Regulation of Mre11/Rad50 by Nbs1EFFECTS ON NUCLEOTIDE-DEPENDENT DNA BINDING AND ASSOCIATION WITH ATAXIA-TELANGIECTASIA-LIKE DISORDER MUTANT COMPLEXES*

Received for publication, August 6, 2003, and in revised form, September 3, 2003Published, JBC Papers in Press, September 8, 2003, DOI 10.1074/jbc.M308705200

Ji-Hoon Lee‡, Rodolfo Ghirlando§, Venugopal Bhaskara¶, Michaela R. Hoffmeyer�, Jian Gu�,and Tanya T. Paull‡�**

From the ‡Department of Molecular Genetics and Microbiology, the �Institute of Cellular and Molecular Biology,and the ¶Department of Chemistry and Biochemistry, University of Texas, Austin, Texas 78712 and the §NIDDK,National Institutes of Health, Department of Health and Human Services, Bethesda, Maryland 20892

The Mre11/Rad50 complex is a critical component ofthe cellular response to DNA double-strand breaks, inorganisms ranging from archaebacteria to humans. Inmammalian cells, Mre11/Rad50 (M/R) associates with athird component, Nbs1, that regulates its activitiesand is targeted by signaling pathways that initiateDNA damage-induced checkpoint responses. Muta-tions in the genes that encode Nbs1 and Mre11 areresponsible for the human radiation sensitivity disor-ders Nijmegen breakage syndrome (NBS) and ataxia-telangiectasia-like disorder (ATLD), respectively,which are characterized by defective checkpoint re-sponses and high levels of chromosomal abnormalities.Here we demonstrate nucleotide-dependent DNA bind-ing by the human M/R complex that requires the Nbs1protein and is specific for double-strand DNA du-plexes. Efficient DNA binding is only observed withnon-hydrolyzable analogs of ATP, suggesting that ATPhydrolysis normally effects DNA release. The alleles ofMRE11 associated with ATLD and the C-terminal Nbs1polypeptide associated with NBS were expressed withthe other components and found to form triple com-plexes except in the case of ATLD 3/4, which exhibitsvariability in Nbs1 association. The ATLD 1/2, ATLD3/4, and p70 M/R/N complexes exhibit nucleotide-de-pendent DNA binding and exonuclease activity equiv-alent to the wild-type enzyme, although the ATLD com-plexes both show reduced activity in endonucleaseassays. Sedimentation equilibrium analysis of the re-combinant human complexes indicates that Mre11 is astable dimer, Mre11 and Nbs1 form a 1:1 complex, andboth M/R and M/R/N form large multimeric assembliesof �1.2 MDa. Models of M/R/N stoichiometry in light ofthis and previous data are discussed.

The mre11 and rad50 mutants in Saccharomyces cerevisiaewere originally named for the deficiencies in meiotic recombi-nation and ionizing radiation survival observed in mutantstrains (1, 2). We now know that the protein products of thesetwo genes associate together with a third component, Xrs2, inbudding yeast, and that the complex plays important roles

in homologous recombination, non-homologous end-joining,telomere maintenance, S-phase checkpoint control, and meioticrecombination (3–15). Homologs of Mre11 and Rad50 havebeen identified in every organism that has been geneticallycharacterized, as well as some bacteriophage, indicating thatthe functions of the complex are fundamental to DNA transac-tions. Proteins analogous to the Xrs2 component have not beenidentified in prokaryotes or archaebacteria, however, suggest-ing that this part of the complex is unique to eukaryotic cells.In mammals, the Mre11 and Rad50 proteins associate with aprotein called Nbs1 (also nibrin and p95) (16) into a largecomplex, Mre11/Rad50/Nbs1 (M/R/N).1

Nbs1 has little if any sequence similarity to Xrs2, but hasclinical significance because mutations in the NBS1 gene havebeen identified as causal factors in the human autosomal re-cessive genetic disorder Nijmegen breakage syndrome (NBS)(17). Patients with NBS exhibit radiation sensitivity, immunesystem deficiency, and a high rate of malignancy (18). At thecellular level, abnormalities include a defective S-phase check-point response and an elevated rate of chromosomal breakageand translocations, although few overt deficiencies in DNArepair are found. It was recently demonstrated that the mostcommon NBS allele, 657del5, can generate a C-terminalpolypeptide through the use of an internal ribosome entry siteupstream of the 5-nucleotide deletion (19). This translationproduct (p70) is capable of binding Mre11; thus, the 657del5NBS allele is very likely a hypomorphic allele that can supplya subset of the normal functions of Nbs1.

The clinical presentation of NBS patients constitutes a sub-set of the phenotype of patients with ataxia-telangiectasia (A-T), a related radiation sensitivity disorder. A-T is caused bymutations in the A-T-mutated gene (ATM), which encodes alarge protein kinase that initiates DNA damage signaling inresponse to DNA double-strand breaks in eukaryotic cells. Abiochemical connection between the two disorders was estab-lished with the demonstration that the ATM protein kinasephosphorylates the Nbs1 protein, in addition to many othertargets, and that Nbs1 phosphorylation by ATM is essential fora normal S-phase checkpoint response (20–23). An additionalconnection between M/R/N and ATM arose with the identifica-tion of two families with A-T-like disorder (ATLD), clinicallyidentical to A-T, yet caused by mutations in the MRE11 gene

* This work was supported by the Kimmel Cancer Foundation and byNational Institutes of Health Grant CA940008-01. The costs of publi-cation of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked “advertisement”in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

** To whom correspondence should be addressed: Dept. of MolecularGenetics and Microbiology, University of Texas, 1 University Station,A4800, Austin, TX 78712. Tel.: 512-232-7802; Fax: 512-232-3432;E-mail: tpaull@icmb.utexas.edu.

1 The abbreviations used are: M, Mre11; R, Rad50; N, Nbs1; NBS,Nijmegen breakage syndrome; A-T, ataxia-telangiectasia; ATLD, atax-ia-telangiectasia-like disorder; ATM, ataxia-telangiectasia-mutatedgene; MOPS, 4-morpholinepropanesulfonic acid; GST, glutathione S-transferase; AMP-PNP, adenosine 5�-(�,�-imino)triphosphate; ATP�S,adenosine 5�-O-(thiotriphosphate).

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 278, No. 46, Issue of November 14, pp. 45171–45181, 2003Printed in U.S.A.

This paper is available on line at http://www.jbc.org 45171

by guest on April 28, 2018

http://ww

w.jbc.org/

Dow

nloaded from

(24). The two known ATLD mutations are quite different; theATLD 1/2 allele contains a premature stop codon resulting in atruncation of the C-terminal domain, whereas the ATLD 3/4allele contains a missense mutation within the nuclease do-main in the N terminus. The biochemical basis of the abnor-malities seen in ATLD patients has not yet been elucidated.

Using a recombinant baculovirus system, we have expressedthe human Mre11, Rad50, and Nbs1 proteins together andfound that they form a large protein complex that exhibitsseveral distinct enzymatic activities on DNA substrates. TheMre11 protein contains highly conserved phosphoesterase mo-tifs that are responsible for the manganese-dependent nucleaseactivities of M/R/N. By itself and in association with Rad50 andNbs1, Mre11 exhibits a distributive, 3� to 5� exonuclease activ-ity on blunt and 3� recessed ends (25, 26), as well as a weakendonuclease activity on distorted DNA substrates such ashairpin structures. Association of Nbs1 with Mre11 stimulatesthe endonuclease function to act on hairpin structures and on3� overhangs (27), although Nbs1 itself has no apparent enzy-matic activities.

The Rad50 protein contains conserved Walker A and WalkerB ATPase motifs that are closely related to the ABC trans-porter family of membrane-associated ATPases. A long coiled-coil region separates the N-terminal and C-terminal ATP bind-ing domains, which are thought to associate with each other viaintramolecular, antiparallel association of the coiled-coil regionalong its length, a hypothesis supported by microscopy dataand by analogy with the SMC family of related coiled-coilproteins (28–30). M/R can catalyze a limited DNA unwindingreaction on DNA ends that is stimulated by ATP and alsorequires Nbs1 (27).

