Phosphoprotein Phosphatase Inhibitor-2 IDENTIFICATION AS A SPECIES OF MOLECULAR WEIGHT 31,000 IN RABBIT MUSCLE, LIVER, AND OTHER TISSUES*

Specific antibodies, raised to purified rabbit skeletal muscle inhibitor-2, were used to analyze for the pres- ence of inhibitor-2 in extracts of rabbit skeletal, cardiac, and diaphragm muscles, liver, kidney, brain, and lung. Direct analyses of the extracts by “Western blot- ting” revealed several immunoreactive species, apparent molecular weights in the range 26,000-136,000, as well as species with the electrophoretic mobility of inhibitor-2, apparent molecular weight 31,000. When supernatants from boiled extracts were similarly analyzed, most of the immunoreactive material was lost and the species corresponding to inhibitor-2 became prominent. Liver and muscle were studied in more detail; immunoprecipitates from either boiled or unboiled extracts were analyzed by Western blotting. The dominant polypeptide now was the species of apparent molecular weight 3 1,000, corresponding to inhibitor- 2. Higher molecular weight species (1 15,000 in muscle and 136,000 in liver) were also detectable. The amount of inhibitor-2 detected in immunoprecipitates was not greatly different whether unboiled or boiled tissue ex- tracts were used. In addition, extraction of the precip-itates by boiling released material that inhibited puri- fied type 1 protein phosphatase. The results suggest that inhibitor-2 is widely distributed in rabbit tissues and is found predominantly as a form of apparent molecular weight 31,000. In particular, the study pro-vides direct demonstration of a species in rabbit liver with similar properties to rabbit muscle inhibitor-2.

* This work was supported in part by National Institutes of Health Diabetes Foundation (to A. A. D.-R.), and the Grace M. Showalter 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 "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$Currently a medical student a t Indiana University School of Medicine. Supported in this research by a Student Research Fellowship from the Indiana Affiliate of the American Diabetes Association.

Recipient of Research Career Development Award AM01089
from the National Institutes of Health. the protein in cells or even tissue extracts. Indeed, a t least some inhibitor-2 may exist in noncovalent association with other proteins. Several groups (7-12) have suggested that inhibitor-2 is a component of the ATP-Mg-dependent protein phosphatase described by Merlevede and colleagues (13). The exact native structure of this phosphatase has yet to be unequivocally elucidated.
Although considerable information has accumulated regarding skeletal muscle inhibitor-2, the status of phosphatase inhibitors in other tissues is less certain. Studies of liver have indicated the presence of low molecular weight protein inhibitors of phosphatase activity (14)(15)(16)(17). The relationship of these species to the better studied muscle inhibitors is not completely clear and in the only case where purification to homogeneity was described, a species of apparent M, approximately 15,000 was obtained (15). The conclusion of the present investigation is that inhibitor-2 can be detected mainly as a species of apparent molecular weight 31,000 in rabbit skeletal, diaphragm, and cardiac muscle, liver, brain, lung, and kidney.

EXPERIMENTAL PROCEDURES
Antibodies to Inhibitor-2"Female guinea pigs (400-500 g) were injected subcutaneously with an emulsion of Freund's complete adjuvent (1 ml) containing 80 pg of purified rabbit skeletal muscle inhibitor-2. After 4 weeks, a booster injection was administered and 2 weeks later the animals were bled. Preimmune serum was obtained from heart puncture before inoculation. For some experiments, antibodies were purified from serum by affinity chromatography. Purified inhibitor-2 (0.8 mg) was coupled to 1 g of CNBr-activated Sepharose 4B (see Ref. 18). Immune serum (6 ml) was applied to a column (2.5 ml) of inhibitor-2-Sepharose. After washing with 20 mM Tris-HCI, pH 7.0, plus 0.1 M NaCl and then 20 mM Tris-HC1, pH 7.0, plus 0.45 M NaCI, specific antibodies were eluted with 3 M NaSCN and then dialyzed against 10 mM sodium phosphate, pH 7.4, plus 0.13 M NaCl.
From gel electrophoretic analysis in the presence of SDS' and binding to protein A, the antibodies were identified as IgG.

