Regular Article

PLATELETS AND THROMBOPOIESIS

cAMP signaling regulates platelet myosin light chain (MLC)phosphorylation and shape change through targeting the RhoA-Rhokinase-MLC phosphatase signaling pathwayAhmed Aburima, Katie S. Wraith, Zaher Raslan, Robert Law, Simbarashe Magwenzi, and Khalid M. Naseem

Centre for Cardiovascular and Metabolic Research, Hull York Medical School, University of Hull, Hull, United Kingdom

Key Points

• Protein kinase A (PKA)phosphorylates RhoA onserine188 to inhibit RhoAmembrane translocation andRhoA kinase (ROCK)signaling.

• Inhibition of RhoA/ROCK2promotes myosin lightchain (MLC) phosphataseactivity, which prevents thephosphorylation of MLC andplatelet shape change.

Cyclic adenosine monophosphate (cAMP)-dependent signaling modulates platelet shape

change through unknown mechanisms. We examined the effects of cAMP signaling on

platelet contractile machinery. Prostaglandin E1 (PGE1)-mediated inhibition of thrombin-

stimulated shape change was accompanied by diminished phosphorylation of myosin

light chain (MLC). Since thrombin stimulates phospho-MLC through RhoA/Rho-

associated, coiled-coil containing protein kinase (ROCK)-dependent inhibition of MLC

phosphatase (MLCP), we examined the effects of cAMP on this pathway. Thrombin

stimulated the membrane localization of RhoA and the formation of a signaling com-

plex of RhoA/ROCK2/myosin phosphatase-targeting subunit 1 (MYPT1). This resulted in

ROCK-mediated phosphorylation of MYPT1 on threonine 853 (thr853), the disassociation

of the catalytic subunit protein phosphatase 1d (PP1d) from MYPT1 and inhibition of

basal MLCP activity. Treatment of platelets with PGE1 prevented thrombin-induced

phospho-MYPT1-thr853 in a protein kinase A (PKA)-dependent manner. Examination of

the molecular mechanisms revealed that PGE1 induced the phosphorylation of RhoA

on serine188 through a pathway requiring cAMP and PKA. This event inhibited the

membrane relocalization of RhoA, prevented the association of RhoA with ROCK2 and

MYPT1, attenuated the dissociation of PP1d from MYPT1, and thereby restored basal MLCP activity leading to a decrease in

phospho-MLC. These data reveal a new mechanism by which the cAMP-PKA signaling pathway regulates platelet function.

(Blood. 2013;122(20):3533-3545)

Introduction

Platelet shape change, which is critical for spreading and stableadhesion, requires the dynamic remodeling of the actin cytoskel-eton. This is a complex temporal sequence driven by signalingevents that regulate actin dynamics and by proteins that bind actindirectly to facilitate its polymerization.1 The foremost regulator ofactin function in platelets is myosin IIa, and the phosphorylation ofthe regulatory myosin light chains (MLCs) on serine19 (ser19) endowsadenosinetriphosphatase activity that facilitates myosin interactionwith actin filaments required for shape change and secretion of plateletgranules.2 Increased intracellular Ca21 in response to numerousplatelet agonists activates Ca21/calmodulin-dependent MLC kinase(MLCK), resulting in the phosphorylation of MLC-ser19, shape change,and secretion.3-8 Thrombin, thromboxane A2, and oxidized low-density lipoproteins (oxLDLs) also activate a Rho-associated,coiled-coil containing protein kinase (ROCK) pathway that drives therearrangement of the actin cytoskeleton by inhibition of Ca21-independent MLC phosphatase (MLCP).9-12 MLCP is composedof a 38-kDa protein phosphatase 1d (PP1d) catalytic subunit, a130-kDa myosin phosphatase-targeting subunit 1 (MYPT1), anda 20-kDa subunit of unknown function.13 Activated ROCK forms

a complex withMYPT1 leading to ROCK-mediated phosphorylationand inhibition of MLCP.9 Together with the activation of MLCK, theinhibition of MLCP promotes net phosphorylation of MLC requiredfor platelet shape change, spreading, and thrombus stability.

The cyclic adenosine monophosphate (cAMP) pathway, acti-vated by prostacyclin (PGI2), prostaglandin E1 (PGE1), and adenosine,inhibits multiple aspects of platelet function through protein kinase A(PKA), including reduced platelet aggregation in vitro and thrombosisin vivo.14-19 However, the precise molecular mechanism by whichcAMP signaling inhibits platelet activatory signaling cascades remainsunresolved. PKA phosphorylates Ga13, inositol trisphosphatereceptor, vasodilator-stimulated phosphoprotein (VASP), actin-binding protein, and caldesmon, although the physiological rel-evance of the phosphorylation of these targets remains unclear.14 Theidentification of new substrates that are phosphorylated underphysiological conditions and linked to the inhibition of specificplatelet functions is required in order to understand the precisemechanisms that control unwanted platelet activity. With this inmind, we explored the mechanism underlying the ability of cAMPsignaling to modulate the myosin contractile machinery that controls

Submitted March 6, 2013; accepted September 18, 2013. Prepublished online

as Blood First Edition paper, October 7, 2013; DOI 10.1182/blood-2013-03-

487850.

The online version of this article contains a data supplement.

The publication costs of this article were defrayed in part by page charge

payment. Therefore, and solely to indicate this fact, this article is hereby

marked “advertisement” in accordance with 18 USC section 1734.

© 2013 by The American Society of Hematology

BLOOD, 14 NOVEMBER 2013 x VOLUME 122, NUMBER 20 3533

platelet shape change. We wanted to establish the effects of cAMPsignaling on the RhoA/ROCK pathway in platelets and determine howthis influenced the multimeric structure and activity of MLCP. Ourstudy identifies RhoA as a novel target for cAMP/PKA signaling inplatelets. cAMP signaling inhibits RhoA and ROCK activity tofacilitate sustained activation of MLCP, which in turn preventsthe phosphorylation ofMLC required for platelet shape change. Thesefindings provide evidence for a new mechanism of platelet regulationby cAMP signaling.

Methods

Reagents

These studies were approved by the Hull York Medical School ResearchEthics Committee and conducted in accordance with the Declaration ofHelsinki. The following antibodies were used; anti–phospho-VASP-ser157, anti-MYPT1, anti–phospho-MYPT1-thr853(threonine 853), anti-ROCK1 andROCK2 (Cell Signaling Technology, Hitchin, UK); anti-PKAc antibody(BD Transduction Laboratories, Lexington, KY); anti–b-tubulin (Up-state Biotechnology, Dundee, UK); anti-PP1d (Millipore, Watford, UK);anti-phospholipase Cg2 (anti-PLCg2), anti–integrin b3, and phospho-RhoA-ser188 (Santa Cruz Biotechnology, Calne, UK). Y27632, 1,2-bis-(o-aminophenoxy)ethane-tetra-acetic acid tetra-(acetoxymethyl) ester (BAPTA-AM), and GGTI-298were from Calbiochem (Nottingham, UK). 8-(4-Chlorophenylthio)adenosine-39,59-cyclic monophosphorothioate Rp isomer (Rp-8-CPT-cAMPS), and 8-(4-chlorophenylthio)-N6-phenyladenosine-39,59-cyclicmonophosphate (8-CPT-6-Phe-cAMP) were fromBiolog (Breman, Germany).RhoA pull-down kit and Rho Activator I were from Cytoskeleton (Peter-borough, UK). Para-nitrophenyl phosphate was from New England Biolabs(Hitchen, UK). All other reagents were obtained from Sigma Ltd (Poole, UK).

