FANCJ, also called BACH1/BRIP1, is a 5′-3′ DEAH helicase, whose mutations are known as a risk factor for Fanconi anemia and also breast and ovarian cancer. FANCJ is thought to contribute to DNA double-strand break (DSB) repair and S-phase checkpoint through binding to multiple partner proteins, such as BRCA1 and TopBP1, but its molecular regulation remains unclear. We focused on DNA damage-induced phosphorylation of FANCJ and found that reagents that cause DSB or replication fork stalling induce FANCJ hyperphosphorylation. In particular, camptothecin (CPT) induced rapid and efficient FANCJ hyperphosphorylation that was largely dependent on TopBP1 and ATM-Rad3 related (ATR) kinase. Furthermore, DNA end resection that exposes single-strand DNA at the DSB site was required for hyperphosphorylation. Interestingly, upon CPT treatment, a dramatic increase in the FANCJ–TopBP1 complex was observed, and this increase was not alleviated even when ATR-dependent hyperphosphorylation was suppressed. These results suggest that FANCJ function may be modulated by hyperphosphorylation in a DNA end resection- and ATR-dependent manner and by FANCJ–TopBP1 complex formation in response to replication-coupled DSBs.
The DNA damage response is an essential system for protecting cells from genomic instability caused by DNA damage. One of the most severe types of DNA damage is DNA double-strand break (DSB), which is caused by ionizing radiation and DNA-damaging reagents such as camptothecin (CPT), an inhibitor of DNA topoisomerase I (Top1). CPT traps the Top1 reaction intermediate, Top1–DNA cleavage complex (Top1-cc), and indirectly induces DSBs through collision with the DNA replication fork (Pommier 2006). DSBs mediated by DNA replication activate the DNA damage checkpoint and homologous recombination (HR) to repair DSBs and re-establish the replication fork. The DNA damage checkpoint is organized by phosphatidyl inositol 3-kinase-like protein kinases (PIKKs), including ataxia-telangiectasia mutated (ATM) and ATM-Rad3 related (ATR) (Ciccia & Elledge 2010). When DSBs occur, these kinases are immediately activated and phosphorylate downstream factors such as Chk1 and Chk2, which induce cell cycle arrest to allow for the repair of DSBs. CPT-induced DSBs exhibit a unique DSB structure that harbors a single DNA end, different from that of IR- or etoposide (topoisomerase II inhibitor)-induced DSBs, which have two DNA ends. This kind of DSB, described as a ‘double-strand end’ (DSE), is repaired only by HR and not by nonhomologous end joining (Arnaudeau et al. 2001; Klein & Kreuzer 2002). To initiate HR, DNA end resection is required to expose single-strand DNA at DSE sites in a CtIP-dependent manner, and the replication protein A (RPA) complex and Rad51 are sequentially recruited to a single-stranded DNA region and the DNA replication fork is eventually re-established through a strand exchange reaction (Klein & Kreuzer 2002; Pommier 2006; Takeda et al. 2007).
FANCJ, also known as BACH1 or BRIP1, is a DEAH type 5′-3′ DNA helicase that was isolated based on its interaction with the product of the breast cancer–susceptible gene, BRCA1, through its BRCT domain at the C-terminus (Cantor et al. 2001). Mutations in FANCJ have been reported to be associated with an increased risk of breast and ovarian cancer (Cantor et al. 2001; Rutter et al. 2003; Seal et al. 2006; Rafnar et al. 2011). FANCJ is one of the 15 Fanconi anemia genes, in which cells are hypersensitive to DNA-cross-linking agents (Levitus et al. 2005; Levran et al. 2005; Litman et al. 2005). FANCJ has been reported to contribute to DSB repair mediated by HR and S-phase checkpoint activation (Cantor et al. 2001; Bridge et al. 2005; Litman et al. 2005; Greenberg et al. 2006), although the absolute underlying mechanism remains unclear. FANCJ forms complexes with some partner proteins, including BRCA1 and TopBP1, and complex formation is regulated by FANCJ phosphorylation. Extensive studies have shown that phosphorylation of Ser990 on FANCJ is required for interaction with BRCA1 and that Thr1133 phosphorylation is required for interaction with TopBP1 (Rodriguez et al. 2003; Yu et al. 2003; Gong et al. 2010). Both sites are followed by a proline residue, which creates a consensus motif for phosphorylation by cyclin-dependent kinases (CDKs). The formation of these complexes likely defines the function of FANCJ for the maintenance of genomic integrity. In response to IR, hydroxyurea (HU) and ultraviolet light (UV) exposure, phosphorylated FANCJ is detected as a slower-migrating band by Western blotting (Peng et al. 2006). It is also reported that dephosphorylation of FANCJ activates its intrinsic helicase activity, which is required for timely progression through S-phase (Kumaraswamy & Shiekhattar 2007). These findings indicate that the cellular functions of FANCJ are regulated by phosphorylation.
