Abstract
To establish whether the thromboxane A2 (TXA2) receptor (TP) functionally couples to the Gq family of heterotrimeric G proteins in vivo, we have coexpressed the cDNAs coding for the human platelet/placental TP alpha isoform (TPα) and the α subunits of Gq or G11 in human embryonic kidney (HEK) 293 cells. TP activation in response to ligand stimulation was monitored by analyzing mobilization of intracellular calcium (Ca++ i) in FURA2/AM-loaded transfected HEK 293 and in platelets. Second, we wished to examine the possible interaction of the isoprostane 8-epi prostaglandin F2α with the TPα, in transfected HEK 293 cells and with the TPs expressed in platelets. Thus both the prostaglandin endoperoxide/TXA2 analog (U46619) and the 8-epi PGF2α were utilized as ligand probes of TPα activation. The results demonstrate that each ligand induced elevations of Ca++ i levels in HEK 293 cells, cotransfected with either the TPα and Gαq or the TPα and Gα11, and also in platelets. Initial stimulation of these cells with U46619 or 8-epi PGF2α desensitized a subsequent rise in [Ca++]i in response to U46619 or 8-epi PGF2α, respectively. Moreover, prestimulation with U46619 desensitized a subsequent rise in Ca++ iconcentration in response to 8-epi PGF2α, and vice versa. These responses were blocked by the TP antagonist SQ29,548 in both cell types. In contrast, prestimulation of the transfected HEK 293 cells or platelets with thrombin did not desensitize a subsequent rise in [Ca++]i in response to U46619 or 8-epi PGF2α. After stimulation with either U46619 or 8-epi PGF2α, no significant rise in Ca++ i levels was observed in HEK 293 cells transfected with the TPα receptor only or in control cells transfected with the vector pCMV5. These results demonstrate that the TPα isoform functionally couples with either Gq or G11 in vivo, whether activated by a PG/TXA2 analog or by the F2 isoprostane 8-epi PGF2α.
TP is a G protein-coupled receptor (Hirata et al., 1991) that, on ligand stimulation, results in activation of PLC and subsequent increases in inositol IP3, DAG and Ca++ i concentrations (Brass et al., 1987). The human TP cDNA was originally cloned from placenta and a platelet-like megakaryocyte cell line (Hirata et al., 1991). In a result consistent with previous pharmacologic and biochemical evidence pointing to the existence of TP isoforms (Dorn, 1989; Takahara et al., 1990; Furci et al., 1991), a second form of TP has recently been cloned from human umbilical vein endothelial cells (HUVECs) (Raychowdhury et al., 1994, 1995). The endothelial receptor, termed TPβ, and the platelet/placental TP, termed TPα, are derived by an alternative splicing mechanism; they are identical for the first 328 amino acids but differ in their carboxyl terminal cytoplasmic tails. TPα activity is regulated in vivo both by direct phosphorylation and by regulated expression of the TP gene (Kinsella et al., 1994a). Polymerase chain reaction (PCR) analyses have confirmed that HUVECs contain only TPβ, whereas both TPαand TPβ are expressed in placental tissues and in platelets (Raychowdhury et al., 1994, 1995; Hirata et al., 1996; Miggin and Kinsella, unpublished). DNA sequence analysis has also pointed to sequence polymorphisms in the promoter of the gene coding for the human TPs (Kinsella et al., 1994b).
Using a variety of in vitro approaches, involving reconstitution studies (Shenker et al., 1991), copurification experiments (Knezevic et al., 1993) or photo cross-linking studies with GTP analogs (Offermanns et al., 1994), various investigators have previously proposed that the platelet TP(s) might couple to the heterotrimeric G proteins Gq and G12. However, there has been no direct demonstration of functional coupling between the TPα receptor and these G proteins in vivo.
