Changes in Regulatory Phosphorylation of Cdc25C Ser287 and Wee1 Ser549 during Normal Cell Cycle Progression and Checkpoint Arrests

Abstract

Entry into mitosis is catalyzed by cdc2 kinase. Previous work identified the cdc2-activating phosphatase cdc25C and the cdc2-inhibitory kinase wee1 as targets of the incomplete replication-induced kinase Chk1. Further work led to the model that checkpoint kinases block mitotic entry by inhibiting cdc25C through phosphorylation on Ser287 and activating wee1 through phosphorylation on Ser549. However, almost all conclusions underlying this idea were drawn from work using recombinant proteins. Here, we report that in the early Xenopus egg cell cycles, phosphorylation of endogenous cdc25C Ser287 is normally high during interphase and shows no obvious increase after checkpoint activation. By contrast, endogenous wee1 Ser549 phosphorylation is low during interphase and increases after activation of either the DNA damage or replication checkpoints; this is accompanied by a slight increase in wee1 kinase activity. Blocking mitotic entry by adding the catalytic subunit of PKA also results in increased wee1 Ser549 phosphorylation and maintenance of cdc25C Ser287 phosphorylation. These results argue that in response to checkpoint activation, endogenous wee1 is indeed a critical responder that functions by repressing the cdc2-cdc25C positive feedback loop. Surprisingly, endogenous wee1 Ser549 phosphorylation is highest during mitosis just after the peak of cdc2 activity. Treatments that block inactivation of cdc2 result in further increases in wee1 Ser549 phosphorylation, suggesting a previously unsuspected role for wee1 in mitosis.

INTRODUCTION

Entry into mitosis is initiated by activation of cyclin B/cdc2. Preformed complexes of cyclin B/cdc2 accumulate during interphase, but their activity is repressed by inhibitory phosphorylations on cdc2 at Tyr15 (catalyzed by wee1 and myt1) and Thr14 (catalyzed by myt1). These phosphorylations are removed by the phosphatase cdc25C (reviewed in Berry and Gould, 1996; Lew and Kornbluth, 1996). Early work led to the conclusion that cdc2 and cdc25C activities both increase rapidly during the G₂/M transition as the result of positive feedback loops between cyclin B/cdc2 and cdc25C, ultimately leading to the full activation of both cdc25C and cyclin B/cdc2 (Izumi et al., 1992; Hoffmann et al., 1993; Izumi and Maller, 1993; Strausfeld et al., 1994). The mechanisms that trigger this abrupt activation are still not well understood (Margolis and Kornbluth, 2004; Perdiguero and Nebreda, 2004).

Over the past decade, there has been a large increase in the molecular understanding of how checkpoint pathways become activated, and how they affect the basic cell cycle regulatory machinery. Genetic and biochemical approaches have identified sensors that detect incompletely replicated or damaged DNA and that stimulate signal transduction pathways that lead to activation of the kinases Chk1 and Chk2/Cds1 (reviewed in Sancar et al., 2004). Both Chk1 and Chk2 can phosphorylate cdc25C on Ser287 (human Ser216), which creates a binding site for 14-3-3 proteins; this is thought to inhibit cdc25C activation and thus the G₂/M transition (Peng et al., 1997; Sanchez et al., 1997; Kumagai et al., 1998b; Matsuoka et al., 1998; Zeng et al., 1998). The mechanisms by which phosphorylation of this residue and 14-3-3 binding suppress cdc25C's ability to dephosphorylate cdc2, including changes in its subcellular localization in some types of cell cycles (Dalal et al., 1999; Kumagai and Dunphy, 1999; Lopez-Girona et al., 1999; Yang et al., 1999; Graves et al., 2001), are still being worked out.

Kumagai et al. (1998b) were the first to suggest that phosphorylation of cdc25C on Ser287 might also play a role in normal cell cycles. This idea was based on observations that a significant amount of endogenous cdc25C coprecipitated with 14-3-3 from unperturbed Xenopus egg interphase extracts and that association of 14-3-3 with recombinant cdc25C protein was dependent on cdc25C phosphorylation on Ser287. In addition to the checkpoint kinases, several others can phosphorylate cdc25C on Ser287. C-TAK1, identified as a human Ser216 phosphorylating activity from mammalian somatic cells, was the first to be described (Ogg et al., 1994; Peng et al., 1998). Protein kinase A (PKA), a well-known inhibitor of the G₂/meiosis I transition, helps to maintain Xenopus oocytes in their natural G₂ arrest through phosphorylation of cdc25C on Ser287 (Duckworth et al., 2002; Schmitt and Nebreda, 2002). In fertilized Xenopus eggs, calmodulin-dependent protein kinase II (CaMKII) seems to be responsible for the majority of Ser287 phosphorylation during interphase of the first mitotic cell cycle (Hutchins et al., 2003). PRK/PLK3, RSK, and MAPKAP kinase-2 have also been implicated as cdc25C Ser287 (human Ser216) kinases (Ouyang et al., 1999; Chun et al., 2005; Manke et al., 2005).

The complete removal of the inhibitory Ser287 phosphorylation on cdc25C seems to require phosphorylation of cdc25C by cdk2 during interphase and by other kinases during the G₂/M transition. Phosphorylation of recombinant cdc25C on Thr138 by cdk2 is required for the subsequent release of 14-3-3, which is thought to make Ser287 more accessible to dephosphorylation by the phosphatases PP1 and/or PP2A (Hutchins et al., 2002; Margolis et al., 2003). However, cdk2 phosphorylation of cdc25C is not sufficient for 14-3-3 removal. During mitosis, cdc25C becomes phosphorylated at multiple sites, leading to the formation of hyperphosphorylated, electrophoretically retarded forms (Izumi et al., 1992; Kumagai and Dunphy, 1992). The cdc2-dependent mitotic phosphorylation of cdc25C on Ser285 (human Ser214) seems to prevent rephosphorylation of Ser287 during mitosis, presumably strengthening the positive feedback loop (Bulavin et al., 2003a,b).

So far, virtually all conclusions about the inhibitory role of cdc25C Ser287 phosphorylation have come from studies that relied almost exclusively on recombinant cdc25C proteins (Peng et al., 1997; Sanchez et al., 1997; Kumagai et al., 1998a,b; Matsuoka et al., 1998; Peng et al., 1998; Zeng et al., 1998; Dalal et al., 1999; Kumagai and Dunphy, 1999; Hutchins et al., 2002; Margolis et al., 2003). Remarkably, almost nothing is known about how the Ser287 phosphorylation status of endogenous cdc25C changes during checkpoint arrest. Moreover, nothing is known about the kinetics of Ser287 dephosphorylation during the G₂/M transition or about the timing or the regulation of the rephosphorylation of cdc25C on Ser287 during mitotic exit.

Even less is known about the regulation of endogenous wee1 Ser549. In mammalian somatic cells, as in Xenopus eggs, a portion of wee1 has been reported to bind 14-3-3 during interphase, but not during M phase, and this binding requires phosphorylation of wee1 on Ser549 (human Ser642) (Honda et al., 1997; Wang et al., 2000; Lee et al., 2001; Rothblum-Oviatt et al., 2001). Phosphorylation of this residue does not seem to directly increase wee1's kinase activity; instead, it promotes 14-3-3 binding, which then enhances the activity of recombinant wee1 two- to threefold (Lee et al., 2001; Rothblum-Oviatt et al., 2001). The wee1 Ser549Ala recombinant protein, or equivalent mammalian mutant protein, is less able than wild-type wee1 to maintain the cell cycle in interphase, suggesting that phosphorylation of this site during interphase is important for regulating mitotic entry (Lee et al., 2001; Rothblum-Oviatt et al., 2001). Chk1 can phosphorylate wee1 on Ser549 in vitro (Lee et al., 2001), leading to the idea that activation of Chk1 during a checkpoint response would increase the level of active wee1. However, the status of Ser549 phosphorylation on endogenous wee1 has never been examined.

