Cdk5 activator-binding protein C53 regulates apoptosis induced by genotoxic stress via modulating the G2/M DNA damage checkpoint.

In response to DNA damage, the cellular decision of life versus death involves an intricate network of multiple factors that play critical roles in regulation of DNA repair, cell cycle, and cell death. DNA damage checkpoint proteins are crucial for maintaining DNA integrity and normal cellular functions, but they may also reduce the effectiveness of cancer treatment. Here we report the involvement of Cdk5 activator p35-binding protein C53 in regulation of apoptosis induced by genotoxic stress through modulating Cdk1-cyclin B1 function. C53 was originally identified as a Cdk5 activator p35-binding protein and a caspase substrate. Importantly, our results demonstrated that C53 deficiency conferred partial resistance to genotoxic agents such as etoposide and x-ray irradiation, whereas ectopic expression of C53 rendered cells susceptible to multiple genotoxins that usually trigger G(2)/M arrest. Furthermore, we found that Cdk1 activity was required for etoposide-induced apoptosis of HeLa cells. Overexpression of C53 promoted Cdk1 activity and nuclear accumulation of cyclin B1, whereas C53 deficiency led to more cytoplasmic retention of cyclin B1, suggesting that C53 acts as a pivotal player in modulating the G(2)/M DNA damage checkpoint. Finally, C53 and cyclin B1 co-localize and associate in vivo, indicating a direct role of C53 in regulating the Cdk1-cyclin B1 complex. Taken together, our results strongly indicate that in response to genotoxic stress, C53 serves as an important regulatory component of the G(2)/M DNA damage checkpoint. By overriding the G(2)/M checkpoint-mediated inhibition of Cdk1-cyclin B1 function, ectopic expression of C53 may represent a novel approach for chemo- and radio-sensitization of cancer cells.

Mammalian cells have developed sophisticated mechanisms to cope with DNA damage induced by either normal metabolic processes or genotoxic stresses. In response to genotoxic stress, cells may either undergo cell cycle arrest and DNA repair or commit suicide if the damage is beyond repair. Among many factors that influence cellular decision-making under genotoxic stress, DNA damage checkpoints play crucial roles in regulation of cell cycle arrest and cell death. DNA damage checkpoints are biochemical pathways that delay or arrest cell cycle progression in response to DNA damage (1). Based upon the interphase transition that is inhibited by DNA damage, three major checkpoints are defined as the G 1 /S, intra-S, and G 2 /M checkpoints (2). In response to DNA damage, these checkpoints inhibit cell cycle progression from G 1 to S (the G 1 /S checkpoint), DNA replication (the intra-S checkpoint), or G 2 to mitosis (the G 2 /M checkpoint), respectively (2). Although DNA damage checkpoints are crucial for maintaining DNA integrity in normal cells, they may also present a great challenge to cancer treatment because they may arrest cell cycle and trigger DNA repair in tumor cells, thereby reducing the effectiveness of many conventional therapies that cause DNA damage. Therefore, many efforts have been made to target specific checkpoints to sensitize tumor cells to conventional therapies (3).
