DNA Damage-Induced Activation of p53 by the Checkpoint Kinase Chk2

See allHide authors and affiliations

Science  10 Mar 2000:
Vol. 287, Issue 5459, pp. 1824-1827
DOI: 10.1126/science.287.5459.1824


Chk2 is a protein kinase that is activated in response to DNA damage and may regulate cell cycle arrest. We generated Chk2-deficient mouse cells by gene targeting. Chk2−/− embryonic stem cells failed to maintain γ-irradiation–induced arrest in the G2 phase of the cell cycle. Chk2−/−thymocytes were resistant to DNA damage–induced apoptosis. Chk2−/− cells were defective for p53 stabilization and for induction of p53-dependent transcripts such as p21 in response to γ irradiation. Reintroduction of the Chk2 gene restored p53-dependent transcription in response to γ irradiation. Chk2 directly phosphorylated p53 on serine 20, which is known to interfere with Mdm2 binding. This provides a mechanism for increased stability of p53 by prevention of ubiquitination in response to DNA damage.

The maintenance of genomic integrity after DNA damage depends on the coordinated action of the DNA repair system and cell cycle checkpoint controls. The failure of such controls leads to genomic instability and a predisposition to cancer (1). Chk2 is a mammalian homolog of the Saccharomyces cerevisiae Rad53 and Schizosaccharomyces pombe Cds1 checkpoint genes. Chk2 is a protein kinase that acts downstream of ataxia teleangiecstasia mutated (ATM) and may regulate cell cycle arrest (2–4).

To investigate the physiological role of Chk2, we generated Chk2-deficient cells by gene targeting in embryonic stem (ES) cells (5). The genotype of Chk2−/− ES clones was confirmed by Southern (DNA) blotting (Fig. 1A), and complete loss of Chk2 protein in Chk2−/− cells was confirmed by protein immunoblotting (Fig. 1B) (6). Because Chk2 is thought to have a role in the prevention of entry into mitosis (2–4), we examined arrest of the cell cycle in ES cells, a cell type in which γ irradiation induces arrest in the G2 phase but not arrest in the G1 phase or apoptosis (7). Twelve hours after 10 grays (Gy) of γ irradiation, about 90% of both Chk2+/+ and Chk2−/− ES cells were arrested with a G2 DNA content (Fig. 1C). However, at later time points, substantially more Chk2−/− cells entered G1 and S relative to controls. Examination of only S phase cells marked by a bromodeoxyuridine (BrdU) pulse label gave similar results (Fig. 1C, middle panel). We additionally investigated whether cells were able to enter M phase after γ irradiation by treatment with nocodazole, a microtubule-disrupting agent that traps cells in mitosis (8,9). In the absence of γ irradiation, about 35% of cells of both genotypes were trapped in mitosis after 12 hours of nocodazole treatment (Fig. 1D). When cells were subjected to γ irradiation with nocodazole treatment, cells of both genotypes arrested in G2 for 12 hours (Fig. 1D). However, after 18 hours, significantly more Chk2−/− cells entered mitosis. p53−/− ES cells did not show a defect in cell cycle arrest after γ irradiation (Fig. 1C), as previously shown (7), indicating that the defect in G2arrest observed in Chk2−/− ES cells is p53-independent.

Figure 1

Failure of maintenance of γ irradiation–induced G2 arrest in Chk2−/− ES cells. (A) Gene targeting of Chk2. Southern blot of Hind III– and Nhe I–digested genomic DNA from Chk2+/+(+/+), Chk2+/ (+/−), and Chk2−/− (−/−) ES cells hybridized to the 5′ flanking probe. (B) Protein immunoblot showing the expression of mouse Chk2 in ES cells either left untreated or subjected to γ irradiation (IR) (10 Gy, 3 hours). β-Actin, loading control. (C) Kinetics of cell cycle progression after γ irradiation. Wild-type and Chk2−/− (left) or p53−/− (R. Jaenisch, Whitehead Institute) (right) ES cells were subjected to 10 Gy of γ irradiation and stained with propidium iodide (PI; Sigma) as described (11) at the indicated times. The S cells were labeled with 10 μM BrdU for 45 min and then washed with phosphate-buffered saline, followed by 10 Gy of γ irradiation. DNA content profiles for BrdU-positive cells are shown (middle). (D) Mitotic indices of ES cells treated either with nocodazole alone (left) or with 10 Gy of γ irradiation plus nocodazole (right). The mitotic index was determined at the indicated times. Results shown represent the mean ± SD of five independent experiments. *, P < 0.01 (t test). •, Chk2−/−; ○, Chk2+/ ; □, Chk2+/+. (E) Kinase activity of Cdc2 in Chk2+/+ and Chk2−/− ES cells at the indicated times after 10 Gy of γ irradiation. In vitro kinase assays were done as described (8). Phosphorylation of histone H1 (top) and protein levels of Cdc2 coimmunoprecipitated with cyclin B1 (bottom) are shown.

