Protein kinase C–catalyzed phosphorylation of an inhibitory phosphoprotein of myosin phosphatase is involved in human platelet secretion

Phosphorylation of the 20-kd myosin light chain (MLC₂₀) at Ser-19 is thought to be one of the primary steps in the activation of actomyosin contractile events, which involve platelet shape change and secretion.1-7 The extent of MLC₂₀ phosphorylation is regulated not only by protein kinases, such as Ca⁺⁺-dependent MLC kinase8and Ca⁺⁺-independent Rho-kinase,9 but also by myosin phosphatase.10,11 Platelet shape change, the earliest response induced by agonists, is a prerequisite for full platelet activation, including secretion and aggregation.1,2 Platelet shape change is regulated by MLC₂₀ phosphorylation either through RhoA/Rho-kinase activation or through Ca⁺⁺-dependent MLC kinase activation.4 Activation of protein kinase C (PKC) is not likely to be involved in platelet shape change.1Inhibition of myosin phosphatase activity appears to be an important mechanism for increasing the Ca⁺⁺ sensitivity of MLC₂₀ phosphorylation and platelet secretion, as is the case for smooth muscle.10,11 Myosin phosphatase is composed of 3 subunits—a catalytic subunit of the type 1 protein phosphatase δ isoform (PP1δ) and 2 noncatalytic subunits (a 130-kd myosin-binding subunit (MBS) and a 20-kd subunit).12-14Different mechanisms have been proposed to account for the inhibition of myosin phosphatase,10,11 including dissociation of subunits of the holoenzyme by arachidonate,15,16phosphorylation of MBS by RhoA/Rho-kinase,17,18 or phosphorylation of CPI-17 through PKC activation.19-23 The regulation of myosin phosphatase occurs through phosphorylation of its regulatory subunit, the 130-kd MBS, by Rho-kinase,18 a downstream target of the small GTPase RhoA.24Phosphorylation of MBS by Rho-kinase inhibits myosin phosphatase activity,18 and we reported that Thr-695 on chicken MBS (Thr-696 in human MBS) is the major Rho-kinase phosphorylation site involved in the inhibition of phosphatase activity.25 With use of an antibody that specifically recognizes MBS phosphorylated at Thr-695, we have shown that phosphorylation of Thr-695 is increased in Swiss 3T3 cells after stimulation by lysophosphatidic acid, which is thought to activate the RhoA pathway.25 Treatment of intact platelets with agonists, including thrombin, thromboxane A₂ (TXA₂) mimetic STA₂, and epinephrine, leads to rapid phosphorylation of MBS and to inactivation of myosin phosphatase.26 27 However, it remains to be clarified whether phosphorylation of MBS at Thr-696 is increased in intact human platelets in response to various agonists.

Another possible candidate for the inhibition of myosin phosphatase in platelets is PKC, an event that may be independent of the RhoA/Rho-kinase pathway. In permeabilized platelets, GTPγS and phorbol ester enhanced Ca⁺⁺ sensitivity of secretion without inositol phospholipid hydrolysis,7 suggesting that PKC activation contributes to Ca⁺⁺ sensitivity of platelet secretion. It has been suggested that PKC inhibits myosin phosphatase in smooth muscle by the phosphorylation of CPI-17, a 17-kd PKC-potentiated inhibitory protein of myosin phosphatase. CPI-17 is a phosphorylation-dependent inhibitory protein for myosin phosphatase selectively expressed in smooth muscle.19-21Phosphorylation of CPI-17 at Thr-38 by PKC enhances the inhibitory potency with respect to myosin phosphatase, approximately 1000-fold in vitro,20,21 and the phosphorylated CPI-17 induces contraction in permeabilized smooth muscle.22,23,28 There has been no evidence that CPI-17 or its homologous protein is present and is involved in the in vivo regulation of myosin phosphatase in human platelets. Moreover, the relative importance of the Rho-kinase–mediated pathway and the PKC-mediated pathway for regulating myosin phosphatase activity in platelet secretion has remained to be elucidated. We report here that platelet CPI has a slightly higher apparent molecular mass than does smooth muscle CPI-17.19 We also provide evidence that phosphorylation of platelet CPI by PKC regulates myosin phosphatase activity in vitro and in vivo; this is in addition to the inhibition that results from MBS phosphorylation at Thr-696 by Rho-kinase.

