BRCA2: a Universal Recombinase Regulator

T Thorslund and SC West

London Research Institute, Cancer Research UK, Clare Hall Laboratories, South Mimms, Herts, UK

Homologous recombination has a dual role in eukaryotic In yeast, it has been shown that a single un-repaired organisms. Firstly, it is responsible for the creation of double-strand break (DSB) in DNA can lead to cell death, genetic variability during meiosis by directing the forma- indicating that DSBs pose one of the most dangerous tion of reciprocal crossovers that result in random forms of DNA lesions (Bennett et al., 1993). As a combinations of alleles and traits. Secondly, in mitotic consequence, eukaryotic cells contain several DNA repair cells, it maintains the integrity of the genome by promoting pathways that can specifically recognize DSBs and initiate the faithful repair of DNA double-strand breaks (DSBs). the efficient repair of the DNA backbone (Figure 1). There In vertebrates, it therefore plays a key role in tumour are two primary mechanisms of DSB repair in vertebrates, avoidance. Mutations in the tumour suppressor protein known as non-homologous end joining and homologous BRCA2 are associated with predisposition to breast and recombination (HR). Non-homologous end joining occurs ovarian cancers, and loss of BRCA2 function leads to by a relatively simple end-splicing mechanism, but is genetic instability. BRCA2 protein interacts directly with potentially error-prone and can lead to the loss of genetic the RAD51 recombinase and regulates recombination- information (Hoeijmakers, 2001). It is primarily utilized mediated DSB repair, accounting for the high levels of during the G1 phase of the cell cycle. In contrast, DSB spontaneous chromosomal aberrations seen in BRCA2- repair by HR is generally error free and functional during defective cells. Recent observations indicate that BRCA2 the S and G2 phases. The difference in cell cycle timing also plays a critical role in meiotic recombination, this time reflects the role that HR plays in the repair of collapsed through direct interactions with the meiosis-specific recom- replication forks, whereas both HR and non-homologous binase DMC1. The interactions of BRCA2 with RAD51 end joining can promote the repair of DSBs caused by and DMC1 lead us to suggest that the BRCA2 tumour ionizing radiation (IR) or genotoxic stress. HR therefore suppressor is a universal regulator of recombinase actions. plays a critical role in the maintenance of genome integrity Oncogene (2007) 26, 7720–7730; doi:10.1038/sj.onc.1210870 in replicating cells. Two simplified mechanisms for HR are shown in Keywords: tumour suppressor; RAD51; DMC1; DNA Figure 1. The process is initiated by the resection of the repair; genome instability; meiotic recombination DSB termini to expose protruding 30-single-stranded DNA (ssDNA) tails, to which the essential recombinase RAD51 will bind and formactive nucleoprotein filaments (Figure 1c). RAD51 loading is facilitated by the ssDNA-binding protein RPA and by RAD52. The The dual functions of recombination RAD51 nucleoprotein filament mediates a search for homology within the sister chromatid, which in turn The structural integrity of DNA is continually being serves as a template for DNA synthesis to restore any challenged due to the presence of physical and chemical lost sequences at the break site and ultimately regenerate carcinogens in our environment. In addition to DNA intact DNA (West, 2003; Sung and Klein, 2006). The lesions caused by exogenous agents, DNA suffers tumour suppressor and breast cancer-susceptibility gene spontaneous decay, replication errors, and oxidative BRCA2 is required for normal levels of HR-mediated and other damages that result from normal metabolic DSB repair (Yu et al., 2000; Moynahan et al., 2001), processes (Lindahl, 1993). The repair of damaged DNA providing an important link between the ability of a cell is therefore crucial for the maintenance of genome to maintain genome stability and tumorigenesis. integrity, and as a result, all organisms have evolved a The creation of a DSB in DNA is not always an wide variety of DNA repair pathways that can restore accidental and problematic event. A large number of DNA structure and the information encoded within it DSBs are deliberately created in the DNA of germ-line (Hoeijmakers, 2001). cells undergoing meiosis, where they serve as sites for the initiation of genetic recombination between parental chromosomes. Meiosis is a key event in the life cycle of Correspondence: Dr SC West, London Research Institute, Cancer Research UK, Clare Hall Laboratories, South Mimms, Herts EN6 all sexually reproducing organisms. It promotes the 3LD, UK. transition fromdiploid to haploid phase, and during this E-mail: [email protected] process it is essential that the maternal and paternal Regulation of recombination by BRCA2 T Thorslund and SC West 7721 NHEJ HR Predominant in G1 Predominant in S/G2 Error-prone

Resection RPA MRN complex

5’ 3’

Filament RAD51 RAD52/BRCA2 Formation

Invasion into RAD51 RAD52/RAD54 sister chromatid SDSA

DNA synthesis & RAD51 RAD52/RAD54 Second-end capture Polymerase

DNA synthesis & Polymerase Ligation Ligase

Holliday Junction Resolution

Figure 1 Schematic illustration of DNA double-strand break (DSB) repair. (a) In the S/G2 phases of the cell cycle, DSB repair is predominantly carried out by homologous recombination (HR). (b) DNA resection produces 30 single-stranded DNA (ssDNA) tails that serve as the initiation site for RAD51 ﬁlament formation (c). (d) Homologous pairing and strand invasion reactions take place with the sister chromatid to form a D-loop structure. Two possible mechanisms of HR can then take place involving the formation and resolution of Holliday junctions (e–g) or synthesis-dependent strand annealing (SDSA) (i–k). The products of mitotic recombination are generally non-crossovers. Non-homologous end joining can take place at all times throughout the cell cycle but is particularly useful in G1 when HR is inactive (h). MRN complex, MRE11/RAD50/NBS1 complex; NHEJ, non-homologous end joining

chromosomes recombine before the production of Cellular recombinases progeny (Neale and Keeney, 2006). The basic mechanism of recombination during meiosis is remarkably In all organisms, proteins belonging to the conserved similar to that which can occur in somatic cells, but a family of RecA/RAD51 recombinases mediate the few key features mark important differences (Figure 2). DNA–DNA interactions required for mitotic and meiotic Firstly, the introduction of DSBs during meiotic recombination. They have the unique ability to search for recombination is a deliberate and tightly controlled homologous sequences and to catalyse the exchange of process that is promoted by the SPO11 DNA endo- DNA strands fromone molecule to another. Most nuclease. Secondly, meiotic recombination involves the eukaryotes possess two recombinases: the ubiquitously actions of two recombinases, with DMC1 supplement- expressed RAD51 (Rad51 in Saccharomyces cerevisiae, ing the actions of RAD51. Thirdly, the search for Rhp51 in S. pombe, and RAD51 in vertebrates), and its homologous sequences by nucleoprotein ﬁlaments is meiosis-speciﬁc homologue DMC1. RAD51 is essential directed towards homologous chromosomes rather than in mammals, as indicated by the embryonic lethality the sister chromatids. Finally, while mitotic recombina- associated with the Rad51À/À k/o in the mouse (Lim and tion usually leads to the formation of non-crossover Hasty, 1996; Tsuzuki et al., 1996). In contrast, Dmc1À/À products, the processing of meiotic recombination mice develop normally but are infertile due to defects in intermediates results in at least one crossover per chromosome synapsis during meiotic recombination chromosome pair. (Pittman et al., 1998; Yoshida et al., 1998).

