XRCC2 Is a Nuclear RAD51-like Protein Required for Damage-dependent RAD51 Focus Formation without the Need for ATP Binding*

The human XRCC2 gene was recently identified by its ability to complement a hamster cell line, irs1, which is sensitive to DNA-damaging agents and shows genetic instability. The XRCC2 protein is highly conserved in mammalian species and has structural features, including a putative ATP-binding domain (P-loop), consistent with membership of the RecA/RAD51 family of recombi-nation-repair proteins. We show that a hybrid XRCC2-green fluorescent protein, which was found to be functional by complementation, localizes to the nucleus. We have established a functional link between XRCC2 and RAD51 by looking at damage-dependent RAD51 focus formation in the irs1 cell line. Little or no formation of RAD51 foci occurred in irs1. This effect was specific to the loss of XRCC2 because transfection of the gene into irs1 restored normal levels of focus formation. Surpris-ingly, XRCC2 genes carrying site-specific mutations in P-loop residues were found to be able to complement the XRCC2-deficient irs1 line for a number of different end points. We conclude that XRCC2 is important in the early stages of homologous recombination in mammalian cells to facilitate RAD51-dependent recombination repair but that it does not make use of ATP binding to promote this function.

The repair of DNA damage by homologous recombination has an important function in maintaining genetic stability in cells. In bacteria, the RecA protein has a central role in the recombination process, and in the last decade RecA-like proteins have been discovered in eukaryotes. In particular the RAD51 protein is highly conserved from yeast to humans and has been shown to have similar attributes to RecA (1,2). Mutations in both RecA and RAD51 cause severe defects in recombination and extreme sensitivity to DNA-damaging agents. RecA acts directly in recombination processes in which, in the presence of ATP, it forms a polymer on single-stranded DNA and promotes strand exchange with a homologous sequence (3). Using molecular recombination assays, yeast (Saccharomyces cerevisiae) and human RAD51 proteins have been shown to promote strand exchange similarly, although some of the biochemical properties of RAD51 differ from those of RecA.
For example, purified RecA preferentially binds to singlestranded DNA and hydrolyzes ATP at a relatively high rate, whereas the yeast and human RAD51 proteins bind equally to single-and double-stranded DNA and show a much lower rate of ATP hydrolysis (1). All members of this family have a highly conserved sequence motif, first described by Walker et al. (4), that has been linked to ATP binding. The flexible loop of this motif (Walker box A) interacts with the phosphates of ATP and is therefore sometimes called the P-loop.
In S. cerevisiae, two further members of the RecA/ RAD51family of proteins facilitate homologous recombination in mitotic cells; Rad55p and Rad57p form a heterodimer and stimulate RAD51-mediated recombination (5). Yeast mutants that lack these recombination proteins are also extremely sensitive to agents causing severe forms of damage to DNA, such as double strand breaks and interstrand cross-links (6). These additional recombination proteins also have homologues in somatic mammalian cells, in which five RAD51-like proteins have recently been discovered (XRCC2, XRCC3, RAD51L1, RAD51L2, and RAD51L3) (7).
The XRCC2 gene was identified and cloned by its ability to complement the damage-sensitive phenotype of the irs1 hamster cell line (8 -10). The irs1 line, isolated in this laboratory several years ago (11), was found to be sensitive to a variety of agents including ionizing radiation, ultraviolet light, alkylating agents, and especially DNA cross-linking agents such as mitomycin-C. It also shows spontaneous genetic instability with increased frequencies of mutations (12), chromosomal aberrations (13), and chromosome nondisjunction (14). It has recently been shown that the repair of a site-specific double strand break by recombination is severely reduced in irs1 compared with the paternal V79 cells (15). The results of twohybrid interactions involving XRCC2 have led to the speculation of an indirect association with RAD51 through other recombination proteins (16). However, as yet there is no direct evidence for the involvement of XRCC2 in RAD51-dependent recombination processes.
