BRCT domains

Abstract

BRCA1 C-terminal (BRCT) domains are integral signaling modules in the DNA damage response (DDR). Aside from their established roles as phospho-peptide binding modules, BRCT domains have been implicated in phosphorylation-independent protein interactions, DNA binding and poly(ADP-ribose) (PAR) binding. These numerous functions can be attributed to the diversity in BRCT domain structure and architecture, where domains can exist as isolated single domains or assemble into higher order homo- or hetero- domain complexes. In this review, we incorporate recent structural and biochemical studies to demonstrate how structural features allow single and tandem BRCT domains to attain a high degree of functional diversity.

BRCT Domains in DNA Damage Response

The DNA damage response (DDR) is an elaborate signaling network that orchestrates the cellular response to maintain genomic integrity. Such signaling cascades rely heavily on protein phosphorylation by serine/threonine kinases that include the phosphoinositide-3-kinase-related kinases (PIKKs) (ATM, ATR, DNA-PK)1,2 and checkpoint effector kinases (Chk1, Chk2 and MK2).3 Phosphorylated proteins are recognized by specialized 14-3-3, forkhead associated (FHA) and BRCA1 C-terminal (BRCT) domains present in downstream DDR proteins. Interestingly, these phospho-peptide binding modules commonly match the conserved phospho-peptide motifs targeted by their cognate kinases. 14-3-3 domains preferentially bind the targets of checkpoint kinases, while the BRCT and FHA domains can bind to the phosphorylated targets of ATM/ATR.4 While the kinases involved show little or no substrate specificity for serine or threonine residues, the binding modules can discriminate between their phospho-threonine or phospho-serine targets. For example, FHA domains show profound specificity for pT-peptides, while certain BRCT domains show specificity for pS-peptides.

The BRCT domain was originally identified in the tumor suppressor protein, Breast cancer associated 1 (BRCA1). Truncation and missense mutations in this region correlate with an increased risk for breast and ovarian cancers, underscoring its importance in the tumor suppressor function of BRCA1.5,6 Subsequent bioinformatics analysis revealed that BRCT domains exist in a myriad of proteins, most of which have roles in DNA metabolism and repair.7,8 The conserved phospho-peptide binding function was later identified in BRCA1 and other tandem BRCT domains, as well as various single BRCT domains.9,10 For example, the tandem BRCT domain pair in BRCA1 recognizes the phosphorylated peptide motif pS-P-T-F in various protein partners such as BRCA1-associated C-terminal helicase/Fanconi anemia group J protein (BACH1/FANCJ),9,10 CtIP11 and Abraxas/Coiled-coil domain-containing protein 98 (CCDC98),12–14 and the ability of BRCA1 to recognize different binding partners in DDR signaling regulates BRCA1 recruitment and function.

It is now clear that BRCT domains possess an array of activities other than phosphorylated protein interactions. Besides the conserved phospho-peptide recognition exhibited by tandem BRCT domains, BRCT domains have also been implicated in phosphorylation-independent protein interactions, DNA binding and poly(ADP-ribose) (PAR) binding. This diversity can be credited not only to variations in sequence and structure within a single BRCT domain, but also to the unique ability of BRCT domains to assemble as multi-domain complexes with other BRCT domains or even other functional domains, adding another level of complexity, specificity and regulation. In light of structural and biochemical findings in recent years, we analyze the structural aspects of single and tandem BRCT domains that contribute to their functional versatility in DDR signaling.

BRCT Domain Structure and Assembly

The BRCT domain fold was first revealed in the crystal structure of the X-ray repair cross-complementing protein 1 (XRCC1) N-terminal BRCT domain.15 The fold is comprised of a central 4-stranded β-sheet flanked by a single α-helix (α₂) on one side and two α-helices (α₁ and α₃) on the opposite side. Comparison of various BRCT domain structures illustrate that deviations in structure and sequence are mainly localized to the connecting loops, whereas conserved residues are situated in the hydrophobic core and in residues involved in recognizing the phosphorylated amino acid in phospho-peptide targets.16,17

