Regulation of Platelet Plasma Membrane Ca2+-ATPase by cAMP-dependent and Tyrosine Phosphorylation*

(Received for publication, October 24, 1996, and in revised form, March 24, 1997)

William L. Dean Dagger , Dong Chen §, Paul C. Brandt § and Thomas C. Vanaman §

From the Dagger  Department of Biochemistry, University of Louisville School of Medicine, Louisville, Kentucky 40292 and the § Department of Biochemistry, University of Kentucky College of Medicine, Lexington, Kentucky 40536-0084

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
REFERENCES


ABSTRACT

As a consequence of its central role in the regulation of calcium metabolism in the platelet, the plasma membrane Ca2+-ATPase (PMCA) was assessed for cAMP-dependent and tyrosine phosphorylation. Addition of forskolin or prostaglandin E1, agents known to elevate platelet cAMP and calcium efflux, to platelets pre-labeled with [32P]PO4 resulted in the direct phosphorylation of platelet PMCA. Similarly, addition of the catalytic subunit of protein kinase A to platelet plasma membranes resulted in a 1.4-fold stimulation of activity. Thus, the previously reported inhibition of platelet activation by elevated intracellular cAMP may be accomplished in part by stimulation of PMCA, likely resulting in a decrease in intracellular calcium.

Treatment with thrombin evoked tyrosine phosphorylation of platelet PMCA, while PMCA from resting platelets exhibited little tyrosine phosphorylation. Phosphorylation of platelet plasma membranes by pp60src resulted in 75% inhibition of PMCA activity within 15 min. Similarly, membranes isolated from thrombin-treated platelets exhibited 40% lower PMCA activity than those from resting platelets. Phosphorylation of erythrocyte ghosts and purified PMCA by pp60src also resulted in up to 75% inhibition of Ca2+-ATPase activity, and inhibition was correlated with tyrosine phosphorylation. Sequencing of a peptide obtained after 32P labeling of purified erythrocyte PMCA in vitro showed that tyrosine 1176 of PMCA4b is phosphorylated by pp60src. These results indicate that tyrosine phosphorylation of platelet PMCA may serve as positive feedback to inhibit PMCA and increase intracellular calcium during platelet activation.


INTRODUCTION

Ca2+ signaling is an integral component of activation in platelets (1). The intracellular Ca2+ concentration is regulated by Ca2+-ATPases and Ca2+ channels located in both the plasma membrane and in the dense tubular system, a modified endoplasmic reticulum in the platelet (2). Strong agonists such as thrombin promote release of Ca2+ from internal stores and influx through the plasma membrane. Conversely, platelet antagonists such as prostaglandin E1 that elevate platelet cAMP, reduce cytosolic Ca2+ levels. Since the plasma membrane Ca2+-ATPase (PMCA)1 is the main agent of Ca2+ removal at resting Ca2+ concentrations (2), it is a key point for regulation of platelet Ca2+ metabolism. The activity of PMCA has been shown to be activated by several mechanisms including Ca2+/calmodulin, protein kinases A and C, acidic phospholipids, and proteolytic attack (3). In fact, Johnson et al. (4) showed indirectly that cAMP increases the rate of Ca2+ extrusion in platelets. However, direct phosphorylation of the platelet PMCA has not been reported previously.

In addition to increased cytosolic Ca2+, platelet activation is also accompanied by tyrosine phosphorylation of several proteins including focal adhesion kinase (pp125FAK) and the non-receptor tyrosine kinase pp60src followed by their association with the cytoskeleton (5, 6). Ca2+ appears to be intimately involved in both tyrosine phosphorylation and cytoskeletal rearrangement in the platelet (7). It has recently been demonstrated that the IP3 receptor can be phosphorylated on tyrosine residues leading to increased release of Ca2+ from internal stores (8). We report here that tyrosine phosphorylation of PMCA inhibits its activity further enhancing generation of the calcium signal.

In the present work we have assessed direct phosphorylation of platelet PMCA in vivo after exposure to thrombin, an agonist, or prostaglandin E1, an antagonist. In addition, the effects of both cAMP-dependent and tyrosine kinase-mediated phosphorylation of platelet PMCA in vitro on Ca2+-ATPase activity were measured. Finally, in vitro tyrosine phosphorylation of erythrocyte PMCA was used to determine the phosphorylated tyrosine residues. Purified erythrocyte PMCA was used as a model for platelet PMCA since we demonstrate in this work that the platelet contains the same isoforms of PMCA as found in the erythrocyte, PMCA1 and PMCA4 (9).


EXPERIMENTAL PROCEDURES

Materials

Outdated human platelet concentrates and whole blood were obtained from the Louisville Chapter of the American Red Cross and from the Central Kentucky Blood Center (Lexington, KY). Polyclonal antibodies against human erythrocyte PMCA were raised in rabbits as described previously (10), and monoclonal anti-PMCA antibody 5F10 (11) was obtained from Affinity BioReagents (Neshanic Station, NJ). Isoform-specific anti-PMCA antibodies (9) were kindly provided by Dr. Ernesto Carafoli (Swiss Federal Institute of Technology, Zurich, Switzerland). Mouse monoclonal anti-phosphotyrosine (PY-20) was purchased from Transduction Laboratories (Lexington, KY). All secondary antibodies, electrophoresis and Western blotting reagents were obtained from Bio-Rad. Thrombin, prostaglandin E1, ATP, phosphoenolpyruvate, NADH, pyruvate kinase, lactate dehydrogenase, catalytic subunit of cAMP-dependent kinase, and egg yolk phosphatidylcholine were purchased from Sigma. Triton X-100 and protein A-agarose were products from Pierce. Genistein was purchased from Calbiochem. 32P-Labeled Na3PO4 and ATP were products from DuPont NEN. Endoproteinase Lys-C was purchased from Boehringer Mannheim. Purified recombinant pp60src and protein phosphatase type 2A were products from Upstate Biotechnology Inc. (Lake Placid, NY).