It has been shown that RAD51 can be detected in discrete nuclear foci following DNA damage to mammalian cells (17)(18)(19). In the present study, we considered the localization of XRCC2 in mammalian cells and established a functional link between XRCC2 and RAD51 by looking at damage-dependent RAD51 focus formation in the XRCC2-deficient cell line irs1. Additionally, we have created a number of mutations of XRCC2 in the Walker A motif to determine the importance of this site in the functioning of XRCC2.

EXPERIMENTAL PROCEDURES
Mammalian Cell and Bacterial Culture Methods-Wild-type V79 and XRCC2-deficient irs1 cells were grown as monolayers at 37°C in minimal essential medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum and antibiotics. To assay the sensitivity of individual clones to the DNA cross-linking agent mitomycin-C (MMC) 1 , cells were respread into 24-well plates at 100 -300 cells/well containing graded concentrations of MMC (0 -100 nM). After 10 -12 days of incubation, the density of cell growth was assessed visually (see Table I). Bacterial cultures were grown in Luria-Bertani medium supplemented with appropriate antibiotics; plasmids were propagated in strain DH5␣. Plasmid DNA was prepared using a Qiaprep kit (Qiagen); sequencing reactions were performed using a BigDye kit (PerkinElmer Life Sciences).
Expression of XRCC2-GFP Protein in irs1 Cells-XRCC2 was amplified from a human cDNA library using Pfu polymerase (Stratagene) with primers to incorporate restriction sites for cloning into pEGFP-N1 (CLONTECH) to create an in-frame fusion of the EGFP gene to the 3Ј-end of XRCC2. Approximately 4 g of plasmid DNA was transfected into irs1 cells by electroporation (Bio-Rad Gene Pulser at 400 V/500 microfarads) in Cytomix buffer (20). Cells were selected for 14 days in 500 g/ml G418 (Life Technologies, Inc.). G418-resistant clones were tested for MMC resistance using 24-well plates (see above) to demonstrate XRCC2 function. MMC-resistant cells were grown on glass coverslips, and the fusion protein was visualized at 488 nm using a Bio-Rad MRC600 laser scanning confocal microscope.
Site-directed Mutagenesis-Human XRCC2 cDNA was cloned into pBS as a template for site-directed mutagenesis. Mutagenic polymerase chain reaction was carried out using the QuikChange site-directed mutagenesis protocol (Stratagene) with primers designed to create changes at the highly conserved lysine 54 residue (K54A, K54R, and ⌬G53K54). In addition to these mutations, a cDNA with a fortuitous stop codon in the same region was derived from the mutagenic polymerase chain reaction, and this was used as a negative control. Inserts carrying mutations were subcloned into the NheI and BamHI sites of the pIRESneo2 vector (CLONTECH). Plasmid DNA was transfected into irs1 cells (as above), and cells were selected for 14 days in 500 g/ml G418. Six separate clones for each construct were picked and tested for MMC resistance using the 24-well plate assay (see above). The internal ribosome entry site vector was used to select against integration events that disrupted the XRCC2 gene. In control experiments, ϳ85% of G418resistant clones also expressed the XRCC2 construct (data not shown).
Cytogenetic Analysis-Metaphases were collected from cells grown in flasks by exposure to colcemid (0.05 g/ml) for 2 h. Cells were trypsinized and incubated in hypotonic solution (1:1; 0.075 M KCl, 0.034 M trisodium citrate) for 8 min at 37°C. Cells were fixed in 3:1 ethanol: acetic acid and dropped onto slides before staining with Giemsa for analysis. Approximately 100 metaphases/cell type were analyzed for both chromatid and chromosome aberrations.
