Role of Phosphorylation in Mediating the Association of Myosin with the Cytoskeletal Structures of Human Platelets*

The effect of myosin light chain phosphorylation on the association of myosin with the cytoskeletal structures of platelets was quantitated. In unstimulated platelets, little myosin light chain was phosphorylated and myosin remained in solution when cytoskeletons from Triton X-100 lysates of platelets were sedimented by centrifugation. In platelets activated by thrombin, the calcium ionophore A23187, or collagen, the rate and extent of myosin light chain phosphorylation paralleled the association of myosin with platelet cytoskeletal structures. Dephosphorylation of myosin light chain and myosin dissociation from the cytoskeleton oc- curred at comparable rates at longer times after addition of the stimulating agents to platelets. Quantitation of radioactive phosphate in the cytoskeleton-associated myosin and in the soluble myosin showed that the phosphorylated myosin light chain was selectively isolated with the Triton-insoluble cytoskeletons, whereas nonphosphorylated myosin was not associated. Inhibition of the light chain kinase with the calmodulin an- tagonist trifluoperazine inhibited myosin light chain phosphorylation and incorporation of myosin into the platelet cytoskeletons. Inhibition of light chain phosphorylation by prostaglandin El and prostacyclin pro- 3MM chromatography paper and exposed to Trimax x-ray film for 1 to 5 days. Autoradiographs were developed in a Kodak RP X- Omat processor. The relative amounts of radioactivity in phosphopolypeptides were determined either by integration of densitometric scans of autoradiograms, as described above for stained protein bands, or by cutting out the areas of dried gels that superimposed with the bands of interest on the processed autoradiographs. These gel slices were incubated at 37 "C for 18 h with 0.4 ml of Protosol (New England Nuclear) and 50 pl of water. Samples were neutralized with 40 pl of glacial acetic acid and radioactivity determined by scintillation count- ing in 3 ml of scintillation fluid (ACS, New England Nuclear).

The effect of myosin light chain phosphorylation on the association of myosin with the cytoskeletal structures of platelets was quantitated. In unstimulated platelets, little myosin light chain was phosphorylated and myosin remained in solution when cytoskeletons from Triton X-100 lysates of platelets were sedimented by centrifugation. In platelets activated by thrombin, the calcium ionophore A23187, o r collagen, the rate and extent of myosin light chain phosphorylation paralleled the association of myosin with platelet cytoskeletal structures. Dephosphorylation of myosin light chain and myosin dissociation from the cytoskeleton occurred at comparable rates at longer times after addition of the stimulating agents t o platelets. Quantitation of radioactive phosphate in the cytoskeleton-associated myosin and in the soluble myosin showed that the phosphorylated myosin light chain was selectively isolated with the Triton-insoluble cytoskeletons, whereas nonphosphorylated myosin was not associated. Inhibition of the light chain kinase with the calmodulin antagonist trifluoperazine inhibited myosin light chain phosphorylation and incorporation of myosin into the platelet cytoskeletons. Inhibition of light chain phosphorylation by prostaglandin El and prostacyclin produced similar effects. Thus, phosphorylation of the myosin light chain stabilizes the association of myosin with the contractile structures within platelets.
When platelets are stimulated with thrombin or other stimuli such as collagen or the ionophore A23187, they change shape, release the contents of their granules, and, in the presence of Ca2' ions, aggregate. Platelets contain high concentrations of the contractile proteins actin and myosin, and it has long been thought that platelet responses to stimulation involve contractile mechanisms (1,2). Since contractile processes within platelets presumably depend on force-generating interactions of myosin with actin filaments, two fundamental questions can be asked, What changes occur in these two proteins during platelet activation? What regulates their interaction? Recent studies have shown that actin undergoes pronounced changes during platelet activation. In unstimulated platelets, about 40% of the platelet actin is filamentous. In platelets activated with thrombin for only 30 t o 60 s, this value increases to about 60% (3)(4)(5)(6). Actin filaments are insoluble in Triton X-100 and can be sedimented from Triton X-100 extracts of either control or activated platelets by low speed centrifugation (8,700 X g for 4 min), permitting ready evaluation of their structure and composition (5). Filaments from unstimulated platelets are dispersed but those from thrombinactivated platelets are more structured, often retaining the shape of the platelet that existed before it was extracted with Triton X-100. The protein composition of the filamentous structures from unstimulated platelets consists primarily of actin (41% of the total actin in platelets), actin-binding protein (6% of the total), myosin (14% of the total), and an undetermined amount of a 31,000 M, polypeptide. The filamentous structures from platelets activated with thrombin for 30 to 60 s contain more actin (60% of the total), actin-binding protein (20% of the total), and essentially all of the platelet myosin (SO% of the total). While the amounts of actin and actinbinding protein in cytoskeletons remain at these levels at various times after thrombin addition, the amount of myosin decreases (to 60% of the total within 30 min). Thus, thrombin activation of platelets causes actin to polymerize into filaments. These filamentous structures, or cytoskeletons,' can be readily isolated and contain several proteins found in the cytoskeletons of a variety of cell types (4,5).
