REVIEWS ATM AND RELATED PROTEIN KINASES: SAFEGUARDING GENOME INTEGRITY. Yosef Shiloh

Transcription

1 ATM AND RELATED ROTEIN KINASES: SAFEGUARDING GENOME INTEGRITY Yosef Shiloh Maintenance of genome stability is essential for avoiding the passage to neoplasia. The DNAdamage response a cornerstone of genome stability occurs by a swift transduction of the DNA-damage signal to many cellular pathways. A prime example is the cellular response to DNA double-strand breaks, which activate the ATM protein kinase that, in turn, modulates numerous signalling pathways. ATM mutations lead to the cancer-predisposing genetic disorder ataxiatelangiectasia (A-T). Understanding ATM s mode of action provides new insights into the association between defective responses to DNA damage and cancer, and brings us closer to resolving the issue of cancer predisposition in some A-T carriers. The David and Inez Myers Laboratory for Genetic Research, Department of Human Genetics and Molecular Medicine, Sackler School of Medicine, Tel Aviv University, Tel Aviv 69978, Israel. yossih@post.tau.ac.il doi: /nrc1011 Cancer is a genetic disease of the somatic cell. This cliché makes increasing sense with the discovery of every new oncogene and tumour-suppressor gene. The road to cancer is paved with alterations in the sequence and organization of the cellular genome that range from single-nucleotide substitutions to gross chromosomal aberrations. These represent deviations from a prime requisite for cellular homeostasis: maintenance of genome stability. As the neoplastic cellular phenotype progresses, genomic stability continues to deteriorate, leading to a vicious cycle of genomic aberrations and advancing malignancy 1. Although sequence variation of germline DNA is essential for maintaining genetic variability, changes in DNA sequence in somatic cells are usually unwanted, and cells possess strict safeguards against such changes 2. The crucial event that signifies the onset of malignancy is thought to be a single genomic alteration, the outcome of which might be as subtle as a slight change in the amount of a protein, or the substitution of a single amino acid. The occurrence and fixation of such an alteration signifies the failure of a mechanism that should have detected the DNA lesion or mismatch that caused the mutation and evoked the response that is required to restore the original sequence and leave the cellular life cycle unperturbed. Sequence alterations in DNA arise from spontaneous chemical changes in DNA constituents, replication errors and damage inflicted on the DNA 3.The greatest challenge to genome stability comes from DNA-damaging agents that can be either endogenous (they form during normal cell metabolism) or exogenous (they come from the environment). Damaging agents such as radiation and reactive chemicals are capable of inducing a plethora of DNA lesions. Some are extremely cytotoxic if not repaired, whereas others are mutagenic and can affect the production, structure and function of cellular proteins, with consequences ranging from malfunction of the cell to malignant transformation (FIG. 1). It is not surprising, therefore, that many mutagens are also carcinogens, and that there is a high correlation between their carcinogenic and mutagenic potencies 4. How does the cell defend itself against this serious existential threat? The basic cellular response is to repair the damage, but the type and amount of damage might overwhelm the survival response machinery to the extent that programmed cell death (apoptosis) is initiated instead (FIG. 1). The mechanism of this important choice between attempts at survival and programmed death is not entirely clear, but here we will focus on the road taken by the cell when chances for NATURE REVIEWS CANCER VOLUME 3 MARCH

2 Summary The DNA-damage response is crucial for cellular life and for avoiding neoplasia. It occurs by rapidly transducing the damage signal to many cellular systems, such as the DNA-repair machinery and the cell-cycle checkpoints. Double-strand breaks (DSBs) in the DNA deadly DNA lesions mobilize an intricate signalling network by activating the ATM protein kinase, which, in turn, orchestrates this network by phosphorylating one or more key proteins in each of its branches. Other protein kinases related to ATM carry out similar functions in response to other genotoxic stresses, and some of them collaborate with ATM in the DSB response. ATM deficiency leads to ataxia-telangiectasia (A-T), a genomic instability syndrome, the hallmarks of which neurodegeneration, immunodeficiency, radiation sensitivity and cancer predisposition show the intimate connection between maintenance of genome stability, cellular and tissue functioning, and cancer prevention. Certain types of ATM mutations seem to increase cancer predisposition in heterozygous carriers. This adds ATM to the list of genes that have sequence variations with important implications for public health, especially with regards to cancer epidemiology. CELL-CYCLE CHECKOINTS Regulatory mechanisms that do not allow the initiation of a new phase of the cell cycle before the previous one is completed, or temporarily arrest cell-cycle progression in response to stress. DNA damage activates specific checkpoints at the G1 S and G2 M boundaries and in the S phase, with each one based on a different mechanism. survival are better when the cell activates an intricate network of pathways in order to resume its normal life cycle, unharmed. The DNA-damage response DNA damage initiates a response that switches on various repair mechanisms that recognize specific DNA lesions. These are repaired in situ or removed and the DNA is restored to its original sequence 3. A wealth of information has accumulated on DNA-repair mechanisms over the past 50 years. These mechanisms are different for different types of DNA lesions. Genetic defects that perturb these mechanisms almost invariably cause severe syndromes that are characterized by the degeneration of specific tissues (especially the nervous and immune system), sensitivity to specific DNA-damaging agents and predisposition to cancer 5 8 (BOX 1). Researchers are becoming aware, however, that the DNAdamage response is considerably broader than DNA repair, and actually encompasses additional processes 9. This is especially true for the response to highly cytotoxic DNA lesions, such as double-strand breaks (DSBs) A hallmark of this response is the activation of CELL-CYCLE CHECKOINTS 13,14. The sudden arrest of the cell cycle must involve marked alterations in numerous physiological processes. Indeed, DNA damage leads to changes in gene-expression profiles 15 17, and probably also in protein synthesis, degradation and trafficking. This means that the DNA-damage alarm must be conveyed swiftly and precisely to numerous pathways across the cell. What is the signal emanating from a DNA lesion and how does it reach so many destinations at the same time? Answers were found by studying the sophisticated damage response to DSBs, which has been conserved from yeast to mammals. This is a fine-tuned, multi-branched network of signalling pathways that act together to repair the deadly DSB, while temporarily arresting the cell cycle and balancing cellular metabolism accordingly. It turns out that this intricate network is mobilized primarily through the action of a single protein kinase ATM. The DSB model of the DNA-damage response DSBs are naturally formed and sealed during processes such as meiotic recombination and the assembly of the T-cell receptor and immunoglobulin genes via V(D)J recombination, in T cells and B cells, respectively. It is safe to assume that cellular DSB repair mechanisms (BOX 2) maintain continuous, low-level activity, ensuring that the occurrence and resealing of these breaks leave the cell unharmed. But when DSBs are inflicted on the genome by damaging agents, such as free radicals or ionizing radiation, their threat to cell life is sufficiently serious to set in motion, within minutes, a rapidly mounting, decisive DNA-damage response Recent models 64 depict the DSB response as developing through a series of steps (FIG. 2). According to these models, DSBs might first be detected by sensor proteins that recognize the DNA lesion itself or possibly chromatin alterations that follow DNA breakage. The broken ends are then processed their chemical nature is random, so they cannot serve directly as substrates for repair mechanisms. Then, the transducers are brought into action; these convey the damage signal to downstream effectors. It is this relay system from transducers to effectors that enables a single transducer to quickly affect the operation of many pathways. The transducers might also be involved in the assembly of DNA-repair complexes at the site of the damage (BOX 2), so DSB repair and signalling are probably concomitant and functionally linked (FIG. 2). In the case of DSBs, the initial and primary transducer is ATM although related protein kinases are also involved which transmits the message via a standard signalling mode: protein phosphorylation. The I3K-related protein kinase family ATM belongs to a conserved family of proteins, most of which possess a serine/threonine kinase activity Interestingly, all of these proteins contain a domain with motifs that are typical of the lipid kinase phosphatidylinositol 3-kinase (I3K) (FIG. 3), so they are dubbed I3K-like protein kinases (IKKs). The I3K domain harbours the catalytic site of the active protein kinases of the IKK family. The mammalian members of this family, at present, include five protein kinases ATM, ATR, ATX/SMG-1, mtor/fra and DNA-Kcs (FIG. 3) and TRRA, which is a component of histone acetyltransferase complexes 23,24. The active protein kinases in the family which are conserved from yeast to mammals respond to various stresses by phosphorylating key proteins in the corresponding response pathways. They could therefore simultaneously affect numerous processes, depending on the spectrum of their substrates. Four mammalian IKKs are known to be involved in the DNA-damage response: the DNAdependent protein kinase (DNA-K), ATM, ATR and ATX (REFS 17 21; and R. T. Abraham, personal communication). Whereas ATM and DNA-K respond primarily to DSBs, ATR and ATX respond to both ultraviolet (UV) light damage (possibly UV-light-induced replication arrest) and DSBs, and ATR also responds to stalled replication forks. mtor/fra is the only active kinase in this family that is not involved in responding to DNA 156 MARCH 2003 VOLUME 3

