Introduction

Top Summary Introduction Experimental Procedures Results Discussion References

Prostaglandin E₂ (PGE₂) is a ubiquitous autocoid that exerts a variety of physiological effects through interactions with specific cell-surface receptors. Molecular cloning has identified four subtypes of PGE₂ receptors, referred to as E-prostanoid (EP)₁, EP₂, EP₃, and EP₄ (Coleman et al., 1994). These receptors belong to the seven-transmembrane (TM) G protein-coupled receptor (GPCR) superfamily and are classified based on their ligand binding and signal transduction characteristics. Activation of the EP₁ receptor elicits elevation of intracellular calcium, the EP₃ receptor mediates inhibition of cAMP generation, and the EP₂ and EP₄ receptors mediate receptor-evoked increases in intracellular cAMP generation (Funk et al., 1993; Namba et al., 1993; Bastien et al., 1994; Regan et al., 1994).

The G_s-coupled EP₂ and EP₄ receptors are distinguished by their ligand selectivity and differential desensitization (Coleman et al., 1994; Nishigaki et al., 1996). Butaprost and AH13205 are selective EP₂ agonists, whereas AH23848B is a weak EP₄-selective antagonist. In addition, [³H]PGE₂ binds to the EP₄ receptor with at least 10-fold higher affinity than the EP₂ receptor. Structurally, the EP₄ receptor has a much longer carboxyl-terminal sequence than the EP₂ receptor and has been shown to undergo short-term agonist induced desensitization, which is absent in the EP₂ receptor (Nishigaki et al., 1996; Bastepe and Ashby, 1997). Northern blot analysis has revealed that the mRNA encoding the EP₄ receptor is highly expressed and widely distributed in the body, whereas the mRNA encoding the EP₂ receptor is expressed at lower levels (Honda et al., 1993; Bastien et al., 1994; Regan et al., 1994).

Extensive mutagenesis studies performed on biogenic amine binding GPCRs suggest that their ligand binding pocket is embedded in the membrane-spanning regions of these receptors (for review, see Savarese and Fraser, 1992). In contrast, studies on calcitonin, vasopressin, and neurokinin peptidergic receptors provide evidence that the extracellular domains are important for the structure and function of this class of peptide binding GPCRs (Fong et al., 1992a; Bergwitz et al., 1996; Howl and Wheatley, 1996). Although prostaglandins are small molecules like the biogenic amines, their receptors share the greatest sequence similarity to a subclass of peptide receptors that includes the vasopressin and gonadotropin-releasing hormone peptidergic receptors (Kolakowski, 1994). Thus, prostanoid receptor-ligand interactions may represent a new paradigm, in which extracellular sequences play a critical role in receptor structure and/or function for these small-molecule-binding receptors. Previous studies have identified residues within the TM regions (Arg in TMVII, Ser in TMVI; Negishi et al., 1995; Huang and Tai, 1996; Audoly and Breyer, 1997b) as well as conserved residues in the second extracellular loop critical for receptor ligand binding and signal transduction (Audoly and Breyer, 1997a; Stillman et al., 1998). The hypothesis of this study is that, as for the related peptidergic receptors, the extracellular regions of the prostanoid receptors are important for receptor structure and function. We explore this hypothesis by creating chimeras replacing EP₂ extracellular regions with the corresponding EP₄ residues. The findings presented here suggest that certain extracellular sequences play a role in receptor structure-function.

	Experimental Procedures

Top Summary Introduction Experimental Procedures Results Discussion References

Materials. The human EP₂ receptor cDNA was a gift from Dr. Daniel Gil (Allergan, Irvine, CA). The human EP₄ receptor cDNA was a gift from Dr. Mark Abramovitz (Merck-Frosst, Montreal, Canada). PGE₂, PGD₂, and PGF₂ were purchased from Cayman Chemical (Ann Arbor, MI). Butaprost free acid was a gift from Dr. Jilly Evans (Merck-Frosst). M&B28767 was a gift from Dr. M. P. L. Caton (Rhone Poulenc Ltd, Dagenham, Essex, United Kingdom). AH13205 was a gift from Dr. Robert Coleman (Glaxo Research Group, Greenford, Middlesex, United Kingdom). [³H]PGE₂ and ³⁵S protein-labeling mix were purchased from DuPont-New England Nuclear (Boston, MA). Lipofectamine and Optimem were purchased from Life Technologies (Grand Island, NY). The 12CA5 monoclonal antibody was purchased from Boehringer Mannheim (Indianapolis, IN).

