Thrombin-Induced Protein Phosphorylation in Human Platelets

Thrombin-induced protein phosphorylation in human platelets.

R M Lyons, … , N Stanford, P W Majerus

J Clin Invest. 1975;56(4):924-936. https://doi.org/10.1172/JCI108172.

Research Article

Intact human platelets loaded with 32PO4 contain multiple phosphorylated proteins. Thrombin treatment of intact 32PO4- loaded platelets results in a 2-6-fold increase in phosphorylation of a platelet protein (designated "peak 7" protein) of approximately 40,000 mol wt as determined by sodium dodecyl sulfate polyacrylamide gel electrophoresis and by gel filtration on Sephadex G-150. A similar increase in phosphorylation was observed in a platelet protein (designated "peak 9" protein) of approximately 20,000 mol wt. The time for half-maximal phosphorylation of peak 7 and peak 9 protein was 10-14 s. The concentration of thrombin at half-maximal phosphorylation was 0.25 U/ml for both proteins. Prior incubation of platelets with dibutyryl cyclic adenosine 3',5'-monophosphate or prostaglandin E1 inhibited thrombin-induced peak 7 and peak 9 protein phosphorylation. The erythroagglutinating phytohemagglutinin of Phaseolus vulgaris, a non-proteolytic release-inducing agent, induced peak 7 and peak 9 protein phosphorylation. Thus, the characteristics of peak 7 and peak 9 protein phosphorylation are similar to those of the platelet release reaction, suggesting that the phosphorylation of these proteins may play a role in the platelet release reaction. When platelet sonicates or the supernatant fraction from platelet sonicates were incubated with [gamma-32P]ATP there was phosphorylation of both peak 7 and peak 9 proteins. This phosphorylation was unaffected by either added thrombin or adenosine 3',5'-cyclic monophosphate (cAMP) despite the presence of the phosphodiesterase inhibitor […]

Find the latest version: https://jci.me/108172/pdf Thrombin-Induced Protein Phosphorylation in Human Platelets

ROGER M. LYONS, NANCY STANFORD, and PHMIIP W. MAJERUS From the Divisi'ons of Hematology and Oncology, the Departments of Internal Medicine and Biochemistry, Washington University School of Medicine, St. Louis, Missouri 63110

A B S T R A C T Intact human platelets loaded with "PO, bin-dependent phosphorylation depends upon intact contain multiple phosphorylated proteins. Thrombin platelets. When the supernatant fraction from platelet treatment of intact "PO-loaded platelets results in a sonicates was fractionated by histone-Sepharose affinity 2-6-fold increase in phosphorylation of a platelet pro- chromatography, two distinct protein kinase enzymes tein (designated "peak 7" protein) of approximately were resolved, one a cAMP-dependent holoenzyme and 40,000 mol wt as determined by sodium dodecyl sulfate the other a cAMP-independent enzyme. The isolated polyacrylamide gel electrophoresis and by gel filtration cAMP-dependent enzyme fraction catalyzed the cAMP- on Sephadex G-150. A similar increase in phosphoryla- (but not thrombin-) stimulated phosphorylation of a tion was observed in a platelet protein (designated protein that co-electrophoresed with peak 7 protein. "peak 9" protein) of approximately 20,000 mol wt. The time for half-maximal phosphorylation of peak 7 and INTRODUCTION of peak 9 protein was 10-14 s. The concentration Evidence has been presented in recent years suggesting thrombin at half-maximal phosphorylation was 0.25 U/ml and the release reaction may be of with that platelet aggregation for both proteins. Prior incubation platelets controlled by variations in intracellular adenosine 3',5'- dibutyryl cyclic adenosine 3',5'-monophosphate or prosta- (cAMP)' levels (2). Thrombin, 7 and peak 9 cyclic monophosphate glandin E1 inhibited thrombin-induced peak a potent inducer of the platelet release reaction, has protein phosphorylation. The erythroagglutinating phyto- been shown to bind to receptors on the platelet surface hemagglutinin of Phaseolus vulgaris, a non-proteolytic 9 (3, 4) and to decrease the activity of platelet adenylate release-inducing agent, induced peak 7 and peak cyclase without a change in phosphodiesterase activity protein phosphorylation. Thus, the characteristics of cAMP are mediated similar (5). In other tissues the effects of peak 7 and peak 9 protein phosphorylation are through changes in protein phosphorylation resulting to those of the platelet release reaction, suggesting that alteration in protein kinase and/ role from a cAMP-induced the phosphorylation of these proteins may play a or phosphoprotein phosphatase activity (648). In many in the platelet release reaction. is associated with increased fraction tissues increasing cAMP When platelet sonicates or the supernatant protein kinase activity although recently the opposite from platelet sonicates were incubated with ['y-"P]ATP of protein kinase activity 9 response, namely inhibition there was phosphorylation of both peak 7 and peak by rising cAMP has also been described (9, 10). The proteins. This phosphorylation was unaffected by either cAMP-dependent protein kinases are relatively non- added thrombin or adenosine 3',5'-cyclic monophosphate specific, and their physiologically important endogenous (cAMP) despite the presence of the phosphodiesterase difficult to identify in cell homog- throm- substrates have been inhibitor 1-methyl-3-isobutylxanthine. Thus, the enates (6). Thus we have studied protein phosphoryla- This work was reported in part at the American Society tion in both intact platelets and platelet homogenates in for Clinical Investigation Meeting, Atlantic City, NC. J., May 1974 (1). adenosine 3',5'- Dr. Lyons' present address: Division of Hematology, 'Abbreviations used in this paper: cAMP, Department of Medicine, Veterans Administration Hospital, cyclicmonophosphate; DB-cAMP, N-2',O'-dibutyrylcAMP; University of Texas Health Science Center at San Antonio, DFP, diisopropyl fluorophosphate; E-PHA, erythroagglu- 78284. tinating phytohemagglutinin of Phaseolus vulgaris; PGEL, Tex. revised prostaglandin E; SDS, sodium dodecyl sulfate; TPCK, Received for publication 20 March 1975 and in ketone. form 9 June 1975. L- (tosylamido 2-phenyl) ethyl-chloromethyl 924 The Journal of Clinical Investigation Volume 56 October 1975 924-936 an attempt to define potential physiologic substrates fered saline, pH 6.5, containing 0.113 M NaCl, 0.0043 M of the platelet protein kinases. K2HPO4, 0.0043 M Na2HPO4, 0.0244 M NaH2PO4, and 0.0055 M glucose. The isolated platelets were resuspended We report here that thrombin treatment of intact at 0.1-1 X 10' platelets/ml in 35 ml isotonic Tris-buffered platelets results in a 2-6-fold increase in phosphorylation saline, pH 7.4, containing 0.14 M NaCl, 0.015 M Tris-HCl, of two proteins of mol wt 40,000 (peak 7) and 20,000 and 0.0055 M glucose ("resuspension buffer") and incu- (peak 9) and that agents which elevate intracellular bated for 15 min to deplete them of phosphate. The plate- cAMP block this reaction. In disrupted platelets proteins lets were then sedimented at 2,250 g for 15 min at room temperature and resuspended in 35 ml of the same buffer of molecular weight similar to peak 7 and 9 proteins for a further 15-min incubation. The platelets were sedi- are phosphorylated by [e-_-P]ATP. However, these mented again at 2,250 g for 15 min and resuspended in a latter phosphorylation reactions are unaffected by either small volume of resuspension buffer, and the platelet con- thrombin or cAMP, suggesting that the normal control centration was estimated from the packed cell volume of the platelet suspension by assuming that a 1% packed cell mechanisms of platelet protein phosphorylation are dis- volume equals 1 X 10' platelets/ml (3, 15). The platelet rupted during cell lysis. suspension was adjusted to contain 2 X 10' platelets/ml Further, we have studied the characteristics of two and 0.5 mCi/ml '3PO4 adjusted to pH 7-7.5 with NaOH. platelet protein kinases resolved from platelet homog- Incubation at room temperature for 60 min served to load enates, one of which appears to be a typical cAMP- the platelets with 'PO4. For most experiments reactions were carried out in a total volume of 0.025-0.125 ml in dependent holoenzyme while the other is unaffected by either 500-,ul polyethylene or 1.5-ml polypropylene micro- cAMP. The cAMP-dependent enzyme fraction catalyzes test tubes (Brinkmann Instruments, Inc., Westbury, N.Y.). the phosphorylation of a protein that co-electrophoreses Platelets prepared from fresh whole blood were used for with peak 7 protein. This phosphorylation is stimulated most studies of intact platelets, and the use of outdated platelets for a particular experiment is indicated in the text. twofold by cAMP. Preparation of crude supernate for protein kinase studies. Platelets for protein kinase studies were isolated in isotonic phosphate-buffered saline, pH 6.5, from either fresh blood METHODS or 72-h-old platelet concentrations as described above. They