Oncogene Regulation of recombination by BRCA2 T Thorslund and SC West 7722

E. coli RecA SPO11 Human RAD51 339 aa

Human DMC1 340 aa Identity RAD51/DMC1 35% 40% 60%

N-terminal Helix-hairpin-Helix RecA/RAD51 core domain

E. coli C-terminus Walker A+B motifs DMC1/RAD51

Homologous Chromosome

Figure 3 The RecA/RAD51 family of recombinases. (a) Sche- matic representation of E. coli RecA, human RAD51 and human DMC1, showing their characteristic motifs. (b) These proteins are active as nucleoprotein ﬁlaments that form on DNA, as indicated schematically and in this electron microscope image.

represent sites on chromatin where DNA repair reactions take place. Similarly, RAD51 forms foci on meiotic chromosomes, where it colocalizes with DMC1 (Barlow et al., 1997; Tarsounas et al., 1999). These foci appear similar to the RAD51 foci observed in somatic cells after DNA damage and presumably identify sites of Crossover formation meiotic recombination (Haaf et al., 1995; Raderschall et al., 1999). It is not known why meiotic recombination requires two recombinases, while RAD51 is sufficient to mediate mitotic recombination. Indeed, the proteins appear to be Figure 2 Schematic illustration of meiotic recombination. Meiotic so similar, certainly in terms of their molecular structure recombination is similar to mitotic recombination, as described in and reactions catalysed, that it has been difficult to Figure 1a–g; however, there are a few important differences: during meiosis, double-strand breaks (DSBs) are introduced by SPO11 elucidate the specific role played by DMC1 during endonuclease (a), it involves two recombinases, RAD51 and meiotic recombination. What is so special about DMC1 DMC1 (b) and occurs between homologous chromosomes, not that permits recombination between homologous chro- sister chromatids, leading to crossover formation (c). mosomes rather than sister chromatids? Much of our current understanding of the mechanism of meiotic recombination comes from studies carried out in yeast, Human RAD51 and DMC1 are relatively small partly because rad51 mutants are viable in this organ- proteins of 339 and 340 amino acids, respectively. They ism. In S. cerevisiae, the appearance of Dmc1/Rad51 contain a large core domain that is homologous to the foci coincides with the formation of Spo11-induced catalytic domain of Escherichia coli RecA protein and DSBs, and the foci tend to disappear as the chromo- contains the Walker A and B motifs responsible for somes synapse (Shinohara et al., 2000). ATP binding and hydrolysis (Figure 3a). This domain This review summarizes our current understanding of is very similar in DMC1 and RAD51, whereas the how BRCA2 contributes to the control of the recombi- N-terminal regions of the two recombinases are more nation process, both during DSB repair in somatic cells divergent. Recombinant RAD51 and DMC1 proteins and at meiosis. We speculate on how BRCA2 may formoligomeric ring structures of seven or eight function as a universal regulator of both RAD51 and monomers, respectively (Shin et al., 2003; Kinebuchi the meiosis-specific DMC1 recombinase during recom- et al., 2004). However, they are functionally active when bination. associated with DNA in the formof a highly ordered right-handed helical nucleoprotein filament (Figure 3b) (Benson et al., 1994; Conway et al., 2004; Sehorn et al., 2004). It is within this nucleoprotein filament that DNA BRCA2 – a mediator of homologous recombination interactions take place between homologous sequences. In undamaged cells, RAD51 is normally distributed Breast cancer represents one of the most common throughout the nucleus, but during S phase or after cancers and accounts for about one in three of all treatment with DNA-damaging agents such as IR or cancers in women. Approximately 5–10% of individuals mitomycin C, RAD51 forms distinct nuclear foci (Haaf that develop breast cancers are genetically predisposed et al., 1995; Scully et al., 1997). These foci are thought to to the disease, and of these 10–30% have been attributed