Crystal structures of the catalytic cores of Mre11 and Rad50homologs from the archaebacterium Pyrococcus furiosus illu-minate the essential components of the active sites for both thephosphoesterase and the ATPase (31, 32). Biochemical studieswith the P. furiosus Rad50 protein also indicated that ATPbinding induced dimerization of the Walker A/Walker B cata-lytic unit, and that this dimerization interface promoted ATP-dependent DNA binding by the protein. S. cerevisiae Rad50 hadpreviously been shown to exhibit nucleotide-dependent bindingto DNA as well (33). Mutations in the ATP-binding motifs ofRad50 in budding yeast have been shown to be equivalent tonull mutations with respect to meiotic recombination, mitoticrecombination, and growth rate (3), so ATP binding (and likelyalso hydrolysis) is essential to Rad50 function.

In this study, we have investigated the nucleotide-dependentbiochemical properties of the human M/R/N complex, focusingon the role of the Nbs1 protein. Nbs1 stimulates nucleotide-de-pendent DNA binding by Mre11/Rad50 (M/R), and these com-plexes are promoted and stabilized by the presence of non-hydrolyzable ATP analogs. Nucleotide-bound M/R/N bindsspecifically to double-stranded DNA duplexes, but does notshow an obvious preference for DNA ends. Association of Nbs1with the rest of the complex is destabilized in one of the ATLDM/R/N complexes (ATLD 3/4) but not in the other (ATLD 1/2),but both ATLD complexes show reduced levels of endonucleaseactivity compared with the wild-type enzyme. The M/R/N(p70)complex exhibits enzymatic activities essentially equivalent tothe wild-type enzyme. Finally, sedimentation equilibrium anal-ysis of the complex shows that Mre11 is clearly a dimer insolution, whereas both M/R and M/R/N are extremely largeprotein assemblies of �1.2 MDa.

EXPERIMENTAL PROCEDURES

Plasmid Expression Constructs—Baculovirus expression constructsfor human Mre11, Rad50, and Nbs1 have been described previously (25,27): pTP17, pTP11, and pTP36, respectively. The ATLD 1/2 allele of

Mre11 was generated by PCR from pTP17, generating a 6-histidine tagand stop codon at arginine 633, to form pTP219. A bacmid was madefrom this plasmid using the Bac-to-Bac system (Invitrogen), to formpTP221. The ATLD 3/4 allele of Mre11 was generated using theQuikChange system (Stratagene), introducing the N117S mutation intopTP17 to make pTP131 and subsequently the bacmid form, pTP133.The S1202R version of Rad50 was also generated using QuikChange inthe hRad50 gene in pFastBac1 to make pTP140, and the bacmid ver-sion, pTP141. The p70 version of Nbs1 was constructed in two steps.First, a 5-nucleotide deletion was made at position 657 in the Nbs1 geneusing QuikChange to re-create the 657del5 Nbs1 allele. Second, anN-terminal truncated version of this allele was generated by PCR, usingthe ATG at position 602 of the original gene as the initiator codon. Thisversion of Nbs1 was cloned into pFastBac1 (Invitrogen) without anaffinity tag, generating pTP270 and the bacmid form, pTP271. GST-Nbs1 was constructed by insertion of the glutathione S-transferasegene from pGEX-4T-1 (Amersham Biosciences) into pTP36 (27) to cre-ate a fusion protein with GST at the N terminus of Nbs1. All of thealleles generated by PCR were sequenced in their entirety. Each of thebacmids was used to make recombinant baculovirus according to theinstructions from the manufacturer.

Substrate DNA—The substrate in the gel mobility shift assays and inthe 3� overhang cutting assay consisted of TP423 (CTGCAGGGTTTT-TGTTCCAGTCTGTAGCACTGTGTAAGACAGGCCAGATC) annealedto TP424 (CACAGTGCTACAGACTGGAACAAAAACCCTGCAGTACT-CTACTCATCTC). TP423 was labeled with 32P at the 5� end for the gelmobility shift assays, or at the 3� end for the 3� overhang cutting assay.Labeling was performed with T4 polynucleotide kinase on the 5� end(New England Biolabs) using [�-32P]ATP, or with terminal deoxynu-cleotidyltransferase on the 3� end (Roche Molecular Biochemicals) using[�-32P]cordycepin. The single-stranded DNA used in Fig. 2A consisted oflabeled TP423 only. The plasmid DNA used as a competitor in Fig. 2was a derivative of the topo 2.1 vector (Invitrogen), unrelated in se-quence to the substrate oligonucleotides. The substrate used in Fig. 6A forthe exonuclease assay consisted of TP74 annealed to TP124 (34), withTP74 labeled with 32P at the 5� end. The hairpin substrate consisted ofTP355 (CATCCATGCCTACCTGAGTACCAGTAGCTACTGGTACTCA-GGTAGGCATGGATGCCAGATCGAC), labeled with 32P at the 5� end.

Proteins—M/R/N complexes were purified as previously described(25). The mutant M/R/N complexes were purified using the same pro-tocol as with the wild-type complex. All of the Mre11 proteins andRad50 contain a C-terminal histidine tag. Nbs1 (wild-type and p70)expressed with Mre11 or with both Mre11 and Rad50 did not contain anaffinity tag. GST-Nbs1 was expressed by itself in the baculovirus sys-tem and purified over a glutathione-Sepharose column, followed by ionexchange chromatography on Q Sepharose (Amersham Biosciences).GST-Nbs1 eluted from the Q Sepharose at �0.2 M NaCl. Protein con-centrations were determined by Bradford assay (Pierce) and confirmedby comparison with protein standards in Coomassie-stained SDS-PAGE gels. Western blotting of Mre11, Rad50, and Nbs1 was performedwith antibodies PC388 (Oncogene), MS-RAD10 (Genetex), and MS-NBS10 (Genetex), respectively, on polyvinylidene difluoride membrane(Millipore) using standard immunoblotting techniques.

Assay Conditions—Gel mobility shift assays were performed in avolume of 10 �l with 25 mM MOPS, pH 7.0, 50 mM NaCl, 1 mM

dithiothreitol, 0.1% Tween 20, 100 �g/ml bovine serum albumin, 5 mM

magnesium chloride, 0.5 mM AMP-PNP or ATP as indicated, and 1 nM

oligonucleotide substrate, with M/R/N complex added as indicated inthe figure legends (between 20 and 120 ng, which is 1.7–10 nM assum-ing 1.2 � 106 g/mol for M/R/N). In Fig. 2 (B and C), unlabeled plasmidDNA or single-stranded DNA (TP423) was added to the reaction beforethe addition of protein. M/R/N was incubated with the DNA and otherreaction components for 15 min at room temperature before the addi-tion of 1 �l of 50% glycerol and separation on a 0.7% agarose, 0.5� Trisborate-EDTA (TBE) gel at 5.7 V/cm for 100 min. Gels were dried andanalyzed using a phosphorimager (Amersham Biosciences). In the com-petitor assays, the amount of complex formed by M/R/N in the absenceof competitor was considered 100%, and the decreases in complex for-mation relative to that amount were determined by quantitative phos-phorimager analysis.

