Phosphorylation of the Thromboxane Receptor α, the Predominant Isoform Expressed in Human Platelets*

A single gene encodes the human thromboxane receptor (TP), of which there are two identified splice variants, α and β. Both isoforms are rapidly phosphorylated in response to thromboxane agonists when overexpressed in human embryonic kidney 293 cells; this phenomenon is only slightly altered by inhibitors of protein kinase C. Pharmacological studies have defined two classes of TP in human platelets; sites that bind the agonist I-BOP with high affinity support platelet shape change. Low affinity sites, which irreversibly bind the antagonist GR 32191, transduce platelet activation and aggregation. Isoform-specific antibodies permitted detection of TPα, but not TPβ, from human platelets, although mRNA for both isoforms is present. A broad protein band of 50–60 kDa, reflecting the glycosylated receptor, was phosphorylated upon activation of platelets for 2 min with I-BOP. This was a rapid (∼30 s) and transient (maximum, 2–4 min) event and was inhibited by TP antagonists. Both arachidonic acid and low concentrations of collagen stimulated TPα phosphorylation, which was blocked by cyclooxygenase inhibition or TP antagonism. Blockade of the low affinity TP sites with GR 32191 prevented I-BOP-induced TPα phosphorylation. This coincided with agonist-induced platelet aggregation and activation but not shape change. Also, activation of these sites with the isoprostane iPF2α-III induced platelet shape change but not TPα phosphorylation. Heterologous TP phosphorylation was observed in aspirin-treated platelets exposed to thrombin, high concentrations of collagen, and the calcium ionophore A 23187. Both homologous and heterologous agonist-induced phosphorylation of endogenous TPα was blocked by protein kinase C inhibitors. TPα was the only isoform detectably translated in human platelets. This appeared to correspond to the activation of the low affinity site defined by the antagonist GR 32191 and not activated by the high affinity agonist, iPF2α-III. Protein kinase C played a more important role in agonist-induced phosphorylation of native TPα in human platelets than in human embryonic kidney 293 cells overexpressing recombinant TPα.

A single gene encodes the human thromboxane receptor (TP), of which there are two identified splice variants, ␣ and ␤. Both isoforms are rapidly phosphorylated in response to thromboxane agonists when overexpressed in human embryonic kidney 293 cells; this phenomenon is only slightly altered by inhibitors of protein kinase C. Pharmacological studies have defined two classes of TP in human platelets; sites that bind the agonist I-BOP with high affinity support platelet shape change. Low affinity sites, which irreversibly bind the antagonist GR 32191, transduce platelet activation and aggregation. Isoformspecific antibodies permitted detection of TP␣, but not TP␤, from human platelets, although mRNA for both isoforms is present. A broad protein band of 50 -60 kDa, reflecting the glycosylated receptor, was phosphorylated upon activation of platelets for 2 min with I-BOP. This was a rapid (ϳ30 s) and transient (maximum, 2-4 min) event and was inhibited by TP antagonists. Both arachidonic acid and low concentrations of collagen stimulated TP␣ phosphorylation, which was blocked by cyclooxygenase inhibition or TP antagonism. Blockade of the low affinity TP sites with GR 32191 prevented I-BOP-induced TP␣ phosphorylation. This coincided with agonist-induced platelet aggregation and activation but not shape change. Also, activation of these sites with the isoprostane iPF 2␣ -III induced platelet shape change but not TP␣ phosphorylation. Heterologous TP phosphorylation was observed in aspirin-treated platelets exposed to thrombin, high concentrations of collagen, and the calcium ionophore A 23187. Both homologous and heterologous agonist-induced phosphorylation of endogenous TP␣ was blocked by protein kinase C inhibitors. TP␣ was the only isoform detectably translated in human platelets. This appeared to correspond to the activation of the low affinity site defined by the antagonist GR 32191 and not activated by the high affinity agonist, iPF 2␣ -III. Protein kinase C played a more important role in agonist-induced phosphorylation of native TP␣ in human platelets than in human embryonic kidney 293 cells overexpressing recombinant TP␣.