Oncogene Regulation of recombination by BRCA2 T Thorslund and SC West 7723 to mutations in the tumour suppressor gene BRCA2. homologues. The common presence of this motif The penetrance of having BRCA2 mutations is severe – indicates its critical importance, which is thought to be there is a 50% risk of developing cancer before the age due to its ability to bind directly to RAD51. Interactions of 50 years and an 80% risk before the age of 70 years. between RAD51 and various BRCA2 fragments con- The BRCA2 gene was identified in 1995 (Wooster taining different BRC motifs have been studied using et al., 1995), and it soon became apparent that BRCA2 the yeast two-hybrid systemand also by pull-down is intimately involved in maintaining genomic stability analyses using tagged proteins. The first yeast two- via interactions with, and regulation of, the RAD51 hybrid studies demonstrated the interaction of RAD51 recombinase. The direct association of mouse BRCA2 with six of the BRC motifs, namely BRC1/BRC2/ with RAD51 was first observed in 1997 by yeast two- BRC3/BRC4 and BRC7/BRC8 (Wong et al., 1997). hybrid analyses, using a 100 amino-acid-long C-terminal Using a semiquantitative yeast two-hybrid assay, it was region of BRCA2 as bait for RAD51 (Sharan et al., estimated that a fragment containing BRC motifs 1–4 1997). The interaction between RAD51 and BRCA2 interacted approximately 5 Â stronger with RAD51 was further established by co-immunoprecipitation of than a fragment containing BRC motifs 5–8 (Chen et al., endogenous BRCA2 and RAD51 (Chen et al., 1998a, b; 1998b), and that the affinity of BRC4 for RAD51 was Marmorstein et al., 1998). RAD51 and BRCA2 were 5 Â stronger than that seen with BRC1 (Chen et al., also shown to colocalize in nuclear foci in somatic cells 1999). Others have reported yeast two-hybrid studies after IR treatment (Tarsounas et al., 2004). Interest- with individual BRC motifs showing strong interactions ingly, the formation of RAD51 foci after DNA damage between BRC motifs 1–4 and RAD51, but saw only was shown to be dependent upon functional BRCA2, a weak evidence for interactions between RAD51 and result that led to suggestions that BRCA2 is required for BRC motifs 5–8 (Thorslund et al., 2007). the recruitment of RAD51 to sites of DSBs, and thus for When analysed in pull-downs, using glutathione the cellular function of RAD51 (Sharan et al., 1997; S-transferase-tagged BRCA2 fusion proteins, it was Yuan et al., 1999). Consistent with this view, BRCA2- shown that glutathione S-transferase-BRC1, a fragment deficient murine and human cells exhibit a spontaneous of BRCA2 containing BRC2, and also a larger fragment genome instability phenotype that involves the accumu- containing both BRC3 and BRC4 interact with RAD51, lation of chromosome breaks and radial chromosomes, whereas a multitude of different glutathione S-transfer- presumably due to defects in the repair of replication- ase-tagged fragments of BRCA2 containing BRC5– associated DSBs (Patel et al., 1998; Yu et al., 2000; BRC8 did not interact with RAD51 (Chen et al., 1998b). Moynahan et al., 2001; Tutt et al., 2001). Similarly, it was shown that a glutathione S-transferase- Perhaps surprisingly, orthologues of the BRCA2 gene tagged fragment of BRCA2 containing BRC1–2, and a have been found in a variety of evolutionary diverse taxa fragment containing BRC3–5 interacted with RAD51, (Figure 4a), such as the smut fungus Ustilago maydis whereas a fragment containing BRC5–8 failed to pull- (UmBRH2) (Kojic et al., 2002), the nematode Caenor- down RAD51 (Esashi et al., 2005; Thorslund et al., 2007). habditis elegans (CeBRC-2) (Martin et al., 2005), the Hence, the ability of different BRC motifs to bind to plant Arabidopsis thaliana (AtBRCA2) (Siaud et al., RAD51 varies fromone motifto another and detection of 2004), as well as vertebrates such as chicken (cBRCA2) an interaction is dependent upon the method of analysis. (Takata et al., 2002) (Figure 4a). Human BRCA2 is a The differences in affinity may reflect the degree of very large polypeptide of 3418 amino acids and it divergence fromthe critical residues of the BRC consensus contains a number of recognizable motifs, some of sequence (Figure 4b). which can be found in all BRCA2 homologues. Most The generally conserved nature of BRC motifs indicates characteristic are the BRC motifs and the ssDNA- that they are likely to interact with a unique motif or region binding region. In human BRCA2, the ssDNA-binding in RAD51. It has been demonstrated that both BRC3 and region is composed of a conserved helical domain and BRC4 interact with the RAD51 core domain (a region of three OB folds. The structure of this region has been RAD51 lacking the N-terminal domain), suggesting that determined by X-ray crystallography and is thought to some aspect of the core represents the site of interaction be important for the recruitment of RAD51 to DSBs (Wong et al., 1997; Esashi et al., 2007). Further insight into (Yang et al., 2002). this interaction was provided when the structure of a protein comprising BRC4 fused to the RAD51 core was solved by X-ray crystallography. It was revealed that BRC4 interacts with RAD51 by mimicking the structural motif in RAD51 Binding and regulation of RAD51 by the BRC motifs of that mediates the protomer–protomer contacts necessary BRCA2 for oligomerization (Pellegrini et al., 2002). Consistent with this, it has been shown that BRC4 interacts with RAD51 to All BRCA2 proteins contain one or more BRC motif, formstable BRC4-RAD51 heterodimersthat cannot bind which is a degenerate repetitive sequence approximately DNA (Davies et al., 2001). The structural studies also 70 amino acids in length, containing a core of 25 amino showed that the conserved F-TASGK residues (a critical acids (Figures 4a and b). Human BRCA2 has eight BRC part of the BRC consensus sequence) of BRC4 (Figure 4b) motifs, six are present in chicken BRCA2, four in are involved in mediating hydrophobic and polar interac- AtBRCA2, whereas only a single BRC motif appears to tions with RAD51. The flanking non-conserved residues, be present in the U. maydis and C. elegans BRCA2 however, could be responsible for the higher affinity that

Oncogene Regulation of recombination by BRCA2 T Thorslund and SC West 7724 RAD51DMC1 ssDNA RAD51 hBRCA2 3418 aa

mBRCA2 3328 aa

cBRCA2 3397 aa

AtBRCA2 (IV+V) 1151/1155 aa

CeBRC-2 394 aa

UmBRH2 1075 aa

BRC motif PhePP motif Helical domain OB fold TR2 region

Human BRC1 1009 SFRTASNKEIKLSEHNIKKSKMFFKDI 1035 Human BRC2 1218 GFYSAHGTKLNVSTEALQKAVKLFSDI 1244 Human BRC3 1427 FFQTASGKNISVAKESFNKIVNFFDQK 1453 Human BRC4 1523 GFHTASGKKVKIAKESLDKVKNLFDEK 1549 Human BRC5 1570 AFYTSCSRKTSVSQTSLLEAKKWLREG 1596 Human BRC6 1843 AFRIASGKIVCVSHETIKKVKDIFTDS 1869 Human BRC7 1977 IFSTASGKSVQVSDASLQNARQVFSEI 2003 Human BRC8 2057 GFSTASGKQVSILESSLHKVKGVLEEF 2083 Mouse BRC1 987 SFRTASNKEIKLSEHNVKKSKMFFKDI 1013 Mouse BRC2 1198 GFCSALGTKLSVSNEALRKAMKLFSDI 1224 Mouse BRC3 1400 SFQTASGKNTRVSKESLNKSVNIFNRE 1426 Mouse BRC4 1497 SFHTASGKKVKIMQESLDKVKNLFDET 1523 Mouse BRC5 1629 AYYTEDSRKTCVRESSLKSGRKWLREQ 1655 Mouse BRC6 1812 VFITAH------SQET-ERTKIEVTDN 1831 Mouse BRC7 1930 IFSTASGKAIQVSDASLEKARQVFSEM 1956 Mouse BRC8 2010 GFSTAGGKLVTVSESALHKVKGMLEEF 2036 Chicken BRC1 964 GFQTASNKQIKFSESSIAKGKMLFKDI 990 Chicken BRC2 1201 GFTSAGGKKINISKAALTRSAELFKDL 1227 Chicken BRC3 1501 GFCTASGKKITIADGFLAKAEEFFSEN 1527 Chicken BRC4 1802 VFSTAKGKAVSVSESALASIRQMFQTD 1828 Chicken BRC5 1941 FFSTASGKPVQLSEESLKKARQLFSEM 1967 Chicken BRC6 2020 GFSTASGKQVTISESAYQKAKAILKEA 2046 AtBRC1 69 MFRTGLGKSVVLKESSIAKAKSILAEK 95 AtBRC2(IV) 122 MFRTASGKSVPLKESSIAKAMSILGSD 148 AtBRC2(V) 122 MFRTALGKTVPLKESSIAKPLSILGSD 148 AtBRC3 169 LFQTASNKKVNVSSAGLARAKALLGLE 195 AtBRC4 263 KFQTAGGKSLSVSAEALKRARNLLGDP 289 CeBRC1 34 VFSTAAGIRIDVKQESIDKSKKMLNSD 60 UmBRC1 293 GFQTGHGKQVKLSDKALEKARKLMMQL 319 Consensus F-TASGK-I-IS---L-K----L-D VFV E

