Purification and Characterization of a Novel Protein Phosphatase

THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1989 by Tho American Society for Biochemistry and Molecular Biology, Inc.

VOl. 264, No. Issue of January 5, pp. 151-156,1983 Printed in U.S.A.

Purification and Characterization of a Novel Protein Phosphatase Highly Specific for Ribosomal Protein S6*

(Received for publication, July 7 , 1988)

Janet L. Andres and James L. MallerS From the Department of Pharmacology, University of Colorado School of Medicine, Denver, Colorado 80262

Ribosomal protein 56 is the principal phosphoprotein of the eucaryotic ribosome that becomes multiply phosphorylated on serine residues in response to a wide variety of mitogenic stimuli. In this paper the principal protein phosphatases able to dephosphorylate S6 were characterized in Xenopus laevis ovary and eggs. Two enzymes termed peak I and peak I1 were found to account for most 56 phosphatase activity in both oocytes and eggs. The peak I enzyme had an apparent M, of 200,000 on gel filtration, dephosphorylated the /3 subunit of phosphorylase kinase and phosphorylase a, and was inhibited by inhibitor 1 and inhibitor 2, suggesting it was similar to protein phosphatase 1. The peak I1 enzyme was purified over 12,000-fold and had an apparent M. = 66,000 on glycerol gradient centrifugation. This phosphatase could dephosphorylate all sites in 56 but was unable to dephosphorylate phosphorylase a or phosphorylase kinase. However, it was inhibited by nanomolar concentrations of inhibitor 1 and inhibitor 2. These results indicate the peak I1 enzyme represents a new class of highly specific protein phosphatase and suggest that inhibition of dephosphorylation in cellular extracts by inhibitor 1 and inhibitor 2 is not a sufficient criterion for implicating protein phosphatase 1 in a cellular process.

Regulation of the phosphorylation of ribosomal protein S6 has been the subject of extensive investigation in recent years. S6, a component of the 40 S ribosomal subunit, becomes phosphorylated on serine residues at a stoichiometry of up to 5 mol/mol in response to a number of mitogenic agents in many different cell types (1-9). Although the function of this modification in the control of protein synthesis is unclear, S6 phosphorylation has remained of interest because it is an easily measured proximal response to growth-promoting stimuli. Consequently, it can be used as a model system for elucidating the mechanisms involved in mitogenic stimulation. Most studies have focused on the protein kinases which phosphorylate S6. Several specific S6 kinases have been par- tially purified from cultured cells (10-15) and an S6 kinase activated during oocyte maturation has been purified to ho- mogeneity from unfertilized Xenopus eggs (16,17). In contrast to S6 protein kinases, relatively little attention has been given to the protein phosphatases which dephosphorylate S6, although the possibility that protein phosphatases may also be regulated has been suggested by studies in several laboratories (18-21).

* This work was supported by National Institutes of Health Grant DK28353. 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.

$ To whom correspondence should be addressed.

In general, less is known about protein phosphatases than about protein kinases. Most of what is known has come from studies on the enzymes involved in glycogen metabolism. It has been suggested by Cohen and co-workers (22-25) that four enzymes identified and purified in their studies can account for the dephosphorylation of most other substrates present in a variety of tissues. These enzymes have been classified into two types (see Ref. 24 for review). The type 1 protein phosphatase specifically dephosphorylates phosphorylase a and the (3 subunit of phosphorylase kinase and is inhibited by nanomolar concentrations of two heat-stable proteins called inhibitor 1 and inhibitor 2. The activity of protein phosphatase 1 is not dependent on the presence of any ions. There are three type 2 enzymes. These enzymes preferentially dephosphorylate the a subunit of phosphorylase kinase and are not affected by the two inhibitors. The first of these, protein phosphatase 2A, has no ion dependence and can also dephosphorylate phosphorylase a. Protein phosphatase 2B, also known as calcineurin, is a Ca2'/calmodulin- dependent enzyme with little or no activity against phosphorylase a. The third type 2 enzyme, protein phosphatase 2C, catalyzes Me-dependent dephosphorylation of both phosphorylase a and the a subunit of phosphorylase kinase.