Construction of the Hemagglutinin Epitope-Tagged EP₂ and EP₄ Receptors. The human EP₂ and EP₄ receptor expression plasmids have been previously described (Stillman et al., 1998). These constructs consist of the EP₂ or EP₄ cDNA-coding region, containing no flanking 5' or 3' untranslated regions, in the expression vector pCDNA3 (InVitrogen, San Diego, CA). A DNA sequence coding for a nine-amino acid hemagglutinin (HA) tag (YPYDVPDYA) was fused directly to the start codon. Previous studies indicate that these tagged receptor constructs function identically to the nontagged EP₂ or EP₄ receptor (Stillman et al., 1998). Throughout this paper, the HA-tagged receptor fusion constructs are referred to as the "wild-type" receptor.

Site-Directed Mutagenesis of Receptor cDNAs. Mutant receptors (Table 1) were constructed with a polymerase chain reaction (PCR) method as described previously (Higuchi, 1989). The following chimeric oligonucleotides were used (EP₄ sequences are in bold; silent mutations to add or remove diagnostic restriction sites for screening purposes are underlined):

EP Receptor Expression in Cell Culture. COS1 cells were transiently transfected with pCDNA3 plasmids containing either wild-type or mutant EP receptor cDNAs by the lipofectamine method according to the manufacturer's instructions (Life Science Technologies Inc., Grand Island, NY) with 12 µg of plasmid DNA and 45-µl lipofectamine solution. Cells were cultured for 72 h, and 5 mM sodium butyrate was added to culture medium 16 h before lysis. Total cell membranes were prepared as described previously (Breyer et al., 1994).

Ligand Binding Assays. For saturation binding isotherm experiments, 15 µg or 20 µg of membrane protein was incubated with varying [³H]PGE₂ concentrations, and reactions were stopped by filtration onto glass fiber filters as described previously (Breyer et al., 1994). For competition binding assays, 20 µg of membrane protein was incubated with 1 to 2 nM [³H]PGE₂ and varying concentrations of unlabeled competitors.

Immunoprecipitation. COS1 cells transfected with the wild-type or mutant EP₂ cDNAs were cultured for 72 h. Immunoprecipitation was performed as described previously (Stillman et al., 1998) with an anti-HA monoclonal antibody (12CA5). Immunoprecipitated proteins were resolved on a 10% polyacrylamide gel by SDS-polyacrylamide gel electrophoresis, and proteins were visualized by autoradiography.

Cell-Surface Enzyme-Linked Immunosorbent Assay (ELISA). An indirect cellular ELISA, based on a protocol from Schoneberg et al., (1995) was used to quantify the amount of receptors present on the plasma membrane. COS1 cells were transiently transfected with the expression plasmids encoding EP₂ wild-type or mutant receptors and were plated into 96-well plates. After 72 h, the cells were fixed with 4% paraformaldehyde in PBS for 30 min. In some cases, cells were permeabilized with 0.2% Triton X-100 PBS. The cells were then blocked for 30 min with Dulbecco's modified Eagle's medium/10% FBS. The anti-HA monoclonal antibody (12CA5) was diluted 1:100 in culture medium and added to the cells for 2 h at 37°C. After four washes with PBS, the cells were incubated with a donkey antimouse horseradish peroxidase secondary antibody diluted 1:5000 in culture medium. Cells were washed four times in PBS, and then 100 µl of tetramethylbenzidine substrate (Sigma Chemical Co., St. Louis, MO) was added for 20 min. Reactions were stopped by adding 100 µl of 1 M phosphoric acid, and then absorbances at 450 nm were determined on a microtiter plate spectrophotometer.

cAMP Measurements. COS1 cells transiently transfected with expression plasmids encoding the wild-type or mutant EP₂ receptors were distributed into 24-well plates. The medium was replaced 24 h later with 450 µl of Dulbecco's modified Eagle's medium/20 mM HEPES/0.25 mM 3 isobutyl-1-methylxanthine/40 µM indomethacin and incubated for 1 h at 37°C. Medium containing varying amounts of PGE₂ or butaprost free acid was added to each well and incubated for 5 min. The reactions were stopped by the addition of 500 µl of 10% trichloroacetic acid. cAMP measurements of the cell lysates were performed by an enzyme immunoassay kit, according to manufacturer's instructions (Stratagene, La Jolla, CA).

