REVIEW / SYNTHÈSE

Structural basis for phosphorylation-dependentsignaling in the DNA-damage response1

R. Scott Williams, Nina Bernstein, Megan S. Lee, Melissa L. Rakovszky,Diana Cui, Ruth Green, Michael Weinfeld, and J.N. Mark Glover

Abstract: The response of eukaryotic cells to DNA damage requires a multitude of protein–protein interactions thatmediate the ordered repair of the damage and the arrest of the cell cycle until repair is complete. Two conserved proteinmodules, BRCT and forkhead-associated (FHA) domains, play key roles in the DNA-damage response as recognitionelements for nuclear Ser/Thr phosphorylation induced by DNA-damage-responsive kinases. BRCT domains, first identifiedat the C-terminus of BRCA1, often occur as multiple tandem repeats of individual BRCT modules. Our recent structuraland functional work has revealed how BRCT repeats recognize phosphoserine protein targets. It has also revealed asecondary binding pocket at the interface between tandem repeats, which recognizes the amino-acid 3 residues C-terminalto the phosphoserine. We have also studied the molecular function of the FHA domain of the DNA repair enzyme,polynucleotide kinase (PNK). This domain interacts with threonine-phosphorylated XRCC1 and XRCC4, proteinsresponsible for the recruitment of PNK to sites of DNA-strand-break repair. Our studies have revealed a flexible modeof recognition that allows PNK to interact with numerous negatively charged substrates.

Key words: BRCA1, BRCT, PNK, FHA, polynucleotide kinase, breast cancer, phosphopeptide-protein interactions,DNA damage response.

Résumé : La réponse des cellules eucaryotes à des dommages d’ADN requiert une multitude d’interactions protéine–protéinequi règlent une réparation ordonnée du dommage et l’arrêt du cycle cellulaire jusqu’à ce que la réparation soit complète.Deux modules protéiques conservés, les domaines BRCT et FHA, jouent un rôle clé dans la réponse aux dommages àl’ADN en tant qu’éléments de reconnaissance lors de la phosphorylation sur Ser/Thr induite par des kinases répondantaux dommages à l’ADN. Les domaines BRCT,d’abord identifiés à l’extrémité C-terminale de BRCA1, existent souventsous forme de répétitions en tandem de modules BRCT individuels. Nos travaux récents sur les relations structure/fonctionde BRCT ont montré comment les répétitions BRCT reconnaissent les protéines cibles à phosphosérine et ont révélé laprésence d’une poche de liaison secondaire située à l’interface des répétitions en tandem qui reconnait les 3 résidusd’acides aminés localisés en position C-terminale de la phosphosérine. Nous avons aussi étudié la fonction moléculairedu domaine FHA d’une enzyme de réparation d’ADN, la PNK. Ce domaine interagit avec XRCC1 et XRCC4 phos-phorylés sur thréonine, les protéines responsables du recrutement de PNK aux sites de réparation de bris à l’ADN. Nosétudes ont révélé l’existence d’un mode flexible de reconnaissance qui permet à PNK d’interagir avec de nombreuxsubstrats chargés négativement.

Mots clés : BRCA1, BRCT, PNK, FHA, polynucléotide kinase, cancer du sein, interactions phosphopeptide-protéine,réponse aux dommages à l'ADN.

[Traduit par la Rédaction] Williams et al. 727

Biochem. Cell Biol. 83: 721–727 (2005) doi: 10.1139/O05-153 © 2005 NRC Canada

721

Received 11 April 2005. Revision received 23 August 2005. Accepted 13 September 2005. Published on the NRC Research PressWeb site at http://bcb.nrc.ca on 29 October 2005.

R.S. Williams, N. Bernstein, M.S. Lee, M.L. Rakovszky, D. Cui, R. Green, and J.N.M. Glover.2 Department of Biochemistry,University of Alberta, Edmonton, AB T6G 2H7, Canada.M. Weinfeld. Experimental Oncology, Cross Cancer Institute, and Department of Oncology, University of Alberta, Edmonton, ABT6G 1Z2, Canada.

1This paper is one of a selection of papers published in this Special Issue, entitled CSBMCB – Cellular dynamics, and hasundergone the Journal’s usual peer review process.

2Corresponding author (e-mail: mark.glover@ualberta.ca).