Preparation of Tissue Extracts-Rabbits
(male New Zealand White, 2.5-3.5 kg) were sacrificed by injection of sodium pentobarbital in the marginal vein of the ear. Portions of the following tissues were excised: skeletal, diaphragm and cardiac muscle, liver, kidney, brain, and lung. These were frozen in liquid nitrogen and stored a t of buffer, pH 7.2, composed of 50 mM Tris-HCI, 2 mM EGTA, 5 mM -70 "C until processing. Tissue was homogenized in a ratio l 'g: 8 ml EDTA, 100 mM NaF, 2 mM benzamidine, 0.5 mM phenylmethylsulfonyl fluoride, 0.1 mM N"-p-tosyl-t-lysine chloromethyl ketone HCI, 0.2 mM L-1-tosylamido-2-phenylethyl chloromethyl ketone, and 2 pg/ ml of leupeptin. After disruption using a Polytron, setting 6 for 30 s, the homogenate was centrifuged for 15 min a t 13,000 X g and the supernatant was removed. A portion was retained (unboiled extract) while another was heated to 100 "C for 10 min before centrifugation as above to yield a supernatant (boiled extract). Protein Phosphatase Inhibitor-2 6315 1 g of tissue/5 ml of buffer for these experiments, were incubated with 19 pg of affinity purified antibodies or control IgG for 30 min a t room temperature before addition of 60 p1 of a 5% (w/v) suspension o f inactivated Staphylococcus aureus of strain Cowan 1 (Pansorbin, Calbiochem-Behring). After 30 min at 0 "C, the bacterial pellet was phosphate, 1 mM EDTA, 0.4 M NaC1, 0.1% Triton X-100, pH 7.5. In harvested by centrifugation and washed twice with 10 mM sodium the final wash the NaCl concentration was reduced to 0.1 M. The bacterial pellet was resuspended in 200 pl of 50 mM Tris-HC1, 0.1 mM EDTA, 50 mM @-mercaptoethanol, pH 7.0, and treated at 100 "c for 10 min to release inhibitor-2. The bacteria were removed by centrifugation. Gel Electrophoresis, Western Transfers, and Immunoblotting-Electrophoresis on 6-20% gradient polyacrylamide gels in the presence of SDS followed a modification of the method of Laemmli (19) as described previously (20). Gels 0.75-mm thick were used. Immunohlotting followed a modification of the method of Burnette (21). After electrophoresis, the gels were submerged in a solution of 25 mM Tris, 190 mM glycine, pH 8.3, with methanol added to 15% (v/v). Transfer to nitrocellulose (BA85, Schleicher & Schuell) occurred in the same buffer at 240 V and 0.8 A for either 2 or 4 h. The nitrocellulose filter was washed for 45 min in PBS-T (20 mM phosphate, 115 mM NaCl, 0.1% Tween 20, pH 7.35). After changing the buffer (35 ml), antibodies were added and incubated for 16 h a t room temperature. Either immune serum (35 pl) or affinity purified antibodies (9 pg) were used. A second identical nitrocellulose transfer was incubated with an equivalent amount of nonimmune serum or IgG to act as a control. The filters were washed, 10 min each, with PBS-T, PBS-T plus 0.1% (v/v) Triton X-100, Tris-saline (20 mM Tris-HC1, 150 mM NaC1, pH 7.4) plus 0.075% (v/v) Nonidet NP-40 and 1 M NaC1, Tris-saline, and PBS-T. The filters were exposed to 1 pCi of ""I-Protein A (New England Nuclear, specific activity 8.9 pCi/pg) for 60 min, after which the same washes as described above were applied. Autoradiograms were prepared from dried filters. When required, nitrocellulose filters were stained for protein using Amido Black 10-B. In preliminary experiments, the effective transfer of inhibitor-2 to the nitrocellulose was established using 32P-inhibitor. Variations in transfer time, stringency of the washes, inclusion of bovine serum albumin in the "blocking" step, and reduction of the time of exposure to antibodies all had no qualitative effect on the results using tissue extracts.
Other Methods and Materials-Inhibitor-2 was purified close to homogeneity from rabbit skeletal muscle as described by DePaoli-Roach (12). Casein kinase I1 (PCO.,) and phosphorylase were isolated from the same source by published methods (12, 18). Homogeneous type 1 phosphatase was purified from rabbit muscle by modification of published methods (22);' the enzyme contained a single subunit of M , = 38,000. Phosphatase assays using [32P]phosphorylase a as substrate, phosphorylation of inhibitor-2 with casein kinase 11, and other materials and methods can be found in a recent publication (12). A unit of phosphatase activity releases 1 nmol of phosphate from phosphorylase a a t 30 "C.

RESULTS
Characterization of Antibodies-Initial characterization of the immune serum utilized inhibitor-2 that had been labeled with '*P after phosphorylation by casein kinase I1 (see Ref.

