Molecular Cloning and Characterization of a Novel Human G-protein-coupled Receptor, EDG7, for Lysophosphatidic Acid*

Lysophosphatidic acid (LPA), together with sphingosine 1-phosphate, is a bioactive lipid mediator that acts on G-protein-coupled receptors to evoke multiple cellular responses, including Ca2+ mobilization, modulation of adenylyl cyclase, and mitogen-activated protein (MAP) kinase activation. In this study, we isolated a human cDNA encoding a novel G-protein-coupled receptor, designated EDG7, and characterized it as a cellular receptor for LPA. The amino acid sequence of the EDG7 protein is 53.7 and 48.8% identical to those of the human functional LPA receptors EDG2 and EDG4, respectively, previously identified. LPA (oleoyl) but not other lysophospholipids induced an increase in the [Ca2+] i of EDG7-overexpressing Sf9 cells. Other LPA receptors, EDG4 but not EDG2, transduced the Ca2+response by LPA when expressed in Sf9 cells. LPAs with an unsaturated fatty acid but not with a saturated fatty acid induced an increase in the [Ca2+] i of EDG7-expressing Sf9 cells, whereas LPAs with both saturated and unsaturated fatty acids elicited a Ca2+ response in Sf9 cells expressing EDG4. In EDG7- or EDG4-expressing Sf9 cells, LPA stimulated forskolin-induced increase in intracellular cAMP levels, which was not observed in EDG2-expressing cells. In PC12 cells, EDG4 but not EDG2 or EDG7 mediated the activation of MAP kinase by LPA. Neither the EDG7- nor EDG4-transduced Ca2+ response or cAMP accumulation was inhibited by pertussis toxin. In conclusion, the present study demonstrates that EDG7, a new member of the EDG family of G-protein-coupled receptors, is a specific LPA receptor that shows distinct properties from known cloned LPA receptors in ligand specificities, Ca2+ response, modulation of adenylyl cyclase, and MAP kinase activation.

Lysophosphatidic acid (LPA) 1 and sphingosine 1-phosphate (S1P) are lipid mediators with diverse biological properties (1)(2)(3). The cellular responses elicited by LPA vary widely. The effects of LPA on the cell cycle are either mitogenic or antimitotic. LPA stimulates phospholipase C (PLC) activation and consequent Ca 2ϩ mobilization, inhibits adenylyl cyclase, activates mitogen-activated protein (MAP) kinase, and stimulates the transcription of serum response element transcriptional reporter genes, such as c-fos, in various types of cells. It also exerts an effect on the cytoskeleton that can lead to changes in cell shapes and motility, which include inducing stress fiber production and stimulating chemotaxis, cell migration, and tumor cell invasiveness. These actions of LPA are believed to be mediated by seven-transmembrane G-protein-coupled receptor(s) (GPCR) on the cell surfaces. Some functional studies have suggested that multiple subtypes of LPA receptors with distinct signaling properties mediate the diverse cellular effects of LPA (4 -6). Indeed, several subtypes of LPA receptors, which are GPCRs, were identified recently.
The EDG (endothelial cell differentiation gene) family of orphan receptors comprises EDG1 (7), EDG2/Rec1.3/Vzg-1 (8,9), EDG3 (10), EDG4 (11), AGR16/H218 (12,13), and EDG6 (14), and their amino acid sequences show 36 -58% homology with one another. Hecht et al. (9) first reported that EDG2/ Rec1.3/Vzg-1 increased responsiveness to LPA in cell rounding and adenylyl-cyclase inhibition assays when overexpressed in cerebral cortical cell lines, showing that EDG2/Rec1.3/Vzg-1 is a receptor for LPA. Because LPA and S1P are structurally related, this finding has enabled scientists to examine whether members of EDG family function as receptors for LPA and S1P. At present, according to their amino acid sequence homologies, ligand specificities, and genomic structures (15), these GPCRs of the EDG family fall into two major groups that interact either with LPA (EDG2 (9) and EDG4 (11)) or with S1P (EDG1 (16 -18), EDG3 (19), and AGR16/H218 (20)). The ligand of EDG6, which is expressed predominantly in lymphoid tissue (14), has not been elucidated yet. A novel GPCR, named PSP24, which does not show significant sequence similarity with any member of the EDG family, has also been isolated from Xenopus oocytes as a functional receptor for LPA (21). Because some LPA-responsive cells do not express known LPA receptors (EDG2, EDG4, and PSP24), it was expected that unidentified subtypes of LPA receptors were suggested to be present in mammals (11,22). To understand the biological functions of LPA fully, we attempted to identify novel subtypes of LPA receptors. In this study, we identified and characterized a novel GPCR, EDG7, the third functional LPA receptor belonging to the EDG family.
