cell line by western blot (Fig. 2a). Note we do not detect expression of RAD51C-Q178E (c.532 C > G) (Fig. 2a).

**Fig. 2: Protein levels and HR analysis of RAD51C ovarian cancer variants expressed in *RAD51C* knockout U2OS cells.**

Next, we determined whether the RAD51C variants are impaired for HR activity. For this, we performed sister chromatid recombination assays on the 27 remaining ovarian cancer variants identified to date. In this assay, a nonfunctional copy of GFP has the unique I-SceI cut site inserted with an upstream GFP fragment. Upon transient expression of the I-SceI enzyme by transfecting the I-SceI endonuclease expressing plasmid, a DSB is created and GFP expression can be restored by use of the GFP fragment as a repair template. Using this assay, we find that five of the 27 RAD51C ovarian cancer variant cell lines were reduced in their HR capabilities when compared to the RAD51C complemented wild-type cell line (Fig. 2b; Q178E, Q181R [c.542 A > C], L238R, A279P, and D318N [c.952 G > A]). Note that while RAD51C-Q181R expressing cells are HR deficient, they are not reduced to the level of a RAD51C knockout cell line (Fig. 2b; indicated in yellow). Furthermore, RAD51C-Q178E is not expressed and therefore is unable to complement the HR deficiency of a RAD51C knockout cell.

Testing RAD51C ovarian cancer variant expressing cells for cell growth

Recently, a saturation mutagenesis screen was used to examine the viability of RAD51C variants in a HAP1 cell line by bulk screening²⁵. Therefore, we asked whether changes in cellular growth is indicative of HR proficiency in a representative subset of the variants (including both wild-type and HR deficient mutants). We analyzed the growth rate of each RAD51C variant cell line by monitoring cell number over time (Supplementary Fig. 3a). Using the growth rate, we also calculated the doubling time for each RAD51C variant tested (Supplementary Fig. 3b). Most of the RAD51C variant expressing cells grew at comparable rates to wild-type complemented RAD51C knockout U2OS cells. These findings are largely consistent with the results from the HAP1 cell line and our previous work demonstrating that RAD51C loss results in decreased cell proliferation^25,26 (Supplementary Fig. 3c). However, Q181R, D318N (c.952 G > A), M10R (c.29 T > G), and V166G (c.497 T > G) from this study and D108Y (c.322 G > T), G125S (c.373 G > A), and V140G (c.419 T > G) from Prakash et al. do not correlate with the findings in HAP1 cells, which may be a functional limitation of the screening approach, a difference between cell lines, or a discrepancy between our findings (²⁵; Supplementary Fig. 3c). Together, these results suggest that viability alone is not a surrogate for functional analysis. Note that comparison between our functional studies and the saturation mutagenesis analysis cannot be used as a basis for calibration in variant classification according to the American College of Medical Genetics guidelines.

Clustering of dysfunctional RAD51C ovarian cancer variants around the Walker B

To analyze where the RAD51C HR deficient ovarian cancer variants are located on the RAD51C structure, we modeled the RAD51C variants on the recently solved cryo-EM, BCDX2-ssDNA complex structure (Fig. 3a, b; PDB ID:8GBJ)^27,28. We previously identified the Walker A region as being crucial for RAD51C HR function¹². Importantly, our current results affirm this finding of HR deficient ovarian cancer variants clustering in the Walker A region (Fig. 3b, c, d[Site II])¹². In addition, we were surprised to find the newly characterized RAD51C variants (A279P, L238R, V239G) to be located in and around the Walker B motif (Fig. 3b, c, d[Site I]). The Walker B motif of RAD51C comprises a beta-strand that spans amino acids 238–242 and includes a highly conserved aspartate residue (D242) that plays a role in ATP binding and typically is involved in ATP hydrolysis. Importantly, two of the variants we identified reside within the motif itself (L238R, V239G) and an additional severely HR deficient variant, A279P, resides within a beta-strand that pairs with the Walker B motif. Therefore, the A279P mutation may disrupt strand pairing in the conserved beta-sheet of the RecA-like fold of RAD51C (Fig. 3c, d[Site II]). Notably, we identified additional HR deficient ovarian cancer variants in the literature in this Walker B region (L219S [c.656 T > C], L226P [c.677 T > G], R237P[c.710 G > C], and D242N [c.724 G > A]) (Fig. 3c, d[Site I])^12,14. Collectively, the deleterious effects of these substitutions are likely due to perturbation of the structure/positioning of the Walker B motif, disruption of the conserved beta-sheet in the RAD51C fold, and/or substitution of residues known to be important for ATP binding and hydrolysis (e.g. D242).

**Fig. 3: The BCDX2 structure reveals a RAD51C variant cluster in the Walker B region.**

Ovarian cancer RAD51C Walker B motif variant, V239G, possesses reduced ATPase activity