Y-27632 [(+)-(R)-trans-4-(1-aminoethyl)-N-(4-pynidyl) cyclohexanecarboxamide dihydrochloride, monohydrade] was a kind gift from WelFide (Osaka, Japan). Other drugs and suppliers were as follows: GF109203X (2-[1-(3-dimethylaminopropyl)-1H-indol-3-yl]-3-(1H-indol-3-yl)-maleimide) and ionomycin from Calbiochem-Novabiochem (La Jolla, CA); STA₂ [9, 11-epithio-11, 12-methano-thromboxane A₂] from Ono Pharmaceutical (Osaka, Japan); fura 2-am from Dojindo Laboratory (Kumamoto, Japan); phorbol 12-myristate 13-acetate (PMA) from Sigma Chemical (St Louis, MO); and adenosine triphosphate (ATP)γS from Boehringer Mannheim GmbH (Mannheim, Germany); [γ-³²P]ATP (111 τBq/mmol) from NEN Life Science Products (Boston, MA). Polyclonal antibodies against the PP1δ catalytic subunit and 130 kd MBS were raised with synthesized fragments of respective proteins and used after purification.26,29 Polyclonal anti–CPI-17 antibody raised against recombinant CPI-17 was affinity-purified.30Polyclonal antimyosin light chain (anti-MLC) antibody was raised against mixed light chains of chicken gizzard myosin. These anti-MLC antibodies recognized both the 20-kd and the 17-kd light chains of human platelet myosin. A monoclonal antibody against MLC phosphorylated at Ser-19 (anti–pSer19-MLC₂₀)31 was kindly provided by Y. Sasaki (Kitasato University, Tokyo, Japan). Polyclonal rabbit antibodies against MBS phosphorylated at Thr 695 (anti–pThr695-MBS)25 and CPI-17 phosphorylated at Thr-38 (anti–pThr38–CPI-17)30 were generated, absorbed onto the respective unphosphorylated peptide resins, and affinity-purified. PKC (a mixture of α- and β-isoforms) was prepared to apparent homogeneity from fresh platelets, as described.32 CPI-17 was purified from the porcine aorta, as described.30 

We estimated the amount of PKC-potentiated inhibitory phosphoprotein of myosin phosphatase (CPI) in human platelets by immunoblot analysis, using an anti–CPI-17 antibody. Immunoblot analysis was carried out as described.27,28 Quantitative estimation of the levels of CPI in human platelets was performed densitometrically, using a Personal Scanning Imager (Amersham Pharmacia Biotech AB, Uppsala, Sweden) in conjugation with the Imagequant program run on a Dell Personal Computer (Austin, TX) by scanning the immunoreactive band after immediately photographing the visualized band.29 The signal was then compared with that of known amounts of purified porcine smooth muscle CPI-17.

Platelet suspension (10⁹/mL) was dissolved in ice-cold lysis buffer (1% NP-40, 20 mM Tris-HCl, pH 7.5, 0.15 M NaCl, 2 mM EDTA, 2 mM sodium vanadate, and an appropriate amount of protease inhibitor cocktail tablets [Boehringer Mannheim GmbH]). Immunoprecipitation of platelet lysates was performed after clarification of the samples by centrifugation at 15 000gfor 10 minutes, and the soluble fraction, precleared with protein A–Sepharose CL-4B (Amersham Pharmacia Biotech AB), was incubated with 30 μL anti–CPI-17 antibody at 4°C for 1 hour.27 28The immune complex was precipitated by adding protein A–Sepharose CL-4B (Amersham Pharmacia Biotech AB), incubating it for an additional 1 hour at 4°C, and washing the beads 3 times with lysis buffer. These preparations were used to detect the in vitro phosphorylation of CPI and for assay of the phosphatase activity.

