QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Kennedy, R. D. Articles by D'Andrea, A. D. Search for Related Content PubMed PubMed Citation Articles by Kennedy, R. D. Articles by D'Andrea, A. D. Related Content DNA Replication, Recombination, and Repair Cancer and Disease Models Social Bookmarking                   What's this?

REVIEW

The Fanconi Anemia/BRCA pathway: new faces in the crowd

Department of Radiation Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts 02115, USA

	Abstract

Top Abstract The basic FA pathway Recently identified components... The structure of the... The structure of FA... Activation of the FA... Functions of the FA... Turning off the FA... Cancer risk in heterozygous... Epigenetic disruption of the... The FA pathway and... Unanswered questions Conclusions Acknowledgments References

 
Over the past few years, study of the rare inherited chromosome instability disorder, Fanconi Anemia (FA), has uncovered a novel DNA damage response pathway. Through the cooperation of multiple proteins, this pathway regulates a complicated cellular response to DNA cross-linking agents and other genotoxic stresses. In this article we review recent data identifying new components of the FA pathway that implicate it in several aspects of the DNA damage response, including the direct processing of DNA, translesion synthesis, homologous recombination, and cell cycle regulation. We also discuss new findings that explain how the FA pathway is regulated through the processes of ubiquitination and deubiquitination. We then consider the clinical implications of our current understanding of the FA pathway, particularly in the development and treatment of malignancy in heterozygous carriers of FA mutations or in patients with sporadic cancers. We consider how recent studies of p53-mediated apoptosis and loss of p53 function in models of FA may help explain the clinical features of the disease and finally present a hypothesis to account for the specificity of the FA pathway in the response to DNA cross-links.

[Keywords: BRCA; cancer; chemotherapy; DNA cross-links; DNA helicase; Fanconi Anemia]

FA is a genetically heterogeneous disease and has been traditionally classified using cell fusion experiments that correct sensitivity to DNA cross-linking agents such as MMC. To date at least 12 complementation groups (A, B, C, D1, D2, E, F, G, I, J, L, M) have been identified (Table 1). Complementation groups A, C, and G are the most prevalent accounting for approximately 60%, 15%, and 10% of patients, respectively (Strathdee et al. 1992a; Lo Ten Foe et al. 1996; de Winter et al. 1998). The genes for 11 of the complementation groups have been cloned, and 10 of the proteins have been demonstrated to cooperate in a common biochemical pathway. Disruption of the FA pathway results in the common clinical and cellular phenotype observed in FA.

	The basic FA pathway

Top Abstract The basic FA pathway Recently identified components... The structure of the... The structure of FA... Activation of the FA... Functions of the FA... Turning off the FA... Cancer risk in heterozygous... Epigenetic disruption of the... The FA pathway and... Unanswered questions Conclusions Acknowledgments References

We will now give a brief overview of the entire Fanconi pathway and recently identified components before considering the structures and functions of complex 1 and complex 2 in more detail. Complex 1, the monoubiquitin E3 ligase complex, or core complex, consists of 8 of the 11 identified FA proteins (A, B, C, E, F, G, L, M). The putative catalytic element of this multisubunit complex is the L protein. The L subunit contains a PHD domain, and has homology to other E3 ubiquitin ligases (Meetei et al. 2003a). Complex 1 is required for the monoubiquitination of FANCD2 on Lys 561 during S phase of the normal cell cycle. Complex 1 is also essential for the inducible monoubiquitination of FANCD2 following cellular exposure to DNA damaging agents, such as MMC, ultraviolet (UV), or ionizing radiation (IR) and deoxynucleotide depletion due to hydroxyurea (Garcia-Higuera et al. 2001; Gregory et al. 2003). Following monoubiquitination, FANCD2 relocalizes from a soluble nuclear compartment to the chromatin fraction (Wang et al. 2004) where it colocalizes in nuclear foci with other DNA damage response proteins, including FANCD1/BRCA2 (Hussain et al. 2004; Wang et al. 2004) BRCA1 (Garcia-Higuera et al. 2001), RAD51 (Taniguchi et al. 2002a; Hussain et al. 2004), Proliferating Cell Nuclear Antigen (PCNA) (Hussain et al. 2004), and Nijmegen Breakage Syndrome (NBS1) (Nakanishi et al. 2002). Following relocalization to chromatin, monoubiquinated FANCD2 (FANCD2-Ub) promotes chromatin loading of the FANCD1/BRCA2 protein, forming a second complex (complex 2), where it may facilitate the formation of nuclear foci containing the homologous recombination (HR) DNA repair protein RAD51 (Wang et al. 2004). FA cells that fail to form either complex 1 or complex 2 are hypersensitive to cross-linking agents such as MMC and exhibit the characteristic broken and radial chromosome forms of FA.

	Recently identified components of the FA pathway

Top Abstract The basic FA pathway Recently identified components... The structure of the... The structure of FA... Activation of the FA... Functions of the FA... Turning off the FA... Cancer risk in heterozygous... Epigenetic disruption of the... The FA pathway and... Unanswered questions Conclusions Acknowledgments References

 
Three additional FA pathway components have recently been reported. Traditional approaches, based on complementation studies of cells derived from FA patients, have identified the groups FANCI and FANCJ (Levitus et al. 2004). Experimental approaches involving the knockout of genes in chicken and human cell lines have identified FANCM (Meetei et al. 2005; Mosedale et al. 2005). We will now consider where these new components may function within the context of the basic FA pathway.

FANCI, an unknown component of complex 1?

FANCI cells express the multisubunit FA complex (A/B/C/E/F/G/L/M complex), but fail to monoubiquitinate FANCD2 (Levitus et al. 2004). Accordingly, the FANCI protein may be another loosely associated protein of complex 1, such as an E2 ubiquitin conjugating protein, which is not required for complex assembly but rather is required for E3-mediated monoubiquitination of FANCD2. To date, however, the FANCI gene has not been cloned.

