Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The use of recombinant receptor cells lines is an established approach in the search for selective compounds for G-protein-coupled receptors (GPCRs). Indeed, since the advent of molecular cloning and the consequent discovery of an unexpected degree of receptor multiplicity within GPCR superfamilies (e.g., 5-hydroxytryptamine [5-HT]) combined with subtype colocalization within tissue types it is hard to imagine how selective compounds could be identified without such techniques. Great care, however, must be taken in the choice of host cell lines for such heterologous receptor expression systems. For example, it is essential that the receptor subtype under investigation is not endogenously expressed by the host cell line. To measure the assay endpoint with confidence, it is necessary to ensure that the host cell is completely unresponsive to the transfected receptor's endogenous ligand or closely related analogs. Generally this might, at worst, severely restrict the availability of suitable host cell types requiring perhaps the use of a less than ideal cell line in terms of, for example, growth characteristics. In the case of orphan receptors, i.e., receptors with no known endogenous ligand, or, as is the case in the present study, receptors whose endogenous ligand ubiquitously affects mammalian cells, such considerations become more problematic. Here we describe an approach to develop a functional assay for the human Edg2 (hEdg2) receptor that uses lysophosphatidic acid (LPA) as its endogenous ligand. Evidence is emerging to suggest that Edg2 receptors play a crucial role in nerve cell myelination and might offer a potential target in the treatment of demyelinating diseases such as multiple sclerosis. For example not only are Edg2 receptors expressed exclusively in adult oligodendrocytes and Schwann cells (Weiner et al., 1998) but also LPA protects against apoptosis in cultured Schwann cells (Weiner and Chun, 1999).

The Edg2 receptor belongs to a GPCR family so called because the first family member to be cloned was encoded by an endothelial differentiation gene (Hla and Maciag, 1990). Eight family members have been identified sharing closest resemblance to the cannabinoid receptor family. They play an important role in cell growth, development, and maintenance and in modulating cytoskeletal-dependent changes. These effects are achieved by activating multiple secondary messenger pathways involving coupling to numerous trimeric guanine nucleotide-binding proteins (Gi, Gq/11, G12/13) and/or the small G-protein Rho (Goetzl and An, 1998).

LPA, the endogenous ligand for the Edg2 receptor, is the simplest of all glycerophospholipids and one of a growing list of lipid extracellular messengers. It is found in 0.5 µM concentrations in serum being formed either by de novo synthesis from glucose-derived dihydroxyacetone phosphate and a fatty acid-coenzyme A but more probably by enzymatic conversion of precursor glycerophospholipids (Goetzl and An, 1998). It is largely associated with, and released from, activated platelets and mediates its effects, at least in part, by Edg2, Edg4, and Edg7 receptor stimulation.

The ubiquitous responsiveness of mammalian cells lines to LPA has led not only to a limited choice of suitable host cell lines for recombinant receptor cell line studies but also to difficulties in identifying it as the Edg2 receptor endogenous ligand. Erickson et al. (1998) sought to overcome this problem by transfecting the hEdg2 receptor in Saccharomyces cerevisiae and monitoring the activation of the pheromone-inducible mitogen-activated protein kinase cascade. The main disadvantage of this approach is the very weak potency of compounds tested, presumably due to inefficient coupling of the mammalian receptor with the yeast G-protein, limiting potential structure activity relationship studies.

