ARCHIVES
OF BIOCHEMISTRY
AND
BIOPHYSICS
289, No. 1, August 15, pp. 103-108, 1991
Vol.
Hormone-Induced Phosphorylation of the Plasma Membrane Calcium Pump in Cultured Aortic Endothelial Cells Kevin
K. W. Wang,’
Department
of
Pathology,
Yeng Sheng Du, Clement
Diglio,
Wayn,e State University
of
School
Wayne
Tsang,
and Tuan
Medicine, Detroit, Michigan
H. Kuo2
48201
Received March 7, 1991, and in revised form April 16, 1991
The regulation of the plas:ma membrane Ca2+ pump by hormones via phosphorylation in intact cells has not been clearly established. We now present evidence that the Ca” pump is phosphorylated on both serine and threonine residues in unstimulated and stimulated cultured rat aortic endothelial cells. Among the stimuli tested, the protein kinase C activator phorbol 1%myristate 13-acetate (PMA) was most potent and increased the level of phosphorylation threefold, while the CAMP-dependent protein kinase activator 6(4+hlorophenylthio)-CAMP (CPT-CAMP) stimulated the phosphorylation 1.6-fold. Two-dimensional tryptic phosphopeptide maps of the Ca2+ pump from unstimulated and CPT-CAMP-stimulated cells have identical patterns (five phosphopeptides) while PMA-stimulated cells have three additional phosphopeptides. Isoproterenol-, ATP-, angiotensin II-, and bradykinin-stimulated cells also have increased levels of Ca2+ pump phosphorylation.. Stimuli-induced phosphorylation of the Ca”+ pump was rapid (5-10 min) and was concomitant with stimulated calcium efflux from the same cells. This is the first direct evidence that the plasma membrane Ca2+ pump in intact cells is regulated by various hormones or agonists via CAMP-dependent protein o 1991 ~eademi~ kinase or protein kinase C phosphorylation. Press,
Inc.
The plasma membrane Ca2+ pump is important in maintaining submicromolar intracellular Ca2+ concentration in mammalian cells (1). This enzyme is thought to be stimulated by endogenous calmodulin directly (2). The enzyme can potentially be irreversibly activated by endogenous calcium-dependent protease (calpain) (3). i Research fellow, American Heart Association/Michigan Current address: Parke-Davis Pharmaceutical Research Warner-Lambert Company, Ann Arbor, MI 48106. ’ To whom correspondence should be addressed. 0003-9861/91
53.00 1991 by of reproduction
Copyright 0 All
rights
Academic in any
Press, form
Inc. reserved.
(1990). Division,
Furthermore, under in vitro conditions, the membranebound or the purified Ca” pump from human erythrocytes (4-S) or cardiac plasma membrane (5,9) was found to be activated following phosphorylation by CAMP-dependent protein kinase (protein kinase A)3 or protein kinase C. It is therefore conceivable that the Ca2+ pump is regulated by phosphorylation in living cells. In an earlier report, we showed that the Ca2+ pump is indeed phosphorylated in living cells (10). In this report, we present evidence that phosphorylation of the Ca2+ pump is hormone- or agonistinduced. MATERIALS
AND
METHODS
Materials. Phorbol esters and ionomycin were from Sigma. CPTCAMP was from Boehringer Mannheim. [32P]orthophosphate (3000 Ci/ mmol) was from Amersham. A monoclonal anti-Ca’+ pump antibody (5FlO) was a gift from Dr. J. T. Penn&on (Mayo Clinic). Molecular weight standards were from Bio-Rad. All reagents used were of analytical grade.
32P-lalwlingof endothelial cells and immurwprecipitation. Rat thoracic aorta endothelial cells were initiated and cloned as described previously by Diglio et al. (11). The cell cultures were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, and 25 pg/ml gentamicin. Confluent cultures (5 X 10s cells/lOO-mm dish) were washed three times in phosphate-free MEM (Sigma) and incubated with the same medium containing 100 @i/ml [32P]orthophosphate (Amersham) for 4 h at 37°C under 5% CO,/95% air. Various agents were added for 10 min before the end of the [32P]phosphate incubation. Incubation was stopped by removing the medium and rapid washing with 10 mM sodium phosphate, 150 mM sodium chloride (pH 7.4), and the addition of 1.3 ml lysis buffer (1%
3 Abbreviations used: protein kinase A, CAMP-dependent protein kinase; protein kinase C, Ca’+-, phospholipid-dependent protein kinase; PMA, phorbol 12-myristate 13-acetate; CPT-CAMP, 8-(4-chlorophenylthio)-CAMP; 4(~-PDD, 4a-phorbol didecanoate; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; EGF, epidermal growth factor; EGTA, ethylene glycol bis(@aminoethyl ether)N,,hr’-tetraacetic acid; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid, Tris, tris(hydroxymethyl)aminomethane; PMSF, phenylmethylsulfonyl fluoride; AII, angiotensin II; BSA, bovine serum albumin.
104
WANG
Nonidet-P40,lO mM Tris-HCl, pH 7.4,150 mM NaCl, 1% bovine serum albumin (BSA), 1 mM phenylmethylsulfonyl fluoride, 5 mM EDTA, 5 mM EGTA, 50 mM NaF, 100 pM sodium vanadate, 5 rg/ml leupeptin, and 2 mM dithiothreitol). Cells were lysed for 20 min at 4°C. After cell debris was removed by centrifugation, lysates were cleared by incubating with 5 pg/ml rabbit anti-mouse immunoglobulin, 1% protein A-Sepharose, 3 ~1 normal ascetic fluid for 40 min at 4°C. The cleared lysate was then incubated at 4°C with 3 ~1 monoclonal antibody (5FlO) against the plasma membrane Gas+ pump. After 20 min, 5 yg rabbit anti-mouse immunoglobulin was added. Immunocomplexes were then adsorbed to 1% protein A-Sepharose at 4°C for 40 min with rocking. The immunoprecipitate-protein A-Sepharose complex was then centrifuged and washed five or six times with lysis buffer (as above, except without BSA). The complexes were then dissociated by treatment of 20% trichloroacetic acid at 4°C. The centrifuged protein precipitates were then neutralized with Tris base and solubilized in sample buffer (2% SDS, 10% glycerol, 100 mM Tris-HCl, pH 6.8, 0.002% bromophenyl blue, 0.1 mM p-mercaptoethanol) and electrophoresed in SDS-gradient (5-20%) polyacrylamide gel. The 32P-labeled Ca*+ pump band was identified by Western blotting (12, 13) using the anti-Ca*+ pump antibody and the alkaline phosphatase-conjugated goat anti-mouse IgG. Phosphoamino acid analysis. The method of Cooper et al. (14) for phosphoamino acid analysis was followed. Gel strips containing the 32Plabeled Ca*+ pump bands were first digested with 50 pg trypsin at 37°C overnight and then treated with 7 M HCl at 105°C for 60 min. The supernatant was dried with a stream of nitrogen and the pellet washed five times with 100 ~1 of water. Usually, one-third of the final samples dissolved in 10 ~1 water were applied to a thin layer cellulose plate with plastic backing (Kodak) and 8 pg each of phosphoserine, phosphothreonine, and phosphotyrosine were added as internal standards. Electrophoresis at 700 V for 80 min was performed with cooling (10°C) in pyridine:acetic acid:water (5:45:950) (pH 3.6). The plates were then exposed to X-ray film before spraying with 0.25% ninhydrin in acetone to visualize the standards. Phosphopeptide analysis. The method of Nishikawa et al. (15) was followed with some modifications. Gel strips containing the ‘“P-labeled Ca*+ pump band from SDS-PAGE were excised and cut into pieces which were then digested with 50 pg of trypsin in 3 mM ammonium bicarbonate overnight at 37°C. The supernatant was dried with a stream of nitrogen and washed once with 100 bl water. The final samples were dissolved in 10 ~1 of water and aliquots of the samples were applied to silica plates (Analtech, silica gel H). The plates were subjected first to thin layer electrophoresis in acetic acidformic acid:water (15:5:80) at 750 V for 60 min. The air-dried plates were then subjected to ascending chromatography in n-butanol:pyridine:acetic acidwater (32.5:25:5:50) until the solvent front reached the top edge (4 h). The dried plates were then processed for autoradiography. Analyticalprocedures. SDS-polyacrylamide gel electrophoresis was done according to Laemmli (16). Densitometry analysis of autoradiogram and gels were performed with a video densitometer (Bio-Rad, Model 620).