In this study, we found that FANCJ is efficiently hyperphosphorylated in response to CPT treatment. FANCJ hyperphosphorylation was dependent on TopBP1 and ATR. CtIP-dependent DNA end resection was also required for CPT-induced FANCJ hyperphosphorylation. In addition, we found that CPT treatment enhances the formation of the FANCJ–TopBP1 complex in a fashion distinct from FANCJ hyperphosphorylation. These results show novel insight into the regulation of FANCJ phosphorylation in response to DNA damage.
FANCJ hyperphosphorylation is induced by DNA damage
FANCJ is phosphorylated in S-phase, and IR, HU and UV exposure induce FANCJ mobility shift in SDS-PAGE due to DNA damage-induced phosphorylation (Peng et al. 2006). We also observed that CPT, neocarzinostatin (NCS) as well as HU treatment induced a shift in FANCJ mobility that was detectable by Western blotting (Fig. 1A). NCS is a radiomimetic reagent that directly induces DSBs. HU, a ribonucleotide reductase inhibitor, leads to stalling of the DNA replication fork. CPT indirectly causes DSEs mediated by DNA replication, resulting in collapse of the replication fork. After 4 h of treatment, the mobility shift induced by NCS was slight, whereas the shift induced by HU was more significant. By comparison, CPT treatment dramatically induced FANCJ mobility shift even at 1 h of treatment. The results suggest that FANCJ is efficiently modified by the coupling of DNA replication stall with the induction of DSBs.
To confirm whether the CPT-induced mobility shift of FANCJ reflects phosphorylation, as in the response to UV exposure (Peng et al. 2006), immunoprecipitated FANCJ was treated with a protein phosphatase in vitro. Protein phosphatase treatment abrogated the FANCJ mobility shift and in fact increased the mobility of FANCJ above that of the nontreated sample (Fig. 1B). This result indicates that the CPT-induced FANCJ mobility shift is the result of phosphorylation, and suggests that FANCJ exists in a constitutively phosphorylated state and is further hyperphosphorylated after the induction of DNA damage. The mobility shift caused by DNA damage-induced hyperphosphorylation was confirmed using a different antibody against FANCJ provided by Sigma (Fig. 1C). Moreover, pre-treatment with aphidicolin or HU abolished CPT-induced FANCJ phosphorylation (Fig. 1D), and S-phase-synchronized cells showed efficient phosphorylation of FANCJ in comparison with G1-phase-synchronized cells (Fig. S1 in Supporting Information). NCS treatment induced only a slight mobility shift in both S-phase-synchronized cells and asynchronized cells (Fig. 1A and Fig. S1 in Supporting Information). These results confirm that FANCJ is phosphorylated in response to DSEs caused by DNA replication.
