www.elsevier.com/locate/yabbi

Archives of Biochemistry and Biophysics 465 (2007) 16–25

ABB

Involvement of SNARE proteins in thrombin-inducedplatelet aggregation: Evidence for the relevance of Ca2+ entry

Isaac Jardin a,1, Nidhal Ben Amor c,1, Juan M. Hernandez-Cruz b,Gines M. Salido a, Juan A. Rosado a,*

a Department of Physiology, Cellular Physiology Research Group, University of Extremadura, Av. Universidad s/n,

Caceres 10071, Spainb Clinical Analysis Laboratory, Caceres, Spain

c Unite de Recherche de Biochimie, Institute Superieur de Biotechnologie, 5019 Monastir, Tunisia

Received 4 April 2007, and in revised form 27 April 2007Available online 21 May 2007

Abstract

Thrombin induces platelet activation through a variety of intracellular mechanisms, including Ca2+ mobilization. The protein of theexocytotic machinery SNAP-25, but not VAMPs, is required for store-operated Ca2+ entry, the main mechanism for Ca2+ influx in plate-lets. Hence, we have investigated the role of the SNAP-25 and VAMPs in thrombin-induced platelet aggregation. Platelet stimulationwith thrombin or selective activation of thrombin receptors PAR-1, PAR-4 or GPIb-IX-V results in platelet aggregation that, exceptfor GPIb-IX-V receptor, requires Ca2+ entry for full activation. Depletion of the intracellular Ca2+ stores using pharmacological toolswas unable to induce aggregation except when cytosolic Ca2+ concentration reached a critical level (around 1.5 lM). Electrotransjectionof cells with anti-SNAP-25 antibody reduced thrombin-evoked platelet aggregation, while electrotransjection of anti-VAMP-1, -2 and -3antibody had no effect. These findings support a role for SNAP-25 but not VAMP-1, -2 and -3 in platelet aggregation, which is likelymediated by the regulation of Ca2+ mobilization in human platelets.� 2007 Elsevier Inc. All rights reserved.

Keywords: Thrombin; SNAP-25; VAMP; Actin cytoskeleton; Aggregation; Platelets; Intracellular stores

Thrombin is a physiological agonist that stimulateshuman platelets by activation of two G-protein-coupledprotease-activated receptors (PAR)2, PAR-1 and PAR-4and the leucine-rich glycoprotein receptor GPIb-IX-V [1–3].PAR-1 and PAR-4 are cleaved by thrombin at specificsites in the extracellular domain, resulting in a new N-ter-minal sequence that acts as a ‘‘tethered ligand’’ that acti-vates transmembrane signaling [2]. Peptides containingthe new N-terminal sequence of cleaved PAR-1 and

0003-9861/$ - see front matter � 2007 Elsevier Inc. All rights reserved.

doi:10.1016/j.abb.2007.04.038

* Corresponding author. Fax: +34 927 257110.E-mail address: jarosado@unex.es (J.A. Rosado).

1 These authors contributed equally to this work.2 Abbreviations used: PAR, protease-activated receptor; VAMP, vesicle-

associated membrane protein; SNARE, soluble N-ethylmaleimide-sensi-tive factor attachment protein receptor; TG, thapsigargin.

PAR-4 receptors, such as SFLLRN and AYPGKF, respec-tively, are potent and selective activators of PAR-1 andPAR-4 and mimic thrombin-evoked platelet responses[4,5].

The events associated with platelet activation by throm-bin include activation of membrane receptors, shapechange, granular secretion, cytoskeletal remodeling, andaggregation. Changes in shape, and aggregation thataccompanies platelet activation, are dependent on theassembly and reorganization of the actin cytoskeleton,which has been reported to be necessary for the irreversibil-ity of platelet aggregation induced by strong agonists, suchas thrombin [6]. Using inhibitors of actin polymerization,such as latrunculin A, the actin cytoskeleton has beenreported to differentially regulate a-granule and densegranule secretion. In addition, latrunculin A-facilitated

I. Jardin et al. / Archives of Biochemistry and Biophysics 465 (2007) 16–25 17

a-granule secretion was inhibited by antibodies directed atvesicle-associated membrane protein (VAMP), demon-strating that the actin cytoskeleton supports soluble N-eth-ylmaleimide-sensitive factor attachment protein receptor(SNARE) protein-dependent membrane fusion [7]. Humanplatelets also express the synaptosome-associated protein(SNAP-25), which has been shown to associate to the cor-tical actin cytoskeleton through the SNAP-25 interactingprotein SNIP [8].

Thrombin-induced platelet activation is mediated bycytosolic Ca2+ mobilization, which consists of Ca2+ releasefrom the intracellular stores and entry through plasmamembrane channels. Two independent agonist-releasableCa2+ stores have been described in human platelets, whichare differentiated by the distinct sensitivity of the sarco/endoplasmic reticulum Ca2+ ATPase (SERCA) isoformsexpressed in each store to thapsigargin (TG) and 2,5-di-(tert-butyl)-1,4-hydroquinone (TBHQ) [9–12]. The mainstore, the dense tubular system (DTS), is an analogue ofthe endoplasmic reticulum and is releasable by physiologi-cal agonists like thrombin, ADP or vasopressin [13]. TheDTS expresses SERCA2b, which shows a high sensitivityto TG and is insensitive to TBHQ [9,14]. The second storeis acidic in nature and releasable by thrombin but not byADP or vasopressin [15]. The acidic stores expressSERCA3, which has been reported to be sensitive toTBHQ and show a lower sensitivity to TG than SERCA2b[13–17]. Ca2+ mobilization has been shown to be requiredfor platelet secretion [3], a mechanism regulated bySNARE proteins [18,19] that has been shown to play animportant role in aggregation. Among other SNAREs,human platelets express SNAP-25 and VAMP-1, -2, and-3 [18–20]. We have previously found that SNAP-25, butnot VAMPs, is involved in the regulation of cytosolicCa2+ mobilization through the activation and maintenanceof store-operated Ca2+ entry in human platelets [21].

In the present study we have investigated the role of theSNARE proteins SNAP-25 and VAMP-1, -2 and -3 inplatelet aggregation and the role of cytosolic Ca2+ concen-tration and actin filament reorganization in thrombin-induced platelet aggregation.

