The tumor suppressor BRCA1 functions in DNA homologous recombination, and mutations in BRCA1 increase the risk of breast and ovarian cancers. RAP80 is a component of BRCA1-containing complexes that is required for recruitment of BRCA1 to sites of DNA damage. To evaluate the role of RAP80 in DNA damage repair, we genetically disrupted both RAP80 alleles in the recombinogenic avian DT40 cell line. The resulting RAP80−/− cells were proficient at homologous recombination and nonhomologous end-joining (NHEJ), but were specifically sensitized to the topoisomerase II inhibitor etoposide. Notably, doubly mutant RAP80−/−BRCA1−/− cells were more sensitive to etoposide than were BRCA1−/− cells, revealing that RAP80 performs a BRCA1-independent repair function. Moreover, jointly impairing the function of CtIP, a distinct BRCA1 effector protein, rendered RAP80−/− cells more sensitive to etoposide compared with singly mutant cells, again illustrating a BRCA1-independent role of RAP80. Based on our findings, we propose that RAP80 exerts a specific function in repair of the topoisomerase-cleavage complex, such as the removal of covalently bound polypeptides from double-strand break ends independently of BRCA1. Cancer Res; 70(21); 8467–74. ©2010 AACR.
Antimalignant therapy such as radiotherapy and treatment with inhibitors against topoisomerase I and II (Top1 and Top2) kills cycling cells by inducing double-strand breaks (DSB; ref. 1). DSBs are repaired by two major pathways: homologous recombination and nonhomologous end-joining (NHEJ; ref. 2). Homologous recombination repairs DSBs by using the intact homologous sequence as a template, and is initiated by a process called resection, which generates single-strand tails at DSBs. This initial resection is promoted by the Mre11/RAD50/NBS1 complex and CtIP (3–5). Subsequently, Rad51 assembles with the single-strand DNA with the help of Rad51 mediators such as BRCA1 and BRCA2 (6, 7). NHEJ does not involve the resection step and is initiated by the assembly of Ku70/Ku80 at DSBs (8).
Top1 and Top2 induce transient DNA breaks in distinctly different manners. Top1 creates single-strand breaks and binds to the 3′ end of DNA, whereas Top2 creates DSBs and binds to the 5′ ends of DNA (1). Topoisomerase inhibitors are divided into two categories: the topoisomerase poisons and the topoisomerase catalytic inhibitors (9, 10). Top2 poisons such as etoposide stabilize the covalent bond between a tyrosine of Top2 and the 5′ ends of DSBs (the Top2-cleavage complex), whereas Top2 catalytic inhibitors such as ICRF-193 prevent Top2 enzymatic activity without generating noticeable DNA damage. It is only after the long-term exposure of catalytic inhibitors that they induce DNA damage responses (9). The Top1 poisons such as camptothecin stabilize the covalent bond between a tyrosine of Top1 and the 3′ ends of single-strand breaks (the Top1-cleavage complex), which are subsequently converted into DSBs during replication (10).
The choice of repair mechanisms is dependent on the type of topoisomerase inhibitors used to induce the DNA damage. ICRF-193–induced DNA damage is repaired exclusively by NHEJ (11), whereas DSBs induced by etoposide are repaired mainly by NHEJ, with a minor fraction repaired by homologous recombination (11–13). In contrast, camptothecin-induced DSBs are repaired primarily by homologous recombination (14). Topoisomerase poison–induced DSBs, but not catalytic inhibitor–induced DSBs, require an additional step prior to NHEJ or homologous recombination, that is, the elimination of the topoisomerase bound covalently to the DNA ends (Supplementary Fig. S1). This step is carried out by tyrosyl-DNA-phosphodiesterases 1 and 2 (Tdp1 and Tdp2), as well as by the CtIP nuclease. Tdp1 and Tdp2 resolve tyrosyl-DNA covalent bonds that are formed on the 3′- and 5′-side of the DNA ends, respectively (15, 16). CtIP, on the other hand, liberates the covalently bound polypeptides together with the oligonucleotides by making an incision near the cleavage complex (17, 18). CtIP has two functions: the removal of polypeptides from DSBs and the resection of DSBs. Only the former function requires the physical interaction of CtIP with BRCA1 through the phosphorylation of Ser332 in chicken CtIP (corresponding to Ser327 in human CtIP; refs. 6, 19).