Orderly progression of normal cell cycle is closely monitored and tightly controlled by multiple cell cycle checkpoints. A recent study indicates that the same biochemical pathways and relevant proteins are involved in both cell cycle control and DNA damage checkpoints. One of the major players of the G 2 /M checkpoint is cyclin-dependent kinase 1 (Cdk1, 1 formerly called cdc2)-cyclin B1 complex that is absolutely required for the entry of mitosis (4 -7). Cdk1 activation is a multistep process that begins with the increase of cyclin B1 transcription during G 2 phase. Meanwhile, Cdk1 is inhibited by phosphorylation on residues Thr-14 and Tyr-15 (by Myt1 and Wee1 kinases, respectively). At the onset of mitosis, Cdk1 is activated through dephosphorylation of Thr-14 and -15 that is carried out by dual phosphatase CDC25C as well as phosphorylation on Thr-161 by Cdk-activating kinase (a heterodimer of Cdk7 and cyclin H). Once active, Cdk1-cyclin B1 complex phosphorylates CDC25C to enhance its phosphatase activity, resulting in a positive feedback loop. Meanwhile, active Cdk1-cyclin B1 translocates from cytoplasm to nucleus during early mitosis and phosphorylates nuclear substrates, leading to nuclear envelope disassembly, chromosome condensation, and separation of sister chromatids. At the end of metaphase, cyclin B1 is destroyed by anaphase promoting complex to allow mitosis to proceed. In response to DNA damage, the G 2 /M checkpoint system functions to arrest cell cycle in part by inhibiting Cdk1-cyclin B1 activity. The phosphatidylinositol 3-kinase family of protein kinases including ATM (ataxia telangiectasia-mutated) and ATR (A-T and rad3-related) transduce the signal of DNA damage and phosphorylate and activate checkpoint kinases checkpoint kinase 1 and 2, which in turn phosphorylate CDC25C (3). Phosphorylated CDC25C is sequestered by 14-3-3 away from Cdk1, thereby preventing dephosphorylation and activation of Cdk1 (8,9). Another mechanism of the G 2 /M checkpoint control is nuclear exclusion of cyclin B1 (10,11). Phosphorylation of cyclin B1 is critical for subcellular localization of Cdk1-cyclin B1 complex (10,(12)(13)(14)(15). At the onset of mitosis, phosphorylation of four serine residues (Ser-126, -128, -133, and -147) at the N-terminal region of human cyclin B1 that contains both cytoplasmic retention sequence and nuclear export sequence causes an inhibition of nuclear export and enhances nuclear import, resulting in a net accumulation of cyclin B in nucleus.
Cdk1-cyclin B1 has also been demonstrated to play a critical role in regulation of apoptosis in a number of experimental systems (16). Increased Cdk1 activity has been observed in various apoptotic conditions (17)(18)(19). Furthermore, overexpression of active Cdk1-cyclin B1 promotes mitotic catastrophe and apoptosis (10,20), and inhibition of the Cdk1-cyclin B1 complex by dominant-negative cdk1 mutant, antisense constructs, or chemical inhibitors prevents apoptosis (18,21,22). Yet the molecular mechanism of Cdk1-mediated apoptosis remains largely unclear.
In this study we report here that Cdk5 activator p35-binding protein C53 plays a critical role in regulating genotoxic stressinduced apoptosis through modulating the G 2 /M DNA damage checkpoint. By overcoming the G 2 checkpoint-mediated inhibition of Cdk1-cyclin B1 function, ectopic expression of C53 may represent a novel approach for chemo-and radio-sensitization of cancer cells.

MATERIALS AND METHODS
Tissue Culture Cells and Reagents-HeLa cells (from ATCC) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and antibiotics. TNF␣, genotoxic reagents, and chemicals were purchased from Sigma. Purified Cdk1-cyclin B1 was purchased from Upstate.
In Vitro Expression Cloning-Construction of mouse spleen cDNA small pool library and in vitro expression cloning was described (23).
cDNA Cloning, Construction of Expression Vectors, and Site-specific Mutagenesis-Human C53 cDNA was amplified using primers containing SalI and KpnI sites and EST clone BE336801 (Invitrogen) as a template. Amplified C53 cDNA was sequenced and subcloned in-frame into pCMV-5a (Sigma), and the resulting construct was used for expression of the C-terminal FLAG-tagged C53 in mammalian cells. Sitespecific mutagenesis was performed using the QuikChange site-specific mutagenesis kit (Stratagene) according to the manufacturer's protocol. Mutations were confirmed by DNA sequencing.
