Phosphorylation of glycogen synthase in the perfused rat heart.

This study was initiated in order to elucidate further the role of protein phosphorylation in the regulation of glycogen synthase in ho. Rat hearts were perfused with inorganic [32P]phosphate and subsequently extracted under conditions which stabilized the glycogen synthase I/D ratio during the isolation of the enzyme. Throughout the standard Langendorff perfusion of up to 60 min, rat heart glycogen synthase existed 80% in the already phosphorylated D-form. We have previously shown (England, P. J., and Walsh, D. A. (1976) Anal. Biochem. 75, 429-435) that inorganic [32P]phosphate in the perfusate equilibrates slowly with intracellular inorganic [32P]phosphate and [Y-~~P]ATP, reaching only 5 to 10% equilibration in 60 min. In this present study it has been shown that perfusion with inorganic [32P]phosphate led to a slow accumulation of 32P covalently bound to glycogen synthase. Epinephrine addition (2 pM) did not modify the extent of 32P covalently bound to glycogen synthase. When purified rat heart glycogen synthase was phosphorylated with the catalytic subunit of CAMP-dependent protein kinase, sodium dodecyl sulfate gels of the resultant CNBr hydrolysates exhibited two 32P-phosphopeptides. These 32P-phosphopeptides co-migrated with those derived from either rat heart or rabbit skeletal muscle glycogen synthases that had been phosphorylated either by the rat heart CAMP-independent synthase kinase or the catalytic subunit of CAMP-dependent protein kinase. The CAMP-dependent and CAMPindependent enzymes, however, did preferentially phosphorylate the two sites of rat heart glycogen synthase at different rates. The same two 32P-phosphopeptides were obtained from rat heart glycogen synthase that had been phosphorylated in ho.

did not modify the extent of 32P covalently bound to glycogen synthase. When purified rat heart glycogen synthase was phosphorylated with the catalytic subunit of CAMP-dependent protein kinase, sodium dodecyl sulfate gels of the resultant CNBr hydrolysates exhibited two 32P-phosphopeptides.
These 32P-phosphopeptides co-migrated with those derived from either rat heart or rabbit skeletal muscle glycogen synthases that had been phosphorylated either by the rat heart CAMP-independent synthase kinase or the catalytic subunit of CAMP-dependent protein kinase. The CAMP-dependent and CAMPindependent enzymes, however, did preferentially phosphorylate the two sites of rat heart glycogen synthase at different rates. The same two 32P-phosphopeptides were obtained from rat heart glycogen synthase that had been phosphorylated in ho.
Glycogen synthase (UDPglucose:glycogen a-4-glucosyltransferase, EC 2.4.1.11) is the rate-limiting enzyme for the synthesis of glycogen in mammalian tissues (1). Glycogen synthase exists in two interconvertible forms which can be distinguished by their dependence on glucose 6-phosphate for activity (1,2). While the I-form has activity in the absence of any effector, the D-form is dependent on glucose &phosphate for maximum activity. In vitro studies have shown that rabbit heart and skeletal muscle glycogen synthases can be converted from the I-form to the D-form by phosphorylation reactions catalyzed by CAMP-dependent protein kinase (3)(4)(5). Rabbit skeletal muscle glycogen synthase can also be converted from the I-form to the D-form by phosphorylation reactions catalyzed by cyclic nucleotide-independent kinases (6)(7)(8) and by * This work was supported by Grant AM 13613 from the National Institutes of Health. 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.
the cGMP-dependent protein kinase (9). These three types of kinases apparently phosphorylate the same sites on the skeletal muscle enzyme but at differing rates (9,10).
Epinephrine administration to the perfused rat heart elicits a rapid elevation of CAMP levels (11) with a concomitantly rapid activation of phosphorylase b kinase (12). However, conflicting reports have indicated that either there is no change in the activation state of glycogen synthase (l&13) or that there is either an increase or a decrease in the percentage of the enzyme in the I-form, depending on whether or not glucose is included in the perfusate (14).
In this presentation, we report that in rat hearts perfused with inorganic [32P]phosphate there is a slow accumulation of 32P covalently bound to glycogen synthase. We have compared the sites phosphorylated in vivo with the sites on both heart and skeletal muscle glycogen synthases phosphorylated in vitro. In in vitro studies the 32P-labeled cyanogen bromide peptides obtained from heart glycogen synthase coincided with those previously characterized from skeletal muscle (10). In vivo phosphorylation occurred in the same CNBr fragments as were phosphorylated in vitro.

