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

In humans, thromboxane A2 signals through two thromboxane A2 receptor (TP) isoforms termed TPα and TPβ. Signaling by TPα, but not TPβ, is subject to prostacyclin-induced desensitization mediated by direct protein kinase (PK) A phosphorylation where Ser329 represents the phosphotarget (Walsh, M. T., Foley, J. F., and Kinsella, B. T. (2000) J. Biol. Chem. 275, 20412-20423). In the current study, the effect of the vasodilator nitric oxide (NO) on intracellular signaling by the TP isoforms was investigated. The NO donor 3-morpholinosydnonimine, HCl (SIN-1) and 8-bromo-guanosine 3′,5′-cyclic monophosphate (8-Br-cGMP) functionally desensitized U46619-mediated calcium mobilization and inositol 1,4,5-trisphosphate generation by TPα whereas signaling by TPβ was unaffected by either agent. NO-mediated desensitization of TPα signaling occurred through a PKG-dependent, PKA- and PKC-independent mechanism. TPα, but not TPβ, was efficiently phosphorylated by PKG in vitro and underwent NO/PKG-mediated phosphorylation in whole cells. Deletion/site-directed mutagenesis and metabolic labeling studies identified Ser331 as the target residue of NO-induced PKG phosphorylation of TPα. Although TPαS331A was insensitive to NO/PKG-desensitization, similar to wild type TPα its signaling was fully desensitized by the prostacyclin receptor agonist cicaprost occurring through a PKA-dependent mechanism. Conversely, signaling by TPαS329A was insensitive to cicaprost stimulation whereas it was fully desensitized by NO/PKG signaling. In conclusion, TPα undergoes both NO- and prostacyclin-mediated desensitization that occur through entirely independent mechanisms involving direct PKG phosphorylation of Ser331, in response to NO, and PKA phosphorylation of Ser329, in response to prostacyclin, within the unique carboxyl-terminal tail domain of TPα. On the other hand, signaling by TPβ is unaffected by either NO or prostacyclin.

The local control of hemostasis is regulated by a variety of vasoconstrictory and vasodilatory autocoids that act to modulate platelet, vascular endothelium, and vascular smooth muscle (VSM) 1 function (1). The prostanoids thromboxane (TX)A 2 and prostacyclin (prostaglandin (PG)I 2 ) play key, yet opposing, roles in the maintenance of vascular hemostasis (2); TXA 2 , synthesized primarily by platelets and activated macrophages, mediates platelet aggregation and acts as a potent vasoconstrictor whereas prostacyclin, mainly synthesized by the vascular endothelium, inhibits platelet activation and aggregation and induces vasodilation (2). TXA 2 and prostacyclin signal through their respective G protein-coupled receptors (GPCRs) termed TP and IP and are primarily coupled to Gq-dependent activation of phospholipase (PL)C␤ and to Gs-dependent activation of adenylyl cyclase, respectively (2). The main inhibitory actions of IP signaling in platelets and VSM are believed to be largely mediated through its regulation of cAMP-dependent protein kinase A (PKA) activity; established PKA substrates within platelets include the low molecular weight G protein rap1B, caldesmon, myosin light chain kinase, PLC, thrombolamban, G␣ 13 , and actin-binding protein (3)(4)(5).
In humans, TXA 2 signals not through one but through two TP receptors or isoforms termed TP␣ and TP␤ that arise by differential splicing. TP␣ and TP␤ are identical for their aminoterminal 328 amino acids but differ exclusively in their carboxyl-terminal (C)-tail domains (6 -8). The physiological significance for two TP receptors in humans, but not in other species thus far characterized, remains to be determined, but it is now evident that they exhibit distinct patterns of mRNA and protein expression (9); they are under the differential control of two distinct promoters within the single TP gene (10) and exhibit functional differences in terms of secondary effector signaling and receptor desensitization (11,12). Although both TP␣ and TP␤ display identical ligand binding and activation of PLC␤ (7,(13)(14)(15), they oppositely regulate adenylyl cyclase activity (16), and TP␣, but not TP␤, mediates activation of the novel G protein/tissue transglutaminase Gh (17).
Receptor desensitization, defined as the attenuation of receptor signaling irrespective of continued agonist stimulation, is a common feature among GPCRs and provides mechanisms for both agonist-induced homologous desensitization and intramolecular cross-talk/heterologous desensitization (18). Receptor desensitization may involve phosphorylation of the receptor by specific Ser/Thr kinases and/or sequestration or internalization of the receptor to intracellular stores where they are unavailable for interaction with their coupling G protein (19). Although both TPs undergo agonist-induced phospho-rylation (13,20), TP␤, but not TP␣, is subject to internalization and down-regulation following prolonged exposure to the TXA 2 mimetic U46619 (21,22). In studies investigating cross-talk between TXA 2 and other prostanoids, we have established that TP␣, but not TP␤, is subject to prostacyclin-and PGD 2 -mediated desensitization/inhibition of signaling involving direct PKA phosphorylation of TP␣ itself within its unique C-tail domain (11,23). Hence, these studies identified TP␣ as a novel target of prostacyclin-induced PKA phosphorylation and inhibition of platelet activation/vasodilation and imply an essential role for TP␣ in prostanoid-regulated vascular hemostasis but point to a redundant, or an as yet unidentified, role for TP␤ in this essential physiologic process (11,23).