Introduction

Cells have evolved to deal with a bewildering complexityof DNA damage, from relatively small base lesions to single-and double-strand breaks in the DNA backbone. In general,repair of DNA damage involves multiple steps and severalenzymatic activities. It is believed that the DNA intermedi-ates generated enroute to repair are often more mutagenicthan the original damage, and it appears that scaffold pro-teins are used to efficiently shuttle these DNA intermediatesbetween the various repair enzymes that are required at suc-cessive steps in the repair process. In addition, specific cell-cycle checkpoints can effectively delay progression throughthe cell cycle in response to DNA damage, allowing the celltime to repair damage before replication.

Recent work has uncovered unique protein-signalingmodules that are specific to DNA-damage response. Here,we review recent structural and functional work from ourlaboratory that reveals the fundamental principles of phospho-protein recognition used by BRCT and forkhead-associated(FHA) proteins. Together with the 14-3-3 family (Yaffe andSmerdon 2001), these proteins seem to regulate a network ofphosphorylation-dependent protein–protein interactions thatare key to the regulated cellular response to DNA damage.

Our studies reveal the structural basis for the reason specificmutations in BRCA1 BRCT repeats are associated with breastand ovarian cancer, and the findings could be used to de-velop new cancer therapies that modulate these interactions.

BRCT repeats

BRCT repeats were first discovered at the extreme C-terminus of the breast cancer - associated protein BRCA1,but were later shown to exist in a large family of proteinsthat are linked to the cellular response to DNA damage(Koonin et al. 1996; Callebaut and Mornon 1997; Bork et al.1997; Glover et al. 2004) (see http://www.sanger.ac.uk/cgi-bin/Pfam/getacc?PF00533 for the Pfam listing of the BRCTfamily). The BRCT repeats in BRCA1 are essential for thetumor-suppressor function of the protein; protein truncationand missense variants within the BRCT have been shown tobe associated with human breast and ovarian cancers (Wil-liams et al. 2003).

We determined the X-ray crystal structure of the 2 BRCTrepeats of BRCA1 (Fig. 1) (Williams et al. 2001). The struc-ture revealed that the 2 repeats adopt similar folds and packtogether in a head-to-tail manner. This interaction is stabi-lized by the packing of a single helix from the N-terminal

© 2005 NRC Canada

722 Biochem. Cell Biol. Vol. 83, 2005

Fig. 1. Structures of the BRCA1 BRCT domain and interactions with phosphopeptides. (Top panel) Structure of the BRCA1 BRCTdomain bound to a pSer-x-x-Phe peptide target. The N- and C-terminal BRCT repeats are shown in blue and green, the inter-repeatlinker is shown in orange, the bound peptide is in yellow. Residues important for peptide recognition are shown as sticks. (Bottompanels) Details of the Phe +3 binding pocket in the wild-type structure (left) and the M1775R variant (right). Note that the M1775Rand V1809F structures were established in the absence of a bound peptide. The yellow peptide shown in the M1775R panel is overlaidfrom the wild-type structure.

repeat against a pair of helices from the C-terminal repeatand the inter-repeat linker. Although this structure alone didnot reveal the molecular function of the protein, we wereable to use it to predict the effects of specific mutations, theunderlying hypothesis being that the function would be dis-rupted by mutations that disrupt the structure of the domain.We also developed a simple proteolytic assay to directly testthe conformational stability of BRCA1 BRCT variants(Williams et al. 2003).

Although most of the cancer-causing missense variantscause a dramatic destabilization of the BRCT fold, some ofthe mutations lead to a more modest increase in proteolyticsusceptibility. One of these mutations, M1775R, is particu-larly interesting; it was one of the first characterized BRCTmutations to be linked to cancer (Miki et al. 1994; Futreal etal. 1994). We were able to crystallize and determine thestructure of this variant, which revealed a subtle rearrange-ment of side chains around the site of the mutation, and aperturbation of the surface features of the protein (Fig. 1)(Williams and Glover 2003). However, a key question re-mained: Is the defect associated with this mutation due tothe somewhat reduced stability of this variant, or does the al-teration of the protein surface perturb an important interac-tion surface?