Methods
Heart Perfusion-Hearts from male Sprague-Dawley rats (240 to 260 g) were perfused essentially as described by England (15,16 Enzyme Activity Assays-Glycogen synthase was assayed by a modification of the method of Thomas et al. (20). The extract from powdered frozen tissue was diluted to a final concentration equivalent to 1 g wet weight of tissue in 50 ml of extraction buffer containing 1 mg/ml of bovine serum albumin. The reaction was initiated by the addition of 0.02 ml of diluted extract to 0.10 ml of a solution containing 60 mru Tris (pH 7.75), 40 mM NaF, 12 mg/ml of glycogen, 5 mM EDTA, 5 mM ['?]UDP-Glc, and a 10 mrvr concentration of either Na2S04 or glucose-6-P.
14C incorporation into glycogen was determined by pipetting 0.05-ml aliquots onto squares (1 X 1 cm) of filter paper (Whatman ET31); the papers were immersed in 66% ethanol at 0°C and subsequently washed three times (30 min each) with 66% ethanol and once (5 min) with acetone (20). The dried squares were counted in toluene scintillation fluor. The rate of 14C incorporation into glycogen was linear between 2 and 45 min but nonlinear for the first 2 min. Glycogen synthase activity was determined routinely from the extent of incorporation between 2 and 32 min.  Weber and Osborne (22). The gels were fixed in 7.5% acetic acid and then sliced in 2-mm sections. The slices were placed on filter paper squares, dried at lOO"C, and counted in toluene scintillation fluor. Purification of Rat Heart Glycogen Synthase-Rat heart glycogen synthase was partially purified from 10 g of tissue obtained from 12 animals.
The minced tissue was homogenized in a Waring Blendor with 60 ml of 50 mM Tris (pH 7.5), 100 mM NaCl, 2 mM EDTA, and 45 mM 2.mercaptoethanol (Buffer A). The homogenate was clarified by centrifugation for 20 min at 40,000 x g. The supernatant then was ' The abbreviation used is: SDS, sodium dodecyl sulfate. centrifuged for 90 min at 78,000 X g. Glycogen synthase was precipitated from the 78,000 x g supernatant by ammonium sulfate (20 g/ 100 ml). Following centrifugation for 15 min at 10,000 X g, the pellet was solubilized in 150 ml of 50 mM Tris (pH 7.5), 2 mM EDTA, 45 mM 2-mercaptoethanol, 5% sucrose (Buffer B). The enzyme was incubated for 1 h at 4°C with 180 ml packed volume of DE52 resin that had been equilibrated previously with Buffer B. The resin was poured in a column (2.5 x 45 cm) and the protein subsequently was eluted with a 500-ml linear gradient of 0 to 0.4 M NaCl in Buffer B. Glycogen synthase was eluted at a conductivity of between 5.0 and 7.0 mmhos and was recovered from the pooled fractions by the addition of ammonium sulfate (20 g/100 ml). Following centrifugation for 15 min at 10,000 x g, the pellet was solubilized in 2 ml of 50 mM Tris (pH 7.5), 2 mM EDTA, 45 mru 2-mercaptoethanol, and 10% sucrose (Buffer C) and then applied to a Sepharose 4B column ( kinases was concentrated to 1.5 ml with an Amicon filtration system, using a PM-10 filter, and stored at 4'C. The inhibitor protein of CAMP-dependent protein kinase had no effect on glycogen synthase kinase activity. Phosphorylation of Glycogen To ensure that all the CNBr phosphopeptides were derived from glycogen synthase and not from any contaminating phosphoproteins, samples of the phosphorylated glycogen synthases isolated by the procedures described above were routinely characterized by electrophoresis on 6% SDS gels (10 cm) as per Weber and Osborne (22). In all experiments gels of pure rabbit skeletal muscle glycogen synthase exhibited a single, sharp peak of radioactivity co-migrating with the enzyme. However, gels of partially purified rat heart glycogen synthase that had been phosphorylated in Phosphorylation of Glycogen Synthase vitro or of an immunoprecipitate prepared from rat heart perfused with inorganic [""PIphosphate occasionally exhibited small amounts of contaminating '"P-phosphoproteins. Therefore, the glycogen synthase in partially purified rat heart glycogen synthase phosphorylated in vitro and in immunoprecipitates of a rat heart perfused with inorganic r'"P]phosphate was routinely further purified by preparative electrophoresis prior to CNBr digestion.
