Rapid, agonist-dependent phosphorylation in vivo of human thromboxane receptor isoforms. Minimal involvement of protein kinase C.

Thromboxane A2 (TxA2) is a potent vasoconstrictor and platelet agonist. Its biological function is tightly regulated. G protein-coupled membrane receptors transduce the effects of TxA2. However, although a single thromboxane receptor (TP) gene has been identified, two splice variants have been cloned from human placenta and megakaryocytic lines (TPα) and from human endothelial cells (TPβ). These differ in the length of their carboxyl-terminal extensions (15 versus 79 residues), which contain multiple potential sites for receptor phosphorylation. Given that TP agonists activate protein kinase C (PKC), it would seem possible that PKC-dependent phosphorylation of TPs might play a central role in homologous desensitization of these receptors. To determine if the TP isoforms were differentially phosphorylated in response to agonist in vivo, human embryonic kidney (HEK) 293 cells were stably transfected with TPα and TPβ. Isoform-specific anti-peptide antibodies were developed and used to immunoprecipitate the phosphorylated receptors. U46619, a PGH2/TxA2 mimetic, induced specific phosphorylation of both isoforms. Phosphorylation of the two isoforms was similar in dose and time dependence, reaching a plateau at around 100 nM U46619. Inhibition of PKC with either GF 109203X (5 μM) or RO 31-8220 (5 μM) or of protein kinase A with H-89 (50 μM) marginally influenced agonist-dependent phosphorylation of either isoform and failed to modulate homologous desensitization of agonist-induced stimulation of inositol phosphate formation. Similar results were obtained when PKC was down-regulated by long term incubation with the phorbol ester, phorbol myristate acetate. Although short term stimulation with phorbol myristate acetate caused PKC-dependent phosphorylation of TPs in vivo, thrombin stimulation of the TP-transfected HEK cells in vivo failed to phosphorylate either of the TP isoforms. Thus, despite the capacity of PKC to phosphorylate TPs in HEK 293 cells and the likely activation of PKC by TP stimulation, this enzyme, like protein kinase A, contributes marginally to rapid, agonist-induced phosphorylation of either TP isoform.

ally to rapid, agonist-induced phosphorylation of either TP isoform.
Thromboxane A 2 (TxA 2 ) 1 is a product of the sequential metabolism of arachidonic acid by the cyclooxygenases and TxA 2 synthase (1). It is formed upon activation of a variety of cells, including platelets, macrophages, and vascular smooth muscle cells and exhibits potent biological activity, causing platelet aggregation and secretion, mitogenesis, and vasoconstriction (2,3). These effects are transduced via membrane receptors, identified initially with a variety of diverse structural ligands (4). Although such pharmacological studies suggested diversity among thromboxane receptors (TPs), a single gene, encoding a member of the heptahelical G protein-coupled receptor (GPCR) family, has been cloned (5). However, two variants, based on alternative splicing of the carboxyl-terminal tail of the receptor, have been identified. The first, TP␣, was cloned from a cDNA library derived from placental (6) and also from megakaryocytic cell lines (7,8). The second, TP␤, was cloned from a human umbilical endothelial cell cDNA library (9). The precise biological functions subserved by these isoforms is presently unknown. For example, mRNAs for both isoforms exist in human platelets (10); however, it is unknown whether they represent the two forms of functional TPs identified by ligand binding in human platelets (11). Similarly, the functional response to TP stimulation in endothelial cells remains ill defined. The capacity of the isoforms to subserve distinct biological functions is illustrated by differential coupling of the expressed isoforms to adenylate cyclase in COS cells (10).
One aspect of TP function that, a priori, may differ between the isoforms is agonist-dependent desensitization. TP␤ differs from the a isoform in having a longer (79 versus 15 amino acids) carboxyl-terminal extension, which contains an additional 11 serine and 4 threonine residues. It also contains a tyrosine residue, absent from the TP␣ isoform (9). These amino acids are potential targets for phosphorylation, which is likely to be intrinsic to the process of desensitization. Given the rapid formation of TxA 2 by activated cells and its potency, regulation of the response to this eicosanoid by homologous receptor desensitization is of likely biological importance. Biosynthesis of additional ligands, such as thrombin and growth factors (12)(13)(14)(15)(16) may amplify the response to TxA 2 , perhaps via cross-talk with TPs. Similarly, TxA 2 may evoke formation of counterregulatory ligands, such as prostacyclin, which may also modulate TP function (17).
Presently, our understanding of the molecular events that underlie TP desensitization in intact cells is limited. These have largely been confined to pharmacological studies (18), which suggest differences between cells but have not discriminated between the cloned isoforms. We have previously demonstrated (19) that a fusion protein, including the carboxylterminal tail of the TP␣, may be phosphorylated by purified PKC and, to a lesser extent, PKA, in vitro. Okwu et al. (20) have provided evidence that TPs in human platelets may be subject to phosphorylation.
