Phosphoprotein Phosphatases from Rat Cerebral Cortex I~ISTRIBIJTION ANI> CHARACTERIZATION*

The subcellular distribution of phosphoprotein phosphatases which release orthophosphate from phosphoprotein was in rat cerebral

10, 1971) HIROO Nb~xo AND ~'AUL GREENGARD &ml the Deprtnlent of Pharmacology, Yale I;niversity School of Medicine, New Haven, Connecticut 0651G SUMMARY The subcellular distribution of phosphoprotein phosphatases which release orthophosphate from phosphoprotein was studied in rat cerebral cortex.
In contrast to several other tissues examined, more than 50% of the total protein phosphatase activity in rat cerebral cortex was found in the particulate fractions; the activity was especially high in the crude mitochondrial fraction. Further subfractionation of the crude mitochondrial fraction by sucrose density gradient centrifugation showed that, among the membrane fractions, the specific activity of protein phosphatase was highest in the fractions rich in synaptic membranes and lowest in the mitochondria.
A considerable amount of the enzyme activity in synaptic and microsomal membranes existed in a latent form which could be partially unmasked by treatment with Triton X-100.
The specific activity of the protein phosphatase of the cell sap and of the synaptoplasm was considerably higher than that of the membrane fractions. Column chromatography on DEAE-cellulose resolved the protein phosphatase activity of the cell sap into three distinct protein phosphatases, which were clearly distinguished from membrane enzyme by differences in substrate specificity and metal ion requirements.
The soluble protein phosphatases, but not the enzyme released from the synaptic plasma membrane fraction by Triton X-100 treatment, were specifically activated by manganese chloride. Endogenous membrane protein was found to be the best among several phosphorylated proteins examined as substrates for membrane-bound protein phosphatase.
Both membrane-bound and soluble protein phosphatase exhibited pH optima in the neutral range.
Protein phosphatase catalyzed the stoichiometric release of orthophosphate from a phosphoserine residue of protamine.
This was the only amino acid residue in protamine and histones which appeared to be phosphorylated by an adenosine 3',5'-monophosphate-dependent protein kinase purified from bovine brain.
Studies of the subcellular distribution in cerebral cortex of the enzymes related to the metabolism and fullctioll of cyclic  (1)) cyclic ,4MP* phosphodiesterase (l-3) and cyclic AMP-dependent protein kinase (4), have shown them to be considerably enriched in the synaptic membrane fractions. Furthermore, the subcellular distribution of endogenous proteins able to act as substrates for a partially purified cyclic AMP-dependent protein kinase from brain was found to parallel the distribution of endogenous cyclic AMP-dependent protein kinase activity (5). The localization in synaptic membrane fractions of these various enzymes and protein substrates of the cyclic AMP system, along with several other lines of evidence (e.g. 6-8), suggests an intimate involvement of the cyclic AMP system in the process of transmission at certain neural synapses.
Following the discovery of cyclic AMP-dependent ljrotein kinases, initially in muscle (9), and then in liver (10) and brain (ll), the hypothesis was suggested (6,(12)(13)(14) that the diverse biochemical and physiological effects of cyclic AMP may be mediated through regulation of cyclic SMPdependent protein kinase activity, the specificity of the action of the cyclic nucleotide residing in the specificity of the enzyme and of its substrates in the various tissues. The dephosphorylation of the phosphoprotein products of protein kinase activity can, within this framework, be expected to be of importance in regulating the magnitude and duration of the effects of cyclic AMP.
It, therefore, seemed of considerable importance, in evaluating the role of cyclic AMP in neural function, to study the subcellular location and properties of phosphoprotein phosphatase of neural tissue.
Phosphoprotein phosphatase capable of releasing phosphate from phosphoprotein has been reported in several tissues (15-21), including mammalian brain (18,19,21). In one study (20), a liver phosphoprotein phosphatase was found which is apparently specific for histone and other basic proteins.
