Phosphorylation of a 22,000-Dalton Component of the Cardiac Sarcoplasmic Reticulum by Adenosine 3’ : S-Monophosphate-dependent Protein Kinase*

SUMMARY Cardiac microsomes were incubated with [y-azP]ATP and a cardiac adenosine 3’:5’-monophosphate (cyclic AMP)-dependent protein kinase in the presence of ethylene glycol bis(&aminoethyl ether)-A’, IV’-tetraacetic acid. After solubilization in sodium dodecyl sulfate and fractionation by polyacrylamide gel electrophoresis, a single microsomal protein component of approximately 22,000 daltons was found to bind most of the a2P label. The a*P labeling of this component increased several fold when NaF was included in the incubation medium. No other component of cardiac microsomes, including sarcoplasmic reticulum ATPase protein, contained significant amounts of azP label. This 22,000-dalton phosphoprotein formed by cyclic AMP-dependent protein lrinase had stability characteristics of a phosphoester rather than an acyl phosphate. Washing of microsomes with buffered KC1 did not decrease the amount of azP labeling to the 22,000-dalton protein, suggesting that this protein is associated with the membranes of sarcoplasmic reticulum rather than being a contaminant from other soluble proteins. The

Chem. 249, 61666173; TADA, M., KIRCHBERGER, M. A., REPKE, D. I., AND KATZ, A. M. (1974) J. Biol. Chem. 249, 6174-6180) that cyclic AMP-dependent protein hinase-catalyzed phosphorylation is associated with stimulation of calcium transport in the cardiac sarcoplasmic reticulum, and further indicate that this phosphorylation occurs at a component of low mass (22,000 daltons) of the cardiac sarcoplasmic reticulum which, while separable from the calcium transport ATPase protein (100,000 daltons) by sodium dodecyl sulfatepolyacrylamide gel electrophoresis, has the ability to regulate calcium transport by the cardiac sarcoplasmic reticulum.
In previous communications we have reported that a cardiac cyclic adenosine 3' : 5'-monophosphate-dependent protein kinase enhances both calcium uptake and Ca2+-activated ATPase activity of a dog cardiac microsomal preparation (4,5) which consists mainly of fragmented sarcoplasmic reticulum (6). These functional alterations in calcium transport are correlated with phosphorylation of microsomes by cyclic AMP'-dependent protein kinases (7). The microsomal phosphoprotein has stability characteristics of a phosphoester in which the phosphate is incorporated largely into serine (7). Its formation does not require Ca2+ (7). Cardiac microsomes can form another type of phosphoprotein, an intermediate of calcium transport ATPase (8,9) with molecular weight of approximately 100,000 (10). Formation of the latter requires Ca2+ (S-10) and the phosphate appears to be present as acyl phosphate (9, lo), like its skeletal muscle counterpart (11)(12)(13). To elucidate the mechanism by which cyclic AMP-dependent protein kinase modulates calcium transport by the cardiac sarcoplasmic reticulum, we have investigated the relationship between the calcium transport ATPase protein and the phosphoprotein whose formation is catalyzed by cyclic AMP-dependent protein kinase. The present communication reports the protein kinase-catalyzed formation of a phosphoprotein of approxi-mately 22,000 daltons, which can be separated from ATPase protein by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. This 22,000-dalton protein appears to have the ability to modulate calcium transport by the cardiac sarcoplasmic reticulum.

Materials
Cardiac microsomes were prepared from dog heart ventricle according to the procedures of Harigaya and Schwartz (6) with minor modifications (7). Cyclic AMP-dependent protein kinase was purified through the DEAE-cellulose chromatography step from bovine hearts, according to the method of Miyamoto et al. (14). [ Reactions were started by the addition of trypsin. Aliquots taken at various time intervals were added to tubes containing trypsin inhibitor that gave a trypsin to trypsin inhibitor ratio of 1:2 by weight. For zero time, a mixture of trypsin and trypsin inhibitor was added. Phosphoprotein formation in trypsin-treated microsomes was determined by incubation in Reaction Mixture A or B and elcctrophoresis on sodium dodecyl sulfate-polyacrylamide gels as described above. Calcium uptake and Ca2+-act ivttted ATPase activity of trypsin-treated microsomes were measured by the procedures described previously (5).
In the experiments in which phosphorylated microsomes were treated with trypsin, microsomcs were phosphorylated in Reaction Mixture A or C for 10 min at 25", after which trypsin was added at a microsomal protein to trypsin ratio of 2O:l. In some experiments, phosphorylated microsomes were washed with buffer solution before treatment with trypsin. At time intervals after the addition of trypsin, trypsin inhibitor was added (trypsin to trypsin inhibitor = 1:2). The amount of phosphoprotcin formed in these microsomes (Reaction Mixture A) was determined by gel electrophoresis as described above, and calcium uptake and Cat+activated ATPase activity (Reaction Mixture C) were measured as described previously (5).