We now report the characterization of specific, peptide-based antibodies to TP␣ and TP␤ and demonstrate that both receptor isoforms are predominantly localized at the plasma membrane of human embryonic kidney (HEK) 293 cells stably overexpressing TP receptors. Both isoforms are phosphorylated in response to stimulation with the prostaglandin H 2 /TxA 2 mimetic, U46619 (21). Interestingly, the dose and time dependence of agonist-induced receptor phosphorylation appeared similar for both isoforms in vivo, and neither PKC nor PKA played a major role in this phenomenon.

Stable Expression of Human Thromboxane A 2 Receptors in HEK 293
Cells-The cDNA encoding the TP␣ isoform was subcloned into the EcoRI-XbaI sites of the mammalian expression vector pcDNA3. The cDNA for the TP␤ isoform subcloned into the EcoRI-EcoRI sites of the same vector (22) was further digested by BamHI and Bsu36I and then ligated after creating a blunt end. This treatment resulted in the truncation of 195 base pairs in the 5Ј-untranslated region, leaving one ATG start codon. This enhanced the expression of the receptors in stably transfected cells 4 -5-fold relative to the undigested construct (data not shown). Following truncation, the stably transfected HEK 293 cells expressed similar amounts of the TP␣ and TP␤ isoforms as assessed by binding studies (see "Results").
HEK 293 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum, 2 mM glutamine, 50 units/ml penicillin, 100 g/ml streptomycin. HEK 293 cells stably overexpressing both isoforms were prepared as described previously (22). Briefly, cells were plated at 0.7 ϫ 10 6 cells/100-mm culture dish and transfected with 10 g of TP␣-pcDNA3, TP␤-pcDNA3, or pcDNA3 using a cationic liposome-mediated transfer according to the manufacturer's instructions (DOTAP; Boehringer Mannheim). After 8 h of transfection, the culture medium was changed, and selection of the clones was performed in the presence of 1 mg/ml Geneticin (G418). Receptor expression was assessed by binding of the specific TP antagonist [ 3 H]SQ29,548 as well as by the agonist (U46619)-stimulated increase in total inositol phosphate formation.
Binding Assays-Binding of [ 3 H]SQ29,548 was performed on intact cells as follows. Briefly, subconfluent adherent cells in 24-well plates (2.5 ϫ 10 5 cells) were incubated for 30 min at 37°C in 250 l of serum-free DMEM, 0.2% bovine serum albumin (BSA) in the presence of increasing concentrations of [ 3 H]SQ29,548. Cells were washed twice in ice-cold phosphate-buffered saline (PBS), containing 0.02% BSA and lysed in 0.5 N NaOH. The protein content was determined by micro-BCA ® assay (Pierce) with the microbicinchoninic acid reagent and BSA as a standard. Cell number was assessed in parallel wells. Nonspecific binding was determined by the addition of 10 M unlabeled SQ29,548 and did not exceed 5-10% of the total binding.
Alternatively, crude membrane fractions were prepared as described (19) with minor modifications. Briefly, cells at confluence in 100-mm dishes were scraped in Hepes buffer (10 mM Hepes, pH 7.6, containing 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 0.5 mM dithiothreitol, and 10 M indomethacin), sonicated, and centrifuged for 15 min at 1,000 ϫ g at 4°C. Supernatants were further centrifuged at 125,000 ϫ g at 4°C for 1 h, and pellets were resuspended in the same buffer, containing 10% glycerol and 100 mM NaCl. Radioligand binding was performed using 50 g of membrane protein and increasing concentrations of [ 3 H]SQ29,548. Reactions were carried out in a total volume of 100 l at 4°C for 2 h. The reaction was stopped with 4 ml of 10 mM Tris, pH 7.4, followed by filtration through Whatman GF/C glass filters. Filters were subsequently washed twice with the same buffer and counted for radioactivity. Nonspecific binding was determined in the presence of excess unlabeled SQ29,548 (25 M). It did not exceed 15% of the total binding. For competition binding experiments, 50 g of membrane protein was incubated with 40 nM [ 3 H]SQ29,548 and increasing concentrations of U46619 (0.1-30 M). Reactions were carried out under the same conditions as above. The data were subject to Scatchard analysis, and an apparent K d and B max were determined using a computer program (Radlig Biosoft ® , Cambridge, UK). The same program was used for analysis of the competition experiments and for calculation of K i values for U46619.
Total Inositol Phosphate Formation-Inositol phosphate formation was measured as described previously (23). Briefly, confluent cells (4 -5 ϫ 10 5 cells in 12-well plates) were labeled with 2 Ci of myo-[2-3 H]inositol (18.3 Ci/mmol) for 20 -24 h in serum-free, inositol-free DMEM, containing 0.5% BSA, 20 mM Hepes. The culture medium was changed, and cells were incubated in the same medium containing 20 mM LiCl. After a 10-min incubation, U46619 was added for 10 min unless otherwise indicated. When SQ29,548 was used, it was added for 10 min prior to the addition of U46619. The reaction was stopped by aspiration of the supernatant and the addition of 0.75 ml of 10 mM formic acid. After incubation for 30 min at room temperature, the solution was collected in 3 ml of 5 mM ammonia hydroxide, giving a pH of 8.5-9. Samples were subjected to an anion exchange AG 1X-8 column. Free inositol and glycerophosphoinositol were washed with a 40 mM concentration of a formate/formic acid buffer, pH 5. Total inositol phosphates were eluted with 4 ml of the 2 M formate/formic acid buffer from which 1 ml was counted. Results were expressed as a percentage of the increase in agonist-stimulated total inositol phosphate formation compared with unstimulated cells.