Materials and methods

Materials

Apyrase (grade VII), EGTA, aspirin, bovine serum albumin, throm-bin, thapsigargin (TG), paraformaldehyde, ionomycin (Iono), NonidetP40, fluorescein isothiocyanate-labeled phalloidin, were from Sigma(Madrid, Spain). SFLLRN and AYPGKF were from Bachem (Mersey-side, UK). TBHQ was from Alexis (Nottingham, UK). Anti SNAP-25antibody (C-18), anti-VAMP antibody (FL-118), horseradish peroxidase-conjugated donkey anti-goat IgG antibody and horseradish peroxidase-conjugated goat anti-rabbit IgG antibody were from Santa Cruz (SantaCruz, CA, USA). Protein A–agarose was from Upstate Biotechnology Inc.(Madrid, Spain). Enhanced chemiluminescence detection reagents werefrom Pierce (Cheshire, UK). Hyperfilm ECL was from Amersham(Arlington Heights, IL, USA). All other reagents were purchased fromPanreac (Barcelona, Spain). The thrombin preparation (specific activity

P2000 NIH units/mg protein) was predominantly a-thrombin, containingminimum autolytic digestion products, according to the manufacturer’sinstructions; therefore, most of the effects shown in the present studyshould be attributed to a-thrombin.

Platelet preparation

Blood was obtained from healthy volunteers and mixed with one-sixthvolume of acid/citrate dextrose anticoagulant containing (in mM): 85sodium citrate, 78 citric acid and 111 D-glucose. Platelet-rich plasma wasthen prepared by centrifugation for 5 min at 700g and aspirin (100 lM)and apyrase (40 lg/mL) added. Cells were then collected by centrifugationat 350g for 20 min and resuspended in Hepes-buffered saline (HBS) con-taining (in mM): 145 NaCl, 10 Hepes, 10 D-glucose, 5 KCl, 1 MgSO4, pH7.45, and supplemented with 0.1% w/v bovine serum albumin and 40 lg/mL apyrase.

Cell viability

Cell viability was assessed using calcein and trypan blue. For calceinloading, cells were incubated for 30 min with 5 lM calcein-AM at 37 �C,centrifuged and the pellet was resuspended in fresh HBS. Fluorescence wasrecorded from 2 ml aliquots using a spectrophotometer (Varian Ltd.,Madrid, Spain). Samples were excited at 494 nm and the resulting fluo-rescence was measured at 535 nm. The results obtained with calcein wereconfirmed using the trypan blue exclusion technique. 95% of cells wasviable in our platelet suspensions.

Platelet aggregation

The percentage, rate and lag-time of aggregation in washed plateletswere monitored using a Chronolog (Havertown, Pa, USA) aggregometerat 37 �C under stirring at 1200 rpm [22]. The percentage of aggregation oramplitude is estimated as the percentage of the difference in light trans-mission between the platelet suspension in HBS and HBS alone andindicates the percentage of platelets that aggregate in response to anagonist. Resting platelets in suspension are arbitrarily considered by theaggregometer as 0% aggregation and HBS is considered to be 100%aggregation. The rate, or slope, of the aggregation is the % change ofaggregation per minute.

Measurement of intracellular free calcium concentration ([Ca2+]i)

Human platelets were loaded with fura-2 by incubation with 2 lMfura-2/AM for 45 min at 37 �C. Fluorescence was recorded from 2 mLaliquots of magnetically stirred cellular suspension (2 · 108 cells/mL) at37 �C using a Cary Eclipse spectrophotometer (Varian Ltd., Madrid,Spain) with excitation wavelengths of 340 and 380 nm and emission at505 nm. Changes in [Ca2+]i were monitored using the fura-2 340/380fluorescence ratio and calibrated according to a established method [23].

Measurement of F-actin content

The F-actin content of resting and activated platelets was determinedaccording to a previously published procedure [24]. Briefly, washedplatelets (2 · 107 cells/mL) were activated in HBS. Samples of plateletsuspensions (200 lL) were transferred to 200 lL of ice-cold 3% (w/v)formaldehyde in PBS for 10 min. Fixed platelets were permeabilized byincubation for 10 min with 0.025% (v/v) Nonidet P40 detergent dissolvedin PBS. Platelets were then incubated for 30 min with FITC-labeledphalloidin (1 lM) in PBS supplemented with 0.5% (w/v) BSA. Plateletswere then collected by centrifugation in a Galaxy 7D centrifuge (VWRInternational, Fontenay sous Bois, France) for 60 s at 3000g and resus-pended in PBS. Staining of 2 · 107 cells/mL was measured using a fluo-rescence spectrophotometer (Shimadzu, Japan). Samples were excited at496 nm and emission was at 516 nm.

Fig. 1. Platelet aggregation by activation of thrombin receptors. Humanplatelets were suspended in HBS containing 1 mM Ca2+ (black traces) orin a Ca2+-free HBS (grey traces) and then stimulated with 1 U/mLthrombin (a), 10 lM SFLLRN (b), 500 lM AYPGKF (c) or 10 lMSFLLRN (SF) in combination with 500 lM AYPGKF (AYP) followed bytreatment with either 1 U/mL (d) or 0.1 U/mL (e) thrombin. Aggregationof human platelets was induced at a shear rate of 1200 rpm at 37 �C in anaggregometer as described in Materials and methods. Traces shown arerepresentative of 6–10 separate experiments.

18 I. Jardin et al. / Archives of Biochemistry and Biophysics 465 (2007) 16–25

Reversible electroporation procedure

The platelet suspension was transferred to an electroporationchamber containing antibodies at a final concentration of 1 lg/mL, andthe antibodies were transjected according to published methods [25,26].Reversible electropermeabilization was performed at 4 kV/cm at a set-ting of 25-lF capacitance and was achieved by 7 pulses using a Bio-Rad Gene Pulser Xcell electroporation system (Bio-Rad, Hercules, CA,USA). Following electroporation, cells were incubated with antibodiesfor an additional 60 min at 37 �C and were centrifuged at 350g for20 min and resuspended in Hepes-buffered saline (HBS) prior to theexperiments.