FANCM is the mammalian ortholog of Hef

Very recently, two groups have identified the human ortholog of helicase-associated endonuclease for forked structured DNA (Hef) as a component of FA complex 1. Hef was originally described in archaebacteria as a protein that could process blocked DNA replication forks into double-strand DNA breaks (Komori et al. 2002). Meetei et al. (2005) identified the human ortholog of Hef as a 250-kDa component of complex 1 using coimmunoprecipitation experiments. Small interfering RNA (siRNA)-mediated depletion of this protein in HeLa cells resulted in a failure to monoubiquitinate FANCD2, confirming a functional role in complex 1. Genotyping of an individual with previously unclassified Fanconi Anemia identified a nonsense mutation in the Hef gene on one allele and a partial deletion of the gene on the other. They concluded that mutation in Hef represents a new FA complementation group, FANCM. However, it is important to note that FA-M cells have not yet been complemented with a wild-type FANCM gene construct. This may be due to difficulty in ectopically expressing the 250-kDa protein (Meetei et al. 2005).

In a separate study, Mosedale et al. (2005) used knockout studies to characterize the function of the Hef ortholog in DT40 chicken cells. They noted that Hef-knockout cells are highly sensitive to MMC and demonstrate chromosomal instability. When Hef was knocked out of FANCC disrupted cells, no further sensitivity was noted, indicating that Hef functions within the FA pathway. Consistent with this hypothesis, Hef coimmunoprecipitated with labeled FANCF and FANCC, suggesting that it is a component of complex 1. Furthermore, Hef was required for the relocalization of complex 1 to chromatin following DNA damage. They also noted that Hef was required for DNA damage-induced FANCD2 monoubiquitination, indicating that it has a role in the monoubiquitin ligase activity of complex 1 (Mosedale et al. 2005). Interestingly, the FANCM-deficient DT40 cells were defective in the HR mediated repair of DNA cross-links due to MMC but demonstrated normal HR repair of double-strand breaks (DSBs). This observation suggests some specificity for the FA pathway in the processing of DNA cross-links as opposed to other forms of DNA damage and could explain the specific high sensitivity to cross-linking agents seen in cell lines derived from FA patients.

In bacterial cells, Hef has both endonuclease and helicase domains that function together to unwind and cleave synthetic complexes that mimic stalled replication forks (Komori et al. 2002). Similarly, the N-terminal region of human FANCM is homologous to the DEAH-box ATP-dependent helicase family, whereas the C terminus is homologous to the endonuclease region of the nucleotide excision repair protein ERCC4. In contrast to bacterial Hef, the endonuclease domain of FANCM appears to be degenerate and nonfunctional (Meetei et al. 2005). Indeed, mutation of the endonuclease domain of FANCM in DT40 cells does not markedly affect cross-linker sensitivity (Mosedale et al. 2005). However, FANCM does demonstrate an in vitro DNA-dependent ATPase function with an affinity for single-strand and forked structures that is lost when the helicase domain is mutated (Meetei et al. 2005; Mosedale et al. 2005). Mutation of the helicase domain also results in marked cisplatin sensitivity in DT40 cells. Importantly, no in vitro helicase activity has been demonstrated for FANCM. Rather, FANCM has been shown to function as a DNA translocase in an in vitro triple-helix displacement assay (Meetei et al. 2005). This observation supports a model where FANCM moves the FA complex along DNA, perhaps in association with a replication fork. FANCM may be specifically involved in the detection of a cross-link at a stalled replication fork and may then respond by activating the rest of the FA pathway. From the recent studies it appears that FANCM function may be ATP dependent, and it would be interesting in the future to investigate if mutation of the ATPase region of this protein affects the chromatin association of complex 1 and monoubiquitination of FANCD2 following DNA cross-linker treatment.

FANCJ is the DNA helicase BRIP1

Three groups have recently reported that FANCJ is the BRCA1-Interacting Protein C-terminal helicase BRIP1 (also known as BACH1). Interestingly, in common with BRCA1 and FANCD1/BRCA2, heterozygous mutation of BRIP1 predisposes to early onset breast cancer, further supporting a role for the FA pathway in breast cancer development (Cantor et al. 2001).

Levitus et al. (2005) used positional cloning and a microcell-mediated chromosome complementation approach to study eight cell lines derived from FA-J individuals. They identified eight different mutations affecting the helicase domain of BRIP1, of which five were predicted to result in a protein truncation and the remainder a full-length protein with an altered amino acid residue. Three patients were homozygous for the same mutation whereas five carried a different mutation on each allele. Complementation of FANCJ mutant cells was achieved using part of chromosome 17 but not specifically with expression of a wild-type BRIP1 gene or cDNA construct.

Similarly, Levran et al. (2005) used genetic mapping to initially study 5 patients assigned to complementation group FA-J and then a further 18 unclassified FA patients. They also identified homozygous mutations in the BRIP1 gene in FA-J patients and reported an additional two truncating mutations, an insertion mutation and a missense mutation predicted to disrupt the helicase domain of the BRIP1 protein.

Two other groups also recently identified BRIP1 as FANCJ by using knockout approaches in cell lines. Litman et al. (2005) used siRNA to silence BRIP1 expression in MCF7 human breast cancer cells whereas Bridge et al. (2005) used a BRIP1 knockout approach in chicken DT40 cells. Both studies reported that abrogation of BRIP1 expression resulted in a FA cellular phenotype with profound cross-linker sensitivity and an accumulation in the G2/M phase of the cell cycle following DNA damage. Litman et al. also demonstrated a truncating mutation of BRIP1 in the FA-J AG656 cell line and showed that lentiviral-mediated expression of BRIP1 in this cell line corrected the G2/M cell cycle abnormality. Interestingly these papers reported differing results as regards the role for BRIP1 in HR. Litman et al. found a defect in the HR of a I-Sce1 induced DSB in the BRIP1-deficient MCF7 cells whereas Bridge et al. reported normal repair of the same lesion in the BRIP1-deficient DT40 cells. The reason for these differing results is unclear but may relate to differences between human and chicken cells. Indeed, chicken BRIP1 is structurally different from its human counterpart. It is only 54% homologous to human BRIP1 and does not contain the BRCT-binding consensus sequence (Bridge et al. 2005). This sequence appears to be required for the response to DNA damage in mammalian cells by mediating an association with BRCA1 (Yu et al. 2003).