To evaluate the kinetics of FANCJ phosphorylation, a 2-h time course analysis was performed. FANCJ phosphorylation was weakly observed at 40 min and a marked bandshift was observed at 60 min (Fig. 1E). These phosphorylation kinetics are similar to those of replication protein A 34 (RPA34). The DNA damage checkpoint kinases Chk1 and Chk2, downstream factors of ATR and ATM, respectively, were rapidly phosphorylated after the addition of CPT (Fig. 1E). These results suggest the possibility that multiple intracellular processing steps are required for both FANCJ hyperphosphorylation and RPA34 phosphorylation. During processing, single-strand DNA may be exposed at the site of the collapsed replication fork.
End resection-ATR pathway is required for CPT-induced FANCJ hyperphosphorylation
Several proteins other than BRCA1 and TopBP1, such as MLH1 and BLM, have been reported to interact with FANCJ (Cantor et al. 2001; Peng et al. 2007; Gong et al. 2010; Suhasini et al. 2011). To assess the effect of these interactions on FANCJ hyperphosphorylation, the expression of these genes was suppressed using specific siRNAs. No significant effect on FANCJ hyperphosphorylation was observed in MLH1- or BLM-knockdown cells (Fig. S2 in Supporting Information), whereas knockdown of TopBP1 and BRCA1 expression suppressed the FANCJ mobility shift markedly and moderately, respectively (Fig. 2A). TopBP1 is required for ATR activation during the DNA damage response, which is considered to require a single-stranded DNA region (Kumagai et al. 2006). BRCA1 is a multifunctional protein that also plays a major role in the DNA damage response. Some recent reports indirectly show that BRCA1 may play a role in modulating DNA end resection to expose single-strand DNA for DSB repair by HR through association with CtIP and the Mre11–Rad50–Nbs1 complex (Yu & Chen 2004; Chen et al. 2008; Bunting et al. 2010). When considered with these findings, our results raise the possibility that TopBP1-dependent ATR activation after DNA end resection is a critical step for FANCJ hyperphosphorylation. To test this possibility, cells were treated with ATR and ATM inhibitors. Caffeine, an inhibitor of both ATR and ATM, suppressed CPT-induced FANCJ hyperphosphorylation. This suppressive effect was also observed in cells treated with the ATM inhibitor KU55933 (Fig. 2B). Moreover, when ATR and ATM expression was knocked down using specific siRNAs, ATR knockdown significantly suppressed FANCJ hyperphosphorylation, whereas hyperphosphorylation was partially suppressed in ATM-knockdown cells (Fig. 2C). These results suggest that ATR is a critical kinase for FANCJ hyperphosphorylation. It has been reported that FANCJ is phosphorylated on Ser/Pro or Thr/Pro sites that were predicted to be phosphorylated by CDKs (Yu et al. 2003; Dephoure et al. 2008; Gong et al. 2010). Treatment with the CDK2 inhibitors roscovitine and CVT-313 resulted in marked suppression of CPT-induced FANCJ hyperphosphorylation (Fig. 2D), suggesting that this hyperphosphorylation requires constitutive phosphorylation by CDK2, consistent with the results of protein phosphatase shown in Fig. 1B.
To investigate the contribution of DNA end resection to FANCJ hyperphosphorylation, the expression of CtIP, a regulatory factor of end resection, was knocked down using specific siRNA. FANCJ hyperphosphorylation as well as RPA34 phosphorylation, generally considered as a marker of single-strand DNA produced by DNA end resection, was strongly suppressed by CtIP knockdown, suggesting that DNA end resection promotes CPT-induced FANCJ hyperphosphorylation (Fig. 3A). The effects of ATM inhibition and knockdown suggest that ATM might be involved in FANCJ hyperphosphorylation. This possibility is supported by the finding that ATM also contributes to DNA end resection (You et al. 2009).