Nuclease assays were performed in a volume of 10 �l with 25 mM

MOPS (pH 7.0), 50 mM NaCl, 1 mM dithiothreitol, 1 mM manganesechloride, 1 nM oligonucleotide substrate, with M/R/N complex added asindicated at 20–25 nM (hairpin and 3� overhang assays) or 2.0–2.5 nM

(exonuclease assays). M/R/N was incubated with the DNA and otherreaction components at 37 °C for 90 min (hairpin and 3� overhangassays) or 15–30 min (exonuclease assays) before the addition of 1 �l of2% SDS, 0.1 M EDTA. ATP (0.5 mM) was included in the 3� overhang

M/R/N Hypomorphic Alleles and Stoichiometry45172

by guest on April 28, 2018

http://ww

w.jbc.org/

Dow

nloaded from

nuclease assays as indicated in Fig. 5C. Reactions were lyophilized,resuspended in 7 �l of formamide loading buffer, separated on a dena-turing, 20% polyacrylamide sequencing gel, and analyzed byphosphorimager.

Analytical Ultracentrifugation—Mre11, Rad50, and Nbs1 proteincomplexes were dialyzed overnight against buffer A (50 mM NaCl, 25mM Tris, pH 8.0, and 5 mM 2-mercaptoethanol) (Mre11) or buffer B (100mM NaCl, 25 mM Tris, pH 8.0, 5 mM 2-mercaptoethanol, and 5% (v/v)glycerol) (Mre11/Rad50, Mre11 ATLD 1/2/Nbs1 and Mre11/Rad50/Nbs1) at 4 °C. Samples were loaded and studied at nominal loadingconcentrations of 0.35 (Mre11), 0.11 (Mre11/Rad50), 0.31 (Mre11ATLD1/2/Nbs1), and 0.08–0.20 (Mre11/Rad50/Nbs1) A280. In the case ofthe double (Mre11/Rad50) and triple complexes loaded at 0.11 and 0.08A280, respectively, data were also collected at shorter wavelengths of230 and 225 nm.

Sedimentation equilibrium experiments were conducted at 4.0 °C ona Beckman Optima XL-A analytical ultracentrifuge. Mre11 sampleswere studied at 6,000, 8,000, and 10,000 rpm. Complexes of Mre11 withRad50 and with both Nbs1 and Rad50 were studied at rotor speeds of3,000, 3,500, and 4,000 rpm, whereas complexes of Mre11 ATLD1/2with Nbs1 were studied at rotor speeds of 6,000, 7,000, and 8,000 rpm.Data were acquired as an average of eight absorbance measurements at280 nm and a radial spacing of 0.001 cm. Scans were collected at 6-hintervals until equilibrium was reached. Data were analyzed in terms ofa single ideal solute to obtain the buoyant molecular mass, M1(1 � v�),using the Optima XL-A data analysis software (Beckman, MicrocalOrigin 3.78). In the case of Mre11 ATLD1/2 with Nbs1, data wereanalyzed in terms of two non-interacting solutes. Residuals were cal-culated. A random distribution of the residuals around zero was ob-tained as a function of the radius. A global analysis in terms of a singleideal solute performed in Sigma Plot 2001 as described (49) yieldssimilar results. Values of the molecular mass, M, were obtained fromthe buoyant molecular mass, M1(1 � v�), and calculated using densities,�, at 4.0 °C obtained from standard tables. Values of v � 0.7303, 0.7326,and 0.7339 ml g�1 were calculated for Mre11, ATLD1/2, and Nbs1,respectively, based on the amino acid composition using consensus datafor the partial specific molar volumes of amino acids (50). Otherwise, aconsensus protein value of 0.73 ml g�1 was utilized.

RESULTS

Nucleotide-dependent DNA Binding by M/R/N—The crystal-lographic structure of the P. furiosus Rad50 enzyme suggestedthat heterodimeric Rad50 catalytic domains further dimerizeinto a tetrameric unit in the presence of ATP (31). This changein multimeric state, stabilized by ATP, was hypothesized tofacilitate DNA binding by the enzyme. Rad50 from S. cerevisiaehas also been reported to bind DNA in an ATP-dependentmanner (33). Despite this compelling evidence, we have notobserved any significant ATP dependence in DNA binding bythe human M/R/N enzyme (Fig. 1A, lanes 4 and 5 and data notshown).

When we measured DNA binding by wild-type M/R/N in thepresence of magnesium and the non-hydrolyzable ATP analogAMP-PNP, however, we observed �20–40-fold higher levels ofDNA binding compared with reactions with ATP or withoutadded nucleotides (Fig. 1A, lanes 6 and 7). The protein concen-trations used here are 2–3-fold lower than the concentrationspreviously used to demonstrate M/R/N binding of DNA in theabsence of nucleotides (27). We have also observed a similarphenomenon with the non-hydrolyzable ATP analog ATP�S,and in each case, magnesium is absolutely required for forma-tion of the protein-DNA complex (data not shown).

We also tested subcomplexes of M/R/N for AMP-PNP-de-pendent DNA binding and found that Mre11/Rad50 (M/R)was incapable of binding DNA under these conditions, sug-gesting that the presence of Nbs1 in the complex is necessaryfor nucleotide-dependent DNA binding (Fig. 1B). We con-firmed this to be the case by performing DNA-binding assayswith M/R plus varying amounts of GST-Nbs1, which wasexpressed and purified in the absence of Mre11 and Rad50.As shown in Fig. 1C, the addition of GST-Nbs1 restoredAMP-PNP-dependent DNA binding to the M/R complex (Fig.

1C, lanes 5 and 6), yet did not exhibit any DNA-bindingcapability alone (lane 7).

The Rad50 protein is the only component of the complex thatcontains ATP-binding motifs and is therefore the likely sourceof nucleotide-dependent DNA binding by M/R/N. To confirmthis, we expressed a mutant form of Rad50 containing a singleamino acid change (S1202R) that is equivalent to the mutationmade in the P. furiosus enzyme (S793R), which was previouslyshown to abolish ATP-driven dimerization of the ATP-bindingdomain (31). The S1202R Rad50 mutant in association withwild-type Mre11 and Nbs1 failed to bind DNA in the presenceof AMP-PNP, verifying that the ATP-binding domain of theRad50 protein is required for the protein-DNA complex ob-served with the wild-type enzyme (Fig. 1D).

The AMP-PNP-dependent M/R/N complex is only formed ondouble-stranded DNA duplexes and not on single-stranded ol-igonucleotides, as shown in Fig. 2A. Addition of increasingamounts of M/R/N to the labeled double-stranded DNA in thepresence of magnesium and AMP-PNP actually generates twotypes of protein-DNA complex; the low mobility species previ-ously indicated (complex I), and a higher mobility species thatmigrates close to the unbound DNA (complex II). Formation ofeach complex is highly cooperative, normally appearing at athreshold of �30–60 ng of M/R/N per 10-�l reaction (equiva-lent to 2.5–5 nM, considering a molecular mass of 1.2 MDa). Therelative amounts of labeled DNA in the two species do notchange with increasing M/R/N concentration, although the mo-bility of both complexes decreases, suggesting that more pro-tein can be incorporated into each complex with additionalM/R/N. Under these conditions, essentially 100% of the DNA isincorporated into one or the other of these forms, with a ma-jority of the DNA in complex II. Preliminary off-rate experi-ments suggest that complex I has an off-rate on the order ofseveral minutes, whereas the off-rate of complex II is less than30 s (data not shown).

Competition studies with M/R/N on DNA duplexes indicatedthat both linear and uncut plasmid DNA could compete effec-tively with short DNA duplexes for M/R/N binding in complexI (Fig. 2B). Thus, DNA ends are not essential for nucleotide-dependent binding by M/R/N. Single-stranded oligonucleotide(Fig. 2B) and tRNA (data not shown) did not compete forbinding. In contrast, single-stranded DNA was able to partiallycompete with the double-stranded duplex in complex II (Fig.2C), indicating that complex II is less specific for duplex DNA.Mre11 protein alone is capable of forming a complex similar tocomplex II in gel mobility shift assays (27), also suggesting thatthis complex does not require functional Rad50 protein. Wehave not observed any sequence specificity of M/R/N binding toDNA substrates, although an exhaustive analysis of this ques-tion has not been performed.