Several potential mechanisms exist for the regulation of protein phosphatases. Inhibitor 1 is only active when phosphorylated by the CAMP-dependent protein kinase. Studies (25-28) have shown that the phosphorylation state of inhibitor 1 is altered by hormones acting in vivo, providing pre- sumptive evidence for regulation of protein phosphatase 1. In addition, microinjection of inhibitor 1 or inhibitor 2 into Xenopus l aev i s oocytes inhibits the dephosphorylation of co- injected phosphorylase a and delays the maturation of pro- gesterone-treated oocytes (29, 30), providing evidence that these proteins function as phosphatase inhibitors in vivo. The catalytic subunit of protein phosphatase 1 is phosphorylated in vitro on a tyrosine residue by pp6OV-"", the transforming protein of Rous sarcoma virus, inactivating the phosphatase (31). This finding provides a potential mechanism for the direct regulation of a protein phosphatase by an oncogene product. Finally, protein phosphatase 2B is a Ca2+/calmodulin-dependent enzyme and may be regulated by changes in calcium levels.

In a previous study, we investigated the nature of the phosphatases which affect S6. We reported (32) that protein phosphatase 1 and protein phosphatase 2B purified from rabbit skeletal muscle could function as S6 phosphatases in vitro while protein phosphatase 2A was unable to dephosphorylate S6. In addition, enzymes similar to these appeared to exist in Xenopus oocyte extracts as well as in living oocytes microinjected with labeled phosphatase substrates. In this paper we report the partial purification of the two enzymes responsible for the majority of the S6 phosphatase activity in Xenopus oocytes and eggs. One of these enzymes appears to

151

152 A Novel S6 Phosphatase

be a high molecular weight form of protein phosphatase 1. The second enzyme, which has been highly purified, resembles protein phosphatase 1 in that it is inhibited by nanomolar concentrations of inhibitor 1 and inhibitor 2, but it is unable to dephosphorylate phosphorylase a or phosphorylase kinase. It therefore represents a specific class of protein phosphatase not previously described.

EXPERIMENTAL PROCEDURES

Materiak-DEAE-Sephacel, SP-Sephadex, Sephacryl S-200, Mono S and Mono Q columns were obtained from Pharmacia LKB Biotechnology Inc. Proteins used as molecular weight standards for gel filtration were ferritin (Mr = 450,000), alcohol dehydrogenase (Mr = 145,000), ovalbumin (Mr = 42,000), and soybean trypsin inhibitor (M, = 20,100), all obtained from Boehringer Mannheim. Proteins used as molecular weight standards for gel electrophoresis were phosphorylase b (Mr = 97,400), bovine serum albumin (Mr = 67,000), ovalbumin ( M , = 43,000), carbonic anhydrase (M, = 30,000), and soybean trypsin inhibitor ( M , = 20,100) and were obtained from Pharmacia. Alcohol dehydrogenase, ovalbumin, and cytochrome c (Mr = 12,400) purchased from Boehringer Mannheim were used as molecular weight standards for glycerol gradients. DTT, pepstatin A, and leupeptin were purchased from Boehringer Mannheim. Unless otherwise specified, all other reagents were obtained from Sigma. Female X. laeuis were obtained from Xenopus I (Ann Arbor, MI). Ribosomes were prepared from ovaries of X. laeuis according to the method of Cox et al. (33), and 40 S ribosomal subunits were isolated as described previously (32). Homogeneous phosphorylase b (34), phosphorylase kinase (35), inhibitor 1 (35), inhibitor 2 (36), and the catalytic subunit of CAMP-dependent protein kinase (37) were isolated as previously described. Protein concentrations were measured by the method of Bradford. Polyacrylamide gels were run according to the method of Laemmli (38) and silver stained by the method of Oakley et al. (39).