Immunoprecipitation and Western blotting

The immunoprecipitation and Western blotting were performed asdescribed previously [27]. Briefly, 500 lL aliquots of platelet suspension(2 · 109 cell/mL) non-permeabilized or electropermeabilized and incu-bated in the presence of anti-SNAP-25 or anti-VAMP antibodies werelysed with an equal volume of RIPA buffer, pH 7.2, containing316 mM NaCl, 20 mM Tris, 2 mM EGTA, 0.2% SDS, 2% sodiumdeoxycholate, 2% Triton X-100, 2 mM Na3VO4, 2 mM phenylmethyl-sulfonyl fluoride, 100 lg/mL leupeptin and 10 mM benzamidine. Ali-quots of platelet lysates (1 mL) were incubated with 25 lL of proteinA–agarose overnight at 4 �C on a rocking platform. Samples werecentrifuged for 60 s at 3000g and the proteins in the pellet wereresolved by 15% SDS–PAGE and electrophoretically transferred ontonitrocellulose membranes for subsequent probing. Blots were incubatedovernight with 10% (w/v) BSA in Tris-buffered saline with 0.1% Tween20 (TBST) to block residual protein binding sites. Immunodetection ofSNAP-25 and VAMP was achieved using the anti-SNAP-25 or anti-VAMP antibodies diluted 1:1000 in TBST for 2 h. The primary anti-body was removed and blots were washed six times for 5 min eachwith TBST. To detect the primary antibody, blots were incubated for45 min with horseradish peroxidase-conjugated ovine anti-rabbit IgGantibody or horseradish peroxidase-conjugated donkey anti-goat IgGantibody diluted 1:10,000 in TBST and then exposed to enhancedchemiluminescence reagents for 4 min. Blots were then exposed tophotographic films. The density of bands on the film was measuredusing a scanning densitometry.

Statistical analysis

Analysis of statistical significance was performed using one-wayanalysis of variance combined with the Dunnett or Tukey test. Forcomparison between two groups Student’s t test was used. p < 0.05 wasconsidered to be significant for a difference.

Results

Platelet aggregation induced by activation of thrombin

receptors

As previously reported [28], treatment of human plate-lets suspended in a medium containing 1 mM Ca2+ with1 U/ml thrombin induced rapid aggregation characterizedby an initial and rapid change in platelet shape, indicatedby a small decrease in light transmission, followed by alarge increase in light transmission as the platelets aggre-gated (Fig. 1a; n = 10). The effect of thrombin on plateletaggregation was found to be significantly reduced in theabsence of extracellular Ca2+ (100 lM EGTA was addedto the medium; Fig. 1a and Tables 1 vs 2; p < 0.05Student’s t test).

SFLLRN corresponds to the new amino terminus of thehigh-affinity thrombin receptor PAR-1 after cleavage bythrombin. The concentration of SFLLRN (10 lM) waschosen because, by itself, it caused maximal cellular effects[29]. In agreement with previous studies, treatment ofplatelets in the presence of 1 mM extracellular Ca2+ withSFLLRN stimulated only a primary phase of aggregation[30] (Fig. 1b; Table 1; n = 10). As for thrombin, both therate and amplitude of SFLLRN-induced aggregation wasreduced in the absence of extracellular Ca2+ (Fig. 1b;Tables 1 vs 2; p < 0.05; n = 10).

Treatment of human platelets in the presence of 1 mMexternal Ca2+ with 500 lM AYPGKF, a potent and selec-tive PAR-4-activating peptide [4,31], induced a greater

Table 1Platelet aggregation by occupation of thrombin receptors in a medium containing 1 mM Ca2+

Stimulatory agent Lag-time (min) % rate % aggregation

Thrombin 1 U/mL 0.03 ± 0.01 57.40 ± 4.38 91.00 ± 2.22SFLLRN 10 lM 0.38 ± 0.10* 31.20 ± 2.16* 15.75 ± 0.67*

AYPGKF 500 lM 0.19 ± 0.09* 35.40 ± 1.64* 54.75 ± 1.70*

SFLLRN + AYPGKF 0.20 ± 0.01* 27.00 ± 5.23* 48.02 ± 2.60*

[PARs-desensitized cells] Thrombin 1 U/mL 1.05 ± 0.01* 46.00 ± 8.14 39.00 ± 6.65*

Human platelets were suspended in a HBS containing 1 mM Ca2+ and then stimulated with 1 U/mL thrombin, 10 lM SFLLRN, or 500 lM AYPGKF. Inaddition, platelets have been treated with SFLLRN (10 lM) in combination with AYPGKF (500 lM) to desensitize PAR-1 and PAR-4 and 10 min laterwere stimulated with thrombin 1 U/mL. Platelet aggregation was determined as described under Materials and methods. Values given are presented asmeans ± SEM of 10 separate determinations.

* p < 0.05 (ANOVA) compared to thrombin-induced response.

Table 2Platelet aggregation by occupation of thrombin receptors in a Ca2+ -free medium

Stimulatory agent Lag-time (min) % rate % aggregation

Thrombin 1 U/mL 0.05 ± 0.01 58.42 ± 6.05 86.85 ± 0.96*

SFLLRN 10 lM 0.36 ± 0.23 18.25 ± 3.40* 12.75 ± 0.90*

AYPGKF 500 lM 1.23 ± 0.39* 18.60 ± 3.74* 17.80 ± 3.52*

SFLLRN + AYPGKF 0.13 ± 0.07 46.33 ± 9.79 36.00 ± 9.50*

[PARs-desensitized cells] Thrombin 1 U/mL 0.28 ± 0.10 41.50 ± 3.18 46.33 ± 10.24

Human platelets were suspended in a Ca2+-free HBS (100 lM EGTA was added) and then stimulated with 1 U/mL thrombin, 10 lM SFLLRN or 500 lMAYPGKF. In addition, platelets have been treated with SFLLRN (10 lM) in combination with AYPGKF (500 lM) to desensitize PAR-1 and PAR-4 and10 min later were stimulated with thrombin 1 U/mL. Platelet aggregation was determined as described under Materials and methods. Values given arepresented as means ± SEM of 10 separate determinations.

* p < 0.05 compared to the results in the presence of 1 mM extracellular Ca2+ (Table 1).

I. Jardin et al. / Archives of Biochemistry and Biophysics 465 (2007) 16–25 19

platelet aggregation than SFLLRN (Fig. 1c and Table 1;n = 10). As for SFLLRN, 500 lM AYPGKF was chosento induce maximal cellular responses [29]. AYPGKF-induced response was significantly reduced when plateletswere suspended in a Ca2+-free medium (Fig. 1c and Tables1 vs 2; p < 0.05).

