Abstract
An assay for measuring agonist-stimulated [35S]guanosine-5′-O-(3-thio)triphosphate (GTPγ35S) binding to heterotrimeric GTP binding proteins was developed for use in 96-well format using commercially available anti-G protein antibodies captured by anti-IgG-coated scintillation proximity assay beads. Use of an anti-Gαq/11 antibody to measure GTPγ35S binding mediated by M1, M3, and M5 receptors stably expressed in Chinese hamster ovary (CHO) cells resulted in a marked increase in agonist-stimulated/basal binding ratio compared with whole membrane binding. Pertussis toxin (PTX) treatment of CHO M1 cells before membrane preparation resulted in a marked reduction in agonist-stimulated GTPγ35S binding to whole membranes. Direct coupling of M1 receptors in CHO cells to inhibitory G proteins was demonstrated using an anti-Gαi(1–3) antibody, and this binding was inhibited by 76% following PTX treatment. However, PTX had no effect on M1-mediated binding determined using anti-Gαq/11. CHO M2 receptors mediated robust agonist-stimulated GTPγ35S binding measured with anti-Gαi(1–3), but coupled only weakly to Gαq/11. Using membranes from rat striatum, GTPγ35S binding stimulated by oxotremorine M was demonstrated using anti-Gαq/11, anti-Gαi(1–3), and anti-Gαo antibodies. Agonist-stimulated binding to striatal membranes showed a marked antibody-dependent GDP requirement with robust signals obtained using 0.1 μM GDP for anti-Gαq/11 compared with 50 μM GDP for anti-Gαi(1–3) and anti-Gαo. The potencies observed for pirenzepine and AFDX 116 blockade of agonist-stimulated GTPγ35S binding to striatal membranes determined with anti-Gαq/11 and anti-Gαo suggested mediation of these responses primarily by M1 and M4 receptors, respectively. Antibody capture GTPγ35S binding using scintillation proximity assay technology provides a convenient, productive alternative to immunoprecipitation for exploration of receptor-G protein interaction in cells and tissues.
Agonist binding to G protein-coupled receptors stimulates the exchange of GTP for GDP bound to the α subunit of coupled heterotrimeric GTP binding proteins (Kaziro et al., 1991). Binding of the stable GTP analog, [35S]guanosine-5′-O-(3-thio)triphosphate (GTPγ35S) to membranes using rapid filtration techniques has been used as a functional assay for muscarinic receptors in cells and tissues (Hilf et al., 1989; Lazareno and Birdsall, 1993;Burford et al., 1995; Olianas and Onali, 1996). GTP binding coupled with immunoprecipitation using specific antibodies to G protein α subunits has been used as a method to identify which GTP binding proteins are coupled to receptors of interest (Okamoto et al., 1992;Offermanns et al., 1994; Murthy and Makhlouf, 1996; Barr et al., 1997). By using immunoprecipitation with anti-Gαq/11, Reever et al. (1995) demonstrated an enhanced agonist-stimulated/basal binding ratio for GTPγ35S binding to membranes from Chines hamster ovary (CHO) cells transfected with M1 receptors compared with the rapid filtration method. Thus the use of specific antibodies to G protein α subunits may be used to identify which GTP binding proteins couple to individual subtypes of muscarinic receptors in cells and tissues, and may also improve signal to noise for GTPγ35S binding under certain circumstances. In the present study we developed a GTPγ35S binding assay using anti-G protein antibodies coupled with antibody capture via anti-IgG-coated scintillation proximity assay (SPA) beads. This approach is much more convenient than conventional immunoprecipitation and allows for development of medium throughput automated assays. In the present study we demonstrate the use of antibody capture GTPγ35S binding to explore muscarinic receptor-G protein interaction in transfected cells and native brain tissue.
Experimental Procedures
Cell Culture.