Here, we investigated how the phosphorylation of endogenous cdc25C Ser287 and wee1 Ser549 changes during the normal early cell cycles of Xenopus eggs, and after induction of the DNA replication and damage checkpoints that result in G₂ arrest. We find that phosphorylation of cdc25C Ser287 is high during interphase of the normal cell cycle and shows no obvious increase after checkpoint activation. By contrast, wee1 Ser549 phosphorylation is very low during interphase and increases substantially in response to checkpoint activation. This checkpoint-induced increase in Ser549 phosphorylation is accompanied by a slight increase in wee1's kinase activity toward cdc2. Surprisingly, wee1 phosphorylation is highest in mid-mitosis, peaking sharply right after cdc2 inactivation, a time when wee1's kinase activity toward cdc2 is even lower than in interphase. These results raise the possibility that, in addition to increasing wee1 activity during DNA checkpoint arrest, Ser549 phosphorylation plays other roles during normal mitotic progression as well.

MATERIALS AND METHODS

Xenopus Egg Extracts

Egg and extract protocols were based on Murray (1991). Xenopus laevis females from the colony in the Cell Biology Department (Harvard Medical School, Boston, MA) were primed with 50 U of pregnant mare serum gonadotropin (PMSG, Sigma-Aldrich, St. Louis, MO) at least 3 d before human chorionic gonadotropin (HCG, Sigma-Aldrich) injection. Ovulation was then induced by injection of 500 U of HCG. Frogs were placed in individual tanks containing 1× MMR (100 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 2 mM CaCl₂, 0.1 mM EDTA, and 5 mM HEPES, pH to 7.8 [NaOH]). Laid eggs were used to make extracts. Because egg quality deteriorates over time, eggs were used within 17 h of HCG injection. All buffers used in making the extract were prepared fresh on the day of the experiment. Dejellying solution was prepared no more than an hour before use [100 mM KCl, 0.1 mM CaCl₂, 1 mM MgCl₂, and 2% (wt/vol) cysteine, free base, pH 7.8]. Eggs were gently washed in 1× MMR to remove detritus and were dejellied.

For extracts of metaphase II-arrested eggs (cytostatic factor [CSF] extracts), eggs were washed in XB (100 mM KCl, 0.1 mM CaCl₂,1 mM MgCl₂, 10 mM potassium HEPES, pH 7.7, and 50 mM sucrose), followed by washing in CSF-XB (100 mM KCl, 0.1 mM CaCl₂, 2 mM MgCl₂, 10 mM potassium HEPES, pH 7.7, 50 mM sucrose, and 5 mM EGTA, pH 7.7). Eggs were washed with CSF-XB + protease inhibitors (leupeptin, chymostatin, pepstatin A, and 10 μg/ml final concentration; Sigma-Aldrich), and then pipetted into Ultra-Clear centrifuge tubes (MI344057, Beckman Coulter, Fullerton, CA) containing CSF-XB + protease inhibitors (10 μg/ml final concentration) + cytochalasin B (100 μg/ml final concentration; Sigma-Aldrich). Eggs were packed by spinning at 860 rpm in a Sorvall HB-6 rotor for 1 min, and excess buffer was aspirated. The crushing spin was performed at 16°C, 11,500 rpm, 15 min in a Sorvall HB-6 swinging bucket rotor. The cytoplasmic layer was removed by puncturing the tube with an 18-gauge needle directly above the yolk layer, and allowing this layer to drain into a collection tube. Protease inhibitors were added at 1:1000, and cytochalasin B at 1:1000. Extracts were mixed by gentle inversion of the tube and were clarified by spinning at top speed in a tabletop centrifuge for 15 min at 4°C. The remaining lipid layer was removed by aspiration. The cytoplasmic layer was transferred using a wide bore pipette tip into a prechilled tube. Then, 1/20th the volume of CSF energy mix (150 mM creatine phosphate, 20 mM ATP, and 20 mM MgCl₂, pH to 7.4) was added to the extract, and this was mixed by gentle inversion. To follow the behavior of cdc25C and wee1 during release from metaphase II arrest, CSF extracts were shifted to 20°C and activated by calcium by addition of calcium chloride in sperm dilution buffer (1 mM MgCl₂, 100 mM KCl, 150 mM sucrose, and 5 mM HEPES, pH 7.6-7.7) to a final concentration of 0.44 mM. To prepare extracts of activated of eggs that were arrested in interphase, CSF extracts were activated by calcium in the presence of 100 μg/ml cycloheximide and collected 45 min later. For preparation of extracts arrested in mitotic M phase, interphase extract was prepared as above and then supplemented with 40 μg/ml MBP-Δ90 Cyclin B protein.

Cycling extracts were prepared as follows. Eggs at 16°C were activated with Ca²⁺ ionophore (A231187 [Fisher Scientific, Pittsburgh, PA], 1/10,000 dilution of a 1 mg/ml stock in dimethyl sulfoxide, diluted in 1× MMR, and mixed rapidly to prevent precipitation). After 5 min, eggs were washed three times in XB. Twenty minutes later, eggs were washed in XB plus protease inhibitors (1:1000 dilution). Then, 10 μl of cytochalasin B (1:1000 dilution) was added to each Ultra-Clear centrifuge tube containing 1 ml of XB plus protease inhibitors (1:1000 dilution) and mixed immediately to prevent precipitation. As described above, eggs were pipetted into the prepared tubes and subjected to a packing spin and crushing spin, except both spins were done at 4°C. The cytoplasmic layer was harvested and was subjected to a clarifying spin as described above. Then, 1/20th the volume of cycling energy mix (150 mM creatine phosphate, 20 mM ATP, 20 mM MgCl₂, and 2 mM EGTA, pH 7.4) was added, and extracts were mixed by gentle inversion. Unless otherwise indicated, all experiments were performed at 20°C.

Sperm Nuclei Preparation

Sperm nuclei were prepared as described previously (Murray, 1991). Four male frogs were primed with 25 U of PMSG 3 d before sperm collection. The day before collection, frogs were primed with 125 U of HCG. Frogs were killed by injection with 100 μl of MS-222 (Sigma-Aldrich) in 0.5 ml of H₂O. Twenty minutes later, testes were dissected. The rest of the procedure was performed exactly as described in Murray (1991), except sperm were incubated in lysolecithin (Sigma-Aldrich) for 8 min.

DNA Damage and Incomplete Replication Checkpoint Arrests

To induce the DNA damage checkpoint, cycling extracts containing 200 sperm nuclei/μl were supplemented 20 ng/μl EcoRI-linearized, phenol-chloroform extracted pGAPT7 (Promega, San Luis Obispo, CA) plasmid, based on methods described in Guo and Dunphy (2000). The incomplete replication checkpoint was stimulated in cycling extracts containing at least 2000 nuclei/μl by the addition of 0.1 μg/μl aphidicolin (Sigma-Aldrich).