Thromboxane (Tx) 1 A 2 is formed in platelets by the sequential metabolism of arachidonic acid by cyclooxygenases and TxA 2 synthase (1) following activation by agonists, such as arachidonic acid, thrombin, collagen, or ADP. Although a weak agonist itself, TxA 2 plays an important role in amplifying the response to other, more potent platelet agonists (2). Studies with pharmacological ligands in human platelets have suggested the presence of two distinct populations of receptors (3,4). Distinct functions have been attributed to these subtypes: (i) aggregation and granule secretion appear to be mediated by receptors with low affinity for the agonist ligand I-BOP, which are bound irreversibly by the antagonist GR 32191; and (ii) shape change appears to be mediated by receptors with high affinity for I-BOP, which are bound reversibly by GR 32191 (3,4). Despite these observations, the molecular basis for this functional segregation of pharmacological TP subtypes is unknown. TP purified from human platelets consists of a broad protein band of 57 kDa (5,6). Initial cloning of the TP, from megakaryoblastic cell lines and human placental cDNA libraries (referred to as the placental TP or the TP␣ isoform) implied its membership in the G protein-coupled receptor superfamily (GPCR) (7)(8)(9)(10). Only one TP gene has been cloned to date (11). However, an alternatively spliced form of TP, TP␤, was cloned from an endothelial cDNA library (12). The mRNA for both splice variants have been demonstrated in platelets (13). Because no pharmacological ligand can presently distinguish between these two isoforms, it is still unknown how they relate to the pharmacological subtypes of the TP in human platelets. We (14,15) and others (16) have also shown that the isoprostane, iPF 2␣ -III (formerly known as 8-iso-PGF 2␣ ) (17), induces platelet shape change, calcium mobilization (15) and reversible aggregation at high concentrations. Although these effects are inhibited by TP antagonists (18), it is unknown whether the isoprostane acts solely via TPs. Evidence consistent with the possibility of receptors specific to iPF 2␣ -III was described in platelets and vascular smooth muscle cells (14,19).
Phosphorylation is an important mechanism in rapid desensitization of many GPCRs, as exemplified by the ␤ 2 -adrenergic receptors (20,21). Different kinases can phosphorylate these receptors: for example, G-protein receptor kinases (GRKs) are receptor-specific kinases that phosphorylate the agonist-occupied receptor, whereas PKC or PKA can be activated by other ligands and participate in heterologous receptor desensitization (20,22). Although it is appreciated that the role of distinct kinases in the phosphorylation of a particular GPCR may vary according to cell type, study of this process has largely been confined to heterologous expression systems (23)(24)(25). There are actually few reported studies of agonist-induced phosphorylation of endogenous receptors (26), probably because they are usually expressed in relatively low abundance. We have previously described isoform-specific antibodies for TPs (27). Using these reagents, we now report that only TP␣ is detected in human platelets. Upon activation of the platelets with a TP agonist or arachidonic acid as a source of endogenous TxA 2 a rapid, transient, and PKC-dependent phosphorylation of the TP␣ occurs. This involves the low affinity form of the TP, as defined by irreversible binding of GR 32191. Furthermore, TP␣ may also be phosphorylated in a PKC-dependent manner in response to platelet activation by thrombin in aspirin-treated platelets. The major role of PKC in rapid, agonist-dependent phosphorylation of endogenous TP␣ in platelets contrasts with our previous observations when recombinant TP␣ was overexpressed in HEK-293 cells (27).
Platelet Preparation and Labeling-Peripheral blood from healthy volunteers, who had not received any medication for at least 10 days, was collected into ACD-A anticoagulant (National Institute of Health formula: 0.8% citric acid, 2.2% trisodium citrate, 2H 2 O, 2.45% glucose) and 1 mM of aspirin unless otherwise indicated. Informed consent was obtained from all donors in conformity with the French Etablissement de Transfusion Sanguine committee. The blood was centrifuged at 120 ϫ g for 20 min at 20°C, and platelet-rich plasma was collected, acidified with ACD-A, and further centrifuged at 1200 ϫ g for 20 min. The platelets were washed in a phosphate-free modified tyrode buffer without calcium (Buffer A, pH 6.8: 136 mM NaCl, 2.7 mM KCl, 12 mM NaHCO 3 , 2 mM MgCl 2 , and 5 mM glucose) in the presence of 0.1 M PGE 1 . Platelets were resuspended at 10 9 /ml in the same buffer and labeled with 1 mCi/ml of [ 32 P]orthophosphate for 1 h 30 min at room temperature (28). After further washing in the same buffer, platelets were resuspended in the reaction buffer (Buffer A containing 2 mM CaCl 2 and 0.4 mM NaH 2 PO 4 , pH 7.4) at 4 ϫ 10 8 /ml and allowed to rest at room temperature for 1 h before aggregation was performed.