To assess the functional relevance of the RAD51C Walker B region, we biochemically analyzed RAD51C-L238R, V239G, and A279P variants for RAD51 paralog complex formation and potential defects of mutant protein complexes in DNA binding and ATP hydrolysis (Fig. 4). Notably unlike wild-type BCDX2, BC^L238RDX2 complex displayed frequent dissociation into BC (RAD51B/RAD51C) and DX2 (RAD51D/XRCC2) subcomplexes during the affinity purification (Fig. 4a). However, we were able to purify BCDX2 complexes with the RAD51C- V239G, and A279P variants as stable complexes (Fig. 4b). Next, we compared ssDNA binding by purified mutant BCDX2 complexes with wild-type BCDX2 using DNA mobility shift assays in the presence of ATP and MgCl₂ (Fig. 4c, d). BCDX2 complexes that harbor the Walker B motif variant V239G exhibit wild-type DNA binding activity, whereas the more distal RAD51C-A279P has a moderate DNA binding defect (Fig. 4c, d). ATPase activity of BCDX2 complex is stimulated by DNA²⁹. Recently, we reported that the BC, but not DX2 subcomplex, can hydrolyze ATP, and BCDX2 complex harboring the RAD51C Walker A motif K131A substitution displays a strong defect in ATP hydrolysis²⁷. Here, we first determined that ATP hydrolysis by the BCDX2 complex also requires a functional RAD51C Walker B motif as RAD51C-D242A substitution greatly impairs ATP hydrolysis by BC^D242ADX2 both in the absence and presence of DNA (Fig. 4e, f). We further discovered that BCDX2 complex containing the Walker B motif variant V239G, has a 50% reduction in ATPase activity in the presence of DNA (Fig. 4e, f). In contrast, BCDX2 containing RAD51C-A279P, which is distal to the Walker B motif, maintains wild-type level of ATPase activity (Fig. 4e, f). Next, we assessed whether the RAD51C Walker B region variants would exhibit defects in CX3 complex formation. Interestingly, affinity purified CX3 complexes containing RAD51C-A279P and RAD51C-L238R emerges only in the void volume in size exclusion chromatography, suggesting that these mutant CX3 complexes form a soluble aggregate. In contrast, C^V239GX3 fractionates between soluble aggregates as well as a monodispersed form, similar to wild-type CX3 complex. These results suggest that RAD51C-V239G can form more stable CX3 complex (Fig. 4g). Together, these results demonstrate that the Walker B region is critical for RAD51C biochemical function.

**Fig. 4: RAD51C Walker B variant impact on DNA binding, ATPase activity, and CX3 complex formation.**

RAD51C Walker B ovarian cancer variant expressing cell lines, L238R and A279P, are sensitive to cisplatin and olaparib

HR deficient tumors are treated with platinum-based drugs like cisplatin, which introduce DNA crosslinks, as well as with PARP inhibitors such as olaparib, which causes PARP trapping and accumulation of ssDNA^30,31,32. Therefore, we asked whether cell lines expressing RAD51C Walker B variants are sensitive to either cisplatin or olaparib. We treated cell lines stably expressing the RAD51C variants of interest (L238R, V239G, and A279P) with increasing concentrations of cisplatin or olaparib and performed clonogenic survival assays (Fig. 5). Even with a low drug concentration range, L238R and A279P expressing cells were killed upon introduction of these drugs. These severely HR deficient variant complemented cells showed exquisite sensitivity to both cisplatin and olaparib with low IC50 values comparable to the RAD51C knockout cell (Fig. 5, Supplementary Data 2; IC50 values L238R = 11.5 ± 3.52 nM, A279P = 48.5 ± 10.6 nM, KO = 32.9 ± 13.8 nM; olaparib IC50 values L238R = 2.28 ± 0.73 nM, A279P = 6.54 ± 1.40 nM, KO = 4.89 ± 3.64 nM). In contrast, RAD51C-V239G does not confer a significant sensitivity to either cisplatin or Olaparib (Fig. 5d). We also performed clonogenic survival assays on a subset of variants in the Walker A region that exhibit a reduction in their interaction with RAD51D and find that this reduction does not confer significant sensitivity to cisplatin or olaparib (Fig. 5). RAD51C-V166G maintains 67% of its RAD51D interaction. Although not significant, we find that T120A, F164L and V239G consistently trend towards olaparib sensitivity (Fig. 5c, d, Supplementary Data 2; IC50 values T120A = 191 ± 46.7 nM, F164L = 187 ± 111 nM, V239G = 112 ± 33.3 nM compared to WT = 696 ± 339 nM)). Note that these variants exhibit HR efficacy of 77%, 73% and 70% HR, respectively (Fig. 2b). Together these results suggest that HR proficiency is largely correlated with olaparib resistance.

**Fig. 5: RAD51C Walker B disrupting variants, L238R and A279P, are sensitive to cisplatin and Olaparib.**

Analysis of replication fork dynamics reveals separation-of-function alleles in the Walker B region

RAD51C functions during HR to promote RAD51 activity as part of the BCDX2 and CX3 complexes and at the same time exhibits HR-independent functions during replication fork dynamics including fork regression^33,34,35. Therefore, we asked whether RAD51C variants in the Walker B region would exhibit defects in replication fork regression. To address this, we performed DNA fiber analysis where a RAD51C knockout U2OS cell uncomplemented or complemented with RAD51C or the variant of interest was first labeled with CldU followed by IdU in the presence or absence of the Topoisomerase I inhibitor camptothecin (CPT; Fig. 6a, Supplementary Fig. 4). Indicative of replication fork regression and reversal defects and consistent with previous findings, loss of RAD51C results in a significant increase in IdU/CldU tract lengths upon CPT in comparison to WT-complemented RAD51C knockout cells, indicating impaired fork arrest and reversal (³³; Fig. 6b). Importantly, cells expressing the HR-proficient RAD51C-V239G variant similarly have a significant decrease in replication fork regression and reversal whereas the two HR deficient variants, L238R and A279P, do not (Fig. 6b). These results suggest that RAD51C HR and replicative functions can be uncoupled and linked to distinct enzymatic activities (Summarized in Fig. 6c).

**Fig. 6: *RAD51C KO* and V239G complemented cell lines have impaired replication fork reversal resulting in ssDNA gap formation.**

Lastly, we sought to determine if the increased tract lengths observed upon loss of RAD51C or upon expression of RAD51C-V239G were due to increased replication fork gaps. To address this, we repeated the IdU and CldU labeling upon CPT in the presence of the S1 nuclease, which digests ssDNA. We find that indeed, both loss of RAD51C and expression of RAD51C-V239G results in ssDNA gaps, thus explaining the increased tract lengths observed (Fig. 6b). In contrast, the HR deficient variants, L238R and A279P, do not accumulate ssDNA gaps (Fig. 6b).