An alternative approach described in this study and based on the Edg2 receptor/Gi signaling pathway has been to construct a fusion protein in which the N terminus of the rGi1 subunit is linked, in-frame, to the C terminus of the Edg2 receptor (Bertin et al., 1994; Stables et al., 1997; Wise et al., 1997a,b). In addition, the rGi1 subunit has been rendered pertussis toxin (PTX)-insensitive by the mutation of cysteine 351 to glycine (Milligan, 1988). An agonist-induced [³⁵S]guanosine 5'[-thio] triphosphate ([³⁵S]GTPS) binding assay with membranes from HEK cells transiently expressing the hEdg2 receptor-rGi1 fusion protein has then been developed (Stanton and Beer, 1997). Pretreating the cells with PTX ensures that the agonist-induced [³⁵S]GTPS binding is, by definition, mediated by the transfected receptor. This fusion protein approach has enabled us to confirm that LPA activates hEdg2 receptors in mammalian cells. This assay gives a measure of potency and efficacy and is amenable to high throughput application. Identification of compounds with a range of intrinsic efficacies should be invaluable in exploring the physiological/pathophysiological role of the Edg2 receptor.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. [³⁵S]GTPS (>1000 Ci/mmol) was obtained from Amersham Pharmacia Biotech (UK). All tissue culture reagents were from Sigma (St. Louis, MO) or Life Technologies Inc. (Rockville, MD). PTX and sphingosine-1-phosphate (S-1-P) were purchased from Sigma and oleoyl 1-acyl-2-hydroxy-sn-glycero-3-phosphate, sodium salt (LPA), oleoyl (and palmitoyl) 1-acyl-2-hydroxy-sn-glycero-3-phosphoethanolamine (LPE), oleoyl 1-acyl-2-hydroxy-sn-glycero-3-[phospho-rac-(1-glycerol)], sodium salt (LPG), and oleoyl 1-acyl-2-hydroxy-sn-glycero-3-phosphocholine (LPC) were from Avanti Polar Lipids Inc. (Alabaster, AL). Primers were purchased from Life Technologies Inc. and restriction enzymes from New England BioLabs Inc. (Beverly, MA).

Cloning of hEdg2 Receptor. The cDNA encoding the hEdg2 receptor was cloned with polymerase chain reaction (PCR) techniques with a klenTaq polymerase mix (Advantage cDNA PCR kit from Clontech, Palo Alto, CA). PCR reactions were performed on human substantia nigra cDNA (Clontech) with the primers 5'-GTCATGGCTGCCATCTCTACT-3' (sense) and 5'-GTTCTAAACCACAGAGTGGTC-3' (antisense). The PCR-amplified fragment was subcloned into pCR-Blunt II-TOPO (Invitrogen, Carlsbad, CA) and confirmed as that encoding for the hEdg2 receptor by sequencing. For use in the cotransfection studies the hEdg2 receptor cDNA was excised with the restriction enzymes XhoI and HindIII and ligated into the mammalian expression vector pcDNA3.1() (Invitrogen). For the hEdg2 receptor-Cys³⁵¹  Gly rGi1 fusion studies the open reading frame (ORF) of the hEdg2 receptor underwent further PCR with the following primers containing HindIII and NcoI restriction sites, respectively (underlined), 5'-AAGCTTATGGCTGCCATCTCTACTTCCATC-3' (sense) and 5'-CCATGGCAACCACAGAGTGGTCATTGCTGT-3' (antisense). The PCR-amplified fragment was then subcloned into pCRII-TOPO (Invitrogen).

Construction of Cys³⁵¹  Gly rGi1. Site-directed mutagenesis on the rGi1 subunit [with a QuikChange kit (Stratagene, La Jolla, CA)], was carried out to remove a NcoI site (Ala⁸⁷, GCCGCG) within the ORF. This allowed a NcoI site at the beginning of the ORF to be used as the site of fusion of the hEdg2 receptor to the rGi1 subunit. To render the rGi1 subunit PTX-insensitive, the cDNA underwent PCR with the Advantage cDNA PCR kit and the primers 5'-GAATTCCCATGGGCTGCACACTGAGCGCTGAGGACA-3' (sense) and 5'-GCGGCCGCTCTAGATTAGAAGAGACCACCGTCTTTTAGGTTATT-3' (antisense) (Cys³⁵¹  Gly in bold; restriction sites for EcoRI, Nco 1, NotI, and XbaI underlined) resulting in a Cys³⁵¹  Gly mutation in the 3' end of the rGi1 subunit and hence to PTX insensitivity. This PCR product was then reinserted into the EcoRI/XbaI site of pBluescript SK(), and the mutation was confirmed by sequence analysis.