Since actin-binding protein binds to actin filaments (7-ll), the small increase in concentration of the actin-binding protein in the cytoskeletons obtained from activated platelets is not surprising. However, the almost total conversion of myosin from a Triton-soluble form to a form isolated by low speed centrifugation and the subsequent reversal of this conversion is unexpected. Clearly, platelet activation has transiently altered the physical state of myosin within these cek.
Platelet myosin has been purified to homogeneity and has

MATERIALS AND METHODS
Preparation of Suspensions of Labeled Platelets-Venous blood from healthy adult donors was collected into ?h volume of a solution containing 85 mM sodium citrate, 111 mM dextrose, and 71 mM citric acid (23). The blood was centrifuged at 160 X g for 20 min, and the platelet-rich plasma removed and centrifuged at 730 X g for 10 min to sediment the platelets. The platelet pellet was washed two times by resuspension in the original plasma volume of a buffer containiig 120 m~ sodium chloride, 13 mM trisodium citrate, and 30 mM dextrose, pH 7.0, and isolated by centrifugation at 730 X g for 10 min. They were then resuspended at about 2 X lo9 platelets/ml in a buffer, pH 7.4, containing 150 mM sodium chloride, 10 mM 4-(2-hy&oxyethyl)-lpiperazineethanesulfonic acid, 1 m~ EDTA, and about 0.8 mCi of carrier-free (32P)phosphate/ml (New England Nuclear). The suspension was incubated for 30 min, then the platelets were isolated by centrifugation, washed once in the same buffer, but without (32P)phosphate, and finally resuspended at 2 X IO8 platelets/ml in a buffer containing 138 mM sodium chloride, 2.9 m~ potassium chloride, 12 mM sodium bicarbonate, 0.36 mM sodium phosphate, 5.5 m~ glucose, and 1 m~ EDTA, pH 7.4. All steps in the washing procedure were performed in polystyrene tubes at 22 f 2 "C. Platelet counts were determined with an Electrozone/Celloscope Counter (Particle Data, Inc., Elmhurst, IL). Platelet lysis was monitored by the release of lactate dehydrogenase assayed by the method of Bergmeyer et al. (24).
Incubations-Samples of platelet suspension (0.4 ml) were incubated in micmfuge tubes (1.5 ml) with either 0.1 NIH unit of thrombin/ml (a generous gft from Dr. J. W. Fenton, 11, of the New York Department of Health, Albany, NY), 20 pg of collagen/ml (Horn, Munich, West Germany), 50 p~ ADP, or 0.4 p~ ionophore A23187 (Calbiochem). Thrombin, collagen, and ADP were each added in 8 pl of buffer. Ionophore A23187 was added to the platelet suspension in 0.8 pl of dimethyl sulfoxide. This volume of solvent alone had no effect on any of the parameters studied. Incubations with collagen were performed with constant agitation of the platelet suspension. In some experiments, platelet suspension was preincubated for 2 min with 5 to 50 p~ trifluoperazine (a gift from Dr. Robert Wallace, St. Jude Children's Research Hospital, Memphis, TN), 2 p~ PGEI,Z or 50 n~ PGIz (both supplied by Dr. J. Pike of the Upjohn Co., Kalamazoo, MI).
In experiments in which protein phosphorylation was studied in intact platelets, incubations were terminated by the addition of an equal volume of a 2 times concentrated solubilization buffer. All samples for electrophoresis were incubated at 100 "C for 5 min.