Platelet and Erythrocyte Plasma Membrane Preparation

Platelet plasma membranes were prepared by the glycerol lysis method described by Harmon et al. (12), except that membranes from lysed platelets were separated on a sucrose step gradient (33 and 66% w/v) in a Beckman SW-28 rotor at 26,000 rpm for 5 h (13). Membranes collected at the 0-33% sucrose interface were concentrated by centrifugation at 100,000 × g, resuspended in 0.01 M TES buffer, pH 7.4, containing 0.1 M KCl, 20% (w/v) glycerol, 1 mM EGTA, 10 µg/ml leupeptin, and 10 µg/ml pepstatin A, frozen in liquid nitrogen, and stored at -70 °C.

Erythrocyte ghosts were obtained from washed cells by hypertonic lysis according to Steck and Kant (14). Final membrane pellets were suspended in 0.01 M TES buffer, pH 7.4, containing 0.1 M KCl, and 20% (w/v) glycerol, frozen in liquid nitrogen, and stored at -70 °C.

Purification of Erythrocyte PMCA

Calmodulin-free erythrocyte ghosts were prepared from outdated whole blood and PMCA purified by calmodulin-Sepharose affinity chromatography according to the method of Niggli et al. (15). The final preparation exhibited a major band at ~140 kDa and a minor band at ~100 kDa on SDS-polyacrylamide gels. The specific activity of the purified material was 3.8 µmol of ATP hydrolyzed/min/mg, and a 2.2-fold stimulation of activity was obtained with 50 nM calmodulin at 10 µM Ca2+. Purified PMCA was frozen and stored at -70 °C.

Ca2+-ATPase Assay

Ca2+-ATPase activity was assayed using a system in which ATP hydrolysis is coupled to NADH oxidation by pyruvate kinase using phosphoenolpyruvate and pyruvate kinase (16). Purified erythrocyte PMCA was assayed in 0.01 M TES buffer at pH 7.2 containing 20 mM HEPES buffer, 5 mM MgCl2, 1 mM ATP, 50 nM calmodulin, 2.1 mM CaCl2, 130 mM NaCl, 2 mM EGTA, 2 mM dithiothreitol, 0.5 mg/ml Triton X-100, 5% (w/v) glycerol, 0.5 mg/ml phosphatidylcholine, and the coupled assay components. The difference in the rate of ATP hydrolysis in the presence and absence of 10 mM EGTA was used to calculate the Ca2+-ATPase activity. Assay of erythrocyte ghosts and platelet plasma membranes was performed in 0.01 M TES buffer at pH 7.5, containing 0.1 M KCl, 5 mM ATP, 5 mM MgCl2, 0.01 mM CaCl2, and coupled assay components.

Labeling of Whole Platelets with [32P]PO4 and Immunoprecipitation of PMCA

Platelet concentrates (5 × 1010 cells/ml) in citrate anticoagulant were centrifuged at 5,000 × g and resuspended in the same volume of Tyrode's buffer (2 ml) containing 10 mM HEPES buffer at pH 7.4, 0.1 mM aspirin, 0.2 units/ml apyrase, and 2 mM EGTA. [32P]Na3PO4 (0.2 mCi) was added, and the platelets were gently mixed for 2 h at room temperature. At the end of the incubation period, the labeled platelets were pelleted at 5,000 × g for 3 min and resuspended in Tyrode's buffer containing the same added components used for 32P labeling except that EGTA was omitted when prostaglandin E1 (5 µM) or forskolin (1 mM) was added. After a 2-min incubation period at room temperature, the platelets were pelleted by centrifugation at 5,000 × g for 3 min and were solubilized with 250 µl of 0.4% Triton X-100 (w/v) containing 10 mM sodium orthovanadate, 10 mM EGTA, 2 mM phenylmethylsulfonyl fluoride, and 10 µg/ml of antipain, leupeptin, and pepstatin A. All subsequent steps were carried out at 4 °C with incubation on a rotary mixing device. After incubation for 1 h, insoluble material was removed by centrifugation at 16,000 × g for 5 min. To detect insoluble PMCA, the pellet was solubilized with 100 µl of 1% (w/v) SDS and diluted 4-fold with the above solubilization buffer. Nonspecific material was then removed from both the Triton-soluble and Triton-insoluble fractions by incubation for 30 min with 20 µg of mouse IgG and 50 µl of protein A-agarose. After removal of the protein A-agarose by centrifugation at 1,000 × g for 3 min, 40 µl of polyclonal anti-PMCA was added followed by 2 h of incubation. Protein A-agarose (50 µl) was then added and incubation continued for an additional 1.5 h. The protein A-PMCA complex was then collected by centrifugation for 3 min at 1,000 × g, followed by washing with 200 µl of solubilization buffer two times and centrifugation at 1,000 × g for 3 min. A final wash with 10 mM Tris/0.15 M NaCl was followed by addition of Laemmli sample dilution buffer (17) and incubation at 40 °C for 20 min. Samples were frozen at -20 °C or electrophoresed immediately. Detection of immunoprecipitated PMCA was accomplished by Western blotting as described below. The presence of 32P in immunoprecipitated PMCA was measured by phosphorimaging using a Molecular Dynamics model PSF PhosphorImager with ImageQuant version 3.3 software.

