Dephosphorylation of Human Cyclin-dependent Kinases by Protein Phosphatase Type 2Cα and β2 Isoforms*

We previously reported that the activating phosphorylation on cyclin-dependent kinases in yeast (Cdc28p) and in humans (Cdk2) is removed by type 2C protein phosphatases. In this study, we characterize this PP2C-like activity in HeLa cell extract and determine that it is due to PP2Cβ2, a novel PP2Cβ isoform, and to PP2Cα. PP2Cα and PP2Cβ2 co-purified with Mg2+-dependent Cdk2/Cdk6 phosphatase activity in DEAE-Sepharose, Superdex-200, and Mono Q chromatographies. Moreover, purified recombinant PP2Cα and PP2Cβ2 proteins efficiently dephosphorylated monomeric Cdk2/Cdk6 in vitro. The dephosphorylation of Cdk2 and Cdk6 by PP2C isoforms was inhibited by the binding of cyclins. We found that the PP2C-like activity in HeLa cell extract, partially purified HeLa PP2Cα and PP2Cβ2 isoforms, and the recombinant PP2Cs exhibited a comparable substrate preference for a phosphothreonine containing substrate, consistent with the conservation of threonine residues at the site of activating phosphorylation in CDKs.

Eukaryotic cell cycle progression is driven by the ordered activation and inactivation of cyclin-dependent protein kinases (CDKs). 1 To precisely control the cell cycle engine, extracellular and intracellular signals control CDK activities through a variety of mechanisms, including association with regulatory proteins (cyclins, inhibitors, and assembly factors), subcellular localization, transcriptional regulation, selective proteolysis, and reversible protein phosphorylation (1)(2)(3)(4)(5)(6)(7). In the budding yeast Saccharomyces cerevisiae, Cdc28p is the main CDK involved in regulating the cell division cycle. On the other hand, Cdc2 (Cdk1), Cdk2, Cdk4, and Cdk6 control cell cycle progression in higher eukaryotes. Full activation of CDKs, which is necessary for normal cell cycle progression, requires binding of a cyclin, removal of inhibitory phosphorylations, and the presence of an activating phosphorylation. The cyclins are transcribed, synthesized, and degraded periodically during the cell cycle (1,4,8,9). The inhibitory phosphorylations are carried out by the Wee1-like protein kinases, and removed by members of the Cdc25 family of dual specificity protein phosphatases (for reviews, see Refs. 6 and 7). Activating phosphorylation occurs within the "T-loop" (7, 10) on a conserved threonine residue corresponding to Thr 161 in human Cdc2 and Thr 160 in human Cdk2. Mutation of the equivalent site to alanine in Cdc2 from a variety of species abolishes kinase activity and biological function (11)(12)(13)(14)(15)(16). This activating phosphorylation is carried out by Cdk-activating kinases (CAKs). Higher eukaryotic CAK, primarily localized to the nucleus, is composed of p40 MO15 / Cdk7, cyclin H, and an assembly factor, MAT1. These proteins also function as components of basal transcription factor IIH (17)(18)(19)(20)(21). In contrast to CAK in higher eukaryotes, CAK from budding yeast (Cak1p or Civ1p) is only distantly related to p40 MO15 (22)(23)(24), functions as a monomer, and is predominantly cytoplasmic (25).
Despite the large body of knowledge on CAK, there is relatively little information about the protein phosphatases that reverse the activating phosphorylation of the CDKs. A dualspecificity human protein phosphatase termed KAP (also Cdi1 and Cip2), which was identified by its interaction with Cdc2, Cdk2, and Cdk3 (26 -28), was shown to dephosphorylate Thr 160 on Cdk2 (29). However, there is no obvious KAP homologue in budding yeast. We found recently that two budding yeast type 2C protein phosphatases (PP2Cs), Ptc2p and Ptc3p, are the predominant and physiological enzymes that dephosphorylate Thr 169 on the Cdc28p cyclin-dependent kinase (30). We also observed that PP2C-like activities were responsible for Ͼ99% of the phosphatase activity in HeLa cell extracts acting on Thr 160 of Cdk2, indicating that the ability of type 2C protein phosphatases to remove the activating phosphorylation of CDKs is evolutionarily conserved (30).