Amplification of the Novel G-protein-coupled Receptor with Degenerate Primers-Total RNA was prepared from about 10 7 human Jurkat T cells, and 5.0 g was reverse-transcribed into DNA using the cDNA Cycle Kit (Invitrogen, Carlsbad, CA) with an oligo(dT) primer. The degenerate PCR primers were designed based on the amino acid sequences of the second and sixth of the seven transmembrane regions of the G-protein-coupled receptors EDG2 and EDG4. The oligonucleotides used were: GCIGCIGCIGA(C/T)(C/T)T(C/T)TTCGC (DP2; based on the second transmembrane region) and ACIAGICCIGGIGTCCAGCA (DP6; based on the sixth transmembrane region). Each PCR was carried out using 2.5 units Ex-Taq DNA polymerase (Takara Shuzo Co. Ltd., Kyoto, Japan) and 25% of the reverse-transcriptase reaction mixture in a 100-l reaction mixture containing 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 2 mM MgCl 2 , 200 M dNTP, and 50 pmol of each degenerate oligonucleotide primer (DP2 and DP6). 35 PCR cycles, each consisting of 45 s at 94°C (denaturation), 2 min at 55°C (annealing), and 3 min at 72°C (elongation) were performed, with the first cycle including an extended denaturation period (5 min), during which the polymerase was added. The 554-bp PCR products were purified by agarose gel electrophoresis and subcloned into a T-vector using the original TA-Cloning Kit (Invitrogen, Carlsbad, CA). Plasmid DNA was prepared using the Wizard Minipreps DNA Purification System (Promega, Madison, WI), and sequencing was performed by the dideoxy chain termination method using the ABI377 system (Perkin-Elmer, Branchburg, NJ).
5Ј-and 3Ј-RACE-A Marathon TM cDNA Amplification Kit (CLON-TECH, Palo Alto, CA) was used to perform 5Ј-and 3Ј-RACE. Doublestranded cDNA was prepared from poly(A)ϩ RNA of human Jurkat T cells, and nested PCR was carried out using the cDNA as a template and AP1, AP2 (supplied in the Marathon 228 cDNA Amplification Kit), and internal oligonucleotide primers (FW1, FW2, RV1, and RV2, see below) as PCR primers under the conditions described above. The sequences of the oligonucleotides were: FW1, AACCAACGTCTT-GTCTCCGCATAC (nucleotide positions 690 -713); FW2, AGCTAAT-GAAGACGGTGATGACTG (nucleotide positions 749 -772); RV1, GTC-CAGAAGCCCCTGACGGAGAAA (nucleotide positions 349 -372); and RV2, AGCAAGTTGGTGAGGGAAGCAGTC (nucleotide positions 381-404). The resulting DNA fragments were subcloned, and their DNA sequences were determined. Then the DNA fragment covering the open reading frame of EDG7 was amplified by reverse transcription-PCR using the cDNA from human Jurkat cells and oligonucleotide primers corresponding to the 5Ј-and 3Ј-noncoding regions and Pfu DNA polymerase (Toyobo, Tokyo, Japan). At least three independent reverse transcription-PCRs were carried out.
Cell Culture-Sf9 insect cells were grown in a serum-free ExCell-420 insect cell medium (Nichirei, Tokyo, Japan) at 27°C. PC12 cells were maintained in Dulbecco's modified Eagle's medium (high glucose) containing 10% horse serum and 5% fetal bovine serum in an atmosphere of 5% CO 2 .
