Activation of the ATM Kinase by Ionizing Radiation and Phosphorylation of p53

See allHide authors and affiliations

Science  11 Sep 1998:
Vol. 281, Issue 5383, pp. 1677-1679
DOI: 10.1126/science.281.5383.1677


The p53 tumor suppressor protein is activated and phosphorylated on serine-15 in response to various DNA damaging agents. The gene product mutated in ataxia telangiectasia, ATM, acts upstream of p53 in a signal transduction pathway initiated by ionizing radiation. Immunoprecipitated ATM had intrinsic protein kinase activity and phosphorylated p53 on serine-15 in a manganese-dependent manner. Ionizing radiation, but not ultraviolet radiation, rapidly enhanced this p53-directed kinase activity of endogenous ATM. These observations, along with the fact that phosphorylation of p53 on serine-15 in response to ionizing radiation is reduced in ataxia telangiectasia cells, suggest that ATM is a protein kinase that phosphorylates p53 in vivo.

Ataxia telangiectasia (A-T) is a rare autosomal recessive disorder characterized by clinical manifestations that include progressive cerebellar ataxia, neuronal degeneration, hypersensitivity to ionizing radiation (IR), premature aging, hypogonadism, growth retardation, immune deficiency, and an increased risk for cancer (1). The gene mutated in A-T, ATM (ataxia telangiectasia–mutated), encodes a 370-kD protein that is a member of a family of proteins related to phosphatidylinositol 3-kinase (PI-3-K) that have either lipid or protein kinase activity. The subset of this family with the greatest identity to ATM functions in DNA repair, DNA recombination, and cell-cycle control (2, 3). Cell lines derived from A-T patients exhibit hypersensitivity to IR and defects in several IR-inducible cell-cycle checkpoints, including a diminished irradiation-induced arrest in the G1 phase of the cell-cycle mediated by the p53 tumor suppressor gene product (4, 5). In response to DNA damage, cells with wild-type ATM accumulate p53 protein and show a subsequent increase in p53 activity, whereas cells with defective ATM show a smaller increase in the amount of p53 protein in response to IR (4,6). Therefore, ATM appears to act upstream of p53 in a signal transduction pathway initiated by IR.

IR induces rapid, de novo phosphorylation of endogenous p53 at two serine residues within the first 24 amino acids of the protein, one of which was identified as Ser15(7, 8). Phosphorylation of p53 at Ser15 in response to DNA damage correlates with both the accumulation of total p53 protein as well as with the ability of p53 to transactivate downstream target genes in wild-type cells (8). Furthermore, phosphorylation of p53 on Ser15 in response to IR is diminished in cell lines derived from A-T patients, suggesting that ATM participates in this response (8).

The PI-3-K–related protein, DNA-activated protein kinase (DNA-PK) phosphorylates p53 in vitro at two different Ser-Gln motifs, Ser15 and Ser37 (9). However, cells with diminished DNA-PK activity still normally accumulate p53 protein and undergo G1 arrest in response to IR (10). We tested whether ATM might also phosphorylate p53 on Ser15and whether the activity of ATM toward p53 as a substrate is regulated by IR.

Most naturally occurring ATM mutant proteins are unstable (11). Because a catalytically inactive ATM mutant is a critical control for in vitro kinase assays, we constructed such a mutant that can be stably expressed. The putative kinase domain of ATM resides in the COOH-terminus of the protein. In related proteins, three critical amino acids within this domain are necessary for phosphotransferase activity (2, 12). Thus, a recombinant, FLAG peptide–tagged, wild-type ATM was used as a source of ATM protein, and a FLAG peptide–tagged, mutant ATM expression construct was generated in which two of the three critical amino acid residues required for catalysis were mutated (Asp2870 → Ala and Asn2875 → Lys) (13). Wild-type and mutant recombinant ATM proteins were individually expressed in 293T cells, and in vitro kinase activity was assessed (14). Equivalent amounts of wild-type (wt) and mutant (kd) ATM recombinant proteins were immunoprecipitated and incubated with [γ-32P]adenosine triphosphate (ATP) and recombinant glutathione S-transferase (GST)–conjugated p53 protein containing the first 101 amino acids of p53 (GSTp531–101) (Fig. 1A). Only the wild-type enzyme phosphorylated GSTp531–101 (Fig. 1B).