In this study, we further characterized the PP2C activities capable of dephosphorylating human Cdk2 and Cdk6 in a HeLa cell extract. Fractionation of proteins by DEAE-Sepharose, Superdex-200, and Mono Q chromatographies demonstrated that Cdk2 and Cdk6 phosphatase activities co-purified with PP2C␣, a previously reported PP2C isoform, and PP2C␤2, a novel human PP2C␤ isoform apparently resulting from alternative splicing. Recombinant PP2C␣ and PP2C␤2 effectively dephosphorylated monomeric-but not cyclin-bound Cdk2 and Cdk6 in vitro, confirming previous studies with yeast PP2Cs and human Cdk2 in a HeLa cell extract (30). Further enzymatic characterization showed that crude HeLa cell extract, partially purified PP2C␣ and PP2C␤2, and recombinant PP2C␣ and PP2C␤2 all exhibited a substrate preference for wild-type (Thr 160 ) Cdk2 compared with a mutant Cdk2 protein containing an altered site of activating phosphorylation (Ser 160 ). These results support the conclusion that PP2C␣ and PP2C␤2 are the PP2C isoforms that dephosphorylate human CDKs in vivo.
[␥-32 P]ATP (3000 Ci/mmol) was from PerkinElmer Life Sciences. Horseradish peroxidase-conjugated secondary antibodies and SuperSignal TM ECL reagents were from Pierce (Rockford, IL). Pfu Turbo DNA polymerase and pBlueScript II KS(Ϫ) were from Stratagene (La Jolla, CA). All other chemicals were from Sigma unless indicated otherwise. 1 ϫ protease inhibitors contained 1 mM phenylmethylsulfonyl fluoride and 10 g/ml each of leupeptin, chymostatin, and pepstatin.
Fractionation of HeLa Cell Extracts-HeLa S3 cells were lysed with a Dounce homogenizer in hypotonic buffer (10 mM Hepes, pH 7.9 (at 4°C), 1.5 mM MgCl 2 , 10 mM KCl, 1 mM dithiothreitol, 1 ϫ protease inhibitors) and clarified by centrifugation at 100,000 ϫ g for 1 h in a 60-Ti rotor (46). The supernatant was frozen in liquid nitrogen and stored at Ϫ80°C until use. Clarified extract (25 ml, protein concentration 6.3 mg/ml) was applied to a DEAE-Sepharose Fast Flow column (1.5 ϫ 10 cm) pre-equilibrated with buffer A (20 mM Tris-HCl, pH 7.4, 0.1 mM EDTA, 0.1 mM EGTA, 0.01% Chaps, 0.1% ␤-mercaptoethanol, 1 ϫ protease inhibitors) at a flow rate of 1 ml/min. The column was washed with buffer A until the absorbance of the elute at 280 nm was Ͻ0.05. Bound proteins were eluted with a 20-ml linear gradient from 0 to 1.0 M NaCl at a flow rate of 1 ml/min. One-ml fractions were collected, and Cdk2/Cdk6 phosphatase activities in the fractions were assayed (see below). The active fractions were pooled, concentrated using a Vivaspin concentrator (10,000 MWCO; Vivascience LTD., Binbrook Hill, United Kingdom), and loaded onto a Superdex-200 column preequilibrated with buffer B (20 mM triethanolamine-HCl, pH 7.0 (at 25°C), 5% glycerol, 0.01% Chaps, 0.1 mM EDTA, 0.1 mM EGTA, 0.1% ␤-mercaptoethanol, 1 ϫ protease inhibitors) at a flow rate of 0.5 ml/min. One-ml fractions were collected, and Cdk2/Cdk6 phosphatase activities in the fractions were assayed (see below). The active fractions (ϳ45-kDa) were pooled, loaded onto a Mono Q HR5/5 column pre-equilibrated with buffer B, and developed with a linear salt gradient from 0 to 700 mM NaCl in buffer B at a flow rate of 0.5 ml/min as described (47). 0.5-ml fractions were collected, assayed for activity, and immunoblotted (see below).