RAP80 was characterized as a component in RAP80-Abraxas-BRCA1 (the A complex) and RAP80-CtIP-BRCA1 (the C complex), two of the three multiprotein complexes containing BRCA1 (20–23). RAP80 promotes the recruitment of BRCA1 to DNA-damage sites by physically interacting with ubiquitylated proteins at DSBs through two ubiquitin-interacting motifs (UIM) at the NH2-terminal portion of RAP80. RAP80 is also known to form a complex with the BRCC36 de-ubiquitinating enzyme, and to antagonize RNF8-UBC13–mediated ubiquitination at DSBs (24). However, due to the absence of RAP80-disrupted cells, the role of RAP80 and the BRCA1 subcomplexes that contain RAP80 in the DNA-damage response remains undefined.
To analyze the role of RAP80, we disrupted RAP80 in chicken DT40 cells. Remarkably, RAP80−/− cells were proficient at homologous recombination and NHEJ, and exhibited a marked increase in sensitivity to the Top2 poison etoposide, whereas the sensitivity of RAP80−/− cells to other DNA damaging agents was at near-normal level. This was in marked contrast with BRCA1−/− cells, which exhibit a defect in homologous recombination and are accordingly hypersensitive to a variety of damaging agents such as γ-rays and cisplatin. To assess the functional interaction between RAP80 and BRCA1, we generated RAP80−/−BRCA1−/− and RAP80−/−CtIPS332A/−/− clones (CtIPS332A abolishes physical interaction between CtIP and BRCA1). The two mutants displayed a more pronounced hypersensitivity to the topoisomerase poisons than did the relevant single-gene-disrupted clones. We therefore conclude that RAP80 works independently of BRCA1 and CtIP in the repair of DNA damage caused by the Top2-cleavage complex.
Materials and Methods
Cell culture and counting
DT40 cells were cultured at 39.5°C in RPMI-1640 medium supplemented with 50 μmol/L β-mercaptoethanol, 10% FCS, and 1% chicken serum (Sigma-Aldrich). They have been maintained by S. Takeda since 1991 (25), and were tested routinely for various criteria such as morphology, growth rate, and karyotype. The cell numbers were determined as described previously (26). Briefly, the cultured cells were stained with propidium iodide (PI) and mixed with the fixed number of polystyrene microspheres (Polysciences, Germany). The mixture was subjected to fluorescence-activated cell sorter (FACS) analysis (FACSCalibur, Becton, Dickinson and Company), in which the ratio of gated PI-negative cells to the microspheres was determined and used for calculation of the total cell number in the culture.
Construction of RAP80 targeting vector
Two RAP80 disruption constructs, RAP80-hyg and RAP80-bsr, were generated from genomic PCR products of DT40 cells combined with hyg- and bsr-selection marker cassettes. Genomic DNA sequences were amplified using the primers 5′-ATGGCGAGTTTGGTGTGCTTGTAAATCACG-3′ and 5′-GTGCCATGGCCTTAGGCCTCTATTAAAAAC-3′ for the 3.2-kb 5′ arm of the targeting construct. Amplified PCR products were cloned into pCR2.1 vector (Invitrogen). For the 3′ arm of the targeting construct, a 2.3-kb fragment was amplified by PCR using the primers 5′-CTCTACTTCTGATGATGAACCAACCACGAG-3′ and 5′-ACTCATCATTAACATAAAATACCAACAGCTCCTCC-3′, and was digested with SacI. The resulting 1.4-kb fragment was cloned into the SacI site of the pCR2.1 vector containing the 3.2-kb 5′ arm. The BamHI site was used to integrate the marker gene cassettes.
Generation of RAP80-deficient clones
To generate RAP80-deficient cells, the linearized targeting constructs RAP80-bsr and RAP80-hyg were sequentially transfected into DT40 cells by electroporation (Gene Pulser, BioRad). The genomic DNA of the transfectants derived from each colony was digested with EcoRV and BamHI, and the targeted clones were confirmed by Southern blot analysis. The probe for Southern blot was obtained by digestion of the 2.3-kb fragment described above with EcoRV and XhoI. The 0.88-kb fragment that does not overlap with the 3′ arm was used as a probe. The lack of RAP80 expression was confirmed by reverse transcriptase-PCR (RT-PCR) using the primers 5′-AGTTCGGTAACCCAATTGTGCCC-3′ and 5′-ATCTCCACCAGCCTGCATAGG-3′.