The present report deals with t,he subcellular distribution, hcterogeneity, and some enzymological properties of protein phos- The protein phosphatase of cell sap and synaptoplasm (11-3) was further purified by chromatography on a DEAE-cellulose column (2.2 x 11 cm), which had previously been equilibrated with 0.01 M Tris-HCl, pH 7.2. Protein phosphatase was eluted with a 300-ml linear gradient of NaCl (0 to 0.7 M) containing 0.01 M Tris-HCl, pH 7.2, and 1 mrvr dithiothreitol.
Cyclic AMP-dependent protein kinase and succinic dehydrogenase were assayed as described previously (4). Protein was determined by the method of Lowry et al. (23) with bovine serum albumin as standard.
To prepare substrates for protein phosphatase, protamine and various histones were phosphorylated by incubating with [Y-~~P],~TP in the presence of cyclic AMP-dependent protein kinase, the kinase having been partially purified from bovine brain by column chromatography on DEAE-cellulose according to the method of Miyamoto et al. (11). To prepare phosphorylated substrates, 1 ml of incubation mixture contained 46 pg of protein kinase; 50 pmoles of sodium acetate buffer, pH 6.4; 1 mg of protamine or histone; 5.0 nmoles of [y-32P]ATP (5 to 10 X lo6 cpm) ; 10 pmoles of magnesium acetate; 10 pmoles of sodium fluoride; 2.0 pmoles of theophylline; 0.3 pmole of ethylene glycol bis(@-aminoethyl ether)-N , N'-tetraacetic acid; and 5.0 nmoles of cyclic A;\'lP. The incubation was carried out at 30" for 45 min. The phosphorylation reaction was terminated by the addition of 0.25 ml of 100% trichloroacetic acid. The resulting precipitate was centrifuged, washed twice by dissolving in water and reprecipitating with 20% trichloroacetic acid, and then dialyzed against distilled water for 24 hours.
Protein phosphatase activity, with protamine or histone as substrate, was assayed by measuring the release of radioactive orthophosphate from 32P-labeled protein.
For routine assays, the reaction mixture contained, in a total volume of 0.1 ml, 10 pmoles of Tris-HCl buffer, pH 7.2, 0.1 pmole of dithiothreitol, 100 pg of 32P-labeled protamine (containing 5 to 10 X lo5 cgm), and 1 to 20 pg of enzyme protein.
The incubation was performed at 30" for 10 min and terminated by the addition of 0.4 ml of 259;) trichloroacetic acid. After the addition of 0.1 ml of of 0.625c/, bovine serum albumin as a carrier for the precipitation, protein was removed by rentrifugation.
Orthophosphate was extracted from the deproteinized supernatant by a modification of the method of Plaut (24). To 0.4 ml of deproteinized supernatant were added 0.05 ml of lop2 nf KHzPOd and 0.15 ml of 57; ammonium molybdate. The resulting phosphomolybdate complex was extracted with 1.0 ml of isobutyl alcohol and the radioactivity of the isobutyl alcohol extract was measured.
The amount of protein labeled with 321' and the amount of phosphate released from this substrate protein have been calculated from the specific activity of the radioactive ATP used as precursor in the protein phosphorylation reaction and neglecting the phosphate present in the protein substrate prior to its phosphorylation by radioactive ATP.
One unit of protein phosphatase activity was defined as 1 pmole of 32P released under the above assay conditions.
In order to identify the site of phosphate incorporation, 3%labeled proteins were hydrolyzed in 6 N HCl at 108" in sealed ampoules for 6 hours. The hydrolysates were dried in uacuo over KOH and then subjected to high voltage electrophoresis on Whatman No. 3MM paper at pH 1.8 in formicacetic acid buffer for 80 min (5, 25). After electrophoresis, the paper was dried and stained with ninhydrin to locate the amino acids. The paper was cut into l-cm strips and the radioactivity of the strips was measured by liquid scintillation spectrometry.