PCR Cloning of Human PP2C␣ and PP2C␤2 Isoforms-To identify novel PP2C␤ isoforms, a BLASTN search of a human EST data base was performed using the last 420 nucleotides in the 3Ј-coding region of human PP2C␤ as the query. In addition to PP2C␤, a second group of EST clones was found that only matched the first ϳ120 nucleotides of the query, indicating that they encode a novel PP2C␤ isoform, which was named PP2C␤2. Total RNA was isolated from HeLa cells with TRIzol reagent according to the manufacturer's protocol. The full-length coding regions of human PP2C␣ and PP2C␤2 isoforms were amplified from total RNA by RT-PCR with the following primers (start and stop codons are underlined): PP2C␣: 5Ј-CCCCATATGGGAGCAT-TTTTAGACAAG-3Ј and 5Ј-CCCCAAGCTTTTACCACATATCATCTGT-TG-3Ј; PP2C␤2: 5Ј-GCCCCATGGGTGCATTTTTGGATAAACC-3Ј and 5Ј-CGGGCTCGAGCTACCATGGGTCTTCTAGATC-3Ј. PCR fragments were inserted into pBlueScript II KS(Ϫ) and sequenced. For PP2C␤2, a second COOH-terminal primer was used to eliminate an internal NcoI site before the stop codon without changing the amino acid residue.
Preparation of 32 P-Labeled CDKs and Protein Phosphatase Assays-Cdk2, GST-Cdk2, GST-Cdk6, and GST-Cdk2 Ser-160 were phosphorylated by GST-Cak1p in the presence of [␥-32 P]ATP as described (30). GST-Cdk2 Thr-160 and GST-Cdk2 Ser-160 were labeled by Cak1p to comparable and nearly saturated levels. The CDK phosphatase activity and casein phosphatase activity were determined as described (30,47). The protein phosphatase activities of HeLa cell lysate and recombinant PP2Cs were assayed in the presence of 5 and 20 mM MgCl 2 , respectively. Briefly, 5 l of each fraction was incubated with 50 ng of 32 Plabeled CDKs in a 20-l reaction for 15 min at room temperature. Reactions were terminated by addition of 10 l of 3 ϫ sample buffer, separated by 10% SDS-PAGE, and analyzed by autoradiography and PhosphorImager (Molecular Imager GS-250, Bio-Rad).
Immunoblotting-Samples were resolved by SDS-PAGE (10% total acrylamide) and transferred to a polyvinylidene difluoride membrane (Immobilon-P, Millipore) with a semi-dry blotting apparatus (Trans-Blot-SD, Bio-Rad). After blocking at room temperature for 2 h in Blotto (1 ϫ TBST containing 5% nonfat dry milk), the membranes were incubated with anti-PP2C polyclonal antibody (1 g/ml in Blotto) overnight at 4°C followed by horseradish peroxidase-conjugated goat anti-rabbit secondary antibody or rabbit anti-sheep secondary antibody (Pierce, 1:2000 dilution in Blotto) at room temperature for 2 h. Antibodies were detected with SuperSignal ECL reagents (Pierce).

Partial Purification of CDK Phosphatases from HeLa Cell
Extract-We previously found that PP2Cs are responsible for the vast majority (Ͼ99%) of phosphatase activity toward Thr 160 of Cdk2 in a HeLa cell extract (30). The activity required Mg 2ϩ and was insensitive to okadaic acid or sodium vanadate. To identify the responsible phosphatase(s), we fractionated a HeLa cell lysate on DEAE-Sepharose Fast Flow, Superdex-200, and Mono-Q columns and assayed fractions for their abilities to dephosphorylate Cdk2 phosphorylated on Thr-160 ( Fig. 1) and Cdk6 phosphorylated on Thr 177 (data not show). The Cdk2 and Cdk6 phosphatase activities were present in the same fractions and showed only a single sharp peak in the first two columns. Mono-Q chromatography, however, showed a broader distribution of Cdk2/Cdk6 phosphatase activity (Fig. 1D), suggesting the presence of more than one CDK phosphatase.