Consistent with these results, asynchronous cells irradiated with 10 Gy showed a reduction in Cdc2-associated H1 kinase activity after 12 hours. However, Chk2−/− cells were unable to maintain this reduction (Fig. 1E). These results indicate that Chk2 is required for the maintenance, but not the initiation, of G2 arrest induced by DNA damage. Furthermore, they support our previous model, in which Chk2 functions to inhibit Cdc2 activation through inactivation of Cdc25C (2–4).

In S. cerevisiae, Rad53 controls the transcriptional response to DNA damage, a role played by p53 in mammalian cells (1). If Rad53's role is conserved, one might expect Chk2 to function upstream of p53. To test this hypothesis, we evaluated p53-dependent thymocyte apoptosis (10). To obtain Chk2−/− thymocytes, we generated somatic chimeras by Rag1−/− blastocyst complementation (11). The size and cellularity of thymi of Chk2−/− chimeras (95 × 106 ± 6 × 106 cells) were comparable to those of age-matched 129/J mice (101 × 106 ± 9 × 106 cells) and Chk2+/+ chimeras (105 × 106 ± 8 × 106 cells). Flow cytometric analyses of Chk2−/− thymocytes revealed normal development of CD4+ and CD8+ cells (Fig. 2A). Lymph nodes and spleens of Chk2−/− chimeras had normal numbers and ratios of CD4+ and CD8+ TCRαβ+ T cells (Fig. 2A), indicating that, unlike ATM, Chk2 is not critical for T cell development.

Figure 2

Impaired DNA damage–induced apoptosis of Chk2−/− thymocytes. (A) Reconstitution of the T cell compartment by Rag complementation. Thymocytes and spleen cells were isolated from Chk2+/+/Rag1−/− and Chk2−/−/Rag1−/− chimeric mice and stained with antibody to CD4 (PharMingen) and antibody to CD8 (PharMingen). Flow cytometric analysis was performed as described (11). Percentages of positive cells within a quadrant are indicated. (B and C) Apoptosis of thymocytes in response to γ irradiation. Thymocytes from Chk2+/+/Rag1−/− and Chk2−/−/Rag1−/− chimeras and p53−/− mice were treated with γ irradiation at the indicated dose. Cells were stained either with annexin V and PI after 24 hours (B) or with antibodies to CD4 or CD8 and 7-AAD after 24 and 48 hours (C). The percentage of viable cells (negative population for both annexin and PI) for each sample is shown in (B), whereas the percentage of CD4+CD8+ cells that remain viable is shown in (C). Open bars, Chk2+/+; filled bars, Chk2−/−; hatched bars, p53−/−. (D) Apoptosis of thymocytes treated with various stimuli (12), followed by staining with annexin V and PI. Values are normalized to the number of viable cells remaining in untreated cultures derived from the same animal. ○, Chk2+/+; •, Chk2−/−; ▵, p53−/−. Data are representative of three independent trials for each experiment. No differences between Chk2+/+ chimeras and wild-type littermates of p53−/− mice were observed.

To investigate the role of Chk2 in apoptosis, thymocytes from Chk2+/+ and Chk2−/− chimeras and from p53−/− mice were treated with various apoptotic stimuli (12). Immature wild-type CD4+CD8+thymocytes were highly susceptible to γ irradiation in a dose-dependent manner (Fig. 2, B and C) (13). Only 14 and 5% of wild-type thymocytes were viable after 24 and 48 hours of treatment, respectively, with 4 Gy of γ irradiation. In contrast, 68% (24 hours) and 40% (48 hours) of Chk2−/− thymocytes were viable after exposure to the same amount of γ irradiation (Fig. 2C). Chk2−/− thymocytes were also resistant to apoptosis induced by other agents that create double-strand DNA breaks, such as adriamycin (Fig. 2D), but not to apoptotic stimuli that induce single-strand DNA breaks or that activate pathways other than those mediating DNA damage, such as ultraviolet (UV) irradiation, dexamethasone, FasL, or staurosporine (Fig. 2D).