The isoprostanes (O’Connor et al., 1984) are biologically potent prostanoids that are primarily generated in vivo by nonenzymatic, free radical-catalyzed lipid peroxidations (Morrowet al., 1990). D-ring, E-ring and F-ring isoprostanes are generated in vivo (Morrow et al., 1990; Morrowet al., 1994a, 1994b; Takahashi et al., 1992). One of these compounds, 8-epi PGF2α, may also be synthesized as a product of COX enzymes in human platelets and monocytes (Pratico and Fitzgerald, 1995; 1996). A potent vasoconstrictor in both lung and kidney, 8-epi PGF2αsignificantly reduces renal blood flow and glomerular filtration rates (Morrow et al., 1990, Morrow et al., 1994a). In rabbits, the vasoconstriction of pulmonary vasculature induced by 8-epi PGF2α appears to be due to the activation of the SQ29,548 responsive TP(s) (Banerjee et al., 1992). Whereas 8-epi PGF2α induces a shape change in human platelets at 10−6 M and 10−5 M, at higher concentrations (10−4 M) it induces reversible but not irreversible aggregation (Morrow et al., 1992). All of these actions of 8-epi PGF2α were blocked by the TP antagonist SQ 29,548. However, both 8-epi PGF2α and its structural isomers 9α, 11β-PGF2 and PGF2α also inhibit platelet aggregation induced by such PG endoperoxide/TXA2analogs as U46619, I-BOP {5-heptenoic acid, 7-[3-[3-hydroxy-4-(4-iodophenoxy)-1-butenyl]-7-oxabicyclo [2.2.1] hept-2-yl]-,[15-[1α, 2α(Z), 3β(1E,3S*), 4α]]} and by arachidonic acid. This suggests that 8-epi PGF2α may function as an antagonist of the platelet TPs (Morrow et al., 1992; Yin et al., 1994). More recently, it has been shown that whereas 8-epi PGF2α can inhibit platelet aggregation induced by low-dose collagen but not thrombin, it can also potentiate reversible platelet aggregation in response to low-dose ADP (Yin et al., 1994). This suggests that 8-epi PGF2α may have partial agonist activity, mediated through the platelet TP(s) (Yin et al., 1994). Furthermore, it has been suggested that 8-epi PGF2α may exert its biological actions in vascular smooth muscle through activation of receptor sites related to but distinct from TPs (Fukunaga et al., 1993). However, in competition binding studies, both we (Kinsella et al., 1994b; Pratico et al., 1996) and others (Yinet al., 1994) have shown that very high concentrations of 8-epi PGF2α are necessary to displace the radiolabeled TXA2 antagonist SQ 29,548 or the agonist I-BOP from either the cloned TPα isoform expressed in human embryonic kidney 293 cells or the TPs expressed on platelets, 8-epi PGF2α being approximately 1000 times less effective than the unlabeled analogs of either TXA2 (I-BOP) or PGH2 (U46619) or the antagonist SQ29,548 (Kinsella et al., 1994b). These observations, coupled with the discrepancies between the EC50 values for ligand displacement and the concentrations of 8-epi PGF2α in the circulation (Morrowet al., 1990) even in syndromes of oxidant stress, call into question the role of this compound as an endogenous TP ligand in vivo.
Thus, in order to establish directly whether the TPαreceptor isoform can couple to members of the heterotrimeric Gq family, resulting in activation of phospholipase C, we have coexpressed the human TPα isoform and the α subunits of Gq or G11 in HEK 293 cells, and have analyzed changes in Ca++ i concentrations in FURA2/AM-loaded HEK 293 transfectants in response to stimulation by the stable TXA2 mimetic U46619. HEK 293 cells express very low levels of endogenous TPs (Kinsella et al., 1994b) or Gαq or Gα11 (Conklin et al., 1992) and therefore provide an ideal background to define the individual components of the TP receptor-mediated signal transduction pathways. To address the question whether the isoprostane 8-epi PGF2α can functionally activate the platelet/placental TPα isoform, we have also analyzed the ability of this ligand to induce mobilization of Ca++ i in transfected HEK 293 cells and in platelets. Given the recent study that demonstrated that receptor affinity for synthetic TP mimetics might be modulated by cotransfected G proteins Gα13 and Gαq (Allan et al., 1996), we also wished to address the possibility that G protein coexpression enhances the affinity of TPα for 8-epi PGF2α, relative to the mimetic U46619. Our results demonstrate that the human TPα isoform functionally couples to the G proteins Gq and G11 in vivo and that 8-epi PGF2α may indeed directly activate the cloned TPα isoform expressed in the HEK 293 cells and also the TP receptor(s) expressed in platelets. Furthermore, although either TP receptor: G protein complex may be activated by 8-epi PGF2α, it remains a considerably less potent ligand than either U46619 or I-BOP, structural mimetics of PG endoperoxides and TXA2.