Characterization of platelet aggregation upon activationof the GPIb-IX-V receptor by thrombin was investigatedby desensitizing PAR-1 and PAR-4 receptors as previouslydescribed [32]. Desensitization was achieved by platelettreatment with SFLLRN (10 lM) in combination withAYPGKF (500 lM). Desensitization of PAR-1 andPAR-4 was confirmed by the inability of SFLLRN andAYPGKF to induce further aggregation in PAR-desensi-tized platelets (data not shown). Platelet stimulation withthrombin (1 U/mL) when PAR-1 and PAR-4 receptorshad been desensitized induced further aggregation, whichshows characteristics of irreversible aggregation and,together with the effect induced by prior treatment withSFLLRN and AYPGKF, reached the amplitude observedafter platelet stimulation with thrombin alone in non-desensitized cells (Fig. 1d and Table 1; n = 10). To investi-gate whether thrombin-induced response in PAR desensi-tized cells involved the competitive displacement of thePAR agonists from their respective receptors (due to thehigher affinity of thrombin [33]), we repeated these experi-ments using a lower concentration of thrombin (0.1 U/mL). As shown in Fig. 1e, stimulation of PAR desensitizedcells with 0.1 U/mL thrombin induced a response that iscomparable with that observed after addition of 1 U/mL,

which suggests that under our experimental procedure theeffect of thrombin is mostly mediated by binding to thehigh affinity GPIb-IX-V receptor (n = 6).

Our results indicate that stimulation of platelets in amedium containing 1 mM Ca2+ with SFLLRN in combi-nation with AYPGKF induce a response that was similarto that observed with AYPGKF alone (Fig. 1d; Tables 1and 2), which suggests that aggregation-induced by activa-tion of PAR-1 and PAR-4 might share a common signaltransduction pathway. In the absence of extracellularCa2+ the effect of treatment with SFLLRN + AYPGKFwas significantly smaller than in the presence of 1 mMCa2+ (p < 0.05), which is consistent with the previousresults in the absence of extracellular Ca2+, but furtherstimulation with thrombin was similar in the absence orpresence of external Ca2+ (Fig. 1d; Table 2).

To further investigate the role of cytosolic Ca2+ on plate-let aggregation we have investigated the effect of selectivedepletion of the intracellular Ca2+ stores. Treatment ofplatelets with 10 nM TG, which specifically inhibits SER-CA2b located in the DTS [10,11], was unable to induceaggregation, at least after 500 s of stimulation both in theabsence and presence of extracellular Ca2+ (Fig. 2a;n = 10). Similar results were observed when cells were stim-ulated with 20 lM TBHQ to deplete the acidic stores [9–11];Fig. 2b; n = 6). TBHQ induced a small decrease in lighttransmission, which might be indicative of platelet shapechange, but was unable to stimulate platelet aggregationboth in the absence and presence of extracellular Ca2+.Treatment of platelets with TG induced a sustained increase

Fig. 2. Platelet aggregation by depletion of the intracellular Ca2+ stores. Human platelets were suspended in HBS containing 1 mM Ca2+ (black traces) orin a Ca2+-free HBS (grey traces) and then stimulated with 10 nM TG (a and d), 20 lM TBHQ (b and e) or 1 lM TG in combination with 50 nM Iono (cand f). Aggregation of human platelets was induced at a shear rate of 1200 rpm at 37 �C in a Chronolog aggregometer as described in Materials andmethods. Elevations in [Ca2+]i were monitored using the 340/380 nm ratio and traces were calibrated in terms of [Ca2+]i. Traces shown are representativeof 6–10 separate experiments.

20 I. Jardin et al. / Archives of Biochemistry and Biophysics 465 (2007) 16–25

in [Ca2+]i, reaching a plateau at 106 ± 5 and 161 ± 12 nM inthe absence and presence of extracellular Ca2+, respectively(means ± SEM; Fig. 2d). After addition of TBHQ the acidicstores discharged more rapidly, as shown in the absence ofexternal Ca2+, then [Ca2+]i reached a plateau at 107 ± 7and 208 ± 15 nM in the absence and presence of extracellu-lar Ca2+ (Fig. 2e). In the presence of 1 mM extracellularCa2+, rapid depletion of both stores using a combinationof 1 lM TG plus a low concentration of Iono (50 nM,needed for extensive depletion of both Ca2+ pools, which,as reported above, show high and low Ca2+ leakage rates[9,15]) resulted in an initial and small decrease in light trans-mission, indicative of shape change, followed by a slowincrease (16 ± 4% rate) in light transmission as the plateletsaggregated reaching a 45 ± 9% of aggregation (Fig. 2c).Treatment of platelets with 1 lM TG + 50 nM Iono induceda transient increase in [Ca2+]i, reaching an initial peak [Ca2+]ielevation of 1568 ± 45 nM and then [Ca2+]i reduced reach-ing a plateau at 1235 ± 65 nM (Fig. 2f). Removal of extracel-lular Ca2+ abolished the effect of TG + Iono on plateletaggregation (Fig. 2c). In the absence of extracellular Ca2+

treatment with TG + Iono resulted in a transient increasein [Ca2+]i, inducing an initial peak [Ca2+]i elevation of294 ± 16 nM and returning back to basal levels (Fig. 2f).These findings suggest that release of Ca2+ accumulated inthe stores per se is not sufficient to stimulated platelet aggre-gation, which, instead, might require a more relevant andsustained increase in [Ca2+]i.

To assess this possibility cells were treated with a higherconcentration of Iono (500 nM), to facilitate calcium entry,and we increased the extracellular free Ca2+ concentration([Ca2+]o) at the rate of 1 mM every 90 s, in order to test thecytosolic level of Ca2+ that initiates platelet aggregation.As shown in Fig. 3b, addition of 1 mM extracellularCa2+ to platelet suspensions previously treated with500 nM Iono resulted in a continuous rise in [Ca2+]i asthe driving force for Ca2+ entry increased. No change inlight transmission was detected at [Ca2+]o 6 8 mM([Ca2+]i = 1390 ± 65 nM; see Fig. 3a and b). However,when [Ca2+]o exceeded this level platelet aggregation initi-ated at a rate of 40%, reaching a 73% of maximal aggrega-tion (Fig. 3a). Under these conditions [Ca2+]i reached alevel of 1560 ± 76 nM (Fig. 3b). Interestingly, treatmentof platelets with Iono in the presence of 9 mM extracellularCa2+ initiates aggregation only when the [Ca2+]i reached asimilar level (platelet aggregated at a rate of 39%, reachinga maximum at 69%; Fig. 3c and d). These findings suggestthat platelet aggregation might be induced by high levels of[Ca2+]i perhaps through the activation of a low-affinityCa2+-dependent intracellular pathway.