Both RAD51C HR and replicative functions are needed to efficiently form RAD51 filaments by EM

While RAD51C is known to promote RAD51 filament formation²⁷, whether RAD51C HR or replicative functions contribute to its mediator activity has not been addressed to date. Using separation-of-function alleles in RAD51C that uncouple its replicative and HR activities, we sought to determine if either of these functions are required for RAD51 filament formation by negative stain EM. As expected, recombinant WT BCDX2 complex promotes assembly of interconnected RAD51 nucleoprotein filaments in EM analysis by increasing the local concentration of RAD51 and ssDNA (Fig. 7; No vs WT BCDX2). In contrast to WT BCDX2, we find that BCDX2 complexes with either the replication deficient RAD51C variant (V239G) or the HR deficient RAD51C variant (A279P) have significantly lower RAD51 filaments per frame and fewer junctions (Fig. 7a, b, d; WT vs. mutant BCDX2). However, both mutant BCDX2 complexes retain partial ability to promote RAD51 filament assembly from no BCDX2 levels and RAD51 filament length is unchanged in either mutant complex in comparison to wild-type BCDX2 (Fig. 7a–c). Together these findings suggest that both RAD51C HR and replicative functions are necessary to efficiently form RAD51 filaments and uniquely contribute to RAD51 filament assembly.

**Fig. 7: BCDX2 complexes with RAD51C variants, V239G and A279P are defective in RAD51 filaments assembly.**

Discussion

RAD51C is a tumor suppressor and is included on high-risk breast and ovarian cancer screening panels^{11,12,36,37,38}. Fifty RAD51C missense variants have been identified in patients with ovarian cancer and 27 of these variants have little to no functional analysis. Therefore, how each individual variant contributes to cancer development has yet to be determined. Furthermore, how RAD51C distinct functions in HR-dependent DSB repair and HR-independent role during replicative damage contribute to cancer and response to therapy is largely unknown. Here we identified a mutation cluster in the Walker B region of RAD51C. We find that ovarian cancer variants in this region can differentially impact its role in HR and replication-associated repair. By identifying separation-of-function alleles, we find that RAD51C ATPase activity is critical for its role in replication fork regression and reversal. Lastly, we find that RAD51C HR and replication-associated functions both contribute to RAD51 filament formation by EM. Together our results suggest that RAD51C HR and replication-associated functions may uniquely contribute to ovarian cancer development and response to HR-deficient targeted therapy.

The classification of RAD51C variants is critical for risk management, early surveillance, and effective cancer treatment^38,39. In this study, we identified five HR deficient RAD51C variants including one hypomorphic or intermediate RAD51C variant. When we consider our results with our previously published data, the emergence of several patterns in variant clustering are identified¹². First, our results support previous findings of a dysfunctional variant cluster surrounding the Walker A motif. The addition of our new variant analysis establishes the boundaries of this region as being between approximately amino acids 121 and 162. Secondly, through this comprehensive functional analysis, we identified a second critical region for RAD51C in the Walker B motif that is important for its cellular and biochemical functions. Together, our findings uncover new RAD51C HR deficient variants while helping to pinpoint and define new regions and activities of RAD51C that may be important for tumor suppression.

We find the RAD51C Walker B region to be functionally critical for its DNA repair function and response to therapy. Interestingly, our previous analysis of ovarian cancer RAD51C variants did not identify functionally deficient variants in the Walker B motif¹². With the newly published BCDX2 complex structure, we can infer this to be due to the differences in the local structure between these two regions. While, the Walker B motif resides within aβ-pleated sheet and is therefore prone to allosteric effects from more distant variants (e.g. A279P), the Walker A motif resides within a solvent exposed flexible loop that is less susceptible to such allosteric effects from substitutions at positions outside of the motif (Fig. 3b, c, d)^27,28. In this study, we identified two functionally deficient variants in the Walker B (L238R and V239G) as well as a proximal Walker B disrupting variant (A279P). Although A279 does not appear to be located by the Walker B motif based on the primary sequence, the structure reveals that A279 is immediately proximal to the Walker B motif and its mutation likely disrupts the beta pleated sheet structure of this motif (Fig. 3a, d; [Site II]). This finding is further supported by previous analysis which identified two deleterious Walker B region variants R237P and D242N that disrupt HR¹⁴. Additionally, HR deficient variants that alter hydrophobic interactions with V239 in the Walker B motif (L219S and L226P) were also found to be HR deficient¹². Upon further analysis, we find that the V239G variant within the Walker B has a direct impact on ATPase activity and that this variant links this biochemical activity to its replicative function (Fig. 4e, f; Fig. 6). Surprisingly, the Walker B variant, V239G shows a reduction in ATPase activity and replication fork regression but still retains HR efficiency (70%). This is in stark contrast to the more distal Walker B disrupting variant, A279P, which exhibits near wild-type ATPase activity and replication fork progression, but a reduced HR efficacy of 3% (Figs. 2b, 6). Additionally, Walker B variant L238R shown deficient protein-protein interactions both through yeast-three hybrid analysis and upon BC^L238RDX2 and C^L238RX3 purification. These results suggest that the RAD51C Walker B motif not only links its ATPase activity to replication fork regression but is also responsible for a host of functions including protein-protein interactions and DNA binding, which ultimately influence HR efficacy and replicative repair. Recently the CX3 structure was solved from the metazoan Alvinella pompejana³⁵. When analyzing the conserved residues in RAD51C with humans, similar to the BCDX2 complex, the HR deficient variants are clustering between the CX3 interphase and around the ATP binding pocket (Supplementary Fig. 6;³⁵). Future studies are needed to determine how these variants impact CX3 function.

Determining the impact of the amino acid change and intermediate HR phenotypes is critical for understanding cancer risk and patient outcomes. This study identified five HR deficient RAD51C variants (T120A, G162A, F164L, V166G and V239G) with intermediate HR efficiency (70–88% of wild-type RAD51C). These variants are in the Walker A (G162A, F164L, V166G) and Walker B region (T120A, V239G). In addition to variant location and conservation, the amino acid change itself can also be critical for predicting HR function. For example, RAD51C-G162A and G162E have very different HR outcomes (Supplementary 3; 78% and 14%, respectively)¹². Since both glycine and alanine are nonpolar amino acids, an alanine substitution (as in G162A) is likely better tolerated than glutamic acid (as in G162E), which is negatively charged. Therefore, it is possible that a more significant change in charge can create a severe impairment in HR function at amino acid residues where intermediate HR variants currently exist. How a moderate reduction in HR affects ovarian cancer risk is not well understood, but a recent study examining hypomorphic variants in BRCA2 has determined that these variants produce a moderate impact on breast cancer risk⁴⁰. Additionally, our clonogenic survival studies identified an insignificant sensitivity to olaparib in the RAD51C-V239G cell line. In the clinic, an ovarian cancer patient harboring this Walker B variant responded well to platinum-based chemotherapy and is in complete remission with no sign of disease (13+ years; Supplementary Table 1). Additional studies are needed to understand how hypomorphic HR variants and replication deficient variants impact disease development and response to treatment.

