Edg-2/Vzg-1 Couples to the Yeast Pheromone Response Pathway Selectively in Response to Lysophosphatidic Acid*

We have functionally expressed the human cDNA encoding the putative lysophosphatidic acid (LPA) receptor Edg-2 (Vzg-1) in Saccharomyces cerevisiae in an attempt to determine the agonist specificity of this G-protein-coupled receptor. LPA activated the pheromone response pathway in S. cerevisiaeexpressing Edg-2 in a time- and dose-dependent manner as determined by induction of a pheromone-responsiveFUS1::lacZ reporter gene. LPA-mediated activation of the pheromone response pathway was dependent on mutational inactivation of the SST2 gene, the GTPase-activating protein for the yeast Gα protein (the GPA1gene product). This indicates that, in sst2Δ yeast cells, Edg-2 can efficiently couple to the yeast heterotrimeric G-protein in response to LPA and activate the yeast mitogen-activated protein kinase pathway. The Edg-2 receptor showed a high degree of specificity for LPA; other lyso-glycerophospholipids, sphingosine 1-phosphate, and diacyl-glycerophospholipids did not activateFUS1::lacZ. LPA analogs including a cyclic phosphoester form and ether-linked forms of LPA activatedFUS1::lacZ, although fatty acid chains of 6 and 10 carbons did not activate FUS1::lacZ, suggesting a role for the side chain in ligand binding or receptor activation. These results indicate that Edg-2 encodes a highly specific LPA receptor.

The lysophospholipid lysophosphatidic acid (LPA, 1 1-acyl-snglycero-3-phosphate) has been shown to be an important extracellular signaling molecule in a variety of systems (1). LPA has been shown to induce mitogenesis in certain cell types, exert an anti-mitogenic effect in other cells types, activate platelets, activate MAP kinase, stimulate ion transport, block apoptosis, and induce morphological changes (Refs. 2-14; for a recent review on these functions, see Refs. 1 and 15). Recently, two putative receptors for LPA have been identified, suggesting that functionally different LPA receptors may exist that dictate the particular cellular response of LPA (16 -19). Most cell types respond to LPA, making it difficult to characterize the receptor dependence of a particular response to LPA since the response cannot be solely attributed to a single LPA receptor. In particular, it is difficult to assess ligand binding specificity of an LPA receptor because other lipid receptors may exist with overlapping ligand specificity. We have therefore used the yeast Saccharomyces cerevisiae to study the human LPA receptor Edg-2 (also called Vzg-1). S. cerevisiae contains no endogenous LPA receptors and is therefore a potentially useful organism in which to functionally express LPA receptors and analyze their ligand specificity. Other mammalian receptors have been functionally expressed in S. cerevisiae including the somatostatin receptor, the A 2a adenosine receptor and the ␤ 2 -adrenergic receptor (20 -22).
S. cerevisiae contains a heterotrimeric G-protein that is activated by mating factor binding to a specific receptor (23) (for review, see Ref. 24). Upon stimulation by an occupied receptor, the ␣ subunit of the heterotrimeric G-protein (G ␣ , the GPA1 gene product; Ref. 25) becomes bound to GTP and dissociates from the ␤␥ dimer (26). In yeast, it is the ␤␥ dimer that transduces the signal to Ste11 (the MEKK equivalent; Ref. 27) and Ste7 (the MEK equivalent; Ref. 28). The active GTP-bound G ␣ is inactivated by hydrolysis of GTP to GDP at which time, G ␣ can reassociate with G ␤␥ and attenuate the signal (25,26,29,30). The yeast MAP kinases, Fus3 and Kss1, activate a transcriptional activator, the STE12 gene product (31). Activated Ste12 in turn activates the transcription of several mating factor-inducible genes such as FUS1 (32). To study the Edg-2 receptor using the yeast G-protein/MAP kinase system, a strain was used that has a mutation in the FAR1 gene. This mutation has the effect of uncoupling the MAP kinase cascade from cell cycle arrest, allowing the yeast to continue growing during MAP kinase activation (33,34). The strain used also contains the bacterial lacZ gene fused to the mating pheromone-inducible promoter from the FUS1 gene, which allows quantification of the receptor-ligand interaction. We mutationally inactivated SST2 gene in this strain to increase the sensitivity of the strain to G-protein activation. The SST2 gene encodes a GTPase-activating protein for the G ␣ subunit (35). By inactivating the SST2 gene product, G ␣ remains in the GTP-bound state longer and thus increases the steady-state concentration of the signal transducing ␤␥ dimer.