Late in 2003, a major breakthrough in our understanding

of BRCT function came when work from the laboratories ofJunjie Chen (Yu et al. 2003) and Mike Yaffe (Manke et al.2003) revealed that the tandem BRCT repeats of BRCA1function as phosphopeptide-binding modules. The BRCA1BRCT is highly selective for the sequence pSer-X-X-Phe.These interactions have been shown to mediate the associa-tion between BRCA1 and the DNA helicase BACH1, whichis essential for the correct functioning of the G2/M cell-cyclecheckpoint (Cantor et al. 2001; Yu et al. 2003), and betweenBRCA1 and the transcriptional corepressor CtIP (Yu andChen 2004). These authors also showed that BRCT repeatsfrom other proteins also functioned as phosphopeptide-bindingmodules, indicating that phosphopeptide binding might be aconserved function for BRCT repeats in DNA-damage re-sponse.

To understand how the BRCA1 BRCT recognizes its spe-cific phosphopeptide target, we established that the structureof the BRCA1 BRCT bound to a high-affinity peptide, de-rived from an in vitro peptide-selection experiment that con-tained the pSer-X-X-Phe motif (Figs. 1 and 2) (Williams etal. 2004). Similar structures have also been found by othergroups (Shiozaki et al. 2004; Clapperton et al. 2004). Thestructure revealed a phosphoserine-binding pocket in the N-terminal BRCT repeat, and a phenylalanine-binding pocketin a groove formed at the interface between the 2 repeats.

© 2005 NRC Canada

Williams et al. 723

Fig. 2. Sequence alignment of BRCT repeat proteins. The amino-acid-sequence alignment of the tandem BRCT repeats from BRCA1,MDC1 PTIP, and BARD1 is shown, along with the secondary structure from the BRCA1 BRCT crystal structure. The residues thatcontact the pSer are shaded blue, and the residues that form the Phe-binding pocket are marked with red circles. Missense mutationsthat have been assayed for interactions with the peptide are displayed above the sequence. Residues involved in inter-repeat BRCTinteractions are boxed. (From Williams et al. 2004, adapted with permission of Nat. Struct. Mol. Biol., Vol. 11, p. 520, © 2004 Na-ture Publishing Group.)

Key residues that constitute the phosphoserine pocket in-clude Ser 1655, Gly 1656, and Lys 1702, which all directlyrecognize the phosphate moiety. Thr 1700 also plays an im-portant role in phosphate recognition; it hydrogen bondswith Ser 1655, keeping the serine hydroxyl in a rigid orien-tation appropriate for phosphate recognition. This [Ser-Gly…Thr-X-Lys] motif is conserved in a variety of otherBRCT repeats, suggesting that many of these proteins willalso bind phosphoserine-containing peptides (Fig. 2) (Gloveret al. 2004; Williams et al. 2004). Indeed, a number of otherBRCT-containing proteins, including PTIP, 53BP1, andBARD1, have demonstrated phosphopeptide-binding activity(Rodriguez et al. 2003; Yu et al. 2003; Manke et al. 2003).

To directly test the phosphoserine-binding capacity of ourset of 25 BRCT missense variants, we assayed the ability ofa pSer-X-X-Phe peptide to specifically pull down in vitrotranscribed–translated BRCT variants, and compared thefindings with those of a nonphosphorylated control. Our re-sults demonstrated that the structural integrity of the BRCTdomain was required for phosphopeptide recognition; noneof the missense variants that highly destabilized the proteinfold specifically bound the pSer peptide. The results con-firmed the importance of the residues of the phosphate-recognition pocket; mutation of any of these residues toalanine completely destroyed the ability of the protein tobind phosphopeptide (Williams et al. 2004).

The Phe at the +3 position relative to the pSer is bound ina deep groove at the interface between the 2 repeats (Fig. 1).This interaction explains the fact that both BRCT repeats areneeded for peptide binding; loss of part of the C-terminal re-peat is associated with hereditary breast cancer. BRCT do-mains are most commonly found as tandem repeats in otherproteins involved in the DNA-damage response, and sequenceanalysis suggests that the head-to-tail packing of the BRCTrepeats in BRCA1 is conserved in other repeats (Williams etal. 2001; Glover et al. 2004). Thus, it is likely that this sec-ondary recognition groove at the repeat interface is con-served in other BRCT repeat proteins. However, the lack ofsequence conservation of residues that line the groove sug-gests that the peptide-binding specificity of other BRCT-repeat-binding proteins may be different.

