INTRODUCTION

Cell division cycle phases are controlled and coordinated by a family of protein kinases, the cyclin-dependent kinases (cdks),¹ in complex with a family of regulatory subunits, the cyclins (reviewed by Meijer et al., 1995; Morgan, 1995; Nigg, 1995; Pines, 1995). Among these kinases, p34^cdc2 plays an important role in cell cycle progression. Genetic studies in the yeast Schizosaccharomyces pombe demonstrate that the cdc2 gene is responsible for both G₁-S and G₂-M transitions (Nurse and Bissett, 1981; reviewed by Nurse, 1990). Its homologue in Saccharomyces cerevisiae, the CDC28 gene, is also required for ``start'' and mitosis (Hartwell et al., 1974; Beach et al., 1982). The cdc2 gene encodes a 34-kDa protein present in all eukaryotic cells investigated thus far. This protein is the catalytic subunit of the M phase-promoting factor (MPF), the universal intracellular factor responsible for entry into M phase. The regulatory subunit of p34^cdc2 is cyclin B, a 47-52-kDa protein encoded by the cdc13 gene (Booher et al., 1989; Labbé et al., 1989; Meijer et al. 1989). Three types of mechanisms are responsible for activation of MPF at the onset of mitosis: binding of p34^cdc2 to cyclin B, a series of phosphorylation/dephosphorylation events on these two subunits (see, for example, Pondaven et al., 1990; Meijer et al., 1989, 1991), and translocation of the complex to the nucleus (Ookata et al., 1992) or to microtubules (Ookata et al., 1995).

Activation of p34^cdc2 kinase is associated with modifications of the phosphorylation of three residues: Thr-14, Tyr-15, and Thr-161. Phosphorylation of the Thr-161 residue of p34^cdc2 is necessary for activation (Ducommun et al., 1991; Gould et al., 1991). The Thr-161 kinase, also called cdc2-activating kinase, has been identified as a complex between cdk7 (MO15) and cyclin H (Fisher and Morgan, 1994; reviewed by Solomon, 1994; Schuttleworth, 1995). The Thr-14 and Tyr-15 residues also play a critical role in p34^cdc2 activation (Krek and Nigg, 1991; Norbury et al., 1991; Pickham et al., 1992; Atherton-Fessler et al., 1994). These residues are localized in the ATP-binding site of the kinase (De Bondt et al., 1993; Endicott et al., 1994; Jeffrey et al., 1995; Schulze-Gahmen et al., 1995). Following cyclin B binding (Solomon et al., 1990; Meijer et al., 1991), p34^cdc2 is phosphorylated on these two residues, leading to an inactive complex, pre-MPF (Gautier and Maller, 1991). In yeast, only Tyr-15 is phosphorylated in G₂ (Gould and Nurse, 1989). Tyr-15 phosphorylation is carried out primarily by the wee1 and mik1 kinases (Lee et al., 1991; Lungren et al., 1991), whereas Thr-14 is phosphorylated by a dual specificity, membrane-bound kinase encoded by Myt1 (Atherton-Fessler, 1994; Kornbluth et al., 1994; Mueller et al., 1995b). Neither wee1 (Honda et al., 1992; Parker and Piwnica-Worms, 1992; McGowan and Russell, 1993) nor mik1 (Lee et al., 1994) phosphorylate the Thr-14 residue.

Dephosphorylation of the Tyr-15 residue of p34^cdc2 at the G₂-M transition is carried out by the p80^cdc25 dual-specificity phosphatase initially identified in S. pombe (Russell and Nurse, 1986; Millar and Russell, 1992). The Tyr phosphatase encoded by the pyp3 gene, in fission yeast, is also able to dephosphorylate Tyr-15 (Millar et al., 1992). The pyp3 and cdc25 phosphatases may thus act cooperatively to activate p34^cdc2 at the G₂-M transition. In higher eukaryotes, where both Thr-14 and Tyr-15 are phosphorylated, p80^cdc25 may dephosphorylate both residues (Gautier and Maller, 1991; Kumagai and Dunphy, 1991; Lee et al., 1992; Honda et al., 1993; reviewed by Millar and Russell, 1992; Hoffmann et al., 1994). In humans, three cdc25 genes (A, B, and C) have been cloned (Sadhu et al., 1990; Galaktionov and Beach, 1991; Nagata et al., 1991). cdc25 A is more expressed in G₁ (Hoffmann et al., 1994), in contrast to cdc25 B (Honda et al., 1993) and cdc25 C (Hoffmann et al., 1993), which are more expressed in G₂-M. The cdc25 C phosphatase is activated by phosphorylation on Ser and Thr residues (Kumagai and Dunphy, 1992; Izumi et al., 1992; Kuang et al., 1994) catalyzed by cdc2/cyclin B (Hoffmann et al., 1993; Izumi and Maller, 1993; Strausfeld et al., 1994a) and other non-cdc2 kinases (Ogg et al., 1994; Izumi and Maller, 1995). The phosphorylation of cdc25 by cdc2/cyclin B has been proposed as a mechanism explaining the ``autocatalytic amplification'' of MPF, the property of a small amount of MPF to activate a large amount of MPF (Hoffmann et al., 1993; Strausfeld et al., 1994b). Phosphorylation/inactivation of wee1 may also contribute to MPF amplification (Tang et al., 1993; Honda et al., 1995; McGowan and Russell, 1995; Mueller et al., 1995a; Parker et al., 1995; Watanabe et al., 1995).