Data Analysis. All binding assays and cAMP measurements were analyzed with PRISM software (GraphPad, San Diego, CA). Statistical analysis was performed with Instat software (GraphPad). K_i values were calculated with the method of Cheng and Prusoff (1973).

	Results

Top Summary Introduction Experimental Procedures Results Discussion References

Exchange of Various EC Sequences Has Differential Effects on Ligand Binding. The putative extracellular sequences of the EP₂ receptor each were replaced by the corresponding EP₄ receptor sequence (Table 1). When transiently expressed in COS1 cells, the third extracellular loop chimera (EP_2/4ECIII) demonstrated specific [³H]PGE₂ binding, with a K_d value of 6.0 ± 2.2 nM. Chimeric receptors exchanging the amino terminus (EP_2/4NT), first extracellular loop (EP_2/4ECI), and second extracellular loop (EP_2/4ECII), demonstrated no [³H]PGE₂-specific binding (Fig. 1A).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Saturation isotherms of the binding of [3H]PGE2 by COS1 cell transfectants expressing the chimeric EP2 receptors. The data shown here is from a single experiment performed in duplicate and is representative of three to four independent experiments. A: , EP2; , EP4; , EP2/4NT; , EP2/4ECI; , EP2/4ECII; , EP2/4ECIII. B: , EP2/4ECII-; , EP2/4ECII-C. C: , EP2; , EP2/4NT-ECI; , EP2/4ECII-III; *, EP2/4NT-ECIII.

All Chimeric Receptors are Expressed in COS-1 Cells at Wild-Type Levels. To determine whether chimeric receptors were expressed at levels similar to the wild-type EP₂ receptors, each of the receptor proteins was immunoprecipitated from COS-1 cells with a monoclonal antibody to the HA tag. Proteins were detected near the predicted molecular mass for each of the receptor constructs (Fig. 2). The presence of multiple bands clustered near this size may be caused by variable post-translational modification, such as glycosylation or phosphorylation of the receptors.

View larger version (35K): [in this window] [in a new window]   Fig. 2.   Immunoprecipitation of EP2 wild-type and chimeric receptor constructs. Metabolically labeled wild-type and mutant receptors expressed in COS1 cells were immunoprecipitated with a monoclonal antibody (12CA5) recognizing the HA epitope fused to the amino terminus of the receptors. This figure is representative of three independent experiments. Lane 1, vector-only transfected cells; lane 2, EP2 wild-type; lane 3, EP2/4NT; lane 4, EP2/4ECI; lane 5, EP2/4ECII; lane 6, EP2/4ECIII; lane 7, EP2/4ECII-N; lane 8, EP2/4ECII-C; lane 9, EP2/4NT-ECI; lane 10, EP2/4ECII-III; lane 11, EP2/4NT-ECIII.

ECII-C and ECIII Chimeras Successfully Traffic to the Cell Surface. The ability of the chimeric receptors to traffic to the cell surface may affect receptor function. Cell surface expression of chimeric receptors was assayed by a whole-cell ELISA technique (Schoneberg et al., 1995) with the 12CA5 anti-HA tag monoclonal antibody. The extracellular location of the amino-terminal HA tag renders it inaccessible to antibody unless the receptor is expressed on the cell surface. As a control for total receptor expression for a particular chimeric construct, cells were permeabilized with Triton X-100. The receptor constructs that were able to bind PGE₂ (wild-type, EP_2/4ECIII, EP_2/4ECII-C) were each detected on the cell surface at levels similar to that observed for the wild-type receptor. Of the chimeras with no detectable PGE₂ binding, EP_2/4ECI was detected on the surface at 25% of wild-type levels, whereas expression of the EP_2/4NT, EP_2/4ECII, and EP_2/4ECII-N chimeras was not detectable at all. ELISA performed in the presence of Triton X-100 detergent demonstrated that each of the receptor proteins was expressed at similar levels (Fig. 3) in agreement with immunoprecipitation experiments.

View larger version (35K): [in this window] [in a new window]   Fig. 3.   Cell surface ELISA of wild-type and chimeric EP2 receptors transfected into COS1 cells. In the absence of Triton X-100 (A), only those receptors expressed on the cell surface were detected, whereas in the presence of 0.2% Triton X-100 (B), surface and internal receptors were detected. Absorbance values are displayed as a percentage of the A450 value observed for the wild-type EP2 receptor, with each data point in triplicate. For EP2 wild-type, 100% A450 = 0.733 (+ triton) and 0.402 ( triton). Data were averaged from three independent experiments on each receptor construct (*p < .05). N-term, amino terminus.

