|Home | About | Journals | Submit | Contact Us | Français|
The rare hereditary disorder Fanconi anemia (FA) is characterized by progressive bone marrow failure, congenital skeletal abnormality, elevated susceptibility to cancer, and cellular hypersensitivity to DNA cross-linking chemicals and sometimes other DNA-damaging agents. Molecular cloning identified six causative genes (FANCA, -C, -D2, -E, -F, and -G) encoding a multiprotein complex whose precise biochemical function remains elusive. Recent studies implicate this complex in DNA damage responses that are linked to the breast cancer susceptibility proteins BRCA1 and BRCA2. Mutations in BRCA2, which participates in homologous recombination (HR), are the underlying cause in some FA patients. To elucidate the roles of FA genes in HR, we disrupted the FANCG/XRCC9 locus in the chicken B-cell line DT40. FANCG-deficient DT40 cells resemble mammalian fancg mutants in that they are sensitive to killing by cisplatin and mitomycin C (MMC) and exhibit increased MMC and radiation-induced chromosome breakage. We find that the repair of I-SceI-induced chromosomal double-strand breaks (DSBs) by HR is decreased ~9-fold in fancg cells compared with the parental and FANCG-complemented cells. In addition, the efficiency of gene targeting is mildly decreased in FANCG-deficient cells, but depends on the specific locus. We conclude that FANCG is required for efficient HR-mediated repair of at least some types of DSBs.
Fanconi anemia (FA), a rare autosomal recessive disorder, is clinically characterized by progressive bone marrow failure, congenital skeletal abnormality, and elevated susceptibility to malignant tumors (reviewed in references 1, 19, and 24). FA is one of the genetic diseases in which the capacity to maintain genomic integrity is severely impaired (24). Cells derived from FA patients exhibit spontaneous chromosomal instability (24), which is intensified following treatment with DNA cross-linking agents such as mitomycin C (MMC) and cisplatin (42). The DNA cross-linker sensitivity is a hallmark of FA and has been used as a simple diagnostic test (2).
FA is caused by mutations in multiple genes. Somatic cell fusion studies defined at least eight FA complementation groups, and six of them were used to clone the complementing genes, which are designated FANCA, -C, -D2, -E, -F, and -G (1, 19, 24). These FA genes do not have any obvious domains or motifs that would suggest possible biochemical functions. Given the similar clinical and cellular features of the different complementation groups, the FA genes are thought to share at least one common function in a single biochemical pathway. Indeed, two-hybrid and immunoprecipitation analyses indicated that FA proteins (A, C, E, F, and G) form a multisubunit nuclear complex (reviewed in references 1, 19, 21, and 24), and the complex has recently been reported to interact with FANCD2 protein (38).
It has been suggested that the FA pathway works in a DNA damage response pathway (1, 19, 24). The occurrence of spontaneous chromosome breakage in FA cells is consistent with a potential role in repairing double-strand breaks (DSBs). In eukaryotic cells, there exist two major DSB repair pathways, nonhomologous end joining and homologous recombination (HR) (29, 57). Defects in either DSB repair pathway cause spontaneous chromosome breakage (27, 28, 39, 46, 51-53; reviewed in reference 55). Furthermore, the repair of interstrand cross-link damage, which may involve the generation of a DSB as an intermediate, requires HR in both prokaryotes and yeast (11, 55). However, FA cells generally exhibit normal levels of sister chromatid exchange (SCE) (31, 60), suggesting that gross defects in cellular HR activity are rather unlikely in FA cells (10, 24, 47).
Nevertheless, recent studies support the view that FA genes may have a link to the HR process. For example, the FANCD2 protein appears to function in association with the breast cancer susceptibility protein BRCA1 (17), mutations in which produce defects in HR (35, 44). These two proteins physically interact and colocalize in distinct nuclear foci following DNA damage. FANCD2 focus formation is dependent on intact BRCA1, the FA nuclear complex, and monoubiquitination of D2 protein at amino acid residue K561 (17). Most importantly, another link to recombination comes from the observation that biallelic causal mutations of BRCA2 are present in human FANCD1 cells (23). BRCA2 is thought to be vital in HR, perhaps by helping to form the Rad51 nucleoprotein filament (6, 40, 62; reviewed in references 56 and 58), a key structure that initiates homologous pairing (5, 29, 57). Thus, the fact that BRCA2 is actually an FA gene suggests that at least a subset of FA patients may have defective HR (49).
