Induction of Cdc25B Regulates Cell Cycle Resumption after Genotoxic Stress

Introduction

The Cdc25 subfamily of the protein tyrosine phosphatases are important regulators of mammalian cell cycle checkpoints, which control proliferation and genomic integrity ( 1). Cdc25 was initially identified in fission yeast as a mitotic inducer ( 2). The three mammalian homologues Cdc25A, Cdc25B, and Cdc25C positively regulate cell cycle progression by activating their cyclin-dependent kinase (Cdk)/cyclin substrates. Cdc25A primarily activates G₁-S specific Cdks ( 1) and is required for entry of cells into the S phase and also regulates G₂-M progression ( 3). Cdc25B and Cdc25C have a more restricted role in promoting progression from G₂ phase to mitosis. Thus, microinjection of neutralizing Cdc25B and Cdc25C antibodies into cells results in mitotic delay ( 4). Despite the seemingly similarity in functions, Cdc25B and Cdc25C have temporally distinct roles in cells with Cdc25B activity peaking before Cdc25C ( 4). Cdc25B seems to activate a cytoplasmic pool of Cdk1/cyclin B, triggering centrosomal microtubule nucleation ( 5). Moreover, Cdc25B undergoes Aurora-A dependent phosphorylation at the centrosome, which may permit entry into mitosis by providing the first stimulus of Cdk1 activity ( 6). Therefore, Cdc25B potentially has a unique role in initiating mitosis that warrants further investigation.

Cdc25A and Cdc25B but not Cdc25C have documented oncogenic properties ( 7). Overexpression of Cdc25B in human tumors correlates with poor prognosis ( 8– 10) and ectopic Cdc25B overexpression in mammary glands increases susceptibility to breast cancer induction by 9,10-dimethyl-1,2benzanthracene ( 11). Previously, we found that benzo(a)pyrene diol epoxide, the ultimate carcinogen found in cigarette smoke, increases Cdc25B expression in lung cancer cells, indicating that Cdc25B could contribute to benzo(a)pyrene diol epoxide induced lung carcinogenesis ( 12).

Eukaryotic cells have evolved surveillance mechanisms to guard their DNA from genotoxic stress ( 13). One critical mechanism is inhibition of cell cycle progression, thus allowing cells to repair damaged DNA and preventing replication or transmission of altered DNA to the next generation ( 13). To induce cell cycle arrest, a complex signaling network is used to activate checkpoint kinases Chk1 and Chk2 ( 14). These kinases induce proteasome-mediated degradation of Cdc25A ( 15– 20) and cytoplasmic retention of Cdc25C mediated by phosphorylation of Ser²¹⁶ ( 21). Cdc25B has also been reported to be affected by at least some forms of genotoxic stress. Thus, in response to UV, activation of the p38/mitogen-activated protein kinase-activated protein kinase 2 signaling pathway leads to phosphorylation of Cdc25B at Ser³⁰⁹ and subsequent 14-3-3 binding in vitro ( 22). Binding to 14-3-3 is thought to restrict Cdc25B to the cytoplasm ( 23, 24). However, the nature of the genotoxic insult as well as the magnitude of the DNA damage may influence the final fate of Cdc25 family members ( 25, 26).

Although DNA damage clearly can cause cell cycle arrest, the factors that control resumption of cell cycle have only recently become the subject of attention. It seems that several components of the eukaryotic checkpoint signaling pathway are essential for resumption of cell cycle progression after genotoxic stress in a process known as “recovery” ( 27). Specifically, polo-like kinase 1 and the serine/theronine phosphatase PPM1D were shown to regulate cell cycle resumption after DNA damage–induced G₂ arrest ( 28, 29). Chk1 and claspin, a regulator of Chk1 activation, are targeted for degradation after replication stress to facilitate resumption after genotoxic stress ( 30– 32). The tyrosine phosphatase Cdc25B is also thought to be essential in regulating resumption of cell cycle progression after genotoxic stress ( 28, 33). Thus, we have reevaluated the effects of different types of DNA-damaging agents on Cdc25B. We now report that mechanistically distinct DNA-damaging agents rapidly induced Cdc25B expression and that levels of Cdc25B determined the kinetics of cell cycle resumption after DNA damage–induced cell cycle arrest.