3 damage. It responds to nutrient levels and mitogenic stimuli by signalling to the protein translation machinery and probably to protein degradation systems, and it is also involved in controlling cellular growth 25,26.ATM is the prototype transducer of the DNA-damage signal. ATM: at the top of a web of DSB signalling Ataxia telangiectasia. The ATM protein was identified as the product of the gene that is mutated (lost or inactivated) in the human genetic disorder ataxia-telangiectasia (A-T) 27. A-T belongs to a group of human diseases that are collectively known as genomic instability syndromes, each of which results from a defective response to a specific DNA lesion (BOX 1). A-T is characterized by cerebellar degeneration, which leads to severe, progressive neuromotor dysfunction, immunodeficiency, genomic instability, thymic and gonadal atrophy, a striking predisposition to lymphoreticular malignancies and extreme sensitivity to ionizing radiation and DSB-inducing agents 28,29. This devastating human disorder typically combines most of the hallmarks of a defective DNAdamage response, clearly pointing to the DSB as the DNA lesion that elicits this defect. Indeed, cultured cells from A-T patients show a broad defect in responding to DSBs that span almost all of the known branches of this response 19. The striking clinical and cellular phenotype that is caused by ATM loss clearly places this protein at a top position in the DSB-response cascade. Activation of ATM. ATM resides predominantly in the nucleus in dividing cells, and responds swiftly and vigorously to DSBs by phosphorylating numerous substrates (see below). ATM-mediated phosphorylation either enhances or represses the activity of its targets, thereby affecting specific processes in which these proteins are involved. Similar to other active IKKs (with the exception of mtor/fra), ATM targets serine or threonine residues followed by glutamine (the SQ/TQ motif) 30,31. The hallmark of ATM s response to DSBs is a rapid increase in its kinase activity immediately following DSB formation 32,33. Researchers have long been impressed by the rapid phosphorylation of the many ATM substrates, which converts them within minutes to phosphorylated derivatives. A marked change in the activity of ATM would account for this massive process. Initial evidence indicated that ATM activation might involve autophosphorylation 34. A breakthrough in our understanding of this process came in a landmark publication by Bakkenist and Kastan 35.They found that ATM molecules are inactive in undamaged cells, being held as dimers or higher-order multimers. In this configuration, the kinase domain of each molecule is blocked by the FAT domain (FIG. 3) of the other. Following DNA damage, each ATM molecule phosphorylates the other on a serine residue at position 1981 within the FAT domain a phosphorylation that releases the two molecules from each other s grip, Damage type Intrinsic Spontaneous base modifications, replication errors Chemical agents Endogenous Exogenous (oxygen (chemical radicals) mutagens) Radiations Ultraviolet radiation Ionizing radiation DNA lesions Base-pair mismatches, insertions, deletions, strand breaks Base modifications, abasic sites, strand breaks DNA adducts, crosslinks, strand breaks yrimidine dimers, base modifications Abasic sites, base modifications, strand breaks Damage extent and severity Amount and type of damage can be handled Damage is excessive and/or irreparable Cellular response Activation of the survival response network Activation of the apoptotic pathway Stress responses Cell-cycle checkpoints DNA repair Low fidelity repair Mutations, chromosomal aberrations Consequence Cell survival Malignant transformation Cell death Figure 1 Cellular responses to DNA damage. Different types of DNA damage cause different types of lesions, and these, in turn, are handled by the cell in different ways. The outcome could be cell survival and resumption of the normal life cycle of the cell, cell death or malignant transformation. The mechanism of choice between attempt at survival and programmed cell death is not completely understood. The survival response is elaborate and encompasses many signalling pathways. NATURE REVIEWS CANCER VOLUME 3 MARCH

4 turning them into fully active monomers. Within minutes after the infliction of as few as several DSBs per genome, most ATM molecules become vigorously active. Bakkenist and Kastan 35 provide evidence that the signal for ATM activation might be chromatin alterations rather than direct contact of ATM with the broken DNA. However, it has previously been shown that, soon after damage infliction, a portion of the Box 1 Genomic instability syndromes Genetic defects that affect specific DNA damage response pathways lead to syndromes that combine various degrees of tissue degeneration, growth and developmental retardation, premature signs of ageing, chromosomal instability, sensitivity to the corresponding DNA-damaging agents and cancer predisposition. The prominent genomic instability syndromes 3,5 8 listed below are all autosomal recessive and represent defects in the main damage-response pathways, each of which is activated by a different class of damaging agent. Three diseases xeroderma pigmentosum, Cockayne s syndrome and trichothiodystrophy are associated with defects in the nucleotideexcision-repair (NER) pathway that deals with bulky, helix-distorting DNA lesions such as those inflicted by ultraviolet-light (UV) radiation and certain chemicals. Xeroderma pigmentosum (X) X patients are extremely sensitive to the UV-light component of sunlight and show accelerated ageing of the skin and a striking tendency to skin cancers. X is genetically heterogeneous. Its classical form entails defects in various proteins that act together in the nucleotide-excision-repair (NER) pathway. NER has a global repair arm and a transcriptioncoupled repair arm, and one or both might be defective in different X patients. Some X proteins are involved in DNA repair and are also part of the transcription complex TFIIH, and one of them is a DNA polymerase that is responsible for trans-lesion DNA synthesis. Cockayne s syndrome This UV-light sensitivity syndrome is characterized by striking dwarfism and skeletal malformations, severe mental retardation, deafness and photosensitivity, but no apparent cancer predisposition. The defective proteins are involved in transcription-coupled repair of DNA damage that is induced by UV light and probably oxidative DNA damage. Trichothiodystrophy (TTD) This rare disease is characterized by brittle hair, dry skin, dysmorphic face, mental retardation and a certain degree of photosensitivity. The mutations that are responsible for this occur in a specific region of the XB and XD genes that encode DNA helicases involved in transcription other mutations in these genes cause X or Cockayne s syndrome. This is an interesting example of different mutations within the same gene leading to different phenotypic outcomes. Bloom s syndrome (BS) BS patients have short stature, sun sensitivity, facial erythema, immunonodeficiency, decreased fertility and a predisposition to various cancers. BS cells show striking elevation of spontaneous sister-chromatid exchanges, chromosomal breakage and mild sensitivity to various DNA-damaging agents. The defective protein, BLM, is a DNA helicase that resembles the bacterial recq protein, which probably participates in homologous recombination repair and repair of damage that occurs at stalled replication forks. Werner s syndrome (WS) WS is characterized primarily by accelerated ageing, and also by a tendency for diabetes, delayed sexual development, chromosomal instability and a predisposition to various cancers. WS cells show chromosomal instability and are hypersensitive to camptothecin an inhibitor of type I DNA topoisomerase and certain DNA alkylating and crosslinking agents. Of note, both BS and WS cells are hypersensitive to hydroxyurea-induced apoptotic killing. The defective protein WRN is another DNA helicase of the recq-like family with a unique 3 5 exonuclease activity. Rothmund Thompson syndrome (RTS) RTS patients are recognized by their patchy skin coloration, stunted growth, skeletal abnormalities, early cataracts, accelerated ageing, chromosomal instability and cancer predisposition. The culprit gene RECQL4 encodes yet another recq-like DNA helicase, the role of which in the DNA-damage response is unclear. Fanconi s anemia (FA) The main characteristics of FA are bone-marrow depletion, which leads to aplastic anaemia (insufficient formation of all blood cell types), skeletal malformations, deformities of various internal organs, reduced fertility, cutaneous abnormalities and marked predisposition to myeloid leukaemias and squamous-cell carcinomas. FA cells show characteristic chromosomal aberrations and sensitivity to DNA crosslinking and DNA-breaking agents. The disease is genetically heterogeneous, with mutations affecting eight different FA genes, the products of which interact with each other and with various components of the DNA-damage response systems, such as BRCA1 and ATM (see text and recent review by D Andrea and Grompe 167 ). Recently, it was found that homozygosity for certain BRCA2 mutations is responsible for one form of FA 168. Others Additional genome instability syndromes are ataxia-telangiectasia (A-T), Nijmegen breakage syndrome (NBS) and ataxia-telangiectasia-like disease (ATLD); these are discussed in detail in the main text. 158 MARCH 2003 VOLUME 3

5 nuclear ATM is recruited to DSB sites and strongly adheres to them, possibly serving as a platform for further enzymatic reactions that take place at those sites 36. Both chromatin-bound and free ATM are autophosphorylated on serine 1981 (L. Moyal, unpublished observations). So, following DSB induction, activated ATM seems to divide between two fractions: one is chromatin bound and the other is free to move throughout the nucleus. ATM substrates. The list of published ATM substrates more than a dozen at this time is far from complete, and many ATM-dependent responses are likely to involve ATM targets that are unknown at present. However, the study of these pathways is gradually disclosing a remarkably broad cellular response to DSBs that is meticulously orchestrated by ATM. A close look at the network of ATM-mediated pathways that is responsible for activation of the Box 2 Repair of DNA double-strand breaks The repair of DNA double-strand breaks 10,11,49,50,64 (DSBs) is carried out by two different mechanisms: a rapid, error-prone mechanism dubbed nonhomologous end-joining (NHEJ) that quickly seals the breaks at the expense of creating local microdeletions, and a high-fidelity repair process based on homologous recombination (HR) between sister chromatids. These highly structured processes stem from the concerted action of several multi-protein complexes. The predominant repair mode in mammalian cells is NHEJ. The exposed ends of the DNA strands are detected by the KU70 KU80 heterodimer that recruits the catalytic subunit of the DNA-dependent protein kinase (DNA-Kcs). DNA-Kcs, in turn, might recruit additional proteins to the damaged site and probably phosphorylates some of them. The XRCC4 ligase IV heterodimer finally seals the breaks. Despite the disadvantage of its low fidelity, this pathway can act quickly, as required of an emergency mechanism, and, unlike HR, it does not depend on sister DNA molecules, which exist in the cells only after DNA replication. NHEJ is also important in sealing the breaks formed during V(D)J recombination, in which the T-cell receptor and immunoglobulin genes are rearranged 169. The high-fidelity, HR-based pathway is mediated by the RAD51-associated proteins that include several RAD51 paralogues, RAD52, the RAD54 helicase and the BRCA2 tumour suppressor. Following initial end resection and binding of RAD52 to the single-stranded end, RAD51 forms a nucleoprotein filament on the exposed strand. This process is probably initiated by a RAD51 BRCA2 complex 170,171 and is essential for the main step in this pathway: strand invasion and strand displacement. This is mediated by RAD51 and RAD54, and allows use of the undamaged sister molecule as a template for the resynthesis of the missing portions in the broken molecule. The MRE11 RAD50 NBS1 (MRN) complex (reviewed by D Amours and Jackson 64 ) seems to be essential for the HR pathway 172, probably by carrying out the initial processing of the DSB ends, and is also involved in meiotic recombination, telomere maintenance and checkpoint signalling. Following DSB induction, the MRN complex rapidly forms prominent foci at the damaged sites. These foci include additional players in the DSB response, such as RAD51 and BRCA1. MRE11 is a nuclease; RAD50 is an ATase; and the NBS1 protein, which is essential for the assembly of the complex, interacts with histone H2AX 173 and is further involved specifically in the HR pathway 172. Non-homologous end-joining (NHEJ) XRCC4 Ligase IV KU80 KU70 Inaccurately repaired DNA DNA-Kcs DSB DNA-Kcs KU80 KU70? RAD50 MRE11 NBS1 Homologous recombination (HR) Sister chromatids Nucleasemediated resection Strand invasion BRCA2 DNA synthesis; branch migration Ligation; junction resolution DSB RAD52 BRCA1 RAD51 BRCA2 RAD54 RAD51 Accurately repaired DNA DNA polymerase DNA ligase NATURE REVIEWS CANCER VOLUME 3 MARCH