Preparation of Phosphoprotein Sub~trates-~~P-Labeled phosphorylase a was prepared from phosphorylase b using phosphorylase kinase (40). 32P-Labeled phosphorylase kinase containing equal amounts of phosphate in the a and 0 subunits was prepared using the catalytic subunit of CAMP-dependent protein kinase (41). 40 S subunits were phosphorylated as follows: oocyte ribosomal subunits (168 pmol/reaction) were incubated for 10 min at 35 "C in a buffer containing 20 mM Hepes, pH 7.0, 0.1 mg/ml bovine serum albumin, 1 mM DTT,' 20 mM MgC12, and 130 p~ ATP in order to reduce nonspecific binding of radiolabeled ATP. The ATP concentration was then increased to 280 p M containing [Y-~'P]ATP (5000 cpm/ pmol), and 50 units of Xenopus unfertilized egg S6 kinase 2 (16, 17) were added in a final volume of 300 pl. The reaction was incubated for 90 min at 35 "C and terminated by dilution with 600 p1 of a buffer containing 40 mM Hepes, 2 mM DTT, 10 mM MgC12, 200 p~ EDTA, and 100 mM NaCl. The labeled ribosomes were washed by repeated precipitation with cold ethanol (42). Incorporation of phosphate into S6 varied from 1 to 3 mol of phosphate/mol of S6 and S6 was the only protein that significantly incorporated radiolabel. The 32P-labeled subunits were then diluted with an equal volume of ethylene glycol and stored at -20 "C. This procedure has been previously demonstrated to label the same phosphopeptides observed in vivo in maximally phosphorylated S6 after mitogenic stimulation (16).

Assays of Protein Phosphatase Activity-S6 phosphatase assays were carried out as previously described (32) except that all reactions were terminated by trichloroacetic acid precipitation, and release of inorganic phosphate was quantitated by Cherenkov counting of an aliquot of the supernatant. In addition, the M F concentration was 2 mM. Phosphorylase a phosphatase and phosphorylase kinase phosphatase assays were carried out under conditions identical to the S6 phosphatase assays.

Partial Purification of S6 Protein Phosphatase from Xenopus Eggs and Ouary-Female X. kzevis were primed and eggs collected as previously described (43). Dejellied eggs or ovarian tissue (60-120 g) were homogenized in 3 volumes of a buffer containing 10 mM bis-tris

' The abbreviations used are: DTT, dithiothreitol; Hepes, 442- hydroxyethy1)-1-piperazineethanesulfonic acid; bis-tris, 2-[bis(2-hy- ~oxyethyl)amino]-2-(hydroxymethyl)-propane-1,3-diol; CHAPS, 3- ((3-cholamidopropyl)dimethylammonio]-l-propanesulfonic acid; EG- TA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid; Pipes, 1,4-pi- perazinediethanesulfonic acid.

propane, pH 7.0, 0.1% 2-mercaptoethanol, 2 mM EGTA, 0.5 pg/ml leupeptin, 200 p~ phenylmethylsulfonyl fluoride, and 0.7 pg/ml pepstatin. After centrifugation for 30 min at 30,000 X g, supernatants were loaded onto a 100-ml DEAE-Sephacel column equilibrated in 10 mM bis-tris propane, pH 7.0, 25 mM NaCl, 0.1% 2-mercaptoethanol, 10% glycerol, 100 p~ EDTA, 200 p~ phenylmethylsulfonyl fluoride, and 2 mM CHAPS. The column was washed with the same buffer and proteins were eluted with an 800-ml linear gradient to 500 mM NaC1. Fractions (7.5 ml) were collected and assayed for S6 phosphatase and phosphorylase a phosphatase activity. Active fractions were collected and dialyzed against a buffer containing 10 mM Pipes, pH 6.8, 25 mM NaCl, 0.1% 2-mercaptoethanol, 10% glycerol, 0.01% Brij-