Interestingly, the domain architecture of BRCT domains is remarkably diverse. BRCT domains range from isolated individual domains to multiple tandem BRCT repeats, or even as fusions with other functional domains (Table 1). Single BRCT domains represent a large class of BRCT domains that exist as single copies in proteins such as Poly(ADP-ribose) polymerase 1 (PARP-1) and DNA Ligase III (Lig3). They can also be found in multiple but isolated copies, such as in XRCC1, where two distinct single BRCT domains are separated by 135 amino acids in sequence. In tandem BRCT repeats, multiple BRCT domains are separated by a variable linker region. The initial crystal structure of the tandem BRCT domains in BRCA1 shed light on the canonical BRCT-BRCT domain packing, which occurs through a hydrophobic interface consisting of the α₂ helix of the N-terminal BRCT and α₁′ and α₃′ helices of the C-terminal BRCT.18 Mutations at this hydrophobic interface in BRCA1 (M1775R and A1708E) are linked to breast and ovarian cancer, highlighting the requirement for an intact tandem BRCT interface for normal function.5,6 This conserved BRCT-BRCT domain packing has since been observed in numerous BRCT repeat structures such as in Mediator of DNA damage checkpoint protein 1 (MDC1), BRCA1-associated RING domain protein 1 (BARD1), Topoisomerase IIβ binding protein 1 (TopBP1) BRCT7/8, S. pombe Crb2 and S. pombe Brc1, and their functional role in phospho-peptide binding is well established.19–25 However, variations in tandem BRCT domain structure also exist. For example, the tandem BRCT domains of DNA Ligase IV (LigIV) do not pack together and are separated by a significantly longer inter-BRCT linker, which ultimately ensures its unique mode of recognition for X-ray repair cross-complementing protein 4 (XRCC4).26–28 Recently, a triple BRCT domain module was confirmed in TopBP1 BRCT0/1/2, suggesting even higher order of tandem BRCT domain architecture. Interestingly, the packing of BRCT0/BRCT1 and BRCT1/BRCT2 at the interface are significantly different from canonical tandem BRCT domains.29,30 This unconventional BRCT domain interface is likely a consequence of the shorter inter-BRCT linker regions, suggesting a role for the inter-BRCT linker in driving the packing of tandem BRCT domains. Whereas longer linker lengths ranging from 30–60 amino acids permit the parallel juxtaposition characteristic of canonical tandem BRCT domain packing, shorter linkers restrict the packing of adjacent BRCT repeats such that they interact in a twisted orientation.

The diversity in BRCT domain architecture extends to other functional domains, although the significance of these hetero-domain modules is still unclear. A single BRCT domain is packed alongside an N-terminal fibronectin type III (FN3) domain in S. cerevisiae Chs5p, a component of the exomer coat complex. Recent studies in yeast suggest that both the FN3 and BRCT domain act as a single module and are both required for localization and cargo delivery function.31 The tandem BRCT repeats in the DNA double-stranded break repair protein Nijmegen breakage syndrome protein 1 (NBS1) are coupled to an N-terminal FHA domain, another phospho-peptide binding domain.32,33 This unique FHA-BRCT-BRCT domain setup may present an intriguing platform for cross-talk between the two adjacent functional phospho-peptide interacting modules.

Phosphorylation-Dependent Interactions

Since the original discovery of single and tandem BRCT domains as phospho-peptide binding modules,9,10 the ability for tandem BRCT domains to recognize phospho-peptides has been well characterized, both functionally and structurally, in a number of proteins. However, the basis of phospho-peptide recognition by single BRCT domains remains unclear.

Single BRCT domains.

The implication of phospho-peptide binding mediated by single BRCT domains was initially demonstrated in vitro using a degenerate peptide library. Using single BRCT domains in Terminal deoxynucleotidyl transferase (TDT), S. cerevisiae REV1 and Lig3, Yu et al. showed that these domains preferentially interact with a phospho-peptide library compared with the non-phosphorylated counterpart and that TopBP1 BRCT6 and hFCP1 BRCT specifically binds to cognate E2F1 and RNA polymerase II phospho-peptides, respectively.10 However, subsequent work on FCP1 and TopBP1 single BRCT domains provides conflicting evidence for their roles as phospho-peptide binding modules. Although studies in S. pombe and A. thaliana indicate that the FCP1 BRCT domain is important for FCP1 C-terminal domain (CTD) phosphatase activity,34,35 structural and biochemical data from Ghosh et al. revealed that the BRCT phosphate-binding pocket is dispensable for FCP1 function.36 Rather as part of a multi-domain complex, the BRCT domain may have an indirect role in supporting critical residues in the neighboring catalytic N-terminal FCP1 homology (FCPH) domain. The interaction between TopBP1 BRCT6 and E2F1 phosphorylated at S31 represses E2F1 activity in response to DNA damage.37,38 Conversely, the crystal structure of TopBP1 BRCT6 suggests that it lacks a functional phosphate-binding pocket and is unlikely to foster a stable phospho-peptide interaction on its own.39

