Cloning and Expression of a cDNA for Mouse Prostaglandin E Receptor EP3 Subtype*

A functional cDNA clone for mouse EP3 subtype of prostaglandin (PG) E receptor was isolated from a mouse cDNA library using polymerase chain reaction based on the sequence of the human thromboxane A2 receptor and cross-hybridization screening. The mouse EP3 receptor consists of 365 amino acid residues with putative seven-transmembrane domains. The sequence revealed significant homology to the human thromboxane A2 receptor. Ligand binding studies using mem- branes of COS cells transfected with the cDNA revealed specific [3H]PGE2 binding. The binding was displaced with unlabeled PGs in the order of PGE2 = PGEl > iloprost > PGF2, > PGD2. The EP3-selective agonists, M&B 28,767 or GR 6379912, potently com- peted for the ['H]PGE2 binding, but no competition was found with EP1- or EP2-selective ligands. PGEz and MQB 28,767 decreased forskolin-induced CAMP for-mation in a concentration-dependent manner in Chinese hamster ovary cells permanently expressing the cDNA. Northern blot analysis demonstrated that the EP3 mRNA is expressed abundantly in kidney, uterus, and mastocytoma P-815 cells and in a lesser amount in brain, thymus, lung, heart, stomach, and spleen.

I To whom correspondence should be addressed.
The abbreviations used are: PG, prostaglandin; G protein, heterotrimeric GTP-binding protein; ICso, drug concentration to inhibit response by 50%; PCR, polymerase chain reaction; TX, thromboxane; CHO, Chinese hamster ovary. receptors are pharmacologically subdivided into three subtypes, EP1, EPz, and EP3 (3, 4), and these subtypes are suggested to be different in their signal transduction; they are presumed coupled to stimulation of phospholipase C, and stimulation and inhibition of adenylate cyclase, respectively (4)(5)(6)(7)(8). Pharmacological actions of these subtypes have been well characterized, and the EP3 receptor has been suggested to be involved in inhibition of gastric acid secretion (7), modulation of neurotransmitter release in central and peripheral neurons (9), and inhibition of sodium and water reabsorption in kidney tubulus (8,10,11). In spite of this information, none of the receptors has been isolated, and their molecular characterization has been carried out only poorly. Recently we cloned a cDNA for the human TXAZ receptor (12). Based on its sequence we carried out PCR to amplify a homologous sequence from mouse cDNA and, using this fragment as a hybridization probe, performed homology screening. Analysis of nucleotide sequence and expression of the isolated clone revealed that it encodes the mouse EP3 receptor. We report here the complete nucleotide and deduced amino acid sequences of this receptor, and its ligand binding and biochemical properties analyzed in several mammalian expression systems. This study will be of help in understanding similarity and divergence of eicosanoid receptors.
Amplification of a Mouse cDNA Fragment Homologous to the Human TXA2 Receptor by PCR-First strand cDNA was synthesized from mouse lung total RNA by using random hexanucleotides as primers. PCR primers were designed based on the human TXA, receptor cDNA sequences corresponding to the putative third and sixth transmembrane domains of the receptor (12). Mouse lung cDNA served as template in 30 cycles of PCR with 1 min of denaturation at 95 "C, 0.5 min of annealing at 60 "C, and 1.5 min of extension at 72 "C on a Zymoreactor (Atto Corp., Tokyo, Japan). A single 418base pair cDNA fragment was amplified and subcloned into pBluescript SK(+) (Stratagene). A clone isolated (LT3) showed a cDNA.
sequence 78% homologous to the corresponding region of the human Molecular Cloning by Cross-hybridization-Mouse lung cDNA prepared by an oligo(dT) priming method was size-selected (>1.8 kilobases) and inserted into the EcoRI site of XZAPII DNA (Stratagene) with EcoRI adaptors (New England Biolabs, Inc.). The 1.9 X IO5 clones derived from the cDNA library were screened by hybridization with LT3. Hybridization was carried out at 58 "C in 6 X SSC (900 mM NaCl and 90 mM sodium citrate) containing 5 X Denhardt's solution (0.1% Ficoll, 0.1% polyvinylpyrrolidone, and 0.1% bovine serum albumin) and 0.5% sodium dodecyl sulfate, and filters were washed at 60 "C in 2 X SSC containing 1% sodium dodecyl sulfate. Among several clones hybridizing positively to LT3, we picked up one showing a signal apparently weaker than others and subjected it to further screening. Nucleotide sequencing of the isolated clone (ML64) revealed that it was a partial clone. Using this clone as a hybridization probe, we then screened the cDNA library of mouse mastocytoma P-815 cells for a full-length clone. From 7.2 X lo5 clones of the P-815 XZAPII library, nine clones were isolated and subjected to sequence analysis. Nucleotide sequencing was carried out on double-stranded templates using the dideoxy chain termination method. One clone (MP660) was a full-length clone with a 1,095-base pair open reading frame.
cDNA Expression in COS-1 Cells and Binding Assay-The EcoRI insert of MP660 was subcloned into pcDNAI (Invitrogen), a modified eukaryotic vector, and the resultant plasmid DNA was transfected into COS-1 cells by the DEAE-dextran method (13). After culture for 72 h, the cells were harvested, and crude membranes were prepared as described (12). Using these membranes, [3H]PGEz binding was determined, and protein was measured as described previously (14).
Stable Expression of the Cloned Receptor and CAMP Assay-cDNA transfection and cell line establishment were performed essentially as described by Nakajima et al. (15). The EcoRI fragment of MP660 was inserted into pdKCR-dhfr, a eukaryotic expression vector containing a mouse gene as a selection marker (16). The plasmids were transfected to CHO cells deficient in dihydrofolate reductase activity (CHO-dhfr-) (17) by the calcium phosphate method (18). Cell populations expressing the cDNA together with dihydrofolate reductase were selected in a-modification of Eagle's medium lacking ribonucleosides and deoxyribonucleosides and containing penicillin (100 units/ ml), streptomycin (100 pglml), and 10% dialyzed fetal bovine serum (Cell Culture Laboratories) (16). From these cell populations, clonal cell lines were isolated by single-cell cloning. Expression of the cDNA was assessed by RNA blotting. As a control, CHO cells were mocktransfected (transfected only with the vector) and isolated. These cells gave no signal on RNA blotting. Cyclic AMP levels in these cells were determined as described previously (19). Inositol phosphates were measured as described (15).
Northern Blots-Total RNAs from various mouse tissues were isolated by the acid guanidinium thiocyanate-phenol-chloroform method (20), and poly(A)+ RNAs were purified using Oligotex dT30 (Takara Shuzo, Kyoto, Japan). Poly(A)+ RNAs (10 pg) from each tissue were separated by electrophoresis on a 1.2% agarose gel, transferred onto nylon membranes (Hybond-N, Amersham Corp.), and hybridized with a 32P-labeled EcoRI/BamHI fragment of MP660 clone. Hybridization was carried out at 68 "C in 6 X SSC, and filters were washed at 68 "C in 1 X SSC. Fig. 1 shows nucleotide and deduced amino acid sequences of MP660. The amino acid sequence was assigned from the longest open reading frame of the cDNA. The nucleotide sequence surrounding the initiation codon agrees reasonably well with the consensus sequence (21). The polypeptide consists of 365 amino acid residues with an estimated molecular weight of 40,077. The hydropathicity profile determined by the Kyte and Doolittle method (22) and the sequence homology analysis indicated that it possesses seven hydrophobic segments and shares a significant sequence similarity with other members of G protein-coupled receptors (23) The deduced amino acid sequence is shown above the nucleotide sequence using single-letter code. Positions of the putative transmembrane segments I-VI1 are indicated by overlines above the amino acid sequence. The termini of each segment are tentatively assigned on the basis of a hydropathicity profile and comparison with other G protein-coupled receptors. Asterisks, potential Nglycosylation sites in the extracellular regions; stars, potential phosphorylation sites by CAMP-dependent protein kinase.