Isotopes were purchased from the following sources: [y- were then washed an additional time with resuspension 'P]ATP (5 Ci/mmol), serotonin [2-'4C]binoxalate (27.5 buffer and suspended in the same buffer at a concentration mCi/mmol), [fy-_P]GTP (26 Ci/mmol), and [3H]cAMP of 20-30 X 10' platelets/ml. The platelet suspension was (20 Ci/mmol), New England Nuclear, Boston, Mass.; car- then frozen in liquid nitrogen and stored at -90'C until rier-free H3'PO4 (50 mCi/ml) and carrier-free ['SI]so- a sufficient quantity had been accumulated. To obtain a dium iodide (50 mCi/ml), Mallinckrodt Chemical Works, crude supernatant fraction, the frozen suspensions were St. Louis, Mo. Millipore filters (RAWP-025-000, 1.2-Am thawed at 37'C and sonicated on ice with constant stirring. pore size, 25-mm diameter) were purchased from Millipore The particulate fraction was pelleted by centrifugation at Corp., Bedford, Mass., and Gelman filters type GN6, 0.45- 4VC at 48,000 g for 30 min. Preliminary experiments ,m pore size, 25-mm diameter, were obtained from Fisher showed that approximately three-quarters of the protein Scientific Co., Pittsburgh, Pa. RP/R-54 Royal X-Omat kinase activity when assayed with exogenous substrates is medical X-ray film was obtained from Eastman Kodak in the supernatant fraction after this treatment. Higher Co., Rochester, N. Y. Highly purified (>94% pure) yields of supernatant protein kinase activity (as much as Escherichia coli alkaline phosphatase was a gift of Dr. M. twofold higher from equal amounts of platelets) were Schlesinger (11). Dowex 1 (AG1-X8, 100-200 mesh, chlo- obtained when the platelets were frozen before sonication ride form) and AG 501-X8 mixed bed ion exchange resin than when not frozen but sonicated immediately after iso- were purchased from Bio-Rad Laboratories, Richmond, lation. The particulate fraction from platelets frozen be- Calif. 1-Methyl-3-isobutylxanthine was a gift from Dr. fore sonication still contained residual protein kinase ac- Lewis Chase. Prostaglandin E1 (PGE1) was a gift from tivity when assayed with exogenous substrates. Dr. John Pike, the Upjohn Co., Kalamazoo, Mich. Sodium dodecyl sulfate (SDS) polyacrylamide gel elec- Bovine thrombin, purified from topical thrombin (Parke, trophoresis. Platelets to be fractionated by SDS poly- Davis & Company, Detroit, Mich.) by the method of Glover acrylamide gel electrophoresis were solubilized by boiling and Shaw (12), yielded a clotting activity of 2,700 U/mg for 5 min in 2-4%o SDS, 0.1 M 2-mercaptoethanol, and thrombin. Homogeneous human thrombin with a clotting 0.02 M sodium phosphate buffer, pH 7.4, ("SDS buffer"). activity of 2,600 U/mg was a gift from Dr. John Fenton. After boiling in SDS buffer, the samples remained stable L- (Tosylamido 2-phenyl) ethyl-chloromethyl ketone (TP- for over 30 min at room temperature and could be stored CK) trypsin was purified by the method of Kostka and Car- at 4VC for several hours or overnight at - 120'C as judged penter (13). Erythroagglutinating phytohemagglutinin of by comparison of the radiochromatograms of SDS gels of Phaseolus vulgaris (E-PHA) was purified from Bacto phy- fresh and stored samples. tohemagglutinin P (Difco Laboratories, Detroit, Mich.), as Cylindrical SDS polyacrylamide gel electrophoresis was previously described (14). performed essentially as described by Weber and Osborn Platelet preparation and "P04 loading. Platelets were (16). Bromphenol blue in 80% sucrose was added to isolated by differential centrifugation as previously de- samples of 5 X 10' solubilized platelets (100 ,ug protein) scribed from fresh EDTA-anticoagulated human blood or which were layered on 5%o polyacrylamide gels with 2.3% f rom acid citrate dextrose-anticoagulated 72-h-old platelet bisacrylamide and 0.1% SDS in 0.1 M sodium phosphate concentrates (for large-scale preparations) beginning with buffer, pH 7.4, in 10-cm-long, 5-mm inside diameter glass the first 2,250-g centrifugation (15). The isolation was car- tubes. Electrophoresis was carried out at 8 mA/gel for 3.5 ried out at room temperature in isotonic phosphate-buf- h with buffers of 0.1 M sodium phosphate, pH 7.4, 0.1% Platelet Protein Phosphorylation 925 SDS containing 0.1 M 2-mercaptoethanol. The gels were supernate from platelets prepared from 2 U fresh platelet- stained overnight in 12% trichloroacetic acid, 50% methanol, rich plasma (4 ml, 2.88 X 105 U protein kinase activity) and 0.03% Coomassie Brilliant Blue R and subsequently was diluted to 100 ml with 0.01 M imidazole chloride, pH destained by diffusion over 24 h in 7% acetic acid, 10% 6, and applied to a histone-Sepharose column (bed volume methanol, and 0.85% phosphoric acid. The gels were frozen of 25 ml) that had previously been equilibrated with 0.01 and divided into 2-mm slices with a razor blade type slicer M imidazole chloride, pH 6. After loading, the column was (Bio-Rad Laboratories), and the slices were assayed for washed with 70 ml 0.01 M imidazole chloride, pH 6, and 'PO4 in a liquid scintillation counter after the addition of the enzyme fractions were then eluted with a linear 0-0.6 Bray's solution (17). M NaCl gradient in 0.01 M imidazole chloride, pH 6 SDS polyacrylamide slab gel electrophoresis was per- (total gradient volume was 240 ml). All steps were performed by a modification of the methods of Laemmli (18) formed at 4°C. 7-ml fractions were collected. Protein and of Weber and Osborn (16) in an apparatus similar to kinase activity with protamine as substrate and cAMP that described by Reid and Bielski (19). The stacking and binding activities were determined on each fraction, as well-forming gel consisted of 3% acrylamide with 3.2% bi- described below. For a large-scale preparation, the super- sacrylamide and 0.1% SDS in 0.05 M sodium phosphate nate from platelets prepared from 150 U 72-h-old platelet buffer, pH 6.4. The separating gel and the buffers were concentrates (158 ml, 1.31 X 107 U protein kinase activity) the same as described above for the cylindrical gels. Sam- was diluted with 4 vol 0.01 M imidazole chloride, pH 6, ples of 25-100 ug platelet protein were layered onto the and chromatographed on a 375-ml-bed volume histone- acrylamide in each of the 12 wells of a 2-mm thick ana- Sepharose affinity column. The total gradient volume was lytical slab gel. Electrophoresis was carried out at 30 V 8 liters. Every fifth fraction (20 ml each) was assayed until the tracking dye had entered the separating gel and for protein kinase activity with protamine as substrate. then at 60 V for 3 h. Staining and destaining were per- Pooled fractions of protein kinase after histone-Sepharose formed as for cylindrical gels. The gels were dried on chromatography were concentrated by using a DEAE- Whatman no. 50 hardened filter paper and exposed to cellulose column equilibrated with 0.01 M imidazole chlo- Kodak RP Royal X-Omat medical X-ray film that was ride, pH 6, (2.5 X 3 cm for up to 3 X 109 U enzyme activ- subsequently developed with a Kodak RPX-Omat developer. ity) at 4°C. The enzyme fraction was loaded onto the When indicated, the autoradiograph was scanned with a column by diluting the histone-Sepharose fraction with 2 spectrophotometer equipped with linear transport and film vol 0.01 M imidazole chloride, pH 6, in portions as they holding accessories (Gilford Instrument Laboratories, Inc., were applied to the column. After washing with the dilut- Oberlin, Ohio). ing buffer, the enzyme was eluted by washing the column Molecular weight determination by SDS polyacrylamide with 0.01 M imidazole chloride, 1 M NaCl, pH 6. Frac- gel electrophoresis (16) employed the following standards: tions (2 ml) were collected and analyzed for protein E. coli. p-galactosidase, 130,000; bovine fibrinogen alpha kinase activity. The peak tubes were pooled. This method chain, 73,000; bovine serum albumin, 68,000; bovine fibrino- usually gave a 10-15-fold concentration of activity with gen beta chain, 60,000; bovine fibrinogen gamma chain, a recovery of approximately 80%. No changes in the spe- 53,000; E. coli alkaline phosphatase subunit, 43,000; rabbit cific activity of the enzymes were observed. muscle aldolase, 40,000; hog stomach mucosa pepsin, 35,000; Preparation of regulatory subunit from rabbit muscle the B chain of human a-thrombin, 33,000; and sperm whale protein kinase. Rabbit muscle cAMP-dependent protein myoglobin, 17,200. kinase was prepared by a slight change in the modification Partial purification of peak 7 protein by preparative SDS (21) of a previously described (22) method. The method polyacrylamide gel electrophoresis. 