I I 20 40 60 80

Fraction Number

I 20

Fraction Number 40 60

I 1 I I I 20 40 60

Fraction Number

FIG. 1. Chromatography of the 30,000 X g supernatant of Xenopus ovary and eggs on DEAE-Sephacel. Fractions of 7.5 ml were collected and assayed for S6 phosphatase activity and phosphorylase a phosphatase activity as described under "Experimental Procedures." Fractions 24-58 in A and fractions 24-34 in B were pooled for further purification. A, Xenopus ovary; B, Xenopus eggs. S6 phosphatase activity is indicated by solid circles, phosphorylase a activity by open circles, Am by solid line. C, elution of a substrate- specific inhibitor of phosphorylase a phosphatase activity. The percent inhibition of rabbit skeletal muscle protein phosphatase 1 by DEAE fractions from oocytes is indicated by open triangles superim- posed on B. Aliquots of the eluted fractions were diluted 1/300 in a protein phosphatase 1 assay and the resulting mixture assayed for phosphorylase a phosphatase activity. Values are reported as percent inhibition of phosphorylase a phosphatase activity as compared to a buffer control.

A Novel S6 Phosphatase 153

TABLE I Chromatography of the inhibitor

The ability of boiled aliquots of each fraction listed in the first column to inhibit the phosphorylase a phosphatase activity of rabbit skeletal muscle protein phosphatase 1 is reported in the second column. The activity that each of these fractions possessed as S6 phosphatases before boiling is reported in the third column. The fourth column lists the phosphorylase a phosphatase activity of each of the fractions before boiling. Data are reported as percent release of phosphate under assay conditions described under “Experimental Procedures.”

Phosphorylase Q Intrinsic s6 Intrinsic phos- Pooled phosphatase ac- phorylase Q

fraction tivity of protein p h ~ ~ ~ ~ ~ ~ phosphatase phosphatase 1 activity

Buffer 15 DEAE 2 7 0 SP-Sephadex 2 5 0

SP-Sephadex 13 13 9 flow through

eluate

Fraction M e r

30 -

20 -

10- - Fractloo Number

FIG. 2. Chromatography of the S6 phosphatases on Sepha- cry1 5-200. Active phosphatase fractions obtained from SP-Sepha- dex were precipitated by the addition of (NH4)$04 as described under “Experimental Procedures” and subjected to gel filtration on Sephacryl S-200. 0, S6 phosphatase activity. -, A2m. The peaks are designated I and I1 in reference to their order of elution from the column. A, Xenopus ovary; B, Xenopus eggs.

35 (Pierce Chemical Co.), 100 p~ EDTA, and 200 p~ phenylmethylsulfonyl fluoride. The dialyzed sample was then applied to a 50-ml SP-Sephadex column equilibrated in the same buffer. The column was washed with the buffer and proteins were eluted by washing with 500 mM NaCl in the same buffer. The eluted sample was adjusted to 70% (NH4)k304 by the addition of saturated (NH4)&304, pH 7.0. After precipitation for 1 h at 4 “C, the sample was centrifuged for 10 min at 8000 X g, and the resulting pellet was resuspended in 0.5 ml of a buffer containing 10 mM Hepes, pH 7.0, 50 mM NaCl, 0.1% 2- mercaptoethanol, 10% glycerol, 0.01% Brij-35, 100 p~ EDTA, and 200 p~ phenylmethylsulfonyl fluoride. The sample was then gel filtered on a Sephacryl S-200 column (60 X 1.5 cm) at a flow rate of 10 ml/h and fractions (1 ml) were assayed for S6 phosphatase activity. The activity eluting at an M, of approximately 67,000 was pooled and diluted 5-fold into a buffer containing 10 mM Pipes, pH 6.5, 25 mM NaCl, 0.1% 2-mercaptoethanol, 5% glycerol, 0.01% Brij-35, and 100 pM EDTA. The sample was then applied to a Mono S column equilibrated in the same buffer and eluted with a linear gradient to 600 mM NaC1. Fractions (0.5 ml) were assayed for S6 phosphatase activity, and active fractions were pooled and diluted 5-fold with a buffer containing 10 mM bis-tris propane, 25 mM NaCl, 0.1% 2- mercaptoethanol, 5% glycerol, 0.01% Brij-35, and 100 p~ EDTA. The sample was then loaded on a Mono Q column, washed with the same buffer, and eluted with a 600 mM linear NaCl gradient. Active fractions were pooled and dialyzed into a buffer containing 10 mM Hepes, pH 7.0,25 mM NaC1,l mM DTT, and 50% glycerol for storage at -20 “C. An aliquot of the Mono Q fraction was subsequently

FIG. 3. Chromatography of the peak I1 enzyme on Mono S (top) and Mono Q (bottom) columns. Fractions were assayed for both S6 and phosphorylase a phosphatase activity. -, AZW.