Actin filament reorganization induced by activation of

thrombin receptors

Platelet aggregation is a process that requires extensiveremodeling of the actin cytoskeleton resulting in loss of

Fig. 3. Effect of rises in cytosolic free Ca2+ concentration in platelet aggregation. Human platelets were suspended in a Ca2+-free HBS and stimulated with500 nM Iono followed by either addition of 1 mM CaCl2 to the platelet suspension at intervals of 90 s after the addition of Iono until aggregation started(after 9 additions of 1 mM CaCl2; a and b) or a single addition of 9 mM CaCl2 30 s after Iono (c and d). Aggregation of human platelets was induced at ashear rate of 1200 rpm at 37 �C in a Chronolog aggregometer as described in Materials and methods. Elevations in [Ca2+]i were monitored using the 340/380 nm ratio and traces were calibrated in terms of [Ca2+]i. Traces shown are representative of four separate experiments.

I. Jardin et al. / Archives of Biochemistry and Biophysics 465 (2007) 16–25 21

platelet discoid shape and transformation to a sphere withlong and thin filopodia extending out from the cell [3]. Herewe have investigated the role of PAR-1 and PAR-4 recep-tors in actin filament polymerization. Platelet stimulationwith 1 U/mL thrombin for 3 min resulted in a rise in theF-actin content of 182 ± 1.2% of basal, in agreement withprevious studies [24]. Treatment of platelets with 10 lMSFLLRN induced a time-dependent and biphasic increasein actin filament polymerization. As shown in Fig. 4a,SFLLRN-induced a rise in F-actin content that reached amaximum within 10 s with an increase of 132.9 ± 5.3% ofbasal. At later times the F-actin content decreased suchthat it was near basal 3 min after SFLLRN stimulation(n = 6).

Similar results were observed when platelets were stim-ulated with the PAR-4 agonist AYPGKF, although thepeak F-actin level elevation was delayed in comparisonwith SFLLRN-induced response. Treatment of plateletswith 500 lM AYPGKF induced a time-dependent increasein F-actin content. An increase was detectable 10 s aftertreatment with AYPGKF and reached a maximum within30 s with an increase of 136.2 ± 3.8% of basal. At latertimes the F-actin content decreased such that it was nearbasal 3 min after the addition of AYPGKF (Fig. 4b;n = 6).

We have found that platelet treatment with 20 lMTBHQ induce a rapid transient increase in F-actin contentreaching a maximum after 10 s of stimulation with anincrease of 139.5 ± 6.7% and returning back to levels closeto basal after 30 s of stimulation (Fig. 4c). In contrast,treatment of platelets with 10 nM TG resulted in an initialdecrease in the F-actin content, reaching a minimum 30 safter stimulation at 94.8 ± 7.3% of the resting level. TheF-actin content then increased, exceeding basal levels andreaching an increase of 124.6 ± 4.7% of the resting levelafter 3 min of stimulation (Fig. 4d).

Role of SNAP-25 and VAMP in platelet aggregation

To assess the role of SNAP-25 and VAMP in plateletaggregation anti-SNAP-25 and anti-VAMP antibodieswere introduced into platelets using an electropermeabili-zation technique with the aim of inducing functional knockdown of these proteins. The anti-SNAP-25 antibody (C-18)binds to the C-terminus of SNAP-25, where a leucine resi-due (Leu203), which has been shown to be critical for exo-cytosis [34], is located. The anti-VAMP antibody (FL-118)recognizes the amino acids 1–118 of VAMP-1,-2 and -3,which contain the calmodulin- and phospholipid-bindingdomain (VAMP77–94) [35], the SNAP-25 binding domain

Fig. 4. Actin filament polymerization by activation of PAR-1 and PAR-4 thrombin receptors and depletion of platelet Ca2+ stores. Human platelets weresuspended in HBS containing 1 mM Ca2+ and then stimulated with 10 lM SFLLRN (a), 500 lM AYPGKF (b) 20 lM TBHQ (c) or 10 nM TG (d) forseveral periods (10–180 s) before mixing with formaldehyde (3% in PBS). Actin filament content was determined as described in Materials and methodssection. Results shown are presented as percentage of the F-actin content in resting cells and expressed as means ± SEM of six independent experiments.*p < 0.05 (ANOVA) compared with F-actin content in resting cells.

22 I. Jardin et al. / Archives of Biochemistry and Biophysics 465 (2007) 16–25

(amino acids 41–60) and the syntaxin 1A binding sequence(amino acids 31–70) [36]. We have previously used revers-ible electroporation to transfer antibodies into plateletswhile maintaining the physiological integrity of the cells[26]. Human platelets were reversibly electroporated asdescribed in Materials and methods. The presence of thisantibody inside platelets was investigated in samples fromcontrol (non-electropermeabilized) or electropermeabilizedcells, both incubated with 2 lg/mL anti-SNAP-25 or anti-VAMP antibody, by immunoprecipitation without addingany additional anti-SNAP-25 or anti-VAMP antibodiesand subsequent Western blotting with the anti-SNAP-25or anti-VAMP antibody, respectively. As shown inFig. 5c and d, SNAP-25 and VAMP were clearly detectedin cells that had been previously electropermeabilized.Electropermeabilization allowed the anti-SNAP-25 oranti-VAMP antibody to enter the cells and immunoprecip-itate the SNAP-25 and VAMP proteins that were thendetected by Western blotting, which confirms the efficacyof the electrotransjection.

As shown in Fig. 5a, platelets electroporated but nottreated with antibodies showed a greater lag-time betweencell stimulation and the initiation of aggregation; however,the percentage of aggregation in response to thrombin wasnot significantly different from control cells. To investigate

the specificity of electrotransjection of antibodies, the effectof incubation with an antibody directed to a protein notrelated to SNAREs or any other platelet protein wastested. We used both goat and rabbit IgG since this is thenature of the anti-SNAP-25 and anti-VAMP antibodiesused, respectively. Electrotransjection with 2 lg/mL rabbitor goat IgG did not modify platelet aggregation in electro-porated cells (Fig. 5b). These data were used as control forthe experiments performed using cells electrotransjectedwith anti-SNAP-25 or anti-VAMP antibodies. Electro-transjection of anti-SNAP-25 antibody by incubation ofelectroporated platelets with 2 lg/mL anti-SNAP-25 anti-body for 60 min significantly reduced thrombin-inducedplatelet aggregation (Fig. 5b and Table 3; n = 6). Incontrast, electrotransjection of anti-VAMP antibody (FL-118), which specifically recognizes VAMP-1, -2 and -3,following the same procedure used for the anti-SNAP-25antibody, had a negligible effect, if any, on platelet aggrega-tion in electroporated cells (Fig. 5b and Table 3; n = 6).