RESULTS
AND
DISCUSSION
To study the phosphorylation of plasma membrane Ca2+ pump, we employed two techniques: (i) [32P]phosphate incubation of cultured cells and (ii) immunoprecipitation of the Ca2+ pump protein. For these studies, cultured rat aortic endothelial cells were utilized. The cells were first allowed sufficient time (4 h) for synthesizing 32P-labeled ATP after which the cells were lysed and the membrane proteins solubilized. Subsequently, the Ca2+ pump protein was immunoprecipitated employing a monoclonal antibody (5FlO) against the human erythrocyte Ca2+ pump (17). When the immunoprecipitates
ET AL.
were analyzed by SDS-polyacrylamide gel electrophoresis (SDS-PAGE), the Ca2+ pump was identified as a 150kDa band barely visible by Coomassie blue staining but easily observed with silver staining (results not shown). This represented most likely the isoform 1 protein (10). A smaller minor band (140 kDa) occasionally seen was probably a degradation product or an alternatively spliced protein of the Ca2+ pump isoform 1 (10). Aside from this minor band, no other protein bands within this molecular weight range were observed (results not shown). A strong band of about 50 kDa identified as the anti-Ca2+ pump immunoglobulin was also observed (results not shown) (10). The identity of the 150-kDa band as the Cazt pump was confirmed as follows: proteins were electrotransferred from a SDS-gel to nitrocellulose paper. Subsequent Western blotting with the anti-Ca2+ pump antibody and an alkaline phosphatase-linked second antibody (goat anti-mouse) showed that the 150 kDa was indeed the band to react with the anti-Ca2+ pump antibody (results not shown). Furthermore, the same 150-kDa band was recognized by several other independent monoclonal antibodies raised against the erythrocyte Ca2+ pump. In this series of studies, when endothelial cells were incubated (4 h) with [32P]phosphate (0.1 mCi/ml) in a serum-free medium, we found that the Ca2+ pump was phosphorylated (Fig. 1A). We consider this the basal level of phosphorylation as it does not require added stimulus or hormone. The Ca2+ pump phosphorylation appeared to be universal since it was also demonstrated in hamster cardiac myocytes and rat adipocytes (result not shown). Since protein kinase A and protein kinase C can phosphorylate the Ca2+ pump in h-0, the endothelial cells were treated for 10 min with CPT-CAMP (a protein kinase A activator) or with PMA (a protein kinase C activator). We found that both CPT-CAMP and PMA increased the level of Ca2+ pump phosphorylation by an average of 1.6fold (n = 3) and 3.0-fold (n = 4), respectively. The PMA effect in enhancing the Ca2+ pump phosphorylation was likely mediated by protein kinase C since a biologically inactive phorbol ester ricu-phorbol didecanoate (4a-PDD) was not effective (Fig. 1A). Forskolin (at 100 PM), which activates adenylate cyclase directly and thereby increases the level of CAMP, was also capable of increasing the phosphorylation level (results not shown). These results suggested the involvement of protein kinase A and protein kinase C in regulating the Ca2+ pump. It should be mentioned that while PMA consistently stimulated the phosphorylation by about 3-fold, the effect of CPT-CAMP was more variable, ranging from 1.4- to 2.0-fold (to be discussed later). In addition to serine/threonine-directed protein kinase A and C, we explored the possibility that there may be a receptor-linked tyrosine kinase involvement. Our results however suggest that under the conditions used, insulin (Fig. 1C) and epidermal growth factor (EGF) (results not shown) did not significantly alter the level of Ca2+ pump phosphorylation in endothelial cells.
IN
SITU
PHOSPHORYLATION
B
Time
(min)
FIG. 1. Immunoprecipitation of 32P-phosphorylated Ca*+ pump. (A) Autoradiogram of Coomassie blue-stained SDS-PAGE gel. The samples were immunoprecipitated from unstimulated cells (lane 1) or cells stimulated with 20 pM CPT-CAMP (lane 2), 150 nM PMA (lane 3), 1 pM 4w PDD (lane 4), 100 nM AI1 (lane 5), or 20 yM isoproterenol (lane 6). The molecular weight standards (not shown) were visualized by Coomassie blue staining. Results shown are typical of three separate experiments, except that the CPT-CAMP result is at the high end of the variation. (B) Densitometric quantitation of s*P-label on immunoprecipitated Ca2+ pump from cells treated with 150 nM PMA for 0, 5, 10, or 30 min (as indicated). (C) Densitometric quantitation of %label on immunoprecipitated Ca2+ pump from untreated cells (basal) (1) or cells treated with 20 pM CPT-CAMP (2), 150 nM PMA (3), 20 pM isoproterenol (4), 100 nM AI1 (5), 250 pM ATP (6), 100 nM bradykinin (7), 1 pM ionomycin (8), 10 pM histamine (9), 40 pM EGF-fragment (lo), or 100 nM insulin (11) (as indicated). Immunoprecipitates were subjected to SDS-PAGE. Average relative levels (n = 2-4) are shown in brackets, using the basal level of phosphorylation as 1.0.
The time course of stimulus-mediated increase of Ca2+ pump phosphorylation was studied. Since PMA produces a large overall increase in phosphorylation level, it was chosen for this study. The increase in phosphorylation was rapid, reaching 65% of the maximum after 5 min of PMA treatment, and plateaued at 10 min (Fig. 1B). It is generally accepted that different hormones or agonists may stimulate cells by activation of different protein kinases (18, 19). For example, angiotensin II (AII) and bradykinin have been implicated in protein kinase C activation in smooth muscle cells and endothelial cells,
OF CALCIUM
PUMP
105
respectively (20-22). On the other hand, a P-adrenergic agonist such as isoproterenol activates protein kinase A by increasing the CAMP level in many cell types (19,23). In this study, we attempted to demonstrate that the phosphorylation of the Ca2+ pump can be linked directly to hormone/agonist stimulation of the cells. Among the stimuli tested, the P-adrenergic agonist isoproterenol indeed enhanced the Ca2+ pump phosphorylation by 1.4fold (Fig. 1C). We also confirmed that both protein kinase C-activating hormones (AI1 and bradykinin) indeed enhanced the Ca2+ pump phosphorylation (Fig. 1C). Interestingly, the addition of ATP or calcium ionophore ionomycin also increased the Ca2’ pump phosphorylation to 1.8-fold and 1.5-fold, respectively, while histamine was ineffective (Fig. 1C). Since the hormone pretreatment was very brief (5-10 min), it seems unlikely that these hormones induced an increase in the specific activity of the radioactive ATP pool thereby affecting the level of 32Plabel on the immunoprecipitated Ca2+ pump protein. To gain further information on the basal and stimulusenhanced in situ phosphorylation of the Ca2+ pump by protein kinases, phosphoamino acid analysis of the 32Pphosphorylated Ca2+ pump was performed. With the unstimulated cells, basal phosphorylation appeared to occur more strongly on serine residue(s) than on threonine residues. The ratio of 32P-labeling on Ser:Thr was 3:l (Fig. 2). With both the PMA-stimulated and the CPT-cAMPstimulated cells, phosphorylation on both serine and threonine residues was again observed. A shift on Ser/ Thr phosphorylation ratio to 4:l was observed after stimulation by PMA (Fig. 2). This would suggest that an additional serine residue(s) was being phosphorylated. On the other hand, CPT-CAMP-stimulated phosphorylation did not significantly shift the labeled Ser:Thr ratio (Fig. 2). Both unstimulated and EGF-stimulated cells did not show any 32P-labeled phosphotyrosine (Fig. 2), which suggests that the Ca2+ pump is not likely to be regulated directly by tyrosine phosphorylation in this cell type. Thus the phosphoamino acid data indicate that the plasma membrane Ca2’ pump is phosphorylated minimally at one serine and one threonine residue. To further characterize the sites of phosphorylation, we compared the 32P-labeled tryptic peptide map pattern of immunoprecipitates from unstimulated cells with that from stimulated cells (Fig. 3). Basal phosphorylation from unstimulated cells was found to be associated with five distinct 32P-phosphopeptides from the Ca2+ pump (peptide spots 1 to 5, Fig. 3A). If each labeled phosphopeptide contained one distinct site of phosphorylation, our result would then suggest the presence of five distinct phosphorylation sites in unstimulated cells. However, it is possible that some of the phosphopeptides represented products of alternative tryptic cleavages near the same phosphorylation site. In this case, two or more peptides could contain the same phosphorylation site (24). Consequently, the actual number of phosphorylation sites may
106
WANG
4 P-S + P-T -
P-Y
ET AL.