In addition to prostacyclin, nitric oxide (NO), also referred to as endothelial-derived relaxing factor or EDRF (24), acts as a second endogenous vasodilator and further protects the blood vessel wall by inhibiting platelet aggregation, secretion, adhesion, and fibrinogen binding to its receptor glycoprotein (Gp)IIb/IIIa (24 -29). Endothelium-derived NO, synthesized by the endothelial isoform of NO synthase (eNOS/NOS III) in response to diverse agonists, such as bradykinin and acetylcholine (24, 30 -32) and various hemodynamic stimuli, especially fluid shear stress (33), activates its target receptor soluble guanylyl cyclase (sGC) leading to increases in intracellular cGMP concentrations (34,35). One of the key molecular targets of cGMP are the cGMP-dependent PKG 1␣ and 1␤ isozymes that are abundantly expressed in platelets and various types of smooth muscle (36,37). Although the precise molecular mechanism(s) of NO-mediated vasodilation and inhibition of platelet activation is currently unknown, like prostacyclin, its main inhibitory actions are mediated through its second messengerregulated kinase PKG (4). Key molecular targets of NO-regulated PKG thus far identified include the IP 3 receptor (38), thereby inhibiting PLC activation and IP 3 -evoked Ca 2ϩ release from intracellular stores (39), IP 3 receptor-associated cGMP kinase substrate or IRAG (40,41), myosin binding substrate (42), and the vasodilator-stimulated phosphoprotein or VASP (43). In a recent study (44), it was reported that TP receptor(s) expressed in the platelet-like human erythroleukemic cell line showed increased incorporation of [ 32 P]orthophosphate in response to cGMP implying that the TP(s) themselves may be direct targets of NO-regulated vascular hemostasis (44).
Thus, in view of our findings that the TP␣, but not the TP␤, isoform of the human TP is subject to direct PKA phosphorylation and desensitization of signaling in response to the inhibitory prostanoid prostacyclin (23), in the current study we sought to investigate NO-mediated regulation of TP signaling and to establish whether TP␣ or TP␤ or both are subject to NO-induced desensitization in whole cells. Hence, this study was designed to investigate the intermolecular cross-talk mediated by the NO donor 3-morpholinosydnonimine (SIN-1) or the non-hydrolyzable analogue of cGMP, 8-Br-cGMP, on signaling by the individual TP␣ and TP␤ isoforms stably overexpressed in human embryonic kidney (HEK) 293 cells, and to identify the molecular mechanism of that desensitization, should it occur. Our results demonstrate that signaling by TP␣ is subject to desensitization in response to both SIN-1 and 8-Br-cGMP and that these effects are mediated by direct PKG phosphorylation of TP␣ whereby Ser 331 was identified as the specific phosphotarget residue. On the other hand, signaling by TP␤ remains unaffected by either SIN-1 and 8-Br-cGMP, and TP␤ was not a substrate for PKG phosphorylation either in vitro or in whole cells. Furthermore, we established that prostacyclin-and NO-regulated desensitization of TP␣ signaling are mediated through distinctly independent mechanisms involving direct PKA phosphorylation of Ser 329 , in response to prostacyclin, and PKG phosphorylation of Ser 331 , in response to NO, respectively, within the unique C-tail domain of TP␣.

Materials
, cAMP-dependent protein kinase A, and peak III from rabbit muscle (350 units/mg protein; 1 unit/l; P3891) were obtained from Sigma. Amylose resin, anti-maltose-binding protein polyclonal antiserum, and pMal-C vector were obtained from New England Biolabs. The QuikChange TM site-directed mutagenesis kit was purchased from Stratagene. Cicaprost was a gift from Schering AG (Berlin, Germany). All oligonucleotides were synthesized by Genosys Biotechnologies.
Conversion of S329A of TP␣ to generate TP␣ S329A has been described previously (23). Conversion of S331A of TP␣ to generate TP␣ S331A was performed using the QuikChange TM site-directed mutagenesis kit employing pHM6:TP␣ as template and mutator oligonucleotides 5Ј-G CCC AGG TCG CTG GCC CTC CAG CCC C-3Ј (sense primer) and 5Ј-G GGG CTG GAG GGC CAG CGA CCT GGG-3Ј (antisense primer), where the sequence corresponding/complementary to mutator Ser (TCC)-Ala (GCC) codon are in italics, to generate the plasmid pHM6:TP␣ S331A . Similarly, conversion of S329A,S331A to generate TP␣ S329A,S331A was performed using the QuikChange TM site-directed mutagenesis kit employing pHM6:TP␣ S331A as template and mutator oligonucleotides 5Ј-CC CGG CCC AGG GCG CTG GCC CTC C (sense primer) and 5Ј-G GAG GGC CAG CGC CCT GGG CCG GG (antisense primer), where the sequence corresponding/complementary to mutator Ser (TCG)-Ala (GCG) codon is in italics, to generate the plasmid pHM6:TP␣ S329A,S331A . Deletion of the amino acids carboxyl to Leu 336 of TP␣ was achieved by conversion of Thr 337 codon to a Stop codon (Thr 337 , ACG to Stop 337 , TAA). Site-directed mutagenesis was performed by PCR mutagenesis using pCMV:TP␣ as template and oligonucleotides 5Ј-GAGAAGCTTG ATG TGG CCC AAC GGC AGT TCC-3Ј (sense primer; nucleotides ϩ1 to ϩ21 of TP␣ sequence are underlined) and 5Ј-CT CTA AGC TTA GAG CTG GGG CTG GAG GGA-3Ј (antisense primer; sequences complimentary to nucleotides ϩ993 to ϩ1008 of TP␣ sequence are underlined, and the mutator in-frame stop codon is in boldface italics). PCR amplifications were performed using Expand High Fidelity ® Taq DNA polymerase, and the resulting PCR-amplified cDNA was subcloned into the HindIII-EcoRI site of pHM6 to generate pHM:TP␣ ⌬336 . The plasmids pHM6:TP␣ S331A and pHM6:TP␣ S329A,S331A and mutations therein were verified by automated double-stranded DNA sequencing. The plasmid pCMV:G␣q, coding for G␣ q , has been described previously (45,47).