Discussion

Circulating platelets can be activated in response to avariety of physiological agonists that become availableafter vascular injury. Agonists, such as thrombin, collagen,

Fig. 5. Effect of SNAP-25 and VAMP-1, -2 or -3 on thrombin-evokedplatelet aggregation. (a) Human platelets (109 cells/mL) were reversiblyelectropermeabilized in a Gene Pulser as described in Materials andmethods (electroporated) or left untreated (Control) and then werestimulated in the presence of 1 mM extracellular Ca2+ with 1 U/mLthrombin. (b) Electroporated cells were incubated with 2 lg/mL rabbitIgG (anti-IgG) or with 2 lg/mL anti-SNAP-25 antibody (anti-SNAP-25)or anti-VAMP antibody (anti-VAMP) for an additional 60 min at 37 �C.Cells were then stimulated with 1 U/mL thrombin in the presence of 1 mMextracellular Ca2+. Aggregation of human platelets was induced at a shearrate of 1200 rpm at 37 �C as described in Materials and methods. Tracesshown are representative of six separate experiments. (c and d) Restingplatelets (lane 1) or platelets electropermeabilized in a Gene Pulser asdescribed in Materials and methods (lane 2) were incubated in the presenceof 2 lg/mL anti-SNAP-25 antibody (c) or anti-VAMP antibody (d) for60 min as indicated, and then lysed. Whole cell lysates were immunopre-cipitated in the absence of antibodies but adding protein A–agarose, andimmunoprecipitated proteins were analyzed by Western blotting usinganti-SNAP-25 antibody (a-SNAP; c) or anti-VAMP antibody (a-VAMP;d). Positions of molecular mass markers are shown on the right. Theseresults are representative of three independent experiments.

I. Jardin et al. / Archives of Biochemistry and Biophysics 465 (2007) 16–25 23

ADP and serotonin, differ in their ability to induce aggre-gation, and activate several transduction mechanismsincluding release of arachidonic acid, activation of proteinkinase C and tyrosine kinases and cytosolic Ca2+ mobiliza-

Table 3Effect of electrotransjection of anti-SNAP-25 and anti-VAMP antibodies on t

Electrotransjected antibody Lag-time (min)

Non-electroporated cells 0.06 ± 0.02Rabbit or goat IgG 4.21 ± 1.07**

Anti-SNAP-25 antibody 3.12 ± 0.27**

Anti-VAMP antibody 3.75 ± 1.39**

Human platelets (109 cells/mL) were reversibly electropermeabilized in a Gene P2 lg/mL rabbit or goat IgG, anti-SNAP-25 antibody or anti-VAMP antibodythrombin in the presence of 1 mM extracellular Ca2+. Aggregation of humanMaterials and methods. Values given are presented as means ± SEM of six se

* p < 0.05 (ANOVA) compared to the results observed in cells electrotransje** p < 0.05 (ANOVA) compared to the results observed in non-electroporate

tion [3]. Platelet activation results in the secretion of a num-ber of prothrombotic factors, such as ADP, serotonin andthromboxane A2, which are able to signal to and activateapproaching platelets [37,38]. Platelet secretion is a criticalstep in hemostasis since the secreted factors are importantfor activation, recruitment, and hemostatic plug formation[39]. This regulated-exocytosis is mediated by integralmembrane proteins known as SNAREs [40]. Human plate-lets have been reported to express SNAP-25 and severalVAMP isoforms, including VAMP-1, -2 and -3 [18–20].Our results indicate for the first time that functional knockdown of SNAP-25 by electrotransjection of cells with anti-SNAP-25 antibody reduces the percentage and rate ofplatelet aggregation by 40% (see Table 3), which providesfurther evidence for the functional significance of SNAP-25 in human platelets. In contrast, electrotransjection ofplatelets with anti-VAMP antibody (FL-118), which recog-nizes the VAMP isoforms 1, 2 and 3, was without effect onthrombin-evoked platelet aggregation. Anti-VAMP-1 anti-bodies have been successfully used to block granule dock-ing to the plasma membrane in mouse pituitary AtT-20cells [41]. The lack of effect of electrotransjection of anti-VAMP antibody, together with those obtained by electro-transjection of rabbit or goat IgG, strongly suggest that theinhibitory effect observed after introduction of anti-SNAP-25 antibody into cells is not likely due to an effect of elec-trotransjection of antibodies itself.

The role of SNAP-25 but not VAMP-1, -2 or -3 in plate-let aggregation might be mediated by different mechanisms.Among them, the inhibition of platelet secretion is one ofthe most relevant actions of impairing SNAP-25 function.Human platelets show three types of regulated exocytosisthat occur upon activation: (a) secretion of dense coregranules, resulting in the release of ADP, serotonin, andcalcium; (b) release of a-granules content, including vonWillebrand’s factor, thromboglobulin, and platelet derivedgrowth factor; (c) secretion of lysosomal degradativeenzymes [42]. SNAP-25, several VAMP isoforms, includingVAMP-3 and -8, and syntaxin-4 have been shown to berequired for granule secretion in platelets [18,19].

The role of SNAP-25 in aggregation might also beattributed to its role in the regulation of [Ca2+]i, which,in turn, has been shown to be required for membrane

hrombin-induced platelet aggregation

% rate % aggregation

57.00 ± 2.47 86.00 ± 5.5685.00 ± 5.13** 82.57 ± 1.4242.57 ± 6.47* 50.77 ± 6.13*

79.80 ± 7.95** 80.57 ± 2.16

ulser as described in Materials and methods and then were incubated withfor an additional 60 min at 37 �C. Cells were then stimulated with 1 U/mLplatelets was induced at a shear rate of 1200 rpm at 37 �C as described inparate determinations.cted with rabbit or goat IgG.d (control) cells.