nous CAMP. This would then account for the variable stimulation of Ca2+ pump phosphorylation by CPT-CAMP as mentioned above. It has been reported that the Ca2+ pump purified from erythrocytes was in vitro phosphorylated by protein kinase A only on serinell-is at the Cterminal (7). However, cDNA studies have shown that in the three Ca2+ pump isoforms from rat brain, there exist several protein kinase A censensus sequences thereby suggesting other potential phosphorylation sites (25,26). Our results from phosphoamino acid analysis (Fig. 2) also suggested that both serine and threonine residues may be phosphorylated by protein kinase A in intact endothelial cells. Tryptic peptide map of the Ca2+ pump from PMA-
A) Basal
3.0
2.8
4.1
3.1
P-S/P-T ratio
FIG. 2. Phosphoamino acid analysis of 3ZP-phosphorylated Ca2+ pump. Immunoprecipitates for the Ca2+ pump from (1) unstimulated (basal), (2) PMA-treated (PMA), (3) CPT-CAMP-treated (CPT-CAMP), or (4) EGF-treated cells (EGF) were isolated and subjected to SDS-PAGE as described under Materials and Methods. Gel strips containing the s2Plabeled Ca2+ pump bands were then subjected to phosphoamino acid analysis. Shown here is the autoradiogram of the thin layer cellulose plate. P-S, P-T, P-Y, and Org represent the positions of phosphoserine, phosphothreonine, phosphotyrosine, and the origin, respectively. The ratio of labeled P-S/P-T in each case is indicated below each lane. Results shown here are reproduced in at least two experiments.
be lower than the number of phosphopeptides. In addition, we cannot rule out the possibility that some of the phosphopeptides may originate from a minor alternatively spliced form of the calcium pump. Such distinctions can only be made by direct sequencing of the phosphopeptides concerned. However, our sequencing efforts so far have been unsuccessful due to the very low concentration of this pump protein in the cell membrane. The insufficient amount of the phosphopeptides produced from the tryptic digest has rendered their purification (for sequencing purpose) a difficult, if not insurmountable task. 32P-labeled phosphopeptides of the Ca2+ pump from CPT-CAMP-stimulated cells (four experiments) showed a similar pattern to that from the unstimulated cells (compare panel B with panel A, Fig. 3). Since CPT-CAMP increased the overall level of Ca2+ pump phosphorylation without increasing the number of 32P-labeled phosphopeptides or altering the phosphoserine:phosphothreonine ratio, it is possible that protein kinase A phosphorylation sites were similar to the phosphorylation sites present in unstimulated cells. We suggest that in the basal state, some cells were already partially stimulated by endoge-
B) CPT-CAMP
C) Isowoterenol
F) ATP
FIG. 3. Two-dimensional tryptic phosphopeptide mapping of the Ca2+ pump. Immunoprecipitates for the Ca*+ pump from unstimulated (A), CPT-CAMP(B), isoproterenol(C), PMA- (D), AII- (E), ATP- (F), and ionomycin-treated cells (G) were isolated and subjected to SDSPAGE as described under Materials and Methods. Gel strips containing the 32P-labeled Ca2+ pump band from SDS-PAGE were subjected to tryptic peptide mapping. Shown here are the autoradiograms of the dried silica plates. The origins are marked by (0). The directions of the electrophoresis (E) (from + to -) and ascending chromatography (C) are as illustrated. The pointers and numbers indicated the positions of observable 32P-labeled phosphopeptide spots. Different numbers were assigned for phosphopeptide with different positions. The patterns shown here are reproduced in at least two to four separate experiments.
IN
SITU PHOSPHORYLATION TABLE
Effect of Various Stimuli
Ca2+ efflux (nmol/min/106
cells)
on the Initial
OF CALCIUM
107
PUMP
I
Rate of Ca2+ Efflux by Endothelial
Cells
CPT-CAMP
Basal
PMA
AI1
0.17 f 0.03 n = 11
0.38 zk0.05
0.34 t- 0.04
0.30 2 0.06
n = 13
n = 12
n = 10
P < 0.0025
P < 0.0025
P < 0.05
Isoproterenol
0.30 f 0.03 n=4 P < 0.025
Note. Ca*+ efflux from cultured cells (5 X 105, 35 mm dish) was measured essentially as described by Furukawa et al. (27). Cells were rinsed three times at 37°C with BSS medium (146 mM NaCl, 4 mM KCl, 2 mM MgC12, 0.5 mM CaCl,, 10 mM glucose, 0.1% bovine serum albumin, and 10 mM Hepes-Tris, pH 7.4) and then incubated in 1 ml of BSS containing 10 &i of *Ca’+ for 4 h at 37°C. Following the treatment, cells were rinsed 10 times with the BSS media without CaCl* and NaCl (efflux media) for 1 min. In this efflux media, NaCl was replaced by 146 mM choline chloride. Time courses of basal “Ca’+ efflux were followed by replacing the efflux media (1.0 ml) every 20 s. For measuring hormone-stimulated efflux, cells were incubated with hormone (100 nM PMA, 100 nM AII, 20 pM CPT-CAMP, or 20 @M isoproterenol) for 5 min prior to the start of time course, and the media used for replacement were switched to those containing the appropriate agonist or hormone. The amount of “Ca*+ extruded from cells during each time interval was measured by liquid scintillation counting. The rate of Ca*+ efflux was calculated from the linear portion of the time course and corrlected for any variation in the cell number. Results presented here are means f SEM. Statistics were calculated with Student’s t test (one-tail). Significance was defined as P i 0.05.