To date, except for BRCA2 (which could vary depending on the specific mutant allele), formal proof is still lacking for involvement of FA proteins in HR. To evaluate HR activity of cells defective in FA genes, we established FANCG-deficient cells by using gene targeting technology with the chicken B-cell line DT40 (4). The fancg DT40 mutant has a phenotype resembling that of human and rodent FANCG mutants. In particular, we show that our fancg knockout cells have not only decreased HR capacity for repairing enzymatically induced, site-specific chromosomal DSBs but also elevated ionizing radiation (IR)-induced chromosomal aberrations in cells irradiated in late S and G2 phases.
A partial chicken cDNA sequence for FANCG was identified in the chicken WebBursalEST database and was subsequently used as a probe in screening a chicken intestinal mucosa cDNA library (Stratagene, La Jolla, Calif.). To isolate full-length FANCG cDNA, the N-terminal portion of chicken FANCG was PCR amplified from DT40 cDNA on the basis of the sequence of another chicken expressed sequence tag (EST) clone. Identity of the cDNA clones was confirmed by sequencing. An alignment of the predicted protein sequences from four species (human, mouse, hamster, and chicken) was performed with Genetyx-Mac version 10 (Software Development Co. Ltd., Tokyo, Japan).
Expression vectors were constructed by inserting the full-length chicken or human FANCG cDNA into either pApuro vector (50) or pCR3-loxP-IRES-EGFP-loxP vector (12). The human Rad51 expression vector (pAneo-hRad51) was previously described (34). The I-SceI expression vector pcBASce was kindly provided by Maria Jasin (Sloan-Kettering Institute, New York, N.Y.).
Genomic DNA clones of FANCG were isolated by screening a chicken genomic library (Stratagene), and the gene disruption constructs were generated as previously described (4). Gene targeting with this construct was expected to replace the genomic region that encodes chicken FANCG amino acids 169 to 399 with selection markers.
Cells were maintained in RPMI 1640 medium supplemented with 10−5 M β-mercaptoethanol, 10% fetal calf serum, and 1% chicken serum (Sigma, St. Louis, Mo.) at 39.5°C. DNA transfections and selection were performed as described previously. To stably express human Rad51 or chicken FANCG in FANCG-deficient cells, expression vector pAneo-hRad51 or pCR3-loxP-chFANCG/IRES-EGFP-loxP was transfected by electroporation. Expressing clones were identified by Western blot analysis or by measuring enhanced green fluorescent protein (EGFP) fluorescence with a FACSCalibur (Becton Dickinson, Mountain View, Calif.), respectively.
Cells (106) were lysed in 20 μl of sodium dodecyl sulfate (SDS) sample buffer (25 mM Tris-HCl [pH 6.5], 1% SDS, 0.24 M β-mercaptoethanol, 0.1% bromophenol blue, 5% glycerol). Boiled aliquots were subjected to SDS-10% polyacrylamide gel electrophoresis. After transfer to a membrane, proteins were detected with rabbit anti-human Rad51 serum (kindly provided by Akira Shinohara) and horseradish peroxidase-conjugated goat anti rabbit immunoglobulin (Amersham Biosciences, Piscataway, N.J.) by using ECL Western blotting detection reagents (Amersham Biosciences).
Total RNA was isolated from wild-type and FANCG-deficient cells by using RNAsol (Invitrogen, Carlsbad, Calif.). cDNA was synthesized from 1 μg of total RNA by using random hexamers and Superscript reverse transcriptase (Invitrogen). GoldTaq polymerase (Applied Biosystems, Inc., Foster City, Calif.) was used for amplification. Sequences of specific primers for chicken FANCG were 5′-GAACCTCTGGGCAGGACCCTAGCACACAAA-3′ and 5′-TGACCTCATCACACACTGTTA-3′. Primers for chicken RAD51 were 5′-GCAGCCATGGCCATGCAGGT-3′ and 5′-AAGTCTTCATTCTTTTGCAT-3′.