Materials and Methods

Cell culture and chemicals. A549 cells (ATTC, Manassas, VA) were maintained in basal medium Eagle supplemented with 1% fetal bovine serum (FBS), 1% l-glutamine, and 1% penicillin-streptomycin; p53^+/+, p53^−/−, and Chk2^−/− HCT116 cells (a gift from Prof. Bert Vogelstein, Johns Hopkins University) in McCoy's medium with 10% FBS and 1% penicillin-streptomycin; HeLa (ATTC) in DMEM with 10% FBS, 1% penicillin-streptomycin; wild-type and knockout Cdc25B mouse embryonic fibroblasts (MEF; a gift from Dr. Peter Donovan, Johns Hopkins University) in DMEM with 20% FBS, 1% l-glutamine, and 1% penicillin-streptomycin; and U2OS-expressing HA-Cdc25B3 under tetracycline-regulated promoter (a gift from Prof. Bernard Ducommun, Université Paul Sabatier) in DMEM supplemented with 10% FBS, G418 (100 μg/mL), 1% penicillin-streptomycin, and 2 μg/mL tetracycline. Etoposide, cisplatin, and bleomycin were obtained from Sigma-Aldrich Co. (St. Louis, MO). Caffeine, nocodazole, MG132, and tetracycline were purchased from Calbiochem (La Jolla, CA). Cells were plated and incubated for 24 h at 25% to 30% density before UV or compounds exposure (Stratagene, UVC Cross linker, La Jolla CA).

Antibodies and western blotting. Cdc25B was detected with a monoclonal antibody from BD Transduction Laboratories (Lexington, KY). Antibodies for p53, Chk1 (Ser³⁴⁵), and poly(ADP-ribose) polymerase were obtained from Cell Signaling Technology (Danvers, MA). Cdc25A, Cdc25C, and Chk1 were detected with antibodies from Santa Cruz Biotechnology (Santa Cruz, CA). Phospho-histone H3 antibody was purchased from Upstate Biotechnology (Lake Placid, NY), anti-HA antibody from Covance (Berkeley, CA), and β-tubulin antibody from Cedarlane Laboratories (Hornby, Ontario, Canada). Bound primary antibodies were detected with either horseradish peroxidase (HRP)–goat anti-mouse antibody or HRP-goat anti-rabbit antibody (Jackson ImmunoResearch, West Grove, PA), and proteins were visualized by chemiluminescence using the enhanced chemiluminescence reagent (Amersham Pharmacia, Piscataway, NJ). Cell were harvested and lysed in modified radioimmunoprecipitation assay buffer [50 mmol/L Tris (pH 7.6), 1% Triton X-100, 0.1% SDS, 150 mmol/L NaCl, 1 mmol/L EDTA, 2 mmol/L Na₃VO₄, 12 mmol/L β-glycerol phosphate, 10 mmol/L NaF, 10 μg/mL aprotinin, 10 μg/mL leupeptin, 100 μg/mL 4-(2-aminoethyl)-benzenesulfonylfluoride hydrochloride, 10 μg/mL soybean trypsin inhibitor, and 1 mmol/L phenylmethylsulfonyl fluoride], incubated on ice for 30 min with brief vortex mixing every 10 min, and centrifuged at 13,000 × g for 15 min to clear the lysates. Western blotting was done as described ( 12) with the exception that proteins were transferred overnight to a nitrocellulose membrane at 30 V to allow maximum protein transfer.

RNA interference and RNA measurements. The vector for creating Cdc25B knockdown HCT116 cells was obtained from Ambion (Austin, TX; pSilencer 4.1 containing a gene for puromycin resistance). Oligonucleotides encoding short hairpin RNA interference (shRNAi) targeting Cdc25B (target sequence for Cdc25B4 vector AAAGGCGGCTACAAGGAGTTC or Cdc25B3 vector GTTCAGCAACATCGTGGATAA) were ligated into the vector according to the manufacturer's instructions. Vector expressing interference RNA with limited homology to any known sequence (scramble vector) was provided by the manufacturer and used as a negative control. The pSilencer plasmid was transfected in HCT116 using LipofectAMINE Plus (Invitrogen, Carlsbad, CA) according to manufacturer's instructions, and 24 h after transfection, cells were replated and allowed to attach overnight. Thereafter, clones were selected with 0.5 μg/mL puromycin. Clones were picked 2 weeks later, expanded by culturing in the presence of 0.5 μg/mL puromycin, and screened for Cdc25B knockdown using Western blotting. For Cdc25B smartpool small interfering RNA (siRNA; Dharmacon, Inc., Lafayette, CO) transfection, HCT116 cells were plated and incubated for 24 h to yield 50% to 55% density. Cells were transfected according to manufacturer's recommendation (Invitrogen). Briefly, 100 nmol/L of Cdc25B smartpool siRNA was transfected using LipofectAMINE 2000, and 24 h later cells were split to yield 30% to 35% density the next day when harvested.