Protein (10 to 50 pg) was electrophoresed through slab gels according to the method of Laemmli (25), using a 5 to 20% exponential gradient of acrylamide in the resolving gel and 3% acrylamide in the stacking gel. Protein was stained with Coomassie brilliant blue. The relative amounts of myosin heavy chain present were determined from scans of stained wet gels, using an Ortec densitometer linked to a Hewlett-Packard 9&15S computer, through which areas under peaks were integrated. To determine the distribution of radioactivity in phosphopolypeptides, gels were dried under vacuum on Whatman No. 3MM chromatography paper and exposed to Trimax x-ray film for 1 to 5 days. Autoradiographs were developed in a Kodak RP X-Omat processor. The relative amounts of radioactivity in phosphopolypeptides were determined either by integration of densitometric scans of autoradiograms, as described above for stained protein bands, or by cutting out the areas of dried gels that superimposed with the bands of interest on the processed autoradiographs. These gel slices were incubated at 37 "C for 18 h with 0.4 ml of Protosol (New England Nuclear) and 50 pl of water. Samples were neutralized with 40 pl of glacial acetic acid and radioactivity determined by scintillation counting in 3 ml of scintillation fluid (ACS, New England Nuclear).   4 5 6 1 2 3 4 5 6 1 2 3 4 5 6 phopolypeptide. Incubation of "'P-labeled intact platelets with thrombin results in a rapid increase in the labeling of two major polypeptides, one of 47,000 M, and one of 20,000 M, (16,17). Fig.  2A shows the "'P-labeling of polypeptides in control and thrombin-treated platelets. The 20,000 M, polypeptide that was phosphorylated during thrombin activation appeared to be a homogeneous peak. Based on the work of Daniel et al. (17,20), who showed that myosin light chain is the only polypeptide of this molecular weight that incorporates increased phosphate during thrombin activation of platelets, we identified this 20,000 M, polypeptide as the myosin light chain. Fig. 1 shows the incorporation of radioactivity into this polypeptide at various times after thrombin was added to intact platelets. The extent of labeling of all platelet polypeptides varied with the three different platelet suspensions used for these experiments. Thus, the amount of radioactivity incorporated into myosin light chain a t each time point was expressed as a percentage of the maximum incorporated. The extent and time course of myosin light chain phosphorylation closely paralleled the amount of myosin association with the cytoskeletons. Maximum phosphorylation was observed approximately 60 s after thrombin addition. After this time, the amount of radioactivity declined in parallel with the amount of sedimentable myosin in the Triton extracts. Unlike that in myosin light chain from thrombin-activated platelets (see later), the radioactivity present in the 20,000 M, region of SDS gels from unstimulated platelets was not recovered in cytoskeletons nor was it a substrate for phosphatase in platelet lysates. Thus, this radioactivity is probably in a polypeptide other than myosin light chain, and to compensate for this background we have adjusted the axes of Figs. 1, 3, and 4 to maximize the overlap of curves.

Time Course
Selective Recovery of Phosphorylated Myosin with Cytoskeletons-We next wanted to determine whether the phosphorylated myosin light chain was associated with the cytoskeletons of Triton-lysed platelets. However, we were aware of the fact that a phosphatase active against myosin light chain has been observed in platelet extracts (26, 27). Therefore, in preliminary experiments, we treated thrombin-activated platelets with the Triton-lysis buffer and determined the level of phosphate in platelet polypeptides by SDS gel electrophoresis a t various times after lysis. It was observed that within 10 s after addition of Triton buffer at 22 k 2 "C, essentially all of the light chain was dephosphorylated, while the 47,000 M, polypeptide was unaffected (data not shown). To prevent dephosphorylation of myosin during isolation of cytoskeletons, a variety of compounds known to inhibit the myosin light chain phosphatase of other cells were added to the Triton extraction buffer (28). We found that these inhibitors (20 mM potassium phosphate, 40 mM sodium pyrophosphate, and 10 mM sodium molybdate), completely inhibited the dephosphorylation of myosin light chain. However, these inhibitors not only inhibited the phosphatase activity, but also