Cell growth was determined as previously described (52). Briefly, cells were stained with 5 μg of propidium iodide (PI) per ml, mixed with fixed numbers of plastic beads (Polysciences, Inc., Warrington, Pa.), and then counted with a FACSCalibur. For cell cycle analysis, cells were cultured in the presence of bromodeoxyuridine (BrdUrd) for 10 min and then fixed and stained with anti-BrdUrd antibody (PharMingen, San Diego, Calif.) and PI. EGFP expression was analyzed with a FACSCalibur, with dead cells gated out by PI staining.
To analyze targeted integration events at the KU70, OVALBUMIN, XRCC2, and XRCC3 loci, we used previously described disruption constructs (13, 52, 53). Some of them were modified by swapping drug resistance marker cassettes so that appropriate drugs could be used. DNAs were transfected into cells, and Southern blot analysis was performed following selection of clones resistant to the appropriate drug.
Chromosome analysis was done with or without X-ray (2 Gy) or MMC (40 or 20 ng/ml, 24 h) treatment as previously described (51, 52). To analyze MMC-induced SCEs, cells were incubated in medium containing 50 ng of MMC per ml for 12 h. Colcemid was added to 0.1 μg/ml for the last 1.5 h of this incubation before harvest. Detection of SCE events was performed as previously described (51).
Transfections were done by electroporating (250 V, 975 μF) cells (107) suspended in complete culture medium with each of the following plasmid DNAs (30 μg) without linearization: I-SceI expression vector (pcBASce) together with empty vector pApuro or human FANCG or chicken FANCG expression vector (pApuro-hFANCG or pApuro-chFANCG, respectively). pBluescript SK was also transfected as a negative control. At 24 h after electroporation, cells were transferred to 96-well cluster trays containing 2.0 mg of G418 per ml. Cells were grown for 10 to 14 days, and surviving G418-resistant colonies were counted to measure the frequency of HR-mediated repair of I-SceI-induced DSBs. To determine the nature of repair events, genomic DNA was extracted from expanded G418-resistant colonies, restricted by KpnI and SacI double digestion, and examined by genomic Southern blotting with a neo fragment as a probe.
Serially diluted cells were plated in medium containing methylcellulose and then irradiated with 4 MV of X rays (linear accelerator; Mitsubishi Electric Inc., Tokyo, Japan) at a dose of 100 cGy per min. To measure sensitivity to MMC (Kyowa-Hakkou, Tokyo, Japan), cells were incubated at 39.5°C in complete medium containing the compound for 1 h, washed three times with warm medium, and then plated in medium containing methylcellulose. Sensitivity to cisplatin (Nihon-Kayaku, Tokyo, Japan) was measured by plating cells onto the methylcellulose plates containing cisplatin.
Cells were harvested at the indicated time points after X irradiation (8 Gy) or MMC exposure (500 ng/ml for 1 h). Cytospin slides were prepared by using Cytospin 3 (Shandon, Pittsburgh, Pa.). Visualization of subnuclear foci was performed as previously described (61) with anti-hRad51 rabbit antiserum. Images were captured by TCS-NT laser scanning confocal microscopy (Leica Microsystems, Bannockburn, Ill.), and processed with Photoshop software (version 6; Adobe Systems Inc., San Jose, Calif.). Cells with more than four brightly fluorescing foci were scored as positive. At least 100 morphologically intact cells were examined at each time point.
The chicken FANCG cDNA sequence has been deposited to DDBJ under accession number AB096039.