Although the function of BRIP1 is unclear, it seems likely that it has a role in the processing of intermediate DNA structures during repair. BRIP1, in common with Bloom helicase (BLM), is a member of the recQ DEAH helicase family. However, BLM in contrast to BRIP1 has been reported to suppress sister chromatid exchange (SCE) in HR (Onclercq-Delic et al. 2003). BRIP1 demonstrates an in vitro 5'-3' DNA unwinding activity of forked DNA duplexes whereas BLM demonstrates a 3'-5' helicase activity (Karow et al. 1997; Gupta et al. 2005). BRIP1, in common with BLM, also associates with DNA substrates mimicking Holliday Junctions (Litman et al. 2005) although, in contrast to BLM, no unwinding of these structures has been demonstrated (Gupta et al. 2005). Therefore, it has been hypothesized that BRIP1 may promote HR by blocking the access of BLM to intermediate DNA structures or by providing a scaffold for other HR proteins (Litman et al. 2005). Although the role for BACH-associated helicase activity remains unclear, its importance is demonstrated by the observation that disease-related mutations all disrupt the helicase domain (Levitus et al. 2005; Levran et al. 2005). Moreover, specific mutation of the helicase domain of BRIP1 in chicken DT40 cells results in hypersensitivity to cisplatin (Bridge et al. 2005).

Another question that both these papers raise is the role of BRCA1 in the FA pathway. BRIP1 is phosphorylated and associates with the BRCA1 C Terminus (BRCT) motifs of BRCA1 during S phase of the cell cycle (Rodriguez et al. 2003; Yu et al. 2003). The association between BRCA1 and BRIP1 has been reported to be essential for DNA repair and cell cycle checkpoint functions following radiation-mediated DNA damage (Cantor et al. 2001; Yu et al. 2003). However, Litman et al. (2005) reported that BRIP1 was not required for RAD51 foci formation after hydroxyurea treatment whereas BRCA1 was essential, indicating that these proteins may have some distinct functions. Bridge et al. (2005) also reported that BRIP1 functioned independently of BRCA1 in the DT40 cell line. One possible explanation for these discrepancies is that the role of BRIP1 in the response to replication fork arrest is distinct from that in radiation damage repair and independent of BRCA1. Indeed, BRIP1 knockout DT40 cells are not very sensitive to radiation damage but are hypersensitive to cross-linking agents (Bridge et al. 2005). It is also important to note that the chicken ortholog of BRCA1 may not function identically to the human counterpart. Chicken BRCA1 protein is only 33% homologous to human BRCA1 (Orelli et al. 2001). Abrogation of chicken BRCA1 does not result in hypersensitivity to cisplatin or to radiation damage (Bridge et al. 2005) unlike mammalian BRCA1 in which loss of function has been reported to result in marked sensitivity to both forms of DNA damage (Xia and Powell 2002; Quinn et al. 2003; Kennedy et al. 2004). Moreover, the observations that functional BRCA1 is required for DNA damage-induced FANCD2 foci (Garcia-Higuera et al. 2001) and that loss of BRCA1 function results in marked cross-linker sensitivity, chromosomal breakage, and a defect in HR (Moynahan et al. 2001a) further support a model in which BRCA1, in common with BRIP1, has a role in the FA pathway.

At present the role of BRIP1 in the FA pathway is uncertain. FA-J cells, in common with FA-D1 (BRCA2 mutant) cells, express an intact FA complex 1 and exhibit normal to elevated levels of FANCD2-Ub (Levitus et al. 2004; Litman et al. 2005). These observations suggest that BRIP1 functions independently or downstream of FANCD2, perhaps as another subunit of the FANCD2-Ub/BRCA2 chromatin complex (complex 2 in Fig. 1). It is clear that future studies are required to elucidate the role of BRIP1 in the HR of DNA cross-linking lesions and answer if there is a requirement for the interaction with BRCA1. It is also important to note that no study to date has reported the correction of chromosomal breakage in FA-J cells through the exogenous expression of wild-type BRIP1. The generation of isogenic mutant and fully corrected cell lines would be expected to help answer many of the questions about the role of BRIP1 in the FA pathway.

	The structure of the FA complex 1

Top Abstract The basic FA pathway Recently identified components... The structure of the... The structure of FA... Activation of the FA... Functions of the FA... Turning off the FA... Cancer risk in heterozygous... Epigenetic disruption of the... The FA pathway and... Unanswered questions Conclusions Acknowledgments References

 
At least eight of the FA proteins (subunits A, B, C, E, F, G, L, M) interact in a large nuclear complex. This complex is shown schematically in Figure 2. While no crystal structures are available for subunits of the complex, several interactions are evident from yeast two-hybrid and coimmunoprecipitation studies.

View larger version (36K): [in this window] [in a new window]   Figure 1. Schematic interaction of the FA pathway. Protein complex 1 contains the A, B, C, E, F, G, L, M, and possibly I subunits. Protein complex 2 contains FANCD2-Ub, BRCA2, and possibly FANCJ. During S phase, when a replication fork encounters a DNA cross-link, the FA complex (complex 1) is activated. This activation leads to the monoubiquitinaiton of FANCD2, which is then targeted to chromatin containing the cross-link. FANCD2-Ub then interacts with BRCA2 in complex 2, leading to repair of the cross-link possibly through HR and TLS. FANCD2 is deubiquitinated by USP1, thereby inactivating the pathway.

