Introduction

Top Summary Introduction Procedures Results Discussion References

Signal transduction cascades involving seven-transmembrane element, GPCRs require the obligate activation of heterotrimeric guanine nucleotide binding proteins (G proteins) (Birnbaumer et al., 1990; Bourne et al., 1990) Interactions between GPCRs and G proteins have been assessed and measured by a variety of means. Among the most widespread and popular are assays that examine GPCR-induced exchange of GDP for GTP in the nucleotide binding pocket of the G protein and the subsequent timer-controlled hydrolysis (Gierschik et al., 1994) of the terminal phosphate of GTP by the GTPase activity intrinsic to the  subunits of all G proteins (Milligan, 1988) or the binding of a poorly hydrolysed analogue of GTP such as [³⁵S]GTPS (Wieland and Jakobs, 1994). Such approaches have been particularly useful for GPCRs that interact predominantly with the subfamily of pertussis toxin-sensitive G_i-like G proteins (Milligan, 1988; Gierschik et al., 1994; Wieland and Jakobs, 1994). This reflects a combination of the higher intrinsic guanine nucleotide exchange and GTPase activity of these G proteins and their relatively high levels of expression compared with other G proteins. As such, although agonist-induced turnover of GTP is accepted to occur on other G proteins, direct demonstration of this often is difficult in membrane systems (Koski and Klee, 1981). This occurs despite some of the earliest evidence for a obligate role for GTP hydrolysis in signal transmission and termination being derived from analysis of -adrenoceptor activation of G_s in avian erythrocytes (Cassel and Selinger, 1976, 1977).

Prostaglandins and thromboxanes are locally acting mediators that are produced from arachidonic acid by the action of cyclooxygenases. Five primary active prostanoid metabolites, the prostaglandins D₂, E₂, F₂, and I₂ and thromboxane A₂, function via a family of GPCRs to produce a range of physiological actions. Receptors from all of these groupings have been cloned (Pierce et al., 1995), including the human IP prostanoid receptor (Boie et al., 1994). Agonist activation of the IP prostanoid receptor results in stimulation of adenylyl cyclase via activation of the G protein G_s (McKenzie and Milligan, 1990b). Although not studied as extensively as certain other G_s-linked adenylyl cyclase stimulatory receptors, such as the -adrenoceptors, the IP prostanoid receptor is of major potential therapeutic significance. Indeed, the recent production of mice lacking the IP prostanoid receptor (Murata et al., 1997) has confirmed its central role in the regulation of platelet aggregation and vasodilation and hinted at contributions to the regulation of inflammation and pain perception. Furthermore, its endogenous expression by platelets (Armstrong, 1996) and NG108-15 neuroblastoma X glioma hybrid (Kelly et al., 1990; McKenzie and Milligan, 1990b; Williams and Kelly, 1994) and other related cell lines (Kelly et al., 1990) resulted in substantial characterization of its regulation and desensitization before the availability of the receptor cDNA.

Chimeric G protein  subunits have been generated that allowed certain agonist activated GPCRs to switch output direction. For example, the alteration of as few as the last three amino acids of G_q with those from G_i1/G_i2 can allow certain GPCRs that normally mediate inhibition of adenylyl cyclase to stimulate phosphoinositidase C (Conklin et al., 1993). A similar chimeric strategy involving substitution of the carboxyl-terminal amino acids of G₁₃ with those from G_z has been used to define a role for G₁₃ in Na⁺-H⁺ exchange (Voyno-Yasenetskaya et al., 1994).

Equally, conversion of the carboxyl-terminal five amino acids of G_s to that of G_i1/G_i2 has been shown to allow the somatostatin SSTR3 receptor to mediate stimulation of cAMP production, although this receptor normally is considered to mediate inhibition of adenylyl cyclase activity (Komatsuzaki et al., 1997). Generic G protein  subunit carboxyl-terminal switching also has been applied to attempt to define the specificity of GPCR/G protein interactions (Liu et al., 1995; Komatsuzaki et al., 1997).

In the current study, we adapted this basic approach to generate a chimeric G protein  subunit (G_i1/G_s6), which has the backbone of G_i1 except that the carboxyl-terminal six amino acids have been converted to those of G_s. We predicted that this G protein should allow predominantly G_s-coupled receptors to produce robust agonist-dependent regulation of high affinity GTPase activity and binding of [³⁵S]GTPS. Using transient transfection of the chimeric G protein into a cell line stably expressing a FLAG-tagged form of the human IP prostanoid receptor, we demonstrate the usefulness of this system and compare it with the effectiveness of the authentic G proteins.

