Differential Roles for Checkpoint Kinases in DNA Damage-dependent Degradation of the Cdc25A Protein Phosphatase

Abstract

In response to DNA damage, cells activate a signaling pathway that promotes cell cycle arrest and degradation of the cell cycle regulator Cdc25A. Cdc25A degradation occurs via the SCF^β-TRCP pathway and phosphorylation of Ser-76. Previous work indicates that the checkpoint kinase Checkpoint kinase 1 (Chk1) is capable of phosphorylating Ser-76 in Cdc25A, thereby promoting its degradation. In contrast, other experiments involving overexpression of dominant Chk2 mutant proteins point to a role for Chk2 in Cdc25A degradation. However, loss-of-function studies that implicate Chk2 in Cdc25A turnover are lacking, and there is no evidence that Chk2 is capable of phosphorylating Ser-76 in Cdc25A despite the finding that Chk1 and Chk2 sometimes share overlapping primary specificity. We find that although Chk2 can phosphorylate many of the same sites in Cdc25A that Chk1 phosphorylates, albeit with reduced efficiency, Chk2 is unable to efficiently phosphorylate Ser-76. Consistent with this, Chk2, unlike Chk1, is unable to support SCF^β-TRCP-mediated ubiquitination of Cdc25A in vitro. In CHK2^–/– HCT116 cells, the kinetics of Cdc25A degradation in response to ionizing radiation is comparable with that seen in HCT116 cells containing Chk2, indicating that Chk2 is not generally required for timely DNA damage-dependent Cdc25A turnover. In contrast, depletion of Chk1 by RNA interference in CHK2^–/– cells leads to Cdc25A stabilization in response to ionizing radiation. These data support the idea that Chk1 is the primary signal transducer linking activation of the ATM/ATR kinases to Cdc25A destruction in response to ionizing radiation.

In response to DNA damage and replication blocks, cells activate an elaborate signaling pathway called the DNA damage stress response pathway. This pathway is composed of two parallel but partially overlapping protein kinase cascades, the ATM5-Chk2 and ATR-Chk1 kinase modules, which are responsible for transducing damage signals to downstream effectors controlling cell division, DNA repair, and apoptosis (1–3). Previous studies indicated that the ATM kinase primarily responds to double-stranded DNA breaks, whereas ATR responds to blocks in replication in addition to DNA breaks. Recent work indicates that ATM is required to activate the ATR kinase in response to double-strand breaks, providing a mechanism for Chk1 activation in response to γ-irradiation (3). The branches of these pathways that control cell cycle arrest are called checkpoints. Primary targets of these checkpoints are cyclin-dependent kinases, major regulators of the cell division cycle. Cdk2, in complexes with G₁- and S-phase cyclins, is a primary target of the DNA damage response. Activation of p53 in response to DNA damage leads to induction of the p21 Cdk inhibitory protein, which blocks Cdk activity (4, 5). In addition, pathways also exist to control the activities of a family of Cdc25 protein phosphatases, which normally remove inhibitory phosphorylation events from Cdks (6–11). Three Cdc25 proteins, A, B, and C, exist in mammals. Cdc25B and -C are not required for mouse development or checkpoint function (12), and inhibition of these phosphatases by DNA damage primarily involves sequestration mechanisms. In contrast, Cdc25A interacts with a highly complex regulatory system, and its degradation via the ubiquitin-proteasome pathway is a primary control pathway both in dividing cells and in response to DNA damage (13). Degradation of Cdc25A in response to DNA damage is critical for cell cycle arrest; cells that are unable to degrade Cdc25A display defects in checkpoint arrest, including radiation-resistant DNA synthesis and elevated levels of Cdk activity (14–17).