This pattern is similar to that observed in p53−/−thymocytes (14, 15), suggesting that Chk2 may function in the same pathway as p53 during DNA damage–induced apoptosis. To test this, we monitored p53 abundance in γ irradiated Chk2−/− thymocytes. γ irradiation of 5 Gy induced a substantial increase in the amount of p53 protein in Chk2+/+ cells, whereas Chk2−/− cells showed no increase (Fig. 3A). Similar results were obtained with activated T cells from Chk2−/− spleens (Fig. 3B) and primary mouse embryonic fibroblasts (MEFs) derived from Chk2−/− ES cells (Fig. 3C) (16). Increased amounts of Bax and p21, which are targets for p53 expressed during apoptosis and G1 arrest, respectively, triggered by DNA damage (17, 18), were observed in wild-type but not Chk2−/− or p53−/− cells exposed to γ irradiation (Fig. 3, A and E). In contrast, UV irradiation resulted in increased amount of p53 in both wild-type and Chk2−/−MEFs (Fig. 3C, right panel), consistent with results obtained for ATM−/− cells (10). This result indicates that, like ATM, Chk2 is not required for increased amount of p53 in response to UV irradiation. Furthermore, Chk2 protein was phosphorylated normally in response to γ irradiation in p53−/− cells (Fig. 3A), indicating that phosphorylation of Chk2 does not depend on p53 function.

Figure 3

Regulation of p53 activation by Chk2. (A) Protein immunoblots of p53, Chk2, and Bax protein levels in wild-type, Chk2−/−, and p53−/−thymocytes after 5 Gy of γ irradiation for the indicated times. Cell lysate proteins were fractionated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE), followed by immunoblotting with antibodies to p53, Chk2, Bax, or β-actin (control). (B) Protein immunoblot of p53 in wild-type and Chk2−/− activated T cells after 5 Gy of γ irradiation for the indicated times. Activated spleen T cells (2 × 106/ml) were activated by incubation with concanavalin A (5 μg/ml; Sigma) for 48 hours before irradiation. (C) Protein immunoblot of p53 in wild-type and Chk2−/− primary MEFs after 10 Gy of γ irradiation (left) or 60 J/m2 of UV irradiation (right) for the indicated times. (D) Protein immunoblot of p53 in Chk2−/− primary MEFs infected for 2 days with retrovirus carrying either the Chk2 gene or GFP (control) followed by 10 Gy of γ irradiation for the indicated times. (E) Northern blots of wild-type and Chk2−/− primary MEFs (left) and Chk2−/− primary MEFs infected for 2 days with retrovirus carrying either the Chk2 gene or GFP (right), followed by 10 Gy of γ irradiation for the indicated times. Ten micrograms of total RNA isolated by TRIZOL Reagent (Gibco) was transferred to membrane (GeneScreen Plus; NEN Life Science Products) and hybridized with mouse p21 and β-actin cDNAs.

To confirm that Chk2 regulates expression of p53 after γ irradiation, we introduced the Chk2 gene or a green fluorescent protein (GFP) control construct into Chk2−/− primary MEFs by retroviral transfer (19). In the presence of Chk2, increased expression of p53 and p21 in response to γ irradiation was restored (Fig. 3, D and E). These findings indicate that Chk2 acts upstream of p53, regulating the activation of p53 induced by DNA damage.

Phosphorylation of p53 Ser15 and Ser20 has been detected in response to DNA damage. Ser15 is phosphorylated by ATM (20,21), whereas the kinase for Ser20 is unknown. Ser20 phosphorylation but not that of Ser15 is required to increase the abundance of p53 in response to DNA damage (22, 23). Ser20 lies directly on the surface used by Mdm2 to bind p53 and target it for ubiquitination, and phosphorylation of Ser20 interferes with Mdm2 binding (22). To investigate the mechanism for p53 regulation by Chk2, we performed in vitro kinase assays. Chk2 phosphorylated Ser20 on p53, whereas ATM phosphorylated Ser15 (Fig. 4, A and B). These results suggest that the increase of abundance of p53 after DNA damage is regulated directly by Chk2 phosphorylation in response to ionizing radiation. These results could not be confirmed in vivo because the antibody that recognizes human phosphorylated Ser20does not recognize mouse Ser23, which is the equivalent of human Ser20.