6 Double-strand break (DSB) Sensors damage detection rocessed DSB Transducers Activated transducers Effectors Cell-cycle checkpoints Activation Stressresponse pathways DNA repair Cell survival Chromatin alterations rocessing enzymes (e.g. nucleases) initial processing (e.g. end resection) Repair enzymes, assembly of repair complexes Repaired DNA Figure 2 Basic steps in a current model of the cellular survival response to DSBs. The damage sensors might sense the DNA lesions themselves and/or alterations in chromatin structure in their vicinity. The activation of the transducers is probably mediated by post-translational modifications (see text). cell-cycle checkpoints (FIG. 4) explains how ATM precisely and decisively controls these pathways using several sophisticated strategies. The first strategy is to approach the same effector from several different directions. A main component in the G1 S cell-cycle checkpoint is mediated by activation and stabilization of p53, which, in turn, activates transcription of the gene that encodes the CDK2 cyclin-e inhibitor WAF1 (also known as p21 and CI1). The main target of ATM in this pathway is p53, which is phosphorylated by ATM on Ser15 (REFS 32,33,37). This contributes primarily to enhancing the activity of p53 as a transcription factor ATM also phosphorylates and activates CHK2, a checkpoint kinase 41,42 that phosphorylates p53 on Ser20. This interferes with the p53 MDM2 interaction. The oncogenic protein MDM2 is both a direct and indirect inhibitor of p53, as it serves as a ubiquitin ligase in p53 ubiquitylation, which mediates its proteasome-mediated degradation 43,44. ATM also directly phosphorylates MDM2 on Ser395, which interferes with nuclear export of the p53 MDM2 complex, and hence the degradation of p53 (REFS 38,45). Finally, it has been reported that phosphorylations of p53 on Ser9 and Ser46 (REF. 46), and dephosphorylation of Ser376 (REF. 47), are ATM dependent as well, although the function of these changes is unknown. This series of ATM-dependent modifications that activate and stabilize p53 although perhaps not complete illustrates the elaborate way in which ATM handles a single effector, and indicates that ATM might regulate several effectors within the same pathway. This principle is also seen in ATM-mediated activation of the BRCA1 tumour-suppressor protein following DNA damage. BRCA1 was found to be associated with large protein complexes that contain DSBrepair and mismatch-repair enzymes as well as ATM 48, but its mode of action in DSB repair is unclear at present. It also activates the expression of certain damage-responsive genes and is involved in the S-phase and G2 M checkpoints ATM phosphorylates BRCA1 on several sites Importantly, whereas ATMmediated phosphorylation of BRCA1 on Ser1387 activates BRCA1 as a regulator of the intra-s-phase checkpoint 56, its phosphorylation by ATM on Ser1423 spurs its involvement in the G2 M checkpoint 57.So, phosphorylation on different sites might direct an effector to act in different pathways. The CHK2 kinase, which is activated by ATM, adds yet another phosphorylation on the BRCA1 molecule 58, while at the same time ATM phosphorylates CtI an inhibitor of BRCA1 on two residues, which inhibits its function and further stimulates BRCA1 (REF. 59). ATM s hold on downstream pathways is shown not only by its multi-pronged grip of specific pathways, but also by its ability to approach the same end point from several different directions. A notable example is, again, the intra-s checkpoint: several ATM-controlled pathways converge to control this process, as it is one of the most crucial cellular responses to DSBs (FIG. 4). At least five ATM-mediated pathways seem, at present, to be involved in this checkpoint. In addition to BRCA1, ATM phosphorylates NBS1 (REFS 60 63), a component of the multi-functional MRE11 RAD50 NBS1 (MRN) complex (see BOX 2) that is involved in DSB repair 64.Ofseveral ATM phosphorylation sites on NBS1, Ser343 and Ser278 seem to be particularly important for its as yet unknown role in this checkpoint 60,61. Another ATM substrate in the intra-s checkpoint pathway is the SMC1 (structural maintenance of chromosomes 1) protein. This protein known primarily for its involvement in sister-chromatid cohesion is phosphorylated by ATM on two serine residues, and interference with this phosphorylation abrogates the S-phase checkpoint 65,66. A recently identified effector of ATM in the intra-s checkpoint is the FANCD2 protein, which is phosphorylated by ATM on Ser222 following DSB induction and undergoes BRCA1-mediated mono-ubiquitylation 67. FANCD2 is a member of a multi-protein complex, defects of which lead to another genomic instability syndrome Fanconi s anaemia (BOX 1). At the same time, CHK2 and another ATM/ATR-activated kinase, CHK1, take care of several checkpoints. Both of them phosphorylate the checkpoint phosphatase CDC25A, which marks it for degradation CDC25A s duty is to 160 MARCH 2003 VOLUME 3

7 dephosphorylate and hence maintain the activity of the cyclin-dependent kinases CDK2 and CDK1. CDK2 drives both the G1 S transition and S phase, and CDK1 mobilizes the G2 phase onto mitosis. So, the destruction of CDC25A contributes to the G1 S, intra-s and G2 M checkpoints (FIG. 4). Of interest is that the CDC25A-mediated component of the G1 S checkpoint is rapid and, unlike the p53-mediated component, is not dependent on gene activation and protein synthesis 70. Notably, in addition to ATM s versatility as a protein kinase with numerous substrates, the ATM web contains protein kinases that are themselves capable of targeting several downstream effectors simultaneously, and so concomitantly control subsets of pathways. CHK2, for example, is known to phosphorylate p53, BRCA1, CDC25A and CDC25C 41,42 (FIG. 4). It has been proposed that the CHK2 CDC25C pathway is similar to the previously mentioned CHK2 CDC25A pathway, but might act instead at the G2 M transition. Here, phosphorylation of CDC25C by activated CHK2 is thought to lead to cytoplasmic sequestration of CDC25C, which prevents activation of CDK1. Noteworthy, the CHEK2 gene that encodes this protein was suggested to act as a tumour suppressor in subsets of patients with the cancer-predisposing Li Fraumeni syndrome 73, but this contention has not been substantiated by other reports 74. Germline CHEK2 mutations have also been implicated in familial breast cancer 75,76, and CHEK2 shows somatic mutations in various sporadic solid tumours (reviewed in REFS 38,39). Most of these pathways have not been completely characterized, and the involvement of ATM substrates in them has been inferred simply from defective activation of specific checkpoints following abrogation of ATM ATR ATX/SMG1 DNA-Kcs mtor/fra TRRA FAT I3K FATC Figure 3 Size and common motifs in the human members of the IKK family. The number of residues is indicated for each protein. This family comprises six proteins, all of which (except for TRRA) possess protein kinase activity. These proteins share three motifs: the FAT and FATC domains are of unknown functional significance, but the FAT domain of ATM contains serine 1981 the site that is autophosphorylated during ATM activation (see text). The I3K domain, which contains the phosphatidylinositol 3-kinase motifs, harbours the catalytic site in the active kinases of the family. ATM-mediated phosphorylation of these proteins. An interesting pathway is mediated by phosphorylation of RAD17 the human orthologue of Rad17 from the fission yeast, Schizosaccharomyces pombe. RAD17 is phosphorylated in an ATM/ATR-mediated manner on two serine residues, and this phosphorylation was reported by different laboratories to be important for proper G1 S and G2 M checkpoints 77,78. This protein is involved in a very early event in the DSB response: it loads a trimolecular complex onto the DNA that consists of the human orthologues of the fission yeast proteins Rad9, Hus1 and Rad1 ( the complex ). This complex acts as a sliding clamp, probably a damage sensor Although this process might be regarded as being upstream of the transducer recruitment, two essential proteins in this process RAD17 and RAD9 are also phosphorylated in an ATM/ATR-dependent manner, and seem to be involved in downstream checkpoint pathways 77,78,83 (FIG. 4). It is possible that such proteins have a dual role in processes upstream and downstream of ATM. Such is the case with NBS1 as well: on one hand, it is a component of the MRN complex that is thought to be involved in the initial processing of the DSB, and, on the other hand, it is a downstream effector of ATM in a checkpoint pathway. Upstream and downstream mingle. Indeed, there is a growing notion that the DSB signal might be initially amplified by means of a cyclic process rather than by a series of steps with a linear hierarchy. One of the first processes that is initiated by DSBs is the massive phosphorylation of the tail of a histone protein variant called H2AX (reviewed by Redon et al. 84 ). Foci of phosphorylated H2AX are rapidly formed at the DSB sites and are thought to be essential for further recruitment of repair factors, such as the MRN complex, RAD51 and BRCA1 (REF. 85). H2AX phosphorylation a very early event in the cascade induced by DSBs was reported to be ATM dependent following DSB induction 86, and ATR dependent following replicative stress 87. This process could therefore serve as a rapid and powerful mechanism for amplifying the damage signal via repeated cycles of H2AX phosphorylation, and recruitment of processing factors. These facilitate further recruitment of damage transducers to the damaged sites, along with repair proteins. Another process of signal amplification might characterize the phosphorylation of certain ATM substrates. The NBS1 protein, itself an ATM substrate, seems to serve as an adaptor in the phosphorylation of other ATM substrates, such as CHK2 (REFS 88,89) and SMC1 (REFS 65,66), particularly at low damage levels. So, in this interesting mechanism, the phosphorylation of one substrate is required for the phosphorylation of others. This is in addition to the role of NBS1 in the MRN complex that acts as an early damage processor 64. Another protein that seems to facilitate the phosphorylation of certain ATM substrates is 53B1. Like NBS1, 53B1 contains a BRCT (breast cancer carboxyl terminus) domain, is recruited to DSB-induced nuclear foci, is required for proper activation of certain cell-cycle checkpoints, and is itself an ATM substrate Yet NATURE REVIEWS CANCER VOLUME 3 MARCH

