Agonist-dependent phosphorylation of an epitope-tagged human prostacyclin receptor.

An epitope-tagged human prostacyclin receptor (HAhIP) was constructed and stably transfected into human embryonic kidney 293 cells. The receptor exhibited high (Kd = 0.4 ± 0.08 nM, Bmax = 0.7 ± 0.2 pmol/mg protein; n = 4) and low (Kd = 75 ± 27.4 nM, Bmax = 7.1 ± 3.6 pmol/mg protein; n = 4) affinity for iloprost and coupled to both cAMP (EC50 = 0.1 ± 0.03 nM) and inositol phosphate (EC50 = 43.1 ± 10 nM) production. The receptor resolved on SDS-polyacrylamide gel electrophoresis as a broad complex with a molecular mass of 44-62 kDa and is glycosylated and phosphorylated. Stimulation of transfected cells with iloprost induced a rapid time- and concentration-dependent phosphorylation of HAhIP. Pretreatment of cells with a protein kinase C (PKC) inhibitor (GF109203X; 5 μM) abolished basal phosphorylation and dramatically reduced iloprost-induced HAhIP phosphorylation. A protein kinase A (PKA) inhibitor (H89) was largely ineffective under the same conditions. HAhIP phosphorylation was stimulated by receptor-dependent (thrombin, 2 units/ml) or receptor-independent (phorbol 12-myristate 13-acetate, 5 μM) PKC activation; both were abolished by pretreatment of cells with GF109203X. In contrast, receptor-independent (forskolin (5 μM) or dibutyryl cAMP (1 μM)) activation of PKA did not induce HAhIP phosphorylation. These results indicate that the human prostacyclin receptor may be regulated by agonist-dependent phosphorylation. This appears to be mediated, in part, by activation of PKC but not by PKA.

gether with the other prostaglandin receptors (6). Expression studies indicate that the recombinant IP may couple to adenylate cyclase activation via Gs (4,5). This pathway is thought relevant to inhibition of platelet aggregation and vascular smooth muscle relaxation (2). However, IPs have also been reported to couple to an increase in inositol phosphates and intracellular calcium via a pertussis toxin-insensitive G protein (5,7). The biological relevance of this pathway is unknown.
The factors that regulate tissue responsiveness to PGI 2 are poorly understood. Conventionally, activation of GPCRs may induce a loss of responsiveness to ligand via two processes (8).
Receptor phosphorylation may result in a rapid decline in its response to subsequent exposure to agonists, due to uncoupling of the receptor from its attendant G protein(s). Subsequently, receptor internalization and degradation and/or recycling may influence the rate of recovery of response to the ligand in question.
Two main classes of kinases have been implicated in phosphorylation of GPCRs; second messenger-dependent kinases, such as protein kinases A (PKA) and C (PKC) (8) and GRK, which require receptor occupancy by agonist ligand for activation (9). Both classes of kinase may act in tandem. Thus, ␤-adrenoreceptors (10) may be phosphorylated by PKA and GRK 2 and 3, (␤-adrenoreceptor kinases 1 and 2), while the thrombin receptor (11) may be phosphorylated by PKC and GRK kinases. Phosphorylation by either class of kinase may result in receptor desensitization; the relative contribution of each appears to vary between cell types (12).
It is likely that similar mechanisms may govern eicosanoid receptor regulation. Indeed, we have previously shown that the human thromboxane receptor (TP) may be phosphorylated by PKA and PKC in vitro (13) and that the sequence of events that accompanies receptor desensitization appears similar to that described for adrenoreceptors (14,15). However, the absence of appropriate reagents has limited insight into the molecular events that accompany desensitization of the receptors for these biologically active lipids. Given the potent biological actions of PGI 2 and the poor understanding of its role in vivo, we have focused upon elucidating the factors that regulate response to this eicosanoid. We have utilized an epitope tag to localize the recombinant hIP to the cell membrane and to immunoprecipitate it from human embryonic kidney cells. The receptor is subject to post-translational modifications, which include glycosylation and phosphorylation. Exposure of the receptor to a stable PGI 2 analog results in rapid receptor phosphorylation. Surprisingly, this appears to be largely dependent on activation of PKC rather than PKA.  1 The abbreviations used are: PGI 2 , prostacyclin; IP, prostacyclin receptor; hIP, human IP; GPCR, G protein-coupled receptor; GRK, GPCR kinase; PKA, protein kinase A; PKC, protein kinase C; HA, hemagglutinin; HAhIP, hemagglutinin-tagged human prostacyclin receptor; TBS-T, Tris-buffered saline with Tween 20; HEK, human embryonic kidney; PMA, phorbol 12-myristate 13-acetate; HAHEK, HEK expressing the HahIP-pcDNAIII construct; VECHEK, HEK expressing the pcDNAIII vector alone.