Construction of the hEdg2 Receptor-Cys³⁵¹  Gly rGi1 Fusion Protein. The hEdg2 cDNA in pCRII-TOPO was digested with HindIII and NcoI and ligated into the Cys³⁵¹  Gly rGi1 subunit containing pBluescript SK(), itself having been digested with HindIII and NcoI. Introduction of the NcoI site at the 3' end of the ORF of the receptor resulted in the removal of the stop codon and insertion of an extra alanine between the receptor and rGi1 subunit. The full fusion construct was then excised from pBluescript SK() with HindIII and NotI, ligated into the mammalian expression vector pcDNA3.1(+) and confirmed by sequence analysis.

Construction of the myc and Hemaglutinin (HA) Epitope-Tagged rGi1 Subunits. The myc (EQKLISEED) and HA (DVPDYA) epitope tags were engineered into the rGi1 subunit after the sixth amino acid residue from its N terminus. The 5'-oligonucleotides containing the myc and HA tags (underlined) and an EcoRI (GAA TTC) restriction site were ATC GAA TTC ATG GGC TGC ACA CTG AGC GCT GAG CAG AAG CTG ATC TCC GAG GAG GAC CTG GAG GAC AAG GCG GCC GTG GAG and ATC GAA TTC ATG GGC TGC ACA CTG AGC GCT GAC GTC CCC GAC TAC GCG GAG GAC AAG GCG GCC GTG GAG, respectively. The 3'-oligonucleotide containing the stop codon and an XbaI restriction site (ATC TAG) was AT ATC TAG ACT AGC AGA GCT TAG AAG AGA CCA CAG TCT. Introduction of the epitope tags was achieved with standard PCR methodologies. The PCR products were then restriction enzyme-digested and subcloned into pcDNA3.1(+). The sequence of these constructs was confirmed by an ABI 375 automated fluorescence sequencer.

Transient Transfection of Human Embryonic Kidney (HEK) 293 Cells. HEK cells were plated out at a density of 3 × 10⁵ cells/ml in cell culture dishes containing 10 ml of Dulbecco's modified Eagle's medium (with 8% fetal calf serum) and incubated overnight. The following day 0.5 ml of 0.25 M CaCl₂, 0.5 ml 2× BBS buffer (50 mM N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid, 280 mM NaCl, 1.5 mM Na₂HPO₄), pH 6.95, and DNA (7.5 µg of hEdg2 receptor plus 7.5 µg of rGi1 subunit [tagged, wild type, or Cys³⁵¹  Gly-mutated] for cotransfection studies and 15 µg for the hEdg2 receptor-rGi1 fusion protein) was made up, left at room temperature (RT) for 15 min, and added to each cell culture dish. The HEK cells were incubated at 37°C in 3% CO₂ for 48 to 60 h before being harvested in phosphate-buffered saline and stored at 70°C. For the PTX studies 100 ng/ml PTX was added to the cell medium 20 h before harvesting.

Agonist-Induced [³⁵S]GTPS Binding. Untransfected HEK 293, CHO-K1, Ltk, HeLa, COS-7, PC12, or HEK 293 cells transiently expressing the hEdg2 receptor either alone or cotransfected with the rGi1 subunit (HA or myc epitope tagged at the N terminus of the subunit, wild type, or Cys351Gly-mutated) were homogenized in ice-cold 20 mM HEPES buffer containing 10 mM EDTA (pH 7.4 at RT) with a Kinematica polytron (setting 5, 10 s) and centrifuged at 48,000g, 4°C for 15 min. The pellet was resuspended in ice-cold 20 mM HEPES buffer containing 0.1 mM EDTA (pH 7.4 at RT) and recentrifuged at 48,000g, 4°C for 15 min. The final pellet was resuspended in 20 mM HEPES buffer containing 100 mM NaCl and 10 mM MgCl₂, pH 7.4, at RT.