The role of FANCJ as a DNA helicase and its DNA end resection-dependent hyperphosphorylation prompted us to examine whether FANCJ hyperphosphorylation affects its recruitment to chromatin. Cell extracts were fractionated into NP-40 soluble and insoluble fractions. In CPT-treated cells, the soluble fraction contained both shifted (hyperphosphorylated) and nonshifted bands, whereas only shifted bands were evident in the insoluble fraction (Fig. 3B). This suggests that hyperphosphorylated FANCJ is enriched on the chromatin in the presence of CPT and that the affinity of FANCJ to chromatin might be modulated by its phosphorylation state. However, it remains unknown whether FANCJ is recruited to chromatin in response to CPT. In addition, we could not detect FANCJ nuclear foci, although it is previously reported that FANCJ forms foci in response to several DNA-damaging agents (Cantor et al. 2001; Peng et al. 2006; Gupta et al. 2007).
CPT promotes FANCJ–TopBP1 complex formation
Protein phosphorylation is a definitive factor for protein–protein interaction. Greenberg et al. (Greenberg et al. 2006) reported that, in response to IR, FANCJ contributes to the cellular response in a complex with TopBP1 and BRCA1. To investigate the effect of FANCJ hyperphosphorylation on complex formation with TopBP1 and BRCA1, immunoprecipitation with TopBP1 and BRCA1 was performed and the co-precipitation of FANCJ was analyzed by Western blotting. The amount of FANCJ that co-precipitated with TopBP1 was dramatically increased after CPT treatment, in contrast to the amount precipitated by BRCA1 immunoprecipitation (Fig. 4A). FANCJ that co-precipitated with TopBP1 showed mobility identical to that of the hyperphosphorylated form of FANCJ. Reciprocal immunoprecipitation with FANCJ also showed an increase in the amount of co-precipitated TopBP1 (Fig. 4B). These results suggest that FANCJ–TopBP1 complex formation is promoted after CPT treatment. Next, the dependency of FANCJ hyperphosphorylation on the FANCJ–TopBP1 complex formation was examined using ATR/ATM inhibitors and ATR RNAi. Cells were treated with CPT with or without caffeine or KU55933, and cell extracts were immunoprecipitated with TopBP1 antibody. Levels of co-precipitated FANCJ were still increased in caffeine-treated cells, even though hyperphosphorylation was suppressed (Fig. 4C). ATR-knockdown cells also showed an increase in co-precipitated nonshifted FANCJ levels (Fig. 4D). These results indicated that CPT treatment increased FANCJ–TopBP1 complex formation in a FANCJ hyperphosphorylation-independent manner.
FANCJ is a DNA helicase that participates in DSB repair and the DNA damage checkpoint. Here, we found that DNA damage, especially CPT-induced DSE mediated by DNA replication, induced FANCJ hyperphosphorylation in an end resection- and ATR-dependent manner. CPT treatment also enhances FANCJ–TopBP1 complex formation, even though ATR is inactivated.
Camptothecin causes the accumulation of a unique initial DNA damage mediated by ‘Top1-cc’ harboring a DNA single-strand break, which is potentially capable of inducing a secondary DNA strand break mediated by DNA replication (Pommier 2006). Efficient hyperphosphorylation of FANCJ was induced to a greater degree by CPT compared to NCS and HU, which implies that DSB or a stalled replication fork alone is not sufficient for efficient FANCJ hyperphosphorylation (even though ATR is quickly activated in response to HU) and that the structure formed after DSE resection at a collapsed replication fork can be suitable for FANCJ hyperphosphorylation. Previous comprehensive study of the relationship between protein phosphorylation and regulation of the cell cycle showed that several serine and threonine residues on FANCJ are phosphorylated (Dephoure et al. 2008). Some of the sites including Ser990 are followed by proline, a known putative CDK target sequence. Phosphorylation of Ser990 on FANCJ is crucial for interaction with BRCA1 via BRCT domains (Yu et al. 2003). Recent study also showed that phosphorylation of Thr1133 on FANCJ by CDK during S-phase is required for interaction with TopBP1 (Gong et al. 2010). These reports imply that FANCJ is constitutively phosphorylated at least by CDKs. We showed here that application of a specific CDK2 inhibitor has a suppressive effect on FANCJ hyperphosphorylation, suggesting that DNA damage-independent phosphorylation of FANCJ, probably by CDK2 during S-phase, is required for DNA damage-induced FANCJ hyperphosphorylation. In addition, FANCJ has four Ser/Gln and two Thr/Gln sites that are putative targets of PIKKs. Although we attempted to identify the residues phosphorylated by ATR using amino acid–substituted mutants of FANCJ, our attempts were not successful because transiently expressed FANCJ did not shift after CPT treatment (results not shown). FANCJ in CPT-treated cells showed much slower mobility than in nontreated cells, which may possibly indicate that multiple sites were phosphorylated in response to CPT-induced DNA damage. Previous study also showed that, in addition to serine and threonine residues, tyrosine residues are also phosphorylated (Dephoure et al. 2008), which suggests that tyrosine kinases may also contribute to FANCJ phosphorylation.