Pathways controlling the degradation of Cdc25A in the cell cycle and in response to multiple types of DNA damage have been explored in detail (13). In certain cancers, overproduction of Cdc25A is reported to stem from inactivation of GSK3β in G₁ that primes delivery of Cdc25A to dedicated degradation machinery (18). Indeed, overexpression of Cdc25A promotes transformation in mice, and mice lacking one copy of the CDC25A gene have lower rates of transformation in vitro and in vivo (19, 20). Multiple lines of evidence indicate a role for Chk1 in turnover of Cdc25A in S-phase or G₂ arrest mediated by clinically relevant topoisomerase inhibitors or in response to UV (21–23). UCN-01, a Chk1 inhibitor, blocks Cdc25A degradation in S-phase, and loss of Chk1 by RNAi leads to defects in Cdc25A degradation in response to topoisomerase inhibitors as well as IR (15, 24). Chk1 phosphorylates multiple sites in Cdc25A, including Ser-124 and Thr-507, both of which are capable of leading to recruitment of 14-3-3 protein (6, 25–28). Recruitment of 14-3-3 appears to be the primary mechanism employed to inhibit Cdc25A function during G₂ and mitosis. In contrast, phosphorylation of Cdc25A on Ser-76 by Chk1 promotes its ubiquitination by the SCF^β-TRCP ubiquitin ligase during S-phase (16, 17, 29). SCF complexes are modular ubiquitin ligases that employ F-box proteins to recruit phosphorylated substrates to a Cul1/Rbx1 catalytic core (30–33). Phosphorylation of Ser-76 is thought to serve as a priming phosphorylation that facilitates phosphorylation of Cdc25A on Ser-82 by an as yet unidentified kinase (16). Phosphorylation of Ser-82 recruits β-TRCP, thereby promoting Cdc25A ubiquitination and degradation by the proteasome. Consistent with this model, depletion of β-TRCP by RNAi leads to Cdc25A stabilization, defects in the intra-S phase checkpoint, and increased Cdk2 activity (16, 17). Interestingly, none of the phosphorylation sites in Cdc25A linked to 14-3-3 binding (Ser-124 and Thr-507) (28) are required for Chk1-dependent ubiquitination of Cdc25A in vitro nor are they required for recruitment of β-TRCP in vivo (16).

Although evidence of a role for Chk1 in this process is strong, a previous report suggested that Chk2 is the primary kinase responsible for controlling Cdc25A degradation in response to ionizing radiation via phosphorylation of Ser-124 (14). The kinetics of Chk2 activation more closely mimicked Cdc25A degradation than did Chk1 activation, as assessed by alterations in electrophoretic mobility. In addition, overexpression of Chk2 mutant proteins (including a catalytically defective mutation D347A as well as a mutant in the forkhead-associated domain derived from a sporadic colon cancer (R145W) or from a germ line mutation in cancer-prone Li-Fraumeni syndrome (I157T)) led to stabilization of Cdc25A in response to DNA damage (14). This study also examined the role of Ser-124 in degradation of Cdc25A by Chk2. Ser-124 is phosphorylated by both Chk2 and Chk1 in vitro. Cdc25A^S124A displayed reduced turnover rates when compared with Cdc25A, but the turnover rates were essentially indistinguishable in the presence or absence of ionizing radiation (IR). Thus, whereas Chk2 is considered to be important for IR-induced Cdc25A turnover via phosphorylation (14, 34), there have been no loss-of-function studies that have verified an essential role for Chk2 in Cdc25A turnover. Moreover, degradation of Cdc25A via the SCF^β-TRCP pathway requires phosphorylation of Ser-76, and mutation of Ser-124 has little effect on this process (16, 29), raising the question of whether Chk2 has any role in phosphorylation of Ser-76.

In this work we have systematically examined the role of Chk2 in phosphorylation, ubiquitination, and degradation of Cdc25A. Although Chk1 and Chk2 phosphorylate many sites in Cdc25A with similar efficiency, Chk2 is unable to efficiently phosphorylate Ser-76 in Cdc25A. As such, Chk2 does not support ubiquitination of Cdc25A by SCF^β-TRCP in vitro. Moreover, HCT116 cells engineered to lack Chk2 display no defects in Cdc25A degradation in response to IR, but depletion of Chk1 in these cells leads to strong stabilization in response to ionizing radiation. Taken together, these data suggest that Chk1 is the major checkpoint kinase controlling Cdc25A degradation in response to ionizing radiation.

