T H E J O U R N A L OF BIOLOGICAL C H E M I S T R Y Vol. 2.58, No. 17. Issue of September 10, pp. 10710-10719.1983 Pnnled in U.S.A.

Hormonal Control of Glycogen Synthase in Rat Hemidiaphragms EFFECTS OF INSULIN AND EPINEPHRINE ON THE DISTRIBUTION OF PHOSPHATE BETWEEN TWO CYANOGEN BROMIDE FRAGMENTS*

(Received for publication, September 10, 1982, and in revised form, February 9, 1983)

John C. Lawrence, Jr.$ and Jeffrey F. Hiken From the Department of Pharmacology, Washington University School of Medicine, St. Louis, Missouri 63110

Anna A. DePaoli-Roach5 and Peter J. Roach7 From the Department of Biochemistry, Indiana University School of Medicine, Indianapolis, Indiana 46223

The effects of insulin and epinephrine on the phos- phorylation of glycogen synthase were investigated using rat hemidiaphragms incubated with [ 3 2 P ] p h ~ ~ - phate. Antibodies against rabbit skeletal muscle gly- cogen synthase were used for the rapid purification of the "'P-labeled enzyme under conditions that pre- vented changes in its state of phosphorylation. The purified material migrated as a single radioactive spe- cies (MaPp = 90,000) when subjected to electrophoresis in sodium dodecyl sulfate.

Insulin decreased the [32P]phosphate content of gly- cogen synthase. This effect occurred rapidly (within 15 min) and was observed with physiological concen- trations of insulin (25 microunits/ml). The amount of [32P]phosphate removed from glycogen synthase by either different concentrations of insulin or times of incubation with the hormone was well correlated to the extent to which the enzyme was activated. Epi- nephrine (10 JLM) inactivated glycogen synthase and increased its content of [32P]phosphate by about 50%.

Cleavage of the immunoprecipitated enzyme with cyanogen bromide yielded two major ""P-labeled frag- ments of apparent molecular weights equal to approx- imately 28,000 and 15,000. The larger fragment (Fragment 11) displayed electrophoretic heterogeneity similar to that observed with the corresponding CNBr fragment (CB-2) from purified rabbit skeletal muscle glycogen synthase phosphorylated by different protein kinases. Epinephrine increased [32P]phosphate content of both fragments; however, the increase in the radio- activity of the smaller fragment (Fragment I) was more pronounced. Insulin decreased the amount of [""PI phosphate present in Fragments I and I1 by about 40%. The results presented provide direct evidence that both insulin and epinephrine control glycogen synthase ac- tivity by regulating the phosphate present at multiple sites on the enzyme.

* 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.

$ Recipient of National Institutes of Health Grant AM28312 in support of this work.

0 Recipient of Fellowship AM06208 from the National Institutes of Health.

TI Recipient of Research Career Development Award AM01089 from the National Institutes of Health. Recipient of National Institutes of Health Grants AM27221 and AM27240 and a grant from the Grace M. Showalter Foundation.

Over two decades ago the observation was made that insulin acted on rat diaphragm to increase glycogen synthase activity when measured in the absence of glucose-6-P but not in its presence (1). This finding led to the proposal that the enzyme existed in active (I) and inactive (D) forms. Epinephrine was later shown to decrease the percentage of glycogen synthase I activity ( 2 ) . This provided a scheme for hormonal control of glycogen synthesis via modulation of glycogen synthase activ- ity by insulin and epinephrine. The discovery of CAMP- dependent protein kinase in skeletal muscle which phospho- rylated and inactivated glycogen synthase provided a mecha- nism by which the stable effects of epinephrine could be explained (3). By activating adenylate cyclase and increasing CAMP, epinephrine would activate CAMP-dependent protein kinase resulting in phosphorylation and inactivation of gly- cogen synthase. Insulin action has been more difficult to fit into the scheme as no effect of insulin can be demonstrated on the concentration of CAMP in skeletal muscle (4,5).

A major criticism of most previous studies of the hormonal regulation of glycogen synthase is that changes in the phos- phorylation of the enzyme have not been directly measured but inferred from changes in kinetic properties of the enzyme. Glycogen synthase has multiple sites per subunit that can be phosphorylated in vitro (6). As many as 7 sites are phospho- rylated in vivo, and the sequences of amino acids surrounding these sites have been identified ( 7 ) . In addition to CAMP- dependent protein kinase, at least four other protein kinases have now been shown to phosphorylate glycogen synthase in vitro (8-10). These kinases exhibit different specificities for the different sites on glycogen synthase and their actions may generate a wide variety of phosphorylated forms of the enzyme (8). Furthermore, there is no guarantee that all of the protein kinases or phosphorylation sites important in the regulation of glycogen synthase have been discovered. Therefore, it is not possible to deduce the actual phosphorylation state of the enzyme from measurements of enzymatic activities.

It is important to determine which sites on glycogen syn- thase are phosphorylated or dephosphorylated as a result of hormone action. Together with the information that exists on the site specificities of the different protein kinases and phosphatases, it might be possible to identify which enzymes mediate the actions of insulin and epinephrine. Several ap- proaches could be taken to identify changes in the phospho- rylation states of individual sites on glycogen synthase pro- moted by hormones. We have used immunological procedures for the rapid purification of ["P]glycogen synthase from hormonally treated rat hemidiaphragms which had been in- cubated with [32P]phosphate. After cleavage with CNBr two

10710

by guest on June 1, 2018http://w

ww

.jbc.org/D

ownloaded from

Hormonal Control of Site-specific Phosphorylation 10711

"P-labeled fragments were obtained as have been described in studies of purified rabbit skeletal muscle glycogen synthase (7, 10-12). More [32P]phosphate was associated with both fragments after treatment of muscles with epinephrine, al- though the hormone caused a greater percentage increase in the radioactivity of the smaller fragment. Insulin caused a 40 to 50% decrease in the content of [32P]phosphate in both fragments. The implications of multisite control of glycogen synthase by insulin and epinephrine are discussed in relation to proposed mechanisms of action of the hormones.

EXPERIMENTAL PROCEDURES

Incubation of Hernidiaphragms-Hemidiaphragms from 80- to 100- g male rats (Sprague-Dawley) were trimmed of extraneous tissue leaving a thin border of ribs as described by Goldberg et al. (13). Muscles (typically from 6 animals) were incubated a t 23 "C in 250 ml of medium A which was composed of 0.1 mM sodium phosphate and essentially the same concentrations of amino acids, vitamins, and other components as Dulbecco's modified Eagle's medium (14). All incubation media were gassed with a mixture of 95% O Z / ~ % C 0 2 and were supplemented with penicillin G (31 pg/ml) and streptomycin (10 pglml). After 1 h hemidiaphragms were transferred to a single vessel containing medium A (2 ml/hemidiaphragm) and, where ap- propriate, [32P]phosphate (0.2-0.4 mCi/ml). The tissue was incubated at 36 "C for 5 h and then transferred to multiple incubation vessels (3 hemidiaphragms each) containing 6 ml of medium B. This medium was composed of 119 mM NaCI, 4.6 mM KCI, 2.4 mM CaC12, 1.2 mM MgSO,, 24.8 mM NaHC03, 5 mM glucose, 0.1 mM potassium phos- phate and [32P]phospbate (same specific activity as medium A), Tissue was incubated for 1 h in medium B before being frozen in liquid nitrogen. Unless otherwise specified, insulin (25 mU/ml) was added for the last 30 min and epinephrine (10 p ~ ) for the last 10 min before the incubation was terminated.