* hBRCA2 3264 NCKKRRALDFLSRLPLPPPVSPICTFVSPAAQKAFQPPRSCGTKYETPIKKKELNSPQMTPFKKFNE 3330 mBRCA2 3188 TCRKRRALDFLSRLPLPSPVSPICTFVSPAAQKAFQPPRSCGTKYATPIKK-EPSSPRRTPFQKTSG 3249 cBRCA2 3212 NSKKRKAVDFLSCIPAPPPLTPLCSIISPSLKKAFQPPRRLGSQHSKLSKETNPNAGCVTPSRKLRE 3277

Human BRCA2 2386 SATRNEKMRHLITTGRPTKVFVPPFK 2411 Mouse BRCA2 2334 SATRTERTRHS---GKSTKVFVPPFK 2356 Chicken BRCA2 2335 VAKDSEETRSLCKSGKAVKTFVPPFK 2360 AtBRCA2(IV) 322 SANRGYIAHEEKTSNKHTNSFVSPLW 348 Ce BRC - 2 65 SSSKGGFSSPLVRKNNGSSAFVSPFR 90 UmBRH2 468 LPRAEIGGSSSTGSKRTLPRFVTPFK 493 Figure 4 BRCA2 proteins. (a) Schematic representation of BRCA2 proteins from human (hBRCA2), mouse (mBRCA2), chicken (cBRCA2), A. thaliana (AtBRCA2), C. elegans (CeBRC-2), and U. maydis (UmBRH2). Key motifs are indicated, and the hatched purple boxes show putative diverged PhePP motifs. (b) Alignment of the core BRC motifs from different species. (c) Alignment of the TR2 regions of vertebrate BRCA2. Blue amino acids represents those conserved in mammalian species. Serine 3291 is indicated by the red asterisk. (d) Alignment of the PhePP motif from vertebrates, and the putative but diverged PhePP motifs in lower eukaryotes. Blue amino acids represent conserved amino acids.

Oncogene Regulation of recombination by BRCA2 T Thorslund and SC West 7725 BRC4 shows for binding RAD51 compared to some of the cells can also be, at least partially, rescued by artificial other BRC motifs (Pellegrini et al., 2002). BRCA2 fusions containing just BRC3 or BRC4 fused to the Further insights into the interactions of the BRC motifs ssDNA-binding domain of RPA (Saeki et al., 2006). These and RAD51 have been provided fromstudies carried out fusions also restored the ability to formRAD51 foci after with small synthetic peptides. A unifying theme that comes DNA damage, supporting the notion that the primary role out of these experiments is that there are differences that of BRCA2 is to target RAD51 to ssDNA generated at a depend on which BRC motif is being analysed, and under DSB, and that both the BRC motifs and the DNA-binding what reaction conditions. Due to the availability of the region of BRCA2 are indispensable for this RAD51- BRC4-RAD51 core structure, many studies have focused targeting activity. on a peptide corresponding to BRC4 itself. The BRC4 Taken together, these results shed new light on what peptide was found to disrupt RAD51 filaments, presum- appear to be complex associations between the BRC ably by binding RAD51 monomers in solution, thereby motifs of BRCA2 with RAD51 and on how these shifting the equilibriumtowards BRC4-RAD51 hetero- interactions are necessary for regulation of RAD51 dimer formation (Davies and Pellegrini, 2007; Esashi et al., function. It is clear that not all BRC motifs are equal in 2007). However, this disruptive effect was only observed terms of their interaction with RAD51, and the same is when BRC4 was present in molar excess compared to implied in terms of how the individual BRC motifs might RAD51, and at lower concentrations of BRC4 it was contribute to RAD51 function. Indeed, the BRC motifs found to associate with RAD51 filaments formed on may have dual, maybe even multiple, non-exclusive roles dsDNA in the presence of the non-hydrolysable ATP in regulating RAD51 within the cell (Figure 5). The analogue 50-adenylyl-b, g-imidodiphosphate (AMP-PNP) binding of monomeric, and thereby inactive, RAD51 to (Galkin et al., 2005). When similar experiments were the BRC motifs may be important for the sequestration carried out with BRC3, it was found that excess BRC3 of RAD51 at certain times during the cell cycle, could again dissociate RAD51 filaments, but only in the particularly when there are no DSBs to be repaired. presence of ATP, while sub-stoichiometric amounts of Upon DSB formation, however, the BRCs may play a BRC3 were found to formstable associations with the critical role as BRCA2 elicits a rapid response by RAD51 filaments. In the presence of a non-hydrolysable recruiting RAD51 to the breaks. The recombinase then ATP analogue, BRC3 could formstable associations with needs to be released such that it forms active nucleo- the RAD51 filaments even when in excess over RAD51 protein filaments that can initiate recombinational (Galkin et al., 2005). That the different BRC motifs are repair. Whether the BRCs are positioned on BRCA2 in non-equivalent in their ability to interact with RAD51 is such a way that it helps to facilitate the establishment of further strengthened by the observation that a single a nucleoprotein filament remains to be determined. In mutation in RAD51-A190V prevents RAD51 interaction this regard, it is interesting to note that a large fragment with BRC3 but not BRC4 (Esashi et al., 2007). of BRCA2, containing all eight BRC motifs was recently That BRC3 and BRC4 can interact with RAD51 purified and shown to stimulate RAD51-mediated strand filaments or can adversely affect their stability may exchange in vitro (Shivji et al., 2006). indicate a potential regulatory function for the BRC motifs. However, this ‘yin and yang’ concept does not appear to correlate well with the positive role that BRCA2 protein plays in RAD51-mediated recombina- Stabilization of RAD51 nucleoprotein filaments by tion. To understand the molecular function of BRCA2, BRCA2 C-terminus is regulated by phosphorylation it will be necessary to purify the protein to homogeneity. Unfortunately, at present, this appears to be an In addition to its interaction with the BRC repeats, RAD51 exceedingly difficult task, due in part to the large size also interacts with an unrelated C-terminal region of of the protein (384 kDa) and its propensity to degrade. BRCA2 (Mizuta et al., 1997; Sharan et al., 1997). This A number of in vivo studies have helped to define the region has been mapped and is referred to as the TR2 contribution of the various domains of BRCA2 in relation region (Esashi et al., 2005) (Figure 4c). TR2 is present within to its biological functions. For example, it has been shown the region encoded by exon 27 and is highly conserved in that the BRCA2-deficient pancreatic cancer cell line called vertebrate BRCA2 proteins, but is not found in the BRCA2 CAPAN-1 is hypersensitive to methyl methanesulphonate homologues from more diverse taxa. Deletion of exon 27 in and other DNA-damaging agents. These cells have lost one the mouse leads to increased tumour incidence (McAllister BRCA2 allele and contain a truncating (6174delT) mutation et al., 2002; Donoho et al., 2003). within BRC7 that leads to the production of a truncated It is thought that the C-terminal RAD51 interaction BRCA2 protein containing just the first six BRC motifs domain within BRCA2 serves a regulatory role in without the ssDNA-binding domain (Goggins et al., 1996). recombinational repair. It has been shown that serine Despite the presence of these BRC motifs, CAPAN-1 cells 3291 in the TR2 region is phosphorylated by cyclin- fail to formRAD51 foci in response to IR (Yuan et al., dependent kinases at late G2 phase of the cell cycle, and 1999), indicating the important roles played by the ssDNA- that S3291-phosphorylated TR2 is unable to bind binding domain and the C-terminal region of BRCA2. RAD51. These results indicate that the inability of The methyl methanesulphonate sensitivity of CAPAN-1 RAD51 to bind the C-terminal region of BRCA2 cells can be rescued by transient expression of wild-type provides a mechanism by which the cell can down- BRCA2 (Chen et al., 1998b). Remarkably, BRCA2-deficient regulate recombination as it approaches mitotic division.

