Abstract
BRCA1, BRCA2, and PALB2 are key players in cellular tolerance to chemotherapeutic agents, including camptothecin, cisplatin, and PARP inhibitor. The N-terminal segment of BRCA2 interacts with PALB2, thus contributing to the formation of the BRCA1–PALB2–BRCA2 complex. To understand the role played by BRCA2 in this complex, we deleted its N-terminal segment and generated BRCA2ΔN mutant cells. Although previous studies have suggested that BRCA1–PALB2 plays a role in the recruitment of BRCA2 to DNA-damage sites, BRCA2ΔN mutant cells displayed a considerably milder phenotype than did BRCA2−/− null-deficient cells. We hypothesized that the DNA-binding domain (DBD) of BRCA2 might compensate for a defect in BRCA2ΔN that prevented stable interaction with PALB2. To test this hypothesis, we disrupted the DBD of BRCA2 in wild-type and BRCA2ΔN cells. Remarkably, although the resulting BRCA2ΔDBD cells displayed a moderate phenotype, the BRCA2ΔN+ΔDBD cells displayed a very severe phenotype, as did the BRCA2−/− cells, suggesting that the N-terminal segment and the DBD play a substantially overlapping role in the functionality of BRCA2. We also showed that the formation of both the BRCA1–PALB2–BRCA2 complex and the DBD is required for efficient recruitment of BRCA2 to DNA-damage sites. Our study revealed the essential role played by both the BRCA1–PALB2–BRCA2 complex and the DBD in the functionality of BRCA2, as each can compensate for the other in the recruitment of BRCA2 to DNA-damage sites. This knowledge adds to our ability to accurately predict the efficacy of antimalignant therapies for patients carrying mutations in the BRCA2 gene. Cancer Res; 74(3); 797–807. ©2013 AACR.
Introduction
Homologous recombination plays a central role in maintaining genomic DNA as well as cellular tolerance to various DNA-damaging agents (1, 2). Specifically, homologous recombination restores stalled replication at damaged template strands by using the intact sister chromatid as a template for DNA synthesis (3–6) and repairs any double-strand breaks (DSB) that may occur in one of the sister chromatids during replication or as a result of chemotherapeutic treatments with camptothecin, cisplatin, and PARP inhibitor. Individual homologous recombination factors contribute in different ways to various types of homologous recombination (7, 8). In homologous recombination–dependent DSB repair, nucleases initiate homologous recombination by processing DSBs to generate 3′ overhangs, followed by the polymerization of RAD51 recombinase on the overhang (9–11). The resulting RAD51 polymer is responsible for homology search and subsequent strand exchange with intact homologous duplex DNA. RAD51 is required for every type of homologous recombination reaction, and inactivation of RAD51 in the chicken DT40 cell line causes numerous spontaneous chromosomal breaks (12). The polymerization of RAD51 is strictly regulated by a number of RAD51 mediator proteins, including BRCA1, BRCA2, Gemin2 (13), the 5 RAD51 paralogs, RAD52, SFR1, and SWS1 (14, 15).
The breast-cancer-susceptibility gene 2 (BRCA2) encodes a tumor-suppressor protein carrying 3,418 amino acids. Although the overall amino-acid identity between the human and the chicken BRCA2 protein is only ∼40%, several regions are conserved, including the N-terminal segment (approximately 100 amino acids), 8 BRC motifs, a carboxy-proximal region containing a DNA/DSS1-binding (DDB) domain, and the C-terminal end (16). The BRC motifs, the DDB domain, and the C-terminal end may involve RAD51 loading onto DNA, an association with damaged DNA sites, and an interaction with RAD51, respectively (17–22). We and others have previously reported on BRCA2-null–deficient DT40 cells (hereafter called BRCA2−/− cells), the DT40 cells that express a mutant BRCA2 protein truncated at the C-terminal of the BRC3 motif (BRCA2BRC3tr cells; refs. 14, 15, 23, and 24), as well as the DT40 cells that carry the deletion of the C-terminal end (BRCA2ΔC cells; ref. 17). BRCA2BRC3tr and BRCA2ΔC cells displayed a significantly milder phenotype than did BRCA2−/− cells. Likewise, BRCA2BRC3tr mice can survive to adulthood, whereas BRCA2−/− mice died early in embryogenesis (25). These observations, as well as, other reports (26) suggest that neither the DDB domain nor the C-terminal end are absolutely essential for promoting homologous recombination by BRCA2. The contribution of the N-terminal segment to DNA-damage tolerance is unclear, because mutant cells lacking only this segment in the endogenous BRCA2 gene have not yet been documented.
