A Novel Protein Phosphatase That Dephosphorylates and Regulates Ca2+/Calmodulin-dependent Protein Kinase II*

A synthetic peptide corresponding to the autophosphorylation site of Ca2+/calmodulin-dependent protein kinase II (CaMKII) (residues 281–289) was conjugated to paramagnetic particles, and phosphorylated by a constitutively active CaMKII fragment. Using this phosphopeptide conjugate as a substrate, a calyculin A-insensitive, Mn2+-dependent, and poly-l-lysine-stimulated protein phosphatase activity was detected in the crude extract of rat brain. The protein phosphatase (designated as CaMKII phosphatase) (CaMKIIPase) was purified to near homogeneity from rat brain. CaMKIIPase showed apparent molecular weights of 54,000 and 65,000, on SDS-polyacrylamide gel electrophoresis and gel-filtration analysis, respectively. It was not inhibited by 100 nm calyculin A or 10 μmokadaic acid. Mn2+, but not Mg2+, was absolutely required for activity. CaMKIIPase was potently activated by polycations. Autophosphorylated CaMKII was dephosphorylated by CaMKIIPase, whereas phosphorylase kinase, mixed histones, myelin basic protein, and α-casein (which had been phosphorylated by cAMP-dependent protein kinase) and phosphorylasea (phosphorylated by phosphorylase kinase) were not significantly dephosphorylated. No other proteins than CaMKII in rat brain extract which had been phosphorylated by CaMKII were dephosphorylated. The stimulated Ca2+-independent activity of autophosphorylated CaMKII was reversed by the action of CaMKIIPase. Thus, CaMKIIPase appears to be a specialized protein phosphatase for the regulation of CaMKII.

Calmodulin-dependent protein kinase II (CaMKII) 1 occurs abundantly in the brain, and has a broad substrate specificity (1). It plays a number of roles in the functioning of the central nervous system in response to intracellular Ca 2ϩ (reviewed in Refs. [2][3][4]. The possible involvement of CaMKII in the regulation of neuronal functions such as neurotransmitter synthesis (5,6), neurotransmitter release (7,8), long term potentiation (9 -11), and the formation of spatial learning (12) has so far been suggested. CaMKII is known to be regulated by autophosphorylation at multiple sites. Among them, autophosphorylation at Thr 286 results in generation of Ca 2ϩ /calmodulin-independent activity (13)(14)(15)(16)(17), full activation of the total activity (18 -20), and trapping of Ca 2ϩ /calmodulin (21,22). Therefore, the protein phosphatases that dephosphorylate the residue may be important for regulation of CaMKII activity. It has so far been reported that protein phosphatase 1 (PP1) (23), protein phosphatase 2A (PP2A) (14,16,24,25), and protein phosphatase 2C (PP2C) (26) dephosphorylate Thr 286 in vitro and that the dephosphorylation is catalyzed by distinct phosphatases in distinct subcellular compartments (27). Recently, we developed a novel in-gel protein phosphatase assay, using polyacrylamide gels containing phosphorylated peptide conjugates, and found that at least three distinct phosphatases existed in the rat brain extract which catalyzed dephosphorylation of the residue corresponding to Thr 286 of CaMKII in the gels (28).
In the present study, we achieved purification and characterization of the protein phosphatase with an apparent molecular weight of 54,000, which had been previously detected by the in-gel assay (28). The purified phosphatase, which was designated as CaMKII phosphatase (CaMKIIPase), was highly specific for autophosphorylated CaMKII, and actually reversed the activated Ca 2ϩ /calmodulin-independent activity of autophosphorylated CaMKII.