The importance of several of the residues that line thephenylalanine-binding pocket were also tested using thepull-down assay (Williams et al. 2004). Mutations of manyof these residues were found to disrupt specific phospho-peptide interactions. For example, mutation of Arg 1699,which interacts with the backbone of the phenylalanine, toeither Trp or Gln, completely abrogated binding. The cancer-associated mutation M1775R also resulted in a completeloss of peptide binding. Interestingly, M1775 lies at the bot-tom of the inter-repeat groove and makes close van derWaals contact with the phenylalanine. Superposition of theM1775R mutant structure on the structure of the BRCT–peptide complex reveals that the substituted arginine sidechain occupies the phenylalanine-binding pocket, suggestingthat this mutant is impaired when it comes to binding pSer-X-X-Phe targets (Fig. 1) (Williams et al. 2004).

Significant advances have been made in the understandingof the molecular function of the BRCA1 BRCT repeats, andstudies have allowed us to understand the molecular defectsassociated with a number of previously uncharacterized se-

quence variants in this region of the protein. However, thefunction of phosphoserine binding in the context of the in-tact protein remains to be determined. Do these interactionsfacilitate rearrangements in the overall structure and organi-zation of the BRCA1 complex? Do these interactions affectthat ubiquitin ligase activity of BRCA1? Answers to thesequestions will not only help to detail the molecular mecha-nisms that underlie hereditary breast cancer, they will alsohelp to define mechanisms that regulate the cellular responseto DNA damage.

FHA domains

FHA domains are ~80–100 amino-acid domains that playkey signaling roles in multiple cellular processes, such assignal transduction, transcription, vesicular transport, andprotein degradation, in addition to the DNA-damage re-sponse (Durocher and Jackson 2002). The structures of anumber of FHA domains are now known, both free and incomplex with phosphopeptide targets, and studies of theFHA domains from the DNA-damage-response proteinsRad53 and Chk2 have been particularly informative in re-vealing the peptide-binding specificities of these proteins. Ingeneral, FHA proteins differ from BRCT domains in thatthey are highly specific for phosphothreonine, as opposed tophosphoserine-containing targets.

However, like BRCT domains, the Rad53 and Chk2 FHAdomains exhibit marked selectivity for specific side chain 3residues C-terminal to the phosphorylated residue. In thecase of the N-terminal FHA domain from Rad53, the speci-ficity is for negatively charged residues at this position,which is recognized by an arginine residue (Arg 83) (Fig. 3)(Durocher et al. 2000). In contrast, the Chk2 FHA domainshows selectivity for hydrophobic residues at the pThr +3position, which is bound in a largely hydrophobic pocket (Liet al. 2002).

FHA function in polynucleotide kinasePolynucleotide kinase (PNK) is a bifunctional 5′-kinase –

3′-phosphatase that is responsible for the processing of dam-aged DNA ends at both double- and single-stranded breaksand in base excision repair pathways (Chappell et al. 2002;Rasouli-Nia et al. 2004; Koch et al. 2004; Loizou et al.2004). PNK activity is often essential to restore the 5′-phosphate – 3′-hydroxyl termini that are required for DNApolymerase and ligase activities to complete the repair ofthese lesions.

PNK contains, in addition to its catalytic kinase andphosphatase domains, an N-terminal FHA domain, whichtargets the enzyme to sites of repair and exhibits a substrateselectivity different than that of previously studied FHA do-mains. The FHA domain of PNK specifically recognizesCK2 phosphorylated forms of the single-strand break-repairscaffold protein XRCC1 and the double-strand break-repairprotein XRCC4, interactions that are required to direct PNKto sites of damage (Koch et al. 2004; Loizou et al. 2004).Peptide-array binding experiments have demonstrated thatthe PNK FHA domain, unlike other FHA domains, exhibitsits highest degree of selectivity for sequences N-terminal,rather than C-terminal, to the phosphothreonine (Koch et al.2004).

© 2005 NRC Canada

724 Biochem. Cell Biol. Vol. 83, 2005

To understand the structural basis for the recognition ofphosphorylated XRCC4 and XRCC1 by the PNK FHA domain,we determined the structure of this domain bound to anXRCC4-derived phosphopeptide (Ac-YDES(pT)DEESEKK-CONH2 (Fig. 3) (Bernstein et al. 2005). The structure re-veals that the phosphothreonine residue is recognized in thesame way as it is in other FHA domains, using the invariantArg 35 and Ser 47 and the highly conserved Arg 48, each ofwhich ligates the phosphate group.