PALB2 was originally identified as a protein that physically interacts at the N-terminal segment of BRCA2 (27, 28). Mutations in the BRCA2 and PALB2 genes cause the Fanconi anemia subtypes FA-D1 and FA-N, respectively (28–31). These observations reveal a possible functional interaction between BRCA2 and PALB2. Moreover, PALB2 binds directly to BRCA1 (32, 33) irrespective of DNA damage, indicating the stability of the BRCA1–PALB2–BRCA2 complex. However, it is unclear whether or not the role of PALB2 in homologous recombination is entirely dependent on BRCA2. In fact, purified PALB2 is able to promote RAD51-dependent strand exchange in the absence of BRCA2 (34, 35). The biological significance of the BRCA1–PALB2–BRCA2 complex in homologous recombination remains to be investigated.
We undertook a comprehensive genetic study to address the following 2 questions: (i) Does the functionality of PALB2 totally depend on BRCA2? (16) What role does the BRCA1–BRCA2–PALB2 complex play in the promotion of homologous recombination by BRCA2? (16) To address the first question, we created PALB2−/− and BRCA2−/−/PALB2−/− cells. The data indicate that the role played by PALB2 in homologous recombination is totally dependent on BRCA2. This echoes studies showing that the role played by BRCA1 in homologous recombination is dependent on BRCA2 (14). To address the second question, we deleted the BRCA2 N-terminal segment in wild-type, BRCA2BRC3tr, and BRCA2ΔDBD cells. Although the resulting BRCA2ΔN cells displayed only modest phenotypes, the BRCA2ΔN+BRC3tr, BRCA2ΔN+ΔDBD, and BRCA2−/− cells displayed very severe (and indistinguishable) phenotypes. In summary, our genetic data reveal an interdependency among BRCA1, BRCA2, and PALB2, wherein the promotion of homologous recombination by PALB2 as well as by BRCA1 requires BRCA2, and the full functionality of BRCA2 is enhanced by the BRCA1–PALB2–BRCA2 complex. Moreover, the BRCA1–PALB2–BRCA2 complex and the DNA-binding domain (DBD) of BRCA2 play substantially overlapping roles in the recruitment of BRCA2 to DNA-damage sites.
Materials and Methods
Cell lines, cell culture, and DNA transfection
DT40 cells were cultured as described by Sonoda and colleagues (12). DNA transfection and selection were performed as described previously (36). A description of DT40 mutants used in this study is provided in Table 1.
Cell line . | Selection marker for gene disruption . | Plating efficiency (%) . | Reference . |
---|---|---|---|
Wild-type | – | 100 | Ref. 36 |
BRCA2−/− | Puro-Bsr | 6 | Ref. 14 |
PALB2−/− | Bsr-Puro | 80 | This study |
BRCA2−/−/PALB2−/− | -/Bsr-Puro | 5 | This study |
BRCA2ΔN | His | 28 | This study |
BRCA2BRC3tr | His | 27 | Ref. 24 |
BRCA2ΔDBD | Bsr | 35 | This study |
BRCA2ΔN+BRC3tr | His-Bsr | 8 | This study |
Cell line . | Selection marker for gene disruption . | Plating efficiency (%) . | Reference . |
---|---|---|---|
Wild-type | – | 100 | Ref. 36 |
BRCA2−/− | Puro-Bsr | 6 | Ref. 14 |
PALB2−/− | Bsr-Puro | 80 | This study |
BRCA2−/−/PALB2−/− | -/Bsr-Puro | 5 | This study |
BRCA2ΔN | His | 28 | This study |
BRCA2BRC3tr | His | 27 | Ref. 24 |
BRCA2ΔDBD | Bsr | 35 | This study |
BRCA2ΔN+BRC3tr | His-Bsr | 8 | This study |
Western blot
The antibodies used to detect the green fluorescent protein (GFP) in the immunoblotting assay were polyclonal rabbit anti-GFP (MBL) and donkey anti-rabbit HRP (Santa Cruz Biotechnology) as primary and secondary antibodies, respectively. The antibodies used to detect the BRCA2 protein were the same as those used previously (24).