Visualization of Sperm Nuclei Morphology

Sperm nuclei were added to extracts to a final concentration of 500/μl unless otherwise indicated. One microliter of extract containing 500 nuclei/μl was removed per time point studied and was added to 4 μl of fix [60% (vol/vol) glycerol, 1 μg/ml Hoechst 33342/33258, and 10% formaldehyde, in 1× MMR] on a precleaned slide. Corning coverglass (VWR, Bridgeport, NJ), no. 1.5, 18 mm², was used to cover the sample. Slides were examined using a Nikon (Garden City, NY) E800 Eclipse upright microscope in the Nikon imaging facility in the Department of Cell Biology (Harvard Medical School). Photos of the sperm nuclei morphology were taken with the Hamamatsu Orca 100 cooled charge-coupled device camera, and images were manipulated using MetaMorph software (Molecular Devices, Sunnyvale, CA).

Constructs and Protein Expression

Recombinant viruses were used to generate glutathione S-transferase (GST)-cyclin B1-Cdc2 (K33R) kinase-inactive complexes in Sf9 cells (Atherton-Fessler et al., 1993; Lee et al., 1994), which were then purified as described previously (Liu et al., 1997). pFastBac-His6-GST-Xwee1 and pFastBac-His6-GST-Xwee1-S549A expression vectors (Lee et al., 2001) were used to express Xenopus wee1 in Sf9 cells (Invitrogen, Carlsbad, CA). Briefly, 50 ml of 10⁶ Sf9 cells (Invitrogen) were infected with recombinant virus and harvested after 48 h. The cell pellet was resuspended in 1.2 ml of Sf9 lysis buffer (20 mM Tris-HCl, pH 8.5, 120 mM NaCl, 1% CHAPS, 10 mM β-mercaptoethanol, and 1 Complete EDTA-free protease inhibitor tablet per 50 ml [Roche Diagnostics, Indianapolis, IN]), vortexed vigorously, and spun at 20,000 × g at 4°C for 10 min. Then, 100 μl of a 50% slurry of nickel-nitrilotriacetic acid agarose beads (QIAGEN, Valencia, CA) that had been prewashed in lysis buffer was added to the supernatant and incubated with rocking at 4°C for 1 h. Beads were washed three times with wash buffer (20 mM Tris-HCl, pH 8.5, 500 mM KCl, 1% CHAPS, 10 mM β-mercaptoethanol, and 1 Complete EDTA-free protease inhibitor tablet per 50 ml). If proteins were λ-phosphatase treated, beads were resuspended in wash buffer and were split into two aliquots. One aliquot was left in wash buffer until the elution step. The other aliquot was washed three times with λ-phosphatase buffer (50 mM Tris, pH 7.5, 0.1 mM EDTA, 0.01% CHAPS, 2 mM MnCl₂, 10 mM β-mercaptoethanol, and 1 EDTA free Complete protease inhibitor tablet per 50 ml, pH to 7.5). Beads were resuspended in 300 μl of λ-phosphatase buffer. Then, 1 μl of λ-phosphatase per 100 μl of buffer was added, and this was incubated for 30 min at 30°C, mixing every 5 min. Beads were washed three times with wash buffer. Proteins were eluted with 3 μl of elution buffer (20 mM Tris-HCl, pH 8.5, 100 mM KCl, 250 mM imidazole, 1% CHAPS, 10% glycerol, 100 μg/ml ovalbumin, 10 mM β-mercaptoethanol, and 1 Complete EDTA-free protease inhibitor tablet per 50 ml) per each microliter of beads.

Immunoblotting

All samples were analyzed by SDS-PAGE by using the Anderson gel system unless otherwise indicated (Anderson et al., 1973). Blots were incubated in specific blocking solutions for 1 h at room temperature (RT), washed three times with Tris-buffered saline/Tween 20 (TBST) (50 mM Tris, 150 mM NaCl, and 1% Tween 20), incubated in primary antibody overnight at 4°C, washed three times with TBST, incubated in secondary antibody for 1 h at RT, washed three times with TBST, developed with ECL (GE Healthcare, Little Chalfont, Buckinghamshire, United Kingdom) or ECL Plus (GE Healthcare), and visualized by chemiluminescence. All primary and secondary antibodies were prepared in TBST. For α-phospho-MAPK antibodies (#9101S; Cell Signaling Technology, Beverly, MA), blots were blocked with 3% milk, primary antibody was diluted 1:5000 in 1% bovine serum albumin (BSA), and secondary antibody was donkey-α-rabbit (#NA 934; GE Healthcare) diluted 1:5000. For α-phospho-cdc2 antibodies (#9111L; Cell Signaling Technology), blots were blocked with 3% milk, primary antibody was diluted 1:1000 in 1% BSA, and secondary antibody was donkey-α-rabbit diluted 1:5000 in 3% milk. For α-cdc25C-phospho-S287 antibodies (Duckworth et al., 2002), blots were blocked with 2% BSA, primary antibody was diluted 1:1000 in 1% BSA, and secondary antibody was donkey-α-rabbit diluted 1:10,000 in 2% BSA. For α-cdc25C antibodies (Kumagai and Dunphy, 1992), blots were blocked with 5% milk, primary antibody was diluted to a concentration of 1 μg/ml in 5% milk, and secondary antibody was donkey-α-rabbit diluted 1:5000 in 5% milk. For α-CycB1 (Gautier and Maller, 1991), blots were blocked with 3% milk, primary antibody was diluted 1:10,000 in 1% BSA, and secondary antibody was donkey-α-sheep (#713-035-147; Jackson ImmunoResearch Laboratories, West Grove, PA) diluted 1:50,000 in 2.5% donkey serum (#017-000-121; Jackson ImmunoResearch Laboratories). For α-GST (#sc-138; Santa Cruz Biotechnology, Santa Cruz, CA), blots were blocked with 3% milk, primary antibody was diluted 1:1000 in 2% BSA, and secondary antibody was sheep-α-mouse diluted 1:5000 in 1% BSA. For α-wee1 antibodies (#51-1700; Zymed Laboratories, South San Francisco, CA), blots were blocked with 3% milk, primary antibody was diluted 1:500 in 1% BSA, and secondary antibody was donkey-α-rabbit (#4050-05; Southern Biotechnology Associates, Birmingham, AL) diluted 1:5000 in 3% milk. For α-wee1-phospho-Ser549 antibodies, blots were blocked with 1% milk, primary antibody was diluted 1:100 in 1% BSA, and secondary antibody was goat-α-rabbit (#4050-05; Southern Biotechnology Associates) diluted 1:5000 in 1% milk. Blots were washed three times for 10 min each wash after primary and secondary antibody incubation. Blots were developed using ECL Plus (GE Healthcare).

Generation of Wee1 Phospho-Ser549 Antibodies

Phospho-specific antibodies to Xenopus embryonic wee1 Ser549 were generated by Invitrogen, using the phospho-peptide sequence AKNTRSL-pS-FTCGGY conjugated to keyhole limpet hemocyanin. Two rabbits were injected with this conjugated peptide on days 1, 13, 43, and 56, and then once a month for the next 3 mo the rabbits were boosted and bled. The serum from the third extension bleed from one of the rabbits was affinity purified.

Immunodepletion of Wee1

Thirty microliters of Dynabeads (#100.01; protein A, Dynal, Brown Deer, WI) were incubated with 40 μl of wee1 antibody (#51-1700, 0.25 mg/ml concentration; Zymed Laboratories) in 60 μl of XB (100 mM KCl, 0.1 mM CaCl₂, 1 mM MgCl₂, 10 mM potassium HEPES, pH 7.7, and 50 mM sucrose) + 0.1% Triton X-100, for 2 h with inverting at 4°C. The supernatant was spun off, and 100 μl of CSF extract was added. This was incubated at 4°C for 1 h with mixing to deplete wee1. This protocol was derived from suggestions from Sun Young Kim and Joe Pomerening in Jim Ferrell's laboratory (Stanford University, Stanford, CA).