	Experimental Procedures

Top Summary Introduction Procedures Results Discussion References

Materials. All materials for tissue culture were supplied by Life Technologies Europe (Paisley, Scotland, UK). The IP prostanoid receptor agonist [³H]iloprost (15.3 Ci/mmol) was purchased from Amersham International (Buckinghamshire, UK). [-³²P]GTP (30 Ci/mmol) was obtained from DuPont-New England Nuclear (Boston, MA). [³⁵S]GTPS (1250 Ci/mmol) was purchased from DuPont-New England Nuclear. Pertussis toxin (240 µg/ml) and all other chemicals were from Sigma Chemical (Poole, Dorset, UK) and Boehringer-Mannheim (Mannheim, Germany). Oligonucleotides were purchased from Genosys (Cambridge, UK).

Construction of a FLAG epitope-tagged human IP prostanoid receptor cDNA. An hIPR cDNA was obtained from Dr. Mark Abramovitz (Department of Biochemistry and Molecular Biology, Merck Frosst Center for Therapeutic Research, Quebec, Canada). To incorporate the FLAG epitope (Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys) at the amino terminus of the cDNA, a set of PCR oligonucleotide primers was used. The sense oligonucleotide was 5'-AAGGATCCGCCACCATG(GACTACAAGGACGACGATGATAAG)GCGGATTCGTGCAGGAACC-3'; the underlined bases refer to restriction sites for BamHI and NcoI, respectively, with FLAG epitope bases in parenthesis. The antisense oligonucleotide was 5'-AAGAATTCTCAGCTTGAAATG(TCA)GCAGAG-3'; the underlined bases refer to EcoRI restriction site, with the stop codon in parenthesis. The PCR-amplified fragment was digested with BamHI and EcoRI and ligated to pcDNA3 (InVitrogen, San Diego, CA) through these restriction sites. Introduction of the NcoI site at the start codon allowed the selection of positive clones on NcoI digestion and agarose gel electrophoresis. The cDNA construct was fully sequenced.

Construction of G_i1/G_s6 cDNA. G_i1/G_s6 was generated from a pertussis toxin-resistant form of rat G_i1 in which the Cys351 was mutated to aspartic acid rat G_i1 by PCR amplification using the primers 5'-ACGT-GAA-TTC-GCC-ACC-ATG-GGC-TGC-ACA-CTG-AGC-GC-3' (primer 1) and 5'-CCA-CGT-GAA-TTC-TTA-TAA-GAG-TTC-ATA-TTG-CCT-TAG-GTT-ATT-CTT-TAT-3' (primer 2).

Cell culture and stable transfection. HEK 293 cells were maintained in Dulbecco's modified Eagle's medium containing 10% (v/v) newborn calf serum and 2 mM L-glutamine. Cells were seeded into 100-mm culture dishes and grown to 60-80% confluency (18-24 hr) before transfection with 5 µg of FhIPR cDNA using DOTAP reagent (Boehringer-Mannheim). At 48 hr after transfection, the cells were split 1:4 into 800 µg/ml geneticin (G418; Calbiochem, San Diego, CA)-containing medium. A 100-mm dish of untransfected HEK 293 cells also was split into the same medium as a control dish.

Transient transfections. Transient transfections into HEK 293 and cells stably expressing the FhIPR (clone 13) were essentially similar to that described. DNA at 5 µg/100-mm culture dish was used. At 48-72 hr after transfection, cells were rinsed and harvested in PBS (0.27 mM KCl, 0.15 mM KH₂PO₄, 137 mM NaCl, 4 mM Na₂HPO₄, pH 7.5), followed by centrifugation at 1000 × g in a refrigerated centrifuge. The cell pastes obtained were kept at 80° until ready for membrane preparation.

