|Home | About | Journals | Submit | Contact Us | Français|
Fanconi anemia (FA) is a genetic disease featuring genomic instability and cancer predisposition1. Nine FA genes have been identified, and their products participate in a DNA damage response network involving BRCA1 and BRCA22,3. We have previously purified a FA core complex containing the FANCL ubiquitin ligase and 6 other FA proteins4–6. Each protein in this complex is essential for monoubiquitination of FANCD2, a key reaction in the FA DNA damage response pathway2,7. Here we show that another component of this complex, FAAP250, is mutated in FA patients of a new complementation group (FA-M). FAAP250, renamed FANCM, has sequence similarity to known DNA repair proteins, including archaeal Hef, yeast Mph1 and human ERCC4/XPF. FANCM can dissociate DNA triplex, possibly due to its ability to translocate on duplex DNA. FANCM is essential for FANCD2 monoubiquitination and becomes hyperphosphorylated in response to DNA damage. Our data suggest an evolutionary link between FA proteins and DNA repair; FANCM may act as an engine that translocates the FA core complex along DNA.
We have previously shown that 7 out of 9 components of the FA core complex are FA proteins (FANC-A, B, C, E, F, G, and L)4–6. Using mass spectrometry, we identified another component, FAAP250 (Fig. 1a), as KIAA1596, a hypothetical protein with unknown function. Antibodies raised against KIAA1596 specifically recognized the 250 kD polypeptide of the FA core complex immunopurified by a FANCA antibody, supporting the identity of KIAA1596 as FAAP250 (Fig. 1b).
Several lines of evidence suggest that FAAP250 is an integral component of the FA core complex. First, FAAP250 was detected in the FA core complex immunoisolated by an anti-Flag antibody from cells expressing either Flag-tagged FANCA, or Flag-tagged FANCL (Fig. 1b). Second, FAAP250 was coimmunoprecipitated by antibodies against multiple FA core components components (FANCA, C, and F) from lymphoblastoid cells of a normal individual, but not from patient cells deficient in the corresponding FA proteins (Fig. 1c). Third, reciprocal immunoprecipitation in HeLa cells using the FAAP250 antibody showed co-precipitation of multiple FA core complex components, such as FANCL, FANCA, and FANCG (Fig. 1d and data not shown).
Importantly, depletion of FAAP250 in HeLa and HEK293 cells by siRNA drastically reduced the levels of monoubiquitinated FANCD2 under both basal and mitomycin C (MMC)-induced conditions (Fig. 2a and data not shown). This is similar to observations following depletion of other core complex components5,6,8. FAAP250-depleted cells also showed increased MMC-sensitivity (Supplementary Fig. 1), a feature that resembles FA cells. We infer that FAAP250, like the other core complex components, has an essential function for FANCD2 monoubiquitination, and its gene likely underlies a distinct FA complementation group.
We noticed that a long form of FAAP250, termed FAAP250-L, was induced when cells were treated with MMC, hydroxyurea, and cisplatin (Fig. 2a, 2b, and data not shown), implying that FAAP250 is post-translationally modified in response to DNA damage or replication block. FAAP250-L was also detected when FA cells deficient in FANCA, FANCB, FANCE and FANCL were treated with MMC (Fig. 2c and data not shown), suggesting that this modification occurs independently of the ubiquitin ligase activity of the FA core complex. When treated with a phosphatase, both short (FAAP250-S) and long forms (FAAP250-L) showed increased mobility on SDS-gel (Fig. 2d), suggesting that FAAP250 is a phosphoprotein that becomes hyperphosphorylated in response to genotoxic stress.
FAAP250 has a single homologue in other vertebrate species. It contains two potential enzymatic domains: one homologous to a superfamily of DEAH-box helicases or DNA-stimulated ATPases; the other homologous to the endonuclease domain of ERCC4/XPF (Fig. 3). Interestingly, FAAP250 is closely related to archaeal protein Hef (E value 10−43), which possesses both helicase and endonuclease activities, and can resolve stalled replication forks 9,10. Hef has also been proposed to be the archaeal homolog of ERCC4/XPF11,12, an endonuclease essential for nucleotide excision repair in eukaryotes13,14. FAAP250 and ERCC4/XPF are the only proteins in the human genome with a helicase/ATPase domain at the N-terminus and an endonuclease domain at the C-terminus, suggesting that they may have a common ancestor.