Ca 2ϩ Release-Intracellular Ca 2ϩ levels ([Ca 2ϩ ] i ) were measured in TP␣-or TP␤-transfected HEK 293 cells in suspension as described previously (24). Briefly, cells were washed twice with PBS, loaded for 1 h at 37°C with 5 M Fura-2/AM (Molecular Probes, Eugene, OR) in phenol red free-RPMI 1640 culture media. Cells were further washed and incubated for 5 min in PBS, containing 1 mM EDTA and 5 mM EGTA, harvested, washed again, and resuspended at 10 6 cells/ml in RPMI 1640. Cells were allowed to sit for 1 h. Fluorescence was detected in suspended cells diluted at 10 6 cells/ml, using an SLM/Aminco AB2 spectrophotometer (Urbana, CA), and approximate values of [Ca 2ϩ ] i were calculated using a K d of 224 nM for Fura-2.
Cells (1 ϫ 10 6 /ml, 1.5 ml) were incubated with 1 M U46619 for the indicated times in the desensitization experiments. They were then washed and further stimulated with 300 nM U46619. Control cells were treated with vehicle (i.e. ethanol, 0.01%) alone and further assessed in parallel for mobilization of [Ca 2ϩ ] i .
Generation of Specific Anti-peptide Antibodies-Rabbit polyclonal antibodies were raised to immunoprecipitate the TP receptor isoforms. These were directed against the last 15 amino acids of the carboxylterminal ends of the isoforms: NH 2 -SLSLQPQLTQRSGLQ-COOH (amino acid sequence 327-341, referred to as Ab␣) for TP␣ and NH 2 -(C)PFEPPTGKALSRKD-COOH (amino acid sequence 355-369, referred to as Ab␤) for TP␤. The extra cysteine residue was added in the TP␤ peptide to allow coupling to BSA and acetylcholinesterase (see below). Immunization procedures and the generation of peptides were performed as described previously (25). Boosters were administered every 2 months. Sera from the same rabbit and within the same booster were pooled and kept sterile at 4°C in the presence of 0.01% sodium azide. The titer of the different sera was tested in a competitive enzyme immunoassay, using the corresponding TP receptor peptides covalently linked to acetylcholinesterase. The antisera Ab␣ and Ab␤ had titers of 1/5000 -1/10000 and 1/3000 -1/5000, respectively. Immunoaffinity columns of each antibody were prepared for immunoprecipitation analysis. Antisera were first partially purified using the E-Z-SEP ® kit (Pharmacia) and further incubated with CNBr-activated Sepharose according to the manufacturer's instructions. Normal rabbit immunoglobulins, coupled to CNBr-activated Sepharose, were used for preclearing of cell lysates.
Immunoblot analysis of total cell lysates or of the immunoprecipitates was performed to characterize the antibodies. Cell membranes were prepared as described previously, and 50 g were mixed with 1 volume of 2 ϫ Laemmli buffer (1 ϫ Laemmli buffer: 4% SDS, 5% glycerol, 60 mM Tris, pH 6.8, and 0.005% bromphenol blue) under reducing conditions (75 mM dithiothreitol) and vortexed. Samples were treated under nonreducing conditions for the immunoprecipitation experiments. SDS-polyacrylamide gel electrophoresis was performed using 10% acrylamide for the separating gel. Proteins were electrotransferred onto nitrocellulose membranes (Schleicher and Schuell, Keene, NH). Blots were saturated for 2 h in Tris-buffered saline (50 mM Tris, pH 7.5, 250 mM NaCl, 0.1% Tween 20) containing 2% BSA. Membranes were further incubated overnight at 4°C with Ab␣ (1/2000) or Ab␤ (1/1000) in Tris-buffered saline containing 0.5% BSA. Membranes were washed five times for 10 min in the same buffer without BSA and incubated for 1 h at room temperature with a 1/5000 dilution of a donkey anti rabbit antibody coupled to horseradish peroxidase (Jackson ImmunoResearch, West Grove, Pa). Excess antibody was washed, and positive bands were revealed by ECL (Amersham) according to the manufacturer's instructions.