To isolate chicken FANCG cDNA, we examined the WebBursalEST database (http://swallow.gsf.de/dt40Est.html) and found several EST sequences that have significant homology to human FANCG/XRCC9 (8, 32). Library screening combined with RT-PCR based on these sequences led to the identification of full-length chicken FANCG cDNA. Of note, RT-PCR products of N-terminal chicken FANCG were found to contain several alternatively spliced variants (data not shown). Most of them appeared to be out of frame, but we identified two sequences as candidate full-length chicken FANCG. One of them could indeed complement chicken FANCG-deficient cells (see below), whereas the other could not (data not shown). The latter transcript lacks nucleotides encoding 11 consecutive amino acids that overlap the putative leucine zipper (Fig. (Fig.1).1). These N-terminal variants were not detected during the course of human XRCC9 cloning (L. H. Thompson, unpublished data) and were not found in silico by BLAST searching among more than 100 EST FANCG sequences deposited in the human EST database.
This isolated chicken cDNA, the first nonmammalian FANCG sequence reported, encodes a protein with 648 amino acids, compared with the 622-amino-acid human homolog. The percent amino acid sequence identity of chicken FANCG with the human, mouse, and hamster homologs is only 39% in each case. Thus, overall conservation is rather low, whereas the identity between mouse and human FANCG is 71% and that between mouse and hamster FANCG is 83%. Similar to the case for mammalian FANCG homologs, no recognizable motifs besides a putative leucine zipper-like region are present in chicken FANCG protein (Fig. (Fig.1).1). Comparison of FANCG amino acid sequences revealed conserved regions or residues across the four species. It is noteworthy that all five non-disease-causing FANCG polymorphisms listed in the Fanconi anemia mutation database (http://www.rockefeller.edu/fanconi/mutate/default.html) are at residues that are not conserved. Furthermore, this database lists three bona fide missense mutations, two of which (L71P and L303P) lie at residues conserved across the four species. The third mutation is adjacent to a conserved residue, and the change is nonconservative (Arg to Pro at amino acid 22). Since human FANCG cDNA could not complement chicken fancg mutant cells (see below), some amino acid residues in the chicken sequence that are not conserved across the chicken counterparts are likely critical for chicken FANCG's function.
Chicken FANCG genome clones were obtained by library screening with a chicken FANCG cDNA probe. We made a FANCG targeting vector by replacing ~4.5 kb of genomic segment with a histidinol resistance (His) (Fig. (Fig.2A)2A) or puromycin resistance gene cassette. The upstream and downstream homology arms are ~6 and ~1 kb in length, respectively. Single transfection with this targeting vector eliminated the wild-type band in Southern blot analysis with two different genomic probes (Fig. (Fig.2B),2B), indicating that there is only one FANCG allele in DT40 cells. This finding is consistent with the extensive synteny between chicken chromosome Z and human chromosome 9 (36) (the human FANCG locus is on chromosome 9p13 ) and with the fact that DT40 cells carry a single Z chromosome (46). Furthermore, we found a neighboring gene in the 5′ region of the chicken FANCG locus that has significant homology to the phosphatidylinositol glycan class o (PIGO) gene, which lies next to FANCG in the human genome sequence as well (data not shown). These data strongly indicate that we have deleted a phylogenetically conserved bona fide chicken FANCG gene.
The resulting fancg (FANCG-deficient) cells are phenotypically null, since no FANCG transcript was detected by RT-PCR analysis (Fig. (Fig.2C).2C). Northern blot analysis was uninformative, perhaps because of low expression of FANCG. To facilitate our analysis of the role of FANCG in HR, we also targeted FANCG in wild-type cells that carry an integrated recombination substrate, SCneo (26), at the OVALBUMIN locus (13).
We characterized the proliferative properties of wild-type and fancg cells by monitoring growth curves; an xrcc3 HR-defective DT40 line (53) was monitored in parallel. DT40 fancg mutant cells grew at almost the same rate as wild-type cells and definitely more rapidly than xrcc3 mutant cells (Fig. (Fig.3A).3A). The proportion of dead cells in fancg mutant cultures was not increased, and the cell cycle distribution was indistinguishable from that of the wild type (Fig. (Fig.3B).3B). The efficiencies of plating of cells in methylcellulose plates were ~100% for wild-type cells and ~60% for fancg cells.