Baculovirus System-The cDNA encoding the coding region of EDG7 (nucleotide positions 38 -1104) was inserted into the EcoRI/NotI site of the baculovirus transfer vector pFASTBAC1 (Life Technologies, Inc.). To achieve expression of FLAG-tagged EDG7, cDNA with a FLAG tag at its N-terminal was generated by the PCR using the following oligonucleotides: ATGCGAATTCATGGACTACAAAGACGATGACGATA-AAAATGAGTGTCACTATGACAAGCAC and ATGCGCGGCCGCTCC-AGAGTTTAGGAAGTGCTTTTA. The resulting DNA fragment was digested by EcoRI/NotI and ligated into pFASTBAC1. Recombinant viruses were prepared using the Bac-to-Bac system (Life Technologies, Inc.) according to the manufacturer's protocol. For infection, 6 ϫ 10 5 cells/ml were mixed with recombinant or wild-type Autographa californica nuclear polyhedrosis virus to produce a multiplicity of infection of 10 and incubated for 48 h at 27°C. Expression of the FLAG-tagged protein was detected by Western blotting with an anti-FLAG monoclonal antibody (M5; Eastman Kodak Company, New Haven, CT). For preparation of recombinant baculoviruses to express FLAG-tagged EDG2 and FLAG-tagged EDG4, cDNAs with FLAG tag at their Nterminal were generated by the reverse transcription-PCR using human brain cDNA as template DNA and the following oligonucleotides: ATGCGAATTCATGGACTACAAAGACGATGACGATAAAGCTGCCA-TCTCTACTTCCATCCCT, ATGCGCGGCCGCCTAAACCACAGAGTG-ATCATTGCT (for EDG2) and ATGCGAATTCATGGACTACAAAGAC-GATGACGATAAAGTCATCATGGGCCAGTGCTACTAC, ATGCGCG-GCCGCTCAGTCCTGTTGGTTGGGTTGAGC (for EDG4). The resulting DNA fragments were digested by EcoRI/NotI and ligated into pFAST-BAC1, and baculoviruses were prepared as described above. DNA sequences of cDNAs prepared by reverse transcription-PCR were confirmed by DNA sequencing.
Ca 2ϩ Measurements-Sf9 cells were harvested 2 days after baculovirus infection, washed gently with an HBS buffer (20 mM Hepes, pH 7.4, containing 120 mM NaCl, 4.7 mM KCl, 1.2 mM MgCl 2, 1.25 mM CaCl 2 , 1.2 mM KH 2 PO 4, and 10 mM glucose), and loaded with 2 M Fura-2 acetoxymethyl ester (Fura-2 AM; Molecular Probes Inc., Eugene, OR) for 45 min. Free Fura-2 AM was washed out, and the cells were resuspended in the HBS buffer to produce a concentration of 10 6 cells/ml. Agonist-induced Fura-2 AM fluorescence of samples in quartz cuvettes kept at 27°C was monitored at excitation wavelengths of 340 and 380 nm and an emission wavelength of 300 nm using a CAF-110 spectrofluorimeter (Japan Spectroscopy, Inc., Tokyo, Japan). Fluorescence was recorded before and after addition of LPA and other phospholipids, which were dissolved in phosphate-buffered saline with 0.01% (w/v) of fatty acid-free bovine serum albumin (Sigma).
Pertussis Toxin and U73122 Treatment-24 h after baculovirus infection of the Sf9 cells, PTX (100 ng/ml; Calbiochem, La Jolla, CA) was added to the culture medium, and incubation was continued for an additional 24 h. Then the cells were collected, and their Ca 2ϩ responses were tested as described above. The phospholipase C inhibitor U73122 (3 M; Calbiochem, La Jolla, CA) was added to Sf9 cells 3 min before LPA was added.
[ 3 H]LPA Binding-Sf9 cells (5 ϫ 10 5 ) infected by each baculovirus for 48 h were washed with a binding buffer (phosphate-buffered saline containing 0.25% bovine serum albumin) and incubated for 60 min at 0°C in the same buffer containing various concentration of [ 3 H]LPA in a 96-well membrane filter plate (pore size, 65 nm; Millipore, Orlando, FL). Then the cells were washed three times with the wash buffer (phosphate-buffered saline containing 1% bovine serum albumin) using a Multi-screen Filtration System (Millipore), and the radioactivity bound to the cells was quantified using a ␤-counter. Total and nonspecific binding was evaluated in the absence and presence of 10 M nonradioactive LPA, respectively. To examine the specificity of LPA binding, the amounts of LPA bound (using 10 nM of cAMP Measurements-Sf9 cells were infected with recombinant baculoviruses and harvested 2 days after infection. Cells were incubated with 5 M forskolin in the presence of the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (0.5 mM) for 10 min followed by a 20-min stimulation with 2.5 M LPA with various acyl chains in an HBS buffer. cAMP levels were determined using a cAMP enzyme immunoassay system (Biotrak; Amersham Pharmacia Biotech), following the instructions of the manufacturer.