Given the less extensive binding surface, one might suspect that the recognition of phospho-peptides by single BRCT domains could be weaker than tandem BRCT domains. Single BRCT domain interactions may participate in more dynamic signaling processes or may require additional contacts provided within oligomeric complexes and/or from other protein domains, to stabilize these interactions. Clearly, more structural characterization will be required to elucidate the mechanism of single BRCT domain phospho-peptide interactions.

Tandem BRCT domains.

Through biochemical and structural studies of a number of BRCT repeats with their cognate phospho-peptides, we now understand not only the conserved mode of phospho-peptide binding in BRCT repeats, but also the differences that allow tandem BRCT domains to achieve such a high degree of specificity for a wide range of phospho-peptide targets.

Crystal structures of the BRCA1, MDC1, S. pombe Brc1, S. pombe Crb2 and TopBP1 BRCT7/8 BRCT repeats bound to their respective phospho-peptides reveal a conserved bipartite recognition involving two distinct pockets: a pS/pT binding pocket in the N-terminal BRCT and a secondary specificity pocket at the BRCT-BRCT interface that generally provides selectivity for a hydrophobic residue at the +3 position in the phospho-peptide with respect to the pS (Fig. 1A).16,21–24,40–44 Two different classes of phosphate binding pockets have been identified to date. The phosphate-binding pockets of BRCA1 and MDC1 serve as prototypes for the first class, which are selective for pS (Fig. 1B; left part). In this class, the phosphate moiety interacts with two side chains (S1655/T1898, K1702/K1936 in BRCA1/MDC1, respectively) and a main chain NH (G1656/G1899 in BRCA1/MDC1). The recent investigations of the interaction of TopBP1 BRCT7/8 with phosphorylated BACH1 revealed a second class of phosphate binding pocket with additional specificity for pThr peptide targets.45 In addition to interactions that are conserved with the class I pocket that involve S1273, K1317 and the main chain NH of S1274, the phosphate group is also recognized by the side chain guanidinium of R1280 (Fig. 1B; right part). The combination of differences in the peptide backbone orientation and coordination of the pThr residue, likely facilitated by additional interactions with the conserved TopBP1 R1280, allows incorporation of the γ-methyl group without steric hindrance in the pocket.22

The conserved minimal phospho-peptide sequence bound by BRCT repeats consists of four residues (pS/pT-X-X-X), in which the +3 residue dictates the majority of the binding specificity.46 The +3 residue sits in a cleft situated at the hydrophobic interface between the two BRCT domains, and conserved residues at the BRCT-BRCT interface determine the binding specificity for the +3 residue. In particular, a conserved arginine residue (R1699/R1933 in BRCA1/MDC1, respectively) has a pivotal role in recognizing the peptide main chain at the +3 position (Fig. 1C). In BARD1, the conserved arginine is substituted for a serine residue (S616), suggesting that the phosphopeptide binding properties of BARD1 may be different.20,25,47,48 Interestingly, both BRCA1 and MDC1 have peptide targets that are internal (BACH1/CtIP/ACCA for BRCA1, MMSET for MDC1)9–11,49,50 or terminate (Abraxas for BRCA1, γH2AX for MDC1)51,52 at the +3 position. Peptide binding studies have shown that the presence of a C-terminal carboxylate increases binding affinity to BRCA1 and MDC1, although the differences between the carboxy and amidated phospho-peptides are more profound for MDC1.19,24,46 Comparison of the peptide-bound structures suggests that positioning of the +3 side chain determines the angle of incidence of the C-terminal carboxylate, which is more favorable in MDC1, and thereby tunes the degree of preference for the carboxylate.46

Alterations in the hydrophobic packing interface between BRCT repeats may also present a means to alter phosphopeptide specificity. Due to the extent of hydrophobic interactions between α₂-α₁′-α₃′ helices, the BRCT domain pair is typically held rigid and only show small deviations in domain movement upon peptide binding. This is perhaps reflected in the absolute conservation of the cognate phospho-peptide motif sequence for BRCA1, where all validated binding partners contain the consensus motif, pS-P-T-F.6,9–11,13,14,49,52 In contrast, the significantly smaller hydrophobic interface in TopBP1 BRCT7/8 allows for a dramatic domain rotation (23°) at the BRCT-BRCT interface upon phospho-peptide binding in order to shape the larger +3/+4 pocket, as well as the phosphate-binding pocket to a lesser extent.22 Since TopBP1 BRCT7/8 has also been implicated in binding another TopBP1 phospho-peptide motif with a divergent sequence (pS-N-L-Q-W-P-S),38 the plasticity at the BRCT-BRCT interface in TopBP1 BRCT7/8 could potentially cater to a larger variety of peptide targets.