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 saturation radioligand binding isotherms, or in the presence of 20 nM [ 3 H]SQ29,548 essentially as described previously (45). Protein determinations were carried out using the Bradford assay (49).
Calcium Measurements-Measurements of intracellular calcium ([Ca 2ϩ ] i ) mobilization in transfected HEK 293 cells was carried out in FURA2/AM preloaded cells essentially as described previously (45). Cells were stimulated with 1 M U46619, 1 M cicaprost, 10 M 8-Br-cGMP, or 10 M SIN-1 unless otherwise specified. The PKG inhibitor KT5823 (50 nM), the PKA inhibitor H-89 (10 M), or the PKC inhibitor GF 109203X (50 nM) were added at times and concentrations as specified in the figure legends. In all cases, the drug (agonists/kinase inhibitors in ethanol or Me 2 SO) was diluted 1:1000 in the vehicle HBSSHB (modified Ca 2ϩ /Mg 2ϩ -free Hank's buffered salt solution, containing 10 mM HEPES, pH 7.67, 0.1% bovine serum albumin), 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 (50), assuming a K d of 225 nM Ca 2ϩ for FURA2/AM. The results presented in the figures are representative data from at least three/four independent experiments, and representative profiles are plotted as changes in intracellular Ca 2ϩ mobilized (⌬[Ca 2ϩ ] i (nM)) as a function of time (s) upon ligand stimulation.
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 (51). Briefly, cells were harvested by scraping and washed twice in ice cold phosphate-buffered saline, and 2 ϫ 10 6 cells were resuspended in 200 l of HEPES-buffered saline (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; HBS) supplemented with 10 mM LiCl. Cells were pre-equilibrated in this buffer at 37°C for 10 min and stimulated for 1 min at 37°C in the presence of U46619 (1 M) or 8-Br-cGMP (10 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 radio-competition assay essentially as described (51). Protein determinations were carried out using the Bradford assay (49). Levels of IP 3 produced by ligand or vehicletreated cells were calculated as pmol IP 3 /mg cell protein Ϯ S.E. and are expressed as -fold increases in IP 3 levels generated in agonist-treated cells relative to basal (vehicle)-treated cells. Data presented are the mean data from four independent experiments, each performed in duplicate.
In Vitro Phosphorylation of Purified Recombinant Fusion Proteins-The maltose-binding protein (MBP) and recombinant MBP:TP␣ 220 -343 (MBPTXR-C (52); fusion protein containing the carboxyl-terminal 220 -343-amino acid residues of TP␣ fused to the MBP protein were expressed in Escherichia coli and purified as described previously (52). Similarly, the cDNA sequence encoding the carboxyl-terminal 220 -407-amino acid residues of TP␤ were cloned in-frame into pMal-C (New England Biolabs) to generate the plasmid pMal-C:TP␤ 220 -407 . The plasmid pMal-C:TP␤ 220 -407 was verified by automated double-stranded DNA sequencing. Thereafter, the recombinant MBP:TP␤ 220 -407 fusion protein was expressed in E. coli JA221 and purified by affinity chromatography on amylose resin, as described previously (52).
Thereafter purified MPB, MBP:TP␣ 220 -343 , and MBP:TP␤ 220 -407 fusion proteins were used as substrates in cAMP-dependant PKA and cGMP-dependant PKG in vitro phosphorylation assays. PKA in vitro phosphorylation assays employing 2 g of purified protein substrates and rabbit muscle PKA (350 units/mg protein; 1 unit/l; 1 l) were performed as described previously (52). Alternatively, purified recombinant protein substrates (2 g/assay) were phosphorylated in vitro with PKG (specific activity, 10,200 units/g; 20,000 units/l; 1 l) at 37°C for 30 min in 50-l reactions containing 50 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , 0.1 mM cGMP, and 5 Ci of [␥-32 P]ATP (6000 Ci/mmol, 10 mCi/ml). In vitro phosphorylation reactions were terminated by acetone precipitation; protein precipitates were collected by centrifugation, resuspended and solubilized in Laemmli sample buffer, and analyzed via SDS-PAGE, on 10% polyacrylamide gels, followed by Western transfer onto PVDF membrane. Membranes were subject to Xomat XAR film autoradiographic exposure for 24 -48 h to detect phosphorylated proteins. Thereafter, membranes were successively screened using anti-MBP antibody, anti-TP␣ antibody, and TP␤ antibody (53) as primary antibodies followed by screening with horseradish peroxidase-conjugated goat anti-rabbit IgG, as secondary antibody, and chemiluminescence detection, as described by the supplier (Roche Applied Science).