In order to perform preparative electrophoresis the proteins were solubilized in 200 ~1 of 10 mM sodium phosphate (pH 7.0), 1% SDS, 2.5% 2-mercaptoethanol, 5%, glycerol, and 0.02% bromophenol blue and electrophoresed on three gels. The region in the gels corresponding to the glycogen synthase was dissected (R, = 0.25 to 0.40); each section was placed on top of a 6% SDS gel plug (1 X 0.6 cm) and electrophoresed 20 h at 8 mA/gel, and the eluted protein was collected in dialysis tubing. Rat heart glycogen synthase was precipitated at 4°C by the addition of 100% trichloroacetic acid to a final concentration of 20%. Following centrifugation for 30 min at 10,000 X g the pellet was washed at 4°C with 5 ml of 10% trichloroacetic acid and two times with 7.5 ml of ether. The pellets were solubilized in 0.5 ml of 70% formic acid. Cyanogen bromide digestion was performed by the method of Soderling et al. (10) and the phosphopeptide products were identified by SDS electrophoresis on 10% gels (IO cm) by the method of Weber and Osborne (22). CAMP Deterninatiorzs-Samples of frozen tissues (50 mg) were homogenized in 0.5 ml of 0.1 N HCl, and 0.15 ml of water was added. The samples were placed in a boiling water bath for 2 min and then centrifuged for 10 min at 27,000 X g. A 0.5-ml aliquot of the supernatant was neutralized by the additions of 0.05 ml of 0.6 N NaOH and 0.01 ml of 2 M sodium acetate. The sample was clarified by centrifugation, and the supernatant was assayed for CAMP by the method of Brostrom and Kon (25).

Materials
The following chemicals were from Sigma Chemical

RESULTS
["'P]Glycogen synthase, labeled during perfusion of rat heart, was isolated by immunoprecipitation for either control or epinephrine-stimulated tissue. In agreement with the results of others (11,13,14) glycogen synthase in extracts from unstimulated perfused heart exists mainly in the already phosphorylated D-form. (As indicated subsequently (uide Fig.  4) the phosphorylation (and, in consequence, activation) state of the enzyme, as measured, is not a result of covalent modification occurring during the extraction procedure.) Immunoprecipitation of glycogen synthase and activity assays were performed with rat heart extracts clarified by centrifugation for 15 min at 15,000 x g. Under these conditions 85% of the total activity was soluble and the percentage of the enzyme in the I-form was the same as for the enzyme measured in the total homogenate ( Fig. 1). At higher centrifugation speeds an increasing percentage of total activity was sedimented and the D-form was selectively removed from the supernatant.
Effect of Epinephrine on CAMP, Phosphorylase, and Glycogen Synthase-As reported by others for perfused rat heart (11) epinephrine stimulated CAMP accumulation, increased phosphorylase activity, but had negligible effect on glycogen synthase activity (Fig. 2). The control level of CAMP was 0.02 nmol/g (wet weight) of tissue. This level increased rapidly reaching a maximum of 1.11 nmol/g at 20 s after epinephrine administration.
The percentage of phosphorylase a increased from a basal level of 12.4% to a maximum of 79.1% at 20 s. In the absence of epinephrine glycogen synthase was observed to be predominantly in the inactive form. The basal level of 79.7% D-form did not vary more than +3.6% during 30 s of 2 pM epinephrine exposure, even though this concentration was shown to produce a maximal change in phosphorylase kinase activity.* The basal level of glycogen synthase in the D-form was not altered when unmodified Krebs-Henseleit buffer (i.e. 1.17 ITIM KHZPOI instead of 0.12 mM) was used, when glucose was deleted from the perfusate, nor when the perfusate temperature was varied from 35 to 38°C (data not shown).
Glycogen Synthase Immunoprecipitation-In the extract, rat heart glycogen synthase is bound to glycogen. The presence of this complex prevented quantitative and reproducible immunoprecipitation of the synthase; however, prior treatment of the extract with cY-amylase for 30 min at 0°C circumvented this difficulty. The characteristics of immunotitration and immunoprecipitation of glycogen synthase from cy-amylase-treated rat heart extract are presented in Fig. 3. Rat heart glycogen synthase activity was inhibited by the addition of rabbit anti-skeletal muscle glycogen synthase (Fig. 3A), but quantitative precipitation of the enzyme was obtained only with the subsequent addition of carrier skeletal muscle glycogen synthase (Fig. 3, B and C). In the experiment presented in Fig. 3A the a-amylase and enable the subsequent assay of glycogen synthase to be performed; in all other experiments soluble (Yamylase was used. In a comparison experiment, approximately 27-fold more antiserum was required to inactivate 1 unit of rat heart glycogen synthase in an extract than to inactivate 1 unit of pure rabbit skeletal muscle enzyme (data not shown).
