ATM and the Mre11 Complex Combine to Recognize and Signal DNA Double-Strand Breaks

Oncogene (2007) 26, 7749–7758 & 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc REVIEW ATM and the Mre11 complex combine to recognize and signal DNA double-strand breaks

MF Lavin1,2

1Radiation Biology and Oncology Laboratory, Queensland Institute of Medical Research, Brisbane, Queensland, Australia and 2Faculty of Health Sciences, University of Queensland, Brisbane, Queensland, Australia

The recognition and repair of DNA double-strand breaks rearrangement of immunoglobulin and T-cell receptor (DSBs) is a complex process that draws upon a multitude genes (Livak, 2004; Dudley et al., 2005). of proteins. This is not surprising since this is a lethal Induced DNA DSBs are potentially lethal to the cell and lesion if left unrepaired and also contributes to genome therefore must be rapidly recognized and repaired to avoid instability and the consequential riskof cancer and other genetic damage (Agarwal et al., 2006). Not unexpectedly pathologies. Some of the key proteins that recognize these there are cellular mechanisms in place to recognize and breaks in DNA are mutated in distinct genetic disorders signal the presence of DNA DSB to the cell-cycle that predispose to agent sensitivity, genome instability, checkpoints to delay the passage of cells through the cycle cancer predisposition and/or neurodegeneration. These and facilitate DNA repair (Zhou and Elledge, 2000). include members of the Mre11 complex (Mre11/Rad50/ Signalling also occurs to the DNA repair machinery. Two Nbs1) and ataxia-telangiectasia (A-T) mutated (ATM), majormechanisms exist forthe repairofthe DNA DSB, mutated in the human genetic disorder A-T. The mre11 homologous recombination repair and non-homologous (MRN) complex appears to be the major sensor of the end joining (NHEJ) (Valerie and Povirk, 2003). Ulti- breaks and subsequently recruits ATM where it is mately, it is paramount that the break be repaired by one activated to phosphorylate in turn members of that of these mechanisms, but this is preceded by detection, complex and a variety of other proteins involved in cell- recognition and signalling of the break to other cellular cycle control and DNA repair. The MRN complex is also processes such as cell-cycle control and transcriptional upstream of ATM and ATR (A-T-mutated and rad3- events (Zhou and Elledge, 2000). Most progress on the related) protein in responding to agents that blockDNA recognition and signalling of DNA DSB has been acquired replication. To date, more than 30 ATM-dependent from a series of rare human genetic disorders sharing substrates have been identiﬁed in multiple pathways that sensitivity to agents that cause breaks in DNA (Lukas maintain genome stability and reduce the riskof disease. et al., 2006). Prominent among these are ataxia-telangiec- We focus here on the relationship between ATM and the tasia (A-T) with A-T-mutated (ATM) being the mutated MRN complex in recognizing and responding to DNA protein; Nijmegen breakage syndrome (NBS), Nbs1 DSBs. defective; A-T like syndrome (A-TLD), Mre11 defective; Oncogene (2007) 26, 7749–7758; doi:10.1038/sj.onc.1210880 and A-T and Rad3 related (ATR) (Seckels), ATR defective. (Table 1). A single case with a defect in Rad50 Keywords: ATM; Mre11 complex; DNA double strand protein has also been described (Dork et al., in prepara- breaks; signal transduction; functional consequences tion). It should be pointed out that while the gene products in all of these syndromes recognize and signal DNA DSB, they do not show dramatic reductions in capacity to repair these breaks (Foray et al., 1997; Girard et al., 2000; Riballo et al., 2004). On the otherhand, defects in DNA-PKcs, Introduction Ku70/80 and ligase IV, all of which participate in NHEJ, are associated with more marked defects in the capacity to Exposure of cells to ionizing radiation and radiomimetic repair DNA DSB (Featherstone and Jackson, 1999; Meek chemicals gives rise to a variety of DNA damage, including et al., 2004; Orii et al., 2006). In this review, the emphasis double-strand breaks (DSB) (Ward, 1985). DNA damage will be on the mre11 (MRN) complex and ATM. As we during S phase can lead to the collapse of DNA replication shall see the MRN complex is the primary sensor of DNA forks also generating DNA DSB (Paulsen and Cimprich, DSB. It recruits ATM to the break where it is activated 2007). DSBs also arise under normal conditions during and subsequently phosphorylates members of the complex and a variety of other proteins that signal to different cellular processes. This ensures that DNA repair complexes Correspondence: Dr MF Lavin, Radiation Biology and Oncology Laboratory, Queensland Institute of Medical Research, Brisbane, efﬁciently remove and repair the break. Failure to do so Queensland 4029, Australia. results in genome instability, which may give rise to cancer, E-mail: [email protected] neurodegeneration and other pathologies. Recognition and signalling of DNA DSB MF Lavin 7750 Table 1 DNA damage recognition/repair syndromes defective in DNA double-strand break repair Syndrome Defective Mutant Cancer Neurological Developmental/ Agent/sensitivity gene protein susceptibility changes growth delay

Ataxia-telangiectasia (A-T) ATM ATM Yes Neurodegeneration No Ionizing radiation A-T-like disorder (ATLD) Mre11 Mre11 No Neurodegeneration No Ionizing radiation Nijmegen breakage syndrome (NS) Nbs1 Nbs1 Yes Microcephaly Yes Ionizing radiation Rad50-deﬁcient patient Rad50 Rad50 ? Microcephaly ? Ionizing radiation Ataxia oculomotorapraxiatype 1 Aptx Aprataxin No Neurodegeneration No H2O2 (AOA1) Ataxia oculomotorapraxiatype 2 Setx Senataxin No Neurodegeneration No H2O2 Spinocerebellar ataxia with axonal Tdp1 TDP1 No Neurodegeneration No H2O2 neuropathy (SCAN1) A-T and Rad30 related disorder ATR ATR No Neurodegeneration No HU, UV (Seckels) DNA ligase Lig IV DNA No Microcephaly Yes Ionizing radiation Ligase IV

Abbreviations: ATM, ataxia-telangiectasia mutated; ATR, ataxia-telangiectasia mutated and Rad3 related; NBS, Nijmegen breakage syndrome.

