The α, but Not the β, Isoform of the Human Thromboxane A2 Receptor Is a Target for Prostacyclin-mediated Desensitization*

In this study, we examined the effects the prostacyclin receptor (IP) agonist cicaprost exhibited on U46619-mediated thromboxane A2 receptor (TP) signaling in platelets and compared it to that which occurs in human embryonic kidney (HEK) 293 cells stably overexpressing the individual TPα or TPβ isoforms. Consistent with previous studies, cicaprost abrogated U46619-mediated platelet aggregation and mobilization of intracellular calcium ([Ca2+] i ). In HEK 293 cells, signaling by TPα, but not TPβ, was subject to IP-mediated desensitization in a protein kinase A-dependent, protein kinase C-independent manner. Desensitization of TPα signaling was independent of the nature of the IP agonist used, the level of IP expression, or the subtype of Gq protein. Signaling by TPΔ 328, a truncated variant of TP devoid of the divergent residues of the TPs, or by TPαS329A, a site-directed mutant of TPα, were insensitive to IP agonist activation. Whole cell phosphorylations established that TPα, but not TPβ or TPαS329A, is subject to IP-mediated phosphorylation and that TPα phosphorylation is inhibited by H-89. Thus, we conclude that TPα, but not TPβ, is subject to cross-desensitization by IP mediated through direct protein kinase A phosphorylation at Ser329 and propose that TPα may be the isoform physiologically relevant to TP:IP-mediated vascular hemostasis.

The prostanoids thromboxane A 2 (TXA 2 ) 1 and prostacyclin play key, yet opposing, roles in the maintenance of vascular hemostasis (1). TXA 2 , which is synthesized mainly by platelets, mediates platelet shape change and aggregation and constriction of vascular and bronchial smooth muscle, whereas prostacyclin, which is synthesized mainly by the vascular endothelium, is a potent inhibitor of platelet aggregation and induces vasodilation (2). TXA 2 may also induce prostacyclin release from endothelial cells in vivo (3). Perturbations in the levels of TXA 2 or prostacyclin, or their synthases or receptors, have been implicated in various cardiovascular disorders (4 -8). However, the molecular mechanisms underlying the counter-regulation of TXA 2 and prostacyclin signaling are poorly understood.
Both TXA 2 and prostacyclin exert intracellular effects by interaction with specific members of the G protein-coupled receptor (GPCR) family, termed TP and IP, respectively (9,10). There are two isoforms of TP in humans, termed TP␣ and TP␤, as recommended by the International Union of Pharmacology classification on prostanoid receptors (11,12). These receptors, which are identical for their first 328 amino acids and differ exclusively in their carboxyl-terminal cytoplasmic tail (C-tail) regions, arise due to alternative splicing in exon 3 of the TP gene (12,13). The physiologic relevance for the existence of two receptors for TP is currently unknown. Wide cell and tissue distribution of the mRNA for both TP isoforms was recently confirmed by selective reverse transcription-PCR procedures (14). Isoform specific antibodies permitted detection of TP␣, but not TP␤, in human platelets, leading to the suggestion that TP␣ may be the predominant isoform in platelets (15), despite the presence of mRNA for both isoforms in platelets (16). The major signaling pathway used by TP in vivo is G protein-dependent stimulation of the ␤-isoforms of phospholipase C (PLC␤), resulting in increased intracellular concentrations of diacylglycerol and inositol 1,4,5-trisphosphate (IP 3 ) and mobilization of intracellular calcium ([Ca 2ϩ ] i ) (17). Using a variety of in vitro approaches, various investigators have proposed that the platelet TPs might couple to the heterotrimeric G proteins G q , G 12 , G 13 , G 16 , and G i2 (18 -25). It was recently demonstrated that the cloned TP␣ can functionally couple to both G q and G 11 following stimulation with the selective TXA 2 mimetic U46619 and the isoprostane 8-epi-prostaglandin F 2␣ to mobilize [Ca 2ϩ ] i (26). Coupling to G 11 was more efficient than that to G q . Both TP isoforms couple similarly to G 11 in stably transfected HEK 293 cells (27) but oppositely regulate adenylyl cyclase activity in transfected Chinese hamster ovary cells (16), suggesting a possible role for the C-tail in determining G protein specificity. Moreover, G h , the novel high molecular weight G protein that may also function as a transglutaminase (28 -31) can mediate agonist activation of TP␣, but not TP␤, leading to inositol phosphate production due to PLC activation (32).
A single receptor, termed IP, appears to mediate the actions of prostacyclin leading to activation of adenylyl cyclase via G s and elevation of intracellular cAMP (33), a signaling system thought to be important in control of both vascular tone and platelet aggregation (34). However, IP may also couple to multiple G protein/effector systems including phosphoinositide turnover via a pertussis toxin insensitive G protein (35,36). In human erythroleukemia cells, IP has even been proposed to differentially couple to both G s and G i (37). Iloprost, a stable carbacyclin analogue of prostacyclin, can stimulate opening of ATP-sensitive K ϩ channels, leading to hyperpolarization and relaxation of canine carotid artery (38). IP is unique among the family of GPCRs in that it undergoes posttranslational modification by carbon-15 farnesyl isoprene groups (39). This isoprenylation is absolutely required for receptor activation of adenylyl cyclase via G S and for efficient coupling to PLC via G q or G 11 (39).
A commonly observed phenomenon among GPCRs is desensitization, defined as reduced receptor responsiveness to repeated agonist challenge (40). GPCR desensitization consists of two key mechanisms, namely phosphorylation of the receptor by specific serine/threonine kinases and sequestration or internalization of receptors to intracellular vesicles where they are unavailable for interaction with G proteins. GPCRs can be subject to either homologous (41,42) or heterologous (42)(43)(44)(45)(46)(47) desensitization, mediated via phosphorylation by G proteincoupled receptor kinases or the second messenger-activated protein kinases, including cAMP-dependent PKA and PKC. Such desensitizations provide mechanisms for feedback regulatory loops following receptor activation and signaling and also for cross-talk between different second messenger systems (43). TP␣ may be phosphorylated in vitro, in the third extracellular loop and the C-tail, by both PKA and PKC (48). Differences in the complement and distribution of serine (Ser) and threonine (Thr) residues within the divergent C-tails of TP␣ and TP␤ could affect their sensitivity to phosphorylation. Both TPs may be phosphorylated in response to stimulation with the TXA 2 mimetic U46619 in transfected HEK 293 cells (49), and recent studies indicate that TP␤ but not TP␣ undergoes agonist induced internalization (50). Like TP, IP is sensitive to desensitization by second messenger kinases following stimulation with the IP agonist iloprost, with a single PKC phosphorylation site being critical for its desensitization (51,36).