Generation of RAP80−/−BRCA1−/− and RAP80−/−CtIPS332A−/− clones
Gene disruption of BRCA1 was carried out as described previously (27). RAP80−/− cells were transfected with BRCA1 targeting vectors (kind gifts from Douglas Bishop, University of Chicago, Chicago, IL) that eliminate exon 3 through exon 11 of chicken BRCA1 gene including the translation start site. The transfected cells were selected against puromycin or histidinol. The BRCA1-disrupted RAP80−/− cells were identified by Southern blot analysis. The CtIPS332A−/− cells were transfected with RAP80 targeting vectors as described previously (6).
Measurement of cellular sensitivity to DNA-damaging agents
Methylcellulose colony formation and liquid-culture cell survival assays were performed as described previously (28, 29). In the latter assay, we measured the amount of ATP in the cellular lysates to determine the number of live cells. Cells (1 × 104) were treated with each DNA-damaging agent in 1 mL medium using 24-well plates and incubated at 39.5°C for 48 hours. We transferred 100 μL of medium containing the cells to 96-well plates and measured the amount of ATP using CellTiter-Glo (Promega), according to the manufacturer's instructions. Luminescence was measured by Fluoroskan Ascent FL (Thermo Fisher Scientific Inc.).
Immunofluorescent visualization of subnuclear foci formation
To visualize Rad51 accumulation at DNA-damage sites in DT40 cells, DSBs were induced in a subnuclear area using 137Cs (Gammacell 40, Nordion), then stained with Rad51 antibodies as described previously (30).
Targeted integration frequencies
The gene targeting frequencies at Ovalbumin and CENP-H loci were determined by Southern blot analysis of G418-resistant colonies as described previously (6).
Plasmid-based NHEJ assay
Cells (1.5×106) were transfected with 2 μg of linearized pmaxGFP vector and 2 μg of circular DsRed expression vector using Nucleofector I (Amaxa Biosystems; ref. 31). Twenty-four hours after transfection, expression of green fluorescent protein (GFP) was analyzed using FACSCalibur (Becton, Dickinson and Company). Data were corrected for transfection efficiency as measured by the percentage of cells that express DsRed.
Synchronization of the cells was carried out by centrifugal counterflow elutriation as described previously (26).
RAP80-deficient DT40 cells grow at a normal rate
We first isolated RAP80 cDNA from chicken DT40 cells based on the information in the chicken genome database. The predicted chicken RAP80 protein shared 42.1% identity with its human counterpart, with a total of 668 amino acids in length (data not shown). The most conserved are the two UIMs at the NH2 terminus, with 91.3% identity between the two species. The chicken RAP80 gene spans 31.7 kb on the genome and consists of 10 exons. We generated gene-targeting constructs so as to delete the two UIMs, which were encoded by exons 4 and 5 (Fig. 1A). Disruption of the two allelic RAP80 genes was confirmed by Southern blot analysis (Fig. 1B) and by RT-PCR (Fig. 1C). The RAP80−/− cells proliferated with normal kinetics (Fig. 1D).
RAP80−/− cells are hypersensitive to etoposide
To evaluate the role of RAP80 in the DNA-damage response, we analyzed the sensitivity of RAP80−/− cells to various DNA-damaging agents using a colony-formation assay. Remarkably, the RAP80−/− cells did not display an increased sensitivity to γ-ray, cisplatin (DNA cross-linker), UV, or H2O2 (Supplementary Fig. S2, data not shown), and showed only a marginal increase in sensitivity to camptothecin compared with wild-type cells (Fig. 2A). This phenotype is in marked contrast with that of the BRCA1−/− cells, which showed the hypersensitivity to all these agents. On the other hand, the RAP80−/− cells showed a significant increase in sensitivity to etoposide, the Top2 poison (Fig. 2B). To further investigate the role of RAP80 in the tolerance of Top2 inhibitors, we tested for the sensitivity to ICRF-193, another type of Top2 inhibitor, which does not involve persistent Top2-cleavage complex formation. The sensitivity of the RAP80−/− cells to ICRF-193 was comparable with that of wild-type cells (Fig. 2C). To confirm the selective hypersensitivity of RAP80−/− cells to etoposide, we measured the cellular survival at 48 hours after continuous exposure to etoposide using the liquid culture assay (29). The RAP80−/− cells again showed an increased sensitivity to etoposide (Fig. 2D). In this assay, the BRCA1−/− cells, with a reduced colony-formation frequency (Supplementary Fig. S3), exhibited a higher relative cellular tolerance to etoposide than in the colony-survival assay. We conclude that RAP80 may be required to repair DSBs when the Top2-cleavage complex is involved.