View larger version (41K): [in this window] [in a new window]   Figure 2. Schematic structure of the Fanconi Anemia E3 ligase complex (complex 1). The core complex contains subunits A, B, C, E, F, G, L, M, and possibly I. According to this model, ubiquitin is transferred from an E1 to an E2 (unknown for the FA complex) to FANCD2. The FANCD2 substrate may have direct binding interactions with FANCL and FANCE, and these interactions may be mediated, at least in part, by ATR-dependent phosphorylations on the FANCD2 protein. Importantly, recent studies indicate that complex 1 is associated with chromatin and may function as a sensor of DNA damage as well as an E3 ligase.

Coimmunoprecipitation experiments support the presence of subcomplexes (for example, the FANCA/FANCG and FANCE/FANCF/FANCG/FANCC subcomplexes), which may preform and translocate intracellularly before the assembly of the complete complex. Specific domains of these subunits are likely to play a role in subunit assembly, such as the tetratricopeptide repeat (TPR) domains of the FANCG subunit (de Winter et al. 1998) and the WD40 domains of the FANCL subunit (Meetei et al. 2003a). Phosphorylation of some of the FA subunits also appears to contribute to complex 1 assembly (Kupfer et al. 1999).

The FANCL protein has a C-terminal RING-finger-like plant domain (PHD), believed to be directly involved in the catalytic activity of the complex. The PHD domain may interact directly with an E2 protein of the FA pathway. Interestingly, a truncated form of FANCL, lacking the C-terminal PHD domain, still binds and stabilizes the FA complex, through its three WD40 domains, but does not complement the FANCD2 monoubiquitination defect in FANCL cells (A. Gurtan and A.D. D'Andrea, unpubl.). WD40 domains of FANCL may be required for its interaction with phosphorylated residues on other FA complex subunits or perhaps with phosphorylated residues on the FANCD2 substrate. It has previously been reported that the WD40 domains of other E3 ligases are important in recruiting specific substrates to the ligase complex for ubiquitination (Orlicky et al. 2003).

While the FA complex is the putative E3 ligase for FANCD2, the in vitro reconstitution of this ligase reaction has not been successful to date. In fact, the only E3 ligase that has been demonstrated to monoubiquitinate FANCD2 in vitro is the BRCA1/BARD1 complex (Vandenberg et al. 2003). Additional subunits of the FA complex, such as the putative FANCI protein, may be required for reconstitution of the FANCL E3 ligase activity in vitro. However, the FANCL subunit does undergo autoubiquitination in vitro, and its PHD domain is required for this activity (Meetei et al. 2003a).

As will be discussed later, Bloom DNA helicase and topoisomerase III have also been reported to associate with complex 1, which, along with recent identification of FANCM as a DNA helicase, suggests that this complex has a direct role in DNA processing as well as having a monoubiquitin E3 ligase activity.

In conclusion, complex 1 has an E3 ubiquitin ligase activity directed toward FANCD2 and a DNA processing activity of uncertain purpose. Important questions remain about complex 1. First, is complex 1 associated with the advancing replication fork and does it bind to helix-distorting lesions, thereby acting as a sensor of damage? Second, do other specific subunits of complex 1, in addition to FANCM, bind directly to DNA? Third, while direct interactions among the subunits of complex have been defined by coimmunoprecipitation and yeast two-hybrid experiments, what are the precise crystallographic interactions?

	The structure of FA complex 2

Top Abstract The basic FA pathway Recently identified components... The structure of the... The structure of FA... Activation of the FA... Functions of the FA... Turning off the FA... Cancer risk in heterozygous... Epigenetic disruption of the... The FA pathway and... Unanswered questions Conclusions Acknowledgments References

 
The downstream function of the FA pathway is largely unknown. Accumulating evidence suggests a major function of complex 1 is to generate monoubiquitinated FANCD2 (FANCD2-Ub), required for the assembly of the FANCD2-Ub/BRCA2 chromatin complex (complex 2). Following monoubiquitination, FANCD2 relocalizes to DNA damage-induced S-phase foci where it associates with the DNA repair proteins BRCA2/FANCD1 (Wang et al. 2004), BRCA1 (Garcia-Higuera et al. 2001), NBS1 (Nakanishi et al. 2002), RAD51 (Hussain et al. 2004), and PCNA (Howlett et al. 2005). The formation of these DNA damage-induced FANCD2 foci is dependent on the presence of functional BRCA1 (Garcia-Higuera et al. 2001), thereby linking this DNA repair protein to the FA pathway. As discussed earlier, BRIP1 (FANCJ) mutant cells demonstrate normal FANCD2 monoubiquitination, indicating that it does not form part of complex 1. Furthermore, BRIP1, in common with FANCD2, localizes to DNA damage-induced nuclear foci along with BRCA1 (Cantor et al. 2001), suggesting that it may function as part of complex 2. A diagrammatic representation of FA complex 2 is shown in Figure 1.

Perhaps the simplest model proposes that FA complex 2 contributes to the stabilization of a stalled DNA replication fork. Indeed BRCA2 (FANCD1) has been demonstrated to be required for the stabilization of the stalled replication fork following treatment with hydroxyurea (Lomonosov et al. 2003). By stabilizing the fork, the complex may facilitate DNA repair, and may be required for the recruitment of repair proteins. Accordingly, it will be interesting to determine whether disruptions of FANCD2 or BRIP1/FANCJ also result in destabilization of the replication fork.

In conclusion, complex 2 may contain a helicase activity (BRIP1) and appears to play a more downstream role in DNA repair. Important data about this complex still remain to be elucidated. First, as will be discussed further later, current data indicate that the FA pathway interacts with the process of HR repair and translesion synthesis (TLS) (Niedzwiedz et al. 2004). How complex 2 engages in these two processes remains unclear. Second, whether complex 2 has a direct function in cell cycle checkpoint activation remains unknown. Third, the molecular interactions among the protein subunits of complex 2 remain to be elucidated. Finally, it is not clear whether BRCA1 and its associated E3 ubiquitin ligase activity also play a role in the function of this complex.