Thus, both TP and IP are potentially vulnerable to "heterologous desensitization" by elements of intracellular cascades induced by activation of other receptors; for example, the IP: adenylyl cyclase system is essential to the control of platelet responses and may be manifested at different levels of the signaling system (52). Indeed, cross-talk occurs between TP and IP in human platelets, with prior U46619 stimulation enhancing iloprost-mediated generation of cAMP (52). Similarly, in the megakaryoblastic cell line MEG-01, TXA 2 mimetics U46619 and STA 2 dose-dependently augment subsequent iloprost-induced cAMP formation in a PKC-dependent manner (53). In view of the interplay between TXA 2 and prostacyclin in the maintenance of vascular homeostasis, we considered the potential influence that the intracellular signaling processes induced by IP may have on TP function. In particular, we examined the effect that the selective, high affinity IP agonists cicaprost and iloprost exhibited on U46619-mediated TP responses in platelets. More specifically, we investigated whether cross-talk between IP and TP signaling exists and considered whether such cross-talk may have differential impacts on signaling by the individual TP␣ and TP␤ isoforms. Our results indicate that TP␣, but not TP␤, appears to be a direct target for cross-talk between IP: TP responses. Furthermore, H-89, a selective inhibitor of PKA (54,55), but not the PKC inhibitor GF 109203X (56), reduced IP-mediated desensitization of TP␣ and platelet TP(s), whereas TP␤ was insensitive to this desensitization pathway. Prior exposure of HEK.TP ⌬328 cells, stably overexpressing a variant of TP truncated at amino acid 328 at the point of divergence of TP␣ and TP␤, to cicaprost or iloprost did not affect subsequent U46619-mediated TP signaling, implying that the C-tail region is a crucial determinant of heterologous desensitization of TP␣ by IP-mediated signaling. TP␣ and TP␤ are predicted to contain 9 and 10 putative PKA sites, respectively; however, 8 are conserved between both isoforms, and thus, TP␣ and TP␤ contain 1 and 2 putative PKA sites, respectively, within their unique C-tail sequences. Thus, TP␣ is predicted to contain a unique PKA consensus site within its divergent C-tail, where Ser 329 represents the putative target residue for phosphorylation. U46619-mediated [Ca 2ϩ ] i mobilization by HEK.TP␣ S329A cells stably overexpressing a sitedirected mutant of TP␣ was insensitive to IP (cicaprost or iloprost)-mediated desensitization, confirming that Ser 329 is a target for IP-mediated desensitization. Finally, whole cell phosphorylation assays established that TP␣, but not TP␤ or TP␣ S329A , is subject to IP-mediated phosphorylation and that phosphorylation of TP␣ is abrogated in the presence of H-89. Thus, taken together, our results establish that TP␣, but not TP␤, is subject to cross-desensitization by IP that is mediated through direct PKA phosphorylation of Ser 329 and therefore imply that TP␣ may be the isoform physiologically relevant to the maintenance of vascular hemostasis.  (15.3 Ci/mmol) and iloprost were purchased from Amersham Pharmacia Biotech. Ultraspec total RNA isolation system was obtained from Biotecx Laboratories (Houston, TX); Moloney mouse leukemia virus reverse transcriptase, RNasin, deoxyribonucleotides, and Taq DNA polymerase were obtained from Promega. Expand High Fi-delity® Taq DNA polymerase, Chemiluminescence Western blotting kit, polyvinylidene difluoride membrane, and rat monoclonal 3F10 anti-HA-horseradish peroxidase-conjugated antibody were obtained from Roche Molecular Biochemicals. Mouse monoclonal 101R anti-HA-peroxidase antibody (5-7 mg/ml) was obtained from Babco; horseradish peroxidase-conjugated goat anti-mouse secondary antibody was from Santa Cruz Biotechnology; protein G-Sepharose 4B Fast Flow was obtained from Sigma. All oligonucleotides were synthesized by Genosys Biotechnologies.
Deletion of the amino acids carboxyl to Arg 328 , at the point of divergence between TP␣ and TP␤ was achieved by conversion of Ser codon 329 to a stop codon (Ser 329 TCG to stop 329 TAA). Site-directed mutagenesis was performed by PCR mutagenesis using pCMV:TXR as template and oligonucleotides 5Ј-CTCTAAGCTTATG TGG CCC AAC GGC AGT-3Ј (sense primer; nucleotides ϩ1 to ϩ18 of TP sequences are underlined) and 5Ј-CTCTGGATCCTTATCTGGGCCGGGTGCTGAG-3Ј (antisense primer; sequences complementary to nucleotides ϩ 967 to ϩ984 of TP sequences are underlined, and the mutator in-frame stop codon is in boldface italics). The resulting PCR-amplified cDNA was subcloned into the HindIII-BamHI site of pcDNA3 (Invitrogen) to generate pcDNA3:TP ⌬328 . Conversion of Ser 329 of TP␣ to Ala 329 was performed by PCR mutagenesis using pCMV:TXR as template and oligonucleotides 5Ј-GAGAAGCTTG ATG TGG CCC AAC GGC AGT TCC-3Ј (sense primer; nucleotides ϩ1 to ϩ21 of TP sequences are underlined) and 5Ј-CTCT AAGCTT CTA CTG CAG CCC GGA GCG CTG CGT GAG CTG GGG CTG GAG GGA CAG CGC CCT GGG CCG GG-3Ј (antisense primer; nucleotides complementary to nucleotides ϩ 974 to ϩ 1032 of TP␣ sequences are underlined, and the sequence complementary to mutator Ser (TCG) to Ala (GCG) codon is in boldface italics). PCR amplifications were performed using Expand High Fidelity® Taq DNA polymerase, and products were subcloned into the HindIII site of pHM6 to generate pHM:TP␣ S329A . The plasmids pcDNA3:TP ⌬328 and pHM: TP␣ S329A were verified by double-stranded DNA sequencing using Sequenase Version 2.0 (United States Biochemical Corp.). The plasmids pCMV:G␣ 11 , pCMV:G␣q, and pcDNA3:mIP, coding for G␣ 11 , G␣ q , and mouse prostacyclin receptor (IP), respectively, have been described previously (26,39).