Platelet isolation

Human blood was taken from drug-free volunteers into acid citrate dextrose(29.9 mM Na3C6H5O7, 113.8 mM glucose, 72.6 mM NaCl, and 2.9 mMcitric acid [pH 6.4]). Platelet-rich plasma was obtained by centrifugation ofwhole blood at 200g at 20°C for 20 minutes. Platelet-rich plasma was treatedwith citric acid (0.3 mM) and indomethacin (20 mM) and was centrifuged at800g for 12 minutes. The platelet pellet was then suspended in wash buffer(36 mM citric acid, 10 mM EDTA, 5 mM glucose, 5 mM KCl, 9 mM NaCl)and centrifuged at 800g for 12 minutes. Platelets were resuspended (33 108

platelets per milliliter) in modified Tyrode’s buffer (150 mM NaCl, 5 mMN-2-hydroxyethylpiperazine-N9-2-ethanesulfonic acid, 0.55 mM NaH2PO4,7 mM NaHCO3, 2.7 mM KCl, 0.5 mM MgCl2, 5.6 mM glucose [pH 7.4])supplemented with apyrase (2 U/mL), indomethacin (10 mM), or EGTA(1 mM), unless otherwise stated.

Platelet shape change

Washed platelets (2.5 3 108 platelets per milliliter) were preincubated withEGTA (1 mM), apyrase (2 U/mL), and indomethacin (10 mM). PGE1 (0.2to 100 nM)was added to platelets for 1minute prior to stimulationwith thrombin(0.05 U/mL), and shape change was recorded as previously described.10,12

LDL preparation and oxidation

LDL was prepared from fresh human plasma by sequential densityultracentrifugation and was oxidized as described previously (supple-mental Methods).20

RhoA pull-down assay

Washed platelets (53 108 platelets per milliliter) were treated with agonistsin the presence and absence of PGE1 at 37°Cwith stirring for 1 minute beforestopping the reaction with an equal volume of lysis buffer. Lysates (300 mg)were incubated for 90 minutes at 4°C with Rhotekin-RBD beads (25 mg).Bead pellets were washed once, and Laemmli buffer was added prior toimmunoblotting.

In vitro kinase assay

Recombinant human full-length active PKAcb was incubated with re-combinant human His-tagged RhoA (55 ng) in kinase buffer (25 mM4-morpholinepropanesulfonic acid; pH 7.2], 12.5 mM glycerol-2-phosphate,25 mM MgCl2, 5 mM EGTA, 2 mM EDTA, and 0.25 mM dithiothreitol)supplemented with adenosine triphosphate (400 mM) at 37°C for 15 minutes.Reaction was stopped by addition of Laemmli buffer to the mixture for furtherimmunoblot analysis.

Measurement of cAMP

Washed platelets (2 3 108 platelets per milliliter) were treated with PGE1

(100 nM) for 1 minute in the presence and absence of inhibitors, and reactionswere terminated by addition of lysis buffer. cAMP levels were assayed witha commercial enzyme immunoassay system and expressed as fmol cAMP per1 3 107 platelets.21

Subcellular fractionation

Washed platelets (53 108 platelets per milliliter) were treated with agonistsfor the appropriate time, and reactions were stopped with an equal volumeof fractionation buffer (320 mM sucrose, 4 mM N-2-hydroxyethylpiper-azine-N9-2-ethanesulfonic acid, and 0.5 mM Na3VO4 [pH 7.4]) supple-mented with phosphatase and protease inhibitor cocktail.22 Samples weresubjected to 5 freeze-thaw cycles. Intact platelets were removed by cen-trifugation at 5000g for 10 minutes at 4°C before centrifugation of lysates at100 000g for 60 minutes at 4°C. The supernatant (soluble fraction) wasremoved, and the pellet (particulate fraction) was resuspended in lysis buffer.Protein concentrations were quantified by using a DC protein assay kit(Amersham Biosciences, UK). Equal protein concentrations of the fractionswere resolved by sodium dodecyl sulfate polyacrylamide gel electrophoresisand immunoblotted for PLCg2 and integrin b3 to validate fractionation.

Immunoprecipitation and immunoblotting

Washed platelets (83 108 platelets per milliliter) were treated with agonistsin the presence or absence of PGE1 and incubated at 37°Cwith stirring for theappropriate time. Reactions were stopped by the addition of an equal volumeof ice-cold lysis buffer (150 mM NaCl, 10 mM tris(hydroxymethyl)aminomethane, 1 mM EDTA, 1 mM EGTA, 1% Igepal, 1 mM phenyl-methanesulfonyl fluoride, 2.5 mM Na3VO4, phosphatase cocktail inhibitor, andprotease cocktail inhibitor [pH 7.4]). Lysates were incubated overnight at 4°Cwith anti-RhoA (1mg), anti-ROCK1 (1.5mg), anti-ROCK2 (1.5mg), anti-PP1d(5 mg), or immunoglobulin G control antibodies, with constant rotation in thepresence of protein A Sepharose beads. To prepare whole-cell lysates, plateletswere then treated with the appropriate reagents before termination of thereactions with Laemmli buffer. Immunoprecipitates or lysates were separatedby sodium dodecyl sulfate polyacrylamide gel electrophoresis and analyzed byimmunoblotting as previously described.21 The following primary antibodieswere used; anti–phospho-VASP-ser157, anti–phospho-RhoA-ser188 (both1:1000), anti-RhoA, anti-ROCK1, anti-ROCK2, anti-PP1d (all 1:500), anti–phospho-MYPT1-thr853, and anti-MYPT1 (both 1:250).

Measurement of phosphatase activity

PP1d was immunoprecipitated from unstimulated or thrombin-stimulatedplatelets in the presence or absence of PGE1 or Y27632. ImmunoprecipitatedPP1d was incubated for 1 hour at room temperature with 5 mM para-nitrophenyl phosphate in reaction buffer (100mL; 4mM tris(hydroxymethyl)aminomethane-base, 2 mM KCl, 3 mM MgCl2, 0.1 mM MnCl2, 0.1 mg/mLbovine serum albumin, and 2 mM dithiothreitol [pH 8.1]). The reaction wasstopped with 5 mM NaOH, and absorbance was measured at 405 nm.23

Specific PP1d activity was obtained by subtracting the activity in the nonimmuneimmunoglobulin G immunoprecipitate from the activity in the PP1d immuno-precipitate. To verify that each sample had equal amounts of PP1d, samples werecollected post-assay and immunoblotted for PP1d.

Statistical analysis

Results are expressed as means 6 standard error of the mean and wereanalyzed by using the Student t test or analysis of variance. The results wereconsidered significant when P values were less than .05.

3534 ABURIMA et al BLOOD, 14 NOVEMBER 2013 x VOLUME 122, NUMBER 20

Results

PGE1 inhibits shape change and phosphorylation of MLC

induced by thrombin

Under conditions that abrogated the effects of secondary signalingthrough adenosine 59-diphosphate, thromboxane, and integrins, throm-bin (0.05 U/mL) induced shape change, which was also associatedwith a rapid increase in phospho-MLC-ser19; phosphorylation

was maximal at 1 minute and was maintained for up to 60 minutes(longest time tested) (Figure 1A-B). When platelets were treatedwith PGE1 (100 nM) prior to stimulation with thrombin, shapechange was abolished and MLC phosphorylation was maintainedat basal levels (Figure 1C). Confirmation that the cAMP/PKAsignaling pathway was active under these conditions was evidencedby an increase in both cAMP from 98 6 30 to 333 6 101 fmols per1 3 107 platelets (P , .05) (Figure 1D) and phospho-VASP-ser157

(Figure 1E), a marker of PKA signaling. The increased cAMP and

Figure 1. PGE1 modulates thrombin-induced shape change and

MLC phosphorylation. (A) Washed platelets (3 3 108 platelets per

milliliter) were preincubated with apyrase (2 U/mL), indomethacin (10 mM),

and EGTA (1 mM) followed by stimulation with thrombin (0.05 U/mL) in the

presence and absence of PGE1 (0.2 to 100 nM), and traces were recorded

for 2 minutes. Shown are representative traces of 3 separate experiments.