We have studied here the regulation of p34^cdc2 activity by dephosphorylation of its Thr-14 and Tyr-15 residues at the prophase/metaphase transition, using the highly synchronized starfish oocyte model (reviewed by Meijer and Mordret, 1994). We first set up a method for identification of the four possible phosphorylation states of p34^cdc2. Two intermediate states of phosphorylation of p34^cdc2 (phosphorylated only on Thr-14 or only on Tyr-15) are obtained in vitro after treatment with a GST-pyp3 phosphatase or the Ser/Thr phosphatase 2A, respectively. We demonstrate that both in vivo and in vitro (cdc25) dephosphorylations of p34^cdc2 occur in two steps; Thr-14 dephosphorylation is followed by Tyr-15 dephosphorylation. The intermediate form of p34^cdc2 (dephosphorylated on Thr-14, phosphorylated on Tyr-15) displays significant kinase activity, showing that the Thr-14 residue carries more powerful inhibitory potential than the Tyr-15 residue. Furthermore, this transient intermediate form may participate to the cdc2 amplification loop through phosphorylation of cdc25.

EXPERIMENTAL PROCEDURES

Chemicals and Reagents

Sodium orthovanadate, 1-methyladenine (1-MeAde), EGTA, EDTA, MOPS, -glycerophosphate, dithiothreitol, sodium fluoride, p-nitrophenylphosphate, leupeptin, aprotinin, soybean trypsin inhibitor, benzamidine, vitamin K3, LB broth base, ampicillin, isopropyl-1-thio--D-galactopyranoside, glutathione-agarose beads, glutathione, Tween 20, and LB broth base were obtained from Sigma. Protein A-Sepharose beads CL-4B were purchased from Pharmacia Biotech Inc. and Nonidet P-40 from Fluka. Bovine serum albumin (fraction V) was obtained from Boehringer Mannheim. Horseradish peroxidase-coupled secondary antibodies (anti-mouse and anti-rabbit), [-³²P]ATP (PB168; 3000 Ci/mmol; 1 mCi/ml), ACS scintillation fluid, hyperfilm MP and -max, ECL detection reagents, were purchased from Amersham Life Science. Calyculin A was kindly provided by Dr. H. Tosuji (Funabashi).

Purified protein phosphatase 2A1 (PP2A) was generously donated by Dr. R. W. MacKintosh (Dundee). The Escherichia coli strains expressing GST-cdc25 A and GST-pyp3 were provided by Drs. K. Galaktionov and D. Beach (Cold Spring Harbor Laboratory) and Drs. J. Millar and P. Russell (La Jolla), respectively. PP2A and its Tyr phosphatase activator were kindly provided by Dr. J. Goris (Leuven).

Monoclonal anti-PSTAIRE antibodies (raised against the NH₂-EGVSLLKEGGC-COOH peptide), polyclonal anti-GEGTYG antibodies (raised against the NH₂-VEKIVVYKARHKLS-COOH peptide), polyclonal anti-Tyr(P) antibodies and polyclonal anti-cyclin B^cdc13 (starfish) antibodies were generously donated by Drs. M. Yamashita (Sapporo), H. Y. L. Tung (Austin), J. Y. J. Wang (La Jolla), and T. Kishimoto (Tokyo), respectively.