RAD51 Immunofluorescence-Cells were grown on glass coverslips to subconfluent levels and irradiated with 10-Gy x-rays. After 0 -5 h of incubation at 37°C, cells were fixed for 2 h at 4°C in 1% paraformaldehyde in PBS. Fixed cells were permeabilized for 10 min at room temperature in 0.1% Triton X-100/0.1ϫ SSC and then blocked for 1 h at room temperature in PBS/5% normal horse serum. Rabbit anti-hRAD51 antibody (FBE1, a generous gift from S. C. West, Imperial Cancer Research Fund Clare Hall Laboratories) was applied at a dilution of 1:100 in PBS/5% horse serum and incubated overnight at 4°C. Cells were treated with anti-rabbit IgG/Cy5 (Jackson ImmunoResearch) at 1:100 dilution for 1 h at room temperature in the dark, and then the coverslips were mounted in Vectashield anti-fade medium. Cells were analyzed by confocal microscopy at 633 nm, and images were selected at random. Immunofluorescence images were scored blindly by two independent scorers who recorded the number of discrete strongly fluorescing nuclear foci present in each cell. Approximately 50 cells were counted per data point, and data were compared statistically to assess homogeneity using the Mann-Whitney U test. One scorer consistently scored about 30% more foci than the other, but identical trends of increased focus formation with irradiation and with time after irradiation (of cell lines with wild-type XRCC2) were found for each scorer.
Cell Cycle Analysis-Cells were seeded at 5 ϫ 10 5 /75-cm 2 flask at 16 h before x-irradiation. Cells were then harvested at intervals of up to 7 h, washed three times with PBS, and fixed in 70% ethanol at 4°C for 30 min. Fixed cells were incubated for 60 min at 37°C with 0.5% RNaseA (Sigma) before the addition of propidium iodide to 50 g/ml final concentration. Flow cytometric analysis (Ͼ10,000 cells/sample) was performed on a FACSort (Becton Dickinson).
Western Blotting-Cells were grown to subconfluent levels in tissue culture flasks, irradiated with 10-Gy x-rays, and harvested after a further 5 h growth. Cells were resuspended in distilled water and heated for 15 min at 95°C, and total protein was measured using the Bradford assay (Bio-Rad). The lysate was resuspended in an equal volume of 2ϫ Laemmli/glycerol buffer and treated for a further 5 min at 95°C. A 20-g sample was separated on a 7.5% SDS-polyacrylamide gel and transferred to polyvinylidene difluoride membrane (Millipore). RAD51 was detected using the FBE1 antibody at a dilution of 1:2500 in PBS, 0.1% Tween 20, 5% bovine serum albumin, 5% nonfat dried milk. Anti-rabbit IgG/horseradish peroxidase conjugate was used as the secondary antibody (Promega) at a dilution of 1:5000, and bands were visualized using ECL (Perkin Elmer Life Sciences Renaissance) according to the manufacturer's protocol) and Biomax film (Eastman Kodak Co.). Anti-␣-tubulin antibody (1:2000) (Sigma) was used as loading control.

XRCC2 Is Localized in the Nucleus of Mammalian Cells-
The human XRCC2 protein was expressed as a fusion to the N terminus of the green fluorescent protein (GFP) in XRCC2defective irs1 hamster cells. This construct was tested for functional ability by transfection into the irs1 line and exposure to a range of concentrations of the DNA cross-linking agent MMC. As shown previously (9), the expression of the wild-type human XRCC2 gene in irs1 cells gives a good, although not complete, level of complementation of MMC sensitivity (Table I). In the same test, the XRCC2-GFP hybrid gene had no loss of complementing ability when compared with the XRCC2 cDNA (Table  I). Although the expression of GFP alone occurred throughout the cell (data not shown), its expression was primarily nuclear when associated with XRCC2 (Fig. 1).