(B) Washed platelets (3 3 108 platelets per milliliter) were stimulated with

thrombin (0.05 U/mL) for the indicated time points and the levels of

MLCser19 were assessed by immunoblotting. (i) Representative immuno-

blots (IB) from 3 independent experiments. (ii) Densitometric analysis of

MLCser19 phosphorylation of 3 different experiments expressed as arbi-

trary units (AU). **P , .01 compared with basal levels. (C) Washed

platelets (3 3 108 platelets per milliliter) were stimulated with thrombin

(0.05 U/mL) for 1 minute in the presence or absence of PGE1 (0.2 to 100

nM) and phospho-MLCser19 was assessed by immunoblotting. (i)

Representative immunoblots from 4 independent experiments. (ii) Densi-

tometry of MLCser19 phosphorylation from 4 different experiments. **P, .01

compared with thrombin alone. (D)Washed platelets (23 108 platelets per

milliliter) were treated with PGE1 (100 nM) in the presence and absence of

RO1138452 (1 mM) or Rp-8-CPT-cAMPS (RP; 500 mM)/KT-5720 (20 mM),

and cAMP levels were measured. Data are mean 6 standard error of the

mean of 3 experiments and are expressed as fmol cAMP/13 107 platelets.

**P, .01 compared with basal levels. (E) Same as in (D), except platelets

were lysed and immunoblotted for phospho-VASP-ser157. Blot is repre-

sentative of 5 independent experiments. (F) Same as in (A), except that

experimentswereperformed in the presence ofRO1138452 (1 mM) or Rp-8-

CPT-cAMPS (500 mM)/KT-5720 (20 mM). Shown are representative

traces of 3 separate experiments.

BLOOD, 14 NOVEMBER 2013 x VOLUME 122, NUMBER 20 cAMP ACTIVATES MYOSIN LIGHT CHAIN PHOSPHATASE 3535

phosphorylation of VASP-ser157 were ablated by the immunopre-cipitate receptor antagonist RO1138452,24 while two structurallydiverse inhibitors of PKA—Rp-8-CPT-cAMPS (500 mM)25 andKT-5720 (20 mM)—blocked only phospho-VASP-ser157. Impor-tantly, these inhibitors also prevented PGE1 from inhibiting thrombin-induced shape change (Figure 1F). Consistent with these data, PGI2,which signals through cAMP, also inhibited thrombin-induced MLCphosphorylation (supplemental Figure 1). Thus, PGE1 (100 nM)abolished thrombin-induced shape change and MLC phosphoryla-tion through a mechanism that may involve cAMP signaling.

cAMP signaling independently targets both Ca21-dependent

and RhoA/ROCK pathways

Thrombin stimulates the Ca21-dependent activation of MLCK andRhoA-ROCK (Ca21-independent) –dependent inhibition of MLCPin platelets to induce MLC phosphorylation.10,11 We examinedwhether PGE1 targeted these signaling cascades independently ofeach other by using BAPTA-AM to chelate intracellular Ca21

and the widely used ROCK inhibitor Y27632 (10 mM).26,27 Acombination of BAPTA-AM (20 mM) and Y27632 (10 mM)ablated thrombin-stimulated phospho-MLC-ser19, confirming a rolefor both pathways. Platelets were then preincubated with BAPTA-AM (20 mM) alone and stimulated with thrombin. We reasoned thatunder these conditions, only the ROCK pathway would be active,allowing examination of the effects of cAMP on ROCK signaling inisolation. BAPTA-AM reduced, but did not block, MLC phosphor-ylation. However, the combination of BAPTA-AM with PGE1

abolished the phosphorylation of MLC (Figure 2A), suggestingthat ROCK signaling is blocked by cAMP. We next inhibited theROCK pathway with Y27632 (10 mM) to study the effects ofcAMP on the isolated Ca21-dependent pathway stimulated bythrombin. Under these conditions, phospho-MLC-ser19 was re-duced significantly but not blocked. The combination of Y27632and PGE1 abolished phosphorylation of MLC-ser19 (Figure 2).Similar results were obtained with HA-1077, a structurally distinctROCK inhibitor (data not shown). These data suggest that physi-ological activation of cAMP signaling can potentially target theRhoA/ROCK–dependent pathways independently of effects onCa21 mobilization to modulate platelet function.

PGE1 inhibits thrombin-induced activation of RhoA through

phosphorylation of ser188

The pharmacologic inhibition of the ROCK and PKA pathways haslimitations because of the potential off-target effects of kinaseinhibitors.26,28 Therefore, to clarify the mechanism by which PGE1

regulated RhoA-ROCK–dependent signaling, we investigated acti-vation of RhoA by using a guanosine triphosphate (GTP) –RhoApull-down assay.29 Thrombin (0.05 U/mL) increased the levels ofGTP-bound RhoA (Figure 3A), which was abolished if plateletswere pretreated with PGE1 (100 nM). To confirm that our observationwas cAMP mediated, we show that the cell-permeable nonhydrolyz-able cAMPmimetic 8-CPT-6-Phe-cAMP (50mM) and direct activatorof PKA also inhibited formation of GTP-RhoA (Figure 3A). Inhibitionof RhoA activation by PGE1 was prevented by the PKA inhibitorRp-8-CPT-cAMPS and KT-5720 (Figure 3B), suggesting that PGE1inhibits the activation of RhoA through a cAMP/PKA pathway. Incontrast, Y27632 and BAPTA-AM had no effect on RhoA activation,indicating that cAMP signaling targets RhoA activation independentlyof any potential effects on ROCK activity and Ca21 flux (Figure 3A).

PGI2 may induce phosphorylation ofGa12/13 to inhibit its activity.30

To confirm that cAMP targets RhoA independently ofGa12/13, we used

two different approaches: Rho Activator I, which activates RhoAindependently of G-protein coupled receptors (GPCRs),31,32 andoxLDL, which we have recently shown to activate RhoA down-stream of CD36 and independently of GPCRs.12 Incubation ofplatelets with Rho Activator I (30 mM) and oxLDL (50 mg/mL) ledto shape change (supplemental Figure 2), GTP loading, and activationof RhoA, which was inhibited by PGE1 (Figure 3C-D). The abilityof PGE1 to block RhoA activation was prevented by Rp-8-CPT-cAMPS/KT-5720. Thus, PGE1 inhibits the activation of RhoAdownstream of several independent pathways through a PKA-dependent mechanism.

The phosphorylation of RhoA on ser188 may negatively regulateRhoA activity either through the inhibition of the GTPase or pre-vention of its membrane compartmentalization.33,34 Since PKAphosphorylates RhoA under in vitro conditions,28,35,36 we hypoth-esized that PKA-mediated phosphorylation of RhoA may account,at least in part, for the inhibition of RhoA by cAMP signaling. PGE1

(0.2 to 100 nM) induced a concentration-dependent increase inphospho-RhoA-ser188. Phosphorylation was maximal at 100 nMand was consistent in profile with the phosphorylation of VASP-ser157 (Figure 4A). Phosphorylation of RhoA in response to PGE1

occurred within 30 seconds and was maintained for 60 minutes(supplemental Figure 3). To determine whether phosphorylationof RhoAwasmediated by PKA, we used a four-pronged strategy. First,PGE1-induced RhoA phosphorylation was reproduced by 8-CPT-6-Phe-cAMP (50 mM), which caused a significant phosphorylation

Figure 2. PGE1 inhibits Ca21 and RhoA/ROCK-dependent phosphorylation of

MLC.Washed platelets (33 108 platelets per milliliter) were stimulated with thrombin

(0.05 U/mL) for 1 minute in the presence or absence of Y27632 (10 mM), BAPTA-AM

(20 mM), or PGE1 (100 nM), and phospho-MLCser19 was assessed by immunoblotting.

