A Reporter Gene Assay for High-Throughput ... - Science Direct

Analytical Biochemistry 284, 316 –326 (2000) doi:10.1006/abio.2000.4698, available online at http://www.idealibrary.com on

A Reporter Gene Assay for High-Throughput Screening of G-Protein-Coupled Receptors Stably or Transiently Expressed in HEK293 EBNA Cells Grown in Suspension Culture Yves Durocher,* ,1 Sylvie Perret,* Eric Thibaudeau,* Marie-Helene Gaumond,* Amine Kamen,* Rino Stocco,† and Mark Abramovitz† *Bioprocess Sector, Biotechnology Research Institute, 6100 Royalmount Avenue, Montreal, Quebec, Canada H4P 2R2; and †Department of Biochemistry and Molecular Biology, Merck Frosst Center for Therapeutic Research, P.O. Box 1005, Pointe-Claire-Dorval, Quebec, Canada H9R 4P8

Received February 22, 2000

We describe in detail a robust, sensitive, and versatile functional assay for G-protein-coupled receptors (GPCRs) expressed in human embryonic kidney (HEK) 293-EBNA (Epstein–Barr virus nuclear antigen) (designated 293E) cells. The ability to grow these cells in suspension, in conjunction with the use of the secreted form of the human placental alkaline phosphatase (SEAP) as the reporter enzyme transcriptionally regulated by 5-cyclic AMP (cAMP) response elements (CREs) (Chen et al., Anal. Biochem. 226, 349 –354 (1995)), makes this CRE-SEAP assay potentially attractive for high-throughput screening (HTS). A 293E clonal cell line, stably transfected with the CRE-SEAP plasmid, was initially characterized with compounds known to activate intracellular signal transduction pathways similar to those activated by GPCRs. Forskolin and cAMP analogues were potent at inducing SEAP expression but calcium ionophores (A23187 and ionomycin) were without effect. The forskolin response was also potentiated by the protein kinase C activator phorbol myristate acetate as well as the phosphodiesterase inhibitor isobutylmethylxanthine. Previously established cell lines expressing the G ␣scoupled DP or the G ␣q-coupled-EP 1 prostanoid receptors were stably transfected with the reporter gene construct and clones were selected based on their ability to secrete SEAP upon agonist challenge. Pharmacological characterization of the DP and EP 1 receptors displayed a similar rank order of potency for several known prostanoids and related compounds to that

1 To whom correspondence should be addressed. Fax: (514) 4966785. E-mail: [email protected].

316

previously reported using classical binding assays or other functional assays. The CRE-SEAP assay was also used to characterize the EP 1 receptor antagonists SC51322, SC-51089, and AH6809. In summary, we have established a reporter gene assay for GPCRs that couple to both G ␣s and G ␣q and is amenable to HTS of both agonists and antagonists. © 2000 Academic Press Key Words: prostanoid receptors; cyclic AMP response element; secreted alkaline phosphatase; gene transcription; antagonists.

G-protein-coupled receptors (GPCRs) 2 are the largest known family of transmembrane receptors, comprising over 200 cloned members (1). They represent important therapeutic targets in the pharmaceutical industry; in fact of the top 100 pharmaceutical drugs, 18 of them interact with GPCRs as either agonists or antagonists (Nature 384(6604 Suppl), 1–5 (1996)). In the past, drug-screening strategies have mainly relied upon binding assays on membrane preparations using radiolabeled ligands or assays measuring second mes2

Abbreviations used: cAMP, cyclic adenosine monophosphate; CaMK, calcium/calmodulin kinase; CRE, cyclic AMP response element; CREB, cyclic AMP response element binding protein; GPCR, G protein coupled receptor; HEK, human embryonic kidney; HTS, high-throughput screening; IBMX, isobutylmethylxanthine; OA, okadaic acid; PGD 2, prostaglandin D 2; PGE 2, prostaglandin E 2; PGF 2␣, prostaglandin F 2␣; PKA, protein kinase A; PKC, protein kinase C; PMA, phorbol myristate acetate; pNPP, para-nitrophenyl phosphate; PP-1, protein phosphatase-1; SEAP, secreted alkaline phosphatase; Sp-cAMPS, Sp-adenosine-3⬘,5⬘-cyclic monophosphorothioate; VIP, vasoactive intestinal peptide. 0003-2697/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