	Activation of the FA pathway

Top Abstract The basic FA pathway Recently identified components... The structure of the... The structure of FA... Activation of the FA... Functions of the FA... Turning off the FA... Cancer risk in heterozygous... Epigenetic disruption of the... The FA pathway and... Unanswered questions Conclusions Acknowledgments References

 
Now that we have described the components of the basic FA pathway, we will address how the pathway is activated following DNA damage. Several findings indicate that the Fanconi pathway is activated and functions primarily in the S phase of the cell cycle. First, cell cycle synchronization studies reveal that FANCD2 monoubiquitination occurs predominantly in S phase. FANCD2 is monoubiquitinated at the G1/S boundary, remains monoubiquitinated throughout S phase, and becomes deubiquitinated at the end of S phase as the synchronized cell population enters G2 (Taniguchi et al. 2002a). Second, during S phase, monoubiquitinated FANCD2 protein specifically colocalizes with replication foci containing BRCA1, RAD51 (Taniguchi et al. 2002a; Hussain et al. 2004), and PCNA proteins (Howlett et al. 2005). These foci are believed to represent sites of homologous repair. Third, the FA complex 1 associates with chromatin during S phase, and is released in mitosis (Mi and Kupfer 2005).

Recent studies indicate that the FA pathway is activated during S phase in response to DNA replication fork arrest. Replication fork arrest leads to the local generation of single-strand DNA, perhaps through helicase unwinding of the DNA duplex. Single-strand DNA binds RPA (Replication Protein A) and then ATR (Ataxia Telangiectasia and RAD3-related protein) and ATRIP (ATR-Interacting Protein). After recruitment to the stalled fork, ATR phosphorylates multiple substrates, leading to additional cell cycle checkpoint and DNA repair responses (Zou and Elledge 2003). One principal substrate of ATR is CHK1 (Checkpoint Kinase 1). Recent studies implicate ATR/CHK1 in the activation of the FA pathway. First, agents that activate ATR kinase, such as hydroxyurea, cross-linkers, or ultraviolet light are potent activators of FANCD2 monoubiquitination (Gregory et al. 2003). Second, cells from patients with Seckel syndrome express a mutant ATR and fail to monoubiquitinate FANCD2 following DNA damage. Similarly, siRNA-mediated silencing of ATR expression abrogates FANCD2 monoubiquitination (Andreassen et al. 2004). Third, knockdown of ATR also results in a chromosome breakage phenotype and cellular hypersensitivity to MMC, consistent with an upstream role for ATR in the FA pathway (Andreassen et al. 2004). Moreover, disruption of the CHK1 kinase either by siRNA knockdown or by specific inhibitors of CHK1 kinase leads to DNA cross-linker sensitivity and a disruption of the FA pathway (X. Wang and A.D. D'Andrea, unpubl.). Fourth, the recently identified FANCM protein is hyperphosphorylated following DNA damage. This protein has multiple ATR consensus sites, although the ATR-mediated phosphorylation of these sites have yet to be investigated (Meetei et al. 2005).

Taken together, these results support a model in which ATR and CHK1, in response to DNA replication fork stalling, modulate the FA pathway. This may occur through the direct phosphorylation of one or more subunits of complex 1. For example, phosphorylation of FANCM may promote the DNA translocation of complex 1 and facilitate the identification of DNA crosslinks, leading to subsequent activation of the rest of the FA pathway, which in turn may be involved in the processing of the lesion that has given rise to the failure in DNA replication.

	Functions of the FA pathway

Top Abstract The basic FA pathway Recently identified components... The structure of the... The structure of FA... Activation of the FA... Functions of the FA... Turning off the FA... Cancer risk in heterozygous... Epigenetic disruption of the... The FA pathway and... Unanswered questions Conclusions Acknowledgments References

 
Despite extensive research into this area, the specific functions of the FA pathway following activation by DNA replication fork arrest are uncertain. However, recent studies suggest that it may function to coordinate several aspects of DNA damage response including (1) direct enzymatic DNA processing, (2) recruitment of DNA repair proteins involved in HR and TLS, and (3) cell cycle checkpoint activation (Fig. 3).

View larger version (53K): [in this window] [in a new window]   Figure 3. A schematic diagram depicting how the FA pathway may function to coordinate the cellular response to replication fork arrest following DNA damage. ATR is activated in response to replication fork arrest and in turn may activate the FA pathway, which detects the DNA lesion and coordinates its repair.

As discussed earlier, the most recently identified FA proteins (J and M) have DNA helicase domains (Levitus et al. 2005; Meetei et al. 2005). BLM and topoisomerase III have also been found to associate with FA complex 1 (Meetei et al. 2003b). BLM, topoisomerase III, and replication protein A associate with constituents of the FA complex 1 (the BRAFT complex) (Meetei et al. 2003b). These proteins have been shown to cooperate in DNA unwinding and have been implicated in the response to replication fork arrest (Chan et al. 1987; Brosh Jr. et al. 2000; Wu and Hickson 2002). Mutation of BLM in common with other FA genes results in chromosomal breakage (Risinger and Groden 2004), although it is important to note that BLM-deficient cells also exhibit increased SCE, a phenotype not observed in FA cells.

FA complex 2 colocalizes with NBS1, a component of the MRE11/RAD50/NBS1 (MRN) complex that has 3'-to-5' exonuclease activity and is required for DNA processing prior to repair through HR (Nakanishi et al. 2002; Kobayashi et al. 2004). Loss of function of the FA pathway has been associated with a failure of the MRN complex to form at sites of DNA damage (Pichierri et al. 2002).

Overall, these observations linking the FA pathway to DNA processing suggest that the FA proteins not only function in DNA damage signaling pathway but may also have active roles in the detection and processing of stalled replication forks. Consequently, loss of the FA pathway may lead to chromosomal instability through the failure to process DNA repair intermediates.