MATERIALS AND METHODS

Protein Expression—GST-Chk1 (7) and GST-Chk2 (11) were expressed in Sf9 cells using recombinant baculoviruses. Infected cells were cultured for 40 h before lysis in lysis buffer (25 mm Tris-HCl (pH 7.5), 1 mm dithiothreitol, 0.5% Nonidet P-40, 125 mm NaCl, 1 mm EDTA, 10 mm β-glycerol phosphate, 5 mm sodium fluoride, and 5 mm p-nitrophenylphosphate). Cleared lysates were incubated with GSH-Sepharose (GE Healthcare) and washed three times with lysis buffer. Proteins were eluted with 40 mm glutathione, 100 mm Tris-HCl (pH 87.5, containing 100 mm NaCl and then dialyzed extensively with 25 mm Tris-HCl (pH 7.5), 100 mm NaCl, 1 mm dithiothreitol, 50% glycerol before storage at –80 °C. GST-Cdc25C (residues 200–256) (7) was prepared in bacteria as described previously. Cdc25A was expressed in bacteria using Intein technology. The Cdc25A open reading frame lacking a stop codon or an analogous point mutant in which Ser-76 was replaced by alanine) was cloned into pET-C-Intein, allowing inducible expression of a Cdc25A-Intein fusion in bacteria. Protein was expressed in BL21(DE3) cells, and cells were lysed using 25 mm Tris-HCl, 0.5% Nonidet P-40, 125 mm NaCl. Cleared lysates where subjected to purification using chitinagarose (New England Biolabs), and Cdc25A was released by incubating the resin with 25 mm Tris-HCl and 125 mm NaCl containing 10 mm dithiothreitol.

Protein Kinase Assays and Peptide Mapping—Kinase reactions with 5 μg of GST-Cdc25C (residues 200–256) were performed in 25 mm Tris-HCl (pH 7.5), 10 mm MgCl₂, 50 μm ATP containing 5 μCi of [γ-³²P]ATP and the indicated quantity of purified GST-Chk1 or GST-Chk2 (30 min, 30 °C). Reaction mixtures were separated by SDS-PAGE before Coomassie staining and autoradiography. Two-dimensional tryptic peptide mapping was performed on equal numbers of cpm (∼3000) of Cdc25A phosphorylated with either GST-Chk1 or GST-Chk2 using established procedures (36).

In Vitro Ubiquitination—Cdc25A ubiquitination was performed in vitro using [³⁵S]methionine-labeled Cdc25A (2.5 μl) in the presence of insect cell-derived GST-Chk1 or GST-Chk2 as indicated, ubiquitin (1 mg/ml), 1 μm ubiquitin aldehyde, 2.3 μl of in vitro translated β-TRCP (35), and 4 mm ATP in a total volume of 10 μl (30 min, 30 °C) (16). Reaction mixtures were subjected to SDS-PAGE on a 4–12% gradient gel and visualized by autoradiography.

Antibodies—The phospho-Ser-76 antibody targeting Ser-76 in human Cdc25A was produced by a method previously described for generation of phospho-specific antibodies (36) using as antigen a synthetic keyhole limpet hemocyanin-coupled Cdc25A phosphopeptide of the sequence CNSNLQRMG-pSSEST, where pS marks the phosphorylated serine. Immunoglobulins from the immunized rabbits were purified by protein A-Sepharose; phospho-Cdc25A-reactive antibodies were then affinity-purified by selecting for antibodies that recognize phosphorylated Cdc25A using columns carrying an immobilized phosphorylated Cdc25A peptide. For immunoblotting, the indicated Cdc25A sample was subjected to electrophoresis and blotted using anti-phospho-Ser-76 antibodies (1:500) in TBST containing bovine serum albumin. Detection was accomplished using horseradish-conjugated goat anti-rabbit IgG. Blots were stripped and re-probed with an antibody directed against Cdc25A (Neomarkers, Inc.).

The Cdc25A antibody employed were either from Neomarkers (AB-3) or Santa Cruz. Chk2 antibodies were from Santa Cruz or provided by Stephen Elledge (Harvard Medical School). Chk1 (G-4) and Cdk2 (M2) antibodies were from Santa Cruz. Chk2 and Cdc25A antibodies differed between Fig. 4A and Fig. 4B.

FIGURE 4.

Cells lacking Chk2 retain their ability to degrade Cdc25A in a Chk1-dependent manner. A, HCT116 cells and HCT116 cells lacking the CHK2 gene (HCT116 CHK2^–/–) were either left untreated or treated with 10 gray (Gy) before the immediate addition of cycloheximide (CHX) to block ongoing protein translation. Cells were harvested at the indicated times after IR treatment. Cells were lysed and subjected to immunoblotting with indicated antibodies; αCul1 is used here as a loading control. B and C, HCT116 or HCT116 CHK2^–/– cells were depleted for Chk1 with two distinct small interfering RNA sequences. Forty-eight hours post-transfection cells were treated with 10 gray before the addition of cyclohexamide. Cell lysates were resolved on SDS-PAGE and analyzed using chemiluminescent imaging and CCD detection, which allowed for the quantitative analysis of relative protein levels in C. In C, turnover rates for three independent experiments are shown.