Determination of RNA
Cardiac microsomes were digested in KOH and RNA content was determined spectrophotometrically by the method of Fleck and Munro (20).
X major protein band, a, five secondary bands, b to f, and a diffuse band, z, were found when small amounts (42 pg) of microsomal protein were applied to tbc gel (A, 1). Washing of the microsomes with 0.6 M KC1 caused no obvious change in this pattern, and no low molecular weight components were found in the supernatant, after IX1 washing. When larger amounts of cardiac microsomal protein were applied to the gels, less distinct tcrt.iary bands, including Bands W, s, and y, became visible (Fig. 2B). When the periodic acid-Schiff method was applied, only Band z was stained. liovine cardiac protein kinase also produced several protein bands when examined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (B, Fig. 1). Electrophoresis of a mixture of cardiac microsomal protein and A; illY0, myoglobin; CYT, cytochrome c. The proteins were incubated for 2 hours at 37" in 1% sodium dodecyl sulfate, 170 p-mercaptoethanol, and 10 mM sodium phosphate buffer (pH 7.2), and aliquots (0.02 ml) of the solutions containing 3 gg of each protein marker were applied to the gels. Electrophoresis was performed as described in the text. Each point represents an average of four determinations.
B and C, microsomes (1.6 mg per ml) were phosphorylated in Reaction Mixture A (see "Methods") with 0.8 mg per ml of bovine cardiac protein kinase and 1 PM cyclic AMP in the presence (0) and absence (0) of 25 mM TaF. Iteactions were terminated by Procedure I, and electrophoresis was carried out, as described in the text, with application of 0.112 mg of microsomal protein + 0.056 mg of protein kinase for determining protein distributions (B), and 0.05G mg of microsomal protein + 0.028 mg of protein kinase for determining the distributions of radioactivity (C). In U, only the electrophoretogram of microsomal protein incubated in the presence of 25 mM NaF is shown since omission of NaF resulted in similar protein distributions.
protein kinase gave the pattern expected from the combination of both proteins (Fig. 2B).
Phosphorylation of B,OOO-dalton Component by Protein Kinase -Cardiac microsomes phosphorylated by cyclic AMP and protein kinase showed a single significant peak of radioactivity (Peak ZZ, Fig. 2C) when azP labeling was determined by measurement of the radioactivity of sliced gels. When protein kinase and cyclic AMP were omitted from the incubation medium, no clear peaks were seen at this area or any other area of the gel. A peak of much higher radioactivity was found at this same location (Peak ZZ, Fig. 2C) when 25 mM NaF was included in the reaction medium in order to inhibit phosphoprotein phosphatase activity present in microsomal preparations.* The increase of azP labeling due to NaF was approximately 5-fold when the amounts of azP were estimated from the area of Peak II. In addition to the major peak, two peaks of much less radioactivity were noted in the presence of NaF (Peaks Z and ZZZ, Fig. 2C). Peak I may represent a2P labeling of one of the subunits of protein kinase since a2P labeling of the same extent at the same location as Peak I was found when protein kinase was incubated in the absence of cardiac microsomes under identical conditions. This peak of radioactivity corresponded to the protein Band 3 of approximately 55,000 daltons seen in Fig. 1B. The major peak of radioactivity (Peak ZZ, Fig. 2C) corresponded to one of the minor components (w, Fig. 2B) of microsomes. Among 12 determinations, using 5 different microsomal preparations, Peak II was seen at RF 0.62 f 0.02 (S.D.), whereas the minor protein band w was found at RF 0.63 f 0.01 (S.D.). Based on the calibration curve ( Fig. 2A), the apparent molecular weight of this phosphoprotein component was estimated from the latter RF value to be 22,000 f 1,000. An additional minor peak of radioactivity (Peak ZZZ, Fig. 2C) was seen when microsomes were phosphorylated in the presence of NaF. The distribution of radioactivity was similar when the specific activity of the [y-a2P]ATP was increased IOO-fold. When the total ["'PIphosphate found in Peak II in the presence of NaF (0.035 nmol) is compared with the total microsomal protein applied (0.056 mg) in Fig. 2C, approximately 0.63 mnol of phosphate were found to be incorporated per mg of microsomal protein. This value varied among a number of microsomal preparations within the range of 0.5 to 1.0 nmol of phosphate per mg of microsomal protein, which is in good agreement with the value obtained under similar conditions by the trichloroacetic acid precipitation procedure (7).