Immunofluorescence Staining-HEK 293 cells overexpressing TP receptor isoforms were subcultured in a two-well Lab-Tek slide culture chamber (Nunc, Naperville, IL) coated with 10 g/ml of human fibronectin (Life Technologies, Inc.). Subconfluent cells were washed once with PBS buffer and fixed in 100% methanol. Fixed specimens were blocked for 30 min at room temperature in PBS containing 2% BSA. Specimens were incubated with the receptor antibodies (diluted at 1/1000 for Ab␣ and 1/500 for Ab␤) in PBS, containing 0.5% BSA for 1 h at room temperature followed by fluorescein isothiocyanate-conjugated donkey anti-rabbit IgG (Jackson ImmunoResearch) at a dilution of 1/1000 in the same PBS solution. Slides were mounted in a mounting solution (Vector Laboratories Inc., Burlingame, CA). Immunofluorescence staining was examined by an Olympus ® inverted fluorescence microscopy.
Phosphorylation of Intact Cells-Whole cell phosphorylation was performed as described (26) with minor modifications. Briefly, subconfluent cells in 60-mm dishes (2-3 ϫ 10 6 cells) were washed once in phosphate-free DMEM media, containing 20 mM Hepes, 0.5% BSA as above. Cells were further incubated for 45 min in the same medium, containing 100 Ci/ml of [ 32 P]orthophosphate (6,000 -7,000 Ci/mmol). Kinase inhibitors, SQ29,548, or vehicle were added during the labeling period. Stimuli or vehicle were added for 10 min, unless otherwise indicated. Me 2 SO and ethanol concentrations did not exceed 0.1% and did not modify the pattern of phosphorylation of the TP isoforms. The reaction was quenched by transferring the dishes on ice and aspirating the supernatants. Cells were washed with 2 ml of ice-cold PBS/dish and lysed with 0.8 ml of radioimmune precipitation buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40 (v/v), 0.5% sodium deoxycholate (w/v), 0.1% SDS (w/v) containing 10 mM sodium fluoride, 25 mM sodium pyrophosphate, 10 mM ATP, 1 g/ml leupeptin, 10 g/ml soybean trypsin inhibitor, 1 mM benzamidine hydrochloride, and 0.5 mM phenylmethylsulfonyl fluoride). Cells were scraped with a rubber policeman after 15 min on ice, passed through a 23 1 ⁄2 gauge needle, and centrifuged for 15 min at 10,000 ϫ g at 4°C. Supernatants were collected, and total cell protein was measured by micro BCA ® , using BSA as a standard. Lysates (0.8 mg of total cell protein in 0.8 ml) were precleared for 1 h at 4°C with 50 l of normal rabbit IgG covalently coupled to Sepharose CL-4B. The preclearing step was critical in the phosphorylation assays, since it removed a phosphorylated doublet, which migrates at the same molecular weight as the TP isoforms.
Samples were centrifuged for 3 min at 10,000 ϫ g, and supernatants were further immunoprecipitated overnight at 4°C using 50 l of immunoaffinity Sepharose for either Ab␣ or Ab␤, as described above. The beads were washed four times with 1 ml of radioimmune precipi-tation buffer and resuspended in 100 l of 1 ϫ Laemmli buffer under nonreducing conditions. Samples were vigorously vortexed for 15 min, centrifuged for 5 min at 10,000 ϫ g, and loaded onto SDS-polyacrylamide gels, as described above. Quantitative analysis of radioactivity in the samples was performed using a PhosphorImager 445-SI (Molecular Dynamics Inc., Sunnyvale, CA) after the gels were dried. Identical areas were integrated, and results were represented as percentages of the control values.

Characterization of the HEK 293 Cells
Overexpressing the TP Isoforms-To determine whether TP␣ or TP␤ couples differentially to downstream signals, we developed HEK 293 cells stably overexpressing both isoforms. Normal HEK 293 cells as well as cells transfected with the pcDNA3 vector alone (HEK 293-VEC) showed no detectable amounts of endogenous TPs in binding assays. Among several HEK 293 clones stably overexpressing the receptors, we selected clones that exhibited similar levels of receptors as expressed per number of binding sites per cell or per pmol of receptor per mg of total protein (Table I) Total Inositol Phosphate Formation and Calcium Release-We assessed the functionality of the expressed isoforms by measurement of agonist-induced total inositol phosphate (IP) formation and calcium mobilization. Both the TP␣-5 and TP␤-17 cell lines exhibited an increase in IP formation in the presence of U46619, a prostaglandin H 2 /TxA 2 mimetic. IP formation was similar in the presence (883 Ϯ 33.2 dpm) and the absence of U46619 (901 Ϯ 180 dpm) in HEK 293 cells transfected with the pcDNA3 vector alone and corresponded to values observed in unstimulated TP␣-5 and TP␤-17 clones.
The addition of increasing amounts of U46619 (3-3000 nM) for 10 min resulted in a 5-8-fold increase in IP formation, which reached a plateau at around 100 nM for TP␣ and 300 nM for TP␤ (Fig. 1). The EC 50 values for U46619 were higher for TP␤ than TP␣ (57 Ϯ 6.7 nM versus 11.1 Ϯ 2 nM, respectively, p Ͻ 0.002). U46619 rapidly increased IP formation via both isoforms (4 min). The increase in IP formation was linear up to 60 min (data not shown). Preincubation of either TP␣-5 or TP␤-17 with the antagonist SQ29,548 abolished the ability of either to transduce IP formation in response to U46619 stimulation (data not shown).