3 ml of platelets con- of Corbin et al. (21) was followed except that a histone- taining 2 X 109 platelets/ml were incubated with 'PO4 as Sepharost affinity column was used and the buffer was 0.01 described above. Thrombin, 1 U/ml, was added to the M imidazole chloride-0.001 M EDTA, pH 6. Frozen rab- platelet suspension, and after a 5-min incubation the plate- bit muscle was used. The regulatory and catalytic subunits lets were solubilized in SDS buffer and the entire volume of this protein kinase were then separated by incubation was electrophoresed in a single well of a 4-mm thick SDS with [3H] cAMP and chromatography on a histone-Sepha- polyacrylamide slab gel. The protein in the region of the rose column as described previously (23) where casein- gel containing peak 7 protein was eluted as previously Sepharose was used. The regulatory subunit eluted after described (20) and adjusted to contain 10 mM diisopropyl the catalytic subunit on this column. fluorophosphate (DFP). The solution was lyophilized, dis- Preparation of partially-purified, heat-stable inhibitor of solved in 1 ml of H20, and then dialyzed for 24 h with cAMP-dependent protein kinases. The heat-stable inhibi- one change of dialysis fluid against 500 ml of resuspension tor was prepared as previously described (24) from 1 lb buffer containing 0.01% SDS. Single peaks of Coomassie of frozen rabbit muscle. The purification was carried Blue staining and radioactivity with the same mobility as through the trichloroacetic acid precipitation step, subse- peak 7 protein were seen on SDS polyacrylamide gel elec- quent resolubilization, and dialysis. trophoresis of this material, suggesting that it had remained Determination of molecular weights by Sephadex G-150 intact during purification. Approximately 40% of peak 7 chromatography. A 2.5 X 61-cm upward flow Sephadex protein was recovered in the elution step. G-150 column was calibrated by determining the elution Preparation of histone-Sepharose 4B. Histone-Sepharose volume (V.) of known molecular weight standards: rabbit was prepared as previously described (21) except that after muscle lactate dehydrogenase, 132,000; bovine serum albu- coupling of the histone to the activated Sepharose, the min monomer, 69,000; 12.I-ovalbumin, 45,000; 'I-diisopropyl histone-Sepharose was washed with 10 vol of 0.01 M phosphoryl-thrombin, 37,000; and sperm whale myoglobin, imidazole chloride, 1 M sodium chloride (pH 6), then 10 17,800. Blue Dextran (mol wt > 2 X 106) was used to cal- vol of 0.01 M imidazole chloride, pH 6, and stored in the culate the void volume (V.), and dinitrophenyl-glycine was same buffer. utilized to calculate the included volume (V6). The pro- Histone-Sepharose affinity chromatography of crude plate- teins were monitored by their absorbance at 280 nm or by let supernate. In a typical small-scale preparation, the measuring their radioactivity. Dinitrophenyl-glycine was 926 R. M. Lyons, N. Stanford, and P. W. Majerus monitored at 410 nm. The radioactive proteins were prepared by the chloramine-T method of Hunter (25). A standard curve was obtained by plotting log molecular weight vs. (V.-VV)/(V --Vo) (26). Protein kinase assav. The standard reaction mixture contained in a total volume of 0.1 ml: 0.05 MI Tris chloride (final pH = 8.31), 7 mM magnesium acetate, 5 tM cAMP Cy- (where added), 0.125 mM [,y-3P]ATP (sp act 5-30 cpm/ 300 pmol), 5 mM dithiothreitol, 0.1 mg protamine or 0.1 mg QC histone fraction f2b (Sigma Chemical Co.), and the enzyme to be assayed. In the study of the rabbit skeletal muscle 0 enzyme, potassium phosphate pH 6.8 was substituted for the z 300 10 2 0 4 Tris chloride. The reaction was initiated by the addition .1 of enzyme and incubated for 5-20 min (depending on level 0'1,000 of enzyme activity) at 37'C. The reaction was terminated by adding 4 ml of cold 20% (wt/vol) trichloroacetic acid and placing the tubes on ice for at least 10 min. The samples were filtered through Gelman type GN-6 (0.45-,um 0 10 20 30 40 pore size, 25-mm diameter) filters with a Millipore 3025 FRACTION NUMBER manifold filtration system, and the precipitate was washed wvith two additional 4-ml portions of cold 20% trichloro- FIGURE 1 SDS polyacrylamide gel radiochromatogram of acetic acid (27). The filters and precipitates were dissolved the effect of thrombin on platelet protein phosphorylation. in 10 ml Bray's solution (17) and counted in a Packard Platelets were loaded with 'PO4 (0.5 mCi/ml) as described Tri Carb liquid scintillation spectrometer (Packard In- in the text and then electrophoresed on SDS polyacrylamide strument Co., Inc., Downers Grove, Ill.). "1 U" of protein gels. (*) 'PO4 incorporation into control platelets; (0) kinase activity was defined as the amount of enzyme that '3PO4 incorporation into platelets treated with thrombin, catalyzed the incorporation of 1 pmol of 'PO4 into protein 1 U/ml, for 5 min. in 1 min. All assays were shown to be linear with respect to time and enzyme concentrations. The product of the protein kinase reaction was stable to boiling in 0.1 M so- (two 15-min incubations) and subsequent loading with dium phosphate, pH 7.4, for 15 min, which would be ex- pected to exchange any noncovalently bound phosphate. `2P04 did not damage the platelets with respect to their Likewise, the product was labile to heating in 0.1 M so- ability to undergo thrombin-induced [14C] serotonin re- dium hydroxide, suggesting the presence of a phosphomono- lease. Shorter periods of phosphate depletion or loading ester linkage. diminished the incorporation of 'PO4 into platelet pro- cAMP binding assay. cAMP binding was measured by teins. While we refer a modification of the method of Walton and Garren (28). to this process as "phosphate load- Saturating levels of ['H]cAMP were used so that the ing," we do not know what changes, if any, occur in assay was proportional to the amount of binding proteins. intracellular phosphate. Rather, these conditions are Phosphodiesterase assay. Cyclic nucleotide 3',5'-phospho- those that we found to result in maximal protein label- diesterase activity was measured by the method of Beavo ing in control platelets. et al. (29) except that Crotalus adaniantents venom was used as the source of 5'-nucleotidase and Dowex 1 was used There was prominent protein phosphorylation ob- to separate the nucleoside from the nucleotide. Incubations served in control 'PO4-loaded platelets as shown in the were at 37'C for 10 min and were terminated by boiling. radiochromatogram of the cylindrical SDS polyacryla- Blank values were higher than those reported by B 2avo mide gel shown in Fig. 1. 9 radioactive peaks were con- et al. (29) in that 10-20% of the counts of [3H]cAMP eluted in the Tris wash when no enzymes were added. sistently seen in radiochromatograms of cylindrical gels, Misccllaneons methods. Platelet suspensions were dis- while approximately 20 radioactive proteins could be rupted by sonication for 15 s on ice with the microprobe distinguished in the autoradiographs of SDS polvacryla- of a Biosonik II sonifier set at 70% intensity for volumes mide slab gels. The increased number of radioactive pro- less than 3 ml and the standard probe under the same con- teins seen with the latter ditions for volumes greater than 3 ml (15). [14C] Serotonin technique is due to the in- release studies were performed as previously described, the creased resolution of radioactivity seen on autoradio- platelets having been incubated with ["C] serotonin before graphs as compared to sliced gels counted in a liquid the phosphate depletion step (3). Alkaline and acid phos- scintillation counter. When 3PO4-loaded platelets were phatase were assayed as described by Lindhardt and Walter treated with thrombin, 1 U/ml, for 5 min before boiling (30). Thrombin was assayed as described by Seegers and Smith (31). Protein was estimated by the method of Lowry in SDS buffer, a different pattern Of 'PO4 incorporation et al. (32). was observed. Either bovine or human thrombin induced a 2-6-fold increase (mean 3.28-fold-+-0.27 SEM, n = 13) in phosphate incorporation into a peak of 'PO4 corre- RESULTS sponding to the darkest staining protein band on the Effect of thrombin on protein phosphorylation in intact SDS polyacrylamide gel. This protein had an apparent platelets. In preliminary experiments we determined that molecular weight of 40,000 and was designated "peak 7." the prolonged period required for phosphate depletion A similar increase in phosphorylation (2.84-fold±0.4 Platelet Protein Phosphorylation 927 the stability of phosphomonoesters of serine and threonine (35) and indicate that the phosphate is not incorporated into RNA, which is labile in hot acid (36). 0L U Partial acid hydrolysis of "PO4-labeled delipidated plate- 1,000 o let protein followed by high voltage paper electrophoresis (37) disclosed the presence of labeled phosphoserine and LUZ phosphothreonine residues. When corrected for hydro- ~_< i lysis of phosphoserine and phosphothreonine (37), over 80% of the label was determined to be in the form of 0 0 phosphoserine. When peak 7 protein, eluted from a 4-mm thick SDS z polyacrylamide gel as described in Methods, was treated 0 0.5 1.0 with hot alkali or hot acid, results similar to those de- THROMBIN U/mi scribed above for the delipidated platelet protein were FIGURE 2 Effect of increasing concentrations of thrombin observed. Treatment of peak 7 protein with TPCK on 3"P0 incorporation into peak 7 protein. 