Oncogene Regulation of recombination by BRCA2 T Thorslund and SC West 7726 RAD51 binding by BRCA2

Stimulation of filament disassembly BRC-mediated Recruitment of RAD51 to resected DSBs RAD51 regulation

Stimulation of filament assembly Figure 5 Putative functions of the BRCA2 BRC motifs in RAD51 regulation. Binding of monomeric RAD51 to the BRC motifs of BRCA2 could inactivate and sequester RAD51 at specific stages of the cell cycle in the absence of DNA damage. Upon DNA damage, the BRCA2–RAD51 complex can relocalize to the sites of damage. The ssDNA-binding domain of BRCA2 is important for this recruitment, as are BRC–RAD51 interactions. Once at the resected DSB, the structural and topological organization of monomeric RAD51 bound to the BRC motifs may help to facilitate the initiation of RAD51 filament formation. After successful strand invasion, the ability of BRC motifs to bind monomeric RAD51 and destabilize RAD51 filaments may be important for filament disassembly. DSB, double-strand break.

Importantly, S3291 is de-phosphorylated in response to Phosphorylation of Ser3291 IR, thereby providing a positive mechanism for the activation of HR in response to DNA damage (Esashi et al., 2005). Thus, the phosphorylation status of S3291 is Activation of HR Inactivation of HR important for modulating interactions between the BRCA2 C-terminus and RAD51, and may provide dual Dephosphorylation of Ser3291 functions (inhibition and activation) during HR. TR2 51 TR2 51 New light relating to the regulatory functions of TR2 P was recently provided by observations showing that the TR2-RAD51 interaction No TR2-RAD51 interaction interaction of TR2 with RAD51 is physically and functionally distinct fromthat mediated by the BRC motifs. Using synthetic peptides corresponding to the TR2 region, and mutant derivatives of RAD51 protein, it RAD51 filament stabilization No stable RAD51 filaments was shown that TR2 interacts specifically with multimeric RAD51 (filaments or rings) but cannot bind RAD51 Figure 6 Stabilization of RAD51 filaments by the C-terminal monomers (Davies and Pellegrini, 2007; Esashi et al., region of BRCA2. During the S and early G2 phases of the cell 2007). Furthermore, TR2 was found to stabilize RAD51 cycle, or after DNA damage, the C-terminal TR2 region of BRCA2 is non-phosphorylated at serine 3291 and thus capable of filaments from the disruptive actions of BRC4, leading to interaction with RAD51. Interaction of TR2 with RAD51 the suggestion that the C-terminal domain of BRCA2 filaments leads to their stabilization, and as a consequence HR is plays a positive role in the establishment of functional active in the repair of DSBs resulting fromreplication fork collapse RAD51 nucleoprotein filaments. These observations are or genotoxic stress. As cells approach mitosis, serine 3291 is phosphorylated, serving to downregulate HR, as TR2 cannot important as they imply that the stabilization/destabiliza- interact with and stabilize RAD51 filaments. HR, homologous tion of RAD51 filaments, mediated by the combined recombination. actions of the BRC and TR2 motifs of BRCA2, may be regulated by the phosphorylation status of S3291 (Figure 6). Non-phosphorylated TR2 can bind and stabilize an existing RAD51 filament, whereas phosphorylation of S3291 blocks recombination and the maintenance of genome stability. RAD51 binding to the C-terminal region of BRCA2, BRH2 is also required for the formation of viable haploid leaving nucleoprotein filaments unprotected to the destabi- progeny fromteliospores, showing that it is essential for lizing influence of BRC motifs. The TR2 and BRC motifs meiosis (Kojic et al., 2002; Yang et al., 2005). in BRCA2 may thereby promote a coordinated response In the nematode C. elegans, a null mutation in BRC-2 that is dependent upon the cell cycle and/or DSB formation. results in defects in meiotic recombination (Martin et al., 2005). Similarly, evidence for a role in meiosis came from plants. Two conserved and almost identical BRCA2-like genes are found in A. thaliana, AtBRCA2(IV) and Actions of BRCA2 in meiotic recombination AtBRCA2(V) (roman numerals refer to the chromosome on which the gene is found). Both genes are expressed The first indications of a role for BRCA2 in meiosis came and encode AtBRCA2 homologues of 1151 amino acids fromstudies of unicellular eukaryotes. Like the human (IV) and 1155 amino acids (V), respectively. The protein, U.maydisBRH2 is required for mitotic AtBRCA2 homologues contain four BRC motifs, and