Reverse transcription–PCR. Total RNA was extracted from cells using RNeasy (Qiagen, Valencia, CA). cDNA was synthesized from 2 μg of RNA using random hexamer (Amersham, Buckinghamshire, United Kingdom) with Superscript RNase H reverse transcriptase (Life Technologies, Inc., Gaithersburg, MD). The reverse-transcribed cDNA from each sample was subjected to PCR amplification using Taq polymerase (Promega, Madison, WI) and primers. The sequence of the primers used was as described previously ( 12). The PCR conditions for the amplification of Cdc25B and β-actin genes were 24 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C for 1 min, followed by a final incubation at 72°C for 7 min. Amplified products were separated by 2% agarose gel electrophoresis, and bands were visualized by staining with ethidium bromide.

G₂ checkpoint recovery assay and flow cytometry. For G₂ checkpoint recovery assay, we plated U2OS cells for 24 h and induced Cdc25B by thoroughly washing cells and culturing in the absence of tetracycline for 16 h. After the 16-h induction, cells were mock or UV treated (as indicated) and cells were either harvested at the indicated time points (asynchronous cells) or were trapped in mitosis with a 23-h nocodazole (1 μmol/L) incubation. For HCT116 Scr and Cdc25B4#2 cells, cells were plated for 24 h and were either mock or UV treated (as indicated). Cycling cells were then trapped in mitosis with an 18-h nocodazole (1 μmol/L) incubation.

Identification of mitotic cells was carried out by simultaneously staining cells with propidium iodide (PI) and phospho-histone H3 (Ser¹⁰) as described previously ( 34). Briefly, cells were fixed overnight in 70% ethanol and permeabilized with PBS, 0.25% Triton X100 for 8 min on ice. After washing in PBS–1% bovine serum albumin, cells were incubated with 1 μg of anti–phospho-histone H3 for 2 h. Cells were again washed before incubating with antirabbit IgG Alexa 488 (1:75 for U20S and 1:100 for HCT116 cells) for 40 min in the dark. After cell washing and resuspension in 500 μL PI for 15 min, 20,000 cells were analyzed by fluorescence-activated cell sorting and WinMDI to determine phospho-histone H3 positive cells.

Results

Cdc25B induction by DNA damage. After an extensive investigation of commercially available Cdc25B antibodies to detect endogenous Cdc25B, we determined that an anti-Cdc25B monoclonal antibody from BD Transduction Laboratories was most specific. As shown in the Fig. 1A , this Cdc25B monoclonal antibody detected two bands in the lysates of wild-type Cdc25B MEFs in a region of the predicted molecular mass (∼65 kDa) of Cdc25B, whereas in the Cdc25B null MEF lysates only the lower band was detected even after loading more lysate ( Fig. 1A, lane 3). These results confirmed that the upper band was Cdc25B. Other commercially available antibodies to Cdc25B cross-reacted with many additional bands and were unaffected in the MEFs lacking Cdc25B or after shRNAi (data not shown). With the BD Transduction antibody, the Cdc25B band ran closer to the recombinant His-tagged human Cdc25B2 protein, further supporting that the upper band was indeed Cdc25B. This was further corroborated by a decreased expression of Cdc25B in the clones picked after Cdc25B-specific shRNAi treatment ( Fig. 1B). Similar results were obtained when we used Cdc25B smartpool (Dharmacon) in transient transfection assays (Supplementary Fig. S1). Notably, Cdc25B specific shRNAi/siRNA did not decrease the levels of the nonspecific lower band.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Agents causing different forms of DNA damage–induced endogenous Cdc25B protein levels. A, early passage of Cdc25B wild-type (+/+) and null (−/−) MEFs were harvested to analyze the Cdc25B expression. RP, recombinant full-length purified human His-tagged Cdc25B2. Cdc25B expression was also analyzed in cells selected for stable knock down of Cdc25B. Arrow, Cdc25B; asterisks, nonspecific band; LE, long exposure. B, A549 cells were treated with DMSO or etoposide (30 μmol/L), cisplatin (25 μmol/L), and bleomycin (25 μmol/L) and harvested at different time points. HCT116 cells were treated with IR (10 Gy) and harvested at different time points. Levels of indicated proteins were examined by Western blotting. Representative of n = 4-6.