C C The MRN complex acts as a sensor of DNA DSB X X X Zn X C C The MRN complex is a highly conserved protein complex involved in the following DNA repair mechanisms: both DNA DNA B homologous recombination repair and NHEJ, DNA B A M M A A A B M N N B replication, telomere maintenance and in signalling to the M cell-cycle checkpoints (D’Amours and Jackson, 2002; van C C den Bosch et al., 2003). The complex is rapidly localized X Zn X X X to nuclear foci in response to radiation exposure (Maser C C et al., 1997), which were shown to be sites of DNA Figure 1 Schematic representation of the binding of the Mre11 damage using irradiation masks and ultrasoft X-rays complex to DNA free ends. These ends are tethered by interaction (Nelms et al., 1998). It was subsequently shown that of the hinge regions of Rad50 molecules, assisted by co-ordination ATM, which phosphorylates Nbs1 for activation of the with Zn þ 2 ions. S-phase checkpoint, was not required for association of the MRN complex with sites of DNA damage (Mirzoeva and Petrini, 2001). However, the complex also binds Artemis, also play a role in DSB processing at least in a tightly to chromatin in the absence of DNA damage subgroup of breaks with damaged termini (Riballo et al., during S phase (Mirzoeva and Petrini, 2003). The Mre11/ 2004). Thus, the initial event in recognizing and Rad50 complex binds to DNA as a heterotetramer, responding to DNA DSB is the binding of the Mre11 tethering broken ends of a DSB (de Jager et al., 2001). complex, which tethers the broken ends together as a The binding appears to be achieved through the two means of preparing for repair. DNA-binding motifs of Mre11 (van den Bosch et al., 2003). This is arranged as a globular domain with Rad50 WalkerA and B motifs (ATPase domains) and the bridging of DNA molecules is achieved through CXXC ATM is recruited to and activated by the MRN complex sequences in the middle of Rad50 (Figure 1). These sequences are displayed at the ends of coiled-coil regions Evidence using rare genetic disorders and mouse models and appearto dimerize by the coordination of a Zn 2 þ ion While the recruitment of the MRN complex to damaged (Hopfner et al., 2002). Upon binding to DNA, the DNA is rapid, so too is the activation of ATM dynamic architecture of the MRN complex is altered to (Bakkenist and Kastan, 2003; Kozlov et al., 2003), give rise to parallel orientation of the coiled-coils of making it difﬁcult to discern the sequence of events Rad50, preventing intracomplex interaction and favour- involved. If the MRN complex were the sensor of DNA ing intercomplex association (Moreno-Herrero et al., breaks then it might be expected that it would be 2005). Association with Rad50 stimulates both the upstream of ATM activation. Evidence for this was exonuclease and endonuclease activities of Mre11 (Paull provided from studies with NBS and A-TLD cells, and Gellert, 1998; Trujillo and Sung, 2001) and Nbs1 hypomorphic for members of the complex; in cells in stimulates its endonuclease activity (Paull and Gellert, which the MRN complex was depleted during viral 1999). The complete complex can also partly unwind or infection; in vitro investigations using recombinant dissociate a short DNA duplex with a 30 overhang and proteins and in Xenopus extracts reconstituted for this activity is stimulated by ATP (Paull and Gellert, DNA damage signalling. In these experiments, ATM 1999). It seems likely that these and otheractivities may activation was determined by autophosphorylation on be responsible for the processing of DNA DSB prior to S1981 or indirectly through its capacity to phosphor- repair. However, it is evident that other nucleases, such as ylate downstream substrates. ATM activation was

Oncogene Recognition and signalling of DNA DSB MF Lavin 7751 retarded in NBS cells in response to neocarzinostatin the phenotype of NBS cells and that from Nbs1tr735 treatment (Uziel et al., 2003). The extent of reduction in (Stracker et al., 2007). Nbs1DC/DC cells showed normal activity was comparable in a moderate A-TLD patient, Atm S1987, p53 and Chk2 phosphorylation, but were compound heterozygous for nonsense and missense defective in structural maintenance of chromosomes mutations in Mre11, and was severely affected in a (SMC1) and BH3 interacting domain death agonist patient homozygous fora nonsense Mre11mutation (BID) phosphorylations. The latter was compatible with that markedly reduced the level of truncated Mre11 an attenuated apoptotic response in Nbs1DC/DC and in protein. Cerosaletti and Concannon (2004) showed that agreement with the results of Difilippantonio et al. (2007) an Nbs1 construct (N6FR5), retaining the Mre11- on the importance of the C-terminus of Nbs1 for binding site, gave rise to nuclear expression of Mre11/ apoptosis. However, as for NBS, Nbs1DC/DC were Rad50 and stimulated ATM activation. This is also the defective in the intra-S-phase checkpoint. Overall, the case forthe most common NBS mutation (657 D5, 5 bp data suggest that ATM recruitment is not mediated deletion giving rise to hypomorphic allele), which solely by the Nbs1 C-terminus. This is supported by the generates two fragments of the Nbs1 protein (Digweed evidence that ATM makes multiple contacts with the et al., 1999). The N-terminal fragment contains the MRN complex (Lee and Paull, 2004; Stracker et al., Mre11 binding site while the C-terminal fragment 2007). What is evident from these studies is the difference contains a conserved domain responsible for recruit- in phenotype, which cannot be explained solely by the ment of ATM (Falck et al., 2005). Cerosaletti and difference between human and mouse. The variability in Concannon (2004) showed that Mre11/Rad50 alone the extent of the defect in the different systems might be stimulated ATM activation afterlow doses of radiation. explained by the cell type and/orcompensatorymechan- However, in these experiments, this complex was isms. Understanding in more detail how the MRN retained in the nucleus by an Nbs1 construct, NbFR5, complex recruits and activates ATM will assist in that contained the N-terminal Mre11-binding site, but resolving these apparent discrepancies between systems/ was lacking in the FHA and BRCT domains. The species. importance of the ATM-binding domain on Nbs1 was also confirmed when it was shown that NBS cells expressing NbFR5DAtm, lacking ATM binding site, Viral infection and in vitro models to investigate the role had dramatically reduced levels of ATM activation of MRN complex in ATM activation (Cerosaletti et al., 2006). Infection of cells with wild-type adenovirus gives rise to Surprisingly, as with other facets of ATM activation, altered localization and proteosome degradation of Nbs mouse mutants behaved differently. Difilippantonio members of the MRN complex (Stracker et al., 2002). et al. (2007) reconstituted Nbs1-knockout mice with In E4 early gene region mutants of adenovirus, concata- human bacterial artificial chromosome transgenes carry- merization is observed which requires functional Mre11 ing specifically engineered mutants in different Nbs1 and Nbs1 and these proteins are localized to foci domains to investigate ATM activation and signalling. adjacent to viral replication centres. Carson et al. Two of these mutants, Nbs1657D5 corresponding to the (2003) subsequently showed that this end-joining of viral most common mutation in NBS patients and Nbs1tr735 genomes forE4-deleted adenoviruswas accompanied by lacking 20 amino acids at the C-terminus, showed Nbs1 phosphorylation reminiscent of that seen in normal levels of Atm activation as determined by response to DNA damage (Gatei et al., 2000). These S1987 (corresponding to human S1981) phosphoryla- results suggested that ATM was being activated as tion. Atm signalling was reduced somewhat in both determined by ATM, pS1981 localization to E4-deleted mutants at low radiation (2 Gy) dose but was normal at a viral centres (Carson et al., 2003). Degradation of the higherdose (8 Gy) and no defects wereevident in cell- MRN complex prevented this autophosphorylation of cycle checkpoint activation. Co-immunoprecipitation of ATM. These results also show that the MRN complex Atm with Nbs1tr735 was observed and Atm was recruited operates upstream from ATM. E4-deleted adenovirus to laser-induced sites of DNA damage in Nbs1tr735- infection not only activated ATM but also ATR transfected cells. However, it was demonstrated that the signalling, suggesting that the complex may also be C-terminus of Nbs1 was required for efficient radiation- upstream from ATR. Additional evidence in support of induced apoptosis. Thus, unlike the data forhuman cells, this was reported by Zhang et al. (2005) who showed that the C-terminus of Nbs1 in these specially engineered the MRN complex is required for ATR-dependent mice was dispensable for Atm recruitment and activa- phosphorylation of SMC1; for cell survival; the intra- tion. Since conditional deletion of Nbs1 almost abolishes S-phase checkpoint and genomic stability afterexposure Atm activation (Difilippantonio et al., 2005) the authors to UV. Furthermore, recruitment of ATR to sites of suggest that the lack of requirement for the C-terminus radiation-induced DNA damage and activation requires can be explained by otherinteractions between Atm and ATM and components of the MRN complex (Adams members of the complex. As support for this they cite et al., 2006; Myers and Cortez, 2006). data from Lee and Paull (2004) that provide evidence for The ability of the MRN complex to function as a multiple independent contacts between ATM and the DNA DSB sensorhas also been demonstrated in vitro MRN complex. Anothermouse model of Nbs1, (Lee and Paull, 2004, 2005). They demonstrated direct prepared by the more conventional approach, lacking activation of ATM by the MRN complex using the C-terminal 24 amino acid (Nbs1DC/DC) shared some of Baculovirus expressed complex members in the presence