Preparation of Extracts-Frozen muscles were manually ground using a porcelain mortar and pestle chilled with liquid nitrogen. Typically, powdered tissue from 3 hemidiaphragms was homogenized in homogenization buffer (1 ml) composed of 100 mM KF, 10 mM EDTA, 2 mM EGTA,' 2 mM potassium phosphate, 10 mM benzami- dine, 0.1% Triton X-100, and 50 mM Tris/HCI (pH 7.0 a t 0 "C). After centrifugation of homogenates at 20,000 X g for 30 min, the super- natants were removed for assays of enzymatic activities and immu- noprecipitations of 32P-labeled glycogen synthase.

Measurement of Glycogen Synthase and Phosphorylase Activities- Samples of the extracts were diluted &fold with homogenization buffer. Glycogen synthase activities were measured essentially as described by Thomas et al. (15) and phosphorylase activities as described by Gilboe et al. (16). For glycogen synthase activity, diluted samples (30 pl) were added to 60.~1 aliquots of reaction mixtures composed of 20 mM EDTA, 25 mM KF, 10 mg/ml of rabbit liver glycogen, 7.5 mM UDP-glucose and 50 mM Tris/HCl (pH 7.8 at 30 "C) plus or minus 15 mM glucose-6-P. Phosphorylase reaction mixtures were composed of 200 mM KF, 10 mg/ml of glycogen, and 100 mM glucose-1-P (pH 6.1 a t 30 "C) plus or minus 3 mM AMP. Reactions were run a t 30 "C for 20 min in the absence and presence of UDP- [U-"C]glucose (600 pCi/mmol) or [U-"C]glucose-l-P (16.7 pCi/ mmol). Enzyme activities were determined from the amount of I4C. labeled substrates incorporated into glycogen. Incubations performed in the absence of 14C-labeled substrates enabled a direct correction for the amount of 32P detected in the I4C channel of the liquid scintillation spectrometer. Results for glycogen synthase activity are presented as total activity (defined as that activity measured in the presence of 10 mM glucose-6-P) or as an activity ratio (activity minus glucose-6-P/activity plus glucose-6-P). Phosphorylase activity is ex- pressed as total activity (measured in the presence of 2 mM AMP) or as an activity ratio (activity minus AMP/activity plus AMP). Activ- ities are based on the amount of protein in the extracts as determined by the method of Lowry et al. (17).

Measurements of Concentrations of ATP and Specific Activities of

' The abbreviations used are: EGTA, ethylene glycol bis(B-amino- ethyl ether)-N,N,N',N'-tetraacetic acid; SDS, sodium dodecyl SUI- fate; CNBr, cyanogen bromide; Mapp, apparent molecular weight determined hy polyacrylamide gel electrophoresis in the presence of SDS (see Ref. 12); anti-GS Ab-, purified IgG fraction from guinea pigs immunized with glycogen synthase.

[~-~'P]ATP"Powdered tissue (approximately 50 mg) was homoge- nlzed a t 0 "C in 1 ml of 0.5 N perchloric acid using a Polytron homogenizer (Brinkmann Instruments) set at top speed for 15 s. After centrifugation a t 10,000 X g for 30 min, the pellets were dissolved in 1.5 ml of 6.7% (w/v) SDS in 0.13 N NaOH and the supernatants were neutralized with KHCO,. After 30 min the neutralized samples were centrifuged to remove potassium perchlorate. The specific activity of [ Y - ~ ~ P ] A T P was determined by procedures described by England and Walsh (18). The concentration of ATP was measured by the method of Williamson and Corkey (19) and is expressed in terms of tissue protein measured in the detergent-solubilized pellets (17).

Preparation of Antibodies to Glycogen Synthase-Male guinea pigs (400-600 g) were immunized by subcutaneous injection of Freund's complete adjuvant emulsified with glycogen synthase (100 pg per animal) purified from rabbit skeletal muscle in a dephosphorylated form (20). A booster injection was administered after 4 weeks and animals were bled 2 weeks later. The blood was allowed to clot and the sera were collected. Unfractionated sera contained both phospha- tase and protease activities and were not suitable in the immunopre- cipitation of glycogen synthase. The following procedure resulted in a highly purified IgG fraction in which neither of these activities were detected. Sera from 12 animals were pooled and incubated a t 56 "C for 30 min. After centrifugation a t 20,000 X g for 1 h, 25 ml of saturated ammonium sulfate were added to 30 ml of the supernatant, and the suspension was centrifuged at 10,000 X g for 30 min. The pellet was suspended in 12 ml of 10 mM potassium phosphate (pH 8) and ammonium sulfate was added to 40% saturation. After centrif- ugation a t 10,000 x g for 30 min, the pellet was dissolved in 10 ml of 10 mM potassium phosphate (pH 8) and dialyzed for 20 h against several changes of 50 mM potassium phosphate. After centrifugation a t 20,000 x g for 1 h, the supernatant was applied a t 15 ml/h to a column containing 200 ml of DEAE-Affi-Gel blue (Bio-Rad) equili- brated with 50 mM potassium phosphate (pH 8). The first 80 ml containing protein were collected and the IgG was precipitated with ammonium sulfate (45% saturation). The purified antibodies were collected by centrifugation, dissolved in 50 mM sodium phosphate (pH 7.41, and dialyzed against the same buffer before storage a t -20 "C in small aliquots a t 10 mg/ml prior to use. Nonimmune IgG was prepared by the same procedure from the sera of nonimmunized guinea pigs. To assess purity, samples were subjected to electropho- resis using polyacrylamide gels (15%) in the presence of SDS before staining with Coomassie blue. Only bands corresponding to the heavy and light chains of IgG were observed, and anti" Ab- and control IgG were indistinguishable.2

Binding of Antibodies to Glycogen Synthase-Antibodies were in- cubated a t 23 "C with samples of phosphorylated and nonphospho- rylated glycogen synthase purified from rabbit skeletal muscle (Fig. 1A). Glycerol (25%) was used in these incubations with the purified enzymes to prevent aggregation (21). After 20 min the samples were cooled a t 0 "C and a suspension of inactivated Staphylococcus aureus of strain Cowan I (Pansorbin, Calbiochem-Behring) was added to bind the IgG (22). The bacteria were centrifuged and resuspended several times in fresh homogenization buffer before use to remove fines which accumulated during storage. After 30 min the samples were centrifuged a t 3000 X g for 5 min and the amount of glycogen synthase activity removed from solution was determined. The results presented in Fig. 1A show that the same concentrations of anti- glycogen synthase antibodies were effective in removing the activity of both phosphorylated and nonphosphorylated glycogen synthase. These results demonstrate that the antibodies bind to both unphos- phorylated glycogen synthase and enzyme phosphorylated by C sub- unit of CAMP-dependent protein kinase. In the experiment presented in Fig. lB, antibodies were incubated a t 23 "C with extracts of hormonally treated hemidiaphragms. After addition of bacteria and centrifugation, glycogen synthase activities remaining in the super- natants were measured. The same concentrations of anti-glycogen synthase antibodies were effective in removing glycogen synthase activity from extracts of control and insulin-treated hemidiaphragms. Slightly lower concentrations were required in extracts of tissues that had been incubated with epinephrine. However, glycogen synthase activity was completely removed by appropriate concentrations of antibodies in all three extracts. These results directly demonstrate tbat the anti-glycogen synthase antibodies recognize the different forms of the enzyme which exist in muscle following incubation with

J. C. Lawrence, Jr., J. F Hiken, A. A. DePaoli-Roach, and P. J. Roach, unpublished observation.

by guest on June 1, 2018http://w

ww

.jbc.org/D

ownloaded from

10712

IgG(pg/ml) IgG(pg/ml )