8 another newly discovered, BRCT-containing protein, NFBD1, might have similar characteristics The emerging complex relationships between ATM, these three and other ATM substrates are drawing new flow charts for the DNA-damage signal that deviate from the traditional linear ones and assign to several proteins more than one role upstream or downstream in this chart. ATM a direct activator of repair? ATM s command over the damage response raises the obvious question as to whether ATM directly controls the repair process itself (BOX 2). A-T cells do show a subtle but distinct defect in DSB repair, which was ascribed to suboptimal assembly and functioning of the RAD51-associated protein complexes in the HR arm of DSB repair 96,97 (BOX 2). Activation of this pathway has been suggested to be associated with phosphorylation of the RAD51 and RAD52 proteins by the ABL tyrosine kinase 96,98,99, which physically interacts with ATM and is activated in an ATM-dependent manner following DSB induction The essential rapidity of the damage response on one hand, and the numerous functional interactions it entails on the other, indicate that the relevant proteins constantly remain in close physical proximity. Indeed, a large complex containing BRCA1, ATM and numerous repair proteins was identified in human cells 48. The physical interactions within such complexes are M G2 635 RAD (ATR) 272 RAD9 De CDK1 CDC25C 216 Ser BRCA CHK1 345 (ATR) ATM (ATR) Thr 68 CHK2 Additional TMs 15 p MDM2 WAF1 G1 Ser Ctl FANCD SMC NBS1 123 CDC25A De CDK2 S Figure 4 ATM-mediated activation of cell-cycle checkpoints in response to DSBs. Arrows indicate stimulation; T-shaped lines mark inhibition. Inhibitory phosphorylations are marked by a line through the arrow. Dephosphorylation is marked by De. The positions of the phosphorylated residues are indicated by their number. In most cases, the phosphorylated residues are serines, except for ATM-mediated phosphorylation of CHK2, which targets threonine 68. Most of these pathways are regulated primarily by ATM, particularly in the early stage of the DNA-damage response, with ATR probably becoming important at later stages to maintain these pathways. However, the pathways mediated via RAD17 and CHK1 phosphorylation are primarily under ATR control. In some of these pathways, the mechanism by which the cell-cycle machine is perturbed is not clear, but ATM-mediated phosphorylation of the indicated proteins is important for these processes. Note the different BRCA1 phosphorylations directing this protein to act at different checkpoints. 162 MARCH 2003 VOLUME 3

9 probably dynamic and respond rapidly to DNA damage. A recent example is the interaction between two players in the damage response BRCA1 and ABL which is quickly disrupted following DNA damage in an ATM-dependent manner 103. ATM and gene transcription. Many damage responses end up by modulating gene expression. Indeed, one of the main ATM targets p53 is a transcription factor. Another transcription factor that has a central role in cell-cycle control E2F1 has been reported to be phosphorylated and stabilized in an ATM-dependent manner 104. In addition, it is becoming evident that stress-response pathways that are best known for responding to other triggers in an ATM-independent manner, such as those mediated by the mitogen-activated protein kinases or the transcription factor NF-κB, also contain a DNA-damage-responsive component 14. Interestingly, when these pathways are activated by DSBs, their response becomes ATM dependent The direct targets of ATM in these pathways have not been elucidated, but these observations considerably expand the ATM-dependent network. ATM and damage-independent genomic stability. ATM helps maintain genomic stability by mechanisms other than responding to DSBs caused by damaging agents. A-T cells have abnormally shortened telomeres and show abnormal association of chromosome ends, telomere clustering and altered interactions between the telomeres and the nuclear matrix (reviewed by andita 108 ). Evidence is accumulating that ATM is functionally linked to the maintenance of telomere length and integrity, a process that is crucial to ageing and cancer Here, ATM might be responding continuously to an ongoing process rather than abruptly and vigorously to an acute insult 112. The functional relationships between ATM and the telomere-maintenance machinery have recently been illuminated by a study that generated mice that are doubly null for Atm and the telomerase RNA component (Terc) 113. These animals showed increased genomic instability, enhanced ageing and premature death, with a general proliferation defect extending into stem- and progenitor-cell compartments. Interestingly, the rate of T-cell malignancies in these mice was reduced rather than enhanced. A substrate of ATM-dependent phosphorylation in the telomere maintenance system is TRF1. TRF1 negatively regulates telomere elongation and its phosphorylation seems to suppress TRF1-mediated apoptosis following DNA damage 114. Interestingly, inhibition of TRF1 in A-T cells rescues telomere shortening and decreases the radiosensitivity of these cells, providing further evidence of the link between telomere metabolism and the DNA-damage response 115. Further evidence of ongoing activity of ATM at sites of normally occurring breakage and reunion of DNA was provided by its presence at sites of V(D)J recombination 116. p53 phosphorylated at the ATM target residue was also present at those sites, indicating that ATM continuously surveys the V(D)J recombination process, and can induce a damage response if DSBs are not sealed in a timely manner. Indeed, A-T patients often show clonal translocations involving the sites of the T-cell receptor and immunoglobulin genes, showing the consequences of absence of this surveillance. These translocations often herald the onset of lymphoid malignancy 117. Cross-talk among genome instability syndromes Defective responses to specific types of DNA lesions underlie the genome instability syndromes (BOX 1).Not surprisingly, the proteins that are defective in clinically similar syndromes were found to maintain functional interactions. Such is the case for A-T and disorders that are caused by deficiencies of the different members of the MRN complex. As mentioned above, the NBS1 protein is an ATM substrate in the intra-s checkpoint pathway (FIG. 4), and the gene that encodes it is mutated in patients with Nijmegen breakage syndrome (NBS) 118. NBS is characterized by immunodeficiency, small head, mental deficiency, genomic instability, radiation sensitivity and acute predisposition to lymphoid malignancies. The NBS phenotype shows significant overlap with that of A-T, with the important exception that NBS patients do not have cerebellar degeneration. Another member of the MRN complex RAD50 was recently found to be mutated in a patient with an NBS phenotype (T. Doerk, personal communication). Importantly, however, patients with defective MRE11 show yet another phenotype that is even closer to that of A-T. This disorder was therefore called A-T-like disease (ATLD) 119. ATLD patients develop most of the hallmarks of A-T, albeit onset at a later age and with slower progression. The different phenotypes that are associated with deficiencies in NBS1 and MRE11 might indicate that these two proteins have specific roles in the DNA-damage-response pathway, with different functional links to ATM. Furthermore, the striking similarity between A-T and ATLD raises the possibility that MRE11 deficiency might somehow affect ATM activation and hence cause defective phosphorylation of ATM substrates. If this contention proves to be true, it will place the MRN complex upstream of ATM in the DNA-damage response, in addition to NBS1 s role downstream of ATM as an ATM target. Less expected have been the functional links that were recently reported between the ATM and the MRN complex, and the Fanconi s anaemia (FA) proteins. The ATM FANCD2 connection (FIG. 4) links the DSBresponse network to the FA complex, known primarily for its involvement in the response to DNA interstrand crosslinks 120,167. Conversely, the MRN complex known for its involvement in DSB repair has recently been linked to the DNA crosslinks response, as cells from NBS patients are hypersensitive not only to DSB-inducing agents, but also to chemicals that induce interstrand crosslinks 121. Furthermore, in response to this damage, the MRN complex forms distinct nuclear foci that are similar to those found in the radiation response, but that also contain some of the FA proteins 121,122. Finally, ATM was recently reported to phosphorylate the BLM protein, a DNA helicase that is defective in Bloom s syndrome 123 NATURE REVIEWS CANCER VOLUME 3 MARCH

10 (BOX 1). Another link between BLM and the IKK family was indicated by findings that the phosphorylation of BLM following replication arrest is dependent on ATR 124. Furthermore, the same study also provided evidence that the relocalization of the MRN complex at sites of replication arrest requires functional BLM. These examples highlight functional links between apparently separate signalling pathways that might finally converge to a unified, global DNA-damage response. Other IKK-family members ATR. ATM-mediated pathways mainly depend on ATM during the initial, rapid phase of the damage response, which lasts for 1 2 hours. These pathways are therefore not completely abolished in A-T cells, but rather are attenuated and are turned on slowly later. Such a pattern indicates that other protein kinases act with ATM, but at different kinetics and with different vigour, and their activity is uncovered in the absence of ATM. Not surprisingly, one of these protein kinases is thought to be ATR, which shares several substrates with ATM 20. Suppression of endogenous ATR activity using overexpression of recombinant, inactive ATR indicated that ATR is responsible for the phosphorylation of certain ATM substrates at late stages following damage induction 125,126. So, in the DSB response, ATM reacts first, being responsible for the immediate, rapid phase of the response, whereas ATR joins in later and maintains the phosphorylated state of specific substrates. This important redundancy adds further finesse and complexity to the fine-tuned DSB response. Taking care of the late stage of the DSB response is not, however, the main role of ATR. It also responds to UV-light treatment, stalled replication forks and hypoxia 18,20,127,128 (which do not activate ATM) by phosphorylating at least a subset of ATM substrates, among them p53 and BRCA1 (REFS 125,129). Indeed, ATR was recently shown to preferentially bind to UV-light-damaged DNA 130. ATR therefore links the ATM-mediated web to signals other than DSBs. Not all downstream effectors are shared equally by ATM and ATR. For example, whereas CHK2 is a preferred ATM substrate in the cell-cycle-checkpoint pathways, CHK1 is a preferred ATR target in the G2 M checkpoint pathway (FIG. 4). This could have significant downstream consequences for example, if ATM, but not ATR, activates CHK2, p53 might not be stabilized by phosphorylation of Ser20 in cells lacking ATM, despite the presence of ATR. Indeed, recent work has shown that ATR does not have a significant influence on p53 stabilization and the activation of the G1 S checkpoint. On the other hand, ATR is important for preventing premature chromatin condensation following DNA damage 134. This observation highlights the functional distinction between ATM- and ATRdependent pathways and indicates that the relationship between them might be less straightforward than merely serving the same purpose at a different time. Another important ATR substrate is the previously mentioned RAD17 (REF. 126) (FIG. 4). ATR-mediated phosphorylation of RAD17 and CHK1 is dependent on the complex member HUS1 (REFS 126,135), although recruitment of the complex to chromatin does not depend on ATR or on the phosphorylation by ATR, of RAD17, which is the loader of the clamp onto chromatin 126. The binding of ATR and the complex therefore seem to be independent of each other, but once they are bound to the damaged site, the complex might act to enable ATR to recognize its substrates on the chromatin. Recent work based on analysis in Xenopus egg extracts indicates that ATR might be involved in the important replication checkpoint that continuously ensures that mitosis is not initiated before replication is completed 136. Involvement in such a crucial process might explain why, unlike ATM, loss of ATR is embryonic-lethal in mice 137,138 and is not compatible with cellular viability in culture 139. Interestingly, in Xenopus extracts, the activation of the ATR-dependent damage checkpoint requires initiation of DNA replication 140. Further support for the importance of ATR in the response to replicative stress came recently from a study showing that deficiency of ATR, but not of ATM, markedly destabilizes common chromosomal fragile sites, the fragility of which are usually expressed under such stress 141. Another important characteristic of ATR is its need for an accessory protein, ATRI (ATR-interacting protein). ATR and ATRI maintain physical interaction and depend on each other for their stability, and ATRI is one of the immediate substrates of ATR following its activation 139. This system is conserved through evolution, as ATRI is the functional homologue of the fission yeast Rad26 protein, which interacts with the ATR orthologue Rad3 (REF. 142). In budding yeast, the ATR orthologue, Mec1p, needs the ATRI equivalent, Ddc2/ie1/Lcd1, for its recruitment to the damaged sites on the DNA and its subsequent activation ATX/SMG1. The ATX/SMG1 protein is the most recent addition to the IKK family. Interestingly, the first physiological process that was found to involve this protein was not the DNA-damage response, but mrna surveillance, also known as nonsense-mediated mrna decay (NMD). NMD is a conserved surveillance mechanism that eliminates mrna species that contain premature termination codons, thereby preventing the production of truncated proteins 148. The human SMG1 protein is an important component of this system it phosphorylates another key player in NMD, the hupf1/smg2 protein 149,150. Unexpected of a protein associated with NMD, but perhaps not surprising given its I3K motif, human ATX/SMG1 is also involved in the DNA-damage response. Recent work from R.T. Abraham s laboratory (personal communication) indicates that ATX/SMG1 also phosphorylates Ser15 of p53, in response to UV light and ionizing radiation treatment. Deficiency of this protein causes a rapid decline in cell viability even in the absence of extrinsic genotoxic stress. These exciting findings raise important questions about the nature of the functional link between NMD and the DNA-damage response, and the degree of overlap/collaboration between this newcomer to the IKK family and the veteran members. 164 MARCH 2003 VOLUME 3