Authentic phosphoserine and phosphothreonine were hydrolyzed under the same conditions in order to correct for their decomposition durin g acid hydrolysis; the amount of orthophosphate formed during hydrolysis was determined calorimetrically according to Rockstein and Herron (26).

RESULTS
In view of the possible existence of multiple protein phosphatases with different substrate specificities, two different proteins, protamine and arginine-rich histone, were used as substrates in studying the distribution of activity in cerebral cortical subfractions. Table  I depicts the distribution of protein phosphatase activity among the four primary subfractions.
The crude mitochondrial fraction marked by the highest activity of succinic dehydrogenase and the high speed supernatant fraction (cell sap) each accounted for about 40y0 of the total activity.
The specific activity of the phosphatase in the cell sap was much higher than that from any of the particulate fractions.
Cyclic AMP-dependent protein kinase activity was also measured in these experimenbs; no significant difference in subcellular distribution of ljrotein kinase was observed between the present studies, in which cerebral cortex was used, and the earlier studies (4) in which whole cerebrum was used. Differences in the ratio of protamine phosphatase activity to arginine-rich histone phosphatase activity among the different fractions suggested the existence of more than one type of protein phosphatase in rat cerebral cortex.
After osmotic shock of the crude mitochondrial fraction, about 40% of the protamine phosphatase activity remained in the particulate fraction, M-l, which contains synaptic membranes and mitochondria.
The soluble fraction (aynaptoplasm), M3, contained a somewhat higher portion (567,) of the total activity. Arginine-rich histone phosphatase activity, on the other hand, was found slightly but reproducibly higher in the Ml fraction than in the M-3 fraction (Table II).
The fraction enriched in synaptic vesicles, RI-2, exhibited very little of the total activity seen with either substrate.
The specific activity of the phosphatase in the synaptoplasm was much higher than that in the Ml or M-2 particulate mitochondrial subfractions. It is of interest that the specific activity of protamine phosphatase was slightly higher and that of arginine-rich histone phosphatase was slightly lower in the synaptoplasm than in the cell sap. The ratio of the two activities, accordingly, differed markedly in these two soluble fractions, synaptoplasm being relatively richer in protamine phosphatase activity. These differences between the enzymes from synaptoplasm and cell sap were not large, but they were consistent. Further fractionation of the crude mitochondrial subfraction, hi-l, by discontinuous sucrose density gradient centrifugation and identificat.ion of the subfractions were carried out as described previously (4). The data indicate (a) that the specific activity of protein phosphatase for both of the substrates was considerably lower in the mitochondrial subfraction than in the fractions rich in myelin and synaptic membranes, (b) that the M-l (1 .O) subfraction, which is enriched with synaptic membrane fragments, had the highest specific activity of protein phosphatase for both substrates, and (c) that the ratio of protamine phosphatase to arginine-rich histone phosphatase activity varied considerably among the subfractions of M-l. However, the recovery of enzyme activity from the sucrose gra.dient was low, ranging from 40 to 45?;, with protamine as substrate and from 55 to 60% with arginine-rich histone as subst.rate in three experiments, in contrast to a recovery from t,he gradient of about 75% (4) for cyclic AMP-dependent protein kinase activity, so that the pattern of distribution of protein phosphatase might have been altered by a selective loss of activity in some fractions.
Therefore, data on the distribution of activity in subfractions of M-1 are not presented.
In some experiments, fractionation of the crude mitochondrial Tanks III Solubilization of protein phosphatase from particulate fractions by Triton X-100 The indicated amounts of part,iculat.e fractions were incubated with 0.1% Triton X-100 for 30 min in ice and then cent.rifuged at 150,000 X 9 for 60 min in a Beckman ultracentrifuge.