Buffers

The following buffers were used. Calcium-free artificial sea water contained 452.2 mM NaCl, 10.08 mM KCl, 29.8 mM MgCl₂ (6H₂O), 17.2 mM MgSO₄ (7H₂O), and 5 mM Tris-HCl, pH 8 (Shapiro, 1941). Homogenization buffer contained 60 mM -glycerophosphate, 15 mM p-nitrophenylphosphate, 25 mM MOPS, pH 7.2, 15 mM EGTA, 15 mM MgCl₂, 2 mM DTT, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 1 mM disodium phenylphosphate, 10 µg/ml leupeptin, 10 µg/ml aprotinin, 10 µg/ml soybean trypsin inhibitor, and 100 mM benzamidine. Bead buffer contained 50 mM Tris-HCl, pH 7.4, 5 mM NaF, 250 mM NaCl, 5 mM EDTA, 0.1% Nonidet P-40, 5 mM EGTA, 10 µg/ml leupeptin, 10 µg/ml aprotinin, 10 µg/ml soybean trypsin inhibitor, and 100 mM benzamidine. Phosphate-buffered saline, pH 7.2-7.4, contained 140 mM NaCl, 2.7 mM KCl, 1.5 mM KH₂PO₄, and 8.1 mM Na₂HPO₄. Lysis buffer contained 1% Nonidet P-40, 1 mM EDTA, 1 mM DTT, 10 µg/ml leupeptin, 10 µg/ml aprotinin, 10 µg/ml soybean trypsin inhibitor, and 100 mM benzamidine in phosphate-buffered saline. Tris buffer A contained 50 mM Tris, pH 8, 50 mM NaCl, 1 mM EDTA, and 1 mM DTT. Tris buffer B contained 50 mM Tris, pH 8, 50 mM NaCl, 1 mM EDTA, and 20 mM DTT. Elution buffer contained Tris buffer B with 20 mM glutathione. Buffer C contained 60 mM -glycerophosphate, 30 mM p-nitrophenylphosphate, 25 mM MOPS, pH 7.0, 5 mM EGTA, 15 mM MgCl₂, 1 mM DTT, and 0.1 mM sodium orthovanadate. Transfer buffer contained 39 mM glycine, 48 mM Tris, 0.037% SDS, and 20% methanol. Tris buffered saline Tween 20 (TBST) contained 50 mM Tris, pH 7.4, 150 mM NaCl, and 0.1% Tween 20.

Preparation of Gametes

Asterias rubens and Marthasterias glacialis were collected in Northern Brittany and kept under running sea water until use. The gonads were dissected out of the starfish and gently torn open in ice-cold calcium-free artificial sea water. Oocytes were then filtered through cheese cloth, and washed four times in calcium-free artificial sea water to remove the 1-MeAde-producing follicle cells. They were resuspended in the same medium as a 10% (v/v) suspension. Oocyte maturation was triggered by addition of 1-MeAde to a final concentration of 1 µM. During maturation time-course experiments, 1-ml aliquots of the oocyte suspension were withdrawn at regular intervals after hormonal stimulation and centrifuged in microtubes; then the oocyte pellets were frozen in liquid nitrogen.

Vitamin K3 is a powerful inhibitor of the cdc25 phosphatase. It inhibits hormone-induced oocyte maturation (Kerns et al., 1995).² Oocytes were treated with vitamin K3 (final concentration, 0-250 µM) for 15 min prior to 1-MeAde addition. After a 20-min incubation, aliquots were centrifuged, and oocyte pellets were frozen in liquid nitrogen. The rate of nuclear envelope breakdown (germinal vesicle breakdown) was recorded by light microscopy examination.

Preparation of G₂ and M-phase Oocytes

One-ml aliquots of an oocyte suspension before (G₂ phase oocytes) or 20 min after 1-MeAde addition (M-phase oocytes) were rapidly centrifuged, the supernatant was removed, and the oocyte pellets were frozen in liquid nitrogen.

Purification of p34^cdc2 on p9^CKShs1-Sepharose Beads

p34^cdc2 was purified by affinity chromatography on p9^CKShs1-Sepharose beads, prepared as described in Azzi et al. (1994). Four hundred µl of homogenization buffer were added per 100 µl of G₂ or M-phase oocyte pellets. After sonication, extracts were centrifuged at 14,000 × g for 10 min at 4 °C. The supernatant was then incubated at 4 °C for 30 min and under constant rotation with 10 µl of p9^CKShs1-Sepharose beads in the presence of 400 µl of bead buffer. After removal of the supernatant, the beads were washed three times with ice-cold bead buffer, and the bound proteins were recovered with 50 µl of 2 × Laemmli sample buffer prior to analysis by Western blotting.

Preparation and Purification of GST-pyp3 and GST-cdc25 Fusion Proteins

E. coli strains were transformed by plasmids encoding the gene fusion constructs of GST and yeast pyp3 or human cdc25 A (or cdc25 C). Bacteria were first grown overnight at 37 °C in the presence of 100 µg/ml ampicillin in LB medium. Four ml of this preculture were inoculated per liter of LB containing 100 µg/ml ampicillin. Incubation was continued at 30 °C until the culture absorbance at 600 nm reached 0.8-1. At this time, 0.4 mM isopropyl-1-thio--D-galactopyranoside was added, and the culture was incubated at 25 °C for at least 5 h (GST-pyp3) and 7 h (GST-cdc25 A and GST-cdc25 C). Cells were then harvested by a 3000 × g centrifugation for 15 min at 4 °C. Pellets were kept frozen at 80 °C until extraction.