We performed karyotype analysis of fancg cells before and after MMC treatment, since both spontaneous and MMC-induced chromosomal breakages usually characterize the FA phenotype. DT40 fancg cells did not have appreciably elevated levels of spontaneous chromosomal aberrations (Table (Table1),1), whereas the xrcc3 mutant does (53). After treatment with MMC (40 ng/ml, 24 h), fancg cells exhibited more chromosomal aberrations than wild-type cells, and the magnitude was consistent with the survival data (see below). With a lower dose of MMC (20 ng/ml), the level of aberrations in fancg cells was comparable to that of wild-type cells treated with 40 ng of MMC per ml. Also, as for human FA cells (24), our fancg cells exhibited a normal level of MMC-induced SCE. The level of spontaneous SCE was 2.1 ± 1.7 and 2.7 ± 2.0 (mean ± standard deviation; n = 50) for wild-type and fancg cells, respectively. The level of MMC-induced SCE was 4.8 ± 2.7 per cell (mean ± standard deviation; n = 50) for both wild-type and fancg cells.
The ability of fancg mutant cells to form colonies following exposure to DNA-damaging agents was assessed. We define sensitivity as the dose reduction factor, based on the dose giving 10% survival, of the survival curve. Asynchronous fancg cells are not sensitive to X irradiation as measured by cell killing (Fig. (Fig.4A),4A), but below we show modestly increased sensitivity based on induced chromosomal aberrations when cells are exposed in late S and G2 phases. fancg cells are only slightly sensitive to MMC (~1.5-fold) but show ~3-fold sensitivity to cisplatin compared with the parental cells (Fig. 4B and C). This pattern somewhat resembles that for human FA-G cells (8), CHO UV40 and NM3 fancg mutants (32, 59), and mouse knockout mutants of Fancg (30, 63). Human fancg cells show 10- to 20-fold MMC sensitivity (reference 8 and our unpublished data), but there are little or no data for other agents with gene-complemented mutant cells. One study found IR sensitivity with fancg mouse splenocytes (63), but this was not seen with embryonic fibroblasts (30). Compared with our previously reported rad51 paralog mutants (51, 53), the fancg mutant exhibits milder sensitivity toward each DNA-damaging agent tested. Indeed, xrcc3 mutant cells were more sensitive to cisplatin than fancg cells, as shown in Fig. Fig.4C4C.
To prove that the observed defects were indeed caused by specific gene disruption of FANCG, we expressed chicken or human FANCG cDNA in fancg mutant cells. FANCG expression was verified by EGFP fluorescence that was bicistronically expressed from the same construct. Full-length chicken FANCG cDNA was able to normalize cisplatin sensitivity (Fig. (Fig.4D),4D), while human FANCG cDNA gave no correction (data not shown).
We evaluated one aspect of HR capacity of fancg cells by measuring the efficiency of integration of transfected gene-targeting vectors. Transformants were selected with appropriate drugs, and targeting events were examined in each clone by Southern blot analysis. As summarized in Table Table2,2, depending on the targeting vector or locus, we observed modest reductions (ranging from ~2- to 9-fold) in the ratio of targeted to total (targeted plus random) integration events in fancg cells at three loci, but not for OVALBUMIN. The extremely high targeting efficiency at OVALBUMIN may have a bearing on this result. In fancg cells complemented with chicken FANCG cDNA, the efficiency of targeting at these three loci was restored, confirming that defective gene targeting was caused by the FANCG gene disruption. These data suggest that FANCG might participate in the HR process that mediates targeted integration.
To directly test whether fancg cells are defective for HR-mediated repair of DSBs, we used a transformant that carries the artificial recombination substrate SCneo (26) at the OVALBUMIN locus (13). The growth properties and sensitivities to genotoxic agents were indistinguishable between fancg cells with or without the integrated SCneo (data not shown). By using transient transfection of the plasmid encoding the rare restriction enzyme I-SceI, a DSB is produced in one of two tandem nonfunctional neo genes and can be repaired by HR, resulting in expression of a functional neo gene. Thus, the frequency of G418-resistant colonies represents the HR-directed DSB repair capacity.
