A Mg(2+)-dependent, Ca(2+)-inhibitable serine/threonine protein phosphatase from bovine brain.

The Mg2+-dependent serine/threonine protein phosphatases, also known as type 2C phosphatases (PP2C), belong to a gene family distinct from the other serine/threonine phosphatases and tyrosine phosphatases. Here we report the purification to apparent homogeneity of a novel Mg2+-dependent, Ca2+-inhibitable serine/threonine protein phosphatase from bovine brain. It is a type 2C enzyme in view of its Mg2+ requirement, resistance to okadaic acid and calyculin A, inability to use phosphorylase a as substrate, and a segment of amino acid sequence typical of all PP2C type phosphatases known to date. However, it differs from the other PP2C enzymes, particularly the mammalian PP2Cα and -β isoforms, in that its molecular weight, 76,000, is considerably larger and that it is inhibited by Ca2+, NaF, and polycations, but not by orthovanadate. The Ca2+ inhibition may not be related to its cellular regulation because of Ki values in the 20-90 μM range, but this property permits distinction of this enzyme from the other phosphatases. Although the precise physiological role of this phosphatase is not yet known, its ability to dephosphorylate a wide variety of phosphoproteins and its broad distribution, as shown by a survey of mouse tissues for its activity, suggest that it may serve an important cellular function.

Protein phosphorylation-dephosphorylation is a universal mechanism by which numerous cellular events are regulated. Of the enzymes catalyzing the reversible reactions, it has become apparent that there may exist as many as 1,000 phosphatases which, like the kinases, are just as elaborately and rigorously controlled (1)(2)(3)(4)(5). The serine/threonine-specific phosphatases have been classified into four main types according to their in vitro specificity for selected substrates and sensitivity to activators and inhibitors (6). Sequence analyses revealed that they can be sorted into two major gene families. The first one includes type 1 (PP1), 1 type 2A (PP2A), and type 2B (PP2B) phosphatases, which share 37-59% sequence identity (7) in their catalytic domains and are inhibited by okadaic acid (8). The second family, the Mg 2ϩ -dependent phosphatases, also designated type 2C (PP2C), share little sequence similarity with the first family and are insensitive to okadaic acid.
The Mg 2ϩ -dependent phosphatases have been identified in animals (9 -15), plants (30), and yeast (31), and several have been purified to homogeneity (10 -12, 32, 35). cDNA sequences of PP2C␣ and -␤ from mammalian sources showed Ͼ90% identity (16 -19, 34), whereas those from yeast were 35% identical (13,20,33) and showed 21-24% identity with the mammalian ones. The mammalian PP2C␣ and -␤ are monomeric cytosolic proteins with molecular weights in the 42,000 -45,000 range, and yeast PP2Cs are 31,500 -51,400 monomeric enzymes. PP2Cs of larger sizes, however, have been reported. A rabbit myosin light chain phosphatase has a molecular weight of 70,000 (32), and a bovine pyruvate dehydrogenase phosphatase is a dimer with a 50,000 catalytic subunit complexed to a 97,000 subunit (35). The physiological roles of PP2Cs are unclear (1,2), although they have been implicated in the regulation of fatty acid and cholesterol biosynthesis (21) and heat shock response (13,20). Recently, a M r ϭ 200,000 PP2C-like phosphatase from HeLa cells has been reported (15) which dephosphorylates the C-terminal region of RNA polymerase II. Conceivably, PP2Cs can exist in different molecular sizes and serve diverse biological functions.
We report here the purification to homogeneity and characterization of a novel Mg 2ϩ -dependent, Ca 2ϩ -inhibitable protein phosphatase (MCPP) from bovine brain. Although the Ca 2ϩ inhibition (K i ϭ 20 -90 M) may not be related to its cellular control, this unique property permits the differentiation of this PP2C-type phosphatase from other phosphatases during its isolation. While MCPP possesses many characteristics common to other PP2Cs, it is a monomer with a larger molecular weight of 76,000 and responds dissimilarly to several substrates and inhibitors.