While the helicase/ATPase domains of FAAP250 proteins are highly conserved in the seven motifs present in the SF2 family of helicases (Supplementary Fig. 2), their endonuclease domains are degenerate at several residues that are absolutely conserved in the ERCC4/XPF and the Mus81 family of endonucleases15 (Supplementary Fig. 3). Mutation of the corresponding residues in ERCC4/XPF either inactivates or strongly reduces its endonuclease activity15. It seems likely that during evolution, the nuclease domains of FAAP250 proteins may have diverged, becoming dispensable and non-functional. Indeed, proteins with high homology to the helicase domains of FAAP250 are present in lower eukaryotic species such as Drosophila (CG7922) and yeast (MPH1) (E values: 10−78 and 10−60, respectively). These proteins may have the same antecedents as FAAP250, but along with the loss of nuclease domains during evolution. MPH1 protein has a DNA helicase activity16 and participates in an error-free DNA damage bypass pathway to protect the genome from spontaneous and chemically-induced damage17,18. The similarity between FAAP250 and MPH1 suggests that they may be part of common mechanisms that guard genome stability.
The helicase domains of ERCC4/XPF proteins are also degenerate and perhaps nonfunctional 11,12 (Supplementary Fig. 2). Thus, FAAP250 and ERCC4/XPF may have derived from a common ancestor that, like archaeal Hef, has both functional helicase and nuclease domains. They may have subsequently diverged, with ERCC4/XPF retaining its endonuclease activity and FAAP250 maintaining its helicase/ATPase activity.
To assess possible biochemical activity of FAAP250, we generated cell lines that stably express FAAP250 tagged with a Flag-epitope and immunoisolated the tagged protein by using the flag antibody (Fig. 4a). As a control, we made a point mutant that substitutes arginine for a conserved lysine within the helicase motif I (Supplementary Fig. 2; Fig. 4a). This substitution in other helicase domains completely inactivates ATPase and helicase activities19,20. Consistent with previous instances, the wildtype FAAP250 protein, but not its mutant, exhibited single- and double- stranded DNA-stimulated ATPase activity (Fig. 4b). The fact that the mutant protein has no detectable ATPase activity indicates that this activity is derived from the helicase domain of FAAP250.
We found that the FAAP250 protein displayed no detectable helicase activity in strand-displacement assays using substrates consisting of double-stranded DNA, 4-way junction, or replication fork, which can be resolved by either MPH1 or Hef10,16 (Fig. 4c and Supplementary Fig. 4). The absence of detectable helicase activity in FAAP250 argues that it may function differently from MPH1 and Hef. However, we cannot rule out the possibility that FANCM may be a structure-specific helicase, and the absence of a detectable helicase activity can be due to lack of normal physiological environment in our assays.
Many proteins with SF2 helicase domains can use the energy of ATP-hydrolysis to translocate on DNA21–23. One way to demonstrate translocase activity is the triple-helix displacement assay. In this assay, a triple helix (H-DNA) is formed in which each nucleotide of the third strand forms Hoogsteen base pairs with Watson-Crick base pairs of the duplex. When the translocase proceeds through the triplex, it displaces the third strand. Wildtype FAAP250, but not its mutant, displayed triple-helix displacement activity (Fig. 4d). To exclude the possibility that FAAP250 may displace the triplex through direct binding and dissociation, we used a blunt triplex in this assay, and found that FAAP250 failed to displace this substrate (Fig. 4e). Together, these results suggest that FAAP250 can dissociate DNA triplex, possibly due to an ability of the protein to translocate on duplex DNA.
We screened FA patient derived cells for FAAP250 mutations by immunoblotting. FAAP250 was detected in cells from all known complementation groups (Fig. 5a and data not shown). This result implied that if FAAP250 is mutated in FA patients, it should be in a new group. FAAP250 was undetectable in a cell line from patient EUFA867 that had been excluded from many known groups (A, B, C, E, F, G) by linkage analysis, cell fusion and cDNA transfection (data not shown). Sequencing of the patient’s cDNA revealed a nonsense mutation (2171C>A; S724X) in exon 13, and the absence of the sequence encoded by exon 15, which was caused by a 2554 bp deletion spanning a part of intron 14 and almost the entire exon 15 (Fig. 5b; Supplementary Fig. 5). In theory, these mutations could lead to truncated products of 80 kD and 160 kD, respectively. However, neither product was detectable (data not shown), suggesting that they may be unstable.