ATLD Mutant Complexes—Mutations in Mre11 have beenidentified in human patients with the inherited genomic insta-bility syndrome ATLD (24). Two distinct mutations, ATLD 1/2and ATLD 3/4, were identified in different family groups. TheATLD 1/2 allele contains a premature stop codon, truncatingthe protein at amino acid 633, and eliminating the last 76amino acids. The ATLD 3/4 allele contains a missense muta-tion, N117S, but encodes a polypeptide of normal length. Thesealleles confer a clinical phenotype virtually identical to A-T.

We expressed the R633stop (ATLD1/2) and N117S (ATLD3/4) alleles of Mre11 together with wild-type Nbs1 and Rad50and isolated complexes with each protein (Fig. 3A). TheR633stop version of Mre11 purified with Nbs1 and Rad50 iden-tically to wild type, yielding a triple complex of similar appar-ent stoichiometry to the wild-type complex. In contrast, theN117S Mre11 usually failed to bind Nbs1, but did form a

M/R/N Hypomorphic Alleles and Stoichiometry 45173

by guest on April 28, 2018

http://ww

w.jbc.org/

Dow

nloaded from

FIG. 1. Nbs1 alters the DNA binding properties of Mre11/Rad50. A, gel mobility shift assays were performed with M/R/N in the presenceof magnesium and either ATP or AMP-PNP, as indicated. Reactions in lanes 2, 4, and 6 contained �20 ng of M/R/N, and reactions in lanes 3, 5,and 7 contained �60 ng of M/R/N (1.7 and 5 nM, respectively). Proteins were mixed with a 32P-labeled double-stranded DNA substrate containing3� overhangs at each end, and incubated for 15 min at room temperature before electrophoresis in a 0.7% 0.5� TBE-agarose gel. B, gel mobilityshift assays were performed as in A with 60 ng of M/R/N or 120 ng of M/R, and ATP or AMP-PNP, as indicated. C, gel mobility shift assays wereperformed as in A with 180 ng of M/R, and ATP or AMP-PNP, as indicated. 150 ng of GST-Nbs1 protein was also included in reactions 5–7. In thereaction shown in lane 5, the GST-Nbs1 was pre-incubated on ice with the M/R complex for 10 min before the addition of the other reactioncomponents. In the reactions shown in lanes 6 and 7, the GST-Nbs1 was added to M/R simultaneously with the labeled DNA and reaction mix. D,gel mobility shift assays were performed as in A with 60 ng of wild-type M/R/N or 60 ng of M/R(S1202R)/N, and ATP or AMP-PNP, as indicated.

M/R/N Hypomorphic Alleles and Stoichiometry45174

by guest on April 28, 2018

http://ww

w.jbc.org/

Dow

nloaded from

complex with Rad50 (Fig. 3A, lane 3, ATLD 3/4). The N117SMre11 mutant was unusual, however, in that it did retain someNbs1 binding in some protein preparations, as shown in Fig.3A, lane 4 (ATLD 3/4(N)). At this point it is not clear whatconditions during expression or during purification are respon-sible for retaining this association. In our experience, wild-typeMre11 has never exhibited this variability in Nbs1 binding.The identity of the R633stop protein and the absence of Nbs1 inthe ATLD 3/4 complex was verified by Western blotting, asshown in Fig. 3B.

In the nucleotide-dependent DNA-binding assay, the ATLD1/2 M/R/N complex bound the DNA fragment similarly to wildtype, and the ATLD 3/4 M/R complex did not bind (Fig. 3C),consistent with the requirement for Nbs1 in this assay asdemonstrated in Fig. 1. When GST-Nbs1 was added to theATLD 3/4 complex, however, it induced formation of a protein-DNA association similar to that observed with wild-type M/R inthe presence of GST-Nbs1. The ATLD 3/4(N) complex alsobound the DNA substrate, but at a reduced level comparedwith the wild-type complex (Fig. 3C, lanes 4–6). The reductionin binding is likely the result of the lower levels of Nbs1 in thispreparation relative to M/R. The ATLD 3/4 Mre11 mutant isthus capable of binding Nbs1 and forming nucleotide-depend-ent complexes with DNA, although it is apparently reduced inits affinity for Nbs1 such that it exhibits variability in Nbs1association.

M/R/N(p70) Complex—A large majority of identified NBSpatients are homozygous for an allele of the NBS1 gene thatcontains a 5-nucleotide deletion just downstream of the fork-head and BRCT domains (657del5) (16, 17). This deletioncauses a frameshift at codon 219 and a premature stop at codon233. Petrini and colleagues (19) have observed that, in additionto this truncated product, lymphoblast cells from NBS patientsalso express a novel C-terminal polypeptide that utilizes eitherof two alternative initiation codons just upstream of the dele-tion. This C-terminal protein, p70, is able to bind to Mre11 andwas subsequently shown by the same group to support viabilityin the mouse, in contrast to the lethality caused by null allelesof NBS1 (35). All of the available evidence suggests that the657del5 mutation creates a hypomorphic allele that can stillfulfill the essential functions of the M/R/N complex but lacksother nonessential functions because of the missing N terminusof the protein.

To study the biochemical properties of the M/R/N(p70) com-plex, we expressed a 657del5 allele of NBS1 utilizing the first ofthe two alternative ATG codons upstream of the deletion (19).This construct generates a polypeptide of 553 amino acids, ofwhich the C-terminal 535 residues are identical to the C ter-minus of wild-type Nbs1. Co-infection of insect cells with bacu-lovirus containing the p70 construct and wild-type Mre11 andRad50 yielded a triple complex of all three proteins (Fig. 4A).The p70 band migrates in the SDS gel very close to a polypep-tide present in the wild-type preparation that is likely a deg-radation product of Rad50 (data not shown), but the p70 bandis clearly an Nbs1 product based on Western blotting withanti-Nbs1 antibody (Fig. 4B).

M/R/N(p70) complexes are capable of binding DNA in theAMP-PNP-dependent assay, as shown in Fig. 4C. The migra-tion of the protein-DNA complex is very similar to the wild typein this assay (data not shown), suggesting that the overall massof the p70 M/R/N complex when bound to DNA is similar to thenormal complex and that the N terminus of Nbs1 is dispensablefor stimulation of Rad50-mediated DNA binding.

Nuclease Activity of ATLD and NBS Complexes—The M/R/Ncomplex exhibits exonuclease activity on blunt and 3�-recesseddouble-stranded DNA ends, and endonuclease activity on hair-

FIG. 2. The nucleotide-dependent protein-DNA complexformed by M/R/N is specific for double-stranded DNA. A, gelmobility shift assays were performed as in Fig. 1 with varying amountsof wild-type M/R/N in the presence of AMP-PNP and a 32P-labeleddouble-stranded DNA substrate (*TP423⁄TP424; lanes 1–4) or a 32P-labeled single-stranded DNA substrate (*TP423 only; lanes 5–8). Reac-tions contained 30 ng (lanes 2 and 6), 60 ng (lanes 3 and 7), or 120 ng(lanes 4 and 8) of M/R/N, as indicated. B, gel mobility shift assays wereperformed as in Fig. 1 with 60 ng of M/R/N and varying amounts ofunlabeled competitor DNA: linearized plasmid, uncut plasmid, or asingle-stranded oligonucleotide. 0.1, 1, 10, or 100-fold excess (mol nu-cleotides) of competitor DNA was mixed with the 32P-labeled DNAbefore addition of the protein. The amount of low mobility complex (I)without competitor was quantified and set to 100%, and the decrease inthe formation of this complex with increasing competitor was quantifiedand shown here. C, assays were performed with the same DNA com-petitors as in B except quantifying the higher mobility complex (II).