CHO cells transfected with human M1–M5 receptors were grown either in suspension or in monolayer. For suspension cultures cells were grown in roller bottles with constant agitation at 37°C and 5% CO2 using Dulbecco’s modified Eagles medium/F-12 (3:1) culture medium supplemented with 5% fetal bovine serum, 50 μg/ml tobramycin, and 20 mM HEPES. Monolayer cultures were grown in T-225 flasks at 37°C and 5% CO2 in Dulbecco’s modified Eagles medium supplemented with 10% fetal bovine serum and 100,000 U/liter of penicillin/streptomycin. Cells were harvested using trypsin-free dissociation media at 95% confluence and were collected by centrifugation and stored at −80°C. Cells stably expressing human muscarinic receptors were obtained from the National Institutes of Health.
Membrane Preparation.
Cell pellets were thawed and resuspended in 20 volumes of 20 mM sodium phosphate buffer, pH 7.4, and were homogenized twice for 30 s at high speed using a Tissuemizer. Homogenates were centrifuged at 200g for 15 min at 4°C. The supernatant was removed and reserved on ice. This procedure was repeated twice and the pooled supernatants were then centrifuged at 40,000g for 45 min at 4°C. Membranes were suspended at 5 mg protein/ml and were stored at −80°C. Unless indicated otherwise in the figure legends, membranes from M1, M2, and M4 cells were prepared from cells grown in suspension, whereas those from M3 and M5 cells were from cells grown in monolayer. Receptor densities (pmol mg−1 membrane protein) were 9.3, 0.7, 0.6, 0.9, and 4.8 for M1–M5receptors, respectively.
Striatal tissue from male Sprague-Dawley rats was homogenized by hand in 10 volumes of 10 mM HEPES and 1 mM EGTA, pH 7.4, containing Complete protease inhibitor cocktail, 1 mM dithiothreitol, and 10% sucrose. The homogenate was diluted 6-fold and centrifuged at 1000g for 10 min at 4°C. The supernatant was saved and the pellet rehomogenized and centrifuged as above. The combined supernatants were centrifuged at 11,000g for 20 min. The resulting pellet was homogenized in 40 volumes of 10 mM HEPES and 1 mM EGTA, pH 7.4, containing 1 mM dithiothreitol and 1 mM MgCl2, and was centrifuged at 27,000g for 20 min. The resulting pellet was suspended in the same buffer at a protein concentration of 1.5 mg/ml and aliquots were frozen and stored at −80°C.
GTPγ35S Binding.
Assays were run in 20 mM HEPES, 100 mM NaCl, and 5 mM MgCl2 at pH 7.4 in a final volume of 200 μl in 96-well Costar plates at 25°C. One hundred microliters of membrane preparation (25 μg protein per well for cell membranes and 9–15 μg per well for brain membranes) containing the appropriate concentration of GDP was added followed by addition of 50 μl of buffer ± agonists and antagonists being tested followed by 50 μl of GTPγ35S to provide a final concentration in the assay of 200 pM for CHO membranes and 500 pM for brain membranes. For CHO membranes, 0.1 μM GDP was used for M1, M3, and M5 receptor assays, whereas 1 μM GDP was used for M2 and M4 assays. For brain membranes 0.1 μM GDP was used in assays carried out with anti-Gαq/11, whereas 50 μM GDP was used for assays usng anti-Gαi(1–3) and anti-Gαo. CHO cell membranes were incubated for 30 min at 25°C with agonists and antagonists followed by addition of GTPγ35S and incubation for an additional 30 min. Brain membranes were incubated for 20 min at 25°C with agonists and antagonists followed by addition of GTPγ35S and incubation for an additional 60 min. Preincubation was employed to ensure that agonists and antagonists were at equilibrium during the labeling period.