Western Blotting

Proteins were electrophoresed on 7.5% SDS-polyacrylamide gels (17) followed by electrophoretic transfer to nitrocellulose membranes at 100 V in Tris/glycine buffer at pH 8.3 containing 10% methanol. After transfer, the blots were washed in TTBS (30 mM Tris at pH 7.5, 150 mM NaCl, and 0.1% Tween 20 containing 1 mM sodium o-vanadate) and blocked at room temperature for 1 h in 5% (w/v) nonfat dry milk dissolved in this same TTBS solution. After washing two times in TTBS for 10 min, primary antibody dissolved at the appropriate dilution in TTBS containing 2% (w/v) bovine serum albumin was incubated with the membrane for 1 h. The membrane was washed one time for 15 min and four times for 5 min in TTBS and then exposed to the appropriate secondary antibody conjugated to horseradish peroxidase at a dilution of 1:20,000 for 1 h. The membranes were washed as described after the primary antibody incubation and exposed to chemiluminescence reagents (Amersham Corp.). Immunoreactive proteins were visualized by exposure of the membrane to x-ray film. PMCA was assessed using 5F10 (anti-PMCA), phosphotyrosine with PY20 (anti-phosphotyrosine), and specific isoforms of PMCA with anti-PMCA 1N, 2N, 3N, and 4N (9). In some experiments membranes were stripped of bound antibody and probed with a different primary antibody after incubation of the membrane in 62.5 mM Tris at pH 6.8, containing 2% (w/v) SDS and 100 mM 2-mercaptoethanol at 60 °C for 1 h. The blot was then washed with TTBS, and the Western blotting procedure was repeated. The mass of PMCA present was determined by Western blotting known amounts of pure erythrocyte PMCA on the same gel with unknowns followed by ECL-based immunodetection and quantitation of scans of the resulting lumigrams using SigmaGel software.

pp60src-dependent Phosphorylation and Ca2+-ATPase Activity Measurement

Purified erythrocyte PMCA was phosphorylated with pp60src (0.4 units pp60src/µg PMCA) at 30 °C in 5 mg/ml Triton X-100, 20 mM HEPES buffer at pH 7.2, 130 mM NaCl, 31 mM MgCl2, 0.1 mM ATP, 23 mM MnCl2, 2 mM sodium orthovanadate, 2 mM dithiothreitol, 2 mM EGTA, and 0.5 mg/ml phosphatidylcholine. Reactions were terminated by rapid chromatography (5 min) of 50-µl assay mixtures on 5-ml Sephadex G-50 columns equilibrated with 20 mM HEPES buffer, 3 mM MgCl2, 0.05 mM CaCl2, 130 mM NaCl, 2 mM EGTA, 2 mM dithiothreitol, 0.5 mg/ml Triton X-100, 5% (w/v) glycerol, and 0.5 mg/ml phosphatidylcholine to remove the sodium orthovanadate, which is an inhibitor of PMCA. The eluate was then assayed immediately for Ca2+-ATPase activity.

Erythrocyte and platelet plasma membranes were phosphorylated with 0.3 unit of pp60src/µg of membrane protein at room temperature in 10 mM Tris at pH 7.2, 20 mM MgCl2, 1 mM ATP, 0.2 mM sodium orthovanadate, 1 mM EGTA, 5 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, and 10 µg/ml of pepstatin A, antipain, and leupeptin. After the specified incubation time, vanadate was removed by centrifugation of the membranes (50 µl) in a Beckman Airfuge at 100,000 × g for 3 min, followed by resuspension in 100 µl of 10 mM Tris at pH 7.2, 5 mM dithiothreitol, and 10% (w/v) glycerol and recentrifugation for 3 min. The final washed pellet was resuspended in 50 µl of the centrifugation buffer and assayed immediately for Ca2+-ATPase activity.

32P Labeling and Isolation of pp60src-treated PMCA

Large scale labeling of purified erythrocyte PMCA was carried out to identify peptides containing phosphorylated tyrosine residues. Approximately 1 nmol of PMCA in 350 µl was labeled with 90 units of pp60src overnight at room temperature in 5 mg/ml Triton X-100, 20 mM HEPES buffer at pH 7.2, 130 mM NaCl, 31 mM MgCl2, 0.1 mM [gamma -32P]ATP at a specific activity of 2.5 µCi/nmol, 2.5 mM sodium orthovanadate, 2 mM dithiothreitol, 4 mM EGTA, and 0.5 mg/ml phosphatidylcholine. The labeled protein was applied to two SDS-polyacrylamide gels and electrophoresed according to Laemmli (17). The protein was then transferred to polyvinylidene difluoride membrane and stained briefly with 1% (w/v) aqueous Coomassie Blue to visualize the protein bands. The 140-kDa PMCA band was excised, dried, and stored at 4 °C.

Peptide Mapping and Sequencing

Peptide mapping and isolation by HPLC and automated sequence analyses were performed by the University of Kentucky Macromolecular Structure Analysis Facility as follows. 32P-Labeled PMCA bound to polyvinylidene difluoride was reduced with 7 mM dithiothreitol in 100 mM Tris at pH 8.5 for 1 h at 37 °C and then alkylated with 20 mM iodoacetamide at room temperature for 15 min in the dark. The reduced and alkylated protein was then digested with endoproteinase Lys-C in 100 mM Tris at pH 8.5 containing 1% (w/v) hydrogenated Triton X-100 and 10% CH3CN for 18 h at 37 °C. Released peptides were recovered in 40% (v/v) CH3CN at 37 °C for 30 min and then pooled with a wash of 0.05% (v/v) trifluoroacetic acid in 40% CH3CN. Reverse phase HPLC was performed on a Vydac C18 TP silica column with a pellicular C18 guard column. Absorbance was monitored at 214 nm with a mobile phase of 0.06% trifluoroacetic acid versus 0.054% trifluoroacetic acid in 70% CH3CN. Fractions were collected and counted without scintillation fluid. The appropriate fractions were pooled, applied to a precycled polybrene-coated glass fiber disc, and sequenced on an Applied Biosystems 477A sequencer with pulsed liquid phase chemistry and on-line PTH-derivative analysis.