When TP␣-5 and TP␤-17 were loaded with Fura-2, U46619 induced a rapid increase in intracellular [Ca 2ϩ ] i (Fig. 2). In TP␣-5 ( Fig. 2A, left part), the increase in [Ca 2ϩ ] i mobilization was sustained up to 10 min and was dependent on extracellular sources, since chelation of extracellular Ca 2ϩ by pretreating the cells with 2 mM EGTA for 45 s led to a rapid return to the base line. In contrast, U46619 induced a more transient Ca 2ϩ mobilization in TP␤-17 (Fig. 2B, left part), although the response did not return completely to base line. Basal [Ca 2ϩ ] i was 46 Ϯ 13 nM (mean Ϯ S.D., n ϭ 21) for TP␣ and 51 Ϯ 18 nM (mean Ϯ S.D., n ϭ 13). For both isoforms, the presence of EGTA depressed Ca 2ϩ mobilization by roughly 60 -70%. Thus, Ca 2ϩ mobilization is dependent on extracellular, as well as intracellular, pools of Ca 2ϩ . Similarly, preincubation with 1 M U46619 abolished the increase of Ca 2ϩ in response to a second addition of U46619 in both cell lines (Fig. 2, A and B, right parts).
Homologous desensitization of the Ca 2ϩ response to U46619 was demonstrable for concentrations of U46619 as low as 10 nM and for periods of preincubation as short as 2 min (data not shown). No significant increase was observed when cells were incubated with vehicle alone or when they were pretreated with SQ29,548. HEK 293-VEC showed no Ca 2ϩ mobilization (data not shown).
Immunoblotting and Immunoprecipitation of the Thromboxane Receptors-Anti-peptide antibodies were raised against amino acid sequences unique to either the human TP␣ or the human TP␤ isoform. Both antibodies demonstrated specific immunoreactivity toward the receptors, as assessed by immunoblot analysis and immunoprecipitation of TP␣-5 or TP␤-17. Fig. 3A represents an immunoblot analysis of thromboxane receptors in crude membrane fractions of the TP␣-5 and TP␤-17 cells. The antiserum Ab␣ reacted with a broad protein band ranging from 55 to 65 kDa in TP␣-5 (Fig. 3A, lane 2). Similar results were obtained for the Ab␤ with the TP␤-17 membrane fractions (Fig. 3A, lane 4). No immunoreactivity was revealed in membranes from cells transfected with vector alone, by either Ab␣ (Fig. 3A, lane 1) or Ab␤ (data not shown). Preincubation of the Ab␣ or Ab␤ with the corresponding peptides suppressed the immunoreactivity in TP␣-5 (Fig. 3A, lane  3) or TP␤-17 (Fig. 3A, lane 5), respectively. The Ab␣ and Ab␤ antibodies did not cross-react (data not shown), indicating their specificity. The capability of these antibodies to immunoprecipitate the TP isoforms was demonstrated using immunoaffinity columns prepared with either Ab␣ or Ab␤ as described under "Experimental Procedures." Immunoaffinity columns prepared with these antibodies immunoprecipitated a broad protein band from TP␣-5 and TP␤-17 (Fig. 3B, lanes 3 and 4,  respectively), with an apparent molecular weight similar to that obtained by direct immunoblotting of the membrane fractions. It is important to note that since Ab␣ and Ab␤ are different sera and could have different characteristics, it is not  2 and 3), and TP␤-17 (lanes 4 and 5) cells were subjected to SDS-PAGE and electrotransferred to nitrocellulose. The membranes were further saturated in Tris-buffered saline buffer, containing 2% BSA. They were incubated with specific antibodies for TP␣ (lanes 1-3) or TP␤ (lanes 4 and 5). Immunoreactivity was revealed as described under "Experimental Procedures." The antisera were preadsorbed with the corresponding peptides prior to the incubation with the membranes in lanes 3 and 5. B, HEK-293 cells or washed human platelets were lysed, and 0.8 mg of total cell protein were immunoprecipitated using the immunoaffinity columns of Ab␣ and Ab␤ for TP␣ and TP␤ isoforms. Samples were processed as described above. Lanes 1, 3, and 5 correspond to HEK 293-VEC, TP␣-5, and human platelets, respectively, and were blotted with Ab␣; lanes 2 and 4 correspond to HEK 293-VEC and TP␤, respectively, and were blotted with Ab␤. possible to compare the quantities of the two isoforms that are immunoprecipitated. Similarly, immunoprecipitation of a human platelet cell lysate with the Ab␣ revealed a broad protein band with an apparent molecular weight of 45-50 (Fig. 3B, lane 5). No immunoreactivity could be detected when immunoprecipitation of the platelet cell lysate was performed with the Ab␤.