! ’ 145K 4x 1%

1 1 n

20 10 .~

Fraclmlrlmbsr

FIG. 4. Sedimentation of the S6-specific phosphatase on a glycerol gradient (0). An aliquot of the Mono Q peak was applied to a glycerol gradient and sedimented as described under “Experi- mental Procedures.” - , Am (Mr = 145,000 peak) or A406 (Mr = 42,000 and 12,400 peaks) of molecular weight markers under identical conditions.

dialyzed into the same buffer containing only 7.5% glycerol for sedimentation on a glycerol gradient. The sample was layered onto a 10-30% gradient and centrifuged for 20 h at 3 “C in an SW-55 rotor a t 54,000 rpm. Samples (0.1 ml) were collected and assayed for S6 phosphatase and phosphorylase a phosphatase activity.

RESULTS

Previous studies (32) indicated that most S6 phosphatase activity in Xenopus ovary was inhibited by the heat-stable protein inhibitor 2, suggesting that most of the activity present was due to a protein phosphatase 1-like enzyme. Subse- quent studies have therefore concentrated on the characterization of this ion-independent, inhibitor 2-sensitive activity.

In oocytes, S6 is largely dephosphorylated while in eggs it is fully phosphorylated (8). Therefore, initial purification steps were carried out on both eggs and ovarian tissue in order to compare the nature of the S6 phosphatases present in each

154 A Novel S6 Phosphatase TABLE I1

Summary of the purification of the peak 11 enzyme Results of a typical preparation of the peak I1 enzyme. The SP-Sephadex and S-200 columns were evaluated as

a single step. Phosphorylase a phosphatase activity of the glycerol peak was too low to be determined. Total Total Specific Ratio SG/phospho- Purification Yield

protein activity activity rylase a

mg units unitslmg -fold 5% Cytosol 3,600 300 0.08 1 DEAE 883 132 1.5 1.8 1.8 44 SP/S-200 3.5 9.8 2.8 1.4 34 3.4 Mono S 0.85 7.5 8.8 3.5 106 2.5 Mono Q 0.08 8.2 102 17 1,280 2.7 From 0.1 ml of Mono Q fraction

Glycerol gradient 0.0003” 0.29 967 ND* 12,083 0.7 a Estimated from silver staining. * ND, not determined.

6 8 5

C Y T O S O L

W E

SPIS 200

MONOS MONO Q

GLYCEROL

FIG. 5. Silver-stained polyacrylamide gel of proteins present during purification of the S6-specific enzyme. The arrow marks the position of an M, = 55,000 protein correlated with activity in the glycerol gradient fractions.

and to ascertain the relative levels of activity. In Fig. 1, DEAE profiles of S6 phosphatase and phosphorylase a phosphatase activity in Xenopus ovary and eggs are shown. Two peaks of S6 phosphatase activity were observed in both tissues. The peak eluting at 160 mM NaCl was sensitive to inhibitor 2, whereas the peak eluting at 250 mM NaCl, which was some- what variable in height in different preparations, was not affected (not shown). While the peak at 160 mM appeared to be specific for S6, subsequent analysis revealed that the apparent lack of phosphorylase a phosphatase activity was due to the presence of a substrate-directed inhibitor (Fig. 1C). This inhibitor, which coelutes on DEAE with the major peak of S6 phosphatase activity, was assayed by its ability to inhibit the phosphorylase a phosphatase activity of protein phosphatase 1 and protein phosphatase 2A, two structurally distinct enzymes (24). Boiled, dialyzed preparations of these inhibi- tory fractions did not affect the S6 phosphatase activity of protein phosphatase 1 or of S6 phosphatases found in Xenopus eggs, although they retained their ability to affect phosphorylase a phosphatase activity (not shown). This heat-stable inhibitor was separated from the S6 phosphatase activity by chromatography on SP-Sephadex, resulting in the reappear- ance of phosphorylase a phosphatase activity at this step. While the S6 phosphatases bound to the column, the inhibitor did not and was found in the flow-through fraction (Table I).