(A) Representative immunoblots from 5 independent experiments. (B) Densitometry of

MLCser19 phosphorylation of 5 different experiments. ns, not significant. *P , .05

compared with thrombin only; **P , .01 compared with thrombin only.

3536 ABURIMA et al BLOOD, 14 NOVEMBER 2013 x VOLUME 122, NUMBER 20

of RhoA (Figure 4B). Second, Rp-8-CPT-cAMPS (500 mM)/KT-5720 (20 mM) prevented the phosphorylation of RhoA by bothPGE1 (100 nM) and 8-CPT-6-Phe-cAMP (50 mM) (Figure 4B).Third, immunoprecipitation of RhoA from platelets demonstratedthat the PKA catalytic subunit (PKAc) is associated with RhoAunder basal conditions (Figure 4C), suggesting that RhoA couldbe a direct target for PKA in platelets. Fourth, to confirm that thisassociation under physiological conditions could lead to the phos-phorylation of RhoA by PKA, we performed an in vitro kinase

assay. Incubation of recombinant human RhoA with recombinantactive PKAc resulted in the phosphorylation of RhoA on ser188

(Figure 4D).

PGE1 inhibits membrane compartmentalization of RhoA

In unstimulated cells, RhoA is localized predominantly in the cytosolwith a small fraction associated with the plasma membrane.35,37,38

We examined whether subcellular localization of RhoA was affected

Figure 3. cAMP/PKA signaling modulates the

activation of RhoA. (A) Washed platelets (5 3 108

platelets per milliliter) were stimulated with thrombin

(0.05 U/mL) for 1 minute in the presence or absence of

Y27632 (10 mM), BAPTA-AM (20 mM), 8-CPT-6-Phe-

cAMP (50 mM), or PGE1 (100 nM), and activated RhoA

(GTP-RhoA) was assessed by using a pull-down assay

with Rhotekin-RBD beads. (i) Representative immuno-

blots from 3 independent experiments. (ii) Densitomet-

ric analysis of activatedRhoA (GTP-RhoA) of 3 different

experiments. *P, .05 comparedwith thrombin only. (B)

Same as in (A), except platelets were treated with PKA

inhibitors Rp-8-CPT-cAMPS (500 mM)/KT-5720 (20 mM)

for 15minutes prior to addition of thrombin and PGE1. (C)

Same as in (B), except platelets were treated with Rho

Activator I (30 mM) for 2 minutes instead of thrombin. (D)

Same as in (B), except platelets were treated with oxLDL

(50 mg/mL) for 15 seconds instead of thrombin.

BLOOD, 14 NOVEMBER 2013 x VOLUME 122, NUMBER 20 cAMP ACTIVATES MYOSIN LIGHT CHAIN PHOSPHATASE 3537

by PGE1. Small amounts of RhoA were observed in the particulatefraction with the majority in the soluble fraction. However, uponstimulation with thrombin (0.05 U/mL), there was a rapid trans-location of RhoA to the particulate fraction, which peaked at 15seconds before returning to basal levels at 60 seconds (Figure 5A).Preincubation with PGE1 (100 nM) ameliorated thrombin-inducedRhoA membrane localization (Figure 5B). The inhibition of move-ment of RhoA to the membrane by PGE1 was mirrored by anincreased phospho-RhoA-ser188 in the soluble fraction (Figure 5B).Similar effects were observed when platelets were stimulated withoxLDL (supplemental Figure 4), indicating that the inhibition ofRhoAmembrane localization was independent of GPCR-mediatedplatelet activation. Therefore, consistent with data on other cells,39,40

our data indicate that the phosphorylation of RhoA-ser188 in response tocAMP signaling may inhibit RhoA membrane compartmentalization.

To confirm that membrane compartmentalization of RhoAwas critical for its activation, we used an inhibitor of proteingeranylgeranyltransferase type I (GGTase I), GGTase I inhibitor298 (GGTI-298), which prevents the association of RhoA andRas with the plasma membrane through inhibition of proteinisoprenylation.41 If translocation of RhoA preceded its activa-tion, then GGTI-298 would inhibit both translocation and GTP

loading. Conversely, if the loading of GTP occurs prior to membranelocalization, then GGTI-298 would inhibit only RhoA translocationbut not GTP loading. Pretreatment of platelets with GGTI-298(50 mM) inhibited RhoA membrane translocation, GTP loadingof RhoA, and shape change in response to thrombin (Figure 5B-D).To confirm that cAMP-mediated phosphorylation of RhoA occurred inthe cytosol, we examined phospho-RhoA-ser188 in the presence ofGGTI-298. PGE1-induced VASP-ser157 and RhoA-ser188 phosphor-ylation were unaffected by GGTI-298, suggesting that phospho-RhoA-ser188 is generated in the cytosolic fraction and that theinhibition of isoprenylation has no effect on activation of the cAMPsignaling cascade (Figure 5E). These data demonstrate that PGE1

prevents RhoA activation by blocking its membrane translocation,potentially through phosphorylation of RhoA on ser188.

PGE1 inhibits formation of RhoA/ROCK2/MYPT1

multiprotein complex

ROCK is a downstream effector of RhoA that is proposed to driveplatelet shape change and secretion through inhibition of MLCP.9,11

Having established that cAMP signaling phosphorylated and inhibitedRhoA, we wanted to determine the downstream consequences. ROCK

Figure 4. The phosphorylation of platelet RhoA in

response to PGE1 is cAMP-PKA–dependent. (A)

Washed platelets (3 3 108 platelets per milliliter) were

incubated with PGE1 (0 to 100 nM) for 1 minute, and the

phosphorylation of RhoA-ser188 and VASP-ser157 were

determined by immunoblotting. (i) Representative

immunoblots from 3 independent experiments. (ii)

Densitometry for phospho-RhoA-ser188 from 3 dif-

ferent experiments. *P, .05 for untreated compared

with PGE1 (100 nM). (B) Washed platelets (3 3 108

platelets permilliliter) were treatedwithPGE1 (100 nM) for

1 minute or 8-CPT-6-Phe-cAMP (50 mM) for 2 minutes in

the presence or absence of Rp-8-CPT-cAMPS (500mM)/

KT-5720 (20 mM). RhoA-ser188 and VASP-ser157 phos-

phorylation were assessed by immunoblotting. (i) Repre-

sentative immunoblots from 3 independent experiments.

(ii) Densitometry for phospho-RhoA-ser188 (black bars)

and VASP-ser157 (white bars) from 3 different experi-

ments. *P , .05 PGE1 only compared with PGE1 with

PKA inhibitors. §P, .05 8-CPT-6-Phe-cAMP compared

with 8-CPT-6-Phe-cAMP with PKA inhibitors. (C) Washed

platelets (8 3 108 platelets per milliliter) were lysed, and

RhoA was immunoprecipitated from the lysate overnight.

RhoA and PKAc in the immunoprecipitates (IP) were

detected by immunoblotting. Representative blots for 3

experiments. (D) In lanes 1 to 4, recombinant active

PKAcb (0.1 to 0.5 mg) was incubated with recombinant

His-tagged RhoA (55 ng) in the presence of adenosine

triphosphate (ATP; 400 mM) for 15 minutes at 37°C.Reactions were stopped with 23 Laemmli buffer, and

RhoAser188 phosphorylation was assessed by immu-

noblotting. In lane 4, ATP was omitted from the reaction

mixture. In lane 5, whole-cell lysates from PGE1

(100 nM) –treated platelets were immunoblotted for

phospho-RhoA-ser188. Shown are representative

immunoblots from 3 independent experiments.