Purification of CaMKIIPase from Rat Brain Stem
All procedures were carried out at 4°C. Approximately 20 g of rat brain stems, which had been frozen in liquid nitrogen immediately after * This work was supported by grants-in-aid for scientific research from the Ministry of Education, Science, Sports and Culture of Japan and by grants from the Byotai Taisha Research Foundation, the Mitsubishi Foundation, and the Uehara Memorial Foundation. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 The abbreviations used are: CaMKII, calmodulin-dependent protein kinase II; BSA, bovine serum albumin; CaMKIIPase, CaMKII phosphatase; DTT, dithiothreitol; EMCS, N-(6-maleimidocaproyloxy)succinimide; HPLC, high performance liquid chromatography; MBP, myelin basic protein; PAGE, polyacrylamide gel electrophoresis; PKA, catalytic subunit of cAMP-dependent protein kinase; PMC, peptide-magnetic particle conjugate; PMSF, phenylmethanesulfonyl fluoride; poly(Lys), poly-L-lysine; poly(Arg), poly-L-arginine; poly(Glu,Lys,Tyr); poly-L-(glutamic acid,lysine,tyrosine); PP, protein phosphatase. excision and stored at Ϫ80°C, were homogenized in 70 ml of 50 mM Tris-HCl (pH 7.5) containing 10 mM EDTA, 10 mM EGTA, 1 mM dithiothreitol (DTT), 1 mM phenylmethanesulfonyl fluoride (PMSF), and 5 g/ml each of pepstatin, leupeptin, antipain, and chymostatin with a Teflon/glass homogenizer. The homogenate was centrifuged at 100,000 ϫ g for 60 min, and the supernatant was applied to a phosphocellulose column (Whatman P11, 5.2 ϫ 3.0 cm), which had been equilibrated with the homogenization buffer. The flow through fraction was pooled and applied to a DEAE-cellulose column (Whatman DE52, 5.2 ϫ 3.0 cm), which had been equilibrated with the homogenization buffer without PMSF, pepstatin, leupeptin, antipain, and chymostatin. The column was washed with buffer A (comprising 25 mM Tris-HCl (pH 7.5), 1 mM EGTA, 0.5 mM DTT, 0.05% Tween 40) containing 1 mM PMSF and 0.1 M NaCl, and then eluted with buffer A containing 0.25 M NaCl. To the active fraction eluted, solid ammonium sulfate was added to give 60% saturation, and it was allowed to stand for overnight while being stirred. The precipitate, collected by centrifugation at 100,000 ϫ g for 30 min, was dissolved in 25 mM Tris-HCl (pH 7.5) containing 1 mM EGTA and 0.5 mM DTT (buffer B). To this solution, an equal volume of buffer B containing 2 M ammonium sulfate was added. The supernatant, obtained by centrifugation at 100,000 ϫ g for 30 min, was filtrated by a cellulose acetate membrane filter (0.45 m, Advantec) and applied to a TSKgel phenyl-5PW column (0.5 ϫ 5.0 cm, Tosoh) attached to a non-metal HPLC system (Tosoh), previously equilibrated with buffer B containing 1 M ammonium sulfate. The column was eluted with a linear gradient from 1 to 0 M ammonium sulfate. Fractions eluted as a major peak of CaMKIIPase activity were pooled and made 60% saturation in ammonium sulfate. After overnight stirring, the precipitate was collected by centrifugation at 100,000 ϫ g for 40 min, and it was dissolved in buffer B containing 0.05% Tween 40 and 0.1 M NaCl, followed by centrifugation at 30,000 ϫ g for 20 min. Subsequently, the supernatant was applied to a Hiload 26/60 Superdex 200 prep grade column (Pharmacia) attached to a non-metal HPLC system (Tosoh), which had been equilibrated with the same buffer, and active fractions eluted with the same buffer were pooled. Ethylene glycol was added to the pooled active solution to give 5% (v/v), and it was centrifuged at 30,000 ϫ g for 10 min. The supernatant was applied to a DEAE-NPR column (0.46 ϫ 10 cm, Tosoh) attached to a HPLC system (Tosoh), previously equilibrated with buffer C (comprising 40 mM Tris-HCl (pH 7.5), 0.5 mM EDTA, 0.5 mM EGTA, 0.05% Tween 40, 5% ethylene glycol, and 0.5 mM DTT) containing 50 mM NaCl. The column was eluted, at a flow rate of 0.3 ml/min, with linear gradients of NaCl in buffer C as follows: 0 -15 min, 50 mM NaCl; 15-20 min, 50 -120 mM NaCl; 20 -85 min, 120 -175 mM NaCl; 85-105 min, 175-300 mM NaCl. Fractions of 0.3 ml were then collected. The purities of the active fractions were checked by SDSpolyacrylamide gel electrophoresis (SDS-PAGE), and essentially pure fractions were combined. The purified enzyme was frozen in aliquots at Ϫ80°C. The enzyme can be stored for at least several months without detectable loss of activity under the storage conditions.