Measurement of Agonist-mediated TP Phosphorylation-Whole cell phosphorylation assays were performed essentially as described previously (23) 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 media (1.5 ml/60-mm dish) containing 100 Ci/ml [ 32 P]orthophosphate (8,000 -9,000 Ci/mmol) at 37°C, 5% CO 2 . Where appropriate, kinase inhibitors (KT5823, 50 nM; H-89, 10 M; GF 109203X, 50 nM) or vehicle were added during the labeling period. Thereafter, specific ligands, or vehicle, were added for 10 min. Reactions were terminated by transferring the dishes to ice and aspirating the labeling media. 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), 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 plus 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 (18-, 21-, 23-, and 26-gauge), and soluble cell lysates were harvested by centrifugation for 15 min at 13,000 ϫ g at room temperature. HAepitope TP receptors were immunoprecipitated using the anti-HA antibody (1:300 dilution; 101R) 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 Laemmli sample buffer (10% ␤ mer- 50 mM Tris-HCl, pH 6.8; 40 l). Samples were loaded without boiling onto 10% polyacrylamide gels and analyzed by SDS-PAGE and thereafter were electroblotted onto PVDF membranes, essentially as described previously (47). 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 are screened by immunoblot analysis using the anti-HA 3F10 horseradish peroxidase conjugate; immunoreactive proteins were visualized using the chemiluminescence detection system, as described by the manufacturer (Roche Applied Science).
Data Analyses-Radioligand binding data was 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 unpaired 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.

Effect of the NO/cGMP Pathway on U46619-mediated Signaling in HEK 293
Cells-In the current study we sought to establish whether TP␣ or TP␤ or both are subject to NOinduced desensitization by investigating the effect of the NO donor SIN-1 or the cGMP analogue 8-Br-cGMP on U46619mediated signaling ([Ca 2ϩ ] i mobilization and IP 3 generation) by the individual TP␣ (HEK.TP␣10 cells) and TP␤ (HEK.TP␤3 cells) isoforms stably over-expressed in HEK 293 cells.
Consistent with previous studies (15) 1G, and compare Fig. 1 To further investigate the differential effects of the NO/ cGMP pathway on TP␣ and TP␤ signaling, the effect of 8-Br-cGMP and SIN-1 on U46619-mediated IP 3 generation was also examined. Stimulation of HEK.TP␣10 and HEK.TP␤3 cells with U46619 (1 M) yielded 5-to 7-fold increases in IP 3 levels ( Fig. 1, D and H, respectively). Although 8-Br-cGMP did not yield a significant increase in IP 3 generation in either cell type, pre-incubation of HEK.TP␣10 cells with 8-Br-cGMP significantly reduced U46619-mediated IP 3 generation by TP␣ (Fig.  1D, p Ͻ 0.005). In contrast, 8-Br-cGMP had no significant effect on U46619-mediated IP 3 generation in HEK.TP␤3 cells (Fig.  1H, p ϭ 0.41). Similarly, SIN-1 reduced U46619-mediated IP 3 generation by TP␣ but had no significant effect on IP 3 generation by TP␤ (data not shown).
To investigate the effect of the second messenger PKs on NO/cGMP-mediated cross-desensitization of TP signaling, the effect of KT5823, a PKG inhibitor (54), H-89, a PKA inhibitor (55,56), and GF 109203X, a PKC inhibitor (57) and HEK.TP␤3 (panel H) cells, respectively, transiently co-transfected with pCMV:G␣ q were stimulated at 37°C for 1 min with 1 M U46619 (U46619) or were pre-stimulated for 1 min with 10 M 8-Br-cGMP prior to stimulation for 1 min with 1 M U46619 (8-Br-cGMP, U46619). In each case, basal IP 3 levels were determined by exposing the cells to the vehicle HBS under identical incubation conditions. Levels of IP 3 generated by ligand or vehicle-treated cells were calculated as pmol IP 3 /mg protein Ϯ S.E. (n ϭ 4) and are expressed as mean -fold increases in IP 3 levels generated in agonist-treated cells relative to basal (vehicle)-treated cells (-fold increase in IP 3 Ϯ S.E., n ϭ 4). The asterisks (***, p Ͻ 0.005) indicate that U46619-induced IP 3 generation in HEK.TP␣10 cells was significantly lower in cells pre-stimulated with 8-Br-cGMP compared with cells stimulated with U46619 alone. restored U46619-mediated IP 3 levels to normal levels ( Fig. 2D, p ϭ 0.001), whereas neither H-89 nor GF 109203X had any effect (Fig. 2D). KT5823 (Fig. 2E), H-89, or GF 109203X (data not shown) did not affect U46619-mediated [Ca 2ϩ ] i mobilization or IP 3 generation (Fig. 2F) by TP␤ expressed in HEK.TP␤3 cells.