Membrane Localization of the TP Receptor Isoforms-The antibodies Ab␣ and Ab␤ recognized the receptors in situ as assessed by immunofluorescence staining of TP␣-5 and TP␤-17 clones (Fig. 4, A and C, respectively), Almost all of the cells stained positively for the TPs. Immunofluorescence staining was uniformly distributed over the cell surface. No staining was detected in the absence of Ab␣ or Ab␤ (Fig. 4, B and D,  respectively). Furthermore, no immunofluorescence was observed when anti-peptide antibodies were substituted with nonimmune rabbit serum or when the antibodies were saturated with the corresponding peptides (data not shown).
Specific Phosphorylation of Human TPs by U46619 -Incubation of [ 32 P]orthophosphate-prelabeled TP␣-5 or TP␤-17 with 1 M U46619 resulted in phosphorylation of a broad radioactive protein band that migrated at the same molecular weight as that of the TP isoforms described above (Fig. 5, A and B, for TP␣-5 and TP␤-17, respectively). A very weak signal was detected in the unstimulated cells (Control) or when HEK 293-VEC were incubated with U46619 (data not shown). The increase in U46619-dependent phosphorylation over unstimulated cells was 4 -5-fold (n ϭ 12) and 3-4-fold (n ϭ 8) for TP␣ and TP␤, respectively. Preincubation of the cells with 50 M SQ29,548 prevented phosphorylation of the TPs, whereas SQ29,548 alone did not induce any phosphorylation. The nature of the weakly phosphorylated band observed in the unstimulated cells is difficult to determine, since preclearing of the cell lysates with the nonimmune rabbit serum eliminated a nonspecific but strongly radioactive doublet that migrates exactly at the same molecular weight as that of TP␣ and TP␤. Immunoblot analysis of the immunoprecipitated receptors in these samples showed no difference in the quantity of the TP isoforms immunoprecipitated (data not shown). This indicates that the antisera may recognize both the phosphorylated and the unphosphorylated isoforms of the receptors. This is important, since the antibodies are directed against a polypeptide sequence present in the carboxyl-terminal end, where phosphorylation of the receptors may occur (9).
Dose-and Time-dependent Phosphorylation of the TP␣ and the TP␤ Isoforms by U46619 -32 P-Labeled TP␣-5 and TP␤-17 were stimulated with increasing concentrations of U46619 (1-1000 nM). Phosphorylation of the TP isoforms was detectable in TP␣-5 cells at concentrations as low as 3 nM U46619 and reached a plateau at ϳ 100 nM (Fig. 6A). Similar results were obtained for TP␤-17 cells (Fig. 6B). The EC 50 values for U46619 for phosphorylation of the TP␣ and TP␤ isoforms were similar (12 Ϯ 2 nM for TP␣, n ϭ 3; 11.5 Ϯ 1.3 nM for TP␤, n ϭ 4) (Fig.  6B). Phosphorylation of the TP␣ and TP␤ (Fig. 7, A and B, respectively) receptors was rapidly (1 min) detected at a saturating concentration of U46619 (300 nM) and reached a plateau at ϳ 30 min. Incubation times of up to 90 min resulted in no further change in the phosphorylation pattern. Basal phosphorylation was detected at longer incubation times (Ͼ30 min). were lysed in radioimmune precipitation buffer and precleared with a nonimmune rabbit serum affinity Sepharose column. Supernatants were further immunoprecipitated using the specific TP antibody affinity columns as described under "Experimental Procedures." Samples were subjected to SDS-PAGE, and the gels were dried. The signal was revealed by PhosphorImager analysis. The data are representative of at least three experiments. 66 and 49 correspond to the protein molecular weight markers. D) were blocked in PBS, containing 2% BSA, and incubated with anti-peptide antibodies specific for the TP receptor isoforms. Immunoreactivity was detected using a secondary antibody anti-rabbit coupled to fluorescein isothiocyanate and examined using an fluorescence microscope. The magnification is ϫ 150. TP␣-5 and TP␤-17 cells were incubated in the presence (A and C) or absence (B and D) of specific antibodies. rylation was observed with 4␣-PMA, an isomer of PMA that does not activate PKC. Increasing concentrations of GF 109203X (2-10 M) inhibited the PMA-dependent phosphorylation of the TP isoforms by 70 -90% (Fig. 8, A and B). Although high concentrations of GF 109203X reduced basal phosphorylation of the both TP isoforms, they inhibited U46619-dependent phosphorylation by only about 30%. Similar percentages of inhibition were obtained at lower concentrations of U46619. Prolonged preincubation with GF 109203X for up to 2 h or the use of a distinct PKC inhibitor, RO-31-8220 (28), in TP␣-5 cells gave similar results (data not shown). Furthermore, thrombin (2 units/ml) failed to induce phosphorylation of the TP isoforms, although it did cause a significant increase (80 -100 nM) in [Ca 2ϩ ] i mobilization (data not shown). Similar results were obtained when the cells were transiently transfected with the cDNA for the human thrombin receptor despite a 2-fold increase in thrombin receptor expression as analyzed by flow cytometry using specific anti-peptide antibody against human thrombin receptor (data not shown).