Cell Culture and Transfections-Human embryonic kidney (HEK) 293 cells were obtained from the American Type Culture Collection and were grown in minimal essential medium containing 10% fetal bovine serum.
Preparation of Platelets-Blood was drawn via venipuncture from normal human volunteers, who had not taken any medication for at least 10 days, into syringes containing indomethacin (10 M) and 3.8% sodium citrate (9:1 v/v) (final concentration, 0.38% sodium citrate). The blood was centrifuged for 10 min at 160 ϫ g; and the platelet-rich plasma was removed and recentrifuged for 10 min at 160 ϫ g to remove contaminating red blood cells. Where necessary, platelet-poor plasma was prepared by spinning the remaining blood at 900 ϫ g for 15 min. For aggregation studies, platelets in platelet-rich plasma were diluted to approximately 10 8 platelets/ml in platelet-poor plasma; 0.5-ml aliquots were preincubated at 37°C for 2 min before addition of the aggregating agent (1 M U46619, 1 M cicaprost) or vehicle, and the extent of aggregation was monitored by light transmission in a Biodata Pap 4 aggregometer. TXB 2 formation was routinely measured in platelet-rich plasma and platelet-poor plasma by enzyme immunoassay using a thromboxane B 2 EIA kit (Cayman Chemical Co.).
Calcium Measurements-Measurements of intracellular calcium either in transfected HEK 293 cells or in platelets were made by monitoring the intensity of Fura2 fluorescence as described previously (26). Cells were stimulated with 1 M U46619, 1 M cicaprost, 1 M iloprost unless otherwise specified. The PKA inhibitor H-89 or PKC inhibitor GF 109203X was added at times and concentrations specified in the figure legends. In all cases, the drug (agonists/kinase inhibitors in ethanol or DMSO) was diluted 1:1000 in vehicle (modified Ca 2ϩ /Mg 2ϩfree Hank's buffered salt solution, containing 10 mM HEPES, pH 7.67, 0.1% bovine serum albumin) for HEK 293 cells and platelet resuspension buffer (10 mM HEPES, 145 mM NaCl, 5 mM KCl, 5.5 mM glucose, pH 7.4) for platelets, and 20 l of the vehicle (containing an equivalent volume of the drug solvent) or drug in vehicle was added to 2 ml of cells; the vehicle had no effect on [Ca 2ϩ ] i mobilization by either TP isoform and had no effect on experimental data. The ratio of the fluorescence at 340 nm to that at 380 nm is a measure of [Ca 2ϩ ] i (59), assuming a K d of 225 nM Ca 2ϩ for Fura2/AM. The results presented in the figures are representative data from at least four independent experiments and are plotted as changes in intracellular Ca 2ϩ mobilized as a function of time upon ligand stimulation.
Radioligand Binding Studies-TP radioligand binding assays were carried out at 30°C for 30 min in 100-l reactions in the presence of 0 -40 nM [ 3 H]SQ29,548 for Scatchard analyses or in the presence of 20 nM [ 3 H]SQ29,548 for saturation radioligand binding experiments (26). IP radioligand binding assays were carried out on nonfractionated HEK 293 cells, as described previously (39). Protein determinations were carried out using the Bradford assay (60).
Measurement of IP 3 Levels-Measurement of IP 3 levels in HEK 293 cells was made on the basis of competition between unlabeled IP 3 and a fixed concentration of [ 3 H]IP 3 for binding to an IP 3 -binding protein derived from bovine adrenal glands, essentially as described by Godfrey (61). Briefly, cells were harvested by scraping, washed twice in ice-cold phosphate-buffered saline, and 2 ϫ 10 6 cells were resuspended in 200 l of HEPES-buffered saline (HBS) (140 mM NaCl, 4.7 mM KCl, 2.2 mM CaCl 2 , 1.2 mM KH 2 PO 4 , 11 mM glucose, 15 mM HEPES-NaOH, pH 7.4) supplemented with 10 mM LiCl. Cells were preequilibrated in this buffer at 37°C for 10 min and stimulated for 1 min at 37°C in the presence of cicaprost (1 M) or U46619 (1 M), in the presence of cicaprost (1 M) for 1 min followed by U46619 (1 M) for 1 min, or, to determine basal IP 3 levels in cells, in the presence of an equivalent volume (50 l) of HBS vehicle. IP 3 extraction and quantification was determined by radioimmunoassay essentially as described by Godfrey (61). Levels of IP 3 produced by ligand-stimulated cells over basal stimulation, in the presence of HBS, were expressed in pmol of IP 3 /10 6 cells Ϯ S.E. and as fold stimulation over basal (fold increase Ϯ S.E.). The data presented are representative of four independent experiments, each performed in duplicate.
Measurement of cAMP-cAMP produced was quantified by radioimmunoassay using the cAMP-binding protein from bovine adrenal medulla (62) as described previously (39). Levels of cAMP produced by cicaprost-stimulated cells over basal stimulation, in the presence of HBS, were expressed in pmol of cAMP/mg of cell protein Ϯ S.E. and as fold stimulation over basal (fold increase Ϯ S.E.). The data presented are representative of four independent experiments.
Reverse Transcriptase-Polymerase Chain Reaction-Total RNA isolated from human erythroleukemia 92.1.7 or HEK 293 cells using the Ultraspec RNA isolation procedure was converted to first strand cDNA with Moloney murine leukemia virus reverse transcriptase, as described previously (14). Aliquots (3.5 l) of each first strand cDNA were used as templates in PCRs (25 l) using the following primers specific for the human IP cDNA: primer A, 5Ј-GCTCCCTGCCTCTCACGATC-CGCTGCTTCACCC-3Ј (sense primer); and primer B, 5Ј-GTGGGGATC-CAAGCTTTCAGCAGAGGGAGCAGGC-3Ј (antisense primer). Primers were designed to span across intron 2 of the human IP gene (63) to distinguish products derived from first strand cDNA from trace genomic DNA present in the total RNA.