PP2C␣ and PP2C␤ isoforms are the mammalian PP2Cs most closely related to Ptc2p and Ptc3p, the budding yeast phosphatases responsible for dephosphorylating Thr 169 of Cdc28p. We speculated that the CDK phosphatases in HeLa cells might be PP2C␣ and/or PP2C␤ isoforms. This suggestion was supported by the observation that overexpression of human PP2C␣ in yeast led to synthetic lethality in cak1-22 ts cells at a semipermissive temperature. 2 We determined whether the active Mono Q fractions contained PP2C␣ and/or PP2C␤ proteins by immunoblotting. 0.5-ml fractions were collected to provide higher resolution than in Fig. 1D. Affinity purified antibodies that specifically recognize both PP2C␣ and ␤ detected two proteins in a HeLa cell lysate (Fig. 2B). Their sizes of about 45 and 42 kDa were similar to the calculated molecular mass of PP2C␣ (ϳ42-kDa). Immunoblotting of Mono-Q fractions showed that the 45-and 42-kDa PP2C␣/␤s co-purified with the Cdk2/Cdk6 phosphatase activities (Fig. 2, A and B). The total amounts of these isoforms correlated well with Cdk2/Cdk6 phosphatase activity, suggesting that they were likely candidates for the CDK phosphatases in HeLa cells. The elution profiles of the 45-and 42-kDa PP2C␣/␤s on Mono Q chromatography were similar to those of the previously described rabbit PP2C1 and PP2C2 isoenzymes, respectively (48). In addition, the 45-and 42-kDa PP2C␣/␤ isoforms were partially separable by Mono-Q chromatography (Fig. 2B): fractions 30 and 31 contained only the 45-kDa PP2C␣/␤ and factions 34 and 35 contained almost exclusively the 42-kDa PP2C␣/␤. Since the CDK phosphatase activity in fraction 34 was similar to that in fraction 30 ( Fig. 2A), it appears that both the 45-and 42-kDa PP2C␣/␤ proteins are capable of dephosphorylating Cdk2 and Cdk6.
Identification of PP2C␣ and PP2C␤2 Isoforms in HeLa Cell Lysate-We identified the 45-and 42-kDa PP2C isoforms using a combination of immunoblotting and molecular cloning. An antibody specific for PP2C␣ isoforms recognized the 45-kDa PP2C but not the 42-kDa PP2C (Fig. 2C), indicating that the 45-kDa protein is a PP2C␣ isoform and that the 42-kDa protein is a PP2C␤ isoform. Moreover, both the 45-kDa PP2C␣ isoform and recombinant PP2C␣ exhibited the same mobility in SDS-PAGE (Fig. 2D, lanes 1 and 2), indicating that the 45-kDa PP2C was likely the previously reported human PP2C␣.
In contrast, the 42-kDa PP2C␤ isoform was much smaller than the previously reported human PP2C␤, which has 479 amino acid residues and an apparent molecular mass of ϳ55-kDa in SDS-PAGE (49). Instead, the size of the 42-kDa PP2C␤ isoform was similar to those of reported PP2C␤s from rabbit, mouse, and rat (34 -36). We speculated that the 42-kDa PP2C␤ might be an unreported human ortholog of PP2C␤s in these other mammals. By searching an EST data base, we identified and cloned a novel human PP2C␤ isoform ("PP2C␤2") from HeLa cells by RT-PCR. The encoded protein is predicted to be 387 amino acids, compared with 479 amino acids for human PP2C␤ and 390 amino acids for mouse and rat PP2C␤. Since the first 1134 nucleotides (378 amino acids) in the human PP2C␤ and PP2C␤2 coding regions were identical, PP2C␤ and PP2C␤2 appear to arise via alternative splicing. Human PP2C␤2 showed ϳ95% identity to mouse and rat PP2C␤s (Fig.  3). Recombinant human PP2C␤2 exhibited the same mobility as the 42-kDa PP2C on SDS-PAGE (Fig. 2D, lanes 3 and 4), suggesting that the 42-kDa HeLa PP2C is PP2C␤2. Interestingly, the mobility of PP2C␣ was less than that of PP2C␤2 on

FIG. 2. Identification of PP2C␣ and PP2C␤2 isoforms as Cdk2 and Cdk6 phosphatases.