Preparation of membranes. Plasma membrane-containing P2 particulate fractions were prepared from cell pastes that had been stored at 80° after harvesting. Cell pellets were resuspended in Tris/EDTA buffer (10 mM Tris·HCl, pH 7.5, 0.1 mM EDTA), and rupture of the cells was achieved with 25 strokes of a hand-held Teflon-on-glass homogenizer. Unbroken cells and nuclei were removed by centrifugation at low speed (2,000 rpm) in a refrigerated microcentrifuge. The supernatant fraction then was centrifuged at 75,000 rpm for 30 min in a Beckman Instruments (Palo Alto, CA) Optima TLX Ultracentrifuge with a TLA100.2 rotor. The pellets were resuspended in Tris/EDTA buffer to a final protein concentration of 1-3 mg/ml and stored at 80° until required.

[³H]Iloprost binding studies. Binding assays were initiated by the addition of 20 µg of membrane protein to an assay buffer (50 mM Tris·HCl, pH 7.5, 5 mM MgCl₂) containing [³H]iloprost (0.1-50 nM in saturation assays and 10 nM in competition assays). Nonspecific binding was determined in the presence of 10 µM unlabeled iloprost. Reactions were incubated for 30 min at 30°, and bound ligand was separated from free ligand by vacuum filtration through GF/C filters. The filters were washed three times with ice-cold wash buffer (50 mM Tris·HCl, pH 7.5, 0.25 mM EDTA), and bound ligand was determined by liquid scintillation counting.

Immunological studies. M5 anti-FLAG antibody (IB13091; Kodak IBI) was used to blot for the FLAG epitope. It was raised in mouse and specifically recognizes the amino-terminal FLAG (MDYKDDDDK) epitope. Antiserum CS was produced in a New Zealand White rabbit using conjugates of synthetic peptide (RMHLRQYELL) and keyhole-limpet hemocyanin (Calbiochem) as antigen. It was used to blot for both authentic G_s and G_i1/G_s6 because it recognizes the carboxyl-terminal decapeptide of G_s (both the short and long isoforms) (Milligan and Unson, 1989). Antisera I1C and SG were produced in similar fashion using peptides LDRIAQPNYI (I1C), which corresponds to amino acids 159-168 of G_i1, and KENLKDCGLF (SG), which corresponds to amino acids 341-350 of transducin 1 and differs by a single amino acid from the carboxyl-terminal decapeptide of G_i1 and G_i2 (Green et al., 1990).

[³⁵S]GTPS binding studies. [³⁵S]GTPS binding studies were performed according to Wieland and Jakobs (1994). Briefly, 25 µg of membranes was incubated at 25° for 60 min in a final assay volume of 100 µl. The assay mix consisted of 20 mM HEPES, pH 7.4, 5 mM MgCl₂, 100 mM NaCl, and 5 µM GDP. [³⁵S]GTPS was used at 50 nCi/assay tube, giving a final assay concentration of 0.3-0.5 nM. Agonist-driven binding of [³⁵S]GTPS was determined in the presence of 1 µM iloprost, whereas nonspecific binding was performed in the presence of 20 µM unlabeled GTPS. The binding was stopped by the addition of 2.5 ml of ice-cold wash buffer (20 mM HEPES, pH 7.4, 5 mM MgCl₂), filtered through Whatman GF/C filters, and followed by three quick washes of 5 ml of wash buffer. Filters were air-dried; 5 ml of ULTIMA-Gold XR scintillant was added and left overnight before liquid scintillation counting.

Whole-cell adenylate cyclase assays. Whole-cell adenylate cyclase assays were performed as described by Merkouris et al. (1997). Clone 13 cells were split into the wells of a 24-well plate and incubated in medium containing [³H]adenine at 0.5 µCi/well for 16-24 hr. The generation of [³H]cAMP in response to iloprost or forskolin was assessed. Results are expressed as the ratio of [³H]cAMP to total [³H]adenine nucleotides (×100).