Data Analysis. Background filter counts, i.e., residual radioactivity bound to the filter in the absence of membrane, were subtracted from each sample count. Test results are represented as either percentage increase in binding above basal (that seen in the absence of test compound) or stimulated  basal dpm. Dose-response curves were plotted and analyzed by nonlinear, least-squares regression analysis with an iterative curve-fitting routine (Marquardt-Levenberg method) provided by the data manipulation software RS/1 (Software Products Corp., Cambridge, MA). Potency values are expressed as pEC₅₀ values (log₁₀ concentration of agonist required to give 50% of its own maximal stimulation). Efficacy values are given as a percentage of the maximal response seen with LPA. Data are expressed as arithmetic mean ± S.E. from 3 experiments.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

All assays were carried out in the presence of 1 µM GDP. Addition of GDP favors binding to Gi subunits, which have a relatively high spontaneous rate of GDP dissociation. As expected, LPA dose dependently increased [³⁵S]GTPS binding in all the nontransfected cell lines tested with varying degrees of efficacy (Fig. 1), and as anticipated this response was abolished after pretreatment of membranes (HEK cells) with PTX (100 ng/ml) due to ADP-ribosylation of endogenous Gi/o-proteins (data not shown). The following rank order of responsiveness was seen: Ltk = CHO PC12 > HeLa = HEK = COS-7. HEK cells were used for all subsequent transient expression studies, because they displayed a relatively low endogenous LPA response and were known to have good transfection efficiency.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   LPA-induced [35S]GTPS binding, in the presence of 1 µM GDP, mediated by endogenous receptors in HEK 293 (), CHO-K1 (), Ltk (), HeLa (), COS-7 (), and PC12 () cells. Data are expressed as percentage increase in binding above basal. Curves are a composite of three or more separate experiments, mean ± S.E.

The LPA-mediated [³⁵S]GTPS binding in the hEdg2 receptor-expressing cell line, either alone or together with the various Gi constructs or as a PTX-insensitive hEdg2 receptor-rGi1 fusion protein, is shown in Fig. 2. The maximal response to LPA after transient transfection of the hEdg2 receptor alone was only modestly improved compared with that seen in the mock or untransfected cells (10%). The response was improved by cotransfection of the receptor with the wild-type rat Gi1 (rGi1) subunit (25 ± 1%), was further enhanced when the wild-type rGi1 subunit was replaced with HA (27 ± 2%) or myc (30 ± 5%) N-terminal-tagged rGi1 subunits and even more so when replaced with the mutant rGi1 subunit (48 ± 3%). The transient expression of the PTX-insensitive hEdg2 receptor-rGi1 fusion protein yielded the most efficacious response (69 ± 10%). The potency values of LPA, under the various assay conditions, are presented in Table 1 as pEC₅₀ values.

View larger version (17K): [in this window] [in a new window]   Fig. 2.   LPA-induced [35S]GTPS binding, in the presence of 1 µM GDP, mediated by mock (), hEdg2 receptor (), hEdg2 receptor + wild-type rGi1 subunit (), hEdg2 receptor + rGi1HA(N) subunit (), hEdg2 receptor + rGi1 myc(N) subunit (), hEdg2 receptor + mutant rGi1 subunit (), and the hEdg2 receptor-rGi1 fusion protein () transiently transfected HEK cells. For potency (pEC50) and efficacy (Emax) values see Table 1. Data are expressed as percentage increase in binding above basal. Each curve is a composite of four experiments, mean ± S.E.

View this table: [in this window] [in a new window]   TABLE 1 Basal constitutive activity and LPA-induced [35S]GTPS binding following transient transfections Levels of basal activity, i.e., in the absence of agonist, seen in membranes prepared from HEK cells transiently transfected with the hEdg2 receptor (alone or together with the various Gi constructs) or the hEdg2 receptor-rGi1 fusion protein or mock-transfected HEK cells are presented as dpm bound. Potency values for LPA-induced [35S]GTPS binding carried out in membranes prepared from each condition are given as pEC50 values, and efficacy values (Emax) are the maximal stimulation achieved. Results are arithmetic means ± S.E. of four experiments.