Wee1 Kinase Assays

Wee1 immunoprecipitates were prepared and tested for their ability to phosphorylate cdc2 on Tyr15 as follows. Twenty microliters of Dynabeads (#100.01; protein A, Dynal, Brown Deer, WI) were incubated with 10 μl of wee1 antibody (#51-1700, 0.25 mg/ml concentration; Zymed Laboratories) in 50 μl of XB (100 mM KCl, 0.1 mM CaCl₂, 1 mM MgCl₂, 10 mM potassium HEPES, pH 7.7, and 50 mM sucrose) + 0.1% Triton X-100, for 2 h with inverting at 4°C. Supernatant was removed from the beads, and 50 μl of extract and 50 μl of IP buffer (10 mM HEPES-KOH, pH 7.5, 150 mM NaCl, and 20 mM β-glycerophosphate) were added to the beads, which were mixed with inverting for 1 h at 4°C. Beads were rinsed twice with IP wash buffer (10 mM HEPES-KOH, pH 7.5, 150 mM NaCl, 20 mM β-glycerophosphate, and 0.1% NP-40), followed by two washes with kinase wash buffer (20 mM Tris, pH 7.5, 10 mM MgCl₂, 20 mM β-glycerophosphate, and 0.1 mM sodium orthovanadate). After removal of the last wash, 10 μl of kinase buffer (20 mM Tris, pH 7.5, 10 mM MgCl₂, 1 mM dithiothreitol, and 5 μM ATP) was added to each tube. Then, 0.2 μg of GST-cycB/cdc2 (K33R) complex was added, and samples were incubated for 25 min at 30°C. Reactions were terminated with 10 μl of SDS sample buffer. Samples were run analyzed by SDS-PAGE followed by blotting with the anti-phospho-Y15 antibody. This protocol represents slight modifications of previously described methods (Liu et al., 1997; Lee et al., 2001; Mueller and Leise, 2005) and of suggestions from Sun Young Kim and Joe Pomerening in the laboratory of Jim Ferrell (Stanford University).

RESULTS

Oscillations in Phosphorylation of Endogenous Cdc25C Protein during the Normal Early Embryonic Cell Cycles

In previous work, we developed and characterized a phospho-specific antibody that specifically recognizes the Ser287-phosphorylated form of cdc25C but not the unphosphorylated form (Duckworth et al., 2002). In that work, we examined changes in Ser287 phosphorylation across the early cell cycles of fertilized Xenopus eggs and found that endogenous cdc25C Ser287 phosphorylation is low in M phase and high in interphase. However, the detailed timing of those changes relative to well-characterized biochemical and morphological markers of cell cycle progression was not investigated. To examine this more carefully, we used two types of egg extracts, CSF extracts and cycling extracts (see Materials and Methods), which go through cell cycle transitions with high synchrony. To follow Ser287 phosphorylation as eggs exit from metaphase II arrest, eggs were crushed, a low-speed supernatant was prepared (CSF extract), and calcium was added to break the metaphase arrest and initiate entry into the first interphase. Samples were taken at the indicated times and analyzed by SDS-PAGE followed by immunoblotting with the phospho-cdc25C Ser287 antibody. As shown in Figure 1A, cdc25C Ser287 phosphorylation first became detectable within 10 min after cyclin B destruction, a marker for cdc2 inactivation. This phosphorylation was first evident on the hyperphosphorylated form of cdc25C (see short and long exposures), and Ser287 phosphorylation increased as the M-phase-specific hyperphosphorylations were lost. By 30 min, when the extract was in the first interphase, cdc25C Ser287 phosphorylation had reached its highest intensity and then remained constant throughout interphase.

Figure 1.

Timing of cdc25 Ser287 phosphorylation and dephosphorylation. (A) CSF extracts of metaphase II-arrested eggs were activated by the addition of calcium. Extract samples were analyzed by SDS-PAGE and immunoblotting with the antibodies shown. (B) Intact eggs were activated by the addition of calcium ionophore, and cycling extracts were prepared and sampled as described in the text. Samples were fixed, stained with Hoechst 33342, and visualized by fluorescence microscopy at 40× magnification. (C) Samples from the extract shown in B were analyzed by SDS-PAGE followed by blotting with indicated antibodies. The diagram at the bottom is representative of the sperm nuclei morphologies observed in the extracts, shown in detail in B. The thin line represents decondensed DNA, and the shaded shape represents condensed chromosomes.

Cycling extracts were used to follow oscillations of cdc25C Ser287 phosphorylation across the mitotic cycles. Briefly, eggs were activated by the addition of calcium ionophore and incubated at 16°C for 30 min, during which time cyclin B is degraded, cdc2 is inactivated, mitogen-activated protein kinase (MAPK) activity declines, and the eggs enter interphase of the first mitotic cell cycle. At 30 min, eggs were shifted rapidly to 4°C to stop cell cycle progression, buffer was removed, and eggs were crushed by centrifugation at 11,500 rpm in a Sorvall HB-6 rotor. The cytosolic fraction was collected and kept on ice. A low concentration of demembranated sperm nuclei was added. The extract was then transferred to 20°C to initiate resumption of cell cycle progression. The first time point after the extracts were transferred to 20°C is designated as the 30-min time point to reflect the time that has passed since egg activation. Because exact timing of cell cycle events varies among extracts, biochemical and morphological markers of cell cycle state, rather than absolute time, were used to follow the stages of cell progression.

At 50 min, the extract shown in Figure 1, B and C, was in interphase of the first cell cycle, as judged by the presence of round nuclei containing decondensed chromatin, and remained in interphase until 85-90 min. By 90 min, nuclear envelope breakdown (NEBD) and chromosome condensation had occurred. By 110 min, the extract had exited from mitosis into interphase of the second cell cycle. Immunoblotting with the phospho-specific cdc25C Ser287 antibody revealed a single strong band at 50 min. The intensity and electrophoretic mobility of this band were approximately constant during the remainder of interphase, where it ran at the same position as total cdc25C protein. At 85 min, around the time of the G₂/M transition, the cdc25C Ser287 phosphorylation signal changed dramatically. It underwent a strong retardation in its electrophoretic mobility, which paralleled that seen for total cdc25C protein. This is a rapid event that was not captured in all experiments (Figures 2 and 3). By 90 min, when NEBD and chromosome condensation had occurred, phosphorylation of cdc25C Ser287 was no longer detected, although the total cdc25C protein level remained constant, as seen previously (Gabrielli et al., 1996; Nishijima et al., 1997; Lammer et al., 1998). Destruction of cyclin B, a marker for the metaphase/anaphase transition, occurred between 90 and 95 min. The tyrosine-phosphorylated (active) form of MAPK first occurred around this time, which is important for the maintenance of chromosome condensation during the latter stages of mitosis (Bitangcol et al., 1998; Chau and Shibuya, 1998, 1999). MAPK phosphorylation increased and then dropped at 105-110 min, when chromosome decondensation was being completed, and the extract entered interphase of the second cell cycle. Rephosphorylation of cdc25C Ser287 first occurred during the latter stages of mitotic exit after the downshift in total cdc25C protein and increased during interphase of the second cell cycle.

Figure 2.