11 DNA-K. DNA-K had been recognized as a protein kinase that was activated by DSBs long before the other IKK-family members were identified. Despite its apparently narrow spectrum of substrates, investigation of this robust protein kinase has helped to shape the concept of protein kinases as damage transducers. DNA-K is part of the non-homologous end-joining arm of DSB repair 151 (BOX 1). The DNA-K holoenzyme contains a large (450-kDa) catalytic subunit (DNA- Kcs) bearing the I3K motif, and two smaller accessory proteins, KU70 and KU80, which form a heterodimer. The KU heterodimer is thought to be a sensor of DSBs that adheres to the damaged site and recruits the catalytic subunit. DNA-Kcs, in turn, binds and brings together the broken ends 152, then undergoes autophosphorylation on several serine residues that are essential for its activation 153,154. Reduced activity of DNA-Kcs in mice leads to a profound defect of the immune system called severe combined immunodeficiency (SCID), which is coupled with extreme hypersensitivity to ionizing radiation 151. Complete loss of DNA-Kcs increases the risk of developing lymphomas, highlighting the strong link between DSB repair, maintenance of genomic integrity, the V(D)J recombination process and cancer formation. The identification of DNA-K s downstream substrates is clearly important for understanding how it transduces the DNA-damage signal, but, unfortunately, the plethora of in vitro substrates of DNA-K does not represent its in vivo targets. Two proteins the WRN DNA helicase (BOX 1) and Artemis were, however, recently reported to be phosphorylated in cells in a DNA-K-dependent manner. hosphorylation affects WRN s helicase activity, which might act in DSB repair 155,156. Artemis normally possesses a singlestranded 5 3 exonuclease activity, but following its physical interaction with DNA-K and its DNA-Kmediated phosphorylation, it becomes capable of opening the hairpin structures that are typically formed at the ends of DSBs in the course of V(D)J recombination 157. Loss of Artemis in humans causes SCID and radiation sensitivity 158. This pathway provides an elegant example of a signalling process that is initiated by a DNA lesion and ends up by activating an enzyme that is involved in the direct processing of that lesion. ATM mutations and cancer Like the other genomic instability syndromes, A-T is a cancer-predisposing disorder. Atm-deficient mice show a striking predisposition to lymphoid malignancies, particularly thymic lymphomas, to which they succumb before the age of 1 year (reviewed in REFS 17 19). However, much of the literature on ATM mutations and cancer is not about A-T patients, but is, instead, on heterozygous carriers of A-T mutations. For more than two decades, ATM has been of interest to cancer epidemiologists and the public because of observations of cancer predisposition among A-T carriers. These studies pointed to a high incidence of malignancies, particularly breast cancer, among unaffected members of A-T families (reviewed by Khanna 159 ). In view of the estimated 1 2% frequency of A-T carriers in the general population, this observation has important implications for public health. For example, should ATM mutations be regarded as a genetic marker for cancer predisposition? Should people that might potentially be exposed more than others to radiation be screened for ATM mutations? Intriguingly, epidemiological studies that were carried out in different communities to confirm this important observation have often produced conflicting results, and the extent of cancer predisposition among A-T carriers seemed to be low at best. Important findings have finally emerged that seem to help resolve this apparent conflict. A-T mutations lead, in most cases, to truncated, unstable protein products, and these ATM alleles therefore fail to produce any ATM at all. Carriers of such mutations have a reduced amount of otherwise functional ATM. Other A-T mutations lead to amino-acid substitutions (missense mutations) or in-frame deletions that produce a catalytically inactive protein. Cells of carriers of these mutations are expected to contain both functional and inactive ATM molecules in various ratios. A meticulous study of the type of ATM mutations in A-T families with high incidence of cancers disclosed a high frequency of missense mutations 160.The importance of missense mutations in predisposing carriers to cancer can be explained simply by the dominant-negative effect of the inactive version of the protein leading to a reduction in ATM function that is greater than in carriers of truncation mutations 161. This model also explains why this important phenomenon was not uniformly observed in many studies that presumably included families with the truncation mutations. An elegant experimental demonstration of this principle was obtained by Scott et al. 162, who expressed, in cells, recombinant ATM that had been manipulated to contain missense variations that were previously found in a cohort of sporadic breast cancer patients 163. Some of these mutant ATM proteins indeed suppressed the kinase activity of endogenous ATM and caused radiosensitivity and genomic instability. Strong support for this notion came from a recent study that used an animal model of A-T 164. Despite the tremendous cancer predisposition of Atm-knockout mice, animals that carry the null Atm mutations in a heterozygous state show no cancer predisposition. However, Spring et al. 165 established a knock-in mouse mutant in which an inframe deletion that was previously found to cause A-T in humans was induced. Mice homozygous for this mutation produce small amounts of inactive Atm and usually show the hallmarks of the Atm-knockout phenotype. Significantly, mice heterozygous for this mutation are predisposed to various cancers, unlike the animals that carry a single knockout allele that does not produce any protein 164. Further evidence of the importance of ATM missense mutations as cancer-causing genomic alterations came from the search for somatic ATM mutations in sporadic human tumours. Indeed, ATM was found to undergo somatic mutations in sporadic lymphoid tumours behaving like a tumour-suppressor gene in these malignancies (reviewed by Stankovic et al. 166 ). NATURE REVIEWS CANCER VOLUME 3 MARCH