Phosphorylation of the anti–CPI-17 immunoprecipitate was carried out using purified platelet PKC in the presence of 50 μM ATP with 10 mM MgCl₂, 0.5 mM CaCl₂, 50 μg/mL phosphatidylserine, 12.5 μg/mL diolein, 0.5 μM okadaic acid (Wako Pure Chemicals, Osaka, Japan), and 20 mM Tris-HCl, pH 7.5, at 30°C for 2 minutes.32 The kinase reaction was initiated by adding ATP and terminated by adding sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer. Samples were subjected to SDS-PAGE, and CPI phosphorylation was analyzed by immunoblotting, using an anti–pThr38–CPI-17 antibody.

To analyze phosphatase activities, thiophosphorylation of the anti–CPI-17 immunoprecipitate was carried out using platelet PKC in the presence of 1 mM ATPγS with 10 mM MgCl₂, 0.5 mM CaCl₂, 50 μg/mL phosphatidylserine, 12.5 μg/mL diolein, and 20 mM Tris-HCl, pH 7.5, at 30°C for 60 minutes. The kinase reaction initiated by adding the ATPγS solution was terminated by adding 12 mM ethyleneglycotetraacetic acid (EGTA) and 2 mM EDTA; then phosphatase activity was immediately measured. Activities of PP1, PP2A, PP2B, and PP2C were determined using [³²P]-phosphorylated MLC of chicken gizzards as a substrate, as described elsewhere.26 33-35 PP1 and PP2A activities were measured in a reaction mixture (50μL) containing 20 mM Tris-HCl, pH 7.4, 50 mM NaCl, 0.1 mM EGTA, 2 mM sodium vanadate ± 10 nM okadaic acid ± heat-stable phosphatase inhibitor-2, 4 mM phenylmethylsulfonyl fluoride, and 200 μg/mL leupeptin. PP2A was completely inhibited with 10 nM okadaic acid, whereas PP1 at this concentration was little affected. PP1 can be taken as the activity that is sensitive to the inhibitor-2, and PP2A can be taken as the activity that is blocked by 10 nM okadaic acid. PP2B and PP2C activities were determined as activity in the presence of 0.5 mM CaCl₂, 5 mM MgCl₂, 0.5 μM okadaic acid, 1 μM calmodulin, and ± 0.15 mM trifluoperazine. PP1 and PP2A are completely inhibited by 0.5 μM okadaic acid; therefore, this activity was taken as the total activity of PP2B and PP2C. Trifluoperazine (0.15 mM) was added to completely inhibit PP2B, but PP2C was unaffected by trifluoperazine, even at 0.15 mM. PP2B activity was determined by subtracting PP2C (the activity in the presence of 0.15 mM trifluoperazine) from the total activity. Assays were made in duplicate, and the phosphatase activity was indicated as pmol ³²Pi released per minute in cell samples.

Venous blood was freshly drawn from a healthy donor who had not taken any drugs for at least 2 weeks. Human platelets (10⁹/mL) were resuspended in Ca⁺⁺-free Tyrode-HEPES buffer that contained a final concentration of 0.14 M NaCl, 2.7 mM KCl, 1 mM MgCl₂, 0.1% d-glucose, 3.75 mM NaH₂PO₄, and 15 mM HEPES, pH 7.4. Washed platelets (300 μL) were stimulated with various agonists, without stirring, at 37°C in an PA-200 aggregometer (Kowa, Tokyo, Japan), as described.5 29 Precipitates treated with 10% trichloroacetic acid were subjected to SDS-PAGE for determination of the extent of MLC₂₀, MBS, and CPI phosphorylation. Quantitative estimation of the level of phosphorylation was made densitometrically by scanning the immunoreactive bands after immediately photographing the visualized band. The area of an individual peak was measured above the background of the densitometric tracing and was estimated as an arbitrary unit.