REPORTER GENE ASSAY FOR G-PROTEIN-COUPLED RECEPTORS

sengers or enzymes modulated by these receptors. The ability to conduct high-throughput screens based upon functional activity of a given GPCR in a cell-based assay offers a more direct way of identifying lead agonists or antagonists. Activation of GPCRs can impact signaling cascades ultimately leading to transcriptional modulation. Cellbased assays relying on transcriptionally controlled reporter genes are well suited to monitor the cellular response induced by GPCRs. Indeed, these assays are quite attractive as in addition to their high-throughput capabilities, they allow for discrimination between agonistic and antagonistic compounds as well as between partial, full, or inverse agonists. They might also provide some information vis-a`-vis toxicity of the studied compounds as the assay is performed in a cellular context. Three major signal transduction pathways mediated by G-protein ␣ subunits are known to be modulated by GPCRs (see Ref. 2 for a review), namely, stimulation (G ␣s) or inhibition (G ␣i) of adenylyl cyclase or activation of phospholipase C (G ␣q). Activation of these effectors results in the modulation of the second messengers cAMP, inositol 1,4,5-trisphosphate (IP 3), diacylglycerol (DAG), and calcium, which, in turn, can regulate protein kinase activities, ultimately leading to phosphorylation and hence activation of various transcription factors. One of the most widely used response elements in reporter gene assays is the cyclic AMP response element (CRE), a pivotal target for these signaling pathways. Indeed, the transcription factor CREB (cAMP response element binding protein), a major regulator of CREs, can be activated (phosphorylated) by protein kinase A (PKA) and by some members of the calcium/ calmodulin kinase (CaMK) family in vitro (for a review, see Ref. 3). Chen and co-workers (4) showed that both G ␣s- and G ␣q-coupled receptors could signal through CREs; however, it is still unclear as to how G ␣q-coupled receptors activate CREB. Various reporter genes whose transcription is under the control of CRE-containing promoters have been used to monitor cell surface receptor activation in a number of different cell lines (for reviews, see Refs. 5, 6). In this paper we describe a robust cell-based reporter gene assay for GPCRs. For assay development we chose the human-derived 293E cell line that is ideally suited to stable or transient heterologous expression of GPCRs and is easily grown in suspension culture, a key feature for increased throughput. We have selected the secreted form of the human placental alkaline phosphatase (SEAP) (7) under the control of a modified fragment of the vasoactive intestinal peptide (VIP) promoter containing a total of five CREs (4). This reporter gene was chosen since it has several advantages over other intracellular enzymes typically used as reporter

317

genes. The G ␣s-coupled DP and G ␣q-coupled EP 1 prostanoid receptors stably expressed in 293E cells were used to validate the assay for both agonists and antagonists. Combining key features from previously described reporter gene assays with new ones makes this assay highly attractive for both basic pharmacological studies and HTS applications. MATERIALS AND METHODS

Reagents. para-Nitrophenyl phosphate (pNPP) was from Diagnostic Chemicals Ltd. (Charlottetown, PEI, Canada), L-homoarginine, diethanolamine, and Pluronic acid F-68 (F68) were from Sigma–Aldrich (Oakville, ON, Canada), and A23187, AH6809, dibutyryl-cAMP, forskolin, isobutylmethylxanthine (IBMX), okadaic acid (OA), phorbol myristate acetate (PMA), prostaglandin E2 (PGE 2), PGD 2, PGF 2␣, SC-51089, SC-51322, Sp-adenosine-3⬘,5⬘-cyclic monophosphorothioate (Sp-cAMPS), and U46619 were from BioMol (Plymouth Meeting, PA). Sulprostone was from Cayman Chemical (Ann Arbor, MI) and iloprost from Amersham Life Science (Oakville, ON, Canada). Culture media, geneticin, Lipofectamine, and Lipofectamine 2000 were from GIBCO/BRL (Burlington, ON, Canada) Cells. HEK293-EBNA (293E) cells (Invitrogen) were maintained in suspension culture in shake flasks or spinner flasks (100 –130 rpm) in low-calcium hybridoma serum-free media (LC-HSFM) containing 1% ironenriched bovine calf serum (BCS, Intergen, Purchase, NY), 0.1% F68, and 10 ␮g/ml geneticin in a humidified incubator at 37°C with 5% CO 2. Construction of CRE-SEAP plasmid. The pSEAPBasic plasmid (Clontech, Palo Alto, CA) was linearized with ClaI and overhangs blunted with Klenow. SEAP was then released with KpnI and subcloned in pcDNA3 (Invitrogen, Carlsbad, CA) linearized with KpnI and EcoRV. The pcDNA3/SEAP construct was then digested with BglII and NotI and the SEAP DNA insert was cloned in pZeoSV2(⫹) (Invitrogen) digested with BamHI and NotI, resulting in plasmid pZeo/SEAP. The VIP promoter fragment (⫺93 to ⫹146; Ref. 33) containing four additional synthetic CREs was released from the pCRE/␤-Gal plasmid (generous gift of Dr. Roger Cone) by HindIII digestion and ligated in HindIIIlinearized pZeo/SEAP, resulting in the plasmid pCRESEAP. Correct orientation and functionality of the promoter were verified by sequencing and forskolin inducibility following transient transfection in 293E cells, respectively. Development of stable cell lines. Stable transfection of pCRE-SEAP in 293E cells or 293E cells stably expressing the human DP (8) or EP 1 (9) prostanoid receptors was achieved by using Lipofectamine according to the manufacturer’s instructions. A population of stable transfectants was obtained by selection with anti-

318

DUROCHER ET AL.