Measurement of Agonist-mediated TP Phosphorylation-Whole cell phosphorylation assays were performed essentially as previously (49) with certain modifications. Briefly, cells (2-3 ϫ 10 6 cells in 60-mm dishes) were washed once in phosphate-free Dulbecco's modified Eagle's medium, 10% dialyzed fetal bovine serum and were metabolically labeled for 60 min in the same medium (1.5 ml/60-mm dish) containing 100 Ci/ml [ 32 P]orthophosphate (8000 -9000 Ci/mmol) at 37°C, 5% CO 2 . Where appropriate, kinase inhibitor (H-89, 10 M) or vehicle was added during the labeling period. Thereafter, specific ligand or vehicle was added for 10 min. Reactions were terminated by transferring the dishes to ice and aspirating the labeling medium. Cells were washed once in ice-cold phosphate-buffered saline (2 ml/dish) and were lysed with 0.6 ml of radioimmune precipitation buffer (50 mM Tris-Cl, 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, 0.5 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate. Following 15 min of incubation on ice, cells were harvested using a rubber policeman and disrupted by sequentially passing through hypodermic needles of decreasing bore size (G20, G21, G23, and G26), and soluble cell lysates were harvested by centrifugation for 15 min at 13,000 ϫ g at room temperature. HA epitope TP receptors were immunoprecipitated using the anti-HA antibody (101R, 1:300 dilution) at room temperature for 2 h followed by the addition of 10 l of protein G-Sepharose 4B (Sigma) and further incubation at room temperature for 1 h. Immune complexes were collected by centrifugation at 13,000 ϫ g at room temperature for 5 min and were washed three times in 0.5 ml of radioimmune precipitation buffer and finally resuspended in 1ϫ solubilization buffer (10% ␤ mercaptoethanol (v/v), 2% SDS (w/v), 30% glycerol (v/v), 0.025% bromphenol blue (w/v), 50 mM Tris-HCl, pH 6.8; 40 l). Samples were loaded without boiling onto 10% polyacrylamide gels, analyzed by SDS-polyacrylamide gel electrophoresis, and thereafter electroblotted onto polyvinylidene difluoride membranes, essentially as described previously (39). Electroblots were then exposed to Eastman Kodak Co. Xomat XAR film to detect 32 P-labeled proteins. Thereafter, blots were subject to phosphorimage analysis, and the intensities of phosphorylation relative to basal phosphorylation were determined and were expressed in arbitrary units of intensity relative to basal levels. In parallel experiments, cells were incubated under identical conditions in the absence of [ 32 P]orthophosphate; HA-TP receptors were immunoprecipitated from those cells, and immunoblots were screened using the anti-HA antibody to check for quantitative recovery of each receptor type. Thereafter, membranes were screened by immunoblot analysis using the anti-HA 3F10 horseradish peroxidase conjugate; immunoreactive proteins were visualized using the ECL detection system, as described by the manufacturer (Amersham Pharmacia Biotech).
Data Analyses-Radioligand binding data were analyzed using GraphPad Prism V2.0 (GraphPad Software Inc.) to determine the K d and B max values. Statistical analyses were carried out using the un-paired Student's t test using the Statworks Analysis Package. p values of less than or equal to 0.05 were considered to indicate a statistically significant difference. ] i , indicating that the IP does not exhibit significant coupling to PLC␤ isozymes in human platelets. Platelet aggregation studies indicated that although platelets aggregated irreversibly in response to 1 M U46619, this aggregation was completely blocked by prior stimulation with 1 M cicaprost (Fig. 1, C and  D). To ensure that the observed effects in platelets were due to the addition of external ligands, rather than due to the production of endogenous cyclooxygenase-derived products, effective inhibition of platelet cyclooxygenase by 10 M indomethacin was confirmed by routine measurement of levels of TXB 2 in platelet-rich plasma before and after external agonist activation (data not shown).

Differential Effects of Cicaprost on U46619 Signaling via TP␣ and TP␤ Isoforms-To
To assess whether the abrogation of U46619-induced [Ca 2ϩ ] i mobilization by cicaprost observed in human platelets may actually involve possible molecular cross-talk between IP activation and TP signaling and, if so, whether it may be targeted toward a specific TP isoform or both isoforms, we utilized HEK 293 cell lines stably expressing either TP␣ (HEK.␣10) or TP␤ (HEK.␤3) (27). TP signaling was assessed by measurement of [Ca 2ϩ ] i mobilization in Fura2/AM-loaded cells in response to U46619 (1 M). In the case of both cell lines, for efficient mobilization of [Ca 2ϩ ] i in response to U46619, it was necessary to co-transfect the cells with the ␣ subunit of a member of the G q family of heterotrimeric G proteins (G␣ 11 , for example) (Fig. 2, A and E).
HEK.␣10 cells co-transfected with G␣ 11 showed efficient mobilization of [Ca 2ϩ ] i in response to 1 M U46619 ( Fig. 2A). Cicaprost (1 M) stimulated a 2-fold elevation of intracellular cAMP levels over vehicle-treated cells (Fig. 2C), thereby confirming the presence of IP in HEK.␣10 cells. The presence of mRNA encoding IP in HEK 293 cells was also confirmed by reverse transcription-PCR (Fig. 2D). Cicaprost at 1 M (Fig. 2B) or 10 M (data not shown) did not stimulate [Ca 2ϩ ] i mobilization in HEK.␣10 cells transiently co-transfected with G␣ 11 , indicating that the endogenous IP receptors present in these cells lack the ability to couple to PLC␤. However, following prior stimulation with 1 M cicaprost, mobilization of [Ca 2ϩ ] i induced by U46619 was significantly reduced to 44.3 Ϯ 7.3% of that generated by U46619 stimulation only (Fig. 2B, p Ͻ 0.001). This implies that, as in human platelets, activation of IP leads to desensitization of TP␣ signaling in HEK 293 cells, confirming cross-talk between the cAMP signaling system induced by IP activation and the IP 3 dependent Ca 2ϩ mobilization system induced by TP␣ activation.