In the experiments presented in Fig. 3, B and C, the heart extract was treated with normal sera (followed by centrifugation) prior to the addition of anti-glycogen synthase.
Without this treatment the antibody. antigen. glycogen synthase complex also contained co-precipitating low molecular weight compounds. This observation is illustrated in Fig. 4A. The immunoprecipitated '"P-labeled protein was examined by SDS-gel electrophoresis.
For immunoprecipitates isolated without prior treatment with normal sera, in addition to the band of radioactivity that co-migrated with glycogen synthase (88,000 daltons), two highly labeled low molecular weight compounds of less than 40,000 were detected. The same two low molecular weight compounds were present in samples treated with normal sera alone (Fig. 4A) or in a sample that had been incubated under identical conditions without addition of sera (data not shown); under neither of these latter two conditions was a band of 32P-labeled glycogen synthase present. Presumably the low molecular weight phosphocompounds are rapidly denaturing (or precipitating, or both) components not removed during the initial centrifugation but labile under the incubation conditions required for the immunoprecipitation.
These compounds could be precipitated by centrifugation of the original homogenate at 40,000 X g for 15 min, but this procedure also removed a large amount of the glycogen synthase (vide Fig. 1). The addition of normal sera, although not essential for precipitation of the low molecular weight phosphocompounds, did enhance their removal. Routinely for the isolation of 32P-labeled glycogen synthase from rat heart the extracts were initially incubated with normal sera, the precipitate was removed by centrifugation, and glycogen synthase subsequently was precipitated by the addition of antisera. Under these conditions only one band of 32P-protein was detected in the immunoprecipitate isolated from 32P-perfused rat heart; this band co-migrated with both skeletal muscle glycogen synthase (identified by Coomassie protein stain, Fig. 4) and with rat cardiac glycogen synthase (detected by Coomassie stain and by 32P from in vitro labeling; data not presented).
The peak of [32P]glycogen synthase was absent from the SDS gel of the immunoprecipitate when antiserum was pretreated with pure skeletal muscle glycogen synthase, thus adding additional confirmation that the 32P label identified by the SDS gel is covalently associated with the cardiac glycogen synthase and not coincidently co-migrating.
In all experiments reported in this manuscript the extent of 32P incorporation into glycogen synthase was determined from the SDS-gel electrophoresis profiles. As indicated in Fig. 3 quantitative immunoprecipitation of rat heart glycogen synthase required the subsequent addition of skeletal muscle glycogen synthase as carrier. The possibility that either the added pure skeletal muscle enzyme or the endogenous glycogen synthase of the heart extract was being phosphorylated during the extractions and incubations required for immunoprecipitation was examined (Fig. 4B). A lg rat heart contains 4.15 pmol of ATP (28). Therefore, with the assumption that there was no hydrolysis of endogenous ATP, exogenous high specific activity [Y-~P]ATP was added to a homogenate of a nonradioactive heart to give a final specific activity of the total ATP pool of 30 cpm/pmol, a value equivalent to the specific activity of the endogenous [T-~*P]-ATP obtained on perfusion of the heart with inorganic [""PIphosphate. Glycogen synthase was isolated from this supplemented extract by the standard immunoprecipitation technique. No 32P-protein was observed in the resultant immunoprecipitate (Fig. 4B). In a second control experiment 3 pg of 32P-labeled pure skeletal muscle glycogen synthase was added to 0.4 ml of a nonradioactive heart extract; 103% of the of Glycogen Synthase radioactivity was recovered in the resultant immunoprecipitate in a single sharp peak (Fig. 4C) In A, hearts were perfused with inorganic [3"P]phosphate and extracted as described under "Experimental Procedures." Three milliliters of normal serum (o---O) or 2.5 ml of antiserum (a---A) were added to 0.4 ml of extract and the solution was incubated 15 min at 30°C and then for 90 min at 4°C. The immunoprecipitates were sedimented by centrifugation for 15 min at 15,000 x g. The supernatant from the normal serum-extract incubation was added to 2.5 ml of antiserum (u), and the solution was incubated for 15 min at 3O"C, then for 90 min at 4°C. The immunoprecipitate was sedimented by centrifugation for 15 min at 10,000 X g. The immunoprecipitates were washed and electrophoresed as described under "Experimental Procedures" (see "Standard Isolation of Glycogen Synthase by Immunoprecipitation").