XRCC2 Is Required for the Formation of RAD51 Foci in Response to DNA Damage-To assess the function of XRCC2 specifically in recombination-repair processes, we measured the formation of radiation-induced RAD51 foci in wild-type (V79) and irs1 hamster cells. Cells were irradiated with 10-Gy x-rays, and the numbers of foci were counted at hourly intervals for up to 5 h. In agreement with previous studies (17,21,22), a low level of focus formation was seen in unirradiated cells and in cells fixed immediately after irradiation (generally Ͻ2 foci/cell; see Fig. 2A). In V79 cells the numbers of foci increased significantly following x-irradiation ( Fig. 2A). Although there was some fluctuation in the numbers of foci seen in V79 cells with time following irradiation (Fig. 2B), this variation was not significant over 1-5 h after irradiation (p ϭ 0.8). In the XRCC2deficient irs1 cells, little focus formation was found ( Fig. 2A), although there is some evidence for a slow accumulation of foci with time. However, this level is on average about 5-fold lower (p ϭ 10 Ϫ6 ) than that seen for the XRCC2-proficient V79 cells (Fig. 2B). It has been shown that RAD51 foci in unirradiated cells form primarily in S-phase (21), raising the question of whether differences in cell cycle distribution may influence the ability to detect foci in irs1 after irradiation. However, cell cycle profiles for irradiated V79 and irs1 cells were very similar with most cells blocked in G 2 phase at 5 h after a dose of 10 Gy (Fig. 3A). It is also possible that the formation of RAD51 foci is affected by the levels of RAD51 protein in irs1 relative to V79, but we found that RAD51 levels were similar in the two cell lines, both before and after irradiation (Fig. 3B).
Transfection with the Human XRCC2 Gene Restores RAD51 Focus Formation-The stable transfection of the irs1 cells with human XRCC2 genomic (P1-artificial chromosome) DNA (8) or cDNA (Table I) gave clonal lines with a good level of correction of DNA-damage sensitivity. These transfected lines were also found to have a restored ability to form RAD51 foci after irradiation ( Fig. 2A), whereas RAD51 protein levels were similar to those in irs1 and V79 (Fig. 3B). This finding implicates the XRCC2 gene specifically in correct focus formation. The time course for the development of foci was apparently slower in the XRCC2-transfected lines than for V79, possibly because the human gene is not as efficient in a hamster cell background (Fig. 2B).

Site-directed Mutations of the Walker Box A of XRCC2 Have Little Effect on Complementation of Mitomycin-C Resist-
ance-To gain further information on the function of XRCC2, we mutated the highly conserved Walker box A (putative ATP binding) motif in several different ways. It has been shown previously that the invariant lysine residue of this motif is critical to the function of RecA or RAD51 proteins (see "Discussion"). Initially therefore we mutated this residue to either a conservative (K54R) or nonconservative (K54A) alternative in XRCC2 as well as deriving a mutant form in which the lysine was deleted along with an adjacent glycine (⌬G53K54) (Fig. 4). Negative controls in this experiment were provided by a mutation leading to a stop codon in this region of the XRCC2 sequence and by the cloning vector alone. Each mutant gene was verified by sequencing and was then transfected into irs1 cells to assess the functional complementation. We initially used MMC survival to test the response of the mutated genes because XRCC2-deficient cells show an extreme sensitivity to this DNA-damaging agent. The ability to complement MMC sensitivity was first tested in the simple growth assay from a standard cell inoculum in 24-well plates (see "Experimental Procedures"). As seen in Table I, each of the mutants tested apart from the truncation showed little or no reduction in the ability to complement the MMC sensitivity of irs1 relative to the wild-type cDNA. To check that there was no loss of complementing ability by more stringent procedures, allowing for the cloning efficiency of each transfected cell population, multipoint survival curves were carried out with pooled clones for several of the mutants (Fig. 5A). It is seen that the K54A mutant shows little or no loss of complementing ability compared with the wild-type cDNA, whereas the ⌬G53K54 mutant has a small but consistent reduction in this capacity.
The Influence of Walker Box A Mutants on X-ray Survival, Chromosomal Aberrations, and RAD51 Focus Formation-To generalize our findings with MMC sensitivity, we tested some of the mutant XRCC2 genes for other responses. As expected  the differential between wild-type and mutant XRCC2 genes was much less for x-rays than for MMC treatment (11), and in this case the ⌬G53K54 mutant showed no significant difference in sensitivity from the wild-type cDNA (Fig. 5B).