3"P0-loaded trypsin (0.3 mg/ml) for 21 h at 370C, pH 7.4, resulted platelets were incubated with various thrombin concentra- in a 63% decrease in radioactivity in peak 7 protein tions for 5 min at room temperature before the platelets and a similar decrease in Coomassie Blue staining upon were solubilized and electrophoresed on cylindrical SDS re-electrophoresis of the protein on SDS gels. Incubation polyacrylamide gels as described in Methods. The radioactivity in each peak was calculated by summing the counts of peak 7 protein for 15 h at 450C, pH 8.0, with per minute of the three highest fractions in the peak. E. coli alkaline phosphomonoesterase, 7 U/ml, resulted in 78% loss of radioactivity from peak 7 protein as SEM, n = 9) was consistently seen in a peak of approxi- determined by re-electrophoresis of the protein on SDS mately 20,000 mol wt designated "peak 9." There was gels. no consistent thrombin-induced alteration of 32P04 Treatment of the delipidated platelet protein with incorporation into radioactive peaks other than 7 and sodium hydroxide or alkaline phosphatase as described 9 when the radioactive peaks were examined by either above resulted in a 65 and 73% decrease, respectively, cylindrical or slab SDS polyacrylamide gel electrophore- in radioactivity in peak 7 protein as assessed by sub- sis. In contrast, thrombin that was inhibited by DFP at sequent SDS polyacrylamide gel electrophoresis of the its active serine residue, had no effect on the incorpora- samples. These experiments indicate that the radio- tion of "POt into platelet protein. activity in peak 7 and peak 9 is in the protein in the To examine the possibility that release-inducing agents form of phosphomonoesters of serine and threonine other than thrombin could initiate platelet protein phos- residues. phorylation, we studied the effects of E-PHA (14, 33) Effect of thrombin concentration on peak 7 protein on platelets. Exposure of 3"P04-loaded platelets to E- and peak 9 protein phosphorylation. The effect of PHA (0.13 mg/ml for 15 min) resulted in a 3.2-fold thrombin on peak 7 protein and peak 9 protein phos- and 3.9-fold increase in phosphorylation of peak 7 and phorylation was dependent upon thrombin concentration. 9 proteins, respectively. The thrombin concentration that produced half-maximal Characterization of peak 7 and peak 9 protein. We phosphorylation was 0.25 U/ml for both peak 7 protein initially demonstrated that the radioactivity in peak 7 (Fig. 2) and peak 9 protein. was covalently bound to protein. 3"POt-loaded platelets Time-course of thrombin-induced peak 7 protein and were precipitated with cold 10% trichloroacetic acid and peak 9 protein phosphorylation. Thrombin had no effect then delipidated as described by Cohen et al. (34). The on the phosphorylation of peak 7 protein or peak 9 pro- SDS polyacrylamide gel Coomassie Blue staining and tein when platelets were disrupted before thrombin treat- radiochromatogram patterns of the material remaining ment (see below). Thus, by lysing the platelets with after lipid extraction were similar to those described SDS, we were able to stop the interaction between plate- above for whole solubilized platelets, suggesting that the lets and thrombin rapidly, allowing us to determine the radioactivity in peaks 7 and 9 is not associated with time-course of the effect of thrombin on peak 7 and phospholipids. When samples of the delipidated tri- peak 9 protein phosphorylation. The effect of thrombin, chloroacetic acid precipitate from thrombin-treated plate- 1 U/ml, was rapid with the time for half-maximal phos- lets were boiled for 15 min in 0.2 M sodium hydroxide phorylation of peak 7 protein being 10 s (Fig. 3). The we found a 90% loss of 32POt as measured by trichloro- time for half-maximal phosphorylation of peak 9 pro- acetic acid precipitation. In contrast, treatment in 10% tein (14 s) was similar (data not shown). trichloroacetic acid for 15 mmn at 90°C resulted in only The thrombin concentration curve and the time-course 8% hydrolysis of "2PG4. These results are consistent with for thrombin-induced phosphorylation of peak 7 and

928 R. M. Lyons, N. Stanford, and P. W. Majerus peak 9 proteins were similar to those parameters previously observed for the thrombin-induced platelet release reaction and for the inhibition of platelet adenylate cyclase by thrombin (5, 33, 38, 39). Effect of raising intracellular cAMP concentrations on phosphorylation of peak 7 and 9 proteins. PGE1, 1 ,ug/ml, and N6-2',0'-dibutyryl cAMP (DB-cAMP), 1mM, compounds that raise the intracellular cAMP levels in platelets (2) and partially block the thrombin- Peak 7 - x4 induced release reaction (5, 40), did not alter platelet protein phosphorylation significantly in the absence of thrombin. However, treatment of intact platelets with Peak 9 these agents before the addition of thrombin resulted in a 48 and 53% inhibition, respectively, of peak 7 protein phosphorylation, as estimated from densitometry of the 1 2 3 4 5 6 7 8 9 10 11 12 autoradiographs as described in Methods. Similarly, thrombin-induced peak 9 protein phosphorylation was in- FIGURE 4 Autoradiograph of an SDS polyacrylamide slab hibited 55 and 70% by PGE1 and DB-cAMIP, respec- gel electrophoresis showing the effect of PGE1 (1 ug/ml) tively. The inhibition of thrombin-induced peak 7 and or DB-cAMP (1 mM) on platelet protein phosphorylation. Platelets were loaded with '0PO4 (0.5 mCi/ml) as described peak 9 protein phosphorylation by PGE1 and DB-cAMP in Methods. Gels 1 and 2: no additions; 3 and 4: throm- is shown in Fig. 4. bin, 1 U/ml, for 5 min; 5 and 7: PGE1 alone for 20 min; Molecular weight of peak 7 protein by gel filtration. 6 and 8: PGE1 for 15 min with the addition of thrombin, Platelets (70 ml of 4 X 109 were isolated 1 U/ml, for a further 5 min; 9 and 11: DB-cAMP alone platelets/ml) for 20 min; 10 and 12: DB-cAMP for 15 min with the from 10 U of 72-h-old platelet concentrates, loaded with addition of thrombin, 1 U/ml, for a further 5 min. The 0.02 mCi/ml 32P4, and treated with 4 U/ml thrombin. radioactivity in peak 7 and 9 proteins was assessed from These platelets retained the ability to aggregate and to the A500 of the bands on the autoradiograph of an SDS release ["C]serotonin when exposed to thrombin. The slab gel electrophoresis of the fractions. The peak ab- was dis- sorbance of each band was less than 2.0. At these levels platelets were sedimented, and the supernate the peak absorbance of each band was proportional to the carded before sonication of the platelets in 12.5 ml radioactivity of the same band on the SDS polyacrylamide 0.1 mM sodium phosphate buffer, pH 7.4, containing gel as assessed in a liquid scintillation spectrometer. 10 mM DFP. The particulate material was sedimented at 165,000 g X 60 min at 4VC, and 100 mg of the super- Three peaks of radioactivity were eluted from the col- natant protein in 10 ml (containing 75% of the total umn. The first peak was detected at the void volume, platelet peak 7 protein) was chromatographed at 4VC while the third peak eluted just ahead of the included in 0.1 Ml sodium phosphate buffer, pH 7.4, on a cali- volume. The second peak of radioactivity eluting from brated Sephadex G-150 column as described in Methods. the column, subsequently identified as containing peak 7 protein by SDS polyacrylamide gel electrophoresis of the fractions, had an apparent molecular weight of 49,000. This value is similar to the molecular weight of 2,000 peak 7 protein determined by SDS polyacrylamide gel electrophoresis. Furthermore, the apparent molecular weight of the radioactive protein in peak 7 as determined