Another FA patient in the same family, a sibling of EUFA867, was found to carry the identical mutations in his blood DNA, providing additional evidence for FAAP250 as the gene responsible for FA in these patients. The nonsense mutation was detected in DNA from blood and from a lymphoblastoid cell line of the patients’ healthy mother, consistent with maternal inheritance of this allele. The genomic deletion was not detected in DNA from blood or a lymphoblastoid cell line of the healthy father, but linkage analysis with tightly linked flanking markers (D14S259 and D14S1027) showed that both siblings had inherited the same paternal allele, and correct paternity was established with multiple markers (data not shown). Because it is unlikely that the same deletion has occurred independently twice in EUFA867 and her sibling, the most likely explanation is that the father is mosaic for this deletion: his germ cells, but not blood cells, carry the mutation, which was thus transmitted to his children. In summary, our results show that FAAP250 is mutated in patients of a new FA complementation group (FA-M), and is therefore termed FANCM.
Numerous attempts to use FANCM cDNA to complement FA-M (EUFA867) cells for MMC-sensitivity (Supplementary Fig. 1) and defective FANCD2 monoubiquitination (see below) have failed. The main reason is that we were unable to ectopically express FANCM in FA-M (EUFA867) cells using both plasmid and viral vectors with either constitutive or inducible promoters. As an alternative, we used other approaches to evaluate the importance of FANCM in the FA core complex and pathway. First, FA-M cells display drastically reduced levels of FANCA and FANCG, and modestly reduced level of FANCL (Fig. 5c). This feature resembles FA cells lacking FANCA or FANCB, in which the levels of FANCG and FANCL are correspondingly reduced. The fact that FANCA and FANCG are unstable in FA-M cells underscores the importance of FANCM for the integrity of the core complex. Second, FA-M cells exhibit deficient nuclear localization of FANCA and FANCL (Fig. 5d), a characteristic also observed in FA-B cells. Third, the interaction between FANCM and FANCA is defective in FA cells lacking other core complex components (Fig. 5e, 5f), and the interaction between FANCM and FANCL is absent in FA-A and FA-B cells (Fig. 5f), providing additional evidence for FANCM as part of the core complex.
Notably, we found that the FA-M cells are defective in FANCD2 monoubiquitination (Fig. 5c), which is a common feature of all FA patient cells with a mutated core complex component. This result is also consistent with siRNA data in HeLa and HEK293 cells where depletion of FANCM drastically reduced levels of monoubiquitinated FANCD2 (Fig. 2a and data not shown). Together, these data indicate that FANCM, like the other 7 FA core complex proteins, is necessary for FA core complex assembly and FANCD2 monoubiquitination.
Although FA proteins have been implicated in DNA repair, the interactions between FA proteins and DNA are poorly understood, because the known FA proteins lack DNA-related enzymatic domains or activities. Here we demonstrated that FANCM has a conserved helicase domain and a DNA-translocase activity. A companion paper identified another FA protein, FANCJ, as BACH1/BRIP124, a known DNA helicase25. The discovery of two FA proteins with helicase domains or activities suggests a mechanism of direct participation in DNA repair by the FA proteins.
We propose that FANCM may have at least three roles in this process. First, it may have a structural role to allow assembly of the FA core complex; because in its absence, the nuclear localization and stability of several FA proteins are defective (Fig. 5c and d). Second, FANCM may act as an engine that translocates the core complex along DNA. Speculatively, this translocation may allow the core complex to sense and locate to the damaged DNA, which could be an important step either before or after FANCD2 monoubiquitination. Third, FANCM is hyperphosphorylated in response to DNA damage, suggesting that it may serve as a signal transducer through which the activity of the core complex is regulated. A DNA damage checkpoint kinase, ATR, has been shown to act upstream of the FA pathway26,27, but its substrate has not been defined. FANCM contains multiple predicted ATR phosphorylation sites (data not shown), and may serve as a substrate through which ATR regulates FANCD2 monoubiquitination.
In contrast to FANCM which acts upstream of the FA pathway (in terms of FANCD2 monoubiquitination) and displays no helicase activity, FANCJ functions downstream28 and possesses a 5′ to 3′ helicase activity25. Possibly, FANCJ may directly repair damaged DNA with this activity. It remains to be determined how the function of FANCJ is connected to the FA pathway. Notably, BLM, a helicase with 3′ to 5′ polarity (opposite of FANCJ), associates with FANCM (Fig. 1d) and the rest of the FA core complex4. The presence of three proteins with distinct helicase properties in the FA-related complexes suggests a coordinated action of FA proteins in DNA repair.