	Results

Top Summary Introduction Procedures Results Discussion References

The hIPR was amino-terminally modified by the addition of a FLAG (Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys) epitope. The modified IP prostanoid receptor cDNA (FhIPR) was expressed stably in HEK 293 cells. A number of individual antibiotic resistant clones were expanded and analyzed. Specific binding of [³H]iloprost was observed in membranes derived from a number of the clones (data not shown), whereas no detectable specific binding was observed in the membranes of parental HEK 293 cells. Clone 13 expressed the highest levels of specific [³H]iloprost binding. The specific binding of [³H]iloprost (5 nM) to membranes of clone 13 cells could be competed for with high affinity by unlabeled iloprost, under the assay conditions used, in an apparently monophasic manner (n_H = 1.00 ± 0.12) (Fig. 1a) with an IC₅₀ value of 9.5 ± 1.1 nM. Applying the formalisms of DeBlasi et al. (1989), this would correspond to a K_d value for [³H]iloprost of 4.5 nM under these assay conditions. Immunoblotting membranes of clone 13 cells with the anti-FLAG monoclonal antibody M5 confirmed expression of a FLAG-tagged protein (Fig. 1b), which was not present in parental HEK 293 cells. The predominant FLAG antibody reactive species migrated through SDS-PAGE with an apparent molecular mass of 45 kDa. There was some evidence for heterogeneity in that a somewhat more slowly migrating polypeptide also was observed (Fig. 1b). These may represent differentially glycosylated forms of the receptor.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Stable expression of the FLAG-tagged human IP prostanoid receptor in clones of HEK 293 cells. a, Competition for the specific binding of [3H]iloprost in membranes of clone 13 by unlabeled iloprost. The binding of [3H]iloprost (5 nM) to membranes of clone 13 cells was measured in the presence of varying concentrations of unlabeled iloprost. Data are taken from a representative experiment. Through the use of the formalisms of De Blasi et al. (1989), these data estimate a Kd value of 4.5 nM for iloprost in these assay conditions. b, Immunodetection of the FLAG-tagged human IP prostanoid receptor in clone 13 cells. Membranes from parental HEK 293 cells (1) or clone 13 cells (2) were resolved by SDS-PAGE, transferred to nitrocellulose, and immunoblotted using the anti-FLAG monoclonal antibody M5.

The functionality of the expressed FhIPR was demonstrated by measuring the ability of iloprost to stimulate cAMP production in intact cells. The capacity of forskolin to stimulate cAMP production was measured in parallel; based on a combination of levels of expression of the FhIPR and because iloprost and forskolin gave similar maximal levels of cAMP production, clone 13 was selected for detailed analysis. Concentration-response curves to iloprost indicated an EC₅₀ value of 1.5 ± 0.4 × 10¹⁰ M for stimulation of cAMP production (Fig. 2).

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Agonist occupation of the FLAG-tagged human IP prostanoid receptor results in stimulation of cAMP production. Basal cAMP levels and their regulation by varying concentrations of iloprost were measured in intact clone 13 cells. Data represent mean ± standard deviation of quadruplicate assays from a typical experiment. Forskolin (50 µM) produced a similar stimulation of adenylyl cyclase activity in clone 13 cells as maximally effective concentrations of iloprost (data not shown).

The addition of iloprost to membranes of clone 13 cells resulted in a small stimulation of high affinity GTPase activity (see below). However, although this increase in activity in response to the agonist was not observed in membranes of parental HEK 293 cells, it was too limited for detailed pharmacological analysis. Historically, it has been much easier to measure agonist stimulation of high affinity GTPase activity when studying GPCRs that interact with members of the G_i family of pertussis toxin-sensitive G proteins (Koski and Klee, 1981; Milligan, 1988; Gierschik et al., 1994), and, for example, we recently examined the interaction of the _2A-adrenoceptor and the  subunit of G_i1 (Wise et al., 1997) using this approach.