RESULTS

Cyclic AMP-dependent Phosphorylation of Human Platelet PMCA

The work of Johnson et al. (4) demonstrated that an increase in platelet cAMP results in increased efflux of Ca2+, suggesting that the platelet PMCA is stimulated by cAMP-dependent phosphorylation. Neyses et al. (18) demonstrated that both sarcolemmal and erythrocyte PMCAs are phosphorylated by cAMP-dependent kinase resulting in stimulation of Ca2+-ATPase activity. Furthermore, Johnson and Haynes (19) showed that the Na-Ca2+ exchanger makes a negligible contribution to platelet calcium efflux at resting calcium concentrations, thus implying that the main effect of cAMP on platelet Ca2+ efflux is on PMCA. Consequently, we attempted to demonstrate that elevated cAMP leads to direct phosphorylation of PMCA in the platelet. The result of increasing cytosolic platelet cAMP with prostaglandin E1 or forskolin is shown in Fig. 1. Platelets were labeled with [32P]PO4 and then treated with forskolin or prostaglandin E1 prior to immunoprecipitation of PMCA. The results demonstrate that both agents cause a significant increase in incorporation of 32P, presumably resulting from cAMP-dependent phosphorylation. The effect of cAMP-dependent phosphorylation on Ca2+-ATPase activity was determined using purified platelet plasma membranes. Addition of protein phosphatase type 2A inhibited Ca2+-ATPase activity, and subsequent addition of the catalytic subunit of cAMP-dependent protein kinase resulted in 1.4-fold stimulation (Table I). These results confirm that platelet PMCA can be stimulated by cAMP-dependent phosphorylation and show that the protein in isolated membranes is partially phosphorylated. Studies presented in a later section demonstrate that PMCA1 and -4 are the two isoforms present in human platelets, and these are most likely PMCA1b and -4b.


Fig. 1. cAMP-mediated phosphorylation of platelet PMCA in vivo. Platelets (5 × 1010 cells/ml) were labeled with [32P]Na3PO4 and then treated with either forskolin (1 mM) or prostaglandin E1 (PGE1) (5 µM) for 5 min followed by immunoprecipitation as described under "Experimental Procedures." Forskolin and prostaglandin E1 were omitted in the control. The immunoprecipitates were submitted to SDS-polyacrylamide gel electrophoresis and blotted onto a nitrocellulose membrane. Radioactivity was detected by phosphorimaging, and then immunodetection was performed with anti-PMCA (5F10). PMCA was visualized by chemiluminescence.
[View Larger Version of this Image (19K GIF file)]

Table I. Effects of protein phosphatase 2A and cAMP-dependent kinase on platelet plasma membrane Ca2+-ATPase

Platelet plasma membranes were dephosphorylated with protein phosphatase 2A by preincubating membranes (70 µg of total protein) with 0.5 units of phosphatase in 10 mM TES buffer containing 1 mM dithiothreitol and 1 mM EDTA for 10 min at 30 °C prior to assaying for Ca2+-ATPase activity. When kinase treatment followed phosphatase treatment, 63 units of catalytic subunit of cAMP-dependent protein kinase were added to the membranes along with 0.1 mM ATP, 50 mM K3PO4, and 10 mM MgCl2, and the membranes were incubated for an additional 10 min.

Addition to membranes Specific activity

nmol/min/mg
No addition 12.6  ± 3.4 (n = 4)
Phosphatase 9.2  ± 0.9 (n = 4) p < 0.035a
Phosphatase then kinase 14.7  ± 1.1 (n = 4) p < 0.002b

a Phosphatase treated compared with untreated membranes.
b Phosphatase treated then kinase treated compared with phosphatase treated.

Tyrosine Phosphorylation of PMCA in Human Platelets in Vivo

During platelet activation, several proteins become specifically phosphorylated on tyrosine residues and associate with the platelet cytoskeleton. Since it is not known if platelet PMCA is associated with the cytoskeleton in resting platelets, it was necessary to look for PMCA in the cytoskeleton of resting platelets to determine a baseline for PMCA distribution between the membrane and cytoskeleton. The results shown in Fig. 2 demonstrate that PMCA is not associated with the cytoskeleton or other Triton X-100-insoluble material in the resting platelet. The band labeled "heavy chain" represents nonspecific binding of the primary or secondary antibodies to the heavy chain of IgG used to immunoprecipitate PMCA.


Fig. 2. Immunoprecipitation of PMCA from resting platelets. Platelets (1.5 ml of a suspension containing 5 × 1010 cells/ml) resuspended in Tyrode's buffer containing aspirin, apyrase, and prostaglandin E1 were solubilized with 0.4% Triton X-100 containing EGTA and protease inhibitors (see "Experimental Procedures"). After centrifugation at 16,000 × g, the Triton-insoluble material was solubilized in 1% SDS and diluted 4-fold with Triton solubilization buffer. PMCA was then immunoprecipitated with polyclonal anti-PMCA, and the protein A-PMCA complex was solubilized in SDS and electrophoresed; Western blotting was then performed with anti-PMCA 5F10. Antibody binding was visualized by chemiluminescence. PMCA indicates a lane loaded with 1.0 µg of purified erythrocyte PMCA, and cytoskeleton indicates the lane containing the immunoprecipitate from the Triton-insoluble fraction.
[View Larger Version of this Image (57K GIF file)]

Immunoprecipitation followed by immunoblotting with anti-phosphotyrosine (PY20) permitted the detection of PMCA tyrosine phosphorylation as a function of the platelet activation state. Activation of platelets with thrombin or spontaneous aggregation (labeled ADP because of the involvement of secreted ADP in spontaneous aggregation) resulted in tyrosine phosphorylation of PMCA as shown in Fig. 3. Considerably less tyrosine phosphorylation was observed in platelets treated with prostaglandin E1 to block spontaneous aggregation. Furthermore, the tyrosine kinase inhibitor genistein totally eliminated all tyrosine phosphorylation and prevented spontaneous aggregation. Stripping and reprobing the immunoblot membrane with monoclonal 5F10 anti-PMCA showed that differences in tyrosine phosphorylation were not simply due to the difference in amounts of PMCA applied to the SDS gel.