Methods

Plasmids, yeast strains and cell lines

A complete list of plasmids and primers used in this study are found in Supplementary Data 1. RAD51C variants were created in the pGAD-RAD51C or pCMV-2B-RAD51C vector using site-directed mutagenesis as in Kondrashova et al. or made by Gene Universal. Yeast-3-hybrid experiments were performed with S. cerevisiae strain PJ69-4α. The RAD51C CRISPR/Cas9 U2OS knockout cell line containing a sister chromatid recombination assay (SCR#18) was generously provided by Mauro Modesti²⁶. These cells were grown in DMEM with 10% fetal bovine serum. U2OS cell lines were cultured in DMEM with 10% FBS and 1% penstrep with or without plasmocin prophylactic (InvivoGen). Cells were grown at 37 °C with 5% CO₂. Polyclonal stable U2OS cell lines containing pCMV-2B-RAD51C or the pCMV-2B-RAD51C variant of interest plasmids were selected by culturing the parental RAD51C knockout U2OS cell line with the indicated plasmid harboring G418 resistant marker in G418-containing medium for two weeks.

Yeast-3-hybrid assays

The indicated pGAD-C1, pGAD-RAD51C wild-type or variant plasmids (1 µl of miniprepped plasmid, roughly 100–500 ng/µl) were transformed into competent yeast, PJ69-4α, with RAD51C binding partners, pGBD-C1-RAD51D and pRS-ADH-416-RAD51B. Transformants were plated on synthetic complete media lacking leucine, tryptophan, and uracil (SC-Leu-Trp-Ura) and incubated for 48–72 h at 30 °C. Three colonies of each variant were inoculated and grown in 5 mL YPD overnight (30° C). Overnight cultures were diluted to 0.2 OD₆₀₀. The cultures were incubated for 3–4 h at 30 °C to 0.5 OD₆₀₀. The 5 µl of 0.5 OD units of culture were plated on SC-Leu-Trp-Ura medium and synthetic complete medium lacking leucine, tryptophan, uracil and histidine (SC-His-Leu-Trp). The plates were incubated for 72 h at 30° C. Plates were imaged and edited for brightness and contrast using Adobe photoshop and yeast spots were quantified using ImageJ software. The relative growth of each RAD51C variant was measured by subtracting the growth of the RAD51C variant with the empty pGBD plasmid. This value was normalized to wild-type RAD51C which was set to 100% as the maximum value.

Yeast western blot analysis

Three colonies of yeast expressing the indicated plasmids were inoculated for growth overnight at 30 °C in 3 mL of YPD. The 0.75 OD₆₀₀ of cells were pelleted and washed in 1 mL of water. The cells were lysed by TCA precipitation and the protein pellet was resuspended in 50 µl of loading buffer (HU) as described (Knop, 1999) except that the HU buffer was supplemented with DTT. Fifteen µL of protein was run on a 4–20% gradient SDS-PAGE gel at 120 V for 75 min. The protein was transferred to a PVDF membrane at 100 V for 2 h and the membrane was blocked for 1 h using Oddessy Licor TBS blocking buffer. RAD51C and Kar2 loading control was visualized on a Licor CLX scanner using a RAD51C antibody (Abcam Cat# ab55728, RRID:AB_945135, 1:500), Kar2 antibody (sc-33630, 1:2000) and secondary antibodies (Licor IRDye 680RD goat anti-rabbit 926-68071 and Licor IRDye 800CW goat anti-rabbit 926-32211).

Mammalian western blot analysis

One million cells of each RAD51C knock out or variant complemented cell line were seeded in a 60 mm plate and collected 24 h later. Cells were lysed in RIPA buffer supplemented with PMSF, 1X PhosStop and 1X protease inhibitor cocktails (ThermoFischer) and 0.2 µL Benzonase (Sigma). Cells were incubated on ice for 30 min. Lysates were pelleted and 20–40 µL of supernatants were run on a 10% SDS-PAGE gel at 120 V for 75 min. The protein was transferred to a PVDF membrane at 100 V for 2 h and the membrane was blocked for 1 h using Oddessy Licor TBS blocking buffer. RAD51C and αTubulin loading control was visualized on a Licor CLX scanner using a RAD51C antibody (Abcam Cat# ab55728, RRID:AB_945135, 1:500), αTubulin antibody (Cell Signaling Antibody#2144S, 1:1000) and secondary antibodies (Licor IRDye 680RD goat anti-rabbit 926-68071 and Licor IRDye 800CW goat anti-rabbit 926-32211).

Sister chromatid recombination assays

RAD51C knockout U2OS cells alone or stably expressing wild-type RAD51C or the RAD51C variant of interest were transfected with 4 µg of plasmid expressing I-SceI (pCBSceI) using 5 µL Lipofectamine2000. Three days post-transfection, the cells were collected and immediately analyzed or fixed in 2% PFA for analysis. The percentage of HR performed was determined by measuring GFP expression of 20,0000 events by flow cytometry (CytoFlex 6 L;). Percentage of GFP was normalized to wild-type RAD51C and RAD51C knockout cell line was ran as a negative control.