AGCGTGCAGTGGCCGGGCACGTGGTGCTTCATCAGCACCGGGCCGGCGGGCAACGAGACAGACCCTGCGCGCGAG S V Q W P G T W C F I S T G P A G N E T D P A R E P G S V A F A S A F A C L G L L A L V V T F A C N CCGGGCAGCGTGGCCTTTGCCTCCGCCTTCGCCTGCTTGGGCTTGCTGGCTCTGGTGGTGACCTTTGCCTGCAAC
IV v

ATAATGATGTTGAAAATGATCTTCAATCAGATGTCGGTTGAGCAATGCAAGACACAGATGGGAAAGGAGAAGGAG I M M L K M I F N Q M S V E Q C K T Q M G K E K E C N S F L I A V R L A S L N Q I L D P W V Y L L L TGCAATTCCTTTCTAATTGCAGTTCGCCTGGCTTCGCTGAACCAGATCTTGGATCCCTGGGTTTATCTGCTGCTA
w AGAAAGATCCTTCTTCGGAAGTTCTGCCAGATCAGAGACCACACCAACTATGCTTCCAGCTCCACCTCCTTGCCC glycosylation sites are seen at the amino-terminal and the second extracellular loop regions, and nine serine and threonine residues at the carboxyl-terminal region as possible phosphorylation sites.