Production and Purification of C53 Antibodies-His-tagged murine C53 fusion protein was purified using a nickel column (Novagen) and injected into three rats to generate polyclonal antibodies from rat. Rat polyclonal antibody (used for immunostaining assay) was affinity-purified. Two rabbit polyclonal antibodies were generated against two peptides (residues 242-256, KRGNSTVYEWRTGTE, and residues 491-506, SKRYSGRPVNLMGTSL). Sera were further purified using affinity column cross-linked with respective peptides (Pierce).
Co-immunoprecipitation, Immunoblotting, Immunostaining, and Antibodies Used in This Study-For co-immunoprecipitation assay, HeLa cells (5 ϫ 10 7 ) were harvested and resuspended in the lysis buffer (10 mM HEPES, pH7.4, 150 mM NaCl, 5 mM MgCl 2 , 1 mM dithiothreitol, 1 mM EDTA, 0.2% Nonidet P-40, 50 mM NaF supplemented with protease inhibitor cocktails). After clearance, 5 g of cyclin B1 antibody was incubated with cell lysate for 2 h at 4°C followed by the addition of 10 l of protein A/G beads (Pierce). After three washes of lysis buffer, the bound proteins were eluted with 0.1 M glycine, pH 2.5. The samples were subjected to SDS-PAGE and immunoblotting.
Caspase Activity Assay-Caspase activity was analyzed by Apo-one caspase assay kit (Promega) according to the manufacturer's instruction.
Cell Death and Cell Viability Assays-For trypan blue exclusion assay, cells were trypsinized, combined and harvested with floating cells, and subsequently incubated with 0.4% trypan blue solution (Sigma) for 5 min. Blue cells were scored as dead cells. Three independent assays were performed for each experiment, and at least more than 200 cells were scored. For annexin V assay, cells (10 6 ) were partially trypsinized, harvested, and incubated with annexin V and propidium iodine (Oncogene) in the binding buffer according to the supplier's instruction. The samples were subjected to flow cytometry analysis. Cell viability was measured by CellTiter-Blue assay (Promega) according to the supplier's instruction. For ectopic expression of C53 in normal and C53-deficient cells, the constructs expressing C53 and its mutant were transfected into HeLa cells along with EGFP marker using Lipofectamine 2000 (Invitrogen). After a 24-h incubation, cells were treated with a variety of agents at the indicated concentrations for either 24 or 48 h. The percentage of cell death was scored by counting GFP-positive cells with DNA staining. Small and round cells with condensed, and fragmented nuclei were scored as dead cells, whereas flat and big cells with normal nuclei were scored as live cells. Each experiment was independently repeated as a triplicate.
Cell Cycle Analysis-Cells were trypsinized, washed in phosphatebuffered saline and fixed with 100% ethanol at Ϫ20°C. Cells were then washed in phosphate-buffered saline again and resuspended in phosphate-buffered saline containing RNase A (5 g/ml) and propidium iodide (10 g/ml). A total of 10,000 labeled nuclei were analyzed in a FACScan Flow Cytometer (BD Biosciences).
Cdk1 Kinase Assay-Cells were detached from dishes with a rubber policeman, harvested by centrifugation, and resuspended in lysis buffer (10 mM HEPES, pH 7.4, 150 mM NaCl, 5 mM MgCl 2 , 1 mM dithiothreitol, 1 mM EDTA, 0.2% Nonidet P-40, 50 mM NaF, supplemented with protease inhibitor cocktails). After 30 min of incubation on ice, lysates were clarified by centrifugation (10,000 ϫ g for 20 min at 4°C). For the histone H1 kinase assay, cell lysates (150 g) was incubated for 2 h at 4°C with anti-cyclin B1 polyclonal antibody (Santa Cruz). Protein A/G beads were then added to the lysates and incubated for 1 h. Immunoprecipitates were washed 3 times with lysis buffer, once with HBS (10 mM HEPES, pH 7.4, 150 mM NaCl). The beads were incubated with 5 g of histone H1 in HBS (20 l) containing 15 mM MgCl 2 , 50 M ATP, 1 mM dithiothreitol, and 1 Ci of [␥-32 P]ATP. After 10 and 30 min of incubation at 30°C, the reaction products were analyzed by SDS-PAGE and autoradiography.