biotic over a period of 2 weeks. Clonal cell lines (293E/ EP 1/CRE-SEAP and 293E/DP/CRE-SEAP) were then isolated by plating one cell per well in 96-well plates in LC-HSFM containing 2% BCS, F68, and antibiotics. The B max and K d for [ 3H]PGE 2 on the 293E/EP 1/CRESEAP clone were 1.51 ⫾ 0.10 pmol/mg and 6.3 ⫾ 1.1 nM, respectively (10), and 0.5 ⫾ 0.1 pmol/mg and 0.6 ⫾ 0.2 nM for [ 3H]PGD 2 on the 293E/DP/CRE-SEAP receptor (11). Agonist and antagonist assays for stable cell lines. Cells were plated in 96-well plates at densities varying between 5 ⫻ 10 4 and 1 ⫻ 10 5 cells/well (for 16 or 6 h of incubation, respectively) in 90 ␮L of Ham’s F12 supplemented with 0.2% BCS and 0.02% F68 (HBF) and challenged with 10 ␮L of 10⫻-concentrated compounds dissolved in HBF. For antagonist assays, compounds were preincubated for 15 min with cells prior to agonist addition. The SEAP assay was subsequently performed (see below). Transient transfection assays. For transient transfection assays in 96-well plates, cells were seeded the day before at 5 ⫻ 10 4 per well in 100 ␮l of LC-HSFM containing 1% BCS and 0.1% F68. Transfection was performed by the addition of 50 ␮L of OPTI-MEM containing 300 ng of plasmid and 800 nL of Lipofectamine 2000 (GIBCO/BRL) to each well. Fifteen nanograms of the green fluorescent protein (GFP) expression vector pIRESpuro-EGFP (Clontech) was included to normalize for transfection efficiency. After 6 h, 100 ␮L of LC-HSFM containing 1% BCS and 0.1% F68 was added and cells were incubated overnight. Medium was then replaced by 100 ␮L of HBF containing the appropriate compounds and cells were incubated for 16 h. The SEAP assay was subsequently performed (see below). SEAP assay. Following the incubation period of either stably or transiently transfected cells, culture medium, 50 ␮L, was transferred to a new 96-well plate and mixed with an equal volume of SEAP assay solution containing 20 mM para-nitrophenyl phosphate (pNPP), 1 mM MgCl 2, 10 mM L-homoarginine (optional; see below), and 1 M diethanolamine, pH 9.8. Absorbance was read at 410 nm at 1- to 15-min intervals to determine pNPP hydrolysis rates, which were expressed as an increase in absorbance units per minute. End point readings following the addition of ethylenediaminetetraacetic acid (EDTA, 10 mM final) to stop the reaction gave the same results. Heating samples at 65°C for 30 min to inhibit endogenous alkaline phosphatase activities (7) was not necessary, since no such activity was detectable in 293E cells (data not shown). Data analysis. For agonist and antagonist assays, EC 50 values were determined by nonlinear regression using Prism version 2.0 software (GraphPad, San Di-

ego, CA). Antagonist potency (K B value) was measured by Schild plot analysis. Agonist concentration–response curves were first generated in the absence and presence of fixed concentrations of antagonist. The EC 50 values for the agonist response were determined and the dose ratios (DR) calculated, i.e., EC 50 plus antagonist/EC 50 minus antagonist. Log DR-1 is expressed as a function of ⫺log[antagonist] to generate a Schild plot, which determines both the log K B (the intercept on the x axis) that is converted to K B and the slope. A slope of 1 is indicative of compounds that behave as competitive antagonists (produces rightward shifts in the agonist concentration–response curves without depressing of the maximal response to agonist). RESULTS

Characterization of the 293E/CRE-SEAP stable cell lines. Clonal 293E cell lines stably transfected with the CRE-SEAP reporter construct (Fig. 1) were initially screened for robust SEAP induction following a 6-h incubation with 10 ␮M forskolin, a direct activator of adenylyl cyclase. One of these highly responsive cell lines (293E/CRE-SEAP) displayed a signal-to-noise ratio greater than 10 (data not shown). The kinetics of SEAP accumulation in the culture medium following forskolin treatment in this cell line is shown in Fig. 2. SEAP activity was detectable 4 h following forskolin addition to 293E/CRE-SEAP cells and accumulated in the medium in a linear fashion from 6 to 16 h. SEAP production was also efficiently induced by the membrane-permeable cAMP analogue Sp-cAMPS or dibutyryl-cAMP (Fig. 3A). On the other hand, no induction was observed with 100 nM of the protein kinase C (PKC) activator PMA nor with the calcium ionophore A23187 (1 ␮M) or a combination of PMA and A23187. PGE 2 (10 ␮M) also induced a weak but consistent activation of the promoter (Fig. 3A; see below). As expected for a cAMP-mediated response, the nonspecific phosphodiesterase inhibitor IBMX increased the forskolin-induced maximal effect (by 40%) and potency (⬃5-fold). In a similar manner, PMA also increased the forskolin-mediated response (by 40%) and potency (⬃2fold); however, A23187 (1 ␮M) blocked completely the forskolin-induced response (Fig. 3B). The same inhibition was observed using 2 ␮M of the calcium ionophore ionomycin or 100 nM of the endoplasmic reticulum Ca 2⫹-ATPase inhibitor thapsigargin (data not shown). Protein phosphatase-1 (PP-1), the major regulator of CREB activity in cAMP-responsive cells, has been shown to induce transcriptional attenuation by dephosphorylation of CREB at Ser-133 (12, 13). The effect of okadaic acid (OA), a highly potent and specific inhibitor of PP-1, on forskolin-induced transcriptional activation was monitored (Fig. 3B). In the 293E/CRE-