prevented the sedimentation of actin-binding protein and partially inhibited the sedimentation of myosin with the Triton-insoluble cytoskeletons. This may have resulted from the high ionic strength of the lysates (approximately 0.4), or from the known inhibitory effect of pyrophosphate on the actinmyosin interaction (29). Table I shows results in which we lowered the EGTA concentration, eliminated Tris from the Triton extraction buffer, and attempted to find a combination of phosphatase inhibitors that prevented dephosphorylation without inhibiting the sedimentation of myosin with cytoskeletons. Lower concentrations of the inhibitors did not affect the sedimentation properties of myosin but they did not inhibit the phosphatase either. Most of the myosin was recovered in cytoskeletons even though it was extensively dephosphorylated during the isolation procedure. The lowest concentrations of inhibitors that completely inhibited dephosphorylation (20 mM potassium phosphate, 40 mM sodium pyrophosphate, and 10 mM sodium molybdate) still partially inhibited myosin sedimentation (only 62% of the total platelet myosin sedimented with cytoskeletons compared with the usual 90 to 100%). These concentrations also prevented the sedimentation of actin-binding protein with cytoskeletons (data not shown). Electron microscopy of negatively stained preparations showed that cytoskeletons of individual platelets were present in preparations containing the phosphatase inhibitors (data not shown). However, these structures were not as tightly condensed as those isolated by our previous procedures (5). We decided to use the modified lysis buffer in which dephosphorylation was completely inhibited to determine whether phosphorylated myosin was preferentially associated

TABLE I Inhibition of dephosphorylation of myosin light chain in Triton lysates and the effect of inhibitors on the recovery of myosin in
cytoskeletons P-labeled platelets were incubated at 25 "C either alone or with shown. Lysates were immediately centrifuged as described in the text 0.1 unit of thrombin/ml for 60 s. Platelets were solubilized in SDS for except that all steps were performed at 4 "C. Recovery of myosin and electrophoresis or lysed by the addition of an equal volume of an ice-of 32P-labeled myosin light chain in the Triton-soluble and -insoluble cold buffer, pH 7.4, containing 2% Triton X-100, 2 mM EGTA, 2 mM fractions were determined as described in the text. Values given are N-ethylmaleimide, and inhibitors to give the final concentrations mean k S.E. from   Platelets that had been labeled with (32P)phosphate were treated at 25 "C with 0.1 unit of thrombin/ml for increasing times then lysed by the addition of an equal volume of ice-cold buffer containing 2% Triton X-100, 2 mM EGTA, 2 mM Nethylmaleimide, 40 mM potassium phosphate, 80 mM sodium pyrophosphate, and 20 mM sodium molybdate, pH 7.4. Cytoskeletons were isolated at 4 "C by centrifugation and solubilized immediately with SDS. The amount of myosin and the amount of radioactivity associated with myosin light chain in the platelet cytoskeletons was determined after electrophoresis of the samples. with cytoskeletons. We argued that if phosphorylation of myosin in intact platelets increased the stability of myosin with cytoskeletons, then the ratio of (32P)phosphate/unit of myosin in the cytoskeletons would be constant at all levels of phosphorylation of myosin in intact platelets (i.e. at all time points after thrombin addition), even under lysis conditions in which sedimentation of some of the myosin with cytoskeletons was inhibited. Fig. 2A shows the phosphopolypeptides present in unstimulated platelets and the time-dependent increases in phosphorylation of myosin light chain and the 47,000 M, polypeptide during thrombin activation. When Triton-soluble ( Fig. 2 0 and -insoluble (Fig. 2B) fractions were isolated in the presence of phosphatase inhibitors, most of the major phosphopolypeptides, including the one of 47,000 M, and the one that comigrated with actin-binding protein, were recovered in the Triton-soluble fraction. Myosin light chain was the only major phosphopolypeptide recovered in cytoskeletons. In the experiment shown in Fig. 2, approximately 80%

Myosin Phosphorylation in Platelets
protein composition of cytoskeletons were examined. Collagen, the divalent cation ionophore A23187, and ADP caused an increase in the amount of actin, actin-binding protein, and the 31,000 M, polypeptide in cytoskeletons (data not shown).