treated cells showed three additional phosphopeptides, besides the five found in unstimulated and CPT-cAMPstimulated cells (Fig. 3). The pattern was reproducible in at least six experiments. Two of the peptides (peptides 6 and 7) were strongly phosphorylated while the third one (peptide 8) only incorporated 32P weakly (Fig. 3D). These data suggested that protein kinase A and protein kinase C phosphorylate the Ca” pump at different sites. It should be mentioned that the tryptic peptide map of the Ca2+ pump from CPT-CAMP-treated cells sometimes displayed two to three faintly labeled spots which most likely correspond to spots 6, 7, and 8 (Fig. 3B). However, these spots, if present, were usually rather weak and could not account for the total increased level of phosphorylation induced by CPT-CAMP. We suggest that the phosphorylation site(s) in peptides 6,7, and 8 may represent a good substrate site(s) for protein kinase C but a poor substrate site(s) for protein kinase A, or that protein kinase A activation could have indirectly and partly activated the protein kinase C pathway. When the cells were subjected to a time course of PMA treatment (5,10,15,30, and, 180 min), the phosphopeptide pattern for the Ca2+ pump at any time point was identical to the one shown here (Fi:g. 3D). Recently, it was demonstrated that long term treatment of cells by PMA (3 h) also induces synthesis of new Ca2+ pump protein (10). Taken together, this would suggest that the newly synthesized Ca2+ pump molecules were also phosphorylated on the same site(s). When the AII-stimulated cells were used, the phosphopeptide maps of the Ca2+ pump (consisting of peptides 1 to 8) were found to be identical to that from PMA-stimulated cells (Fig. 3). This is consistent with an earlier observation that AI1 treatment of the cells induces protein kinase C activation (20). On the other hand, isoproterenol-stimulated cells exhibited a phosphopeptide map showing peptides 1 to 5, characteristic
of the basal or CPT-CAMP-stimulated phosphorylation, as expected. These results confirmed that the hormoneinduced phosphorylation of the Ca2+ pump was indeed mediated by the respective protein kinase pathway. ATPstimulated cells showed a phosphopeptide pattern identical to that of PMA-treated cells (Fig. 3), suggesting that ATP, as an agonist, triggers protein kinase C activation. Interestingly, ionomycin-stimulated phosphorylation also has a phosphopeptide map similar to that of PMA-stimulated cells. Finally we attempted to correlate the rapid stimuliinduced phosphorylation of the plasma membrane Ca2+ pump to the effects of the same stimuli on the Ca2+ efflux from the cells. We examined 45Ca2+ efflux using intact endothelial cells in culture since it best matched the conditions of in situ phosphorylation. Typically, cells were preloaded with 45Ca, challenged with hormone, and Ca2+ efflux was measured. It was previously established that when the external medium is depleted of sodium (Naof) and calcium ion (Cd+), the plasma membrane Na+-Ca2+ exchanger and the Ca2+-Ca2+ exchange process are not active and therefore the Nao+/C$-independent Ca2+ efflux occurs via the plasma membrane Ca2+ pump (27). Using this method, it has been suggested that PMA stimulates the Ca2+ pump in vascular smooth muscle cells (27). It has also been reported that PMA stimulated calcium efflux from human platelets and this stimulated efflux may be mediated by the Ca2+ pump (28). In our study, the basal Ca2+ efflux was determined to be 0.17 nmol/ min/106 cells (n = 11) (Table I). Both PMA and CPTCAMP treatment of the cells significantly activated the rate of Ca2+ efflux to 0.38 and 0.30 nmol/min/lO’j cells, respectively (Table I). Furthermore, AI1 and isoproterenol also increased the Ca2+ efflux rate (to 0.34 and 0.30 nmol/ min/106 cells, respectively), most likely through their respective protein kinase activation pathways (Table I). It
108
WANG
also appears that higher phosphorylation level was accompanied by larger increase in Ca2+ efllux rate (compare Fig. 1C and Table I). Linear regression analysis of the data suggested a correlation between the Ca2+ pump phosphorylation and the stimulated Ca2+ e&x (correlation coefficient = 0.88). This stimulated Ca2+ efflux is likely to be mediated by the Ca2+ pump. We have also attempted to assay the plasma membrane Ca2+ pump activity directly by measuring Ca2+ uptake in the crude membrane vesicles (29). Our results indicated that Ca2+ uptake rates (30 PM trifluoperazine-sensitive) were higher with membranes from both CPT-CAMP-treated and PMA-treated cells as compared to unstimulated cells (data not shown). Furthermore, it was demonstrated previously that both protein kinase A and protein kinase C phosphorylation led to activation of the purified Ca2+ pump under in. vitro conditions (6, 7). Taken together, these data suggest that upon hormone stimulation, the Ca2+ pump is regulated directly by phosphorylation via protein kinase A or protein kinase C. However, we do not rule out possible involvement of other protein kinases such as calmodulin-dependent or cGMP-dependent kinases. In summary, our results illustrated that activators of both protein kinase C and protein kinase A led to phosphorylation of the plasma membrane Ca2+ pump protein in intact rat aortic endothelial cells. Phosphopeptide mapping of the Ca2+ pump suggested that protein kinases A and C phosphorylate the protein at different sites. Furthermore, the stimuli that produced phosphorylation of the pump protein also induced an increase in Ca2+ efllux from intact cells, most likely due to an activation of the Ca2+ pump. ACKNOWLEDGMENTS The authors acknowledge the support of National Institutes of Health (HL-39481) and the American Heart Association/Michigan. We also thank Dr. John Penniston (Mayo Clinic) for the anti-Ca*+ pump antibody and Ms. Leena Cooria and Ms. Wei Quin Liu for technical assistance. REFERENCES 1. Carafoli, E. (1987) Anna Rev. Biochem. 56, 395-433. 2. James, P. H., Pruschy, M., Vorherr, T. E., Penniston, J. T., and Carafoli, E. (1989) Biochemistry 28, 4253-4258. 3. Wang, K. K. W., Villalobo, A., and Roufogalis, B. D. (1988) Arch. Biochem. Biophys. 260, 696-704.
ET AL. 4. Caroni, P., and Carafoli, 5. Neyses, L., Reinlib, 10,283-10,287.
E. (1981) J. Biol. Chem. 256,9371-9373.
L., and Carafoli,
E. (1985) J. Biol. Chem. 260,
6. Smallwood, J. I., Giigi, B., and Rasmussen, H. (1988) J. Bid. Chem.
262,2195-2202. 7. James, P., Maeda, M., Fischer, R., Verma, A., Krebs, J., Penn&on, J. T., and Carafoli, E. (1988) J. Biol. Chem. 263, 2905-2910. 8. Wang, K. K. W., Wright, L. C., Machan, C. L., Allen, B. G., Conigrave, A. D., and Roufogalis, B. D. (1991) J. Biol. Chem. 266, 9078-9085. 9. Ogurusu, T., Wakabayashi, S., Furukawa, Tawada-Iwata, Y., Imagawa, T., and Shigekawa, M. (1990) J. Biochem. 108, 222-229. 10. Kuo, T. H., Wang, K. K. W., Carlock, L., Diglio, C., and Tsang, W. (1991) J. Biol. Chem. 266, 2520-2525. 11. Diglio, C. A., Grammas, P., Giacomelli, F., and Wiener, J. (1989) Lab. Invest. 60, 523-531. 12. Towbin, H., Staehelin, T., and Gordon, J. (1979) Proc. Natl. Acad. Sci. USA 76,4350-4354. 13. Blake, M. S., Johnson, K. H., Russell-Jones, E. C. (1984) Anal. Biochem. 136,175-179.
G. J., and Getschlich,
14. Cooper, J. A., Sefton, B. M., and Hunter, T. (1983) in Methods in Enzymology (Corbin, J. D., and Hardman, J. G., Eds.), Vol. 99, pp. 387-402, Academic Press, San Diego. 15. Nishikawa, M., Shirakawa, Chem. 260,897s8983. 16. Laemmli,
S., and Adelstein,
R. S. (1985) J. Biol.
U. K. (1971) Nature 227,680-685.
17. Barke, J. L., Caride, A., Verma, A. K., Penniston, R. (1989) Am. J. Physiol. 257, F&12-F849.
J. T., and Kumar,
18. Nishizuka, Y. (1984) Nature 308.693-698. 19. Edelman, A. M., Blumenthal, D. H., and Krebs, E. G. (1987) Annu. Rev. Biochem. 56, 567-613.
20. Taubman, M. B., Berk, B. C., Izumo, S., Tsudo, T., Alexander, R. W., and Nadal-Ginard,
B. (1989) J. Bid. Chem. 264,526-530.
21. Dixon, B. S., Breckon, R., Fortune, J., Sutherland, E., Simon, F. R., and Anderson,
R. J. (1989) Am. J. Physiol. 257, F808-F817.
22. Issandou, M., and Rozengurt, E. (1990) J. Biol. Chem. 265,11,89011,896.
23. Hausdorff,
W. P., Caron, M. G., and Lefkowitz, J. 4, 2881-2889.
24. 25. 26. 27.
R. J. (1990) FASEB
Haycock, J. W. (1990) J. Biol. Chem. 265, 11,682-11,691. Shull, G. E., and Greeb, J. (1988) J. Biol. Chem. 263,8646-8657. Greeb, J., and Shull, G. E. (1989) J. Bid. Chem. 264, 18,569-l&576. Furukawa, K-I., Tawada, Y., and Shigekawa, M. (1988) J. Bid. Chem.
263,8058-8065. 28. Rink, T. J., and Sage, S. 0. (1987) J. Physiol. 393, 513-524. 29. Kuo, T. H., and Tsang, W. (1987) Biochem. Biophys. Res. Commun. 152,1111-1116.