Ratio peak 7 radioactivity Supernatant Thrombin (1 U/ml) cAMP (1 pM) Gel Whole sonicates fraction during labeling during labeling peak 7 protein 1 + - - - 2.1 2 + - - + 2.3 3 + - + - 2.2 4 + - + + 2.0 5 - + - - 3.5 6 - + - + 3.3 7 - + + - 3.4 8 - + + + 3.3 Platelet samples (4 X 109/ml) were sonicated and centrifuged to obtain the supernatant fraction where indicated. Endogenous phosphorylation was measured by incubation with 125 OM [Ly-"2P]ATP (244 ,/Ci/mol), 5 mM magnesium acetate-0.5 mM 1-methyl-3-isobutylxanthine for 10 min at 220C. 10-iul portions of each reaction mixture were subjected to SDS polyacrylamide slab gel electrophoresis. Autoradiograph film exposure 7 h. The autoradiograph and a negative photograph of the stained gel were scanned at 500 nm in a Gilford spectrometer as described in Methods. Units are peak absorbance at 500 nm as described in the legend of Fig. 4. Results are expressed as the ratio of peak absorbance on the autoradiograph/peak absorbance of a negative photograph of the stained gel. tion amounted to less than 5% of the total activity and cated platelets and the supernatant fraction of sonicated has been subtracted in all activities determined with platelets were studied using higher specific activity exogenous substrates reported hereafter. Crude sonicates [y-"P]ATP (200-1,000 cpm/pmol), we found no of fresh platelets contained protein kinase activity that change in total incorporation of label into trichloroacetic was only slightly stimulated by added cAMP when acid-precipitable material with or without 1 U/ml bovine assayed with protamine as substrate (1.1-1.2-fold) and thrombin present during the labeling period. When the was stimulated 1.5-fold when histone f2b was used as samples were electrophoresed on an SDS slab gel, an substrate. After centrifugation at 48,000 g for 30 min, autoradiograph showed most of the label was contained 65-75% of the total protein kinase activity remained in in peak 7 and 9 proteins, and no significant change in the supernatant fraction, and the cAMP dependence re- labeling pattern was observed when thrombin was added mained unchanged. We attempted to alter either the level during labeling. Likewise, addition of 1 AM cAMP in of activity or degree of cAMP stimulation in the crude the presence of the phosphodiesterase inhibitor 1-methyl- sonicates or supernatant fraction by treating platelets 3-isobutylxanthine with or without 1 U/ml thrombin with thrombin, 1 U/ml, before sonication. There was showed no difference in incorporation of "PO4 into no observed change in total protein kinase activity or trichloroacetic-precipitable material. The autoradiograph ratio of activity -cAMP/+cAMP despite the presence also showed no significant change in labeling pattern. of the phosphodiesterase inhibitor 1-methyl-3-isobutyl- There was a slight decrease in phosphorylation of peak xanthine with either protein substrate (41). Similar 7 and 9 proteins when platelets were treated with experiments using adipose tissue have demonstrated hor- thrombin before disruption and measurement of endo- mone-induced alterations in the state of activation of genous phosphorylation (data not shown). This result protein kinases (41). In experiments where platelets could be explained by increased saturation of protein were incubated with PGE1 for 20 min before sonication, substrates due to phosphorylation induced by thrombin the total amount of exogenous protein kinase activity before disruption; however, we have no direct evidence with each substrate was unchanged, but the activity in to support this hypothesis. the absence of cAMP with histone f2b as substrate was In experiments without the phosphodiesterase inhibi- elevated 1.6-fold, suggesting that if there had been a tor, identical results were obtained, indicating that de- change in the degree of cAMP stimulation of protein spite the high levels of phosphodiesterase in platelets kinase activity by thrombin, our method could have (370 pmol cAMP hydrolyzed/min per 10' platelets) detected it. cAMP does not appear to regulate endogenous phos- Endogenous phosphorylation of crude platelet extracts. phorylation in crude extracts of disrupted platelets. When the endogenous phosphorylation activities of soni- Table I shows results of experiments in which the 9.30 R. M. Lyons, N. Stanford, and P. W. Majerus labeling was carried out by using either the whole TABLE II sonicate(l platelet fraction or the supernatant fraction lr(i(-I onl(/ inoi IlHurni n Platelet P'roltein KiIoINs front sonicated platelets. The lower labeling of peak 7 protein in the whole sonicate compared to the separated Protein kinase activity cAMP )inlding supernatant fraction cannot be explained simply. The Crude stlj)ernatant prepareti from: Total SI) act atc tivitv particulate fraction may inhibit phosphorylation or contain phosp)hoprotein phosphatases. When the separated U l /'ng prolfin pi7ol particulate was 2 U 72-h-old platelet 4.5 X l05+ 8,385+ 468 fraction incubated with [y-y-P]ATP, no concentrates 2.0 X 105* 3,719* significant labeling was observed on the autoradiograph. cAMP-dependent enzyme 1.5 X 1(5+ 9.846+ 262 WAhen the whole platelet sonicate was labeled and 3.2 X 1(1* 2,146* then cAMP-independent enzyme 2.7 X 10'+ 8,234+ 13 separated into supernatant and particulate fractions and 4.4 X 104* 13,439* electrophoresed, approximately 10-15% of the label appearing in peaks 7 and 9 was located in the particulate Protein kinase activities in the presence of cAMP are expressed in units at pH 8.3 with protamine as substrate (*) or histone f2b as substrate (+) as fraction. The labeling pattern of the supernatant frac- described in Methods. cAMP binding was measured at pH 8.4 as ilescribed tion upon electrophoresis was identical to that obtained in Methods. when the separated supernatant fraction was labeled. Thus, (a) the particulate fraction contains very little experiments with platelets treated with thrombin before protein kinase but some substrate, (b) the particulate sonication showed an essentially identical chromatogranm. fraction does not contain additional protein kinase(s) The two enzyme fractions were designated "cAMP-de- that phosphorylate additional supernatant proteins, and pendent" and "cAMP-independent," referring to the (c) the supernatant fraction contains protein kinase earlier (at low salt) and later (at higher salt) eluting plus the peak 7 and 9 protein substrates. peaks, respectively. Initially, protamine was used as a Platelet protein kinase fractionation on histonie- screening substrate to monitor column elution. Since Sepharose. In an attempt to characterize the protein the preferential substrate for the cAMIP-dependent kinase enzymes in the platelet supernatant fraction, we enzyme was histone f2b, an experiment was performed in chromatographed this material on histone-Sepharose as which both protamine and histone f2b in the presence of described in Mlethods. A typical elution pattern of a cAMP were used as substrates in analyzing a histone- small-scale preparation (crude supernate prepared from Sepharose affinity column. The same two enzyme frac- 2 U of platelet-rich plasma) is shown in Fig. 5. Parallel tions were found as in Fig. 5, with the relative peak enzyme activities reversed with the two substrates (i.e., with histone f2b as substrate, the major peak was the cAMAIP-dependent enzyme fraction; with protamine as substrate, the cANIP-independent enzyme was the larger peak). A summary of enzyme activities recovered for a typical small-scale experiment is shown in Table II. The true recoveries of the enzyme activities are greater than the values shown since approximately 20% more 0 activity was in the region between the two peaks which E was not pooled because of overlap of the two enzymes. -15 E- In a large-scale experiment requiring 24 h to complete, E -10 a only about A-4 of the total protein kinase activity was oZ recovered after histone-Sepharose 0 chromatography. Smaller-scale experiments generally gave higher -5 o- yields E since they could be completed within a few hours. Stor- u age of portions of the crude extracts or separated enzyme FRACT ION fractions at 4VC for 24 h resulted in loss of up to 3 of protein kinase activity, suggesting that the poor recovery- FIGURE 5 Elution pattern of protein kinase activity of in large-scale experiments may result crude platelet supernate from histone-Sepharose affinity from enzyme in- column. Supernate from platelets prepared from 2 U plate- activation during fractionation. let-rich plasma was chromatographed as described in Further studies and purification of these enzymes have Methods. (0) Protein kinase activity in the presence of proved difficult due to their instability. Thus, neither cAMP with protamine as substrate; (0) protein kinase enzyme could be concentrated by Amicon activity in the absence of cAMP with protamine as sub- ultrafiltration. strate; (0) [3H]cAMP binding activity; the solid line is Overnight dialysis against a variety of buffers resulted absorbance at 280 nm. The sodium chloride gradient was in approximately 50% inactivation of the cAM\P-de- begun at fraction 27. pendent enzyme and over 90% inactivation of the