Oncogene Recognition and signalling of DNA DSB MF Lavin 7752 of immunoprecipitated ATM. Under these conditions, activation and autophosphorylation. As pointed out ATM was shown to phosphorylate p53, Chk2 and above the sequence of events is different to that H2AX; however, the importance of the integrity of the proposed for human cells where the monomerization complex forthis activation was evident fromonly occurs as a consequence of autophosphorylation and partial stimulation of ATM kinase when Nbs1 was left activation. out. Adding DNA to this mixture revealed MRN complex binding, unwinding of the free ends, recruitment of ATM and dissociation of the inactive dimer followed by substrate phosphorylation (Lee and Paull, Focusing on the DNA DSB for ATM activation 2005). While Mre11/Rad50 could recruit ATM to DNA ends, it was incapable of stimulating ATM kinase Exposure of cells to ionizing radiation gives rise to activity. A phosphorylation site mutant of ATM approximately one DNA DSB per 30 cGy. Since these (S1981A) was capable of forming dimers, dissociated are distributed in a largely random fashion across the into monomers in the presence of MRN complex and genome, it is not possible to report the molecular events DNA, and was active with p53 and Chk2 substrates. occurring at individual breaks. Chromatin immunopre- This lack of requirement for autophosphorylation for cipitation followed by real-time PCR provides a sensitive activation in vitro was similar to that reported earlier by measurement of the kinetics and spatial distribution of Bakkenist and Kastan (2003). Indeed, there are several changes to chromatin and recruitment of DNA repair reports of ATM activation in vitro and in vivo without proteins in the proximity of the break site. Tsukuda et al. the requirement for S1981 autophosphorylation (Kozlov (2005) employed Saccharomyces cerevisiae with an et al., 2003; Hamer et al., 2004; Powers et al., 2004). inducible HO endonuclease to introduce a DSB at Xenopus cell-free extracts have been shown to mating type (MAT) that is only receptive to repair by recapitulate signalling pathways induced by DNA NHEJ. As observed previously (Schroff et al., 2004) g- damage and have also contributed to delineation of H2A accumulated rapidly and extensively around the the roles of Mre11 and ATM in this process (Costanzo break but g-H2A levels were lower closer to the break. et al., 2000, 2001). Depletion of Mre11 from these This was followed by loss of H2B and H3 histones, extracts abrogated the response to DNA DSB. While exposing DNA to nuclease susceptibility. Histone loss full-length Mre11 restored this response, Mre11 lacking was dependent on the S. cerevisiae Mre11 complex, the C-terminal DNA-binding domain failed to do so. MRX (Mre11/Rad50/Xrs2) (Usui et al., 2006) and the This group subsequently showed that fragmented DNA ATP-dependent nucleosome remodelling complex assembles with proteins into macromolecular structures INO80 (Shen et al., 2000). Using an introduced IScel enriched in the MRN complex and activated ATM site or radiation doses generating 2 breaks per S. (Costanzo et al., 2004). Assembly of these structures was cerevisiae nucleus, Lisby et al. (2004) demonstrated that dependent on Mre11 but not on ATM and Mre11 was Mre11 is the first protein detected at the break and phosphorylated. This phosphorylation of Mre11 may deletion of either Rad50 or Xrs2 interferes with the play a role in processing of DNA ends, since this assembly of Mre11 foci. This complex was required for modification has been shown to correlate with increased the recruitment of Tel1 kinase (homologue of ATM) and nuclease activity (Costanzo et al., 2001). In the absence it was suggested that this interaction may play a of the MRN complex, some activation of ATM was determining role as to whether homologous recombina- observed and this was dependent on the concentration tion repair or NHEJ operates to repair the break (Lisby of DNA free ends (Dupre et al., 2006). In this system, et al., 2004). Recruitment of HR proteins Rad51 and ATM activation as determined by monomerization, did Rad52 only occurs when Mre11 and Tel1 foci disas- not occur in MRN complex-depleted extracts. However, semble. IScel cleavage has also been utilized in human at higher concentrations of DNA ends the requirement cells to demonstrate recruitment of Mre11 and other forthe MRN complex forATM monomerizationwas repair proteins to the cleavage site (Rodrigue et al., bypassed. While the ATM monomergeneratedby only 2006). To understand the roles of ATM and Nbs1 in DNA ends was active and capable of phosphorylating chromatin structure modulation at a specific DNA DSB, H2AX, it was only weakly autophosphorylated on Berkovich et al. (2007) used the eukaryotic homing S1981. These data show that the dimerto monomer endonuclease I-Ppol (15-bp recognition sequence) to transition can occur without autophosphorylation un- cleave DNA at endogenous target sites in the human like that reported for ATM activation in human cells genome. Expression of this enzyme in human cells gives (Bakkenist and Kastan, 2003), but it should be kept in rise to approximately 30 DSB per cell. In this case, context since these experiments were performed in vitro. expression was targeted to the nucleus with 4-hydro- Dupre et al. (2006) showed that the DNA-tethering xytamoxifin by fusing a mutant oestrogen receptor activity of the MRN complex promotes the formation of hormone-binding domain to I-Ppo1 in a retroviral DNA damage signalling complexes, ATM monomeriza- vector. Addition of 4-hydroxytamoxifin to MCF7 cells, tion and ATM activation. The overall model derived infected with this construct, caused cleavage of DNA, from this studies envisages a two-step mechanism, the ATM activation and ATM-mediated downstream signal- first is an MRN complex-dependent DNA-tethering ling. Underthese conditions, dimeric ATM dissociated activity that increases the local concentration of DNA to the monomeric form, and this was dependent on ATM ends to trigger ATM monomerization followed by ATM kinase activity and the S1981 autophosphorylation site.