OF BIOCHEMISTRY
AND
BIOPHYSICS
289, No. 1, August 15, pp. 103-108, 1991
Vol.
Hormone-Induced Phosphorylation of the Plasma Membrane Calcium Pump in Cultured Aortic Endothelial Cells Kevin
K. W. Wang,’
Department
of
Pathology,
Yeng Sheng Du, Clement
Diglio,
Wayn,e State University
of
School
Wayne
Tsang,
and Tuan
Medicine, Detroit, Michigan
H. Kuo2
48201
Received March 7, 1991, and in revised form April 16, 1991
The regulation of the plas:ma membrane Ca2+ pump by hormones via phosphorylation in intact cells has not been clearly established. We now present evidence that the Ca” pump is phosphorylated on both serine and threonine residues in unstimulated and stimulated cultured rat aortic endothelial cells. Among the stimuli tested, the protein kinase C activator phorbol 1%myristate 13-acetate (PMA) was most potent and increased the level of phosphorylation threefold, while the CAMP-dependent protein kinase activator 6(4+hlorophenylthio)-CAMP (CPT-CAMP) stimulated the phosphorylation 1.6-fold. Two-dimensional tryptic phosphopeptide maps of the Ca2+ pump from unstimulated and CPT-CAMP-stimulated cells have identical patterns (five phosphopeptides) while PMA-stimulated cells have three additional phosphopeptides. Isoproterenol-, ATP-, angiotensin II-, and bradykinin-stimulated cells also have increased levels of Ca2+ pump phosphorylation.. Stimuli-induced phosphorylation of the Ca”+ pump was rapid (5-10 min) and was concomitant with stimulated calcium efflux from the same cells. This is the first direct evidence that the plasma membrane Ca2+ pump in intact cells is regulated by various hormones or agonists via CAMP-dependent protein o 1991 ~eademi~ kinase or protein kinase C phosphorylation. Press,
Inc.
The plasma membrane Ca2+ pump is important in maintaining submicromolar intracellular Ca2+ concentration in mammalian cells (1). This enzyme is thought to be stimulated by endogenous calmodulin directly (2). The enzyme can potentially be irreversibly activated by endogenous calcium-dependent protease (calpain) (3). i Research fellow, American Heart Association/Michigan Current address: Parke-Davis Pharmaceutical Research Warner-Lambert Company, Ann Arbor, MI 48106. ’ To whom correspondence should be addressed. 0003-9861/91
53.00 1991 by of reproduction
Copyright 0 All
rights
Academic in any
Press, form
Inc. reserved.
(1990). Division,
Furthermore, under in vitro conditions, the membranebound or the purified Ca” pump from human erythrocytes (4-S) or cardiac plasma membrane (5,9) was found to be activated following phosphorylation by CAMP-dependent protein kinase (protein kinase A)3 or protein kinase C. It is therefore conceivable that the Ca2+ pump is regulated by phosphorylation in living cells. In an earlier report, we showed that the Ca2+ pump is indeed phosphorylated in living cells (10). In this report, we present evidence that phosphorylation of the Ca2+ pump is hormone- or agonistinduced. MATERIALS
AND
METHODS
Materials. Phorbol esters and ionomycin were from Sigma. CPTCAMP was from Boehringer Mannheim. [32P]orthophosphate (3000 Ci/ mmol) was from Amersham. A monoclonal anti-Ca’+ pump antibody (5FlO) was a gift from Dr. J. T. Penn&on (Mayo Clinic). Molecular weight standards were from Bio-Rad. All reagents used were of analytical grade.
32P-lalwlingof endothelial cells and immurwprecipitation. Rat thoracic aorta endothelial cells were initiated and cloned as described previously by Diglio et al. (11). The cell cultures were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine, and 25 pg/ml gentamicin. Confluent cultures (5 X 10s cells/lOO-mm dish) were washed three times in phosphate-free MEM (Sigma) and incubated with the same medium containing 100 @i/ml [32P]orthophosphate (Amersham) for 4 h at 37°C under 5% CO,/95% air. Various agents were added for 10 min before the end of the [32P]phosphate incubation. Incubation was stopped by removing the medium and rapid washing with 10 mM sodium phosphate, 150 mM sodium chloride (pH 7.4), and the addition of 1.3 ml lysis buffer (1%
3 Abbreviations used: protein kinase A, CAMP-dependent protein kinase; protein kinase C, Ca’+-, phospholipid-dependent protein kinase; PMA, phorbol 12-myristate 13-acetate; CPT-CAMP, 8-(4-chlorophenylthio)-CAMP; 4(~-PDD, 4a-phorbol didecanoate; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; EGF, epidermal growth factor; EGTA, ethylene glycol bis(@aminoethyl ether)N,,hr’-tetraacetic acid; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid, Tris, tris(hydroxymethyl)aminomethane; PMSF, phenylmethylsulfonyl fluoride; AII, angiotensin II; BSA, bovine serum albumin.
104
WANG
Nonidet-P40,lO mM Tris-HCl, pH 7.4,150 mM NaCl, 1% bovine serum albumin (BSA), 1 mM phenylmethylsulfonyl fluoride, 5 mM EDTA, 5 mM EGTA, 50 mM NaF, 100 pM sodium vanadate, 5 rg/ml leupeptin, and 2 mM dithiothreitol). Cells were lysed for 20 min at 4°C. After cell debris was removed by centrifugation, lysates were cleared by incubating with 5 pg/ml rabbit anti-mouse immunoglobulin, 1% protein A-Sepharose, 3 ~1 normal ascetic fluid for 40 min at 4°C. The cleared lysate was then incubated at 4°C with 3 ~1 monoclonal antibody (5FlO) against the plasma membrane Gas+ pump. After 20 min, 5 yg rabbit anti-mouse immunoglobulin was added. Immunocomplexes were then adsorbed to 1% protein A-Sepharose at 4°C for 40 min with rocking. The immunoprecipitate-protein A-Sepharose complex was then centrifuged and washed five or six times with lysis buffer (as above, except without BSA). The complexes were then dissociated by treatment of 20% trichloroacetic acid at 4°C. The centrifuged protein precipitates were then neutralized with Tris base and solubilized in sample buffer (2% SDS, 10% glycerol, 100 mM Tris-HCl, pH 6.8, 0.002% bromophenyl blue, 0.1 mM p-mercaptoethanol) and electrophoresed in SDS-gradient (5-20%) polyacrylamide gel. The 32P-labeled Ca*+ pump band was identified by Western blotting (12, 13) using the anti-Ca*+ pump antibody and the alkaline phosphatase-conjugated goat anti-mouse IgG. Phosphoamino acid analysis. The method of Cooper et al. (14) for phosphoamino acid analysis was followed. Gel strips containing the 32Plabeled Ca*+ pump bands were first digested with 50 pg trypsin at 37°C overnight and then treated with 7 M HCl at 105°C for 60 min. The supernatant was dried with a stream of nitrogen and the pellet washed five times with 100 ~1 of water. Usually, one-third of the final samples dissolved in 10 ~1 water were applied to a thin layer cellulose plate with plastic backing (Kodak) and 8 pg each of phosphoserine, phosphothreonine, and phosphotyrosine were added as internal standards. Electrophoresis at 700 V for 80 min was performed with cooling (10°C) in pyridine:acetic acid:water (5:45:950) (pH 3.6). The plates were then exposed to X-ray film before spraying with 0.25% ninhydrin in acetone to visualize the standards. Phosphopeptide analysis. The method of Nishikawa et al. (15) was followed with some modifications. Gel strips containing the ‘“P-labeled Ca*+ pump band from SDS-PAGE were excised and cut into pieces which were then digested with 50 pg of trypsin in 3 mM ammonium bicarbonate overnight at 37°C. The supernatant was dried with a stream of nitrogen and washed once with 100 bl water. The final samples were dissolved in 10 ~1 of water and aliquots of the samples were applied to silica plates (Analtech, silica gel H). The plates were subjected first to thin layer electrophoresis in acetic acidformic acid:water (15:5:80) at 750 V for 60 min. The air-dried plates were then subjected to ascending chromatography in n-butanol:pyridine:acetic acidwater (32.5:25:5:50) until the solvent front reached the top edge (4 h). The dried plates were then processed for autoradiography. Analyticalprocedures. SDS-polyacrylamide gel electrophoresis was done according to Laemmli (16). Densitometry analysis of autoradiogram and gels were performed with a video densitometer (Bio-Rad, Model 620).