Activating the incomplete replication or DNA damage checkpoint maintains but does not increase the interphase level of cdc25C Ser287 phosphorylation. Five hundred sperm nuclei/μl (control), 4000 nuclei/μl, and 0.17 μg/μl aphidicolin (incomplete replication checkpoint) or 500 nuclei/μl and 20 ng/μl linearized plasmid (DNA damage checkpoint) were added to cycling extracts on ice at 30 min. Extracts were then moved to 20°C to reinitiate cell cycle progression. Samples were taken at indicated times for analysis by blotting and microscopy. The diagram at the bottom of each figure represents the cell cycle stage at each time point, as judged by nuclear morphology.

Figure 3.

Dephosphorylation of cdc25C Ser287 requires cdk activity. Cycling extracts were prepared in the absence (control) or presence of 100 μg/ml cycloheximide (+CHX). After the crushing spin at 4°C, and addition of sperm nuclei (500/μl), water, or 40 μg/ml MBP-Δ90 cyclin B (+ΔB), extracts were moved to 20°C to resume cell cycle progression. Samples were taken at indicated times and analyzed as described in Figure 2 legend.

Activation of the Incomplete Replication or DNA Damage Checkpoint Results in the Maintenance, but not an Increase, of Cdc25C Ser287 Phosphorylation

When Xenopus egg extracts are arrested by the incomplete replication or DNA damage checkpoint, the addition of recombinant cdc25C Ser287Ala mutant protein was more potent than wild type at overriding the arrest and inducing mitotic entry (Kumagai et al., 1998b). Thus, cdc25C Ser287 phosphorylation was thought to be important for preventing inappropriate entry into mitosis. However, the phosphorylation status of endogenous cdc25C Ser287 under checkpoint conditions has not been examined. To stimulate the incomplete replication checkpoint, cycling extracts were supplemented with 4000 sperm nuclei/μl, a concentration that is sufficient for checkpoint activation, and then with aphidicolin (0.17 μg/μl) to block DNA synthesis. In the control extract, cdc25C Ser287 dephosphorylation and mitotic entry occurred ≈70-80 min (Figure 2, top). The aphidicolin-treated extract arrested in interphase: chromosomes did not condense (our unpublished data), cyclin B was not degraded, and MAPK rephosphorylation did not occur (Figure 2, middle). Phosphorylation of cdc25C Ser287 was maintained throughout the interphase arrest, indicating that the checkpoint blocked dephosphorylation of endogenous cdc25C. However, checkpoint activation did not lead to any significant increase in the level of cdc25C Ser287 phosphorylation. Activation of the DNA damage checkpoint, by addition of linearized plasmid DNA (20 ng/μl) to extracts, blocked entry into mitosis and resulted in the maintenance of interphase levels of Ser287 phosphorylation on endogenous cdc25C but not an increase in phosphorylation of Ser287 (Figure 2, bottom).

Cyclin B/Cdc2 Activity Is Required for Full Dephosphorylation of Cdc25C Ser287

Previous work carried out with recombinant cdc25C protein indicated that cdk2 activity is required for dephosphorylation of cdc25C Ser287 (Margolis et al., 2003). In agreement, we find that inhibition of cdk activity by roscovitine completely blocked dephosphorylation of Ser287 on endogenous cdc25C (our unpublished data). In further examining the behavior of endogenous cdc25C (Figure 3), we found that when cdc2 activation was prevented by blocking synthesis of cyclins and other newly made proteins with cycloheximide, Ser287 dephosphorylation did not occur. Addition of nondegradable cyclin B (cyclin ΔB) to the cycloheximide-arrested extract, which forces activation of cdc2, induced dephosphorylation of cdc25C Ser287 and mitotic entry. Together, these results support the idea that even though a small amount of cdc25 activity is required for the initial activation of a small amount of cdc2, the majority of cdc25C Ser287 dephosphorylation is a consequence of cdc2 activity.

Phosphorylation of Wee1 on Ser549 Oscillates across the Cell Cycle

Wee1 Ser549 (human wee1 Ser642) has been described as a second important interphase target of Chk1 (Lee et al., 2001; Rothblum-Oviatt et al., 2001). Thus, it has been predicted that activation of the DNA replication or damage checkpoint would lead to an increase in wee1 Ser549 phosphorylation and wee1 activity. Furthermore, by analogy to cdc25C, it might have been assumed that at least a portion of endogenous wee1 is phosphorylated on Ser549 during interphase of the normal cell cycle and that inactivation of wee1 may be accompanied by dephosphorylation of Ser549. However, no studies have looked at the phosphorylation of endogenous wee1 on Ser549, either during the normal cell cycle or under checkpoint arrest.

To examine these questions, we developed a phospho-specific wee1 Ser549 antibody. As shown in Figure 4A, the antibody recognized recombinant wee1 protein that had been expressed in Sf9 cells, but it failed to react with either phosphatase-treated wee1 or with wee1 Ser549Ala mutant protein. On blots of total endogenous egg extract, the phospho-wee1 Ser549 antibody recognized a protein that comigrated with total wee1 protein (Figure 4B). Depletion of wee1 protein removed almost all of the material that reacted with the phospho-specific wee1 Ser549 antibody. These results indicate that the phospho-Ser549 signal on blots of total egg extract is due to wee1 rather than a cross-reacting protein.

Figure 4.

Wee1 Ser549 phosphorylation is low during interphase and rises sharply during mid-mitosis. (A) A wee1 Ser549 phospho-specific antibody, generated as described in Materials and Methods, was blotted onto 25 ng of the following GST-tagged wee1 proteins prepared from Sf9 cells. Lane 1, wt wee1; lane 2, λ-phosphatase-treated wild-type protein; and lane 3, Ser549Ala mutant protein. Samples were analyzed by SDS-PAGE and blotting with affinity-purified phospho-Ser549 wee1 antibodies. The blots were stripped and reprobed for anti-GST to verify the amounts of wee1 protein loaded. (B) CSF extracts were depleted with 10 μg of total wee1 antibody (Zymed Laboratories) or control IgG per 100 μl of extract. Nondepleted, IgG- and wee1-depleted samples were analyzed by SDS-PAGE and blotted with either total wee1 (left) or wee1 phospho-Ser549 antibodies (right). (C) CSF extracts were activated with calcium, and samples were taken at the indicated times and analyzed as described in Figure 2 legend. (D) Sperm nuclei were added at a concentration of 500/μl to cycling extracts on ice. Extracts were then moved to 20°C to resume cell cycle progression. Samples taken at the indicated times were analyzed by immunoblotting and microscopy. (E) Top, samples from metaphase II-arrested extracts (MII; CSF extracts), interphase extracts (Int; CSF + calcium + cycloheximide) or mitotic extracts (M1; CSF + calcium, + cycloheximide, + nondegradable cyclin B) were analyzed by SDS-PAGE and blotting for wee1 phospho-Ser549. Bottom, Wee1 protein was immunoprecipitated with 2.5 μg of wee1 antibody from each of these extracts. Immunoprecipitated wee1 was then used to phosphorylate kinase dead cdc2-cyclin B protein complexes for 25 min at 30°C. Samples were analyzed by SDS-PAGE and blotting for cdc2 anti-phospho-Y15.