Although the rate of the responses varied, collagen and ionophore A23187, like thrombin, also stimulated phosphorylation of myosin light chain and the association of myosin with cytoskeletons. Fig. 4 shows the correlation between phosphorylation of myosin and myosin sedimenting with cytoskeletons after addition of A23187, collagen, or ADP. Addition of 0.4 p~ A23187 to intact platelets caused both a rapid incorporation of radioactivity into the myosin light chain and a rapid association of myosin with cytoskeletons. The rate and extent of phosphorylation closely correlated with the rate and extent of association of myosin with platelet cytoskeletons. Collagen (20 pg/ml) caused a slower phosphorylation of myosin light chain than did the ionophore but this rate closely paralleled that for the association of myosin with cytoskeletons. In contrast, ADP had little effect on the incorporation of phosphate into myosin light chain or on the association of myosin with cytoskeletons.   dependent inhibition of thrombin-induced phosphorylation of myosin light chain (Fig. 5). This drug also inhibited the association of myosin with cytoskeletons. The extent of inhibition of phosphorylation resembled that of association of myosin with cytoskeletons at all concentrations of trifluoperazine used. When platelets were centrifuged from suspension, the supernatants from control and trifluoperazine-treated platelets contained 7.7 and 8.5%, respectively, of the total platelet lactate dehydrogenase activity (data not shown), indicating that trifluoperazine did not cause extensive platelet lysis. Table I1 shows that thrombin-induced phosphorylation of myosin was also inhibited by preincubation of platelets with PGEI or PGL. These drugs also inhibited the thrombin-induced association of myosin with cytoskeletons and again the extent of inhibition of these two processes was similar.

DISCUSSION
Several groups have demonstrated that adding thrombin to platelets incubated with (32P)phosphate increases incorporation of radioactivity into myosin light chain (16)(17)(18)(19)(20). Four observations in the present study show that phosphorylation of the light chain of myosin stabilizes the association of myosin with the contractile structures of platelets.
( a ) The time course of phosphorylation of myosin was similar to the time course of association of myosin with platelet cytoskeletons isolated by low speed centrifugation after Triton-lysis. This temporal relationship was observed even when we used a variety of stimulating agents that mediate phosphorylation of myosin at very different rates. Thus, the Ca2+ ionophore A23187 caused maximum phosphorylation of myosin and maximum association of this protein with cytoskeletons within 10 s. Thrombin induced maximum phosphorylation after 30 to 60 s, and myosin incorporation into cytoskeletons reached a maximum at a similar time and followed a similar time course. Collagen activates platelets more slowly than either A23187 or thrombin and induces phosphorylation of myosin at a much slower rate (18). We found that the collagen-induced association of myosin with cytoskeletons also occurred at this slower rate. Unlike the other stimuli, ADP causes platelets to change shape, but w i l l not induce secretion unless fibrinogen and calcium are added and the suspension is agitated to induce aggregation (see Ref. 30 for review). We found, as have others (18), that ADP did not induce phosphorylation of myosin in unstirred suspensions, Although the amount of actin, actin-binding protein, and the 31,000 M , polypeptide increased in the cytoskeletons following ADP stimulation, the amount of myosin remained constant.
( b ) The dephosphorylation of the myosin light chain that occurs with time after stimulus addition also caused a dissociation of myosin from cytoskeletal structures. Platelets contain a light chain phosphatase that appears to be active within platelets since the amount of phosphate bound to myosin in activated platelets decreases with time (16)(17)(18)(19)(20). We have previously shown that the amounts of actin, actin-binding protein, and a 31,000 M , polypeptide in cytoskeletons increase to stable levels after the addition of thrombin to platelets, while the amount of cytoskeleton-associated myosin peaks and then declines (4,5). The similarity of the rate and extent of myosin dephosphorylation within platelets to the loss of myosin from cytoskeletal structures is consistent with the idea that myosin light chain phosphorylation stabilizes the association of myosin with platelet cytoskeletons. The failure of myosin to dissociate from cytoskeletons when dephosphorylated in Triton lysates may be due to loss of a control mechanism during platelet lysis.

Myosin Phosphorylation in Platelets 4125
(c) Phosphorylated myosin light chain was selectively isolated with the Triton-insoluble cytoskeletons. Previously, it has been shown that essentially all myosin light chain is phosphorylated during thrombin-induced activation of platelets (20) at a time when essentially all myosin is isolated with the cytoskeletons (5). The constant specific activity of myosin light chain in cytoskeletons at time points before and after the maximum phosphorylation indicates that only phosphorylated myosin is associated with cytoskeletons during thrombin activation of platelets.
(d) Inhibitors of myosin light chain phosphorylation also inhibited myosin association with platelet cytoskeletons. Trifluoperazine inhibits calmodulin-dependent reactions (31). We found that this compound, when added to intact platelets, caused an identical dose-dependent inhibition of the phosphorylation of myosin light chain and the amount of myosin associated with the cytoskeletons. PGEl and PGIz inhibit platelet responses, possibly by lowering the calcium ion concentrations in the cytosol (27). They both inhibited thrombininduced myosin phosphorylation and association of myosin with the cytoskeletons to similar degrees.