Fig. 3. Tyrosine phosphorylation of platelet PMCA. Platelets (1.5 ml of a suspension containing 5 × 1010 cells/ml) were resuspended in Tyrode's buffer containing thrombin (5 units/ml), no additions (labeled ADP due to the occurrence of spontaneous aggregation accompanied by secretion of ADP), genistein (0.5 mM), or prostaglandin E1 (1.0 µM) for 5 min at room temperature, and immunoprecipitation was then carried out as described in Fig. 2. Western blots were first probed with anti-phosphotyrosine (PY20), stripped, and reprobed with anti-PMCA (5F10).
[View Larger Version of this Image (32K GIF file)]

The results in Fig. 3 indicate that PMCA is directly phosphorylated on tyrosine residues during platelet activation. The kinetics of tyrosine phosphorylation of PMCA after addition of thrombin are shown in Fig. 4. Increased phosphorylation induced by thrombin was seen at 5 min, and by 10 min the level of phosphorylation returned nearly to the baseline level observed in the absence of thrombin.


Fig. 4. Time course of tyrosine phosphorylation of PMCA in thrombin-stimulated platelets. Platelets were pelleted and resuspended in Tyrode's buffer at 5.5 × 1010 cells/ml. Either thrombin (1 unit/ml) or EGTA (2 mM) was added (0 min), and aliquots were removed and solubilized with Triton X-100 at the indicated times as described under "Experimental Procedures." Immunoprecipitated PMCA was electrophoresed and blotted onto a nitrocellulose membrane. The membrane was first probed with anti-phosphotyrosine followed by chemiluminescence detection. The nitrocellulose membrane was then stripped with SDS and reprobed with anti-PMCA (5F10). Film was scanned and analyzed using SigmaGel. The peak area of the phosphotyrosine signal was then divided by the area determined for PMCA to yield the phosphotyrosine/PMCA ratio for each time point. The ratio presented on the y axis does not reflect the absolute ratio of phosphotyrosine to PMCA but only the ratio of the intensities of PY20- and 5F10-stained PMCA bands.
[View Larger Version of this Image (12K GIF file)]

Effects of in Vitro Phosphorylation of PMCA with Purified pp60src

Phosphorylation of PMCA by both cAMP-dependent protein kinase and protein kinase C stimulates PMCA activity (3). The effect of tyrosine phosphorylation on platelet PMCA was assessed by the addition of pp60src to purified platelet plasma membranes. This is a reasonable model for platelet-tyrosine phosphorylation since pp60src is the most abundant non-receptor tyrosine kinase in the platelet and is associated with the plasma membrane (21). The results shown in Fig. 5A demonstrate that platelet plasma membrane Ca2+-ATPase activity was rapidly inhibited by tyrosine phosphorylation. Within 15 min, the platelet PMCA activity was reduced by 75% of the initial activity (square ). Furthermore, membranes isolated from thrombin-treated platelets also exhibited inhibition of PMCA activity. Membranes were isolated from thrombin- (5 min) or EGTA-treated platelets by sonication and differential centrifugation and assayed for Ca2+-ATPase activity in the presence of 200 nM thapsigargin to inhibit internal membrane Ca2+-ATPase activity (22). Membranes obtained from EGTA-treated platelets exhibited PMCA activity of 4.1 ± 0.9 nmol/min/mg protein, while thrombin treatment yielded membranes with an activity of 2.5 ± 0.5, an inhibition of 40% with p < 0.021 (two-tail Student's t test).


Fig. 5. Inhibition of platelet and erythrocyte PMCA by tyrosine phosphorylation with pp60src. Purified erythrocyte PMCA, erythrocyte ghosts, and platelet plasma membranes were incubated with pp60src for the indicated times as described under "Experimental Procedures." Panel A, the fraction of Ca2+-ATPase activity remaining was determined by comparing the Ca2+-ATPase in pp60src-treated assays with controls prepared in identical fashion except that the kinase was omitted. Points are the averages of two to four measurements. Open circles indicate activity of purified erythrocyte PMCA in the absence of kinase, while filled circles show the result of addition of 0.4 units of kinase/µg of PMCA. Filled triangles indicate the effect of the addition of 0.3 units of kinase/µg of protein on erythrocyte ghosts, while open squares show the results of the addition of 0.3 units of kinase/µg of protein to platelet plasma membranes. The initial activity of purified erythrocyte PMCA was 3.8 µmol of ATP hydrolyzed/min/mg; activity of erythrocyte ghosts was 20 nmol/min/mg; and activity of platelet plasma membranes was 16 nmol/min/mg. Panel B shows the results of CaM affinity chromatography of the purified human erythrocyte PMCA preparation after extensive pp60src phosphorylation. Purified erythrocyte PMCA was incubated ±pp60src for 6 h under the conditions described in panel A above. The resulting reaction mixtures were treated with CaM-Sepharose by batch absorption using detergent-containing buffers identical to those used for PMCA purification as described under "Experimental Procedures." Samples containing 50 ng of PMCA were applied to 50 µl of CaM-Sepharose beads (100 µl of 1:1 slurry) in 1 mM Ca2+ purification buffer in a microcentrifuge tube and mixed gently for 1 h at 4 °C; then the bound and unbound fractions were separated by centrifugation in a microcentrifuge. After three washes with 100 µl of the calcium-containing buffer, bound proteins were eluted by treating the resin pellet with 50 µl of the purification elution buffer containing 2 mM EGTA for 1 h. The bound and unbound fractions were again isolated by centrifugation, and 20 µl of the supernatant fraction were resolved by SDS-polyacrylamide gel electrophoresis and immunoblotted with anti-phosphotyrosine (PY20, left panel) or anti-PMCA (5F10, right panel) as described under "Experimental Procedures."
[View Larger Version of this Image (29K GIF file)]

The results shown in Fig. 5A also demonstrate that PMCA in erythrocyte ghosts is rapidly inhibited by pp60src, reaching 60% inhibition by 1 h (black-triangle). Purified PMCA is also inhibited to a similar extent, but the process is much slower, possibly because of the lack of membranes or the presence of the detergent Triton X-100 (bullet ). The inhibition of activity seen, at least in the case of the purified PMCA, was not due to loss of ability to interact with calmodulin. As shown in Fig. 5B, the phosphotyrosine-containing form of PMCA bound as well, if not better, to CaM-Sepharose in the presence of Ca2+ as compared with the dephospho- form.