A, analysis of Cdk2 and Cdk6 phosphatase activities in the Mono Q fractions. 5 l of each fraction was incubated with 50 ng of 32 P-Cdk2 or 32 P-Cdk6 at room temperature for 15 min, separated by SDS-PAGE, and analyzed by autoradiography. B, immunoblotting analysis of a HeLa cell lysate and of partially purified Cdk2/ Cdk6 phosphatases using a PP2C␣/␤-specific antibody. Proteins from 40 l of HeLa S3 lysate or 20 l of the indicated Mono Q fractions were resolved in 10% SDS-PAGE, transferred to a membrane, and detected with a PP2C␣/␤ specific antibody. C, immunoblotting analysis of partially purified Cdk2/Cdk6 phosphatases using a PP2C␣-specific antibody. Proteins from 20 l of the indicated Mono Q fractions were resolved in 10% SDS-PAGE, transferred to a membrane, and detected with a PP2C␣-specific antibody. D, comparison of partially purified Cdk2/Cdk6 phosphatases with recombinant PP2C␣ and PP2C␤2 isoforms. Mono Q fractions 31 and 34, non-tagged recombinant PP2C␣, and non-tagged recombinant PP2C␤2 were separated in 10% SDS-PAGE and analyzed by immunoblotting with PP2C␣/␤-specific antibodies.
Substrate Preference of PP2C␣ and PP2C␤2-PP2C␣ has been observed to show a 20-fold preference for a phosphothreonine peptide substrate compared with an equivalent phosphoserine substrate in vitro (51), leading to the suggestion that PP2C substrates are generally phosphorylated on threonine residues (32). We, therefore, compared the abilities of PP2Cs to dephosphorylate wild-type Cdk2 Thr-160 and mutant Cdk2 Ser-160 . We previously showed that a HeLa cell extract dephosphorylated Cdk2 Thr-160 about 4 times as fast as Cdk2 Ser-160 (52). Fig. 5A confirms this preference. Fig. 5B shows that the partially purified HeLa PP2Cs (Mono-Q fractions 31 and 34) exhibited the same qualitative preference for Cdk2 Thr-160 . To quantitate this effect, we used recombinant PP2C␣ and PP2C␤2 and performed assays within the linear range (when less than 30% of the substrate was dephosphorylated). This analysis showed that PP2C␣ and PP2C␤2 have 3-and 2.7-fold preferences for Cdk2 Thr-160 over Cdk2 Ser-160 (Fig. 5C), confirming the qualitative observations with the native enzymes (Fig. 5B).