Thus, for all subsequent experiments, we used the BD Transduction monoclonal antibody to detect endogenous Cdc25B in cells. To investigate the effects of DNA-damaging agents on endogenous Cdc25B expression, we treated asynchronous A549 cells with different clinically used anticancer drugs: topoisomerase II poison etoposide (30 μmol/L), DNA cross-linker cisplatin (25 μmol/L), and radiomimetic bleomycin (25 μmol/L) for 0, 1, 4, and 24 h ( Fig. 1B). These drug concentrations were based on previous published findings revealing significant DNA damage ( 35, 36). Unexpectedly, all of these agents increased cellular Cdc25B protein levels within 1 h ( Fig. 1B). With etoposide treatment, Cdc25B was elevated for at least 24 h, whereas, with cisplatin and bleomycin treatment, the increase in Cdc25B was more transient and basal levels were seen at 24 h. Treatment with anticancer drugs did not affect the levels of the lower band (data not shown). As expected, treatment with these drugs decreased Cdc25A and transcriptionally increased p53 levels, although the kinetics for p53 induction was slower than that for Cdc25B. Additionally, Cdc25B levels increased within 1 h after treatment with 10 Gy of ionizing irradiation, a prototypical DNA-damaging agent, with levels remaining high for at least 4 h ( Fig. 1B). The kinetics of Cdc25B increase after ionizing irradiation was similar to that of p53. Therefore, Cdc25B induction was a promiscuous response shared by mechanistically distinct DNA-damaging agents, and this observation supports our previous finding that the DNA-damaging carcinogen, benzo(a)pyrene diol epoxide, increases Cdc25B expression in these lung cancer cells.

p53 independence of Cdc25B induction by UV. We next examined the effect of DNA damage by UV irradiation on Cdc25B expression. As shown in Fig. 2A , A549 cells treated with UV (60 J/m²) had increased Cdc25B protein levels within 30 min after irradiation and the levels were persistently increased for 24 h posttreatment. Because of a previous report ( 22) showing that Cdc25B levels were unaffected in HeLa cells exposed to UV, we examined the effect of UV irradiation on endogenous Cdc25B expression in HeLa cells to exclude cell type–specific effects. We observed increased Cdc25B levels within 30 min after UV exposure, which remained elevated for at least 24 h, similar to A549 cells ( Fig. 2A). Furthermore, Cdc25B induction was independent of p53, as indicated by using isogenic HCT116 cells that expressed or lacked p53 ( Fig. 2B). We also studied the effect of UV on Cdc25B wild-type and null MEFs. Cdc25B expression was increased in the Cdc25B wild-type MEFs, whereas no Cdc25B was detected in Cdc25B null cells as expected ( Fig. 2C). Furthermore, Cdc25B induction was attenuated in cells stably expressing Cdc25B shRNAi (Cdc25B4#2) compared with cells with scrambled shRNAi ( Fig. 2C). Finally, to study the effect of UV irradiation on overexpressed Cdc25B, we treated U20S cells expressing a tetracycline-regulated HA-tagged Cdc25B. After removal of tetracycline from the medium, HA-Cdc25B expression was induced and cells were either mock or UV treated. Cells treated with UV expressed more HA-Cdc25B at 24 h and 48 h compared with the mock-treated cells. Induction of exogenous Cdc25B was confirmed using an antibody to the HA-epitope tag ( Fig. 2C).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Cdc25B induction by UV is independent of p53 levels. A, A549 and HeLa cells were washed with PBS twice and then irradiated with 60 J/m² of UV in the absence of medium. After exposure, cell culture medium was added and cells were harvested at different time points for determination of endogenous Cdc25B expression by Western blotting. B, HCT116 (p53^+/+) and HCT116 (p53^−/−) cells were treated as described for (A). C, Cdc25B^+/+ and Cdc25B^−/− MEFs were treated as described above. HCT116 Scr and Cdc25B4#2 cells were treated with either mock or UV (60 J/m²) for 16 h as indicated. NS, nonspecific band. U2OS cells were cultured either in the presence or absence of tetracycline to induce HA-Cdc25B3, and the effect of either mock or UV (15 J/m²) irradiation on exogenous Cdc25B was analyzed at the indicated time points. D, A549 cells were treated with different doses of UV, and 1 h later cells were harvested. We plated 800 HCT116 (p53^+/+) cells in 6-cm dish, and 24 h later cells were treated with different doses of UV. The surviving clones were harvested 8 d posttreatment for Cdc25B expression. Levels of indicated proteins were examined by Western blotting. Representative of n = 4-6.