Oncogene Regulation of recombination by BRCA2 T Thorslund and SC West 7727 BRC2 is the only one that differs between the two plant described (Thorslund et al., 2007). Using a combination BRCA2 proteins, which overall share 95% identity of yeast two-hybrid analysis, pull-downs with recombi- (Figure 4b). Using a BRCA2 RNAi construct, which nant tagged fusion proteins, and peptide-arrays, the inactivates both AtBRCA2(IV) and AtBRCA2(V), expres- primary DMC1 interaction site on BRCA2 was mapped sed under the control of a meiosis-specific promoter, it to a motif located at BRCA22386–2411. This novel DMC1 was shown that BRCA2 inactivation results in partial interaction motif has been named the PhePP motif, due sterility. This phenotype was dependent on SPO11, as the to the critical importance of amino acids Phe2406, BRCA2 knockdown had no effect in a SPO11-deficient Pro2408, and Pro2409 (Figure 4d). The PhePP motif is background (Siaud et al., 2004), indicating that BRCA2 is unrelated to the BRC repeats and located away fromthe required for the processing of SPO11-generated DSBs in RAD51 interaction sites on BRCA2. It promotes meiotic recombination. specific interactions between BRCA2 and DMC1, but The role of BRCA2 in meiosis in mammals has been not RAD51. more difficult to study due to the embryonic lethality The PhePP motif is highly conserved in BRCA2 from a associated with inactivating BRCA2 mutations (Ludwig variety of vertebrate species, but has diverged in A. et al., 1997; Sharan et al., 1997; Suzuki et al., 1997). thaliana, U.maydis,andC. elegans (Figure 4d). The However, indications of its involvement in meiosis were identification of this novel DMC1-interacting motif in provided by observations showing that BRCA2 is highly BRCA2 was surprising as, based on the observed expressed during spermatogenesis in mice (Connor interactions between AtBRCA2 and AtDMC1, it was et al., 1997) and localizes to meiotic chromosomes expected that interactions between hBRCA2 and during early meiotic prophase I when homologous hDMC1 would be mediated though one or more of the chromosomes undergo synapsis (Chen et al., 1998a). BRC motifs. Yeast two-hybrid analyses did reveal an Viable mice that exhibit partial loss of function association between DMC1 and the BRC4 motif of truncations in BRCA2 exon 11 displayed a complete hBRCA2, but this was so weak that it could not be failure in spermatogenesis (Connor et al., 1997), whereas substantiated in pull-down experiments (Thorslund et al., no meiotic defects were observed in mice where the 2007). Moreover, interactions could not be observed C-terminal region, including TR2, was deleted (McAllister between hDMC1 and a larger fragment of BRCA2 et al., 2002; Atanassov et al., 2005). Conclusive evidence containing all eight BRC motifs. The interactions of for the involvement of BRCA2 in meiosis was confirmed hBRCA2 with hDMC1, mediated through the PhePP when it was shown that BRCA2 k/o mice, rescued by a motif, therefore contrast with those that take place with bacterial artificial chromosome expressing human hRAD51. BRCA2, which exhibit low expression of the BRCA2 hDMC1 was also found to associate with the TR2 transgene in the gonads, were infertile as their spermato- region of BRCA2, but this interaction was weak in cytes failed to progress beyond early prophase I. In these comparison to that which takes place between TR2 and mice, SPO11-induced DSBs occur normally, but persist RAD51, and was not affected by phosphorylation of without being repaired (Sharan et al., 2004). Importantly, S3291 (Thorslund et al., 2007). The biological signifi- in these mice, both RAD51 and DMC1 failed to localize to cance of this interaction may be minor as transgenic foci on meiotic chromosomes, indicating that BRCA2 is mice with C-terminal BRCA2 deletions fail to exhibit required for localization of both recombinases to SPO11- meiotic defects (McAllister et al., 2002; Atanassov et al., induced DSBs. 2005). This is important because it indicates that the TR2 region, which is required for the stabilization of RAD51 filaments during HR in somatic cells, is dispensable for HR during meiosis. Whether there are Interactions between BRCA2 and DMC1 other factors that stabilize RAD51 filaments during meiosis, or whether at meiosis the bias is switched from Two-hybrid analyses, together with pull-down studies RAD51 stabilization to DMC1 filament stabilization, with tagged protein constructs, showed that the remains to be investigated. AtBRCA2 proteins interact directly with both RAD51 Although the DMC1-interacting PhePP motif identi- and DMC1 (Siaud et al., 2004). These interactions occur fied in hBRCA2 is divergent in non-vertebrate eukar- through the N-terminal BRC motif-containing region of yotic species, some of the highly conserved amino acids BRCA2 (AtBRCA2(IV)1–784). Interestingly, interactions found in mammals that are important for DMC1 with RAD51 were blocked by the presence of a polypeptide interaction (for example, Phe2406 and Pro2409 of corresponding to the human BRC3 motif, while this peptide hBRCA2) are also present in AtBRCA2, UmBRH2, hadnoeffectonAtBRCA2(IV)–AtDMC1 interactions and CeBRC-2 (Figure 4d). It is possible that this (Dray et al., 2006). Furthermore, the BRC2 motif of divergent putative PhePP motif in AtBRCA2 may, in AtBRCA2(IV) was found to interact with AtDMC1, addition to the AtBRCA2(IV)-BRC2 motif, be involved whereas the binding to RAD51 was mediated through the in interactions with AtDMC1, but this remains to be BRC4 motif, indicating that the AtBRC motifs exhibit investigated. Interestingly, the diverged PhePP motif different specificities for the two recombinases (Dray et al., in C. elegans was recently shown to interact with 2006). CeRAD51 (Petalcorin et al., 2007). C. elegans is unusual Details of the interactions that take place between in that it lacks DMC1, and promotes meiotic recombi- human BRCA2 and DMC1 protein were recently nation through the high-level expression of CeRAD51

Oncogene Regulation of recombination by BRCA2 T Thorslund and SC West 7728 alone (Takanami et al., 2000). That this region of Perhaps the greatest challenge will be to develop the BRCA2 in humans interacts specifically with DMC1, present studies, carried out with BRCA2 peptides and whereas in C. elegans it interacts with RAD51, the only fragments, in terms of understanding how this large recombinase in this organism, may suggest a general protein functions as a whole, and towards this goal it regulatory role. Evidence for this proposal was recently will be a remarkable achievement to purify the full-size obtained when it was shown that a peptide correspond- protein. As techniques improve, we hope that it will be ing to CeBRC-260–89, containing the PhePP-like motif, possible to prepare recombinant BRCA2 for biochem- binds and stabilizes CeRAD51 nucleoprotein filaments ical studies. The availability of full-length BRCA2 will (Petalcorin et al., 2007). allow the analysis of BRC- and TR2-mediated RAD51 complex formation and permit us to develop in vitro assays for BRCA2-mediated nucleoprotein filament formation. No doubt we will get answers that explain Conclusions and perspectives how the BRC motifs can both bind and disrupt RAD51 nucleoprotein filaments, and it is entirely possible that One priority for the future will be to understand how these apparently antagonistic effects will not be appar- BRCA2 coordinates the binding of RAD51 and DMC1, ent with the full-size protein. and regulates their function during meiotic recombina- The identification of a novel DMC1 interaction site in tion. Given that RAD51 and DMC1 have distinct and BRCA2 suggests that mutations in BRCA2 may not non-overlapping binding sites on BRCA2, it is assumed only be associated with breast cancer susceptibility, but that a BRCA2–RAD51–DMC1 complex may exist may also be associated with fertility problems. We in vivo. Data to support this concept, however, is yet suggest that these two phenotypes are not necessarily to be obtained. We suggest that the simultaneous linked, and that specific mutations in the PhePP motif binding of both RAD51 and DMC1 is likely to be will affect interactions with DMC1 but not RAD51. The necessary for the localization of both recombinases to consequences will relate to defects in meiotic recombi- meiotic chromosomes following SPO11-induced DSB nation without necessarily causing the genomic instabil- formation during meiotic recombination. But how the ities that lead to tumour predisposition. recombinases are loaded, whether it occurs sequentially or via the formation of mixed nucleoprotein filaments Acknowledgements remains to be determined. Little is presently known about the regions in DMC1 that mediate interactions We thank Dr Uli Rass for critical reading of the manuscript. with the PhePP domain of BRCA2, and this informa- This work was supported by Cancer Research UK, the EU tion may be valuable in understanding how BRCA2 DNA Repair Consortium, the Breast Cancer Campaign, and regulates the coordinated actions of DMC1 and RAD51 the Louis-Jeantet Foundation. TT is a recipient of a post- during HR at meiosis. doctoral fellowship fromthe Carlsberg foundation.