Cell Culture and Small Interfering RNA Depletion—HCT116 cells or HCT116 cells lacking the Chk2 gene (a generous gift from Dr. B. Vogelstein) (37) were maintained in McCoy's 5A media supplemented with 10% fetal bovine serum (5% CO₂). Cells were reverse-transfected in 60-mm dishes using RNAimax (Invitrogen) with individual Stealth small interfering RNAs targeting Chk1 (siChk1-1, UUCAUAUCUACAAUCUUCACUGCGA; siChk1-2, UAUACCGAAAUACUGUUGCCAAGCC) or Stealth universal negative RNAi control (Invitrogen) and then the next day split into six 35-mm dishes. To measure Cdc25A turnover in response to IR, half of the cells were subjected 48 h post-transfection to 10 gray, and then cycloheximide (25 μg/ml) was added to all cells immediately after irradiation to inhibit protein synthesis. At the indicated times cells were harvested and lysed in lysis buffer. Equal quantities of proteins were subjected to immunoblotting or chemiluminescent imaging (Alpha Innotech) using indicated antibodies. Quantification of Cdc25A levels were determined using a CCD camera by determining the number of pixels associated with the Cdc25A protein.

RESULTS AND DISCUSSION

Chk1, but Not Chk2, Efficiently Phosphorylates Ser-76 in Cdc25A in Vitro—We previously provided evidence based on mass spectrometry that Chk1 is capable of phosphorylating Ser-76 in Cdc25A (16). This phosphorylation event was found to be required for recognition of β-TRCP and for subsequent ubiquitination and degradation of Cdc25A (16, 29). Although Chk2 has been implicated in control of Cdc25A turnover (14), its ability to phosphorylate Ser-76, a key site for Cdc25A degradation, has not been examined. To examine this issue, we purified both Chk1 and Chk2 in active form from insect cells as GST fusion proteins (Fig. 1A) and then established the specific activity of these kinases toward GST-Cdc25C (residues 200–256). This well characterized Chk1/Chk2 substrate contains a single major phosphorylation site Ser-216, which is analogous to Ser-124 in Cdc25A. The phosphorylation site in this fragment (Arg-Ser-His-Ser²¹⁶-Met-Pro) conforms to motifs recognized by both Chk1 and Chk2 (38). Comparable incorporation of phosphate into GST-Cdc25C^200–256 was achieved with Chk2 levels ∼10-fold higher than Chk1 (Fig. 1B) (similar phosphate incorporation was observed with 50 nm Chk1 and 550 nm Chk2). Using similar ratios of GST-Cdc25C^200–256 kinase activity, we then examined the extent of phosphorylation of Cdc25A and the Cdc25A^S76A mutant purified from bacteria by GST-Chk1 and GST-Chk2 (Fig. 1C). Under these conditions, the extent of phosphorylation of wild type Cdc25A observed with Chk2 was ∼10-fold lower than that observed with Chk1. Importantly, the extent of phosphorylation with wild type Cdc25A was indistinguishable from that observed with the S76A mutant, indicating that sites other than Ser-76 are phosphorylated in this in vitro reaction, as expected based on previous studies.

FIGURE 1.

Phosphorylation of Cdc25A by Chk1 and Chk2 in vitro. A, GST-Chk1 and GST-Chk2 proteins purified from insect cells. Proteins were purified as described under “Materials and Methods.” Two μg of each protein was subjected to SDS-PAGE and the gel-stained with Coomassie Brilliant Blue. B, GST-Cdc25C^200–256 (5 μg) was incubated with 0.1 μg of GST-Chk1 (lane 2), 1 μg of GST-Chk2, or with the appropriate volume of control buffer (lane 1) in the presence of [³²P]ATP as described under “Materials and Methods.” Reaction mixtures were subjected to SDS-PAGE and Coomassie staining, and the dried gel was subjected to autoradiography to detect phosphorylated GST-Cdc25C^200–256. C, Cdc25A or Cdc25A^S76A (1 μg, purified from bacteria) was incubated with 100 ng of GST-Chk1 and 1 μg of GST-Chk2, conditions that provide equivalent phosphorylation of Cdc25C^200–256 in the presence of [³²P]ATP (see “Materials and Methods”). Reaction mixtures were subjected to SDS-PAGE and Coomassie staining, and the dried gel was subjected to autoradiography to detect phosphorylated Cdc25A. WT, wild type.