In another layer of complexity, tandem BRCT domain phospho-peptide binding may also be further regulated by interactions with neighboring domains, such as in the NBS1 FHA-BRCT-BRCT hetero-domain module. Structural studies of S. pombe NBS1 with the Ctp1 phospho-peptide reveal a structural rearrangement in the tandem BRCT domain interface, reminiscent of TopBP1 BRCT7/8 but in an orthogonal direction.33 The NBS1 conformational change is triggered by an arginine switch located at the FHA-BRCT1 domain interface that is activated upon FHA domain binding to its phospho-peptide. Since both the NBS1 FHA domain and BRCT repeat are capable of binding to the phosphorylated MDC1 SDT motifs,32,53–56 interaction from the FHA domain may regulate the phospho-peptide binding ability of the adjacent BRCT repeat by initiating a tandem BRCT domain structural rearrangement.

There is also the rare scenario where the putative phosphate-binding pocket exists in the C-terminal BRCT rather than the N-terminal BRCT domain of a BRCT repeat, such as in TopBP1 BRCT4/5 and PTIP BRCT1/2. TopBP1 BRCT5 is required for TopBP1 localization to IR-induced foci57 and interacts with phosphorylated SDTD motifs in MDC1 at stalled replication forks.58 In addition, TopBP1 BRCT4/5 has been shown to bind 53BP1, an interaction that is important for 53BP1-mediated checkpoint control in response to DSBs in G₁.59 It is likely that the phosphatebinding pocket in BRCT5 is involved in these interactions. However, the absence of a downstream +3 hydrophobic pocket implies a smaller peptide binding surface and potentially a weaker interaction compared with canonical tandem BRCT domain interactions.

The structure of the N-terminal region of TopBP1 shows an unexpected packing of BRCT1/2 with another BRCT domain at the N-terminus, forming a triple BRCT repeat module. This region interacts with Rad9, a component of the trimeric Rad9-Hus1-Rad1 (9-1-1) clamp, which is a prerequisite for stimulating ATR activity and subsequent DNA replication checkpoint activation.60,61 Interestingly, both BRCT1 and BRCT2 contain putative phosphate-binding pockets, but peptide binding data suggests that only the BRCT1 pocket is responsible for the phosphorylation-dependent interaction with Rad9.29,30 The presence of a distal functional phosphate-binding pocket in BRCT2, as indicated by the coordination of a sulfate ion in the crystal structure,29 may imply an intriguing basis for the recognition of other phospho-peptide targets for TopBP1 BRCT0/1/2. In addition, deletion of BRCT0 only slightly decreases binding to the Rad9 phospho-peptide, suggesting that BRCT0 may only have a minor role in phospho-peptide interactions.30 Indeed, further work will be needed to elucidate the potential significance of BRCT0 in the triple BRCT domain.

Phosphorylation-Independent Protein Interactions

Single BRCT domains.