Platelet Protein Phosphorylation 931 of this protein kinase by the relatively crude preparation 68r A of heat-stable inhibitor is not due to the inhibitor's acting 64 A was no increased in- A as a substrate since there 32PO4 A corporated into trichloroacetic acid-precipitable material 60 o A from - when protamine (protein acceptor) was omitted c I. the reaction mixture. > 24 Effect of thrombin on the activity of the platelet pro- U 20 tein kinases. When we incubated both separated enzyme fractions with varying concentrations of thrombin (0.1- u 16 10 U/ml) for varying periods of time (1-60 min) at Z 12 room temperature, we observed no alteration of activity or change in cAMP dependence when assayed with exo- 8 0 genous substrates in any of these experiments when 0 4~ comparing thrombin-treated to untreated enzyme. Properties of platelet protein kinases. The platelet 0 10 20 30 40 50 60 70 protein kinases were unlike in their lability upon heating REGULATORY SUBUNIT ADDED of the DEAE-concentrated enzymes (6,000 U/ml) at (Pg) 37° C. Thus, the cAMP-dependent enzyme lost 30% FIGURE 6 Effect of rabbit muscle regulatory subunit on of its activity after 1 h while the cAMP-independent the platelet cAMP-dependent and cAMP-independent en- enzyme lost 85% of its activity. Both enzymes require zymes in the presence and absence of cAMP. The assays Mg"+ for activity, and neither enzyme showed any activ- were performed as described in Methods except that the ity when Ca"+ was the only divalent cation present. In enzymes were incubated with the indicated amount of regulatory subunit for 30 min at 0'C before initiating the the presence of 7 mM Mg"+ ions, calcium markedly in- protein kinase assay. Protamine was the substrate with the hibited protein kinase activity with 50% inhibition of cAMP-independent enzyme and histone f2b was the sub- activity of both enzymes occurring at 1-2 mM Ca". strate with the cAMP-dependent enzyme. (0) cAMIP- The platelet protein kinase enzymes were also readily dependent enzyme in the absence of cAMP; (0) cAMP- dependent enzyme in the presence of cAMP; (A) and (A) cAMP-independent enzyme in the absence and presence of cAMP, respectively. 320 5 cAMP-independent enzyme. Numerous attempts to stabi- 28 -- - lize the enzymes proved unsuccessful although they were more stable after concentration on small DEAE-cellulose 2: columns (see Methods). Characterization of platelet protein kinases with > 20 u rabbit muscle proteins. Experiments using rabbit < 16 muscle regulatory subunit and rabbit muscle heat- LL stable inhibitor were carried out to establish that the cAMP-dependent enzyme was a typical cAMP- Ez 128S regulated protein kinase and also to determine whether the cAMP-independent enzyme was the free catalytic 4 subunit of a cAMP-regulated holoenzyme. In the presence of the rabbit muscle regulatory subnuit, the cAMP- O 50 1 00 150 260 250 dependent enzyme was inhibited in the absence of cAMP RABBiT MUSCLE INHIBITOR ADDED (pg) but was unaffected in the presence of cAMP as shown in FIGURE 7 Effect of rabbit muscle heat-stable protein on Fig. 6. This result suggests that this enzyme is a cAMP- the platelet cAMP-dependent and cAMP-independent en- dependent holoenzyme. This conclusion was further sup- zymes in the presence and absence of cAMP. The assays in where inhibition were performed as described in Methods except that the ported by results shown Fig. 7, by enzymes were incubated with the indicated amount of rabbit muscle heat-stable inhibitor in the presence and inhibitor at 37'C for 10 min before initiating the protein to some extent in the absence of cAMP is demonstrated. kinase assay. The cAMP-dependent enzyme was assayed In contrast, the cAMP-independent enzyme does not with histone f2b as substrate and the cAMP-independent to the subunit of the enzyme with protamine as substrate. (0) cAMP-dependent apparently correspond catalytic enzyme in the absence of cAMP; (0) cAMP-dependent cAMP-regulated enzyme. Thus, it is not inhibited by enzyme in the presence of cAMP; (A) and (-) cAMP- regulatory subunit (Fig. 6) nor is it inhibited by the independent enzyme in the absence and presence of cAMP, heat-stable inhibitor (Fig. 7). The apparent stimulation respectively.

932 R. M. Lyons, N. Stanford, and P. W. Majerus distinguished by substrate specificity as can be seen in DISCUSSION Table II. Thus, the enzyme was cAMP-dependent most Our results indicate that in intact platelets thrombin with histone as active fLb substrate while the cAMP- induces a rapid 2-6-fold increase in phosphorylation of active this independent enzyme was less using substrate two platelet proteins of approximately 40,000 and the substrate. 20,000 with protamine being preferred Because mol wt as determined by SDS polyacrylamide gel electro- the substrate specificities were so it is different, highly phoresis. The phosphate is incorporated into protein as is improbable that the cAMP-independent enzyme the the phosphomonoesters of serine and threonine. free catalytic of We subunit the cAMIP-dependent enzyme, have demonstrated that peak 7 protein migrates the same supporting the evidence shown with the rabbit skeletal on SDS gels when either unreduced or reduced with muscle inhibitors 6 and (Figs. 7). 2-mercaptoethanol. Furthermore, the molecular weight When [y-32P]GTP was used as the phosphate donor of undenatured, unreduced peak 7 protein as determined in place of ATP, there was no activity with either by gel filtration on Sephadex G-150 (49,000) was simi- enzyme. Neither enzyme fraction contained any acid lar to the molecular weight of 40,000 determined by SDS phosphatase activity. When assayed for cANIP-phos- gel electrophoresis. Thus, peak 7 protein is not a co- phodiesterase activity, both enzyme fractions showed valently linked subunit of a larger protein. activity that was sufficient to destroy only 5-10% of There are many similarities between the characteristics the cAMP a standard kinase during protein assay. of the platelet release reaction and phosphorylation of of the Thus, the lack of stimulation by cAMP cAMP- peak 7 and peak 9 proteins. The published time of was not due to destruction of independent enzyme the 10-30 s for half-maximal thrombin-induced release cAMP a in the by phosphodiesterase present protein (33, 38, 39) is similar to the observed for kinase fraction. time half- maximal thrombin-induced phosphorylation of protein Endogenous phosphorylation using fractionated pro- peaks 7 and 9 (10-14 s). Likewise, the concentration tein kinases. Since the two enzyme fractions were of thrombin required for half-maximal release (5, 38, relatively crude preparations, we studied their endo- 39) is similar to the observed concentration of thrombin genous phosphorylation with high specific activity (0.25 U/ml) required for half-maximal phosphorylation [e--"P]ATP (200-1,000 cpm/pmol). The cAMP-depend- of peak 7 and peak 9 proteins. DFP-treated thrombin, ent enzyme showed increased incorporation of label into which binds normally to platelets without inducing the trichloroacetic acid-precipitable material when cAMIP release reaction (3), does not induce peak 7 protein or was added. When the reaction product was electro- peak 9 protein phosphorylation. The lectin E-PHA, phoresed on an SDS polyacrylamide slab gel and an which binds to the platelet surface, decreases adenylate autoradiograph was obtained, cAMP-stimulated in- cyclase activity, and causes the platelet release reaction corporation of label into a protein of mol wt 98,000 in washed human platelets (14-33), induces peak 7 was seen. Incorporation of label into a protein that co- protein and peak 9 protein phosphorylation. Thrombin electrophoresed with peak 7 protein was also observed, and E-PHA are presumably related in their mode of and this incorporation was increased twofold when action on platelets since elevation of intracellular plate- 2 gNM cAM1P was added (data not shown). This result is let cAMIP blocks their initiation of the release reaction in that the thrombin-induced paradoxical phosphoryla- (5, 33, 40). We observed that PGE1 or DB-cAMP in- tion reaction in intact platelets is inhibited by agents hibited thrombin-induced phosphorylation of peak 7 and that increase intracellular cAMP. There was no "PO, 9 proteins. These similarities between the platelet release label in the of region the gel corresponding to peak 9 reaction and phosphorylation of peak 7 and peak 9 protein. Whether this results from absence of substrate suggest a relationship between the two phenomena. or of the enzyme to inability phosphorylate peak 9 pro- However, our data do not indicate whether peak 7 be a tein could not determined. When similar experiment or peak 9 protein the phosphorylation is part of the release- using cAMP-independent enzyme was performed, inducing mechanism, a result of the release reaction, there was no of either 7 or 9 labeling peak peak proteins. or a simultaneous but unrelated event. Again, it was not possible to determine whether this In most cAMP-modulated systems an increase in pro- related to lack of substrates in the fraction or to spe- tein phosphorylation follows activation of a cAMP- cificity of the enzyme. These experiments indicate that dependent protein kinase by an increase in intracellular the cAMP-dependent enzyme may act on peak 7 protein cAMP. However, in the intact platelet system the usual although the results with the isolated enzyme indicate positive correlation between cAMNP levels and protein that other factors must be present in intact cells to phosphorylation does not appear to hold since an in- account for the clear stimulation of phosphorylation by crease in intracellular cAMP blocks the thrombin- thrombin and concomitant inhibition of phosphorylation induced peak 7 and peak 9 protein phosphorylation. by agents that elevate