We next determined the minimum dose of UV required to increase endogenous Cdc25B expression. In A549 cells, doses as low as 15 J/m² were sufficient to induce Cdc25B within 1 h ( Fig. 2D). To examine the effect of UV treatment on endogenous Cdc25B expression in a long-term assay, we treated HCT116 (p53^+/+) cells with UV doses ranging from 5 to 60 J/m² and harvested clones 8 days later. In this assay, doses at ≥5 J/m² induced significant levels of Cdc25B ( Fig. 2D). The effect of doses >15 J/m² could not be evaluated due to toxicity and lack of sufficient colonies for analysis. These results support previous findings, demonstrating an increase in Cdc25B levels after chronic exposure to ionizing radiation ( 37).

Checkpoint regulates Cdc25B induction. Checkpoint activation in response to DNA damage results in the Chk1-dependent phosphorylation of Cdc25A and Cdc25C ( 15, 18, 21). Because Cdc25B has canonical Chk1/Chk2 phosphorylation sites and Chk1 has been shown to phosphorylate Cdc25B ( 38, 39), it is reasonable to hypothesize that Cdc25B induction might be checkpoint-regulated. Indeed, Chk1 was rapidly phosphorylated and activated in HCT116 cells after UV treatment ( Fig. 3A ). To investigate the role of ATR/Chk1 pathway, we pretreated cells with caffeine (5 mmol/L) to block ATR/ATM activation for 30 min before mock or UV (60 J/m²) irradiation. As illustrated in Fig. 3A, pretreatment with caffeine blocked the increase in Cdc25B seen 1 h after UV exposure. Similar results were also observed in A549 cells (data not shown). To determine the role of Chk1 in Cdc25B up-regulation, we pretreated cells with UCN-01 (300 nmol/L) for 30 min and measured Cdc25B levels 1 h after mock or UV treatment (60 J/m²). UCN-01 blocked the increase in Cdc25B expression with UV ( Fig. 3B). UCN-01 prevented loss of Cdc25A after UV (Supplementary Fig. S2), confirming the efficacy of UCN-01 in inhibiting Chk1 activity. These results also highlight the mechanistically distinct regulation of Cdc25A and Cdc25B isoforms by checkpoint signaling after DNA damage. We also examined Cdc25B induction in HCT116 Chk2^−/− cells. Cells were either mock or UV (60 J/m²) treated and harvested 1 h later. As shown in Supplementary Fig. S3, the increase in Cdc25B level was independent of Chk2. We also noted similar Cdc25B mRNA levels 4 h after UV treatment using semiquantitative reverse transcription-PCR (Supplementary Fig. S4). The labile nature of Cdc25B was confirmed by MG132-mediated proteasomal inhibition, which revealed increased Cdc25B levels by 24 h (Supplementary Fig. S4C). These results support previous findings, showing proteasome-mediated degradation of Cdc25B ( 40). Collectively, our results support a model that the increase in Cdc25B protein levels after UV is regulated by ATR/Chk1 pathway via posttranscriptional mechanism, potentially by affecting Cdc25B protein stability.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Cdc25B induction after UV irradiation is regulated by ATR/Chk1. A, HCT116 cells were pretreated with vehicle or caffeine (5 mmol/L) for 30 min before UV (60 J/m²) or mock irradiation for 1 h. B, HCT116 cells were pretreated with UCN-01 (300 nmol/L) for 30 min or vehicle before UV (60 J/m²) or mock irradiation for 1 h. Representative of n = 4.