Materials
(Beverly, MA). Isobutylmethylxanthine and deoxycholic acid were purchased from Sigma. All cell culture reagents, G418, and Albumax were obtained from Life Technologies, Inc. Protein G-Sepharose was purchased from Pharmacia Biotech Inc., and protease inhibitors were from Boehringer Mannheim. pcDNA III was purchased from Invitrogen (San Diego, CA). AG 1-X8 resin (formate form) was obtained from Bio-Rad. The hIP cDNA was generously donated by Dr. Kathleen Metters (Merck Frosst, Quebec, Canada).
Epitope Tagging of the hIP-The 9-amino acid hemagglutinin epitope (HA; YPYDVPDYA) was inserted between the N-terminal initiator methionine and the second amino acid of the hIP by polymerase chain reaction, as described previously (16). The 5Ј-oligonucleotide contained 3 miscellaneous bases, 6 bases encoding a HindIII site, the 3 bases immediately 5Ј of the hIP initiator methionine, 3 bases encoding a methionine, 27 bases encoding the 9-amino acid HA peptide sequence, and 18 bases encoding the first 6 amino acids of the hIP, following the initiator methionine. The 3Ј-oligonucleotide contained 3 miscellaneous bases, 6 bases encoding a BamHI site, 1 miscellaneous base, a stop codon, and 18 bases complementary to the final 6 amino acids of the hIP. Using the hIP cDNA as a template, a polymerase chain reaction was carried out using these primers. The resulting HA-tagged hIP (HAhIP) was purified after electrophoresis using a QIAquick gel extraction kit (Qiagen, Chatsworth, CA) and digested with HindIII and BamHI. This was cloned into M13mp18 and mp19 for complete sequencing on both strands. An M13 clone with the correctly tagged hIP sequence was then digested with HindIII and BamHI again, and the isolated fragment was subcloned into the mammalian expression vector pcDNA III to generate the HAhIP-pcDNA III construct.
Cell Culture and Transfection-HEK 293 cells (American Type Tissue Culture Collection; Rockville, MD) were maintained in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum, 50 units/ml penicillin, 50 g/ml streptomycin, 25 mM HEPES, and 2 mM glutamine. Cells were seeded at 1.5 ϫ 10 6 cells/ 100-mm dish and, the next day, transfected with 100 g/dish HAhIP-pcDNA III or pcDNA III by liposome-mediated transfer (DOTAP; Boehringer Mannheim) according to the manufacturer's instructions. The media were replaced after approximately 6 h with fresh Dulbecco's modified Eagle's medium containing 1.5 mg/ml G418. Several resistant clones, arising from single cells, were selected within 15-20 days and expanded. The G418 concentration was maintained at 1 mg/ml in the cell culture medium.
Radioligand Binding-Membranes were prepared from confluent 100-mm dishes as follows. Cells were washed once with phosphatebuffered saline and scraped into 20 mM Tris, pH 7.4, containing 4 mM EDTA, 10 g/ml aprotinin, 10 g/ml leupeptin, and 0.2 mM phenylmethanesulfonyl fluoride. Cells were lysed by sonication on ice, and membrane fractions were collected by centrifugation at 115,000 ϫ g for 1 h at 4°C. The resulting pellet was resuspended in the same buffer and stored at Ϫ80°C for further use.