To examine the specificity of Cdc25A phosphorylation by Chk1 and Chk2, we then performed two-dimensional tryptic peptide mapping (39). Proteins were excised from gels and digested with trypsin, and equal quantities of radioactivity subjected to two-dimensional mapping (see “Materials and Methods”) (Fig. 2A). Cdc25A phosphorylated by Chk2 displayed a pattern of phosphorylation reminiscent of that observed with Chk1. In particular, prominent peptides seen in Chk1-phosphorylated Cdc25A (Fig. 2A, a, c, d, e, g, and h) were present in Cdc25A phosphorylated by Chk2. However, peptide b was absent from Cdc25A phosphorylated by Chk2. Peptide b was absent from Cdc25A^S76A, indicating that peptide b reflects phosphorylation of Ser-76 in Cdc25A. However, we were unable to recover sufficient amounts of peptide b for Edman degradation, making it impossible to unequivocally designate this peptide to phosphorylation of Ser-76.

FIGURE 2.

Chk1, but not Chk2 phosphorylates Ser-76 in Cdc25A. A, two-dimensional tryptic peptide mapping of Cdc25A phosphorylated by Chk1 and Chk2 in vitro. Equal counts/min (∼3000) of Cdc25A or Cdc25A^S76A phosphorylated by GST-Chk1 or GST-Chk2 (Fig. 1C) were excised from dried gels and subjected to tryptic peptide mapping as described under “Materials and Methods.” Peptides were visualized by autoradiography. The position of a peptide absent from Cdc25A^S76A is indicated with a circle (peptide b). B, an epitope recognized by an antibody against phosphoSer-76 in Cdc25A is produced in the presence of Chk1 but not Chk2. Kinase reactions were performed with the indicated quantity of Chk1 or Chk2 and 1 μg of Cdc25A (or Cdc25A^S76A) purified from bacteria. Blots were probed with anti-phospho-Ser-76 (1:500) and subsequently stripped and probed with a monoclonal antibody against Cdc25A as a control for loading. WT, wild type.

Given the results of these peptide mapping studies, we sought to further evaluate the ability of Chk2 to phosphorylate Ser-76 in Cdc25A through the use of phosphospecific antibodies directed against Ser-76. Peptide antibodies were generated in rabbits and affinity-purified. Bacterial Cdc25A alone displayed weak reactivity with the anti-phospho-Ser-76 antibodies (Fig. 2B, lane 12), but reactivity was dramatically stimulated by prior treatment with 2 nm Chk1 (lane 11). Importantly, reactivity against Chk1-phosphorylated Cdc25A was reversed upon mutation of Ser-76 to alanine (lane 10), demonstrating the specificity of the antibody for phospho-Ser-76. We next examined the ability of Chk2 to generate an epitope recognized by this antibody. Cdc25A or Cdc25A^S76A was incubated with 5 or 10 nm Chk2 and then subjected to immunoblotting with the anti-phospho-Ser-76 antibody (lanes 6–9). The signal generated by Chk2 was indistinguishable from that seen when Cdc25A alone was immunoblotted with anti-phospho-Ser-76 (compare lanes 6 and 12). Moreover, increasing the concentration of Chk2 from 5 to 10 nm had no effect on the reactivity of the anti-phospho-Ser-76 antibody. In stark contrast, 1 nm Chk1 led to a robust phosphorylation of Ser-76 (lane 3). Control experiments employing Chk1 alone indicated that autophosphorylation of Chk1 did not generate a phospho-epitope that co-migrated with Cdc25A, thereby ruling out a contribution of Chk1 to the reactivity with the anti-phospho-Ser-76 antibody (lanes 5 and 13). Taken together, these data indicate that whereas Chk2 is capable of phosphorylating Cdc25A on a subset of sites that are also phosphorylated by Chk1, it is unable to efficiently phosphorylate Ser-76, a residue required for SCF^β-TRCP-mediated degradation of Cdc25A.