To investigate whether the "cross-talk" between the IP and TP signaling systems extended to TP␤, the effects of cicaprost on subsequent U46619-induced mobilization of [Ca 2ϩ ] i by HEK.␤3 cells was monitored. HEK.␤3 cells stimulated with 1 M U46619 showed efficient mobilization of [Ca 2ϩ ] i , which was dependent on the presence of G␣ 11 (Fig. 2E). Similar to that observed in platelets and HEK.␣10 cells, HEK.␤3 cells cotransfected with G␣ 11 did not support [Ca 2ϩ ] i mobilization in response to cicaprost (Fig. 2F). However, in contrast to both platelets and HEK.␣10 cells, prior stimulation with 1 M cicaprost showed no reduction in U46619-mediated mobilization of [Ca 2ϩ ] i (Fig. 2F). To determine whether the difference in IPmediated desensitization of TP␣ and TP␤ could be accounted for by an inability of HEK.␤3 cells to support elevation in cAMP in response to cicaprost, the presence of functional endogenous IP receptors in HEK.␤3 cells was confirmed by analyzing cicaprost-mediated increases in cAMP formation (Fig. 2G). The observed elevation of cAMP levels was not significantly (p Ͼ 0.1) different between HEK.␤3 and HEK.␣10 cells (Fig. 2, C  and G). The higher levels of cAMP elevation observed using 10 M cicaprost were still insufficient to reduce subsequent U46619-induced [Ca 2ϩ ] i mobilization in HEK.␤3 cells (data not shown).
Similar to the results with cicaprost, prior exposure of HEK.␣10 cells with iloprost (1 M) reduced subsequent U46619-mediated mobilization of [Ca 2ϩ ] i by TP␣ to 58.4 Ϯ 4.1% (Fig. 3A), whereas mobilization of [Ca 2ϩ ] i by TP␤ was unaffected by iloprost (Fig. 3B). There was no significant difference in iloprost-or cicaprost-mediated desensitization of TP␣ (p ϭ 0.168). Moreover, whereas HEK 293 cells do contain endogenous IP (Table I), which is coupled to activation of adenylyl cyclase (Fig. 2, C and G), it is formally possible that the levels of endogenous IP expressed are not sufficiently high to mediate efficient desensitization of the TP isoforms, which might also account for the failure to observe desensitization of the TP␤ isoform in these cells. Thus, to address this, HEK.␣10 or HEK.␤3 cells were transiently co-transfected with the cDNA encoding the mouse IP. Overexpression of IP was initially confirmed by saturation radioligand binding studies using [ 3 H]iloprost (Table I). Thereafter, the effect of overexpression of IP on iloprost-mediated (Fig. 3C) or cicaprost-mediated (data not shown) desensitization of the TP isoforms was investigated. Similar to previous data involving endogenous IPs, prior stimulation of HEK.␣10 cells with iloprost significantly reduced U46619-mediated [Ca 2ϩ ] i mobilization by TP␣ to 49.8 Ϯ 5.78% (Fig. 3C), whereas iloprost had no effect on U46619-mediated [Ca 2ϩ ] i mobilization in HEK.␤3 cells (Fig. 3D). Moreover, overexpression of IP in HEK.␣10 cells did not significantly enhance iloprost-mediated (p ϭ 0.29) or cicaprost-mediated (data not shown) desensitization of TP␣ signaling, indicating that the endogenous levels of IPs expressed in HEK 293 cells are sufficient to mediate efficient desensitization of TP␣ in those cells.
To investigate whether IP-mediated desensitization of TP signaling in HEK 293 cells may be dependent on the nature of the coupling G protein, we extended our studies to investigate the effect of co-expression of the ␣ subunit of G q as a substitute to G␣ 11 . Prestimulation of cells with iloprost ( Fig. 3E) or cicaprost (data not shown) reduced subsequent U46619-mediated [Ca 2ϩ ] i mobilization in HEK.␣10 cells to 55.1 Ϯ 2.2% (Fig. 3E), whereas U46619-mediated [Ca 2ϩ ] i mobilization in HEK.␤3 cells was unaffected (Fig. 3F). Moreover, IP-mediated desensitization of TP␣ was not significantly different in the presence of G q compared with G␣ 11 (p ϭ 0.52). Thus, we conclude that TP␣ but not TP␤ is subject to IP-mediated desensitization in HEK 293 cells and that this desensitization is independent of the nature of the IP agonist, is independent of the level of IP expression, and is independent of the coupling G protein.
Differential Effects of Cicaprost on U46619-mediated IP 3 Generation via TP␣ and TP␤ Isoforms-To further investigate the differential effects of IP activation on TP␣ and TP␤ signaling, U46619-induced IP 3 generation was measured in HEK.␣10 and HEK.␤3 cells in the presence or absence of prestimulation with cicaprost. Stimulation of HEK.␣10 and HEK.␤3 cells with U46619 (1 M) resulted in 1.5-and 1.4-fold increases in IP 3 levels, respectively, comparable with previously reported data (16). However, preincubation of HEK.␣10 cells with cicaprost (1 M) significantly (p ϭ 0.024) reduced U46619-mediated IP 3 generation by TP␣ (Fig. 4A). However, in contrast to HEK.␣10 cells, preincubation of HEK.␤3 cells with cicaprost (1 M) did not significantly (p ϭ 0.42) reduce U46619-mediated IP 3 generation by TP␤ (Fig. 4B). Stimulation of HEK 293 cells with U46619 or HEK.␣10 and HEK.␤3 cells with cicaprost alone failed to generate any increase in IP 3 , further indicating that endogenous IP receptors in HEK 293 cells do not couple to PLC.
H-89, an Inhibitor of PKA, Prevents Cicaprost-induced Inhibition of TP Signaling-Second messenger protein kinases, such as PKA and PKC, have been implicated in heterologous desensitization and in mediating cross-talk between different G protein-coupled receptor signaling systems (42)(43)(44)(45)(46)(47). Moreover, both IP and TP are subject to phosphorylation by these kinases (48,49,51). Thus, to investigate whether PKA and/or PKC is involved and provide a potential mechanism whereby IP activation cross-desensitizes TP signaling in platelets and HEK.␣10 cells, we used H-89, a specific inhibitor of PKA (54,55), and GF 109203X, a specific PKC inhibitor (56). Preincubation of platelets with 50 nM GF 109203X for 2 min prior to agonist stimulation had no effect on cicaprost induced desensitization of U46619-mediated [Ca 2ϩ ] i (Fig. 5, A and B). Similarly, in HEK.␣10 cells, whereas pretreatment with 50 nM GF 109203X for 2 min had no effect on cicaprost inhibition of U46619-induced [Ca 2ϩ ] i mobilization (Fig. 6, A and B), pretreatment with H-89 (10 M, 1 min) significantly (p Ͻ 0.001) rescued cicaprost-induced inhibition of U46619-induced [Ca 2ϩ ] i mobilization from 45 to 86% (Fig. 6C). In HEK.␤3 cells, in which the U46619-induced changes in [Ca 2ϩ ] i mobilization are impervious to prestimulation by cicaprost, addition of GF 109203X or H-89 had no effect on signaling by TP␤ (Fig. 6,  D-F). These results indicate that the observed reduction of U46619-induced changes in [Ca 2ϩ ] i by cicaprost or iloprost in HEK.␣10 cells is largely due to activation of PKA and subsequent phosphorylation, either of TP␣ directly or of some other element of its signaling pathway. Mobilization of [Ca 2ϩ ] i via TP␤ is, on the other hand, unaffected by PKA in this crossdesensitization pathway.