Binding of Cyclin Prevents the Dephosphorylation of CDKs by PP2Cs-Since the binding of cyclins to CDKs blocked the dephosphorylation of yeast Cdc28p and human Cdk2 by Ptc2p, Ptc3p, and KAP (29 -30), we tested whether the binding of cyclin could also inhibit the dephosphorylation of Cdk2 and Cdk6 by human PP2C isoforms. Preincubation of Cdk2 with excess GST-cyclin A blocked the dephosphorylation of Cdk2 by 3  PP2C␣, PP2C␤2, PP2C␣⌬C, and PP2C␤2⌬C (Fig. 6A). Similarly, the dephosphorylation of GST-Cdk6 by PP2C␣ and PP2C␤2 was blocked by the binding of a viral D-type cyclin, KSHV-cyclin (Fig. 6B). In contrast, these cyclins had no effects on the casein phosphatase activities of PP2C␣, PP2C␤2, PP2C␥, PP2C␣⌬C, and PP2C␤2⌬C (Fig. 6C), indicating that cyclins did not inhibit PP2C activity directly. DISCUSSION We previously reported that two yeast PP2Cs, Ptc2p and Ptc3p, are the major physiological protein phosphatases for the Cdc28p cyclin-dependent kinase in budding yeast and that PP2C-like activity was also responsible for Ͼ99% of the Cdk2 phosphatase activity in a HeLa cell extract (30). We now provide evidence that the PP2C-like activities in HeLa cell extract are due to PP2C␤2, a novel PP2C␤ isoform, and to PP2C␣. PP2C␣/␤-specific antibodies detected two proteins with apparent molecular masses of 45-and 42-kDa in HeLa cell lysate. Mg 2ϩ -dependent Cdk2 and Cdk6 phosphatase activity co-purified with 45-and 42-kDa PP2C␣/␤ isoforms during chromatography on DEAE-Sepharose, Superdex-200, and Mono-Q columns. The 45-kDa PP2C was also recognized by a PP2C␣specific antibody and had the same electrophoretic mobility as recombinant PP2C␣. The 42-kDa PP2C␣/␤ did not react with the PP2C␣-specific antibody and exhibited the same electrophoretic mobility as a novel PP2C␤ isoform (PP2C␤2). PP2C␤2 possesses 387 amino acid residues and shows ϳ95% identity to PP2C␤'s from mouse, rat, and rabbit. We found that recombinant PP2C␣ and PP2C␤2 efficiently dephosphorylated monomeric Cdk2 and Cdk6 in vitro but that two other PP2C isoforms, PP2C␥ and Wip1, did not. Further biochemical analysis demonstrated that Cdk2 Ser-160 was a relatively poor substrate compared with wild-type Cdk2 Thr-160 for HeLa cell extract, the partially purified 45-and 42-kDa PP2Cs, and recombinant PP2C␣ and PP2C␤2. Similar to budding yeast Ptc2p and Ptc3p, PP2C␣ and PP2C␤2 could not dephosphorylate cyclin-bound CDKs. These results indicate that human PP2C␣ and PP2C␤2 represent the PP2C activity responsible for removing the activating phosphorylation from Cdk2 in HeLa cells. These studies also provide evidence that Cdk6, like Cdk2, is a substrate for PP2Cs.
Besides CDKs, a number of MAP kinases appear to be substrates for type 2C protein phosphatases. For instance, PP2Cs have been implicated in negatively regulating stress-responsive protein kinase cascades in eukaryotic cells. In both budding yeast and fission yeast, genetic studies have shown that PP2C-like enzymes oppose the activation of the MAP kinase pathway that is activated in response to osmotic and heat shocks (53)(54)(55). In human cells, PP2C␣ can reverse the activation of the p38 and JNK MAPKs induced by stress and cytokines (44). Biochemically, a human PP2C␣ isoform dephosphorylated a similar threonine within the activation loop of the p38 MAPK (44). Recently, Ptc1 and Ptc3 in Schizosaccharomyces pombe were shown to dephosphorylate Thr 171 of the p38 homolog (Spc1) in its activating loop (56). Given the similarity between the sites of activating phosphorylation in MAPKs and CDKs, we have proposed that PP2C-like enzymes could be general T-loop protein phosphatases (30).