12 Importantly, a significant proportion of somatic ATM mutations in these tumours are of the missense type, indicating that such mutations might indeed be important in cancer formation. These observations seem to reconcile the debate on the role of ATM mutations in genetic predisposition to cancer 159,161, and place the gene encoding the ATM protein well within the list of genes that are involved in cancer morbidity in the general population. Future directions The ATM protein stands out as a prime model of a damage signal transducer. As such, its investigation is expected to continue to yield valuable new information about the DNA-damage response, one of the cardinal processes for maintaining cellular homeostasis and keeping the cell from taking the neoplastic path. The DNA-damage response signifies a crossroads at which tissue development, ageing and cancer converge. Understanding the ATM-mediated network, from the top upstream of ATM through the mechanism of ATM activation, down to the last ATM effector will certainly have an impact on numerous fields of medicine. Moreover, the contribution of ATM gene mutations to cancer predisposition in human populations will keep this gene in the spotlight of cancer epidemiology. 1. Hahn, W. C. & Weinberg, R. A. Modelling the molecular circuitry of cancer. Nature Rev. Cancer 2, Levitt, N. C. & Hickson, I. D. Caretaker tumour suppressor genes that defend genome integrity. Trends Mol. Med. 8, Hoeijmakers, J. H. Genome maintenance mechanisms for preventing cancer. Nature 411, (2001). 4. Davidson, J. F., Guo, H. H. & Loeb, L. A. Endogenous mutagenesis and cancer. Mutat. Res. 509, Vessey, C. J., Norbury, C. J. & Hickson, I. D. Genetic disorders associated with cancer predisposition and genomic instability. rog. Nucleic Acid Res. Mol. Biol. 63, (1999). 6. Moses, R. E. DNA damage processing defects and disease. Annu. Rev. Genom. Hum. Genet. 2, (2001). 7. Thompson, L. H. & Schild, D. Recombinational DNA repair and human disease. Mutat. Res. 509, O Driscoll, M. & Jeggo,. Immunological disorders and DNA repair. Mutat. Res. 509, Rouse, J. & Jackson, S.. Interfaces between the detection, signaling, and repair of DNA damage. Science 297, Jackson, S.. Sensing and repairing DNA double-strand breaks. Carcinogenesis 23, Khanna, K. K. & Jackson, S.. DNA double-strand breaks: signaling, repair and the cancer connection. Nature Genet. 27, (2001). 12. van Gent, D. C., Hoeijmakers, J. H. & Kanaar, R. Chromosomal stability and the DNA double-stranded break connection. Nature Rev. Genet. 2, (2001). 13. Bartek, J. & Lukas, J. Mammalian G1- and S-phase checkpoints in response to DNA damage. Curr. Opin. Cell. Biol. 13, (2001). 14. earce, A. K. & Humphrey, T. C. Integrating stress-response and cell-cycle checkpoint pathways. Trends Cell Biol. 11, (2001). 15. Amundson, S. A., Bittner, M., Meltzer,., Trent, J. & Fornace, A. J., Jr. Induction of gene expression as a monitor of exposure to ionizing radiation. Radiat. Res. 156, (2001). 16. Jelinsky, S. A., Estep,., Church, G. M. & Samson, L. D. Regulatory networks revealed by transcriptional profiling of damaged Saccharomyces cerevisiae cells: Rpn4 links base excision repair with proteasomes. Mol. Cell. Biol. 20, (2000). 17. Sesto, A., Navarro, M., Burslem, F. & Jorcano, J. L. Analysis of the ultraviolet B response in primary human keratinocytes using oligonucleotide microarrays. roc. Natl Acad. Sci. USA 99, Shiloh, Y. ATM and ATR: networking cellular responses to DNA damage. Curr. Opin. Genet. Dev. 11, (2001). 19. Shiloh, Y. & Kastan, M. B. ATM: genome stability, neuronal development, and cancer cross paths. Adv. Cancer Res. 83, (2001). 20. Abraham, R. T. Cell cycle checkpoint signaling through the ATM and ATR kinases. Genes Dev. 15, (2001). 21. Durocher, D. & Jackson, S.. DNA-K, ATM and ATR as sensors of DNA damage: variations on a theme? Curr. Opin. Cell. Biol. 13, (2001). 22. Khanna, K. K., Lavin, M. F., Jackson, S.. & Mulhern, T. D. ATM, a central controller of cellular responses to DNA damage. Cell Death Differ. 8, (2001). 23. McMahon, S. B., Wood, M. A. & Cole, M. D. The essential cofactor TRRA recruits the histone acetyltransferase hgcn5 to c-myc. Mol. Cell. Biol. 20, (2000). 24. Nikiforov, M. A. et al. TRRA-dependent and TRRAindependent transcriptional activation by Myc family oncoproteins. Mol. Cell. Biol. 22, roud, C. G. Regulation of mammalian translation factors by nutrients. Eur. J. Biochem. 269, Dennis,. B. & Thomas, G. Quick guide: target of rapamycin. Curr. Biol. 12, R Savitsky, K. et al. A single ataxia telangiectasia gene with a product similar to I-3 kinase. Science 268, (1995). 28. Crawford, T. O. Ataxia telangiectasia. Semin. ediatr. Neurol. 5, (1998). 29. Becker-Catania, S. G. & Gatti, R. A. Ataxia-telangiectasia. Adv. Exp. Med. Biol. 495, (2001). 30. Kim, S. T., Lim, D. S., Canman, C. E. & Kastan, M. B. Substrate specificities and identification of putative substrates of ATM kinase family members. J. Biol. Chem. 274, (1999). 31. O Neill, T. et al. Utilization of oriented peptide libraries to identify substrate motifs selected by ATM. J. Biol. Chem. 275, (2000). 32. Banin, S. et al. Enhanced phosphorylation of p53 by ATM in response to DNA damage. Science 281, (1998). References 32, 33 and 37 provided the first complete description of an ATM target Ser15 of p53. References 32 and 33 provided the first evidence for ATM activation following DNA damage. 33. Canman, C. E. et al. Activation of the ATM kinase by ionizing radiation and phosphorylation of p53. Science 281, (1998). 34. Kozlov, S., Gueven, N., Keating, K., Ramsay, J. & Lavin, M. F. AT activates ATM in vitro: importance of autophosphorylation. J. Biol. Chem. (in the press). 35. Bakkenist, C. J. & Kastan, M. B. DNA damage activates ATM through intermolecular autophosphorylation and dimer dissociation. Nature 421, (2003). This seminal work solved the basic mechanism of ATM activation: ATM lies dormant as dimers in undamaged cells, with its kinase domain being blocked by this interaction. Rapid intermolecular phosphorylation that follows DSB induction immediately releases very active ATM monomers. 36. Andegeko, Y. et al. Nuclear retention of ATM at sites of DNA double strand breaks. J. Biol. Chem. 276, (2001). This work described the tight adherence of ATM to the damaged sites in the DNA following DSB induction: a fraction of ATM seems to be allocated to duties to be carried out at the site of the damage. 37. Khanna, K. K. et al. ATM associates with and phosphorylates p53: mapping the region of interaction. Nature Genet. 20, (1998). 38. Khosravi, R. et al. Rapid ATM-dependent phosphorylation of MDM2 precedes p53 accumulation in response to DNA damage. roc. Natl Acad. Sci. USA 96, (1999). 39. Ashcroft, M., Kubbutat, M. H. & Vousden, K. H. Regulation of p53 function and stability by phosphorylation. Mol. Cell. Biol. 19, (1999). 40. Dumaz, N. & Meek, D. W. Serine15 phosphorylation stimulates p53 transactivation but does not directly influence interaction with HDM2. EMBO J. 18, (1999). 41. McGowan, C. H. Checking in on Cds1 (Chk2): a checkpoint kinase and tumor suppressor. Bioessays 24, Bartek, J., Falck, J. & Lukas, J. CHK2 kinase: a busy messenger. Nature Rev. Mol. Cell. Biol. 2, (2001). 43. Ryan, K. M., hillips, A. C. & Vousden, K. H. Regulation and function of the p53 tumor suppressor protein. Curr. Opin. Cell. Biol. 13, (2001). 44. Oren, M. et al. Regulation of p53: intricate loops and delicate balances. Biochem. harmacol. 64, Maya, R. et al. ATM-dependent phosphorylation of Mdm2 on serine 395: role in p53 activation by DNA damage. Genes Dev. 15, (2001). 46. Saito, S. et al. ATM mediates phosphorylation at multiple p53 sites, including Ser(46), in response to ionizing radiation. J. Biol. Chem. 277, Waterman, M. J., Stavridi, E. S., Waterman, J. L. & Halazonetis, T. D. ATM-dependent activation of p53 involves dephosphorylation and association with proteins. Nature Genet. 19, (1998). 48. Wang, Y. et al. BASC, a super complex of BRCA1- associated proteins involved in the recognition and repair of aberrant DNA structures. Genes Dev. 14, (2000). 49. Venkitaraman, A. R. Cancer susceptibility and the functions of BRCA1 and BRCA2. Cell 108, Jasin, M. Homologous repair of DNA damage and tumorigenesis: the BRCA connection. Oncogene 21, Yarden, R. I., ardo-reoyo, S., Sgagias, M., Cowan, K. H. & Brody, L. C. BRCA1 regulates the G2/M checkpoint by activating Chk1 kinase upon DNA damage. Nature Genet. 30, El-Deiry, W. S. Transactivation of repair genes by BRCA1. Cancer Biol. Ther. 1, Cortez, D., Wang, Y., Qin, J. & Elledge, S. J. Requirement of ATM-dependent phosphorylation of BRCA1 in the DNA damage response to double-strand breaks. Science 286, (1999). ATM-mediated phosphorylation of BRCA1 was demonstrated in references 53 55, connecting two cancer-predisposing genes in the web of the DNAdamage response. 54. Gatei, M. et al. Role for ATM in DNA damage-induced phosphorylation of BRCA1. Cancer Res. 60, (2000). 55. Gatei, M. et al. Ataxia telangiectasia mutated (ATM) kinase and ATM and Rad3 related kinase mediate phosphorylation of Brca1 at distinct and overlapping sites. In vivo assessment using phospho-specific antibodies. J. Biol. Chem. 276, (2001). 56. Xu, B., O Donnell, A. H., Kim, S. T. & Kastan, M. B. hosphorylation of serine 1387 in Brca1 is specifically required for the Atm-mediated S-phase checkpoint after ionizing irradiation. Cancer Res. 62, Work described in references 56 and 57 showed how ATM-mediated phosphorylation of BRCA1 on different sites directs it to act at different cell-cycle checkpoints. These results show the finesse of ATM s control of downstream pathways. 57. Xu, B., Kim, S. & Kastan, M. B. Involvement of Brca1 in S-phase and G(2)-phase checkpoints after ionizing irradiation. Mol. Cell. Biol. 21, (2001). 58. Lee, J. S., Collins, K. M., Brown, A. L., Lee, C. H. & Chung, J. H. hcds1-mediated phosphorylation of BRCA1 regulates the DNA damage response. Nature 404, (2000). 59. Li, S. et al. Functional link of BRCA1 and ataxia telangiectasia gene product in DNA damage response. Nature 406, (2000). 166 MARCH 2003 VOLUME 3