Comparison of the sequences surrounding the phospho-threonine in the XRCC1- and XRCC4-target sequences re-veals a predominance of acidic residues both N- and C-terminal to the phosphothreonine. In XRCC1, there is theadditional possibility of multiple phosphorylations by theacidophilic kinase CK2 in the vicinity of the primary site ofphosphorylation (Loizou et al. 2004). This, together withpeptide-binding array studies, indicates that the FHA domainpreferentially binds acidic target peptides. The structure ofthe PNK FHA domain explains this preference (Fig. 3B).The peptide-binding surface is highly positively charged; keybasic residues within this surface, Arg 44, Lys 45 and Arg48, are presented on 2 loops, which together comprise apeptide-binding cradle. Arg 48, in addition to its role inphosphate recognition, is poised to select for the acidic Aspat the peptide pThr –3 position. Arg 44 adopts different ori-entations in each of the 3 distinct FHA domain–peptidecomplexes in the crystallographic asymmetric unit, indicat-

ing a significant degree of flexibility in this residue, whichcould allow it to interact with either the pThr –2 or the pThr+1 positions. In the XRCC4 target, both the –2 and +1 resi-dues are negatively charged. Arg 44 and Arg 48 are essentialfor peptide recognition because mutation of these residues(to alanine for Arg 44 and asparagine for Arg 48) com-pletely destroys the ability of the PNK FHA domain to bindthe XRCC4 phosphopeptide. Lys 45 could provide addi-tional electrostatic recognition for negatively charged resi-dues at pThr +1 and pThr +2; however, mutation of thisresidue to alanine has a negligible affect on the peptide-binding affinity, suggesting that this residue plays only aminor role, if any, in phosphopeptide selection. There is nointerpretable electron density for the peptide chain C-terminalto the pThr +2 position, indicating that this region of the tar-get is highly mobile and, in contrast to the peptide targets ofother FHA domains, is not bound by the FHA domain.

The PNK FHA domain is quite distinct at the amino acid -sequence level from most other FHA domains; this is re-flected in its unusual mode of peptide recognition. The PNKFHA domain is, however, highly similar to the FHA domainof aprataxin (APTX), a protein associated with the neurologicaldisorder ataxia-oculomotor apraxia, and is likely also involvedin DNA repair (Fig. 3D) (Moreira et al. 2001; Gueven et al.2004). Interestingly, APTX also associates with both XRCC1and XRCC4, and it has been proposed that competition be-tween APTX and PNK for XRCC1 or XRCC4 may provide

© 2005 NRC Canada

Williams et al. 725

Fig. 3. Phosphothreonine peptide recognition by the forkhead-associated (FHA) domain of polynucleotide kinase (PNK). (A) Structuralalignment of the FHA domain–peptide complexes from Rad53 and PNK. Strands are numbered according to the order in the PNK FHAdomain. (B and C) Details of PNK-phosphopeptide (B) and Rad53-phosphopeptide (C) recognition. FHA side chains that make critical in-teractions with the peptide are shown as sticks; the FHA domains are in the same orientation for comparison. Note that the criticalinteraction involving Arg 83, which specifies an aspartate at the pThr +3 position in Rad53, is missing in PNK. (D) Sequence alignmentof the FHA domains from PHK and APTX. P indicates residues that recognize the target pThr, and the green circles indicate otherresidues that contact the peptide.

a mechanism for the regulation of PNK activity (Luo et al.2004; Clements et al. 2004).

The peptide-contacting residues, including Arg 44, Lys45, and Arg 48, are all conserved in APTX, suggesting thatthe mode of peptide recognition in this protein is similar tothat seen in PNK. However, the pThr +3 position appears tobe important in the APTX:XRCC1 interaction; mutation ofGlu to Ala at this site in XRCC1 abolished binding to APTX(Luo et al. 2004). APTX has a single positively charged resi-due, Lys 75 (corresponding to Pro 81 in PNK), which mightapproach the pThr +3 residue to provide sequence selectivityat this position.

Interestingly, although the PNK FHA domain recognizesits negatively charged targets using a complementary posi-tively charged binding surface, all of the electrostatic inter-actions that we have identified between the FHA domain andthe acidic residues N- or C-terminal to the pThr are rela-tively long (>3.5 Å), and are assumed to be relatively weak(Fig. 3B). This may allow the PNK FHA domain to recog-nize several similar but nonidentical acidic target peptides inboth XRCC1 and XRCC4. It may also be that we have notyet identified the highest-affinity target for the PNK FHAdomain. Peptide selection studies have not revealed dramaticbinding preferences, such as the preference of the BRCA1BRCT for phenylalanine–tyrosine at the pSer +3 position.The fact that CK2, the kinase responsible for phosphorylationof both XRCC1 and XRCC4, phosphorylates clusters of res-idues in these targets raises the possibility that the PNKFHA domain may bind most tightly to multiply phos-phorylated targets.