Gel filtration of the S6 phosphatase activity eluted from the SP-Sephadex column on Sephacryl S-200 (Fig. 2) produced two peaks of activity, one of M, approximately 200,000

B 24 - - I 1 - 12

-e 20 Inhibitor Concentration (nM)

60

FIG. 6. Substrate specificity of the S6-specific enzyme. A, activity of the enzyme as an S6 phosphatase (01, phosphorylase a phosphatase (O), and phosphorylase kinase phosphatase (A) was assessed as described under “Experimental Procedures.” B, dose- response curve for the effect of inhibitor 1 and inhibitor 2 on the S6 phosphatase activity of the enzyme, carried out as described under “Experimental Procedures.”

(peak I), the second of M, approximately 67,000 (peak 11). Further purification of the M, = 200,000 phosphatase by Mono Q and Mono S chromatography gave a preparation that possessed a substrate specificity similar to rabbit skeletal muscle protein phosphatase 1 in that it dephosphorylated both phosphorylase a and the p subunit of phosphorylase kinase. In addition, it was inhibited by nanomolar concentrations of inhibitor 1 and inhibitor 2 (data not shown). The M, = 67,000 enzyme (peak 11) proved to have a narrow substrate specificity, and further efforts were concentrated on the characterization of this activity. At this stage, further attention was also focused on Xenopus eggs, rather than ovary. Early results (Figs. 1 and 3) suggested that the enzymes present in the two tissues were identical, and eggs proved to be a more convenient source of enzyme.

A Novel S6 Phosphatase 155

I FIG. 7. Time course of the dephosphorylation of Sf3 by the

SB-specific enzyme showing removal of phosphate from all labeled sites. 0, S6 phosphatase activity.

The M, = 67,000 enzyme migrated as a single peak of activity during further purification on Mono S and Mono Q columns (Fig. 3). Relative activities against S6 and phosphorylase a are shown for the Mono Q peak in Fig. 3. The virtually complete absence of activity against phosphorylase a indicates that this enzyme differs in its substrate specificity from both the rabbit skeletal muscle protein phosphatase 1 and the peak I phosphatase found in Xenopus tissue (Fig. 2). Further purification of an aliquot of the Mono Q peak by sedimentation through a glycerol gradient produced a single peak of activity that sedimented with an M, of approximately 55,000 (Fig. 4). The purification steps are summarized in Table 11. The S6- specific enzyme was purified approximately 12,000-fold over the crude supernatant. Although not homogeneous (Fig. 5), the highly purified enzyme appears to be a single activity uncontaminated by any other phosphatase and exhibits by the Mono Q step a 17-fold preference for S6 relative to phosphorylase a. The enzyme functioned neither as a phos- photyrosyl casein phosphatase nor as a p-nitrophenyl phosphate phosphatase (not shown), suggesting that it was not related to the tyrosine phosphatases or to the acid or alkaline phosphatases.

Because the activity eluted as a single peak on both S-200 and Mono Q chromatography, further characterization uti- lized enzyme preparations purified 1000-3000-fold with these columns. Initially, experiments were designed to determine if the oocyte enzyme could be classified using the standard criteria of Cohen’s laboratory. However, as shown in Fig. 6A, the enzyme had very little activity against either phosphorylase a or phosphorylase kinase labeled in the a and /3 subunits as compared to its activity as an S6 phosphatase. This result suggested that the purified enzyme did not fit the classification criteria, since it was inactive against both the substrates normally used to classify protein phosphatases. On the other hand, the S6 phosphatase was completely inhibited by inhibitor 1 and inhibitor 2 (Fig. 6B) with ICSo values of 63 nM for inhibitor 1 and 27 nM for inhibitor 2.