12)
. Immunoprecipitation was carried out by harvesting antigen-antibody complexes with S. aureus. As seen in Fig. 1, almost complete removal of inhibitor-2 (70 ng) from solution was achieved with 2 p1 of immune serum. No inhibitor-2 was precipitated with the control serum. Analysis of the immunoprecipitate by polyacrylamide gel electrophoresis in the presence of SDS indicated that all the radioactivity removed was associated with inhibitor-2 (not shown). Also, addition of unphosphorylated inhibitor-2 to the 32P-labeled inhibitor before exposure to antibodies reduced the amount of 32P immunoprecipitated (Fig. l ) , indicating competition for the antibodies. The results demonstrate first the ability of the antibodies to recognize purified inhibitor-2. Secondly, the antibodies do not distinguish between unphosphorylated in- pg  FIG. 1. Characterization of antibodies to inhibitor-2. Immunoprecipitation of purified rabbit muscle inhibitor-2 (i-2), labeled with 32P using casein kinase 11, was effected by methods similar to those described under "Experimental Procedures" for tissue extracts. A , inhibitor-2 (70 ng, 22,400 cpm) was incubated for 30 min a t room temperature with the indicated amount of immune (0) or control (0) serum. Excess of S. aureus was added and, after removing the bacterial pellet, by centrifugation, the radioactivity remaining in solution was determined. B , unlabeled purified inhibitor-2, as indicated in the figure, was mixed with 3ZP-labeled inhibitor before exposure to a fixed amount (2 pl) of immune serum. Increasing 3zP in solution then represents competition by the unlabeled inhibitor-2 for antibody binding.
hibitor-2 and inhibitor that was phosphorylated by casein kinase 11.
Analysis of Inhibitor-2 by Zmmunoblotting-Extracts were prepared from various rabbit tissues, as described under "Experimental Procedures," either with or without heat treatment of the extracts. These samples, together with a standard of purified rabbit muscle inhibitor-2, were subjected to polyacrylamide gel electrophoresis in the presence of SDS, transferred to nitrocellulose, and probed with antibodies against inhibitor-2. In Fig. 2 are shown the results of such an experiment using affinity purified antibodies. The corresponding control, using nonimmune IgG as a probe, had no radioactive species detectable on the autoradiogram (not shown). With unboiled extracts, several immunoreactive species were found for each of the tissues analyzed. The patterns were to some degree tissue specific. Prominent species with apparent M , approximately 39,000 were seen in all of the tissues except lung. In skeletal muscle, diaphragm, and liver, another polypeptide of apparent M , 58,000 was visible and a species of apparent M, 50,000 was common to liver, kidney, lung, and perhaps brain. Brain was characterized by a strongly interacting species of M, 60,000-66,000. A few very high molecular weight polypeptides were also apparent. Of interest are species of approximately 115,000 in skeletal and diaphragm muscle, and of 136,000 in liver, kidney, and lung. Some other species can be seen in the figure. Relatively less prominent but still clearly detectable in all tissues, however, was a species with the electrophoretic mobility of inhibitor-2, apparent M , 31,000.
The parallel analysis of boiled extracts indicated much simpler patterns. The species of M , 31,000, corresponding to inhibitor-2, was now prominent in all tissues (Fig. 2) though present perhaps in lower amount than in unboiled extracts. The dominant brain polypeptide(s) of 60,000 to 66,000 daltons remained. The only other species detected were the very high molecular weight ones, 115,000 for skeletal and diaphragm muscle and 136,000 for liver, kidney, and lung.
The basic features of the results just described were not altered by a number of variations in the procedures both related to the Western blotting (see "Experimental Procedures") and in sample preparation. Omission of protease inhibitors and EGTA from the homogenization buffer did modify the patterns slightly where tested (with liver, skeletal muscle, and diaphragm) but did not lead to any increase in the amount of the 31,000-dalton species corresponding to inhibitor-2. In fact, in the boiled liver extract, the amount of inhibitor-2 was clearly decreased if the protease inhibitors were omitted, with the appearance of smaller immunoreactive polypeptides (not shown). Initially, immune serum was used instead of purified antibodies. No major differences in the results were found except that a prominent species was detected, at around 60,000, in all the tissue extracts. Presumably, this species was not due to antibodies that bound inhibitor-2 during affinity purification.
Immunoprecipitation of Inhibitor-2 from Liver and Muscle-Further experiments were carried out using immunoprecipitation which allowed physical separation of the material recognized by the antibodies. After exposure of extracts, boiled or unboiled, to antibodies, the antibodies were harvested with S. aureus, as described under "Experimental Procedures." Material released from the bacterial pellets by boiling was analyzed by gel electrophoresis followed by immunoblotting (Fig. 3) or assayed for phosphatase inhibitory activity ( Table I). The effective detection of purified inhibitor-2 (66 ng) by immunoblotting is indicated by track 1 (Fig. 3). From Fig. 3, it is apparent that the main species immunoprecipitated from either liver or muscle extracts had apparent M, 31,000. No such species was present when nonimmune IgG were used in the immunoprecipitation. Similar results were obtained using material released from the bacterial pellet by 70% formic acid rather than boiling (data not shown).
Traces of the very high M , polypeptides were also present. It should be noted that the relative abundance of these low mobility polypeptides was somewhat variable in these experiments. In addition, comparison of the results using boiled or unboiled extracts indicated that the yield of 31,000-dalton polypeptide was not greatly affected by heat treatment of the tissue extracts. Table I indicates that the material released from the immunoprecipitates by boiling, from liver or muscle