The activities of PP2Cs toward a variety of substrates can be affected by whether the site to be dephosphorylated is a serine or a threonine and by regulatory factors that bind to the substrates. Biochemically, PP2C␣ has been seen to dephosphorylate a phosphothreonine substrate 20-fold more efficiently than the corresponding phosphoserine substrate in vitro (51) and PP2C substrates have been proposed to be phosphorylated on threonine residues in vivo (32). Using human Cdk2, a likely physiological substrate for PP2Cs, we confirmed that PP2C␣ FIG. 5. Substrate preferences of PP2C␣ and PP2C␤2 isoforms. A, kinetics of dephosphorylation by Cdk2 Thr-160 and Cdk2 Ser-160 by PP2C-like activities in HeLa cell extract. 50 ng of 32 P-Cdk2 Thr-160 or 32 P-Cdk2 Ser-160 were incubated with 2.5 g of HeLa cell extract at room temperature. The reactions were stopped at different times (0, 1, 2, 5, 10, 15, and 20 min) and analyzed by SDS-PAGE followed by autoradiography. B, relative activities of partially purified CDK phosphatases using Cdk2 Thr-160 and Cdk2 Ser-160 as substrates. 50 ng of 32 P-Cdk2 Thr-160 (solid bars) or 32 P-Cdk2 Ser-160 (open bars) was incubated with 5 l of buffer or Mono Q fractions 31 or 34 at room temperature for 15 min. The reactions were stopped and analyzed by SDS-PAGE followed by Phos-phorImager analysis. C, relative protein phosphatase activities of recombinant PP2C␣ and PP2C␤2 using Cdk2 Thr-160 and Cdk2 Ser-160 as substrates. 100 ng of 32 P-Cdk2 Thr-160 or 32 P-Cdk2 Ser-160 was incubated with 10 ng of PP2C␣ or PP2C␤2 at room temperature. The reactions were stopped at different times (0, 1, 2, 5, 10, 15, and 20 min) and analyzed by SDS-PAGE followed by PhosphorImager analysis. The ratios of activities toward Cdk2 Thr-160 and Cdk2 Ser-160 are shown. Data represent the mean and standard deviations of six experiments. The samples were then incubated with buffer or 100 ng of PP2Cs at room temperature for 15 min. Cdk2 (A) and Cdk6 (B) were separated by 10% SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and analyzed by autoradiography (AR) and immunoblotting (IB) with anti-PSTAIR antibodies (A). C, cyclins have no effect on the casein phosphatase activity of PP2Cs. 50 ng of PP2Cs was preincubated with 250 ng of cyclin A or KSHV-cyclin ("V-cyc") at room temperature for 30 min, followed by determination of casein phosphatase activities. and PP2C␤2 removed the phosphate from a phosphoserine substrate slower (ϳ3-fold) than from the phosphothreonine substrate (Fig. 5). Indeed, many PP2C substrates, including all known CDKs undergoing activating phosphorylation, AMPK (57)(58)(59), moesin (62), and p38 MAPK (44), are phosphorylated on threonine residues. However, some PP2C substrates such as axin (60) and CFTR (61) may be phosphorylated on serines. Since exchanges of serines and threonines often have little effect on protein functions, the replacement of a threonine with a serine could be used to control the duration of phosphorylation. In addition, the binding of ligands or regulatory proteins to substrates also influences the rate of dephosphorylation by PP2Cs. For example, the dephosphorylation of AMPK is inhibited in the presence of 5Ј-AMP (59), and the dephosphorylation of CDKs is blocked by the binding of cyclins (Fig. 6, A-B, and Ref. 30). The binding of ligands and regulatory proteins could induce conformational changes, block dephosphorylation, and preserve the phosphorylated state of the substrate for a period of time. Dephosphorylation could occur rapidly following removal of the ligand or of the regulatory proteins, such as happens with CDKs after cyclin degradation.
Given that no regulatory subunits for PP2C␣/␤ have been found and that PP2C␣/␤ isoforms only differ significantly in their COOH-terminal segments, it is tempting to speculate that the diverse COOH-terminal regions are involved in regulating PP2C activity. Our studies showed that the COOHterminal truncated forms of PP2C␣ and PP2C␤2 dephosphorylated Cdk2 and Cdk6 as well as the full-length enzymes (Fig.  4), indicating that the carboxyl regions may not directly regulate enzymatic activity or substrate specificity. Additional work will be required to determine whether these segments might facilitate interactions with substrates, localization within the cell, or other properties of these enzymes. In addition to the COOH-terminal regions, mammalian PP2C␣ and -␤ isoforms possess potential N-myristoylation sites similar to those in budding yeast Ptc2p and Ptc3p. In the crystal structure of human PP2C␣, the NH 2 -terminal glycine residue is close to its catalytic center (32), therefore, it is possible that mammalian PP2C␣ and PP2C␤ are regulated by N-myristoylation in vivo. Further experiments will be necessary to test this possibility.