Figure 1

Phosphorylation of Ser15 of p53 by ATM and ATR/FRP1 in vitro. We transfected 293T/17 cells with expression vectors encoding FLAG-tagged wild-type (wt) or catalytically inactive (kd) ATM or ATR/FRP1. After 48 hours, ATM or ATR was immunoprecipitated with antibody to FLAG and used in an in vitro kinase assay with [γ-32P]ATP and either wt, S6A, S9A, or S15A GSTp531–101 as substrates (14). Proteins from each reaction were separated by SDS-PAGE (7% gel), transferred to nitrocellulose, and analyzed either on a PhosphorImager or by immunoblotting. (A) Amounts of FLAG-tagged ATM or ATR in each kinase reaction as measured by immunoblotting with anti-FLAG M2 (top panel) and amount of [γ-32P]phosphate incorporated into ATM or ATR during the reaction (lower panel). (B) In vitro kinase assay with wt GSTp531–101 or various mutant GSTp531–101 proteins (S6A, S9A, or S15A) as substrates (top panel). Levels of substrate protein present in each reaction were determined by immunoblotting for p53 (lower panel). The upper immunoreactive band represents phosphorylated GSTp53 fusion protein. ATM did not phosphorylate GST alone. The same exposures are shown for ATM, ATR/FRP1, and corresponding substrate proteins in all panels shown in (A) and (B).

Endogenous p53 becomes phosphorylated on Ser15and one other serine residue within the first 24 amino acids of the protein in response to IR (8). We tested whether mutation of each of the four serine residues (S6, S9, S15, S20) within the first 24 amino acids of p53 altered the ability of ATM to phosphorylate the NH2-terminus of p53. Recombinant ATM was immunoprecipitated and used to phosphorylate wt or mutant GSTp531–101. Wild-type recombinant ATM phosphorylated wt p53, Ser6 →Ala (S6A), and S9A mutant p53, but not S15A mutant p53 protein (Fig. 1B). Similar results were obtained with synthetic peptides comprising the first 24 amino acids of p53 (15). Therefore, ATM or a closely associated kinase phosphorylates GSTp531–101 exclusively on Ser15 in vitro. Wild-type ATM kinase also showed autophosphorylation in this assay (Fig. 1A). Because mutation of Asp2870 and Asn2875 within the kinase domain of ATM abolished both phosphorylation of p53 and autophosphorylation of ATM, the kinase activity observed in these assays appears to be intrinsic to the ATM protein. The DNA-PK also phosphorylates Ser15 (9), but unlike DNA-PK, ATM was dependent upon the presence of Mn2+ and did not require the addition of exogenous DNA for activity (15).

ATR/FRP-1 (ataxia telangiectasia and rad3-related/FRAP-related protein 1), another PI-3-K–related family member, may share functional overlap with ATM in cell-cycle checkpoint function (16,17). Conditional expression of catalytically inactive ATR/FRP-1 abrogates G2-M cell cycle arrest in response to IR. Furthermore, overexpression of wild-type ATR/FRP-1 complements the defective IR-inducible S-phase checkpoint in A-T cells (17). Although ATM is required for rapid phosphorylation of Ser15 in response to IR in vivo, ATM appears not to be required when cells are exposed to other genotoxic agents, such as ultraviolet (UV) radiation (8). Thus, other cellular kinases must also phosphorylate p53 on Ser15 in vivo. FLAG-tagged recombinant wt ATR/FRP-1 also showed autophosphorylation in vitro that was dependent upon the integrity of the catalytic domain. Like ATM, ATR/FRP-1 also phosphorylated p53 on Ser15 in a Mn2+-dependent manner (Fig. 1B), though ATR/FRP-1 had at least 20-fold less activity than ATM toward GSTp531–101 when assayed under identical experimental conditions (Fig. 1B). Thus, p53 appears to be a better substrate for ATM than ATR/FRP-1.

To test whether endogenous p53 required ATM for phosphorylation on Ser15 in cells treated with IR in vivo, we generated a monoclonal antibody specific for p53 phosphorylated at Ser15 (Fig. 2A). The p53 protein was immunoprecipitated from normal and A-T lymphoblasts either exposed to 5 Gy IR or treated with the proteosome inhibitor, acetyl-Leu-Leu-norleucinal (ALLN), which causes stabilization of p53 protein (8). Immunoblot analysis with the monoclonal antibody to phosphoserine-15 of p53 demonstrated that p53 became phosphorylated only in normal lymphoblasts exposed to IR (Fig. 2B) (18). Phosphoserine-15 was undetected in normal cells treated with ALLN, although they accumulated equivalent amounts of total p53 protein to those in irradiated cells. Phosphoserine-15 p53 was also undetected in the 1526 A-T line (Fig. 2B, upper panel). Thus, examination of radiation responses in ATM-mutant cells further supports this link between ATM and irradiation-induced phosphorylation of p53.