24 I. Jardin et al. / Archives of Biochemistry and Biophysics 465 (2007) 16–25

fusion and exocytosis in platelets [for review see 3]. Therequirement of functional SNAP-25 for platelet aggrega-tion is similar to that previously found for Ca2+ entry inplatelets and pancreatic acinar cells where we havereported that treatment with botulinum toxins significantlyreduced thrombin- or CCK-8-evoked Ca2+ influx and amy-lase secretion [21,43]. Our results indicate that full plateletaggregation by thrombin requires extracellular Ca2+. Ca2+

entry is necessary for platelet aggregation upon activationof thrombin PAR-1 and PAR-4 receptors, although aggre-gation stimulated through the activation of the GPIb-IX-Vreceptor appeared independent on Ca2+ influx. To furtherinvestigate the role of cytosolic Ca2+ in platelet aggrega-tion, we have used pharmacological tools to discharge theCa2+ stores described in these cells, the DTS and the acidicstores. We have found that selective depletion of the DTSor the acidic stores, as well as depletion of both stores,was unable to induce aggregation in the absence of extra-cellular Ca2+. Aggregation was only found after extensivedepletion of both stores under conditions that allow Ca2+

influx. Store depletion is accompanied by two different phe-nomena: a reduction in Ca2+ concentration in the stores([Ca2+]s), which, by itself, is able to activate a number ofcellular processes, such as the mechanisms underlyingstore-operated Ca2+ entry [22], and the subsequent increasein [Ca2+]i. Our data indicate that the reduction in [Ca2+]s ofone or both stores per se is unable to activate plateletaggregation, which occurs only when this event is accom-panied by a substantial increase in [Ca2+]i that, in our con-ditions, we have estimated around 1.5 lM, asdemonstrated by the effect of Iono, used to facilitateCa2+ entry, on platelet aggregation. At the concentrationused for this experimental procedure, 500 nM, Iono is ableto release Ca2+ from the stores, as demonstrated by the ini-tial increase in [Ca2+]i prior the addition of external Ca2+

(see Fig. 3b), but aggregation started when the [Ca2+]ireached a critical level that might be necessary to activatea low affinity Ca2+-dependent signal transduction mecha-nism. This mechanisms might be sensitized by thrombinor thrombin agonists, which were able to induce aggrega-tion even at lower [Ca2+]i levels.

The involvement of SNAP-25 in platelet aggregationthrough the regulation of [Ca2+]i is supported by the resultsobtained by electrotransjection of anti-VAMP antibody.Our findings indicate that VAMP-1, -2 and -3 are notinvolved in platelet aggregation. These findings are consis-tent with the lack of effect of tetanus toxin-sensitive VAMPisoforms in thrombin-induced Ca2+ entry in human plate-lets [21], although VAMP-2 was found to be required foragonist-induced Ca2+ influx in a more specialized secretorycell type, pancreatic acinar cells [43].

The actin cytoskeleton and cytoskeleton-associated pro-teins has been reported to be involved in platelet aggrega-tion as demonstrated using inhibitors of actinpolymerization [44,45]. Here we show that activation ofPAR-1 and PAR-4 receptors results in actin filamentremodeling. Activation of both thrombin receptors results

in Ca2+ release from both the DTS and the acidic stores[32]; however, depletion of these stores per se was unableto induce aggregation. TBHQ rapidly discharges the acidicstores, suggesting that the acidic store has a high Ca2+

leakage rate [15,46]. The rapid discharge by TBHQ paral-lels the rapid and transient actin filaments remodeling,which suggests that actin reorganization might be associ-ated to Ca2+ release. TG induced an initial decrease inthe F-actin content, which is consistent with the activationof SOCE in platelets [47], and then induced a sustainedincrease in the F-actin content. The actin remodeling pat-tern induced by PAR-1 and PAR-4 clearly differs from thatof TG and TBHQ, suggesting that a timely and sustainedelevation in the F-actin content is necessary for the activa-tion of aggregation.

Our findings suggest that fully functional SNAP-25, butnot VAMP-1, -2 or -3, is necessary for platelet aggregation,a mechanism that is likely mediated by the involvement ofSNAP-25 in the activation and maintenance of store-oper-ated Ca2+ entry, which provide a source of Ca2+ needed forfull activation of aggregation through a low affinity Ca2+-binding mechanism. We cannot rule out the possibility thatgranule secretion mediate the role of SNAP-25 in plateletaggregation, although cytosolic Ca2+ plays a direct rolein membrane fusion events that lead to granule secretion.Activation of PAR-1 and PAR-4 thrombin receptors inplatelets is able to induce aggregation in a different extent,a phenomenon accompanied by a rapid and transient actinfilament remodeling. The involvement of SNAP-25 inthrombin-evoked platelet aggregation is a key feature forthe characterization of the signal transduction mechanismsinduced by this physiological agonist, which, in turn, mightbe essential for the investigation of pathophysiologicalalterations associated to thrombin in human platelets.

Acknowledgments

I.J. is supported by a fellowship from Fundacion Val-hondo Calaff. We thank Mercedes Gomez Blazquez forher technical assistance. This work was supported byMEC-DGI Grant BFU2004-00165, Junta de Extremadu-ra-Consejerıa de Educacion, Ciencia y Tecnologıa andFEDER (2PR04A009) and Consejerıa de Sanidad y Con-sumo SCSS0619.

References

[1] M.L. Kahn, Y. Zheng, W. Huang, V. Bigornia, D. Zeng, S. Moff,R.V. Farese Jr., C. Tam, S.R. Coughlin, Nature 394 (1998) 690–694.

[2] P. Lova, F. Campus, R. Lombardi, M. Cattaneo, F. Sinigaglia, C.Balduini, M. Torti, J. Biol. Chem. 79 (2004) 25299–25306.

[3] L.V. Parise, S.S. Smyth, B.S. Coller, in: E. Beutler, M.A. Lichtman,B.S. Coller, T.J. Kipps (Eds.), Williams Hematology, McGraw-Hill,New York, 2001, pp. 1357–1408.

[4] M.T. Harper, S.O. Sage, Platelets 17 (2006) 134–142.[5] D.T. Hung, Y.H. Wong, T.K. Vu, S.R. Coughlin, J. Biol. Chem. 267

(1992) 20831–20834.[6] M. Torti, E.T. Festetics, A. Bertoni, F. Sinigaglia, C. Balduini,

Thromb. Haemost. 76 (1996) 444–449.

I. Jardin et al. / Archives of Biochemistry and Biophysics 465 (2007) 16–25 25

[7] R. Flaumenhaft, J.R. Dilks, N. Rozenvayn, R.A. Monahan-Earley,D. Feng, A.M. Dvorak, Blood 105 (2005) 3879–3887.

[8] L.S. Chin, R.D. Nugent, M.C. Raynor, J.P. Vavalle, L. Li, J. Biol.Chem. 275 (2000) 1191–1200.

[9] L. Cavallini, M. Coassin, A. Alexandre, Biochem. J. 310 (1995) 449–452.

[10] T. Kovacs, G. Berger, E. Corvazier, K. Paszty, A. Brown, R. Bobe, B.Papp, F. Wuytack, E.M. Cramer, J. Enouf, Br. J. Haematol. 97(1997) 192–203.