Oncogene Recognition and signalling of DNA DSB MF Lavin 7753 The ATM kinase inhibitorKU55933 (Hickson et al., protein XRCC4. In the absence of Nbs1, there was no 2004) inhibited ATM activation and this protein was not recruitment of ATM or loss of H2B. Higher levels of detected at the site of cleavage, but as expected Nbs1 still persistent DNA DSB were evident in cells lacking localized to the break, since this is ATM-independent. functional ATM orNbs1 pointing to a repairdefect, Furthermore, while ectopically expressed ATM localized which agrees with earlier reports of defective repair of to I-Ppo1 DSB in A-T cells, neithera kinase dead noran DNA DSB in both A-T and NBS cells (Foray et al., S1981A mutant form of ATM bound to the DSB. These 1997; Riballo et al., 2001). While Nbs1 largely accumu- data are at variance from those observed in vitro where lated at the break, ATM was enriched not only at the unphosphorylated dimers bind to DNA to become break but also in the flanking regions on chromatin. On activated (Lee and Paull, 2005), and in Xenopus extracts the otherhand, gH2AX was not found at the DNA DSB where ATM monomerization is followed by ATM but rather was enriched adjacent to the break. autophosphorylation and activation (Dupre et al., 2006). The model by Berkovich et al. (2007) envisages an initial binding of the MRN complex to the DNA DSB followed by Nbs1-dependent association of ATM with the break and surrounding area. In parallel, chromatin is DNA damage response: ATM/MRN interdependence disrupted as evidenced by the loss of H2B. As ATM is displaced from the break site, XRCC4 is recruited As mentioned above, the MRN complex and ATM are presumably as part of a NHEJ mechanism to repair the interdependent for the recognition and signalling of DSB. However, displacement of ATM from the break site DNA DSB. The complex binds to DNA DSB indepen- leads to its continued association with flanking regions of dent of ATM, recruits ATM to the break where it is chromatin. What these data also show is that inactive or activated and then acts as a substrate and adapter for phosphorylation site mutants (S1981A) of ATM fail to ATM signalling (Figure 2). While the requirement of bind to DNA DSB. This remains a point of contention MRN forATM activation is not absolute, it is evident since cells from a bacterial artificial chromosome recon- that cells deficient in Nbs1 and Mre11 show some stituted mouse mutant forthis site (S1987A) showed rapid abnormalities. This can be explained by the observations recruitment of wild type and S1987A ATM to sites of that ATM is activated with reduced efficiency in the DNA DSB (Pellegrini et al., 2006). AtmTgS1987A AtmÀ/À B absence of DNA DSB (Bakkenist and Kastan, 2003; cells showed enhanced retention of Atm-S1987A protein Difilippantonio et al., 2005). The stimulus in these cases on chromatin after DNA damage. Furthermore, irra- appears to be alteration in chromatin structure and it is diated cells showed normal phosphorylation of Chk2, possible that such alteration could contribute to ATM SMC1 and p53 and they were not hypersensitive to activity remote from a DSB. Exposure of cells to radiation. As mentioned above, mutant ATM-S1981A chromatin-modifying treatments activates ATM to was also shown to bind to DNA ends and can be phosphorylate p53 but not H2AX, SMC1 and Nbs1, activated in vitro (Lee and Paull, 2005). supporting the importance of the break or rather its Pellegrini et al. (2006) speculate that afterAtm interaction with the MRN complex for the phosphoryla- becomes activated at the DSB where it becomes tion of substrates other than p53. This is further accessible to rapid phosphorylation as a consequence substantiated by the results with Nbs1-deficient cells, of a high local concentration of Atm protein. In this where ATM activation was reduced at low doses of case, autophosphorylation as well as Atm-dependent radiation (0.3 Gy), but as the dose was increased there substrate phosphorylation are downstream from Atm was no significant effect on ATM phosphorylation. activation. Thus, these data suggest that the mechanism However, phosphorylation of downstream substrates of activation is considerably different in mouse and such as SMC1 was deficient even at higherdoses. This human and there are discrepancies between mice can be explained by the observation that ATM pS1981 mutants. showed a diffuse pattern of nuclear staining in Nbs1- deficient cells, failing to form discrete foci as observed in control cells. While ATM is activated in Nbs1-deficient cells, its inability to be anchored to the MRN complex at Perspective sites of DNA damage interferes with normal signalling. This is compatible with the model forATM activation The response to DNA DSB represents a complex, highly outlined in Figure 2. A clearer picture of the series of branched signalling network controlled by the MRN events involved has been provided by Berkovich et al. complex and ATM (Shiloh, 2006). An extensive ATM- (2007) using I-Ppol cleavage at a specific site on regulated network has been established using showcase chromosome 1. The dynamics of recruitment of different forATM-relatedpathways (SHARP; http://www.cs. proteins to the break were initially investigated in tan.ac.il/Bsharp). The substrates included play important G1 phase-synchronized cells (repair by NHEJ). Chroma- roles in DNA repair, cell-cycle control and transcription tin immunoprecipitation analysis revealed a time-depen- to determine the fate of the cell. Thus, it is imperative dent accumulation of ATM at the break accompanied by that the initiating orupstreamevents, in this case ATM loss of histone H2B from the break site, indicating activation, are tightly controlled. As we have seen ATM nucleosome disruption to expose the damaged DNA. is rapidly recruited to DNA DSB via the MRN complex. This was followed by the accumulation of the NHEJ It seems likely that ‘partial’ activation of ATM occurs