Samples were further treated to immunoprecipitate the TPs as described below. The phosphorylation of pleckstrin p-47, a PKC substrate, was assessed by SDS-PAGE of 20 l of total platelet lysate (29,30). TxB 2 , the inactive hydrolysis product of TxA 2 , was measured in the platelet lysates by enzyme immunoassay (31).
Immunoprecipitation of Human TPs from Platelets or Cells Overexpressing TPs-Immunoprecipitation of TPs was performed using specific polyclonal antibodies for human TP isoforms (27). Briefly, these antibodies were directed against peptides located at the end of the carboxyl-terminal tail of either TP␣ or TP␤: NH 2 -SLSLQPQLTQRS-GLQ-COOH (referred to as Ab␣) for TP␣ and NH 2 -(C)-PFEPPT-GKALSRKD-COOH (referred to as Ab␤) for TP␤. Immunoaffinity columns with each antibody were prepared as follows. Briefly, antisera were first partially purified using the E-Z-SEP ® kit and further incubated with CNBr-activated Sepharose, according to the manufacturer's instructions. Immunoglobulins derived from nonimmune rabbit serum, coupled to CNBr-activated Sepharose, were used to preclear the platelet lysates. After preclearing for 1 h at 4°C, using 50 l of normal rabbit IgG covalently coupled to Sepharose CL-4B, samples were immunoprecipitated overnight at 4°C using 50 l of immunoaffinity Sepharose for either antibody. The beads were washed four times with 1 ml of radioimmune precipitation buffer and resuspended in 100 l of 1ϫ Laemmli buffer (4% SDS (w/v), 5% glycerol (v/v), 60 mM Tris, pH 6.8, 2 M urea, and 0.005% bromphenol blue) under nonreducing conditions. Samples were vigorously vortexed for 15 min, centrifuged for 5 min at 10,000 ϫ g, and loaded on SDS-polyacrylamide gels as described previously (27). Analysis of radioactivity in the samples was performed using a Fuji BioImaging Analyzer (Fuji, Tokyo, Japan) after the gels were dried.
In some cases, immunoblot analysis of the TP isoforms was performed. SDS-polyacrylamide gels were transferred onto nitrocellulose membranes. The TPs were visualized using the Ab␣ or Ab␤ followed by a donkey anti-rabbit antibody coupled to horseradish peroxidase (Jackson ImmunoResearch, West Grove, PA). Positive protein bands were revealed by ECL according to the manufacturer's instructions.
PNGase F Digestion-[ 32 P]P i -labeled platelets were incubated with 10 nM I-BOP for 2 min. Labeled HEK-TP␣ or HEK-TP␤ was incubated for 10 min with 300 nM U 46619 as defined previously (27). TP␣ or TP␤ were immunoprecipitated as described above. Immunoprecipitates were further denatured for 10 min at 90°C with SDS 0.5% and ␤-mercaptoethanol 1% prior to the addition of 1250 units of PNGase F per reaction according to the manufacturer's instructions. The reaction was carried on for 1 h at 37°C and then stopped with 1ϫ Laemmli buffer. Samples were subjected to SDS-PAGE, and dried gels were exposed to Biomax MS films.
Stimulation of Tyrosine Phosphorylation-Platelets were exposed to high or low concentrations of I-BOP, thrombin, or pervanadate, an inhibitor of tyrosine phosphatases, which induced strong tyrosine phosphorylation. After a 4-min incubation, platelets were lysed as described above, and tyrosine phosphorylation was assayed in total platelet lysates or after immunoprecipitation of the TP␣ receptor isoform by immunoblot analysis, using a specific P-Tyr monoclonal antibody.