References

Atanassov BS, Barrett JC, Davis BJ. (2005). Homozygous binding and resistance to methyl methanesulfonate treat- germ-line mutation in exon 27 of murine BRCA2 disrupts the ment. Proc Natl Acad Sci USA 95: 5287–5292. FANCD2-BRCA2 pathway in the homologous recombina- Connor F, Bertwistle D, Mee PJ, Ross GM, Swift S, tion-mediated DNA interstrand cross-link repair but does Grigorieva E et al. (1997). Tumorigenesis and a DNA-repair not affect meiosis. Genes Chromosomes Cancer 44: 429–437. defect in mice with a truncating BRCA2 mutation. Nat Genet Barlow AL, Benson FE, West SC, Hulte´ n MA. (1997). 17: 423–430. Distribution of RAD51 recombinase in human and mouse Conway AB, Lynch TW, Zhang Y, Fortin GS, Fung CW, spermatocytes. EMBO J 16: 5207–5215. Symington LS et al. (2004). Crystal structure of a Rad51 Bennett CB, Lewis AL, Baldwin KK, Resnick MA. (1993). filament. Nat Struct Mol Biol 11: 791–796. Lethality induced by a single site-specific double-strand Davies AA, Masson J-Y, McIlwraith MJ, Stasiak AZ, Stasiak break in a dispensable yeast plasmid. Proc Natl Acad Sci A, Venkitaraman AR et al. (2001). Role of BRCA2 in USA 90: 5613–5617. control of the RAD51 recombination and DNA repair Benson FE, Stasiak A, West SC. (1994). Purification and protein. Mol Cell 7: 273–282. characterisation of the human RAD51 protein, an analogue Davies OR, Pellegrini L. (2007). Interaction with the BRCA2 of E. coli RecA. EMBO J 13: 5764–5771. C-terminus protects RAD51-DNA filaments from disas- Chen CF, Chen PL, Zhong Q, Sharp ZD, Lee WH. (1999). sembly by BRC repeats. Nat Struct Mol Biol 14: 475–483. Expression of BRC repeats in breast cancer cells disrupts the Donoho G, Brenneman MA, Cui TX, Donoviel D, Vogel H, BRCA2–RAD51 complex and leads to radiation hypersen- Goodwin EH et al. (2003). Deletion of BRCA2 exon 27 sitivity and loss of G(2)/M checkpoint control. J Biol Chem causes hypersensitivity to DNA crosslinks, chromosomal 274: 32931–32935. instability, and reduced life span in mice. Genes Chromosomes Chen JJ, Silver DP, Walpita D, Cantor SB, Gazdar AF, Cancer 36: 317–331. Tomlinson G et al. (1998a). Stable interaction between the Dray E, Siaud N, Dubois E, Doutriaux MP. (2006). Interaction products of the BRCA1 and BRCA2 tumor suppressor between Arabidopsis BRCA2 and its partners RAD51, genes in mitotic and meiotic cells. Mol Cell 2: 317–328. DMC1, and DSS1. Plant Physiol 140: 1059–1069. Chen PL, Chen CF, Chen YM, Xiao J, Sharp ZD, Lee WH. Esashi F, Christ N, Gannon J, Liu Y, Hunt T, Jasin M et al. (1998b). The BRC repeats in BRCA2 are critical for RAD51 (2005). CDK-dependent phosphorylation of BRCA2 as a