Oncogene Recognition and signalling of DNA DSB MF Lavin 7754

H2AX H2AX DNA ends

Nucleosome disruption

C C X X X Zn X γH2AX C C γH2AX ATM ATM B A M P 367 M B A A P 1893 B M N N A B 1981 P P 1981 M 1893 P 367 P C C X X X Zn X C C

C C X Zn X γ X X γ H2AX C C H2AX ATM ATM P P B A M M B A A B M N P P N A B M PP

C C X X X Zn X C C

C C X Zn X X X γH2AX γH2AX C C P P P P B A M P P M B A A B A B M N P P N M P P

P C C P X X X Zn X C C

Chk2 Chk2 SMC1 SMC1 BRCA1 BRCA1 FANCD2 FANCD2 Figure 2 Schematic representation of the series of events associated with ATM recruitment and activation at the sites of DNA DSB. Exposure of cells to ionizing radiation introduces DSB into DNA. This leads to opening of chromatin and nucleosome disruption, which is dependent upon both ATM and MRN. The MRN complex a sensorof the DSB, localizes to the breakto tetherthe ends and recruit ATM. This induces ATM autophosphorylation on at least three sites, which leads to dissociation of an inactive dimer to form active monomeric ATM. There is not an absolute requirement for MRN for this process and indeed there is some evidence that alteration to chromatin structure alone leads to partial activation of the ATM pool. Once activated by the MRN complex, ATM then phosphorylates members of this complex which act as adaptors for downstream signalling through ATM-dependent phosphorylation of other substrates. A list of some of the greater than 30 substrates is included. It should be pointed out that this is the model that applies to events in human cells. The mouse data are not the same for all steps. ATM, ataxia-telangiectasia mutated; DSB, double- strand break.

prior to localization to the break site and there is some strates such as Chk2, SMC1 and BRCA1 which depend evidence that this is achieved through local alterations in on MRN (Kitagawa et al., 2004). This partially active chromatin structure (Bakkenist and Kastan, 2003). form may represent ATM localized to regions adjacent ATM thus activated has the capacity to phosphorylate to the break (Berkovich et al., 2007). This is further the nucleoplasmic substrate p53, but not other sub- supported by the observation that, in the absence of

Oncogene Recognition and signalling of DNA DSB MF Lavin 7755 Nbs1, active ATM is not present in foci but rather in a DNA DSB for‘full activation’. Once activated it then diffuse pattern in the nucleus and only p53 phosphor- phosphorylates a series of substrates including Nbs1. ylation is observed at short times after irradiation. It is Specific phosphorylation of Nbs1 on S287 and S343 plays possible that H2AX may also be a substrate for ATM a role in the activation of the intra-S-phase checkpoint adjacent to the break and that enhancement of this (Lim et al., 2000). Yazdi et al.(2002)showedthatATM- response is due to further activation of ATM at the dependent phosphorylation of Nbs1 was required for break site. It is difficult to test the former, since normal phosphorylation of SMC1, implying that Nbs1, presum- numbers of gH2AX foci are observed even in the ably as part of the MRN complex, was an adaptor in an absence of ATM (McManus and Hendzel, 2005). In ATM/Nbs1/SMC1 pathway controlling the S-phase addition, while further activation involves MRN, it is checkpoint. An adaptorrole forphosphorylated Nbs1 also evident that MDC1 plays an important role in has also been demonstrated in DNA-damage-induced amplifying ATM-dependent DNA-damage signalling suppression of Tousled-like kinase activity (Krause et al., (Bekker-Jensen et al., 2006; Lou et al., 2006). This 2003). There is also some evidence that a second member may be a later event with MDC1, providing a platform of the MRN complex, Mre11, is phosphorylated in to stabilize the complex and sustain ATM signalling. response to DNA damage (Dong et al., 1999; Yuan et al., As discussed above, ATM autophosphorylation on 2002; Costanzo et al., 2004). More recently, the use of S1981 appears to be a crucial event for recruitment and large-scale and systematic proteomic analysis has identi- activation of ATM in human cells (Bakkenist and fied a site on Rad50 that is phosphorylated in response to Kastan, 2003; Berkovich et al., 2007). However, other DNA damage (Linding et al., 2007; Matsuoka et al., post-translational changes to ATM are an inherent part 2007). However, at this stage there is no evidence that of the activation process. Kozlov et al. (2006) used mass these phosphorylations are functionally significant. spectrometry to identify three phosphopeptides from While the model outlined in Figure 2 describes well the ATM from radiation-exposed cells. Use of tandem mass series of events that apply in response to DNA damage in spectrometry identified the specific phosphorylation human cells, it is at variance with at least some data in sites as S367, S1893 and S1981. Phosphorylation site mouse models. Pellegrini et al. (2006) demonstrated normal mutants (S367A, S1893A and S1981A) were defective in radiation-induced phosphorylation of Chk1, Chk2, p53 and ATM signalling and failed to correct either radio- SMC1 in a bacterial artificial chromosome-reconstituted TgS1987A À/À sensitivity orthe defective G 2/M checkpoint in A-T Atm S1987A mutant mouse (Atm Atm ). Further- cells. These data suggest that all three autophosphoryla- more, enhanced retention of Atm-S1987A was observed at tion sites are important for ATM activation and laserdamage-induced sites, suggesting that the phosphosite- signalling. What is not clearyet is whetherATM mutant could localize normally to DNA breaks. Cells from pS367 and pS1893 coincide with pS1981 at the break these mice had normal intra-S-phase and G2/M checkpoints site orwhetherthey areassociated with movement of and they were not hypersensitive to radiation. They further ATM from the site to the flanking regions. Not showed that while cell-cycle checkpoints and radiation surprisingly, protein phosphatases have also been shown survival were dependent upon the N-terminal FHA domain to play a role in ATM activation (Ali et al., 2004; of Nbs1, neitherthe phosphorylatedresidueson Nbs1 nor Goodarzi et al., 2004). DNA-damaged-induced acetyla- its conserved C-terminus were important for these processes tion of ATM has been reported by Sun et al. (2005) and (Difilippantonio et al., 2007). These data are interpreted to this was shown to run in parallel with pS1981 after mean that Atm phosphorylation on S1987 is not the signal radiation exposure. They identified the Tip60 histone that leads to Atm monomerization and activation and that acetyltransferase as the enzyme responsible and revealed this autophosphorylation and downstream transphosphor- that a dominant negative form of Tip60 reduced both ylation is a consequence rather than a cause of Atm acetylation and autophosphorylation of ATM. This activation. One explanation forthis is that the ATM enzyme was constitutively associated with ATM and its activation in the mouse may be quite different from that in acetylase activity increased in response to DNA damage. the human. This seems unlikely since the three autopho- No dependence forMRN was shown forTip60 sphorylation sites described for human cells are all activation of ATM nor for downstream substrate conserved in the mouse. While efforts were made in the phosphorylation of p53 or Chk2 (Jiang et al., 2006). bacterial artificial chromosome approach to select founder This is somewhat at variance with other results where lines expressing Atm-S1987A at levels similar to endogen- MRN-independent activation of ATM only resulted in ous levels there was some evidence of overexpression p53 phosphorylation (Kitagawa et al., 2004). (Pellegrini et al., 2006). This might create a situation where These data reveal that ATM activation is indeed an increased pool size of Atm in the vicinity of the break led controlled by a complex series of events in human cells. to Atm activation without the need forautophosphoryla- These steps are depicted in Figure 2. An important aspect tion. Activation can occur in vitro by adding ATP orDNA of this model is the interdependence of ATM and the ends to ATM without a requirement for pS1981 (Kozlov MRN complex. Since hypomorphic mutations in Nbs1 et al., 2003; Lee and Paull, 2005). Indeed, there are other and Mre11 do not dramatically affect the ability of the examples of ATM activation without evidence of phos- complex to localize to DNA DSB and recruit and activate phorylation on S1981 (Hamer et al., 2004; Powers et al., ATM, it is possible to define the role of phosphorylation 2004). In addition, it is evident that human ATM S1981A changes in members of the complex for downstream can be activated in vitro (Bakkenist and Kastan, 2003). signalling. ATM is recruited to the complex present at the Underthese conditions, Atm S1987A could be activated to