As TP␣ and TP␤ diverge exclusively within their C-tail domains (Fig. 3), and results presented herein show that TP␣, but not TP␤, is sensitive to NO-mediated desensitization, we then investigated the effect of SIN-1 and 8-Br-cGMP on signaling by TP ⌬328 , a variant of TP devoid of those residues unique to TP␣ and TP␤ (15). Stimulation of HEK.TP␣ ⌬328 cells, transiently co-transfected with G␣ q , showed efficient U46619-mediated [Ca 2ϩ ] i mobilization (Fig. 2, G and H). However, neither SIN-1 (Fig. 2G,  and 220 -407 spanning from the beginning of the third intracellular loop through to the end of the unique C-tail domains of TP␣ and TP␤, respectively, were used as substrates in PKG and, as controls, in PKA in vitro phosphorylation assays (Fig.  4). MBP:TP␣ was efficiently phosphorylated in vitro by both PKG (Fig. 4B, lane 2) and PKA (Fig. 4A, lane 2). In contrast, neither MBP:TP␤ (Fig. 4, A and B, lane 3) nor MBP (Fig. 4, A  and B, lane 1) were phosphorylated by either kinase. The presence of the equimolar concentrations of the MBP and MBP:TP fusion proteins in the phosphorylation assays was confirmed by Western blotting using anti-MBP (Fig. 4C) whereas the identities of the recombinant MBP:TP fusion proteins to be that of TP␣ (lane 2) and TP␤ (lane 3) were confirmed using affinity purified anti-TP␣ (Fig. 4D) and anti-TP␤ (Fig.  4E) antibodies, respectively. Thereafter, [ 32 P]orthophosphate metabolic labeling studies were performed to investigate whether TP␣ and/or TP␤ are direct targets for NO/cGMP-induced PKG phosphorylation in vivo/in whole cells. Initially, the ability of the anti-HA 101R antisera to immunoprecipitate the HA epitope-tagged TP␣ or TP␤ from their respective HEK.TP␣ and HEK.TP␤ cells stable cell lines (23), but not from the control HEK 293 cell line, was confirmed (Fig. 5D). Broad protein bands of ϳ39 -60 kDa (Fig.  5D, lane 2) and 46 -60 kDa (Fig. 5D, lane 3) representing the non-glycosylated and glycosylated forms of TP␣ and TP␤ were immunoprecipitated from HEK.TP␣ and HEK.TP␤ cells, respectively, but were not present in the immunoprecipitates from control HEK 293 cells (Fig. 5D, lane 1). Consistent with previous reports (23), stimulation of HEK.TP␣ (Fig. 5A, lanes 1  and 2) and HEK.TP␤ cells (Fig. 5B, lanes 1 and 2) with U46619 significantly increased the level of TP␣ and TP␤ phosphorylation relative to their basal phosphorylation observed in vehicletreated control cells. Incubation of cells with SIN-1 signifi-cantly increased the level of TP␣ phosphorylation (Fig. 5A, lane 3) but had no effect on TP␤ phosphorylation (Fig. 4B, lane 3). KT5823, but not H-89 or GF 109203X, significantly reduced the level of SIN-1-induced TP␣ phosphorylation (Fig. 5A, lanes  4 -6, respectively). Taken together, these data demonstrate that signaling by TP␣ is subject to desensitization in response to the NO donor SIN-1 and that this occurs through a KT5823sensitive PKG mechanism involving direct phosphorylation of TP␣ within its unique C-tail domain. On the other hand, neither SIN-1 nor 8-Br-cGMP had any effect on signaling by TP␤ or on its basal phosphorylation, and a recombinant form of TP␤ was not a substrate for phosphorylation by PKG in vitro.
Identification of the Mechanism of NO/cGMP-desensitization of TP␣ Signaling-Analysis of the amino acid sequence of the unique C-tail domain of TP␣ reveals the presence of 4 Ser/Thr residues (Ser 329 , Ser 331 , Thr 337 , Ser 340 ) that might represent putative target residues for PKG phosphorylation (Fig. 3). Thus, site-directed/deletion mutagenesis, in combination with metabolic labeling studies, were employed to investigate whether any of these putative phosphorylation sites within TP␣ represent potential target sites for NO/cGMP-mediated desensitization (Fig. 3). Initially, TP␣ ⌬336 , a truncated variant of TP␣ devoid of all C-tail residues distal to Leu 336 , including the putative phosphotargets Thr 337 and Ser 340 , was generated by deletion mutagenesis (Fig. 3), and a HEK 293 stable cell line over-expressing TP␣ ⌬336 was established ( Table I)  Phosphorylated proteins were analyzed by SDS-PAGE and electroblotted onto PVDF membranes, which were subsequently exposed to Xomat XAR-5 film (Kodak) for 24 -48 h. Thereafter, membranes were successively screened with either anti-MBP antisera (panel C) or affinity purified anti-TP␣ (panel D) or anti-TP␤ (panel E) antisera, using anti-rabbit peroxidase-conjugated antibody as the secondary antibody. Immunoreactive bands were visualized by chemiluminescence detection. The relative position of the 45-kDa molecular mass marker (kDa) is indicated to the right of panels A-E.