Flow cytometric analysis
To measure growth kinetics, cells were counted daily using flow cytometric analysis, as described previously (37).
Measurement of cell sensitivity to damaging agents
To measure cellular sensitivity, cells were exposed for 1 day to mitomycin C (Sigma), and for 3 days to camptothecin (TopoGEN), cis-diamminedichloroplatinum (cisplatin; Nippon Kayaku), etoposide (Funakoshi), and olaparib (inhibitor of PARP; JS Research Chemical), or irradiated with ionizing radiation (137Cs). The number of cells was measured at 72 hours using CellTiter-Glo Luminescent Cell Viability Assay (Promega) or counted with FACScan (BD Biosciences). Sensitivity was calculated by dividing the number of surviving cells treated with damaging agents by the number of untreated cells (14, 38).
Immunofluorescent visualization of RAD51 subnuclear focus formation
Cells were collected on a slide glass using Cytospin3 (Shandon). Cells were fixed by 4% paraformaldehyde for 10 minutes, followed by permeabilization with 0.1% NP-40 in PBS for 10 minutes. Fixed cells were stained with rabbit polyclonal anti-human RAD51 (1:1,000, CALBIOCHEM; EMD Biosciences) for 1 hour, then with anti-rabbit Ig Alexa Fluor 488 (1:3,000; Molecular Probes) for 1 hour. Slides were stained and mounted using Vectashield mounting medium (Vector Laboratories). The number of RAD51 foci was counted in a minimum of 100 morphologically intact cells at each time point after gamma irradiation. Images were taken with an Olympus BX-61 microscope (Olympus Corp.) with a CoolSNAP CCD camera unit (Photometrics) and Metamorph Software (Molecular Devices). Subsequent adjustment and resizing of images were done with Adobe Photoshop (Adobe Systems).
Immunofluorescent visualization of GFP-tagged protein recruitment
To visualize BRCA2 in immunocytochemical analysis, we inserted a GFP tag into the C-terminal end of an endogeneous BRCA2 allelic gene. To stain the GFP-tagged proteins, we used rabbit polyclonal anti-GFP (1:1,000; MBL) for 1 hour, then anti-rabbit Ig Alexa Fluor 488 for 1 hour (1:3,000; Molecular Probes).
Chromosome analysis of mitotic cells
To enrich mitotic cells, cells were treated with 0.1 mg/m: colcemid (Invitrogen) for 3 hours and then fixed. Metaphase spreads were prepared as described previously (12). Additional details are provided in Supplementary material and methods.
Results
PALB2−/− cells display a significantly milder phenotype than do BRCA2−/− cells
We generated recombinant plasmids to disrupt the PALB2 gene by inserting selection markers in exon 2 of the PALB2 gene (Supplementary Fig. S1A). The disruption of the PALB2 gene was verified by genomic Southern blotting and reverse transcriptase PCR (RT-PCR) analysis (Supplementary Fig. S1B and S1C). PALB2−/− cells exhibited a milder phenotype than did BRCA2−/− cells in terms of both cell-cycle kinetics and spontaneously arising chromosomal aberrations in mitotic cells (Supplementary Fig. S3 and Table S2; ref. 14).