FIG. 1. Binding of antibodies to purified glycogen synthase and enzymes from hormonally treated hemidiaphragms. A, glycogen synthase (80 pg/ml), purified from rabbit skeletal muscle in a dephosphorylated form (20), was incubated at 30 "C in a solution containing 0.1 mM ATP, 6 mM MgC1, 25% glycerol, 50 mM Tris/ HC1 (pH 7.5), and 1 pg/ml of C subunit of CAMP-dependent protein kinase. The phosphorylation reaction was terminated after 30 min by adding KF and EDTA to final concentrations of 100 and 10 mM, respectively. Under these conditions, glycogen synthase was phosphorylated to a level of 1.4 phosphates/subunit as determined from ["P]phosphate incorporation in a sample incubated under identical conditions except in the presence of [y-"P] ATP (1600 cpm/pmol). Another sample of enzyme was incubated as described above except in the absence of ATP and C subunit and is designated nonphosphorylated glycogen synthase. Increasing concentrations of control IgG (0, V) or anti-glycogen synthase antibodies (0, V) were incubated at 23 "C in 30 p1 of homogenization buffer containing 25% glycerol and 0.4 Kg of either phosphorylated glycogen synthase (V, V) or nonphosphorylated enzyme (0, 0). After 20 min 100 pI of a 40% suspension of inactivated bacteria were added to each sample and incubated at 0 "C. The samples were centrifuged after 30 min to pellet the bacteria, and glycogen synthase activities (in the presence of 10 mM glucose-6-P) were measured in the supernatants. These activities and those observed in supernatants of samples incubated without immunoglobulins were used to estimate the percentage of glycogen synthase activity removed a t each antibody concentration. B, hemidiaphragms were incubated in medium A (without [32P]phosphate) and then transferred to medium B and incubated without ( 0 , O ) or with 250 milliunits/ ml of insulin (A, A) or 10 p~ epinephrine (W, 0) before extracts were prepared as described under "Experimental Procedures." The activity ratios and total activities of glycogen synthase in the different extracts were as follows: control, 0.12, 149 milliunits/ml; epinephrine, 0.06, 153 milliunits/ml; and insulin, 0.26, 155 milliunits/ml. One unit of enzyme activity is that amount catalyzing the utilization of 1 pmol of substrate in 1 min a t 30 "C. Samples (30 p1) of extract were added to solutions (20 pl) of increasing concentrations of control IgG (0, A, 0) or anti-glycogen synthase antibodies (0, A, W) and incubated a t 23 "C for 20 min. Inactivated bacteria (100 p1 of a 20% suspension) were added to each and incubated at 0 "C. After 30 min the samples were centrifuged and glycogen synthase activities (in the presence of 10 mM glucose-6-P) were measured in the supernatants. The results presented represent the percentages of glycogen synthase activities removed by the antibodies.

insulin and epinephrine. Interestingly, slightly lower concentrations of antibodies were effective in removing glycogen synthase from the muscle extracts (Fig. 1B) than were needed to remove the purified enzymes from solution (Fig. Ut). At present we do not have a definitive explanation for this observation. In other experiments we have studied the binding of purified 32P-laheled glycogen synthase by determining the radioactivity retained by the bacteria following sev- eral washes to remove unbound enzyme. All of the 32P-labeled enzyme could be bound by the antibody and the binding could be inhibited by unlabeled glycogen synthase.' Taken together, these results indi- cate that the heteroclonal antibodies obtained by immunizing guinea pigs with unphosphorylated glycogen synthase recognize both phos- phorylated and nonphosphorylated enzyme.

Immunoprecipitation of p2PlGlycogen Synthase from Rat Hemidia- phragms-Extracts (300 pl) of hemidiaphragms incubated with ["PI phosphate were incubated a t 23 "C with 150 pg of either anti-GS antibodies or control IgG. As shown in Fig. lB, the concentration of 0.5 mg/ml of anti-GS antibody is sufficient to bind 80 to 90% of the enzyme present. Although desirable, it was not possible to use a large excess of antibodies because the increase in nonspecific binding observed with increasing concentrations of immunoglobulins led to high levels of "P-labeled contaminants in the immunoprecipitate. After 20 min the samples were placed on ice and to each was added 60 pl of a 10% (w/v) suspension of PansorbinP After 30 min a t 0 "C,

Nonspecific contamination by 32P-labeled proteins is proportional to the amount of bacteria used. We attempted to reduce this contam-

the samples were centrifuged a t 3000 X g for 2 min. The bacteria were resuspended in 1 ml of Ca" and Mgz'-free phosphate-buffered saline containing 1% crystalline bovine serum albumin and then washed 5 times a t 0 "C by sequential steps of centrifugation and resuspension; the last two washes were performed in buffer without albumin.

An estimation of the amount of glycogen synthase activity immu- noprecipitated under the standard conditions can be obtained from the results presented in Table 1. In these experiments extracts of control hemidiaphragms and muscles incubated with insulin and epinephrine were incubated with 0.5 mg/ml of control IgG or anti- GS antibodies before the bacterial suspension was added. After in- cubation and centrifugation to pellet the bacteria, glycogen synthase activity was measured in the supernatant. Essentially the same per- centage of activity was removed from all three extracts by the anti- bodies? Furthermore, the activity ratios in the supernatant were

ination by limiting the bacteria used to the smallest amount sufficient to bind all of the IgG present. Different preparations of S. aureus vary significantly with respect to binding capacity for IgG. The amount of the particular lot (Pansorbin, Lot 130155) used in the standard procedure was not sufficient to bind all of the guinea pig antibodies. This is the reason that the amounts of activity removed in the presence of the antibodies in the experiments presented in Table I are less than those removed by equivalent concentrations of antibody in the experiment shown in Fig. 18, where excess S. aureus was used.

by guest on June 1, 2018http://w

ww

.jbc.org/D

ownloaded from

Hormonal Control of Site-specific Phosphorylation

TABLE I Immunoprecipitation of glycogen synthase activities

Hemidiaphragms were incubated in medium A (without %'P) for 5 h, then transferred to medium B and incubated with and without insulin (25 milliunits/ml) or epinephrine (10 p ~ ) . Extracts were prepared and samples (150 pl) were incubated with antibodies (0.5 mg/ml) at 23 "C for 20 min. S. aureus (30 pl of a 10% suspension) was added and incubated at 0 "C. After 30 min the samples were centrifuged and glycogen synthase activities were measured in the supernatants. The bacterial pellets were suspended and washed a t 0" (4 times) with Ca2+- and Me- f ree phosphate-buffered saline. After a final resuspension in homogenization buffer, glycogen synthase activity was measured. The activity ratios and total activities, respectively, of the different extracts not incubated with immunoglobulins were as follows: control, 0.24 -C 0.03, 298 f 10 milliunits/ml; epinephrine, 0.13 f 0.01, 239 % 14 milliunits/ml; and insulin, 0.38 f 0.04, 246 f 4 milliunits/ml. The results presented are the mean values (fS.E. where indicated) of three experiments and represent activity ratios or total glycogen synthase activity (milliunits per ml relative to the volume of the original tissue extract).