It has recently been reported that H-89 may act as antagonist of certain GPCRs, thereby calling into question its utility as a selective PKA inhibitor (65). Thus, to rule out the possibility that H-89 may act as an antagonist of the IP, cicaprost-   Cicaprost-induced Desensitization of TP Signaling Is Mediated by the TP C-tail-In order to establish whether the C-tail of the TP␣ contains the target regulatory site for phosphorylation by PKA, deletion mutagenesis was utilized to generate a truncated version of TP (TP ⌬328 ), which is devoid of C-tail sequences carboxyl-terminal to amino acid 328 at the point of divergence of TP␣ and TP␤. A stable HEK 293 cell line overexpressing TP ⌬328 was established, and cells were characterized by Scatchard analysis using [ 3 H]SQ29,548 as the specific radioligand. Values obtained for the affinity (K d ) and maximal binding (B max ) for TP ⌬328 (K d ϭ 6.99 Ϯ 0.88 nM; B max ϭ 1.54 Ϯ 0.28 pmol/mg; n ϭ 5) compared well to values previously reported for the wild type TP␣ and TP␤ receptors (26). It is noteworthy that TP ⌬328 exhibited identical affinity for its ligand and retained the ability to mediate specific agonist induced intracellular signaling, albeit at somewhat reduced levels relative to those of the wild type TPs. Intracellular signaling by HEK.TP ⌬328 cells transiently co-transfected with G␣ 11 was investigated by analyzing [Ca 2ϩ ] i mobilization and IP 3 generation in response to the TXA 2 mimetic U46619 and the effect of cicaprost on TP signaling was assessed. Stimulation of HEK.TP ⌬328 cells with U46619 (1 M) led to mobilization of [Ca 2ϩ ] i (Fig. 7A), whereas stimulation with cicaprost (1 M) did not (Fig. 7B). Unlike platelets or HEK.␣10 cells, prestimulation with cicaprost did not reduce subsequent U46619-induced [Ca 2ϩ ] i mobilization (Fig. 7B). Moreover, these effects were independent of the agonist used or the level of IP expression (Fig. 7, C and D; Table I). Similarly, U46619 stimulation of HEK.TP ⌬328 cells generated increases in IP 3 levels, which were not significantly reduced by prior stimulation with cicaprost (data not shown). These data confirm that the increased sensitivity to cicaprost or iloprost observed for TP␣, as opposed to TP␤ and TP ⌬328 , is due to unique elements in the C-tail of TP␣, most likely, given that pretreatment with H-89 alleviated cicaprost induced inhibition of TP␣ signaling, at an important PKA-sensitive phosphorylation site(s).
TP␣ S329A Is Not Subject to IP-mediated Desensitization-Computational analysis of the amino acid sequence of the Ctail of TP␣ for putative protein kinase phosphorylation sites using the PhosphoBase program for sequence analysis (66) identified the presence of a unique consensus PKA phosphorylation site within the sequence RPRSLSL, where Ser 329 was identified as the target residue for phosphorylation. Thus, to investigate whether this consensus PKA phosphorylation site may represent a target site for IP-mediated desensitization of TP␣, the critical Ser 329 was mutated Ala 329 to generate the variant TP␣ S329A . A stable HEK 293 cell line overexpressing a HA epitope-tagged TP␣ S329A was established, and cells were characterized by Scatchard analysis using [ 3 H]SQ29,549 (K d ϭ 9.63 Ϯ 0.94 nM; B max ϭ 6.01 Ϯ 0.16 pmol/mg; n ϭ 3). Stimulation of HEK.TP␣ S329A cells with U46619 (1 M) led to efficient mobilization of [Ca 2ϩ ] i , whereas stimulation of cells with iloprost did not (Fig. 8, A and B). However, initial stimulation of cells with iloprost did not reduce subsequent U46619-induced [Ca 2ϩ ] i mobilization. Moreover, these effects were independent of the agonist used (cicaprost or iloprost), the coupling G protein (G␣ q or G␣ 11 ), or the level of IP expression ( Fig. 8; Table I) IP-mediated Desensitization Involves Phosphorylation of TP␣ at Ser 329 -To further investigate whether the TP isoforms or TP␣ S329A are subject to IP-mediated phosphorylation, additional stable cell lines overexpressing HA epitope-tagged variants of TP␣ and TP␤ were established. HA:TP␣ and HA:TP␤ receptors were characterized by radioligand binding studies, and each cell line exhibited similar levels of [ 3 H]SQ29,548 binding (approximately 3 pmol of [ 3 H]SQ29,548 bound/mg of cell protein); moreover, consistent with previous reports (50), the presence of the HA epitope tag did not affect the ligand binding characteristics or the cell signaling properties of the TPs (data not shown). Initially, the specificity of the anti-HA antibodies to immunoprecipitate TPs from the respective cell lines, but not from the parental HEK 293 cells, was confirmed (Fig. 9D). The presence of a discrete band of approximately 39 kDa and a broad diffuse band of 46 -66 kDa was evident in TP␣ and TP␣ S329A immunoprecipitates; the narrow and broad bands may represent the nonglycosylated and glycosylated forms, respectively, of TP␣ and TP␣ S329A . Similarly, bands of 46 kDa and 50 -66 kDa, possibly representing the nonglycosylated and glycosylated forms of TP␤, respectively, were immunoprecipitated from cells overexpressing TP␤. On the other hand, no immunoreactive bands were evident in immunoprecipitates prepared from nontransfected HEK 293 cells (Fig.  9D).