Immunodetection of TP␣ in Human
Platelets-We used polyclonal antibodies raised against specific sequences of TP␣ or TP␤, to isolate TPs from human platelets. Polyclonal antibodies specific for TP␣ (Ab␣) were used to immunoprecipitate 1 mg of human platelet lysate, which corresponds to 0.3 pmol/mg of protein, as assessed by binding of [ 3 H]SQ 29548. A broad band with a molecular weight of 50 -60 was detected after immunoblot analysis with the same antibody (Fig. 1A). Ab␣ also immunoprecipitated the TP␣ from HEK-293 cells stably transfected with the corresponding cDNA (Fig. 1A) as described previously (27). However, immunoprecipitation of 1 mg of total platelet protein lysate using the TP␤ isoform-specific antipeptide antiserum failed to reveal any detectable band (Fig. 1A), although these antibodies were able to immunoprecipitate the TP␤ receptor isoform from HEK-293 overexpressing these receptors (Fig. 1A). These antibodies were able to immunoprecipitate as little as 50 fmol of receptors/mg of protein from HEK-293 overexpressing either TP␣ or TP␤. Using these cells, we have previously shown that these antibodies were both isoform-specific by Western blotting and by immunofluorescence analysis (data not shown).
We next checked that the broad protein band isolated from platelets with Ab␣ was glycosylated. Results were compared with the digestion of TPs in HEK-293-TP␣ or TP␤. To avoid technical problems subsequent to deglycosylation of the antibodies used for immunoprecipitation, we performed these experiments on phosphorylated TPs, which are obtained after activation with TP agonist, as demonstrated below. Deglycosylation with PNGase F resulted in a shift of the broad protein band from 50 -60 kDa in platelets and from 55-70 kDa HEK-293 cells, to an apparent molecular weight of 28. Deglycosylation of TP␤ also revealed a shift in the molecular weight to 32.5. These results indicate that the broad protein band of 50 -60 kDa observed in platelets corresponds to glycosylated TP␣ and that TP␣ in HEK-293 cells and human platelets are differentially glycosylated. The difference in the apparent molecular weight between deglycosylated TP␣ and TP␤ corresponds to the difference in the number of amino acids between the two isoforms (343 amino acids for TP␣ and 407 for TP␤).
Phosphorylation of the TP␣ Isoform in Human Platelets-Homologous and heterologous desensitization of human platelets in response to U46619, a TxA 2 mimetic, or to thrombin has been reported previously (32,33). Incubation of aspirin-treated platelets with I-BOP, a Tx analog, for increasing periods of time, resulted in the phosphorylation of a broad protein band of 50 -60 kDa ( Fig. 2A). Phosphorylation was rapid (Յ0.5 min) but transient (maximum, 2-4 min). We regularly observed a phosphorylated band of 68 kDa in these samples, with a stronger signal in activated platelets. Detection of this band was not modified when immunoprecipitation of TP␣ was performed in the presence of the specific Ab␣-peptide used for immunization, whereas immunoprecipitation of the broad protein band of 50 -60 kDa was completely abolished (data not shown), suggesting that it is not related to TP receptors.
Since previous pharmacological studies suggested the presence of high and low affinity receptors in human platelets, we investigated the relevance of these observations in the phosphorylation of TPs in human platelets. We used a particular TP antagonist, GR 32191, that dissociates very slowly, if at all, from the low affinity binding sites in human platelets (4). Fig.  2D illustrates this effect. When platelets derived from PRP treated with 1 M of GR32191, no increase in TP␣ phosphorylation was observed with I-BOP 10 nM. TP␣ in platelets derived from untreated-PRP were normally phosphorylated by I-BOP and GR 32191 inhibited this phosphorylation, similarly to SQ 29548.
