An Automated Aequorin Luminescence-Based Functional Calcium Assay for G-Protein-Coupled Receptors

Analytical Biochemistry 272, 34 – 42 (1999) Article ID abio.1999.4145, available online at http://www.idealibrary.com on

An Automated Aequorin Luminescence-Based Functional Calcium Assay for G-Protein-Coupled Receptors Mark D. Ungrin, 1 Laila M. R. Singh, 2 Rino Stocco, Dean E. Sas, and Mark Abramovitz 3 Department of Biochemistry and Molecular Biology, Merck Frosst Center for Therapeutic Research, P.O. Box 1005, Pointe Claire-Dorval, Quebec H9R 4P8, Canada

Received December 8, 1998

We describe in detail an automated and highly sensitive functional assay for calcium-coupled receptors (those receptors whose activation results in an increase in intracellular calcium levels) utilizing coelenterazine-charged aequorin as a probe for intracellular calcium levels ([Ca 21] i). The assay was originally established to investigate Ga q-coupled prostanoid receptors, which are members of the G-protein-coupled receptor (GPCR) superfamily, signaling through elevation of [Ca 21] i, initially focusing on the human EP 1 prostanoid receptor (hEP 1). The parental human embryonic kidney cell line 293-AEQ17, developed by Button and Brownstein (Cell Calcium 14, 663– 671, 1993), constitutively expresses apoaequorin and was used to develop a clonal cell line which stably coexpresses hEP 1. This cell line was used to optimize assay parameters in order to maximize accuracy and throughput in an automated 96-well format with the result that each 96-well plate can be completed in 70 min. Use of this flexible system will greatly simplify the functional analysis of GPCRs and other receptors which when activated result in increases in [Ca 21] i. © 1999 Academic Press

Our interest in prostanoid receptor function has made it desirable to establish a robust and technically simple functional assay for G-protein-coupled recep-

1

Current address: Department of Medical Biophysis, Ontario Cancer Institute/Amgen Research Institute, University of Toronto, 620 University Ave., Suite 706, Toronto, Ontario M5G 2C1, Canada. 2 Current address: Institute of Molecular Biology and Biochemistry, Simon Fraser University, 8888 University Dr., Burnaby, British Columbia V5A 1S6, Canada. 3 To whom correspondence should be addressed. Fax: 514-4288615. E-mail: [email protected]. 34

tors (GPCRs) 4 such as human EP 1 prostanoid receptor (hEP 1) (2) which couple to increases in intracellular calcium ([Ca 21] i). Assessing the potency of various prostanoids at hEP 1 has been problematic, arising from the fact that the functional assay of choice for the EP 1 receptor has been the guinea pig ileum smooth muscle contraction assay (3), which is tedious and of very low throughput. The interpretation of the results is further obscured by the complexity of prostanoid action on this preparation (4). A number of groups (1, 5, 6, 7) have reported the use of the photoprotein aequorin (8) as a sensitive detector of increases in [Ca 21] i (see 9 for kinetics) triggered by agonist-activated GPCRs which can couple to the Ga q class of G-proteins. In these assay systems, cells expressing apoaequorin are charged with the chromophore cofactor coelenterazine, forming holoproteins which are then able to generate photons in response to [Ca 21] i transients. This, in turn, permits quantitative data readout in a luminometer. We describe in this report the detailed characterization of our aequorin luminescence assay that has been modified and extended from the aequorin assay as originally described by Button and Brownstein (1). Characterization was carried out utilizing the heterologous expression of hEP 1 in an HEK 293-AEQ17 cell line (1) constitutively expressing apoaequorin. A streamlined aequorin assay protocol was developed in an automated format, permitting increased throughput, reproducibility, and flexibility. As well, optimal values were obtained for such assay parameters as assay volumes, charging period and conditions, cell number, and cell viability window. This assay should be applicable to any GPCR (and theoretically any other receptor) which 4 Abbreviations used: GPCRs, G-protein-coupled receptors; hEP 1, human EP 1 prostanoid receptor; DMEM, Dulbecco’s modified essential medium; PG, prostaglandin; HBSS, Hanks’ balanced salt solution; LDAM, luminometer data analysis macros.