In B and C, two rat hearts were perfused in the absence of inorganic [32P]phosphate. One heart was stimulated 16 s with 2 pM epinephrine; the other heart was not stimulated. The extracts, prepared as described under "Experimental Procedures," were pooled. In B, 9.3 ~1 of 6.06 mM [y-'*P]ATP (1138 cpm/pmol) were added to 5 ml of extract. In C, 3 pg of pure skeletal muscle glycogen synthase containing 3237 cpm of 32P were added to 0.4 ml of extract. Immunoprecipitates were prepared and electrophoresed as described under "Experimental Procedures" (see "Standard Isolation of Glycogen Synthase by Immunoprecipitation"). The Coomassie-stained gels are of purified skeletal muscle glycogen synthase (upper) and immunoprecipitate from rat heart (lower). glycogen synthase from rat heart extract and in addition demonstrating the absence of post-homogenization phosphatase or protease activity.
The conditions as defined by the experiments depicted in Figs. 3 and 4 formed the basis for the experimental procedure utilized in these studies for the determination by immunoprecipitation of 32P-labeled glycogen synthase from perfused heart; the full protocol is presented under "Experimental Procedures" (see "Standard Isolation of Glycogen Synthase by Immunoprecipitation").
The Specific Activities of ["P/Phosphate, [y"'PJATP, and 13'P JGlycogen Synthase During Perfusion-We have previously shown that equilibration of inorganic [""Plphosphate into the myocardium occurs slowly and that the transport of inorganic ["'PIphosphate across the myocardial sarcolemma is the rate-limiting step in the formation of intracellular [y-"'P]ATP (18). The characteristics of the cardiac perfusion system used in this study are presented in Fig. 5. During perfusion the specific activity of the perfusate inorganic phosphate declined substantially and by 50 min it was 25% of that of the initial value. This occurred without a change in the extracellular inorganic phosphate concentration (0.12 mM), even though the perfusate inorganic phosphate concentration was %oth of the physiological value. Uptake of inorganic [32P]phosphate occurs as an exchange of intracellular and extracellular phosphate with no net efflux from the myocardium and no net loss of total cellular phosphate. Intracellular inorganic phosphate and the y-phosphate of ATP are in rapid equilibrium (18). During 60 min of perfusion the specific activity of [y-32P]ATP continued to increase (Fig. 5), but because of the extensive total exchangeable phosphate pool of the myocardial cell it only reached a specific activity that was 8.5% of that of the extracellular phosphate even after a 50min perfusion.  (29). Intracellular [y-"'P]ATP was measured as described previously (18). The concentration of extracellular inorganic phosphate remained constant during perfusion.
by guest on March 24, 2020 http://www.jbc.org/ Downloaded from glycogen synthase during cardiac perfusion is complex. As indicated by the experiment presented in Fig. 6, following a brief lag (< 10 min) uptake of ["'PIphosphate into glycogen synthase occurred throughout the 60-min perfusion. During this period the I/D ratio of the synthase remains constant; this R2P incorporation presumably represents an exchange rather than a net phosphorylation of the protein. Throughout this period the increase in the "P content of glycogen synthase paralleled the increase in the specific activity of the intracellular [Y-'"P]ATP.
To further elucidate this process hearts were perfused with 0.12 mM inorganic [""PIphosphate for 30 min at which time KH,PO, and NaH2P04 were added to the perfusate so to reduce the specific activity of the perfusate inorganic ["'PIphosphate to a value equal to the specific activity of the intracellular [y-""PIATP. Under these conditions the specific activity of the [y-"*P]ATP remained constant during an additional 30-min perfusion (Fig. 6B). During this subsequent 30-min period, the radioactivity in glycogen synthase continued to increase at an incorporation rate of 0.3 mol of phosphate incorporated/90,000 daltons of enzyme/30 min (Fig. 6A).
The possibility that this slower increase in "'P-incorporated into glycogen synthase represented phosphorylation of newly synthesized enzyme was investigated.
Perfusion for 30 min with 20 pM cycloheximide did not cause any reduction in the specific enzyme activity of glycogen synthase in the resultant heart homogenate (not shown). This level of cycloheximide effectively blocks protein synthesis in the perfused rat heart (30).