CaMKII (112.4 g/ml) was autophosphorylated at 5°C for 1 min in the reaction mixture containing 40 mM Hepes-NaOH (pH 8.0), 5 mM Mg(CH 3 CO 2 ) 2 , 0.1 mM EGTA, 5 M calmodulin, 0.4 mM CaCl 2 , 0.02% Tween 20, and 50 M [␥-32 P]ATP (28,000 cpm/pmol). Excess EDTA (12.3 mM) and BSA (1 mg/ml) were added to stop the reaction, and the reaction mixture was applied to a small column of Sephadex G-50 (Pharmacia Biotech Inc.) equilibrated with 50 mM Tris-HCl (pH 7.5) containing 0.2 M NaCl, 0.05% Tween 40, and 1 mM DTT. The column was eluted with the same buffer by the spin column method as described (35). Fractions of 50 -100 l were collected, and the radioactive fractions eluted near the void volume of the column were pooled and stored in aliquots at Ϫ80°C.
Mixed histones (1 mg/ml), MBP (1 mg/ml), ␣-casein (4 mg/ml), and phosphorylase kinase (1.6 mg/ml) were phosphorylated by PKA in the reaction mixture containing 50 mM Hepes-NaOH (pH 7.0), 5 mM Mg(CH 3 CO 2 ) 2 , 0.01% Tween 20, 2 g/ml PKA, and 50 M [␥-32 P]ATP (28,000 cpm/pmol). For phosphorylase kinase, EGTA (1 mM) and EDTA (0.4 mM) were included in the reaction mixture. The phosphorylation reactions were carried out at 30°C for 4 h for histones, MBP, and ␣-casein, and for 4 min for phosphorylase kinase. The reactions were terminated and the resulting phosphoproteins were desalted, as described above. The phosphoproteins were stored in aliquots at Ϫ80°C. The substrate concentrations presented in the text represent the concentrations of 32 P bound to the substrate proteins.
A crude extract of rat cerebral cortex (407 g/ml) was phosphorylated by CaMKII at 30°C for 1 min in the reaction mixture containing 40 mM Hepes-NaOH (pH 8.0), 5 mM Mg(CH 3 CO 2 ) 2 , 0.5 mM EGTA, 0.4 mM EDTA, 1 M calmodulin, 1.2 mM CaCl 2 , 0.01% Tween 20, 0.96 g/ml CaMKII, and 50 M [␥-32 P]ATP (26,300 cpm/pmol). Termination of the reaction and desalting of the phosphoproteins were carried out as described for autophosphorylated CaMKII except that 100 nM calyculin A was included in the equilibration/elution buffer for the spin column. The phosphorylated sample was stored in aliquots at Ϫ80°C.

Protein Phosphatase Assay
CaMKIIPase activity was determined by the following four assay methods. Unless otherwise stated, poly(Lys) means M r 87,000 species of poly(Lys). Basic proteins and polycations including poly(Lys) were initially dissolved in phosphate-buffered saline to yield 10 mg/ml solutions and they were diluted with H 2 O to appropriate concentrations prior to use.
Immobilized Phosphopeptide Assay-For analysis with the purified enzyme, dephosphorylation of [ 32 P]PMC-KII(281-289) was carried out at 30°C for 5 min in a reaction mixture (30 l) containing an appropriate amount of CaMKIIPase, 50 mM Tris-HCl (pH 7.5), 100 mM KCl, 2 mM MgCl 2 , 2 mM MnCl 2 , 0.1 mM EGTA, 0.01% Tween 20, 10 g/ml poly(Lys), and 89 nM [ 32 P]PMC-KII(281-289) in a polypropylene microtube (Eppendorf). For purification of CaMKIIPase, the dephosphorylation of [ 32 P]PMC-KII(281-289) was carried out at 30°C for 10 min in the presence of 100 nM calyculin A. It should be noted that the peptide conjugate in the tube must be suspended uniformly by vortexing just prior to initiation of the reaction. The reaction was started by adding the enzyme, and terminated by adding 500 l of 5% trichloroacetic acid, followed by vortexing and then brief centrifugation at room temperature. The 32 P radioactivity of the supernatant (450 l) was determined by liquid scintillation counting. The sedimented conjugate was attached to the bottom of the tube by a neodymium iron magnet (Polysciences, Inc.) when the supernatant was withdrawn. One unit of enzyme was defined as the amount that catalyzed the release of 1 mol of phosphate/min at 30°C.