0003-2697/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.

AN AUTOMATED AEQUORIN ASSAY FOR G-PROTEIN-COUPLED RECEPTORS

upon activation results in a rapid transient increase in [Ca 21] i. MATERIALS AND METHODS

Construction of Expression Vector hEP 1-pCEP4 pcDNA-PGQ(Bam) containing the cDNA for the human EP 1 receptor (2) was subjected to cleavage with BamHI, and the purified 1.3-kb band was ligated to BamHI-digested pCEP4 (Invitrogen, La Jolla, CA). The orientation of hEP 1-pCEP4 was verified by restriction digest. Plasmid DNA was prepared using a Qiagen Megaprep plasmid kit (Qiagen, Mississauga, Ontario) according to the manufacturer’s protocol. Production of Stable hEP 1 Expressing 293-AEQ17 Cell Lines 293-AEQ17 cells, obtained under licence from the NIH, were maintained in culture in high glucose DMEM (GIBCO BRL, Mississauga, Ontario) growth medium (DMEM containing 10% heat inactivated fetal bovine serum, 1 mM sodium pyruvate, 20 units/ml penicillin G, 20 mg/ml streptomycin sulfate, and 500 mg/ml G418). Cells were transfected in 10-cm tissue culture dishes with 20 mg of hEP 1-pCEP4 plasmid DNA using a Ca 21 phosphate transfection kit (Stratagene, La Jolla, CA) according to the manufacturer’s instructions. After 24 h the cells were split into a T175 flask and DMEM growth medium was replaced with DMEM selection medium (DMEM growth medium plus 200 mg/ml hygromycin B). The DMEM selection medium was changed 4 days later and resistant colonies were allowed to grow for an additional 10 days. Clonal selection was carried out using the limited dilution technique. Cells from the T175 flask were counted and then diluted into a 50-ml sterile conical tube such that the cell number was approx 1 3 10 4 cells/ml to a total volume of 22 ml, and 0.2 ml was pipetted into each well of a 96-well plate. Tenfold dilutions from the original 50-ml tube were repeated three additional times, resulting in four 96-well plates covering a dilution range of 2 3 10 2 cells/well to 2 3 10 21 cells/well. Wells containing colonies arising from a single cell were then progressively expanded into T75 flasks. Ten clonal cell lines were subsequently assessed on the basis of aequorin luminescence in response to PGE 2 challenge.

35

After charging, the cells were washed from the growth surface by pipetting up and down, spun down for 5 min at 300g, rinsed, recentrifuged as before, resuspended, and maintained in suspension in modified Ham’s F12 medium in a 50-ml sterile Falcon disposable tube containing a small magnetic spin bar at 5 3 10 5 cells/mL. Cells and assays were all maintained at room temperature. Cell viability was determined by Trypan blue exclusion. Aequorin Luminescence Assay Protocol Prostanoid (PGE 2, PGF 2a, PGD 2, PGI 2, and U46619, Cayman Chemical Company, Ann Arbor, MI; iloprost, Amersham Canada Ltd., Oakville, Ontario) dilutions in 2 mL vehicle (dimethyl sulfoxide unless otherwise stated) were added to 98 mL modified Hanks’ balanced salt solution (HBSS) (25 mM Hepes, at pH 7.3) (GIBCO BRL) in a 96-well plate and loaded into a Luminoskan RS luminometer (Labsystems, Needham Heights, MA). In all assays, each row of the 96-well plate contained one positive and one negative control. The remaining 80 wells were divided into four series, numbered 1 to 4 in the order tested, each containing duplicate 10-point serial dilutions of a compound being investigated over a range of four and a half orders of magnitude in concentration. In all plates, series 1 consisted of a PGE 2 curve as an internal standard. Instability issues with prostacyclin (PGI 2) were dealt with by loading this compound and the associated PGE 2 controls into the 96-well plate in 2-mL volumes of ethanol. Immediately preceding the test, 100 mL of modified HBSS was injected into a well using peristaltic pump No. 3, a 5-s delay was allowed to ensure complete mixing, and the well was then immediately tested. Wells were tested sequentially, starting at position A1, by rows. Unless otherwise stated, 5 3 10 4 cells (in 100 mL volume) were injected per well using one peristaltic pump, and light emission was recorded over 30 s (peak 1) as a series of 60 half-second integrations, after which cells were lysed by injection of 25 mL of 0.9% (v/v) Triton X-100 solution in modified HBSS, from a second pump, and light emission was measured for an additional 10 s (peak 2, recorded as 20 half-second integrations). Instrumentation