Stability of Phosphoprotcin-In order to study the nature of phosphate-binding to the 22,000-dalton component, microsomes phosphorylated in the presence of protein kinase and [y-"P]ATP were treated under various conditions, after which they were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and the amounts of a2P bound to this component were estimated from the area of Peak II. The amount of radioactivity found in Peak II when the phosphorylation was terminated by sodium dodecyl sulfate (Procedure I, see "Methods") was virtually the same as that when the phosphorylation was terminated by trichloroacetic acid and the mixture allowed to stand at 0" for 10 min (Procedure II). Incubation in 0.5 N NaOH for 15 min at 0" 37", and 90" demonstrated relative alkali stability at the lower temperatures (Table I) These results were presented at the Second International Conference on Cyclic AMP (1974) in Vancouver, British Columbia, Canada. Control (lOTo trichloroacetic acid, 0") 100 Chloroform-methanol, 0' 99 Acetone, 0' 88 at 30", and also stable in chloroform-methanol and acetone (Table I). Treatment with 10% trichloroacetic acid for 15 min at 90" caused the peak of radioactivity to become more broad although loss of recovery of total radioactivity was only approximately 10 $!&.
Ribosomal Contamination-To evaluate the possibility that the 22,000-dalton component was derived from a ribosomal contaminant, three preparations of cardiac microsomes were digested with KOH and analyzed spectrophotometrically for RNA. The average content of ribosomal protein, estimated from RNA content with the assumption that cardiac ribosomes contain approximately equal weights of protein and RNA, was less than 0.4% of the total microsomal protein.
Washing of Microsomes with KC'-Microsomes were washed repeatedly with 0.6 M KC1 to remove soluble contaminants and myofibrillar proteins. The amount of a2P labeling in Peak II did not change significantly after two washings with 0.6 M KCl. The specific activity of the phosphoprotein increased slightly (approximately 10%) after the first KC1 wash and remained unchanged when microsomes were washed further.
Treatment of Microsomes with Trypsin--Incubation of microsomes with trypsin for 20 min at 25" and pH 6.8 at a microsomes to trypsin ratio of 20 : 1 (w/w) prevented subsequent phosphorylation of the 22,009dalton component by protein kinase (Fig. 3). Under these conditions trypsin uncoupled the calcium transport system in that calcium uptake was markedly decreased while Ca2+-activated ATPase activity was slightly increased (Fig. 4). When trypsin digestion was carried out in the presence of 1.0 M sucrose, however, calcium uptake was only slightly affected (Fig.  4) as shown previously in skeletal microsomes by Ikemoto et al. (21). In the presence of 1.0 M sucrose, where trypsin causes no significant change in calcium transport activity, brief trypsin treatment caused a marked decrease in formation of the phosphoprotein by protein kinase, and less than 5y0 of control phosphorylation was found after 20 min of digestion (Fig. 5) Aliquots (50 ~1 containing 0.21 mg of microsomes and 10.5 pg of trypsin) were taken at the indicated times and added to tubes containing trypsin inhibitor (21 pg in 50 ~1). Each sample (0.1 ml), which had been kept on ice, was diluted with 0.4 ml of an ice-cold solution containing 50 mM KC1 and 20 rnr.4 Tris-HCl (pH 6.8) to give microsomal protein concentration of 0.42 mg per ml. Aliquots of this mixture were subjected to assay for calcium uptake (0, 0) and Ca2+-activated ATPase activity (A, A) under conditions described previously (5) with 1 pM Ca2+ (Ca-EGTA buffer containing 125 PM CaClr) and 37 pg per ml of microsomal protein.
of microsomol calcium transport and protein kinase-catalyzed phosphorylation of the 22,000-dalton protein was investigated by determining whether protein kinase and cyclic AMP could stimulate calcium uptake by microsomes which, after 20 min of trypsin treatment in 1 M sucrose, have retained calcium uptake activity but lost most of the ability to be phosphorylated by protein kinase (Fig. 6). In contrast to control microsomes, Then 0.2 ml-aliquots taken at the indicated times were added to solutions (0.02 ml) containing trypsin inhibitor. Assays for calcium uptake and phosphoprotein formation in these microsomes were subsequently carried out. For measurement of calcium uptake, 50 pl-aliquots were diluted with buffer solutions and added to assay media as described in Fig. 4. The final concentration of microsomal protein in assay media was 50 pg per ml. For determination of phosphoprotein formation, a 100 rl-aliquot was incubated in Reaction Mixture A with 0.5 mg per ml of protein kinase, 1 PM cyclic AMP, and 25 mM NaF in a total volume of 0.6 ml; the final concentration of microsomal protein was 0.56 mg per ml. The phosphorylation was terminated by Procedure II, and electrophoresis was carried out as described under "Methods." The amount of phosphoprotein formed was determined by measuring the area of Peak II. where the rate of calcium uptake was greatly stimulated by treatment with protein kinase and cyclic AMP, calcium uptake by microsomes that had been treated with trypsin was not stimulated after incubation with protein kinase and cyclic AMP.