Recruitment of DNA repair proteins

Homologous recombination (HR) HR is a process by which double-strand DNA breaks are repaired through the alignment of homologous sequences of DNA and occurs primarily during the late S to M phase of the cell cycle. This repair pathway requires the 3'-5' exonuclease activity of the RAD50, MRE11, and NBS1 complex to expose the 3' ends on either side of the DSB, a process that may also require BRCA1 (Zhong et al. 1999). The break can then be repaired by one of two main pathways. Either single-strand annealing, a potentially mutagenic process that uses regions of homology between the complementary strands to align the broken ends of DNA, or strand invasion, a RAD51- and BRCA2-dependent process that typically uses either a homologous sequence in adjacent DNA or the sister chromatid as a template to repair the break. This latter process may result in a SCE of genetic material (Wyman et al. 2004).

During S phase, DNA cross-links result in a replication fork arrest. In the process of repair, the cross-links are excised by an endonuclease leading to the generation of DSB intermediates (Dronkert and Kanaar 2001; Rothfuss and Grompe 2004). The DSB can then be repaired by HR. Following treatment with cross-linkers, DSB formation has been reported to precede FANCD2 monoubiquitination. FANCD2-Ub then colocalizes with Histone 2AX S-phase foci, a marker of DSBs, indicating a role for the FA pathway in the repair of this type of DNA damage. Accordingly, disruption of the FA pathway results in persistence of DNA breakage (Rothfuss and Grompe 2004). It is interesting to note that although XPF/ERCC1 has been reported to colocalize with FANCA, suggesting it may function in the FA pathway (Sridharan et al. 2003), XPF/ERCC1 mutant cells demonstrate normal FANCD2 monoubiquitination (R.D. Kennedy and A.D. D'Andrea, unpubl.). This implies that another endonuclease such as the Mus81-Eme1 complex (Abraham et al. 2003) is, at least partly, involved in the excision of DNA cross-links upstream in the FA pathway.

Monoubiquitinated FANCD2 associates with BRCA1 (Garcia-Higuera et al. 2001), BRCA2 (Hussain et al. 2004; Wang et al. 2004), RAD51 (Hussain et al. 2004), PCNA (Howlett et al. 2005), and NBS1 (Nakanishi et al. 2002) and is required for the relocalization of BRCA2 to chromatin and the subsequent formation of RAD51 foci (Hussain et al. 2004; Wang et al. 2004). Although the purpose of FANCD2 monoubiquitination remains unknown, it seems likely that it has a role in HR as BRCA1, BRCA2, RAD51, and NBS1 have all been implicated in this process (Moynahan et al. 1999; Tutt et al. 2001; Kobayashi et al. 2004). In support of this hypothesis, cell lines deficient in the FA pathway have been reported to have defective HR (Donahue et al. 2003; Yamamoto et al. 2003; Niedzwiedz et al. 2004; Nakanishi et al. 2005). In general, disruption of upstream FA proteins, such as FANCA (Nakanishi et al. 2005) and FANCG (Yamamoto et al. 2003) result in milder HR defects than disruption of BRCA2/FANCD1 (Moynahan et al. 2001b), suggesting that the FA chromatin-associated complex 2 is largely responsible for the DNA repair functions of the FA pathway. Interestingly, recent evidence suggests that DNA cross-links may be preferentially repaired by the single-strand annealing subset of HR (Jonnalagadda et al. 2005). Murine cells deficient in components of complex 1 or FANCD2 demonstrate abnormal single-strand annealing following cross-linker damage, suggesting a specific role for the FA pathway in this subset of HR (Nakanishi et al. 2005; Yang et al. 2005).

Translesion synthesis (TLS) Another proposed role for the FA pathway is TLS. This process can result in point mutations due to the misincorporation of nucleotides. TLS may offer an alternative to HR in order to re-establish replication forks or may work in conjunction with HR in the resolution of certain types of DNA damage (Niedzwiedz et al. 2004). The best characterized DNA damage resulting in TLS is the ultraviolet-induced cyclobutane pyrimidine dimer (CPD), although TLS is also likely to bypass DNA adducts such as those produced by cross-linking agents (Yamashita et al. 2002). CPD damage is either repaired by nucleotide excision repair (NER) or bypassed by TLS (Friedberg et al. 2002; Prakash and Prakash 2002). At the beginning of the process of TLS, a normal replicative DNA polymerase, such as Pol , encounters a CPD and is displaced by a TLS polymerase (Friedberg et al. 2005). The TLS polymerase can replicate past CPD damage with higher fidelity and efficiency than a replicative polymerase (Yu et al. 2001). Evidence for the role of the FA pathway in TLS comes from experiments in the DT40 chicken cell line where loss of the TLS proteins Rev1 and Rev3 along with FANCC are epistatic as regards DNA cross-linker sensitivity. Rev1 and monoubiquitinated FANCD2 are also observed to colocalize in nuclear foci following replication fork arrest (Niedzwiedz et al. 2004). Loss of the FA pathway results in a hypomutable phenotype following cross-linker treatment, further suggesting a relationship between the FA and TLS pathway (Papadopoulo et al. 1990a,b). Furthermore, Rev1 deficient DT40 cells, in common with FA cells, demonstrate hypersensitivity to cisplatin (Ross et al. 2005). Finally, as will be discussed later, recent evidence suggests that the FA pathway and TLS are terminated by the same deubiquitinating enzyme, USP1, further indicating their coordinate regulation (T. Huang and A.D. D'Andrea, unpubl.).

One hypothesis, which may link the HR and TLS functions of the FA pathway, is the requirement for both processes in the repair of DNA cross-links. These lesions are unusual in that they require the excision of adducts on both strands of DNA and the subsequent repair of the resultant DSB in S phase (Rothfuss and Grompe 2004). The FA pathway may function to coordinate the excision and bypass of DNA adducts with the HR of DSBs. Consistent with this model, components of the FA pathway, HR, and TLS have all been shown to be required for the repair of cross-links in chicken DT40 cells (Niedzwiedz et al. 2004). Figure 4 demonstrates a simplified version of the model recently proposed by Niedzwiedz et al.