ATP secretion was measured, using a Lumi-aggregometer (Chrono-Log, Havertown, PA), as described.5 The platelet reaction mixture consisted of 400μL washed platelet suspension (5 × 10⁹/mL) in Ca⁺⁺-free HEPES-Tyrode buffer, 50 μL Chrono-Lumi luciferase luciferin reagent, 50μL inhibitor (or vehicle), and STA₂ or PMA. Washed platelet suspensions pretreated with or without the inhibitor at 37°C for 3 minutes were activated by 0.5 μM STA₂ or 100 nM PMA for 3 minutes, under conditions of nonstirring. For measurement of cytosolic Ca⁺⁺ concentrations, platelets loaded with fura 2-am (1 μM) were resuspended in Ca⁺⁺-free HEPES-saline buffer (145 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 10 mM HEPES, pH 7.4, 5 mM glucose).36 Fluorescence and ATP secretion were monitored simultaneously, in parallel samples. Fluorescence of fura 2 was measured at 340- and 380-nm excitation and at 550-nm emission using a Hitachi F-4000 fluorescence spectrometer, according to the method of Tsien et al.37 Two-milliliter aliquots of fura 2-loaded platelet suspensions (1 × 10⁸/mL) were transferred to quartz cuvettes, maintained at 37°C in a circulating water bath, continually stirred with a small magnetic stirrer for 2 minutes, and stimulated with STA₂ or PMA under conditions of non-stirring. In all experiments, the fluorescence of untreated and unloaded cells was used for subtraction of the autofluorescence. To minimize the time-dependent effects on platelet responsiveness and leakage of fura 2, experiments were designed to be completed within 1 hour.

Data were analyzed using the StatView statistical software package (Version 5; SAS Institute, Cary, NC). Data are shown as mean ± SE and were compared statistically by Student ttest (unpaired). P < .05 was considered significant.

Recent reports have suggested that CPI-17 is almost exclusively expressed in smooth muscle tissues.19-21 We determined whether CPI is present in human platelets by immunoblot analysis using an anti–CPI-17 antibody.30 As shown in Figure1A, the anti–CPI-17 antibody detected a major immunoreactive band with an apparent molecular mass of 22 kd, higher than that in smooth muscle CPI-17 with an apparent molecular mass of 20 kd in SDS-PAGE.19 To determine whether the immunoreactive protein at 22 kd was the CPI present in human platelets, we analyzed immunoprecipitates with anti-CPI antibody using platelet extracts (Figure 1B). The procedure was to subject the protein A–Sepharose precipitates to SDS-PAGE and follow that by immunoblotting with polyclonal antibodies against the PP1δ catalytic subunit and MBS of myosin phosphatase. PP1 δ isoform and MBS of myosin phosphatase were detected at 38 kd and 130 kd, respectively, in the anti–CPI-17 immunoprecipitate of human platelets (Figure 1B). We further examined phosphatase activity of the immunoprecipitates using [γ-³²P]-phosphorylated MLC as a substrate. Approximately 85% of the total phosphatase activity in the precipitate was attributed to PP1 activity (Figure 1C). We then asked whether the 22-kd immunoreactive band was phosphorylated by PKC, using the anti–CPI-17 immunoprecipitate. CPI phosphorylation was analyzed by immunoblot using an antibody that specifically recognizes CPI-17 phosphorylated at Thr-38 (anti–pThr38–CPI-17). As shown in Figure 1D, CPI phosphorylation at a Thr residue corresponding to Thr-38 in CPI-17 was increased approximately 4-fold with the addition of PKC, and this phosphorylation was inhibited by a non-isoform–selective PKC inhibitor, GF109203X.38 We then investigated the effect of CPI phosphorylation on myosin phosphatase activity in the immunoprecipitates. Because thiophosphorylation of proteins, unlike regular phosphorylation, is highly resistant to protein phosphatases,39 we used the thiophosphorylation form of CPI to determine the effect on myosin phosphatase activity. Myosin phosphatase activity was decreased by CPI thiophosphorylation and was recovered by the inhibition of CPI thiophosphorylation with GF109203X (Figure 1E). These results indicate that the immunoreactive protein at 22 kd is human platelet CPI.