Effect of Epinephrine on Phosphorylation of Glycogen Synthase-As indicated previously (Fig. 2) epinephrine administration had no effect on the per cent of glycogen synthase in the D-form. In addition, the administration of 2 PM epinephrine for 10 or 20 s to a rat heart perfused with inorganic ["2P]phosphate caused no significant change in the amount of phosphate incorporated (Fig. 7). Under identical conditions Hearts were perfused with 32P by recycling media under the standard conditions as presented under "Experimental Procedures" (open symbols). 32P content of glycogen synthase was determined as described under "Experimental Procedures" (see "Standard Isolation of Glycogen Synthase by Immunoprecipitation").
For one set of hearts (solid symbols) 184 ~1 of 0.1 M KHPOI, pH 7.0, and 205 ~1 of 0.1 M NaHPO+ pH 7.0, were added to the 17.5 ml of recycling perfusion media and perfusion continued for the indicated periods.
To permit the compilation of the several sets of experiments required to obtain these data, the values of both the '*P content of glycogen synthase and specific activity of  both activation and phosphorylation of phosphorylase kinase* and troponin I (16) are readily demonstrated.
Comparison between Heart Glycogen Synthase Phosphorylated in Vivo, Heart Glycogen Synthase Phosphorylated in Vitro, and Skeletal Muscle Glycogen Synthase Phosphorylated in Vitro-We have compared the in vivo sites of phosphorylation of rat heart glycogen synthase with the sites phosphorylated in vitro by the CAMP-dependent protein kinase and by the CAMP-independent glycogen synthase kinase(s). Additionally, these data have been compared to the already determined sites in skeletal muscle glycogen synthase. Soderling et al. (10,31) have determined by limited trypsin digests and by CNBr digests that there are two major phosphorylation sites in skeletal muscle glycogen synthase. Both sites can be phosphorylated by the CAMP-dependent protein kinase and by the cyclic nucleotide-independent glycogen synthase kinase. However, the CAMP-dependent protein kinase preferentially phosphorylates a site which can be removed from glycogen synthase by mild trypsinization and is therefore termed the "trypsin-sensitive" site. The CAMP-independent glycogen synthase kinase preferentially phosphorylates a site which is not degraded by mild trypsin treatment and, therefore, is termed the "trypsin-insensitive" site. The two sites can also be distinguished by cyanogen bromide digestion of phosphorylated glycogen synthase and the subsequent separation of the phosphopeptides by SDS-gel electrophoresis.
The larger CNBr phosphopeptide contains the "trypsin-sensitive" site and the lower molecular weight phosphopeptide contains the "trypsin-insensitive" site. In agreement with Soderling et al. (lo), SDS gels of CNBr digests of pure rabbit skeletal muscle glycogen synthase that had been phosphorylated by the pure catalytic subunit of CAMP-dependent protein kinase exhibited two peaks of radioactivity; of the radioactivity in the two peaks 67% was in the higher molecular weight phosphopeptide (Fig. 8A). Treatment of intact glycogen synthase with trypsin (1:125 w/w) for 30 min at 30°C prior to precipitation at 4°C with 10% trichloroacetic acid resulted in the elimination of the heavier CNBr phosphopeptide from the ensuing SDS gel (data not presented). An essentially identical pattern was obtained for partially purified rat heart glycogen synthase. The SDS-gel electrophoresis profile (Fig. 8B) of cyanogen bromide phosphopeptides exhibited two major peaks of radioactivity with mobilities identical to those obtained for the skeletal muscle enzyme; 66% of the radioactivity in the two major phospho- peptides was in the larger species. Aliquots of the cyanogen bromide hydrolysates of both rat heart and rabbit skeletal muscle glycogen synthase were combined and electrophoresed (Fig. 8C). The phosphopeptides derived from rat heart glycogen synthase co-migrated with those from the rabbit skeletal muscle enzyme.