The precipitates were sllspended mechanically in 0.01 M Tris-HCl buffer, pH 7.2, by means of a glass homogenizer. were present in the 0.32 M sucrose used for resuspension of M-l and in the sucrose density gradient used for fractionation. Prot.amine phosphatase activity was then measured in the subfractions. Under these conditions, virtually complete recovery of enzyme activity after sucrose density gradient centrifugation was achieved.
Again, the specific activity of the M-l (1.0) subfraction was the highest and the specific activity of the mitochondrial pellet (identified by its high relative specific activity for succinic dehydrogenase) was the lowest of the various subfractions (data not shown). However, the distribution of protein was changed under these conditions.
The alt,eration in prot.ein fractionation pattern observed in the presence of the Tris buffer and dithiothreitol makes it difficult to compare such dat,a with other data presently available in the literature; therefore, such studies were not pursued in the present investigation, in spite of the higher recovery of enzyme.
In order to determine whether particulate-bound protein phosphatase could be solubilized, the microsomal fraction, the synaptic vesicle (M-2) fraction, and a synaptic membrane (RL1 (1.0)) fraction were treated with 0.1% Triton X-100. Triton X-100, at this concentration, exhibited no inhibitory effect on cell sap protein phosphatase.
It is evident from a comparison of protamine phosphatase activity in untreated fractions with activity in Triton X-loo-treated fractions that large amounts of protamine phosphatase in the microsomal and the synaptic membrane fraction exist in a latent form (Table  III).
The M-2 fraction showed no latent activity, although most of the activity was found in the soluble fraction after treatment with the detergent.
It is of interest that after Triton X-100 treatment of the microsomal or the synaptic membrane fraction, the precipitate showed higher specific activity than the corresponding fraction did prior to the treatment,  Combined fractions of tubes 50 to 54, 57 to 60, and 65 to 67 shown in Fig. 1 were designated cell sap fractions I, IT, and III, respect,ively.
The data for the particulate fractions were obtained with Triton-soluble enzvmes prepared from these particulate fractions, as described in ?'ableIIi. suggesting that the enzyme still remaining in the membranes after treatment with the detergent may have become more accessible to substrate. In order to see whether the cell sap protein phosphatase activity could be resolved into more than one component, cell sap was subjected to column chromatography on DEXEcellulose (Fig. 1). Over-all recoveries of protamine and argininerich histone phosphatase activity from this step were 72 and 950/,, respectively.
When protamine was used as substrate, three distinct loci of activity were invariably observed; in the experiment illust,rated, peaks appeared at tube 51 (Fraction I) and at tube 58 (Fraction II), with a shoulder around tube 64 (Fraction III). Arginine~rich histone phosphatase activity showed two clear components; in the experiment shown in Fig. 1, these were manifested by a shoulder at tube 58 and a peak at tube 64. As shown in Table IV, fractions representing the three activity peaks were clearly distinguished from one another by the difference in their relative V,,,, values for dephosphorylation of prot.amine and arginine-rich histone, although apl)arent II, values for either prootamine or arginine-rich histone did not d&r significantly among t,hese three fractions. Chromatogralhy of synaptoplasm (AI-3) on DEAE-cellulose gave an elution pattern in which the position of the peaks was similar to that of crll sap for both types of phosphatase activity (data not shown). hforeover, the K,, for protamine, the K, for argin:ne-rich histone, and the Vmax (protaminr)/ V max (arginine) ratio of the three enzyme fractions from synaptoplasm were similar to t,hose of the corresponding fractions from cell sap. Interestillgly, the synaptoplnsm differed from the cell sap in the relative nmoullts of the three enzyme fractions.
The enzymes which had been solubilizcd from the various particulate fractions resembled one anot,her with respect to t,he values for each of t,hesc t,hree kill&c parameters (Table Iv). It is of considerable interest that the enzymes solubilized from the particulate fractions are distinguishable from t,he enzymes of the cell sap by having a much higher affinity for protamine. The combination of enzyme solubilized from any of the particulate fractions with any of the t,hree cell sap enzymes eshibited only additive effects over a wide range of protamine concentrations.