Aequorin Charging Protocol Holoaequorin was reconstituted in intact cells by charging 85% confluent cultures for 1 or 4 h at 37°C in modified Ham’s F12 0.1% fetal bovine serum, 25 mM Hepes, at pH 7.3) (GIBCO BRL) containing 30 mM reduced glutathione (Sigma, St. Louis, MO) and 8 mM of coelenterazine cp (Molecular Probes, Eugene, OR).

Experiments were performed using a Luminoskan RS luminometer (Labsystems), with three integral peristaltic pumps, allowing for negligible delays between injections and reading. An internal orbital mixer and the velocity of the injected liquids ensured mixing of the solutions. Instrument control software was written in-house using the National Instruments LabView

36

UNGRIN ET AL.

programming language in order to allow maximum flexibility in optimizing the assay protocols and to permit increased automation of both the assay itself and subsequent data analysis. This custom software permits control over injection volumes from any of the three peristaltic pumps, in any order, along with the number and length of integration intervals. Drug concentrations may be entered by well, and serial dilutions may be calculated for a given series of wells. Basic calculations (peak integration and peak ratios) may be applied to the experimental results, and data files are produced in a format suitable for further data analysis. This was accomplished by means of custom Microsoft Excel macros, referred to as the luminometer data analysis macros (LDAM), which include capacity for curve fitting and EC 50 calculations. Instrument control software was run on a Macintosh IIci or a Compaq Deskpro Pentium. Data Analysis Peak integration values were obtained by summing the half-second integrations that made up the raw trace. Fractional luminescence for each well was determined by dividing the area under peak 1 (see above) by the total area under peaks 1 and 2. These calculations were performed in the Lskan Controller program, and a data file was generated, containing both the raw traces and the calculated results for each well, drug concentrations, and the start time of each well. This data file was then subsequently analyzed using the LDAM software, in which a modified version of the Levenberg–Marquardt four-parameter curve-fitting algorithm was employed to calculate EC 50 values for each series, and raw traces were examined for qualitative differences from the expected results. RESULTS

Parental cells, 293-AEQ17, which stably express apoaequorin (1), were used to establish a cell line which stably coexpresses hEP 1. The vector chosen, pCEP4, uses a cytomegalovirus promoter to drive transcription of hEP 1 and has been used previously for stable cell line production from the parental cell line HEK 293 EBNA (stably expressing the Epstein–Barr virus nuclear antigen), achieving satisfactory expression levels for all eight human prostanoid receptors, including hEP 1, with a maximum number of binding sites (B max) in the 1–5 pmol/mg of protein range (10). Initial characterization of the hEP 1 receptor expressing 293-AEQ17 cells was carried out by manually measuring aequorin luminescence in response to agonist challenge. Of the 10 stable clonal cell lines tested with two different concentrations of PGE 2, 100 nM and 1 mM, 3 showed similar relatively robust responses (data