The loss of XRCC2 seriously influences genetic stability even when XRCC2-deficient cells are not exposed to exogenous DNA-damaging agents. We have shown previously that the spontaneous chromosome aberration frequency is high in irs1 cells but is restored to near normal levels in cells transfected with the XRCC2 cDNA (9). Consistent with the survival data, examination of the chromosomes of irs1 cells transfected with the different mutant forms of XRCC2 showed that the alteration or loss of the conserved glycine and lysine residues of the P-loop did not affect the ability of the gene to confer genetic stability (Table II).
Repeat experiments assessing the damage-dependent RAD51 focus formation are shown in Fig. 6. In this assay, as was found for the mitomycin-C response, the K54A mutation does not show a significant difference from the wild-type cDNA, and the ⌬G53K54 mutation shows only a small reduction in focus formation. DISCUSSION We have found that the XRCC2 protein is located in the nucleus, which is consistent with a role in DNA metabolism. To test for the involvement of XRCC2 in repair by homologous recombination, as would be predicted from sequence homologies (7), we measured the requirement for XRCC2 in damagedependent RAD51-focus formation. RAD51 has been shown to be expressed in a cell cycle-dependent fashion with the majority of expression in S-and G 2 -phase cells (23)(24)(25)(26). In previous studies it has been established that some nuclear RAD51 foci will form in the S-phase of untreated somatic cells (21). However, the numbers of foci are substantially increased by treatment with DNA-damaging agents, including ionizing radiation, in a time-and dose-dependent manner (17,19). Although experiments shielding some cells from irradiation (27) have suggested that RAD51 protein does not co-localize with sites of FIG. 3. A, cell cycle distributions of V79 and irs1 cells in response to 10-Gy x-rays. B, Western blot of RAD51 in V79, irs 1, and irs1 cells transfected with XRCC2 with and without 10-Gy irradiation. In a representative experiment using the ␣-tubulin response as a loading control, the RAD51 signal strength showed no significant differences between cell lines or before and after irradiation. radiation damage, more recent studies using microbeams show that focus formation is spatially related to sites of damage in the nucleus of somatic cells (28). Additionally, human RAD51 shows co-localization to sites of DNA damage as represented by single-stranded DNA (29) and histone H2AX phosphorylation (30). V79 hamster cells showed extensive focus formation within 1 h of irradiation, and thereafter the number of foci/cell remained approximately constant. Very little RAD51 focus formation occurred in the XRCC2-deficient irs1 cells, although there was some evidence for a small increase with time. This small increase could be attributable to the accumulation of cells in S-and G 2 phases of the cell cycle (Fig. 3A) because RAD51 is expressed primarily in these phases. In the irs1 cells, the XRCC2 transcript is shortened because of the loss of exon 2, 2 and this will lead to a frameshift in the sequence, giving a predicted protein of only 33 amino acids (of which 13 are correct, compared with the full-length sequence of 278 amino acids). Therefore, this truncated XRCC2 protein is unlikely to have any normal activity. Additionally, we have shown that the loss of RAD51 focus formation in irs1 is specific to XRCC2 because the ectopic expression of full-length human XRCC2 restored normal levels of foci following irradiation.