Single BRCT domains have shown the ability to interact with single BRCT domains in other proteins via interfaces that are distinct from the hydrophobic BRCT-BRCT interface in tandem BRCT domains. Perhaps the best examples are in interactions involving the two single BRCT domains in XRCC1. The interaction between the C-terminal BRCT (BRCT2) of XRCC1 and Lig3 BRCT domain is well established in short-patch base excision repair.62–65 Although a crystal structure is unavailable (see Note added in proof), models have been proposed for the XRCC1-Lig3 complex based on a number of studies. A combination of the homodimer crystal structure of XRCC1 BRCT2 and mutational analysis of key residues in α₁ helix provides a structural model for a dimerization interface.15,66 However, it remains unclear whether this interface is responsible for homo or hetero-dimerization and other data support a hetero-tetrameric complex, comprising of a homodimer of XRCC1 BRCT2 interacting with a homodimer of Lig3 BRCT domains.64,67,68 A less characterized single BRCT-BRCT interaction involves the N-terminal BRCT of XRCC1 (BRCT1) and PARP-1 BRCT, which negatively regulates PARP-1 activity.67,69 Like XRCC1 BRCT1, TopBP1 BRCT6 also interacts with the PARP-1 BRCT domain,70 providing evidence for a common mode of interaction mediated by the BRCT domain of PARP-1. The recent crystal structure of the N-terminal BRCT domain of Microcephalin (MCPH1/BRIT1) provides insight for another potential protein-protein interaction surface that is required for MCPH1 BRCT1-dependent centrosomal localization and prevention of abnormal chromosome condensation.71,72 The conserved phosphate-binding residues (S1655 and K1702 in BRCA1) are absent in MCPH1 (E14 and T59, respectively) and the pocket is instead replaced by a larger, more acidic surface that is conserved in MCPH1 homologs. Mutations at this pocket also correlate to defective regulation of chromosomal condensation. Thus, the distinct surface that substitutes for the phosphate-binding pocket in MCPH1 BRCT1 may represent another contact surface for protein-protein interactions mediated by single BRCT domains.

Tandem BRCT domains.

Tandem BRCT domains can also interact with other proteins independent of phosphorylation. As demonstrated in crystal structures involving the BRCT repeats of LigIV, S. pombe Crb2 and 53BP1 in complex with their respective protein binding partners, a common theme observed in each of these interactions is the presence of a β-hairpin motif in the inter-BRCT linker, which is responsible for extensive contacts at the protein binding interface.73

Interactions between the 53BP1 BRCT repeats and the p53 DNA binding domain (DBD) enhance the transactivation potential of p53.74–76 In the 53BP1 BRCT1/2-p53 DBD complex crystal structure, the 53BP1 BRCT-BRCT linker adopts an extended β-hairpin loop that, together with the α₄ helix, contacts the conserved L₂ and L₃ loops of the p53 DBD (Fig. 2A).77,78 A distal salt bridge between the 53BP1 BRCT1 α₃ helix residue D1833 and p53 R181 also makes a minor contribution to the overall binding interface.

The S. pombe Crb2 checkpoint function relies on the BRCT repeat-mediated homodimerization of Crb2.79,80 Biochemical and structural evidence for Crb2 BRCT1/2 further confirms the oligomeric status of the Crb2 tandem BRCT domains.21 At the interface of the Crb2 dimer, the extended β-hairpin loop of the Crb2 BRCT-BRCT linker from each protomer participates in an extensive symmetrical hydrogen bonding network (Fig. 2B). Mutations in the β-hairpin (S666R and C663R) also confer checkpoint functional defects,81,82 which further supports the biological role of Crb2 homodimerization.

The LigIV-XRCC4 interaction is critical for non-homologous end joining (NHEJ) and in V(D)J recombination.83–86 Interestingly, the tandem BRCT domains of LigIV contain an unusually long linker region, and packing of the tandem BRCT domains is not observed. The crystal structures of the human and yeast ortholog of XRCC4 in complex with the tandem BRCT domains of LigIV provide a possible rationale for the unique LigIV BRCT repeat structure.26–28 Besides contacts mediated by the β-hairpin of the linker region in LigIV and α₁ helix of LigIV BRCT2 on opposite ends of the coiled-coil tails of the XRCC4 homodimer, the extended LigIV linker forms a distinct helix-loop-helix (HLH) clamp structure that wraps around the linear XRCC4 coiled-coil tails to stabilize the interaction (Fig. 2C).

The inter-BRCT domain linker has evolved a role in facilitating protein-protein interactions in addition to its role in dictating the orientation and packing of tandem BRCT repeats. Whereas shorter linker lengths affect the overall tandem BRCT domain structure by restricting the BRCT packing orientation, as evident in the TopBP1 BRCT0/1/2 structure,29,30 the flexibility of longer linkers to form common structure motifs, such as a β-hairpin and/or HLH, can be utilized for protein-protein interactions.

DNA Binding

Although DNA-binding activity has been proposed for a number of different single and tandem BRCT domains, there is still uncertainty as to the specificity of these interactions for different types of DNA structures and their relevance to the biological functions of BRCT proteins. Recent biochemical and structural studies of the replication factor C (RFC) BRCT domain have provided molecular insight into how a subgroup of single BRCT domains are selective for DNA ends.