whereby the critical Ser 329 and Ser 331 were mutated to Ala 329 and Ala 331 either individually or collectively (Fig. 3), and the relative sensitivities of the latter receptors to NO/PKG and prostacyclin/PKA signaling was investigated. Initial saturation radioligand binding studies confirmed that the mutations per se did not affect the ligand properties of the variant TP␣ receptors stably expressed in the respective HEK.TP␣ S329A , HEK.TP␣ S331A , and HEK.TP␣ S329A,S331A cells (Table I). Thereafter, the effect of the 8-Br-cGMP (10 M) and the NO donor SIN-1 (data not shown) on U46619-mediated signaling by TP␣ S329A , TP␣ S331A , and TP␣ S329A,S331A was investigated and was compared with the effect of the prostacyclin receptor (IP) agonist cicaprost on TP signaling. Stimulation of HEK.TP␣ S329A and HEK.TP␣ S331A cells, each transiently cotransfected with G␣ q , with U46619 led to efficient [Ca 2ϩ ] i mobilization (Fig. 6, A and E, respectively). Pre-incubation of cells with 8-Br-cGMP significantly reduced U46619-mediated [Ca 2ϩ ] i mobilization in HEK.TP␣ S329A cells (see Fig. 6B, and compare Fig. 6, A  Thereafter, the effect of 8-Br-cGMP and cicaprost on U46619-mediated IP 3 generation in HEK.TP␣ S329A and HEK.TP␣ S331A cells were investigated. Stimulation of HEK.TP␣ S329A and HEK.TP␣ S331A cells with U46619 yielded 7-fold increases in IP 3 generation above basal levels (Fig. 6, D  and H, respectively). Pre-stimulation of HEK.TP␣ S329A cells with 8-Br-cGMP significantly reduced U46619-mediated IP 3 generation (Fig. 6D, p Ͻ 0.05), whereas cicaprost had no significant effect on U46619-mediated IP 3 by TP␣ S329A (p ϭ 0.13). In contrast, pre-stimulation of HEK.TP␣ S331A cells with 8-Br-cGMP had no significant effect U46619-mediated IP 3 generation (Fig. 6H, p ϭ 0.16) whereas cicaprost significantly impaired IP 3 generation by TP␣ S331A (Fig. 6H, p Ͻ  0.004).
Hence, taken together, these data indicate that TP␣ is subject to NO/8-Br-cGMP-induced desensitization of signaling mediated through its direct PKG-mediated phosphorylation whereby Ser 331 , but not Ser 329 , has been identified as the phosphotarget for that desensitization. On the other hand, consistent with our previous findings (23), TP␣ is also subject to independent cicaprost-induced desensitization that occurs through a mechanism involving PKA phosphorylation of Ser 329 , without targeting Ser 331 .

DISCUSSION
The TXA 2 and NO/cGMP signaling pathways exert opposing actions in platelets and in smooth muscle cells within the vasculature to regulate hemostasis and blood vessel tone (2,24). The physiological significance for the existence of two TP isoforms in humans is unknown, but they exhibit distinct functional differences, particularly in terms of their modes of regulation and desensitization of signaling (12). For example, TP␣, but not TP␤, is desensitized in response to signaling by the inhibitory prostanoids prostacyclin and PGD 2 and has led to the suggestion that TP␣ is the TP isoform involved in prostanoid-regulated vascular hemostasis whereas the role of TP␤, if any, in this critical physiologic process remains to be identified (11,23). Moreover, Wang et al. (44) reported that 8-Br-cGMP inhibited TXA 2 -stimulated GTPase in human platelets, that the TP(s) expressed in megakaryocytic human erythroleukemia cells showed increased phosphorylation in response to cGMP, and that recombinant glutathione S-transferase fusion proteins encoding domains of TP␣ and TP␤ were substrates for PKG phosphorylation in vitro. These latter observations imply that, in addition to their established differential regulation by prostacyclin and PGD 2 , the TP isoform(s) may also be direct targets for NO-regulated desensitization. Thus, the primary aim of the current study was to investigate whether TP␣ or TP␤ or both are subject to NO-induced desensitization in vivo/in whole cells and to establish whether they may be differentially sensitive to such regulation.
Hence, to this end, the effect of NO donor SIN-1 and 8-Br-cGMP on signaling by the individual TP␣ and TP␤ isoforms stably over-expressed in HEK 293 cells was investigated. Both SIN-1 and 8-Br-cGMP impaired U46619-mediated [Ca 2ϩ ] i mobilization and IP 3 generation by TP␣, but neither agent affected signaling by TP␤. Pre-incubation with the PKG inhibitor KT5823, but not the PKA or PKC inhibitors H-89 or GF 109203X, blocked SIN-1-and 8-Br-cGMP-induced desensitization of TP␣ signaling but had no effect on signaling by TP␤. Moreover, neither SIN-1 nor 8-Br-cGMP had any effect on U46619-mediated signaling by TP ⌬328 , the variant of TP␣/TP␤ devoid of the divergent C-tail domains of TP␣ and TP␤ (15).