REPORTER GENE ASSAY FOR G-PROTEIN-COUPLED RECEPTORS

319

SEAP plasmid. More than 75% of the clones were positive upon stimulation with forskolin, and of these, about 10% showed robust responses with signal-tonoise ratios of five to ten. Typically, the EP 1 and DP CRE-SEAP clones showed a similar activation window when challenged with iloprost and PGD 2, respectively (data not shown). The pharmacological profiles of the EP 1 and DP receptors were then assessed using various prostanoid agonists (Fig. 4). Figure 4A shows that PGE 2 was the most potent agonist tested against the human prostanoid EP 1 receptor, with an EC 50 value of 12.3 ⫾ 3.6 nM (n ⫽ 21), followed by iloprost, sulprostone, PGF 2␣, and U46619. PGD 2, up to 10 ␮M, was without effect on this receptor. This is similar to data obtained when measuring calcium fluxes using an aequorin reporter assay (14). For the DP receptor, PGD 2 was the best agonist (Fig. 4B), with an EC 50 value of 1.2 ⫾ 0.4 nM (n ⫽ 21). This EC 50 value is very similar to that obtained by measuring cAMP levels (EC 50 ⫽ 0.5 nM) and to the inhibition constant (K i) of 0.6 ⫾ 0.2 nM measured with [ 3H]PGD 2 (11). The rank order of potency was PGD 2 ⱖ PGE 2 ⬎ PGF 2␣ ⫽ U46619 ⬎ iloprost, similar to previously published data (11). The EP 3/EP 1-selective agonist sulprostone was totally inactive on this receptor. All these effects were prostanoid receptor mediated since the parental cell line was completely refractory to these agonists except for a slight response to micromolar concentrations of PGE 2 (Fig. 4C). This response to PGE 2 in HEK cells is likely to be mediated by an endogenously expressed EP 2 and/or EP 4 G ␣s-coupled

FIG. 1. Map of the CRE-SEAP reporter plasmid. Transcription of the human placental secreted alkaline phosphatase (SEAP) is under the control of a fragment (⫺93 to ⫹146) of the vasoactive intestinal peptide (VIP) promoter containing one endogenous and four synthetic CRE sequences (4). The sequence of the CRE-VIP promoter is also shown. In the sequence, the five CREs are boxed, the VIP promoter fragment (33) is underlined, the TATA box is shadowed, the transcriptional start site (G) is double underlined, HindIII sites are in bold, and the initiating methionine of SEAP is indicated.

SEAP cell line, OA did not potentiate the transcriptional response induced by forskolin, but rather induced a twofold decrease in SEAP production without affecting the EC 50 value. This effect was not due to inhibition of SEAP activity since it was insensitive to micromolar concentrations of this compound (data not shown). Pharmacological characterization of DP and EP 1 receptors using the CRE-SEAP assay. Over 20 zeocinresistant clones were screened for both the 293E/EP 1 and 293E/DP cells stably transfected with the CRE-

FIG. 2. Time course of SEAP production following forskolin activation. The 293E/CRE-SEAP reporter cell line (1 ⫻ 10 5 cells per well) was incubated with 10 ␮M forskolin for the indicated period of time prior to SEAP assay. SEAP activity was measured in a 50-␮L supernatant aliquot as described under Materials and Methods. The data are representative of two independent experiments ⫾ SE done in triplicate. SEAP activity is expressed as the increase in absorbance units at 410 nm per minute.

320

DUROCHER ET AL.

stant both for the EP 1 receptor challenged with iloprost (3.1– 6.5 nM) and for the DP receptor challenged with PGD 2 (0.2– 0.3 nM) (data not shown). The effects of DMSO and serum concentration in the assay were also assessed on the DP receptor (Fig. 6).

FIG. 3. Response of the 293E/CRE-SEAP reporter cell line to various cAMP inducers and other compounds. (A) The reporter cell line (1 ⫻ 10 5 cells/well) was incubated with the indicated compounds for 6 h. The following concentrations were used: forskolin, 10 ␮M; PMA, 100 nM; A23187, 1 ␮M; Sp-cAMPS, 500 ␮M; dibutyryl-cAMP, 1 mM; PGE 2, 10 ␮M. (B) The reporter cell line was preincubated for 15 min with vehicle alone (F), 100 ␮M IBMX (䊐), 100 nM PMA (E), 100 nM OA (〫), or 1 ␮M A23187 (‚) before the addition of forskolin. Cells were incubated for 6 h prior to SEAP activity measurement, which is expressed as a percentage of the activity obtained with 30 ␮M forskolin alone.

prostanoid receptor (8, 15) as the response is pertussis toxin insensitive (data not shown). The effect of the incubation period with agonists was evaluated with the EP 1 and DP receptors (Figs. 5A and 5B, respectively). EC 50 values obtained following 6- or 16-h stimulation were unchanged for both EP 1 and DP receptors. The sensitivity of the assay was tested by obtaining agonist dose–response curves at decreasing cell numbers, from 2 ⫻ 10 5 down to 2.5 ⫻ 10 4 cells per well, in which there was a linear decrease in SEAP activity with respect to cell number (data not shown). The calculated EC 50 values remained relatively con-

FIG. 4. Pharmacological characterization of the EP 1 and DP receptors. 293E/EP 1 (A) or 293E/DP (B) cells (5 ⫻ 10 4/well) stably transfected with the CRE-SEAP reporter plasmid and the 293E/CRESEAP reporter cells (5 ⫻ 10 4/well) (C) were incubated with PGE 2 (E), PGD 2 (〫), iloprost (‚), sulprostone (ƒ), PGF 2␣ (䊐), and U46619 (F) for 16 h. SEAP activity is expressed as a percentage of the activity obtained with 10 ␮M forskolin alone.

