Rat sublingual gland M1 and M3 muscarinic receptors each directly activate exocrine secretion. To investigate the functional role of coreceptor expression, we determined receptor-G protein coupling. Although membrane proteins of 40 and 41 kDa are ADP-ribosylated by pertussis toxin (PTX), and 44 kDa proteins by cholera toxin (CTX), both carbachol-stimulated high-affinity GTPase activity and the GTP-induced shift in agonist binding are insensitive to CTX or PTX. Carbachol enhances photoaffinity labeling ([α-32P]GTP-azidoaniline) of only 42-kDa proteins that are subsequently tractable to immunoprecipitation by antibodies specific for Gαq or Gα11 but not Gα12 or Gα13. Carbachol-stimulated photoaffinity labeling as well as phosphatidylinositol 4,5-bisphosphate (PIP2) hydrolysis is reduced 55% and 60%, respectively, by M1 receptor blockade with m1-toxin. Gαq/11-specific antibody blocks carbachol-stimulated PIP2 hydrolysis. We also provide estimates of the molar ratios of receptors to Gαq and Gα11. Although simultaneous activation of M1 and M3receptors is required for a maximal response, both receptor subtypes are coupled to Gαq and Gα11 to stimulate exocrine secretion via redundant mechanisms.
- salivary glands
- muscarinic cholinergic receptors
- mucous cells
- exocrine secretion
salivary glands of the oral cavity are numerous and diverse and are composed of either one or both of two principal acinar exocrine cell types, serous and mucous cells. In rodents and humans, mucous exocrine cells predominate within acinar structures and receive robust parasympathetic innervation with a paucity of sympathetic innervation. Correspondingly, unlike serous glands, both fluid secretion (20) and exocrine secretion of mucin glycoproteins by mucous glands (5) are primarily under muscarinic cholinergic control. Previously, we identified equivalent amounts of M1 and M3 muscarinic receptor subtypes in membranes from either whole rat sublingual glands or isolated acini (27). Furthermore, both muscarinic receptor subtypes appear to be associated directly with mucous acinar cells and are required for maximal exocrine secretion (5). The coexpression and regulatory function of muscarinic receptors within mucous acini may or may not extend to signaling events at the postreceptor level. For example, M1 and M3receptors may each couple potentially to different guanine nucleotide binding proteins (G proteins), even when expressed in the same cell (8).
At least 19 separate G protein α-subunits have been identified in mammalian tissues and are segregated into four families on the basis of primary sequence homology: Gαs, Gαi, Gαq, and Gα12 (11). Cholera toxin (CTX) catalyzes the ADP ribosylation of the Gαs family and Gαt(1 and 2), resulting in the loss of intrinsic GTPase activity (14). Pertussis toxin (PTX) ADP ribosylates a cysteine residue near the carboxyl terminus of Gαi, Gαo, Gαt, and Gαgus, preventing subsequent G protein and receptor interactions (14). In studies of cell lines expressing M1 or M3 receptors, both receptor subtypes are linked predominantly to Gαq/11 to activate phospholipase C (PLC) (8, 14). In addition, both receptor subtypes may activate PTX-sensitive G proteins (21) and stimulate PLC through Gβγ dimers (16). In rat parotid glands, M3 receptors couple to both Gαq and Gαi1 (6). Thus, in sublingual glands, coupling of either M1 or M3 receptors to other G proteins in sublingual acini cannot be ruled out, especially given the predominance of muscarinic receptors in regulating mucous acinar cell functions.
To provide insights into the functional role of receptor coexpression in sublingual glands, we therefore investigated the coupling of M1 and M3 receptors in sublingual membranes to G protein α-subunits. The coupling and toxin sensitivity (PTX and CTX) of receptors to G proteins was assessed by high-affinity GTPase activity and the GTP-induced shift in carbachol high-affinity binding sites. Photoaffinity labeling by [32P]GTP-azidoaniline and antisera specific for α-subunits was used to detect candidate subunits coupled to muscarinic receptors and to identify α-subunits linked to carbachol-induced phosphoinositide hydrolysis. The contribution of M1 receptors to carbachol-induced responses was determined by using the pseudoirreversible and highly specific M1receptor antagonist m1-toxin, a 64-amino acid peptide isolated from the venom of the Eastern green mamba, Dendroaspis angusticeps(23).