Control IgG Anti-GS Ab- Fraction Hormonal treatment

Activity ratio Total activity Activity ratio Total activity

milliunitslml milliunitslml

Supernatant Control 0.27 f 0.09 246 17 0.24 f 0.06 150 +- 15 Epinephrine, 10 p M 0.19 f 0.07 183 f 2 0.13 f 0.03 107 +- 3 Insulin, 25 milliunits/ml 0.40 f 0.09 196 12 0.37 f 0.07 123 +- 17

Pellet Control 0 0.27 f 0.08 18 +- 2 Epinephrine, 10 pM 0 0.14 f 0.04 14 f 2 Insulin. 25 milliunits/ml 0 0.44 f 0.08 14 f 2

10713

removed Activity

% 39 42 37

TABLE I1 Effects of insulin and epinephrine on glycogen synthase

Hemidiaphragms were incubated with hormones before extracts were prepared. Glycogen synthase activity was measured and 32P-labeled glycogen synthase was immunoprecipitated, treated with CNBr, and subjected to electrophoresis as described under "Experimental Procedures." Autoradiograms of gels were scanned and photo- copies of the profiles were made. The peaks corresponding to intact 32P-labeled glycogen synthase and its two major CNBr fragments were cut out and weighed. Results are expressed as a percentage of the weights of the peaks from hemidiaphragms not exposed to hormones. The results presented are the mean values f S.E. from 4 experiments.

Glycogen synthase 32P-labeled immunoprecipitate Fragment 11/ Activity ratio Total activity Fragment I1 Fragment I Intact fragment I

rmollminlmg Control 0.24 f 0.02 0.027 +- 0.002 Epinephrine, 10 p M 0.16 f 0.02 0.028 f 0.002 Insulin, 25 milliunitsl 0.38 f 0.01 0.025 f 0.004

Insulin, 25 milliunits/ 0.22 2 0.03 0.025 f 0.004 ml

ml plus epinephrine, 10 p M

unchanged from those in the whole extract. Interestingly, a readily detectable amount of glycogen synthase activity was observed in the washed bacterial pellets from samples incubated with anti-glycogen synthase antibodies. A constant percentage of activity was recovered in the different pellets and, again, the activity ratios were the same as in the whole extracts. These results provide additional evidence that the antibodies were not selectively removing specific forms of enzyme from solution and also indicate that the conditions of im- munoprecipitation were sufficient to prevent posthomogenizational changes in the phosphorylation state of the enzyme.

Cyanogen Bromide Cleavages and Electrophoretic Resolution of 32P- labeled Proteins-To elute the 3ZP-labeled glycogen synthase from the antibody-protein A complexes, bacterial pellets were suspended in 70% formic acid (200 p l ) and incubated a t 23 "C for 15 min. After centrifugation at 12,000 X g for 5 min, the supernatants were collected and the pellets extracted once more with formic acid. Supernatants from the two extractions were pooled. Each sample of formic acid extract was divided into two equal volumes. To one was added 3-6 mg of CNBr. Both samples were sealed and incubated for 18 h at 23 "C. The samples were lyophilized and then reconstituted in 0.2 ml of H,O and evaporated to dryness in a centrifugal Bio-Dryer (VirTis). To each tube was added 30 p1 of electrophoresis sample buffer composed of 10% glycerol, 120 mM P-mercaptoethanol, 0.5% SDS, and 12.5 mM Tris/P04 (pH 6.8). After thorough mixing, the samples were incubated at 100 "C for 1 min and then allowed to cool to room temperature. The original extracts contained similar but unequal amounts of glycogen synthase. To correct for these differences sample buffer was added in appropriate volumes to normalize samples to equal amounts of total glycogen synthase activity which is presumably

percentage of control 100 100 100 6.9 f 0.7 150 f 15 247 f 11 150 f 14 3.6 f 0.2 54 f 4 61 f 10 56 f 3 6.4 f 1.0

148 +- 11 198 f 39 141 f 13 5.1 k 0.5

proportional to the amount of enzyme present. Since hormones did not change total synthase activity (Table II), the same results would be expected if samples had been corrected using measurements of extract protein. Polyacrylamide gel electrophoresis was performed using a modification of the procedures of Laemmli (23). Samples (20 pl) were added to slab gels (1.5 mm thick) formed with a linear gradient of acrylamide from 5 to 20% in the resolving gel. Molecular weights were estimated from the mobilities of the following standards: @-galactosidase (130,000), phosphorylase b (94,000), rabbit skeletal muscle glycogen synthase (85,0001, bovine serum albumin (68,000), pyruvate kinase (57,0001, glutamate dehydrogenase (53,000), ovalbu- min (45,000), aldolase (40,000), lactate dehydrogenase (36,000), car- bonic anhydrase (29,000), trypsinogen (24,000), soybean trypsin in- hibitor (21,000), @-lactoglobulin (18,400), myoglobin (17,200), and hemoglobin (15,500). For references on autoradiograms, purified rab- bit skeletal muscle glycogen synthase and CNBr fragments that had been phosphorylated with [r3'P]ATP and catalytic subunit of CAMP- dependent protein kinase were included. After electrophoresis was complete, gels were placed in a solution of Coomassie blue to stain for protein (24) or were placed in 2 liters of 1% glycerol and 10% acetic acid for 90 min and then dried onto filter paper (Whatman 3").

Autoradiography and Analysis of Autoradiograms-Dried gels were placed in x-ray casettes containing films (Kodak XAR-5) which were exposed a t -85 "C for different periods of time depending on the amount of radioactivity present. After developing, the autoradiograms were cut into strips corresponding to individual lanes on the gel. Differences in optical density of strips from samples that had been incubated with immune and nonimmune IgG were determined at 540

by guest on June 1, 2018http://w

ww

.jbc.org/D

ownloaded from

10714 Hormonal Control of Site-specific Phosphorylation

0.4

0.3 I

0. I

0 20 40 60 80 100 120 mm

' i '013 014 d.5 O k di' 0.; C

MOB1 LlTY FIG. 2. Immunoprecipitation of 32P-labeled glycogen syn-

thase from rat hemidiaphragms. Skeletal muscle was incubated in media containing [32P]phosphate for 6 h before extracts were

nm using a dual beam gel scanner (Isco model 1310) in which the track of the sample incubated with nonimmune IgG was used as reference. Thus, the scan represents the radioactivity specifically precipitated by antibodies to glycogen synthase. . That is, the absor- bance due to minor contaminants in both immune and nonimmune samples was substracted (for example, the species of relative mobili- ties equal to 0.42 and 0.58 in Figs. 4 and 7). The profiles of the gel scans were photocopied, and the peaks were cut out and weighed. Results presented were derived from these weights which are directly proportional to the areas beneath the respective peaks.

Other M~teriak-[U-'~C]Glucose-l-P and [U-'*C]glucose were ob- tained from ICN Radioisotope Division. UDP-[U-'4C]glucose was prepared essentially as described by Tan (25) . [y-"PIATP was from Amersham Corp. and ["P]orthophosphate was from New England Nuclear. Cyanogen bromide was obtained from Pierce Chemical Co. Purified catalytic subunit of CAMP-dependent protein kinase from beef heart was kindly supplied by Dr. David Brautigan (Brown University). Proteins used as molecular weight standards and other enzymes were obtained from Boehringer Mannheim. Highly purified insulin was a gift from Lilly. Glucose-1-P, UDP-glucose, rabbit liver glycogen, ATP, benzamidine, and other chemicals were from Sigma.

RESULTS

Electrophoretic Mobility of "'P-labeled Glycogen Synthase from Rat Hemidiaphragms-Antibodies against highly puri- fied rabbit skeletal muscle glycogen synthase were used to immunoprecipitate glycogen synthase from extracts of rat hemidiaphragms that had been incubated with ["Plphos- phate. Samples of the immunoprecipitate were incubated in 70% formic acid in the absence and presence of cyanogen bromide before being subjected to electrophoresis in the pres- ence of SDS. The profile of a scan of the optical density of an autoradiogram corresponding to a sample of uncleaved im- munoprecipitate is shown in Fig. 2A (panel i). Nearly all of the radioactivity was present in a single band, attesting to the specificity of the antibodies. The immunoprecipitate had slightly lower mobility than purified rabbit skeletal muscle glycogen synthase labeled with ["Plphosphate by incubation with [-y-"2P]ATP and the catalytic subunit of CAMP-depend- ent protein kinase. Likewise treatment of the immunoprecip- itate with CNBr resulted in the formation of two fragments of lower mobility than those obtained with the "'P-labeled rabbit muscle enzyme (Fig. 2 A , panel ii). The peak of lower mobility (Fragment 11) of the immunoprecipitate is distinctly broader than that corresponding to the large fragment (CB- 2) of the rabbit enzyme. As will be discussed later, multiple species of slightly different mobilities comprise Peak 11. The mobilities of the immunoprecipitate and CNBr fragments are compared with those of other proteins of known molecular weights in Fig. 2B. The intact immunoprecipitate has an apparent molecular weight of 90,000; the two CNBr fragments have apparent molecular weights of 27,900 and 15,500.