Subcellular Fractionation-Cell pellets collected from individual dishes were resuspended in lysis buffer (20 mM Tris-HCl, pH 7.2, 150 mM NaCl, 2 mM MgCl 2 , 1 mM dithiothreitol, 0.5% Nonidet P-40, 50 mM NaF) supplemented with protease inhibitor cocktails. After 30 min of incubation on ice, lysates were centrifuged at 1000 ϫ g for 10 min at 4°C. The supernatants were cleared again at 13,000 ϫ g for 15 min at 4°C to harvest the cytosolic fraction. The pellets from the first centrifugation were washed twice with lysis buffer and centrifuged again at 1000 ϫ g for 10 min at 4°C. Pellets from this step contained the nuclear fractions.

C53 Was Involved in Modulating Apoptosis Induced by
Genotoxic Stress-C53 was originally isolated as a Cdk5 activator p35-binding protein in a yeast two-hybrid screening using p35 as bait (26) and subsequently identified as a caspase substrate. 2 To investigate the cellular function of C53 in regulation of apoptosis, we took the advantage of siRNA technique. As shown in Fig. 1A, C53 siRNA-1 inhibited overexpression of C53-FLAG fusion protein in HeLa cells. In attempt to achieve a better deficiency of C53, we used this siRNA sequence, which is identical among human, mouse, and rat C53 genes, to construct a plasmid-based shRNA expression vector (pSIREN-RetroQ-C53). After transfection of shRNA expression vector into HeLa cells, drug selection, and single clone isolation, we obtained three clones expressing extremely low levels of C53 (Fig. 1B). Vectorbased C53 shRNA specifically reduced expression of C53 protein but not other proteins such as caspases (casp-2, -3, -7, and -9), Bcl-2 family members (Bcl-2, Bcl-x L , Bim, Bak, Bax), Apaf-1, inhibitor of apoptosis (CIAP1 (cellular inhibitor of apoptosis 1) and XIAP (X-chromosome-linked inhibitor of apoptosis)), iB␣, and tubulin (Fig. 1B). 2  affect cell proliferation and cellular morphology. 3 We further examined the responses of these stable cell lines to DNA topoisomerase II inhibitor etoposide. As illustrated in Fig.  1C, loss of C53 rendered HeLa cells partially but significantly resistant to cell death induced by DNA topoisomerase II inhibitor etoposide. After 48 h of treatment of 20 M etoposide, more than 60% of control HeLa cells underwent cell death, whereas around 30% of C53-deficient cells died. Similar results were observed for E1A-transformed human fibroblast cells IMR90E1A (Supplemental Fig. S1). Moreover, annexin V staining and flow cytometry analysis demonstrated that loss of C53 significantly inhibited both early and late stages of apoptosis induced by etoposide (Fig. 1D), indicating that C53 may modulate etoposide-induced apoptosis.
To further confirm the specificity of C53 shRNA and to test if C53 is indeed required for etoposide-induced apoptosis, we attempted to restore the sensitivity of those stable cell lines to etoposide by expressing C53 ectopically. To avoid inhibition of ectopic expression of C53 protein by C53 shRNA, we introduced two silent mutations in the region of C53 cDNA that were complementary to C53 shRNA. As expected, only the construct of C53-si (C53 cDNA with two silent mutations) could ectopically express wild-type C53 in C53-deficient cells (Fig. 1E), which in turn restored the sensitivity of C53-deficient cell line 6 to etoposide-induced apoptosis (Fig. 1F). Taken together, our results strongly suggest that C53 is likely to play a critical role in modulating etoposide-induced apoptosis.