Further characterization of this phosphorylation showed that okadaic acid, an inhibitor of serine/threonine phosphatases, resulted in an increase in TP␣ phosphorylation. Under these conditions, receptor phosphorylation was sustained for up to 30 min, compared to 4 min in the absence of okadaic acid (data not shown). Immunoblot analysis of immunoprecipitated TP␣ using P-Tyr antibodies did not reveal any phosphorylation of TP␣ in platelets activated with I-BOP, thrombin, or pervanadate, a strong inhibitor of tyrosine-phosphatases (35), despite marked tyrosine kinase-dependent substrate phosphorylation (data not shown).
iPF 2␣ -III Does Not Induce Phosphorylation of the TP␣ Receptor-Previous studies by our group and others have shown that iPF 2␣ -III induces platelet shape change (14,16), Ca 2ϩ mobilization, and reversible platelet aggregation at high concentrations of the agonist (15,18). All of these effects were abolished by TP antagonists. However, iPF 2␣ -III-induced inositol phosphate formation in human platelets was not blocked by TP antagonists (14). Consistent with this observation, we failed to observe TP␣ phosphorylation with 5-50 M iPF 2␣ -III (Fig. 3A). Although iPF 2␣ -III induced platelet shape change, neither prolongation of the incubation time (5 min) (Fig. 3A) nor pretreatment with 1 M okadaic acid induced significant TP␣ phosphorylation as compared with control unstimulated platelets (data not shown). Thus, iPF 2␣ -III appears to favor activation of the high affinity sites, which mediates platelet shape change. In contrast, pretreatment of platelets with 50 M of iPF 2␣ -III reduced I-BOP-induced platelet aggregation (60%) and TP␣ phosphorylation (Fig. 3B), consistent with a competition between I-BOP and high concentrations of iPF 2␣ -III for the occupancy of the low affinity sites, which mediate agonist-induced phosphorylation of TP␣ in human platelets. Moreover, activation of HEK-TP␣ cells with iPF 2␣ -III resulted in TP␣ phospho-FIG. 1. Immunodetection of the TP in human platelets. A, 1 mg of lysate obtained from human platelets (Plt) or HEK-293 overexpressing TP␣ or TP␤ (HEK-TP␣ or HEK-TP␤) was immunoprecipitated with Ab␣ or Ab␤ linked to Sepharose. This corresponded to 0.3 pmol/mg of protein for platelets and 1.6 pmol/mg for HEK-TPa and HEK-TPb, respectively. Samples were subjected to SDS-PAGE, and immunoblot analysis was performed using corresponding antibodies as described under "Experimental Procedures." These data are representative of two similar experiments for HEK-293 cells and at least five experiments for platelets. B, cell lysates obtained from [ 32 P]P i -labeled platelets or HEK-TP␣ or TP␤ activated with Tx analogs were immunoprecipitated using Ab␣ or Ab␤. Immunoprecipitated samples were further incubated in the absence or presence of PNGase F for 1 h at 37°C. Samples were subjected to SDS-PAGE. Electrophoresis gel was dried and exposed to Biomax MS films. rylation (data not shown).
Endogenously Formed TxA 2 Phosphorylates TP␣: Effect of arachidonic acid and low concentrations of collagen-We next tested whether endogenously formed TxA 2 induces phosphorylation of the platelet TP␣. Addition of arachidonic acid (2.5 M) to platelets resulted in the formation of 200 -400 ng/ml of TxB 2 (corresponding to 0.2 ϫ 10 9 platelets). Under these conditions, TP␣ was phosphorylated, to a degree similar to platelets, when incubated with 10 nM I-BOP (Fig. 4). When platelets were pretreated with 10 M of SQ 29548 or flurbiprofen, an inhibitor of cyclooxygenase, TP␣ phosphorylation was inhibited, demon-strating that endogenous TxA 2 (or PGH 2 ) formed by cyclooxygenase-1 was responsible for receptor phosphorylation in response to arachidonic acid (Fig. 4). Similar results were obtained with a low concentration of collagen (Fig. 4). These results demonstrate that TP␣ phosphorylation can occur in activated platelets via endogenous TxA 2 generation. In these samples, platelet aggregation and pleckstrin phosphorylation were also inhibited by SQ 29548 and flurbiprofen treatment.