Incubation of Hemidiaphragms with (:"PJPhosphate-Since

prepared. The "P-labeled glycogen synthase was immunoprecipitated and treated with CNBr before electrophoresis as described under "Experimental Procedures." A , profiles of scans of optical density from autoradiograms of noncleaved samples (i) or samples which were treated with CNBr (ii). The dotted lines were obtained using purified rabbit skeletal muscle glycogen synthase that had been phosphoryl- ated with [y-32P]ATP and C subunit of CAMP-dependent protein kinase. B, estimation of the molecular weights of the immunoprecip- itated glycogen synthase and cyanogen bromide fragments. The standard curve was constructed using proteins of known molecular weights (0, see under "Experimental Procedures" for a description of standards). Rabbit skeletal muscle glycogen synthase (V) and its cyanogen bromide fragments I1 (a, 23,000) and I (A, 13,500) have lower mobilities than the immunoprecipitated enzyme (V, 90,000) and its CNBr fragments I1 (0, 29,000) and I (A, 15,500) from rat diaphragms (the numbers in parentheses denote apparent molecular weights). The molecular weight of purified rabbit skeletal muscle glycogen synthase I has been taken to be 85,000 Da (20).

by guest on June 1, 2018http://w

ww

.jbc.org/D

ownloaded from

Hormonal Control of Site-specific Phosphorylation 10715

0.15-

0.10 -200; - 3

40- @ 140 -

-!

HOURS

0 125-

100 - a W

Y

w a

HOURS

I O - 125-

100 - a W

Y

w a

HOURS FIG. 3. Effect of increasing times of incubation of rat hemidiaphragms on enzyme activities and on

incorporation of [S2P]phosphate into ATP and two cyanogen bromide fragments of glycogen synthase. Hemidiaphragms were incubated in 250 ml of medium A (minus 32P) for 1 h. The muscles were then placed (0 h in the figures) in an incubation vessel containing medium A plus [32P]phosphate (2 ml/hemidiaphragms, 0.2 mCi/ ml) and incubated for increasing periods of time before being transferred to tubes containing medium B plus ["PI phosphate (0.2 mCi/ml). Immediately after transfer, medium A (2 ml/hemidiaphragm) was withdrawn to maintain a constant tissue/medium ratio. After 1 h in medium B the tissue was frozen in liquid nitrogen; the times indicated represent total time of exposure to [32P]phosphate. Frozen tissue was powdered and processed to allow measurement of ( A ) glycogen synthase and phosphorylase activities, ( B ) concentrations of ATP and the specific activities of [ y - "PIATP, and (C) relative amounts of 32P incorporated into glycogen synthase and cyanogen bromide fragments of the enzyme. The values presented in C represent the weights of peaks from photocopies of gel scans analyzed by the procedures described under "Experimental Procedures." All results presented are the average values from two separate experiments.

uptake and incorporation of ["PJphosphate into muscle phos- phoproteins occur slowly, relatively long incubations were required to allow incorporation of sufficient radioactivity into glycogen synthase. Under appropriate conditions, prepara- tions of rat diaphragm may remain viable for several days in organ culture (26). In the experiments presented in Fig. 3, rat hemidiaphragms were incubated for increasing times in en- riched culture media (Medium A, [32P]phosphate added a t 0 h). Total activities of glycogen synthase and phosphorylase were essentially unchanged for over 12 h in uitro (Fig. 3A). The concentrations of ATP, although slowly decreasing with time (Fig. 3B), remained within the range of concentrations observed for rat skeletal muscle fibers in vivo (27). The phosphorylase activity ratio decreased from 0.34 to 0.22 dur- ing the first hour of incubation (-1 h to 0 h) and was then unchanged for 6 h. The activity ratio of glycogen synthase was constant for 7 h in vitro (-1 h to 6 h); however, the activity ratios of both glycogen synthase and phosphorylase were elevated at the 12-h time point. The specific activity of [y'"P]ATP increased until reaching a plateau value a t 6 h (Fig. 3B). The incorporation of ["P]phosphate into both CNBr fragments of glycogen synthase increased throughout the entire incubation period (Fig. 3C).

Hormonal Effects on the '"P Content of Glycogen Synthase in Hemidiaphragms Incubated with p2P]Phosphate-The ef- fects of insulin and epinephrine on the activity ratios of glycogen synthase were determined in hemidiaphragms incu- bated with [''2P]phosphate for 6 h (Table 11). Incubation with insulin for the last 30 min increased the ratio from 0.24 to 0.38, and epinephrine decreased the ratio to 0.16: The activity ratio in extracts from hemidiaphragms incubated with insulin

' As another index of epinephrine action, phosphorylase activity was measured in two of the experiments presented in Table 11. Epinephrine increased the activity ratio of phosphorylase from 0.22 to 0.46; insulin alone was without effect and the hormone did not oppose the effects of epinephrine.

1.0 1.0 0

EPINEPHRINE

- 0.6 - 0.4

0.2 0.2 - -

0.2 - - 0.2

-0 NI

I I I I I I I , , I I I I I

0 20 40 60 80 1 0 0 I200 x) 40 60 80 100 I 2 0 MILLIMETERS

FIG. 4. Electrophoretic analysis of S2P-labeled glycogen synthase following incubation of skeletal muscle with insulin and epinephrine. Hemidiaphragms were incubated without addi- tions (i) and with 25 milliunits/ml of insulin (ii), 10 pM epinephrine (iii), or 25 milliunits/ml of insulin plus 10 pM epinephrine (iu) as described under "Experimental Procedures." After immunoprecipi- tation samples were submitted to electrophoresis. The profiles of optical density scans of autoradiograms from gels of immunoprecipi- tated "P-labeled glycogen synthase are shown. Below each scan is a reproduction of the autoradiogram corresponding to the gel tracts which were scanned. The abbreviations are NI (nonimmune I&) and Ab- (anti" Ab-). Arrows indicate the position of purified rabbit skeletal muscle glycogen synthase.

plus epinephrine was quite similar to control. The results presented in Fig. 4 are consistent with the hypothesis that epinephrine increases phosphorylation of glycogen synthase and that insulin decreases the phosphate content of the en-

by guest on June 1, 2018http://w

ww

.jbc.org/D

ownloaded from

10716 Hormonal Control of Site-specific Phosphorylation

zyme. In this representative experiment, 32P-labeled glycogen synthase was immunoprecipitated from extracts of hormon- ally treated tissues before being dissolved and subjected to electrophoresis in SDS. Autoradiograms were prepared and the tracks corresponding to control samples and samples from different hormonal treatments were scanned for optical den- sity a t 540 nm. Incubation of hemidiaphragms with insulin resulted in a decrease in the size of the peak of "2P-labeled glycogen synthase (Fig. 4). The size of the peak from samples of epinephrine-treated tissue was larger than the control peak, and the peak from samples of tissue incubated with insulin plus epinephrine was between control and epinephrine in size. These results were expressed as a percentage of the size of the control peak and averaged with results from three other experiments (Table 11). These values indicate that epineph- rine increased the 32P content of glycogen synthase by 50%, whereas insulin action resulted in a 44% decrease in the amount of ["P]phosphate bound to the enzyme. Under the conditions used in this experiment, we have observed no effect of insulin or epinephrine on the specific activity of [y3'P] ATP.'