No defect in homologous recombination or NHEJ in RAP80-deficient cells
Previous reports indicate that depletion of RAP80 by siRNA leads to radiosensitivity and reduces focus formation for BRCA1 (20–23), which plays a key role in homologous recombination by promoting the accumulation of Rad51 at DNA-damage sites (27, 30, 32). We therefore tested whether RAP80 promotes Rad51 focus formation at induced DSB sites in chicken DT40 cells. In contrast to the greatly reduced radiation-induced Rad51 focus formation in BRCA1−/− cells, we found no reduction in Rad51 focus formation in RAP80−/− cells after irradiation (Fig. 3). Next, we measured the gene targeting frequency at the Ovalbumin and CENP-H loci by Southern blot analysis. We found no decrease in gene-targeting frequency in RAP80−/− cells compared with wild-type cells at these loci (Table 1). These results indicate that RAP80 does not play an important role in homologous recombination.
|Genotype .||Targeted loci .|
|CENP-H .||Ovalbumin .|
|Wild-type||34/43 (79%)||32/37 (86%)|
|RAP80−/−||40/43 (93%)||33/37 (89%)|
|BRCA1−/−||1/37 (3%)||5/37 (14%)*|
|Genotype .||Targeted loci .|
|CENP-H .||Ovalbumin .|
|Wild-type||34/43 (79%)||32/37 (86%)|
|RAP80−/−||40/43 (93%)||33/37 (89%)|
|BRCA1−/−||1/37 (3%)||5/37 (14%)*|
NOTE: Gene targeting frequency was determined by Southern blot analysis.
*The data of BRCA−/− cells at Ovalbumin locus are from ref 6.
NHEJ is the pathway used to repair damage caused by etoposide (11–13). Because RAP80−/− cells exhibited hypersensitivity to etoposide (Fig. 2B and D), we measured the efficiency of NHEJ in RAP80−/− cells by using a plasmid-based DSB-repair assay, which monitors the NHEJ-mediated religation of DSBs in a transiently transfected GFP expression plasmid. Prior to transient transfection, the plasmid was cut between the promoter and the GFP coding sequence by a restriction enzyme. If the linearized plasmids were correctly recircularized by NHEJ, the cells would express GFP (Fig. 4A). Although KU70−/− cells showed a dramatic decrease in GFP expression, RAP80−/− cells showed the rejoining activity comparable with wild-type cells (Fig. 4B). The GFP expression was rather enhanced in BRCA1−/− cells, presumably due to compensatory upregulation of NHEJ activity. To further evaluate NHEJ in the RAP80−/− cells, we measured radiosensitivity in cells synchronized in the G1 to early S phase, where NHEJ plays a dominant role in DSB repair (26). Cells were separated by centrifugal elutriation with >80% of live cell population found to be in the G1 phase (Supplementary Fig. S4). The G1-enriched RAP80−/− and wild-type cells showed a similar sensitivity to γ-rays (Fig. 4C). Taken together, we conclude that the deletion of RAP80 does not compromise either the homologous recombination or the NHEJ pathway.
RAP80 contributes to the damage response to Top2 poisons independently of BRCA1
Because previous reports indicate the functional interaction between RAP80 and BRCA1 (20–23), we generated DT40 cells deficient in both RAP80 and BRCA1 genes (Supplementary Fig. S5A and B). The growth rate of the resulting RAP80−/−BRCA1−/− cells was similar to that of the BRCA1−/− cells (Supplementary Fig. S5C). To evaluate cellular sensitivity to etoposide, we measured the number of live cells after 48-hour continuous exposure to etoposide (6, 29). We took these measures instead of performing a conventional colony formation assay because plating efficiencies were different among the mutant clones (Supplementary Fig. S3), and the growth and the ability to form a colony from a single cell in the semisolid methylcellulose medium were poor with double mutant clones and BRCA1−/− cells. We found that the sensitivity of RAP80−/−BRCA1−/− cells to camptothecin and ICRF-193 was similar to the BRCA1−/− single-mutant cells (Fig. 5A). On the other hand, the RAP80−/−BRCA1−/− cells showed an increased sensitivity to etoposide when compared with BRCA1−/− cells, suggesting that BRCA1 and RAP80 can contribute to the repair of the Top2-cleavage complex independently of each other.