We assessed the role played by PALB2 in the DNA-damage response by measuring the cellular sensitivity of PALB2−/− cells to various DNA-damaging agents (Fig. 1A). As with BRCA2−/− cells, PALB2−/− cells showed an increased sensitivity to a wide variety of DNA-damaging agents, including ionizing radiation, an inhibitor of PARP (olaparib), the topoisomerase poisons (camptothecin and etoposide) and chemical-crosslinking agents [cisplatin (cis-diamminedichloroplatinum (II)), and mitomycin C; Fig. 1A]. In addition to this hypersensitivity, decreased homologous recombination–mediated DSB repair in the SCneo construct (Fig. 1B; ref. 14) indicated that PALB2 plays a role in general homologous recombination, as does BRCA2 (14, 39–43). It should be noted that the phenotype of the PALB2−/− cells was significantly milder than that of the BRCA2−/− cells.
To analyze the functionality of PALB2 as a RAD51 mediator, we analyzed RAD51 subnuclear focus formation at 3 hours after ionizing radiation (Fig. 1C and D). The loss of PALB2 diminished the induction of RAD51 focus formation, which observation agrees with the phenotype of PALB2-depleted mammalian cells (44). Thus, PALB2 contributes to homologous recombination as a RAD51 mediator in DT40 as well as in mammalian cells. The loss of PALB2 had a lesser impact on RAD51 focus formation than did the loss of BRCA2. In conclusion, PALB2 and BRCA2 play a role in general homologous recombination as a RAD51 mediator, with PALB2 contributing to homologous recombination to a lesser extent than BRCA2.
The deletion of the PALB2 gene has no impact on the phenotype of BRCA2−/− cells
The phenotypic similarity between BRCA2−/− and PALB2−/− clones as RAD51 mediators led us to explore a possible functional interaction between BRCA2 and PALB2. To this end, we generated BRCA2−/−/PALB2−/− cells. The BRCA2−/− and BRCA2−/−/PALB2−/− clones displayed a very similar phenotype in terms of both proliferation kinetics (Supplementary Fig. S3) and spontaneous chromosomal aberrations (Table 2). This suggests that PALB2 requires BRCA2 to maintain chromosomal integrity. Furthermore, the inactivation of PALB2 had no obvious impact on the sensitivity of the BRCA2−/− cells to any of the DNA-damaging agents (Fig. 1A). This observation reveals that the functionality of PALB2 in the DNA-damage response is totally dependent on BRCA2. This also echoes the fact that the functionality of BRCA1, the 5 RAD51 paralogs, RAD52, SFR1, and SWS1 in homologous recombination is totally dependent on BRCA2 (14, 15).
Genotype . | Chromatid type . | Chromosome type . | Exchange . | Total aberrations (per cell ± SE)a . |
---|---|---|---|---|
Wild-type | 2 | 4 | 0 | 0.06 ± 0.02 |
PALB2−/− | 0 | 8 | 0 | 0.08 ± 0.02 |
BRCA2−/− | 6 | 24 | 6 | 0.36 ± 0.06 |
BRCA2−/−/PALB2−/− | 8 | 22 | 2 | 0.32 ± 0.05 |
BRCA2ΔN | 2 | 8 | 0 | 0.10 ± 0.03 |
BRCA2BRC3tr | 6 | 6 | 0 | 0.12 ± 0.03 |
BRCA2ΔDBD | 4 | 10 | 0 | 0.14 ± 0.03 |
BRCA2ΔN+BRC3tr | 5 | 20 | 2 | 0.27 ± 0.05 |
Genotype . | Chromatid type . | Chromosome type . | Exchange . | Total aberrations (per cell ± SE)a . |
---|---|---|---|---|
Wild-type | 2 | 4 | 0 | 0.06 ± 0.02 |
PALB2−/− | 0 | 8 | 0 | 0.08 ± 0.02 |
BRCA2−/− | 6 | 24 | 6 | 0.36 ± 0.06 |
BRCA2−/−/PALB2−/− | 8 | 22 | 2 | 0.32 ± 0.05 |
BRCA2ΔN | 2 | 8 | 0 | 0.10 ± 0.03 |
BRCA2BRC3tr | 6 | 6 | 0 | 0.12 ± 0.03 |
BRCA2ΔDBD | 4 | 10 | 0 | 0.14 ± 0.03 |
BRCA2ΔN+BRC3tr | 5 | 20 | 2 | 0.27 ± 0.05 |
aSE was calculated based on Poisson distribution.