The effects of increasing concentrations of insulin are shown in Fig. 5. The decreases in "P content of the enzyme correlated well with the increases in activity ratio. Concentra- tions of insulin as low as 25 microunits/ml were effective. The effect of the hormone (25 milliunits/ml) was maximal after 15 min and could be observed after as little as 5 min of incubation (Fig. 6) . Thus the dephosphorylation of glycogen synthase produced by insulin was rapid and occurred at phys- iological concentrations of the hormone.

To determine if insulin and epinephrine preferentially af- fected the amount of ["2P]phosphate present in sites found in different CNBr fragments of the enzyme, optical density scans were made of autoradiograms from gels of samples that had been cleaved with CNBr (Fig. 7). Incubation of the muscles

I+' T

I L 0 t+ 0625 2.5 25

INSULIN(mU/ml)

FIG. 5. Dephosphorylation of glycogen synthase by increas- ing concentrations of insulin. Hemidiaphragms were incubated with ["P]phosphate as described under "Experimental Procedures" before increasing concentrations of insulin were added. After 30 min the incubations were terminated and extracts were prepared. Glyco- gen synthase activities were measured and the R2P-labeled enzymes were immunoprecipitated from samples of the extracts. After electro- phoresis, autoradiograms of the gels were scanned and changes in the amounts of radioactivity due to the different concentrations of insulin were determined. The results presented are the mean values f S.E. from 5 experiments and represent the percentage changes in activity ratio (0) or "P contents (A) of glycogen synthase produced by insulin. mu, milliunits.

407

MINUTES

FIG. 6. Effect of increasing times of incubation on the de- phosphorylation of glycogen synthase by insulin. Hemidia- phragms were incubated in vitro in the presence of [""Plphosphate for a total of 6 h. Insulin (25 milliunits/ml) was added at the appropriate time prior to termination of the incubation to give the period of exposure indicated. The results are expressed as described in the legend to Fig. 5 and represent the mean values f S.E. from 3 experiments.

CONTROL

0.3

6"3: 02

0 I EPINEPHRINE i 0.4

6 MILLIMETERS

'1 d 0 20 40 60 d0 I00 I20

FIG. 7. Electrophoretic analysis of the cyanogen bromide fragments of 32P-labeled glycogen synthase following incu- bation of skeletal muscle with insulin and epinephrine. Sam- ples of the immunoprecipitates used in the experiment shown in Fig. 4 were cleaved with cyanogen bromide before being subjected to electrophoresis. Shown below each optical density scan are reproduc- tions of the autoradiograms from the gels of the CNBr cleavage products. The arrows show the location of the two CNBr fragments of glycogen synthase phosphorylated with the catalytic subunit of CAMP-dependent protein kinase.

with insulin resulted in a decrease in the total amount of radioactivity present in the fragments. In insulin samples, three peaks were most clearly resolved in the area correspond- ing to Fragment I1 in Fig. 1. Examination of the other scans reveals a consistent shoulder close to the position of the phosphorylated CB-2 of rabbit muscle glycogen synthase, and clearly more than one band may be seen in the pictures of the autoradiograms below the scans. For quantitation the differ- ent electrophoretic species have been considered collectively

by guest on June 1, 2018http://w

ww

.jbc.org/D

ownloaded from

Hormonal Control of Site-specific Phosphorylation 10717

as Fragment I1 (Table 11). Insulin decreased the [3'P]phos- phate in Fragment I by 39%. The hormone also promoted a 46% reduction in the radioactivity associated with Fragment 11. In samples from control tissue, the ratio of radioactivity present in Fragment I1 to that in Fragment I was about 7:l. Insulin did not significantly affect this ratio, indicating that the hormone did not differentially affect the phosphate con- tent of the two fragments. Epinephrine increased the amount of radioactivity present in both fragments; however, it had a greater effect on the radioactivity present in Fragment I which resulted in a 50% decrease in the ratio of the 32P content of Fragment I1 to that of Fragment I. When epinephrine was added to tissue first incubated with insulin, the amount of radioactivity was increased in both Fragments I and 11. In these experiments insulin had little, if any, effect on opposing epinephrine action on either the phosphorylation or the activ- ity ratio of glycogen synthase.

DISCUSSION

The subunit of rabbit skeletal muscle glycogen synthase is subject to phosphorylation i n vitro a t a minimum of seven different sites through the action of at least five distinct protein kinases (for example, see Refs. (7-12). All seven sites characterized i n vitro are contained in only two CNBr frag- ments of the enzyme (9). One site (site 2) is located in an NH2-terminal fragment (CB-1) and the other six (sites la, lb, 3a, 3b, 3c, 5)5 in a COOH-terminal fragment (CB-2). It is known that the mobility of CB-2 on analysis by polyacryl- amide gel electrophoresis in the presence of SDS is very sensitive to its phosphorylation pattern, and depending on the particular site occupancy, may display significant electro- phoretic heterogeneity (7, 12).

The activity of glycogen synthase is a complex function of the degree and site distribution of its covalent phosphoryla- tion. Though it is almost certain that hormones which cause stable changes in glycogen synthase activity do so by modi- fying its phosphorylation state, the site specificity of such hormonal control has yet to be fully defined. The present report describes experiments designed to study directly hor- monal control of phosphorylation of glycogen synthase in rat diaphragms.

Our experimental approach was to incubate rat hemidia- phragms with [3'P]phosphate to allow labeling of glycogen synthase prior to treatment with insulin and/or epinephrine. '"P-labeled glycogen synthase was then rapidly purified by the use of specific antibodies. Most other studies of rat or rabbit skeletal muscle glycogen synthase phosphorylation in intact tissue have utilized conventional purification of the enzyme (28-35); one report of rat heart glycogen synthase does describe the effective use of 32P-labeling and immuno- precipitation although no effect of epinephrine on the 32P content of the enzyme was detected (36). Our method has several advantages when compared to conventional purifica- tion of glycogen synthase from whole animals. Treatment of isolated tissue avoids some of the complex interactions which are possible when injecting hormones into whole animals and which may obscure interpretation of the results. Rapid im- munopurification is likely to be more reproducible, and cer- tainly simpler, than large scale enzyme preparation, and Some information can be obtained on the rate of phosphate turnover in the enzyme. Since the enzyme is labeled with ["plphos- phate, detection of phosphorylation does not depend on prior chemical knowledge of the site so that novel phosphorylation sites could in principle be detected. Nonetheless, maximum

' For simplicity, we will refer collectively to sites l a and l b and 3a, 3b, and 3c as sites 1 and sites 3, respectively.

interpretation of the data requires reference to what is known of the in vitro phosphorylation of the enzyme. The main disadvantage of our approach was that a steady state labeling of glycogen synthase with [3'P]phosphate was not achieved, even though the specific activity of the terminal phosphate of ATP had reached a plateau level. Thus, the absolute amount of phosphate per enzyme subunit cannot be determined. How- ever, as discussed later, observed changes in [32P]phosphate associated with glycogen synthase were consistent with chem- ical measurements of the phosphate content of purified en- zymes made in other studies.