Materials and Methods
Materials.
The following chemicals were obtained from Cayman Chemical Company: 5-heptenoic acid, 7-[6-(3-hydroxy-1-octenyl)-2-oxabicyclo [2,2,1] hept-5-yl]-[1R-[1α,4α,5β(z), 6α(1E,3S*)]-9,11-dideoxy-9α,11α-methanoepoxy prostaglandin F2α (U46619); 8-epi PGF2α; 5-heptenoic acid, 7-[3-[3-hydroxy-4-(4-iodophenoxy)-1-butenyl]-7-oxabicyclo [2.2.1] hept-2-yl]-,[15-[1α, 2α(Z), 3β(1E,3S*), 4α]] (I-BOP); 5-heptenoic acid, 7-[3-[[Z-[phenylamino)carbonyl]hydrazino]methyl]-7-oxabicyclo [2.2.1] hept-2-yl]-,[1S-[1α,2α(Z),3α,4α]] (SQ29,548). Thrombin and [1-[2-(5-carboxyoxazol-2-yl)-6-aminobenzofuran-2-oxy}-2-(2′-amino-5′-methylphenoxy)-ethane-N,N,N′N′-tetraacetic acid, pentaacetoxymethyl ester] (FURA2/AM) were obtained from Calbiochem. [3H]SQ29,548 (50.4 Ci/mmol) and anti G-protein GA/I antibody were obtained from DuPont NEN.
Plasmids.
The plasmids p3:Gαq and pBluescript(KS+):Gα11, coding for the α subunits of the mouse heterotrimeric G proteins Gq and G11(Strathman and Simon, 1990), respectively, were kindly donated by Dr. Melvin Simon, California Institute of Technology, Pasadena, CA. The plasmid pCMV:Gαq was constructed by subcloning the full-length coding sequence for Gαq from plasmid p3:Gαq into the Kpn1-BamH 1 sites of pCMV5 (Andersson et al., 1989). The plasmid pCMV:Gα11 was constructed by subcloning the full-length coding sequence for Gα11 from plasmid pBluescript(KS+):Gα11 into the Kpn1-HindIII sites of pCMV5. The plasmids pCMV5 and pCMVTXR, containing the full-length cDNA for the human platelet/placental TPα, have been previously described (Andersson et al., 1989; Kinsellaet al., 1994a, respectively).
Cell culture and transfections.
HEK 293 cells, obtained from the American Type Culture Collection, were grown in minimal essential medium containing 10% heat-inactivated horse serum. For transfection studies, HEK 293 cells were plated in 100-mm culture dishes approximately 24 hr before transfection at a density of 1.8 to 2 × 106 cells/dish. Cells were transfected with 10 μg of pADVA (Gorman et al., 1990) and 25 μg of pCMV5 or the pCMV-based vectors using the calcium phosphate/DNA coprecipitation procedure (Graham and van der Eb, 1973). The cells were harvested 48 hr after transfection.
Radioligand binding and Western blot analysis.
Transfected HEK 293 cells were harvested by centrifugation at 500 ×g at 4°C for 5 min, were washed twice in Dulbecco’s phosphate-buffered saline (PBS) and were resuspended in modified Ca++/Mg++-free Hank’s buffered salt solution containing 10 mM HEPES, pH 7.67, and 0.1% bovine serum albumin (HBSSHB buffer). Alternatively, in order to fractionate the cells into their soluble (S100) and membrane (P100) components, washed cells were resuspended and homogenized in HED buffer (20 mM HEPES, pH 7.67, 1 mM EGTA and 0.5 mM dithiothreitol) supplemented with 1 mM phenylmethylsulfonyl fluoride and 10 μM indomethacin. The homogenates were centrifuged at 100,000 × g for 30 min at 4°C, and the membrane fractions (P100) were resuspended in HEDG buffer (20 mM HEPES, pH 7.67, 1 mM EGTA, 0.5 mM dithiothreitol, 100 mM NaCl and 10% glycerol) supplemented with 1 mM phenylmethylsulfonyl fluoride and 10 μM indomethacin. Protein determinations were carried out according to the Bradford assay (Bradford, 1976). For ligand binding studies, protein concentrations in the membrane fractions and whole-cell fractions were diluted to 1 mg/ml in HEDG and HBSSHB buffer, respectively. Saturation radioligand binding experiments with the TP antagonist [3H] SQ 29,548 (20 nM, 50.4 Ci/mmol) were carried out at 30°C for 30 min in 100-μl reactions. Nonspecific binding was determined in the presence of excess nonlabeled SQ 29,548 (10 μM). Reactions were terminated by the addition of 4 ml of ice-cold 10 mM Tris-HCl, pH 7.4, followed by filtration through Whatman GF/C glass filters. Subsequent washing of the filters three times with 10 mM Tris-HCl, pH 7.4, was followed by liquid scintillation counting of the filters in 5 ml of scintillation fluid.