The bacterial pellet was homogenized by sonication in lysis buffer at 4 °C. The homogenate was centrifuged at 100,000 × g for 30 min at 4 °C, and the supernatant was stored in 10-ml aliquots at 80 °C. Fusion proteins were purified by affinity chromatography on glutathione-agarose beads. Ten ml of bacterial extract were incubated with 400 µl of glutathione-beads (equilibrated in lysis buffer) for 30 min at 4 °C under constant rotation. The beads were washed four times with 10 ml of lysis buffer, followed by four washes with 10 ml of Tris buffer B. Fusion proteins were eluted by incubation with 4 × 1 ml of 20 mM glutathione in Tris buffer B (elution buffer). Efficiency of the elution was monitored by a phosphatase assay and by SDS-PAGE. The glutathione-beads were recycled by a wash with 1 M NaCl, followed by equilibration in lysis buffer.

Assays were performed in microtitration plates as described by Baratte et al. (1992). Twenty µl of GST-phosphatase were added to 20 µl of 100 mM DTT in Tris buffer A and 140 µl of Tris buffer A. Plates were preincubated at 37 °C for 15 min in a Denley Wellwarm 1 microplate incubator. Reactions were initiated by the addition of 20 µl of 500 mM p-nitrophenylphosphate in Tris buffer A. After a 30-min incubation at 37 °C, absorbance at 405 nm was measured in a Bio-Rad microplate reader.

In Vitro Dephosphorylation of p34^cdc2 by Purified Phosphatases

P9^CKShs1-Sepharose beads, loaded with G₂ oocyte extracts, were prepared as described above. Following the bead buffer step, the beads were washed three times with Tris buffer A prior to incubation for 30 min at 30 °C with 100 µl of recombinant phosphatases or 2 µl of purified PP2A (in a final volume of 100 µl). The dephosphorylation reaction was stopped by the addition of 1 ml of bead buffer. The beads were washed three times with bead buffer before kinase assays or addition of 50 µl of 2 × Laemmli sample buffer. The phosphorylation status of p34^cdc2 was then analyzed by SDS-PAGE and Western blotting with appropriate antibodies.

Assay of p34^cdc2 Kinase Activity

The kinase activity of p34^cdc2 was measured after its purification on p9^CKShs1-Sepharose beads and various phosphatase treatments. Assays were performed by incubation of 10 µl of packed p9^CKShs1-Sepharose beads for 5 min at 30 °C with 15 µl of buffer C, 5 µl of histone H1 (5 mg/ml), and 10 µl of 45 µM [-³²P]ATP. Assays were terminated by transferring the tubes into ice. After a brief centrifugation, 25 µl of supernatant were spotted on 2.5 × 3-cm pieces of Whatman p81 phosphocellulose paper. Filters were washed five times in 1% phosphoric acid, dried, and transferred in plastic scintillation vials with 1 ml of ACS (Amersham) scintillation fluid. [³²P]Phosphate incorporation in the histone H1 substrate was measured in a Packard counter. Fifty µl of 2 × Laemmli sample buffer were added to the remaining beads and supernatant prior to SDS-PAGE and analysis by autoradiography.

Immunoprecipitation of Different Forms of p34^cdc2 with Anti-cyclin B Antibodies

p34^cdc2 bound to p9^CKShs1-Sepharose beads was eluted by incubation with 500 µl of free p9^CKShs1 (2 mg/ml) for 30 min at 4 °C. The elution products were incubated with 10 µl of anti-cyclin B antibodies for 60 min on ice. Eighty µl of protein A-Sepharose beads (50% (v/v) suspension) were then added, and the mixture was rotated for 30 min at 4 °C. The protein A-Sepharose beads were washed three times with bead buffer before the addition of 50 µl of 2 × Laemmli sample buffer. The immunoprecipitated proteins were analyzed by SDS-PAGE and Western blotting.

Electrophoresis and Western Blotting

Proteins bound to p9^CKShs1-Sepharose beads were recovered with 2 × Laemmli sample buffer. Samples were run in 10% SDS-polyacrylamide gels. For detection of ³²P-labeled proteins, gels were stained with Coomassie Blue and exposed overnight to -max film. For Western blotting, proteins were transferred from the gel to a 0.1 µm nitrocellulose sheet (Schleicher and Schuell) in a milliblot-SDE system (Millipore) for 30 min at 2.5 mA/cm² in transfer buffer. Subsequently, the filter was blocked with 5% low fat milk in TBST for 1 h. The filter was then washed with TBST and incubated for 1 h with the first antibodies (anti-PSTAIRE, 1:2000; anti-Tyr(P), 1:1000; anti-GEGTYG, 1:1000). After four washes (1 × 20 min and 3 × 5 min) with TBST, the nitrocellulose sheet was treated for 1 h with horseradish peroxidase-coupled secondary antibodies diluted in TBST (1:1000). The filter was then washed five times (1 × 20 min and 4 × 5 min) with TBST and analyzed by enhanced chemiluminescence with ECL detection reagents and hyperfilm MP.