Receptor Chimeras Possess Specificity of the EP₂ Receptor. The ligand-binding selectivity of the chimeras that displayed [³H]PGE₂ binding was examined in competition binding assays with the naturally occurring prostaglandins PGE₂, PGD₂, PGF₂, butaprost free acid (EP₂ selective), AH13205 (EP₂ selective), and M&B28767 (EP_3/4 selective) compounds. Each chimeric receptor possessed ligand selectivity indistinguishable from that of the wild-type EP₂ receptor for all drugs tested, with none of the K_i values for the chimeric receptors statistically different from the EP₂ wild-type values (Table 2).

Only Chimeras Expressed on the Cell Surface Regulate cAMP Production. The ability of PGE₂ and the EP₂-selective agonist butaprost free-acid to elicit receptor-evoked increases in intracellular cAMP levels was examined (Fig. 4; Table 3). Although the EP2/4ECIII and EP2/4ECII-C chimeras were able to stimulate cAMP production, the EP2/4NT, EP2/4ECI, EP2/4ECII, and EP2/4ECII-N chimeras demonstrated no receptor-evoked stimulation of cAMP above the vector-only transfected control cells (Fig. 4).

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Receptor-evoked stimulation of intracellular cAMP production in COS1 cells transfected with wild-type and chimeric EP2 receptors. Cells were incubated with various concentrations of PGE2, were lysed, and cell lysates were analyzed for cAMP content. Data shown is from a single experiment performed in duplicate and are representative of three independent experiments. A, , vector only; , EP2 wild-type; , EP2/4NT; , EP2/4ECI; , EP2/4ECII; , EP2/4ECIII. B, , vector only; [sqlo], EP2 wild-type; , EP2/4ECII-N; , EP2/4ECII-C. C, , vector only; , EP2 wild-type; , EP2/4NT-ECI; , EP2/4ECII-III; *, EP2/4NT-ECIII. In B and C, the data from A for EP2 wild-type () is replotted for comparison.

Chimeras with Multiple EC Exchanges Do Not Rescue Receptor Function. The loss of function of the chimeric receptors might be because of incompatibility of the EP₄ receptor EC sequences with the EC sequences of the of the EP₂ receptor. To investigate this possibility, multiple EC sequences of the EP₂ receptor were replaced with the corresponding EP₄ sequences. Three additional chimeric receptors were constructed that had either the first two EC sequences exchanged (EP_2/4NT-ECI), the last two EC sequences exchanged (EP_2/4ECII-ECIII), or all four EC sequences of the EP2 receptor replaced with the corresponding EP₄ sequences, (EP_2/4NT-ECIII; Table 1). Neither ligand binding nor signal transduction could be detected with any of these constructs (Figs. 1C and 4C). As with the other nonfunctional chimeras, although receptor expression could be detected by immunoprecipitation (Fig. 2), receptor expression could not be detected on the cell surface (Fig. 3).

	Discussion

Top Summary Introduction Experimental Procedures Results Discussion References

A phylogenetic tree of the GPCR superfamily has revealed that the prostanoid receptor family is most homologous to a subfamily of peptide receptors that includes the vasopressin receptors (Fig. 5; Kolakowski, 1994). In light of this observed homology, it is of interest that the EC sequences of the EP receptors are critical determinants of receptor structure/function, as has been observed for the peptidergic receptors. Replacement of the amino terminus, ECI, or the conserved sequence motif of ECII of the EP₂ protein resulted in receptors that were unable to bind [³H]PGE₂ or stimulate intracellular cAMP generation. Moreover, replacement of multiple EP₂ EC sequences with the corresponding EP₄ sequences did not rescue receptor function, which suggests that the loss function is not simply because of incompatibility among the EP receptor loops. One possible explanation for the observed phenotypes is that EC sequences form part of the ligand binding surface and, as observed for the peptidergic receptors, mutation of these putative contact surfaces contributes to the loss of ligand binding and signal transduction. However, we found that each of the nonfunctional receptors could not be detected on the plasma membrane as determined by ELISA. It is possible that each of the chimeric receptors is unfolded and this precludes their trafficking to the plasma membrane. Alternatively, these chimeras could represent correctly folded receptors with specific trafficking defects, and this expression in an inappropriate cellular compartment allows neither ligand binding nor participation in ligand-elicited signal transduction. The EP_2/4ECI chimera was expressed on the cell-surface at 25% of the level of the EP₂ wild-type receptor, although [³H]PGE₂ binding was not observed. Detection of EP_2/4ECI on the cell surface suggests that, at least for the EP_2/4ECI chimera, altered receptor trafficking alone does not cause the loss of binding and receptor-evoked signaling.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   A partial phylogenetic tree of the GPCR superfamily. Prostaglandin receptors lie in a sub-branch of peptide receptors, Group V, whereas the prototypical small-ligand GPCRs are in Group II (modified from Table 2 of Kolakowski, 1994).