For Western blot analysis, aliquots of the S100 (25 μg) and the P100 (75 μg) cell fractions were solubilized by boiling at 100°C for 5 min in solubilization buffer (10% β-mercaptoethanol, 2% SDS, 30% glycerol, 0.025% bromophenol blue, 50 mM Tris-HCl, pH 6.8) and were resolved on a 10% SDS-polyacrylamide gel (SDS-PAGE) (Laemmli, 1970), followed by Western blot transfer onto nitrocellulose filters according to standard procedures (Sambrook et al., 1989). The Western blots were screened with the anti G-protein GA/I antibody (1:2000 dilution), essentially as described (Maltese and Sheridan, 1990).
Preparation of platelets.
Blood was drawn viavenipuncture from normal human volunteers, who had not taken any medication for at least 10 days, into syringes containing indomethacin (10 μM) and 3.8% sodium citrate (9:1 v/v) (final concentration 0.38% sodium citrate). The blood was centrifuged for 10 min at 160 × g; the platelet-rich plasma (PRP) was removed and recentrifuged for 10 min at 160 × g to remove contaminating red blood cells. Where necessary, PPP was prepared by spinning the remaining blood at 900 × g for 15 min. The quality of the PRP was routinely checked by monitoring the aggregation properties of the platelets (data not shown). For aggregation studies, platelets in PRP were diluted to approximately 108 platelets/ml in PPP; 0.5-ml aliquots were preincubated at 37°C for 2 min before addition of the aggregating agent (1 μM U46619 or 0.1 U/ml thrombin), and the extent of aggregation was monitored by light transmission in a Biodata Pap 4 aggregometer.
Calcium measurements.
Ca++ imeasurements either in transfected HEK 293 cells or in platelets were made by monitoring the intensity of FURA2 fluorescence. For the transfected cells, 48 hr after transfection the HEK 293 cells were washed twice in PBS, resuspended in HBSSHB buffer at 107cells/ml and incubated in the dark with 5 μM FURA2/AM for 45 min at 37°C. Subsequently the cells were collected by centrifugation (900 × g, 5 min), washed once in an equal volume of HBSSHB and finally resuspended in HBSSHB buffer at 107cells/ml and kept at room temperature in the dark until use. For each measurement of Ca++ i, aliquots of HEK 293 cells were diluted to 0.825 × 106 cells/ml in HBSSHB buffer containing 1 mM CaCl2. For platelet studies, PRP was incubated in the dark with 5 μM FURA2/AM (Calbiochem) at 37°C for 45 min; platelets were harvested by centrifugation (900 ×g, 15 min), washed once in resuspension buffer (10 mM HEPES, 145 mM NaCl, 5 mM KCl, 5.5 mM glucose, pH 7.4) and finally resuspended in resuspension buffer containing 1 mM CaCl2 at a cell density of 3 × 108/ml and kept at room temperature in the dark until use.