[11] B. Papp, A. Enyedi, T. Kovacs, B. Sarkadi, F. Wuytack, O. Thastrup,G. Gardos, R. Bredoux, S. Levy-Toledano, J. Enouf, J. Biol. Chem.266 (1991) 14593–14596.

[12] O. Thastrup, A.P. Dawson, O. Scharff, B. Foder, P.J. Cullen, B.K.Drobak, P.J. Bjerrum, S.B. Christensen, M.R. Hanley, AgentsActions 27 (1989) 17–23.

[13] J.J. Lopez, P.C. Redondo, G.M. Salido, J.A. Pariente, J.A. Rosado,Cell. Signal. 18 (2006) 373–381.

[14] A. Enyedi, K. Paszty, T. Kovacs, B. Sarkadi, G. Gardos, C. Magnier,F. Wuytack, J. Enouf, Biochem. J. 288 (1992) 297–302.

[15] J.J. Lopez, C. Camello-Almaraz, J.A. Pariente, G.M. Salido, J.A.Rosado, Biochem. J. 390 (2005) 243–252.

[16] R. Bobe, R. Bredoux, F. Wuytack, R. Quarck, T. Kovacs, B. Papp, E.Corvazier, C. Magnier, J. Enouf, J. Biol. Chem. 269 (1994) 1417–1424.

[17] F. Wuytack, B. Papp, H. Verboomen, L. Raeymaekers, L. Dode, R.Bobe, J. Enouf, S. Bokkala, K.S. Authi, R. Castells, J. Biol. Chem.269 (1994) 1410–1416.

[18] J. Polgar, S.H. Chung, G.L. Reed, Blood 100 (2002) 1081–1083.[19] G.L. Reed, K.H. Aiilyan, M.L. Fitzgerald, Blood 93 (1999)

2617–2626.[20] A.M. Bernstein, S.W. Whiteheart, Blood 93 (1999) 571–579.[21] P.C. Redondo, A.G. Harper, G.M. Salido, J.A. Pariente, S.O. Sage,

J.A. Rosado, J. Physiol. 558 (2004) 99–109.[22] J.A. Rosado, P.C. Redondo, S.O. Sage, J.A. Pariente, G.M. Salido, J.

Cell. Physiol. 205 (2005) 262–269.[23] G. Grynkiewicz, M. Poenie, R.Y. Tsien, J. Biol. Chem. 260 (1985)

3440–3450.[24] J.A. Rosado, S.O. Sage, Biochem. J. 347 (2000) 183–192.[25] A. Dhar, S.D. Shukla, J. Biol. Chem. 269 (1994) 9123–9127.[26] J.J. Lopez, G.M. Salido, J.A. Pariente, J.A. Rosado, J. Biol. Chem.

281 (2006) 28254–28264.[27] N. Ben-Amor, P.C. Redondo, A. Bartegi, J.A. Pariente, G.M. Salido,

J.A. Rosado, J. Physiol. 570 (2006) 309–323.

[28] P.C. Redondo, I. Jardin, J.M. Hernandez-Cruz, J.A. Pariente, G.M.Salido, J.A. Rosado, Biochem. Biophys. Res. Commun. 333 (2005)794–802.

[29] S. Kim, C. Foster, A. Lecchi, T.M. Quinton, D.M. Prosser, J. Jin, M.Cattaneo, S.P. Kunapuli, Blood 99 (2002) 3629–3636.

[30] M.L. Rand, W. Sangrar, M.A. Hancock, D.M. Taylor, S.M.Marcovina, M.A. Packham, M.L. Koschinsky, Arterioscler. Thromb.Vasc. Biol. 18 (1998) 1393–1399.

[31] T.R. Faruqi, E.J. Weiss, M.J. Shapiro, W. Huang, S.R. Coughlin, J.Biol. Chem. 275 (2000) 19728–19734.

[32] I. Jardin, N. Ben Amor, A. Bartegi, J.A. Pariente, G.M. Salido, J.A.Rosado, Biochem. J. 401 (2007) 167–174.

[33] F.S. London, M. Marcinkiewicz, P.N. Walsh, J. Biol. Chem. 279(2004) 19854–19859.

[34] M. Criado, A. Gil, S. Viniegra, L.M. Gutierrez, Proc. Natl. Acad. Sci.96 (1999) 7256–7261.

[35] S. Quetglas, C. Leveque, R. Miquelis, K. Sato, M. Seagar, Proc. Natl.Acad. Sci. USA 97 (2000) 9695–9700.

[36] J.C. Hao, N. Salem, X.R. Peng, R.B. Nelly, M.K. Bennett, J.Neurosci. 17 (1997) 1596–1603.

[37] J. Gibbins, J. Asselin, R. Farndale, M. Barnes, C.L. Law, S.P.Watson, J. Biol. Chem. 271 (1996) 18095–18099.

[38] G.R. Sambrano, E.J. Weiss, Y.W. Zheng, W. Huang, S.R. Coughlin,Nature 413 (2001) 74–78.

[39] T.D. Schraw, T.W. Rutledge, G.L. Crawford, A.M. Bernstein,A.L. Kalen, J.E. Pessin, S.W. Whiteheart, Blood 102 (2003)1716–1722.

[40] G.L. Reed, M.L. Fitzgerald, J. Polgar, Blood 96 (2000) 3334–3342.[41] E. Chieregatti, J.W. Witkin, G. Baldini, Traffic 3 (2002) 496–511.[42] J.G. White, in: A.L. Bloom, C.D. Forbes, D.P. Thomas, E.G.D.

Tuddenham (Eds.), Haemostasis and Thrombosis, Churchill Living-ston, Edinburgh, 1994, pp. 49–89.

[43] J.A. Rosado, P.C. Redondo, G.M. Salido, S.O. Sage, J.A. Pariente,Am. J. Physiol. Cell Physiol. 288 (2005) 214–221.

[44] H. Falet, G. Chang, B. Brohard-Bohn, F. Rendu, J.H. Hartwig, Am.J. Physiol. Cell Physiol. 289 (2005) 819–825.

[45] J.E. Fox, L. Lipfert, E.A. Clark, C.C. Reynolds, C.D. Austin, J.S.Brugge, J. Biol. Chem. 268 (1993) 25973–25984.

[46] J.A. Rosado, J.J. Lopez, A.G. Harper, M.T. Harper, P.C. Redondo,J.A. Pariente, S.O. Sage, G.M. Salido, J. Biol. Chem. 279 (2004)29231–29235.

[47] P.C. Redondo, M.T. Harper, J.A. Rosado, S.O. Sage, Blood 107(2006) 973–979.