R K I L L R K F C Q I R D H T N Y A S S S T S L
To identify a ligand for this receptor, MP660 was expressed in COS-1 cells, and membranes of the transfected cells were subjected to binding assays using various radioactive PGs. Among the PGs tested, [3H]PGEz specifically bound to the membranes. Scatchard analysis of this binding yielded a dissociation constant (Kd) of 2.9 nM, which agrees well with that previously reported on binding of [3H]PGEz to canine renal medullary membranes (26). The average density of binding sites in three experiments was 770 fmol/mg of protein of the transfected COS cell membranes. Specificity of this binding is shown in Fig. 2a. The binding of [3H]PGE2 was inhibited by unlabeled PGs in the order of PGEz = PGEl > iloprost, a PGIz analogue > PGF2, > PGD2. This characteristic of binding specificity was in good agreement with the PGE receptor previously characterized in various tissues (26,27). Because PGE receptor is pharmacologically subdivided into three receptor subtypes, EPI, EP2, and E P , with different agonist and antagonist profiles (4), we further characterized the specificity of this [3H]PGEz binding using ligands specific for PGE receptor subtypes. As shown in Fig. 26, among various PGE analogues, only EP3-specific agonists, GR 63799X and M&B 28,767, specifically competed for the [3H]PGEz binding with equal potency, and they were more potent than PGEz itself. On the other hand, no competition was found at all with either an EP1-specific antagonist, SC-19220, or an EPz-specific agonist, butaprost.
[3H]PGE2 did not bind to membranes of untransfected cells. These results established that MP660 encodes the EP3 subtype of PGE receptor.
Possible association of the EP3 receptor with inhibition of adenylate cyclase has been indicated (4). We tested this possibility by permanently expressing the cDNA in CHO cells and examining response of the cells to PGE analogues. As shown in Fig. 3, the transfected CHO cells showed a dosedependent decrease to PGE, in forskolin-induced cellular cAMP accumulation. M&B 28,767, an EP3-specific agonist, also inhibited forskolin-induced cAMP synthesis and was more potent than PGEz (ICso of M&B 28, 767 = 1 X 10"' M; ICs0 of PGEz = 1 X 10"' M). This potency of PGEz correlates well with that found in canine kidney and rat uterine membranes (11,28). Either agonist alone did not increase cAMP accumulation. We also examined phosphatidylinositol turnover in the transfected CHO cells. Addition of up to 1 PM M&B 28,767 revealed no significant increase in inositol phosphate content over the basal levels (data not shown). These results demonstrated that the EP3 receptor is coupled exclusively to inhibition of adenylate cyclase.
The mouse EP3 and human TXAz receptors (12) are significantly similar in size and show highly homologous amino acid sequences, especially in the putative seven-transmembrane segments except segments I and V (Fig. 4). The most highly conserved regions are segment VI1 and that spanning the latter half of segment IV to the first 12 amino acids in the second extracellular loop (from Leu-165 to Phe-184), 63.6 and 80.0% homology, respectively. As in the human TXAz receptor, there is no Asp in the third transmembrane segment of the EP3, a residue which is presumed to bind the amino group of ligands in the adrenergic receptors (29). Furthermore, Arg-309 in the EP3 is equivalent to Arg-295 in the TXAz receptor, which are located at the position analogous to Lys-296 of bovine rhodopsin in the seventh transmembrane segment. The latter amino acid residue was assigned for retinal attachment in the rhodopsin molecule (30). These structural features may reflect the acidic nature of the ligand for the prostanoid receptors. EP3 receptor has the two potential phosphorylation sites by CAMP-dependent protein kinase (31) in the first cytoplasmic loop, which may be relevant to the finding that [3H]PGE2 binding is affected by CAMP-dependent phosphorylation in brain membranes (32).
Poly (A)+ RNAs were prepared from various mouse tissues and hybridized with the EcoRI/BarnHI fragment of MP660 (Fig. 5). A positive band was seen at 2.3 kilobases in a number of tissues in which PGEz has pharmacological effects and/or specific binding sites (4). Another hybridizing band was detected at an estimated mRNA size of 7.0 kilonucleotides in kidney, uterus, brain, and mastocytoma P-815 cells. Identity of this latter band is not known at present. The tissue most highly expressing EP3 mRNA was kidney in which PGE2 exerts an inhibitory effect on sodium and water reabsorption by inhibiting adenylate cyclase via Gi (8,33). A significant band was also observed in stomach, suggesting that the receptor we cloned is indeed involved in inhibition of histamineinduced gastric acid secretion in this tissue (7). This analysis also showed that the uterus expressed the mRNA much higher than most tissues. It is known that PGE, exerts contractile response in uterine smooth muscle. The EP3 receptor we found may mediate this contractile action. Uterine contraction has been observed as a side effect of several EP3 agonists used as gastric anti-secretory prostanoid drugs (34). The EP3 receptor was also expressed in heart, lung, thymus, and spleen.
Although PGE, causes inhibition of sympathetic neurotransmitter release in some of these tissues (35), major functions of this receptor remain to be investigated. Our results also showed that EP, mRNA is significantly expressed in brain. The exact function of this receptor in this tissue is again not known at present. On the other hand, EP3 mRNA was not detectable in testis, and little was found in liver and ileum. In summary, we present here the complete amino acid sequence of the mouse EP3 subtype PGE receptor and provide direct proof that this receptor functionally couples to adenylate cyclase in an inhibitory manner. This work will contribute to our understanding on individual functions for three subtypes of PGE receptors and facilitate cloning of other members of eicosanoid receptors.