C53 Was Required for Mitochondrial Damage and Caspase Activation in Etoposide-induced Apoptosis-Because C53 may play a critical role in etoposide-induced apoptosis of HeLa cells, we attempted to understand more about this apoptotic process. Etoposide strongly induced a G 2 /M arrest in HeLa cells after 24 h of treatment (Fig. 4) 3 followed by rapid cell death around 40 h post-treatment ( Fig. 2A). Meanwhile, a basal level of cell death was also induced by etoposide (20 -30% cell death in a 48-h treatment) that is independent of C53 and caspases (Fig.  2, A and B). To investigate the possible working mechanism of C53, we first examined if C53 deficiency resulted in the defects on either drug uptake and/or sensing of DNA damage. As shown in Fig. 2D, loss of C53 did not inhibit up-regulation of p53 induced by etoposide, indicating that C53 may not be involved in drug uptake or sensing of DNA damage. We then 3 H. Jiang and H. Li, unpublished data. attempted to examine the relationship between C53 and the core apoptotic apparatus. As shown in Fig. 2B, a pan-caspase inhibitor zVAD-fmk inhibited etoposide-induced apoptosis to the basal level (around 30% cell death) but did not have any effect on C53-deficient cell lines. Furthermore, C53 deficiency partially inhibited caspase activity induced by etoposide (Fig.  2C). Activation of both initiator caspases (caspase-2 and -9) and executioner caspase-3 in etoposide-induced apoptosis of HeLa cells was partially inhibited by loss of C53 (Fig. 2E), indicating that C53 acts upstream of caspase activation. Finally, loss of C53 partially prevented the release of apoptogenic factor cytochrome c from mitochondria (Fig. 2, F and G). These results suggest that C53 may act upstream of mitochondrial damage and caspase activation during etoposide-induced apoptosis.
Ectopic Expression of C53 Promoted Apoptosis Induced by Genotoxic Stress-Because C53 deficiency partially inhibited etoposide-induced apoptosis, we further examined if overexpression of C53 could sensitize HeLa cells to genotoxic stress. As shown in real-time images (Fig. 3A), C53-transfected cells (indicated by fluorescence of EGFP, panel a and b) underwent rapid cell death (as early as 6 h post-treatment, indicated by shrinkage and blebbing) after etoposide treatment, whereas untransfected cells remained in G 2 /M arrest (indicated by large cell bodies in the center of the images) followed by delayed cell death around 40 h post-treatment ( Fig. 2A). After a 24-h treatment, vector-transfected HeLa cells had only 20% cell death, whereas in the presence of C53 overexpression, more than 60% of HeLa cells underwent cell death (Fig. 4B). Similarly, overexpression of C53 also sensitized IMR90E1A cells to etoposideinduced cell death ( Fig. 3B and Supplemental Fig. S1B). This result strongly suggests that overexpression of C53 sensitizes tumor cells to etoposide-induced cell death. Furthermore, C53promoted cell death was apoptotic cell death, as indicated by more poly(ADP-ribose) polymerase cleavage and caspase-3 ac-tivation in the presence of C53 overexpression (Fig. 3C). Additionally, we investigated whether C53 is involved in ionizing radiation-induced apoptosis. As shown in Fig. 3, C and D, C53 deficiency partially inhibited x-ray-induced apoptosis of HeLa cells, whereas C53 overexpression promoted it, suggesting that C53 may also play a regulatory role in x-ray-induced apoptosis.
Ectopic Expression of C53 Rendered HeLa Cells Susceptible to Apoptosis Induced by Multiple Genotoxic Agents in a Cdk1dependent Manner-We further examined the effect of C53 on apoptosis induced by other death stimuli. As shown in Supplemental Fig. S2, loss of C53 did not lead to inhibition of apoptosis induced by staurosporine and TNF␣ as well as other genotoxins such as cisplatin and doxorubicin. Nonetheless, overexpression of C53 rendered HeLa cells susceptible to multiple genotoxins such as etoposide, doxorubicin, and camptothecin, but not cisplatin, TNF␣ plus cycloheximide, and UV irradiation (Fig. 4A). Because both etoposide and x-ray irradiation normally induce G 2 /M arrest, the discriminating effect of C53 on sensitization of HeLa cells to apoptosis induced by various agents prompted us to examine cell cycle profiles induced by these agents. Interestingly, flow cytometry analysis showed that treatment of etoposide, doxorubicin, and camptothecin, but not cisplatin, TNF␣, and UV radiation induced the G 2 /M arrest of HeLa cells (Fig. 4B), indicative of a possible correlation between C53 function and apoptosis induced by genotoxins that triggers the G 2 /M arrest.