The latter results make it very unl,kely that  it.ate was washed twice with trichloroaeetie acid-tungstate-HzS04, and then washed, in addition, two times with 4.0 ml of chloroform-methanol (2:1), as described previously (5). The washed precipit,ate was dissolved in 0.05 ml of 0.5 N NaOH and the radioactivity was measured.
In other experiments, it was shown that treatment of the precipitate with hydroxylamine, which preferentially hydrolyzes acyl bonds, removed no radioactivity from the precipitate. any loosely binding activators or inhibitors might be responsible for t.he difference in affinity of the various enzymes for prot.amine.
The ability of protein phosphat.ase in the synapt.ic membrane fractions to dephosphorylate prot,amine and hi&ones was compared wit,h the abi1it.y to dephosphorylate phosphoprot.eins intrinsic to the synaptic membrane fractions.
For the measurement of intrinsic substrate activity, the synaptic membrane fractions were first phosphorylated with [rJ2P]ATP by the intrinsic protein kinase in the presence of cyclic AX?.
As shown in Fig. 2 for t'he M-l (1.0) fraction, phosphorylation reached a steady stat,e level within a few minutes.
LLt this point, EDTA, plus a large amount of nonradioactive ATP, was added to stop the incorporation of 32P into the membrane protein.
(EDTA inhibits t,he protein kinase by about 60 to TO%, uuder the conditions used. Therefore, nonradioactive ATP was added to dilute t,he [yJ2P]ATP aud minimize further incorporation of 321' into the protein.) Immediat.ely after the addition of these reagents, rapid dephosphorylation due to intrinsic protein phosphatase activity could be observed. The time curve of the removal of phosphate was biphasic; about half of the phosphate was removed in the first, more rapid phase, and the remainder more slowly.
In order to evaluate the possibility that the release of phosphate was due to reversal of the kinase-cat'alyzed reaction, i.e. transfer of phosphoprotein (0.9) and iC1 (l.O), were dephosphorylated by the intrinsic prot'ein phosphatase much more rapidly than were any of the exogenous substrates tested. The microsomal fraction, as another example of a membranous material, was also compared with various other proteins as potential substrate for its intrinsic protein phosphatase ( Table  V). The microsomal prot'ein was the best among the proteins tested as substrate for the microsomal protein phosphatase, although its substrate activit,y, relative to the other proteins tested as substrate, was not as striking as in the analogous case of the endogenous substrate activity of t,he synaptic membrane fractions. The data of Fig. 2 and Table V indicate a high turnover of phosphate in membrane proteins, particularly in the subcellular fractions rich ill synaptic membranes.
It is important to emphasize that the direct, quantitative comparison of rndogenous with exogenous protein substrates suffers the uncertainty of meaning of t,he term "concentration" with respect to membrane proteins in t,he semisolid membrane phase. The differences found may be attributable to a greater accessibility of intrinsic membrane substrates than of exogenous substrates to the membranebound phosyhatasc.
Nevertheless, the result,s do indicate unequivocally that these neuronal membranes contain lxotein phosphatases which are very effective in removing phosphate from those proteins in the membrane which have been phosphorylated by protein kinases.
The stoichiometry of the protein phosphatase reaction was examined by measuring the radioactivity of the trichloroacetic acid-insoluble protein and of the inorganic phosphate released The M-l (1.0) enzyme had been solubilized with 0.1% Triton X-100 as described in TabIe III.
Activity is expressed as the percentage of that observed in t.he absence of added metal ions. The incubation mixture cou&ed 100 pg of protamine (with 420 pmoles of 3*P) and 9.6 pg of enzyme, solubilized from the M-l (1.0) fraction by treatment with 0.15; Triton X-100. as described in Table III. Incubatiou conditions were as described uuder "Experimental Procedure," except for the variation of pH of the reaction mixture, in which 0.1 hl acetate buffer was used for pH 5 to 7 and 0.1 M Tris-HCl buffer was used Aor pH 7 to 9. during incubation.