Radioligand binding studies were carried out using membrane proteins (50 g/reaction) in 10 mM HEPES, pH 7.4, containing 10 mM MnCl. Reactions were initiated by the addition of [ 3 H]iloprost, allowed to continue for 30 min at 30°C, and terminated by the addition of 4 ml of ice-cold wash buffer (10 mM HEPES, pH 7.4, 0.01% bovine serum albumin), followed by immediate filtration through GF/C filters that had been thoroughly soaked in the same ice-cold buffer. Following one wash with 10 ml of ice-cold wash buffer, radioactivity associated with the filters was quantified by scintillation counting. Nonspecific binding was measured in the presence of a 500-fold excess of unlabeled iloprost. Saturation binding data were analyzed using Radlig version 4 (Biosoft, Cambridge, UK) to calculate K d and B max values for binding and to compare one-and two-site curve-fitting models (partial F-test).
cAMP Measurements-Cells were pretreated with 0.5 mM isobutylmethylxanthine (unless otherwise indicated) for 15 min at 37°C prior to the addition of stimulant (10 min at 37°C) to assess the accumulation of intracellular cAMP. Reactions were terminated by aspiration, and cAMP was extracted with ice-cold 65% ethanol for 30 min. Samples were dried under vacuum and reconstituted in assay buffer, and cAMP was quantified by radioimmunoassay (Amersham) according to the manufacturer's instructions, except that half the recommended amounts of binding protein and [ 3 H]cAMP tracer were used in the samples and standards. Dose-response data were analyzed using the Dose-Effect Analysis program (Biosoft) to calculate EC 50 values.
Inositol Phosphate Production-Cells, grown to 70 -80% confluence in 12-well plates were labeled overnight with 2 Ci/ml [ 3 H]myoinositol in Dulbecco's modified Eagle's medium (without inositol) containing 0.5% Albumax, 50 units/ml penicillin, 50 g/ml streptomycin, and 1 mg/ml G418. Thirty minutes prior to stimulation, cells were treated with 20 mM LiCl at 37°C. After stimulation for 10 min at 37°C, the reactions were terminated by aspiration. Total inositol phosphates were extracted with 750 l of 10 mM formic acid for 30 min at room temperature. Samples were neutralized (final pH 8 -9) with 3 ml of 5 M ammonia. Total inositol phosphates were recovered by anion exchange using Dowex 1-X8 AG anion exchange resin (formate form). Samples were applied to the resin and washed with 50 mM formic acid/ammonium formate (pH 5), and finally the inositol phosphates were eluted with 2 mM formic acid/ammonium formate (pH 5) (17). Dose-response data were analyzed using the Dose-Effect Analysis program (Biosoft) to calculate EC 50 values.
Western Blotting and Immunoprecipitation-Membrane proteins, prepared as outlined above, were resolved (50 g/lane) on 10% sodium dodecyl polyacrylamide gels (SDS-polyacrylamide gel electrophoresis) and transferred to nitrocellulose. The HAhIP was visualized by treating immunoblots with anti-HA (1:1500 dilution) in 5% milk in TBS-T for 60 min at room temperature followed by horseradish peroxidase-conjugated anti-mouse IgG (1:5000 dilution; Jackson Immunology), after first blocking with 5% nonfat milk in TBS-T. Antigen-antibody complexes were visualized by ECL.
All immunoprecipitation procedures were carried out at 4°C. The cells were lysed with radioimmune precipitation buffer (50 mM Tris/5 mM EDTA, pH 8.0, containing 150 mM NaCl, 1% Nonidet P-40, 0.1% SDS, 0.5% deoxycholic acid, 0.5 g/ml aprotinin, 0.5 g/ml leupeptin, 1 mM phenylmethanesulfonyl fluoride, 10 mM NaF, and 10 mM NaH 2 P 2 O 7 ), drawn though a 23-gauge needle six times, rotated for 60 min, and centrifuged at 14,000 rpm. The resulting supernatants were precleared by adding 100 l of 10% (w/v) protein G-Sepharose to each tube and rotating for 60 min. Anti-HA-protein G-Sepharose was prepared by adding 1 l of anti-HA ascites/lysate to 10% protein G-Sepharose and rotating for 60 min. The HAhIP was immunoprecipitated from precleared lysates by adding 100 l of anti-HA-protein G-Sepharose to each lysate and rotating for 2 h. Protein G was precipitated at 14,000 rpm for 1 min, washed three times with radioimmune precipitation buffer, and finally resuspended in 100 l of SDS-polyacrylamide gel electrophoresis sample buffer (60 mM Tris, pH 6.8, 2% SDS, 10% glycerol, 0.002% bromphenol blue, 100 mM dithiothreitol). Samples were then boiled for 10 min and subjected to electrophoresis as outlined above. Gels were either transferred to nitrocellulose or, if radioactive (see below), dried for PhosphorImager analysis.