An identical comparison has been performed of sites in rat heart and rabbit skeletal muscle glycogen synthase phosphorylated in vitro by the CAMP-independent synthase kinase (Fig. 9). The phosphorylation reactions were performed in the presence of the inhibitor protein of CAMP-dependent protein kinase, present at a 300-fold excess of the amount necessary for 90% inhibition of all the kinase activity had it been CAMPdependent. For skeletal muscle glycogen synthase phosphorylated for 6 min at 3O"C, the SDS gel of the resultant cyanogen bromide hydrolysate exhibited one predominant peak of radioactivity, which contained 78% of the radioactivity (Fig. 9A). Following phosphorylation for 30 min at 3O"C, the radioactivity was observed to be in two major phosphopeptides (Fig. 9B); 65% of the radioactivity in these two phosphopeptides was in the lower molecular weight species. The two peptides phosphorylated by the CAMP-indpendent synthase kinase co-migrated with those phosphorylated by the CAMPdependent enzyme (CL Fig. 8A) but, in accord with the data of 50 Purified rat heart and rabbit skeletal muscle glycogen synthases were phosphorylated by rat heart CAMP-independent synthase kinases. Following phosphorylation each sample was prepared, electrophoresed, and counted as in Fig. 8. In A and C and in B and D phosphorylation was for 6 and 30 min at 3O"C, respectively. In A and B and in C and D the subtrates were rabbit skeletal muscle and rat heart glycogen synthases, respectively. Soderling et al. (lo), the CAMP-dependent enzyme preferentially phosphorylated the site in the larger peptide, whereas the CAMP-independent kinase preferentially phosphorylated a site in the smaller species. A similar, although less definitive result was observed for the cardiac glycogen synthase. Following phosphorylation of the cardiac glycogen synthase with the CAMP-independent kinase for 6 min, phosphate was incorporated predominantly into two phosphopeptides (Fig. 9C). The two major phosphopeptides co-migrated with those from skeletal muscle glycogen synthase as catalyzed by the cyclic nucleotide independent kinase (cf. Fig. 9B) and with those from cardiac glycogen synthase as catalyzed by the CAMPdependent protein kinase (cf Fig. 8B). Seventy per cent of the radioactivity in the two phosphopeptides was in the lower molecular weight species (Fig. 9C). Following phosphorylation of cardiac glycogen synthase for a longer time period (30 min) with the cyclic nucleotide-independent kinase the radioactivity in the heavier phosphopeptide increased to over 50% of the total radioactivity in the two phosphopeptides (Fig. 9D). For the cardiac glycogen synthase particularly in comparison to the skeletal muscle enzyme, both the CAMP-dependent and the CAMP-independent protein kinase catalyzed the incorporation of low levels of phosphate into peptides other than those containing the two major sites.
The sites of glycogen synthase phosphorylated in the intact heart have been examined by an identical procedure.
To ensure that the phosphopeptides identified were derived from glycogen synthase the 32P-labeled glycogen synthase was isolated from the perfused heart extract by immunoprecipitation, the intact protein electrophoresed on SDS gel, the band containing the 32P-labeled glycogen synthase was dissected out, and CNBr digestion was performed on the protein eluted from the gel (details of the methods used are given under "Experimental Procedures"). The SDS gels of the resultant CNBr phosphopeptides exhibited two major peaks of radioactivity of R, 0.60 and 0.90 (Fig. lOA); 57% of the radioactivity was in the higher molecular weight species. These phospho-by guest on March 24, 2020 http://www.jbc.org/ Downloaded from synthase kinase activity could be detected with further elution by 1 M NaCl. In contrast, the CAMP-independent synthase kinase studied by Nimmo et al. (6), Brown et al. (38), and Soderling et al. (10) elutes from phosphocellulose at 0.5 to 1.0 M NaCl. For our study, we collected all the cardiac cyclic nucleotide independent glycogen synthase kinases that eluted from phosphocellulose between 0.3 and 1.0 M NaCl. The mechanisms whereby the activities of these kinases are regulated are as yet unknown. TOP Gel Slice Number   FIG. 10. SDS-gel electrophoresis of cyanogen bromide 32Ppeptides. Rat heart [32P]glycogen synthase from 4 ml of an extract of a rat heart that had been perfused with 5 mCi of inorganic [""PIphosphate for 30 min was isolated by immunoprecipitation followed by preparative electrophoresis. Pure rabbit skeletal muscle glycogen synthase was phosphorylated by the catalytic subunit of CAMPdependent protein kinase. Cyanogen bromide phosphopeptides were prepared and electrophoresed. Full details of these methods are presented under "Experimental Procedures." In A, 50 ~1 of a CNBr hydrolysate of rat heart glycogen synthase phosphorylated in viuo were applied to a 10% SDS gel. In B, 25 ~1 of a CNBr hydrolysate of rabbit skeletal muscle glycogen synthase were applied to a 16% SDS gel. In C, 50 ~1 of rat heart and 25 ~1 of rabbit skeletal muscle glycogen synthase CNBr hydrolysates were applied to a gel.