M/R/N Hypomorphic Alleles and Stoichiometry 45175

by guest on April 28, 2018

http://ww

w.jbc.org/

Dow

nloaded from

pins and 3� overhangs (25–27). The ATLD and p70 mutantM/R/N complexes were analyzed for exonuclease activity asshown in Fig. 5A, and each of the mutant complexes exhibitedsteady-state levels of activity nearly indistinguishable from thewild-type enzyme. The initial rate of resection on a 50-bp DNAfragment containing a 4-nucleotide recessed 3� strand was ex-amined for each of the complexes, and the ATLD 3/4 and ATLD1/2 complexes exhibited rates within 5–10% of wild type,whereas the M/R/N(p70) complex was �20% reduced.

The ATLD and p70 mutant complexes were also tested forendonuclease activity on a hairpin substrate (Fig. 5B) and a 3�overhang (Fig. 5C). In comparison to the wild-type enzyme, theATLD1/2 complex exhibited �50% activity on the 3� overhangspecifically, whereas the ATLD 3/4 � N complex showed loweractivity on both substrates. The lower activity of the ATLD

3/4(N) complex may very well be a result of the reduced bindingof Nbs1, as Nbs1 is required for endonuclease activity (27). Thepreparation of ATLD 3/4 without Nbs1 shows no activity inthese endonuclease assays, as expected (data not shown). Thep70 complex exhibits endonuclease activity between 50 and70% of the wild-type enzyme (Fig. 5, B and C). Overall, theATLD complexes show reduced, but not completely abrogated,levels of endonuclease activity. This defect appears to be spe-cific to the endonuclease function, considering that the rate ofexonucleolytic degradation seen with both of the ATLD mutantcomplexes was essentially identical to the wild-type enzyme.

Sedimentation Studies of the Mre11/Rad50/Nbs1 Complex—The larger of the protein-DNA complexes formed by M/R/N issignificantly retarded in mobility compared with the free oli-gonucleotide DNA, even in the low percentage agarose gels

FIG. 3. ATLD M/R/N complexesbind DNA. A, Coomassie-stained SDS-PAGE of ATLD M/R/N complexes.Mre11(R633stop)/R/N (ATLD 1/2) andMre11(N117S)/R/N (ATLD 3/4) com-plexes are shown in comparison to thewild-type (wt) complex. The ATLD 3/4allele of Mre11 associated with Rad50and Nbs1 in one preparation (3/4(N)),but failed to associate in another prepa-ration (3/4). The band marked with anasterisk is a degradation product ofRad50 that is found in all of the proteinpreparations. B, Western blot of ATLDM/R/N complexes. Blots containing wild-type, ATLD 1/2, and ATLD 3/4 com-plexes were probed with antibodies di-rected against Rad50, Nbs1, and Mre11,as indicated. C, gel mobility shift assayswith ATLD complexes. Increasingamounts of ATLD M/R/N complexeswere incubated with 32P-labeled duplexDNA and separated on an agarose gel asdescribed in Fig. 1. Reactions contained30 ng (lanes 1 and 4), 60 ng (lanes 2 and5), or 120 ng (lanes 3, 6, and 8–10) ofM/R/N. GST-Nbs1 protein was also in-cluded in reactions shown in lanes 10and 11.

M/R/N Hypomorphic Alleles and Stoichiometry45176

by guest on April 28, 2018

http://ww

w.jbc.org/

Dow

nloaded from

used for the separation of the complexes. To obtain a betterestimate of the size and stoichiometry of this complex, weanalyzed wild-type M/R/N complexes, as well as Mre11, Mre11/Nbs1 (M/N), and Mre11/Rad50 (M/R) by analytical ultracentrif-ugation. Sedimentation equilibrium experiments on wild-typeMre11 resulted in the gradual loss of material as a result ofsample precipitation; however, the Mre11 mutant H217Y (34)was found to have increased solubility. Studies of H217YMre11 show that the sample is monodisperse with an averagebuoyant molecular mass, M(1 � v�) of 40,360 � 4,100 g mol�1

(Fig. 6A). This corresponds to a measured molecular mass of150,500 � 15,300 g mol�1, indicating that under these condi-tions the Mre11 is dimeric (n � 1.9 � 0.2). Previously it wasobserved that the migration of human Mre11 in gel filtrationexperiments was more consistent with a multimeric stoichiom-etry (25), whereas measurements of yeast Mre11 using glycerolgradient sedimentation suggested that Mre11 is a monomer insolution (36). Our results in this work clearly indicate a dimerconfiguration, however, and at least one other measurement ofyeast Rad50 is consistent with this analysis (37). Gel filtrationexperiments with wild-type and H217Y Mre11 give identical

separation profiles (data not shown); thus, the wild-type pro-tein is very likely to also be a dimer in solution.

Mre11 and Rad50 interact to form a large macromolecularcomplex. Sedimentation equilibrium studies indicate that thiscomplex is monodisperse with a buoyant molecular mass of318,000 � 17,000 g mol�1. Based on a consensus partial spe-cific molar volume for proteins, this corresponds to a complexmolecular mass of (1.24 � 0.06) � 106 g mol�1.

Sedimentation equilibrium studies on the triple complexformed between Mre11, Rad50, and Nbs1 show that thesespecies interact to form a large, discrete, and slightly polydis-perse complex. Data analysis in terms of a single ideal soluteindicated rotor speed dependence for the buoyant molecularmass, with values ranging from 225,000 to 350,000 g mol�1 atspeeds of 4,000 and 3,000 rpm, respectively. Average M(1 � v�)values of 293,000 � 69,000 g mol�1 were calculated. Interest-ingly, global data analysis in terms of a single ideal soluteyielded an excellent fit returning a similar M(1 � v�) value of303,000 g mol�1 (Fig. 6C). This corresponds to a complex mo-lecular mass of (1.2 � 0.2) � 106 g mol�1. The mass observedfor the Mre11/Rad50/Nbs1 complex is thus identical, within the

FIG. 4. M/R/N(p70) complexes bindDNA. A, Coomassie-stained SDS-PAGEof M/R/N(p70) complex in comparison tothe wild-type (wt) complex. The bandmarked with an asterisk is a degradationproduct of Rad50 that is found in all of theprotein preparations. B, Western blot ofthe M/R/N(p70) complex in comparison towild type. Blots containing wild-type andM/R/N(p70) complexes were probed withantibodies directed against Rad50, Nbs1,and Mre11, as indicated. C, gel mobilityshift assays with M/R/N(p70) complexes.Increasing amounts of M/R/N(p70) com-plexes were incubated with 32P-labeledduplex DNA and separated on an agarosegel as described in Fig. 1. Reactions con-tained 30 ng (lane 1), 60 ng (lane 2), or 120ng (lanes 3 and 4) of M/R/N(p70) and ATPor AMP-PNP as indicated.

M/R/N Hypomorphic Alleles and Stoichiometry 45177

by guest on April 28, 2018

http://ww

w.jbc.org/

Dow

nloaded from

precision of the method, to that measured for the Mre11/Rad50complex. From these data, it is not possible to assign a clearstoichiometry to the M/R/N complex, and it is possible that thetriple complex may in fact be a collection of complexes withvarying stoichiometry.

We also tested wild-type M/N by sedimentation analysis, butfound that the complex precipitated over the course of theexperiment, similar to wild-type Mre11 alone. The ATLD1/2form of Mre11 expressed very well in insect cells, however, soan M/N complex of ATLD1/2 Mre11 and Nbs1 was used for thesedimentation analysis and was found to be stable over longperiods in solution. These results show that Mre11 ATLD1/2and Nbs1 interact to form a discrete 1:1 complex. Sedimenta-tion equilibrium studies show that a solution containing thiscomplex has both low and high molecular mass species. Dataanalysis in terms of two single and non-interacting solutesreturn a buoyant molecular mass of 40,260 � 2,000 g mol�1 forthe low molecular mass complex. This represents the predom-inant species (Fig. 6B). The experimental buoyant mass ap-pears to represent the 1:1 sum of the buoyant masses calcu-lated for the Mre11 ATLD1/2 (18,546 g mol�1) and Nbs1(21,387 g mol�1) components, indicating a 1:1 stoichiometry forthe M/N complex. Based on the buoyant molecular mass, acomplex mass of 159,740 � 7,950 g mol�1 is calculated (n �1.01 � 0.05). Because ATLD1/2 M/N separates identically towild-type M/N in gel filtration experiments (data not shown),we conclude that the complex of wild-type Mre11 with Nbs1 isvery likely to have 1:1 stoichiometry as well. All of the sedi-mentation equilibrium results are summarized in Table I.