Chk1, but Not Chk2, Supports SCF^β-TRCP-dependent Ubiquitination of Cdc25A in Vitro—Previously, we developed an assay for Chk1-dependent ubiquitination of Cdc25A in vitro by SCF^β-TRCP (16). Phosphorylation of Cdc25A on Ser-76 by Chk1 leads to formation of a phospho-degron that is recognized by β-TRCP, allowing ubiquitination. Available data indicate that one or more kinases present in reticulocyte extracts may be capable of phosphorylating Ser-82 in a phospho-Ser-76-dependent manner, as Chk1 alone does not phosphorylate Ser-82 (16). To investigate the ability of Chk2 to support Cdc25A ubiquitination in this assay, we phosphorylated in vitro translated Cdc25A using quantities of Chk1 and Chk2 (50 and 550 nm, respectively) that gave equivalent levels of GST-Cdc25C^200–256 phosphorylation. Under these conditions, Chk1 activity led to robust Cdc25A ubiquitination (Fig. 3). In contrast, Chk2 did not support Cdc25A ubiquitination despite the fact that at this concentration of Chk2, numerous sites in Cdc25A are phosphorylated (Fig. 2A). Moreover, the addition of Chk2 to the Chk1-dependent Cdc25A ubiquitination reaction had no effect on the extent of ubiquitin conjugates. The inability of Chk2 to support Cdc25A ubiquitination in this assay is consistent with the inability of Chk2 to phosphorylate Ser-76 in Cdc25A. Interestingly, although Chk1 led to a mobility shift in Cdc25A, Chk2 did not. In a previous study (16) we found that mutation of Ser-76 to alanine led to a large reduction in the mobility shift observed upon phosphorylation of Cdc25A by Chk1 (16). These data suggest that the majority of the Cdc25A mobility shift is because of phosphorylation of Ser-76.

FIGURE 3.

Chk1, but not Chk2, supports SCF^β-TRCP-dependent Cdc25A ubiquitination in vitro. Cdc25A ubiquitination reactions were performed as described under “Materials and Methods” in the presence of Chk1 (100 ng) or Chk2 (1 μg). Reactions mixtures were subjected to SDS-PAGE, and dried gels were visualized by autoradiography.

Depletion of Chk1 by RNAi in Cells Lacking Chk2 Stabilizes Cdc25A in Response to Ionizing Radiation—Evidence of the involvement of Chk2 in Cdc25A degradation rests largely on the phenotype of cells overexpressing Chk2-defective Chk2 mutants (14, 21). To date, no turnover studies have been performed using loss-of-function mutations. To examine this question, we took advantage of existing HCT116 colon cancer cells in which the CHK2 gene was deleted by homologous recombination (37). HCT116 cells provide an excellent system for studying checkpoint functions. The G₁/S and G₂/M checkpoints are operative in these cells, and it has been demonstrated that Chk2 is activated in these cells in response to IR (37). Moreover, targeted deletion of p53, p21, or 14-3-3σ in these cells leads to checkpoint phenotypes, confirming the integrity of DNA damage checkpoints in these cells.

HCT116 cells with endogenous CHK1 and CHK2 loci rapidly degraded Cdc25A with an approximate half-life of less than 30 min in response to ionizing radiation (10 gray) (Fig. 4A, lanes 13–15, slightly faster than observed with non-irradiated samples, lanes 9–12). Importantly, in HCT116 cells lacking Chk2 (HCT116 CHK2^–/–), Cdc25A is degraded with kinetics that are comparable with that seen with HCT116 cells in the presence or absence of DNA damage (Fig. 4, A, lanes 1–8, and C, solid and dotted black lines). Thus, degradation of Cdc25A in response to ionizing radiation does not require Chk2.

To determine the relative contributions of Chk1 and Chk2 in IR-dependent Cdc25A turnover, we depleted Chk1 in both HCT116 and HCT116 CHK2^–/– cells. Using two distinct small interfering RNAs (siChk1-1 and siChk1-2), we observed a dramatic stabilization of Cdc25A in the presence of ionizing radiation (Fig. 4B). This stabilization occurred regardless of the abundance of Chk2 in these cells (lanes 1–9). Moreover, the stabilization mediated by depletion of Chk1 was independent of, and not enhanced by the absence of Chk2 in these cells (Fig. 4B, and turnover rates for three independent experiments are plotted in 4C, solid and dotted gray lines). Note that the apparent turnover of Cdc25A with siChk1-2 seen at 60 min (lane 18) likely reflects incomplete Chk1 depletion in this particular sample and was not observed with siChk1-1 where depletion at this time point was more complete. Together, this suggests that Chk1 activity is the major influence in Cdc25A turnover after ionizing radiation, and there is only a minimal contribution by Chk2.