In Vivo IP Phosphorylation-Cells were plated in 60-mm dishes and grown to 70 -80% confluence. 32 P labeling was carried out using 150 -200 Ci/ml [ 32 P]orthophosphate in phosphate-free Dulbecco's modified Eagle's medium containing 0.5% bovine serum albumin (bovine serum albumin; Fisher) for 60 min at 37°C. Inhibitors (GF109203X or H89) were added for the final 30 min of the labeling reaction where indicated. Labeled cells were treated with stimulant for 10 min (unless otherwise indicated). Dishes were placed on ice, and the overlying medium was removed at the end of the reaction. Cells were lysed with radioimmune precipitation buffer, following one wash with ice-cold phosphate-buffered saline, and the HAhIP was immunoprecipitated as outlined above. Following PhosphorImager analysis of dried gels, rectangles of equal size were drawn around each of the bands observed, and the intensity was quantified using ImageQuant (Molecular Dynamics, Sunnyvale, CA). Data were analyzed using the Dose-Effect Analysis program (Biosoft) to calculate EC 50 values, as appropriate.

Expression of the HAhIP in HEK 293
Cells-HEK 293 cells were transfected with the HAhIP-pcDNA III mammalian expression construct (HAHEK) or with the pcDNA III vector alone (VECHEK). Expression is under the control of the cytomegalovirus promoter, which has been shown to direct high expression of transfected sequences in these cells (18), and successful transfection renders cells resistant to G418 in the culture medium. Positive clones demonstrating G418 resistance were selected for HAhIP expression on the basis of [ 3 H]iloprost binding (10 nM [ 3 H]iloprost, 30 min at 30°C) and immunoblotting with anti-HA (data not shown). One clone that demonstrated high levels of receptor expression was selected for further work. Fig. 1 shows saturation binding of [ 3 H]iloprost to membranes prepared from HAHEK. Scatchard analysis of specific binding revealed the presence of two binding sites, one of high affinity (K d ϭ 0.4 Ϯ 0.08 nM; B max ϭ 0.7 Ϯ 0.2 pmol/mg protein; n ϭ 4) and one of low affinity (K d ϭ 75 Ϯ 27.4 nM; B max ϭ 7.1 Ϯ 3.6 pmol/mg protein; n ϭ 4). The two-site model was a statistically better fit that a one-site model (partial F-test) in each experiment. Both the native and recombinant IP (4,19) have previously been shown to exist in two affinity states, such as we demonstrated for the HA-tagged receptor. The K d for high affinity binding reported here is somewhat lower than that previously documented for iloprost binding in platelets (20) or other nontransfected cells (21)(22)(23) but again is in close agreement with reported K d values in cells overexpressing the receptor (4). Receptor expression was also analyzed by indirect immunocytochemistry (data not shown). Staining, which was undectable in VECHEK, was predominantly at the cell surface in HAHEK.
Stimulation of HAHEK with increasing concentrations of iloprost for 10 min resulted in a concentration-dependent increase in cAMP and inositol phosphate production (Fig. 2). Iloprost has been reported to cross-react with EP1 receptors at higher concentrations (24). However, this is not likely to account for the second messenger production seen in this study, since production of cAMP or inositol phosphates was minimal in VECHEK (Fig. 2). This indicates that the responses seen were due to activation of the HAhIP and not the EP1 receptors endogenously expressed in HEK 293 cells. The EC 50 values for cAMP and inositol phosphate production in HAHEK were 0.1 Ϯ 0.03 nM (n ϭ 3) and 43.1 Ϯ 10 nM (n ϭ 3), respectively. A similar disparity in the EC 50 values for coupling to adenylyl cyclase versus PLC in transfection systems has been reported elsewhere (5) but is not evident in cells that express the receptor endogenously (25,26). The K d for binding of iloprost and the EC 50 values for second messenger coupling in HEK 293 cells expressing the hIP that was not tagged with the HA epitope were not significantly different from those reported above (data not shown). This indicates that the presence of the epitope tag does not alter these aspects of receptor function.