In a second series of experiments, the effects of PTX pretreatment (100 ng/ml) on the hEdg2 receptor-expressing cells cotransfected with the various Gi constructs or as a PTX-insensitive hEdg2 receptor-rGi1 fusion protein were investigated (Fig. 3). Compared with the equivalent untreated cells, PTX treatment led to a dramatic reduction in the measurable LPA-mediated response seen in the rGi1HA(N) subunit and the rGi1 myc(N) subunit cotransfected cell lines (77 ± 8% and 73 ± 5%, respectively) and abolished the response in the wild-type rGi1 subunit-cotransfected cell line. In contrast, PTX pretreatment of the cells transiently transfected with the hEdg2 receptor-rGi1 fusion protein was without effect on either the efficacy or potency values generated.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   LPA-induced [35S]GTPS binding mediated, in the presence of 1 µM GDP, by (A) hEdg2 receptor + rGi1HA(N) subunit, (B) hEdg2 receptor + rGi1 myc(N) subunit, (C) hEdg2 receptor + wild-type rGi1 subunit, and (D) hEdg2 receptor-rGi1 fusion protein transiently transfected HEK cells, in the absence () or presence () of 100 ng/ml PTX. Data are expressed as stimulated  basal dpm. Each curve is a composite of three or more experiments, mean ± S.E. PTX treatment completely abolished the measurable LPA response in the wild-type rGi1 cotransfected cell line and reduced the response seen in the rGi1 HA(N) and the rGi1 myc(N) cotransfected cell lines when compared with the equivalent untreated cells (77% and 73%, respectively). In contrast, PTX pretreatment of hEdg2 receptor-rGi1 fusion protein-expressing cells was without effect on either the efficacy or potency values generated.

The potency and efficacy values (with respect to LPA) for several lysophospholipids in the PTX-pretreated hEdg2 receptor-rGi1 fusion protein-expressing HEK cells are given in Table 2. LPA was, by over an order of magnitude, the most potent compound tested. LPG was marginally less efficacious. LPE (oleoyl) and S-1-P were able to induce hEdg2 receptor-mediated [³⁵S]GTPS binding albeit with much reduced potency and efficacy. LPC (oleoyl) and LPE (palmitoyl) were inactive at up to 100 and 30 µM concentrations, respectively.

View this table: [in this window] [in a new window]   TABLE 2 Potency and efficacy values for LPA and analogs using agonist-induced [35S]GTPS binding mediated by the pertussis toxin-insensitive hEdg2 receptor-rGi1 fusion protein Potency (pEC50) and efficacy values (Emax) for LPA and related phospholipids in the agonist-induced [35S]GTPS binding assay were carried out in membranes prepared from PTX treated (100 ng/ml for 20 h before harvesting) hEdg2 receptor-rGi1 fusion protein transiently expressing HEK cells. Efficacy values are the maximal stimulation achieved and are presented as a percentage of the maximal LPA response. Results are arithmetic means ± S.E. of 3 experiments.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Edg receptors are classified into two subfamilies, depending on their amino acid sequence similarity (within a subgroup, 44-52%; between subgroups, 27-34%) and their preferred endogenous ligand. Edg1, -3, -5, and -8 receptors are activated by S-1-P and Edg2, -4, and -7 receptors by LPA. The orphan Edg6 receptor displays a 30 to 40% amino acid identity with both subgroups. These receptors lack the commonly seen disulfide bridge between extracellular loops 1 and 2, which might be relevant in providing a suitable binding pocket for the bulky ligand.

Members of this family activate multiple secondary messenger pathways (Gi, Gq/11, G12/13, and/or Rho coupled). The effector system(s) activated could conceivably depend on the receptor subtype, the ligand concentration, and, therefore, the level of receptor activation as well as the internal milieu of the host cell type in both native tissue and recombinant receptor cell lines.

Edg receptors are expressed in most mammalian cells, with each subtype having a distinct distribution pattern, raising the possibility of tissue-specific biological roles that could be explored in drug discovery programs. Several subtypes are often expressed within the same tissue and possibly on differing cells types. Expression on the same cell may allow subtle control mechanisms that increase the repertoire of endogenous ligands, perhaps involving synergism between phospholipids.