MATERIALS AND METHODS
Specific pathogen and sialodacryoadenitis virus-free male Wistar rats (2 mo old, 150–175 g) were obtained from Charles River Laboratories (Kingston facility, Stone Ridge, New York). PTX and CTX were from List Biological Laboratories. Carbachol, atropine, ATP, GDP, NAD, phosphocreatine, creatine phosphokinase, 5′-adenylyl-β,γ-imidodiphosphate (AppNHp), thymidine, benzamidine, EDTA, EGTA, 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide-HCl (EDAC-HCl), triethylammonium bicarbonate, 2-(N-morpholino) ethanesulfonic acid (MES), HEPES, 14C-methylated protein markers, and phosphatidylinositol 4,5-bisphosphate (PIP2) were from Sigma Chemical. Dithiothreitol (DTT) was obtained from United States Biochemical, and GTP, guanosine 5′-O-(3-thiotriphosphate) (GTPγS), aprotinin, leupeptin, and phenylmethylsulfonyl fluoride (PMSF) were from Boehringer Mannheim Biochem. Protein A-agarose was obtained from Calbiochem. Immobilon-PVDF (polyvinylidene difluoride) was obtained from Millipore. Cappel Rabbit IgG was obtained from ICN Pharmaceuticals, polyethylenimine-cellulose plates were from J. T. Baker, and 4-azidoaniline-HCl was from Fluka. Phosphatidylserine (PS) and phosphatidylethanolamine (PE) were from Avanti Polar Lipids; [125I]-labeled rProtein A (81.1 μCi/μg), [3H]PIP2 (6.0 Ci/mmol), [α-32P]GTP (3,000 Ci/mmol), [γ-32P]GTP (6,000 Ci/mmol), and [32P]NAD (800 Ci/mmol) were from Dupont NEN; and [3H]-N-methyl-scopolamine (NMS) was from Amersham International. Recombinant proteins, rat Gαq, and human Gα11 were obtained from Chemicon International. Rabbit antibodies (IgG fractions, 200 μg IgG/ml) D17, E17, S20, and A20, as well as their blocking proteins, were obtained from Santa Cruz Biotechnology. E17 was raised to amino acids 13–29 of the amino-terminal domain unique to Gαq. D17 was raised to amino acids 13–29 of the amino-terminal domain unique to Gα11. A20 and S20 were raised to amino acids 2–21 of the amino-terminal domain of Gα13 and Gα12, respectively. Rabbit antisera W082, B825, and Z811 were generous gifts from Dr. Paul Sternweis (Department of Pharmacology, University of Texas Southwestern Medical Center, Dallas, TX). W082 was raised to a synthetic peptide representing the internal amino acid sequence 115–133, unique to Gαq (22). Z811 was raised to a synthetic peptide of the common carboxy-terminal sequence 345–359 for Gαq and Gα11 (22). B825 was raised to a peptide spanning the internal sequence 114–133 of Gα11 (13). Antibody CT92, kindly provided by Dr. Dianqing Wu (Department of Pharmacology, University of Rochester Medical Center, Rochester, NY), was raised against a synthetic peptide to the internal sequence 116–132 of Gα14(1). Antisera AS233 and AS343 (25) were raised to synthetic peptides sharing distinct carboxy-terminal sequences unique to Gα12 (370) and Gα13 (367), respectively, and were kind gifts from Dr. Karsten Spicher (Institut für Pharmakologie, Freie Universität Berlin, Thielallee, Berlin, Germany).
Rats were killed by exsanguination after CO2anesthesia. Sublingual glands were removed quickly and kept in ice-cold L-15 medium until all glands were obtained. The glands were dissected at 4°C to remove remaining fat, connective tissue, and the hilum region and were then rinsed briefly in 20 mM Tris · HCl, pH 7.5, 1 mM EDTA, 10 μg/ml leupeptin, 5 μg/ml aprotinin, and 0.2 mM PMSF (Tris-EDTA buffer), minced, and homogenized in 10 volumes of Tris-EDTA buffer with a glass Dounce homogenizer. The homogenates were filtered through 250-μm nylon mesh (Small Parts) and spun at 25,000g for 25 min. The resulting pellets were further homogenized (6 × 30 s) with a BioHomogenizer (Biospec Products) at a high setting, pelleted as before, and washed with and resuspended in Tris-EDTA buffer. Protein content was determined by the method of Bradford, as described previously (5), with the use of bovine serum albumin (BSA) as standard. Aliquots (2 mg/ml) were stored at −80°C until use.
PTX and CTX were preactivated for 30 min at 30°C with 20 and 50 mM DTT, respectively, dissolved in 20 mM Tris-EDTA buffer (pH 7.5). Activated toxin and [32P]NAD (5 μCi/assay) were mixed with ADP-ribosylation buffer consisting of 1 mM EDTA, 2.5 mM MgCl2, 100 mM NaCl (for PTX only), 1 mM ATP, 10 μM GTP, 2.5 mM DTT, 10 mM thymidine, 10 μM NAD, 10 μg/ml leupeptin, 5 μg/ml aprotinin, 0.2 mM PMSF, and 20 mM Tris · HCl, pH 7.5, and the reaction was initiated by adding the membrane suspension (3 mg/ml resuspended in ADP-ribosylation buffer). The reaction was terminated after 60 min with ice-cold Tris-EDTA buffer, followed by spinning (12,000 g for 10 min, 4°C). Membrane pellets were subjected to SDS-PAGE and autoradiography. In competition, photoaffinity labeling and GTPase assays for bacterial toxin pretreatment of membranes were conducted with and without [32P]NAD at final concentrations of 25 μg/ml PTX or 10 μg/ml CTX, and the ratios of toxins to membrane proteins were kept constant at 10 μg PTX/mg protein and 4 μg CTX/mg protein.