RESULTS
AND
DISCUSSION
To study the phosphorylation of plasma membrane Ca2+ pump, we employed two techniques: (i) [32P]phosphate incubation of cultured cells and (ii) immunoprecipitation of the Ca2+ pump protein. For these studies, cultured rat aortic endothelial cells were utilized. The cells were first allowed sufficient time (4 h) for synthesizing 32P-labeled ATP after which the cells were lysed and the membrane proteins solubilized. Subsequently, the Ca2+ pump protein was immunoprecipitated employing a monoclonal antibody (5FlO) against the human erythrocyte Ca2+ pump (17). When the immunoprecipitates
ET AL.
were analyzed by SDS-polyacrylamide gel electrophoresis (SDS-PAGE), the Ca2+ pump was identified as a 150kDa band barely visible by Coomassie blue staining but easily observed with silver staining (results not shown). This represented most likely the isoform 1 protein (10). A smaller minor band (140 kDa) occasionally seen was probably a degradation product or an alternatively spliced protein of the Ca2+ pump isoform 1 (10). Aside from this minor band, no other protein bands within this molecular weight range were observed (results not shown). A strong band of about 50 kDa identified as the anti-Ca2+ pump immunoglobulin was also observed (results not shown) (10). The identity of the 150-kDa band as the Cazt pump was confirmed as follows: proteins were electrotransferred from a SDS-gel to nitrocellulose paper. Subsequent Western blotting with the anti-Ca2+ pump antibody and an alkaline phosphatase-linked second antibody (goat anti-mouse) showed that the 150 kDa was indeed the band to react with the anti-Ca2+ pump antibody (results not shown). Furthermore, the same 150-kDa band was recognized by several other independent monoclonal antibodies raised against the erythrocyte Ca2+ pump. In this series of studies, when endothelial cells were incubated (4 h) with [32P]phosphate (0.1 mCi/ml) in a serum-free medium, we found that the Ca2+ pump was phosphorylated (Fig. 1A). We consider this the basal level of phosphorylation as it does not require added stimulus or hormone. The Ca2+ pump phosphorylation appeared to be universal since it was also demonstrated in hamster cardiac myocytes and rat adipocytes (result not shown). Since protein kinase A and protein kinase C can phosphorylate the Ca2+ pump in h-0, the endothelial cells were treated for 10 min with CPT-CAMP (a protein kinase A activator) or with PMA (a protein kinase C activator). We found that both CPT-CAMP and PMA increased the level of Ca2+ pump phosphorylation by an average of 1.6fold (n = 3) and 3.0-fold (n = 4), respectively. The PMA effect in enhancing the Ca2+ pump phosphorylation was likely mediated by protein kinase C since a biologically inactive phorbol ester ricu-phorbol didecanoate (4a-PDD) was not effective (Fig. 1A). Forskolin (at 100 PM), which activates adenylate cyclase directly and thereby increases the level of CAMP, was also capable of increasing the phosphorylation level (results not shown). These results suggested the involvement of protein kinase A and protein kinase C in regulating the Ca2+ pump. It should be mentioned that while PMA consistently stimulated the phosphorylation by about 3-fold, the effect of CPT-CAMP was more variable, ranging from 1.4- to 2.0-fold (to be discussed later). In addition to serine/threonine-directed protein kinase A and C, we explored the possibility that there may be a receptor-linked tyrosine kinase involvement. Our results however suggest that under the conditions used, insulin (Fig. 1C) and epidermal growth factor (EGF) (results not shown) did not significantly alter the level of Ca2+ pump phosphorylation in endothelial cells.
IN
SITU
PHOSPHORYLATION
B
Time
(min)
FIG. 1. Immunoprecipitation of 32P-phosphorylated Ca*+ pump. (A) Autoradiogram of Coomassie blue-stained SDS-PAGE gel. The samples were immunoprecipitated from unstimulated cells (lane 1) or cells stimulated with 20 pM CPT-CAMP (lane 2), 150 nM PMA (lane 3), 1 pM 4w PDD (lane 4), 100 nM AI1 (lane 5), or 20 yM isoproterenol (lane 6). The molecular weight standards (not shown) were visualized by Coomassie blue staining. Results shown are typical of three separate experiments, except that the CPT-CAMP result is at the high end of the variation. (B) Densitometric quantitation of s*P-label on immunoprecipitated Ca2+ pump from cells treated with 150 nM PMA for 0, 5, 10, or 30 min (as indicated). (C) Densitometric quantitation of %label on immunoprecipitated Ca2+ pump from untreated cells (basal) (1) or cells treated with 20 pM CPT-CAMP (2), 150 nM PMA (3), 20 pM isoproterenol (4), 100 nM AI1 (5), 250 pM ATP (6), 100 nM bradykinin (7), 1 pM ionomycin (8), 10 pM histamine (9), 40 pM EGF-fragment (lo), or 100 nM insulin (11) (as indicated). Immunoprecipitates were subjected to SDS-PAGE. Average relative levels (n = 2-4) are shown in brackets, using the basal level of phosphorylation as 1.0.
The time course of stimulus-mediated increase of Ca2+ pump phosphorylation was studied. Since PMA produces a large overall increase in phosphorylation level, it was chosen for this study. The increase in phosphorylation was rapid, reaching 65% of the maximum after 5 min of PMA treatment, and plateaued at 10 min (Fig. 1B). It is generally accepted that different hormones or agonists may stimulate cells by activation of different protein kinases (18, 19). For example, angiotensin II (AII) and bradykinin have been implicated in protein kinase C activation in smooth muscle cells and endothelial cells,
OF CALCIUM
PUMP
105
respectively (20-22). On the other hand, a P-adrenergic agonist such as isoproterenol activates protein kinase A by increasing the CAMP level in many cell types (19,23). In this study, we attempted to demonstrate that the phosphorylation of the Ca2+ pump can be linked directly to hormone/agonist stimulation of the cells. Among the stimuli tested, the P-adrenergic agonist isoproterenol indeed enhanced the Ca2+ pump phosphorylation by 1.4fold (Fig. 1C). We also confirmed that both protein kinase C-activating hormones (AI1 and bradykinin) indeed enhanced the Ca2+ pump phosphorylation (Fig. 1C). Interestingly, the addition of ATP or calcium ionophore ionomycin also increased the Ca2’ pump phosphorylation to 1.8-fold and 1.5-fold, respectively, while histamine was ineffective (Fig. 1C). Since the hormone pretreatment was very brief (5-10 min), it seems unlikely that these hormones induced an increase in the specific activity of the radioactive ATP pool thereby affecting the level of 32Plabel on the immunoprecipitated Ca2+ pump protein. To gain further information on the basal and stimulusenhanced in situ phosphorylation of the Ca2+ pump by protein kinases, phosphoamino acid analysis of the 32Pphosphorylated Ca2+ pump was performed. With the unstimulated cells, basal phosphorylation appeared to occur more strongly on serine residue(s) than on threonine residues. The ratio of 32P-labeling on Ser:Thr was 3:l (Fig. 2). With both the PMA-stimulated and the CPT-cAMPstimulated cells, phosphorylation on both serine and threonine residues was again observed. A shift on Ser/ Thr phosphorylation ratio to 4:l was observed after stimulation by PMA (Fig. 2). This would suggest that an additional serine residue(s) was being phosphorylated. On the other hand, CPT-CAMP-stimulated phosphorylation did not significantly shift the labeled Ser:Thr ratio (Fig. 2). Both unstimulated and EGF-stimulated cells did not show any 32P-labeled phosphotyrosine (Fig. 2), which suggests that the Ca2+ pump is not likely to be regulated directly by tyrosine phosphorylation in this cell type. Thus the phosphoamino acid data indicate that the plasma membrane Ca2’ pump is phosphorylated minimally at one serine and one threonine residue. To further characterize the sites of phosphorylation, we compared the 32P-labeled tryptic peptide map pattern of immunoprecipitates from unstimulated cells with that from stimulated cells (Fig. 3). Basal phosphorylation from unstimulated cells was found to be associated with five distinct 32P-phosphopeptides from the Ca2+ pump (peptide spots 1 to 5, Fig. 3A). If each labeled phosphopeptide contained one distinct site of phosphorylation, our result would then suggest the presence of five distinct phosphorylation sites in unstimulated cells. However, it is possible that some of the phosphopeptides represented products of alternative tryptic cleavages near the same phosphorylation site. In this case, two or more peptides could contain the same phosphorylation site (24). Consequently, the actual number of phosphorylation sites may
106
WANG
4 P-S + P-T -
P-Y
ET AL.