Because the extreme carboxyl terminus of G protein  subunits is a key element for GPCR activation of these transducing proteins (Conklin et al., 1993; Liu et al., 1995) and modification of as few as three amino acids from the carboxyl terminus of the  subunit of the phosphoinositidase C-linked G protein G_q has been reported to be sufficient to cause alteration in the nature of GPCRs able to activate these G proteins (Conklin et al., 1993), we used a PCR-based strategy to alter the last six amino acids of G_i1 to those of G_s (Fig. 3a). Confirmation of the success of this strategy was obtained both by direct sequencing of the modified G protein (G_i1/G_s6) cDNA and by a series of immunoblotting strategies. G_i1/G_s6 was expressed transiently in COS-7 cells, and membranes from each of these cells, mock transfected cells, and rat brain cortex (which expresses high levels of G_i1) were immunoblotted with either antiserum I1C, which was generated against a peptide corresponding to amino acids 159-168 of G_i1 and is specific for this G protein (Green et al., 1990), or with antiserum SG, which was generated against a peptide corresponding to the carboxyl-terminal 10 amino acids of G_i1 and which identifies both G_i1 and G_i2 (this epitope is identical in these two G proteins) (Green et al., 1990). When blotting with antiserum I1C, a low level of authentic G_i1 was detected as a 41-kDa protein in mock transfected COS-7 cells, as we reported previously (Galbiati et al., 1996). Successful expression of G_i1/G_s6 was recorded as a marked increase in I1C reactivity, which essentially comigrated with authentic rat brain G_i1 (Fig. 3b, top). In contrast, immunoblotting of the same membranes with antiserum SG failed to identify G_i1/G_s6, although a combination of endogenous G_i1 and G_i2 was detected in both mock and transfected COS-7 cells and in rat brain (Fig. 3b, bottom). Although these results clearly demonstrated the loss of carboxyl-terminal G_i1 immunoreactivity in G_i1/G_s6, the positive appearance of G_s carboxyl-terminal immunoreactivity in this protein was obtained by immunoblotting such transfected cells with antiserum CS, which was generated against a peptide corresponding to the carboxyl-terminal 10 amino acids of the splice variants of G_s. Now, in addition to authentic, endogenously expressed G_s, transient transfection with G_i1/G_s6 resulted in the appearance of a more rapidly migrating CS immunoreactive polypeptide that comigrated with G_i1 (Fig. 3c).

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Generation and expression of Gi1/Gs6. a, Carboxyl-terminal sequence of Gi1 and Gi1/Gs6. The carboxyl-terminal decapeptides of Gi1 and Gi1/Gs6 are shown using standard one-letter codes. Unlike Gi1, Gi1/Gs6 is not a substrate for pertussis toxin-catalyzed ADP-ribosylation because the marked C, which is the target for modification in Gi1, is absent from the chimeric G protein. b, Immunological characterization. Gi1/Gs6 has lost the carboxyl terminus of Gi1. COS-7 cells were mock transfected (1) or transiently transfected to express Gi1/Gs6 (2). Membranes from these cells or from rat brain cortex (3) were resolved by SDS-PAGE, transferred to nitrocellulose, and immunoblotted with antiserum I1C (internal epitope of Gi1) (top) or with antiserum SG (carboxyl terminus of Gi1 and Gi2) (bottom). c, Gi1/Gs6 has an internal epitope derived from Gi1 but the carboxyl-terminal epitope derived from Gs. COS-7 cells were mock transfected (1) or transiently transfected to express Gi1/Gs6 (2). Membranes prepared from these cells were resolved by SDS-PAGE and immunoblotted with antiserum CS (carboxyl-terminal epitope of Gs) or I1C (internal epitope of Gi1). Mock and positively transfected cells endogenously express Gs (upper immunoreactive bands, left), whereas only the positively transfected cells display a Gs-immunoreactive polypeptide (Gi1/Gs6), which comigrates with Gi1.

Clone 13 cells subsequently were either mock transfected or transiently transfected with G_i1/G_s6, G_i1, or G_s. Membranes were prepared, and basal and iloprost stimulation of high affinity GTPase activity was measured. As noted, a small stimulation over the basal high affinity GTPase activity was observed in mock transfected clone 13 cells (Fig. 4). This was markedly increased in cells transiently transfected with G_i1/G_s6 but not after transfection with either G_i1 or G_s (Fig. 4). In contrast, transient transfection of parental HEK 293 cells with G_i1/G_s6 did not result in any measurable iloprost stimulation of high affinity GTPase activity (Fig. 4).

View larger version (46K): [in this window] [in a new window]   Fig. 4.   Expression of Gi1/Gs6 with the FLAG-tagged human IP prostanoid receptor showing the effects on iloprost stimulation of high affinity GTPase activity. Clone 13 cells were either mock transfected (1) or transiently transfected to express Gs (2), Gi1 (3), or Gi1/Gs6 (4). Parental HEK 293 cells also were transiently transfected to express Gi1/Gs6 (5). Membranes from these cells then were used to measure basal and iloprost (1 µM)-stimulated high affinity GTPase activity. The stimulation above basal produced by iloprost is shown. The experiment displayed is representative of four performed on membranes derived from separate transient transfections.