We determined the amount of CPI in whole platelets by immunoblot analysis using purified CPI-17 from aorta smooth muscle as standard. The amount of CPI in platelets was 0.33 ± 0.04 μM (mean ± SE), which is almost the same as that in smooth muscle.21 

To determine whether the agonist-induced phosphorylation of CPI occurs in intact platelets, human platelets were stimulated with various agonists under conditions of nonstirring, and CPI phosphorylation was analyzed by immunoblot analysis using an anti–pThr38–CPI-17 antibody. In parallel, we also examined the phosphorylation of MLC₂₀ at Ser-19 and MBS at Thr-696 by immunoblot analysis using the respective phosphospecific antibodies. When platelets were activated with agonists in the absence of stirring, they changed shape and secreted ATP from dense bodies, without aggregation.5,29 Figure 2shows ATP secretion and cytosolic Ca⁺⁺ elevation response to STA₂ and PMA in a nominally Ca⁺⁺-free medium and under conditions of nonstirring. STA₂ (0.5 μM) treatment led to ATP secretion and a rapid increase in the cytosolic Ca⁺⁺ level. This Ca⁺⁺ increase reached a peak level within 15 to 20 seconds and was followed by a rapid decrease. As expected from the cytosolic Ca⁺⁺ response to STA₂, the maximum extent of MLC₂₀phosphorylation at Ser-19 was obtained within 20 seconds of STA₂ (0.5 μM) stimulation, followed by dephosphorylation, as shown in Figure 3. CPI was rapidly phosphorylated and then dephosphorylated after STA₂stimulation, similar to the phosphorylation of MLC₂₀ at Ser-19, whereas STA₂-induced MBS phosphorylation was slightly delayed. Thus, treatment of intact platelets with STA₂ led to a rapid and transient increase in phosphorylation of CPI at a Thr residue (corresponding to Thr-38 on CPI-17), in addition to MBS phosphorylation at Thr-696. To determine whether phosphorylation of CPI can be induced by agonists other than STA₂ in platelets, CPI phosphorylation was examined after stimulation with thrombin, PMA, and ionomycin. As shown in Figure4, treatment with 0.1 U/mL thrombin led to a rapid increase in phosphorylation of CPI at a Thr residue and of MBS at Thr-696. On the other hand, though the Ca⁺⁺ionophore ionomycin or PMA, which activates PKC, could induce the phosphorylation of CPI, these artificial agonists failed to increase MBS phosphorylation. Ionomycin-induced CPI phosphorylation was almost completely inhibited by 0.1 mg/mL aspirin, suggesting that most of the CPI phosphorylation resulted from the potential of ionomycin to promote cyclooxygenase products.40 It is now apparent that PMA and the Ca⁺⁺ ionophore cannot activate the RhoA signaling pathway. Incubation of fura 2-loaded platelets with PMA led to ATP secretion, with no change in cytosolic Ca⁺⁺ concentrations (Figure 2), as reported,41 and PMA (100 nM)–induced ATP release was slower and less than that induced by STA₂ (0.5 μM). We then examined the time course of phosphorylation of CPI and MLC₂₀ at Ser-19 during PMA-induced platelet activation. As shown in Figure 5, after stimulation with 100 nM PMA, CPI was rapidly phosphorylated, and this phosphorylation preceded MLC₂₀ phosphorylation at Ser-19 and ATP secretion.