Although BRCA2ΔN and BRCA2BRC3tr cells show modest phenotypes, BRCA2ΔN+BRC3tr cells show a very severe phenotype
Because the N-terminal segment physically interacts with PALB2, which thereby mediates a complex formation between BRCA1 and BRCA2 (27, 32, 33), we next explored the function of the N-terminal segment by deleting exons 3 to 7 (encoding amino acids 23 to 209) in the endogenous BRCA2 gene (Supplementary Fig. S2A). We deleted the same N-terminal segment in the BRCA2BRC3tr cells, thereby generating BRCA2ΔN+BRC3tr cells. To assess the stability of the BRCA2BRC3tr and BRCA2ΔN+BRC3tr proteins, we transiently transfected transgenes encoding these truncated proteins tagged at the C-terminal end with the GFP gene into DT40 cells and measured the expression of the transgenes by Western blotting using anti-GFP antibody (Supplementary Fig. S2D). The data indicated that the deletion of the N-terminal region does not affect the stability of the BRCA2BRC3tr protein.
The BRCA2ΔN cells proliferated with nearly normal kinetics, showing only a slight increase in the number of spontaneous chromosomal aberrations (Supplementary Fig. S3 and Table S2). Likewise, the BRCA2ΔN mutant cells displayed a considerably lower sensitivity to the DNA-damaging agents (Fig. 2A) and a more efficient homologous recombination (Fig. 2B) than did the BRCA2−/− cells. Thus, although previous studies have suggested that BRCA1–PALB2 is essential for the recruitment of BRCA2 to DNA-damage sites (27, 32, 33), even in the absence of a tight physical association with BRCA1–PALB2, the BRCA2ΔN protein can still significantly contribute to homologous recombination. The data are reminiscent of the fact that although RAP80 was shown to be essential for the recruitment of BRCA1 to DNA-damage sites by immunocytochemical analysis, even in the absence of RAP80, BRCA1 protein can still fully promote homologous recombination (45). Accordingly, the physical association of BRCA2 with BRCA1–PALB2 might not play an important role in functionality of BRCA2. Alternatively, a part of the BRCA2 protein other than the N-terminal segment might compensate for the missing physical association.
To address the latter possibility, we generated BRCA2ΔN+BRC3tr cells. In contrast with the very mild phenotypes of the BRCA2ΔN and BRCA2BRC3tr cells, the BRCA2ΔN+BRC3tr cells displayed a very severe phenotype, whereas the BRCA2ΔN+BRC3tr and BRCA2−/− clones showed a similar number of spontaneous chromosomal aberrations (Table 2). Moreover, the BRCA2ΔN+BRC3tr and BRCA2−/− clones displayed the same degree of sensitivity to all DNA-damaging agents tested (Fig. 3A and Supplementary Fig. S4). These observations support the idea that either the N-terminal region or the truncated part of the BRCA2BRC3tr protein is sufficient for significant promotion of homologous recombination–mediated repair by BRCA2. Thus, the physical association of BRCA2 with BRCA1–PALB2 does play an important role in functionality of BRCA2, whereas the important role is masked by the functional overlap between the N-terminal region and the truncated part of the BRCA2BRC3tr protein.
Although BRCA2ΔN and BRCA2ΔDBD clones show a modest phenotype, BRCA2ΔN+ΔDBD cells show a very severe phenotype
The next question was, which truncated sequences of the BRCA2BRC3tr protein compensated for the absence of the N-terminal segment in the BRCA2ΔN protein? We hypothesized that the DBD (19) might provide the answer. To test this hypothesis, we disrupted the DBD in BRCA2ΔN as well as in wild-type cells and generated BRCA2ΔN+ΔDBD and BRCA2ΔDBD clones. We achieved the disruption of the DBD by deleting nucleotide sequences encoding 19 amino acids included in two oligonucleotide/oligosaccharide-binding folds (amino acids 3009 to 3017 together with 3074 to 3083) in the endogenous BRCA2 gene. Deletion was confirmed by Southern blotting and RT-PCR analysis (Supplementary Fig. S2A–S2C). These deletions had no effect on the stability and the cellular distribution of the BRCA2ΔN, BRCA2ΔDBD, and BRCA2ΔN+ΔDBD proteins (Supplementary Fig. S2E and S6). The mutation that we generated in DBD domain does not affect the distribution of BRCA2 protein, which is consistent with the recent publication in (46).