After CPT treatment, FANCJ found in the chromatin fraction was in a hyperphosphorylated form. This suggests that functional FANCJ in response to CPT-induced DNA damage probably exists in a hyperphosphorylated form that is likely to contribute to DSE repair or the DNA damage checkpoint. Because FANCJ nuclear foci formation has been reported to be suppressed in BRCA1-deficient cells (Cantor et al. 2001; Gupta et al. 2007), partial suppression of FANCJ hyperphosphorylation in BRCA1-knockdown cells shown in Fig. 2A might be explained by the inefficient recruitment of FANCJ to damaged chromatin in the absence of BRCA1. This and the result shown in Fig. 3B provide a possibility that chromatin-recruited FANCJ might be phosphorylated after the induction of DNA damage.
After CPT treatment, hyperphosphorylated FANCJ efficiently interacted with TopBP1, an important factor for the activation of the DNA damage checkpoint and DNA replication. TopBP1 is required for ATR activation through interaction with the Rad9–Rad1–Hus1 complex and ATR (Kumagai et al. 2006; Delacroix et al. 2007). Our results suggest a CPT-induced enhancement of the FANCJ–TopBP1 complex formation. TopBP1 has eight BRCT domains and constitutively exists in complex with FANCJ through interaction with its BRCT7 and 8 domains (Gong et al. 2010). The increase in the FANCJ–TopBP1 complex allowed us to consider the possibility that the interaction between FANCJ and TopBP1 through the TopBP1 BRCT domains is enhanced by FANCJ hyperphosphorylation after CPT treatment. However, neither caffeine treatment nor ATR knockdown affected CPT-induced FANCJ–TopBP1 complex formation, although the FANCJ mobility shift was suppressed (Fig. 4C,D). The possibility remains that other ATR-independent phosphorylations are required for CPT-induced increase in the FANCJ–TopBP1 complex. As a candidate for the kinase other than ATR required for FANCJ–TopBP1 complex formation, we investigated CDK by using an inhibitor. Interestingly, CDK inhibition did not affect the complex formation either (Fig. S3 in Supporting Information). These results provide another possibility that FANCJ–TopBP1 complex formation is an initial step before FANCJ hyperphosphorylation. TopBP1 also regulates DNA replication by associating with other proteins including CDC45, Treslin and GEMC1, and loading them onto the pre-replication complex leading to replication origin firing (Schmidt et al. 2008; Balestrini et al. 2010; Kumagai et al. 2010). Reportedly, suppression of CDC45 loading onto chromatin after IR irradiation is not observed in FANCJ-knockdown cells (Greenberg et al. 2006). FANCJ is likely to trap TopBP1 by increasing complex formation, which might modulate TopBP1-dependent firing of replication origins until checkpoint recovery. To investigate these hypotheses and further characterize the role of FANCJ in the maintenance of genome integrity, it will be necessary to identify the FANCJ phosphorylation sites and establish phosphorylation-deficient FANCJ mutants.
Cell culture and reagents
HeLa cells were maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum. CPT, HU, roscovitine, NCS and KU-55933 were purchased from Sigma-Aldrich. Caffeine was purchased from Nacalai Tesque. Aphidicolin and CVT-313 were purchased from Calbiochem.