PKC cannot directly phosphorylate MLC₂₀ at Ser-19, the MLC kinase site8,42,43; therefore, the phosphorylation of this site induced by PMA in intact platelets appeared to result from the inhibition of myosin phosphatase through phosphorylation of CPI at a Thr residue. In the next set of experiments, we determined the participation of PKC in PMA- or STA₂-induced MLC₂₀ phosphorylation of intact platelets, using a specific PKC inhibitor, GF109203X (Figure 6). As shown in Figure 6A, preincubation of platelets with GF109203X attenuated the PMA-induced phosphorylation of CPI and MLC₂₀at Ser-19, with similar IC₅₀ values (1.52 ± 0.33 μM, n = 3, for CPI phosphorylation; 1.17 ± 0.21 μM, n = 3, for MLC₂₀ phosphorylation, respectively). Inhibition of PMA-induced MLC₂₀ phosphorylation by GF109203X correlated well with CPI phosphorylation. GF109203X at 2 μM completely abolished PMA-induced ATP secretion, as reported.38 The Rho-kinase inhibitor Y-2763244 diminished neither PMA-induced phosphorylation of MLC₂₀ nor CPI (data not shown). GF109203X also dose dependently inhibited the STA₂-induced phosphorylation of CPI and MLC₂₀. However, GF109203X caused a nearly complete inhibition of STA₂-induced CPI phosphorylation, whereas the drug did not inhibit STA₂-induced MLC₂₀ phosphorylation to the same extent seen with the inhibition of CPI phosphorylation (Figure 6B). GF109203X decreased MLC₂₀ phosphorylation by approximately 60%, even at 2 μM, a concentration at which the drug abolished STA₂-induced CPI phosphorylation. GF109203X pretreatment resulted in a marginal decrease in STA₂-induced phosphorylation of MBS at Thr-696.

We reported that the Rho-kinase inhibitor Y-27632 could inhibit the STA₂-induced MBS phosphorylation dose dependently (1-20 μM) in intact platelets.27 Smooth muscle CPI-17 was shown to be phosphorylated in a cell-free system not only by PKC but also by the downstream kinases RhoA, Rho-kinase,30 and protein kinase N.45 We also examined the effect of the Y-27632 on the STA₂-induced phosphorylation of CPI, MBS, and MLC₂₀ in intact platelets. As shown in Figure7, Y-27632 almost completely inhibited the STA₂-induced phosphorylation of MBS at Thr-696, whereas the drug had only a minor inhibitory effect on the STA₂-induced CPI phosphorylation, which suggests that Rho-kinase preferentially phosphorylates MBS, compared to CPI, in intact platelets. Y-27632 partly inhibited the STA₂-induced MLC₂₀ phosphorylation at Ser-19 in intact platelets. As shown in Figure 8, Y-27632 inhibited STA₂-induced MLC₂₀ phosphorylation at Ser-19 to approximately 40% of the level seen without Y-27632, an inhibition likely because of the inhibition of MBS phosphorylation, though slight inhibition of CPI phosphorylation was evident at 20 μM Y-27632. The addition of GF109203X (5 μM) to Y-27632 (20 μM) led to a significant inhibition of CPI phosphorylation and further suppression of STA₂-induced MLC₂₀ phosphorylation. This additional inhibition of MLC₂₀ phosphorylation by GF109203X appeared to be mainly due to the inhibition of CPI phosphorylation because the additional inhibition of MBS phosphorylation by GF109203X was slight. STA₂-induced ATP secretion was partially inhibited by either Y-27632 (20 μM) or GF109203X (5 μM) treatment, and treatment with Y-27632 or GF109203X led to a further decrease in ATP secretion (Figure 8). Inhibition of STA₂-induced ATP release by these compounds appeared to correlate with the extent of MLC₂₀ phosphorylation at Ser-19.