not shown). One of these cell lines, designated hEP 1-5/ 293-AEQ17, was selected for further characterization and assay development. In initial experiments it was observed that the lowest concentration of PGE 2 which elicited a response resulted in the longest period of increased luminescence before the signal returned to base line, from approximately 5 to 25 s postchallenge. Increases in the agonist concentration resulted in a reduction in both the onset and termination times of the signal, down to 1 and 8 s, respectively, at maximal response levels. The use of aequorin as a reporter of [Ca 21] i transients in real time, therefore, requires that measurement begin immediately upon addition of an activating substance. This necessitates the use of a luminometer with integral injectors. In order to automate the assay it was necessary that the injected components of the assay remain constant during a run, to eliminate the need for manual intervention by the operator. In the case of analysis of a known receptor, agonist type and agonist concentration are the variable parameters, and the agonists must, therefore, be distributed in the assay plate for maximum efficiency. If available, an automated pipetting station may be used to prepare the plate in advance, minimizing manual labor and its inherent variability. This allows the use of two injectors for dispensing cells and Triton X-100, which remain constant during the run, and additional injectors can be used in special cases (see below). A schematic of how the assay works is shown in Fig. 1A. A typical example of the raw data generated by the luminometer using the hEP 1-5/293-AEQ17 cell line is shown in Fig. 1B. The small initial luminescent burst within the first second of recording is the cell injection artifact, which is proportional to the number of cells injected, and is constant within an experiment. The first major peak is due to agonist activation of hEP 1. The time-to-peak is very consistent, approximately 15 s for the minimal agonist response and decreasing to 2.5 s for a maximal agonist response. Similar results were obtained for the human recombinant FP and TP prostanoid receptors heterologously (transiently and stably) expressed in the same 293-AEQ17 cells as well as for an endogenously expressed thrombin receptor (data not shown), all of which are Ga q-coupled receptors. The second peak is due to the injection of TritonX-100 which results in cell lysis and triggers the remaining luminescence. Integration of the area under the curve of peak 1 (0 –30 s) plus that of peak 2 (30 – 40 s) gives the total luminescence per sample (Fig. 1B), from which the fractional luminescence (defined as that portion of the total luminescence potential emitted in response to agonist challenge) is derived (see Materials and Methods). Use of this fractional luminescence value allows for normalization of variations in cell

FIG. 1. (A) Aequorin luminescence assay schematic. Cells constitutively expressing cytosolic apoaequorin and a membrane receptor, which upon stimulation by an agonist results in the dissociation of the Ga q subunit, and exchange of GDP for GTP, from the Gbg subunits of a G-protein heterotrimer. Ga q then activates a PLC-b, which generates IP 3, with the resulting IP 3 receptor-mediated transient release of calcium from intracellular stores. When two molecules of calcium bind to charged aequorin photons of light are emitted and detected by the photomultiplier tube of the luminometer. The fraction of charged aequorin which has not reacted with calcium is subsequently discharged by Triton X-100, which disrupts cellular membrane integrity and allows calcium to pour in from the extracellular medium. PLC-b, phospholipase C-b; IP 3, inositol-(1,4,5)triphosphate; PIP 2, phosphatidylinositol 4,5-biphosphate; DAG, diacylgycerol; PKC, protein kinase C. (B) Determination of fractional luminescence from raw luminescence traces. hEP 1-5/293-AEQ17 cells were stimulated with 1, 10, 100 nM, and 1 mM PGE 2 and luminescence was monitored as relative light units (RLUs) over 40 s. PGE 2 was injected at t 5 0 s and Triton X-100 was injected to 0.1% (v/v) final at t 5 30 s. Light production was measured continuously as a series of 0.5-s integrations. The fraction of total integrated luminescence was determined at each ligand concentration by dividing the area under peak 1 by the sum of the areas under peaks 1 and 2. PGE 2 concentrations, peak 1, and peak 2 are indicated in the figure. The small luminescence signal visible within the first second of recording is the injection artifact. 37

38

UNGRIN ET AL.