Cdk1-cyclin B1 complex is an essential factor for entry and completion of mitosis (4). Previous studies have also demonstrated its regulatory role in apoptotic signaling in a number of experimental paradigms (16). We examined if Cdk1-cyclin B1 is involved in etoposide-induced apoptosis of HeLa cells. Specific Cdk1 inhibitor olomoucine II (IC 50 ϭ 20 nM) at the concentration of 2 M partially inhibited etoposide-induced cell death but had no effect on C53-deficient cells (Fig. 4C). Fur- thermore, C53-promoted etoposide-induced cell death was partially inhibited by Cdk1 dominant negative mutant (Cdk1DN), but not by Cdk2DN and Cdk5DN (Fig. 4D). Taken together, these results are suggestive of a critical role of Cdk1 activity in the signaling of etoposide-induced apoptosis of HeLa cells and the sensitization by C53.
Overexpression of C53 Inhibited the Inhibitory Phosphorylation of Cdk1 and Promoted Nuclear Accumulation of Cyclin B1-The result indicative of the involvement of Cdk1 in our experimental system prompted us to examine the effect of C53 on Cdk1-cyclin B1 activity. The 24-h treatment of etoposide did not lead to an increase in overall Cdk1 activity indicated by histone H1 kinase assay (Fig. 5A). However, overexpression of C53 significantly increased Cdk1 activity in the presence of etoposide treatment (Fig. 5A). As described above, activation of Cdk1-cyclin B1 is subjected to multiple levels of transcriptional and post-translational regulation. In response to DNA damage, the G 2 /M checkpoint system (ATM/ATR and checkpoint kinase 1 and 2) inactivates phosphatase CDC25C, which in turn leads to phosphorylation at residues at Thr-14 and Tyr-15 by kinases Wee1 and Myt1, resulting in inactivation of Cdk1. We then examined the protein level and phosphorylation of Cdk1 and cyclin B1 during the time course of etoposide treatment. As shown in Fig. 5B, the inhibitory phosphorylation of Cdk1 at Tyr-15 was increased when cells were treated with etoposide, indicating proper activation of the G 2 /M DNA damage checkpoint in HeLa cells. Intriguingly, the inhibitory phosphorylation of Cdk1 at Tyr-15 was dramatically reduced in the presence of C53 overexpression, indicating an increased Cdk1 activity (Fig. 5B). This result was consistent with our Cdk1 kinase assay described in Fig. 5A. Furthermore, the C53-dependent increase on Cdk1 activity was not inhibited by caspase inhibitor zVAD-fmk. 2 Taken together, these results indicate that C53 may be involved in regulation of the G 2 /M DNA damage checkpoint. We further examined the protein level and phosphorylation of cyclin B1. During the course of etoposide treatment, cyclin B1 accumulated and peaked at 18 h post-treatment and followed by gradual disappearance, presumably due to anaphasepromoting complex-mediated degradation (Fig. 5B). Interestingly, the cyclin B1 level was significantly lower in the presence of C53 overexpression during the time course (Fig. 5B, cyclin B1 panel), whereas its level was higher in C53-deficient cells (Supplemental Fig. S3). To further investigate the role of C53 in regulation of Cdk1-cyclin B1, we first investigated the possible cause of lower cyclin B1 levels in the presence of C53 overexpression. Because C53 overexpression promoted etoposide-induced apoptosis, it is possible that lower cyclin B1 levels may be caused by nonspecific effect of more cell death. To test this possibility, we used pan-caspase inhibitor zVAD-fmk to inhibit most of cell death induced by etoposide. As shown in Fig. 5C, zVAD failed to prevent the decrease of cyclin B1 level induced by C53 overexpression. In contrast, proteasome inhibitor MG132 fully restored cyclin B1 levels in the presence of C53 overexpression. This result suggests that C53 may promote proteasome-mediated cyclin B1 degradation. Because cyclin B1 degradation usually occurs at the exit of mitosis after Cdk1-cyclin B1 activation, the observation of lower cyclin B1 in the presence of C53 overexpression may indirectly indicate that C53 overexpression promotes Cdk1-cyclin B1 activation.