Unexpectedly, immunoblotting across the first two cell cycles (Figure 4, C and D) revealed that phosphorylation of wee1 on Ser549 was lowest during interphase (when wee1 is more active toward cyclin B/cdc2; Lee et al., 2001) and highest during M phase (when it is less active). Wee1 Ser549 phosphorylation was high in CSF extracts and stayed high for 20-30 min after calcium addition, as the extract proceeded through the prolonged exit from metaphase II arrest (Figure 4C). Wee1 Ser549 phosphorylation was greatly reduced as the extract entered interphase of the first mitotic cell cycle, and then it increased again during the first mitosis, remaining high as the extracts (as is typical) arrested in M phase of the first mitotic cycle.

To examine the behavior of wee1 Ser549 phosphorylation during progression through mitosis 1 and into the second cell cycle, cycling extracts were prepared and sampled at 5-min intervals for both immunoblotting and microscopy (Figure 4D). Again, wee1 Ser549 phosphorylation was very low during the first interphase and was highest during mitosis. Wee1 Ser549 phosphorylation spiked sharply during mid-mitosis, just after the peak in cyclin B levels that mark the highest level of cyclin B/cdc2 kinase activity, and just as MAPK activation was beginning. Thus, during the normal, unperturbed early mitotic cycles, phosphorylation of wee1 Ser549 is very low during interphase and peaks during mid-mitosis.

Given that wee1 Ser549 phosphorylation correlates with a 14-3-3-dependent increase in its activity and that wee1 activity is thought to be highest during interphase, these results were unexpected. We thus assayed wee1 kinase activity by the ability of total wee1 immunoprecipitates to phosphorylate Tyr15 of cdc2 in complexes of cyclin B/cdc2(K33R) that had been produced in Sf9 cells. Activity was assayed by two methods that gave comparable results: incorporation of γ-³²P into the whole complex (our unpublished data) and immunoblotting the cdc2 reaction product with a phospho-specific Tyr15 antibody. As seen in Figure 4E, some wee1 kinase activity was detected in M-phase extracts prepared from metaphase II-arrested eggs and extracts that were arrested in mitosis; however, wee1 activity in interphase extract was slightly but consistently higher than M-phase activity levels. These results suggest that, at least during M phase, phosphorylation of wee1 on Ser549 cannot be used as a surrogate measure of its kinase activity.

The Incomplete Replication and DNA Damage Checkpoints Both Lead to an Increase in Wee1 Ser549 Phosphorylation and a Slight Increase in Kinase Activity during Interphase

In vitro, Chk1 can phosphorylate recombinant wee1 on Ser549; this phosphorylation is required for the subsequent binding of 14-3-3, which in turn leads to a modest increase in wee1's in vitro activity (Lee et al., 2001; Rothblum-Oviatt et al., 2001). We thus examined the phosphorylation status of endogenous wee1 Ser549 under checkpoint conditions (Figure 5A). As noted above, wee1 Ser549 phosphorylation in the control extract was low during interphase and rose sharply during mid-mitosis. Activation of the incomplete replication checkpoint or the DNA damage checkpoint at 30 min each led to an increase in wee1 Ser549 phosphorylation ∼50 min later, which was then maintained in these interphase-arrested extracts. The observation that it takes 50-60 min after activation of the DNA damage or incomplete replication checkpoints to see an increase in wee1 Ser549 phosphorylation can be explained by the activation kinetics of Chk1 and Cds1. After the addition of single-stranded DNA to extracts, it can take as long as 60 min for Cds1 modifications to occur (Guo and Dunphy, 2000). Similar kinetics have been seen after UV damage or aphidicolin treatment (Lupardus et al., 2002).

Figure 5.

Activating the incomplete replication or DNA damage checkpoint leads to an increase in wee1 Ser549 phosphorylation. (A) Five hundred sperm nuclei/μl (control), 2000 nuclei/μl, and 0.1 μg/μl aphidicolin (incomplete replication checkpoint), or 500 nuclei/μl and 20 ng/μl linearized plasmid (DNA damage checkpoint) were added to cycling extracts on ice. Extracts were then moved to 20°C to reinitiate cell cycle progression. Samples were taken at indicated times and analyzed as described in Figure 2 legend. (B) Top, samples from metaphase II-arrested extracts (MII; CSF extracts), interphase extracts (Int; CSF + calcium + cycloheximide), or DNA damage checkpoint induced extracts (DNA damage; cycling extracts + 500 nuclei/μl and 25 ng/μl plasmid incubated for 90 min) were analyzed by SDS-PAGE and blotting for wee1 phospho-Ser549. Bottom, Wee1 kinase activity was analyzed as described in Figure 4E legend.

To test the possibility that the appearance of wee1 Ser549 phosphorylation during the checkpoint-induced interphase arrest was the result of interphase arrest, rather then a checkpoint response, cycloheximide was added to cycling extracts at 30 min to block mitotic entry by preventing cyclin synthesis. Wee1 Ser549 phosphorylation did not rise in cycloheximide-arrested interphase extracts (our unpublished data), indicating that it is not simply a response to sustained interphase arrest. Furthermore, the absolute times at which wee1 Ser549 phosphorylation first became evident in the checkpoint-arrested extracts were earlier than the time at which wee1 Ser549 phosphorylation rose during mitosis in the control extract (Figure 5A). Also, wee1 Ser549 phosphorylation was maintained during checkpoint arrest, in contrast to the transient peak that occurs during mid-mitosis. Finally, the level of wee1 Ser549 phosphorylation during checkpoint arrest was slightly lower than that seen during mitosis in cycling extracts. These results argue that the rise in wee1 Ser549 phosphorylation in checkpoint-arrested interphase extracts is a direct consequence of checkpoint activation.

Previous work found that Chk1-mediated, Ser549-dependent 14-3-3 binding to wee1 results in a two- to threefold increase in the activity of recombinant wee1 (Lee et al., 2001). We thus compared the relative changes in Ser549 phosphorylation of endogenous wee1 and its kinase activity after activation of the DNA damage checkpoint. As seen in the example shown in Figure 5B, the large increase in Ser549 phosphorylation was accompanied by a slight increase in wee1 phosphorylation of cdc2 on Tyr15. These results suggest that, at least under the conditions assayed here, there is a slight but not linear correlation between phosphorylation of wee1 on Ser549 and wee1 inhibitory activity toward cdc2 in interphase.

Elevating PKA Activity, Which Blocks Mitotic Entry, Also Induces Phosphorylation of Wee1 on Ser549

PKA activity is required to maintain Xenopus oocytes in their natural arrest at the G₂/meiosis I border (Maller and Krebs, 1977), and cdc25C Ser287 is a physiologically relevant target of PKA in oocytes (Duckworth et al., 2002). PKA has also been implicated in inhibiting the G₂/M transition in mitotic cell cycles. For example, treatments that elevate cAMP levels in somatic cells result in G₂ arrest (reviewed in Fernandez et al. (1995)) and addition of the catalytic subunit of PKA (PKAc) to cycling egg extracts blocks mitotic entry (Grieco et al., 1994). Given that wee1 Ser549 is part of a PKA consensus phosphorylation sequence, we reasoned that this site might be a relevant target in PKA-mediated cell cycle arrest.

We first examined the effect of PKA on cdc25C Ser287 phosphorylation. Addition of PKAc to cycling extracts at 30 min blocked dephosphorylation of cdc25C Ser287. Because cdk activity seems to be required for cdc25C Ser287 dephosphorylation in the embryonic cell cycles (see above), we asked whether high PKA activity maintained cdc25C Ser287 phosphorylation by preventing activation of cdks. PKA was added at 30 min and cyclin ΔB was then added at 50 min. PKA blocked both the ability of cyclin ΔB to drive the extract into mitosis and the dephosphorylation of cdc25C Ser287 (Figure 6A). Thus, just as in extracts arrested in interphase by DNA damage or incomplete replication checkpoints, cdc25C Ser287 phosphorylation was maintained but not increased when extracts were arrested in interphase by elevating PKA.