3538 ABURIMA et al BLOOD, 14 NOVEMBER 2013 x VOLUME 122, NUMBER 20

activation leads to the phosphorylation ofMYPT1 on thr696 and thr853,the targeting subunit of MLCP, which inhibits the phosphatase.13 Weused the phosphorylation of the best-characterized site—thr85—asa marker of ROCK activity. Thrombin-induced phospho-MYPT1-thr853 was abolished by Y27632 but not by BAPTA-AM (Figure 6A),confirming the role of ROCK. PGE1 (100 nM) blocked thrombin-stimulated inhibitory phosphorylation of MYPT1-thr853 (Figure 6B)which, in turn, was prevented by inhibition of PKA (Figure 6C). Toconfirm that membrane localization of RhoA is required for ROCK-mediated phospho-MYPT1-thr853, we repeated these experiments inthe presence of GGTI-298. Under these conditions, thrombin stimu-lation of phospho-MYPT1-thr853 was abolished (Figure 6B). Inaddition, both Rho Activator I and oxLDL induced the phos-phorylation of MYPT1-thr853, which was prevented by the pres-ence of PGE1 (Figure 6D-E) and was associated with phosphorylationof RhoA-ser188. Together, these data suggest that inhibition ofRhoA by cAMP signaling prevents the activation of ROCK and

phosphorylation of MLCP, regardless of the mechanism of RhoAactivation.

The phosphorylation and inhibition ofMLCP by ROCK requiresthe formation of a multienzyme complex of RhoA, ROCK, andMLCP.42 However, the consequence of RhoA phosphorylation bycAMP/PKA and identity of the ROCK isoform present in this keycomplex is unclear. By using isoform-specific antibodies andimmunoprecipitation, we detected both ROCK1 and ROCK2 inplatelet lysates (supplemental Figure 5).43 Consistent with studies insmooth muscle cells,44 co-immunoprecipitation studies demonstratedthat under basal conditions, RhoA was associated with ROCK2, butthe amount of RhoA associated with ROCK2 increased in response tothrombin. However, in the presence of PGE1 (100 nM), the as-sociation of RhoA/ROCK2 was similar to that observed in restingplatelets (Figure 6F). We next examined the interaction betweenROCK2 and MYPT1. We found no evidence of association betweenROCK2 andMYPT1 under basal conditions; we didfind thatMYPT1

Figure 5. Inhibition of thrombin-stimulated membrane localization of RhoA by PGE1. (A) Washed platelets (53 108 platelets per milliliter) were stimulated with thrombin

for 15 to 60 seconds, and reactions were terminated by the addition of fractionation buffer and snap frozen. Samples were separated into soluble (S) and particulate (P)

fractions by ultracentrifugation and were analyzed for the presence of RhoA, PLCg2, and integrin b3 by immunoblotting. (i) Shown are representative blots of 5 independent

experiments. (ii) Densitometry for RhoA in the particulate fraction from 5 different experiments. *P , .05 unstimulated compared with thrombin (15 seconds). (B) Washed

platelets (5 3 108 platelets per milliliter) were stimulated with thrombin for 15 and 30 seconds in the presence and absence of PGE1 (100 nM) or GGTI-298 (50 mM). Soluble

and particulate fractions were analyzed for the presence of RhoA, RhoA-ser188, PLCg2, and integrin b3 by immunoblotting. (i) Representative blots of 4 independent

experiments. (ii) Densitometry of the presence of RhoA in the particulate fractions at 15 seconds. *P , .05 thrombin compared with thrombin/PGE1 or thrombin/GGTI-298

(15 seconds). (iii) Densitometry of the presence of phospho-RhoAser188 in the soluble fraction at 15 seconds compared with untreated cells. **P, .01 thrombin compared with

thrombin/PGE1. (C) Washed platelets (5 3 108 platelets per milliliter) were stimulated with thrombin (0.05 U/mL) for 1 minute in the presence and absence of PGE1 (100 nM) or

GGTI-298 (50 mM), and reaction was stopped with lysis buffer. Lysates were incubated with Rhotekin-RBD beads (25 ng) for 90 minutes at 4°C. After washing the beads, the

level of activated RhoA (GTP-RhoA) was assessed by immunoblotting. Lysates remaining after pull-down assay were used to examine phospho-RhoA-ser188, VASP-ser157,

and MLC-ser19. (i) Representative immunoblots from 3 independent experiments. (ii) Densitometry of RhoA-GTP from 3 different experiments. *P , .05 thrombin alone

compared with thrombin/PGE1 and GGTI-298. (D) Washed platelets (3 3 108 platelets per milliliter) were preincubated with apyrase (2 U/mL), indomethacin (10 mM), and

EGTA (1 mM) followed by stimulation with thrombin (0.05 U/mL) in the presence and absence of GGTI-298 (50 mM), and traces were recorded for 2 minutes. Shown are

representative traces of 3 separate experiments. (E) Washed platelets (5 3 108 platelets per milliliter) was treated with PGE1 (100 nM) for 1 minute in the presence (white bars)

and absence (black bars) of GGTI-298 (50 mM), and the phosphorylation of RhoA and VASP was assessed by immunoblotting. (i) Representative immunoblots of phospho-RhoA-

ser188 and phospho-VASP-ser157 from 3 independent experiments. (ii) Densitometry of phosphorylation of RhoA from 3 different experiments. NS, not significant.

BLOOD, 14 NOVEMBER 2013 x VOLUME 122, NUMBER 20 cAMP ACTIVATES MYOSIN LIGHT CHAIN PHOSPHATASE 3539

was recruited to the RhoA/ROCK2 complex in response to thrombin.PGE1 attenuated the thrombin-stimulated association of these proteins(Figure 6F). To confirm that the inhibition of ROCK/MYPT1 in-fluenced MLCP activity, we used an anti-PP1d–specific antibodyto immunoprecipitate the phosphatase.9 The basal activity of myosinphosphatase was reduced significantly by stimulation with thrombin(16 0 to 0.36 0.1; P, .05). Pretreatment with PGE1 prevented theinhibition of MLCP and reversed the activity to basal levels (16 0.1)(Figure 6G). In ferret portal smooth muscle cells, the loss of PPIdactivity in response to cell activation is caused by its disassociationfrom MYPT1.45 Consistent with those observations, we found thatunder basal conditions, PP1d is associated with MYPT1 (Figure 6H).However, stimulation with thrombin led to the disassociation of thiscomplex. When platelets were treated with PGE1, the ability ofthrombin to cause the disassociation was lost, and inhibition ofMLCP activity was prevented.

Discussion

The cAMP-PKA signaling cascade modulates a number of importantplatelet functions including Ca21 mobilization and granule secretionand aggregation.15,17,18 An early study indicated that cAMP signalinginhibited platelets independently of its effects onCa21mobilization,15

although the precise mechanisms remain unclear. Here, we presentevidence for a new mechanism by which cAMP regulates plateletfunction. Our novel findings are that (1) cAMP signaling inhibitsplatelet shape change andMLC phosphorylation through activationof MLCP, (2) the activation of MLCP occurs through cAMP-mediated phosphorylation and inhibition of RhoA and downstreamROCK activity, and (3) cAMP prevents the formation of a multien-zyme complex of RhoA/ROCK2/MYPT1 and promotes association ofMYPT1 with the catalytic subunit PP1d.