Regulation of the subcellular localization of cyclin B1 is another control mechanism of the G 2 /M checkpoint. As shown in Fig. 5D, more cyclin B1 were present in the nuclear fraction in the presence of C53 overexpression when cells were treated with etoposide. Moreover, there was more cytoplasmic retention of cyclin B1 in C53-deficient cells (Supplemental Fig. S4A). Phosphorylation of cyclin B1 is required for nuclear accumulation and translocation of Cdk1-cyclin B1 complex from cytoplasm to nucleus during prophase (10, 12-15). C53 overexpres- sion indeed promoted cyclin B1 phosphorylation, as indicated by more slow-migrating phosphorylated cyclin B1 (Supplemental Fig. S4B). Taken together, our results suggest that C53 may also modulate the subcellular localization of cyclin B1.
C53 Associated with Cyclin B1 but Did Not Affect Cdk1 Enzymatic Activity-C53 was also identified as a protein interacting with Cdk5 activator p35 in a yeast two-hybrid screening. Functional similarity between p35 and cyclin B1 prompted us to test whether C53 also interacts directly with cyclin B1 and affects its phosphorylation. The specificity of our anti-C53 antibody (affinity-purified rat polyclonal antibody) for immunostaining was verified by protein competition and staining of C53-deficient cells (Supplemental Fig. S5). Immunostaining of C53 displayed mainly cytoplasmic pattern. Co-immunofluorescence staining and confocal microscopy showed extensive colocalization of cyclin B1 and C53 after etoposide treatment (Fig. 6A), indicating a possible association between C53 and cyclin B1. We further performed a co-immunoprecipitation assay to confirm the interaction. As shown in Fig. 6B, endogenous C53 was co-immunoprecipitated down by cyclin B1 antibody. To exclude the possibility of an antibody cross-reaction, we performed the co-immunoprecipitation assay using overexpressed proteins to further confirm the C53-cyclin B1 interaction (Fig. 6C). Because our results demonstrated that C53 promoted the Cdk1-cyclin B1 activity in response to genotoxic stress, we examined if C53 directly enhances the enzymatic activity of Cdk1-cyclin B1. As shown in Fig. 6C, purified C53 did not affect the enzymatic activity of purified active Cdk1cyclin B1, suggesting that C53 probably acts upstream of Cdk1 in the G 2 /M DNA damage checkpoint. DISCUSSION In response to DNA damage, the cellular decision of life versus death involves an intricate network of multiple factors that play critical roles in regulation of DNA repair, cell cycle, and cell death. Among them, multiple checkpoint proteins such as p53, ATM/ATR, etc., play very important roles in cellular decision-making (2). Here we report the involvement of Cdk5 activator-binding protein C53 in regulation of apoptosis induced by genotoxic stress via modulating the G 2 /M DNA damage checkpoint. Our data demonstrated that C53 deficiency conferred partial resistance to genotoxic agents such as etoposide and x-ray irradiation, whereas ectopic expression of C53 rendered cells susceptible to multiple genotoxins. These results strongly suggest that C53 plays an important role in regulation of genotoxin-induced apoptosis. The finding that C53 sensitized cells only to the genotoxins that trigger G 2 /M arrest prompted us to examine the effect of C53 on the Cdk1-cyclin B1 complex. Cdk1-specific inhibitors (olomoucine II and Cdk1 DN mutant) inhibited etoposide-induced apoptosis of HeLa cells and C53 sensitizing effect, indicating that Cdk1 plays a critical role in initiation of apoptotic signaling in our experimental system. Overexpression of C53 enhanced nuclear accumulation and enzymatic activity of Cdk1-cyclin B1, whereas C53 deficiency leads to more cytoplasmic retention of cyclin B1, suggesting that C53 acts as a pivotal player in modulating the Cdk1-cyclin B1 function. Taken together, our results suggest that in response to genotoxic stress, C53 serves as an important regulatory component for the G 2 /M DNA damage checkpoint.