Growth rate and doubling times

Three hundred thousand RAD51C knockout U2OS cells alone or stably expressing wild-type RAD51C or the RAD51C variant of interest were seeded in a 12 well plate. The cells were collected from individual wells and counted every day for four days. The number of cells collected each day was plotted as the growth rate. Doubling times were calculated for each 24-h period using the following equation:

$${{{\rm{Doubling}}}}\; {{{\rm{Time}}}}= {{{\rm{duration}}}} * {{{\mathrm{ln}}}}(2)/\log ({{{\rm{final}}}}\; {{{\rm{concentration}}}}) \\ -\log ({{{\rm{initial}}}}\; {{{\rm{concentration}}}})$$

(1)

For each cell line, the fastest doubling time, indicative of log phase for that cell line, was graphed.

Clonogenic survival assays

Six hundred RAD51C knockout U2OS cells alone or stably expressing wild-type RAD51C or the RAD51C variant of interest were seeded on a 60 mm plate in triplicate. The next day, cells were treated with cisplatin (Sensitive cell lines; 0, 0.015625, 0.03125, 0.0625, 0.125 and 0.25 µM. Insensitive cell lines; 0.125, 0.25, 0.5, 1, and 2 µM) or olaparib (Sensitive cell lines; 0, 0.015625, 0.03125, 0.0625, 0.125 and 0.25 µM. Insensitive cell lines; 0.125, 0.25, 0.5, 1, and 2 µM) containing media. The cells were incubated in cisplatin-containing media for two days (one cell cycle), after which the cells were grown in fresh media for two weeks. Olaparib containing media was continuously replaced every three days for a two-week treatment. Olaparib containing media was supplemented with equal volumes of DMSO (vehicle). At the end of either the cisplatin or olaparib treatment, the cells were washed with PBS, fixed with 100% methanol for 20 min and stained with Crystal Violet for two hours. Plates were imaged using a BioRad imager, and the colony area was quantified in ImageJ. Colony area data was normalized to the no-treatment control for each cell line. Colony survival was graphed vs cisplatin or olaparib concentration. The data were fit to the following IC50 model:

$$({{{\rm{Y}}}}={{{\rm{Bottom}}}}+({{{\rm{Top}}}}-{{{\rm{Bottom}}}})/(1+{10 ^\wedge }{(({{{\rm{Log}}}}({{{\rm{IC}}}}50)-{{{\rm{Log}}}}({{{\rm{X}}}})))}))$$

(2)

where top is 100% survival and bottom is the final % survival in Graphpad Prism. Data are reported as IC50 values with standard error of the mean.

Protein expression and purification

BCDX2 and CX3 purification was carried out as previously described¹². Briefly BCDX2 (RAD51B-His and XRCC2-Flag) and CX3 (MBP-RAD51C and XRCC3-FLAG) were expressed in Hi5 insect cells using baculoviruses and all the purification steps were carried out at 0–4 °C. Cell pellets were resuspended in T buffer (25 mM Tris-HCl, pH 7.5, 10% glycerol, 0.5 mM EDTA, 1 mM DTT, 0.05% IGEPAL CA-630 (Sigma), 1 mM PMSF and protease inhibitors) containing 300 mM KCl (T300), 2 mM ATP, and 2 mM MgCl₂ and lysed by sonication. BCDX2 complex was affinity purified from Ni-NTA resin (Qiagen) and for further purification protein was fractionated on HiTrap Q HP column with 150-600 mM KCl gradient and concentrated pooled fractions were subject to size exclusion chromatography in a Superose6 Increase 10/300 GL column in T300 buffer with 2 mM each of ATP and MgCl₂. CX3 complex was first affinity purified from amylose resin (NEB) and then anti-FLAG resin (Sigma) and eluted protein complex was concentrated and subject to size exclusion chromatography in a Superdex200 Increase 10/300 GL column in T300 buffer with 2 mM each of ATP and MgCl₂.

DNA binding and ATPase assays

For DNA binding, 1 nM of 5’ Cy5-80-nt ssDNA (Cy5-iSp9-TTATGTTCATTTTTTATATCCTTTACTTTATTTTCTCTGTTTATTCATTTACTTATTTTGTATTATCCTTATCTTATTTA) was incubated with the indicated concentration of purified BCDX2 complex in 10 μl reaction buffer (50 mM Tris-HCl, pH 7.5, 155 mM KCl, 1 mM DTT, 1 mM ATP, 1 mM MgCl₂ and 100 ng/μl BSA) for 10 min at 37 °C. Nucleoprotein complexes were resolved on 5% polyacrylamide gels in Tris-borate buffer (45 mM each). Gels were visualized using the ChemiDoc imaging system (Bio-Rad) and proportion of bound versus unbound DNA was quantified using the ImageJ software.

BCDX2 complexes or its sub-complexes (1 μM) were incubated with or without viral ssDNA (20 μM nucleotides, phiX174 virion) in 10 μl of reaction buffer (20 mM HEPES, pH 7.5, 1 mM DTT, 1 mM MgCl₂ and 30 mM KCl) containing 0.05 mM ATP with 0.25 μCi [γ−³²P] ATP at 37 °C. At indicated time points aliquots of 2 μl were withdrawn and mixed with 2 μl of 0.5 M EDTA to stop the reaction. The ATP hydrolysis was determined by thin layer chromatography on PEI cellulose F sheets (Millipore, 105579) in 0.375 M potassium phosphate (pH 3.5), followed by phosphorimaging analysis with Amersham Typhoon phosphorimager (Cytiva) and the ImageQuant software (Cytiva).