Cdc25B regulates mitotic entry after DNA damage. Recent evidence suggests that Cdc25B, but not Cdc25A or Cdc25C, regulates cell cycle reentry after DNA damage produced by doxorubicin ( 28). Stimulated by our observation that Cdc25B expression was increased within 1 h in cells after exposure to mechanistically distinct DNA-damaging agents, we hypothesized that levels of Cdc25B might be crucial in regulating the rate of cell cycle resumption after DNA damage–induced cell cycle arrest. Thus, asynchronous cells were treated with mock or UV irradiation, and cells exiting G₂ were trapped in mitosis with nocodazole (1 μmol/L) treatment for 18 h (HCT116) or 23 h (U2OS). Mitotic arrest was detected either by probing lysates with a phospho-histone H3 (Ser¹⁰) antibody or by staining cells with PI and phospho-histone H3 (Ser¹⁰) and analyzing cells with a flow cytometer. To examine the effect of Cdc25B in this assay, we reduced intracellular Cdc25B levels by >75% using shRNAi ( Fig. 4A ). In the absence of DNA damage, 69.2 ± 2.8% of control cells (Scr; n = 3) and 59.4 ± 4.0% of Cdc25B depleted cells (n = 3) were trapped in mitosis ( Fig. 4B and C), consistent with previous studies suggesting Cdc25B is not required for normal cell cycle progression ( 41). After 30 J/m² UV, we observed a slight decrease in percentage of cells trapped in mitosis by nocodazole for control cells (60.3 ± 0.5%; n = 4), whereas only 40.5 ± 1.1% of Cdc25B depleted cells (n = 4) were arrested in mitosis. The difference in mitotic trapping was even more pronounced at a higher dose of UV (60 J/m²), in which 25.0 ± 1.1% of control cells (n = 5) and 14.7 ± 0.6% of Cdc25B depleted cells (n = 5) were trapped in mitosis. At both doses of UV, the difference between Scr and Cdc25B depleted cells was statistically significant ( Fig. 4D). Interestingly, with all of the conditions, ∼80% of the cell population was in G₂-M ( Fig. 4D), suggesting that, even after DNA damage, a majority of cells could progress from G₁ and S phase into G₂ phase but entry into mitosis was dependent on levels of Cdc25B and was directly correlated to the intensity of DNA damage. These results were confirmed using different shRNAi against Cdc25B (Supplementary Fig. S5), suggesting that this was not due simply to an off-target effect of a single shRNAi.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Depletion of Cdc25B decreases the entry into mitosis after DNA damage. A, HCT116 cells were transfected with scramble (Scr) and Cdc25BB4#2 shRNAi. Clones were selected to enhance Cdc25B depletion. Asynchronous cells were harvested to analyzed Cd25B levels. B, Western blot analysis of phospho-histone (Ser¹⁰; mitosis marker) in Scr (controls cells) and Cdc25B shRNAi cells after mock or UV (60 J/m²) irradiation and subsequent treatment with 1 μmol/L nocodazole for 18 h to trap cells in mitosis. C, cells were fixed at the end of 18 h of trapping, and phospho-histone H3 positive cells were determined. D, bar graph and histogram representation of the data in (C). Columns, mean (n = 3–5); bars, SE. Statistical significance was determined by two-tailed unpaired t test. ***, P < 0.0001. Histogram: percentage, mean of cells in G₂-M.

To understand how increased levels of Cdc25B affected cellular progression after DNA damage, we exploited a previously described tetracycline-regulated U2OS cell system ( 33). Ectopic HA-Cdc25B expression was induced before treatment with either mock or UV irradiation ( Fig. 5A ). In the absence of DNA damage, 73.5 ± 1.4% of control cells (+tet) were trapped in mitosis by nocodazole, which was similar to Cdc25B overexpressing (−tet) cells (69.4 ± 1.1%; Fig. 5B and C). In the presence of DNA damage (UV at 30 J/m²), however, 4.1 ± 0.3% of UV exposed control (+tet) cells and 8.1 ± 0.5% of UV exposed Cdc25B overexpressing (−tet) cells were trapped in mitosis by nocodazole ( Fig. 5B–D). In our experiments, U20S cells were more sensitive to the toxic effects of UV compared with HCT116 cells (data not shown). Also, HCT116 cells are deficient in the mismatch repair protein hMLH1, which impairs their ability to arrest in G₂-M for prolonged period after DNA damage ( 42).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Cdc25B regulates cell cycle resumption after DNA damage. A, expression of HA-tagged Cdc25B in U2OS cells expressing Cdc25B under the control of tetracycline-regulated promoter for 16 h. Cdc25B was detected using a HA antibody. B, U2OS cells were treated as in Fig. 4B except the dose of the UV was 30 J/m² and cells were trapped for 23 h with nocodazole. C, cells were fixed at the end of 23 h of trapping and phospho-histone H3 levels were determined. D, bar graph and histogram representation of the data in (C). Columns, mean (n = 6); bars, SE. Statistical significance was determined by two-tailed unpaired t test. ***, P < 0.0001. Histogram: percentage, mean of cells in G₂-M.