The XRCC2 protein has few identifiable motifs to give clues on function except for the proposed ATP binding boxes. As noted in the Introduction, the conserved Walker A box (or P-loop) is characteristic of all members of the RecA/RAD51 family. However, we found that the substitution of highly conserved residues in the P-loop has little effect on the ability of XRCC2 to complement a number of end points when introduced into the irs1 line, including survival in response to MMC or x-rays and genetic instability. Even in the case of the ⌬G53K54 mutation, which deletes two key residues within the Walker box A motif, there was only a small negative effect relative to wild-type cDNA when measured by mitomycin-C response (Fig.  5A) and focus formation (Fig. 6). We especially targeted the lysine residue in our XRCC2 mutations because this site is conserved in all NTP-binding proteins with the P-loop motif (31). In ATPases the conserved lysine is probably important to the conformation of the P-loop and for direct interaction with the ␤and ␥-phosphates of the bound ATP. Consistent with this degree of conservation, the mutation of this residue in ATPbinding proteins generally compromises function. This is certainly true for Escherichia coli RecA for which even the relatively conservative K72R mutation gives a null phenotype (32) although the mutant protein retains the ability to promote homologous pairing of DNA strands (33). S. cerevisiae Rad51p altered at this P-loop lysine similarly has a compromised function; the K191A mutation gives a null phenotype, whereas K191R has a partially defective phenotype (34). In human RAD51, again a conservative mutation of this lysine residue (K133R) gives a protein capable of partially restoring function in chick RAD51-defective cells, whereas the nonconservative K133A mutation cannot restore function (35). Therefore, our finding that such mutations do not lead to a compromise of XRCC2 function is perhaps surprising, although not without precedent. Mutation of the P-loop lysine in the S. cerevisiae RAD51-like protein Rad57p, even to a nonconservative change (K133A), does not affect its function in DNA recombination repair. In contrast Rad55p, the yeast RAD51-like protein, which forms a heterodimer with Rad57p, does require an intact lysine residue in the P-loop motif for function (36). We have recently shown that, similarly to Rad55p/Rad57p, XRCC2 forms a heterodimer with another human RAD51-like protein, RAD51L3, and that RAD51L3 has significant DNA-stimulated ATPase activity (37).
It seems likely therefore that in both yeast and humans, only one of the partners in RAD51-like protein heterodimers retains the ATP binding requirement. In recent studies there are a number of precedents for this type of "functional asymmetry" in protein complexes that affect repair and recombination processes. For example, the ATP-dependent deoxyribonuclease complex in Bacillus subtilis possesses two ATP binding sequences. The mutation of the lysine residue in the AddA subunit drastically affects function, whereas the identical mutation in the AddB subunit has only marginal effects (38). Similarly there is evidence that the partner proteins of heterodimers involved in DNA mismatch repair, such as MSH2 and MSH6, make different contributions to ATPase activity (39). The equivalent mismatch repair protein in E. coli, MutS, forms homodimers, but even in this case there is evidence from recent crystal structure analysis that the partners do not participate equally in the ATP binding reaction (40). This asymmetry may promote conformational changes that assist further reactions such as DNA binding. It will be interesting to see whether other mammalian RAD51-like proteins, some of which are proposed to interact as dimers or in higher-order complexes (16), have a similar functional divergence.   Cartwright et al. 1998 (9). Approximately 100 cells were scored per cell line. Our results with XRCC2 support the evolutionary conservation of function in RAD51-like proteins from yeast to man. For meiotic recombination in yeast, Rad55p and Rad57p are required for RAD51 focus formation (41), and purified Rad55p/ Rad57p heterodimer has been shown to facilitate the loading of RAD51 onto single-stranded DNA coated with replication protein A (5). The XRCC2/RAD51L3 heterodimer in mammalian cells may act in a similar manner to Rad55p/Rad57p. However, the process is likely to be more complex in mammalian cells, which include several additional RAD51-like proteins (7) that interact to form heterodimers and possibly larger complexes (5,16,37). Some of these proteins have also been shown to be required for RAD51 focus formation, namely XRCC3 in mammalian cells (19) and RAD51L1 in chick cells (42). Despite their apparent facilitating function, it is clear that the RAD51-like proteins have an important part to play in mammalian development because in the case of XRCC2 (43), RAD51L1 (44), and RAD51L3 (45) gene disruption in mice leads to embryonic lethality. Further, there are parallels between XRCC2 deficiency and the disruption of the breast cancer susceptibility genes BRCA1 and BRCA2. These genes have recently been shown to be required for damage-dependent RAD51 focus formation (46,47) and have reduced levels of homologous recombination (48 -50). The BRCA genes are also required for correct chromosome segregation at mitosis (51,52), and we have shown that XRCC2 and XRCC3 promote the fidelity of chromosome segregation (14). These similarities suggest a role for RAD51-dependent homologous recombination repair and for XRCC2 in particular in maintaining genetic stability in mammalian cells and perhaps in influencing the incidence of cancer.