We found that loss of C53 conferred partial resistance to apoptosis specifically induced by DNA topoisomerase II inhibitor etoposide and x-ray irradiation but not other genotoxic agents. One outstanding question is why loss of C53 confers resistance only to etoposide and x-ray irradiation but not other genotoxins even though the agents such as doxorubicin and camptothecin can also induce G 2 arrest and cell death. One possible explanation may lie in the nature of damages produced by genotoxic agents. Most genotoxic agents can produce not only many types of DNA damage but also damages on non-DNA targets, which can also trigger response of cell death. As an example, doxorubicin induces not only DNA damages such as DNA alkylation, cross-linking, and double-stranded breaks but also free radical generation and lipid peroxidation in cellular membrane, mitochondria, and microsomes (27).
In contrast to other genotoxic agents used in this study that can induce HeLa cells to undergo rapid cell death (within 24 h), both etoposide and x-ray irradiation appeared to induce most cells to undergo lengthened G 2 arrest (as indicated by the enlarged nucleus and cell body and confirmed by flow cytometry) followed by delayed apoptotic cell death in 48 -72 h. Judged by the features of G 2 arrest and delayed cell death, etoposide-induced apoptosis of HeLa cells fits well into the definition of "mitotic death" or "reproductive death" (28,29), a type of cell death usually associated with p53-mutated tumor cells or non-hematopoietic cells as a consequence of DNA damage. Yet the underlying molecular mechanism remains poorly understood. A number of recent studies have indicated that Cdk1 plays a role in regulation of mitotic cell death (16). In this study we observed the increase of Cdk1 activity and nuclear accumulation of Cdk1-cyclin B1 complex when cells were treated with etoposide at late time points. Furthermore, Cdk1specific inhibitors (Cdk1 DN mutant and pharmacological inhibitor) were able to inhibit etoposide-induced apoptosis. Additionally, our real-time images revealed that typical apoptotic cell death occurred at minutes after cells entered metaphase from G 2 phase. 3 Therefore, we postulate that Cdk1 may play a critical role in initiation of mitotic death by driving cells to enter aberrant mitosis, which leads to activation of caspase and apoptosis. If so, what is the link between Cdk1 and the core apoptotic machinery? How do the G 2 /M checkpoint or other checkpoint systems breakdown after lengthened G 2 arrest, thereby resulting in Cdk1 activation? Further studies will provide more definitive answers to these important questions.
DNA damage checkpoints are crucial for maintaining DNA integrity in normal cells but also post as a challenge for conventional cancer treatment. Therefore, abrogation of DNA damage checkpoints appears to be a logical way for chemosensitization of tumor cells. Due to the fact that many cancer cells have defective G 1 checkpoints, it becomes more important to modify the G 2 /M checkpoint for chemosensitization. Our finding that ectopic expression of C53 rendered HeLa cells susceptible to multiple apoptotic stimuli is intriguing with regard to its potential implication on cancer treatment. By overriding the G 2 /M DNA checkpoint-mediated inhibition of the Cdk1-cyclin B1 activity, ectopic expression of C53 may represent a novel approach for chemo-and radio-sensitization of tumor cells, especially tumor cells with defective G 1 checkpoint. Further elucidation of C53 function in both cell cycle and cell death control will facilitate the exploitation of this novel protein in cancer treatment.