Sedimentation equilibrium analysis of ATLD1/2 M/N alsoindicated the presence of a small amount of a high molecularmass species. The preparation of the ATLD1/2 Mre11 complexwith Nbs1 came from a preparation of the ATLD1/2 triplecomplex containing all three recombinant proteins. TheM(ATLD1/2)/N was separated from the M(ATLD1/2)/R/N by gelfiltration, and a small amount of ATLD1/2 M/R/N was presentin the fraction used for the analytical ultracentrifugation (datanot shown). The M/R/N in the fraction accounts for the minorspecies of large complex present, and has a buoyant molecularmass consistent with the presence of an Mre11, Rad50, andNbs1 triple complex present in small amounts (Fig. 6B).

DISCUSSION

Nbs1 Regulates Rad50 Function—The results shown heresuggest that Nbs1 plays an important role in modulating theDNA binding activity and specificity of the Rad50 protein. Itwas previously demonstrated that Nbs1 association alters theendonuclease specificity of Mre11 (27). Here we show that Nbs1also alters Rad50 function, such that nucleotide-dependentDNA binding is strongly promoted when Nbs1 is present in thecomplex. This increased DNA binding activity may explain thepreviously reported requirement for Nbs1 in DNA unwindingby Mre11/Rad50, an activity that is stimulated by ATP (27).

Crystallographic analysis of P. furiosus Rad50 suggestedthat binding of adenine nucleotides by Rad50 promotes dimer-ization of the protein and creates a binding site for a DNA helix(31). The sedimentation equilibrium analysis we report here forthe human M/R/N complex did not yield any evidence for nu-

FIG. 5. Nuclease assays of ATLD and p70 M/R/N complexes. A, exonuclease assays were performed with wild-type (wt), ATLD 1/2, ATLD3/4, and M/R/N(p70) complexes on a 50-bp 32P-labeled oligonucleotide substrate containing a recessed 3� end (diagram). Exonuclease activity byM/R/N proceeds in the 3� to 5� direction (arrow). Reactions were incubated for 15 or 30 min at 37 °C, as indicated. B, endonuclease assays wereperformed with wild-type (WT), ATLD 1/2, ATLD 3/4(N), and M/R/N(p70) complexes on a 50-bp 32P-labeled oligonucleotide substrate containinga hairpin on one end and a 3� overhang on the other (diagram). The predominant sites of cleavage by M/R/N are 1 and 2 nt 3� of the hairpin tip(arrows). Reactions were incubated for 90 min. C, endonuclease assays were performed in the presence of ATP as indicated with wild-type (WT),ATLD 1/2, ATLD 3/4(N), and M/R/N(p70) complexes on a 50-bp 32P-labeled oligonucleotide substrate with 3� overhangs on each end (diagram). Thepredominant site of cleavage by M/R/N in the presence of ATP is at the single-strand/double-strand junction (arrow). Reactions were incubated for90 min. All of the exo- and endonuclease assays were performed in the presence of 1 mM manganese, and the reaction products were separated ondenaturing polyacrylamide gels. The location of the 32P label is indicated with an asterisk (*) in all of the diagrams.

M/R/N Hypomorphic Alleles and Stoichiometry45178

by guest on April 28, 2018

http://ww

w.jbc.org/

Dow

nloaded from

cleotide-induced dimerization, but DNA-binding assays in vitrorevealed a high affinity protein-DNA complex formed by M/R/Nin the presence of non-hydrolyzable ATP analogs. AMP-PNP-dependent binding occurs 20–40-fold more efficiently than inthe absence of nucleotide, and is much less variable than com-plexes occasionally observed in the presence of ATP (data notshown). The most straightforward interpretation of these datais that non-hydrolyzable nucleotides facilitate DNA binding byM/R/N but prevent DNA release, thus locking all of the DNA-bound molecules into this state. The similarity between theconformations of the ATP and AMP-PNP-bound Rad50 crystalstructures (31) predicts that the AMP-PNP-bound state inM/R/N complexes is simply blocked at the hydrolysis step ofcatalysis, and is not a fundamentally different protein-DNAinteraction compared with the ATP-bound state. The require-ment for a non-hydrolyzable ATP analog instead of ATP alsosuggests that ATP hydrolysis by M/R/N must be very rapid.

Although M/R/N forms transient complexes with both singleand double-stranded DNA, the nucleotide-dependent DNA

binding shown in this work is specific for double-strandedduplexes. The effect of DNA end structure was not assessed inthis work, but a recent microscopy study of M/R also showedthat AMP-PNP altered the specificity of DNA binding to favor3� overhang structures (38).

Nbs1 C Terminus Is Sufficient for Stimulation of M/R Enzy-matic Functions—NBS homozygotes have no full-length Nbs1protein (16), but some tissues in NBS patients contain a C-terminal Nbs1 protein generated from an alternative start site(19). This truncated Nbs1 protein, p70, is sufficient for associ-ation with Mre11, and the data presented here indicate thatthis region is also sufficient for the stimulation of Mre11 endo-nuclease activity as well as for the nucleotide-dependent bind-ing by Rad50. Although the exact biochemical activities ofM/R/N are still under investigation, our analysis shows thatevery enzymatic function we can measure in vitro is essentiallyunchanged in a M/R/N(p70) mutant compared with the wild-type complex. This observation is not unexpected, consideringthat NBS cells have no overt defects in DNA repair but primar-

FIG. 6. Sedimentation equilibriumprofiles for Mre11 (H217Y) (A), theMre11(ATLD1/2)/Nbs1 complex (B),the Mre11/Rad50 complex (C), andthe M/R/N triple complex (D) shownas a distribution of A280 at equilib-rium. A, data were collected at 10,000rpm. The best single component fit forMre11 is shown as a solid line, and thecorresponding distribution of the residu-als is shown above the plot. B, data werecollected at 6,000 rpm and analyzed interms of two non-interacting ideal solutesto account for the presence of the Mre11/Rad50/Nbs1 triple complex. The best fit isshown as a solid line. The contributions ofthe double (M/N) and triple (M/R/N) com-plexes are indicated as dashed and dottedlines, respectively. The corresponding dis-tribution of the residuals is shown abovethe plot. C, data were collected at 3,000rpm. The best single component fit for theMre11/Rad50 complex is shown as a solidline, and the distribution of the residualsis shown above the plot. D, base-line cor-rected data for the triple M/R/N complexshown were collected at 3,500 rpm. Thebest single component data fit is shown asa solid line. The corresponding distribu-tion of the residuals is shown above.

TABLE IEstimated mass of Mre11/Rad50/Nbs1 complexes measured by sedimentation equilibrium analysis

Protein Mass Stoichiometrya

g mol�1

Mre11(H217Y) 150,500 � 15,300 (Mre11)n, n � 1.9 � 0.2Mre11(ATLD1/2)/Nbs1 159,740 � 7,950 (Mre11�Nbs1)n, n � 1.01 � 0.05Mre11/Rad50 1.24 � 0.06 � 106 See Fig.7Mre11/Rad50/Nbs1 1.2 � 0.2 � 106 See Fig.7

a Molecular weight of each polypeptide including C-terminal histidine tags on Mre11 and Rad50 (g mol�1): Mre11, 81,963; Mre11(ATLD1/2),73,343; Nbs1, 84,905; Rad50, 155,702.