Figure 6. PGE1 modulates the formation of a RhoA/ROCK/MYPT1 signaling complex. (A) Washed platelets (53 108 platelets per milliliter) were stimulated with thrombin

(0.05 U/mL) for 1 minute in the presence or absence of Y27632 (10 mM) and BAPTA-AM (20 mM); phospho-MYPT1-thr853 was then assessed by immunoblotting. (i)

Representative immunoblots from 3 independent experiments. (ii) Densitometry of phospho-MYPT1-thr853 from 3 different experiments. **P, .01 thrombin comparedwith basal

levels; §P , .05 thrombin alone compared with Y27632. (B) Same as in (A), except platelets were treated with PGE1 (100 nM) or GGTI-298 (50 mM) prior to stimulation with

thrombin, and phospho-MYPT1-thr853 was assessed by immunoblotting. (i) Representative immunoblots from 3 independent experiments. (ii) Densitometry of MYPT1-thr853

phosphorylation of 3 different experiments. **P , .01 thrombin alone compared with thrombin/PGE1 and GGTI-298. (C) Washed platelets (53 108 platelets per milliliter) were

stimulated with thrombin (0.05 U/mL) for 1 minute in the presence or absence of PGE1 (100 nM), and phospho-MYPT1-thr853 was assessed by immunoblotting. In some cases,

platelets were also treated with Rp-8-CPT-cAMPS (500 mM)/KT-5720 (20 mM). (i) Representative immunoblots from 3 independent experiments. (ii) Densitometric analysis of

phospho-MYPT1thr853 of 3 different experiments. *P , .05 compared with thrombin only. (D) Same as in (B), except platelets were treated with Rho Activator I (30 mM) for

2 minutes instead of thrombin. (E) Same as in (B), except platelets were treated with oxLDL (50 mg/mL) for 15 seconds instead of thrombin. (F) Washed platelets (8 3 108

platelets per milliliter) were stimulated with thrombin (0.05 U/mL) for 1minute in the presence and absence of PGE1 (100 nM). Reactionwas stoppedwith lysis buffer andROCK1

was immunoprecipitated. The immunoprecipitates were then immunoblotted for the presence of ROCK2, MYPT1, and RhoA. (i) Representative immunoblots of 4 independent

experiments. (ii) Densitometric analysis of amount of RhoApresent in the immunoprecipitates. *P, .05 thrombin comparedwith thrombin alone. (G) Same as in (F), except PP1d

was immunoprecipitated and was incubated for 1 hour at room temperature with 5 mM para-nitrophenyl phosphate, and absorbance was measured at 405 nm. Graph is

representative of 5 different experiments. **P, .01 thrombin compared with basal activity; *P , .05 thrombin alone compared with thrombin with PGE1. (H) Samples from (G)

were immunoblotted for PP1d and MYPT1. (i) Representative immunoblots of 4 independent experiments. (ii) Densitometric analysis of total MYPT1 of 4 different experiments.

*P , .05 thrombin compared with basal levels; §P , .05 thrombin alone compared with thrombin with PGE1.

3540 ABURIMA et al BLOOD, 14 NOVEMBER 2013 x VOLUME 122, NUMBER 20

A recent study showing that RhoA depletion results in plateletswith altered shape and function46 has placed renewed importanceon understanding the regulation of this small G-protein. In theabsence of the availability of RhoA-deficient mice, we used a lessstringent pharmacologic isolation of ROCK signaling to showthat cAMP targets this pathway independently of any effect onCa21 signaling. Y27632 targets the catalytic adenosinetriphos-phatase domain of ROCK, which shares homology with otherserine-threonine kinases and therefore may not be entirely specific.26,28

Since using this pharmacologic approach leaves open the possibility

that cAMP targets other kinases to regulate shape change, we useda pull-down assay to demonstrate that cAMP signaling preventsthe activation of platelet RhoA upstream of ROCK. In this regard,cAMP signaling may target the activation of platelet Ga12/13 toprevent RhoA activation,30,47 although this remains unconfirmed inother cell types. We found that the activation of cAMP signalingor PKA directly blocks RhoA activation and ROCK signaling inresponse to Ga12/13-dependent, receptor-independent, and CD36-mediated stimulation, suggesting that the actions of cAMP onRhoA are independent of its mode of activation. Moreover, we

Figure 6. (Continued).

BLOOD, 14 NOVEMBER 2013 x VOLUME 122, NUMBER 20 cAMP ACTIVATES MYOSIN LIGHT CHAIN PHOSPHATASE 3541

provide multiple lines of evidence to suggest that RhoA may bea novel physiological target of cAMP-PKA signaling in platelets.PGE1 and the cAMP mimetic 8-CPT-6-Phe-cAMP induce rapidand sustained phosphorylation of RhoA-ser188, which is blockedby two structurally distinct inhibitors of PKA. The possibilitythat RhoA is a physiological substrate for PKA signaling is strength-ened by observations demonstrating that RhoA is phosphorylatedby PKA in vitro and that the catalytic subunit of PKA associateswith RhoA in vivo, which could facilitate a phosphorylationevent similar to those observed in other cells.35 In lymphocytes,fibroblasts, and smooth muscle cells, phosphorylation at ser188

by PKA or expression of a phosphomimetic RhoA-ser188 (RhoA-188A) prevents RhoA movement to the membrane.35,48 Consistentwith the results of these studies, the results of our study showedthat the thrombin-induced membrane compartmentalization ofplatelet RhoA was prevented by PGE1 and was associated withincreased cytosolic phospho-RhoA-ser188. The cytosolic com-partmentalization of RhoA by the pharmacologic inhibition ofisoprenylation also resulted in maintenance of the guanosinediphosphate–bound form. Thus, the ability of cAMP signaling

to prevent membrane relocalization of RhoA likely reduces itsactivation, potentially through guanosine diphosphate dissociationinhibitors (GDIs), since phospho-RhoA-ser188 increases its inter-action with GDIs and enhances the ability of GDIs to extractRhoA from membranes.33,34

RhoA/ROCK signaling induced by thrombin inactivates MLCP,leading to phosphorylation of MLC and platelet shape change.49-51

By using co-immunoprecipitation, we showed for the first time thatplatelet activation by thrombin caused formation of a multienzymeRhoA/ROCK2/MYPT1 complex,52 which led to the inhibitoryphosphorylation of MYPT1. Moreover, cAMP signaling preventsthe formation of the RhoA/ROCK2 complex and, under theseconditions, ROCK2 cannot associate with MYPT1 (Figure 7). Byusing a site-specific antibody to phospho-MYPT1-thr853, wedemonstrated that cAMP prevents thrombin-stimulated inhibitoryphosphorylation of MLCP. Since inhibition of membrane localizationof RhoA by blocking its isoprenylation also inhibited thrombin-stimulated phospho-MYPT1-thr853, our data suggest that the cy-tosolic compartmentalization of RhoA by cAMP signaling preventsthe formation of the multienzyme RhoA/ROCK2/MYPT1 complex

Figure 6. (Continued).

3542 ABURIMA et al BLOOD, 14 NOVEMBER 2013 x VOLUME 122, NUMBER 20

required to inhibitMLCP. These data are consistent with observationsin smooth muscle cells in which cAMP-mediated inactivation ofRhoA leads to activation of MLCP.48 In smooth muscle cells, PKAand protein kinase G prevent or reverse agonist-mediated in-hibitory phosphorylation on thr696 and thr853 and depression ofMLCP activity.49,53 We demonstrate that an analogous mecha-nism exists in platelets. Thrombin causes a significant depression inplatelet PP1d activity, which is associated with the disassociation ofPP1d from MYPT1, a phenomenon already reported in smoothmuscle cells.45 cAMP caused a modest increase in basal PPIdactivity but completely reversed the inhibitory effects of thrombin onPP1d activity. It seems that a major role of cAMP signaling is toprevent this dissociation of PP1d from MYPT1 and maintain theholoenzyme complex and phosphatase activity. Our data suggestthat one mechanism to achieve this is the inhibition of RhoA/ROCKsignaling. In smooth muscle cells,49 cAMP signaling targetsMYPT1 directly to influence phosphatase activity, and it istherefore possible that regulation of RhoA represents only one aspectof MLCP regulation by cAMP signaling.