The BRCA2ΔDBD and BRCA2ΔN cells showed only moderate increases in the number of spontaneous chromosomal aberrations (Table 2) and in sensitivity to DNA damage (Figs. 2A and 3B), which moderate phenotype was in marked contrast to the very severe phenotype of the BRCA2−/− cells. Likewise, significant numbers of RAD51 foci induced by ionizing radiation were detected in these mutant cells, which finding is in marked contrast to the virtual absence of RAD51 foci induced by ionizing radiation in the BRCA2−/− cells (Fig. 2C and D). Remarkably, the combined mutations of BRCA2ΔN and BRCA2ΔDBD (BRCA2ΔN+ΔDBD) caused a phenotype very similar to that of the BRCA2−/− cells in terms of spontaneously arising chromosomal breaks (Table 2), capability of homologous recombination–mediated DSB repair in SCneo (Fig. 2B), and RAD51 focus formation induced by ionizing radiation (Fig. 2C and D).
We measured cellular sensitivity to camptothecin, cisplatin, etoposide, ionizing radiation, mitomycin C, and olaparib, comparing between BRCA2ΔN, BRCA2ΔDBD, BRCA2ΔN+ΔDBD, and BRCA2−/− cells (Figs. 2A, 3B, and Supplementary Fig. S4). BRCA2ΔDBD cells were slightly more sensitive to these DNA-damaging agents than BRCA2ΔN cells and showed a phenotype similar to that of BRCA2BRC3tr cells (Fig. 3A). Importantly, the combined mutations of BRCA2ΔN and BRCA2ΔDBD caused a synergistic increase in cellular sensitivity (Fig. 3B), as seen in the combined mutations of BRCA2ΔN and BRCA2BRC3tr (Fig. 3A). This indicates that the N-terminal segment of BRCA2 and the DBD of BRCA2 have substantially overlapping functions in homologous recombination. Remarkably, the BRCA2ΔN+ΔDBD and BRCA2−/− clones displayed the same degree of sensitivity to the DNA-damaging agents, indicating that either the N-terminal segment or the DBD is sufficient for the promotion of homologous recombination by BRCA2. It is possible that the function shared by the N-terminal segment and the DBD is the physical association with DNA-damage sites, because this association can be achieved via an interaction between the N-terminal segment and the BRCA1–PALB2 complex as well as via the DBD of BRCA2 (19, 33).
The N-terminal segment and the DBD work together to recruit the BRCA2 protein to DNA-damage sites
We hypothesized that the N-terminal segment as well as the DBD might contribute to the recruitment of BRCA2 to DNA-damage sites. To test this hypothesis, we visualized the recruitment of BRCA2 by inserting DNA sequences encoding a GFP tag into the endogenous BRCA2wt, BRCA2ΔN, BRCA2ΔDBD, and BRCA2ΔN+ΔDBD genes, measuring GFP-focus formation in the resulting wild-type, BRCA2ΔN, BRCA2ΔDBD, and BRCA2ΔN+ΔDBD cells after ionizing radiation (Fig. 3C and D). BRCA2 recruitment to DNA-damage sites was impaired in the BRCA2ΔN and BRCA2ΔDBD mutants. However, we observed a few clear foci in the BRCA2ΔN and BRCA2ΔDBD cells, which observation is consistent with the tolerance of these mutant cells to the DNA-damaging agents described above. In agreement with the null-phenotype of BRCA2ΔN+ΔDBD cells, they showed no detectable BRCA2 recruitment. In summary, the BRCA2 protein can accumulate at DNA-damage sites even if it loses either its N-terminal segment or its DBD (Fig. 3C and D). Furthermore, either the PALB2 protein or the DBD is essential for the recruitment of BRCA2 to DNA-damage sites.