To establish whether TP␣ or TP␤ are substrates for PKG, initially in vitro phosphorylation assays were carried out using MBP:TP␣ and MBP:TP␤ recombinant fusion proteins that contain amino acids 220 -343 and 220 -407, which include the third intracellular loop through to the end of the divergent C-tail domains of TP␣ (6) and TP␤ (7,8), respectively. Consistent with previous reports, TP␣ was efficiently phosphorylated in vitro by both PKG (44) and PKA (52) within its C-tail domain. On the other hand, MBP:TP␤ protein was not phosphorylated by either PKG or PKA. Although the observation that TP␤ is not phosphorylated by PKA is in keeping with earlier studies (11,23), the finding that TP␤ is not phosphorylated by PKG is actually contrary to previous reports (44). The reason for the apparent discrepancy between data generated within the current study and those reported previously (44) can be rationalized by the fact that the recombinant TP␤ protein (residues 331-369) employed by the latter investigators is unlikely to truly correspond to correct TP␤ sequences (44). Because of an initial cloning artifact, it was originally reported  that the cDNA for TP␤ encodes a protein of 369 amino acids, where residues 329 -369 were predicted to represent the unique C-tail domain of TP␤, divergent from TP␣ sequences (7); however, it was subsequently established and reported that because of an out of frame mutation in the original cloned cDNA, the TP␤ mRNA actually encodes a protein of 407 amino acids (protein accession number A56194) and not 369 residues as originally reported incorrectly (7,8). Thus, the divergent C-tail domain of TP␤ does not actually contain the 41-amino acid sequence (residues 329 -369) predicted to be present in the TP␤ sequence reported originally (7) but actually contains a unique peptide sequence of 79 residues corresponding to amino acids 329 -407 (8). Hence, it is apparent that the recombinant TP␤ protein employed by Wang et al. (44) did not actually correspond to TP␤ sequence. In the current study reported herein, the identities of the recombinant MBP:TP fusion proteins to truly express TP␣ and TP␤ sequences (6,8) were validated by immunoblot analysis using affinity purified anti-TP␣ and anti-TP␤ peptide antibodies directed to the carboxylterminal 15-amino acid residues of TP␣ (residues 329 -343) and TP␤ (residues 393-407), respectively (53).
Whole cell [ 32 P]orthophosphate metabolic labeling studies established that SIN-1 and 8-Br-cGMP (data not shown) significantly increased the level of TP␣ phosphorylation relative to its basal phosphorylation but had no effect on TP␤ phosphorylation. Moreover, KT5823, but not H-89 or GF 109203X, significantly reduced the level of SIN-1-induced TP␣ phosphorylation. Thus, these data establish that TP␣ is subject to NO-mediated desensitization of signaling that occurs through a KT5823-sensitive, PKG mechanism and that the target sites of this desensitization are located within the unique C-tail regions of TP␣ at site(s) distal to the divergent Arg 328 residue. The observations that TP␣ is phosphorylated both in vitro and in vivo by PKG and that its signaling is subject to desensitization by NO/cGMP-regulated mechanisms are consistent with previous reports (44,58). On the other hand, the fact that neither SIN-1 nor 8-Br-cGMP had any effect on signaling by TP␤ or on its basal phosphorylation and that TP␤ is not a substrate for phosphorylation by PKG in vitro establishes that TP␤ is not subject to direct NO-regulated desensitization. Moreover, these data clearly establish that, similar to their responses to the inhibitory prostanoids prostacyclin and PGD 2 , TP␣ and TP␤ are differentially sensitive to NO signaling.
Thereafter, we sought to define the mechanism of NO/cGMP desensitization of TP␣ signaling by identifying the residue(s) with the unique C-tail domain of TP␣ specifically targeted by PKG phosphorylation. Initial deletion mutagenesis confirmed that, like the wild type TP␣, its truncated variant TP␣ ⌬336 devoid of those residues distal to Leu 336 was subject to both SIN-1-and 8-Br-cGMP-induced desensitization of signaling and thereby confirmed that neither Thr 337 nor Ser 340 represent the target(s) for NO/PKG phosphorylation and desensitization. As stated previously, it has been reported that the prostacyclinand PGD 2 -mediated desensitization of TP␣ signaling occurs through a mechanism involving direct PKA phosphorylation of Ser 329 within the consensus PKA site Arg-Pro-Arg-Ser 329 -Leu-Ser-Leu located within the C-tail domain of TP␣ (11,23). However, in other studies investigating the regulation of TP␣ by agents that activate PKG (58) or PKA (59), it has been also suggested that the latter sequence Arg-Pro-Arg-Ser 329 -Leu-Ser 331 -Leu within TP␣ may act as an overlapping PKA and PKG consensus site and that Ser 331 , and not Ser 329 , may represent the predominant site of phosphorylation by both kinases (58,59). Thus, in the current study, we sought to identify the site of NO-mediated PKG phosphorylation within the C-tail domain of TP␣. Moreover, in view of the suggestion the Ser 331 , as opposed to Ser 329 , might act as a phosphotarget for both PKA and PKG, we sought to clarify whether Ser 329 and Ser 331 are targeted independently by the former kinases, respectively, or, on the contrary, whether they act as overlapping PKA and PKG phosphorylation sites.