After 60-min incubation of solubilized membrane enzyme or of cell sap enzyme with [32P]protamine as substrate, the sum of trichloroacetic acid-insoluble phosphate and inorganic phosphate released from the substrate accounted for all radioactivity in the substrate (Table VI).
No significant radioactivity was detected in the trichloroaeetic acidsoluble fraction after removal of inorganic phosphate (by estraction of the phosphomolybdate complex into isobutyl alcohol).
These results indicate that the protein phosphatase activity is not to be attributed to the degradation by a proteolytic enzyme of substrate phosphoprotein into low molecular weight. peptides or amino acids, followed by the action of a phosphoserine or phosphothreonine phosphatase. Since it is important in the characterization of the protein phosphatases to know the site(s) of phosphorylation in t,he substrate, the Y?labeled protamine used as substrate was hydrolyzed in 6 s HCI at 108" and subject.ed to electrophoresis on paper at high voltage.
After correction for hydrolysis of phosphoserine and phosphothreonine, 917; of the total activity of the acid digest was recovered as phosphoserine and no significant radioactivit,y was detected in t,he phosphothreonine spot. These results indicate that the cyclic Xi\lP-dependent protein kinase from bovine brain had phosphorylated serine residues of protamine primarily or exclusively, and that the various prot.amine phosphatases were capable of slrlitting phosphate from these residues.
In fact, the protein phosl)hatases hydrolyzed the phosphoserine residues of protamine lvith nearly quantitative release of inorganic phosphate, as delricted in Table VI. In other experiments, in which cyclic AMPdependent protein kinase from bovine brain had been used to phosphorylate arginine-rich histone, lgsine-rich histone, and histone mixture, the amino acid residues ~vhich 1la.d been phosphorylated were identified as serine too.
The three cell sap enzymes and the solubilizcd fractions from microsomes, synaptic vesicles (M2), and synaptic membranes (11-l (1.0 4. Effect of manganese chloride concentratiou ou proteiu phosphatase activity. The amouut of the various enzymes prescut in the iucubatiou mixture was as described in Table VII. Arginine-rich histone, 100 hg (containing 420 pmoles of SZP), was used as substrate.
Other incubation conditions were as described under "JSxperimental Procedure." reaction rate at neutral pH. As an example, Fig. 3 illustrates the enzyme activity as a function of pH for t,he solubilized fraction from synaptic membranes. The liver histone ljhosphatase studied by Meisler and Langan (20) and the brain membrane-bound protein phosphatase studied by Weller and Rodnight (21) showed a similar pH dependence.
All of the protein phosphatases from cerebral cortex were inhibited approximately 500/, by lop2 M NaF, as well as by lo-* M orthophosphate, but u-ere unaffected by any of the These results make it seem unlikely that these protein phosphatases from cerebral cortex are identical with any of the known 5'-nucleotidases, "nonspecific" phosphatases, AYIPases, glucose R-phosphatases, or cyclic nucleotide phosphodiesterases.
The effect of various ions on the activities of membrane and cell sap protein phosphatases is shown in Table VII. At the concentration tested (2.5 IIIM), MgClz and CaClz exhibited no significant, effect on any of the four enzymes.
CoC& strongly inhibited the activity of the Triton-solubilized enzyme from the M-l (1.0) fraction and the activity of cell sap Fraction I and slightly inhibited that from Fraction II, but did not inhibit that from Fraction III. ZnSO1, CuCL, and FeC& inhibited almost completely the activity of all fractions.
Interestingly, MnC12 markedly stimulated Fractions I and II but had a lesser effert, on Fraction III and no significant effect on the solubilized membrane eneymc.
The effect of varying concentrations of JInClz on each of these four enzymes is presented in Fig. 4.