Single BRCT domains.

The initial studies on the DNA binding activity of BRCT domains by Yamane et al.87,88 demonstrated that the single BRCT domains in TopBP1 and XRCC1 can bind DNA strand breaks and termini in vitro, although the relevance of these interactions to the cellular function of TopBP1 and XRCC1 is still unclear. In contrast, significant progress has been made to characterize the DNA binding activity in the BRCT domain of Replication factor C (RFC) p140, a subunit of the five protein RFC complex that functions in DNA replication and repair.89–91 The RFC BRCT domain selectively recognizes the 5′-phosphate DNA terminus within a 3′ overhanging DNA end.92,93 Structural investigations by NMR spectroscopy shed light on the mechanism of DNA recognition, and involves not only the phosphate-binding pocket but an α-helix N-terminal to the BRCT domain that inserts into the major groove of the DNA (Fig. 3A and B).93,94 Analogous to the recognition of phosphopeptides, conserved residues in the phosphate-binding pocket of RFC (T415, G416, K458, R423) coordinate the recessed 5′-phosphate (Fig. 3C). These conserved residues are also absolutely conserved with the BRCT domain of the bacterial NAD⁺-dependent DNA ligase (LigA), which is highly similar to RFC (53% identity) and constitutes a distinct BRCT subgroup with RFC and PARP-1.7 As in RFC, the BRCT domain of LigA has DNA binding activity and mutations in these conserved residues (T599, G600) markedly reduce DNA binding and ligation activity.95,96 In light of this evidence, the mechanism of DNA recognition by the RFC BRCT domain may represent a conserved mode of DNA recognition for this subset of single BRCT domains.

Tandem BRCT domains.

As with the single BRCT domains of TopBP1 and XRCC1, tandem BRCT domains in BRCA1 and TopBP1 can also bind DNA strand breaks and ends in vitro.87,88 In the case of BRCA1, BRCT repeat-mediated DNA-binding activity is enhanced by formation of the BRCA1-BARD1 heterodimer. 97 Interestingly, TopBP1 BRCT7/8 also displays unique specificity for damaged DNA containing bulky lesions. Using an in vitro partially reconstituted checkpoint system, Choi et al. demonstrated that a C-terminal fragment of TopBP1, containing BRCT7/8 and the ATR-activation domain (AAD), preferentially binds benzo[a]pyrene diol epoxide (BPDE)-treated DNA and stimulates ATR activity,98 which is critical for DNA replication checkpoint activation in response to DNA replication stress. TopBP1 BRCT7/8 binding also favors longer stretches of DNA and, contrary to the previous studies by Yamane et al.87,88 is not specific for DNA ends. The implication of BRCT domains directly detecting damaged DNA offers intriguing possibilities that proteins with BRCT domains can act as DNA damage sensors in addition to transducers of DDR signaling.

Poly(ADP-Ribose) Binding

Single BRCT domains.

Although still poorly understood, single BRCT domains have been implicated in binding poly(ADP-ribose) (PAR), a post-translational modification catalyzed by poly(ADP-ribose) polymerases in DDR signaling. A consensus PAR-binding motif is found in TopBP1 BRCT6 and the N-terminal BRCT of XRCC1,99 although there is more evidence supporting XRCC1 BRCT1 as a validated PAR-binding module. XRCC1 BRCT1 preferentially recognizes ribosylated PARP-1 and a peptide of the XRCC1 PAR-binding sequence binds to PAR in vitro.69,99–102 In contrast, the PAR?binding activity of XRCC1 does not appear to be conserved in TopBP1, as indicated by the absence of PAR-binding by TopBP1 BRCT6 in vitro.39 Further comparison of the structures of TopBP1 BRCT6 and XRCC1 BRCT1 reveal that their respective PAR-binding motifs are not structurally conserved and that the PAR-binding motif in TopBP1 BRCT6 is partially buried in the protein core (Fig. 4).39

Conclusions

BRCT domains have proven to be remarkably versatile binding modules. Starting from the single BRCT domain building block, the ability to incorporate different functions by varying or adding different structural features grants unprecedented advantages to establish regulation and specificity for various targets such as phospho-peptides, proteins, DNA and PAR (Fig. 5). Although the phospho-peptide and phosphorylation-independent protein binding functions of tandem BRCT domains has been well characterized both structurally and functionally, less is clear about the particular role of single BRCT domains, aside from the DNA binding activity of the RFC BRCT domain. Indeed, further work is needed to elucidate the mechanisms underlying the multitude of functions associated with BRCT domains.