115
1 2 3 4 5 6 7 8 910 FIG. 3. Analysis of immunoprecipitates from rabbit skeletal muscle and liver extracts by Western blotting. Immunoprecipitates prepared from either native (tracks 7-10) or boiled (tracks 3-6) extracts of muscle and liver were analyzed by immunoblotting, as described under "Experimental Procedures." An autoradiogram of the nitrocellulose filter is shown. Track 1,66 ng of purified inhibitor-2 as a standard; track 2, a protein A control in which neither extract nor antibodies were included; tracks 3 and 4, boiled liver extracts immunoprecipitated with immune or control I&, respectively; tracks 5 and 6, boiled muscle extracts immunoprecipitated with immune or control IgG; tracks 7 and 8, unboiled liver extracts immunoprecipitated with immune or control I&; tracks 9 and IO, unboiled muscle extracts immunoprecipitated with immune or control IgG. For tracks 3-10, the equivalent to 44 pl of the material released from the immunoprecipitates was applied to the polyacrylamide gel. The two-digit numbers are apparent molecular weights ( X calculated from the migration of standard proteins. extracts, was able to inhibit purified type 1 protein phosphatase. Therefore, the evidence strongly argues that the species of apparent M, 31,000 visualized in this study indeed corresponds to inhibitor-2.

DISCUSSION
The main experimental result of this investigation is straightforward. Inhibitor-2 can be detected in rapidly processed tissue extracts as a species of apparent M , 31,000. This species is the dominant one recognized by antibodies to inhib-Protein Phosphatase Inhibitor-2 itor-2 in either boiled or unboiled extracts of rabbit liver and muscle in immunoprecipitation experiments and correlated with the presence of phosphatase inhibitory material in the immunoprecipitates. These results demonstrate that liver and muscle contain inhibitor-2 in a form close to that purified from muscle by conventional methods. This observation is consistent with the detection of inhibitor-2 as a component of the muscle ATP-Mg-dependent protein phosphatase purified by procedures that avoid harsh treatments (e.g. Refs. 7 and 11). The significance of the higher M , species detected by immunoblotting of unboiled extracts is not clear. From the immunoprecipitation experiments, either most of these species were not recognized in their undenatured states by the antibodies or else they were not released by boiling or formic acid extraction of the immunoprecipitates. Since formic acid would be expected to solubilize most proteins, one possibility is that antigenic determinants common to inhibitor-2 were exposed only upon denaturation. Of course, an interesting question is whether such species have any functional and/or structural relationship to inhibitor-2. More detailed studies will be required but we can note that our experiments never indicated the generation of inhibitor-2 as a 31,000-dalton species. Boiling of extracts did not lead to the accumulation of inhibitor-2 nor did omission of protease inhibitors from the homogenization buffer. These results are not definitive but do not provide any strong indications that inhibitor-2 is produced by manipulation of the tissue extracts. The most important conclusion of the study relates to the tissue distribution of inhibitor-2 in a form consistent with our knowledge of the muscle inhibitor. As noted in the Introduction, relatively little structural information is available con-cerning inhibitor-2 in non-muscle tissues. From the results, it is evident that rabbit liver contains an antigenically related species of similar molecular weight. In liver extracts, this protein was more susceptible to proteolytic degradation, which could correlate with the identification of a lower molecular weight form of phosphatase inhibitor in liver in an earlier study (15). From immunoblotting of boiled extracts of other rabbit tissues (cardiac and diaphragm muscle, kidney, brain, and lung) we can infer that inhibitor-2 is indeed a widely distributed protein. The function of inhibitor-2 in the control of phosphatase activity is therefore likely to be important for protein dephosphorylation in many, if not all, mammalian tissues.