BRCA1 promotes two different repair reactions: the CtIP-mediated removal of the topoisomerase-cleavage complex from DSBs, and RAD51 assembly at DNA-damage sites (6, 27, 30, 32, 33). Although RAP80 is not involved in the latter BRCA1 function, it is unclear whether it facilitates the CtIP-mediated removal of the topoisomerase-cleavage complex. To clarify this issue, we disrupted the RAP80 gene in CtIPS332A/−/− cells. The S332A mutation abolishes the capability of CtIP to interact with BRCA1 without compromising homologous recombination (6). CtIPS332A/−/− cells exhibited an increase in sensitivity to both camptothecin and etoposide as reported previously (Fig. 5B; ref. 6). Remarkably, RAP80−/−CtIPS332A/−/− cells showed a greater sensitivity to etoposide and camptothecin than did either single mutant, suggesting that RAP80 acts separately from the BRCA1/CtIP-dependent pathway. The increase in sensitivity of RAP80−/−CtIPS332A/−/− cells to camptothecin compared with CtIPS332A/−/− cells could be due to the shift in the equilibrium of the BRCA1 protein to a distinct complex (for example BACH1-BRCA1 complex), which would act in the homologous recombination repair of camptothecin-induced DNA damage in the absence of RAP80. In summary, our results show that RAP80 plays a role in removing the Top2-DNA cleavage complex from DNA ends independently of BRCA1 and the BRCA1-CtIP complex.
In this study, we present genetic evidence that RAP80 contributes to the repair of DNA damage induced by Top2 poison. RAP80-deleted cells showed hypersensitivity to etoposide and, to a marginal extent, camptothecin, but not to ICRF-193, γ-rays, UV, or cisplatin. There are various mechanisms by which cells deal with DNA damage caused by topoisomerase inhibitors (Supplementary Fig. S1). These mechanisms include (a) homologous recombination (14), (b) NHEJ (11–13), (c) nuclease (the BRCA1/CtIP complexes; refs. 6, 17, 18), and (d) Tdp enzymes, which cut the covalent bond between the DNA and the polypeptides of the topoisomerase-cleavage complex (15, 16). The third and fourth mechanisms affect cellular tolerance to topoisomerase poisons by eliminating polypeptides from the DNA ends for subsequent homologous recombination– and NHEJ-dependent DSB repair. We investigated which pathway RAP80 contributed to the repair of etoposide-induced DNA damage using DT40 cell lines. We found that RAP80−/− cells were proficient at homologous recombination and showed normal NHEJ. Furthermore, although RAP80 was identified as a component of the BRCA1 complex, the absence of RAP80 significantly augmented etoposide sensitivity in the BRCA1−/− and CtIPS332A/−/− clones, indicating that RAP80 functions separately from the BRCA1/CtIP-dependent damage response. These observations are in concert with the previous report showing that RAP80 exists in two distinct complexes, one of which is devoid of BRCA1 (34). That will leave us with the fourth possibility that RAP80 acts through Tdp2. Indeed, we did observe an epistatic relationship between RAP80 and Tdp2, which repairs Top2 damage by cleaving the phosphotyrosyl bond linking Top2 to DNA 5′-termini (Supplementary Fig. S6). RAP80 may thus act in concert with Tdp2 to remove Top2 from 5′ ends of DSBs. We are currently pursuing this possibility further.
Although our study clearly shows the BRCA1-independent role of RAP80, it does not necessarily exclude the BRCA1-dependent function for RAP80. We could not observe the BRCA1-dependent function of RAP80, because RAP80-deficient DT40 cells were not hypersensitive to γ-ray and did not show discernible homologous recombination defects. Several groups have reported that the human cells show hypersensitivity to ionizing radiation and reduced homologous recombination capability (20–23). The discrepancy between our results and previous studies may be due to the defective damage checkpoint in DT40 cells (35), which could mask the role of RAP80 in BRCA1-dependent damage-checkpoint signaling (36, 37) or even may stimulate homologous recombination (38). Hence, although we show that RAP80 can play a role in repairing the topoisomerase-cleavage complex independently of BRCA1, our study does not necessarily exclude a BRCA1-dependent function for RAP80.
Although etoposide is an effective anticancer drug, treatment with etoposide is associated with the occurrence of the secondary leukemia through the generation of translocations at chromosome 11q23 (39). Further studies on RAP80-dependent DNA repair may contribute to our understanding of chromosomal translocation and thus to the development of better clinical treatments for this life-threatening disease.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
We thank Douglas Bishop for the gift of chicken BRCA1 gene targeting constructs and Keith Caldecott for the gift of unpublished reagents.
Grant Support: Japanese Science and Technology, Grant-in-Aid for Scientific Research on Priority Areas (17013039).
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.