The inhibitory effects of cAMP/PKA on RhoA signaling providea new mechanism through which endothelial-derived plateletmediators could modulate platelet shape change. The inactiva-tion of RhoA by cAMP/PKA prevents the formation of a RhoA/ROCK2 signaling complex, which attenuates the inhibitoryphosphorylation of MLCP and prevents the disassociation ofPP1d/MYPT1 complex. Under these conditions, the phosphatasecould remain active, thereby preventing phosphorylation of MLCand inhibiting shape change (Figure 7). The use of protein kinaseinhibitors has allowed us to delineate a potentially new pathwaycontrolling platelet shape change. However, given the complexity of

signaling pathway and potential off-target effects of the inhibitors wehave used, the generation of conditional knockouts for platelet-selective deletion of ROCK1, ROCK2, and PKA are required toconfirm the regulation ofMLCP by ROCK2 and the ability of PKA tomodulate ROCK2 activity. In conclusion, the discovery of RhoA asa novel substrate of cAMP signaling and targeting of the RhoA/ROCK2/MLCP pathway represents a new and important aspect of themechanism by which cAMP modulates platelet function.

Acknowledgments

This work was supported by The British Heart Foundation (PG/10/90/28636).

Authorship

Contribution: A.A. performed experiments and analyzed and inter-preted data; K.S.W., Z.R., R.L., and S.M. performed experiments; andK.M.N. designed research, analyzed and interpreted data, and wrotethe manuscript.

Conflict-of-interest disclosure: The authors declare no compet-ing financial interests.

Correspondence: Khalid M. Naseem, Centre for Cardiovascularand Metabolic Research, Hull York Medical School, University ofHull, Cottingham Rd, Hull, United Kingdom; E-mail: khalid.naseem@hyms.ac.uk.

Figure 7. cAMP signaling caused activation of

MLCP through regulation of RhoA/ROCK. Diagram-

matic representation of the cross talk between cAMP

and RhoA/ROCK signaling in platelets. Platelet activa-

tion leads to membrane relocalization and GTP loading

of RhoA, which facilitates the activation of ROCK (1). A

RhoA/ROCK/MYPT1 complex is formed leading to

ROCK-dependent inhibitory phosphorylation of MLCP

(2). Simultaneously, phospholipase C (PLC) is

activated to generate inositol trisphosphate (IP3)

and release intracellular Ca21. The increased

Ca21 concentrations leads to the calmodulin

(CAM) dependent activation of MCLK. Under these

conditions, the ability of MLCP to dephosphorylate

MLC is diminished (gray dotted line; upper panel).

Activated PKA phosphorylates RhoA on ser188, which

prevents membrane relocalization and GTP loading.

ROCK-mediated phosphorylation of MYPT1 is pre-

vented, leaving MLCP in its active state. Here the

dephosphorylation of MLC is the predominant path-

way, while phosphorylation of MLC by MLCK is

diminished (lower panel).

BLOOD, 14 NOVEMBER 2013 x VOLUME 122, NUMBER 20 cAMP ACTIVATES MYOSIN LIGHT CHAIN PHOSPHATASE 3543

References

1. Hartwig JH, Barkalow K, Azim A, Italiano J. Theelegant platelet: signals controlling actinassembly. Thromb Haemost. 1999;82(2):392-398.

2. Siess W. Molecular mechanisms of plateletactivation. Physiol Rev. 1989;69(1):58-178.

3. Klages B, Brandt U, Simon MI, Schultz G,Offermanns S. Activation of G12/G13 results inshape change and Rho/Rho-kinase-mediatedmyosin light chain phosphorylation inmouse platelets. J Cell Biol. 1999;144(4):745-754.

4. Lokeshwar VB, Bourguignon LY. The involvementof Ca21 and myosin light chain kinase incollagen-induced platelet activation. Cell Biol IntRep. 1992;16(9):883-897.

5. Oury C, Sticker E, Cornelissen H, De Vos R,Vermylen J, Hoylaerts MF. ATP augments vonWillebrand factor-dependent shear-inducedplatelet aggregation through Ca21-calmodulinand myosin light chain kinase activation. J BiolChem. 2004;279(25):26266-26273.

6. Riondino S, Gazzaniga PP, Pulcinelli FM.Convulxin induces platelet shape changethrough myosin light chain kinase and Rhokinase. Eur J Biochem. 2002;269(23):5878-5884.

7. Hathaway DR, Eaton CR, Adelstein RS.Regulation of human platelet myosin light chainkinase by the catalytic subunit of cyclic AMP-dependent protein kinase. Nature. 1981;291(5812):252-256.

8. Daniel JL, Molish IR, Rigmaiden M, Stewart G.Evidence for a role of myosin phosphorylation inthe initiation of the platelet shape changeresponse. J Biol Chem. 1984;259(15):9826-9831.

9. Suzuki Y, Yamamoto M, Wada H, et al. Agonist-induced regulation of myosin phosphatase activityin human platelets through activation of Rho-kinase. Blood. 1999;93(10):3408-3417.

10. Bauer M, Retzer M, Wilde JI, et al.Dichotomous regulation of myosinphosphorylation and shape change by Rho-kinase and calcium in intact human platelets.Blood. 1999;94(5):1665-1672.

11. Paul BZ, Daniel JL, Kunapuli SP. Platelet shapechange is mediated by both calcium-dependentand -independent signaling pathways. Role ofp160 Rho-associated coiled-coil-containingprotein kinase in platelet shape change. J BiolChem. 1999;274(40):28293-28300.

12. Wraith KS, Magwenzi S, Aburima A, Wen Y,Leake D, Naseem KM. Oxidized low-densitylipoproteins induce rapid platelet activation andshape change through tyrosine kinase and Rhokinase-signaling pathways. Blood. 2013;122(4):580-589.

13. Hartshorne DJ, Ito M, Erdodi F. Myosin light chainphosphatase: subunit composition, interactionsand regulation. J Muscle Res Cell Motil. 1998;19(4):325-341.

14. Smolenski A. Novel roles of cAMP/cGMP-dependent signaling in platelets. J ThrombHaemost. 2012;10(2):167-176.

15. Pannocchia A, Hardisty RM. Cyclic AMP inhibitsplatelet activation independently of its effect oncytosolic free calcium. Biochem Biophys ResCommun. 1985;127(1):339-345.

16. Graber SE, Hawiger J. Evidence that changes inplatelet cyclic AMP levels regulate the fibrinogenreceptor on human platelets. J Biol Chem. 1982;257(24):14606-14609.

17. Feinstein MB, Fraser C. Human platelet secretionand aggregation induced by calcium ionophores.

Inhibition by PGE1 and dibutyryl cyclic AMP.J Gen Physiol. 1975;66(5):561-581.

18. Salzman EW, Kensler PC, Levine L. Cyclic39,59-adenosine monophosphate in human bloodplatelets. IV. Regulatory role of cyclic amp inplatelet function. Ann N Y Acad Sci. 1972;201:61-71.

19. Sim DS, Merrill-Skoloff G, Furie BC, Furie B,Flaumenhaft R. Initial accumulation of plateletsduring arterial thrombus formation in vivo isinhibited by elevation of basal cAMP levels. Blood.2004;103(6):2127-2134.

20. Gerry ABSL, Satchell L, Leake DS. A novelmethod for production of lipid hydroperoxide- oroxysterol-rich low-density lipoprotein.Atherosclerosis. 2008;197(2):579-587.

21. Roberts W, Magwenzi S, Aburima A, NaseemKM. Thrombospondin-1 induces plateletactivation through CD36-dependent inhibition ofthe cAMP/protein kinase A signaling cascade.Blood. 2010;116(20):4297-4306.

22. Hall KJ, Jones ML, Poole AW. Coincidentregulation of PKCdelta in human platelets byphosphorylation of Tyr311 and Tyr565 andphospholipase C signalling. Biochem J. 2007;406(3):501-509.