To determine total membrane binding, 50 μl of suspended wheat germ agglutinin (WGA)-coated SPA beads was added. After 15 min, plates were centrifuged at 1000g for 15 min and radioactivity was determined using a Wallac plate counter. For determining binding to specific Gα proteins, 35S-labeled membranes were solubilized for 30 min with 0.27% Nonidet P-40 (20 μl/well of a solution containing 1.5 ml of 10% Nonidet P-40 for every 3.5 ml assay buffer) followed by addition of desired antibody (10 μl/well) to provide a final dilution of 1/400 to 1/100 and incubation for an additional 60 min. Fify microliters of suspended anti-IgG-coated SPA beads was added per well, plates were incubated for 3 h, and then were centrifuged and radioactivity determined as above. Each bottle of WGA-coated SPA beads was suspended in 10 ml of assay buffer and each bottle of anti-IgG-coated SPA beads was suspended in 20 ml of assay buffer. Protein was determined using the bicinchoninic acid assay (Smith et al., 1985).
Materials.
35 S-GTPγS (1000–1200 Ci/mmol), anti-rabbit-IgG and anti-mouse-IgG-coated SPA beads, and WGA-coated SPA beads were obtained from Amersham (Arlington Heights, IL). Rabbit anti-Gαq/11 and rabbit anti-Gαi(1–3) were from Santa Cruz Biotechnologies (Santa Cruz, CA). Mouse monoclonal anti-Gαo was from Chemicon (Temecula, CA). Oxotremorine M and pirenzepine were from Research Biochemicals Inc. (Natick, MA). 11-{[2-((Diethylamino)methyl)-1-piperidinyl]acetyl}-5,11-dihydro-6H-pyrido[2,3b][1,4]benzodiazepin-6-one (AFDX 116) was synthesized at Eli Lilly. Complete protease inhibitor cocktail and 10% Nonidet P-40 were from Boehringer Mannheim (Indianapolis, IN).
Data Analysis.
Concentration-response curves were fitted using sigmoidal nonlinear regression with variable slope in Graphpad Prism. Equilibrium K
is for data obtained via antagonist dose responses in the presence of a fixed agonist concentration were calculated from the general Cheng Prusoff relationship (Leff and Dougall, 1993):
Results
Characteristics of Antibody Capture GTPγ35S Binding.
Figure 1 illustrates the effect of antibody concentration on oxotremorine M-stimulated GTPγ35S binding determined with anti-Gαq/11 (CHO M1 membranes), anti-Gαi(1–3) (CHO M2 membranes), and anti-Gαo (rat striatal membranes). No agonist-stimulated binding was observed in the absence of added antibody. The cpm measured in the absence of antibody result from the nonproximity effect of 35S in solution plus a small amount of nonspecific binding. Basal and agonist-stimulated binding increased to a plateau level with increasing antibody concentration. The antibody dilution curves shown in Fig. 1were obtained using the SPA bead dilution stated in Experimental Procedures, i.e., each bottle of reagent obtained from the manufacturer diluted with 20 ml of assay buffer. Using the anti-Gαq/11 dilution that produced maximal agonist-stimulated binding to CHO M1 membranes as shown in Fig. 1, it was observed that, over the range of 25 ml/bottle to 10 ml/bottle, basal and agonist-stimulated binding signals increased with increasing bead density without major effect on signal to noise. A bead density was chosen that provides a sufficiently large agonist-stimulated signal to obtain reproducible concentration- response curves without using more reagent than necessary. We have observed that different lots of antibody may show differences in optimal antibody concentration required for maximal agonist-stimulated binding signal.
Concentration-response curves obtained for oxotremorine M-stimulated GTPγ35S binding to M1–M5 receptors in membranes from CHO cells are shown in Fig.2. Concentration-response curves for agonist-stimulated binding were determined at antibody dilutions that produced maximal binding. As indicated, data were obtained using anti-Gαq/11 for M1, M3, and M5 receptors and anti-Gαi(1–3) for M2 and M4 receptors. Maximal binding was observed with 25 to 50 μg of cell membrane protein per well (data not shown). Figure3 compares basal and oxotremorine M-stimulated GTPγ35S binding mediated by M1–M5 receptors in CHO cell membranes determined by WGA whole membrane binding versus antibody capture. When expressed as percentage of increase over basal, agonist-stimulated binding determined using anti-Gαq/11 for M1, M3, and M5 receptors was increased 12- to 17-fold using antibody capture. For M2 and M4 receptors, signal to noise was only slightly improved by the antibody technique even though basal binding was markedly reduced.