Assay for MAP Kinase-mediated Signal Transduction-To measure MAP kinase-mediated signal transduction, we employed the PathDetect TM Elk1 Trans-Reporting System (Stratagene, La Jolla, CA). This assay employs a fusion protein that contains the DNA-binding domain of GAL4 and the transactivation domain of Elk1 to induce expression of a luciferase reporter driven by an artificial promoter containing five GAL4-binding sites. Phosphorylation of the transactivation domain of Elk1 by MAP kinase, in turn, activates the transcription of the luciferase gene from the reporter plasmid. cDNAs encoding FLAG-tagged EDG7, EDG2, and EDG4 were inserted into EcoRI/NotI sites of the mammalian expression plasmid pcDNA3 (Invitrogen, Carlsbad, CA) (FLAG-EDG7-pcDNA3, FLAG-EDG2-pcDNA3, and FLAG-EDG4-pcDNA3). Transient transfections were performed using SuperFect transfection reagent (Qiagen, Hilden, Germany). Briefly, 100,000 PC12 cells were seeded in 24-well plates 24 h before the transfection. Each point was transfected with 200 ng of FLAG-EDG7-pcDNA3 (or FLAG-EDG2-pcDNA3, FLAG-EDG4-pcDNA3, or empty pcDNA3 for the control experiment), 200 ng of pFR-Luc, 25 ng of pFA-Elk1 (as described by Stratagene), and 575 ng of the empty pcDNA3 vector. 12 h post-transfection, the cells were rinsed twice with phosphate-buffered saline and incubated in a serum-free medium for another 12 h. The cells were then stimulated with 10 M of LPA (oleoyl) for 10 h at 37°C and lysed in 500 l of an extraction buffer, 100 l of which was used to measure luciferase activity, following the instructions of the manufacturer. Expression of each protein was confirmed by immunofluorescence using anti-FLAG monoclonal antibody M5.
Northern Blot Analysis-Human Multiple Tissue Northern blots were purchased from CLONTECH (Palo Alto, CA), DNA probes (nucleotide positions 307-730) were labeled by random priming with [␣-32 P]dCTP, and hybridization was carried out at 65°C for 4 h in a rapid hybridization buffer (Amersham Pharmacia Biotech). Each blot was rinsed with 2ϫ SSC at room temperature for 5 min, washed twice with 0.5ϫ SSC containing 0.1% SDS at 65°C for 40 min, and then autoradiographed using Kodak X-Omat AR film at Ϫ80°C with an intensifying screen for 12 h. Finally, each blot was rehybridized with a glyceraldehyde-3-phosphate dehydrogenase cDNA probe (CLONTECH, Palo Alto, CA) as an internal standard.

RESULTS
PCR Amplification, Cloning, and Sequencing of EDG7 cDNA-We investigated a previously unidentified GPCR belonging to the EDG family by subjecting cDNA of human Jurkat T cells to PCR amplification. Two blocks of conserved amino acid sequences in EDG2 and EDG4, one from the second and one from the sixth transmembrane domain, were chosen for the synthesis of the degenerate oligonucleotide primers DP2 and DP6 (see "Materials and Methods"). The PCR reaction was performed, yielding a 554-bp fragment containing a GPCR-like sequence distinct from those of known members of the EDG family. Using 5Ј-and 3Ј-RACE, we isolated a DNA fragment that covered the entire open reading frame of this novel gene. Complete sequencing of the DNA revealed a 1059-bp open reading frame flanked by a 42-bp 5Ј-untranslated region and a 44-bp 3Ј-untranslated region. The translational initiation site (ATG) was assigned to the first methionine codon (nucleotide positions 43-45), because an in-frame stop codon was present upstream of this methionine residue and flanking sequences (CCACA) were present, fulfilling Kozak's criteria for initiation (25). An in-frame translational termination codon (TAA) was present after nucleotide 1,101. Therefore, we concluded that this new GPCR contains 353 amino acids and that its molecular mass is 40,128 Da (Fig. 1A).