M/R/N Hypomorphic Alleles and Stoichiometry 45179

by guest on April 28, 2018

http://ww

w.jbc.org/

Dow

nloaded from

ily exhibit deficiencies in checkpoint-related DNA damage re-sponses (18). Other studies have also demonstrated that the Nterminus of Nbs1 that includes the forkhead and BRCT do-mains is necessary and sufficient for focus formation at sites ofDNA damage, but that the C terminus is required for survivalof cells following ionizing radiation treatment (39–43). Regu-latory interactions between the C terminus of Nbs1 and Mre11/Rad50 are likely to be necessary for the DNA repair carried outby the complex and are therefore necessary for cell survival.

ATLD Mutant Complexes—The most obvious difference be-tween the ATLD mutant M/R/N complexes and the wild-typeenzyme is seen with ATLD 3/4, which exhibits a high degree ofvariability in Nbs1 binding. As shown here and in previouswork (27), the degree of Nbs1 association affects both nucleotide-dependent DNA binding as well as endonuclease activity ofMre11. Immunoprecipitations of the ATLD 3/4 Mre11 proteinby Stewart et al. (24) show that relatively little Nbs1 proteinassociates with ATLD 3/4 Mre11, and would thus suggest thatthe Nbs1-dependent enzymatic functions of the mutant com-plexes are likely to be compromised in these cells. In contrast toATLD 3/4, the C-terminal truncation mutation in ATLD 1/2has no apparent effect on Nbs1 binding in our recombinantexpression system. Neither of the ATLD mutant M/R/N com-plexes shows any significant reduction in nucleotide-dependentDNA binding, although the ability of ATLD 3/4 to bind DNAwas affected by the level of Nbs1 association. The exonucleaseactivities of the ATLD complexes are similarly unaffected com-pared with the wild-type enzyme, although both mutant com-plexes exhibit a significant reduction in endonuclease function.

ATLD cells, like A-T cells, do not show an overt defect inDNA repair (24), although a partial loss in endonuclease func-tion could result in repair deficiencies that would be difficult toassess using assays for whole-genome recovery. Indirect sup-port for a repair deficiency in ATLD cells comes from a recentobservation that ATLD3 cells are deficient in processing ofadenovirus genome intermediates, an activity that is postu-lated to occur via Mre11 endonuclease activity (44).

Stoichiometry of the M/R/N Complex—Mre11 shows the abil-

ity to interact with itself in two-hybrid assays (45), suggestingthat it forms a multimeric complex in vivo. The sedimentationequilibrium analysis of a nuclease-deficient mutant of Mre11(H217Y) performed in this study indicates that this mutantMre11 is clearly a dimer in solution. Wild-type Mre11 was notsoluble over the long time required for equilibrium centrifuga-tion (several days), but separates identically to the H217Yprotein during gel filtration over a Superdex 200 column; thus,wild-type Mre11 is very likely a stable dimer in solution.

Our sedimentation analysis of the Mre11(ATLD1/2) complexwith Nbs1 indicates that M/N has a 1:1 stoichiometry in solu-tion, because the mutant and wild-type M/N complexes alsoseparate identically in gel filtration. Nbs1 binding to Mre11 istherefore mutually exclusive with Mre11 dimer formation, ei-ther because Nbs1 binding occurs via the same interface, orbecause Nbs1 causes a conformational change in Mre11 thatinhibits dimer formation.

By analogy with the SMC family of coiled-coil ATPases (29,30) and microscopy analysis of the human M/R complex (28,46), it is likely that each Rad50 molecule folds back on itself toform an antiparallel coiled-coil within a monomer unit. Thecentral region of the Rad50 coiled coil has recently been shownto form a zinc-mediated hook structure that can associate withanother zinc hook (46); thus, Rad50 is predicted to form dimersconnected between their hook domains. Mre11 binds to Rad50at the base of the coiled-coil region (36, 47), and thus the M/Rcomplex would be expected to contain two Rad50 molecules andfour Mre11 molecules (assuming a dimer of Mre11 bound toeach Rad50 protein). Support for this type of configuration alsocomes from electron microscopy studies of yeast M/R (36) andEscherichia coli SbcC/SbcD (48). All of these complexes havebeen observed to form extended, dumbbell-like shapes withglobular domains at the ends.

The mass of M/R in solution as determined in this work byequilibrium sedimentation was 1.24 � 0.06 MDa, however,which is significantly larger than a M4R2 configuration. Astoichiometry of M4R2 would yield a complex of 630 kDa, ap-proximately half of the observed value. One possibility is that

FIG. 7. Hypothetical models ofM/R/N stoichiometry. A, Mre11, Nbs1,and Rad50 are symbolized by a circle, ahexagon, and an intramolecular coiled coilwith catalytic domains A and B, respec-tively. B, two possibilities for the configu-ration of M/R consistent with the ob-served mass of the complex are shown:head-to-head (top) and tail-to-tail (bot-tom). C, two possibilities for the configu-ration of M/R/N consistent with the ob-served mass of the complex are shown in atail-to-tail configuration.

M/R/N Hypomorphic Alleles and Stoichiometry45180

by guest on April 28, 2018

http://ww

w.jbc.org/

Dow

nloaded from

the stoichiometry is simply (M4R2)2, which would contain M4R2

complexes held together along the length of the coiled-coils, orassociated at the globular domains with the zinc hooks on theoutside (Fig. 7B). Both of these configurations have been ob-served for human M/R by electron microscopy (46). A recentstudy of human M/R by scanning force microscopy also visual-ized complexes with a very large globular domain at the centerwith arms (coiled-coils) extending outward (28), supporting thehead-to-head model of Mre11 and Rad50 interactions.

Our expectation in analyzing the mass of M/R and M/R/Ncomplexes was that the mass of M/R/N would be larger thanthat of M/R, and that the difference between the values wouldbe a consequence of Nbs1 association. The unexpected conclu-sion from our studies, however, is that the mass of the triplecomplex of human M/R/N is not significantly larger than that ofM/R within the limits of our assay, but exhibits a polydispersecharacter suggesting a heterogeneous collection of protein com-plexes. One explanation for this result is that varying numbersof Mre11 and Nbs1 are associated with each Rad50 dimer (orpair of dimers) such that M/R/N may in fact consist of a heter-ogeneous group of assemblies (two of the most plausible possi-bilities are shown in Fig. 7C). Our finding of a 1:1 stoichiometryof Mre11 with Nbs1 would suggest that the models that main-tain a 1:1 ratio of these proteins would be more likely to becorrect, assuming that M/N associates as a unit with Rad50molecules. Further experiments are clearly necessary to deter-mine which, if any, of the proposed models for the M/R/N triplecomplex are correct, and whether the stoichiometry is alteredunder different conditions.

Acknowledgments—We thank Martin Gellert and members of thePaull laboratory for helpful comments on the manuscript.

REFERENCES

1. Game, J. C., and Mortimer, R. K. (1974) Mutat. Res. 24, 281–2922. Ajimura, M., Leem, S. H., and Ogawa, H. (1993) Genetics 133, 51–663. Alani, E., Padmore, R., and Kleckner, N. (1990) Cell 61, 419–4364. Ivanov, E. L., Korolev, V. G., and Fabre, F. (1992) Genetics 132, 651–6645. Moore, J. K., and Haber, J. E. (1996) Mol. Cell. Biol. 16, 2164–21736. Kironmai, K. M., and Muniyappa, K. (1997) Genes Cells 2, 443–4557. Boulton, S. J., and Jackson, S. P. (1998) EMBO J. 17, 1819–18288. Haber, J. E. (1998) Cell 95, 583–5869. Bressan, D. A., Baxter, B. K., and Petrini, J. H. (1999) Mol. Cell. Biol. 19,

7681–768710. Chen, L., Trujillo, K., Ramos, W., Sung, P., and Tomkinson, A. E. (2001) Mol.