In the absence of structural data, it cannot be determined if the EC domains are directly involved in prostanoid binding. Previous characterization of EC mutations have not assessed, in parallel, surface expression of mutant receptors. It had been shown that mutation of residues within a conserved motif of ECII of prostaglandin E₂ receptors results in receptors with altered EP₃ receptor ligand selectivity (Audoly and Breyer, 1997a) or loss of receptor function for the EP₂ and EP₄ receptors (Stillman et al., 1998). Similarly, other point mutations in ECI or ECII of the human EP₂ receptor (Kedzie et al., 1998) or the related thromboxane receptor (D'Angelo et al., 1996) resulted in a loss of ligand binding and/or signal transduction. Taken together with the current study, these data demonstrate that the amino terminus, ECI, and ECII sequences are critical determinants of prostaglandin receptor structure and/or function. It is possible that the role of the conserved motif across the prostanoid receptor family in the EC sequences is to ensure proper receptor folding, and the phenotypes of EC mutants reported in previous studies are caused by improper folding and/or trafficking.

Overall, the phenotypes observed for the EP_2/4 chimeric receptors are in sharp contrast to those observed for other small-ligand GPCRs. Mutation of the extracellular domains of GPCRs for small ligands have generally not had effects on ligand binding or receptor structure. Dixon et al. (1987) demonstrated that deletion of the extracellular domains of the -adrenergic receptor had no effect on receptor function. Studies with chimeric receptors of small-ligand GPCR subtypes have not revealed binding determinants in the extracellular domains (Frielle et al., 1988; Kobilka et al., 1988; Robinson and Caron, 1996). Replacement of the extracellular domains of the ₂-adrenergic receptor with the corresponding _1a-adrenergic receptor sequences had no effect on the ligand binding selectivity of this receptor (Zhao et al., 1998). Much of the mutagenesis and biochemical evidence obtained so far suggest that the ligand binding pocket of these small molecule-binding GPCRs is embedded within the transmembrane-spanning helices (Savarese and Fraser, 1992). Moreover, these data suggest that the extracellular sequences of the biogenic amine receptors are not critical determinants of receptor folding or trafficking as observed in the present study for the prostanoid receptors. In contrast, studies of peptide GPCRs have supported a role for the extracellular domains in ligand binding and receptor structure. The vasopressin V_1a and V₂ receptors are 20% identical with the EP₂ receptor. Howl et al. demonstrated that the amino terminus, ECI, and ECII of the V_1a receptor participate in the formation of the ligand binding pocket (Howl and Wheatley, 1996), and in the V₂ receptor, a point mutation in ECII eliminated vasopressin binding and signal transduction (Pan et al., 1994). Mutagenesis studies have revealed principal binding site determinants in the EC domains of other peptidergic GPCRs such as the angiotensin (Hjorth et al., 1994), thrombin (Gerszten et al., 1994), and neurokinin receptors (Fong et al., 1992a,b; Yokota et al., 1992; Huang et al., 1994).

Based on the accumulation of mutagenesis data, it is now evident that the GPCR superfamily contains receptor subgroups possessing distinct motifs of receptor-ligand interactions. For example, ligands can bind to the TM core (e.g., small ligand receptors), to both the TM core and the EC (peptide receptors), or to only the amino-terminal domain (metabotropic glutamate receptors; Ji et al., 1998). Prostaglandin receptors are currently classified with the receptors for small ligands, which include the biogenic amine and nucleotide receptors (Ji et al., 1998). Although data presented here do not demonstrate direct interaction of the EC sequences with the prostaglandin ligand, taken together with the prostanoid receptor phylogeny, they suggest that these receptors share structural requirements in the EC sequences similar to the peptidergic GPCRs.