Stimulation of cells with U46619 (1 M, 10 min) led to 5-7fold increases in the phosphorylation of TP␣, TP␤, and TP␣ S329A , confirming that each of these receptors is subject to homologous desensitization (Fig. 9, A-C). Stimulation of cells with iloprost (1 M, 10 min) led to a 5-fold increase in the phosphorylation of TP␣ (Fig. 9A) but no increase in the phosphorylation of TP␤ or TP␣ S329A (Fig. 9, B and C) over basal stimulation. Moreover, H-89 blocked iloprost-mediated phosphorylation of TP␣ (Fig. 9A). Similar findings were observed when cicaprost was used as the stimulating ligand. Thus, taken together, these studies confirm that TP␣, but not TP␤, is subject to IP-mediated desensitization at a PKA-sensitive site within its unique C-tail, that this desensitization involves direct IP-induced PKA phosphorylation of TP␣, and that Ser 329 is the target phosphoamino acid.
Intermolecular cross-talk has been extensively demonstrated between different GPCRs and their intracellular signaling pathways (42)(43)(44)(45)(46)(47) and between GPCRs and members of the tyrosine kinase receptor family (42,43). For example, in A7r5 vascular smooth muscle cells transfected with TP␣, stimulation with the TXA 2 mimetic I-BOP led to activation of the mitogen-activated protein kinase cascade with concomitant ty- rosine phosphorylation not only of phosphoinositide 3-kinase but also of TP␣ itself (70). Such cross-talk has been widely documented to occur between the anti-aggregatory IP/adenylyl cyclase system and the pro-aggregatory TP/phospholipase C system within platelets and vascular smooth muscle (51,34,71). Similarly, in addition to prostacyclin, the vascular endothelium secretes nitric oxide (NO), the most important known endogenous vasodilator (72,73), which further protects the vessel wall by inhibiting platelet aggregation (74 -77), granule secretion (78), adhesion (79), and fibrinogen binding to its platelet GpIIbIIIa receptor (80). A recent study (81) identified the human platelet TP(s) as a substrate for cGMP-dependent protein kinase G and proposed that direct TP phosphorylation by PKG may provide a mechanism for the inhibitory effects of NO on TP responses. Moreover, both TP␣ and TP␤ are substrates for PKG phosphorylation in vitro, suggesting that both TPs may be potential targets for NO-mediated desensitization (81).
In this study, we have investigated the counter-regulation of TP responses by IP signaling and sought to establish whether the TPs themselves may be direct targets in this cross-talk. More specifically, using stably transfected HEK 293 cells, we investigated whether cross-talk resulting from agonist stimulation of IP might have differential impacts on signaling by the individual TP␣ and TP␤ isoforms. Consistent with previous studies involving other IP agonists (64), stimulation of platelets with the selective IP agonist cicaprost completely desensitized U46619-TP-mediated platelet aggregation and activation of PLC, as assessed by measurement of intracellular Ca 2ϩ mobilization. Similarly, prestimulation of endogenous IPs in HEK.␣10 cells, stably transfected with the TP␣ isoform, almost completely desensitized subsequent TP-mediated [Ca 2ϩ ] i mobilization and IP 3 generation in response to U46619 stimulation. The EC 50 for cicaprost to desensitize the TP␣ receptor was estimated at 4.8 ϫ 10 Ϫ7 Ϯ 1.1 ϫ 10 Ϫ7 M. Failure to completely desensitize TP␣ in HEK.␣10 cells is possibly due to the relatively high TP receptor density in those cells (82). On the other hand, signaling by TP␤ was unaffected by prior cicaprost stimulation of endogenous IPs in HEK.␤3 cells, indicating that the TP␣, but not TP␤, is a target for IP-mediated cross-desensitization of TP responses. Moreover, these effects were independent of the IP agonist used, the coupling G protein, or the level of IP expression. In investigating the kinetics of TP␣ desensitization, prestimulation of cells with cicaprost from times ranging from 30 s to 5 min prior to U46619 stimulation resulted in no significant, time-dependent difference in IP-mediated desensitization of TP␣ signaling over that period of time (n ϭ 12, p ϭ 0.662). As the IP itself is widely reported to undergo agonist-mediated internalization, leading to loss of surface receptors (64), longer preincubations with IP agonists make results difficult to interpret.
We have previously reported that in order to support efficient mobilization of [Ca 2ϩ ] i in HEK cells overexpressing TP, it was necessary to overexpress the ␣ subunits of a member of the G q family such as G␣ q or G␣ 11 (26). In the absence of the co-transfected G protein, the levels of [Ca 2ϩ ] i mobilization are low and may reflect the endogenous pool of G␣ subunits, which may not be sufficient to support signaling due to relative abundance of overexpressed TP receptors. Our findings for a requirement of exogenous G protein to mediate efficient receptor signaling are not unique. In the case of the ␣ 2 -adrenoreceptor, stimulation of PLC activity in HEK 293 cells was completely dependent on co-expression of G␣ q (83). Whereas platelets and other hematopoietic cells are reported to express G␣ q and G␣ 16 , these cells are reported not to express significant levels of G␣ 11 (84 -86). To investigate whether IP-mediated desensitization of TP signaling in HEK 293 cells may be dependent on the nature of the coupling G protein, we extended our studies to investigate the effect of co-expression of the ␣ subunit of G q as opposed to G␣ 11 . Substitution of G␣ 11 with G␣ q also supported IPmediated desensitization of TP␣ but not TP␤ in response to cicaprost or iloprost.