Coupling of the M1 Receptor in CHO Cells to Inhibitory G proteins.
As shown in Fig. 4A, treatment of CHO M1 cells for 18 h with 100 ng ml−1 pertussis toxin (PTX) before membrane preparation resulted in a significant 70% reduction in agonist-stimulated GTPγ35S binding to whole membranes, suggesting that the M1 receptor couples to Gi as well as Gq family GTP binding proteins in these cells. The data in Fig. 4B were obtained by antibody capture and demonstrate that M1-mediated binding to Gαq/11 was not significantly affected by PTX pretreatment, whereas binding determined using anti-Gαi(1–3) was reduced by 76%. In contrast to the promiscuous coupling of the M1 receptor in CHO cells, M2 receptor-mediated GTPγ35S binding was largely restricted to inhibitory G proteins (binding determined with anti-Gαq/11 was only 5% of that determined with anti-Gαi(1–3), (data not shown).
GTPγ35S Binding to Rat Striatal Membranes Using Antibody Capture.
Basal and oxotremorine M-stimulated GTPγ35S binding to rat striatal membranes determined with anti-Gαq/11, anti-Gαo, and anti-Gαi(1–3) is shown in Fig. 5A. Agonist-stimulated binding averaged 92%, 62%, and 49% over basal binding with these three antibodies, respectively. Basal binding determined with anti-Gαq/11 was significantly lower than that obtained with the two antibodies against inhibitory G proteins (p < .01). Figure 5B compares the effect of 0.1 μM versus 50 μM GDP on agonist-stimulated GTPγ35S binding to striatal membranes determined with anti-Gαq/11 and anti-Gαo. Using anti-Gαq/11, agonist-stimulated binding at 50 μM GDP was only 6% of that observed at 0.1 μM GDP. In contrast, with the anti-Gαo antibody no detectable agonist stimulation was observed at 0.1 μM GDP (and basal binding was very high), whereas a robust signal was obtained at 50 μM GDP. Concentration-response curves for oxotremorine M and for pirenzepine reversal of oxotremorine M-stimulated GTPγ35S binding to striatal membranes determined with anti-Gαq/11, anti-Gαi(1–3), and anti-Gαo are shown in Fig. 6. Figure7 demonstrates the rightward shifts of oxotremorine M dose responses due to a single concentration of AFDX 116 determined using the three antibodies. Table1 compares the equilibriumK is for pirenzepine and AFDX 116 calculated from the data in Figs. 6 and 7 with the constants determined for these antagonists based on data obtained for M1–M5 receptors in CHO cells. The antagonist constants obtained with CHO cell membranes were determined from concentration responses for pirenzepine- and AFDX 116-mediated reversal of oxotremorine M-stimulated binding.