Figure 4

Chk2 phosphorylates Ser20 on p53. (A) Phosphorylation of p53 by Chk2. Chk2 was immunoprecipitated from 293T cells with antibody to Chk2 as described (2) and then incubated with (+) or without (−) GST-p53(1-80) (30) in 50 mM tris-HCl (pH 7.5) and 10 mM MgCl2, 25 μM adenosine triphosphate (ATP), and 10 μCi [γ-32P]ATP for 30 min at 30°C. Proteins were resolved by SDS-PAGE and visualized by autoradiography. (B) Phosphorylation of Ser20 on p53 by Chk2. Chk2 immunoprecipitated from 293T cells (Chk2) was incubated with (+) or without (−) GST-p53(1-80) in 50 mM tris-HCl (pH 7.5), 10 mM MgCl2, and 25 μM ATP for 30 min at 30°C. FLAG-tagged ATM (ATM) immunoprecipitated with antibody to FLAG from transfected 293T cells was incubated with GST-p53(1-80) as described (20,21). Proteins were resolved by SDS-PAGE and transferred onto nitrocellulose membranes. The membranes were immunoblotted with antibody to phosphorylated Ser20(anti-P-S20) (31) or antibody to phosphorylated Ser15 (anti-P-S15) (New England Biolab). GST-p53 was visualized by reprobing the membranes with antibodies to GST (anti-GST). (C) A model for the role of Chk2 in activation of p53 in response to DNA damage.

Chk2 has been implicated in checkpoint control through its regulation by ATM and its ability to phosphorylate Cdc25C on an inhibitory residue (2–4). In ES cells, we found that Chk2 is required for the maintenance but not for the establishment of G2 arrest in response to DNA damage. In human colorectal cells or fibroblasts, p53, p21, and 14-3-3sigma have been shown to control maintenance as well (8, 24). However, because p53 loss does not affect G2 arrest in ES cells, Chk2 is acting in a p53-independent manner to control arrest. This may work through Cdc25C regulation. There must be an additional mechanism to initiate the arrest and that is likely to operate through Chk1, which can also phosphorylate and inhibit the function of Cdc25C (3, 25).

Our results also show that Chk2 activates the key tumor suppressor p53 after DNA damage. Various forms of cellular stress induce marked posttranslational increases in the amount of p53 protein. ATM functions upstream of p53 in vivo (10) and can phosphorylate p53 directly in vitro (20, 21). However, it has remained unclear whether ATM alone regulates p53 activation in vivo by direct phosphorylation or whether other molecules are also involved. Our study demonstrates that activation of p53 by DNA damage is impaired in the absence of Chk2 and is consistent with the model shown in Fig. 4C. In response to ionizing radiation, ATM is activated and results in the activation of Chk2. Chk2 can then phosphorylate and inhibit Cdc25C, contributing to maintenance of G2 for cells in S or G2, and phosphorylate p53 on Ser20, which prevents Mdm2 binding and results in p53 stabilization. ATM can also phosphorylate p53 on Ser15, which is required for activation of p53 as a transcription factor and may act synergistically with Ser20 phosphorylation (23). In principle, Ser15 phosphorylation could facilitate p53 phosphorylation by Chk2 by a priming-like event. Activated p53 then induces transcription of Bax and other genes to initiate apoptosis in certain cell types and induces p21 to cause inhibition of G1 Cdks and cell cycle arrest. We were unable to analyze the functionality of the G1 DNA damage checkpoint in response to γ irradiation because of technical reasons. However, given the defect in p21 induction in Chk2−/− cells, we think it is highly likely that Chk2 mutants will also be defective for this p53-dependent response.

In response to UV damage, neither ATM nor Chk2 are required for activation of p53. How this works is unknown, but we hypothesize it could be through the ATR-Chk1 pathway. Chk1 and Chk2 share substrate specificity to some degree and have been shown to play partially redundant roles in the S. cerevisiae DNA damage checkpoint and the S. pombe replication checkpoint. Furthermore, we have shown that Chk1 can phosphorylate multiple residues on the NH2-terminus of p53 (26).

Our results indicate that Chk2 is a major effector of the ATM kinase and carries out several of its functions. In addition, because of Chk2's key role in connecting p53 to the response to double-strand breaks, Chk2 is likely to be a tumor suppressor. A recent study identified two heterozygous germ line loss of function mutations in the Chk2 gene in Li-Fraumeni–like syndrome patients without p53 mutation (27). Although loss of the wild-type allele of Chk2 in tumors from these patients was not investigated, it is likely that mutant Chk2 is the causative agent in this cancer predisposition syndrome. Our results provide a mechanistic link between Chk2 and p53 to explain the phenotypic similarity of these two genetically distinct Li-Fraumeni syndrome families. Thus, like p53, Chk2 may contribute to a wide range of human cancers.

  • * Present address: Department of Immunology, Medical Institute of Bioregulation, Kyushu University, 3-1-1 Maidashi, Higashi-Ku, Fukuoka 812-8582, Japan.

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


Stay Connected to Science

Navigate This Article