FURA2 fluorescence was recorded in HEK 293 cells and in platelets (2-ml aliquots) at 37°C with gentle stirring using a Perkin Elmer-Cetus LS50-B spectrofluorimeter at excitation wavelengths of 340 nm and 380 nm and emission wavelengths of 510 nm (Grynkiewiz et al., 1985). For each of the cell types used in this study, a dose-response curve to U46619 or 8-epi PGF2α was determined, and individual ligand EC50 values were found to be identical irrespective of the cell type. Thus each experiment was performed with the dose of agonist corresponding to 1 μM U46619 or 10 μM 8-epi PGF2α, unless otherwise specified. A rapid, transient rise and fall in Ca++ i levels in response to ligand stimulation was interpreted as receptor-mediated Ca++ i mobilization. The calibration of the signal was performed in each sample by adding 0.2% Triton X-100 to obtain the maximal fluorescence (F max) and then adding 1 mM EGTA to obtain the minimal fluorescence (F min). A rapid, transient rise and fall in Ca++ i levels in response to ligand stimulation was interpreted as receptor-mediated Ca++ mobilization. The ratio of the fluorescence at 340 nm to that at 380 nm is a measure of Ca++ i (Grynkiewiz et al., 1985), assuming a K d of 225 nM Ca++ 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++ mobilized (Δ[Ca++]i (nM)) as a function of time (seconds) upon ligand stimulation.
Results
Coexpression of the TPα isoform and Gαqor Gα11 in transfected HEK 293 cells.
To address directly whether it might couple to the heterotrimeric G proteins Gq and/or G11, we transiently expressed the TPα isoform in HEK 293 cells, either in the presence of the α subunits of the heterotrimeric G protein Gq or G11 or in the presence of the control vector pCMV5. Expression of TPα was confirmed by radioligand binding assays using the radiolabeled TP antagonist [3H]SQ29,548 (Ogletree et al., 1985). G-protein expression was confirmed by Western blot analysis using an antibody, GA/I, that recognizes a conserved region within many heterotrimeric G-protein α subunits (Goldsmith et al., 1988). In line with our previous reports (Kinsella et al., 1994a), HEK 293 cells display very low levels of TP expression (66 ± 12.9 fmol/mg protein, table1); thus they provide an ideal background in which to study activation of the transfected TPα. Transfection of HEK 293 cells with the TPα cDNA (pCMVTXR) either in the presence of the control vector, pCMV5, or in the presence of plasmids coding for either Gαq (pCMV:Gαq) or Gα11(pCMV:Gα11) resulted in high-level TP expression (table 1) as compared with the level of TP radioligand binding in cells transfected with the control vector only (table 1). Positive expression of the heterotrimeric G proteins Gαq and Gα11 was observed in the membrane fractions of cells transfected with the corresponding cDNAs coding for Gαq and Gα11, respectively, but not in the control transfected cells or in those cells transfected with TPα only (fig.1).
U46619 and 8-epi PGF2α induce Ca++i mobilization in HEK 293 cells cotransfected with the cDNAs for the TPα isoform and Gαq or Gα11.
Functional coupling of TP to PLC activation was assessed throughout by monitoring mobilization of Ca++ i in FURA/2AM-loaded cells in response to TP selective ligand U46619 or the isoprostane 8-epi PGF2α. Stimulation of HEK 293 cells, transfected with either the vector pCMV5 control or TPα, with 10−6 M U46619 (fig. 2A and B, respectively) or with 10−5 M 8-epi PGF2α (fig.3, A and B, respectively) failed to induce a transient rise in Ca++ i. In contrast, stimulation of HEK 293 cells cotransfected with either TPα and Gαq or TPα and Gα11 by either 10−6 M U46619 (fig. 2C; Δ[Ca++]i = 54.15 ± 8.4 nM and fig.2D; Δ[Ca++]i = 107.6 ± 22.6 nM, respectively) or 10−5 M 8-epi PGF2α (fig.3C; Δ[Ca++]i = 53.2 ± 14.4 nM and fig. 3D; Δ[Ca++]i = 99.4 ± 49.2 nM, respectively) resulted in a rapid, transient rise in Ca++ i levels.