Epstein Barr virus immortalized lymphoblastoid cell lines were established from blood samples of FA patient EUFA867 and her parents who had been referred to the EUFAR cell repository for genetic subtyping. The research was carried out after approval by the Guy’s Hospital Research Ethics Committee (London) and the Institutional Review Board (VU University Medical Center, Amsterdam). We have obtained informed consent on all human subjects involved in our work. Cell lines were cultured in RPMI with 10% fetal calf serum.
We immunopurified FA core complex with a FANCA antibody from HeLa cells as described4. We identified FAAP250 by liquid chromatography-mass spectrometry (LC/MS) analysis. The mass spectrometry data are not shown but available upon request. We made two polyclonal rabbit antibodies (FAM-1Ab and FAM-2Ab) against FAAP250 (FANCM) by using two chimeric proteins containing regions of FAAP250 (aa. 1190–1273 and aa. 1643 to 2048 respectively) fused to maltose-binding protein (MBP) as immunogen. These fusion proteins were expressed and purified from E.coli according to manufacturer’s protocols (New England Biolabs), and used for immunization as well as for purification of the antibody. The immunoprecipitation condition has been described4. We prepared nuclear and cytoplasmic extracts as described4.
HEK293 cells were either treated with hydroxyurea for 24hrs or untreated. They were then extracted with lysis buffer (50mM Tris-HCl, pH 7.5, 300mM NaCl, 1% Triton X-100 and 1mM DTT) supplemented with a protease inhibitor cocktail (Roche) and phosphatase inhibitors(50mM NaF and 2mM Na3V04). FAAP250 was immunoprecipitated using FAM-1Ab. The immunoprecipitate was washed twice with λ phosphatase buffer (New England Biolabs), and resuspended in 50 μl of λ phosphatase buffer either in the presence or absence of phosphatase inhibitors. Four hundred units of λ phosphatase (New England Biolabs) was added to each sample, followed by incubation at 30°C for 1 hr. Samples were separated by 6% SDS-PAGE and immunoblotted for FAAP250 using FAM-1Ab antibody.
We purchased a KIAA1596 cDNA clone from Origen Inc. Sequencing revealed that this clone contains an open reading frame of 2048 aa., longer than the predicted sequence of XP_048128 (2014 aa.) at NCBI. The cDNA was subcloned by standard molecular biology techniques into mammalian expression vectors, pCEP4 and pMEP4 (InVitrogen). The cDNA was also linked with the Flag-epitope at the N-terminus, and cloned into pIRESneo3 vector (Clonetech). The Flag-tagged FAAP250 cDNA has also been cloned into the pCEP4 vector and the retroviral vector using the Gateway recombinant DNA system (InVitrogen). These constructs have been used for complementation and expression studies.
Transfection of HeLa cells using siRNA oligonucleotides was performed as described5. The FAAP250 siRNA oligo used two different 19-nucleotide target sequences (CUAUGCUUAUUGCCAGGUU and AGGCUGUGCAACAAGUUAU) and was purchased from Dharmacon Research. HEK293 cells were transfected with Flag-FANCM cDNA in a pIRESneo3 vector, and were selected for G418-resistant clones. These clones were screened for expression of Flag-FANCM by immunoblotting. Those expressing high levels of Flag-FANCM were grown in large scale for immunopurification of the protein. Patient cell line EUFA867 has been transfected with pIRESneo3, pCEP4 and pMEP4 vectors containing FANCM cDNA and selected on medium containing either G418 (0.4 mg/ml) or hygromycin (0.1 mg/ml). For inducible expression of FANCM in EUFA867 cells, the GeneSwitch system (InVitrogen) was used. Cross-linker sensitivity for EUFA867 cells was determined using a MMC-induced growth inhibition assay, as described6. The MMC-sensitivity assay for FAAP250-depleted HeLa cells was done as described29, with the following modifications. HeLa cell were transfected with either a FAAP250 siRNA oligo or a control oligo, and grown for 60 hrs. They were then plated at a density of 500 cells per 10 cm dish. After 8 hrs, cells were treated MMC at the indicated concentration (Supplementary Fig. 1) for 2 hrs. The drug was then washed away and fresh media were added. After 9–12 days, the colonies were fixed, stained and counted. FANCM expression was determined by either direct immunoblotting or by immunoprecipitation-coupled immunoblotting.