REPORTER GENE ASSAY FOR G-PROTEIN-COUPLED RECEPTORS

321

expression vectors (pcDNA3.1, Invitrogen) and challenged with agonists for 16 h. While the response to the G ␣s-coupled DP receptor was more robust than that obtained with the G ␣q-coupled EP 1 receptor, for each receptor the potency of its cognate ligand was similar to that obtained with the stably transfected clones. To assess whether this assay could be used to screen for antagonists, we evaluated the potency of three commercially available EP 1 receptor antagonists on the EP 1 and DP receptors. SC-51322 was a potent EP 1 receptor antagonist (Fig. 8A), with a pK B value of 8.8 (K B of 1.60 nM, n ⫽ 3) (Fig. 8C), similar to its K i value of 13.8 nM as calculated from a radioligand binding assay (10). The compound did not display any DP receptor antagonist activity (Fig. 8B). The second compound, SC-51089, had a pK B value of 6.9 (K B of 115 nM,

FIG. 5. Effect of incubation time on agonist dose–response curves. 293E/EP 1/CRE-SEAP cells (A) were incubated with PGE 2 (E), iloprost (䊐), or forskolin (〫) for 16 (5 ⫻ 10 4 cells/well; open symbols) or 6 h (1 ⫻ 10 5 cells/well; closed symbols). 293E/DP/CRE-SEAP cells (B) were treated as in A except that PGD 2 (䊐) was used instead of iloprost. SEAP activity is expressed as a percentage of the maximal activity obtained for each curve.

Figure 6A shows that the DMSO concentration had to be maintained at 1% or below in order to minimize its inhibitory effect on PGD 2-induced SEAP activity when cells were incubated for a 6-h period. However, when the cells were incubated overnight with concentrations of DMSO up to 1.5%, no significant inhibitory effect was observed (data not shown). On the other hand, increasing the serum concentration up to 1% enhanced the response to PGD 2 and forskolin by up to 33% (Fig. 6B). Neither DMSO (up to 1%) nor serum had any significant effects on the EC 50 values. Similar results were obtained for the EP 1 receptor (data not shown). To evaluate if this assay could be used with transiently expressed receptors, the 293E/CRE-SEAP cells were transfected with EP 1 (Fig. 7A) or DP (Fig. 7B)

FIG. 6. Effect of DMSO and serum in the CRE-SEAP assay. Plates were seeded with 1 ⫻ 10 5 293E/DP/CRE-SEAP cells and challenged for 6 h with PGD 2 in the presence of an increasing concentration of DMSO (A) or serum (B). (A) DMSO concentrations were 0 (F), 0.5 (E), 1 (䊐), and 2% (〫). (B) Serum concentrations were 0 (F), 0.2 (E), and 1% (䊐).

322

DUROCHER ET AL.

DISCUSSION

Transcriptional-based assays for the characterization and screening of GPCRs are becoming more widely used (18 –21; for a review, see Ref. 22). Our goal was to set up a reporter gene assay that was

FIG. 7. Transient expression of the EP 1 and DP receptors in the 293E/CRE-SEAP reporter cells. 293E/CRE-SEAP cells were transiently transfected with expression vectors coding for the EP 1 (F, A) and DP (F, B) receptors or with empty vectors (E, A and B) as described under Materials and Methods. Thirty hours later, cells were treated with iloprost or PGD 2 for 16 h and SEAP activity was measured. The calculated EC 50 values are 5 and 0.5 nM for the EP 1 and DP receptors, respectively.

n ⫽ 3) at the EP 1 receptor (Figs. 9A and 9C), which is one order of magnitude lower than the reported K i value of 1.3 ␮M (10). Again no activity at the DP receptor was detected (Fig. 9B). Finally, the third antagonist, AH6809, was more potent at the DP receptor (Figs. 10B and 10C) than at the EP 1 receptor (Fig. 10A). This antagonism of AH6809 at the DP receptor has also been reported by others (16, 17). At the highest concentration (10 ␮M) tested, AH6809 induced a slight but reproducible increase (25%) in the maximal SEAP activity attained at the EP 1 receptor with iloprost concentrations of 300 nM and above. The calculated pK B value for AH6809 on the DP receptor was 6.7 (K B of 180 nM) (Fig. 10C), which is fourfold lower than the recently reported K i value of 800 nM (17).

FIG. 8. Antagonism of SC-51322 on agonist-mediated activation of the EP 1 receptor. 293E/EP 1/CRE-SEAP (A) or 293E/DP/CRE-SEAP cells (1 ⫻ 10 5/well; B) were preincubated for 15 min with SC-51322 followed by a 6-h incubation in the presence of iloprost (A) or PGD 2 (B). SC-51322 concentrations (nM) used were 0 (F), 25 (䊐), 125 (〫), and 625 (‚). Schild plot analysis of SC-51322 on the EP 1 receptor is shown in C. Data represent the mean ⫾ SE of two independent experiments done in duplicate.

REPORTER GENE ASSAY FOR G-PROTEIN-COUPLED RECEPTORS

323

1. Cell line. For several reasons, 293E cells were selected as the cell line of choice for the reporter gene assay. First of all, in a report by Chen et al. (4), it was shown that of several cell lines tested, commonly used

FIG. 9. Antagonism of SC-51089 on agonist-mediated activation of the EP 1 receptor. 293E/EP 1/CRE-SEAP (A) or 293E/DP/CRE-SEAP cells (1 ⫻ 10 5/well; B) were preincubated for 15 min with SC-51089 followed by a 6-h incubation in the presence of iloprost (A) or PGD 2 (B). SC-51089 concentrations (␮M) used were 0 (F), 0.4 (䊐), 2 (〫), and 10 (‚). Schild plot analysis of SC-51089 on the EP 1 receptor is shown in C. Data represent the mean ⫾ SE of two independent experiments done in duplicate.

robust, flexible, and amenable to high-throughput screening. Three important parameters had to be considered: (1) the cell line, (2) the inducible element and number thereof, and (3) the reporter gene. These will be discussed in turn.