The recruitment of BRCA2 by the N-terminal segment requires both BRCA1 and PALB2
The N-terminal segment of BRCA2 has been shown to perform intrinsic transcriptional activities (16, 47) and to have an association with PALB2 (33). To selectively analyze the latter function, we generated BRC3-truncated BRCA2 (BRCA2BRC3tr) transgenes carrying point mutations (W31R and W31C) that specifically inhibit the interaction between BRCA2 and PALB2 (27). The ectopic expression of the BRCA2BRC3tr transgene, but not the resulting BRCA2W31R+BRC3tr or BRCA2W31C+BRC3tr mutant transgenes, reversed the phenotype of the BRCA2−/− cells, including decreases in cellular tolerance to camptothecin and cisplatin (Fig. 4A). Thus, the interaction with PALB2 may be essential for the promotion of homologous recombination by the BRCA2BRC3tr protein. To test this idea, we ectopically expressed the BRCA2BRC3tr transgene in BRCA2−/− and PALB2−/−/BRCA2−/− clones. The BRCA2BRC3tr transgene significantly suppressed the mutant phenotype in the BRCA2−/− cells but not in the PALB2−/−/BRCA2−/− cells (Fig. 4A), indicating that PALB2 is required to promote homologous recombination by BRCA2BRC3tr. Taking into account the fact that BRCA2 is required for the promotion of homologous recombination by PALB2 (Fig. 1), we conclude that there is a previously unappreciated functional interdependency between BRCA2 and PALB2.
Because PALB2 mediates between BRCA1 and BRCA2 to form a complex, the BRCA2BRC3tr protein might also require BRCA1 to promote homologous recombination. To test this idea, we ectopically expressed the BRCA2BRC3tr transgene in BRCA1−/−/BRCA2−/− cells. The BRCA2BRC3tr transgene failed to reverse the mutant phenotype of the BRCA1−/−/BRCA2−/− cells (Fig. 4A). We therefore conclude that the contribution of the BRCA2BRC3tr protein to homologous recombination is entirely dependent on the stable complex formed by the BRCA2BRC3tr protein with both BRCA1 and PALB2. These 2 proteins might thus be essential for the recruitment of the BRCA2BRC3tr protein to damaged DNA sites, as suggested previously (32, 33).
To investigate the role played by the BRCA1–PALB2–BRCA2 complex in the recruitment of BRCA2 to DNA-damage sites, we undertook 2 experiments. First, we ectopically expressed the GFP-tagged BRCA2BRC3tr, BRCA2W31R+BRC3tr, and BRCA2W31C+BRC3tr transgenes in BRCA2−/− cells, exposed the resulting cells to irradiation with ionizing radiation and then performed immunostaining with anti-GFP antibody (Fig. 4B and C). Second, we ectopically expressed the GFP-tagged BRCA2BRC3tr transgene in PALB2−/−/BRCA2−/− and BRCA1−/−/BRCA2−/− cells and performed immunostaining. The BRCA2BRC3tr transgene, but not the BRCA2W31R+BRC3tr or the BRCA2W31C+BRC3tr transgene, caused GFP-focus formation, indicating that a BRCA2–PALB2 interaction is required for efficient BRCA2 focus formation. In addition, the BRCA2BRC3tr protein was not recruited to DNA-damage sites in either the PALB2−/−/BRCA2−/− or the BRCA1−/−/BRCA2−/− clone (Fig. 4B and C), which observation agrees with previous studies (32, 33). We conclude that both BRCA1 and PALB2 are required for BRCA2-focus formation when the DBD is absent.
Discussion
We have shown that although BRCA2ΔN and BRCA2ΔDBD clones have only moderate phenotypes, the combined ΔN+ΔDBD mutations totally abolished the functionality of BRCA2 and resulted in synergistic increases in cellular sensitivity to DNA-damaging agents (Fig. 3B). This observation indicates that the deleted N-terminal segment and the DBD might share a common function. This function is likely to be a physical interaction with DNA-damage sites, because both BRCA2ΔN and BRCA2ΔDBD cells showed compromised recruitment of mutated BRCA2 proteins to DNA-damage sites (Fig. 3C and D). In summary, our data reveal that the N-terminal segment and the DBD of the BRCA2 protein have overlapping functions, most likely the physical interaction of BRCA2 with the DNA-damage site. Conceivably, the N-terminal segment and the DBD may collaborate for the most appropriate positioning of BRCA2 at DNA-damage sites, which positioning may allow for the most efficient promotion of RAD51 polymerization (Fig. 5).