Oncogene Recognition and signalling of DNA DSB MF Lavin 7756 phosphorylate downstream substrates giving rise to normal Nbs1 in this process. The differences between the two radiosensitivity response and normal cell-cycle checkpoint mouse models might be explained by the different control in the AtmÀ/À background (Pellegrini et al., 2006). methods of generation of these models. With respect to the requirement for the C-terminus of Much has been achieved in recent years in under- Nbs1 for Atm activation, differences are observed in two standing how it is that ATM is activated. In human cells, mouse models. Pellegrini et al. (2006) reported a lack of it is evident that phosphorylation, dephosphorylation requirement for the C-terminus (22 amino-acid deletion) and acetylation play important roles in ATM activation. of Nbs1 forAtm activation and signalling, cell cycle Amongst these changes, autophosphorylation on at least control and radiation survival. In a second mouse model three sites is an inherent part of the activation mechan- with a closely related deletion (24 amino acids), Nbs1DC/ ism. This does not appearto be important formurine DC Atm activation appeared to be normal but down- Atm activation. However, in both cases the MRN stream signalling through SMC1 was defective (Stracker complex acts as a sensor of the break for the recruitment et al., 2007). In addition, cells from these mice exhibited and activation of Atm. No doubt the differences between a defective intra-S-phase checkpoint, but had normal the mouse and human will be shortly resolved and a radiation sensitivity and chromosomal stability. Radia- greater knowledge of the series of events involved in tion-induced apoptosis was defective in both mouse ATM activation and its dependence on the MRN models supporting a requirement for the C-terminus of complex delineated in the not too distant future.