RESULTS

In this study, we have taken advantage of the excellent synchrony of the starfish oocyte model. Starfish oocytes are naturally arrested in late prophase of the first meiotic division (frequently and incorrectly referred to as ``G₂ arrest''). A follicle cell-derived hormone, 1-MeAde, triggers rapid, protein synthesis-independent entry into meiotic divisions. Nuclear envelope breakdown occurs within 20 min after hormone stimulation. p34^cdc2/cyclin B is present as an inactive complex in the prophase-arrested oocyte. It is activated within 5 min after 1-MeAde addition (reviewed by Meijer and Mordret, 1994).

The cdc2/cyclin B complex can be easily purified by affinity chromatography on p9^CKShs1-Sepharose beads (Pondaven et al., 1990). Most of the work presented below relies on the selective binding of p34^cdc2/cyclin B kinase to p9^CKShs1-Sepharose beads. Although p9^CKShs1-Sepharose beads bind cdc2, cdk2, and cdk3 in mammalian cell extracts (Draetta et al., 1987; Elledge et al., 1992), the case of starfish oocytes differs in two ways: (a) all cells are arrested in late prophase, and this extreme and natural synchrony eliminates potential contaminations by cdks from other cell cycle phases; and (b) these cells accumulate an unusually large amount of p34^cdc2/cyclin B kinase (Pondaven et al., 1990). Immunoprecipitations with anti-cyclin B antibodies eliminate virtually all kinase activity bound to p9^CKShs1-Sepharose, suggesting that the p9^CKShs1-bound PSTAIRE signal and the associated histone H1 kinase activity essentially correspond to p34^cdc2 bound to cyclin B (Azzi et al., 1994).²

Two properties of p34^cdc2 vary according to its phosphorylation state: (a) its electrophoretic mobility; and (b) its cross-reactivity with various antibodies. On SDS-PAGE (Fig. 1), the most phosphorylated form of p34^cdc2 (phosphorylated on Thr-14 and Tyr-15, T^P-Y^P) migrates slowly (``upper form''). The most dephosphorylated form of p34<sup>cdc2</sup> (dephosphorylated on Thr-14 and Tyr-15, T-Y) migrates rapidly (``lower form'') (Fig. 1). We have used three antibodies to recognize the different phosphorylation states of p34^cdc2. The first group, monoclonal antibodies (anti-PSTAIRE), are directed against a conserved sequence in the cdc2 gene family involved in cyclin binding (Yamashita et al., 1991; Jeffrey et al., 1995). These antibodies recognize both upper (T^P-Y^P) and lower (T-Y) forms of p34^cdc2 (Fig. 1A). The second group, polyclonal antibodies, directed against phospho-Tyr only, recognize the upper form of p34^cdc2 (T^P-Y^P) (Fig. 1B). The third group, polyclonal antibodies (anti-GEGTYG), are directed against a conserved sequence involved in ATP binding and comprising the inhibitory residues Thr-14 and Tyr-15. These antibodies only recognize the lower form of p34^cdc2 (T-Y) (Fig. 1C).

To generate intermediately phosphorylated forms of p34^cdc2, the kinase from prophase oocytes (T^P-Y^P) was first purified on p9^CKShs1-Sepharose beads. The p34^cdc2/cyclin B kinase preparation was then split into equal aliquots, which were each individually treated with a specific phosphatase or left untreated. Each intermediate form thus derives from the initial p34^cdc2 form and is not a form present prior to the phosphatase treatment. The treated kinase was then resolved by SDS-PAGE, and its phosphorylation state was analyzed with the three antibodies (Fig. 2).

We first treated cdc2 with a recombinant Tyr phosphatase, GST-pyp3. In S. pombe, pyp3 activates cdc2 by dephosphorylating Tyr-15 (Millar et al., 1992). On the anti-PSTAIRE immunoblot (Fig. 2A), p34^cdc2 migrates in an intermediate position between the upper and and lower forms, suggesting an intermediate phosphorylation state. Following the action of GST-pyp3, p34^cdc2 is dephosphorylated on Tyr-15, since no signal is detected with the anti-Tyr(P) antibodies (Fig. 2B). Thr-14 is still phosphorylated, as suggested by the lack of cross-reactivity with the anti-GEGTYG antibodies (Fig. 2C). Therefore, GST-pyp3 treatment generates a Thr-14 phosphorylated, Tyr-15 dephosphorylated (T^P-Y) form of p34^cdc2 visualized with anti-PSTAIRE antibodies as an intermediate form.