Trichloroacetic Acid Precipitation Assay-This method was used when phosphoproteins were used as substrates. Prior to addition of the substrate, the reaction mixture was preincubated for 30 s at 30°C, and the reaction was started by adding the phosphoprotein substrate. After incubation for 30 s or 1 min, a 10-l aliquot was withdrawn and transferred to a tube containing 100 l of 20% trichloroacetic acid, and then 100 l of 6 mg/ml BSA was added to the tube, followed by vortexing. The mixtures were allowed to stand for 10 min on ice, and then 300 l of ice-cold 5% trichloroacetic acid was added. After centrifugation for 5 min at 4°C in a microcentrifuge at maximum speed, 450 l of the supernatant was withdrawn and counted for 32 P radioactivity.
SDS-PAGE Assay-After preincubation for 30 s at 30°C, the reaction was started by adding the phosphoprotein. After incubation for 1 min, the reaction was terminated by adding 22.7 mM EDTA, and the mixture was mixed with an equal volume of a sample buffer consisting of 125 mM Tris-HCl (pH 6.8), 4% SDS, 20% glycerol, 0.002% bromphenol blue, and 285 mM ␤-mercaptoethanol. The mixture was boiled for 2 min and centrifuged for 2 min at room temperature in a microcentrifuge at maximum speed, and an aliquot of the supernatant was subjected to SDS-PAGE on a 10% acrylamide gel. The gel was dried and visualized by autoradiography.
In-gel Assay-In-gel phosphatase assay of CaMKIIPase was carried out as described previously (28).

Other Analytical Procedures
SDS-PAGE was carried out according to the method of Laemmli (36). Protein concentrations were determined by the method of Lowry et al. (37), as modified by Peterson (38).

RESULTS
Purification of CaMKIIPase-Since our previous in-gel assay for protein phosphatase with a synthetic peptide corresponding to the autophosphorylation site of CaMKII (residues 281-289) as a substrate suggested the existence of two protein phosphatases specific for the autophosphorylation site in rat brain (28), we attempted to purify and characterize them. Our preliminary experiments using immobilized phosphopeptide assay showed that a calyculin A-insensitive, Mn 2ϩ -dependent, and poly(Lys)-stimulated phosphatase activity toward [ 32 P]PMC-KII(281-289) was present in a rat brain extract (data not shown). The possibility that the activity detected under such conditions was due to a protease could be eliminated, since almost all the radioactivities released into the supernatant of the assay mixture during the reaction were extracted with isobutyl alcohol/heptane (1/1, v/v) in the presence of ammonium molybdate/sulfuric acid (data not shown). Thus, we purified a protein phosphatase, which can dephosphorylate [ 32 P]PMC-KII(281-289) in the presence of 100 nM calyculin A, 2 mM Mn 2ϩ , and 10 g/ml poly(Lys), from rat brain stem, as described under "Experimental Procedures." Table I summarizes the results of a typical purification. The purified protein gave a single protein band corresponding to a molecular weight of about 54,000 on SDS-PAGE (Fig. 1A), and in-gel protein phosphatase assay also gave a single transparent band at the same position as the protein staining band (Fig. 1B), indicating that the protein purified to apparent homogeneity was CaMK-IIPase. The molecular weight of the enzyme, as determined by gel filtration, was approximately 65,000 (Fig. 1C), indicating that CaMKIIPase consisted of a single polypeptide chain.