Figure 6.

PKA-mediated G₂ arrest is accompanied by strong phosphorylation of wee1 Ser549 and maintenance of interphase levels of cdc25C Ser287 phosphorylation. (A) Sperm nuclei were added at a concentration of 500/μl, followed by addition of PKA (cPKA; 1 U/μl; Sigma-Aldrich) or water (volume control) to cycling extracts on ice. Extracts were then moved immediately to 20°C to resume cell cycle progression. MBP-Δ90 cyclin B protein (40 μg/ml) was added to the indicated cycling extracts at the 50-min time point. Samples were taken at indicated times and analyzed as described in Figure 2 legend. (B) Extracts were prepared, and PKA addition was exactly as described in A. Samples were taken at indicated times and analyzed as in described in Figure 2 legend.

The effect of PKA on wee1 Ser549 phosphorylation is shown in Figure 6B. This particular extract was slower than most: at 30 min, it was still in the process of completing exit from meiosis II, MAPK phosphorylation and wee1 Ser549 phosphorylation were still declining, and cdc25C Ser287 phosphorylation was just starting to occur. When PKAc was added at 30 min, dephosphorylation of wee1 Ser549 was halted, and phosphorylation was maintained throughout the interphase arrest. Although we cannot distinguish between direct and indirect effects of PKA during the PKA-induced interphase arrest, these results argue that wee1 Ser549 is a relevant target of PKA during this arrest.

Mitotic Phosphorylation of Wee1 on Ser549

Because wee1 Ser549 phosphorylation closely followed the peak of cyclin B protein levels, and thus cyclin B/cdc2 activity, we asked how keeping cdc2 activity high would affect wee1 Ser549 phosphorylation. Cycling extracts were prepared and cyclin ΔB was added to one portion at 30 min. Cyclin ΔB is able to bind and activate endogenous free cdc2 and the extract progresses into mitosis; although the endogenous full-length cyclin B is degraded, the recombinant nondegradable cyclin ΔB keeps cdc2 active, and the cell cycle arrests in mid-mitosis (Murray et al., 1989, 1996; Luca et al., 1991; Holloway et al., 1993). In the control extract, endogenous cyclin B was highest at 80 min and had disappeared by 90 min (Figure 7A). The addition of cyclin ΔB accelerated mitotic entry (and thus subsequent destruction of endogenous cyclin B) by 10-20 min. The extract then arrested in mitosis, as judged by the continuing activation of MAPK and the maintenance of condensed chromosomes. In the presence of cyclin ΔB, phosphorylation of endogenous wee1 protein on Ser549 rose beyond the level seen during the transient, mid-mitotic level seen in cycling extracts and then was sustained.

Figure 7.

Cyclin B-enforced M-phase arrest maintains high levels of Ser549-phosphorylated wee1. (A) Sperm nuclei were added at a concentration of 500/μl to cycling extracts on ice. Extracts were then moved to 20°C to reinitiate cell cycle progression. MBP-Δ90 cyclin B protein (40 μg/ml) was added at 50 min. Samples were taken at indicated times and analyzed as described in Figure 2 legend. (B) Sperm nuclei were added at a concentration of 500/μl to cycling extracts, followed by addition of 20 mM C1 or C2 to the indicated extracts on ice. Extracts were then moved to 20°C to resume cell cycle progression. Samples were taken at indicated times and analyzed as described in Figure 2 legend.

These results demonstrate that the rise in wee1 Ser549 phosphorylation is not dependent on cyclin B destruction, although they do not rule out the possibility that it does require destruction of endogenous cyclin B. To test that idea, two small molecule inhibitors of cyclin B destruction, C1 and C2 (Verma et al., 2004), were added to cycling extracts at 30 min. The disappearance of endogenous cyclin B was delayed by C1 and blocked by C2. In both cases, mitotic phosphorylation of wee1 Ser549 was elevated and sustained (Figure 7B). This argues that the appearance of wee1 Ser549 phosphorylation is not a consequence of cyclin B destruction, and thus it is not likely to be triggered by the inactivation of cdc2 at the metaphase/anaphase transition. Together, these results suggest that the transient mid-mitotic peak of wee1 Ser549 phosphorylation is a normal response to cdc2 activation.

DISCUSSION

The major new findings of this work are the following. 1) Phosphorylation of endogenous wee1 on Ser549 is low during interphase and increases after activation of either the DNA damage or replication checkpoints. This is accompanied by a slight increase in wee1's kinase activity toward cdc2. 2) In contrast to wee1 Ser549 phosphorylation, the level of inhibitory Ser287 phosphorylation on endogenous cdc25C is already high during interphase and does not show any significant increase in response to checkpoint activation. 3) As with the checkpoint-mediated arrests, blocking the cell cycle in interphase by elevating PKA activity results in increased wee1 Ser549 phosphorylation coupled with maintenance of the high interphase levels of cdc25C Ser287 phosphorylation. 4) During the unperturbed early embryonic cell cycles, phosphorylation of wee1 Ser549 rises sharply during mid-mitosis, just after the peak of cyclin B levels, and then drops rapidly, just as MAPK is becoming activated. 5) Treatments that prevent inactivation of cdc2 during mitotic exit result in further increases in mitotic wee1 Ser549 phosphorylation.

Activation of the DNA Damage or Incomplete Replication Checkpoint Interferes with the Cdc2/Cdc25C-Positive Feedback Loop

Some of the earliest biochemical investigations of cdc25C during the G₂/M transition revealed that cdc2 and cdc25C participate in a positive feedback loop, resulting in the rapid activation of cyclin B/cdc2 (Izumi et al., 1992; Hoffmann et al., 1993; Izumi and Maller, 1993; Strausfeld et al., 1994). Although it is clear how cdc25C activates cdc2, what regulates activation of cdc25C during normal cell cycle progression is less well understood. It is now known from work with recombinant cdc25C that cdk2-catalyzed phosphorylation of cdc25C on Thr138 is required for the removal of 14-3-3 from phosphorylated Ser287 of cdc25C (Margolis et al., 2003). However, cdk2 activity alone is not sufficient for 14-3-3 removal and subsequent Ser287 dephosphorylation. In human cells, phosphorylation of a nearby site (Ser214) during mitosis has been reported to prevent rephosphorylation of the 14-3-3 binding site, Ser216 (Bulavin et al., 2003b). Our data show that cdc2 activity is required for the full dephosphorylation of Ser287 during the G₂/M transition. A reflection of this dependency is evident on immunoblots of samples in which favorably spaced time points were taken across the transition from G₂ into mitosis. In such cases, Ser287 phosphorylation is still seen on some of the most electrophoretically retarded forms of cdc25C (Figures 1C and 7B), which arise as a consequence of phosphorylation by cdc2 and other mitotic kinases (Izumi et al., 1992; Kumagai and Dunphy, 1992; Hoffmann et al., 1993; Izumi and Maller, 1993, 1995). Moreover, as shown here, cdc25C Ser287 phosphorylation is maintained during interphase arrest when cdc2 activation is blocked by inhibiting protein synthesis, and subsequent activation of cdc2 by addition of cyclin B drives both cdc25C Ser287 dephosphorylation and entry into mitosis. Thus, although a small amount of Ser287 may either be unphosphorylated or become dephosphorylated before cdc2 activation, full dephosphorylation of Ser287 is a consequence of and follows cdc2 activation, where it functions as part of the positive feedback loop that helps to fully activate both cdc25C and cdc2.