Treatment of the cdc2 kinase with the PP2A phosphatase also leads to an intermediate form of p34^cdc2 recognized by anti-PSTAIRE antibodies (Fig. 2A). This intermediate form cross-reacts with anti-Tyr(P) antibodies (Fig. 2B) and also anti-GEGTYG antibodies (Fig. 2C). Therefore, PP2A treatment generates a Thr-14 dephosphorylated, Tyr-15 phosphorylated (T-Y^P) form of p34^cdc2 visualized with anti-PSTAIRE antibodies as an intermediate form. In the presence of 4 µM calyculin A, an inhibitor of phosphatases 1 and 2A, PP2A is unable to modify p34^cdc2 (data not shown). This demonstrates the specificity of PP2A and also that the intermediate form does not originate from the PP2A preparation. PP2A has been reported to inhibit cdc2 activation (Lee et al., 1991). This is not due to dephosphorylation of the Thr-161 residue, since treatment with both PP2A and pyp3 result in full activation of the kinase. Lee et al. (1994) also reported the lack of effect of PP2A on Thr-161 dephosphorylation.

The T-Y^P form, obtained by PP2A treatment, can be further dephosphorylated on Tyr-15 by GST-pyp3, leading to a completely dephosphorylated (T-Y) form similar to p34^cdc2 obtained from M-phase oocytes. This form cross-reacts with anti-PSTAIRE (Fig. 2A) and anti-GEGTYG (Fig. 2C) but not anti-Tyr(P) antibodies (Fig. 2B). Immunoprecipitation experiments with anti-cyclin B antibodies confirm that the intermediate forms (T^P-Y and T-Y^P) are indeed cyclin B-bound forms of cdc2 (Fig. 3).

We next used the GST-cdc25 A phosphatase to dephosphorylate ``prophase p34^cdc2'' (T^P-Y^P). The recombinant phosphatase generates the totally dephosphorylated form of p34^cdc2 (T-Y), recognized by anti-PSTAIRE and anti-GEGTYG and comigrating with the M-phase oocyte form of p34^cdc2 (Baratte et al., 1992). The dephosphorylation time course shows the transient appearance of an intermediate form detected with anti-PSTAIRE (Fig. 4A) and anti-GEGTYG (Fig. 4C) antibodies. This form did not appear on the anti-Tyr(P) immunoblot (Fig. 4B), but long exposure times allow the detection of the intermediate signal (data not shown). It is, therefore, the Thr-14 dephosphorylated form of p34^cdc2 (T-Y^P). The same results were obtained with GST-cdc25 C. The cdc25 phosphatase thus dephosphorylates cdc2, in vitro, in two steps: first Thr-14, then Tyr-15, generating a transient singly phosphorylated form, T-Y^P.

We next investigated the physiological significance of these intermediate, partially phosphorylated forms of p34^cdc2. Oocyte samples were frozen at regular intervals throughout the prophase/metaphase transition induced by 1-MeAde. The phosphorylation state of p34^cdc2 was analyzed after purification of the cdc2 kinase on p9^CKShs1-Sepharose beads, SDS-PAGE, and Western blotting (Fig. 5). The anti-PSTAIRE immunoblot shows that p34^cdc2 is gradually modified from a totally phosphorylated form (T^P-Y^P, prophase) to a completely dephosphorylated form (T-Y, metaphase) (Fig. 5A). Tyr-15 is dephosphorylated during the first 9 min of maturation, as shown with anti-Tyr(P) antibodies (Fig. 5B). Anti-GEGTYG antibodies monitor the appearance of the fully dephosphorylated form of p34^cdc2 (Fig. 5C). In addition, they show the transient appearance, during the first 6 min, of an intermediate form of cdc2. This short-lived form of cdc2 cross-reacting with anti-GEGTYG is similar to the T-Y^P form obtained after treatment of ``prophase p34^cdc2'' with PP2A (Fig. 2). However, in this experiment, it was not detected with anti-Tyr(P) antibodies (Fig. 5B), in contrast to the PP2A treatment. The low abundance of this form in the time course could explain this. The cdc2 intermediate form was observed in the two starfish species used in this study, M. glacialis and A. rubens. To confirm the phosphorylation state of this transient intermediate form, we next treated p34^cdc2 purified from oocytes taken 3 min after 1-MeAde stimulation with GST-pyp3 (Fig. 6). The anti-GEGTYG immunoblot showed an electrophoretic shift of the intermediate band to a lower position. This down shift confirms that the intermediate band is the T-Y^P form of p34^cdc2.

Vitamin K3 is a good inhibitor of cdc25 phosphatases (Fig. 7A) (Kerns et al., 1995). It inhibits 1-MeAde-induced maturation and the associated dephosphorylation of p34^cdc2 in a dose-dependent manner (Fig. 7B). At an intermediate vitamin K3 concentration (50 µM), maturation is partially inhibited, and cdc2 dephosphorylation is partially blocked. Anti-PSTAIRE immunoblots show the presence of an intermediate form (Fig. 7C). This form is not detected with anti-Tyr(P) antibodies (Fig. 7D) but is easily detected with anti-GEGTYG antibodies (Fig. 7E) and is, therefore, the T-Y^P form of cdc2.