13 60. Lim, D. S. et al. ATM phosphorylates p95/nbs1 in an S-phase checkpoint pathway. Nature 404, (2000). Experiments in references show ATM-mediated phosphorylation of the NBS1 protein and provided the first functional explanation of the partial overlap between A-T and NBS1, two chromosomal instability syndromes. 61. Zhao, S. et al. Functional link between ataxia-telangiectasia and Nijmegen breakage syndrome gene products. Nature 405, (2000). 62. Gatei, M. et al. ATM-dependent phosphorylation of nibrin in response to radiation exposure. Nature Genet. 25, (2000). 63. Wu, X. et al. ATM phosphorylation of Nijmegen breakage syndrome protein is required in a DNA damage response. Nature 405, (2000). 64. D Amours, D. & Jackson, S.. The MRE11 complex: at the crossroads of DNA repair and checkpoint signalling. Nature Rev. Mol. Cell. Biol. 3, Kim, S. T., Xu, B. & Kastan, M. B. Involvement of the cohesin protein, Smc1, in Atm-dependent and independent responses to DNA damage. Genes Dev. 16, Yazdi,. T. et al. SMC1 is a downstream effector in the ATM/NBS1 branch of the human S-phase checkpoint. Genes Dev. 16, Taniguchi, T. et al. Convergence of the fanconi anemia and ataxia telangiectasia signaling pathways. Cell 109, Here, an important functional link between two apparently remote chromosomal instability syndromes was firmly established by showing ATMmediated phosphorylation of one of the Fanconi s anaemia proteins following DNA damage. 68. Mailand, N. et al. Rapid destruction of human Cdc25A in response to DNA damage. Science 288, (2000). 69. Falck, J., Mailand, N., Syljuasen, R. G., Bartek, J. & Lukas, J. The ATM Chk2 Cdc25A checkpoint pathway guards against radioresistant DNA synthesis. Nature 410, (2001). 70. Falck, J., etrini, J. H., Williams, B. R., Lukas, J. & Bartek, J. The DNA damage-dependent intra-s phase checkpoint is regulated by parallel pathways. Nature Genet. 30, This work revealed two distinct mechanisms by which the important intra-s cell-cycle checkpoint is fine-tuned. 71. Mailand, N. et al. Regulation of G(2)/M events by Cdc25A through phosphorylation-dependent modulation of its stability. EMBO J. 21, Zhao, H., Watkins, J. L. & iwnica-worms, H. Disruption of the checkpoint kinase 1/cell division cycle 25A pathway abrogates ionizing radiation-induced S and G2 checkpoints. roc. Natl Acad. Sci. USA 99, Bell, D. W. et al. Heterozygous germ line hchk2 mutations in Li Fraumeni syndrome. Science 286, (1999). 74. Sodha, N. et al. Increasing evidence that germline mutations in CHEK2 do not cause Li Fraumeni syndrome. Hum. Mutat. 20, Meijers-Heijboer, H. et al. Low-penetrance susceptibility to breast cancer due to CHEK2(*)1100delC in noncarriers of BRCA1 or BRCA2 mutations. Nature Genet. 31, Vahteristo,. et al. A CHEK2 genetic variant contributing to a substantial fraction of familial breast cancer. Am. J. Hum. Genet. 71, Bao, S. et al. ATR/ATM-mediated phosphorylation of human Rad17 is required for genotoxic stress responses. Nature 411, (2001). 78. ost, S. et al. hosphorylation of serines 635 and 645 of human Rad17 is cell cycle regulated and is required for G(1)/S checkpoint activation in response to DNA damage. roc. Natl Acad. Sci. USA 98, (2001). 79. Thelen, M.., Venclovas, C. & Fidelis, K. A sliding clamp model for the Rad1 family of cell cycle checkpoint proteins. Cell 96, (1999). 80. Venclovas, C. & Thelen, M.. Structure-based predictions of Rad1, Rad9, Hus1 and Rad17 participation in sliding clamp and clamp-loading complexes. Nucleic Acids Res. 28, (2000). 81. Rauen, M., Burtelow, M. A., Dufault, V. M. & Karnitz, L. M. The human checkpoint protein hrad17 interacts with the CNA-like proteins hrad1, hhus1, and hrad9. J. Biol. Chem. 275, (2000). 82. Roos-Mattjus,. et al. Genotoxin-induced Rad9 Hus1 Rad1 (9-1-1) chromatin association is an early checkpoint signaling event. J. Biol. Chem. 277, Chen, M. J., Lin, Y. T., Lieberman, H. B., Chen, G. & Lee, E. Y. ATM-dependent phosphorylation of human Rad9 is required for ionizing radiation-induced checkpoint activation. J. Biol. Chem. 276, (2001). 84. Redon, C. et al. Histone H2A variants H2AX and H2AZ. Curr. Opin. Genet. Dev. 12, Celeste, A. et al. Genomic instability in mice lacking histone H2AX. Science 296, Burma, S., Chen, B.., Murphy, M., Kurimasa, A. & Chen, D. J. ATM phosphorylates histone H2AX in response to DNA double-strand breaks. J. Biol. Chem. 276, (2001). 87. Ward, I. M. & Chen, J. Histone H2AX is phosphorylated in an ATR-dependent manner in response to replicational stress. J. Biol. Chem. 276, (2001). 88. Buscemi, G. et al. Chk2 activation dependence on Nbs1 after DNA damage. Mol. Cell. Biol. 21, (2001). 89. Girard,. M., Riballo, E., Begg, A. C., Waugh, A. & Jeggo,. A. Nbs1 promotes ATM dependent phosphorylation events including those required for G1/S arrest. Oncogene 21, DiTullio, R. A. Jr et al. 53B1 functions in an ATMdependent checkpoint pathway that is constitutively activated in human cancer. Nature Cell Biol. 4, Fernandez-Capetillo, O. et al. DNA damage-induced G2-M checkpoint activation by histone H2AX and 53B1. Nature Cell Biol. 4, Wang, B., Matsuoka, S., Carpenter,. B. & Elledge, S. J. 53B1, a mediator of the DNA damage checkpoint. Science 298, eng, A. & Chen,. L. NFBD1, like 53B1, is an early and redundant transducer mediating Chk2 phosphorylation in response to DNA damage. J. Biol. Chem. 24 Jan 2003 [epub ahead of print]. 94. Shang, Y. L., Bodero, A. J. & Chen,. L. NFBD1, a novel nuclear protein with signature motifs of FHA and BRCT, and an internal 41 amino acid repeat sequence, is an early participant in DNA damage response. J. Biol. Chem. 9 Dec 2002 [epub ahead of print]. 95. Xu, X. & Stern, D. F. NFBD1/KIAA0170 is a chromatinassociated protein involved in DNA damage signaling pathways. J. Biol. Chem. 23 Dec 2002 [epub ahead of print]. 96. Chen, G. et al. Radiation-induced assembly of Rad51 and Rad52 recombination complex requires ATM and c-abl. J. Biol. Chem. 274, (1999). 97. Morrison, C. et al. The controlling role of ATM in homologous recombinational repair of DNA damage. EMBO J. 19, (2000). 98. Yuan, Z. M. et al. Regulation of Rad51 function by c-abl in response to DNA damage. J. Biol. Chem. 273, (1998). 99. Kitao, H. & Yuan, Z. M. Regulation of ionizing radiationinduced Rad52 nuclear foci formation by c-abl-mediated phosphorylation. J. Biol. Chem. 277, Shafman, T. et al. Interaction between ATM protein and c-abl in response to DNA damage. Nature 387, (1997) Baskaran, R. et al. Ataxia telangiectasia mutant protein activates c-abl tyrosine kinase in response to ionizing radiation. Nature 387, (1997) Shangary, S. et al. Regulation of DNA-dependent protein kinase activity by ionizing radiation-activated abl kinase is an ATM-dependent process. J. Biol. Chem. 275, (2000) Foray, N. et al. Constitutive association of BRCA1 and c-abl and its ATM-dependent disruption after irradiation. Mol. Cell. Biol. 22, Lin, W. C., Lin, F. T. & Nevins, J. R. Selective induction of E2F1 in response to DNA damage, mediated by ATMdependent phosphorylation. Genes Dev. 15, (2001) Bar-Shira, A. et al. ATM-dependent activation of the gene encoding MA kinase phosphatase 5 by radiomimetic DNA damage. Oncogene 21, Wang, X. et al. Involvement of the MKK6-p38gamma cascade in gamma-radiation-induced cell cycle arrest. Mol. Cell. Biol. 20, (2000) Li, N. et al. ATM is required for IkappaB kinase (IKKk) activation in response to DNA double strand breaks. J. Biol. Chem. 276, (2001) andita, T. K. The role of ATM in telomere structure and function. Radiat. Res. 156, (2001) Maser, R. S. & Deinho, R. A. Connecting chromosomes, crisis, and cancer. Science 297, Kim, S. H., Kaminker,. & Campisi, J. Telomeres, aging and cancer: in search of a happy ending. Oncogene 21, Chan, S. W. & Blackburn, E. H. New ways not to make ends meet: telomerase, DNA damage proteins and heterochromatin. Oncogene 21, Karlseder, J., Broccoli, D., Dai, Y., Hardy, S. & de Lange, T. p53- and ATM-dependent apoptosis induced by telomeres lacking TRF2. Science 283, (1999) Wong, K. K. et al. Telomere dysfunction and Atm deficiency compromises organ homeostasis and accelerates ageing. Nature 22 Jan 2003 [epub ahead of print]. Mice doubly deficient for Atm and the telomerase RNA component showed a marked proliferative defect, genome stability and premature ageing, pointing at the importance of the defect in telomere maintenance in some A-T symptoms Kishi, S. et al. Telomeric protein in2/trf1 as an important ATM target in response to double strand DNA breaks. J. Biol. Chem. 276, (2001) Kishi, S. & Lu, K.. A critical role for in2/trf1 in ATMdependent regulation. Inhibition of in2/trf1 function complements telomere shortening, radiosensitivity, and the G(2)/M checkpoint defect of ataxia-telangiectasia cells. J. Biol. Chem. 277, erkins, E. J. et al. Sensing of intermediates in V(D)J recombination by ATM. Genes Dev. 16, Xu, Y. ATM in lymphoid development and tumorigenesis. Adv. Immunol. 72, (1999) Tauchi, H., Matsuura, S., Kobayashi, J., Sakamoto, S. & Komatsu, K. Nijmegen breakage syndrome gene, NBS1, and molecular links to factors for genome stability. Oncogene 21, Stewart, G. S. et al. The DNA double-strand break repair gene hmre11 is mutated in individuals with an ataxiatelangiectasia-like disorder. Cell 99, (1999). Here, the molecular basis for A-T-like disease (ATLD) MRE11 deficiency was identified. This discovery and the striking clinical similarity between A-T and ATLD point to intimate functional relationships between ATM and the MRN complex, which might be related to the very early events in the DSB response Joenje, H. & atel, K. J. The emerging genetic and molecular basis of Fanconi anaemia. Nature Rev. Genet. 2, (2001) Nakanishi, K. et al. Interaction of FANCD2 and NBS1 in the DNA damage response. Nature Cell Biol. 4, References 121 and 122 show the functional interaction between FANCD2 and NBS1 in the interstrand crosslink repair pathway, adding another thread to the damage-response network and drawing an additional line connecting different chromosomal instability syndromes ichierri,., Averbeck, D. & Rosselli, F. DNA cross-linkdependent RAD50/MRE11/NBS1 subnuclear assembly requires the Fanconi anemia C protein. Hum. Mol. Genet. 11, Beamish, H. et al. Functional link between BLM defective in Bloom s syndrome and the ataxia-telangiectasiamutated protein, ATM. J. Biol. Chem. 277, The finding that the Bloom s syndrome gene product, BLM, is an ATM target further intensifies the crosstalk between the chromosomal breakage syndromes Franchitto, A. & ichierri,. Bloom s syndrome protein is required for correct relocalization of RAD50/MRE11/NBS1 complex after replication fork arrest. J. Cell Biol. 157, Tibbetts, R. S. et al. A role for ATR in the DNA damageinduced phosphorylation of p53. Genes Dev. 13, (1999) Zou, L., Cortez, D. & Elledge, S. J. Regulation of ATR substrate selection by Rad17-dependent loading of Rad9 complexes onto chromatin. Genes Dev. 16, Hammond, E. M., Denko, N. C., Dorie, M. J., Abraham, R. T. & Giaccia, A. J. Hypoxia links ATR and p53 through replication arrest. Mol. Cell. Biol. 22, Heffernan, T.. et al. An ATR- and Chk1-dependent S checkpoint inhibits replicon initiation following UVCinduced DNA damage. Mol. Cell. Biol. 22, Tibbetts, R. S. et al. Functional interactions between BRCA1 and the checkpoint kinase ATR during genotoxic stress. Genes Dev. 14, (2000) Unsal-Kacmaz, K., Makhov, A. M., Griffith, J. D. & Sancar, A. referential binding of ATR protein to UV-damaged DNA. roc. Natl Acad. Sci. USA 99, Zhao, H. & iwnica-worms, H. ATR-mediated checkpoint pathways regulate phosphorylation and activation of human Chk1. Mol. Cell. Biol. 21, (2001). NATURE REVIEWS CANCER VOLUME 3 MARCH