DNA fiber assays

DNA fiber assays were performed as previously described (DNA Fiber Analysis: Mind the Gap!, Quinet et al). In short, 10⁵ cells were plated in each well of a 12-well plate. 24 h later, cells were washed with PBS and media was replaced with fresh media mixed with 20 µM Cldu. After 20 min of incubation with Cldu, media was replaced with fresh media containing 200 µM IdU and 100 nM camptothecin (CPT). Cells were incubated for 1 h with IdU/CPT. Media was then removed and cells were washed with PBS and permeabilized with CSK100 buffer (100 mM NaCl, 10 mM MOPS, pH 7, 3 mM MgCl2, 300 mM sucrose, 0.5% Triton X-100) at room temperature for 7 min, followed by S1 nuclease digestion in S1 buffer (30 mM sodium acetate, pH 4.6, 10 mM zinc acetate, 5% glycerol, and 50 mM NaCl) at 37 °C for 20 min. Cells were then collected by scraping, pelleted by centrifugation, and resuspended in 100-200 µl PBS. 2 µl of cell suspension was dropped onto positively charged slides, lysed with lysis buffer for 6 min (0.5% SDS, 200 mM Tris-HCl, pH 7.4, and 50 mM EDTA) and spread by tilting slides at a 45° angle. Slides were then fixed in 3:1 methanol:acetic acid for 10 min, then washed with PBS, and denatured in 2.5 M HCl for 1 h. Slides were then neutralized in 400 mM Tris-HCl (pH 7.4), washed with PBS, and blocked with 5% BSA/10% goat serum. Slides were then treated with primary antibody (1:150-Rat anti-Cldu, 1:40-Mouse anti-IdU) for 1 h. After washing, slides were treated with secondary antibody (goat anti-rat 647, goat anti-mouse 488, 1:100 for both). Slides were then washed with PBS and mounted with ProLong Gold antifade reagent. Slides were then imaged with a Nikon Eclipse Ti2 microscope. Fiber lengths were then measured using DNA Stranding software⁴¹.

Electron microscopy

For presynaptic filament assembly, RAD51 (1 μM) was incubated with 150-nt oligonucleotide (3 μM nucleotides) with or without 25 nM BCDX2 variants in reaction buffer (25 mM HEPES pH 7.5, 25 mM KCl, 1 mM MgCl2, and 1 mM ATP) at 37 °C for 30 min. A droplet of 4 μl of a reaction mixture was blotted onto a glow-discharged carbon-coated grid (EMS CF300 Cu) for 1 min. After blotting with filter paper, the grid was immersed in a 20 μl droplet of the staining solution (2% uranyl acetate). The stain was removed from the grid by blotting with filter paper and the grid was washed twice with water for 1 min. Following excess stain removal, grids were dried in air for about 3 min. Negatively stained samples of hRAD51 presynaptic filaments were examined by transmission electron microscopy (TEM). Images were processed in Fiji ImageJ image analysis package (v2.90/1.53t) and filaments were quantified using ImageJ plugin Ridge detection.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All data supporting the findings of this study are available within the paper and its Supplementary Information. Source data are provided with this paper.