Preparation of 32 P-Labeled Protein Substrates-[ 32 P-Ser]Histone H2B was prepared by incubating 10 mg of the protein with 500 microunits of cAMP-dependent protein kinase catalytic subunit in 1 ml of solution containing 1.5 mM [␥-32 P]ATP (300 Ci), 5 mM MgCl 2 , 50 mM Hepes, pH 7.0, at 30°C for 2 h. Incorporation, usually 0.8 -1 mol of 32 P per mol of histone H2B, was measured by a thin layer chromatography method (23). The reaction mixture was then passed through a Dowex AG 1-X8 column to remove free ATP. MBP, histone H1, and ␣-casein were phosphorylated similarly by cAMP-dependent protein kinase. 32 P-Syntide-2 and -Kemptide were prepared by the same procedure described above except that, after ϳ2 h of incubation, the reaction mixtures were spotted on P-81 phosphocellulose filters, washed with 75 mM phosphoric acid, packed into a column, and then washed extensively with water and eluted with 5 N HCl. Incorporation of 32 P was measured according to Roskoski (24).   Phosphatase Assays-The activity of MCPP was routinely assayed in a 40-l mixture containing 20 M 32 P-labeled histone H2B, 0.1 mM dithiothreitol, 1 mM EGTA, 50 mM KCl, 2.5 mg/ml bovine serum albumin, either with 2 mM MgCl 2 or without added MgCl 2 but with 2 mM CaCl 2 (EGTA keeps free Ca 2ϩ concentration at 1 mM), in 50 mM Hepes buffer, pH 7.8. The presence of CaCl 2 ensured that any MCPP activity in the latter case due to Mg 2ϩ carried over from enzyme and substrate preparations was completely inhibited. Thus, the difference between the ϪCa 2ϩ and ϩCa 2ϩ assays should reflect the activity of MCPP with some contributions from other forms of PP2Cs. The P i released was measured by an acid-molybdenum extraction method (28). All other substrates were assayed in an identical manner. Assay using p-nitrophenyl phosphate was performed as described previously (22). Specific activity of MCPP was expressed as the difference of moles of P i released per min per mg of protein in the absence and presence of Ca 2ϩ . Protein concentrations were determined by the method of Bradford (29) using bovine serum albumin as standard.
Dephosphorylation of ␣ and ␤ Subunits of Phosphorylase b Kinase-The reaction mixture, in a final volume of 130 l, contained 2 M [ 32 P]phosphorylase b kinase, 50 mM KCl, 5 mM MgCl 2 , 1 mM EGTA, 0.1 mM dithiothreitol, 50 mM Hepes buffer, pH 7.8, and 1 g of purified MCPP at 30°C. 15-l aliquots were removed at different time intervals, processed, and run on SDS-PAGE 8 -16% gradient gel. The gel was dried, and dephosphorylation was detected by a PhosphorImager (Molecular Dynamics). PP1 and PP2A were used as controls.
Purification of MCPP-In a typical preparation, 6 bovine brains (1130 g) were homogenized in a blender with 2 volumes per weight (2300 ml) of buffer A (50 mM Tris, pH 7.8, containing 0.5 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 3 mM MgCl 2 , 1 mM phenylmethylsulfonyl fluoride, and 0.1 mM benzamidine-HCl). The homogenate was centrifuged at 1200 ϫ g for 1 h, and the supernatant was filtered through 4 layers of cheesecloth. The extract was mixed with a 500-ml bed volume of DE52 pre-equilibrated in buffer A and gently stirred for 2 h. The DE52 was then collected on a 2-liter scintered glass funnel, resuspended in 500 ml of buffer A, and packed into a 5.5 ϫ 43 cm column already packed with 500 ml of DE52. The column was washed with buffer A ϩ 0.1 M NaCl and then eluted with a linear gradient of 0.1-0.6 M NaCl in the same buffer at a flow rate of 1.2 ml/min (Fig. 1A). The peak phosphatase activity fractions, assayed with histone H2B as described above, were pooled and dialyzed against buffer A. The dialyzed sample was loaded onto an Affi-Gel Blue column (5 ϫ 15 cm) and washed with buffer A at a flow rate of 0.8 ml/min (Fig. 1B). The peak fractions from the wash were collected, concentrated in an Amicon stirred cell fitted with YM-30 membrane, and applied to a Sephacryl S-300 (high resolution) column (2.5 ϫ 91 cm) equilibrated with buffer A ϩ 0.1 M NaCl. The column was eluted with the same buffer at a flow rate of 0.33 ml/min (Fig. 1C). Peak activity fractions were pooled and loaded onto a Mono Q (HR 5/5) column equilibrated in buffer A in a Pharmacia FPLC system. The column was eluted with a linear gradient of 0.1-0.6 M NaCl (Fig. 1D). MCPP activity peak eluted at 0.44 M NaCl. When necessary, last traces of impurities were removed by applying the enzyme sample to a Superose 12 column in the FPLC system. All of the purification steps were performed at 4°C.