FIG. 10. Antagonism of AH6809 on agonist-mediated activation of the EP 1 and DP receptors. 293E/EP 1/CRE-SEAP (A) or 293E/DP/ CRE-SEAP cells (1 ⫻ 10 5/well; B) were preincubated for 15 min with AH6809 followed by a 6-h incubation in the presence of iloprost (A) or PGD 2 (B). AH6809 concentrations (␮M) used were 0 (F), 0.4 (䊐), 2 (〫), and 10 (‚). Schild plot analysis of AH6809 on the DP receptor is shown in C. Schild analysis of the EP 1 receptor is not shown because of the stimulating effect of AH6809 on the receptor at 10 ␮M. Data represent the mean ⫾ SE of two independent experiments done in duplicate.

324

DUROCHER ET AL.

to heterologously express GPCRs, 293 cells responded most robustly to forskolin stimulation. Secondly, 293E cells have been used successfully to express numerous GPCRs, including all eight members of the human prostanoid receptor family (10). Thirdly, 293E cells can be grown both as monolayer cultures, facilitating the isolation of clonal cell lines stably expressing GPCRs, and as suspension cultures. Importantly, when lowcalcium culture medium is used, 293E cells can be easily adapted to suspension growth in shaker or spinner flasks, allowing the production of large quantities of cells needed for HTS. Furthermore, the fact that 293E cells can be transfected with very high efficiency (over 70%, data not shown) directly in 96-well plates makes this system highly suitable for screening of transiently expressed GPCRs (Fig. 7). Finally, the cell line is human derived and is perhaps better suited to assaying human GPCRs. 2. Inducible element. The use of CREs in reporter gene assays has been described in numerous papers. Use of these cis-acting inducible elements allows one to potentially monitor GPCRs that couple to all three major signaling pathways, making this a universal reporter gene assay. It was first shown by Montmayeur and Borrelli (23) that CREs could regulate transcription of a reporter gene by GPCRs that coupled to G ␣i. Subsequently, the use of CREs was extended to both G ␣s-coupled (24) and G ␣q-coupled (4) GPCRs. The promoter used in our system, which contains 5 CREs, gave a robust signal-to-noise ratio following activation of both G ␣s- and G ␣q-coupled GPCRs. The number of CREs used in other reporter constructs, as reported in the literature, has varied from as few as 1 (25) to as many as 12 (26) in conjunction with various minimal promoters. In the case of 1 CRE, the response appeared to be somewhat weak, while in at least one report, 12 CREs were not appreciably better than 6 (26). Although a systematic study has yet to be performed, it would appear that 5 CREs is a reasonable number that works well in our hands with SEAP as the reporter (see below). 3. Reporter gene. In terms of the reporter gene, there are a number of advantages to using SEAP instead of ␤-galactosidase or other intracellular enzymes. SEAP is a stable protein that is secreted from the cell and accumulates in the culture medium, leaving intact viable cells that can be further used for experimentation if need be. In our assay system, SEAP accumulates in the culture medium for incubation periods up to 16 h, without affecting the pharmacological profiles of the receptors (as compared to a 6-h incubation). This allows the assay to be performed overnight, a characteristic that cannot be achieved using other intracellular reporter enzymes, since their activity is transient and rapidly saturates or declines within a

few hours following receptor stimulation (4, 25, 27). For example, the decline in luciferase levels is due to its susceptibility to proteolytic degradation (28). The use of a colorimetric assay to measure SEAP activity, with pNPP as the chromogenic substrate, is both a sensitive measure of GPCR activation and an easy assay to perform. With regard to sensitivity, the assay is more than sensitive enough to detect activation of GPCRs expressed even at relatively low levels. For example, the endogenously expressed ␤ 2-adrenergic receptor could be detected in the assay despite an expression level of only 100 –200 fmol/mg of protein (M. Bouvier, personal communication). However, if need be, sensitivity of the assay can be significantly increased by simply adding the phosphodiesterase inhibitor IBMX in the culture medium (see Fig. 3) or with the use of fluorescent or chemiluminescent substrates (29). In terms of ease of use, the assay does not require any extraction or solubilization steps and it allows for the use of a standard 96-well plate reader and 96-well plates. In addition, the fact that 293E cells do not secrete detectable endogenous alkaline phosphatase eliminates the need to heat the samples at 65°C or to add L-homoarginine, a general phosphatase inhibitor, to the assay buffer prior to performing the assay. All of these factors simplify automation and reduce costs. When stably transfected in the human 293E cell line, this CRE-SEAP reporter system could be used to characterize both the G ␣q-coupled EP 1 and G ␣s-coupled DP prostanoid receptors. Agonist stimulation of cells transiently or stably expressing the G ␣q-coupled FP and TP as well as the G ␣s-coupled EP 2, EP 4, and IP prostanoid receptors also gave rise to robust transcriptional activation of the CRE-SEAP construct (data not shown). The situation was less clear, however, for the G ␣i-coupled EP 3 receptor. Indeed, while stimulation of this receptor with low doses of agonists resulted in partial inhibition of forskolin-elicited SEAP production, higher doses clearly potentiated forskolin action (data not shown). The role of calcium in the signaling pathway linking the G ␣q-coupled EP 1 receptor to transcriptional activation of the CREs is not clear. While it has been shown that CaMK type I and IV phosphorylates and activates CREB in vitro (30), our data suggest that EP 1-mediated transcriptional activation of the CRE-containing promoter is calcium-independent. Indeed, the CRESEAP reporter system was not activated by increasing intracellular calcium using the calcium ionophore A23187 or ionomycin or by the endoplasmic reticulum Ca 2⫹-ATPase inhibitor thapsigargin. This lack of effect of A23187 on CRE-containing promoters has also been shown in CHO (24) and in JEG-3 cells (31) as well. Furthermore, EP 1-mediated transcriptional activation of the CRE-SEAP reporter system by iloprost was not