Proteiil phosphatase activity, measured with histone as substrate, showed a pattern of subcellular distribution in nonneural tissue quite different from that observed with neural tissue (Table VIII).
Thus, the protein phosphat'ase activity of particulate fractions, relative to the total activity of the particulate fraction plus cell sap, was much higher in brain (46 to 547~ for cerebral cortex, cerebellum, caudate nucleus, and medulla) thaii iii nonnervous tissues (18 to 29% for liver, lung, heart, kidney, and spleen).
Recovery of phosphatase activity in the particulate fraction plus cell sap was approximately lOOy& for all tissues studied.

DISCUSSION
The subcellular distribution of protein phosphat'ase in rat cerebral cortex, found in the present study with phosphorylated prot,amine and histone as substrates, appears rat'her similar to the distribution of cyclic AMP I)llosl)llodiesterase i,eportcd by DeRobertis et al. (1). The results of studies of protein phosphatase distribution can be expected to vary, however, dcpending on the nature of the substrate used for the assay. For example, since membrane protein from the WI-1 (1.0) fraction was a far better substrate for the protein phosphatasc of L/I-l (1.0) than any other protein tested, as indicated in Table V, it can be expected that the observed subcellular distribution of the enzyme activity might be greatly changed if this membrane protein could be isolated and used as substrate in t,he dist,ribution studies.
As in the case of cyclic AMP-dependent protein kinase (4), protein phosphatase of membrane fractions is masked to a great extent and, therefore, it seemed desirable to determine the enzyme activity in the presence of the nonionic detergent, Triton X-100, which uncovered much latent activity. The low protein phosphatase activity in the particulate fractions, relative to that of the soluble, reported by Rose in the guinea pig brain (19), may be explained by the absence of unmasking detergent in his assay system. The subcellular dist'ribution of protein phosphatase in guinea pig brain (data not shown), under our assay conditions, was found to be very similar to that reported here for rat brain.
It was shown previously that the synaptic membrane fractions are enriched with respect to cyclic AMP-dependent protein kinase activity (4). The association of protein phosphatase along with cyclic AMP-dependent protein kinaae in these synaptic membrane fractions suggests an active turnover of phosphate in the membrane protein.
Indeed, a rapid turnover of phosphate was observed in U&O, particularly in t,he fractions rich in synaptic plasma membranes, as is illustrated in Fig. 2. Since the completion of these studies, Weller and Rodnight (21) have presented evidence for the turnover of protein-bound phosphorylserine phosphate in membrane preparations from ox brain cortex.
Viewed within the conceptual framework (6, 12-14) that the major biochemical action of cyclic AMP is to stimulate cyclic AMP-dependent protein kinase activity and thereby produce an increase of certain key phosphoproteins, the protein phosphatase can be regarded as a means of terminating the act'ion of cyclic AMP by catalyzing removal of the phosphate group. It is important to emphasize that the presence in the plasma membrane fractions of high concentrations of adenyl cyclase (I), cyclic nucleotide phosphodiesterase (l-3)) cyclic AMPdependent protein kinase (4), substrate for the protein kinase (5), and protein phosphatase appears to be rather unique to brain tissue, supporting the hypothesis (336, 8) that the alterations in excitability of the plasma cell membrane, caused by cyclic *11LIP, result from altering the state of phosphorylation of the membrane protein.
Cell sap protein phosphatase activity was resolved into three fractions by column chromatography on DEAF-cellulose. These fractions could be clearly distinguished by the difference in relative rates of dephosphorylation of protamine and of arginine-rich histone. Moreover, the protein phosphatase of the particulate fractions appears to be different from any of these soluble enzymes, as indicated by the K, values for protamine and arginine-rich histone, as well as by the effect of MnC&. Thus, rat cerebral cortex would appear to contain at least four distinct protein phosphatases. The demonstration in the present study of multiple forms