Electrophoresis-Gel electrophoreses, nondenaturing or in SDS, were run according to the manufacturer's instructions on a minigel system (Novex). Samples used in SDS-PAGE were heated at 100°C for 3 min in SDS sample buffer containing 5% ␤-mercaptoethanol.
cDNA Sequence-The method of polymerase chain reaction was used for amplification of the coding gene of MCPP from bovine brain cDNA library (gt10 library). Polymerase chain reaction primers with BamHI sites were designed according to MCPP peptide sequences determined by Harvard Microchem. Other primers with SalI sites were on the vector gt10. The purified polymerase chain reaction products were digested with BamHI and SalI and then cloned into Bluescript II SK(ϩ). The double-stranded DNA of recombinant clones containing polymerase chain reaction products were sequenced by a 373A DNA Sequencing System (ABI) to obtain cDNA sequence encoding MCPP. Details of the method will be described elsewhere. 2 Table I summarizes the results of a typical purification described under "Experimental Procedures." The ϮCa 2ϩ assay used to identify MCPP turned out to be quite reliable. After the MCPP was purified and its properties were studied, we were able to compare the ϮCa 2ϩ assay with other assays using specific phosphatase inhibitors (cf. Table IV). As can be seen from Fig. 1A, at the DEAE-cellulose chromatography step, MCPP detected by the standard assay eluted with the main PP2C activity peak that was measured by an assay using 1 M okadaic acid to inhibit both PP1 and PP2A. Calcineurin (PP2B) did not contribute much in this assay when Ca 2ϩ was present because exogenous calmodulin was not added and histone H2B is a poor substrate for this enzyme. After the Sephacryl S-300 step, MCPP activity can be measured readily by the ϮCa 2ϩ method. MCPP activity profile obtained by this method agreed well with that using 1 M okadaic acid and 1 mM orthovanadate (see Table III) to inhibit PP1, PP2A, and PP2C␣ and -␤. The peak fraction from the Mono Q step yielded the pure form of MCPP.

Purification of MCPP-
Purity, Molecular Weight, and Subunit Structure-As shown in Fig. 2, fraction 34 from the Mono Q step migrated predominantly as a single band of M r ϳ 76,000 in both 8 -16% gradient SDS-PAGE and 4 -20% gradient nondenaturing PAGE (overloaded). The gel patterns indicated that MCPP is a monomeric protein with Ͼ90% purity. When the 4 -20% gel was sectioned and assayed, a sharp peak of MCPP activity corresponding to the M r ϭ 76,000 band was observed (Fig. 3), confirming that the phosphatase activity is solely associated with this protein band.
Activation and Inhibition by Various Cations-MCPP was found to require Mg 2ϩ for activity with a K a of ϳ0.8 mM, which 2 K. F. Qin and C. Y. Huang, unpublished data. Calmodulin has no effect on this phosphatase. pH Optimum of MCPP-The pH optimum of MCPP was determined by using histone H2B as substrate in the standard assay. MES buffer was used for pH values from 5.0 -7.0 and Hepes buffer from pH 6.6 -8.5. MCPP activity reached optimal values at pH values greater than 7.5, indicating that it is an alkaline protein phosphatase.