Our experiments revealed the existence of a CPI-17–like protein in human platelets and its functional role in regulating MLC₂₀ phosphorylation at Ser-19 in case of agonist-induced platelet secretion. CPI-17 was originally reported to be exclusively expressed in smooth muscle tissues,19-21 but we detected a single major immunoreactive band with an apparent molecular mass of 22 kd, a slightly higher molecular mass than noted in smooth muscle CPI-17 (molecular mass, 20 kd),19 in human platelets. Myosin phosphatase co-immunoprecipitated with CPI from human platelets, suggesting that platelet CPI interacts with myosin phosphatase. Phosphorylation by PKC of smooth muscle CPI-17 at a Thr residue (Thr-38) is sufficient to induce the inhibition of myosin phosphatase activity.20,28 We showed that in vitro phosphorylation of platelet CPI by PKC at a Thr residue, corresponding to Thr-38 on CPI-17, resulted in a decrease in myosin phosphatase activity in the anti–CPI-17 immunoprecipitate of platelets. Inhibition by the PKC inhibitor GF109203X of PKC-induced CPI-17 phosphorylation reduced inactivation of the phosphatase activity. Thus, similar to smooth muscle CPI-17, platelet CPI appears to be a phosphorylation-dependent inhibitor of myosin phosphatase. Concentration of CPI in human platelets was estimated to be approximately 0.3 μM, similar to the concentration in porcine aorta smooth muscle.21 

Stimulus-induced phosphorylation of CPI at a Thr residue appears to be one mechanism by which PKC enhances the intracellular pathway of regulated platelet secretion. PMA can induce ATP secretion in intact platelets without producing any elevation in cytosolic Ca⁺⁺concentration, as determined using the fluorescent Ca⁺⁺probe fura-2,41 albeit the secretion is slower and of lesser amount than the extent of secretion stimulated by agonists such as thrombin and STA₂. Our study shows that PMA increases both CPI phosphorylation at a Thr residue and MLC₂₀phosphorylation at the same site (Ser-19) phosphorylated by MLC kinase8 and Rho-kinase.9 This activating MLC kinase site is not directly phosphorylated by PKC, but phosphorylation of MLC₂₀ on PKC sites in vitro inhibits smooth muscle ATPase activity.42 43 The mechanism of Ca⁺⁺sensitization of platelet secretion by PMA and by GTPγS appears to be mediated by the increased phosphorylation of MLC₂₀, mainly through the inhibition of myosin phosphatase. This notion is strengthened by our findings that the PKC inhibitor GF109203X inhibits the PMA-induced increases in phosphorylation of both CPI and MLC₂₀ at Ser-19, with similar IC₅₀ values. GF109203X completely inhibited PMA-induced ATP secretion. Moreover, PMA or ionomycin did not induce MBS phosphorylation at Thr-696, thus indicating that these nonphysiological agonists do not activate the RhoA/Rho-kinase signaling pathway in human platelets. If PMA stimulation does not activate MLC kinase, MLC₂₀phosphorylation at Ser-19 may slowly increase because of the PKC-mediated inhibition of myosin phosphatase. This possibility is consistent with our findings that the PMA-induced increase in CPI phosphorylation precedes MLC₂₀ phosphorylation and ATP secretion. Therefore, it is likely that in intact platelets, PMA-induced MLC₂₀ phosphorylation at Ser-19 results from the inhibition of myosin phosphatase by the phosphorylation of CPI rather than by the direct phosphorylation of MLC₂₀ by PKC.