Heterologous Phosphorylation of the TP␣-Because heterologous activation of platelets by non-thromboxane agonists may contribute to the desensitization of TP␣ (32), we examined the ability of various agonists to phosphorylate TP␣. We utilized aspirin-treated platelets, thus excluding signaling via endogenous TxA 2 formation . Thrombin, calcium ionophore A23187 and the active phorbol ester PMA, phosphorylated the TP␣ (Fig. 5A). In these experiments, the absence of endogenous TxA 2 was verified by measuring TxB 2 in platelet lysates by enzyme immunoassay (data not shown). Phosphorylation of pleckstrin was also observed (Fig. 5A). In contrast, little phosphorylation of either substrate was obtained with 200 nM PGE 1 or PGE 2 or with 10 M forskolin (Fig. 5A). Although platelet aggregation induced by low concentrations of collagen is dependent on the formation of TxA 2 (Fig. 4), higher concentrations can bypass this inhibition. When collagen was used at 100 g/ml, neither flurbiprofen nor SQ 29548 prevented platelet aggregation and phosphorylation of the TP␣ and pleckstrin (Fig. 5A). Thrombin-induced phosphorylation of TP␣ was transient (Fig. 5B) and resembled kinetics observed with I-BOPinduced phosphorylation (described in Fig. 2A).
Effect of PKC on TP␣ Phosphorylation-Because we observed that TxA 2 and all other agonists tested induced phosphorylation of TP␣ and pleckstrin, we utilized specific PKC inhibitors to address the role of this kinase in TP␣ phosphorylation. Pretreatment of platelets for 30 min at 37°C with two structurally distinct but specific PKC inhibitors, GF 109203X and Ro-31-8220, prior to platelet activation with I-BOP, resulted in a dramatic reduction in TP␣ phosphorylation (ϳ80%) (Fig. 6). Thrombin-induced TP␣ phosphorylation was also inhibited by GF 109203X (Fig. 6, right panel). The effectiveness of these molecules as inhibitors of PKC was assessed by their capacity to inhibit the PMA-dependent phosphorylation TP␣ (Fig. 6, right panel). Our recent studies on the phosphorylation of recombinant TP isoforms stably expressed in HEK-293 cells showed little involvement of PKC in response to TxA 2 mimetics, although PMA could readily induce PKC-dependent TP phosphorylation in this system (27). Thus, agonist-induced phosphorylation of TP␣ in human platelets, in contrast to the HEK-293 expression system, appears largely dependent on PKC.
Involvement of Integrin Gp IIb/IIIa in the Phosphorylation of TP␣-Activation of platelets by U 46619, a stable Tx analog, has been shown to result in the association of pp 60src with the cytoskeleton (36). Such events, related to ligand occupancy of GpIIb/IIIa, may play a role in the phosphorylation of TP␣ via "inside-out" signaling (37). Thus, the influence of platelet aggregation on TP␣ phosphorylation was investigated. The active peptide, RGDS, and two peptide mimetics that are antagonists of GpIIb/IIIa, Ro-43-5054 and Ro-44-9883, were used. The phosphorylation of TP␣ by I-BOP was unaffected in the presence of 50 M RGDS, 0.1 M Ro-43-5054, or 0.2 M Ro-44 -9883 (Fig. 7A). Under these conditions, I-BOP-induced platelet aggregation was totally inhibited (Fig. 7B). Also, phosphorylation of TP␣ induced by I-BOP was unchanged under either stirring or nonstirring conditions (data not shown). In a few blood donors, we observed a small increase (ϳ20%) in TP␣ phosphorylation in nonaggregating conditions (data not shown). Moreover, TP␣ phosphorylation induced by low or high concentrations of collagen was not modified by RGDS (Fig. 7C), thus dissociating platelet TP receptor phosphorylation from aggregation. These results suggest that engagement of the GpIIb/ IIIa complex is downstream of the events leading to agonistinduced phosphorylation of TP␣. DISCUSSION Although mRNA detection for the two recognized human TP isoforms has been reported in human platelets (13), it is unknown whether either or both are translated to protein. It is also unknown whether these isoforms relate to the high and low affinity forms of the receptors that have been characterized pharmacologically (4). Using isoform-specific antibodies, we were able to immunoprecipitate TP␣ as a 50 -60-kDa protein band from platelet lysate. TP␤ could not be detected. We estimate that Ն 50 fmol/mg of protein of TP␤ could be detected with the specific antibodies from binding experiments in HEK-293 cells transfected with the TP␤ isoform. These results suggest that TP␤ is expressed at very low levels, if at all, in human platelets.