The present genetic study provides compelling evidence that BRCA2 and PALB2 are interdependent. This conclusion is supported by the fact that the BRCA2−/− and PALB2−/−/BRCA2−/− cells display the same phenotype (Fig. 1). Possibly, BRCA2 is required for the appropriate positioning of PALB2 at DNA-damage sites in vivo (14). The promotion of homologous recombination by the BRCA2BRC3tr protein totally depends on BRCA2's interaction with PALB2. This conclusion can be drawn from 2 experiments (Fig. 4). First, the BRCA2BRC3tr protein significantly enhanced homologous recombination in BRCA2−/− cells but had no effect at all in PALB2−/−/BRCA2−/− cells. Second, the W31R and W31C mutations of the BRCA2 BRC3-truncated protein prevented the interaction of the resulting mutant proteins with PALB2 and also inhibited the promotion of homologous recombination. In summary, our study highlights the critical role played by the interaction between the N-terminal segment of BRCA2 and PALB2 in the promotion of homologous recombination by both BRCA2 and PALB2 (Fig. 5).
We previously showed that BRCA2 has an epistatic relationship with BRCA1 (14). We here show that the BRCA2BRC3tr protein does not function at all in the absence of BRCA1 or PALB2. These observations demonstrate previously unknown interdependencies between BRCA1 and BRCA2 as well as between PALB2 and BRCA2. Because the BRCA1−/− and PALB2−/− clones (14) displayed a stronger phenotype than did the BRCA2ΔN cells, we concluded that both BRCA1 and PALB2 are able to contribute to homologous recombination, even without the formation of a stable BRCA1–PALB2–BRCA2 complex. Another question is how BRCA1 affects BRCA2- and PALB2-mediated promotion of RAD51 polymerization. BRCA1 seems to play a role in the resection of DSBs, possibly by counteracting the inhibitory effect of 53BP1 and Rif1 (48) on the resection. Presumably, the functional association of BRCA1 with BRCA2 and PALB2 allows for the efficient collaboration between the BRCA1-dependent resection of DSBs and the subsequent loading of RAD51 on the resulting overhang by BRCA2 and PALB2.
In summary, this study, as well as our previous study (14), uncovers the following functional relationships among BRCA1, BRCA2, and PALB2 (Fig. 5). First, the recruitment of BRCA2 to DNA-damage sites is promoted by a stable complex between BRCA2 and both BRCA1 and PALB2. Second, the promotion of RAD51 polymerization by BRCA1 and PALB2 is totally dependent on BRCA2. Third, the DBD can substitute for a defect in the stable complex. These findings contribute to the prediction of the efficacy of antimalignant therapies on malignant tumors carrying a variety of mutations in the BRCA1, BRCA2, and PALB2 genes.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Authors' Contributions
Conception and design: M. Al Abo, D. Dejsuphong, S. Takeda
Development of methodology: M. Al Abo, D. Dejsuphong
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): M. Al Abo, D. Dejsuphong, K. Hirota, Y. Yonetani, H. Kurumizaka
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Al Abo, D. Dejsuphong, S. Takeda
Writing, review, and/or revision of the manuscript: M. Al Abo, S. Takeda
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Al Abo, M. Yamazoe
Study supervision: M. Al Abo, S. Takeda
Acknowledgments
The authors thank Dr. G.-L. Moldovan for his helpful discussions, L. Moreau for technical support, and the members of the Department of Radiation Genetics for their comments.
Grant Support
S. Takeda was financially supported, in part, by the JSPS Core-to-Core Program, Advanced Research Networks, as well as a Grant-in-Aid for Scientific Research (S) from the JSPS.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.