Although changes in cdc25C Ser287 phosphorylation are clearly important for this positive feedback loop during normal cell cycle progression, we found no significant increase in the level of this inhibitory phosphorylation in response to either the DNA damage or replication checkpoints. This finding confirms and extends two earlier observations. First, no increase in the overall level of ³²P incorporation into cdc25C was seen after blocking DNA synthesis (Izumi et al., 1992). Second, a very large fraction of endogenous cdc25C is already bound to 14-3-3 proteins during interphase, and this fraction does not increase upon checkpoint activation (Kumagai et al., 1998b). Activation of the DNA damage or incomplete replication checkpoint during interphase could lead to a switch in the kinase targeting cdc25C Ser287 from an interphase Ser287 kinase such as CaMKII or PKA (Duckworth et al., 2002; Hutchins et al., 2003), to one of the checkpoint kinases. At the same time, the checkpoint-induced increase in wee1 Ser549 phosphorylation and concomitant elevation in its activity due to 14-3-3 binding would both suppress any cdc2 activity that might arise from the increasing concentration of cyclin B-cdc2 complexes that follow the accumulation of newly made cyclin B, and interfere with the positive feedback loop in which a very small amount of active cdc2 can trigger activation of cdc25C through inducing removal of 14-3-3 proteins and allowing Ser287 dephosphorylation to occur. Thus, the pool of wee1 phosphorylated on Ser549 may serve as a critical gate-keeper during checkpoint activation. In further support of this idea are earlier findings that depletion of wee1 allows extracts to enter mitosis in the presence of activated Chk1 and that readdition of wee1, but not the nonphosphorylatable wee1 Ser549Ala mutant, restores the block to mitotic entry (Lee et al., 2001).

Although it seems that wee1 is significantly affected upon checkpoint activation in this system, the relative contributions of cdc25C and wee1 in blocking mitotic entry during checkpoint arrests may be different in somatic cell cycles. For example, in the one study that examined the phosphorylation of endogenous human cdc25C in mammalian somatic cells, DNA damage did in fact lead to an increase in phosphorylation of Ser216 (Zhou et al., 2000).

The inhibitory effect of PKA on the G₂/M transition of the early embryonic cell cycle may also be explained by its effects on wee1. It has long been known that PKA activity is required to maintain oocytes of all species tested in their natural G₂ arrest and that a drop in PKA activity is required for the G₂/meiosis I transition (O'Connor and Smith, 1976; Maller and Krebs, 1977; Speaker and Butcher, 1977; Maller et al., 1979; Mulner et al., 1979; Finidori-Lepicard et al., 1981; Huchon et al., 1981; Sadler and Maller, 1981; Masui, 1985). Phosphorylation of cdc25C Ser287 by PKA has recently been shown to be required to maintain the G₂ arrest (Duckworth et al., 2002; Schmitt and Nebreda, 2002). In the early embryonic cell cycles, PKA activity has been reported to oscillate, being high in mitosis and low in interphase, and high PKA activity blocks entry into mitosis (Grieco et al., 1994, 1996). Like cdc25C Ser287, wee1 Ser549 is a PKA consensus site and as reported here, PKAc addition during interphase leads to a strong increase in wee1 Ser549 phosphorylation and maintenance of cdc25C Ser287 phosphorylation. Thus, it seems very likely that, just as for checkpoint arrests, elevated PKA activity arrests cells in interphase, at least in part by blocking the cdc2-cdc25C feedback loop through direct or indirect phosphorylation of wee1 Ser549. Although endogenous PKA may or may not be a physiologically relevant regulator of the G₂/M transition in the normal cell cycle, the fact that PKA can induce Ser549 phosphorylation suggests that phosphorylation of wee1 on Ser549 may be a common target of mitotic inhibitory kinases.

Regulation of Wee1 at the Level of Proteolysis?

In budding yeast, the wee1 homologue Swe1 is degraded during the G₂/M transition, and stabilizing the protein inhibits mitotic entry (Sia et al., 1998). In mammalian somatic cells, the fate of wee1 is less clear: some studies found that wee1 protein was reduced during mitosis (Baldin and Ducommun, 1995; Watanabe et al., 1995), whereas others found that it was stable (McGowan and Russell, 1995; Parker et al., 1995). In Xenopus egg extracts, immunoblots of endogenous wee1 show that overall wee1 protein levels are essentially constant across the early embryonic cell cycles (Walter et al., 1997), a point that we confirm here. However, two highly cited reports concluded that wee1 is degraded in the early embryonic cell cycles (Michael and Newport, 1998; Ayad et al., 2003). Those studies examined the fate of a radiolabeled in vitro reticulocyte translation product that had been added to egg extracts. Although this approach has been widely used, it should be noted that in vitro translation products do not always mirror the fate of the endogenous proteins, perhaps because the reticulocyte mix provides additional components, or components that are differently modified, from those present in the egg extracts (Crane et al., 2004). Thus, wee1 protein levels remain essentially constant across the early cell division cycles.

Mitotic Phosphorylation of Wee1 Ser549

During the normal cell cycle, wee1 Ser549 phosphorylation is low during interphase and the first part of the first mitosis, and then it spikes sharply after the peak in the amount of cyclin B protein. Wee1 Ser549 is also strongly phosphorylated in two other M-phase settings: in eggs that are naturally arrested at metaphase of meiosis II and under experimental conditions where cyclin B destruction is blocked. It thus seems possible that the transient mid-mitotic peak of wee1 Ser549 phosphorylation is a normal, time-delayed response to cdc2 activation, as is the cdc2-dependent activation of the E3 ubiquitin ligase APC/C (Lahav-Baratz et al., 1995) that is responsible for destruction of cyclin B and other targets during mitotic exit.

Although we cannot rule out transient mitotic activation of a subpopulation of wee1 toward cdc2, several considerations argue against this. First, the Ser549 phosphorylation-linked enhancement of wee1 activity during interphase depends on its binding to 14-3-3 (Lee et al., 2001; Rothblum-Oviatt et al., 2001). However, during mitosis, the majority of wee1 does not seem to associate with 14-3-3, and as shown here and elsewhere, overall wee1 activity toward cdc2 is lower in mitosis than in interphase (Mueller et al., 1995; Lee et al., 2001; Rothblum-Oviatt et al., 2001), despite strong Ser549 phosphorylation. Finally, during mitosis, wee1 is phosphorylated at multiple sites by kinases including cdc2, which is known to inhibit the activity of recombinant wee1 in vitro (Solomon et al., 1990; Devault et al., 1992; Smythe and Newport, 1992; Tang et al., 1993; Honda et al., 1995; Mueller et al., 1995).

What is the role of mitotic wee1 Ser549 phosphorylation? One possibility is that Ser549 phosphorylation regulates the activity of wee1 toward other substrates yet to be identified. For example, it could regulate the activity of wee1 toward a target involved in mitotic exit. Wee1 Ser549 phosphorylation could also be involved in wee1's recently discovered role in apoptotic signaling, a pathway that does not seem to involve the activity of wee1 toward cdc2 (Smith et al., 2000). Alternatively, it could regulate wee1 localization rather than its kinase activity during mitosis. For example, immunofluorescence studies in mammalian somatic cells show that wee1 concentrates at midbody in late mitosis, and on either side of the cleavage furrow during cytokinesis in a microtubule-dependent manner (Baldin and Ducommun, 1995). These are among the many questions to address next.