In the present studies, Tx analogs induced rapid agonistinduced phosphorylation of a broad protein band in platelets. Many arguments support that this broad phosphorylated protein band appears to correspond to the TP␣ isoform. Thus, (i) the broad protein band is specifically immunoprecipitated with the Ab␣ antibodies and migrates at the same molecular weight as that revealed by immunoblot analysis, (ii) digestion of immunoprecipitated TP␣ with PNGase F results in an apparent molecular weight similar to that obtained from HEK-293 cells transfected with recombinant TP␣, and (iii) the phosphorylation of the 50 -60-kDa protein band is associated with TP receptor activation. SQ 29548, a TP receptor antagonist, suppress agonist-induced phosphorylation of this band. Previous results by different groups (5, 38, 39) have detected TP recep- tors as a broad protein band of 50 -57 kDa in human platelets. Other authors have reported a sharp protein band of 55 kDa obtained from oligodendrocytes, neuronal cells (40), rat aorta (39), or human platelets (6). This discrepancy may reflect differential sensitivity of the detection systems involved (ligand affinity or immunoaffinity purification systems).
There is presently no information that relates either TP␣ or TP␤ to the subtypes of TPs that have been defined pharmacologically (4). In the present studies, rapid agonist-induced phosphorylation of TP␣ appeared to involve signaling through low affinity binding sites. Thus, neither low concentrations of the agonist I-BOP, which induce platelet shape change, nor high agonist concentrations on platelets pretreated with GR 32191 (which blocks the low affinity sites) caused TP␣ phosphorylation.
TxA 2 originating in platelets from exogenous (i.e. addition of arachidonic acid) or from endogenous arachidonic acid (i.e. low concentrations of collagen) caused phosphorylation of TP␣. Thus, endogenous TxA 2 (or PGH 2 ) can bind to and activate the receptor, resulting in its phosphorylation in a manner similar to that observed using the synthetic ligand I-BOP.
Other platelet agonists, such as thrombin, high concentrations of collagen, PMA, and A23187, also induce TP␣ phosphorylation in aspirin-treated platelets. It is possible that this phosphorylation relates to homologous or heterologous desensitization of the TP by other platelet agonists (32,41). Examples of heterologous phosphorylation of GPCRs include endothelin-dependent phosphorylation of ␣ 1B -adrenoreceptors (42) and thrombin-dependent phosphorylation of the prostacyclin receptor (43).Our results suggest that I-BOP, thrombin, and PMA-induced TP␣ phosphorylation were dependent on PKC because specific PKC inhibitors suppressed TP␣ phosphorylation.
Phosphorylation of TP␣ occurs in response to PMA in both platelets and transfected HEK-293 cells (27). This indicates that PKC phosphorylation sites are present in TP␣. However, the role of this kinase in mediating agonist-induced TP␣ phosphorylation differs between native receptors in human platelets and recombinant TP␣ stably expressed in HEK-293 (27). Differences in affinities, or in the relative abundance of the receptors, or in the amounts of the kinases in different cells could explain this diversity of response.
Another difference involves the absence of response of this receptor to the isoprostanes in human platelets. iPF 2␣ -III increased phosphorylation of TP␣ in the expression system. However, in human platelets, iPF 2␣ -III failed to cause a dose-dependent increase in TP␣ phosphorylation, despite stimulating inositol phosphate formation as described earlier (44).
In conclusion, we have demonstrated that the TP␣ isoform exists in human platelets; TP␤ is much less abundant, if it is expressed at all. Phosphorylation of TP␣ is consistent with the activation of the low affinity site defined pharmacologically with GR 32191. Our results suggest that human platelet TP␣ is phosphorylated by TxA 2 analogs and by other platelet agonists, such as thrombin, through activation of PKCs. Differences in the regulation of the Tx-dependent TP phosphorylation in the HEK-293 overexpressing system, where PKC is of marginal importance, could derive from differences in cellular contents of kinases and their affinity for the receptors in the presence of their ligands. Thus, heterologous expression systems afford sufficient levels of protein to simplify the study of posttranslational modifications of GPCRs. However, such observations may not accurately mimic the regulation of all endogenous receptors in their native milieu.