REPORTER GENE ASSAY FOR G-PROTEIN-COUPLED RECEPTORS

abrogated by pretreatment of the cells with 1 ␮M of the intracellular calcium chelator BAPTA-AM, a concentration that completely inhibited the calcium surge when monitored using the aequorin reporter system (M. Abramovitz, unpublished data). Another candidate that could be involved in the transcriptional activation of the CRE-containing promoter is PKC. However, while the phorbol ester PMA potentiates the forskolin-induced transcriptional response, it strongly repressed the EP 1 receptor mediated CRESEAP activation (data not shown), suggesting that PKC is perhaps also involved in desensitization of the EP 1 receptor. This has indeed been shown for the rat EP 1 receptor mediated calcium release following shortand long-term PMA treatment of CHO-expressing cells (32). At this time, the EP 1-mediated signal transduction cascade responsible for transcriptional activation of the CRE-SEAP reporter plasmid remains to be elucidated. In conclusion, we have established a robust, flexible, and sensitive transcriptional-based assay that is amenable to HTS for the human EP 1 and DP receptors and that is also applicable to other G ␣q- and G ␣s-coupled GPCRs in general. This has been accomplished by combining the advantageous features, enumerated above, of 293E cells, CREs, and the reporter gene SEAP.

10.

11.

12.

13.

14.

15.

16.

REFERENCES 1. Milligan, G. (1995) Signal sorting by G-protein-linked receptors. in Advances in Pharmacology (August, J. T., Anders, M. W., Murad, F., and Coyle, J., Eds.), Vol. 32, pp. 1–29, Academic Press, San Diego. 2. Gudermann, T., Kalkbrenner, F., and Schultz, G. (1996) Diversity and selectivity of receptor–G protein interaction. Annu. Rev. Pharmacol. Toxicol. 36, 429 – 459. 3. Montminy, M. (1997) Transcriptional regulation by cyclic AMP. Annu. Rev. Biochem. 66, 807– 822. 4. Chen, W., Shields, T. S., Stork, P. J., and Cone, R. D. (1995) A colorimetric assay for measuring activation of G s- and G q-coupled signaling pathways. Anal. Biochem. 226, 349 –354. 5. Joyeux, A., Balaguer, P., Germain, P., Boussioux, A. M., Pons, M., and Nicolas, J. C. (1997) Engineered cell lines as a tool for monitoring biological activity of hormone analogs. Anal. Biochem. 249, 119 –130. 6. Stratowa, C., Himmler, A., and Czernilofsky, A. P. (1995) Use of a luciferase reporter system for characterizing G-protein-linked receptors. Curr. Opin. Biotechnol. 6, 574 –581. 7. Berger, J., Hauber, J., Hauber, R., Geiger, R., and Cullen, B. R. (1988) Secreted placental alkaline phosphatase: A powerful new quantitative indicator of gene expression in eukaryotic cells. Gene 66, 1–10. 8. Boie, Y., Sawyer, N., Slipetz, D. M., Metters, K. M., and Abramovitz, M. (1995) Molecular cloning and characterization of the human prostanoid DP receptor. J. Biol. Chem. 270, 18910 – 18916. 9. Funk, C. D., Furci, L., FitzGerald, G. A., Grygorczyk, R., Rochette, C., Bayne, M. A., Abramovitz, M., Adam, M., and Metters, K. M. (1993) Cloning and expression of a cDNA for the human

17.

18.

19.

20.

21.

22. 23.