Trypsin with trypsin (12 rg) for the indicated 'times under conditions described under "Methods". the digestion was terminated with 24 rg of trypsin inhibitor. Fdr zero time, a mixture of trypsin and trypsin inhibitor was added. After solubilization of microsomal proteins (Procedure II), sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed as described under "Methods." The amount of phosphate bound was determined by measuring the area of Peak II.

conditions
[Reaction Mixture A), after which they were incubated with trypsin for different time intervals. The amount of radioactivity found in Peak II of phosphorylated microsomes was virtually unaffected by subsequent treatment with trypsin (Fig. 7). In order to examine whether the 22,000-dalton component remained associated with microsomal membranes after trypsin treatment, or whether, instead, it was detached from the membranes, microsomes were centrifuged at 105,000 x g for 20 min after phosphorylation and trypsinization, and the supernatant and pellet were subjected to sodium dodecyl sulfatepolyacrylamide gel electrophoresis. While the supernatant was found not to contain the "P label, most of the azP was recovered in the 22,000-dalton component of the pellet. To examine further the functional role of the 22,000-dalton component, we determined the effect of trypsinization on calcium uptake of the phosphorylated microsomes. Microsomes were phosphorylated in Reaction Mixture C and incubated with and without trypsin in the presence of 1.0 M sucrose. The rate of calcium uptake by the phosphorylated microsomes, which was about 2-fold greater than nonphosphorylated microsomes, was only slightly decreased by subsequent treatment with trypsin (Table II). DISCUSSION Cyclic AMP and cardiac cyclic AMP-dependent protein kinase induce marked stimulation of calcium uptake and Ca*+activated ATPase activity of cardiac microsomes (5, 7). Concurrently, a phosphoester phosphoprotein is formed in these membranes which is chemically different from the phosphoprotein intermediate of the calcium transport ATPase (7,22,23). These findings suggest that the protein kinase catalyzes the formation of a phosphoprotein which stimulates the calcium pump of the cardiac sarcoplasmic reticulum. The present study demonstrates that the protein kinase catalyzes phosphorylation under standard conditions (Reaction Mixture C) with 1.0 mg per ml of protein kinase and 1 PM cyclic AMP in a total volume of 1.8 ml. At 9 min after start of the phosphorylation, 1.2 ml of 2.5 M sucrose (pH 6.8) were added to give final sucrose concentration of 1.0 M. Trypsin (90 pg in 20 ~1) was added 1 min after the addition of sucrose. At indicated times after the addition of trypsin, 0.5-ml aliquots were taken and added to 50 ~1 of solution containing 30 pg of trypsin inhibitor.
Of each of the resulting mixtures, 0.4 ml was incubated in 4 ml of calcium uptake assay media containing 1 PM Cat-'; the final concentration of microsomal protein was 54 pg per ml. The rate of calcium uptake of control microsomes, that were incubated in the absence of protein kinase and cyclic AMP, was 0.059 pm01 of calcium taken up per min per mg of protein under conditions described above. of a 22,000-dalton protein3 which is electrophoretically distinct from the calcium transport ATPase protein of approximately 100,000 daltons and provides support for the view that this 22,000-dalton phosphoprotein has a regulatory role in calcium transport by cardiac sarcoplasmic reticulum.
Cardiac microsomes that have been solubilized in sodium dodecyl sulfate can be fractionated into several components by polyacrylamide gel electrophoreeis (Fig. 1A). The major component of 90,000 to 100,000 daltons (Band a) contains the ATPase protein which can undergo phosphorylation to form acyl phosphoprotein, an intermediate of ATPase, in that its mobility is similar to that of the previously described ATPase of cardiac (10) and skeletal (17,24) sarcoplasmic reticulum. Band z may represent proteolipid (24). Some of the other bands of cardiac microsomes shown in Fig. 1 may be analogous to the calcium-binding proteins of skeletal sarcoplasmic reticulum (25,26). However, none of these previously described components served as a substrate for protein kinase-catalyzed phosphorylation (Fig. 2).
The phosphoprotein formed in the presence of protein kinase differs from the phosphoprotein intermediate of the calcium transport ATPase in several aspects. Unlike the ATPase intermediate, this phosphoprotein has the stability characteristics of a phosphoester (Table I) and its formation does not require Cazf (7). The possibility that it is primardy phospholipid can be excluded because of its trypsin sensitivity and because neither chloroform-methanol nor acetone extracted the phosphate (Table I). The 22,000.dalton component was not stained by the periodic acid-Schiff method, indicating that it is not a glycoprotein.
Two peaks of low radioactivity were seen in addition to the main peak of radioactivity that is associated with the 22,000dalton phosphoprotein (Fig. 2C). Peak 1, which was seen when protein kinase was incubated in the absence of microsomes, *We have tentatively named this phosphoprotein "phospholamban" (l-3).