23. McAvoy T, Nairn AC. Chapter 18: Unit 18. Serine/threonine protein phosphatase assays. In:Ausubel FM, Brent R, Kingston RE, et al, eds.Current Protocols in Molecular Biology. Hoboken,NJ: Wiley; 2010.

24. Jones RL, Wise H, Clark R, Whiting RL, Bley KR.Investigation of the prostacyclin (IP) receptorantagonist RO1138452 on isolated blood vesseland platelet preparations. Br J Pharmacol. 2006;149(1):110-120.

25. Gambaryan S, Geiger J, Schwarz UR, et al.Potent inhibition of human platelets by cGMPanalogs independent of cGMP-dependentprotein kinase. Blood. 2004;103(7):2593-2600.

26. Ishizaki T, Uehata M, Tamechika I, et al.Pharmacological properties of Y-27632, a specificinhibitor of rho-associated kinases. MolPharmacol. 2000;57(5):976-983.

27. Shimizu Y, Thumkeo D, Keel J, et al. ROCK-Iregulates closure of the eyelids and ventral bodywall by inducing assembly of actomyosin bundles.J Cell Biol. 2005;168(6):941-953.

28. Davies SP, Reddy H, Caivano M, Cohen P.Specificity and mechanism of action of somecommonly used protein kinase inhibitors.Biochem J. 2000;351(Pt 1):95-105.

29. Ren XD, Kiosses WB, Schwartz MA. Regulationof the small GTP-binding protein Rho by celladhesion and the cytoskeleton. EMBO J. 1999;18(3):578-585.

30. Manganello JM, Djellas Y, Borg C, Antonakis K,Le Breton GC. Cyclic AMP-dependentphosphorylation of thromboxane A(2) receptor-associated Galpha(13). J Biol Chem. 1999;274(39):28003-28010.

31. Schoenwaelder SM, Burridge K. Evidencefor a calpeptin-sensitive protein-tyrosinephosphatase upstream of the small GTPaseRho. A novel role for the calpain inhibitorcalpeptin in the inhibition of protein-tyrosinephosphatases. J Biol Chem. 1999;274(20):14359-14367.

32. Schoenwaelder SM, Petch LA, Williamson D,Shen R, Feng G-S, Burridge K. The proteintyrosine phosphatase Shp-2 regulatesRhoA activity. Curr Biol. 2000;10(23):1523-1526.

33. Forget MA, Desrosiers RR, Gingras D, BeliveauR. Phosphorylation states of Cdc42 and RhoAregulate their interactions with Rho GDP

dissociation inhibitor and their extraction frombiological membranes. Biochem J. 2002;361(Pt2):243-254.

34. Ellerbroek SM, Wennerberg K, Burridge K.Serine phosphorylation negatively regulatesRhoA in vivo. J Biol Chem. 2003;278(21):19023-19031.

35. Lang P, Gesbert F, Delespine-Carmagnat M,Stancou R, Pouchelet M, Bertoglio J. Proteinkinase A phosphorylation of RhoA mediates themorphological and functional effects of cyclic AMPin cytotoxic lymphocytes. EMBO J. 1996;15(3):510-519.

36. Dong JM, Leung T, Manser E, Lim L. cAMP-induced morphological changes are counteractedby the activated RhoA small GTPase and the Rhokinase ROKalpha. J Biol Chem. 1998;273(35):22554-22562.

37. Adamson P, Paterson HF, Hall A. Intracellularlocalization of the P21rho proteins. J Cell Biol.1992;119(3):617-627.

38. Takaishi K, Sasaki T, Kameyama T,Tsukita S, Tsukita S, Takai Y. Translocation ofactivated Rho from the cytoplasm to membraneruffling area, cell-cell adhesion sites andcleavage furrows. Oncogene. 1995;11(1):39-48.

39. Rolli-Derkinderen M, Sauzeau V, Boyer L, et al.Phosphorylation of serine 188 protects RhoA fromubiquitin/proteasome-mediated degradation invascular smooth muscle cells. Circ Res. 2005;96(11):1152-1160.

40. Qiao J, Huang F, Lum H. PKA inhibits RhoAactivation: a protection mechanism againstendothelial barrier dysfunction. Am J PhysiolLung Cell Mol Physiol. 2003;284(6):L972-L980.

41. Casey PJ, Thissen JA, Moomaw JF. Enzymaticmodification of proteins with a geranylgeranylisoprenoid. Proc Natl Acad Sci USA. 1991;88(19):8631-8635.

42. Kawano Y, Fukata Y, Oshiro N, et al.Phosphorylation of myosin-binding subunit (MBS)of myosin phosphatase by Rho-kinase in vivo.J Cell Biol. 1999;147(5):1023-1038.

43. Rowley JW, Oler AJ, Tolley ND, et al. Genome-wide RNA-seq analysis of human and mouseplatelet transcriptomes. Blood. 2011;118(14):e101-e111.

44. Patil SB, Tsunoda Y, Pawar MD, Bitar KN.Translocation and association of ROCK-II withRhoA and HSP27 during contraction of rabbitcolon smooth muscle cells. Biochem BiophysRes Commun. 2004;319(1):95-102.

45. Shin HM, Je HD, Gallant C, Tao TC, HartshorneDJ, Ito M, Morgan KG. Differential associationand localization of myosin phosphatasesubunits during agonist-induced signaltransduction in smooth muscle. Circ Res. 2002;90(5):546-553.

46. Pleines IHI, Hagedorn I, Gupta S, et al.Megakaryocyte-specific RhoA deficiency causesmacrothrombocytopenia and defective plateletactivation in hemostasis and thrombosis. Blood.2012;119(4):1054-1063.

47. Gratacap MP, Payrastre B, Nieswandt B,Offermanns S. Differential regulation of Rho andRac through heterotrimeric G-proteins and cyclicnucleotides. J Biol Chem. 2001;276(51):47906-47913.

48. Murthy KS, Zhou H, Grider JR, Makhlouf GM.Inhibition of sustained smooth muscle contractionby PKA and PKG preferentially mediated byphosphorylation of RhoA. Am J PhysiolGastrointest Liver Physiol. 2003;284(6):G1006-G1016.

3544 ABURIMA et al BLOOD, 14 NOVEMBER 2013 x VOLUME 122, NUMBER 20

49. Nakamura K, Koga Y, Sakai H, Homma K, IkebeM. cGMP-dependent relaxation of smooth muscleis coupled with the change in the phosphorylationof myosin phosphatase. Circ Res. 2007;101(7):712-722.

50. Kimura K, Ito M, Amano M, et al. Regulation ofmyosin phosphatase by Rho and Rho-associatedkinase (Rho-kinase). Science. 1996;273(5272):245-248.

51. Wilson DP, Susnjar M, Kiss E, Sutherland C,Walsh MP. Thromboxane A2-induced contractionof rat caudal arterial smooth muscle involvesactivation of Ca21 entry and Ca21 sensitization:Rho-associated kinase-mediated phosphorylationof MYPT1 at Thr-855, but not Thr-697. Biochem J.2005;389(Pt 3):763-774.

52. Nakai K, Suzuki Y, Kihira H, et al. Regulation ofmyosin phosphatase through phosphorylation of

the myosin-binding subunit in platelet activation.Blood. 1997;90(10):3936-3942.

53. Azam MA, Yoshioka K, Ohkura S, Takuwa N,Sugimoto N, Sato K, Takuwa Y. Ca21-independent, inhibitory effects of cyclic adenosine59-monophosphate on Ca21 regulation ofphosphoinositide 3-kinase C2alpha, Rho, andmyosin phosphatase in vascular smooth muscle.J Pharmacol Exp Ther. 2007;320(2):907-916.

BLOOD, 14 NOVEMBER 2013 x VOLUME 122, NUMBER 20 cAMP ACTIVATES MYOSIN LIGHT CHAIN PHOSPHATASE 3545