nous CAMP. This would then account for the variable stimulation of Ca2+ pump phosphorylation by CPT-CAMP as mentioned above. It has been reported that the Ca2+ pump purified from erythrocytes was in vitro phosphorylated by protein kinase A only on serinell-is at the Cterminal (7). However, cDNA studies have shown that in the three Ca2+ pump isoforms from rat brain, there exist several protein kinase A censensus sequences thereby suggesting other potential phosphorylation sites (25,26). Our results from phosphoamino acid analysis (Fig. 2) also suggested that both serine and threonine residues may be phosphorylated by protein kinase A in intact endothelial cells. Tryptic peptide map of the Ca2+ pump from PMA-
A) Basal
3.0
2.8
4.1
3.1
P-S/P-T ratio
FIG. 2. Phosphoamino acid analysis of 3ZP-phosphorylated Ca2+ pump. Immunoprecipitates for the Ca2+ pump from (1) unstimulated (basal), (2) PMA-treated (PMA), (3) CPT-CAMP-treated (CPT-CAMP), or (4) EGF-treated cells (EGF) were isolated and subjected to SDS-PAGE as described under Materials and Methods. Gel strips containing the s2Plabeled Ca2+ pump bands were then subjected to phosphoamino acid analysis. Shown here is the autoradiogram of the thin layer cellulose plate. P-S, P-T, P-Y, and Org represent the positions of phosphoserine, phosphothreonine, phosphotyrosine, and the origin, respectively. The ratio of labeled P-S/P-T in each case is indicated below each lane. Results shown here are reproduced in at least two experiments.
be lower than the number of phosphopeptides. In addition, we cannot rule out the possibility that some of the phosphopeptides may originate from a minor alternatively spliced form of the calcium pump. Such distinctions can only be made by direct sequencing of the phosphopeptides concerned. However, our sequencing efforts so far have been unsuccessful due to the very low concentration of this pump protein in the cell membrane. The insufficient amount of the phosphopeptides produced from the tryptic digest has rendered their purification (for sequencing purpose) a difficult, if not insurmountable task. 32P-labeled phosphopeptides of the Ca2+ pump from CPT-CAMP-stimulated cells (four experiments) showed a similar pattern to that from the unstimulated cells (compare panel B with panel A, Fig. 3). Since CPT-CAMP increased the overall level of Ca2+ pump phosphorylation without increasing the number of 32P-labeled phosphopeptides or altering the phosphoserine:phosphothreonine ratio, it is possible that protein kinase A phosphorylation sites were similar to the phosphorylation sites present in unstimulated cells. We suggest that in the basal state, some cells were already partially stimulated by endoge-
B) CPT-CAMP
C) Isowoterenol
F) ATP
FIG. 3. Two-dimensional tryptic phosphopeptide mapping of the Ca2+ pump. Immunoprecipitates for the Ca*+ pump from unstimulated (A), CPT-CAMP(B), isoproterenol(C), PMA- (D), AII- (E), ATP- (F), and ionomycin-treated cells (G) were isolated and subjected to SDSPAGE as described under Materials and Methods. Gel strips containing the 32P-labeled Ca2+ pump band from SDS-PAGE were subjected to tryptic peptide mapping. Shown here are the autoradiograms of the dried silica plates. The origins are marked by (0). The directions of the electrophoresis (E) (from + to -) and ascending chromatography (C) are as illustrated. The pointers and numbers indicated the positions of observable 32P-labeled phosphopeptide spots. Different numbers were assigned for phosphopeptide with different positions. The patterns shown here are reproduced in at least two to four separate experiments.
IN
SITU PHOSPHORYLATION TABLE
Effect of Various Stimuli
Ca2+ efflux (nmol/min/106
cells)
on the Initial
OF CALCIUM
107
PUMP
I
Rate of Ca2+ Efflux by Endothelial
Cells
CPT-CAMP
Basal
PMA
AI1
0.17 f 0.03 n = 11
0.38 zk0.05
0.34 t- 0.04
0.30 2 0.06
n = 13
n = 12
n = 10
P < 0.0025
P < 0.0025
P < 0.05
Isoproterenol
0.30 f 0.03 n=4 P < 0.025
Note. Ca*+ efflux from cultured cells (5 X 105, 35 mm dish) was measured essentially as described by Furukawa et al. (27). Cells were rinsed three times at 37°C with BSS medium (146 mM NaCl, 4 mM KCl, 2 mM MgC12, 0.5 mM CaCl,, 10 mM glucose, 0.1% bovine serum albumin, and 10 mM Hepes-Tris, pH 7.4) and then incubated in 1 ml of BSS containing 10 &i of *Ca’+ for 4 h at 37°C. Following the treatment, cells were rinsed 10 times with the BSS media without CaCl* and NaCl (efflux media) for 1 min. In this efflux media, NaCl was replaced by 146 mM choline chloride. Time courses of basal “Ca’+ efflux were followed by replacing the efflux media (1.0 ml) every 20 s. For measuring hormone-stimulated efflux, cells were incubated with hormone (100 nM PMA, 100 nM AII, 20 pM CPT-CAMP, or 20 @M isoproterenol) for 5 min prior to the start of time course, and the media used for replacement were switched to those containing the appropriate agonist or hormone. The amount of “Ca*+ extruded from cells during each time interval was measured by liquid scintillation counting. The rate of Ca*+ efflux was calculated from the linear portion of the time course and corrlected for any variation in the cell number. Results presented here are means f SEM. Statistics were calculated with Student’s t test (one-tail). Significance was defined as P i 0.05.