FIG. 2. Determination of the aequorin assay temporal window. hEP 1-5/293-AEQ17 cells were charged with coelenterazine cp for 1 h and then harvested (requiring an additional 25–30 min; see Materials and Methods for details). The first concentration–response experiment was carried out immediately thereafter (runtime 20 min per curve, n 5 2 at each concentration over a range of 3 3 10 211 to 1 3 10 26 M). Subsequent concentration–response curves were obtained at 1-h intervals up until 7.5 h postharvest. The concentration of agonist required to produce a half-maximal response has been defined as the EC 50. Results from three separate experiments are shown represented by the symbols ■, F, and Œ. The temporal assay window is delimited by the dashed line. (A) EC 50 values from concentration–response curves over time. (B) Maximal fractional luminescence in response to agonist over time.

number on a well-by-well basis, which can otherwise interfere with the reproduceability of the data, resulting in decreased precision (data not shown). A time course of concentration–response curves was used to establish an assay window for the charged cells. As shown in Fig. 2A, the EC 50 values for the potent natural ligand PGE 2 remained relatively constant between 1 and 5 h postharvest (2– 4 nM) as did

the maximum fractional luminescence (Fig. 2B) over the same period of time. Outside of this time period a greater degree of variability was observed. The maximum total luminescence decreased gradually with time in parallel with the decrease in cell viability (data not shown). Most striking, however, was the unexpected decrease in fractional luminescence after 5 h postharvest, which occurred in two of the three exper-

AN AUTOMATED AEQUORIN ASSAY FOR G-PROTEIN-COUPLED RECEPTORS

39

FIG. 3. Representative concentration–response curves from hEP 1 -5/293-AEQ17 cells challenged with endogenous prostanoids. The fractional luminescence responses to PGE 2 , PGF 2a , U46619, and PGD 2 are plotted as a function of their concentrations. Sigmoidal curves were obtained by plotting fractional luminescence at each agonist concentration and analyzed using a modified version of the Levenberg–Marquardt four-parameter curve-fitting algorithm to determine the EC 50 values referred to in the text. Duplicates for each sample are shown.

iments and was an indication that the cells were no longer reliable. It was, therefore, decided to perform all assays within the 1- to 5-h postharvest period to allow the cells 1 h to recover from the charging procedure and maintain maximal reliability. We also tested the parental 293-AEQ17 cells in the aequorin assay for endogenous prostanoid receptor activity. 293-AEQ17 cells were challenged with PGE 2 , PGF 2a , PGD 2 , and PGI 2 as well as U46619, the stable mimetic of TxA 2 , and iloprost, a stable analogue of PGI 2 and a known potent agonist of both the IP and EP 1 receptors (3). PGE 2 ($300 nM) gave rise to a small increase (#10%) in fractional luminescence, while the other prostanoids did not elicit a response (data not shown). The lack of response to the potent EP 1 agonist iloprost suggested that the luminescence observed upon PGE 2 challenge was not related to activation of an endogenously expressed hEP 1 . We then challenged the hEP 1 -5/293-AEQ17 cells with the same prostanoids. Typical 10-point concentration–response curves done in duplicate in a single 96-well plate for PGE 2 , PGF 2a , PGD 2 , and U46619 are shown in Fig. 3. PGE 2 (EC 50 5 2.3 nM), the endogenous ligand, was most potent, followed by PGF 2a (EC 50 5 20.4 nM), approximately 10-fold less

potent, and U46619 (EC 50 5 1.98 mM) and PGD 2 (EC 50 5 2.49 mM), both approximately 1000-fold less potent. Each compound was tested in three independent experiments such that every plate included a PGE 2 standard curve to allow comparison of results across plates and from one experiment to the next. Interassay reproducibility is reflected in the resulting EC 50 values, which are shown along with the associated PGE 2 controls in Table 1. The sensitivity of the aequorin assay was tested by obtaining PGE 2 concentration–response curves at decreasing cell numbers from 5 3 10 4 cells/well down to 500 cells/well (Fig. 4A) in which there was a linear decrease in total luminescence with respect to the number of cells/well (data not shown). The EC 50 values remained relatively constant and ranged from a high of 3.8 nM to a low of 2.3 nM. Raw traces of responses to challenges with the same level of agonist from a 5 3 10 4 cell/well challenge and from a 500 cell/well challenge were virtually superimposable when scaled appropriately (Fig. 4B), demonstrating the sensitivity and high signal-to-noise ratio of the assay.