In this report, we show that Edg-2 expressed in yeast efficiently couples to the endogenous heterotrimeric G-protein in response to LPA. Edg-2 does not respond to other lyso-glycerophospholipids or to diacyl-glycerophospholipids such as phosphatidic acid or to sphingosine 1-phosphate, two phospholipids reported to share receptors with LPA in certain cell types (36,37). The results are consistent with Edg-2 being a functional, specific LPA receptor. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Polymerase Chain Reaction, Subcloning, and Yeast Transformation-The Edg-2 coding region was amplified by RT-PCR using Pfu DNA polymerase under conditions described by the supplier (Stratagene). The template for RT-PCR was cDNA (5 ng) that was reverse transcribed from human fetal brain total RNA (CLONTECH) using SuperScript II Reverse Transcriptase as described by the supplier (Life Technologies, Inc.). 1 M each of the following primers: FP, 5Ј-GCGA-TAGGATCCATCATGGCTGCCATCTCTACTTC-3Ј, and RP, 5Ј-GCGA-TACTCGAGCTAAACCACAGAGTGATCATTGC-3Ј, were used for RT-PCR. The primers were designed based on the human edg-2 cDNA sequence submitted to GenBank by Zondag and Moolenaar (accession no. Y09479) and included restriction site extensions for subcloning into the pYEUra3 vector (Stratagene). This placed the cDNA under the control of a galactose-inducible promoter (UASgal). The resulting plasmid was used to transform JEY5 by the lithium acetate method.
Oligonucleotide Synthesis and DNA Sequencing-RT-PCR primers and DNA sequencing primers were synthesized by the phosphoramidite method with an Applied Biosystems model 394 synthesizer, purified by polyacrylamide gel electrophoresis and desalted on Sep-Pak C 18 cartridges (Waters Associates, Milford, MA). The edg-2 cDNA was sequenced in pYEUra3 by the dideoxy chain termination method using the T7 Sequenase 7-deaza-dGTP sequencing kit as described by the supplier (Amersham Life Science).
LacZ Assays in Response to Phospholipids-JEY5ϩpJE15 was grown on SC medium containing either 2% galactose or 2% glucose lacking uracil to an approximate optical density of 0.1-0.5 before the addition of lipid or ␣-factor. LPA and other glycerophospholipids (Avanti Polar Lipids) were dissolved in chloroform and dried down under argon immediately before experiments and resuspended in BBS/EDTA (50 mM NH 4 HCO 3 , 104 mM NaCl, 250 M EDTA⅐2Na, pH 7.63) at 20 mM with sonication until the solution was clear. Sphingosine 1-phosphate (Matreya) was resuspended in ethanol/water (9:1) pH 3.0 immediately before use. Cyclic LPA and ether-linked forms of LPA were synthesized in G. Tigyi's laboratory. 2 Fatty acid-free bovine serum albumin was obtained from Sigma and used at 0.1 mg/ml in BBS/EDTA. Cells were grown for the indicated time (7 h for dose-response experiments) before assaying. 100 l of yeast culture were then added to 900 l of assay buffer (60 mM Na 2 HPO 4 , 40 mM NaH 2 PO 4 , 10 mM KCl, 0.1 mM MgSO 4 , pH 7.0, plus 270 l of ␤-mercaptoethanol/liter) plus 50 l of 0.1% SDS ϩ three drops of chloroform. Cells were vortexed for 10 s and incubated for 5 min at 28°C. 200 l of 4 mg/ml o-nitrophenol-␤-D-galactopyranoside (Sigma) was added, and the reaction was incubated for 30 min at 28°C. The assay was stopped by the addition of 500 l of 1 M Na 2 CO 3 . Color development was measured at A 420 and normalized to A 600 . Units were expressed as Miller units.