cAMP. Several examples of an inverse correlation between a Platelet Protein Phosphorylation 933 change in cAMP concentration and protein phosphoryla- isolated particulate fraction was incubated with [-y-"P]- tion have been described in the literature (7-10, 42-45). ATP even though this faction contains 1-A of the total In toad bladder mucosa it was observed that antidiuretic protein kinase activity when assayed with exogenous hormone induced a decrease in phosphorylation of mem- substrates. brane and soluble proteins in association with an increase Contradictory information has been presented in the in intracellular cAMP (7-9). In that system, a decrease literature in regard to phosphorylation of the particulate in phosphorylation of a specific membrane protein fraction of disrupted platelets. Booyse et al. have re- appeared to be the result of both the inhibition of protein ported phosphorylation of two proteins of mol wt kinase activity and the stimulation of phosphoprotein 44,000 and 19,000, using [-Y-"P]ATP with particulate phosphatase activity. Our results do not distinguish be- fractions from disrupted platelets (46). They suggested tween thrombin-induced stimulation of protein kinase that thrombin decreases platelet protein phosphorylation activity and thrombin-induced inhibition of phospho- of the 44,000-mol wt band and that cAMP increases protein phosphatase activity. platelet membrane phosphorylation. In contrast, Steiner Other mechanisms unrelated to cAMP could explain has presented preliminary data suggesting that thrombin the thrombin-induced phosphorylation of peak 7 and increases phosphate incorporation into isolated platelet 9 proteins. First, thrombin might alter the activity of an membranes in the presence of ['y-"P]ATP (47). Insuf- as yet undescribed guanosine 3',5'-cyclic monophosphate- ficient details are available in these reports to make dependent protein kinase or phosphatase. Second, direct comparisons to our results. thrombin could rapidly increase the specific activity of Protein kinases can be divided into three major groups the ATP available to the protein kinase(s) responsible (6): (a) cAMP-regulated, including free catalytic sub- for peak 7 and peak 9 protein phosphorylation. This unit and cAMP-stimulated holoenzyme; (b) cAMP- latter suggestion would be difficult to test since changes independent, which includes such enzymes as phos- in the specific activity of total platelet ATP may not phorylase b kinase (48), histone kinase HK2 (49), reflect changes in the specific activity of the ATP avail- phosvitin kinase (50), casein kinase (51), and chromatin able to the kinase. Lastly, we observed that there was protein kinase (52); and (c) the not-so-well char- inhibition of the two protein kinase enzymes by Ca"+ acterized cAMP-inhibited enzymes (7-10, 42-45). Four in the presence of Mg"+ and the complete lack of activity criteria have been established (6) for classifying protein when Mg++ was absent and Ca++ was present. Thus, the kinases. They were designed before reports of cAMP- secretion of Ca++ induced by thrombin during the release inhibited protein kinases had appeared and thus are not reaction (38) could remove the inhibition of the protein as applicable to them. They are useful, however, in kinase(s) responsible for peaks 7 and 9 protein phos- providing a common base for comparison of enzymes phorylation. Possibly the inhibition of phosphorylation isolated from various tissues within an organism and by agents that elevate cAMP is due to inhibition of from various organisms. These parameters are: (a) Ca++ release by cAMP. stimulation by cAMP, (b) binding of cAMP (c) inhibi- We have shown that in the presence of [y-"2P]ATP, tion of activity by the regulatory subunit, and (d) disrupted platelets as well as the supernatant fraction of inhibition of activity by the heat-stable inhibitory disrupted platelets phosphorylate endogenous platelet protein. proteins that co-electrophorese with peak 7 and 9 pro- We have shown that the supernatant fraction of dis- teins. In contrast to the intact platelet system, this rupted platelets can be separated into two distinct pro- phosphorylation is not affected by cAMP or thrombin. tein kinase enzymes. The cAMP-dependent enzyme that We cannot easily explain the difference observed in we have demonstrated is typical on the basis of the four the effect of thrombin and cAMP on protein phos- criteria above-it is stimulated by cAMP, binds ['H]- phorylation in intact platelets versus platelet homoge- cAMP, and is inhibited by the regulatory subunit and nates. However, our results do suggest that the normal by the heat-stable inhibitor protein. This enzyme has control mechanisms of platelet protein kinase and/or been previously described in platelet extracts (2, 46, 53). phosphoprotein phosphatase activity are disrupted during The cAMP-independent enzyme is not affected by the cell lysis. This suggestion is consistent with earlier ob- above-mentioned regulatory proteins. servations which indicated that intact platelets were The different protein acceptor substrate specificities of necessary to observe a variety of physiological effects the two separated protein kinase enzymes now explain of thrombin (5). the results obtained with whole platelet sonicates or the We have observed only minimal phosphorylation supernatant fraction after centrifugation. The protein of peak 7 and 9 proteins in the particulate fraction of kinase activity of these crude preparations when assayed sonicated platelets when the disrupted platelets were with histone f2b as substrate was predominantly that of incubated with [-y-YP]ATP before separation into the cAMP-dependent enzyme while the protein kinase soluble and particulate fractions and none when the activity when assayed with protamine as substrate was 934 R. M. Lyons, N. Stanford, and P. W. Majerus primarily that of the cAMP-independent enzyme. Thus, 3. Tollefsen, D. M., J. R. Feagler, and P. W. Majerus. the protein kinase activity of the crude fractions when 1974. The binding of thrombin to the surface of human platelets. J. Biol. Chem. 249: 2646-2651. assayed with histone f2b as substrate was stimulated by 4. Ganguly, P. 1974. Binding of thrombin to human plate- cAMP while the activity when assayed with protamine as lets. Nature (Lond.). 247: 306-307. substrate was only slightly affected by cAMP. 5. Brodie, G. N., N. L. Baenziger, L. R. Chase, and P. W. The substrates of the two enzymes in intact platelets Majerus. 1972. The effects of thrombin on adenyl cy- have not been determined since the enzymes have not clase activity and a membrane protein from human platelets. J. Clin. Invest. 51: 81-88. been isolated from all potential substrates. The cAMP- 6. Walsh, D. A., and E. G. Krebs. 1973. Protein kinases. dependent enzyme fraction apparently contained one of In, The Enzymes. P. D. Boyer, editor. Academic Press, the substrates of most interest, i.e., peak 7 protein. Inc., New York. 3rd edition. 8: 555-581. Phosphorylation of this protein was stimulated by cAMP 7. DeLorenzo, R. J., and P. Greengard. 1973. Activation by adenosine 3': 5'-monophosphate of a membrane-bound but not by thrombin. We have thus shown that this phosphoprotein phosphatase from toad bladder. Proc. enzyme fraction is capable of phosphorylating a protein Natl. Acad. Sci. U. S. A. 70: 1831-1835. that co-electrophoreses with peak 7 protein, but whether 8. Liu, A. Y-C., and P. Greengard. 1974. Aldosterone- it actually is responsible for the thrombin-induced in- induced increase in protein phosphatase of toad bladder. crease in phosphorylation of peak 7 protein observed in Proc. Natl. Acad. Sci. U. S. A. 71: 3869-3873. 9. DeLorenzo, R. J., K. G. Walton, P. F. Curran, and P. intact platelets is not known. Greengard. 1973. Regulation of phosphorylation of a The fact that phosphorylation of peak 7 protein is specific protein in toad-bladder membrane by antidiu- stimulated by cAMP with the fractionated enzyme while retic hormone and cyclic AMP, and its possible rela- cAMP has no effect on peak 7 protein phosphorylation tionship to membrane permeability changes. Proc. Natl. Acad. Sci. U. S. A. 70: 880-884. in crude sonicates as well as the fact that agents that 10. Kish, V. M., and L. J. Kleinsmith. 1974. Nuclear pro- elevate the intracellular cAMP levels clearly inhibit tein kinases. J. Biol. Chem. 249: 750-760. thrombin-induced phosphorylation with intact 'PO4- 11. Reynolds, J. A., and M. J. Schlesinger. 1967. Conforma- loaded platelets indicates that cAMP regulation of a tional states of the subunit of Escherichea coli alkaline protein kinase is not the sole control mechanism for phosphatase. Biochemistry. 6: 3552-3559. 12. Glover, C., and E. Shaw. 1971. The purification of protein phosphorylation in intact platelets. thrombin and isolation of a peptide containing the active The identities of peak 7 and peak 9 protein are not center histidine. J. Biol. Chem. 246: 4594-4601. known. However, peak 7 protein migrates with actin in 13. Kostka, V., and F. H. Carpenter. 