Fig. 6 shows that there is excellent correlation between tyrosine phosphorylation of purified erythrocyte PMCA and the inhibition of Ca2+-ATPase activity seen in Fig. 5. pp60src is autophosphorylated, and this precedes tyrosine phosphorylation of PMCA as expected (23). The band below PMCA has a molecular mass of approximately 90 kDa and is a proteolysis product of PMCA as shown by Western blotting with the anti-PMCA antibody (data not shown).


Fig. 6. Tyrosine phosphorylation of purified erythrocyte PMCA by pp60src. Purified erythrocyte PMCA was treated with pp60src at 30 °C. At the indicated times, an aliquot was removed (2.7 µg of PMCA) and solubilized in SDS. In the lanes labeled src alone, conditions and procedures were identical except no PMCA was included. The solubilized samples were submitted to SDS-polyacrylamide gel electrophoresis, and Western blotting was carried out with anti-phosphotyrosine (PY20).
[View Larger Version of this Image (69K GIF file)]

PMCA Isoforms in Human Platelets

Since purified erythrocyte PMCA is an excellent source for identification of phosphorylated tyrosine residues, it is essential to determine the isoform content of the platelet. Erythrocytes contain the two housekeeping forms of PMCA, 1 and 4. Probing purified platelet membranes with antibodies specific for the N termini of isoforms 1-4 (9), we show in Fig. 7 that platelets contain the same two isoforms as erythrocytes and lack PMCA2 and -3. Thus erythrocyte PMCA is an excellent model for determination of tyrosine phosphorylation sites. More detailed analysis of the mRNA expression of different PMCA isoforms in platelets and megakaryocytes using reverse transcription-polymerase chain reaction analysis (20) have given similar results and found the isoforms to be PMCA1b and -4b.2


Fig. 7. PMCA isoform content of platelet plasma membranes. Equal amounts of platelet plasma membranes (75 µg/lane) were loaded onto four lanes of an SDS-polyacrylamide gel; after electrophoresis, they were blotted onto nitrocellulose. The lanes were then cut into four strips and blotted individually with anti-PMCA 1N, 2N, 3N, or 4N. Only anti-PMCA 1N and 4N labeled platelet PMCA.
[View Larger Version of this Image (69K GIF file)]

Site(s) of Phosphorylation of PMCA by pp60src

To determine the tyrosine residues phosphorylated in PMCA, we labeled purified erythrocyte PMCA in vitro with 32P by phosphorylation with pp60src and [gamma -32P]ATP. After overnight labeling of 1 nmol of purified erythrocyte PMCA with 90 units of pp60src, the phosphorylated protein was purified by SDS-polyacrylamide gel electrophoresis and blotted onto polyvinylidene difluoride membrane as described under "Experimental Procedures." The labeled protein was reduced and alkylated and then digested with the endoproteinase Lys-C. HPLC analysis of the generated peptides is shown in Fig. 8. A single radioactive peptide peak containing approximately 26 pmol of 32P was detected, and this peptide was isolated and subjected to automated Edman sequence analysis as described under "Experimental Procedures." Extensive losses of 32P-containing material occurred during sample processing, so the recovery of this peptide was not quantitative. The sequence obtained, shown in Table II, corresponded to residues 1162-1177 in hPMCA4b (24), a splice variant of PMCA4 found in most tissues (20). The tyrosyl residue at position 1176 was not detected, indicating that it probably is phosphorylated. Tyr-1176 is preceded by a glutamic acid residue 4 residues N-terminal to the tyrosine as has been observed for many tyrosine phosphorylation sites but is missing the usual lysine or arginine residue 7 residues N-terminal to the tyrosine (25). Since this tyrosine phosphorylation site is C-terminal to the calmodulin binding domain, it might be expected to affect calmodulin binding. However, treatment of purified PMCA with pp60src did not affect the ability of PMCA to bind to calmodulin-Sepharose as shown in Fig. 5B.


Fig. 8. HPLC separation of peptides obtained from pp60src-labeled erythrocyte PMCA. PMCA was phosphorylated with [gamma -32P]ATP and pp60src followed by SDS-polyacrylamide gel purification and Lys-C digestion as described under "Experimental Procedures." Fractions were counted without scintillation fluid to identify the labeled peptides.
[View Larger Version of this Image (22K GIF file)]

Table II. Sequence of phosphorylated PMCA peptide

The radioactive peak eluting from 66-69 min (Fig. 8) containing 26 pmol of 32P was sequenced. Yields of the indicated residues for the first five cycles were 10.2, 17.1, 8.0, 8.1, and 7.4 pmol, respectively.

hPMCA4b 1161FGTRVLLLDGEVTPYA1178
Labeled platelet peptide    FGTRVLLLD#EVTP#A
Tyrosine phosphorylation motifa           KXXEXXXY
          R  DE

a Cooper et al. (23). #, no detected amino acid in that cycle. X, any amino acid.

The affinity of calmodulin for the phosphotyrosine form of PMCA was further assessed by measuring calmodulin stimulation of calmodulin-free erythrocyte ghosts after treatment with pp60src. If tyrosine phosphorylation inhibited PMCA by lowering the affinity for calmodulin, then a higher concentration of calmodulin would be expected to alleviate the inhibition. Membranes were treated with pp60src or pp60src storage buffer (25 mM HEPES buffer, pH 7.0, containing 50% (v/v) glycerol, 0.1% (v/v) Nonidet P-40, and 1.0 mM dithiothreitol) for 30 min at 30 °C as described for Fig. 5A, and PMCA activity was assayed in the absence or presence of 50, 250, 1000, and 3000 µM calmodulin. Calmodulin stimulated PMCA activity 2-fold compared with the activity in the absence of added calmodulin at all four calmodulin concentrations. The inhibition caused by pp60src (based on the comparison between pp60src-treated membranes and the buffer only control) was the same (30%) in the absence of calmodulin as for the four calmodulin concentrations used in this experiment. Thus the inhibition of PMCA activity could not be overcome by a 60-fold increase in calmodulin concentration. This result strengthens the conclusion reached from Fig. 5B that tyrosine phosphorylation does not affect calmodulin binding to PMCA.