40

UNGRIN ET AL. TABLE 1

EC 50 Values (nM) from hEP 1-5/293-AEQ17 Cells Challenged with the Various Prostanoids

Experiment 1

Experiment 2 Experiment 3

Prostanoid

Plate 1

Plate 2

Plate 3

PGE 2 PGF 2a U46619 PGD 2 PGE 2 PGI 2 PGE 2 Iloprost

2.0 21.4 1880 2480 1.7 238 4.7 3.2

2.3 20.9 2020 2440 2.7 278 6.3 3.7

2.3 23.6 2660 2280 3.3 273 1.4 1.0

Average 2.2 6 0.2 22.0 6 1.4 2190 6 416 2400 6 106 2.6 6 0.8 263 6 21.8 4.1 6 2.5 2.6 6 1.4

Note. Each test plate contained a PGE 2 control curve, and the results are grouped accordingly. The average and standard deviation values are calculated for each triplicate experiment.

DISCUSSION

The use of the photoprotein aequorin as a sensitive detector of calcium fluctuations in living cells is well established (9). Its use as a reporter for GPCRs was initially demonstrated in Xenopus oocytes (11) in which the aequorin protein was microinjected directly into the cytoplasm of individual oocytes, which were then monitored for luminescence in response to agonist stimulation. We have used this labor-intensive and inherently variable system for the identification of both the human EP 1 (2) and FP (12) prostanoid receptors. However, any detailed functional analysis of these receptors requires greater throughput and reproducibility. Two reports published in 1993 (1, 5) described for the first time transient and/or stable expression of apoaequorin as a reporter of GPCR-mediated [Ca 21] i mobilization in mammalian cells. Button and Brownstein (1) also showed that apoaequorin was predominantly expressed in the cytoplasm of 293-AEQ17 cells and that by treating cultures with the chromophore cofactor for apoaequorin, coelenterazine, for 4 h, maximal levels of aequorin luminescence could be achieved. We have modified their original assay, as described below, to greatly increase throughput and make the assay more robust and reliable. As a first step toward increased throughput the assay was automated in a 96-well plate format. This involved maintaining charged cells in suspension prior to injection, allowing for advance preparation of the assay plates with the appropriate reagents and dilutions thereof. To increase reproducibility of duplicates and assay robustness it was also desirable to have an internal control for cell count in each well. This required the use of a second injector to dispense the detergent Triton X-100, allowing measurement of the total potential luminescence emission within each sample. The luminometer, therefore, required a minimum of two integral peristaltic pump injectors. The Labsystems Luminoskan RS 96-well plate reader lu-

minometer met these requirements, with the option of mounting one or two additional pumps. For added flexibility in special circumstances, such as when unstable compounds require testing, as in the case of PGI 2, a third pump was installed. Use of this instrument and modifications to the charging procedure have eliminated the problem of unpredictable artifactual light production upon addition of cells or buffer solution first noted by Button and Brownstein (1). Optimization of assay parameters allowed for the generation of concentration response curves such that each assay plate could be completed in 70 min without compromising the integrity of the data. Depending on the experimental requirements, throughput can be further increased. For example, if identification of agonists is the goal, then a shorter 15-s read could be used, decreasing plate reading time to 26 min and nearly trebling throughput. One of the most critical aspects of any whole-cell assay is the health of the cells. In order to maximize the quality of the results, it was necessary to establish the period during which cells could be used reliably. During the first hour postharvest the EC 50 for PGE 2 at hEP 1 was somewhat variable, indicating perhaps that the cells required a recovery period from the stress of the charging/harvesting protocols. The maximum fractional response did not vary over the first 5 h, after which it dropped markedly, coinciding with an increase in the variability of the EC 50 for PGE 2. The reason for this is not obvious as the decrease in total luminescence was linear with time, as was the decrease in the viability of the cells, but it is thought to reflect the onset of changes in the cells brought about by an extended period under suboptimal (room temperature) conditions. For these reasons charged cells were used only between 1 and 5 h postharvest. Another critical aspect of any whole cell assay is interassay variability. Cells may vary from assay to assay due to a variety of factors including passage number, confluency, or culture medium/serum variability, to name a