References

Konstantinopoulos, P. A., Ceccaldi, R., Shapiro, G. I. & D’Andrea, A. D. Homologous Recombination Deficiency: Exploiting the Fundamental Vulnerability of Ovarian Cancer. Cancer Discov. 5, 1137–1154 (2015).
Article CAS PubMed PubMed Central Google Scholar
Maxwell, K. N. et al. BRCA locus-specific loss of heterozygosity in germline BRCA1 and BRCA2 carriers. Nat. Commun. 8, 319 (2017).
Article ADS PubMed PubMed Central Google Scholar
Pennington, K. P. et al. Germline and somatic mutations in homologous recombination genes predict platinum response and survival in ovarian, fallopian tube, and peritoneal carcinomas. Clin. Cancer Res. 20, 764–775 (2014).
Article CAS PubMed Google Scholar
Kondrashova, O. et al. Secondary Somatic Mutations Restoring RAD51C and RAD51D Associated with Acquired Resistance to the PARP Inhibitor Rucaparib in High-Grade Ovarian Carcinoma. Cancer Discov. 7, 984–998 (2017).
Article CAS PubMed PubMed Central Google Scholar
Ollier, M. et al. DNA repair genes implicated in triple negative familial non-BRCA1/2 breast cancer predisposition. Am. J. Cancer Res. 5, 2113–2126 (2015).
CAS PubMed PubMed Central Google Scholar
Baldock, R. A. et al. RAD51D splice variants and cancer-associated mutations reveal XRCC2 interaction to be critical for homologous recombination. DNA Repair (Amst.) 76, 99–107 (2019).
Article CAS PubMed Google Scholar
Jasin, M. & Rothstein, R. Repair of strand breaks by homologous recombination. Cold Spring Harb. Perspect. Biol. 5, a012740 (2013).
Article PubMed PubMed Central Google Scholar
Godin, S. K., Sullivan, M. R. & Bernstein, K. A. Novel insights into RAD51 activity and regulation during homologous recombination and DNA replication. Biochem. Cell Biol. 94, 407–418 (2016).
Article CAS PubMed PubMed Central Google Scholar
Tung, N. et al. Frequency of Germline Mutations in 25 Cancer Susceptibility Genes in a Sequential Series of Patients With Breast Cancer. J. Clin. Oncol. 34, 1460–1468 (2016).
Article CAS PubMed PubMed Central Google Scholar
Rein, H. L., Bernstein, K. A. & Baldock, R. A. RAD51 paralog function in replicative DNA damage and tolerance. Curr. Opin. Genet Dev. 71, 86–91 (2021).
Article CAS PubMed PubMed Central Google Scholar
Sullivan, M. R. & Bernstein, K. A. RAD-ical New Insights into RAD51 Regulation. Genes (Basel) 9 (2018). https://doi.org/10.3390/genes9120629
Prakash, R. et al. Homologous recombination-deficient mutation cluster in tumor suppressor RAD51C identified by comprehensive analysis of cancer variants. Proc. Natl. Acad. Sci. USA 119, e2202727119 (2022).
Article CAS PubMed PubMed Central Google Scholar
Braybrooke, J. P., Spink, K. G., Thacker, J. & Hickson, I. D. The RAD51 family member, RAD51L3, is a DNA-stimulated ATPase that forms a complex with XRCC2. J. Biol. Chem. 275, 29100–29106 (2000).
Article CAS PubMed Google Scholar
Hu, C. et al. Functional and clinical characterization of variants of uncertain significance identifies a hotspot for inactivating missense variants in RAD51C. Cancer Res. https://doi.org/10.1158/0008-5472.CAN-22-2319 (2023).
Article PubMed PubMed Central Google Scholar
LaDuca, H. et al. A clinical guide to hereditary cancer panel testing: evaluation of gene-specific cancer associations and sensitivity of genetic testing criteria in a cohort of 165,000 high-risk patients. Genet. Med. 22, 407–415 (2020).
Article CAS PubMed Google Scholar
Meindl, A. et al. Germline mutations in breast and ovarian cancer pedigrees establish RAD51C as a human cancer susceptibility gene. Nat. Genet. 42, 410–414 (2010).
Article CAS PubMed Google Scholar
Clague, J. et al. RAD51C germline mutations in breast and ovarian cancer cases from high-risk families. PLoS One 6, e25632 (2011).
Article ADS CAS PubMed PubMed Central Google Scholar
Thompson, E. R. et al. Exome sequencing identifies rare deleterious mutations in DNA repair genes FANCC and BLM as potential breast cancer susceptibility alleles. PLoS Genet. 8, e1002894 (2012).
Article CAS PubMed PubMed Central Google Scholar
Osorio, A. et al. Predominance of pathogenic missense variants in the RAD51C gene occurring in breast and ovarian cancer families. Hum. Mol. Genet 21, 2889–2898 (2012).
Article CAS PubMed Google Scholar
Blanco, A. et al. RAD51C germline mutations found in Spanish site-specific breast cancer and breast-ovarian cancer families. Breast Cancer Res. Treat. 147, 133–143 (2014).
Article CAS PubMed Google Scholar
Sullivan, M. R. et al. Long-term survival of an ovarian cancer patient harboring a RAD51C missense mutation. Cold Spring Harb Mol Case Stud 7 (2021). https://doi.org/10.1101/mcs.a006083
Somyajit, K., Subramanya, S. & Nagaraju, G. Distinct roles of FANCO/RAD51C protein in DNA damage signaling and repair: implications for Fanconi anemia and breast cancer susceptibility. J. Biol. Chem. 287, 3366–3380 (2012).
Article CAS PubMed Google Scholar
Masson, J. Y. et al. Identification and purification of two distinct complexes containing the five RAD51 paralogs. Genes Dev. 15, 3296–3307 (2001).
Article CAS PubMed PubMed Central Google Scholar
Schild, D., Lio, Y. C., Collins, D. W., Tsomondo, T. & Chen, D. J. Evidence for simultaneous protein interactions between human Rad51 paralogs. J. Biol. Chem. 275, 16443–16449 (2000).
Article CAS PubMed Google Scholar
Olvera-Leon, R. et al. High-resolution functional mapping of RAD51C by saturation genome editing. Cell 187, 5719–5734 e5719 (2024).
Article CAS PubMed Google Scholar
Garcin, E. B. et al. Differential Requirements for the RAD51 Paralogs in Genome Repair and Maintenance in Human Cells. PLoS Genet 15, e1008355 (2019).
Article CAS PubMed PubMed Central Google Scholar
Rawal, Y. et al. Structural insights into BCDX2 complex function in homologous recombination. Nature 619, 640–649 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Greenhough, L. A. et al. Structure and function of the RAD51B-RAD51C-RAD51D-XRCC2 tumour suppressor. Nature https://doi.org/10.1038/s41586-023-06179-1 (2023).
Article PubMed PubMed Central Google Scholar
Sigurdsson, S. et al. Mediator function of the human Rad51B-Rad51C complex in Rad51/RPA-catalyzed DNA strand exchange. Genes Dev. 15, 3308–3318 (2001).
Article CAS PubMed PubMed Central Google Scholar
Lheureux, S., Braunstein, M. & Oza, A. M. Epithelial ovarian cancer: Evolution of management in the era of precision medicine. CA Cancer J. Clin. 69, 280–304 (2019).
PubMed Google Scholar
Dasari, S. & Tchounwou, P. B. Cisplatin in cancer therapy: molecular mechanisms of action. Eur. J. Pharm. 740, 364–378 (2014).
Article CAS Google Scholar
Cong, K. et al. Replication gaps are a key determinant of PARP inhibitor synthetic lethality with BRCA deficiency. Mol. Cell 81, 3227 (2021).
Article CAS PubMed Google Scholar
Berti, M. et al. Sequential role of RAD51 paralog complexes in replication fork remodeling and restart. Nat. Commun. 11, 3531 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Somyajit, K., Saxena, S., Babu, S., Mishra, A. & Nagaraju, G. Mammalian RAD51 paralogs protect nascent DNA at stalled forks and mediate replication restart. Nucleic Acids Res 43, 9835–9855 (2015).
CAS PubMed PubMed Central Google Scholar
Longo, M. A. et al. RAD51C-XRCC3 structure and cancer patient mutations define DNA replication roles. Nat. Commun. 14, 4445 (2023).
Article ADS CAS PubMed PubMed Central Google Scholar
Kuznetsov, S. G., Haines, D. C., Martin, B. K. & Sharan, S. K. Loss of Rad51c leads to embryonic lethality and modulation of Trp53-dependent tumorigenesis in mice. Cancer Res. 69, 863–872 (2009).
Article CAS PubMed PubMed Central Google Scholar
Pelttari, L. M. et al. RAD51C is a susceptibility gene for ovarian cancer. Hum. Mol. Genet. 20, 3278–3288 (2011).
Article CAS PubMed Google Scholar
Sopik, V., Akbari, M. R. & Narod, S. A. Genetic testing for RAD51C mutations: in the clinic and community. Clin. Genet. 88, 303–312 (2015).
Article CAS PubMed Google Scholar
Liu, J., Konstantinopoulos, P. A. & Matulonis, U. A. Genomic testing and precision medicine–What does this mean for gynecologic oncology?. Gynecol. Oncol. 140, 3–5 (2016).
Article PubMed Google Scholar
Shimelis, H. et al. BRCA2 Hypomorphic Missense Variants Confer Moderate Risks of Breast Cancer. Cancer Res. 77, 2789–2799 (2017).
Article CAS PubMed PubMed Central Google Scholar
Li, L. et al. Automatic DNA replication tract measurement to assess replication and repair dynamics at the single-molecule level. Bioinformatics 38, 4395–4402 (2022).
Article CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