Factors Affecting CaMKIIPase Activity-We initially detected protein phosphatases capable of dephosphorylating a phosphopeptide corresponding to the autophosphorylation site of CaMKII (residues 281-289) in a reaction mixture containing 50 mM Tris-HCl (pH 7.0), 100 mM KCl, 2 mM MgCl 2 , 2 mM MnCl 2 , 0.5 mM EGTA, and 2 mM DTT (28). When purified CaMKIIPase was assayed by immobilized phosphopeptide method as described under "Experimental Procedures," it required Mn 2ϩ (2 mM) for its activity and showed no activity in the absence of Mn 2ϩ (data not shown). No other divalent cations tested, such as Mg 2ϩ , Zn 2ϩ , Co 2ϩ , Ca 2ϩ , Ni 2ϩ , and Ba 2ϩ , could replace Mn 2ϩ but, in the presence of Mn 2ϩ , Mg 2ϩ slightly stimulated the activity (109.7 Ϯ 2.6%). When the pH profile of CaMKIIPase activity was examined using immobilized phosphopeptide assay, the maximal activity was observed over the pH range of 7 to 8. CaMKIIPase activity was markedly stimu-lated in the presence of poly(Lys). Fig. 2 shows the effect of varying the concentration of poly(Lys) on the activation of CaMKIIPase. Maximal activation occurred with 5-50 g/ml poly(Lys), and higher concentrations of poly(Lys) resulted in decrease in the enzyme activity. The activation by poly(Lys) was enhanced 2-3-fold by the presence of 100 M phosphate (data not shown).
Since CaMKIIPase was strongly activated by poly(Lys), the effects of various polycations and basic proteins on the enzyme activity were examined as shown in Fig. 3. Protamine, poly(Lys) (M r 87,000), poly(Lys) (M r 9,600), poly(Arg), mixed histones, and myelin basic protein activated CaMKIIPase at a concentration of 10 g/ml, the highest activation being caused by protamine, followed in the order described above. Poly(Glu, Lys,Tyr) 6:3:1 (10 g/ml) failed to activate the enzyme. L-Lysine (10 g/ml) and spermine (10 g/ml and 100 g/ml) could not substitute for poly(Lys) for the activation of the enzyme (data not shown).
To compare CaMKIIPase with other well known protein phosphatases, effects of various protein phosphatase inhibitors and divalent cations on the activity of CaMKIIPase were examined as shown in Table II. NaF, a common inhibitor of PP1, PP2A, and PP2B, but not PP2C, inhibited the activity of CaMKIIPase. Another common phosphatase inhibitor, ␤-glycerophosphate also inhibited the activity. Orthovanadate, a potent inhibitor of tyrosine phosphatase that also inhibits PP1, PP2A, and PP2C at a millimolar level, rather activated CaMKIIPase in the presence of poly(Lys), although the activation was not observed in the absence of poly(Lys) (data not shown). Heparin, which is known to inhibit PP1, inhibited CaMKIIPase. EDTA completely inhibited the enzyme activity, presumably by removing Mn 2ϩ essential for the activity. Inhibitor 2, a protein inhibitor specific for PP1, and okadaic acid and calyculin A, both of which are known to be potent inhibitors for PP1 and PP2A, did not significantly affect the activity of CaMKIIPase. W-7, a calmodulin antagonist, and Ca 2ϩ /calmodulin had no significant effect. ZnCl 2 , CoCl 2 , or NiCl 2 completely inhibited CaMKIIPase at a concentration of 2 mM, whereas 2 mM BaCl 2 had no effect. CaCl 2 poorly inhibited the activity at  (281-289), a synthetic phosphopeptide conjugate corresponding to residues 281-289 of autophosphorylated CaMKII, as a substrate for assay of CaMKIIPase. The important question of whether the purified enzyme can actually catalyze dephosphorylation of autophosphorylated CaMKII was examined as shown in Fig. 4. CaMKIIPase dephosphorylated autophosphorylated CaMKII in the presence of poly(Lys), and the dephosphorylation was almost