Finally, if Cdc25B were important in the regulation of cell cycle resumption, then cells expressing more Cdc25B should exit the G₂-M checkpoint earlier than control cells. To test this hypothesis, we induced Cdc25B for 16 h by removing tetracycline and treated asynchronous cells with either mock or UV irradiation (15 J/m²). At 2, 4, and 8 h of posttreatment, phospho-histone H3 (Ser¹⁰) staining decreased indicative of a loss of the mitotic population and the engagement of a G₂-M checkpoint, which was independent of Cdc25B levels (Supplementary Fig. S6). In contrast, at 12 and 24 h, Cdc25B overexpressing cells (−tet) recovered from G₂-M checkpoint, whereas in control cells (+tet), the G₂-M checkpoint was still enforced. Similar results were observed by examining phospho-histone H3 (Ser¹⁰) using flow cytometry ( Fig. 6A ). Four hours after UV treatment, control and Cdc25B-overexpressing cells had fewer mitotic cells compared with the corresponding mock-treated cells consistent with the activation of the G₂-M checkpoint. When normalized to unirradiated controls, 0.36 ± 0.01% of control cells and 0.34 ± 0.02% of Cdc25B-overexpressing cells were in mitosis ( Fig. 6B). At 12 h posttreatment, the G₂-M checkpoint was still enforced in control cells (0.22 ± 0.01%) whereas Cdc25B-overexpressing cells resumed cell cycle followed by an increase in the percentage of the population in mitosis (0.86 ± 0.06%; Fig. 6A and B). Collectively, these results show that Cdc25B had a fundamental role in cells progression into mitosis after DNA damage and checkpoint exit.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Overexpression of Cdc25B accelerates resumption of cell cycle. A, Cdc25B was induced for 16 h before either mock or UV (15 J/m²) treatment. Cells were fixed at the indicated time points, and phospho-histone H3 levels were determined. B, bar graph representation of the data in (A). Columns, mean of the percentage of mitotic cells after UV irradiation normalized to unirradiated control (n = 4–5); bars, SE. The tetracycline concentration used is indicated next to the bars. Statistical significance was determined by two-tailed unpaired t test. ***, P < 0.0001.

Discussion

It is well documented that Cdc25s are targeted in response to DNA damage but the exact contributions of the three mammalian Cdc25 homologues toward the resumption of cell cycle progression are not completely understood. Interestingly, Cdc25A stabilization does not seem to be sufficient to overcome ionizing or UV irradiation–induced S phase checkpoints ( 16). In addition, Cdc25C knockout cells have normal G₂-M checkpoint response ( 43). These findings indicate additional regulators cooperate to regulate cell cycle progression after DNA damage. In this report, we describe a previously unrecognized elevation in endogenous Cdc25B after DNA damage generated by diverse types of genotoxic insults. We propose that elevation in Cdc25B is an anticipatory response to resume cell cycle after DNA damage–induced cell cycle arrest.

Our results are in contrast with one report ( 22), in which no increase in Cdc25B levels were observed 1 to 2 h after UV (20 J/m²) irradiation. One potential explanation for the differences in the experimental results could be the reagents used to detect endogenous Cdc25B. The nature and specificity of the antibody used in those experiments are not clear. At the time of the previous publication, Cdc25B null cells and Cdc25B shRNAi were unavailable. We have observed that several commercially available antibodies are useful when Cdc25B is overexpressed but fail to detect endogenous Cdc25B (data not shown). Other possible explanation could be the difference in the response of splice variants to DNA damage. There are five splice variants of Cdc25B ( 44), and it is possible that Bulavin et al. ( 22) were detecting a splice variant which was not regulated in response to DNA damage. This seem unlikely, however, as the monoclonal antibody used in our assay should detect all the splice variants. Nonetheless, this warrants further analysis.