treated cells showed three additional phosphopeptides, besides the five found in unstimulated and CPT-cAMPstimulated cells (Fig. 3). The pattern was reproducible in at least six experiments. Two of the peptides (peptides 6 and 7) were strongly phosphorylated while the third one (peptide 8) only incorporated 32P weakly (Fig. 3D). These data suggested that protein kinase A and protein kinase C phosphorylate the Ca” pump at different sites. It should be mentioned that the tryptic peptide map of the Ca2+ pump from CPT-CAMP-treated cells sometimes displayed two to three faintly labeled spots which most likely correspond to spots 6, 7, and 8 (Fig. 3B). However, these spots, if present, were usually rather weak and could not account for the total increased level of phosphorylation induced by CPT-CAMP. We suggest that the phosphorylation site(s) in peptides 6,7, and 8 may represent a good substrate site(s) for protein kinase C but a poor substrate site(s) for protein kinase A, or that protein kinase A activation could have indirectly and partly activated the protein kinase C pathway. When the cells were subjected to a time course of PMA treatment (5,10,15,30, and, 180 min), the phosphopeptide pattern for the Ca2+ pump at any time point was identical to the one shown here (Fi:g. 3D). Recently, it was demonstrated that long term treatment of cells by PMA (3 h) also induces synthesis of new Ca2+ pump protein (10). Taken together, this would suggest that the newly synthesized Ca2+ pump molecules were also phosphorylated on the same site(s). When the AII-stimulated cells were used, the phosphopeptide maps of the Ca2+ pump (consisting of peptides 1 to 8) were found to be identical to that from PMA-stimulated cells (Fig. 3). This is consistent with an earlier observation that AI1 treatment of the cells induces protein kinase C activation (20). On the other hand, isoproterenol-stimulated cells exhibited a phosphopeptide map showing peptides 1 to 5, characteristic
of the basal or CPT-CAMP-stimulated phosphorylation, as expected. These results confirmed that the hormoneinduced phosphorylation of the Ca2+ pump was indeed mediated by the respective protein kinase pathway. ATPstimulated cells showed a phosphopeptide pattern identical to that of PMA-treated cells (Fig. 3), suggesting that ATP, as an agonist, triggers protein kinase C activation. Interestingly, ionomycin-stimulated phosphorylation also has a phosphopeptide map similar to that of PMA-stimulated cells. Finally we attempted to correlate the rapid stimuliinduced phosphorylation of the plasma membrane Ca2+ pump to the effects of the same stimuli on the Ca2+ efflux from the cells. We examined 45Ca2+ efflux using intact endothelial cells in culture since it best matched the conditions of in situ phosphorylation. Typically, cells were preloaded with 45Ca, challenged with hormone, and Ca2+ efflux was measured. It was previously established that when the external medium is depleted of sodium (Naof) and calcium ion (Cd+), the plasma membrane Na+-Ca2+ exchanger and the Ca2+-Ca2+ exchange process are not active and therefore the Nao+/C$-independent Ca2+ efflux occurs via the plasma membrane Ca2+ pump (27). Using this method, it has been suggested that PMA stimulates the Ca2+ pump in vascular smooth muscle cells (27). It has also been reported that PMA stimulated calcium efflux from human platelets and this stimulated efflux may be mediated by the Ca2+ pump (28). In our study, the basal Ca2+ efflux was determined to be 0.17 nmol/ min/106 cells (n = 11) (Table I). Both PMA and CPTCAMP treatment of the cells significantly activated the rate of Ca2+ efflux to 0.38 and 0.30 nmol/min/lO’j cells, respectively (Table I). Furthermore, AI1 and isoproterenol also increased the Ca2+ efflux rate (to 0.34 and 0.30 nmol/ min/106 cells, respectively), most likely through their respective protein kinase activation pathways (Table I). It
108
WANG
also appears that higher phosphorylation level was accompanied by larger increase in Ca2+ efllux rate (compare Fig. 1C and Table I). Linear regression analysis of the data suggested a correlation between the Ca2+ pump phosphorylation and the stimulated Ca2+ e&x (correlation coefficient = 0.88). This stimulated Ca2+ efflux is likely to be mediated by the Ca2+ pump. We have also attempted to assay the plasma membrane Ca2+ pump activity directly by measuring Ca2+ uptake in the crude membrane vesicles (29). Our results indicated that Ca2+ uptake rates (30 PM trifluoperazine-sensitive) were higher with membranes from both CPT-CAMP-treated and PMA-treated cells as compared to unstimulated cells (data not shown). Furthermore, it was demonstrated previously that both protein kinase A and protein kinase C phosphorylation led to activation of the purified Ca2+ pump under in. vitro conditions (6, 7). Taken together, these data suggest that upon hormone stimulation, the Ca2+ pump is regulated directly by phosphorylation via protein kinase A or protein kinase C. However, we do not rule out possible involvement of other protein kinases such as calmodulin-dependent or cGMP-dependent kinases. In summary, our results illustrated that activators of both protein kinase C and protein kinase A led to phosphorylation of the plasma membrane Ca2+ pump protein in intact rat aortic endothelial cells. Phosphopeptide mapping of the Ca2+ pump suggested that protein kinases A and C phosphorylate the protein at different sites. Furthermore, the stimuli that produced phosphorylation of the pump protein also induced an increase in Ca2+ efllux from intact cells, most likely due to an activation of the Ca2+ pump. ACKNOWLEDGMENTS The authors acknowledge the support of National Institutes of Health (HL-39481) and the American Heart Association/Michigan. We also thank Dr. John Penniston (Mayo Clinic) for the anti-Ca*+ pump antibody and Ms. Leena Cooria and Ms. Wei Quin Liu for technical assistance. REFERENCES 1. Carafoli, E. (1987) Anna Rev. Biochem. 56, 395-433. 2. James, P. H., Pruschy, M., Vorherr, T. E., Penniston, J. T., and Carafoli, E. (1989) Biochemistry 28, 4253-4258. 3. Wang, K. K. W., Villalobo, A., and Roufogalis, B. D. (1988) Arch. Biochem. Biophys. 260, 696-704.
ET AL. 4. Caroni, P., and Carafoli, 5. Neyses, L., Reinlib, 10,283-10,287.
E. (1981) J. Biol. Chem. 256,9371-9373.
L., and Carafoli,
E. (1985) J. Biol. Chem. 260,
6. Smallwood, J. I., Giigi, B., and Rasmussen, H. (1988) J. Bid. Chem.
262,2195-2202. 7. James, P., Maeda, M., Fischer, R., Verma, A., Krebs, J., Penn&on, J. T., and Carafoli, E. (1988) J. Biol. Chem. 263, 2905-2910. 8. Wang, K. K. W., Wright, L. C., Machan, C. L., Allen, B. G., Conigrave, A. D., and Roufogalis, B. D. (1991) J. Biol. Chem. 266, 9078-9085. 9. Ogurusu, T., Wakabayashi, S., Furukawa, Tawada-Iwata, Y., Imagawa, T., and Shigekawa, M. (1990) J. Biochem. 108, 222-229. 10. Kuo, T. H., Wang, K. K. W., Carlock, L., Diglio, C., and Tsang, W. (1991) J. Biol. Chem. 266, 2520-2525. 11. Diglio, C. A., Grammas, P., Giacomelli, F., and Wiener, J. (1989) Lab. Invest. 60, 523-531. 12. Towbin, H., Staehelin, T., and Gordon, J. (1979) Proc. Natl. Acad. Sci. USA 76,4350-4354. 13. Blake, M. S., Johnson, K. H., Russell-Jones, E. C. (1984) Anal. Biochem. 136,175-179.
G. J., and Getschlich,
14. Cooper, J. A., Sefton, B. M., and Hunter, T. (1983) in Methods in Enzymology (Corbin, J. D., and Hardman, J. G., Eds.), Vol. 99, pp. 387-402, Academic Press, San Diego. 15. Nishikawa, M., Shirakawa, Chem. 260,897s8983. 16. Laemmli,
S., and Adelstein,
R. S. (1985) J. Biol.
U. K. (1971) Nature 227,680-685.
17. Barke, J. L., Caride, A., Verma, A. K., Penniston, R. (1989) Am. J. Physiol. 257, F&12-F849.
J. T., and Kumar,
18. Nishizuka, Y. (1984) Nature 308.693-698. 19. Edelman, A. M., Blumenthal, D. H., and Krebs, E. G. (1987) Annu. Rev. Biochem. 56, 567-613.
20. Taubman, M. B., Berk, B. C., Izumo, S., Tsudo, T., Alexander, R. W., and Nadal-Ginard,
B. (1989) J. Bid. Chem. 264,526-530.
21. Dixon, B. S., Breckon, R., Fortune, J., Sutherland, E., Simon, F. R., and Anderson,
R. J. (1989) Am. J. Physiol. 257, F808-F817.
22. Issandou, M., and Rozengurt, E. (1990) J. Biol. Chem. 265,11,89011,896.
23. Hausdorff,
W. P., Caron, M. G., and Lefkowitz, J. 4, 2881-2889.
24. 25. 26. 27.
R. J. (1990) FASEB
Haycock, J. W. (1990) J. Biol. Chem. 265, 11,682-11,691. Shull, G. E., and Greeb, J. (1988) J. Biol. Chem. 263,8646-8657. Greeb, J., and Shull, G. E. (1989) J. Bid. Chem. 264, 18,569-l&576. Furukawa, K-I., Tawada, Y., and Shigekawa, M. (1988) J. Bid. Chem.
263,8058-8065. 28. Rink, T. J., and Sage, S. 0. (1987) J. Physiol. 393, 513-524. 29. Kuo, T. H., and Tsang, W. (1987) Biochem. Biophys. Res. Commun. 152,1111-1116.