Western blotting of membranes prepared from HAHEK demonstrated that the HAhIP resolved as a broad complex with a molecular mass ranging from 44 to 62 kDa (Fig. 3a). Staining was not evident in VECHEK controls. The primary anti-HA antibody was preincubated with the peptide encoding the HA epitope for 60 min at room temperature to assess the specificity of the anti-HA antibody for the HAhIP. Staining in HAHEK was abolished under these conditions (Fig. 3a). This verifies the specificity of the anti-HA antibody for the receptor expressing the HA epitope in these cells. In a similar fashion, the HAhIP immunoprecipitated from HAHEK appeared as a complex of 44 -62 kDa (Fig. 3b). Samples were run under nonreducing conditions to allow visualization of the receptor in the absence of interference from the anti-HA antibody. There was no immunoprecipitated product from VECHEK controls, and the appearance of the hIP was abolished when the peptide contain- ing the HA epitope was included in the immunoprecipitation mix.
Since the molecular mass for the hIP predicted from the cDNA is 41 kDa (4) and the actual molecular mass is 44 -62 kDa, it seemed likely that, since the hIP contains two putative N-linked glycosylation sites on its N-terminal end and in its first extracellular loop (see Fig. 4), glycosylation of the receptor might be contributing to this discrepancy. Membranes prepared from HAHEK were treated with PNGase F, an enzyme that removes N-linked high mannose, hybrid, and complex oligosaccharide sugars. Treatment of membranes resulted in a shift of the receptor to a major band of 39 kDa with minor bands appearing at lower molecular masses (Fig. 3c). These lower molecular mass species may represent degradation products of the receptor or may indicate the presence of other post-translational modifications. These might include isoprenylation, which could occur at the two cysteines present in the C-terminal tail (see Fig. 4), or the addition of O-linked carbohydrates.
Phosphorylation of the HAhIP in Response to Agonist Stimulation-We next tested whether the HAhIP underwent ago-nist-induced phosphorylation. Fig. 5a shows the time course of phosphorylation of the HAhIP stimulated with 1 M iloprost. The receptor underwent rapid phosphorylation in response to stimulation with iloprost; this was evident within 15 s and was maximal within 5 min. A basal level of phosphorylation was evident (no iloprost added). A time point of 10 min was chosen for all further phosphorylation experiments.
Phosphorylation of the receptor was also concentration-dependent (Fig. 5b). Similar to iloprost-induced inositol phosphate production, phosphorylation was only evident at concentrations of iloprost that exceeded 1 nM. The EC 50 for iloprostinduced phosphorylation was 26.6 Ϯ 2.7 nM (n ϭ 3). No IP phosphorylation was seen at concentrations that induced an increase in cAMP production, suggesting that the phosphorylation of the receptor was unlikely to result from cAMP production. The assay was then carried out in the presence of selective PKC (GF109203X; Ref. 27) or PKA (H89; Ref. 28) inhibitors. As shown in Fig. 6, H89 had no effect on basal phosphorylation and a minimal effect on iloprost-induced phosphorylation of IP. In contrast, GF109203X abolished the basal phosphorylation and dramatically reduced the iloprost-induced signal. GF109203X was an effective inhibitor at all concentrations of iloprost tested (Fig. 7).
These data indicate that agonist-induced phosphorylation may be primarily PKC-rather than PKA-mediated. PKC was activated via a receptor-mediated mechanism to investigate further this possibility. Thrombin receptors are endogenously expressed in HEK 293 cells (29), albeit at a low level, and couple to inositol phosphate production and, by inference, PKC activation. Furthermore, thrombin-induced phosphorylation has been demonstrated to be mediated in part by activation of PKC as a result of receptor stimulation (11). Thus, treatment of HAHEK cells with thrombin may be expected to activate PKC. Thrombin stimulation of HAHEK cells resulted in HAhIP phosphorylation (Fig. 8). This was completely abolished by GF109203X. Furthermore, receptor-independent activation of PKC with the phorbol ester PMA (5 M) also resulted in HAhIP phosphorylation (Fig. 9a). Although the level was lower than that seen with iloprost, this was again abolished by GF109203X while being unaffected by H89.