Exploration of the physiological/pathophysiological roles of these receptors requires the identification of selective compounds, preferably with a range of intrinsic efficacies. Here we describe an assay that allows efficient profiling of compounds at the hEdg2 receptor.

The complex intracellular pathways activated by Edg receptors offer several potential biochemical markers of receptor activation. However, monitoring an early event in the signaling cascade should increase the probability of developing an assay design, which could be applied to all Edg receptors, rather than an assay based on a more downstream signaling event, when divergence between receptor subtypes is more likely to have occurred. Agonist-induced [³⁵S]GTPS binding, employing recombinant receptor cell lines, is an established in vitro functional assay for GPCRs and fulfills this condition. The assay measures the accumulation of the nonhydrolyzable GTP analog, [³⁵S]GTPS, which displaces -subunit-bound GDP after receptor activation.

The ubiquitous responsiveness of mammalian cells to LPA restricts the availability of suitable host cell lines, because responses might be due to endogenous receptor activation. We therefore took the approach of developing an assay with a PTX-insensitive hEdg2 receptor-rGi1 fusion protein-expressing cell line (Bertin et al., 1994; Wise et al., 1997b).

Transient transfection of the hEdg2 receptor into HEK cells led to a modest increase in the maximal LPA-mediated response compared with mock-transfected cells. Not unexpectedly, an improved stimulation was seen after coexpression with the wild-type rGi1 subunit. This response was further enhanced by replacing wild-type rGi1 subunits with either HA or myc N-epitope tagged rGi1 subunits. These tags were inserted near the N terminus to minimize possible interference with the secondary signaling cascade and in particular with the conformational changes required for GTPS acceptance in the -subunit binding pocket. The simplest explanation for this enhanced response is a more efficient subunit expression. However, because a similar effect was also seen with the poorly coupled h5-HT_5A receptor but not the efficiently coupled hEdg4 receptor (unpublished data), it is probably due to improved receptor-subunit coupling.

After ligand activation, GPCRs interact with G-proteins to promote an -subunit "open" conformation. The Gi1 subunit N-terminal region is involved in contact sites with the receptor, the effector binding surface, and the - subunit interface (Wall et al., 1995; Lambright et al., 1996) in addition to cell membrane anchorage (Wise and Milligan, 1997). Hence, enhanced efficacy seen with the N-terminal-tagged subunits might be due to either a direct effect at contact sites involved in the conformational changes (Kisselev et al., 1995a,b; Taylor et al., 1996) or to improved maneuverability of the subunit at its point of attachment to the cell membrane, i.e., opposite to fusion protein compartmentalization effects (see below). Similar effects on compound efficacy were reported after _2A-adrenoceptor-Cys³⁵¹  Gly Gi1 fusion protein 12 coexpression studies and argued to be due to enhanced -subunit interactions allowing more effective transduction (Wise et al., 1997a; Wise and Milligan, 1997).

The most efficacious LPA responses were seen after mutant rGi1 receptor cotransfection and hEdg2 receptor-rGi1 fusion protein transfection, although the fusion protein gave a greater response. The direct in-frame fusion of the N terminus of the -subunit to the C terminus of the receptor was achieved by substituting the stop codon of the receptor with alanine and maintaining the -subunit initiator methionine (Bertin et al., 1994; Wise et al., 1997a) and the PTX insensitivity by mutation of the -subunit cysteine 351 to glycine (Hunt et al., 1994; Wise et al., 1997b). These results suggest that the enhanced response seen with the fusion protein is, at least in part, due to the cysteine 351 to glycine mutation.

The LPA response seen after transfection of the fusion protein was unaffected by pretreatment of the cells with PTX (Fig. 3D). In contrast, PTX pretreatment in coexpression studies resulted in a dramatic reduction in the maximal stimulation (Fig. 3, A-C). The residual activity is possibly due to interference of the mutation in the efficiency of ADP-ribosylation.