AN AUTOMATED AEQUORIN ASSAY FOR G-PROTEIN-COUPLED RECEPTORS

41

FIG. 4. Stability of the response in the aequorin assay with respect to cell number. (A) Concentration–response curves from hEP 1-5/293AEQ17 cells challenged with PGE 2 at varying cell counts per 100 ml of injected cell suspension. (B) Similarity in responses of 5 3 10 4 and 500 hEP 1-5/293-AEQ17 cells challenged with 3 mM PGE 2.

few. One means of increasing reliability, therefore, is to include a standard agonist in each assay, compare its EC 50 value with the other agonists tested in that experiment, and derive a relative activity. The rank order of potency thus obtained (iloprost $ PGE 2 . PGF 2a . PGI 2 . U46619 $ PGD 2; see Table 1) is similar to that

obtained from radioligand binding studies conducted on hEP 1 expressed in HEK 293 EBNA cells (10). This method can significantly reduce interassay variability in larger datasets (data not shown). At low concentrations of agonist the peak of the response is observed to occur at a later time point than

42

UNGRIN ET AL.

at higher concentrations (see Fig. 2A). The reason for this shift in time-to-peak is unclear but probably has to do with the complex nature of calcium release and reuptake within the cell (13) and may be related to the time required to build up inositol-(1,4,5)triphosphate and trigger calcium release from internal stores. Similar results were observed for heterologously expressed FP and TP prostanoid receptors, as well as an endogenously expressed thrombin receptor (data not shown). Preparation of a 293-AEQ17 cell line which stably expresses the receptor of interest simplifies the assay and facilitates the detailed analysis of a particular GPCR as the cells may be maintained in culture, and variability from one experiment to next is minimized as receptor expression levels are similar from assay to assay. In addition, preparative work is significantly reduced in that the need for repeated transient transfections is eliminated. This does not, however, preclude the use of transiently expressed receptors (14) or the use of other cell lines where aequorin and the receptor of choice can both be transiently or stably expressed. In this regard, COS7 and U937 cells have also been used successfully in this assay (15 and unpublished data). In terms of its general applicability for Ga q-coupled GPCRs we have been able to use this assay for not only prostanoid receptors but for thrombin, serotonin, galanin, and leukotriene receptors as well (16 and unpublished data). It has also recently been shown that by expressing the promiscuous G-protein, Ga 16 (17), receptors coupling to G-proteins other than Ga q, such as Ga s or Ga i, may also be assayed using aequorin (7), further increasing the versatility of the aequorin assay. The sensitivity of the system was further enhanced by the use of the synthetic cofactor, coelenterazine cp (18), in place of the natural cofactor. As few as 500 cells per well could be used to generate a concentration response curve and picomolar concentrations of agonist were enough to elicit a response in the assay. This exquisite sensitivity also makes this assay ideal for expression cloning projects, where cells can be grown and transfected directly in the assay plate, as well as in the identification of ligands for orphan GPCRs (19). In summary, we have established an automated and sensitive functional aequorin assay for the human EP 1 prostanoid receptor in particular and applicable to receptors coupling to elevation of [Ca 21] i in general. This will greatly simplify the task of functionally assaying any Ga q-coupled receptor in detail, and place their ease-of-analysis on par with Ga s- and Ga i-coupled receptors. ACKNOWLEDGMENTS The authors thank Scott Feighner and Michael Dashkevicz from the Department of Biochemistry and Physiology, Merck Research Laboratories, for providing the 293-AEQ17 cell line and initial cul-

turing protocols. The authors also thank, from the Department of Research Information Systems, Merck Frosst Center for Therapeutic Research, Jerry Ferentinos, Elie Hanna, and Jean Marois for providing the curve-fitting software and Kevin Needen for contributing to the development of the instrument control software.