The stoichiometry of phosphorylation was determined by comparison of 32P incorporation with the mass of PMCA calculated by scanning of Coomassie Blue-stained gels as described under "Experimental Procedures." Labeling for 6 h at room temperature yielded incorporation of 5 pmol of phosphate into 11 pmol of purified PMCA. Since the erythrocyte contains approximately 70% PMCA4 (9, 18) and we only detected a single labeled peptide in PMCA4, this corresponds to 65% incorporation into tyrosine 1176. This result is in reasonable agreement with the 72% inhibition of PMCA activity observed after 6 h of phosphorylation (Fig. 4). It should be noted that analysis of trypsin digests of pp60src-phosphorylated human erythrocyte PMCA on 40% acrylamide gels designed for phosphopeptides (26) also yielded a single radiolabeled peptide band (data not shown).


DISCUSSION

PMCA plays a central role in platelet Ca2+ metabolism. PMCA and the Na+-Ca2+ exchanger are the only proteins capable of removing Ca2+ from the cell. Since published data indicate that the Na+-Ca2+ exchanger does not contribute significantly to platelet Ca2+ efflux at resting Ca2+ levels (4), then PMCA must have the primary role of maintaining the resting level of cytoplasmic Ca2+. Regulation of the activity of PMCAs is an integral and essential component of calcium-signaling pathways. To date, all reported regulatory signals (i.e. calmodulin binding and reversible phosphorylation) have been reported to activate pump activity. However, unlike most other calmodulin-regulated enzymes, PMCAs have a substantial basal unstimulated activity. Thus, activation results in increases in activity of 2-10-fold at most. We report here that tyrosine phosphorylation of PMCA4 in platelets leads to substantial, if not complete, inhibition of its Ca2+-ATPase activity below the basal level. This would allow for a more rapid increase in cellular calcium during initial response phase in the platelet as well as to a much greater differential effect of activators compared with the inhibited state. Because of the important role of Ca2+ in platelet activation, these changes in PMCA activity could have significant effects on intracellular Ca2+ levels and platelet function. The present work clearly supports the central role of PMCA in platelet Ca2+ metabolism since we demonstrated that platelet antagonists stimulate the pump, whereas agonists inhibit its activity. Thus kinases provide positive feedback control of Ca2+ metabolism during activation and inhibition of platelet function.

We demonstrated directly that PMCA is phosphorylated when platelets are treated with agents known to elevate cAMP (Fig. 1). Furthermore, treatment of isolated plasma membranes with the catalytic subunit of cAMP-dependent kinase stimulated Ca2+-ATPase activity if endogenous phosphates were first removed with protein phosphatase type 2A. These results directly confirm the conclusions of Johnson et al. (4) that cAMP stimulates PMCA in the platelet. This stimulation should lead to lower intracellular Ca2+ and a decreased propensity for activation. In addition, it appears that PMCA in membranes isolated from resting platelets is partially phosphorylated as demonstrated by the loss of activity upon dephosphorylation.

Activation of platelets by a strong agonist such as thrombin results in tyrosine phosphorylation of several proteins followed by their incorporation into the Triton-insoluble cytoskeleton. This process is at least partially regulated by Ca2+ (6, 7, 27). The concomitant tyrosine phosphorylation of PMCA would aid in coordination of this process. Here we show for the first time that platelet PMCA is phosphorylated on a tyrosine residue during platelet activation and that this phenomenon is inhibited by the tyrosine kinase inhibitor genistein (Fig. 3). This phosphorylation does not appear to be accompanied by association of PMCA with the Triton-insoluble cytoskeleton since the ATPase is not associated with the cytoskeleton in resting platelets (Fig. 2) and does not decrease in the Triton-soluble fraction after activation (Fig. 3). In addition the kinetics of PMCA phosphorylation are quite similar to that reported for pp125FAK (7).

Even more important is the demonstration that tyrosine phosphorylation inhibits PMCA. Addition of pp60src inhibited PMCA activity in purified platelet membranes 75% in 15 min at room temperature. Furthermore, platelet activation by thrombin decreased PMCA activity in isolated membranes by 40%. The results of pp60src treatment of purified erythrocyte PMCA indicate that the kinase preparation is not contaminated with proteinases, since there is no indication of degradation of PMCA (Fig. 6). Although the rate of kinase-mediated inhibition of PMCA is much slower in the purified membrane-free preparation, the final level of inhibition is the same in erythrocyte ghosts as the purified preparation. Although these results do not prove that pp60src is the actual kinase responsible for PMCA phosphorylation during activation, this kinase is the most abundant non-receptor tyrosine kinase in the platelet and is associated with the plasma membrane in resting platelets (21).

Although a potential pp60src consensus tyrosine phosphorylation site (25) is present in all PMCA1 isoforms at Tyr-596, our tryptic phosphopeptide mapping and sequence analyses yielded no evidence that this isoform was phosphorylated pp60src in vitro. PMCA4 isoforms lack a consensus tyrosine phosphorylation site, but PMCA4b was nevertheless phosphorylated on Tyr-1176, a site lacking the basic residue 7 residues N-terminal of the tyrosine but exhibiting the predicted acidic residue, glutamic acid, 4 residues N-terminal to Tyr-1175. It is possible that PMCA1 was phosphorylated on the predicted tyrosine and that it was missed because PMCA1 is only 30% or less of the total PMCA of the erythrocyte. In platelet plasma membranes, PMCA activity was inhibited approximately 80% by tyrosine phosphorylation, even though PMCA4 is only approximately 50% of the total PMCA (Fig. 6). This suggests that both isoforms may be phosphorylated in platelet membranes. The results presented here are in agreement with the recent observation that the tyrosine phosphatase inhibitor, phenylarsine oxide, increases Ca2+ influx at the plasma membrane in endothelial cells (28) at phenylarsine oxide concentrations that we have subsequently shown have no effect on the activity of PMCA itself.3 This suggests that control of Ca2+ influx by transient activation or inhibition of PMCA through reversible phosphorylation by threonine, serine, and tyrosine kinases may be a general regulatory mechanism in many diverse cell types.