Note added in proof

The structural basis for XRCC1-Lig3 BRCT interactions has recently been revealed in a crystallographic study (Cuneo MJ, Gabel SA, Krahn JM, Ricker MA, London RE. The structural basis for partitioning of the XRCC1/DNA ligase III-a BRCT-mediated dimer complexes. Nucleic Acids Res 2011; In press; PMID: 21652643.

Figures and Tables

Figure 1 Phospho-peptide binding of tandem BRCT domains. (A) Conserved phospho-peptide recognition of BRCA1 BRCT1/2 consists of a phosphate-binding pocket (blue) and +3 binding pocket (brown). The surface representation of BRCA1 BRCT1/2 is shown and the BACH1 peptide pS and +3 side chains are represented as sticks and labeled. (B) pS coordination by BRCA1 (blue) and MDC1 (beige) (left part) and pT coordination by TopBP1 BRCT7/8 (right part). Phosphate contact residues for BRCA1/MDC1 (top/bottom) and TopBP1 are labeled. Equivalent hydrogen bonding and electrostatic interactions are indicated by black dotted lines and additional contacts by TopBP1 R1280 are shown as red dotted lines. (C) +3 phospho-peptide main chain recognition at the BRCT-BRCT interface in BRCA1 (blue), MDC1 (beige) and BARD1 (pink). The equivalent Arg residue in BRCA1/MDC1/BARD1 (top/middle/bottom) is labeled.

Figure 2 Tandem BRCT domain linker-mediated protein interactions. The inter-BRCT linker is shown in orange and contact regions are highlighted in red. Secondary structure motifs are labeled. (A) Structure of 53BP1 BRCT repeats (gray) in complex with p53 DBD (teal). (B) Structure of the S. pombe Crb2 homodimer. Each protomer is labeled (A and B). (C) LigIV (gray) in complex with XRCC4 homodimer (light cyan).

Figure 3 Model of the RFC p140 BRCT domain in complex with dsDNA. (A) Representative ensemble by Kobayashi et al.94 based on the four best structures from HADDOCK. The N-terminal α-helix is shown in purple and the 5′-phosphate is represented as a gray sphere. (B) Surface representation of RFC BRCT domain. Binding surfaces with the DNA major groove (purple) and 5′-phosphate (orange) are marked. (C) Conserved phosphate-binding pocket of RFC BRCT domain, as represented by one model in the ensemble. The 5′-phosphate group and contact residues are represented as sticks and labeled.

Figure 4 The consensus PAR-binding motif is not conserved in TopBP1 and XRCC1. The PAR-binding motif (cyan) is mapped in the structures of TopBP1 BRCT6 (left) and XRCC1 BRCT1 (right).

Figure 5 Summary of BRCT domain variation and function. Each feature is marked and described.

Table 1 BRCT domain architecture found in proteins

	Homo-Domain Assembly
Domain Architecture	Examples	PDB ID *
BRCT	XRCC1, Lig3, PARP-1, TDT, ECT2, DNA Polµ, REV 1, RFC, PES1, DNA Polλ, MCPH1, FCP1, TopBP1	1CDZ, 1IN1, 2COK, 2COE, 2COU1, 2HTF, 2EBW, 2EBU, 2EP8, 2JW5, 2WT8, 3EF02, 3JVE
BRCT-BRCT	BRCA1, 53BP1, MDC1, BARD1, S. pombe Crb2, LigIV, S. pombe Brc1, TopBP1	1JNX, 1KZY, 2ADO, 2NTE, 2VXB, 3II 6, 3L40, 3AL2
BRCT-BRCT-BRCT	TopBP1	2XNK
	Hetero-Domain Assembly
Domain Architecture	Examples	PDB ID
FN3-BRCT	S. cerevisiae Chs5p	Not available
FHA-BRCT-BRCT	NBS1	3I0N/3HUE2

* Unless otherwise specified, listed structures are of the human proteins.

1 Structure of the mouse homolog.

2 Structure of the S. pombe homolog.

Acknowledgments

We thank members of the Glover lab for their helpful discussion and comments on the manuscript. This work was supported by grants from the Canadian Cancer Society, National Institutes of Health, and the Howard Hughes Medical Institute International Scholar Program.