Experiments with purified myosin suggest two mechanisms by which phosphorylation of platelet myosin could cause it to be isolated with the Triton-insoluble cytoskeletons. Adelstein and Conti (32), showed that the actin-activatable ATPase activity of platelet myosin is stimulated by phosphorylation of the 20,000 Mr light chain. This increased activity indicates that phosphorylated myosin interacts with actin filaments differently than unphosphorylated myosin. This property of myosin could account for its Triton-insolubility. Another possibility is suggested by the report of Scholey et aE. (33), who showed that phosphorylation of platelet myosin light chain causes myosin to assemble into filaments at physiological ionic strength and Mg-ATP concentrations. If this assembly occurred within platelets, myosin could become Triton-insoluble because it has polymerized into filaments. The present data are consistent with a mechanism involving increased affiity of myosin to actin filaments, since the low speed centrifugation used for isolating cytoskeletons would not be expected to sediment myosin filaments that were not bound to other cytoskeletal structures (15). Furthermore, we have found that depolymerization of actin in Triton lysates by DNase I or CaZc, also prevented the sedimentation of cytoskeletal-associated myosin by low speed centrifugation? Thus, while our data do not permit us to exclude the possibility that phosphorylation of myosin causes it to polymerize in activated platelets, they do suggest that Triton-insolubility of myosin arises from its increased binding to actin filaments, irrespective of the extent of myosin polymerization.
Although the present data show that the phosphorylated form of myosin has increased association with the cytoskeletal structures, it remains to be determined whether (a) phosphorylation is a prerequisite for myosin association or whether (b) myosin association with cytoskeletons is independent of phosphorylation and that the bound myosin serves as a preferred substrate for the light chain kinase. However, the observations that phosphorylated myosin forms filaments (33) and has a higher actin-activatable ATPase activity (32) suggests that phosphorylation may be a prerequisite. Distinguishing between these mechanisms is clearly relevant to understanding the involvement of myosin in the contractile processes of nonmuscle cells.
It has been reported that there is a direct relationship between the extent of phosaorylation of smooth muscle myosin and its actin-activated ATPase activity (34). Similarly, J. E. B. FOX and D. R. Phillips, unpublished observations. the tension generated by platelet actomyosin threads in vitro was proportional to the level of phosphorylation of myosin (35). In contrast, Persechini and Hartshorne suggest a mechanism in which there is ordered phosphorylation of smooth muscle myosin light chains with phosphorylation of both light chains being required for stimulation of the ATPase activity (36). Our results showed that the extent of myosin light chain phosphorylation closely paralleled the association of myosin with the cytoskeleton, although rate measurements showed that the association of myosin slightly preceded phosphorylation while dephosphorylation preceded dissociation (see Figs.  1 and 4). Thus, our data are most consistent with a mechanism in which phosphorylation and dephosphorylation of the two light chains on a myosin molecule occur randomly and phosphorylation of only one light chain is sufficient for association of a myosin molecule with the cytoskeletal structure.
Experiments by others show that the phosphorylation of myosin correlates with a force-generating system. Lebowitz and Cooke (35) prepared actomyosin threads from human blood platelets and found that the maximum isometric tension of these threads was proportional to the level of phosphorylation of myosin light chain. Likewise, phosphorylation of myosin light chain within cells is associated with contraction. In intact arterial smooth muscle, the contraction-relaxation cycle is coincident with cyclic phosphorylation-dephosphorylation of the myosin light chain (37). Also, in tracheal smooth muscle, the extent of myosin phosphorylation occurred temporally with the increase in isometric tension (28).
With human platelets, agents that mediate the release reaction as well as shape change and aggregation, such as thrombin, collagen, and ionophore A23187, stimulate myosin phosphorylation. ADP, which causes shape change and aggregation but not the release reaction, does not cause phosphorylation (18). With rat platelets, phosphorylation can occur without the release reaction (38) but the converse has not been observed. Thus, phosphorylation of myosin light chain plays a role early in the secretory process. Our finding that phosphorylation regulates the association of myosin with cytoskeletons suggests that the release reaction is a contractile process involving a force-generating interaction of actin with myosin. Shape change and aggregation, however, can presumably occur without the involvement of myosin, but may require newly formed actin filaments and their interaction with actin-binding protein.