14 132. Guo, Z., Kumagai, A., Wang, S. X. & Dunphy, W. G. Requirement for Atr in phosphorylation of Chk1 and cell cycle regulation in response to DNA replication blocks and UV-damaged DNA in Xenopus egg extracts. Genes Dev. 14, (2000) Lopez-Girona, A. et al. Serine-345 is required for Rad3- dependent phosphorylation and function of checkpoint kinase Chk1 in fission yeast. roc. Natl Acad. Sci. USA 98, (2001) Nghiem,., ark,. K., Kim, Y. S., Desai, B. N. & Schreiber, S. L. ATR is not required for p53 activation but synergizes with p53 in the replication checkpoint. J. Biol. Chem. 277, Weiss, R. S., Matsuoka, S., Elledge, S. J. & Leder,. Hus1 acts upstream of chk1 in a mammalian DNA damage response pathway. Curr. Biol. 12, Hekmat-Nejad, M., You, Z., Yee, M. C., Newport, J. W. & Cimprich, K. A. Xenopus ATR is a replication-dependent chromatin-binding protein required for the DNA replication checkpoint. Curr. Biol. 10, (2000) de Klein, A. et al. Targeted disruption of the cell-cycle checkpoint gene ATR leads to early embryonic lethality in mice. Curr. Biol. 10, (2000) Brown, E. J. & Baltimore, D. ATR disruption leads to chromosomal fragmentation and early embryonic lethality. Genes Dev. 14, (2000) Cortez, D., Guntuku, S., Qin, J. & Elledge, S. J. ATR and ATRI: partners in checkpoint signaling. Science 294, (2001). The discovery of ATRI an ATR-associated protein that is essential for its stability and activity was important for understanding how ATR is set in motion following DNA damage Lupardus,. J., Byun, T., Yee, M. C., Hekmat-Nejad, M. & Cimprich, K. A. A requirement for replication in activation of the ATR-dependent DNA damage checkpoint. Genes Dev. 16, Casper, A. M., Nghiem,., Arlt, M. F. & Glover, T. W. ATR regulates fragile site stability. Cell 111, ATR but not ATM was found to be important for preventing the expression of chromosomal fragile sites in the face of replicative stress. This draws another important distinction between ATM and ATR and points out that ATR is involved in the checkpoint associated with DNA replication rather than with DNA damage Edwards, R. J., Bentley, N. J. & Carr, A. M. A Rad3 Rad26 complex responds to DNA damage independently of other checkpoint proteins. Nature Cell Biol. 1, (1999) Rouse, J. & Jackson, S.. LCD1: an essential gene involved in checkpoint control and regulation of the MEC1 signalling pathway in Saccharomyces cerevisiae. EMBO J. 19, (2000) aciotti, V., Clerici, M., Lucchini, G. & Longhese, M.. The checkpoint protein Ddc2, functionally related to S. pombe Rad26, interacts with Mec1 and is regulated by Mec1- dependent phosphorylation in budding yeast. Genes Dev. 14, (2000) Kondo, T., Wakayama, T., Naiki, T., Matsumoto, K. & Sugimoto, K. Recruitment of Mec1 and Ddc1 checkpoint proteins to double-strand breaks through distinct mechanisms. Science 294, (2001) Wakayama, T., Kondo, T., Ando, S., Matsumoto, K. & Sugimoto, K. ie1, a protein interacting with Mec1, controls cell growth and checkpoint responses in Saccharomyces cerevisiae. Mol. Cell. Biol. 21, (2001) Rouse, J. & Jackson, S.. Lcd1p recruits Mec1p to DNA lesions in vitro and in vivo. Mol. Cell 9, Maquat, L. E. & Carmichael, G. G. Quality control of mrna function. Cell 104, (2001) Yamashita, A., Ohnishi, T., Kashima, I., Taya, Y. & Ohno, S. Human SMG-1, a novel phosphatidylinositol 3-kinaserelated protein kinase, associates with components of the mrna surveillance complex and is involved in the regulation of nonsense-mediated mrna decay. Genes Dev. 15, (2001) Denning, G., Jamieson, L., Maquat, L. E., Thompson, E. A. & Fields, A.. Cloning of a novel phosphatidylinositol kinaserelated kinase: characterization of the human SMG-1 RNA surveillance protein. J. Biol. Chem. 276, (2001) Smith, G. C. & Jackson, S.. The DNA-dependent protein kinase. Genes Dev. 13, (1999) DeFazio, L. G., Stansel, R. M., Griffith, J. D. & Chu, G. Synapsis of DNA ends by DNA-dependent protein kinase. EMBO J. 21, Chan, D. W. et al. Autophosphorylation of the DNA-dependent protein kinase catalytic subunit is required for rejoining of DNA double-strand breaks. Genes Dev. 16, Work described in references 153 and 154 shows, for the first time, that DNA-K activation by DNA damage involves autophosphorylation Douglas,. et al. Identification of in vitro and in vivo phosphorylation sites in the catalytic subunit of the DNAdependent protein kinase. Biochem. J. 368, Yannone, S. M. et al. Werner syndrome protein is regulated and phosphorylated by DNA-dependent protein kinase. J. Biol. Chem. 276, (2001) Karmakar,. et al. Werner protein is a target of DNAdependent protein kinase in vivo and in vitro, and its catalytic activities are regulated by phosphorylation. J. Biol. Chem. 277, Ma, Y., annicke, U., Schwarz, K. & Lieber, M. R. Hairpin opening and overhang processing by an Artemis/DNAdependent protein kinase complex in nonhomologous end joining and V(D)J recombination. Cell 108, Here, an important DNA-K substrate was identified after a long search for physiological substrates of this central player in DSB repair. It is the Artemis protein, which is involved in processing the hairpin structures at the site of DSBs formed during V(D)J recombination Moshous, D. et al. Artemis, a novel DNA double-strand break repair/v(d)j recombination protein, is mutated in human severe combined immune deficiency. Cell 105, (2001) Khanna, K. K. Cancer risk and the ATM gene: a continuing debate. J. Natl Cancer Inst. 92, (2000) Stankovic, T. et al. ATM mutations and phenotypes in ataxiatelangiectasia families in the British Isles: expression of mutant ATM and the risk of leukemia, lymphoma, and breast cancer. Am. J. Hum. Genet. 62, (1998) Gatti, R. A., Tward, A. & Concannon,. Cancer risk in ATM heterozygotes: a model of phenotypic and mechanistic differences between missense and truncating mutations. Mol. Genet. Metab. 68, (1999). A model was suggested to reconcile the apparently conflicting data on cancer predisposition of A-T carriers. The model attributes different degrees of cancer predisposition to different hererozygous ATM mutations. Recent results seem to substantiate this model Scott, S.. et al. Missense mutations but not allelic variants alter the function of ATM by dominant interference in patients with breast cancer. roc. Natl Acad. Sci. USA 99, Dork, T. et al. Spectrum of ATM gene mutations in a hospital-based series of unselected breast cancer patients. Cancer Res. 61, (2001) Spring, K. et al. Mice heterozygous for mutation in Atm, the gene involved in ataxia-telangiectasia, have heightened susceptibility to cancer. Nature Genet. 32, This work provided strong support for the model associating different ATM mutations with different degrees of cancer predisposition in heterozygotes Spring, K. et al. Atm knock-in mice harboring an in-frame deletion corresponding to the human ATM 7636del9 common mutation exhibit a variant phenotype. Cancer Res. 61, (2001) Stankovic, T. et al. ATM mutations in sporadic lymphoid tumours. Leuk. Lymph. 43, D Andrea, A. D. & Grompe, M. The Fanconi anaemia/brca pathway. Nature Rev. Cancer 3, (2003) Howlett, N. G. et al. Biallelic inactivation of BRCA2 in Fanconi anemia. Science 297, Gellert, M. V(D)J recombination: rag proteins, repair factors, and regulation. Annu. Rev. Biochem. 71, ellegrini, L. et al. Insights into DNA recombination from the structure of a RAD51 BRCA2 complex. Nature 420, Yang, H. et al. BRCA2 function in DNA binding and recombination from a BRCA2 DSS1 ssdna structure. Science 297, Tauchi, H. et al. Nbs1 is essential for DNA repair by homologous recombination in higher vertebrate cells. Nature 420, Kobayashi, J. et al. NBS1 localizes to gamma-h2ax foci through interaction with the FHA/BRCT domain. Curr. Biol. 12, Acknowledgements The author wishes to thank the members of the David and Inez Myers Laboratory for Genetic Research for their dedication to A-T research. Work in the author s laboratory is supported by research grants from the A-T Medical Research Foundation, The A-T Children s roject, The A-T Medical Research Trust, The National Institutes of Health and the Israel Ministry of Science, Culture and Sport. Online links DATABASES The following terms in this article are linked online to: LocusLink: 53B1 ABL ATM ATRI ATR ATX BLM BRCA1 BRCA2 CDC25A CDC25C CDK2 CHK1 cyclin E E2F1 FANCD2 H2AX KU70 KU80 MDM2 MRE11 mtor NBS1 NF-κB p53 I3K RAD9 RAD17 RAD50 RAD51 RAD52 RAD54 SMC1 TRF1 TRRA WAF1 OMIM: ataxia telangiectasia Bloom s syndrome Fanconi s anaemia Nijmegen breakage syndrome FURTHER INFORMATION The ATM entry in the NCBI web site: Information on A-T at the NIH site: Access to this interactive links box is free online. 168 MARCH 2003 VOLUME 3