Although rat skeletal muscle glycogen synthase has been less well characterized than the rabbit enzyme, the two en- zymes are similar to several respects. Identical values of 85,000 Da have been reported for the molecular weights of both enzymes (20, 28). We have used conventional techniques to obtain a purified preparation of rat skeletal muscle glycogen synthase.6 This enzyme has the same mobility as the rabbit enzyme when subjected to electrophoresis on acrylamide gels in the presence of SDS. Phosphorylation of the rat enzyme with [y-"'P]ATP and the catalytic subunit of CAMP-depend- ent protein kinase, followed by treatment with CNBr, resulted in formation of two "P-labeled fragments of the same electro- phoretic mobility as those from the rabbit enzyme analogously phosphorylated. Furthermore, as demonstrated in this report, antibodies against rabbit skeletal muscle glycogen synthase cross-react with the rat enzyme. Although the assumption of total chemical identity between the two enzymes may be unwarranted, the demonstrated structural homologies justify the working hypothesis that phosphorylation of the rat en- zyme closely resembles that of rabbit muscle glycogen syn- thase.

The "P-labeled glycogen synthase immunoprecipitated from rat hemidiaphragms had Mapp = 90,000 (Fig. 2). Similar values have been obtained with mouse (37) and rabbit dia- phragms.' An initial concern was that limited proteolysis of a larger protein had generated the enzyme of Mapp = 85,000 during purification from either rabbit (20) or rat6 muscle. However, it is now known that the lower mobility form (Mapp = 90,000) can be produced from the higher mobility form (Mapp = 85,000) by phosphorylation of certain sites. Based on mobility differences of different phosphorylated forms of gly- cogen synthase, one might speculate that there was a rela- tively high degree of phosphorylation in sites 3 and possibly 5 in the immunoprecipitated rat enzyme (12). Similar consid- erations can explain the observation that the "P-labeled enzyme isolated from rat diaphragms gave rise to multiple phosphorylated species in the mobility range of CB-2 from purified rabbit enzyme since phosphorylation also critically affects the mobility of this fragment (7, 12). The resolution of three phosphorylated species of CB-2 (Fig. 7) indicates that the glycogen synthase subunit exists in muscle at a minimum of three distinct phosphorylation states, even after hormonal treatment.

A question remains concerning Fragment I of the rat dia- phragm enzyme since it has a slightly lower electrophoretic mobility (Mapp = 15,500) compared with CB-1 from purified rabbit glycogen synthase (Mapp = 13,100) (Fig. 2B). One hypothesis is that these two small fragments are in no way

Glycogen synthase was purified from skeletal muscle of rat hind limbs using methods described by Takeda et al. (20) for purifying the rabbit enzyme. The specific activities in units/mg of 4 preparations were 20, 24, 36, and 37 which represent purifications from crude extracts of 700- to 1000-fold. Based on Coomassie blue staining following electrophoresis in SDS, 90% of the preparation of highest specific activity is a species of Mapp = 85,000; three species of slightly higher mobility constitute the remainder of the staining.

by guest on June 1, 2018http://w

ww

.jbc.org/D

ownloaded from

10718 Hormonal Control of Site-specific Phosphorylation

analogous. If this was correct, then site 2 was neither labeled with [32P]phosphate nor responded to epinephrine in the incubated diaphragm, and a hitherto unknown site(s) exists that was phosphorylated in response to epinephrine. This seems improbable, particularly since Parker et al. (31) pro- vided compelling evidence for stimulation of site 2 phospho- rylation by epinephrine in uiuo. We, therefore, propose that the smaller rat diaphragm CNBr fragment is identifiable with CB-1, although the reason for the slight mobility difference is unknown. Some explanations could be the existence of a novel phosphorylation site in this region, slight NH2-terminal proteolysis, or other unrecognized chemical modifications. Therefore, an important conclusion from this study is that, following incubation of rat hemidiaphragms, [32P]phosphate was confined to the regions of the glycogen synthase molecule corresponding to CB-1 and CB-2, consistent with phospho- rylation sites thus far identified in oitro.

In our experiments, epinephrine caused a 50% increase in the amount of [32P]phosphate present in glycogen synthase (Table 11). Epinephrine has been reported to increase chemi- cally determined phosphate in the enzyme by 58% in perfused rat hindlimbs (28), and 41% (29), 64% (30), and 77% (31) in whole rabbits. Insulin caused, in the present experiments, about a 50% decrease in the 32P content of glycogen synthase. Previous values, again for chemically determined phosphate, were decreases of 35% (32) and 15% (35) in whole rabbits and 21% in perfused rat hindlimbs (33). The similarity of our values to published results argues strongly that changes in the 32P content of glycogen synthase, as determined in the present study, provide a reliable index of changes in phospho- rylation state.

The results presented in the present report provide direct information on the acute effects of insulin on the distribution of phosphate in glycogen synthase. An important conclusion is that insulin decreased the content of phosphate in both CB-1 (containing site 2) and CB-2 (containing sites la, lb, 3a, 3b, 3c, and 5). After insulin treatment, the electrophoretic heterogeneity of CB-2 became more clearly visible (Fig. 7) as though the relative proportions of different forms had been changed. These results are suggestive of dephosphorylation in sites 3 since the mobility of this fragment is most affected by changes in the phosphate content of these sites (7, 12). This interpretation is consistent with the recent finding that injection of rabbits with insulin decreased phosphate in sites 3 (35). Indeed, Parker et al. (35) concluded that insulin action on glycogen synthase was confined to these sites. However, we observed an effect of insulin on decreasing [32P]phosphate in CB-1 (Table 11) indicating that insulin action in rat dia- phragm involves dephosphorylation of at least l other site (presumably site 2).

In earlier work, Sheorain et al. (34) had shown that alloxan- induced diabetes in rabbits resulted in over a 2-fold increase in the phosphate content of what they defined as the “trypsin- insensitive domain” of the enzyme. Treatment of the diabetic animals for 4 days with insulin reversed the effects of diabetes, and administration of insulin to control animals for 4 days decreased the trypsin-insensitive phosphate by about 40%. However, these treatments with insulin must be considered long term and the effects observed may bear little relationship to the acute actions of the hormone. Indeed, it was reported that short term administration of insulin did not affect the phosphate content of glycogen synthase even though the hormone increased the activity ratio (34). This finding is clearly in conflict with previous results (32, 33, 35) and with those of the present report.

Treatment of rat hemidiaphragms with epinephrine led to an increase in the “P contents of both CB-1 and CB-2,

although the greatest increase was CB-1. In recent studies, Parker et al. (31) chemically analyzed the phosphate content of sites on glycogen synthase after rabbits were injected with epinephrine. These investigators concluded that phosphate was introduced into all sites except site 5 . Furthermore, the greatest increase was in site 2 (2.6-fold), a value very similar to that obtained in our study (2.5-fold) for the increase in the 32P content of CB-1, which contains site 2. From the results of Parker et al. (31), the collective increase in phosphate in sites 3, 5, and 1 was 64%, again similar to our value of 50%. In the work of Sheorain et al. (34), in which trypsin sensitivity alone was employed to dissect phosphorylation sites, a 2-fold increase in the trypsin-sensitive phosphate was observed fol- lowing epinephrine treatment but little effect was seen on trypsin-insensitive phosphate. A question may be raised, how- ever, as to whether sites 3 would register quantitatively in the trypsin-sensitive or trypsin-insensitive regions (35), so that direct comparison of our results with those of Sheorain et al. (34) is difficult.