Since S6 is multiply phosphorylated, the ability of the enzyme to act on all phosphorylated sites was investigated. As shown in Fig. 7, the S6 phosphatase was capable of removing over 80% of the 32P from S6 labeled by S6 kinase I1 i n vitro. Since it has previously been shown that the S6 kinase labels all the phosphopeptides seen in maximally phosphorylated S6 i n vivo (16), these results indicate that a single phosphatase may be sufficient to account for S6 dephosphorylation i n vivo.

DISCUSSION

Previously, we reported that the rabbit skeletal muscle enzymes protein phosphatase 1 and protein phosphatase 2B function as S6 phosphatases, and that enzymes similar to these appear to exist in Xenopus oocytes (32). We report here

the partial purification of the two enzymes responsible for the majority of S6 phosphatase activity in Xenopus ovary and eggs. S6 is dephosphorylated in oocytes but fully phosphorylated in eggs, suggesting that there might be differences in the nature of the phosphatases found in each tissue. However, both enzymes were present in similar amounts in ovary and eggs. Both of these enzymes are inhibited by nanomolar concentrations of inhibitor 1 and inhibitor 2, indicating the inhibitor proteins are not specific for protein phosphatase 1. Since the reported physiological concentration in skeletal muscle of inhibitor 1 is 1.8 PM, while that of inhibitor 2 is 0.35 PM (44), this result suggests that the level of S6 phosphorylation may, in part, be regulated through changes in the activity of the inhibitors. One of the enzymes described, which has an apparent molecular weight of 200,000 by gel filtration, has a substrate specificity similar to that of the rabbit skeletal muscle protein phosphatase 1 in that it dephosphorylates S6, phosphorylase a, and the /3 subunit of phosphorylase kinase. It may therefore be a multisubunit form of protein phosphatase 1 as has been reported for skeletal muscle tissue (24).

The second enzyme eluting from Sephacryl S-200 (Fig. 2), which has an apparent molecular weight of 55,000-67,000, has been purified 12,000-fold and appears to have a narrow substrate specificity. Specifically, it has very little activity towards phosphorylase a and phosphorylase kinase, and can therefore be distinguished from previously described phosphatases. Its susceptibility to inhibition by inhibitor 1 and inhibitor 2, coupled with its apparent selectivity, also suggests that it falls outside the classification system for serine/threonine phosphatases described by Cohen and co-workers (22-25).

Although the classification scheme proposed for the serine/ threonine phosphatases has led to acceptance of the concept that protein dephosphorylation is due in most cases to the action of four broad specificity enzymes (24), few rigorous comparisons of enzymes from other tissues to those regulating glycogen metabolism exist. The results presented here indicate that, while enzymes similar to those described in glycogen metabolism do exist in other tissues, sensitivity of dephosphorylation to inhibitor 1 and inhibitor 2 in a tissue extract is not a sufficient criterion for the implication of protein phosphatase 1 in a cellular process. In addition, the assess- ment of phosphatase specificity for substrates in tissue extracts is complicated by the presence of substrate-directed inhibitors such as the one reported here. As a principal conclusion, the data suggest that protein dephosphorylation can result from the action of a highly specific enzyme which may be subject to regulation by proteins that also regulate phosphatase 1. Consequently, the role of protein phosphatases in the regulation of a cellular process by mitogenic stimuli cannot be fully assessed using the protein inhibitors or assay- ing with relatively nonspecific substrates such as phosphorylase a or phosphorylase kinase, which are not substrates for specific enzymes like the S6 phosphatase reported here. In- terestingly, a similar conclusion has been reached in the analysis of mitogenically stimulated S6 kinases, i.e. their absence of activity against conventional substrates such as histone or casein (16, 17) delayed their discovery for some years and made more difficult analysis of the regulatory pathways involved in cell proliferation. It is clear that in the study of S6 phosphorylation, it is important to use 40 S subunits in both kinase and phosphatase assays. We previously showed that microinjection of inhibitor 2 into oocytes increased the phosphorylation state of S6 (32). It is not clear at present which of the two S6 phosphatases reported here is responsible for this effect, but since dephosphorylation of the /3 subunit of phosphorylase kinase was also inhibited (30, 32),

Documents

Purification and Characterization of a Novel Protein Phosphatase