Substrate Specificity-The rates of MCPP-catalyzed dephosphorylation using various substrates and, whenever possible, at comparable concentrations (in terms of phosphoryl groups) The other five were cut into strips, and the protein from each section was eluted and assayed for MCPP activity. are given in Table II. None of the phosphotyrosyl Raytide, MBP, and casein served as substrate for MCPP. Nor was pnitrophenyl phosphate, a widely used substrate for tyrosine phosphatases, dephosphorylated by MCPP. The data shown in Table II showed that MCPP is a phosphoserine/threonine protein phosphatase with diverse specificity. However, MCPP appears to dephosphorylate mostly substrates phosphorylated by cAMP-dependent protein kinase and protein kinase C. Table II also reveals that MCPP prefers basic proteins like MBP and histones as substrates, and MBP phosphorylated by cAMP-dependent protein kinase gave the highest specific activity of ϳ2 mol/min/mg. Phosphorylase b kinase, a substrate used by Ingebritsen and Cohen (6) to designate phosphatase types 1 and 2, was a poor substrate for MCPP. Phosphorylase a, a poor substrate for PP2Cs, was untouched by MCPP. In addition, there are several interesting protein kinase C-phosphorylated substrates of MCPP (not listed in Table II because of inadequate quantitation): a leukemogenic protein called SET (38), the microtubule-associated tau protein, and the 7.5-kDa neurogranin.
Effect of Known Phosphatase Inhibitors-The effects of different known inhibitors of protein phosphatases have been examined, and the results are summarized in Table III. NaF, a common inhibitor of PP1, PP2A, and PP2B, but not of PP2C␣ and -␤, inhibits MCPP. Polylysine and protamine, which activate PP2A, are potent inhibitors of MCPP, although they do not inhibit PP1, PP2B, and PP2C␣ and -␤. Orthovanadate, a potent inhibitor of tyrosine phosphatase that inhibits PP1, PP2A, PP2C (2) at near millimolar levels, has no effect on MCPP at concentrations up to 1 mM. The phosphatase types 1 and 2A inhibitors okadaic acid and calyculin A (27) do not inhibit. Phosphatase (PP1) inhibitors 1 and 2 and heparin also have no effect.
Amino Acid Sequence Homology with Known PP2C-type Phosphatases-Partial cDNA sequence of MCPP obtained by us revealed a segment that is homologous to the same region found in every PP2C reported so far. Fig. 4 compares a 24residue segment of MCPP, which is 90 residues from the C terminus, with those from other sources. The results indicate that MCPP is a member of the PP2C gene family, but it is a new enzyme different from the mammalian PP2C␣ and -␤ isoforms.
Distribution of MCPP in Different Tissues-Since MCPP was isolated from bovine brain, it is of interest to see whether this Ca 2ϩ -inhibitable enzyme or its isoforms may exist in other tissues and other animals. Therefore, we examined the distribution of MCPP activity in various mouse organs by using calyculin A to inhibit PP1 and PP2A and employing the standard ϮCa 2ϩ or Ϯpolylysine assay with histone H2B as substrate to detect the presence of this type of phosphatase. As can be seen from Table IV, the Ca 2ϩ -inhibitable phosphatase activity can be demonstrated in the four selected mouse organs, brain, kidney, lung, and liver, and the results agree very well with those obtained from using polylysine to inhibit MCPP.   DISCUSSION We have purified to homogeneity a new Mg 2ϩ -dependent protein phosphatase from bovine brain. It is named Ca 2ϩ -inhibitable protein phosphatase to differentiate it from other mammalian Mg 2ϩdependent phosphatases which have been called PP2C␣, PP2C␤, or named after their substrates such as phosphofructokinase phosphatase (14), myosin light chain phosphatase (32), and pyruvate dehydrogenase phosphatase (35). Although PP2C-type phosphatases have been identified in various sources, only a handful have been purified to homogeneity, and MCPP is the first 76-kDa enzyme obtained in pure form.