1964. Inhibition of the darkest Coomassie Blue staining region on SDS chymotrypsin activity in crystalline trypsin preparations. J. Biol. Chem. 239: 1799-1803. polyacrylamide gel electrophoresis of platelets (54), 14. Majerus, P. W., and G. N. Brodie. 1972. The binding while a light chain of platelet myosin has been demon- of phytohemagglutinins to human platelet membranes. strated to migrate at a molecular weight similar to peak J. Biol. Chem. 24: 4253-4257. 9 protein (55, 56). Preliminary experiments have sug- 15. Baenziger, N. L., and P. W. Majerus. 1974. Isolation 7 of human platelets and platelet surface membrane. gested that peak protein is not actin but that peak 9 Methods Enzymol. 31: 149-155. protein may be the light chain of platelet myosin. 16. Weber, K., and M. Osborn. 1969. The reliability of Purification of peak 7 and 9 proteins will help to molecular weight determinations by dodecyl sulfate- clarify their identity and location, their interaction with polyacrylamide gel electrophoresis. J. Biol. Chem. 244: the platelet protein kinases, and their role in the platelet 4406-4412. 17. Bray, G. A. 1960. A simple, efficient liquid scintillator release reaction. for counting aqueous solutions in a liquid scintillation counter. Anal. Biochem. 1: 279-285. ACKNOWLEDGMENTS 18. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. We thank the St. Louis Chapter of the American Red Naftnre (Lond.). 227: 680-685. Cross and the Barnes Hospital Blood Bank for making 19. Reid, M. S., and R. L. Bielski. 1968. A simple appara- available the fresh platelet-rich plasma and 72-h-old platelet tus for vertical flat-sheet polyacrylamide gel electro- concentrates. We thank Dr. Lewis Chase for his helpful phoresis. Anal. Biochem. 22: 374-381. discussions about the phosphodiesterase assay. 20. Baenziger, N. L., G. N. Brodie, and P. W. Majerus. This work was supported by grants from the National 1972. Isolation and properties of a thrombin-sensitive Institutes of Health, HE 14147 and HL 16634. protein of human platelets. J. Biol. Chem. 247: 2723- 2731. 21. Corbin, J. D., C. 0. Brostrom, C. A. King, and E. G. REFERENCES Krebs. 1972. Studies on the adenosine 3',5'-monophos- 1. Lyons, R. M., and P. W. Majerus. 1974. The effect of phate-dependent protein kinases of rabbit skeletal muscle. thrombin on protein phosphorylation in human platelets. J. Biol. Chem. 247: 7790-7798. J. Clin. Invest. 53: 48a. (Abstr.) 22. Reimann, E. M., D. A. Walsh, and E. G. Krebs. 1971. 2. Salzman, E. W., and W. Weisenberger. 1972. Role of Purification and properties of rabbit skeletal muscle cyclic AMP in platelet function. Adv. Cyclic Nucleo- adenosine 3'5'-monophosphate-dependent protein kinases. tide Res. 1: 231-247. J. Biol. Chem. 246: 1986-1995. Platelet Protein Phosphorylation 935 23. Reimann, E. M., C. 0. Brostrom, J. D. Corbin, C. A. 40. Wolfe, S. M., and N. R. Shulman. 1970. Inhibition of King, and E. G. Krebs. 1971. Separation of regulatory platelet energy production and the release reaction by and catalytic subunits of the cyclic 3',5'-monophosphate- PGE1, theophylline and cAMP. Biochem. Biophys. Res. dependent protein kinase (s) of rabbit skeletal muscle. Commun. 41: 128-134. Biochem. Biophys. Res. Commun. 42: 187-194. 41. Corbin, J. E., T. R. Soderline, and C. R. Park. 1973. 24. Walsh, D. A., C. D. Ashby, C. Gonzalez, D. Calkins, Regulation of adenosine 3',5'-monophosphate-dependent E. H. Fischer, and E. G. Krebs. 1971. Purification and protein kinase. I. Preliminary characterization of the characterization of a protein inhibitor of adenosine adipose tissue enzyme in crude extracts. J. Biol. Chem. 3',5'-monophosphate-dependent protein kinases. J. Biol. 248: 1813-1821. Chem. 246: 1977-1985. 42. Miyamoto, E., J. F. Kuo, and P. Greengard. 1969. 25. Hunter, W. M. 1967. The preparations of radioiodinated Cyclic nucleotide-dependent protein kinases. III. Puri- proteins of high activity. Their reaction with antibody fication and properties of adenosine 3',5'-monophosphate- in vitro: the radioimmunoassay. In Handbook of Ex- dependent protein kinase from bovine brain. J. Biol. perimental Immunology. D. M. Weir, editor. Blackwell Chem. 244: 6395-6402. Scientific Publications, Ltd., London. 608-654. 43. Kuo, J. F., B. K. Krueger, J. R. Sanes, and P. Green- 26. Ackers, G. K. 1964. Molecular exclusion and restricted gard. 1970. Cyclic nucleotide-dependent protein kinases. diffusion processes in molecular-sieve chromatography. V. Preparation and properties of adenosine 3',5'mono- Biochemistry. 3: 723-730. phosphate-dependent protein kinase from various bovine 27. Sanborn, B. M., R. C. Bhalla, and S. G. Korenman. tissues. Biochiin. Biophys. Acta. 212: 79-91. 1973. The endometrial adenosine cyclic 3',5'-monophos- 44. Johnson, E. M., H. Maeno, and P. Greengard. 1971. phate-dependent protein kinase. J. Biol. Chem. 248: Phosphorylation of endogenous protein of rat brain by 3593-3600. the cyclic adenosine 3',5'-monophosphate-dependent pro- 28. Walton, G. M., and L. D. Garren. 1970. An assay for tein kinase. J. Biol. Chem. 246: 7731-7739. adenosine 3',5'-cyclic monophosphate based on the asso- 45. Kuehn, G. D. 1972. Cell cycle variation in cyclic aden- ciation of the nucleotide with a partially purified bind- osine 3',5'-monophosphate-dependent inhibition of a pro- ing protein. Biochemistry. 9: 4223-4229. tein kinase from Physarum Polycephalum. Biochem. 29. Beavo, J. A., J. G. Hardman, and E. W. Sutherland. Biophys. Res. Cominun. 49: 414-419. 1970. Hydrolysis of cyclic guanosine and adenosine 3',5'- 46. Booyse, F. M., D. Guiliani, J. J. Marr, and M. E. monophosphates by rat and bovine tissues. J. Biol. Chemt. Rafelson, Jr. 1973. Cyclic adenosine 3',5'-monophosphate 245: 5649-5655. dependent protein kinase of human platelets: membrane 30. Linhardt, K., and K. Walter. 1965. Phosphatases. In phosphorylation and regulation of platelet function. Ser. Methods of Enzymatic Analysis. H. U. Bergmeyer, Haematol. 6: 351-366. editor. Academic Press, Inc., New York. 779-787. 47. Steiner, M. 1974. Effect of thrombin on phosphorylation 31. Seegers, W. H., and H. P. Smith. 1942. Factors which of platelet membrane proteins. Blood. 44: 916. (Abstr.) influence the activity of purified thrombin. Am. J. Phys- 48. Krebs, E. G., and E. H. Fischer. 1956. The phosphory- iol. 137: 348-354. lase b to a converting enzyme of rabbit skeletal muscle. 32. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. Biochim. Biophys. Acta. 20: 150-157. J. Randall. 1951. Protein measurement with the folin 49. Langan, T. A. 1971. Cyclic AMP and histone phos- phenol reagent. J. Biol. Chem. 193: 264-275. phorylation. Ann. N. Y. Acad. Sci. 185: 166-180. 33. Tollefsen, D. M., J. R. Feagler, and P. W. Majerus. 50. Rodnight, R., and B. E. Lavin. 1964. Phosvitin kinase 1974. Induction of the platelet release reaction by phy- from brain: activation by ions and subcellular distribu- tohemagglutinin. J. Clin. Invest. 53: 211-218. tion. Biochemt. J. 93: 84-91. 34. Cohen, P., M. J. Broekman, A. Verkley, J. W. W. 51. Bingham, E. W., H. M. Farrell, Jr., and J. J. Basch. Lisman, and A. Derksen. 1971. Quantification of human 1972. Phosphorylation of casein. Role of the Golgi ap- platelet inositides and the influence of ionic environment paratus. J. Biol. Chem. 247: 8193-8194. on their incorporation of orthophosphate-"P. J. Clin. 52. Takeda, M., H. Yamamura, and Y. Ohga. 1971. Phos- Invzest. 50: 762-772. phoprotein kinases associated with rat liver chromatin. 35. Taborsky, G. 1974. Phosphoproteins. Adv. Protein Res. Commun. 42: 103-110. Chem. 28: 1-210. Biochemn. Biophvs. 36. Schmidt, G. 1957. Chemical and enzymatic methods for 53. Marquis, N. R., R. L. Vigdahl, and P. A. Tavormina. the identification and structural elucidation of nucleic 1971. Cyclic AMP-dependent platelet protein kinase. acids and nucleotides. Methods EnZymol. 3: 747-775. Fed. Proc. 30: 423. (Abstr.) 37. Allerton, S. E., and G. E. Perlman. 1965. Chemical 54. Pollard, T. D., S. M. Thomas, and R. Niederman. 1974. characterization of the phosphoprotein phosvitin. J. Biol. Human platelet myosin. I. Purification by a rapid Chem. 240: 3892-3898. method applicable to other nonmuscle cells. Anal Bio- 38. Detwiler, T. C., and R. D. Feinman. 1973. Kinetics of chem. 60: 258-266. the release of Ca (II) by platelets. Biochemistry. 12: 55. Adelstein, R. S., M. A. Conti, and W. Anderson, Jr. 282-289. 1973. Phosphorylation of human platelet myosin. Proc. 39. Detwiler, T. C., and R. D. Feinman. 1973. Kinetics of Natl. Acad. U. S. A. 70: 3115-3119. the thrombin-induced release of adenosine triphosphate 56. Conti, M. A., and R. S. Adelstein. 1975. Platelet myo- by platelets. Comparison with release of calcium. Bio- sin phosphorylation controls actin activation of myosin chemistry. 12: 2462-2468. ATPase activity. Fed. Proc. 34: 670. (Abstr.)

936 R. M. Lyons, N. Stanford, and P. W. Majerus