Oncogene Regulation of recombination by BRCA2 T Thorslund and SC West 7729 regulatory mechanism for recombinational repair. Nature Petalcorin MIR, Galkin VE, Yu X, Egelman EH, Boulton SJ. 434: 598–604. (2007). Stabilization of RAD-51-DNA filaments by a novel Esashi F, Galkin VE, Yu X, Egelman EH, West SC. (2007). interaction domain in Caenorhabditis elegans BRCA2. Proc Stabilization of RAD51 nucleoprotein filaments by the C- Natl Acad Sci USA 104: 8299–8304. terminal region of BRCA2. Nat Struct Mol Biol 14: 468–474. Pittman DL, Cobb J, Schimenti KJ, Wilson LA, Cooper DM, Galkin VE, Esashi F, Yu X, Yang SX, West SC, Egelman EH. Brignull E et al. (1998). Meiotic prophase arrest with failure (2005). BRCA2 BRC motifs bind RAD51-DNA filaments. of chromosome synapsis in mice deficient for DMC1, Proc Natl Acad Sci USA 102: 8537–8542. a germline-specific RecA homolog. Mol Cell 1: 697–705. Goggins M, Schutte M, Lu J, Moskaluk CA, Weinstein CL, Raderschall E, Golub EI, Haaf T. (1999). Nuclear foci of Petersen GM et al. (1996). Germline BRCA2 gene mutations mammalian recombination proteins are located at single- in patients with apparently sporadic pancreatic carcinomas. stranded DNA regions formed after DNA damage. Proc Can Res 56: 5360–5364. Natl Acad Sci USA 96: 1921–1926. Haaf T, Golub EI, Reddy G, Radding CM, Ward DC. (1995). Saeki H, Siaud N, Christ N, Wiegant WW, van BPPW, Nuclear foci of mammalian RAD51 recombination protein Han MG et al. (2006). Suppression of the DNA repair in somatic cells after DNA damage and its localization in defects of BRCA2-deficient cells with heterologous protein synaptonemal complexes. Proc Natl Acad Sci USA 92: fusions. Proc Natl Acad Sci USA 103: 8768–8773. 2298–2302. Scully R, Chen J, Plug A, Xiao Y, Weaver D, Feunteun J et al. Hoeijmakers JH. (2001). Genome maintenance mechanisms (1997). Association of BRCA1 with RAD51 in mitotic and for preventing cancer. Nature 411: 366–374. meiotic cells. Cell 88: 265–275. Kinebuchi T, Kagawa W, Enomoto R, Tanaka K, Miyagawa K, Sehorn MG, Sigurdsson S, Bussen W, Unger VM, Sung P. Shibata T et al. (2004). Structural basis for octameric ring (2004). Human meiotic recombinase DMC1 promotes ATP- formation and DNA interaction of the human homologous- dependent homologous DNA strand exchange. Nature 429: pairing protein DMC1. Mol Cell 14: 363–374. 433–437. Kojic M, Kostrub CF, Buchman AR, Holloman WK. (2002). Sharan SK, Morimatsu M, Albrecht U, Lim SS, Regel E, Dinh C BRCA2 homolog required for proficiency in DNA repair, et al. (1997). Embryonic lethality and radiation hypersensitivity recombination, and genome stability in Ustilago maydis. mediated by RAD51 in mice lacking BRCA2. Nature 386: Mol Cell 10: 683–691. 804–810. Lim DS, Hasty P. (1996). A mutation in mouse RAD51 results Sharan SK, Pyle A, Coppola V, Babus J, Swaminathan S, in an early embryonic lethal that is suppressed by a mutation Benedict J et al. (2004). BRCA2 deficiency in mice leads to in p53. Mol Cell Biol 16: 7133–7143. meiotic impairment and infertility. Development 131: 131–142. Lindahl T. (1993). Instability and decay of the primary Shin DS, Pellegrini L, Daniels DS, Yelent B, Craig L, Bates D structure of DNA. Nature 362: 709–715. et al. (2003). Full-length archaeal RAD51 structure and Ludwig T, Chapman DL, Papaioannou VE, Efstratiadis A. mutants: mechanisms for RAD51 assembly and control by (1997). Targeted mutations of breast cancer susceptibility gene BRCA2. EMBO J 22: 4566–4576. homologs in mice: lethal phenotypes of BRCA1, BRCA2, Shinohara M, Gasior SL, Bishop DK, Shinohara A. (2000). BRCA1/BRCA2, BRCA1/p53,andBRCA2/p53 null zygous Tid1/Rdh54 promotes colocalization of Rad51 and Dmc1 embryos. Genes Dev 11: 1226–1241. during meiotic recombination. Proc Natl Acad Sci USA 97: Marmorstein LY, Ouchi T, Aaronson SA. (1998). The BRCA2 10814–10819. gene product functionally interacts with p53 and RAD51. Shivji MKK, Davies OR, Savill JM, Bates DL, Pellegrini L, Proc Natl Acad Sci USA 95: 13869–13874. Venkitaraman AR. (2006). A region of human BRCA2 Martin JS, Winkelmann N, Petalcorin MIR, McIlwraith MJ, containing multiple BRC repeats promotes RAD51-mediated Boulton SJ. (2005). RAD-51-dependent and independent strand exchange. Nucl Acids Res 34: 4000–4011. roles of a Caenorhabditis elegans BRCA2-related protein Siaud N, Dray E, Gy I, Gerard E, Takvorian N, Doutriaux MP. during DNA double-strand break repair. Mol Cell Biol 25: (2004). BRCA2 is involved in meiosis in Arabidopsis thaliana 3127–3139. as suggested by its interaction with DMC1. EMBO J 23: McAllister KA, Bennett LM, Houle CD, Ward TA, Malphurs J, 1392–1401. Collins NK et al. (2002). Cancer susceptibility of mice with a Sung P, Klein H. (2006). Mechanismof homologous homozygous deletion of the COOH-terminal domain of the recombination: mediators and helicases take on regulatory BRCA2 gene. Cancer Res 62: 990–994. functions. Nat Rev Mol Cell Biol 7: 739–750. Mizuta R, Lasalle JM, Cheng HL, Shinohara A, Ogawa H, Suzuki A, Delapompa JL, Hakem R, Elia A, Yoshida R, Mo Copeland N et al. (1997). Rab22 and Rab163/mouse R et al. (1997). BRCA2 is required for embryonic cellular BRCA2: proteins that specifically interact with the RAD51 proliferation in the mouse. Genes Dev 11: 1242–1252. protein. Proc Natl Acad Sci USA 94: 6927–6932. Takanami T, Mori A, Takahashi H, Higashitani A. (2000). Moynahan ME, Pierce AJ, Jasin M. (2001). BRCA2 is Hyper-resistance of meiotic cells to radiation due to a strong required for homology-directed repair of chromosomal expression of a single recA-like gene in Caenorhabditis elegans. breaks. Mol Cell 7: 263–272. Nucl Acids Res 28: 4232–4236. Neale MJ, Keeney S. (2006). Clarifying the mechanics of Takata M, Tachiiri S, Fujimori A, Thompson LH, Miki F, DNA strand exchange in meiotic recombination. Nature Hiraoka M et al. (2002). Conserved domains in the chicken 442: 153–158. homologue of BRCA2. Oncogene 21: 1130–1134. Patel KJ, Yu VPCC, Lee H, Corcoran A, Thistlethwaite FC, Tarsounas M, Davies AA, West SC. (2004). RAD51 localiza- Evans MJ et al. (1998). Involvement of BRCA2 in DNA tion and activation following DNA damage. Phil Trans R repair. Mol Cell 1: 347–357. Soc Lond B 359: 87–93. Pellegrini L, Yu DS, Lo T, Anand S, Lee MY, Blundell TL Tarsounas M, Morita T, Pearlman RE, Moens PB. (1999). et al. (2002). Insights into DNA recombination from RAD51 and DMC1 form mixed complexes associated with the structure of a RAD51–BRCA2 complex. Nature 420: mouse meiotic chromosome cores and synaptonemal com- 287–293. plexes. J Cell Biol 147: 207–219.

Oncogene Regulation of recombination by BRCA2 T Thorslund and SC West 7730 Thorslund T, Esashi F, West SC. (2007). Interactions between Yang H, Li Q, Holloman WK, Pavletich NP. (2005). The human BRCA2 protein and the meiosis-specific recombi- BRCA2 homologue BRH2 nucleates RAD51 filament nase DMC1. EMBO J 26: 2915–2922. formation at a dsDNS–ssDNA junction. Nature 433: Tsuzuki T, Fujii Y, Sakuma K, Tominaga Y, Nakao K, 653–657. Sekiguchi M et al. (1996). Targeted disruption of the RAD51 Yang HJ, Jeffrey PD, Miller J, Kinnucan E, Sun YT, Thoma NH gene leads to lethality in embryonic mice. Proc Natl Acad Sci et al. (2002). BRCA2 function in DNA binding and USA 93: 6236–6240. recombination from a BRCA2–DSS1–ssDNA structure. Tutt A, Bertwistle D, Valentine J, Gabriel A, Swift S, Ross G Science 297: 1837–1848. et al. (2001). Mutation in BRCA2 stimulates error-prone Yoshida K, Kondoh G, Matsuda Y, Habu T, Nishimune Y, homology-directed repair of DNA double-strand breaks Morita T. (1998). The mouse recA-like gene DMC1 is occurring between repeated sequences. EMBO J 20: required for homologous chromosome synapsis during 4704–4716. meiosis. Mol Cell 1: 707–718. West SC. (2003). Molecular views of recombination proteins Yu VPCC, Ko¨ ehler M, Steinlein C, Schmid M, Hanakahi LA, and their control. Nat Rev Mol Cell Biol 4: 435–445. van Gool AJ et al. (2000). Gross chromosomal rearrange- Wong AKC, Pero R, Ormonde PA, Tavtigian SV, Bartel PL. ments and genetic exchange between non-homologous (1997). RAD51 interacts with the evolutionarily conserved chromosomes following BRCA2 inactivation. Genes Dev BRC motifs in the human breast cancer susceptibility gene 14: 1400–1406. BRCA2. J Biol Chem 272: 31941–31944. Yuan S-SF, Lee S-Y, Chen G, Song M, Tomlinson GE, Wooster R, Bignell G, Lancaster J, Swift S, Seal S, Mangion J Lee EY. (1999). BRCA2 is required for ionizing radiation- et al. (1995). Identification of the breast cancer susceptibility induced assembly of RAD51 complex in vivo. Cancer Res 59: gene BRCA2. Nature 378: 789–792. 3547–3551.

Oncogene