We wanted to assess whether the increased agonist-stimulated high affinity GTPase activity produced by transfection with G_i1/G_s6, but not with authentic G_i1, simply reflected substantially greater levels of expression of G_i1/G_s6. However, immunoblotting membranes of clone 13 with antiserum I1C after transient transfection with either of the two G proteins demonstrated these to be expressed to equivalent levels (Fig. 5).

View larger version (32K): [in this window] [in a new window]   Fig. 5.   The enhanced stimulation of high affinity GTPase by iloprost after transfection of Gi1/Gs6 compared with Gi1 is not related to the level of expression. Membranes of clone 13 cells transfected as in Fig. 4 to express Gi1 (1), Gi1/Gs6 (2), or Gs (3) were resolved by SDS-PAGE and immunoblotted with antiserum I1C. Gi1 and Gi1/Gs6 are identical in the region identified by this antiserum and displayed similar levels of expression.

It was possible that the iloprost stimulation of high affinity GTPase activity in membranes of clone 13 reflected an interaction of the FhIPR with endogenously expressed G_i-like proteins. However, the effect of iloprost was not prevented by pretreatment of the cells with pertussis toxin (25 ng/ml, 16 hr) (Fig. 6). In contrast, pretreatment of the cells with cholera toxin (200 ng/ml, 16 hr) essentially attenuated iloprost-stimulated high affinity GTPase activity (Fig. 6). Because cholera toxin-catalyzed ADP-ribosylation of G_s attenuates the GTPase activity of this G protein, such results indicated the signal to be derived from activation of endogenous G_s.

View larger version (57K): [in this window] [in a new window]   Fig. 6.   Cholera toxin pretreatment abolishes iloprost stimulation of high affinity GTPase activity in membranes of clone 13. Clone 13 cells were treated with vehicle (1) cholera toxin (200 ng/nl, 16 hr) (2), or pertussis toxin (25 ng/ml, 16 hr) (3) before harvest. Membranes from these cells then were used to measure basal high affinity GTPase activity and its stimulation by iloprost. Iloprost stimulation of activity was lost in the cholera toxin-treated cells. Data represent mean ± standard deviation of quadruplicate assays from a representative experiment.

An alternative assay for GPCR-mediated activation of G proteins is to measure agonist-induced binding of [³⁵S]GTPS. Basal specific binding of [³⁵S]GTPS was not increased by iloprost in the membranes of parental HEK 293 cells. As with the GTPase assay, iloprost did cause stimulation of [³⁵S]GTPS binding in membranes of clone 13 cells. However, again, although transient expression of either G_s or G_i1 into clone 13 cells did not result in any greater level of iloprost-mediated stimulation (see below), the binding of [³⁵S]GTPS in response to iloprost was increased substantially in the presence of G_i1/G_s6 (Fig. 7).

View larger version (69K): [in this window] [in a new window]   Fig. 7.   Coexpression of the FLAG-tagged human IP prostanoid receptor and Gi1/Gs6 promotes iloprost-stimulated binding of [35S]GTPS. Membranes were prepared from parental HEK 293 cells (1), clone 13 cells (2), and clone 13 cells that had been transiently transfected to express Gi1/Gs6 (3). Nonspecific, measured in the presence of 10 µM unlabeled GTPS (dots), basal (hatched), and 1 µM iloprost-stimulated (filled) binding of [35S]GTPS was measured as described in Experimental Procedures.

G_i1/G_s6 would be anticipated to be a substrate for neither cholera toxin nor pertussis toxin. Clone 13 cells were either mock transfected or transfected with G_s, G_i1, or G_i1/G_s6 and then treated for the last 16 hr before cell harvest with a combination of cholera toxin (200 ng/ml) and pertussis toxin (25 ng/ml). This treatment essentially attenuated iloprost-stimulation of the binding of [³⁵S]GTPS in each of mock transfected, G_s-transfected, and G_i1-transfected clone 13 cells (Fig. 8). In contrast, in cells transiently transfected with G_i1/G_s6, a substantial stimulation of binding of [³⁵S]GTPS by iloprost was still achieved (Fig. 8). This was not as great as in the absence of pretreatment with cholera and pertussis toxins; however, the degree of reduction in iloprost-stimulated binding of [³⁵S]GTPS herein was similar to the maximal effect of iloprost in mock-transfected clone 13 cells, suggesting that this fraction was derived from activation of the endogenous G_s pool.