References

Adams KE, Medhurst AL, Dart DA, Lakin ND. (2006). De JagerM, van NoortJ, van Gent DC, DekkerC, KanaarR, Recruitment of ATR to sites of ionizing radiation-induced Wyman C. (2001). Human Rad50/Mre11 is a flexible DNA damage requires ATM and components of the MRN complex that can tetherDNA ends. Mol Cell 8: 1129–1135. protein complex. Oncogene 25: 3894–3904. Difilippantonio S, Celeste A, Fernandez-Capetillo O, Chen Agarwal S, Tafel AA, Kanaar R. (2006). DNA double-strand HT, Reina San Martin B, Van Laethem F et al. (2005). Role break and chromosome translocations. DNA Repair 5: of Nbs1 in the activation of the Atm kinase revealed in 1075–1081. humanized mouse models. Nat Cell Biol 7: 675–685. Ali A, Zhang J, Bao S, Liu I, Otterness D, Dean NM et al. Difilippantonio S, Celeste A, Kruhlak MJ, Lee Y, Difilippan- (2004). Requirement of protein phosphatase 5 in DNA- tonio MJ, Feigenbaum L et al. (2007). Distinct domains in damage-induced ATM activation. Genes Dev 18: 249–254. Nbs1 regulate irradiation-induced checkpoints and apopto- Bakkenist CJ, Kastan MB. (2003). DNA damage activates sis. J Exp Med 204: 1003–1011. ATM through intermolecular autophosphorylation and Digweed M, Reis A, Sperling K. (1999). Nijmegen breakage dimerdissociation. Nature 421: 499–506. syndrome: consequences of defective DNA double strand Bekker-Jensen S, Lukas C, Kitagawa R, Melander F, Kastan break repair. BioEssays 21: 649–656. MB, Bartek J et al. (2006). Spatial organisation of the Dong Z, Zhong Q, Chen PL. (1999). The Nijmegen breakage mammalian genome surveillance machinery in response to syndrome protein is essential for Mre11 phosphorylation DNA strand breaks. J Cell Biol 173: 195–206. upon DNA damage. J Biol Chem 274: 19513–19516. Berkovich E, Monnat Jr RJ, Kastan MB. (2007). Roles of Dudley DD, Chaudhuri J, Bassing CH, Atl FW. (2005). ATM and Nbs1 in chromatin structure modulation Mechanism and control of V(D)J recombination versus class and DNA double-strand break repair. Nat Cell Biol 9: switch recombination: similarities and differences. Adv 683–690. Immunol 86: 43–112. Carson CT, Schwartz RA, Stracker TH, Lilley CE, Lee DV, Dupre A, Boyer-Chatenet L, Gautier J. (2006). Two-step Weitzman MD. (2003). The Mre11 complex is required for activation of ATM by DNA and the Mre11-Rad50-Nbs1 ATM activation and the G2/M checkpoint. EMBO J 22: complex. Nat Struct Mol Biol 13: 451–457. 6610–6620. Falck J, Coates J, Jackson SP. (2005). Conserved modes of Cerosaletti K, Concannon P. (2004). Independent roles for recruitment of ATM, ATR and DNA-PKcs to sites of DNA nibrin and Mre11-Rad50 in the activation and function of damage. Nature 434: 605–611. Atm. J Biol Chem 279: 38813–38819. Featherstone C, Jackson SP. (1999). Ku, a DNA repair protein Cerosaletti K, Wright J, Concannon P. (2006). Active role for with multiple cellularfunctions? Mutat Res 434: 3–15. nibrin in the kinetics of Atm activation. Mol Cell Biol 26: Foray N, Priestley A, Alsbeih G, Badie C, Capulas EP et al. 1691–1699. (1997). Hypersensitivity of ataxia telangiectasia fibroblasts Costanzo V, Paull T, Gottesman M, GautierJ. (2004). Mre11 to ionizing radiation is associated with a repair deficiency of assembles linearDNA fragmentsinto DNA damage DNA double-strand breaks. Int J Radiat Biol 72: 271–283. signalling complexes. PLoS Biol 2: E110. Gatei M, Young D, Cerosaletti KM, Desai-Mehta K, Spring Costanzo V, Robertson K, Bibikova M, Kim E, Grieco D, K, Kozlov S et al. (2000). ATM-dependent phosphorylation Gottesman M et al. (2001). Mre11 protein complex prevents of nibrin in response to radiation exposure. Nat Genet 25: double-strand break accumulation during chromosomal 115–119. DNA replication. Mol Cell 8: 137–147. Girard PM, Foray N, Stumm M, Waugh A, Riballo E et al. Costanzo V, Robertson K, Ying CY, Kim E, Avvedimento E, (2000). Radiosensitivity in Nijmegen breakage syndrome Gottesman M et al. (2000). Reconstitution of an ATM- cells is attributable to a repair defect and not cell cycle dependent checkpoint that inhibits chromosomal DNA checkpoint defects. Cancer Res 60: 4881–4888. replication following DNA damage. Mol Cell 6: 649–659. Goodarzi AA, Jonnalagadda JC, Douglas P, Young D, Ye R, D’Amours D, Jackson SP. (2002). The Mre11 complex: at the Moorhead GB et al. (2004). Autophosphorylation of ataxia- crossroads of DNA repair and checkpoint signalling. Mol telangiectasia mutated is regulated by protein phosphatase Cell Biol 3: 317–327. 2A. EMBO J 23: 4451–4461.