Furthermore, prestimulation of HEK 293 cells, cotransfected with TPα and Gα11, with 10−6 M U46619 or with 10−5 M 8-epi PGF2αdesensitized a subsequent rise in Ca++ i upon a second stimulation with the same agonists (fig. 4A; Δ[Ca++]i = 98.3 ± 32.8 nM and fig.4B; Δ[Ca++]i = 67.75 ± 15.5 nM, respectively). In addition, prestimulation with 10−6 M U46619, which resulted in a rapid transient rise in Ca++ i, cross-desensitized a subsequent rise in Ca++ i upon stimulation with 10−5 M 8-epi PGF2α, and vice versa (fig. 4C; Δ[Ca++]i = 100.3 ± 28.6 nM and fig.4D; Δ[Ca++]i = 74 ± 20.2 nM). On the other hand, initial stimulation of HEK 293 cells, cotransfected with the TPα and Gα11, with thrombin (0.1 U/ml) failed to desensitize Ca++ i mobilization in response to secondary stimulation with either 10−6 M U46619 (fig. 5A; Δ[Ca++]i = 113.1 ± 41.7 nM, and vice versa (fig. 5B; Δ[Ca++]i = 139.14 ± 74.1 nM) or 10−5 M 8-epi PGF2α (data not shown).
The specificity of activation of the transfected TPα by both U46619 and 8-epi PGF2α, as monitored by analyses of the mobilization of Ca++ i, was further confirmed by pretreatment with the TP antagonist SQ29,548. In each case, pretreatment with SQ29,548 antagonized the rise in Ca++ i induced by either U46619 (fig. 4E) or 8-epi PGF2α (data not shown). Similar results were observed in HEK 293 cells using the selective TP agonist I-BOP in place of U46619 (data not shown).
8-epi PGF2α induces Ca++imobilization in platelets through activation of TP.
Stimulation of FURA2/AM-loaded platelets with the TP agonist U46619 (10−6 M) induced a rapid, transient rise in Ca++ i (fig. 6A; Δ[Ca++]i = 210.6 ± 25.9 nM). Stimulation of platelets with 8-epi PGF2α(10−5 M) also induced a rapid, transient rise in Ca++ i levels (fig. 6B; Δ[Ca++]i = 58.8 ± 13.5 nM). Preincubation of platelets with SQ29,548 (10−6 M) blocked Ca++ i mobilization in response to subsequent stimulation with either 10−6 M U46619 (fig. 6C) or 10−5 M 8-epi PGF2α (fig. 6D). On the other hand, initial stimulation of platelets with 10−6 M U46619, which resulted in the transient rise in [Ca++]i, desensitized a subsequent rise in Ca++ i upon treatment with 10−6 M U46619 (fig. 7A; Δ[Ca++]i = 222.5 ± 28.8 nM). Similarly, initial activation of platelets with 10−5 M 8-epi PGF2α, which resulted in the elevation of Ca++ i levels, also desensitized a subsequent rise in Ca++ i upon treatment with 8-epi PGF2α (fig. 7B; Δ[Ca++]i = 61.25 ± 17.2 nM). By the same token, pretreatment with 10−6 M U46619α(fig. 7C; Δ[Ca++]i = 210 ± 35.14 nM) or 10−5 M 8-epi PGF2α (fig. 7D; Δ[Ca++]i = 61.75 ± 14.8 nM) desensitized a secondary rise in Ca++ i upon stimulation with 10−5 M 8-epi PGF2α or 10−6 M U46619 (fig. 7, C and D, respectively), whereas initial stimulation of platelets with thrombin (0.1 U/ml) did not desensitize secondary stimulation with either U46619 (fig. 7E; Δ[Ca++]i = 402 ± 49.6 nM for thrombin, Δ[Ca++]i = 223.2 ± 26.35 nM for U46619) or 8-epi PGF2α (data not shown). Thus, in agreement with previously published data, pretreatment of platelets with the TP antagonist SQ 29,548 abolished the rise in Ca++ i induced by either U46619 or 8-epi PGF2α (data not shown). This indicates that the rise in Ca++ i induced by 8-epi PGF2α at these concentrations is mediated by activation of platelet TP(s). Moreover, U46619, a known agonist of platelet TP(s), cross-desensitized the rise in Ca++ i induced by 8-epi PGF2α. Similarly, primary stimulation with 8-epi PGF2α cross-desensitized the U46619-induced elevation of Ca++ i, which is substantial evidence of the activation of platelet TP by the isoprostane 8-epi PGF2α. Similar results were obtained for the TP agonist I-BOP, a closer structural mimetic of TXA2 than U46619 (Dorn, 1989). I-BOP induced a rapid, transient rise in Ca++ i that rapidly desensitized a secondary rise in Ca++ iin response to I-BOP, U46619 or 8-epi PGF2α; 8-epi PGF2α almost completely desensitized a second rise in Ca++ i in response to I-BOP (data not shown).