View larger version (68K): [in this window] [in a new window]   Fig. 8.   The effects of cholera and pertussis toxin on iloprost-stimulated [35S]GTPS binding. Cells of clone 13 stably expressing the FhIPR were mock transfected (1) or transiently transfected to express Gi1 (2), Gs (3), or Gi1/Gs6 (4). For the last 16 hr, these cells were exposed to vehicle (dotted bars) or a combination of cholera toxin (200 ng/ml) and pertussis toxin (25 ng/ml) (lined bars). Membranes were prepared from the cells, and basal and iloprost (1 µM) stimulation of basal [35S]GTPS binding was assessed. The stimulation above basal produced by iloprost is shown.

Immunoblotting membranes from such transfected cells with antiserum I1C demonstrated expression of G_i1/G_s6 and G_i1 (Fig. 9, top). The small reduction in mobility of the transfected G_i1 through SDS-PAGE in the membranes from pertussis toxin-treated cells (Fig. 9, compare lanes 3 and 4) is indicative of toxin-catalyzed ADP-ribosylation of this protein. Very low levels of endogenous G_i1 could be detected in these immunoblots (Fig. 9, see lanes 5-8). Reduced mobility of endogenous G_i1 also was observed after pertussis toxin treatment. In contrast, the mobility of G_i1/G_s6 was unaffected by such treatment because it is not a substrate for pertussis toxin. As anticipated from previous results (Milligan et al., 1989), cholera toxin treatment of clone 13 cells resulted in a substantial reduction in total membrane levels of authentic G_s, even in cells transiently transfected to express excess G_s (Fig. 9, bottom, lanes 5 and 6). However, this treatment did not alter membrane levels of G_i1/G_s6 (Fig. 9, bottom, lanes 1 and 2) because it also is not a substrate for cholera toxin.

View larger version (41K): [in this window] [in a new window]   Fig. 9.   Sustained treatment of clone 13 cells with cholera and pertussis toxin down-regulates levels of Gs but not of Gi1 or Gi1/Gs6. Clone 13 cells were transfected to express either Gi1/Gs6 (1, 2), Gi1 (3, 4), or Gs (5, 6) or mock transfected (7, 8). During the final 16 hr, these cells were exposed to either a combination of cholera toxin and pertussis toxin, as in Fig. 7 (2, 4, 6, and 8), or to vehicle (1, 3, 5, and 7). Membranes prepared from these cells then were immunoblotted with antiserum I1C (top) or antiserum CS (bottom).

	Discussion

Top Summary Introduction Procedures Results Discussion References

Studies on GPCR activation of heterotrimeric G proteins are performed in a number of ways. Probably the two most widely applied methods are to measure agonist-enhanced binding of [³⁵S]GTPS (Wieland and Jakobs, 1994) or the hydrolysis of the terminal phosphate of [³²P]GTP (Gierschik et al., 1994). Despite the widespread use of such assays, they have been applied predominantly to GPCRs that couple selectively with members of the pertussis toxin-sensitive family of G_i-like G proteins. For other GPCRs, the effect of agonist often is small compared with the basal signal, resulting in poor specific signal-to-noise ratios. Systems that improve the response from such GPCRs would provide an attractive way to study agonist pharmacology.

Substantial evidence indicates that the extreme carboxyl terminus of a G protein  subunit is a key site for functional contacts with GPCRs. Pertussis toxin-catalyzed ADP-ribosylation of G_i-like G proteins at a conserved cysteine residue four amino acids from the carboxyl terminus of these  subunits (Milligan, 1988) prevents effective functional interactions between GPCRs and these G proteins. The unc mutation of G_s, which represents a single arginine-to-proline alteration six amino acids from the carboxyl terminus, disallows signal transduction from the ₂-adrenoceptor (Sullivan et al., 1987). Furthermore, anti-peptide antibodies directed toward this region of G proteins often are able to interfere with agonist activation of G proteins (McKenzie and Milligan, 1990a; Shenker et al., 1991). Given the high sequence, and thus probable structural, similarities between individual G protein  subunits, a large number of chimeras have been generated between different G proteins to determine elements of G protein function. Indeed, in recent times, detailed analyses of such chimeras has indicated that only small elements derived from the extreme carboxyl terminus of the G protein  subunit can be sufficient to define receptor specificity (Conklin et al., 1993, 1996; Liu et al., 1995; Gomeza et al., 1996; Ikezu et al., 1996; Komatsuzaki et al., 1997; Tsu et al., 1997). This information has resulted in the use of such chimeric G proteins to transduce signals from GPCRs to second messenger effector systems that may not be their usual ones. This can provide a means to use sensitive assay systems to examine the agonist pharmacology of GPCRs.