Implications for cancer therapy

Inhibitors of proteins involved in DNA repair and DNA-damage-associated cell-cycle checkpoints provide potentialleads for new anticancer therapies. For example, inhibitorsof BRCA1 could provide an avenue of attack for breast andovarian tumors that are BRCA1+ but that have been disabledin an alternative checkpoint system. PNK inhibitors mightprovide a means to increase the efficacy of traditional DNA-targeting drugs by reducing the ability of the tumors to repairdamage induced by therapy. Signaling processes involvingBRCT and FHA domains have been shown to be essential inthese systems, and the detailed analysis of the principles thatunderlie phosphopeptide recognition in these systems couldprovide a basis for the rational design of inhibitors.

References

Bernstein, N.K., Williams, R.S., Rakovszky, M.L., Cui, D., Green, R.,Karimi-Busheri, F. et al. 2005. The molecular architecture of themammalian DNA repair enzyme, polynucleotide kinase. Mol.Cell, 17: 657–670.

Bork, P., Hofmann, K., Bucher, P., Neuwald, A.F., Altschul, S.F.,and Koonin E.V. 1997. A superfamily of conserved domains inDNA damage-responsive cell cycle checkpoint proteins. FASEBJ. 11: 68–76.

Callebaut, I., and Mornon, J.P. 1997. From BRCA1 to RAP1: awidespread BRCT module closely associated with DNA repair.FEBS Lett. 400: 25–30.

Cantor, S.B., Bell, D.W., Ganesan, S., Kass, E.M., Drapkin, R.,Grossman, S. et al. 2001. Bach1, a novel helicase-like protein,

interacts directly with brca1 and contributes to its DNA repairfunction. Cell, 105: 149–160.

Chappell, C., Hanakahi, L.A., Karimi-Busheri, F., Weinfeld, M.,and West, S.C. 2002. Involvement of human polynucleotide kinasein double-strand break repair by non-homologous end joining.EMBO J. 21: 2827–2832.

Clapperton, J.A., Manke, I.A., Lowery, D.M., Ho, T., Haire, L.F.,Yaffe, M.B., and Smerdon S.J. 2004. Structure and mechanismof BRCA1 BRCT domain recognition of phosphorylatedBACH1 with implications for cancer. Nat. Struct. Mol. Biol. 11:512–518.

Clements, P.M., Breslin, C., Deeks, E.D., Byrd, P.J., Ju, L.,Bieganowski, P. et al. 2004. The ataxia-oculomotor apraxia 1gene product has a role distinct from ATM and interacts with theDNA strand break repair proteins XRCC1 and XRCC4. DNARepair (Amsterdam), 3: 1493–1502.

Durocher, D., and Jackson, S.P. 2002. The FHA domain. FEBSLett. 513: 58–66.

Durocher, D., Taylor, I.A., Sarbassova, D., Haire, L.F., Westcott,S.L., Jackson, S.P. et al. 2000. The molecular basis of FHAdomain:phosphopeptide binding specificity and implications forphospho-dependent signaling mechanisms. Mol. Cell. 6:1169–1182.

Futreal, P.A., Liu, Q., Shattuck-Eidens, D., Cochran, C., Harshman,K., Tavtigian, S. et al. 1994. BRCA1 mutations in primary breastand ovarian carcinomas. Science (Washington, D.C.), 266:120–122.

Glover, J.N., Williams, R.S., and Lee, M.S. 2004. Interactionsbetween BRCT repeats and phosphoproteins: tangled up in two.Trends Biochem. Sci. 29: 579–585.

Gueven, N., Becherel, O.J., Kijas, A.W., Chen, P., Howe, O.,Rudolph, J.H. et al. 2004. Aprataxin, a novel protein that protectsagainst genotoxic stress. Hum. Mol. Genet. 13: 1081–1093.