FIG. 4. Immunofluorescence staining of TP␣ or TP␤ receptor isoforms in HEK 293 cells using specific antibodies. Fixed HEK 293 cells stably overexpressing TP␣ (A and B) or TP␤ (C and
PKC down-regulation was accomplished by incubating the cells for 24 h in the presence of PMA (200 nM). Under these conditions, both TP␣ or TP␤ were still phosphorylated by 1 M U46619, but not by 100 nM PMA (Fig. 9).

Effect of Protein Kinase A Inhibition on U46619-dependent
Phosphorylation of the TP Receptors-Incubation of prelabeled cells with 10 M FK in the presence of 0.5 mM IBMX resulted in a weak phosphorylation of the TP␣ isoforms. The capacity for PKA-dependent phosphorylation of both TP isoforms was illustrated when Bt 2 cAMP was incubated for 30 min in the presence of IBMX. The phosphorylation of TP␣ receptors by FK was strongly inhibited by H-89 (29), a competitive inhibitor of PKA (Fig. 10A). Under these conditions, U46619-dependent TP receptor phosphorylation was not inhibited by H-89. Homologous Desensitization of TP Isoform-mediated Inositol Phosphate Formation-Pretreatment of TP␣-5 or TP␤-17 with 300 nM U46619 for 10 min resulted in 70 -90% inhibition of IP formation in response to a subsequent addition of U46619 (Fig.  11, A and B). Desensitization was rapid for both receptors, reaching a maximum after 1-2 min of pretreatment (Fig. 11C). Maximal desensitization was observed when pretreating the cells with ϳ100 nM U46619. Moreover, incubation of cells with 5 M GF 109203X prior to pretreatment of the cells with U46619 did not modify the pattern of desensitization for either the TP␣ or TP␤ isoform, as assessed by IP formation (Fig. 12, A  and B, respectively). Along with these results, the pretreatment of the cells with either 200 nM PMA for 10 min or 0.5 mM Bt 2 cAMP for 30 min did not modify U46619-mediated IP formation (Table II). Neither PMA or Bt 2 cAMP pretreatment modified the extent of homologous desensitization of TPs under these conditions. DISCUSSION TxA 2 is an evanescent biological mediator; it exerts potent effects on platelet function and vascular tone in the immediate microenvironment of its formation (18). Given its critical role in determining vascular patency, it would seem likely that its formation and effects would be tightly regulated. TxA 2 is not stored, preformed in cells. Rather, it is formed and released rapidly in response to cellular (e.g. platelet) activation by diverse stimuli. Its effects are limited by its hydrolysis to the inactive thromboxane B 2 , which has a half-life estimated to be 30 s at physiological pH (2) and by homologous desensitization of its membrane receptor-mediated responses (30,31). This desensitization appears to result initially from uncoupling of the TP from its attendant G proteins; we have previously estimated that the half-life of such a response (coupling to phospholipase C in human platelets) is approximately 120 s. This phenomenon is followed by more gradual loss of binding sites from platelet membranes (30).
A single gene encoding a TP has been cloned (5). This predicts membership of the TPs in the family of GPCRs, consistent with their biochemical characteristics (32). It appears that the TP is subject to alternative splicing in the carboxyl-terminal region, as has been previously described for E prostaglandin receptor type 3 (33). Little is known about the functional significance of this observation. However, pharmacological studies have indicated the possibility of tissue-specific differences in the characteristics of TPs and, indeed, differences in TP binding of ligands within a single cell, platelets (11,34,35). Observations involving E prostaglandin receptor carboxyl-terminal isoforms indicate that they are capable of coupling to distinct downstream signaling systems and that they may differ in the rate and extent to which they are subject to homologous desensitization (36). Narumiya and colleagues have recently reported that mRNAs for the two TP isoforms are expressed in human platelets. When the isoforms are overexpressed in COS cells, they may regulate adenylate cyclase activation differentially (10).
Given these observations, we wished to explore the molecular mechanisms that might underlie homologous desensitization of the two TP isoforms. Importantly, they differ in the length and number of potential target residues for phosphorylation (9), raising the possibility of differential tissue responses to TxA 2 (e.g. in platelets and the vasculature) based on differential rates of desensitization. To address this possibility and to clarify the role of specific kinases in the desensitization process, we characterized specific, peptide-based antibodies for the two receptors. These confirmed the membrane localization of the two receptor isoforms, as inferred by their sequences.
Interestingly, both isoforms were rapidly phosphorylated following exposure to agonist. The time and dose dependence of this phenomenon seemed similar for the two isoforms. Both isoforms coupled to downstream signaling systems. We focused on phospholipase C-dependent events, since these are thought most relevant to TxA 2 -mediated platelet aggregation and vasoconstriction (11,34). Again the TP agonist, U46619, could stimulate an increase in [Ca 2ϩ ] i and total inositol phosphates via both isoforms. However, the pattern of the calcium response evoked via the two isoforms appeared to differ, at least in the stable cell lines that we established. Thus, whereas U46619 induced a rapid transient, followed by a delayed plateau in cells expressing TP␣, the plateau phase was almost absent in the cells expressing TP␤. Activation of both isoforms appeared to involve release of calcium from intracellular stores. However, the plateau phase of the TP␣ response appeared to derive from an extracellular source, raising the possibility of linkage to a receptor-activated calcium channel (37). Further experiments will address this possibility and whether a similar distinction is evident in other cells transfected with the two isoforms.