We used a chimeric G protein in which the last six amino acids of G_i1 were converted to those of G_s (G_i1/G_s6) to allow effective agonist stimulation of high affinity GTPase activity and [³⁵S]GTPS binding by the IP prostanoid receptor. Platelets and neuroblastoma X glioma hybrid NG108-15 cells have been two of the most studied systems that express this receptor. Although this GPCR is expressed to levels of >1 pmol/mg membrane protein in NG108-15 cells, agonists at this GPCR are unable to stimulate high affinity GTPase activity, although they strongly activate adenylyl cyclase. In contrast, agonists at the -opioid receptor, which is expressed in considerably lower levels in these cells, produce a robust GTPase response (Koski and Klee, 1981; McKenzie and Milligan 1990a).

A cDNA encoding the hIPR was FLAG-tagged and then stably expressed in HEK 293 cells. In clone 13, there were good levels of specific [³H]iloprost binding. Because this ligand is an agonist (and no radiolabeled antagonists are currently available), the measured levels of binding can represent only a best estimate. However, using a concentration of [³H]iloprost close to twice its estimated K_d value (Fig. 1a), ~2-3 pmol/mg of membrane protein of the GPCR was detected. In these membranes, an iloprost-mediated stimulation of high affinity GTPase could be measured. However, it represented a small incremental increase compared with the basal activity. This did, however, seem to represent activation of G_s because it was eliminated by pretreatment of the cells with cholera toxin. This conclusion reflects that this toxin eliminates the GTPase activity of G_s by catalyzing the ADP-ribosylation of a key arginine residue, a modification that also can lead to accelerated degradation of this G protein (Milligan et al., 1989). However, by contrast, iloprost-stimulated high affinity GTPase activity and binding of [³⁵S]GTPS were unaffected by pertussis toxin treatment, indicating the GPCR-stimulated GTPase activity did not result from weak activation of the endogenous pool of G_i-like G proteins.

Transient expression of G_i1/G_s6 into clone 13 cells allowed effective stimulation of both high affinity GTPase activity and binding of [³⁵S]GTPS on the addition of iloprost. In contrast, expression of G_i1 was unable to mimic these effects. The availability of an antiserum (I1C) that identifies an internal epitope present in both G_i1 and G_i1/G_s6 demonstrated clearly that these were expressed to equal levels and thus that although the FhIPR displayed no capacity to activate G_i1, interaction with the expressed G_i1/G_s6 could be converted into a simple guanine nucleotide exchange-based assay readout.

Cholera toxin treatment would be anticipated to prevent any GPCR-mediated stimulation of the GTPase activity of G_s, and pertussis toxin is known to prevent functional contacts between GPCRs and the G_i-like G proteins. Because G_i1/G_s6 would not be anticipated to act as a substrate for either pertussis or cholera toxin, pretreatment of clone 13 cells transiently transfected to express G_i1/G_s6 with a combination of both toxins defined that the agonist-induced signal must be generated by the chimeric G protein. Although as noted, in this system no evidence was obtained for an effective interaction of the FhIPR with endogenously expressed G_i-family G proteins and thus the combined cholera and pertussis toxin treatment was not required to isolate and define receptor activation of the chimeric G protein (Fig. 8), this is not always the case. In a number of examples, agonist at GPCRs that do not appear to mediate their primary effector function via activation of pertussis toxin-sensitive G proteins do catalyze measurable GTPase activity or stimulate the binding of [³⁵S]GTPS. In such studies, pertussis toxin pretreatment indicates much of the signal to have been derived from a minor interaction with G_i-like G proteins (Akam et al., 1997).

The current results demonstrate an effective interaction between the IP prostanoid receptor and the chimeric G_i1/G_s6 G protein. It remains to be established whether this chimeric G protein also will be effectively activated by other predominantly G_s-coupled GPCRs. If so, it should provide a sensitive means to analyze agonist pharmacology at such GPCRs using the type of assays that are widely used in both basic research and screening systems designed to detect novel agonists.