U46619 induces mobilization of Ca++ from intracellular stores.
To demonstrate that the Ca++ mobilized in platelets and the transfected HEK 293 cells in response to TP activation was of intracellular origin, before agonist stimulation we preincubated FURA2/AM-loaded cells with EGTA in order to chelate extracellular Ca++. Preincubation of platelets with EGTA did not interfere with the elevation of Ca++ ilevels upon stimulation with U46619 (compare figs. 8A and 8B, Δ[Ca++]i = 208.9 ± 25 nMvs. fig. 8B, Δ[Ca++]i = 209.75 ± 41.5 nM). Similar results were observed in HEK 293 cells transfected with TPα and Gαq (fig. 8C, Δ[Ca++]i = 128 ± 35.3 nM). Taken together, these results demonstrate that U46619 induces the mobilization of Ca++ i stores after activation of TPα.
Discussion
We have investigated the functional coupling of the human placental/platelet TPα isoform to the heterotrimeric G proteins Gq and G11 in transfected HEK 293 cells by monitoring activation of the downstream signal transduction events as reflected by elevation of Ca++ ilevels after activation with the prostaglandin endoperoxide/TXA2 mimetic U46619. In addition, we have explored the structure-function relationships of receptor-G protein action by comparing the effects of U46619 and the isoprostane 8-epi PGF2α on the platelet TP(s) and on the cloned platelet/placental TPα (Hirata et al., 1991), expressed in the mammalian HEK 293 cells. U46619 induced a rapid, transient rise in Ca++ i in platelets and in HEK 293 cells cotransfected with TPα and Gαq or Gα11. It is noteworthy that in the transfected HEK 293 cells, we did not observe a significant rise in Ca++ i in response to U46619 unless the TPα was coexpressed in the presence of a member of the Gαq subfamily of heterotrimeric G proteins. It has been reported that HEK 293 cells lack endogenous Gαq (Conklinet al., 1992) and that alpha-2 adrenoreceptor stimulation of phosphoinositide-specific PLC (PI-PLC) activity in HEK 293 cells was completely dependent on the coexpression of Gαq (Conklin et al., 1992). Using a variety ofin vitro approaches involving reconstitution studies (Shenker et al., 1991), copurification experiments (Knezevicet al., 1993) or cross-linking studies with photoactivated GTP analogs (Offermanns et al., 1994) various investigators have previously proposed that TP might couple to Gq and G12. Our studies in the transfected HEK 293 cells now directly demonstrate that the platelet/placental TPαisoform can functionally couple to Gq or to G11 in vivo, resulting in the downstream mobilization of Ca++ i after stimulation with U46619. In HEK 293 cells, cotransfection of Gα11 produced greater mobilization of Ca++ i than cotransfection of Gαq in response to U46619 stimulation, which suggests a possible preference by TPα for G11 or the preferential coupling of G11 vs. Gqto PLC in this system. In this study, we have also directly established that the isoprostane 8-epi PGF2α induced Ca++ i mobilization in HEK 293 cells cotransfected with TPα and Gαq or Gα11; cotransfection of Gα11 produced greater mobilization of Ca++ i than cotransfection of Gαq in response to 8-epi PGF2α stimulation. This again indicates a possible preference by TPα for G11 in this system. 8-epi PGF2α failed to mobilize Ca++ in HEK 293 cells transfected with the TP receptor alone or with the vector control only. In HEK cotransfected cells and in platelets, higher concentrations of 8-epi PGF2α (10−5 M) than of U46619 (10−6) were required to evoke a half-maximal response in either cell type; however, the maximal Ca++response observed in platelets was 3- to 4-fold greater after stimulation with U46619 than with 8-epi PGF2α.