completely blocked by the omission of poly(Lys). The dephosphorylation was not inhibited by calyculin A, but inhibited by NaF or EDTA, in good agreement with the results obtained with [ 32 P]PMC-KII(281-289) as a substrate (Table II). The dephosphorylation of autophosphorylated CaMKII by CaMKIIPase was also confirmed by trichloroacetic acid precipitation assay. When [ 32 P]autophosphorylated CaMKII (50 nM) was incubated with CaMKIIPase (64 ng/ml) in the presence of poly(Lys) (10 g/ml) and then precipitated with trichloroacetic acid, a rapid release of [ 32 P]ra-dioactivity into the supernatant was observed, while no significant radioactivity was recovered in the supernatant when poly(Lys) was omitted from the incubation mixture (data not shown). When much higher concentrations of CaMKIIPase was used in this assay, however, a slow but significant dephosphorylation was observed in the absence of poly(Lys) (data not shown), indicating that CaMKIIPase possessed a low dephosphorylating activity toward autophosphorylated CaMKII even in the absence of poly(Lys). The activation by 10 g/ml  Approximately 73 nM CaMKII, which had been autophosphorylated with [␥-32 P]ATP as described under "Experimental Procedures," was incubated at 30°C for 1 min as described under "SDS-PAGE Assay" under "Experimental Procedures," with the indicated additions and omissions, and aliquots were analyzed by SDS-PAGE, followed by autoradiography. The positions corresponding to ␣ and ␤ isoforms of CaMKII are indicated. poly(Lys) under the assay conditions was estimated to be approximately 90-fold.
Substrate Specificity of CaMKIIPase-The next important question of whether the action of CaMKIIPase is specific for CaMKII was examined with several phosphorylated proteins known to be dephosphorylated by various protein phosphatases as substrates (39 -41), as shown in Fig. 5A. None of the proteins tested, such as phosphorylase kinase, histones, MBP, and ␣-casein, which had been phosphorylated by PKA, and phosphorylase a, which had been phosphorylated by phosphorylase kinase, was significantly dephosphorylated by CaMKIIPase, while simultaneously added CaMKII which had been autophosphorylated was strongly dephosphorylated, suggesting that CaMKIIPase might be a specific protein phosphatase for autophosphorylated CaMKII. No significant dephosphorylation of the phosphorylase kinase, histones, MBP, ␣-casein, and phosphorylase a by CaMKIIPase was also observed by trichloroacetic acid precipitation assay (data not shown).
To confirm the strict substrate specificity of CaMKIIPase, the dephosphorylation of endogenous proteins in the rat brain crude extract, which had been phosphorylated with [␥-32 P]ATP in the presence of CaMKII, by CaMKIIPase was examined by SDS-PAGE followed by autoradiography, as shown in Fig. 5B. Many proteins were phosphorylated under the experimental conditions, but, among them, only proteins corresponding to molecular weights of about 50,000 and 60,000, which were thought to be autophosphorylated CaMKII, were significantly dephosphorylated only in the presence of poly(Lys) (Fig. 5B,  lane 4). This indicates that only autophosphorylated CaMKII among many endogenous phosphorylated proteins in the brain extract was dephosphorylated by the action of CaMKIIPase. Similar dephosphorylation was also observed when exogenous CaMKIIPase was not added (lane 2), and the dephosphorylation was inhibited by the addition of EDTA (lane 5), presum-ably reflecting the action of endogenous CaMKIIPase. These results, taken together, suggest that CaMKIIPase may be a protein phosphatase highly specific for autophosphorylated CaMKII.