A comparison of the deduced amino acid sequence with those of known EDG sequences revealed that the primary structure of the predicted protein was similar to those of the GPCRs of the EDG family, with overall sequence identities to human EDG1, human EDG2 (Vzg-1), human EDG3, human EDG4, rat H218, and human EDG6 of 34.8, 53.7, 36.3, 48.8, 33.8, and 35.5%, respectively. Therefore, the protein encoded by the cloned cDNA was named EDG7 (Fig. 1B). A high degree of similarity between EDG2 and EDG4 (approximately 50%; Fig.  1B) was observed among EDG7, EDG2, and EDG4. To gain better understanding of the relationships involved in the molecular evolution of the EDG family, a phylogenetic tree was constructed using the neighbor joining method (Fig. 1C). According to this phylogenetic tree, the EDG family can be classified into two distinct groups: EDG1, EDG3, H218/AGR16, and EDG6 belong to one, and EDG2, EDG4, and EDG7 belong to the other. Because EDG1, EDG3, and H218/AGR16 have been reported to function as S1P receptors and EDG2 and EDG4 have been reported to function as LPA receptors, EDG7, EGD2, and EDG4 appear to form a subfamily of the EDG family, and EDG7 may function as a receptor for LPA.
Evaluation of EDG7 as an LPA Receptor by Measuring the Ca 2ϩ Response-To explore the hypothesis that EDG7 is a receptor for LPA, we expressed the recombinant receptor in Sf9 insect cells and measured the increases in the [Ca 2ϩ ] i evoked by LPA. We chose Sf9 cells because, although LPA does not affect their [Ca 2ϩ ] i , it induces rapid increases in the [Ca 2ϩ ] i of almost all types of mammalian cells. Western blotting analysis using anti-FLAG antibody M5 confirmed that an approximately 35-kDa protein containing the FLAG tag was expressed in cells infected with EDG7 baculovirus but not with the wildtype virus (Fig. 2A). The addition of 1 M oleoyl-LPA to Sf9 cells infected with either the FLAG-EDG7-expressing (Fig. 2B) or EDG7-expressing recombinant baculovirus (data not shown)  (Fig. 2C).
We also expressed known LPA receptors EDG2 and EDG4 in Sf9 cells and compared their Ca 2ϩ response. Protein expression was confirmed by Western blotting using anti-FLAG antibody ( Fig. 2A). The expression of EDG2 protein was much higher than those of EDG4 and EDG7, but it failed to transmit a detectable Ca 2ϩ signal in response to LPA (oleoyl) (Fig. 2B), consistently with the observation by Zondag et al. (17). On the other hand, EDG4 mediated a Ca 2ϩ signal like EDG7 in Sf9 cells (Fig. 2B). Thus, both EDG4 and EDG7 but not EDG2 mediate the Ca 2ϩ response by LPA in Sf9 cells.
LPA Binding-We investigated the specific binding of [ 3 H] oleoyl-LPA to Sf9 cells expressing FLAG-EDG7, FLAG-EDG2, and FLAG-EDG4 using various doses of LPA. EDG7-and EDG4-expressing Sf9 cells increased the specific binding of [ 3 H]oleoyl-LPA in comparison with wild-type baculovirus and EDG2-infected cells (Fig. 3A) 3B). Thus, the binding affinity for LPA of EDG7 is relatively lower than that of EDG4. We also examined the competition between the binding of [ 3 H]LPA and related nonradioactive lipids to FLAG-EDG7-expressing Sf9 cells. LPA (10 M) reduced [ 3 H]LPA binding, whereas the other related lipids examined (LPC, LPE, LPS, LPI, S1P, and PAF) did not (Fig. 3C). These data clearly demonstrate that EDG7 represents a specific receptor for LPA.
We also examined the ability of cPA, which was first isolated from the slime mold Physarum polycephalum (27), to increase the [Ca 2ϩ ] i . As shown in Fig. 5, cPA with oleoyl acid at the sn-1 position of the lipid (18:1 cPA, PHYLPA-8) increased the [Ca 2ϩ ] i , whereas 10 M cPAs with palmitic acid (16:0 cPA, PHYLPA-5) and cyclopropane-containing hexadecanoic acid (PHYLPA-1) were inactive. These data demonstrated that EDG7 prefers LPA with unsaturated fatty-acyl chains at the sn-1 or sn-2 position.
Characterization of the EDG7-transduced Ca 2ϩ Response-We next examined the effects of PTX and the PLC inhibitor U73122 on EDG7-transduced Ca 2ϩ responses. As shown in Fig. 6, pretreatment of FLAG-EDG7-expressing Sf9 cells with 100 ng/ml PTX for 24 h did not affect the Ca 2ϩ response evoked by 1 M LPA, whereas U73122 inhibited effectively the Ca 2ϩ response transduced by EDG7, demonstrating that EDG7 is coupled to PTX-insensitive G-protein(s). The Ca 2ϩ response transduced by EDG4 was also inhibited by PLC inhibitor U73122 but unaffected by PTX pretreatment (Fig. 6).