The catalytic and structural characteristics of MCPP indicate that it can be classified as a type 2C phosphatase. 1) It requires Mg 2ϩ (or Mn 2ϩ ) for activity. 2) It catalyzes the dephosphorylation of phosphoseryl/threonyl residues of proteins and peptides phosphorylated by cAMP-dependent protein kinase and protein kinase C (Table II). 3) It is insensitive to inhibitors like okadaic acid and calyculin A, heparin, and PP1 inhibitors 1 and 2. 4) It does not attack phosphorylase a. 5) A segment of its amino acid sequence (24 residues) is homologous to a region found in 12 other PP2C-type enzymes (Fig. 4). In this segment, MCPP is 54.58% identical with and 96% similar to 6 mammalian PP2C␣ and -␤ isoforms; 46 -54% identical with and 88 -92% similar to 4 yeast, isolated or putative, PP2Cs. It should be of interest to note that a CDLLW motif similar to the CDGIW or CDGLW sequence of PP2Cs is found in practically every amino acid sequence of PP1, PP2A, and PP2B (7).
MCPP, however, possesses properties that differ from PP2C␣ and -␤ and other PP2Cs. 1) Its molecular weight, 76,000, is considerably larger than the M r ϭ 42,000 -51,000 mammalian and yeast enzymes, but comparable to the M r ϭ 70,000 myosin light chain phosphatase (32). 2) It is inhibited by polycations and F Ϫ ion which do not inhibit PP2C␣ and -␤, although a Mg 2ϩ -dependent phosphatase from turkey gizzard smooth muscle also is inhibited by F Ϫ (10). 3) PP2Cs are inhibited by millimolar concentrations of orthovanadate, but MCPP is not. 4) PP2Cs preferentially dephosphorylate ␣ subunit of phosphorylase b kinase, whereas MCPP has very low activity with this substrate and shows no preference for either ␣ or ␤ subunit. The M r ϭ 70,000 myosin light chain phosphatase also exhibited little activity with phosphorylase b kinase as substrate (32). 5) MCPP is inhibited by Ca 2ϩ with K i in the 20 -90 M range, depending on the substrate used. The effect of Ca 2ϩ , however, is on the enzyme since the inhibitory effect was observed with every substrate tested so far. The shift in K i suggests that Ca 2ϩ is a noncompetitive inhibitor which affects both V max and K m for the substrates. Since the K i values are much higher than the intracellular Ca 2ϩ level of 0.1-1 M, it appears that the Ca 2ϩ inhibition may not be physiologically important. However, regulation by Ca 2ϩ may not be excluded because, like the synergistic effect of glycogen on the Ca 2ϩ inhibition of glycogensynthetase phosphatase (37), the presence of another compo-nent may be involved in amplifying the Ca 2ϩ signal in vivo. In this regard, it is interesting to note that bovine pyruvate dehydrogenase phosphatase is activated by Ca 2ϩ with K d in the 24 -62 M range in the absence of EGTA buffer (36).
The role of MCPP in the cell, like the other PP2Cs, is not clear at this time. The rather wide distribution of MCPP, as shown in the survey presented in Table IV, suggests that MCPP or its isoforms constitute a new subclass of PP2C and likely serve important cellular functions. It should be of interest to note that a SET protein encoded by a set gene (38) copurified with MCPP until the Mono Q chromatography step. In a case of acute undifferentiated leukemia, the set gene was fused to a can gene as a result of chromosomal translocation (38). Phosphorylation of the SET protein by protein kinase C was blocked by the presence of trace amounts of MCPP, indicating that SET is an excellent substrate for the phosphatase. 3 Since the set gene is expressed in all tissues of the mouse, particularly during embryogenesis, SET may play a key role in the cell and MCPP may regulate the function of SET. The fact that MCPP activity is highest with MBP (10 times better than any other substrates tested so far) may also imply a special function for this phosphatase in the brain. a Activity measured by the standard assay described under "Experimental Procedures" with the addition of 100 nM calyculin A to inhibit PP1 and PP2A. Activity is in picomoles/min. b Activity measured by the method described in Footnote a above Ϯ60 g/ml polylysine (in place of Ϯ1 mM Ca 2ϩ ) in the presence of 100 nM calyculin A. Activity is in picomoles/min. c MCPP activity taken as the difference between the ϮCa or Ϯpolylysine assays.