Western blotting and antibodies
Preparation of cell lysate, SDS polyacrylamide gel electrophoresis and protein transfer were performed as described previously (Sakasai & Tibbetts 2008). For Western blotting, the following antibodies were used: anti-FANCJ (NB100-416; Novus Biologicals, B1310; Sigma-Aldrich), anti-Chk2 pT68 (#2661; Cell Signaling Technology), anti-Chk2 (#3440; Cell Signaling Technology), anti-Chk1 pS345 (#2348; Cell Signaling Technology), anti-Chk1 (sc-8408; Santa Cruz Biotechnology), anti-RPA34 (Cat2461-1; Epitomics), anti-TopBP1 (AB3245; Millipore), anti-BRCA1 (sc-6954; Santa Cruz Biotechnology), anti-ATM (GTX70103; GeneTex), anti-ATR (#2790; Cell Signaling Technology), anti-CtIP (sc-5970; Santa Cruz Biotechnology), anti-histone H3 (H9289; Sigma-Aldrich) and anti-β-actin (ab6276; Abcam).
For immunoprecipitation, HeLa cells were lysed with 0.5% NP-40 buffer containing 20 mm Tris–HCl (pH 8.0), 150 mm NaCl, 1 mm EDTA, and protease and protein phosphatase inhibitor cocktails (Nacalai Tesque). Cell lysates were incubated with anti-TopBP1 or BRCA1 (sc-642; Santa Cruz Biotechnology) antibodies or normal rabbit IgG (sc-2027; Santa Cruz Biotechnology) and precipitated with Protein G Sepharose, Fast Flow (Sigma-Aldrich). After washing, proteins were eluted from the beads with 1× SDS sample buffer.
Protein phosphatase assay
After immunoprecipitation with anti-FANCJ antibody (B1310; Sigma-Aldrich) as described in section 2.3, beads were washed with phosphatase-free buffer and incubated with λ protein phosphatase (New England Biolabs) for 30 min at 30 °C. Proteins were eluted with 1× SDS sample buffer.
siRNAs and transfection
To knock down the expression of each gene, the following siRNA sequences were used: control, UCUUAAUCGCGUAUAAGGC; TopBP1#1, CUCACCUUAUUGCAGGAGA; TopBP1#2, CTCACCTTATTGCAGGAGA (Kim et al. 2005); BRCA1#1, GGACGUUCUAAAUGAGGUA; BRCA1#3, GGAACCUGUCUCCACAAAG (Martin & Ouchi 2005); ATR, CCUCCGUGAUGUUGCUUGA (Yang et al. 2008); ATM, GCGCCUGAUUCGAGAUCCU (Lin & Dutta 2007); CtIP, GCUAAAACAGGAACGAAUC (Yu & Chen 2004). These siRNAs were purchased from Takara Bio, Sigma-Aldrich or Gene Design. siRNA transfection was performed using RNAiMAX transfection reagent (Invitrogen) following the manufacturer's protocol.
Soluble fractions containing cytosolic and nucleoplasmic proteins were extracted by sequential incubation with 0.1% and 0.5% NP-40 buffer containing 20 mm Tris–HCl (pH 8.0), 150 mm NaCl, 1 mm EDTA, and protease and protein phosphatase inhibitor cocktails. Extracted soluble fractions were boiled with 2× SDS sample buffer. Insoluble pellets were lysed and boiled with 2× SDS sample buffer.
This research was supported by the Science and Technology Incubation Program in Advanced Regions from the funding program ‘Creation of Innovation Centers for Advanced Interdisciplinary Research Areas’ from the Japan Science and Technology Agency, commissioned by the Ministry of Education, Culture, Sports, Science and Technology. This work was also supported in part by Grants-in-aid from the Ministry of Education, Science, Sports, and Culture of Japan (H.K. and Y.M.). We would like to specially thank Taiho Pharmaceutical Co. Ltd. for their full cooperation. We also thank the members of the Department of Molecular Oncology at Kyushu University.