Effect on cAMP Level-It has been repeatedly demonstrated that LPA inhibits adenylyl cyclase via a G i -mediated signaling event in mammalian cells. To examine whether EDG7 participates in the inhibition of adenylyl cyclase in response to LPA, Sf9 cells expressing each LPA receptor were pretreated with forskolin to raise the intracellular cAMP level and then treated with LPA. Unexpectedly, LPA did not suppress forskolin-induced cAMP accumulation but rather increased intracellular cAMP level in EDG7-expressing Sf9 cells (Fig. 7A). The in- crease in cAMP level by LPA in EDG7-expressing Sf9 cells was insensitive to PTX (Fig. 7B) and observed only when the cells were pretreated with forskolin (data not shown). The similar response was observed in EDG4-expressing Sf9 cells (Fig. 7B). cAMP level in EDG2-expressing Sf9 cells was unaffected upon stimulation with LPA (Fig. 7A), as described previously (17).
Effect on MAP Kinase-MAP kinase in insect Sf9 cells has not been characterized yet. In addition we could not detect MAP kinase activity using anti-human MAP kinase antibodies or substrates for MAP kinase such as oligopeptide from human EGF-receptor (data not shown). Thus we used the mammalian expression system for this experiment. To determine whether EDG7 mediates MAP kinase activation, we used the PathDetect TM Elk1 trans-Reporting System. Elk1 is a transcription factor that is phosphorylated and activated by MAP kinase (28). PC12 cells were transfected with FLAG-tagged EDG2, EDG4, or EDG7 and subjected to an assay of MAP kinase activation. Expressions of each receptor were confirmed by immunofluorescence in PC12 cells (data not shown). LPA activated the transcription of the luciferase gene through activation of the Elk1 in EDG4-expressing PC12 cells but not in  EDG2-or EDG7-expressing cells (Fig. 8), suggesting that only EDG4 is coupled to MAP kinase activation among the cloned LPA receptors.
Tissue and Cellular Distributions of EDG7-The tissue distribution of EDG7 was examined by subjecting several human tissues and cancer cell lines to Northern blot analysis. EDG7 transcripts of about 4.3 kilobases were detected in the heart, pancreas, prostate, and testis and, to a lesser extent, in the lung and ovary (Fig. 9). EDG7 transcripts were also found weakly in the human cancer cell lines HeLa, K562, SW480, A549, and G361, the last of which, a melanoma cell line, contained the highest level (data not shown).

DISCUSSION
In this study, we isolated a new member of the EDG family of G-protein-coupled receptors, EDG7, and showed that it functions as a cellular receptor for LPA. Interestingly, EDG7 transduce increases in both [Ca 2ϩ ] i and the cAMP level upon stimulating only with LPA containing an unsaturated fatty-acyl chain at the sn-1 or sn-2 position but not with LPA containing any of saturated fatty acids. EDG7 also reacted with cPA with unsaturated fatty acid (Fig. 5). In contrast, EDG4 can be stimulated by LPAs with both an unsaturated and saturated fatty acid. Moreover, Hecht et al. (9) have previously demonstrated that both oleoyl-and stearoyl-LPA induce EDG2-dependent cell rounding in EDG2-transfected cells. Erickson et al. (29) also reported that LPAs with both saturated and unsaturated fatty acid activate EDG2 in the yeast pheromone response pathway. Thus, EDG2 seems to recognize LPAs with both saturated and unsaturated fatty acids like EDG4. It has been previously demonstrated that LPAs with unsaturated fatty acids evoked greater Ca 2ϩ responses of the human epidermoid carcinoma cell line A431 than LPAs with saturated fatty acids (30). In addition, Tokumura et al. (31) reported that degree of unsaturation in the acyl moiety of LPA affects proliferation of cultured vascular smooth muscle cells from rat aorta by LPA. A newly identified LPA receptor, EDG7, may account for these response.