Both the CAMP-dependent protein kinase and the CAMPindependent glycogen synthase kinase catalyze the conversion of skeletal muscle glycogen synthase from the I to the D activity form; each catalyzes the phosphorylation of the enzyme but the site preferentially phosphorylated is different for CAMP-independent and CAMP-dependent kinases. A minimum of four covalently different forms of glycogen synthase exist. These have been termed glycogen synthase a (dephospho form), glycogen synthase bl and b2 (the monophospho forms phosphorylated in the CAMP-dependent site (trypsinsensitive, high molecular weight CNBr digest fragment) or CAMP-independent (trypsin-insensitive, low molecular weight CNBr digest fragment) site, respectively), and glycogen synthase b1,2 (diphospho form). The I/D ratio of each of the forms have been variously reported to be: a, 0.8 to 1 .O; bl, 0.1 to 0. 3; bp, 0.1 to 0.3; and b,,~, 0.01 to 0.1 (2,6,10,37). In addition, Soderling et al. (10) have reported that there is an additional site of phosphorylation on skeletal muscle glycogen synthase catalyzed by the CAMP-dependent kinase. It is apparent, therefore, that an understanding of the regulation of glycogen synthase requires not only the determination of the I/D ratio but also an elucidation of the site phosphorylated.
peptides co-migrated with those obtained by in vitro phosphorylation of the glycogen synthase (Fig. 10, B and C).

DISCUSSION
In vitro studies utilizing glycogen synthase from liver, skeletal muscle, adipose, brain, spleen, and heart (32), and kidney (33) have demonstrated that the enzyme is phosphorylated and inactivated by CAMP-dependent protein kinase. In agreement with the consensus of observations by other investigators (11,13,14,34,35) our results indicate that rat heart glycogen synthase in the basal (unstimulated) tissue exists primarily in the D-form, indicating that it would be phosphorylated even in the absence of elevated CAMP levels. Studies with isolated diaphragm have demonstrated that agents known to elevate intracellular CAMP levels caused a rapid inactivation of glycogen synthase (36); however, conflicting results have been renorted from studies on the intact heart. In the nerfused rat heart epinephrine, which elevates CAMP, has been reported to increase, decrease, or have no effect (11,13,14) on the per cent of glycogen synthase in the I-form. In the in situ rat heart the per cent of glycogen synthase in the I-form was observed to increase (34) or decrease slightly (35) upon epinephrine administration.
Under the experimental conditions used in this investigation epinephrine has no significant effect on glycogen synthase activity nor on the incorporation of phosphate into the enzyme.
Recently, glycogen synthase kinase(s) other than CAMP- The data presented in this manuscript indicate that CNBr phosphopeptides derived from rat heart glycogen synthase phosphorylated in vitro are closely similar (or identical) to those of the skeletal muscle enzyme and that the apparent preferential sites of phosphorylation by the CAMP-dependent and CAMP-independent protein kinases are the same for the cardiac and skeletal muscle glycogen synthases (Figs. 8 and 9). dependent protein kinase and its catalytic subunit have been described (6)(7)(8). These kinases have in common the ability to bind to phosphocellulose. However, they were eluted at differing ionic strengths of buffer. The skeletal muscle CAMPindependent synthase kinase(s) studied by Schlender and Reimann (23) and Itarte et al. (37) was fully eluted by 0.5 M NaCl or KCl. Only trace amounts of CAMP-independent On perfusion of rat hearts with inorganic ["'PIphosphate, there was an increase in radioactivity in glycogen synthase (Fig. 6). This increase in 'lP in glycogen synthase was probably not due to phosphorylation of newly synthesized enzyme since there was no change in the specific enzyme activity of glycogen synthase in perfusions in the presence of cycloheximide at a concentration sufficient to block protein synthesis. The increase in "'P in glycogen synthase may have been due to either net phosphorylation of the protein or turnover of phosphate. Since predominantly only those sites known to regulate enzyme activity were observed to become phosphorylated ( Fig.  10) and there was no measurable effect on enzyme activity (per cent I-form or total activity), this would argue against a change in the net phosphate (i.e. "P + "'P) content; nevertheless, it should be recognized that the per cent I measurement is insensitive to change in phosphate content for enzyme already containing greater than 1 mol of phosphate/subunit. Estimates of the stoichiometry of the ["2P]phosphate incorporated into glycogen synthase due to net turnover are complicated by five factors: (a) the specific activity of intracellular [y-'"P]ATP is changing during the perfusion, (b) the specific activity of phosphate incorporated into protein may be greater than that of the average specific activity of intracellular [y-""P]ATP as it has been reported in skeletal muscle for the phosphorylation of phosphorylase (39), (c) cardiac glycogen synthase has not been obtained in a highly homogeneous form to permit a determination of the specific enzymatic activity, (d) although skeletal muscle glycogen synthase has been