In parallel cultures, a negative control plasmid, pBluescript, was used; <100 colonies from 107 cells arose from spontaneous recombination. However, after transfection of the I-SceI plasmid pcBASce into wild-type cells, a >100-fold induction occurred versus cells transfected with pBluescript. In contrast, we detected ~9-fold fewer G418-resistant colonies in fancg cultures, even after prolonged incubation following I-SceI expression (Fig. (Fig.5).5). Importantly, this defect was corrected nearly to the wild-type level by transient transfection with a plasmid expressing chicken FANCG but not with plasmid carrying human FANCG. These data indicate that FANCG is required for efficient HR repair of I-SceI-induced DSBs.
In mammalian cells, an induced DSB in SCneo undergoes HR-mediated repair through either short tract gene conversion (STGC) or long tract gene conversion (LTGC)/SCE events with the sister chromatid as the repair template (25). To examine how DT40 cells repair induced DSBs, we analyzed genomic DNA from G418-resistant clones by Southern blotting. We found that most repair events occurred by STGC in both parental (85%) and fancg (85%) cells, and relatively few LTGC/SCE events were detected (Fig. 5C and D), indicating that the HR repair proceeds in a normal manner in fancg cells even though the efficiency is reduced.
Given result described above, we investigated IR-induced chromosome aberrations in X-irradiated (2 Gy) cells at 3-h intervals after exposure. fancg cells showed significantly elevated chromosomal aberrations only in the 0- to 3-h interval (Fig. (Fig.6).6). Since the irradiated cells are asynchronous and evenly distributed throughout the cell cycle, cells entering M phase during this time period likely received irradiation in late S and G2 phases. Thus, the result suggests a modest role for FANCG in HR repair between sister chromatids in cells that have replicated their DNA (52). Chromosomal radiation sensitivity was previously reported for FA lymphoblasts and fibroblasts before complementation analysis was developed (3, 20).
The recent report that mutations of BRCA2 can cause FA (23) suggests that other FA proteins might also influence Rad51's activity. DSBs induce subnuclear Rad51 foci that are thought to represent sites of HR. We found that both X-ray- and MMC-induced Rad51 focus formation (Fig. (Fig.7)7) (IR data not shown) were indistinguishable between fancg and wild-type cells. Since BRCA2 is essential for Rad51 focus formation (64), our results suggest that any direct function of FANCG in HR is distinct from that of BRCA2 or Rad51 paralogs.
For all five DT40 rad51 paralog mutants, sensitivity to DNA damage was substantially suppressed by transfecting human RAD51 (51, 53). This finding is probably consistent with the proposed roles of Rad51 paralogs being mediators of Rad51's function (33, 43). To obtain further insights regarding any relationship of FANCG to Rad51, we stably introduced human RAD51 into our fancg cells and measured cisplatin sensitivity. Despite high human Rad51 expression that was comparable to the levels achieved in previous studies (Fig. (Fig.4E),4E), cisplatin sensitivity was actually potentiated in fancg mutant cells (Fig. (Fig.4D).4D). This might be due to overexpressed Rad51 acting in a dominant negative fashion by titrating away an essential factor in the cisplatin damage response pathway and thus further reducing damage tolerance. Together, these results lead us to conclude that FANCG plays a different role than BRCA2 and Rad51 paralogs in the HR repair pathway.
In this study, we provide genetic evidence for a contributory role of FANCG in HR as measured by three different assays. First, HR-mediated repair of I-SceI-induced chromosomal DSBs was reduced in fang mutant cells. Second, the overall efficiency of gene targeting was mildly decreased at three of four loci examined. Third, the levels of chromosome aberrations were elevated in X-irradiated cells in late S to G2 phase compared with cells irradiated in other phases. Given a primary role of HR in repairing IR-induced DSBs in late S and G2 phases in DT40 cells, compared with the secondary role of nonhomologous end joining (52), these findings are consistent with a role of FANCG in HR-mediated repair.
However, other HR-related processes examined were unaffected in the fancg cells. Specifically, the frequency of spontaneous or MMC-induced SCE appears to be normal. SCEs may be considered a surrogate marker of recombination efficiency, in contrast to chromosomal aberrations, which reflect biologically relevant HR repair. Moreover, the IR sensitivity of asynchronous fancg cells was normal, which differs from the case for rad51 paralog mutants. Thus, FANCG in DT40 cells seems to be important for optimal activity of only selected processes that depend on HR.