325

prostaglandin E receptor EP1 subtype. J. Biol. Chem. 268, 26767–26772. Abramovitz, M., Adam, M., Boie, Y., Carriere, M.-C., Denis, D., Godbout, C., Lamontagne, S., Rochette, C., Sawyer, N., Tremblay, N. M., Belley, M., Gallant, M., Dufresne, C., Gareau, Y., Ruel, R., Juteau, H., Labelle, M., Ouimet, N., and Metters, K. M. (2000) The utilization of recombinant prostanoid receptors to determine the affinities and selectivities of prostaglandins and related analogues. Biochim. Biophys. Acta 1483, 285–293. Wright, D. H., Metters, K. M., Abramovitz, M., and FordHutchinson, A. W. (1998) Characterization of the recombinant human prostanoid DP receptor and identification of L-644,698, a novel selective DP agonist. Br. J. Pharmacol. 123, 1317–1324. Hagiwara, M., Alberts, A., Brindle, P., Meinkoth, J., Feramisco, J., Deng, T., Karin, M., Shenolikar, S., and Montminy, M. (1992) Transcriptional attenuation following cAMP induction requires PP-1-mediated dephosphorylation of CREB. Cell 70, 105–113. Alberts, A. S., Montminy, M., Shenolikar, S., and Feramisco, J. R. (1994) Expression of a peptide inhibitor of protein phosphatase 1 increases phosphorylation and activity of CREB in NIH 3T3 fibroblasts. Mol. Cell. Biol. 14, 4398 – 4407. Ungrin, M. D., Singh, L. M. R., Stocco, R., Sas, D. E., and Abramovitz, M. (1999) An automated aequorin luminescencebased functional calcium assay for G-protein coupled receptors. Anal. Biochem. 272, 34 – 42. Sanchez-Yague, J., Rodriguez, M. C., Hernandez-Hernandez, A., and Llanillo, M. (1995) Inhibition of prostaglandin E2-responsive adenylyl cyclase in embryonal human kidney 293 cells by phorbol esters [published erratum appears in J. Lipid Mediat. Cell Signal. 1995 May;11(3):295]. J. Lipid Mediat. Cell Signal. 11, 25–39. Keery, R. J., and Lumley, P. (1988) AH6809, a prostaglandin DP-receptor blocking drug on human platelets. Br. J. Pharmacol. 94, 745–754. Crider, J. Y., Griffin, B. W., and Sharif, N. A. (1999) Prostaglandin DP receptors positively coupled to adenylyl cyclase in embryonic bovine tracheal (EBTr) cells: Pharmacological characterization using agonists and antagonists. Br. J. Pharmacol. 127, 204 –210. Fitzgerald, L. R., Mannan, I. J., Dytko, G. M., Wu, H. L., and Nambi, P. (1999) Measurement of responses from G i-, G s-, or G q-coupled receptors by a multiple response element/cAMP response element-directed reporter assay. Anal. Biochem. 275, 54 – 61. McDonnell, J., Latif, M. L., Rees, E. S., Bevan, N. J., and Hill, S. J. (1998) Influence of receptor number on the stimulation by salmeterol of gene transcription in CHO-K1 cells transfected with the human beta(2)-adrenoceptor. Br. J. Pharmacol. 125, 717–726. Stables, J., Scott, S., Brown, S., Roelant, C., Burns, D., Lee, M. G., and Rees, S. (1999) Development of a dual glow-signal firefly and Renilla luciferase assay reagent for the analysis of G-protein coupled receptor signalling. J. Recept. Signal. Transduct. Res. 19, 395– 410. Wnendt, S., Kruger, T., Janocha, E., Hildebrandt, D., and Englberger, W. (1999) Agonistic effect of buprenorphine in a nociceptin/OFQ receptor-triggered reporter gene assay. Mol. Pharmacol. 56, 334 –338. Naylor, L. H. (1999) Reporter gene technology: The future looks bright. Biochem. Pharmacol. 58, 749 –757. Montmayeur, J. P., and Borrelli, E. (1991) Transcription mediated by a cAMP-responsive promoter element is reduced upon activation of dopamine D2 receptors. Proc. Natl. Acad. Sci. USA 88, 3135–3139.

326

DUROCHER ET AL.

24. Himmler, A., Stratowa, C., and Czernilofsky, A. P. (1993) Functional testing of human dopamine D1 and D5 receptors expressed in stable cAMP-responsive luciferase reporter cell lines. J. Recept. Res. 13, 79 –94. 25. Liaw, C. W., Grigoriadis, D. E., De Souza, E. B., and Oltersdorf, T. (1994) Colorimetric assay for rapid screening of corticotropin releasing factor receptor ligands. J. Mol. Neurosci. 5, 83–92. 26. Vansal, S. S., and Feller, D. R. (1999) An efficient cyclic AMP assay for the functional evaluation of beta-adrenergic receptor ligands. J. Recept. Signal. Transduct. Res. 19, 853– 863. 27. George, S. E., Bungay, P. J., and Naylor, L. H. (1997) Evaluation of a CRE-directed luciferase reporter gene assay as an alternative to measuring cAMP accumulation. J. Biomol. Screening 2, 235–240. 28. Thompson, J. F., Hayes, L. S., and Lloyd, D. B. (1991) Modulation of firefly luciferase stability and impact on studies of gene regulation. Gene 103, 171–177.

29. Bronstein, I., Fortin, J., Stanley, P. E., Stewart, G. S., and Kricka, L. J. (1994) Chemiluminescent and bioluminescent reporter gene assays. Anal. Biochem. 219, 169 –181. 30. Sun, P., Lou, L., and Maurer, R. A. (1996) Regulation of activating transcription factor-1 and the cAMP response element-binding protein by Ca 2⫹/calmodulin-dependent protein kinases type I, II, and IV. J. Biol. Chem. 271, 3066 –3073. 31. Migeon, J. C., and Nathanson, N. M. (1994) Differential regulation of cAMP-mediated gene transcription by m1 and m4 muscarinic acetylcholine receptors. Preferential coupling of m4 receptors to Gi alpha-2. J. Biol. Chem. 269, 9767–9773. 32. Katoh, H., Watabe, A., Sugimoto, Y., Ichikawa, A., and Negishi, M. (1995) Characterization of the signal transduction of prostaglandin E receptor EP1 subtype in cDNA-transfected Chinese hamster ovary cells. Biochim. Biophys. Acta 1244, 41– 48. 33. Tsukada, T., Horovitch, S. J., Montminy, M. R., Mandel, G., and Goodman, R. H. (1985) Structure of the human vasoactive intestinal polypeptide gene. DNA 4, 293–300.