Edg-2 Expressed in Yeast Efficiently Couples to the Endogenous Heterotrimeric G-protein in a Time-and Dose-dependent
Manner-Yeast contain multiple MAP kinase cascades that are functionally analogous to the mitogen-activated protein kinase cascade in mammalian systems (38 -41). A schematic of the S. cerevisiae pheromone response pathway and the relevant genetic components are shown in Fig. 1. The parental yeast strain, SY2069, contains the FUS1 promoter fused to lacZ and HIS3 integrated into different chromosomal loci and carries a far1 Ϫ allele. The FAR1 gene product is required for cell cycle arrest after exposure to mating pheromone (see Fig. 1). By mutating this gene, the cells are able to grow in the presence of activated MAP kinase. In addition, a null mutation in the SST2 gene was created (see "Materials and Methods") because it has been previously reported that the somatostatin receptor can efficiently couple to the endogenous yeast heterotrimeric Gprotein after mutationally inactivating the SST2 gene (20). We created a null allele in the SST2 gene using pBC14 (see "Materials and Methods" for detailed description of the null allele construction) and tested the resultant strains for the supersensitive phenotype by assaying lacZ activity in response to ␣-factor (data not shown). One sst2 Ϫ strain was named JEY5 and used in subsequent experiments.
The edg-2 gene was expressed in either SY2069 or JEY5 under the control of a galactose-inducible promoter (UAS GAL ) in the shuttle vector pYEUra3 (CLONTECH). To test the effects of the SST2 gene product on the Edg-2 response to LPA, JEY5 (sst2⌬) expressing the Edg-2 receptor was compared with the parental SY2069 strain (SST2 ϩ ). Fig. 2 shows that the SST2 ϩ strain was unresponsive to LPA, whereas the sst2 Ϫ derivative was activated by 200 M LPA. This implies that activation of FUS1::lacZ by LPA is dependent on G-protein coupling because a G-protein-independent activation of FUS1::lacZ would not be effected by SST2. Since antibodies specific to the Edg-2 receptor are currently unavailable, the levels of Edg-2 protein expression were not determined. However, as a control, JEY5ϩEdg-2 was assayed in 2% glucose. Under these conditions, the GAL1 promoter would be glucoserepressed and not express the edg-2 gene (for review on glucose repression, see Ref. 42). On glucose, JEY5ϩEdg-2 cells did not respond to LPA, suggesting that the response was dependent on the receptor expression (see Fig. 2). To further characterize the LPA response to Edg-2, the time and dose dependence of LPA activation was tested. As seen in Fig. 3A, LPA resulted in a time-dependent increase in lacZ activity as compared with vector control with a maximal 4-fold stimulation of activity at 7 h. The dose response of Edg-2 to LPA is shown in Fig. 3B (EC 50 ϭ 20 -30 M). LPA concentrations above 600 M could not be tested due to toxicity. However, the dose-response curve can be seen to plateau, suggesting that maximal activity has been reached. The kinetics of FUS1::lacZ induction by LPA and ␣-factor were compared in Fig. 3C. As seen, both LPA and ␣-factor continued to induce lacZ activity over the time course tested, although ␣-factor gave a much more robust response at a much lower concentration, suggesting the Edg-2 receptor simply does not couple to the endogenous G-protein as well as the ␣-factor receptor (the STE2 gene product).
Edg-2 Responds Selectively to LPA and Not to Other Lysoglycerophospholipids or to Diacyl-glycerophospholipids-Yeast do not have endogenous receptors for phospholipids that couple to the pheromone response pathway, so it was of interest to determine the agonist specificity of Edg-2 using this system. Lysophosphatidylethanolamine (LPE), -serine (LPS), -glycerol (LPG) and -choline (LPC) and sphingosine 1-phosphate were tested over the same dose range as LPA for activation of FUS1::lacZ. As shown in Fig. 4A, no other lyso-glycerophospholipids or sphingosine 1-phosphate activated FUS1::lacZ as well as LPA at concentrations up to 200 M, the highest concentration tested due to toxicity. The results of a similar experiment testing the effects of the diacyl-glycerophospholipids is seen in Fig. 4B. In this experiment, no diacyl-glycerophospholipid activated Edg-2.