Hence, to address this issue, the variants TP␣ S329A (23), TP␣ S331A , and TP␣ S329A,S331A were generated whereby the critical Ser 329 and Ser 331 were mutated to Ala 329 and Ala 331 either FIG. 10. Model of nitric oxide-and prostacyclin-mediated regulation of TP signaling. Ligand (TXA 2 ) activation of TP␣ and TP␤ mediates G␣ q -dependent coupling to PLC␤, leading to increases in diacylglycerol (DAG) and IP 3 generation and mobilization of [Ca 2ϩ ] i . NO signaling through its receptor sGC yields increases in cGMP generation and concomitant activation of cGMP-dependent PKG 1␣ and/or 1␤ isozymes. Activated PKG phosphorylates TP␣, but not TP␤, where Ser 331 has been identified as the target residue for PKG phosphorylation and thereby desensitizes signaling by TP␣. PKG-mediated phosphorylation and desensitization of TP␣ signaling can be inhibited by KT5823. Prostacyclin (PGI 2 ) signaling through its receptor IP primarily couples G␣ s -mediated activation of adenylyl cyclase (AC) leading to increases in cAMP generation and, in turn, activation of cAMP-dependent PKA. Activated PKA, in turn, phosphorylates TP␣, but not TP␤, where Ser 329 was identified as the phosphotarget residue and thereby desensitizes signaling by TP␣. PKA-mediated phosphorylation and desensitization of TP␣ signaling can be inhibited by H-89. Hence, signaling by TP␣ is regulated independently by NO-regulated PKG and prostacyclin-regulated PKA mechanisms where Ser 331 and Ser 329 act as phosphotargets, respectively. Signaling by TP␤ is unaffected by NO or prostacyclin.
individually or collectively, and the relative sensitivities of the latter receptors to NO/PKG and prostacyclin/PKA signaling was investigated. Initial saturation radioligand binding experiments confirmed that the mutations per se did not affect the ligand binding properties of the latter TP␣ receptors, and each exhibited efficient U46619-mediated intracellular signaling ([Ca 2ϩ ] i mobilization and IP 3 generation). Although 8-Br-cGMP and SIN-1 (data not shown) significantly reduced U46619-mediated signaling by TP␣ S329A , they had no effect on signaling by TP␣ S331A . Conversely, the prostacyclin mimetic cicaprost did not affect signaling by TP␣ S329A but significantly reduced signaling by TP␣ S331A . Moreover, KT5823, but not H-89 or GF 109203X, blocked 8-Br-cGMP-induced desensitization of TP␣ S329A signaling whereas H-89, as opposed to KT5823 or indeed GF 109203X, blocked cicaprost-induced desensitization of TP␣ S331A signaling. Whole cell phosphorylation assays established that TP␣ S329A underwent increased KT5823-sensitive phosphorylation in response to SIN-1 and 8-Br-cGMP (data not shown), but not in response to cicaprost. In contrast, TP␣ S331A underwent increased H-89-sensitive, KT5823-insensitive phosphorylation in response to cicaprost but not in response to SIN-1 or 8-Br-cGMP (data not shown). Moreover, signaling by TP␣ S329A,S331A was insensitive to desensitization by either the NO donor SIN-1 or 8-Br-cGMP or by the IP agonist cicaprost and did not undergo increased phosphorylation in response to either SIN-1/8-Br-cGMP or cicaprost stimulation (data not shown).
Hence, taken together, these data confirm that TP␣, but not TP␤, is subject to independent regulation or desensitization of signaling by the endothelium-derived vasodilatory/platelet inhibitory autocoids NO and prostacyclin and leads us to propose a model of these desensitizations, as outlined in Fig. 10. In this model, TP␣ expressed, for example, in human platelets or VSM, is subject to desensitization by the endothelium-derived NO mediated through its direct KT5823-sensitive, PKG-mediated phosphorylation whereby Ser 331 , but not Ser 329 , acts as the phosphotarget for that desensitization. On the other hand, consistent with our previous findings (23), TP␣ is also subject to independent cicaprost-induced desensitization that occurs through direct H-89-sensitive PKA phosphorylation of Ser 329 , without targeting Ser 331 . Hence, NO and prostacyclin may act independently or synergistically to regulate signaling by TP␣ through NO-regulated PKG and prostacyclin-regulated PKA mechanisms where Ser 331 and Ser 329 act as the phosphotargets, respectively. Signaling by TP␤ is unaffected by NO or prostacyclin. Each of the molecular components of the NO/sGC/ PKG and prostacyclin/IP/PKA signaling systems are expressed abundantly in platelets and VSM cells where they act to inhibit platelet activation and lead to vasodilation (2,36). Although many of the targets of PKG phosphorylation have been identified, including the IP 3 receptor and vasodilator-stimulated phosphoprotein (VASP) within platelets and SM, phospholambam, and the Ca 2ϩ -activated K ϩ (K Ca ) channel in SM (36), the precise mechanism of NO/PKG actions in platelets and VSM is not fully understood, and it is indeed acknowledged that other unidentified targets may exist (60). NO regulation and synthesis via eNOS activation can occur in response to a diverse array of agonists that signal through GPCRs, including the bradykinin and muscarinic acetylcholine receptors signaling pathways (24,61). Moreover, the dependence by eNOS on Ca 2ϩ for its activation provides a potential feedback mechanism whereby GPCR-mediated Ca 2ϩ mobilization leads to NO activation and subsequent cGMP/PKG signaling (61). In the current study, we have established that TP␣ is a novel direct target for NO/sGC/ PKG phosphorylation and desensitization of signaling and have confirmed direct cross-regulation of the inhibitory NO/ PKG system with the stimulatory TXA 2 /PLC system at the level of the TP␣ receptor, rather than at the level of downstream effectors. Moreover, the finding that signaling by TP␤ is unaffected by NO/PKG inhibitory mechanisms, in addition to its established lack of responses to prostacyclin/PKA mechanisms (23), unveils novel insights into the differential, independent modes of regulation of the individual TP␣ and TP␤ isoforms and points to further potentially critical physiologic differences between the TP isoforms within the vasculature.