Given the role of phospholipase C-mediated responses in the biological consequences of TP activation, it would seem likely that downstream activation of PKC might play a central role in homologous desensitization. Indeed, we have previously shown that PKC may phosphorylate a fusion protein based on the carboxyl-terminal end of the TP␣ receptor (residues 321-343) in vitro (19). Peptide competition experiments suggested that this involved sites in the third extracellular loop and the carboxyl-terminal tail (19). Okwu et al. (20) have recently presented data consistent with TP phosphorylation in human platelets. Phosphorylation plays an important role in regulating both homologous and heterologous desensitization of GPCRs like the TPs. Additionally, other serine/threonine kinases such as PKA, G protein-coupled receptor-dependent kinases, and tyrosine kinases have been implicated in GPCR phosphorylation (38 -40). We have demonstrated that PKA may phosphorylate TP␣ to a minor degree in vitro (19). Similarly, we demonstrate the capacity for PKA to phosphorylate the receptor in the present study in vivo. However, PKA appears to contribute trivially, if at all, to rapid phosphorylation of the TPs induced by U46619. Although there is no information on G protein-coupled receptor-dependent kinase-mediated phosphorylation of eicosanoid receptors, recent studies have demonstrated their ability to phosphorylate angiotensin and adrenoreceptor GPCRs (41,42).
We also demonstrate that PKC appears to play a modest role in agonist-dependent phosphorylation of either TP isoform. Thus, two specific competitive inhibitors of PKC, GF 109203X and RO 31-8220, failed to modify the rate or extent of agoniststimulated phosphorylation of either isoform. Similarly, downregulation of PKC, by prolonged exposure to the phorbol ester, PMA, also failed to alter these phenomena. We demonstrate that homologous desensitization of agonist-evoked increases in IPs is similar for both isoforms. Again, the PKC inhibitor, GF 109203X, failed to modify homologous desensitization of this response evoked via either isoform. Although PKC appears largely irrelevant to the agonist-mediated desensitization response, activation of this enzyme may phosphorylate TPs. Thus, short term exposure to PMA will result in phosphorylation of both isoforms. However, treatment of the cells with thrombin (2 units/ml) did not induce phosphorylation of the TPs, although cells were activated, as assessed by [Ca 2ϩ ] i mobilization. Transient co-expression of the human thrombin receptor and stimulation with thrombin did not result in the phosphorylation of the TP receptor isoforms. Similarly, we have recently shown that phosphorylation of the prostacyclin receptor (43) can be induced by thrombin in a PKC-dependent manner, confirming that the endogenous thrombin receptor in HEK-293 cells is sufficient to mediate PKC-dependent phosphorylation of a related GPCR.
In contrast to agonist-dependent phosphorylation, the PKC inhibitors appeared to diminish basal phosphorylation of both isoforms.
Similar to our findings with PKC, PKA appeared to be of trivial relevance to agonist-dependent phosphorylation of either isoform. Both FK and Bt 2 cAMP treatment of the stable transfectants resulted in phosphorylation of both isoforms under favorable conditions (i.e. pharmacological inhibition of phosphodiesterases), extending our previous in vitro observations (19). However, the PKA inhibitor, H-89, had no apparent effect on U46619-induced phosphorylation of either isoform. Although U46619 is a potent stimulus to phospholipase C activation, higher concentrations may also activate adenylate cyclase (10). However, these concentrations do not appear to result in TP phosphorylation by PKA in vivo. It is likely that, as with PKC, activation of this enzyme might assume more importance in TP phosphorylation during heterologous desensitization.
The use of our antibodies identified TPs as a broad band ranging from 55 to 65 kDa. This molecular mass is somewhat higher than that previously observed by others in human platelets (44,45). Indeed, we identified a 45-50-kDa protein when we immunoprecipitate human platelet TP receptors with Ab␣ (Fig. 3B, lane 5). This may reflect different degrees of glycosylation or palmitoylation (46) of TP isoforms in HEK-293 cells compared with platelets. Interestingly, the molecular weight of the phosphorylated protein is similar to that in the immunoblot analysis. The TP antagonist prevents phosphorylation of that band, further confirming its identity.
In conclusion, we have demonstrated that the rate and extent of agonist-dependent phosphorylation of the two cloned TP isoforms is similar, although their coupling to [Ca 2ϩ ] i and IP formation is somewhat different. Neither PKC nor PKA plays a major role in such phosphorylation of either isoform, implicat-ing the G protein-coupled receptor-dependent kinases in homologous desensitization of these receptors. However, both TP isoforms are substrates for PKC and PKA in vivo, and these enzymes may play a more important role in their heterologous desensitization or in receptor regulation in other cells. The availability of specific antibodies for TP isoforms is likely to facilitate investigation of their comparative distribution and biology.