Platelet agonists such as TXA₂ and thrombin function through heterotrimeric G protein-coupled receptors, which activate G_q-type,46G₁₂/G₁₃-type,47,48 and G_i-type49 G proteins. G_q-mediated phospholipase C activation, stimulating the synthesis of inositol triphosphate and diacylglycerol, is thought to represent the main signal pathway leading to full platelet activation such as secretion, whereas G₁₂/G₁₃ couple receptors to the RhoA/Rho-kinase signaling pathways.2,47,48TXA₂ mimetic STA₂ and thrombin induce an elevation of intracellular Ca⁺⁺concentrations36,41 and most likely lead to Ca⁺⁺/calmodulin-dependent MLC kinase-mediated MLC₂₀ phosphorylation at Ser-19 of platelets.1,2,5,8 In addition to MLC kinase-dependent MLC₂₀ phosphorylation, Rho-kinase–mediated MLC₂₀ phosphorylation (both directly and indirectly through MBS phosphorylation) contributes to the platelet secretion induced by these agonists.27 Our findings suggest that PKC-mediated CPI phosphorylation and the resultant inhibition of myosin phosphatase is also functionally involved in platelet secretion. STA₂and thrombin, at doses sufficient to elicit platelet secretion, led to the phosphorylation not only of CPI but also of MBS at Thr-696 in intact platelets; these findings differ from results with PMA or ionomycin bypassing the receptor-operated signaling pathway. STA₂-induced MLC₂₀ phosphorylation at Ser-19 was diminished, but not abolished, by pretreatment of GF109203X, even at the high concentrations with which STA₂-induced CPI phosphorylation was completely inhibited, in parallel samples. These data suggest that PKC-induced CPI phosphorylation is an important, but not the only, mechanism of the STA₂-induced increase in MLC₂₀ phosphorylation at Ser-19. We reported that G protein–coupled receptor agonists such as epinephrine, STA₂, and thrombin induce MBS phosphorylation and inhibit myosin phosphatase activity and that both reactions are inhibited by a prior treatment with Y-27632.27 We also reported that myosin phosphatase activity is inhibited when chicken gizzard MBS is phosphorylated by Rho-kinase at Thr-695, which corresponds to Thr-696 of human MBS,25 although this site of MBS can be phosphorylated by other, as yet unidentified, kinases.17Phosphorylation of Thr-696 occurred under in vivo conditions in which RhoA/Rho-kinase was thought to be activated after G protein–coupled receptor activation of human platelets. Our data support the hypothesis that Thr-696 is a major functional site of phosphorylation by Rho-kinase in human platelets. The Rho-kinase inhibitor Y-27632 almost completely attenuated the STA₂-induced MBS phosphorylation at Thr-696, whereas this inhibitor slightly inhibited STA₂-induced CPI phosphorylation. Smooth muscle CPI-17 was reported to be an in vitro substrate not only for PKC but also for downstream kinases of RhoA Rho-kinase30 and protein kinase N,45 both of which are sensitive to Y-27632. However, our present findings suggest that platelet CPI may be a substrate, but not a preferential one, for Rho-kinase, compared with MBS in intact human platelets. Y-27632 alone partially inhibited STA₂-induced MLC₂₀ phosphorylation at Ser-19, as was the case with GF109203X. The Y-27632–insensitive component of STA₂-induced MLC₂₀ phosphorylation was further suppressed by the addition of GF109203X, an event mainly attributed to a significant inhibition of CPI phosphorylation and myosin phosphatase suppression. Inhibition of STA₂-induced ATP secretion by Y-27632 or GF109203X apparently correlated with the extent of MLC₂₀ phosphorylation at Ser-19. Combined treatment with Y-27632 and GF109203X substantially inhibited STA₂ (0.5 μM)–induced MLC₂₀ phosphorylation and ATP secretion, suggesting that Ca⁺⁺-dependent activation of MLC kinase contributes to a lesser extent to MLC₂₀ phosphorylation at Ser-19 and to platelet secretion under these conditions. The extent of MLC kinase activation appears to depend on the type and concentration of agonist,27 whereas activation of Rho-kinase and PKC can lead to an increase in Ca⁺⁺ sensitivity of platelet secretion through myosin phosphatase inhibition.

In summary, these pharmacological studies suggest that Rho and PKC participate in parallel pathways to increase MLC₂₀phosphorylation mainly through the inhibition of myosin phosphatase following heterotrimeric G protein–coupled receptor activation, as illustrated in Figure 9. RhoA activates Rho-kinase, which phosphorylates MBS and inhibits myosin phosphatase activity. Activation of PKC can also inhibit myosin phosphatase by phosphorylating CPI. These mechanisms operate independently of the activation of Ca²⁺-dependent MLC kinase. The dual activation of MLC₂₀ phosphorylation by kinase activation and phosphatase inhibition may be important for a rapid and potent platelet response to agonist.

We thank Dr R. S. Adelstein (National Instititutes of Health) for helpful discussions, and we thank M. Ohara for language assistance.