Figure 2

Posttranslational modification of p53 on Ser15 in response to ionizing radiation requires ATM. (A) Monoclonal antibodies against a chemically synthesized p53 phosphoserine-15 peptide (amino acids 9 through 22) were used to immunoblot synthetic peptides (1×, 50 μg) consisting of the first 24 amino acids of p53 with (1–24S15–P) or without (1–24) phosphoserine-15. (B) Normal WT (2184) or AT (1526) lymphoblasts were untreated (C), or were treated with 5 Gy IR (IR) or 20 μM ALLN (AL) for 90 min (18). The p53 was immunoprecipitated, subjected to SDS-PAGE (7.5% gel), transferred to nitrocellulose, and immunoblotted with the monoclonal antibody to phosphoserine-15 p53 (upper panel). Blots were then stripped and immunoblotted with antibodies to p53 (lower panel).

Activation of endogenous ATM was examined in two different normal lymphoblast cell lines exposed to 0 or 5 Gy IR (19). ATM immunoprecipitates were used to phosphorylate GSTp531–101 in vitro. Within 20 min after exposure to IR, ATM protein kinase activity toward GSTp531–101 was increased approximately twofold (Fig. 3, B and C). This appeared to be an increase in the specific activity of ATM because the amount of ATM protein did not change in response to IR (Fig. 3A). Kinase activity toward p53 substrate was minimal in immunoprecipitates from an A-T lymphoblast line (Fig. 3, A and B). The IR-induced activity associated with ATM was directed to Ser15, because the immunoprecipitated endogenous ATM from irradiated cells increased phosphorylation of Ser15 in in vitro kinase assays (Fig. 3B). Therefore, ATM kinase appears to be activated in response to IR and phosphorylates p53 on Ser15.

Figure 3

Activation of endogenous ATM kinase by ionizing radiation in vivo. The 2184 or 536 individual normal lymphoblasts or 1526 AT lymphoblasts (AT) were either untreated or treated with 5 Gy IR and harvested 20 or 60 min later. ATM was immunoprecipitated and assayed (Fig. 1) with wild-type GSTp531–101 protein as a substrate (19). (A) Amounts of ATM present in each reaction were determined by immunoblotting with anti-ATM (Ab-3) (upper panel), and the amount of radiolabel incorporated into ATM during the kinase reaction was visualized with a PhosphorImager (lower panel). (B) Amounts of [γ-32P]phosphate incorporated into GSTp531–101 during each reaction was visualized with a PhosphorImager (upper panel). Serine-15 phosphorylation of GSTp531-101 was determined by immunoblotting with anti-phosphoserine-15 p53 (lower panel). (C) The 2184 and 536 lymphoblasts were treated with IR or 10 J/m2 UV radiation as above. Endogenous ATM was immunoprecipitated and used in an in vitro kinase assay with GSTp531–101 as substrate. The amount of 32P-labeled GSTp531–101 was quantitated with a PhosphorImager and normalized to that obtained with immunoprecipitates from nonirradiated cells. Data are expressed as the mean ± standard error of five independent experiments.

Cells derived from A-T patients are not hypersensitive to UV irradiation (1, 20). Furthermore, such cells respond normally to UV with increased synthesis of p53, phosphorylation of p53 on Ser15, and activation of the stress-activated SAP kinase (JNK) pathway (6, 8, 21). The kinase activity of ATM was not increased in UV-irradiated cells (Fig. 3C). Slight activation of ATM kinase was detected at more than 60 min after exposure, which may be due to signals generated by DNA strand breaks associated with DNA repair (22). These results confirm that ATM plays a minor role in the cellular UV response and suggest that another kinase other than ATM phosphorylates p53 on Ser15 in response to UV irradiation.

Previous genetic and biochemical evidence implicated the ATM gene product in regulating the phosphorylation and induction of p53 in cells exposed to ionizing radiation (4, 6, 8, 23). Our results indicate that ATM is a protein kinase whose activity is increased by ionizing radiation and whose in vivo target may be Ser15 of p53. This conclusion is consistent with the finding that ATM and p53 proteins directly interact with each other (11). The functional ramifications of radiation-induced Ser15 phosphorylation remain to be clearly elucidated. However, phosphorylation of p53 on Ser15 reduces binding of the mdm2 oncogene product to p53 in vitro (7), and binding of mdm2 to p53 promotes rapid degradation of p53 by targeting it for proteolytic degradation, thereby potentially controlling p53 protein levels (24). Because many of the clinical manifestations exhibited by A-T patients cannot be attributed to abnormal regulation of p53 alone, other important targets of the ATM kinase remain to be identified and characterized.

  • * Present address: Department of Hematology-Oncology, St. Jude Children's Research Hospital, 332 North Lauderdale Street, D-1034, Memphis, TN 38105–2794, USA.

  • To whom correspondence should be addressed. E-mail: Michael.Kastan{at}


View Abstract

Navigate This Article