Dependence of Acyl-chain Length and Structural Variations of LPA on Edg-2 Activation-Since Edg-2 appears to show specificity for LPA, the dependence of the acyl-chain length on Edg-2 activation was investigated. Five forms of LPA that varied in chain length and degree of saturation were tested: 18:1 (oleoyl), 18:0 (stearoyl), 16:0 (palmitoyl), 10:0 (capryl), and 6:0 (caproyl). Fig. 5A shows that forms of LPA with chain length of 16 carbons and greater activated Edg-2 but with a preference to the 18:1 (oleoyl) form. The shorter 10:0 and 6:0 forms were inactive. Two structural analogs of LPA were then tested: an 18:1 and a 16:0 2,3-cyclic phosphate form (cLPA) and  (43). The data shown in Fig. 5B shows that cLPA activates FUS1::lacZ with similar potency to LPA, whereas ether-linked 18:1 LPA was slightly less active than the 18:1 cyclic and ester-linked forms. As with the 18:1 forms, the 16:0 ester and cyclic forms were equally active although less active than the 18:1 forms, probably due to solubility. DISCUSSION In this report, we show that the edg-2 gene product, a putative LPA receptor also reported as vzg-1, couples to the yeast heterotrimeric G-protein and activates a MAP kinase when bound to LPA. The response to LPA was quantitated by using the lacZ gene fused to the FUS1 promoter, a mating phero-mone-inducible gene promoter. The yeast strain used in this report was able to grow in the presence of activated G-protein due to a mutation in the FAR1 gene. This mutation has the phenotypic effect of uncoupling G-protein/MAP kinase activation from cell cycle arrest.
The response to LPA is dose-and time-dependent and requires that the SST2 gene be mutationally inactivated. SST2 encodes a GTPase-activating protein for the GPA1 gene product, the G ␣ subunit required for mating pheromone signal transduction. The effect of inactivating SST2 is that Gpa1 remains in the GTP-bound state longer and thus permits signaling through the ␤␥ dimer to proceed at a higher rate, resulting in a higher signal from the receptor. The response of Edg-2 to LPA in this yeast-based assay is markedly reduced compared with the LPA response in mammalian cells. Our results show that LPA activate FUS1::lacZ with an EC 50 of 20 -30 M. In mammalian systems, 1 M LPA is sufficient to induce tyrosine phosphorylation and depolarize membranes in Rat-1 fibroblasts (6,11). This discrepancy is most likely due to receptor/G-protein coupling as suggested in Fig. 3C. It may be possible to increase the response by using a mammalian G ␣i subunit or a Gpa1/G ␣i chimera expressed in place of the endogenous GPA1 gene product (20). However, the response of the FUS1::lacZ reported is specific to cells expressing the Edg-2 receptor as yeast expressing other related receptors such as Edg-1, Edg-3, and H218 as well as the unrelated LPA receptor, PSP24, did not respond to LPA (data not shown).
This yeast cell-based assay for Edg-2 was used to determine the agonist specificity of the receptor for different forms of LPA as well as other glycerophospholipids. The advantage of this system is that yeast contain few G-protein-coupled receptors. It is therefore a simple task to show that the response of the Edg-2 receptor to a particular phospholipid is dependent on the expression of the receptor, since it is expressed from a galactose inducible promoter. This is in contrast to mammalian cells in which the identity and distribution of LPA and other glycerophospholipids receptors is unclear. Our results show that Edg-2 specifically responds to LPA. Edg-2 does not respond to other lyso-glycerophospholipids or to diacyl-glycerophospholipids, in particular phosphatidic acid or to the related lipid messenger sphingosine 1-phosphate. These results are consistent with the response of neocortical neuroblasts, which express the Edg-2 receptor and do not respond to LPC, LPE, LPG, or phosphatidic acid (16). In platelets, LPA shares a surface receptor with sphingosine 1-phosphate (37), and, in human monocytes, phosphatidic acid and LPA both act as chemoattractants and cross-desensitize one another, suggesting that they act through a common receptor (36). Based on the result of our assay, neither of these effects are mediated by the Edg-2 receptor.
Other results show that the acyl-chain length does have an effect on the ability of LPA to activate FUS1::lacZ. It is most likely that the short chain forms of LPA do not bind to the receptor as, in competition experiments, the 6:0 caproyl form did not attenuate the ability of the 18:1 oleoyl form to activate FUS1::lacZ (data not shown). A cyclic phosphoester form of LPA activated FUS1::lacZ as well as ester-linked LPA, whereas an ether-linked form was less active. Interestingly, in A431 cells, the ether-linked LPA was less potent at activating calcium mobilization than was the corresponding ester-linked form (44). Our results demonstrate that expression of Edg-2 in yeast faithfully reconstitutes many of the key properties of an LPA receptor. We conclude that Edg-2 encodes an LPA receptor. This yeast cell-based system should prove useful for studying LPA analogs as well as identifying novel agonists and antagonists of this LPA receptor.