Taken together, these results indicate that, during the prophase/metaphase transition of starfish oocyte maturation, activation of p34^cdc2 occurs in two steps: first Thr-14 dephosphorylation, followed by Tyr-15 dephosphorylation. A short-lived T-Y^P form of p34^cdc2 transiently appears during the exit from prophase.

Kinase activities of the various p34^cdc2 forms were measured and are displayed as direct counts (Fig. 8A) and as an autoradiography (Fig. 8B). Kinase activities of the T^P-Y^P and T-Y forms obtained from prophase and metaphase oocytes, respectively, served as controls. Activity of the T^P-Y form obtained after treatment with GST-pyp3 is very low, consistent with a strong inhibitory effect of phosphorylated Thr-14. In contrast, the activity of the T-Y^P form obtained after treatment with PP2A is significantly higher than those of the T^P-Y^P and T^P-Y forms. Successive treatments with PP2A and GST-pyp3 lead to a p34^cdc2 essentially as active as the M-phase oocyte form. A phosphotyrosyl phosphatase activator of PP2A has been described recently (Cayla et al., 1994; Van Hoof et al., 1994). When added to PP2A, this protein activates the Tyr phosphatase activity of PP2A. No Tyr dephosphorylation of cdc2 was observed when the T^P-Y^P form was treated with PP2A and its phosphotyrosyl phosphatase activator (data not shown), showing that the kinase activity of the T-Y^P form obtained after treatment with PP2A is not due to additional dephosphorylation of the Tyr-15 residue.

DISCUSSION

In higher eucaryotes, p34^cdc2 is phosphorylated on both Thr-14 and Tyr-15 in the G₂ phase, leading to a slowly migrating form of cdc2 on SDS-PAGE. Both sites are inhibitory when phosphorylated (Krek and Nigg, 1991; Norbury et al., 1991; Pickham et al., 1992). At mitosis, these two sites are dephosphorylated, leading to a rapidly migrating form. In numerous studies, both in vivo and in vitro (for an example, see Solomon et al., 1990; Yamashita et al., 1990; Aoki et al., 1992; Atherton-Fessler et al., 1994; Lukas et al., 1992; Kornbluth et al., 1994; O'Connor et al., 1994; Naito et al., 1995; Paules et al., 1995), a third form of cdc2 has been observed, which migrates at an intermediate level between the phosphorylated and the dephosphorylated forms of cdc2. This third form of p34^cdc2 is present in G₂ (Lock, 1992) and in growing HeLa cells (Lukas et al., 1992) and is recognized by an anti-Tyr(P) antibody (Honda et al., 1992). Despite the existence of this third form, possibly representing an intermediately phosphorylated state of cdc2, very few studies have investigated the mechanisms underlying the dephosphorylation of the Thr-14 and Tyr-15 sites (Norbury et al., 1991; Solomon et al., 1990). In this study, we have detected the intermediate migrating form of cdc2 both in vivo, during the starfish oocyte prophase/metaphase transition time course, and in vitro, during cdc2 dephosphorylation by recombinant cdc25 phosphatase. We have identified this third form of cdc2 as a Thr-14 dephosphorylated, Tyr-15 phosphorylated form (T-Y^P). Its identification is based on the use of different tools and methods: (a) three different antibodies; (b) dephosphorylation of the intermediate form with recombinant Tyr phosphatase pyp3, leading to a complete dephosphorylation of cdc2; and (c) generation of the intermediate form using PP2A. These results suggest that dephosphorylation of cdc2 occurs in two steps at the prophase/metaphase transition. Dephosphorylation of the Thr-14 residue happens first, followed by Tyr-15 dephosphorylation. The T-Y^P form of cdc2 is very transient and of low abundance; Tyr-15 is dephosphorylated immediately after Thr-14 dephosphorylation.