Discussion
The present study describes for the first time the use of SPA technology in conjunction with anti-G protein antibodies to measure GTPγ35S binding mediated through activation of G protein- coupled receptors. The method provides a convenient alternative to labor-intensive conventional immunoprecipitation for investigating receptor coupling to specific classes of GTP binding proteins. We have shown, as previously described by Reever et al. (1995) for the muscarinic M1 receptor, that the signal to noise for GTPγ35S binding mediated by muscarinic receptors primarily coupled to phosphatidylinositol hydrolysis (M1, M3, and M5) is markedly increased by isolation of labeled Gq/11 from the total pool of G proteins present in whole membranes. The specificity of the antibodies used in the present investigation is illustrated by the improved signal to noise obtained with anti-Gαq/11, by the PTX block of M1-mediated binding determined with anti-Gαi(1–3) without effect on M1-mediated binding determined with anti-Gαq/11, and by the much greater magnitude of agonist-stimulated binding mediated by M2 receptors (which are coupled to inhibition of adenylate cyclase in CHO cells) when measured with anti-Gαi(1–3) than with anti-Gαq/11. The rank order of potencies determined by antibody capture for oxotremorine M-stimulated GTPγ35S binding mediated by M1-M4 receptors shown in Fig. 2 (M2 > M4 > M1 > M3) was the same as reported by Lazareno and Birdsall (1993) for acetylcholine-stimulated GTPγ35S binding to CHO cell membranes measured by rapid filtration. It is noteworthy that the slopes of agonist dose-response curves shown in Fig. 2 for M1–M5 receptor-mediated GTPγ35S binding were universally steeper than those obtained in the study by Lazareno and Birdsall (1993) using binding to whole membranes, probably reflecting a more restricted variety of G proteins involved in binding due to the use of antibody capture. The direct demonstration of M1 receptor coupling to inhibitory G proteins in CHO cells shown with anti-Gαi(1–3) in this study confirms and extends the observations ofBurford et al. (1995), which were based only on PTX effects in CHO cells. The difference in M1 versus M2 promiscuity observed in these studies may well be due to the fact that M1 receptor expression was 10-fold greater than M2 in our cell lines because we have shown that removal of 80% of [3H]N-methylscopolamine binding to CHO M1 membranes by partial alkylation had no significant effect on Gαq/11 binding while completely eliminating Gαi binding (results not shown). It should be pointed out that not all commercially available antibodies will work in the GTPγ35S binding assay because we have tried a number of anti-Gαq/11 and anti-Gαi/o antibodies that did not allow demonstration of appreciable agonist-stimulated signals.
Using the anti-Gαq/11, anti-Gαi(1–3), and anti-Gαo antibodies employed in this study, we were able to measure significant oxotremorine M-stimulated GTPγ35S binding to membranes prepared from rat striatum. The large difference in GDP requirement for agonist-stimulated binding determined with anti-Gαq/11 compared with anti-Gαi(1–3) and anti-Gαo strongly indicates that these two classes of antibodies are able to separate signaling mediated by receptors in striatal membranes linked to stimulation of phosphatidylinositol hydrolysis from those coupled to the inhibition of adenylate cyclase. With regard to muscarinic receptors, rat striatum is known to be dominated by M1 and M4 receptors (Purkerson and Potter, 1998), but also has significant numbers of M2 receptors, which are found on large cholinergic interneurons in this tissue (Levey et al., 1991). Muscarinic M4 receptors have been shown to be responsible for mediating muscarinic agonist-induced inhibition of adenylate cyclase in membranes from rat striatum (Olianas et al., 1996). The calculated equilibrium K is for pirenzepine and AFDX 116 shown in Table 1 for inhibiting oxotremorine M-stimulated GTPγ35S binding to striatal membranes determined with anti-Gαq/11 are quite similar to those obtained for binding mediated by M1receptors in CHO membranes, strongly suggesting that this binding response is primarily due to the muscarinic M1receptor. It is unlikely that M2 or M4 receptors contributed to the response measured with anti-Gαq/11 based on the equilibriumK i obtained for AFDX 116 and the fact that these receptors do not normally couple to Gαq/11. Any significant contribution of M3 and M5 receptors to the response measured with anti-Gαq/11 would have increased the observedK i for pirenzepine (Table 1). Although we and others have shown that M1 receptors in transfected cells can couple to inhibitory G proteins, such coupling is inhibited by high concentrations of GDP as for coupling