FOOTNOTES

*   This work was supported in part by National Institutes of Health Grant NS21868 (to T. C. V.) and grants from the Kentucky Affiliate of the American Heart Association and the Jewish Hospital Foundation (to W. L. D.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
   To whom correspondence should be addressed: Dept. of Biochemistry, 209 Combs Bldg., University of Kentucky Medical Center, Lexington, KY 40536-0084. Tel.: 606-257-1347; Fax: 606-257-8940; E-mail: vanaman{at}pop.uky.edu.
1   The abbreviations used are: PMCA, plasma membrane Ca2+-ATPase; TES, 2-{[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]amino}ethanesulfonic acid; HPLC, high pressure liquid chromatography.
2   P. C. Brandt and W. L. Dean, unpublished observations.
3   T. C. Vanaman, unpublished observations.

REFERENCES

  1. Heemskerk, J. W. M., and Sage, S. O. (1994) Platelets (Edinb) 5, 295-316
  2. Haynes, D. H. (1993) Platelets (Edinb.) 4, 231-242
  3. Carafoli, E. (1994) FASEB J. 8, 993-1002 [Abstract/Free Full Text]
  4. Johnson, J. S., Nied, L. E., and Haynes, D. H. (1992) Biochim. Biophys. Acta 1105, 19-28 [Medline] [Order article via Infotrieve]
  5. Oda, A., Druker, B. J., Smith, M., and Salzman, E. W. (1992) J. Biol. Chem. 267, 20075-20081 [Abstract/Free Full Text]
  6. Shattil, S. J., Haimovich, B., Cunningham, M., Lipfert, L., Parson, J. T., Ginsberg, M. H., and Brugge, J. S. (1994) J. Biol. Chem. 269, 14738-14745 [Abstract/Free Full Text]
  7. Dash, D., Aepfelbacher, M., and Siess, W. (1995) FEBS Lett. 363, 231-234 [CrossRef][Medline] [Order article via Infotrieve]
  8. Jayaraman, T., Ondrias, K., Ondriasova, E., and Marks, A. R. (1996) Science 272, 1492-1494 [Abstract]
  9. Stauffer, T. P., Guerinin, D., and Carafoli, E. (1995) J. Biol. Chem. 270, 12184-12190 [Abstract/Free Full Text]
  10. Brandt, P., Zurini, M., Neve, R. L., Rhoads, R. E., and Vanaman, T. C. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 2914-2918 [Abstract]
  11. Borke, J. L., Caride, A., Verma, A. K., Penniston, J. T., and Kumar, R. (1989) Am. J. Physiol. 257, F842-F849 [Abstract/Free Full Text]
  12. Harmon, J. T., Greco, N. J., and Jamieson, G. A. (1992) Methods Enzymol. 215, 32-36 [Medline] [Order article via Infotrieve]
  13. Dean, W. L., and Quinton, T. M. (1995) Cell Calcium 17, 65-70 [Medline] [Order article via Infotrieve]
  14. Steck, T. L., and Kant, J. A. (1974) Methods Enzymol. 31, 172-180 [CrossRef][Medline] [Order article via Infotrieve]
  15. Niggli, V., Adunyah, E. S., Penniston, J. T., and Carafoli, E. (1981) J. Biol. Chem. 256, 395-401 [Abstract/Free Full Text]
  16. Dean, W. L. (1988) Methods Enzymol. 157, 351-360 [Medline] [Order article via Infotrieve]
  17. Laemmli, U. K. (1970) Nature 227, 680-685 [Medline] [Order article via Infotrieve]
  18. Neyses, L., Reinlib, L., and Carafoli, E. (1985) J. Biol. Chem. 260, 10283-10287 [Abstract/Free Full Text]
  19. Johnson, J. S., and Haynes, D. H. (1988) J. Membr. Biol. 104, 147-163 [Medline] [Order article via Infotrieve]
  20. Brandt, P., Neve, R. L., Kammesheidt, A., Rhoads, R. E., and Vanaman, T. C. (1992) J. Biol. Chem. 267, 4376-4385 [Abstract/Free Full Text]
  21. Golden, A., Nemeth, S. P., and Brugge, J. S. (1986) Proc. Natl. Acad. Sci. U. S. A. 83, 852-856 [Abstract]
  22. Dean, W. L., Pope, J. E., Brier, M. E., and Aronoff, G. R. (1994) Hypertension 23, 31-37 [Abstract]
  23. Cooper, J. A., and MacAuley, A. (1988) Proc. Natl. Acad. Sci. U. S. A. 85, 4232-4236 [Abstract]
  24. Strehler, E. E., James, P., Fischer, R., Heim, R., Vorherr, T., Filoteo, A. G., Penniston, J. T., and Carafoli, E. (1990) J. Biol. Chem. 265, 2835-2842 [Abstract/Free Full Text]
  25. Cooper, J. A., Esch, F. S., Taylor, S. S., and Hunter, T. (1984) J. Biol. Chem. 259, 7835-7841 [Abstract/Free Full Text]
  26. West, M. H. P., Wu, R. S., and Bonner, W. M. (1984) Electrophoresis 5, 133-138
  27. Negrescu, E. V., de Quintana, K. L., and Siess, W. (1995) J. Biol. Chem. 270, 1057-1061 [Abstract/Free Full Text]
  28. Fleming, I., Fisslthaler, B., and Busse, R. (1996) J. Biol. Chem. 271, 11009-11015 [Abstract/Free Full Text]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.