Preincubation of platelets or HEK.␣10 cells with the PKA inhibitor H-89 almost completely blocked cicaprost-mediated desensitization of TP signaling. On the other hand, H-89 had no effect on TP␤ signaling, and the PKC inhibitor GF 109203X had no appreciable effect on IP desensitization of TP signaling in platelets or on TP␣ or TP␤ signaling in HEK 293 cells. These data indicated that TP␣, but not TP␤, is sensitive to IP desensitization mediated by the second messenger kinase PKA. As the ␣ and ␤ isoforms of TP differ exclusively in their carboxylterminal cytoplasmic tail sequences (11,12), it is possible that TP␣ is subject to IP-stimulated PKA phosphorylation at a site distal to the point of divergence of these TP receptor isoforms. Thus, to investigate this possibility, a stable cell line overexpressing a truncated variant of TP (TP ⌬328 ) was established. Whereas deletion of the divergent carboxyl-terminal amino acids of the TPs had no effect on ligand binding, intracellular signaling by the deletion mutant TP ⌬328 was significantly reduced relative to that of the wild type TP␣/TP␤ receptors, thus supporting the notion that the C-tail of TP may also play a role in G protein coupling (16,32). Consistent with these findings, Spurney and Coffman (68) reported that deletion of the carboxyl-terminal 22 amino acids of the single mouse TP also yielded a receptor with identical ligand binding properties to the wild type receptor but with diminished agonist-mediated IP 3 generation. Unlike that exhibited by the platelet TP(s) or the TP␣ isoform, U46619-mediated signaling by TP ⌬328 was insensitive to prestimulation of IPs (endogenous or overexpressed) with cicaprost or iloprost. Taken together, these data imply that IP-induced desensitization of TP responses is mediated, at least in part, via a PKA target site found in the TP␣ but not in the TP␤ isoform. TP ⌬328 differs from the wild type TP␣ in that it is devoid of the carboxyl-terminal 15 amino acid residues of the latter receptor. We have previously established that the C-tail of TP␣ is phosphorylated in vitro by PKA (48). Computational analysis of the C-tail sequences of TP␣ for putative protein kinase phosphorylation sites (66) identified the presence of a unique consensus PKA phosphorylation site within the sequence RPRSLSL, where Ser 329 was predicted to represent the target residue for phosphorylation. Thus, to investigate whether this consensus PKA site may represent the target site for IP-mediated desensitization of TP␣, the critical Ser 329 was mutated to Ala 329 generating the variant TP␣ S329A . Whereas stable cell lines overexpressing TP␣ S329A exhibited identical ligand binding characteristics and agonist-mediated intracellular signaling to that of the wild type TP␣, U46619mediated signaling by TP␣ S329A was insensitive to prestimulation of IPs with cicaprost or iloprost. Moreover, these effects were independent of the agonist used (cicaprost or iloprost), the level of IP expression (endogenous or overexpressed) or the coupling G protein, G␣ q or G␣ 11 . Finally, to establish whether the TP(s) may be direct a target for IP-mediated phosphorylation, stable cell lines overexpressing HA epitope-tagged forms of TP␣, TP␤, or TP␣ S329A were used in whole cell phosphorylation assays. Whereas each of the TP receptors underwent U46619-mediated phosphorylation, stimulation of cells with iloprost specifically resulted in agonist-dependent phosphorylation of TP␣ but not TP␤ or TP␣ S329A . Moreover, the PKA inhibitor H-89 blocked iloprost-mediated phosphorylation of TP␣. Taken together, these studies confirm that TP␣, but not TP␤, is subject to IP-mediated desensitization at a PKA-sensitive site within its unique C-tail and that this desensitization involves direct IP induced PKA phosphorylation of TP␣. It is also evident that Ser 329 , the very first divergent residue between TP␣ and TP␤, is the target for IP-mediated phosphorylation of TP␣.
In view of the differential desensitization of the TP␣ by IP-selective agonists coupled to the lack of desensitization of TP␤, TP ⌬328 , and TP␣ S329A receptors and our demonstration of IP-mediated PKA phosphorylation of TP␣ and not TP␤, TP ⌬328 , and TP␣ S329A receptors, the mechanism of desensitization involves direct TP␣ phosphorylation. Moreover, in view of the selective involvement of TP␣ and not the other TP␤, TP ⌬328 , and TP␣ S329A receptors, it is unlikely that other components of the signaling system, such as the G␣ subunits, are involved. In terms of the structure-function relationships of GPCRs, many of the target desensitization/phosphorylation sites on the GPCR are not actually located within the G protein interacting domains (mainly believed to be intracellular loop 3 with roles also evoked for intracellular loops 2 and 1, and in some cases, the C-tail) but are mainly located within the C-tail regions of GPCRs. Our findings on TP␣ phosphorylation is consistent with this. Phosphorylation of receptors by the second messenger kinases PKA/PKC are mainly associated with classic feedback mechanisms or in mediating cross-talk whereby phosphorylation is believed to alter the overall conformation of the receptor thereby interfering with receptor: G protein interactions. Other mechanisms are believed to occur for other kinases, such as the G protein-coupled receptor kinases whereby GPCR phosphorylation initiates arrestin binding and may, in turn, lead to GPCR sequestration through dynamin/clathrincoated pit internalization mechanisms (40).
In summary, our results demonstrate that TP␣, but not TP␤, is subject to cross-desensitization by IP in HEK 293 cells and suggest that should this occur in platelets or vascular smooth muscle, TP␣ may be the isoform physiologically relevant to the maintenance of vascular hemostasis. Consistent with these , or vehicle alone, as indicated. HA epitope-tagged receptors were immunoprecipitated using the anti-HA antibody 101R, as described under "Experimental Procedures." Immunoprecipitates were subjected to SDS-polyacrylamide gel electrophoresis and electroblotted onto polyvinylidene difluoride membranes, which were then exposed to Xomat XAR-5 film (Kodak) for 15 days. Thereafter, blots were subject to phosphorimage analysis, and the intensities of phosphorylation relative to basal phosphorylation were determined and expressed in arbitrary units, as follows. findings, based on observations that TP isoform specific antibodies permitted detection of TP␣, but not TP␤ in human platelets, Habib et al. (15) have hypothesized that TP␣ may be the predominant isoform in platelets, despite the presence of mRNA for both isoforms in platelets, which suggests that both isoforms are actually present albeit at different levels (16). It has been recently demonstrated that U46619 induced platelet aggregation required concomitant stimulation of both G q -coupled TP receptors (leading to Ca 2ϩ mobilization and platelet shape change) and G i -coupled P2T AC and ␣ 2A receptors (activated upon secretion of granule contents), leading to platelet aggregation (87). Thus, at least two separate signal transduction cascades are required, one TP-dependent and one TPindependent, to result in platelet aggregation following activation with U46619. Whereas our studies have pinpointed an essential requirement of TP␣ and an apparent redundancy for TP␤ in TP:IP-mediated cross-talk with implications for vascular hemostasis, and other studies have identified differences in the expression, signaling, and desensitization between the TP isoforms (12,14,16,32,50,66), identification of a physiologic requirement for the second ␤ isoform of TP found in humans but not in other species, such as mouse or rat (88,89), requires further investigation. It is also particularly noteworthy that amino acid Ser 329 , the target residue for IP-mediated PKA phosphorylation in TP␣, is absolutely conserved in both monkey TP (90) and bovine TP (91) but not in the mouse or rat TPs; thus, it is possible that the former receptors are also subject to a similar mechanism of IP-mediated desensitization to that of the human TP␣ isoform.