The initial response to U46619 or to 8-epi PGF2α rapidly desensitized a second rise in Ca++ i levels in response to subsequent stimulation by either agent in platelets and in transfected HEK 293 cells. Preincubation of platelets or transfected cells with thrombin, on the other hand, did not desensitize the rise in intracellular Ca++ on subsequent stimulation with either U46619 or 8-epi PGF2α. Thus TP(s) in platelets or the TPα isoform expressed in HEK 293 cells may be subject to homologous desensitization after activation by either U46619 or 8-epi PGF2α. The addition of 8-epi PGF2α(10−5 M) and U46619 (10−6 M) together did not potentiate or antagonize the maximal level of Ca++mobilized in either platelets or transfected HEK 293 cells, which suggests that in platelets at least, 8-epi PGF2α and U46619 may activate the same TP receptors. Moreover, the TP antagonist SQ29,548 was equipotent in abolishing the Ca++ response in both platelets and transfected HEK 293 cells on stimulation with either U46619 or 8-epi PGF2α.
Although 8-epi PGF2α may activate TPs, it is unknown whether this effect is of biological significance. Indeed, we have recently reported that the EC50 concentrations for functional responses in platelets are much higher than the plasma concentrations of the ligand obtained, even in syndromes of oxidant stress (Pratico et al., 1996). The comparative potency of U46619 and 8-epi PGF2α in platelets suggests that in this endogeneous system, 8-epi PGF2α is likely to activate TPsin vivo only when produced in very large amounts as an incidental ligand, or perhaps it has a stronger affinity for a closely related but distinct receptor that has yet to be identified. In conventional terms, whereas the dose-response data do not favor 8-epi PGF2α as an endogeneous TP ligand, receptor activation by 8-epi PGF2α might occur if the ligand were presentedvia an unusual concentrated delivery system, such as microvesicles shed from activated cells, or through selective reincorporation of released isoprostanes into the membrane (Barryet al., 1996).
The transfected HEK system also addresses the capacity of the TPα receptor to couple with different G proteins in a defined system. Whereas TPα readily couples to both Gq and G11 in the transfected cells, we cannot directly extrapolate the actual preference exhibited in an endogeneous system, such as platelets, where both TP isoforms and several G proteins are available. Recently, Allan et al. (1996) have reported that the affinity of the TPα isoform for structural analogs of PG endoperoxides and TXA2 might be modified, depending on the nature of a cotransfected G protein. However, cotransfection of Gαq or Gα11 did not appear markedly to increase the affinity of TPα for 8-epi PGF2α, which remained a less-favored ligand than U46619 in all of our experimental conditions.
In summary, we have demonstrated that human TPαfunctionally couples to members of the Gq family of heterotrimeric G proteins in vivo. Activation of the receptor-G-protein complexes was demonstrated using two structurally distinct ligands, U46619 and 8-epi PGF2α. Although generation of the latter by the COX 1 and COX 2 enzymes lends credence to the possibility that it may function as an autocoid, its biological role in vivo remains to be elucidated. Irrespective of this possibility, we have demonstrated that TPs in human platelets and the TPα isoform may be specifically activated by this compound. Thus, although 8-epi PGF2α represents a potential alternative endogenous ligand for this receptor, it is a considerably less avid ligand than either U46619 or I-BOP, structural mimetics of PG endoperoxides and TXA2. Furthermore, it remains to be established whether incidental activation of TP receptors by 8-epi PGF2α may actually contribute to the adverse effects of oxidant stress in vivo.
Acknowledgments
We are grateful to Catriona Scaife for assistance in the preparation of figures.
Footnotes
-
Send reprint requests to: D. J. O’Mahony, Elan Corporation Research Institute, Trinity College, Dublin 2, Ireland.
-
↵1 This research was supported by grants from the Irish Heart Foundation, Forbairt, the Health Research Board of Ireland (BTK) and the Wellcome Trust (BTK & GAF).
- Abbreviations:
- AOSMC
- aortic smooth muscle cell
- Ca++i
- intracellular calcium
- COX
- cyclooxygenase
- DAG
- diacylglycerol
- 8-epi PGF2α
- 8-epi prostaglandin F2α
- HEK
- human embryonic kidney
- IP3
- inositol 1,4,5-triphosphate
- PAGE
- polyacrylamide gel electrophoresis
- PLC
- phospholipase C
- PPP
- platelet poor plasma
- PRP
- platelet rich plasma
- SDS
- sodium dodecyl sulphate
- TP
- thromboxane A2 receptor
- TXA2
- thromboxane A2
- Received August 26, 1996.
- Accepted January 17, 1997.
- The American Society for Pharmacology and Experimental Therapeutics