This work was supported by the National Institutes of Health grants [R01 ES031796 (K.A.B.), R01 ES030335 (K.A.B.), F31 CA264889 (H.L.R.), R01 ES030396 (P.L.O.), GM115568 (S.K.O), GM128731 (S.K.O), R35 CA241801 (P.S.), R35 GM118026 (E.C.G.), R01 CA236606 (E.C.G.), P01 CA275717 (E.C.G. & P.S.), R01 CA293655 (E.C.G. & S.K.O.), and R01 GM127569 (J.L.Y.)], the American Cancer Society (DBG-24-1243614-01-DMC) to K.A.B., the Penn Center for Genome Integrity, the Basser Center for BRCA and The Intramural Research Program of the Center for CCR of NCI, NIH (ZIA BC011525, to J.-M.L.). A.P. is supported by NIEHS (5T32ES019851-10). S.K.O. is the recipient of a CPRIT Rising Star Award (RR200030). S.R.H. is supported by the NIEHS (K99/R00-ES033738), F31 ES027321 (M.R.S.).

Author information

These authors contributed equally: Hayley L. Rein, Yashpal Rawal.

Authors and Affiliations

University of Pittsburgh, School of Medicine, Department of Pharmacology and Chemical Biology, Pittsburgh, PA, USA
Hayley L. Rein, Reagan Russell & Sarah R. Hengel
University of Texas Health Science Center at San Antonio, Department of Biochemistry and Structural Biology, Greehey Children’s Cancer Research Institute, San Antonio, TX, USA
Yashpal Rawal, Mohammad Afsar, Shaun K. Olsen & Patrick Sung
University of Pennsylvania School of Medicine, Department of Biochemistry and Biophysics, Philadelphia, PA, USA
Anna L. Palovcak-Lightbourn, Gayatri S. Ganesan, Phoebe S. Parker, Kristie E. Darrah & Kara A. Bernstein
University of Pittsburgh, School of Medicine, Department of Microbiology and Molecular Genetics, Pittsburgh, PA, USA
Meghan R. Sullivan
Department of Biology, Tufts University, Medford, MA, USA
Sarah R. Hengel
Division of Gynecologic Oncology, Department of Obstetrics and Gynecology, University of Washington School of Medicine, Seattle, WA, USA
Marc R. Radke & Elizabeth M. Swisher
Department of Environmental and Occupational Health, University of Pittsburgh School of Public Health, Pittsburgh, PA, USA
Patricia L. Opresko
UPMC Hillman Cancer Center, Pittsburgh, PA, USA
Patricia L. Opresko
University of Pittsburgh School of Medicine, Department of Obstetrics, Gynecology, and Reproductive Science, Pittsburgh, PA, USA
Judith L. Yanowitz
Department of Biochemistry and Molecular Biophysics, Columbia University Medical Center, New York, NY, USA
Eric C. Greene
Women’s Malignancies Branch, Center for Cancer Research, National Cancer Institute, Bethesda, MD, USA
Jung-Min Lee
Basser Center for BRCA, Abramson Cancer Centre, University of Pennsylvania School of Medicine, Philadelphia, PA, USA
Susan M. Domchek

Authors

Hayley L. Rein
View author publications
Search author on:PubMed Google Scholar
Yashpal Rawal
View author publications
Search author on:PubMed Google Scholar
Anna L. Palovcak-Lightbourn
View author publications
Search author on:PubMed Google Scholar
Gayatri S. Ganesan
View author publications
Search author on:PubMed Google Scholar
Phoebe S. Parker
View author publications
Search author on:PubMed Google Scholar
Reagan Russell
View author publications
Search author on:PubMed Google Scholar
Kristie E. Darrah
View author publications
Search author on:PubMed Google Scholar
Mohammad Afsar
View author publications
Search author on:PubMed Google Scholar
Meghan R. Sullivan
View author publications
Search author on:PubMed Google Scholar
Sarah R. Hengel
View author publications
Search author on:PubMed Google Scholar
Marc R. Radke
View author publications
Search author on:PubMed Google Scholar
Patricia L. Opresko
View author publications
Search author on:PubMed Google Scholar
Judith L. Yanowitz
View author publications
Search author on:PubMed Google Scholar
Eric C. Greene
View author publications
Search author on:PubMed Google Scholar
Jung-Min Lee
View author publications
Search author on:PubMed Google Scholar
Susan M. Domchek
View author publications
Search author on:PubMed Google Scholar
Elizabeth M. Swisher
View author publications
Search author on:PubMed Google Scholar
Shaun K. Olsen
View author publications
Search author on:PubMed Google Scholar
Patrick Sung
View author publications
Search author on:PubMed Google Scholar
Kara A. Bernstein
View author publications
Search author on:PubMed Google Scholar

Contributions

H.L.R., Y.R., A.L.P., G.G., P.P., R.R. and K.D. all conducted experiments. H.L.R., G.G. and M.S. contributed to manuscript writing. M.A. conducted data analysis and helped with figure creation. S.R.H. helped with data analysis. M.R.R., J.M.L., S.M.D. and E.M.S. supplied patient information. P.L.O., J.L.Y. and E.C.G. provided aid in conceptual experimental design. S.K.O., P.S. and K.A.B. designed experiments, and aided in interpreting results and writing the manuscript.

Corresponding author

Correspondence to Kara A. Bernstein.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

: Nature Communications thanks Matthew Wakefield and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Reporting Summary

Transparent Peer Review file

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Rein, H.L., Rawal, Y., Palovcak-Lightbourn, A.L. et al. Comprehensive RAD51C ovarian cancer variant analysis uncouples homologous recombination and replicative functions. Nat Commun 16, 6539 (2025). https://doi.org/10.1038/s41467-025-61283-2

Download citation

Received: 23 July 2023
Accepted: 17 June 2025
Published: 16 July 2025
DOI: https://doi.org/10.1038/s41467-025-61283-2