Distinct Roles for Chk1 and Chk2 in Control of Cdc25A in Response to Ionizing Radiation—Cdc25A degradation in response to ionizing radiation, replication blocks, and other impediments to DNA replication is a critical component of the DNA damage response and facilitates the cell cycle arrest branch of the response pathway (13). Cells overexpressing Cdc25A that cannot be degraded or cells that lack the machinery required for damage-dependent Cdc25A degradation display defects in the intra-S-phase checkpoint and display elevated levels of Cdk2 activity (16, 17). Elevated Cdk2 activity is the hallmark of a cell cycle checkpoint defect. Previous studies have implicated both the Chk1 and Chk2 branches of the DNA damage signaling system in promoting the degradation of Cdc25A after ionizing radiation. Although loss-of-function studies have been performed with Chk1 (15, 34), the evidence for involvement of Chk2 rests primarily on the use of overexpression of dominant mutant proteins (14), and no loss-of-function studies examining Cdc25A turnover have been reported.

In this paper we have examined the ability of Chk1 and Chk2 to promote phosphorylation of Cdc25A in vitro on sites that support SCF^β-TRCP-dependent ubiquitination of Cdc25A as well as the ability of these kinases to support Cdc25A ubiquitination by SCF^β-TRCP. In addition, we have examined the relative contributions of Chk1 and Chk2 in IR-dependent Cdc25A degradation in HCT116 tissue culture cells engineered to lack Chk2 in combination with RNAi against Chk1. We previously demonstrated that phosphorylation of Ser-76 in Cdc25A by Chk1 supports its ubiquitination in vitro by SCF^β-TRCP, and mutation of this site blocks damage-dependent Cdc25A turnover (16). We find that Chk2, although capable of phosphorylating several sites in Cdc25A, is nevertheless highly inefficient at phosphorylation of the key site Ser-76. Using both two-dimensional tryptic phosphopeptide mapping and phosphospecific antibodies targeting phospho-Ser-76, we found no evidence for phosphorylation of Cdc25A on Ser-76 by Chk2, even under conditions where Chk2 levels were greater than 10-fold higher than Chk1 levels sufficient to efficiently phosphorylate Ser-76 and were sufficient to phosphorylate many other sites in Cdc25A. Consistent with an inability to phosphorylate Ser-76, Chk2 was unable to support SCF^β-TRCP-dependent Cdc25A ubiquitination in an in vitro assay. Previous specificity studies using peptide substrates indicate that the primary specificities of Chk1 and Chk2 are closely related to one another. In particular, basic residues are preferred at position –3, whereas hydrophobic residues are preferred at position +1 (38). The sequence around Ser-76 in Cdc25A (Gln-Arg-Met-Gly-Ser⁷⁶-Ser-Glu) largely conforms to this preference in that an arginine residue occupies the –3 position. Currently, the structural basis for the inefficient phosphorylation of Ser-76 by Chk2 compared with Chk1 is unknown.

To examine whether Chk2 is required for rapid turnover of Cdc25A in vivo, we employed HCT116 cells engineered to lack Chk2 (37). HCT116 cells lacking an intact CHK2 gene have no discernable defect in Cdc25A degradation in response to IR. Importantly, the absence of Chk2 does not further stabilize Cdc25A any more than depletion of Chk1 alone both in the presence of damage, and two classes of Cdc25A turnover rates emerge that appear only to depend on whether Chk1 is present or depleted. Previous studies have also demonstrated that Chk2 has no obvious role in the DNA damage-dependent p53 response (37). Taken together, these data indicate a general role for Chk1, but not Chk2, in promoting Cdc25A degradation in response to IR in HCT116 cells. We have also found that Cdc25A turnover in response to ultraviolet irradiation in HCT116 cells is unaffected by loss of Chk2 (data not shown). Although it is impossible to absolutely rule out a role for Chk2 in Cdc25A in other cell types or in the presence of other types of DNA damage, the available data indicate that Chk2 is not generally required for Cdc25A phosphorylation on sites that promote its degradation via the SCF^β-TRCP pathway.