Niedzwiedz et al. (2004) also reported some interesting data regarding the role of the FA pathway and TLS in the response to endogenous DNA damage. Unlike cross-link repair, in which TLS and HR appear to cooperate, the FA pathway promotes either HR or TLS independently in the repair of abasic sites (Niedzwiedz et al. 2004). Therefore, loss of the FA pathway may be predicted to result in a failure to repair endogenous DNA damage efficiently, thereby promoting chromosomal instability.

In contrast to the Rev1 and Rev3 knockout studies, Rad18, another TLS protein, has been reported to function independently of FANCC in DT40 cells. The Rad6/Rad18 complex mediates PCNA monoubiquitination, thereby facilitating recruitment of a TLS polymerase. The combination of Rad18 and FANCC knockout results in greater cisplatin sensitivity than either knocked out independently (Hirano et al. 2005). These data suggest that FA pathway-mediated TLS may be Rad18 independent. Indeed, Rad18-independent TLS has been reported in vertebrate cells (Okada et al. 2002).

View larger version (36K): [in this window] [in a new window]   Figure 4. A simplified model of FA-mediated cross-link repair that involves NER, TLS, and HR. (1) The replication fork arrests at a DNA cross-link and is detected by FA complex 1. FA complex stabilizes the arrested fork. (2) NER excises one side of the cross-link leaving a single-strand DNA break on one strand and a monoadduct on the opposite strand. (3) FA recruits TLS to bypass the remaining monoadduct. (4) NER excises the monoadduct. (5) FA-mediated HR of double-strand DNA break. (6) Intermediate DNA crossover structures are resolved and the replication fork is re-established, a process that may require FA pathway-associated DNA helicases.

FA cells characteristically demonstrate an accumulation of 4N DNA content either spontaneously or following treatment with DNA cross-linkers (Kubbies et al. 1985). Indeed, this consistent observation has been proposed as a diagnostic test for FA (Miglierina et al. 1991). One hypothesis to explain this phenotype has been that FA cells fail to arrest in S phase due to DNA damage and accumulate at the G2/M damage checkpoint (Centurion et al. 2000; Sala-Trepat et al. 2000). Other investigators have proposed that FA cells arrest in late S phase due to a failure to remove DNA cross-links (Akkari et al. 2001). At present, the function of the FA pathway in cell cycle regulation remains uncertain, largely due to the different methodologies and cell line models used to study this question. Published evidence suggests that the FA pathway may have roles in the DNA replication, intra-S-phase, and G2/M DNA damage checkpoints.

The DNA replication checkpoint is activated during S phase in response to stalled DNA replication and primarily involves the ATR-mediated phosphorylation of DNA damage response kinases such as CHK1. It functions to allow the stabilization and repair of the stalled replication fork or the restoration of a depleted deoxynucleotide pool. It also functions to stop the firing of further DNA replication origins, preventing the formation of new replication forks until the existing damage has been repaired (Bartek et al. 2004). A role for the FA pathway in the replication checkpoint is supported by the observations that FANCD2 monoubiquitination is dependent on ATR activation and that constituents of complex 1 (the E3 ligase complex) and FANCD2 have both been shown to be required for S-phase cell cycle arrest following cross-link damage (Centurion et al. 2000; Sala-Trepat et al. 2000; Pichierri and Rosselli 2004).

In addition to the replication checkpoint, there is evidence that the FA pathway may function in a second S-phase checkpoint, the intra-S-phase checkpoint. This checkpoint is predominantly activated by the ataxia-telangiectasia-mutated (ATM) kinase and the ATM substrate CHK2 in response to double-strand DNA breakage and is not dependent on replication fork arrest (Bartek et al. 2004). Cells derived from FA patients replicate DNA in the presence of radiation damage, a feature of a deficient intra-S-phase checkpoint. ATM-mediated phosphorylation of FANCD2 at Ser 222, a process that also depends on NBS1, has been shown to be required for this checkpoint.

Indeed, recent studies suggest a direct interaction between the C terminus of FANCD2 and ATM (S. Margossian and A.D. D'Andrea, unpubl.). Interestingly, activation of the intra-S-checkpoint is independent of the monoubiquitination of FANCD2 (Taniguchi et al. 2002b).

Recent data also suggest that the FA pathway has a role in maintaining the G2/M boundary DNA damage checkpoint, which exists to prevent progession into mitosis prior to the repair of abnormally replicated DNA. This checkpoint is activated by either ATM or ATR and their downstream kinase targets depending on the type of DNA damage (Liu et al. 2000; Shiloh 2001; Helt et al. 2005). Murine cells deficient in FANCC or FANCD2 undergo an abnormal progression into mitosis following DNA damage when compared with wild-type cells (Freie et al. 2004). Interestingly, loss of BRIP1 (FANCJ) function also results in a defect in the G2/M checkpoint, suggesting that this protein may function in a common pathway with the other FA proteins in G2/M checkpoint regulation (Yu et al. 2003).

	Turning off the FA pathway: the FANCD2 Dub

Top Abstract The basic FA pathway Recently identified components... The structure of the... The structure of FA... Activation of the FA... Functions of the FA... Turning off the FA... Cancer risk in heterozygous... Epigenetic disruption of the... The FA pathway and... Unanswered questions Conclusions Acknowledgments References

 
Inappropriate activation of the FA pathway in undamaged cells may lead to harm by promoting mutagenic TLS. Consequently cells may have developed mechanisms of shutting off the pathway when it is not needed. The Deubiquitinating enzyme (Dub), USP1, has recently been identified as a critical enzyme that regulates the FA pathway. Dub enzymes are cysteine proteases that cleave peptide and isopeptide linkage immediately after the glycine-glycine sequence (i.e., the C terminus of ubiquitin) (D'Andrea and Pellman 1998; Wilkinson 2000; Amerik and Hochstrasser 2004).