Cell 8, 1105–111511. D’Amours, D., and Jackson, S. P. (2001) Genes Dev. 15, 2238–224912. Grenon, M., Gilbert, C., and Lowndes, N. F. (2001) Nat. Cell Biol. 3, 844–84713. D’Amours, D., and Jackson, S. P. (2002) Nat. Rev. Mol. Cell. Biol. 3, 317–32714. Lobachev, K. S., Gordenin, D. A., and Resnick, M. A. (2002) Cell 108, 183–19315. Gonzalez-Barrera, S., Cortes-Ledesma, F., Wellinger, R. E., and Aguilera, A.

(2003) Mol. Cell 11, 1661–167116. Carney, J. P., Maser, R. S., Olivares, H., Davis, E. M., Le Beau, M., Yates, J. R.,

3rd, Hays, L., Morgan, W. F., and Petrini, J. H. (1998) Cell 93, 477–48617. Varon, R., Vissinga, C., Platzer, M., Cerosaletti, K. M., Chrzanowska, K. H.,

Saar, K., Beckmann, G., Seemanova, E., Cooper, P. R., Nowak, N. J.,Stumm, M., Weemaes, C. M., Gatti, R. A., Wilson, R. K., Digweed, M.,Rosenthal, A., Sperling, K., Concannon, P., and Reis, A. (1998) Cell 93,467–476

18. Shiloh, Y. (1997) Annu. Rev. Genet. 31, 635–66219. Maser, R. S., Zinkel, R., and Petrini, J. H. (2001) Nat. Genet. 27, 417–42120. Gatei, M., Young, D., Cerosaletti, K. M., Desai-Mehta, A., Spring, K., Kozlov,

S., Lavin, M. F., Gatti, R. A., Concannon, P., and Khanna, K. (2000) Nat.Genet. 25, 115–119

21. Lim, D. S., Kim, S. T., Xu, B., Maser, R. S., Lin, J., Petrini, J. H., and Kastan,M. B. (2000) Nature 404, 613–617

22. Wu, X., Ranganathan, V., Weisman, D. S., Heine, W. F., Ciccone, D. N.,O’Neill, T. B., Crick, K. E., Pierce, K. A., Lane, W. S., Rathbun, G.,Livingston, D. M., and Weaver, D. T. (2000) Nature 405, 477–482

23. Zhao, S., Weng, Y. C., Yuan, S. S., Lin, Y. T., Hsu, H. C., Lin, S. C., Gerbino,E., Song, M. H., Zdzienicka, M. Z., Gatti, R. A., Shay, J. W., Ziv, Y., Shiloh,Y., and Lee, E. Y. (2000) Nature 405, 473–477

24. Stewart, G. S., Maser, R. S., Stankovic, T., Bressan, D. A., Kaplan, M. I.,Jaspers, N. G., Raams, A., Byrd, P. J., Petrini, J. H., and Taylor, A. M.(1999) Cell 99, 577–587

25. Paull, T. T., and Gellert, M. (1998) Mol. Cell 1, 969–97926. Trujillo, K. M., Yuan, S. S., Lee, E. Y., and Sung, P. (1998) J. Biol. Chem. 273,

21447–2145027. Paull, T. T., and Gellert, M. (1999) Genes Dev. 13, 1276–128828. de Jager, M., van Noort, J., van Gent, D. C., Dekker, C., Kanaar, R., and

Wyman, C. (2001) Mol. Cell 8, 1129–113529. Hirano, M., and Hirano, T. (2002) EMBO J. 21, 5733–574430. Haering, C. H., Lowe, J., Hochwagen, A., and Nasmyth, K. (2002) Mol. Cell 9,

773–78831. Hopfner, K. P., Karcher, A., Shin, D. S., Craig, L., Arthur, L. M., Carney, J. P.,

and Tainer, J. A. (2000) Cell 101, 789–80032. Hopfner, K. P., Karcher, A., Shin, D., Fairley, C., Tainer, J. A., and Carney,

J. P. (2000) J. Bacteriol. 182, 6036–604133. Raymond, W. E., and Kleckner, N. (1993) Nucleic Acids Res. 21, 3851–385634. Paull, T. T., and Gellert, M. (2000) Proc. Natl. Acad. Sci. U. S. A. 97,

6409–641435. Williams, B. R., Mirzoeva, O. K., Morgan, W. F., Lin, J., Dunnick, W., and

Petrini, J. H. (2002) Curr. Biol. 12, 648–65336. Anderson, D. E., Trujillo, K. M., Sung, P., and Erickson, H. P. (2001) J. Biol.

Chem. 276, 37027–3703337. Furuse, M., Nagase, Y., Tsubouchi, H., Murakami-Murofushi, K., Shibata, T.,

and Ohta, K. (1998) EMBO J. 17, 6412–642538. de Jager, M., Wyman, C., van Gent, D. C., and Kanaar, R. (2002) Nucleic Acids

Res. 30, 4425–443139. Cerosaletti, K. M., and Concannon, P. (2003) J. Biol. Chem. 278, 21944–2195140. Desai-Mehta, A., Cerosaletti, K. M., and Concannon, P. (2001) Mol. Cell. Biol.

21, 2184–219141. Kobayashi, J., Tauchi, H., Sakamoto, S., Nakamura, A., Morishima, K., Mat-

suura, S., Kobayashi, T., Tamai, K., Tanimoto, K., and Komatsu, K. (2002)Curr. Biol. 12, 1846–1851

42. Tauchi, H., Kobayashi, J., Morishima, K., Matsuura, S., Nakamura, A.,Shiraishi, T., Ito, E., Masnada, D., Delia, D., and Komatsu, K. (2001) J. Biol.Chem. 276, 12–15

43. Zhao, S., Renthal, W., and Lee, E. Y. (2002) Nucleic Acids Res. 30, 4815–482244. Stracker, T. H., Carson, C. T., and Weitzman, M. D. (2002) Nature 418,

348–35245. Johzuka, K., and Ogawa, H. (1995) Genetics 139, 1521–153246. Hopfner, K. P., Craig, L., Moncalian, G., Zinkel, R. A., Usui, T., Owen, B. A.,

Karcher, A., Henderson, B., Bodmer, J. L., McMurray, C. T., Carney, J. P.,Petrini, J. H., and Tainer, J. A. (2002) Nature 418, 562–566

47. Hopfner, K. P., Karcher, A., Craig, L., Woo, T. T., Carney, J. P., and Tainer,J. A. (2001) Cell 105, 473–485

48. Connelly, J. C., Kirkham, L. A., and Leach, D. R. (1998) Proc. Natl. Acad. Sci.U. S. A. 95, 7969–7974

49. Ghirlando, R., Keown, M. B., Mackay, G. A., Lewis, M. S., Unkeless, J. C., andGould, H. J. (1995) Biochemistry 34, 13320–13327

50. Perkins, S. J. (1986) Eur. J. Biochem. 15, 169–180

M/R/N Hypomorphic Alleles and Stoichiometry 45181

by guest on April 28, 2018

http://ww

w.jbc.org/

Dow

nloaded from

and Tanya T. PaullJi-Hoon Lee, Rodolfo Ghirlando, Venugopal Bhaskara, Michaela R. Hoffmeyer, Jian Gu

DISORDER MUTANT COMPLEXESDNA BINDING AND ASSOCIATION WITH ATAXIA-TELANGIECTASIA-LIKE Regulation of Mre11/Rad50 by Nbs1: EFFECTS ON NUCLEOTIDE-DEPENDENT

doi: 10.1074/jbc.M308705200 originally published online September 8, 20032003, 278:45171-45181.J. Biol. Chem. 

  10.1074/jbc.M308705200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/278/46/45171.full.html#ref-list-1

This article cites 50 references, 18 of which can be accessed free at

by guest on April 28, 2018

http://ww

w.jbc.org/

Dow

nloaded from