Published studies (e.g., Sautel and Milligan, 1998), with _2A-adrenoceptors, indicate that fusion proteins result in compromised coupling efficiency possibly due to physical restrictions inhibiting protein conformational changes or effects on the compartmentalization of signaling elements within specific domains of the plasma membrane interfering with signal transduction (Huang et al., 1997; Sautel and Milligan, 1998). This study supports these findings, because the LPA potency with the fusion protein was 5- and 2-fold weaker than that seen with the wild-type and mutant rGi1 cotransfection studies, respectively.

Finally we have investigated the ability of non-LPA phospholipids to activate the Edg2 receptor. All compounds tested were shown to have an LPA content of less than 0.2%. We have demonstrated with this fusion protein that attaching groups to the phosphate of LPA does not prevent hEdg2 receptor activation. Glycerol and ethanolamine, but not choline, are tolerated albeit with reduced potency and efficacy. Consistent with the findings of other groups (see below) the nature of the carbon chain seems to be critical, because 30 µM LPE (palmitoyl; C 16:0) in contrast to LPE (oleoyl; C18:1) was inactive. Additionally, the study indicates that two acyl side chains are also tolerated. S-1-P yielded an efficacy of 22% compared with LPA.

Erickson et al. (1998), with a reporter gene assay monitoring the pheromone response pathway in Saccharomyces cerevisiae, found LPA (18:1 oleoyl) to have an EC₅₀ of 20 to 30 µM, LPE to be weakly active, and S-1-P to be inactive. These weak activities probably reflect poor coupling of the mammalian receptor to the yeast G-protein and highlight the increased sensitivity of the approach taken in the present study. The yeast study also indicates the importance of ligand chain length. LPA analogs with increased chain length were more efficacious, and those with 6 and 10 carbon chains were inactive.

Consistent with the present study, Fukushima et al. (1998) report LPA (1 µM) induced [³⁵S]GTPS binding after transient transfection of Edg2 receptors into Rh7777 rat hepatoma cells and stable transfection into B103 neuroblastoma cells. LPC, LPE, and LPG (1 µM) were inactive. These apparent discrepant findings might be explained by increased responsiveness and hence sensitivity of the fusion protein assay.

Finally, with hEdg2 receptors transiently expressed in Jurkat cells and a serum response element-driven luciferase expression reporter gene assay, An et al. (1998) report an EC₅₀ value of 10 nM for LPA. This assay therefore appears to have increased sensitivity; however, the authors also report that the Jurkat cells might express an "as yet unidentified LPA receptor subtype," which would prove problematic when interpreting data from novel structurally diverse Edg receptor mimetics. In addition, S-1-P, LPC, and LPE were inactive, although compounds were tested only at concentrations up to 1 µM unlike the present study where compounds were tested up to 30 to 100 µM.

The present study, demonstrating dose-dependent LPA-induced [³⁵S]GTPS binding in PTX-pretreated cells expressing the PTX-insensitive hEdg2 receptor-rGi1 subunit fusion protein, clearly indicates that LPA activates this receptor. In addition to this certainty, the assay has good sensitivity, measures of compound potency, and efficacy and is amenable to high throughput application. The limited number of compounds tested indicates that structural modifications to LPA, in particular attachments to the phosphate group and the length and number of acyl side chains, are tolerated for Edg2 receptor activation.

GPCR fusion proteins have been used for several applications, e.g., to elucidate complex interactions between members of signaling pathways and the nature of agonist efficacy under conditions of fixed receptor/G-protein stoichiometry (Bertin et al., 1994; Wise et al., 1997a,b). Here we demonstrate a further application that, with the use of appropriate -subunits (chimeric or otherwise), toxins, and assays, could be applied to other GPCRs that have a widespread endogenous expression in mammalian cells, e.g., thrombin, adenosine, UTP, and prostaglandin E receptors. Finally, the serendipitous finding that cotransfection of tagged/mutated rGi1 subunits enhance the response compared with the wild-type -subunit might be applicable to assay development for poorly coupled GPCRs.