It has been shown that the FA multiprotein complex's assembly and nuclear localization is impaired in human fancg cells (15, 16). However, at present it is not clear whether defective complex localization underlies the entire phenotype of fancg cells. For example, human FANCG has been reported to interact directly with cytoplasmic CYP2E1 and appears to down-regulate the level of CYP2E1, a potential generator of reactive oxidative species (ROS) (14).
The FANCG protein likely has a function different from that of BRCA2 or RAD51 paralogs in HR for several reasons. First, we found no change in Rad51 focus formation in fancg cells, in keeping with one recent report (18) but not with another (9). Second, overexpression of human Rad51 did not increase cisplatin resistance of fancg mutant cells, whereas the defects in Rad51 paralog mutants were compensated for by elevating Rad51 protein (51, 53). (The effect of Rad51 overexpression in brca2 DT40 cells remains to be clarified.) We suggest that FANCG may function in parallel with BRCA2 by leading to common phenotypic consequences. It will be interesting to test this hypothesis by making mutant cells deficient in both BRCA2 and FANCG. Given the possibility that ATM regulates BRCA2 (58), this relationship may be relevant to the idea that FANCD2, which is also a phosphorylation target of ATM (54), may represent a convergence of two separate pathways, e.g., ATM signaling and the function of the FA nuclear complex.
So far, with the exception of BRCA2, no direct interactions between FA proteins and repair enzymes such as Rad51 and Mre11 have been identified. The defects in fancg cells are consistently milder than those observed in other HR-defective mutants such as the Rad51 paralogs (45, 48). These facts seem to be consistent with the view that FANCG plays an indirect, facilitating role in HR. For example, the FA proteins may function in part through chromatin remodeling (41). FANCA was reported to interact with Brg1 (37), a component of the human SWI/SNF complex, and FANCC interacts with the transcriptional repressor protein FAZF (22). It is possible that defective chromatin remodeling could influence cellular HR capacity (such as gene targeting efficiency) by modulating transcription, or temporal expression, of genes directly involved in HR.
Accumulating evidence supports a role for the FA proteins in determining the level of cytoplasmic ROS (reviewed in reference 1). Specifically, ROS may be elevated in FA group G human lymphoblasts, since they are reported to have elevated oxidative DNA damage, which is reversed in FANCG-complemented cells (14). Importantly, DNA damage by ROS appears to be a major cause of spontaneous chromosome breaks (28). Thus, the extensive chromosome aberrations often seen in FA cells (but not fancg DT40 cells) may be due to increased DNA damage by excessive generation of ROS along with compromised damage response processes linked to HR.
We thank Akira Shinohara (Osaka University) for the anti-Rad51 antibody; Maria Jasin (Sloan-Kettering Institute) for the SCneo substrate and I-SceI expression vector; Masayo Kimura, Keiko Oka, and Mayu Fujii for expert technical assistance; Hiroko Asahara for help with laser confocal microscopy; Y. Imajo and J. Kubota (Department of Therapeutic Radiology, Kawasaki Medical School) for irradiating cells with the linear accelerator; and Kazuko Hikasa and members of the Central Secretarial Office of Kawasaki Medical School for secretarial assistance. We also thank Shunichi Takeda (Kyoto University) and Ashok Venkitaraman (Cambridge University) for critically reading the manuscript.
This work was supported in part by a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology (to M.T. and N.M.) and by grants from the Novartis Foundation, Kobayashi Magobei Foundation, and Yamanouchi Foundation for Research on Metabolic Disorders. Financial support also came from Kawasaki Medical School as project research grants (projects 13-211 and 14-203). A portion of this work was prepared under the auspices of the U.S. Department of Energy by the Lawrence Livermore National Laboratory under contract W-7405-ENG-48 and research funded by the Low Dose Radiation Research Program, Biological and Environmental Research (BER), U.S. Department of Energy.