REFERENCES 1. Button, D., and Brownstein, M. (1993) Cell Calcium 14, 663– 671. 2. Funk, C. D., Furci, L., FitzGerald, G. A., Grygorczyk, R., Rochette, C., Bayne, M. A., Abramovitz, M., Adam, M., and Metters, K. M. (1993) J. Biol. Chem. 268, 26767–26772. 3. Coleman, R. A., Kennedy, I., Humphrey, P. P. A., Bunce, K., and Lumley, P. (1989) in Comprehensive Medicinal Chemistry, Vol. 3, pp. 643–714, Pergamon Press, Oxford. 4. Lawrence, R. A., Jones, R. L., and Wilson, N. H. (1992) Br. J. Pharmacol. 105, 271–278. 5. Sheu, Y.-A., Kricka, L. J., and Pritchett, D. B. (1993) Anal. Biochem. 209, 343–347. 6. Brini, M., Marsault, R., Bastianutto, C., Alvarez, J., Pozzan, T., and Rizzuto, R. (1995) J. Biol. Chem. 270, 9896 –9903. 7. Stables, J., Green, A., Marshall, F., Fraser, N., Knight, E., Sautel, M., Milligan, G., Lee, M., and Rees, S. (1997) Anal. Biochem. 252, 115–126. 8. Shimomura, O., Johnson, F. H., and Saiga, Y. (1962) J. Cell. Comp. Physiol. 59, 223–240. 9. Shimomura, O. (1995) Biochem. Biophys. Res. Commun. 211, 359 –363. 10. Abramovitz, M., Adam, A., Boie, Y., 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. (1999) Biochim. Biophys. Acta., in press. 11. Sandberg, K., Markwick, A. J., Trinh, D. P., and Catt, K. J. (1988) FEBS Lett. 241, 177–180. 12. Abramovitz, M., Boie, Y., Nguyen, T., Rushmore, T. H., Bayne, M. A., Metters, K. M., Slipetz, D. M., and Grygorczyk, R. (1994) J. Biol. Chem. 269, 2632–2636. 13. Clapham, D. E. (1995) Cell 80, 259 –268. 14. Boie, Y., Stocco, R., Sawyer, N., Slipetz, D. M., Ungrin, M. D., Neuschafer-Rube, F., Puschel, G., Metters, K. M., and Abramovitz, M. (1997) Eur. J. Pharmacol. 340, 227–241. 15. Hamdan, F. F., Ungrin, M. D., Abramovitz, M., and Ribeiro, P. (1999) J. Neurochem. 72, 1372–1383. 16. Lynch, K. R., O’Neill, G. P., Liu, Q., Im, D.-S., Sawyer, N., Metters, K. M., Coulombe, N., Abramovitz, M., Figueroa, D. J., Zeng, Z., Connolly, B. M., Bai, C., Austin, C. P., Chateauneuf, A., Stocco, R., Greig, G. M., Kargman, S., Hooks, S. B., Hosfield, E., Williams, D. L., Jr., Pettibone, D. J., Gould, R. J., Ford-Hutchinson, A. W., Caskey, C. T., and Evans, J. F. (1999) Nature, in press. 17. Offermanns, S., and Simon, M. I. (1995) J. Biol. Chem. 270, 15175–15180. 18. Shimomura, O., Musicki, B., and Kishi, Y. (1989) Biochem. J. 261, 913–920. 19. Feighner, S. D., Tan, C. P., McKee, K. K., Palyha, O. C., Hreniuk, D. L., Pong, S.-S., Austin, C. P., Figureoa, D., MacNeil, D., Cascieri, M. A., Nargund, R., Bakshi, R., Abramovitz, M., Stocco, R., Kargman, S., O’Neill, G. P., Van Der Ploeg, L. H. T., Evans, J., Patchett, A. A., Smith, R. G., and Howard A. D. (1999) Science, in press.