Flag-FANCM wildtype and mutant proteins were immunoisolated from HEK293 cell clones that have been selected for their high expression. The protocols were essentially as described 5, but with an important modification: the nuclear extract was depleted with a BLM antibody prior to immunoprecipitation with the Flag antibody, since FANCM, like the other FA core complex components, stably associates with the BLM helicase and Topo IIIα, so that if Flag-FANCM is immunoisolated directly from the extract, the intrinsic helicase activity of BLM is high enough to interfere with our analysis of the FANCM activity. We therefore thoroughly depleted BLM until the Flag-FANCM ATPase mutant protein (K117R) displays no detectable ATPase activity. The Flag-FAAP250 fraction obtained contains low levels of FA core complex, as revealed by presence of other FA core complex components by Western (data not shown). However, the majority flag-FAAP250 did not appear to be assembled into the core complex, possibly due to over-expression of the exogenous protein (which is at least 10-fold higher than that of the endogenous protein; data not shown). The Flag-FANCM proteins obtained were then analyzed for ATPase, helicase, and triple-helix unwinding activities as described4,23,30. About 50 ng of Flag-FANCM was used in the different assays. The final concentration of ATP in the ATPase assay is 1 mM. The reaction was incubated for 30 min at 30 °C. Two DNA duplex substrates (Supplementary Fig. 4) containing 40-nucleotide single stranded DNA tails at either 5′- or 3′- end of the duplex region used the exact sequences as described16.
Total RNA was isolated from lymphoblastoid cell line EUFA867 using a Qiagen RNeasy kit (Qiagen, Venlo, The Netherlands). cDNA was prepared using 10 pmol oligo-dT20 primers and Superscript™ II RT polymerase (Invitrogen, Carlsbad, California, USA). The PCR reactions for amplification of FANCM were performed on cDNA using Platinum Taq polymerase (Invitrogen) and primers that covered the whole cDNA sequence. Genomic DNA (gDNA) was isolated from whole blood and lymphoblastoid cell lines of patients and family members using a Qiagen Blood mini kit (Qiagen). The PCR reactions for amplification of exon 13 of FANCM were performed on 200 ng gDNA using Platinum Taq polymerase (Invitrogen) and primers in intron 12 and 13. The deletion-containing region of the genomic DNA was amplified using Pfx polymerase (Invitrogen) and primers in intron 14 and 16.
PCR products were purified using a SAP/EXO treatment (Amersham Biosciences, Uppsala, Sweden) according to manufactures instructions. Sequencing reactions were prepared using 10 pmol primer and Big Dye terminator cycle sequencing kit (Applied Biosystems, Foster City, CA, USA) and performed with the GeneAmp PCR system 9700 (Applied Biosystems, Foster City, CA, USA). Samples were analysed on ABI 3730 DNA Analyzer (Applied Biosystems, Foster City, CA, USA).
The accession number for human FANCM in this study is DQ140356, which is longer than the predicted KIAA1596 sequence in the database (XP_048128.5). The numbers for FANCM orthologues in other species are: chimpanzee (XP_509928.1), mouse (NP_849243.2), fish (tetraodon nigroviridis; CAG10454.1), rat (XP_234239.2), dog (XP_537429.1); Drosophila (CG7922), yeast (MPH1), and Archaeal Hef proteins (AAB89786, AAB99518, and AAB85892),
We thank the FA patients and families for contributing to this study, N. Sherman for mass spectrometry analysis, S. Brill for providing Mus81, R. Wood for XPF, R. Brosh and M. Kenny for RPA, R.Pattni for genotyping, M. Rooimans for technical assistance, and the National Cell Culture Center for providing cells, D. Schlessinger for critical reading of the manuscript. Financial support was from the FA Research Fund, Eugene, Oregon, U.S.A., Ellison medical Foundation (WW), NIH grant CA112775 (MEH), Daniel Ayling Fanconi Anaemia Trust (CGM) and the intramural program of National Institute of Health, National Institute on Aging (WW). JPW and ALM are supported by Dutch Cancer Society and the Netherlands Organization for Health Research and Development.
Competing Interests Statement The authors declare that they have no competing financial interests.