Several molecular species of LPA with saturated (stearoyl- (18:0), palmitoyl-(16:0)) or unsaturated (oleoyl-(18:1), linoleoyl-(18:2), arachidonoyl-(20:4)) acyl chains have been detected in activated platelets (26). The biochemical pathway for LPA formation has not been established with certainty but may involve, in part, the degradation of phosphatidic acid by a phospholipase A(s). Both phospholipase A2 and A1 may be involved in the production of LPA with saturated and unsaturated fatty acids, respectively. In this connection, it is interesting to note that in our preliminary work, mammalian tissues and possibly cells express PA-specific phospholipase A1, which possesses a primary structure distinct from that of known phospholipases and lipases. Such types of phospholipase A1 may participate in the production of LPA with an unsaturated fatty acid. It is also reasonable to assume that the pathways for the production of each LPA species are regulated differently and that only the process leading to the production of LPA with unsaturated fatty acid can activate EDG7-mediated signaling event.
One of the best documented cellular activities of LPA is its mobilization of [Ca 2ϩ ] i . It has been previously demonstrated that LPA-induced Ca 2ϩ mobilization can be transduced through both PTX-sensitive (32) and PTX-insensitive (33-35) G-protein followed by activation of PLC. An et al. (36) used TAg-Jurkat T cells transiently transfected with human EDG2 or EDG4 to demonstrate that PTX completely blocks LPAinduced Ca 2ϩ mobilization in EDG2-expressing cells, and partially blocks it in EDG4-expressing cells. They speculated from these data that EDG2 is coupled to PTX-sensitive G i , whereas EDG4 is coupled to both G i and G q (36). In this study, we examined EDG7-transduced Ca 2ϩ mobilization using Sf9 cells, because parent Sf9 cells appear not to contain endogenous LPA receptor and therefore does not respond to LPA stimulation. According to our observations, both EDG7 and EDG4 transduced the Ca 2ϩ response by LPA in a PTX-insensitive manner in Sf9 cells, indicating that EDG7 and EDG4 are coupled to a PTX-insensitive G-protein(s), possibly G q . Although the human EDG2 was expressed significantly in Sf9 cells, Ca 2ϩ mobilization has not been observed in the cells, indicating that none of G-proteins in Sf9 cells can be coupled to human EDG2 (Ref. 17 and present study). It is interesting to note that no increase in the specific binding of LPA in EDG2-expressing Sf9 cells was observed either. Consistently with this observation, Figler et al. (37) demonstrated that specific binding site for ligand in GPCR was available only when appropriate G-protein(s) are coupled to the receptor.
The effect of LPA on the cell growth is either proliferative or antiproliferative. In most fibroblastic cell lines, LPA induces mitogenic responses and inhibits adenylyl cyclase via a G imediated signaling event (33). In fact, LPA inhibited adenylyl cyclase in EDG2-transfected mammalian cells (9). On the other hand, in nonfibroblastic cell line Sp2/0-Ag14 myeloma, which does not express either EDG2 or EDG4 (22), LPA induced an increase in cAMP and inhibited cell proliferation in a PTXinsensitive manner, which is accompanied by an increase in cytoplasmic Ca 2ϩ (4). They also indicated that LPA with a higher degree of unsaturation and longer acyl chains increased the antiproliferative effects (4). A newly identified EDG7, which enhanced adenylyl cyclase in a PTX-insensitive manner, may account for those responses. The idea is further supported by the observation that EDG7 is not coupled to MAP kinase activation, which leads to cell proliferation in various types of cells. In EDG7-and EDG4-expressing Sf9 cells increases in cAMP by LPA were observed only after adenylyl cyclase was first stimulated with forskolin (Fig. 7). Felder et al. (38) reported that accumulation of cAMP by CB1 antagonists in CB1expressing Chinese hamster ovary cells was only observed when cells were pretreated with forskolin. Some isoforms of adenylyl cyclase(s) that are coupled to EDG7 and EDG4 in Sf9 insect cells may require an initial priming for activation like type II or IV adenylyl cyclases in mammalian cells as reported by Tang et al. (39). LPA is known to exert an effect on the cytoskeleton that can lead to changes in cell shapes and motility, which includes inducing the production of stress fiber through Rho-GTPase activation. We examined the effect of LPA stimulation on the formation of actin stress fiber in EDG7expressing CHO-K1 and PC12 cells by staining an actin filament with FITC-phalloidin but did not observe any obvious differences between the EDG7-expressing and control cells (data not shown).
In conclusion, EDG7 is a specific LPA receptor that shows distinct properties from known cloned LPA receptors in ligand specificity, Ca 2ϩ response, and modulation of adenylyl cyclase and MAP kinase. Further studies are definitely needed in understanding a physiological role of this receptor.