Koch, C.A., Agyei, R., Galicia, S., Metalnikov, P., O’Donnell, P.,Starostine, A. et al. 2004. Xrcc4 physically links DNA endprocessing by polynucleotide kinase to DNA ligation by DNAligase IV. EMBO J. 23: 3874–3885.

Koonin, E.V., Altschul, S.F., and Bork, P. 1996. BRCA1 proteinproducts…Functional motifs… Nat. Genet. 13: 266–268.

Li, J., Williams, B.L., Haire, L.F., Goldberg, M., Wilker, E., Durocher,D. et al. 2002. Structural and functional versatility of the FHAdomain in DNA-damage signaling by the tumor suppressor kinaseChk2. Mol. Cell, 9: 1045–1054.

Loizou, J.I., El-Khamisy, S.F., Zlatanou, A., Moore, D.J., Chan,D.W., Qin, J. et al. 2004. The protein kinase CK2 facilitates repairof chromosomal DNA single-strand breaks. Cell, 117: 17–28.

Luo, H., Chan, D.W., Yang, T., Rodriguez, M., Chen, B.P., Leng,M. et al. 2004. A new XRCC1-containing complex and its rolein cellular survival of methyl methanesulfonate treatment. Mol.Cell. Biol. 24: 8356–8365.

Manke, I.A., Lowery, D.M., Nguyen, A., and Yaffe, M.B. 2003.BRCT repeats as phosphopeptide-binding modules involved inprotein targeting. Science (Washington, D.C.), 302: 636–639.

Miki, Y., Swensen, J., Shattuck-Eidens, D., Futreal, P.A., Harshman,K., Tavtigian, S. et al. 1994. A strong candidate for the breastand ovarian cancer susceptibility gene BRCA1. Science(Washington, D.C.), 266: 66–71.

Moreira, M.C., Barbot, C., Tachi, N., Kozuka, N., Uchida, E., Gibson,T. et al. 2001. The gene mutated in ataxia-ocular apraxia 1encodes the new HIT/Zn-finger protein aprataxin. Nat. Genet.29: 189–193.

Rasouli-Nia, A., Karimi-Busheri, F., and Weinfeld, M. 2004. Stabledown-regulation of human polynucleotide kinase enhances spon-

© 2005 NRC Canada

726 Biochem. Cell Biol. Vol. 83, 2005

taneous mutation frequency and sensitizes cells to genotoxicagents. Proc. Natl. Acad. Sci. U.S.A. 101: 6905–6910.

Rodriguez, M., Yu, X., Chen, J., and Songyang, Z. 2003. Phos-phopeptide binding specificities of BRCA1 COOH-terminal(BRCT) domains. J. Biol. Chem. 278: 52914–52918.

Shiozaki, E.N., Gu, L., Yan, N., and Shi, Y. 2004. Structure of theBRCT repeats of BRCA1 bound to a BACH1 phosphopeptide:implications for signaling. Mol. Cell, 14: 405–412.

Williams, R.S., Chasman, D.I., Hau, D.D., Hui, B., Lau, A.Y., andGlover, J.N. 2003. Detection of protein folding defects causedby BRCA1-BRCT truncation and missense mutations. J. Biol.Chem. 278: 53007–53016.

Williams, R.S., and Glover, J.N. 2003. Structural consequences ofa cancer-causing BRCA1-BRCT missense mutation. J. Biol. Chem.278: 2630–2635.

Williams, R.S., Green, R., and Glover, J.N. 2001. Crystal structure

of the BRCT repeat region from the breast cancer-associatedprotein BRCA1. Nat. Struct. Biol. 8: 838–842.

Williams, R.S., Lee, M.S., Hau, D.D., and Glover, J.N. 2004.Structural basis of phosphopeptide recognition by the BRCTdomain of BRCA1. Nat. Struct. Mol. Biol. 11: 519–525.

Yaffe, M.B., and Smerdon, S.J. 2001. PhosphoSerine/threoninebinding domains: you can’t pSERious? Structure (Cambridge),9: R33–R38.

Yu, X., and Chen, J. 2004. DNA damage-induced cell cyclecheckpoint control requires CtIP, a phosphorylation-dependentbinding partner of BRCA1 C-terminal domains. Mol. Cell. Biol.24: 9478–9486.

Yu, X., Chini, C.C., He, M., Mer, G., and Chen, J. 2003. TheBRCT domain is a phospho-protein binding domain. Science(Washington, D.C.), 302: 639–642.

© 2005 NRC Canada

Williams et al. 727