Decrease of the Increased Ca 2ϩ /Calmodulin-independent Activity of Autophosphorylated CaMKII by CaMKIIPase-Since autophosphorylation of CaMKII at Thr 286 results in generation of Ca 2ϩ /calmodulin-independent activity (13)(14)(15)(16)(17) and full activation of the total activity (18 -20), dephosphorylation of the autophosphorylated CaMKII by CaMKIIPase should restore the activated activities to the original levels. The facts that autophosphorylated CaMKII is much more labile than nonphosphorylated CaMKII (24, 31) and 10 g/ml poly(Lys) markedly labilized autophosphorylated CaMKII 2 made it difficult to measure changes in the activity of CaMKII produced by dephosphorylation by CaMKIIPase at 30°C. Therefore, changes in the activity of autophosphorylated CaMKII by CaMKIIPase were studied at 5°C and at a low concentration of poly(Lys) (1 g/ml), as shown in Fig. 6. When autophosphorylated CaMKII was incubated with or without 1 g/ml poly(Lys) in the absence of CaMKIIPase at 5°C for 60 min, no significant changes in the autonomy were observed. On the other hand, incubation of autophosphorylated CaMKII with CaMKIIPase in the absence of poly(Lys) caused a relatively small but significant decrease in the autonomy. This decrease in the autonomy is probably due to the low basal activity of CaMKIIPase in the absence of poly(Lys), since a relatively high concentration of CaMKIIPase was used in this experiment. In contrast, the incubation with  1 and 2), mixed histones (lanes 3 and 4), phosphorylase a (lanes 5 and 6), MBP (lanes 7 and 8), and ␣-casein (lanes 9 and 10), which had been phosphorylated with [␥-32 P]ATP as described under "Experimental Procedures," were incubated with (lanes 2, 4, 6, 8, and 10) or without (lanes 1, 3, 5, 7, and 9) 1.6 g/ml CaMKIIPase and 10 g/ml poly(Lys) in the presence of autophosphorylated CaMKII for 1 min at 30°C, and aliquots were analyzed by SDS-PAGE, followed by autoradiography. B, a crude extract of rat cerebral cortex was phosphorylated with [␥-32 P]ATP by CaMKII as described under "Experimental Procedures," and aliquots amounting to 13.5 nM [ 32 P]phosphate incorporated were incubated with the indicated additions and omissions for 1 min at 30°C and analyzed as described above. The concentration of calyculin A brought into the reaction mixture from the phosphorylated sample was calculated to be 20 nM. The positions corresponding to ␣ and ␤ isoforms of CaMKII are indicated.
FIG. 6. The decrease in the increased autonomy of autophosphorylated CaMKII by CaMKIIPase. CaMKII, which had been autophosphorylated with nonradioactive ATP, was incubated at 5°C for 0 or 60 min in a reaction mixture containing 50 mM Tris-HCl (pH 7.5), 100 mM KCl, 2 mM MgCl 2 , 2 mM MnCl 2 , 0.1 mM EGTA, and 0.01% Tween 20, in the presence or absence of 12.8 g/ml CaMKIIPase and/or 1 g/ml poly(Lys), as indicated. Prior to addition of the autophosphorylated CaMKII, the mixture was preincubated for 30 s, and the reaction was started by adding the CaMKII. After incubation for 60 min, a 5-l aliquot was withdrawn and mixed with 20 l of an ice-cold stop buffer consisting of 50 mM Tris-HCl (pH 7.5), 2 mM EDTA, 0.05% Tween 40, and 1 mM DTT, and immediately thereafter the activity of CaMKII was determined in the presence and absence of CaCl 2 as described previously (30). For 0-min incubation, the stop buffer was added before addition of CaMKII. The results are expressed as autonomy, i.e. the ratio (%) of the activity in the absence of Ca 2ϩ to that in its presence. The total activities of the kinase expressed as a percentage of the initial value after 60 min of incubation were as follows: ϪCaMKIIPase and Ϫpoly(Lys), 91.9 Ϯ 7.0%; ϪCaMKIIPase and ϩpoly(Lys), 32.1 Ϯ 1.7%; ϩCaMKIIPase and Ϫpoly(Lys), 88.0 Ϯ 1.1%; ϩCaMKIIPase and ϩpoly(Lys), 64.6 Ϯ 6.6%. Each value represents the average of three independent experiments Ϯ S.D.