H89 was largely ineffective as an inhibitor of iloprost-induced phosphorylation. PKA-mediated phosphorylation of IP would be expected to occur at concentrations of iloprost that induced activation of adenylyl cyclase and, hence, intracellular cAMP accumulation. No iloprost-induced phosphorylation was seen at the EC 50 (0.1 nM) for cAMP production (Fig. 5b). In the absence of the phosphodiesterase inhibitor isobutylmethylxanthine, iloprost-induced cAMP production in HAHEK was observed at equally low concentrations of agonist (EC 50 ϭ 0.12 Ϯ 0.03 nM; n ϭ 3), demonstrating that the discrepancy between the concentrations of iloprost required to stimulate cAMP generation and HAhIP phosphorylation was not due to the presence of isobutylmethylxanthine in the cAMP assays. H89 did not significantly inhibit phosphorylation at any of the concentrations of iloprost used (Fig. 7). It appears, therefore, that agonist-induced phosphorylation of HAhIP was not PKA-mediated. This hypothesis was supported by experiments involving receptor-independent adenylyl cyclase activation with forskolin or dibutyryl cAMP (Fig. 9b). No IP phosphorylation was observed induced by these stimuli. Pretreatment of cells with okadaic acid (1 M), a phosphatase inhibitor, did not augment iloprost-induced IP phosphorylation or induce a component of IP phosphorylation susceptible to H89 inhibition (data not shown). DISCUSSION The results of this study demonstrate for the first time that the hIP undergoes rapid phosphorylation in response to stimulation by agonist. These experiments indicate, perhaps surprisingly, that such hIP phosphorylation is catalyzed primarily by PKC and not by PKA.
We have concentrated on the epitope-tagged IP. However, the presence of the tag does not appear to alter its affinity for ligand or the potency with which these ligands stimulate the receptor to couple with second messenger systems. Previous studies have described the ability of PGI 2 analogs to activate distinct signaling systems, particularly those involving cAMP and inositol phosphates. Mutational analysis of GPCRs has often implicated regions of the putative second and third intracellular loop and the cytoplasmic tail in receptor-G protein interactions (30 -32). Given that splice variation of the cytoplasmic tail has been described for other eicosanoid receptors (33)(34)(35), it has been speculated that such variation in the IP may account for its interaction with multiple G proteins. For example, variation in the EP3 receptor tail results in differential coupling of the receptor to G proteins in expression systems (34,35). However, splice variation in the IP has not been described. Furthermore, the single variant expressed in HEK 293 cells in the present study was capable of interacting with more than one signaling pathway, with apparent differential affinity for the G proteins involved. Interestingly, the expressed IP exhibited two states of affinity for agonist ligand. The functional implications of this observation are presently unknown. We have previously demonstrated that two affinity states of the hTP reflect forms of the receptor with varied preference for signaling pathways and, indeed, functional response, in platelets (36). This raises the possibility that the ligand affinity states of the IP reflect differential efficiency of coupling with cAMP versus inositol phosphate signaling pathways. Cotransfection experiments with the TP suggest that, at least in the case of this transfected eicosanoid receptor, its affinity for agonist may be modulated by the G protein with which it is overexpressed (37). However, similar coupling to cAMP and inositol phosphate production is described for these two signaling systems in cells that endogenously express the IP receptor (25,26), suggesting that this discrepancy in coupling may be an artifact of transfection systems.
Several lines of evidence support the contention that PKC, rather than PKA, plays the predominant role among second messenger kinases in regulating agonist-induced phosphorylation of the hIP. First, the concentration-effect relationships for IP phosphorylation by agonist and agonist-dependent increases in inositol phosphates (and, by implication, activation of PKC) are similar and contrast with the EC 50 for cAMP production in the presence or absence of a phosphodiesterase inhibitor. Second, stimulation of PKA by forskolin or dibutyryl cAMP does not result in IP phosphorylation. Conversely, H89, an inhibitor of PKA, does not alter basal or agonist-stimulated IP phosphorylation. H89 is reportedly a selective inhibitor of PKA (28). Although one must always view results obtained with such inhibitors with caution, the internal consistency of the results lends credence to the present interpretation. Although we have not demonstrated the potential for activation of PKA under the conditions used in our experiments, it would appear extremely likely that this is the case. Forskolin, which increased HAHEK intracellular cAMP by ϳ500-fold over basal level (data not shown) has been shown by others to activate PKA in whole cells at concentrations similar to those employed here (38), while PKA-mediated ␤ 1 -adrenoreceptor phosphorylation has previously been demonstrated in HEK 293 cells (10). It might be argued that cAMP is a short lived intracellular messenger and that phosphodiesterases may have degraded cAMP during the 10-min incubation time in the phosphorylation assay, thus leading us to underestimate cAMP-induced effects. However, this is unlikely, since both iloprost-induced and forskolin-induced cAMP production in HAHEK were maintained above basal over the 10-min period in cells treated with and without