cdc2/cdk2 kinases have been reported as physiological substrates of the dual-specificity cdc25 phosphatases (Russell and Nurse, 1986; Galaktionov and Beach, 1991; Gautier et al., 1991; Kumagai and Dunphy, 1991, 1992; Hoffmann et al., 1993; Sebastian et al., 1993; Rime et al., 1994; reviewed by Hoffmann and Karsenti, 1994). One of the three cdc25 human phosphatases, cdc25C, is activated at mitosis, where it dephosphorylates and activates the cdc2 kinase (Izumi et al., 1992; Hoffmann et al., 1993; Izumi and Maller, 1993, 1995; Strausfeld et al., 1994a). In our in vitro experiments, cdc25 A and C dephosphorylate both Thr-14 and Tyr-15 residues, leading to the T-Y form. We observed that among the three different phosphorylated forms of cdc2 (T^P-Y^P, T-Y^P, and T^P-Y), T^P-Y^P is the preferred substrate for cdc25 compared to the two singly phosphorylated forms of cdc2. Dephosphorylation of T-Y^P and T^P-Y is not complete, whereas the T^P-Y^P form is already totally dephosphorylated (data not shown). Binding of cdc25 to its cdc2 substrate, therefore, appears to be facilitated when the two inhibitory sites are phosphorylated. The dephosphorylations occur successively, first Thr-14, then Tyr-15, as observed in vivo. In vivo experiments, involving inhibition of cdc25 by vitamin K3, have revealed that when cdc25 is inhibited, activation of cdc2 is not possible. In this case, the two inhibitory sites remain phosphorylated. From these results, it is reasonable to assume that cdc25 is: (a) present in starfish oocytes; (b) implicated in cdc2 dephosphorylation; and (c) able to act in vivo by the sequential dephosphorylation of Thr-14, then Tyr-15, as it does in vitro. In addition, these results suggest a potential alternative/additional route for cdc2 activation, involving two different phosphatases, a Ser/Thr phosphatase and a Tyr phosphatase, acting successively on cdc2.

The T-Y^P form of cdc2 displays significant kinase activity. In Xenopus extracts, Tyr phosphorylation of cdc2 by wee1 results in substantial, but not complete, reduction of the kinase activity (Mueller et al., 1995a). This significant activity of the T-Y^P form of cdc2, along with the very reduced activity of T^P-Y^P and T^P-Y, directly demonstrates that Thr-14 carries a stronger inhibitory potential than Tyr-15. From the cdk2 crystal structure and mutational analysis, it has been suggested that Thr-14 phosphorylation interferes with ATP binding to the kinase, whereas Tyr-15 phosphorylation may modulate substrate binding (Atherton-Fessler et al., 1994; Debondt et al., 1993; Endicott et al., 1994).

The existence of the singly phosphorylated T-Y^P form may be physiologically relevant in several instances.

(a) MPF activation without Tyr dephosphorylation has been observed in several models, including in growing immature Xenopus oocytes (Rime et al., 1991) and mollusk oocytes. In such cases, Thr-14 phosphorylation/dephosphorylation could provide the only regulatory mechanism of cdc2 activation. Similarly, Tyr phosphorylation of cdc28 in budding yeast is not crucial for normal cell division (Amon et al., 1992; Sorger and Murray, 1992), and Thr may play the regulating function.

(b) The transient singly phosphorylated cdc2 may act to ``jump start'' the cdc25 phosphatase activity, as well as to down-regulate the wee1 kinase activity. The cdc25 phosphatase is involved in the positive feedback loop activating cdc2 and cdk2 (Izumi et al., 1992; Izumi and Maller, 1993, 1995; Hoffmann et al., 1993, 1994; Hoffmann and Karsenti, 1994; Strausfeld et al., 1994a). Its phosphorylation by cdc2/cyclin B and other kinases leads to an active phosphatase. Similarly, the wee1 kinase is inhibited by phosphorylation by cdc2/cyclin B and other kinases (Tang et al., 1993; Honda et al., 1995; McGowan and Russell, 1995; Mueller et al., 1995a; Parker et al., 1995; Watanabe et al., 1995). Phosphorylation of cdc25 by the T-Y^P form of cdc2 would lead to its activation; phosphorylation of wee1 by the T-Y^P form of cdc2 would lead to its inhibition. The result would be a massive activation of cdc2. In that sense, the T-Y^P form would provide the initial start of the cdc2 autoamplification loop observed at the onset of mitosis.

(c) cdc2/cyclin B may be involved at other stages of the cell cycle; it is increasingly phosphorylated during G₁, S, and G₂ (Gu et al., 1992). A T-Y^P form might be active at these stages of the cell cycle and play other functions. In fact, cdc2 kinase activity has been observed in these phases (for an example, see Geneviève-Garrigues et al., 1995). Thr-14 phosphorylation has been described in S. pombe, where it may transiently occur during DNA replication and early G₂ in a subfraction of cdc2 (Den Haese et al., 1995). It has been shown recently that a low level of cdc2/cyclin B kinase activity can promote S-phase entry in fission yeast (Fisher and Nurse, 1996). The phosphorylation status of this complex has not been described yet.

Dephosphorylation of Thr-14 and Tyr-15 during p34^cdc2 activation at the prophase/metaphase transition occurs in two steps. Thr-14 is dephosphorylated before Tyr-15. Although the dual-specificity phosphatase cdc25 is able to dephosphorylate both residues, an alternative/additional route involving the sequential action of a Ser/Thr phosphatase and a Tyr phosphatase is also possible. The intermediate phosphorylated form (T-Y^P), which appears transiently during the cdc2 activation, displays significant kinase activity and may be involved in MPF amplification by phosphorylating and activating the cdc25 phosphatase as well as by phosphorylating and inhibiting the wee1 kinase.