CaMKIIPase in the presence of poly(Lys) caused a marked decrease in the autonomy. Thus, the Ca 2ϩ /calmodulin-independent activity of CaMKII generated by autophosphorylation at Thr 286 was reversed by incubation with CaMKIIPase in the presence of poly(Lys). When the dephosphorylated kinase was assayed after brief incubation (30°C, 10 s) under the autophosphorylating conditions, generation of Ca 2ϩ /calmodulin-independent activity was observed again (44% of the total activity), suggesting that the dephosphorylation process is reversible. DISCUSSION Since CaMKII, which is involved in controlling a variety of neuronal functions, is markedly activated through autophosphorylation at Thr 286 (2)(3)(4)(13)(14)(15)(16)(17)(18)(19)(20)(21)(22), the dephosphorylation of the Thr 286 is thought to be very important for the regulation of the activity of CaMKII. Recently, the existence of protein phosphatases catalyzing the dephosphorylation of the Thr 286 in the rat brain was suggested by in-gel protein phosphatase assay using a synthetic peptide CaMKII(281-289) (MHRQETVDC) corresponding to the autophosphorylation site of CaMKII (residues 281-289) as a substrate (28). In the present study, we purified and characterized a novel protein phosphatase, designated CaMKIIPase, catalyzing the dephosphorylation of [ 32 P]PMC-KII(281-289), a phosphopeptide conjugate composed of phosphorylated CaMKII(281-289), poly(Lys) and paramagnetic particles. Table III compares some catalytic properties of CaMKIIPase with those of other well known protein Ser/Thr phosphatases, such as PP1, PP2A, PP2B, and PP2C. The most striking characteristics of CaMKIIPase are the activation by polycations and the requirement for Mn 2ϩ . The polycation activation has so far been considered as a characteristic of PP2A (39,40). As judged by requirements for activity and sensitivities to various inhibitors summarized in Table III, CaMK-IIPase is thought to be a protein phosphatase distinct from any of the four well known protein phosphatases. Several protein Ser/Thr phosphatases, which differ in their biochemical properties from PP1, PP2A, PP2B, and PP2C, have recently been isolated from mammalian tissues (39,(41)(42)(43), but no reports describing the properties similar to those of CaMKIIPase have appeared.
It has been reported that some protein phosphatases having broad substrate specificities, such as PP1 (23), PP2A (14,16,24,25), and PP2C (26), can dephosphorylate Thr 286 of autophosphorylated CaMKII in vitro. In contrast to these multifunctional protein phosphatases, CaMKIIPase dephosphorylated autophosphorylated CaMKII but none of the other phosphorylated proteins examined, such as phosphorylase kinase, histones, MBP, ␣-casein, and phosphorylase a (Fig. 5A), and dephosphorylated no other proteins than CaMKII in the rat brain extract (Fig. 5B), suggesting that CaMKIIPase might be a protein phosphatase specific for autophosphorylated CaMKII, in good agreement with our previous finding (28) that the band corresponding to CaMKIIPase detected by the in-gel protein phosphatase assay was observed when CaMKII(281-289) was used as a substrate but not observed when other phosphopeptides such as C-syntide-2 and CAMKAKS peptide were used as substrates. An apparent K m value of CaMKIIPase for autophosphorylated CaMKII estimated from double-reciprocal plots of the rates of the dephosphorylation determined by trichloroacetic acid precipitation assay as functions of the concentrations of CaMKII was in the range of 30 -80 nM (data not shown). This value is much lower than the K m values of PP2A ranging from 1 to 100 M obtained for various substrates (44) or the K m value of PP2C from turkey gizzard smooth muscle of 7.9 M estimated for myosin light chain (45), suggesting that autophosphorylated CaMKII is a good specific substrate for CaMKIIPase.
Since the autophosphorylation of CaMKII at Thr 286 is known to generate the Ca 2ϩ /calmodulin-independent activity, the dephosphorylation of the phosphorylated Thr 286 should result in a decrease in the activity. However, autophosphorylated CaMKII was very labile to poly(Lys) added for the reaction of CaMKIIPase, and therefore the reverse of the activity due to the dephosphorylation of Thr 286 of CaMKII was monitored by a decrease in the autonomy of the activity of CaMKII (Fig. 6). A maximum decrease in the autonomy was observed after incubation with CaMKIIPase in the presence of poly(Lys), suggesting that CaMKIIPase dephosphorylated the autophosphorylated Thr 286 of CaMKII, thereby resulting in a decrease in the autonomy. The results, taken together, suggest the possibility that CaMKIIPase purified in the present study is a specialized protein phosphatase for the regulation of CaMKII.