to Gαq/11 (data not shown). For this reason it is unlikely that M1 receptors made any significant contribution to the binding responses in striatum measured with anti-Gαi(1–3) or anti-Gαo, which were determined at 50 μM GDP. The potency for pirenzepine block of GTPγ35S binding mediated by M4 receptors in CHO cells shown in Table 1 is very similar to the value reported by Lazareno and Birdsall (1993) for GTPγ35S binding in CHO M4cell membranes and also to the value obtained in these cells for pirenzepine block of M4-mediated displacement ofN-methylscopolamine binding (Dorje et al., 1990). These values for pirenzepine block of M4-mediated responses in CHO cells are, however, much lower than the numbers reported in the literature (180–313 nM) for pirenzepine blockade of muscarinic M4 receptor-mediated inhibition of adenylate cyclase in striatal membranes (DeLapp et al., 1996; Olianas and Onali, 1991; Ehlert et al., 1989; McKinney et al., 1989). This discrepancy may reflect the differences in buffer ionic strengths used for radioligand and GTPγS binding compared with conditions used to measure adenylate cyclase activity because M4-mediated inhibition of adenylate cyclase in N1E-115 neuroblastoma cells (McKinney et al., 1991) was blocked by pirenzepine with an affinity (214 nM) in the same range as reported for the values obtained in rat striatum. TheK i for pirenzepine block of oxotremorine M-stimulated GTPγ35S binding to rat striatal membranes determined with anti-Gαo in this study (81 nM) lies between the numbers reported in the literature for M4 receptor-mediated responses in cells versus rat striatum, and is significantly lower than the pirenzepine value for the cloned M2 receptor (Table 1). Because pirenzepine blocked the response determined with anti-Gαo with a potency three times greater than observed for AFDX 116, whereas the reverse order of potency was found for these antagonists at the cloned M2 receptor (Table 1), we concluded that this response was mediated primarily by M4 receptors. Because the values for pirenzepine and AFDX 116 determined with anti-Gαi(1–3) did not differ significantly, it is reasonable to suggest that this response can be attributed to both M2 and M4 receptors. This interpretation of the data suggests that M2 and M4 receptors in striatum may couple differentially to members of the inhibitory G protein family as has been shown for these receptors transfected into JEG-3 cells (Migeon et al., 1995).
Reports of receptor-mediated GTPγ35S binding to whole brain membranes measured by rapid filtration have employed high micromolar concentrations of GDP in the assay to obtain significant agonist stimulation (Olianas and Onali, 1996; Alper and Nelson, 1998). Figure 5B shows that it was not possible in striatal membranes to demonstrate agonist stimulation with the anti-Gαo antibody at 0.1 μM GDP because of the high basal binding. The same observation has been made with striatal preparations for whole membrane binding (determined with WGA beads) and for binding determined with anti-Gαi(1–3) (data not shown). The data in Fig. 5B obtained with anti-Gαq/11 also demonstrate that, at least for muscarinic agonist-stimulated GTPγ35S binding, high micromolar concentrations of GDP inhibit coupling to Gq family proteins. For this reason one cannot reliably measure agonist-stimulated GTPγ35S binding mediated by receptors coupled to stimulation of phospholipase C in striatal membranes using conventional rapid filtration techniques. This is one clear advantage of the antibody capture method which, as we have shown, allowed us to determine oxotremorine M-stimulated binding in striatal membranes using anti-Gαq/11 at 0.1 μM GDP, which was likely mediated primarily through activation of the phospholipase C-coupled M1 receptor. As indicated by the present study, the antibody capture method employing specific antisera to individual G proteins also has the potential for developing receptor subtype-specific binding assays in complex native tissues in situations where differential coupling of receptor subtypes to individual G protein members of the same family may occur.
Footnotes
-
Send reprint requests to: Neil W. DeLapp, Lilly Corporate Center, Indianapolis, IN 46285. E-mail:NWD{at}lilly.com
-
↵1 A preliminary account of this work was presented at the Eighth International Subtypes of Muscarinic Receptors Meeting.
- Abbreviations:
- AFDX 116
- 11-{[2-((diethylamino)methyl)-1-piperidinyl]acetyl}-5,11-dihydro-6H-pyrido[2,3-b][1,4]benzodiazepin-6-one
- CHO
- Chinese hamster ovary cells
- PTX
- pertussis toxin
- SPA
- scintillation proximity assay
- WGA
- wheat germ agglutinin
- Received October 22, 1998.
- Accepted December 30, 1998.
- The American Society for Pharmacology and Experimental Therapeutics