Oncogene Recognition and signalling of DNA DSB MF Lavin 7757 HamerG, Kal HB, Westphal CH, Ashley T, de Rooij DG. Meek K, Gupta S, Ramsden DA, Lees-MillerSP. (2004). The (2004). Ataxia telangiectasia mutated expression and activa- DNA-dependent protein kinase: the director at the end. tion in the testis. Biol Reprod 70: 1206–1212. Immunol Rev 200: 132–141. Hickson I, Zhao Y, Richardson CJ, Green SJ, Martin NM, Mirzoeva OK, Petrini JH. (2001). DNA damage-dependent Orr AI et al. (2004). Identification and characterization of a nucleardynamics of the Mre11complex. Mol Cell Biol 21: novel and specific inhibitorof the ataxia-telangiectasia 281–288. mutated kinase ATM. Cancer Res 64: 9152–9159. Mirzoeva OK, Petrini JH. (2003). DNA replication-dependent HopfnerKP, CraigL, Moncalian G, Zinkel RA, Usui T, nucleardynamics of the Mre11complex. Mol Cancer Res 1: Owen BA et al. (2002). The Rad50 zinc-hook is a structure 207–218. joining Mre11 complexes in DNA recombination and Moreno-Herrero F, de Jager M, Dekker NH, Kanaar R, repair. Nature 418: 562–566. Wyman C, DekkerC. (2005). Mesoscale conformational Jiang X, Sun Y, Chen S, Roy K, Price BD. (2006). The FATC changes in the DNA-repair complex Rad50/Mre11/Nbs1 domains of PIKK proteins are functionally equivalent and upon binding DNA. Nature 437: 440–443. participate in the Tip60-dependent activation of DNA-PKcs Myers JS, Cortez D. (2006). Rapid activation of ATR by and ATM. J Biol Chem 281: 15741–15746. ionizing radiation requires ATM and Mre11. J Biol Chem Kitagawa R, Bakkenist CJ, McKinnon PJ, Kastan MB. 281: 9346–9350. (2004). Phosphorylation of SMC1 is a critical downstream Nelms K, Huang H, Ryan J, Keegan A, Paul WE. (1998). event in the ATM-NBS1-BRCA1 pathway. Genes Dev 18: Interleukin-4 receptor signalling mechanisms and their 1423–1438. biological significance. Adv Exp Med Biol 452: 37–43. Kozlov S, Gueven N, Keating K, Ramsay J, Lavin MF. Orii KE, Lee Y, Kondo N, McKinnon PJ. (2006). Selective (2003). ATP activates ataxia-telangiectasia mutated (ATM) utilization of nonhomologous end-joining and homologous in vitro. Importance of autophosphorylation. J Biol Chem recombination DNA repair pathways during nervous system 278: 9309–9317. development. Proc Natl Acad Sci USA 103: 10017–10022. Kozlov SV, Graham ME, Peng C, Chen P, Robinson PJ, Paull TT, Gellert M. (1998). The 3 to 5 exonuclease activity of Lavin MF. (2006). Involvement of novel autophosphoryla- Mre11 facilitates repair of DNA double-strand breaks. Mol tion sites in ATM activation. EMBO J 25: 3504–3514. Cell 1: 969–979. Krause DR, Jonnalagadda JC, Gatei MH, Sillje HH, Zhou B, Paull TT, Gellert M. (1999). Nbs1 potentiates ATP-driven DNA unwinding and endonuclease cleavage by the Mre11/ Nigg EA et al. (2003). Suppression of Tousled-like kinase Rad50 complex. Genes Dev 13: 1276–1288. activity after DNA damage or replication block requires Paulsen RD, Cimprich KA. (2007). The ATR pathway: fine- ATM, NBS1 and Chk1. Oncogene 22: 5927–5937. tuning the fork. DNA Repair 6: 953–966. Lee JH, Paull TT. (2004). Direct activation of the ATM Pellegrini M, Celeste A, Difilippantonio S, Guo R, Wang W, protein kinase by the Mre11/Rad50/Nbs1 complex. Science Feigenbaum L et al. (2006). Autophosphorylation at serine 304: 93–96. 1987 is dispensable formurineAtm activation in vivo. Lee JH, Paull TT. (2005). ATM activation by DNA double- Nature 443: 222–225. strand breaks through the Mre11-Rad50-Nbs1 complex. Powers JT, Hong S, Mayhew CN, Rogers PM, Knudsen ES, Science 308: 551–554. Johnson DG. (2004). E2F1 uses the ATM signalling Lim DS, Kim ST, Xu B, MaserRS, Lin JH et al. (2000). ATM pathway to induce p53 and Chk2 phosphorylation and phosphorylates p95/Nbs1 in an S-phase checkpoint path- apoptosis. Mol Cancer Res 2: 203–214. way. Nature 404: 613–617. Riballo E, Doherty AJ, Dai Y, Stiff T, Oettinger MA, Jeggo PA Linding R, Jensen LJ, OstheimerGJ, van Vugt MA, Jørgensen C, et al. (2001). Cellularand biochemical impact of a mutation Miron IM et al. (2007). Systematic discovery on in vivo in DNA ligase IV conferring clinical radiosensitivity. JBiol phosphorylation networks. Cell 129: 1415–1426. Chem 276: 31124–33132. Lisby M, Rothstein R. (2004). DNA repair: keeping it Riballo E, Kuhne M, Rief N, Doherty A, Smith GC, Recio MJ together. Curr Biol 14: R994–R996. et al. (2004). A pathway of double-strand break rejoining Livak F. (2004). In vitro and in vivo studies on the generation of the dependent upon ATM, Artemis and proteins locating to primary T-cell receptor repertoire. Immunol Rev 200: 23–35. gamma-H2AX foci. Mol Cell 16: 715–724. Lou Z, Minter-Dykhouse K, Franco S, Gostissa M, Rivera Rodrigue A, Lafrance M, Gauthier MC, McDonald D, MA, Celeste A et al. (2006). MDC1 maintains genomic Hendzel M, West SC et al. (2006). Interplay between human stability by participating in the amplification of ATM- DNA repair proteins at a unique double-strand break dependent DNA damage signals. Mol Cell 21: 187–200. in vivo. EMBO J 25: 222–231. Lukas J, BohrVA, Halazonetis TD. (2006). Cellularresponses Schroff KC, Cowen MS, Koch S, Spanagel R. (2004). Strain- to DNA damage: current state of the field and review of the specific responses of inbred mice to ethanol following food 52nd Benzon Symposium. DNA Repair 5: 591–601. shortage. Addict Biol 9: 265–271. Maser RS, Monsen KJ, Nelms BE, Petrini JH. (1997). hMre11 Shen X, Mizuguchi G, Hamiche A, Wu C. (2000). A chromatin and hRad50 nuclear foci are induced during the normal remodelling complex involved in transcription and DNA cellular response to DNA double-strand breaks. Mol Cell processing. Nature 406: 541–544. Biol 17: 6087–6096. Shiloh Y. (2006). The ATM-mediated DNA-damage response: Matsouka S, Ballif BA, Smogorzewska A, McDonald III ER, taking shape. Trends Biochem Sci 31: 402–410. Hurov KE, Luo J et al. (2007). ATM and ATR substrate Stracker TH, Carson CT, Weitzman MD. (2002). Adenovirus analysis reveals extensive protein networks responsive to oncoproteins inactivate the Mre11-Rad50-NBS1 DNA DNA damage. Science 316: 1160–1166. repair complex. Nature 418: 348–352. McManus KJ, Hendzel MJ. (2005). Using quantitative imaging Stracker TH, Morales M, Couto SS, Hussein H, Petrini JH. microscopy to define the target substrate specificities of (2007). The carboxy terminus of NBS1 is required for histone post-translational-modifying enzymes. Methods 36: induction of apoptosis by the Mre11 complex. Nature 447: 351–361. 218–221.

Oncogene Recognition and signalling of DNA DSB MF Lavin 7758 Sun Y, Jiang X, Chen S, Fernandes N, Price BD. (2005). A role for van den Bosch M, Ronan T, Lowndes BN. (2003). The MRN the Tip60 histone acetyltransferase in the acetylation and complex: coordinating and mediating the response to activation of ATM. Proc Natl Acad Sci USA 102: 13182–13187. broken chromosomes. EMBO Rep 4: 844–849. Trujillo KM, Sung P. (2001). DNA structure-speciﬁc nuclease Ward JF. (1985). Biochemistry of DNA lesions. Radiat Res activities in the Saccharomyces cerevisiae Rad50 Mre11 Suppl 8: 103–111. complex. J Biol Chem 276: 35458–35464. Yazdi PT, Wang Y, Zhao S, Patel N, Lee EY, Qin J. (2002). Tsukuda T, Fleming AB, Nickoloff JA, Osley MA. (2005). SMC1 is a downstream effector in the ATM/NBS1 branch of the human S-phase checkpoint. Genes Dev 16: 571–582. Chromatin remodelling at a DNA double-strand break site Yuan SS, Chang HL, Hou MF, Chan TF, Kao YH, Wu YC Saccharomyces cerevisiae Nature in . 438: 379–383. et al. (2002). Neocarzinostatin induces Mre11 phosphorylation Usui T, Petrini JH, Morales M. (2006). Rad50S alleles of the and focus formation through an ATM- and NBS1-dependent Mre11 complex: questions answered and questions raised. mechanism. Toxicology 177: 123–130. Exp Cell Res 312: 2694–2699. Zhang J, Bao S, Furumai R, Kucera KS, Ali A, Dean NM Uziel T, Lerenthal Y, Moyal L, Andegeko Y, Mittleman L, et al. (2005). Protein phosphatase 5 is required for Shiloh Y. (2003). Requirement of the MRN complex for ATR-mediated checkpoint activation. Mol Cell Biol 25: ATM activation by DNA damage. EMBO J 22: 5612–5621. 9910–9919. Valerie K, Povirk LF. (2003). Regulation and mechanisms of mam- Zhou BB, Elledge SJ. (2000). The DNA damage response: malian double-strand break repair. Oncogene 22: 5792–5812. putting checkpoints in perspective. Nature 408: 433–439.

Oncogene