An important generalization that can be made from the present study is that both insulin and epinephrine control the phosphorylation of multiple (at least two) sites on the enzyme. For epinephrine action, the same conclusion was drawn from studies of mouse diaphragms (37) and rabbit muscle (31). Phosphorylation in both CB-1 and CB-2 is consistent with the specificity of cyclic AMP-dependent protein kinase (9- ll), which is believed to be involved in mediating epinephrine action. However, Parker et al. (31) also demonstrated a sig- nificant increase in the phosphorylation of sites 3, which cannot be phosphorylated by cyclic AMP-dependent protein kinase in vitro. Because we did not resolve sites 3 from sites 1 and 5, our data do not yet address this point and we cannot say whether factors besides cyclic AMP-dependent protein kinase must be invoked to explain epinephrine-mediated phosphorylation of rat diaphragm glycogen synthase.

The observed multisite control of glycogen synthase phos- phorylation by insulin has several interesting implications that must be accounted for in any proposed mechanism by insulin action. Should insulin action be mediated directly by a single protein kinase, then the kinase would have to act on phosphorylation sites in both CB-1 and CB-2. Evidence has been presented that insulin may increase glycogen synthase phosphatase activity in various tissues and several potential mechanisms have been discussed (38-41). Our results do not exclude an involvement of protein phosphatases in the hor- monal regulation of glycogen synthase phosphorylation. In fact, the rapidity of the insulin effect (Fig. 6) relative to the incorporation of [32P]phosphate into glycogen synthase (Fig. 3C) strongly suggests that the hormone increased phosphatase activity. The present results do dictate that, if phosphatase action alone is to explain insulin-mediated control of glycogen synthase phosphorylation, then the phosphatase(s1 must be capable of dephosphorylating sites in both CB-1 and CB-2. I t is, of course, quite conceivable that insulin action could in- volve control of both phosphatases and protein kinases. Con- tinued work on the phosphorylation of glycogen synthase in intact muscle, with improved resolution of the phosphoryla- tion sites, should provide more and more restrictions of the potential mechanisms of control by both insulin and epineph- rine.

Acknowledgment-We thank Dr. David Brautigan for providing US with the C subunit of CAMP-dependent protein kinase.

REFERENCES 1. Villar-Palasi, C., and Larner, J. (1960) Biochim. Biophy~. Acta

2. Craig, J. W., and Larner, J. (1964) Nature (Lond.) 202, 971-973 39, 171-173

by guest on June 1, 2018http://w

ww

.jbc.org/D

ownloaded from

Hormonal Control of Site

3. Villar-Palasi, C., and Larner, J. (1971) Curr. Top. Cell. Regul. 3,

4. Goldberg, N. D., Villar-Palasi, C., Sasko, H., and Larner, J. (1967)

5. Craig, J. W., Rall, T. W., and Larner, J . (1969) Biochim. Biophys.

6. Smith, C. H., Brown, N. E., and Larner, J. (1971) Biochim.

7. Picton, C., Aitken, A., Bilham, T., and Cohen, P. (1982) Eur. J.

8. Roach, P. J. (1981) Curr. Top. Cell. Regul. 20, 24-105 9. Cohen, P., Yellowlees, D., Aitken, A., Donella-Deana, A., Hem-

mings, B. A., and Parker, P. J. (1982) Eur. J. Biochem. 124, 21-35

10. DePaoli-Roach, A. A,, Roach, P. J., and Larner, J. (1979) J. Biol. Chem. 254, 12062-12068.

11. Soderling, T. R., Jett, M. F., Hutson, N. J., and Khatra, B. S. (1977) J. Biol. Chem. 252, 7517-7524

12. DePaoli-Roach, A. A., and Ahmad, A,, Camici, M., Lawrence, J. C., Jr., and Roach, P. J. (1983) J. Biol. Chem. 258, 10702- 10709

13. Goldberg, A. L., Martel, S. B., and Kushmerick, M. J. (1975) Methods Enzymol. 39,82-111

14. Dulhecco, R., and Freeman, G. (1959) Virology 8, 396-397 15. Thomas, J. A,, Schlender, K. K., and Larner, J. (1968) Anal.

Biochem. 25, 486-499 16. Gilhoe, D. P., Larson, K. L., and Nuttal, F. Q. (1972) Anal.

Biochem. 47,20-27 17. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J.

(1951) J . Biol. Chem. 193, 265-275 18. England, P. J., and Walsh, D. A. (1976) Anal. Biochem. 75, 429-

435 19. Williamson, J . R., and Corkey, B. R. (1969) Methods Enzymol.

13,434-497 20. Takeda, Y., Brewer, H. B., Jr., and Larner, J. (1975) J . Biol.

Chem. 250,8943-8950 21. Smith, C. H., Villar-Palasi, C., Brown, N. E., Schlender, K. K.,

Rosenkrans, A. M., and Larner, J. (1972) Methods Enzymol.

195-236

Biochim. Biophys. Acta 148, 665-672

Acta 177, 213-219

Biophys. Acta 242,81-88

Biochem. 124, 37-45

-specific Phosphorylation 10719

48,530-539 22. Goding, J . W. (1978) J. Immunol. Methods 20, 241-253 23. Laemmli, U. K. (1970) Nature (Lond.) 227,680-685 24. Weber, K., and Osborn, M. (1969) J. Biol. Chem. 244,4406-4412 25. Tan, A. W. (1979) Biochim. Biophys. Acta 582,543-547 26. Purves, D., and Sakmann, B. (1974) J. Physiol. (Lond.) 239,125-

153 27. Hintz, C. S., Chi, M. M.-Y., Fell, R. D., Ivy, J. L., Kaiser, K. K.,

Lowry, C. V., and Lowry, 0. H. (1982) Am. J. Physiol. 242, C218-C288

28. Chiasson, J.-L., Aylward, J. H., Shikama, H., and Exton, J. H. (1981) FEBS Lett. 127,97-100

29. Cohen, P. (1978) Curr. Top. Cell. Regul. 14, 117-196 30. Sheorain, V. S., Khatra, B. S., and Soderling, T. R. (1981) FEBS

31. Parker, P. J., Embi, N., Caudwell, B., and Cohen, P. (1982) Eur.

32. Roach, P. J., Rosell-Perez, M., and Larner, J . (1977) FEBS Lett.

33. Uhing, R. J., Shikama, H., and Exton, J . H. (1981) FEBS Lett.

34. Sheorain, V. S., Khatra, B. S., and Soderling, T. R. (1982) J. Biol.

35. Parker, P. J., Caudwell, B., and Cohen, P. (1983) Eur. J. Biochem.

36. McCullough, T. E., and Walsh, D. A. (1979) J. Biol. Chem. 254,

37. Roach, P. J., DePaoli-Roach, A. A., Hiken, J., and Lawrence, J.

38. Lawrence, J. C., Jr., and Larner, J. (1978) J. Biol. Chem. 253,

39. Huang, F. L., and Glinsmann, W. H. (1976) Eur. J. Biochem. 70,

40. Foulkes, J. G., Jefferson, L. S., and Cohen, P. (1980) FEBS Lett.

41. Larner, J., Galasko, G., Cheng, K., DePaoli-Roach, A. A., Huang,

Lett. 127,94-96

J. Biochem. 124,47-55

80,95-98

134, 185-188

Chem. 257,3462-3470

130, 227-234

7336-7344

C., Jr. (1982) Diabetes 31, 124A

2104-2113

419-426

112, 21-24

L., Daggy, P., and Kellog, J. (1979) Science 206, 1408-1410

by guest on June 1, 2018http://w

ww

.jbc.org/D

ownloaded from

J C Lawrence, Jr, J F Hiken, A A DePaoli-Roach and P J Roachfragments.

and epinephrine on the distribution of phosphate between two cyanogen bromide Hormonal control of glycogen synthase in rat hemidiaphragms. Effects of insulin

1983, 258:10710-10719.J. Biol. Chem. 

  http://www.jbc.org/content/258/17/10710Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/258/17/10710.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on June 1, 2018http://w

ww

.jbc.org/D

ownloaded from