The biochemical factors that regulate cell cycle resumption after DNA damage are still being defined. Interestingly, at least some factors that participate in the canonical checkpoint response seem to also regulate cell cycle resumption. Thus, Plk1, a well-characterized checkpoint target, is inhibited in response to DNA damage by ATM/ATR to induce cell cycle arrest ( 45, 46). Plk1 also regulates cell cycle resumption after DNA damage–induced cell cycle arrest ( 28). We now show that Cdc25B, which can be inhibited by p38/mitogen-activated protein kinase-activated protein kinase 2 phosphorylation at Ser³⁰⁹ after DNA damage ( 47), was rapidly induced after DNA damage and propose that this helps regulate cell cycle resumption. One could envisage that Cdc25B phosphorylation at Ser³⁰⁹ is operative only in the G₂ cell population to prevent premature entry into mitosis in the presence of DNA damage. The decrease in mitotic cells at 4 h after UV, which was independent of Cdc25B expression (Fig. 6), could be due to the lack of dephosphorylation of Cdc25B Ser³⁰⁹, which is important for progression through mitosis (48, 49). A key issue that needs to be resolved, however, is how opposing functions are orchestrated in an orderly manner to allow cells to recover from the effects of DNA-damaging agents without compromising genomic integrity.

One possible explanation is that in the period immediately after DNA damage, checkpoints regulating cell cycle arrest are more active, thus allowing time for DNA repair. This would be in agreement with our data showing that cells overexpressing Cdc25B were arrested normally after DNA damage. Nonetheless, they resumed cell cycle progression significantly earlier than control cells ( Fig. 6). As described previously ( 17, 20), degradation of Cdc25A is a key event in arresting cells in G₂ after DNA damage, whereas our data suggest that accumulation of Cdc25B regulates cell cycle resumption. Based on this, we proposed that initial cell cycle arrest after DNA damage might be more dependent on the loss of Cdc25A, whereas Cdc25B accumulation regulates cell cycle resumption. This could explain the absence of checkpoint defects in Cdc25BC double knockouts immediately after exposure to ionizing radiation ( 41). Cell cycle reentry after UV irradiation was not formally studied in this background. Although a rapid increase in Cdc25B expression would seem premature for its role in cell cycle resumption, temporally the Chk1 activity begins to decline 1 h after UV, which we have confirmed by examining Chk1 (Ser³⁴⁵) phosphorylation (activating phosphorylation) after UV (data not shown). These findings suggest a testable hypothesis that increased Cdc25B is a cellular priming factor for cell cycle resumption once DNA damage repair is complete and the ATR/Chk1 pathway prepares the cells for cell cycle resumption before inactivation of checkpoint signaling. Alternatively, gradual accumulation of Cdc25B after DNA damage may also be required to increase Cdk1 activity to a level that would be sufficient to overcome the inhibition accumulated during the DNA damage–induced cell cycle arrest. The dependence on increased Cdc25B levels as a limiting factor only after DNA damage was supported by our result that overexpression of Cdc25B did not significantly increase the number of cells in mitosis in the absence of DNA damage ( Fig. 5).

Van Vugt et al. ( 28) previously found Cdc25B and Plk1 were not required for recovery for DNA damage in Wee1 depleted cells. Cdk1 regulates Wee1 degradation at the onset of mitosis ( 50). Functionally, one attractive hypothesis for how elevated Cdc25B might regulate cell cycle reentry after DNA damage could be by promoting Wee1 degradation through activation of Cdk1, thus ensuring sustained activation of Cdk1.

The important role Cdc25B plays in regulating cell cycle resumption could also help explain the oncogenic properties of this phosphatase. It is possible that overexpression of Cdc25B in tumor cells overwhelms the process of cell cycle resumption and survival, resulting in division of cells with damaged DNA, thereby contributing to genomic instability observed in cancer cells. This could explain the increased susceptibility to breast cancer induction by 9,10-dimethyl-1,2-benzanthracene in mice that ectopically overexpressed Cdc25B in mammary glands ( 11). Finally, our results reinforce the attractive nature of Cdc25B inhibition as an adjuvant approach for anticancer therapy because inhibition of Cdc25B by small molecule inhibitor might impair checkpoint recovery and increase the efficacy of DNA-damaging agents.