The role of USP1 in the FA pathway was originally identified using a retroviral small hairpin RNA (shRNA) library (Nijman et al. 2005). Pools of shRNA molecules were used to abrogate the expression of specific Dub enzymes. ShRNA depletion of USP1 resulted in increased FANCD2-Ub levels, without promoting DNA damage. Moreover, USP1 activity peaks at the end of S phase, perhaps accounting for the observed deubiquitination of FANCD2 in late S phase.

Interestingly, USP1 has also recently been identified as the Dub for PCNA, and therefore has a role in down-regulating TLS (T. Huang and A.D. D'Andrea, unpubl.). Loss of USP1 function results in elevation and prolongation of PCNA ubiquitination. This provides further evidence for the intertwined nature of the FA pathway and TLS. It is important to note that, unlike all the other FA proteins described in Figure 1, mutation of the USP1 protein has not yet been directly linked to the pathogenesis of FA or other human disease.

	Cancer risk in heterozygous carriers of FA gene mutation

Top Abstract The basic FA pathway Recently identified components... The structure of the... The structure of FA... Activation of the FA... Functions of the FA... Turning off the FA... Cancer risk in heterozygous... Epigenetic disruption of the... The FA pathway and... Unanswered questions Conclusions Acknowledgments References

 
Increased cancer risk is a well-recognized clinical feature of FA and has been reviewed extensively in other articles (Alter 2003; Kutler et al. 2003b). However, more recent work has shown that although heterozygous carriers of FA gene mutations do not demonstrate developmental or bone marrow abnormalities, there is a predisposition to cancer in these individuals. In particular, studies have reported a strong association between heterozygous mutation of FA genes and susceptibility to breast or ovarian cancer. Carriers of mutations in BRCA1 or BRCA2/FANCD1 have an 82% lifetime risk of breast cancer and a 54% and 23% risk of ovarian cancer, respectively (King et al. 2003). Mutations in FANCJ (BRIP1) have also been identified in patients with early onset breast cancer (Cantor et al. 2004) although the lifetime risk is unknown. This observation suggests that the FA pathway is important in the prevention of these female cancers and that unidentified mutations of other FA genes may account for some familial breast/ovarian cancer pedigrees not accounted for by BRCA1 or BRCA2/FANCD1.

Why mutations in these components of the FA pathway specifically lead to breast and ovarian cancers is unknown. One possibility is that estrogen in some way promotes the survival of breast and ovarian cells that have undergone critical DNA damage, whereas apoptosis occurs normally in other somatic cells (Elledge and Amon 2002). Consistent with this hypothesis, oophorectomy, which diminishes serum estrogen levels, protects against breast cancer in BRCA1 and BRCA2/FANCD1 mutation carriers (Rebbeck et al. 2002).

Paradoxically, FA patients rarely develop breast or ovarian cancer. This may be because these patients are often hypogonadal. Accordingly, they have decreased estrogen levels, perhaps resulting in a decreased incidence of these hormonally related tumors. Also, in families carrying a BRCA1 or BRCA2/FANCD1 mutation, the peak age of onset of breast or ovarian cancer is in the fourth decade (Ford et al. 1994), whereas FA patients often die from complications of aplastic anemia or other cancers before this age.

In addition to breast and ovarian cancer, heterozygosity for a mutation in BRCA2/FANCD1 predisposes to pancreatic cancer. In one study 19% of families with a history of hereditary pancreatic cancer had either a frameshift mutation or an unclassified variant of BRCA2/FANCD1 (Hahn et al. 2003). Other FA gene mutations have also been identified in pancreatic cancer. One recent study determined that two out of 400 patients with pancreatic cancer carried a heterozygous germline mutation in FANCC (Couch et al. 2005). Interestingly, the pancreatic tumors from these patients exhibited loss of heterozygosity (LOH) at the FANCC locus, indicating that loss of FANCC function resulted from a mutation of one allele and deletion of the other, consistent with Knudson's two-hit hypothesis for tumor suppressor genes (Knudson Jr. 1971).

Other studies have identified human pancreatic tumor lines with biallelic loss of FANCC and FANCG (van der Heijden et al. 2004). These tumor lines were derived from individuals who were heterozygote at these loci. Taken together, these data indicate that heterozygote carriers, carrying one mutant FANCC or FANCG allele, may have increased pancreatic cancer risk, albeit at lower penetrance than in the case for a BRCA2/FANCD1 mutation.

Other cancers that have been reported to be associated with heterozygous mutation of a FA gene include BRCA2/FANCD1 related prostate cancer, gastric cancer, and melanoma (Liede et al. 2004), FANCA related acute myeloid leukemia (Condie et al. 2002; Tischkowitz et al. 2004), and FANCC related acute lymphoblastic leukemia (Rischewski et al. 2000).

	Epigenetic disruption of the FA pathway

Top Abstract The basic FA pathway Recently identified components... The structure of the... The structure of FA... Activation of the FA... Functions of the FA... Turning off the FA... Cancer risk in heterozygous... Epigenetic disruption of the... The FA pathway and... Unanswered questions Conclusions Acknowledgments References

 
The human ovarian cancer cell line, 2008, fails to monoubiquitinate FANCD2 following DNA damage. This results from methylation of the FANCF promoter and a consequent loss of FANCF protein expression. Four out of 19 primary ovarian adenocarcinomas were also demonstrated to have FANCF methylation. Interestingly, prolonged treatment of the 2008 cell line with cisplatin resulted in restoration of FANCF expression, suggesting that correction of the FA pathway may account for acquired resistance to this treatment in a clinical setting (Taniguchi et al. 2003). Recent studies have also identified methylation of the FANCF promoter in 24% of ovarian granulosa cell tumors (Dhillon et al. 2004), 30% of cervical cancer (Narayan et al. 2004), 14% of squamous cell head and neck cancers, and 15% of non-small-cell lung cancers, where in the latter it correlated with a worse prognosis (Marsit et al. 2004