Sublingual membranes were washed twice by centrifugation (12,000g for 10 min, 4°C) with 30 volumes of ice-cold assay buffer (50 mM triethanolamine-HCl, pH 7.4, 0.2 mM EGTA, 10 μg/ml leupeptin, 5 μg/ml aprotinin, and 0.2 mM PMSF). To assay GTPase activity, we used the method described by Ghodsi-Hovsepian et al. (10). Membranes (5 μg) were incubated at 25°C for 5 min with or without muscarinic agonist or antagonist in assay buffer supplemented with 100 mM NaCl, 5 mM MgCl2, 0.2% BSA, 5 mM phosphocreatine, 1 mM App(NH)p, 50 U/ml creatine phosphokinase, 0.25 mM ATP, and 1 mM DTT. The reaction was started by addition of 2 μl of 12.5 μM GTP containing 0.5 μCi [γ-32P]GTP (final volume 50 μl) and stopped with 1 ml of 5% (wt/vol) Norit A suspended in ice-cold 50 mM NaH2PO4 (pH 4.5). Tubes were centrifuged (12,000 g for 10 min, 4°C), and a 60-μl aliquot of each supernatant was counted in 10 ml of scintillation fluid. Results are presented as high-affinity GTPase activity, which is defined as the difference between total and nonspecific hydrolysis of [γ-32P]GTP. Nonspecific GTPase activity was determined in the presence of 50 μM unlabeled GTP. In control experiments, total GTPase activity in the presence of 1 mM carbachol was linear for at least 10 min (not shown). As a result, all subsequent assays were conducted for 10 min. Spontaneous release of32Pi (in the absence of membranes) accounted for <1.5% of total added radioactivity.
Synthesis and purification of [α-32P]GTP-azidoaniline.
[α-32P]GTP-azidoaniline was synthesized and purified as described by Fields et al. (9). Briefly, [α-32P]GTP was evaporated under a mild nitrogen stream and dissolved with 3.6% EDAC-HCl in 0.1 M MES (pH 5.6). Nonradiolabeled GTP was added, and the mixture was allowed to set for 10 min. We then added 4% 4-azidoaniline-HCl suspended in peroxide-free 1,4-dioxane and incubated the solution for 16 h at room temperature in the dark with constant rotation. The reaction was terminated by extraction of unreacted azidoaniline three times with 200 μl of water-saturated ethyl acetate. The water phase was spotted onto a flexible polyethylenimine-cellulose plate (5 × 20 cm), and thin-layer chromatography (TLC) was performed at room temperature in the dark by using 0.8 M triethylammonium bicarbonate buffer (pH 7.5) as the mobile phase. TLC-purified [α-32P]GTP-azidoaniline was detected by autoradiography, eluted with 0.5 M NaCl, lyophilized, and reconstituted. Radioactivity was determined, and aliquots were stored at −70°C. Similar procedures were followed for synthesis and purification of nonradiolabeled GTP-azidoaniline, with the exception that the TLC-separated GTP analog was detected with ultraviolet (UV) light (360 nm).
Photoaffinity labeling was performed as described by Fields et al. (9) with minor modifications. Sublingual membranes were resuspended to 1 mg/ml in photolabeling buffer (20 mM Tris · HCl, pH 7.5, 1 mM EDTA, 10 mM MgCl2, 100 mM NaCl, and 1 mM benzamidine) and incubated for 5 min at 30°C with 100 μM of GDP plus an additional 5 min in the presence or absence of agonist or antagonist, as indicated. In some cases, membranes (1 mg/ml photolabeling buffer) were preincubated for 30 min at room temperature with 100-fold molar excess of m1-toxin (relative to membrane high-affinity pirenzepine sites). We then added 2 μl of 750 μM GTP containing 0.2–0.6 μCi of [α-32P]GTP-azidoaniline to initiate each reaction (final volume 50 μl). After 2 min at 30°C, reactions were terminated by placing tubes on ice and adding 500 μl of ice-cold stopping solution (5 mM DTT in photolabeling buffer). Unbound [α-32P]GTP-azidoaniline was removed by centrifugation (12,000 g for 30 min, 4°C), and supernatants were discarded. Pellets were resuspended in 50 μl of stopping solution and exposed on ice to UV light (300 nm, 60 W; Fotodyne Transilluminator) for 2 min (3 cm from light source) and prepared for SDS-PAGE.
SDS-PAGE and autoradiography.
Electrophoresis was as described previously (4) by using 4% stacking and 10% running gels in a Hoeffer minigel system. In some cases, a lower concentration of bisacrylamide (0.12%) was used to obtain better resolution of the isoforms of α-subunits. Gels were stained with Coomassie blue and/or silver as described (4), dried at 60°C with a gel dryer, and exposed to Kodak XAR-5, Biomax MS, or MR film at −70°C. Autoradiograms were scanned with either a laser densitometer (LKB 2202 UltroScan XL) or a Digital Imaging System (IS-1000; Alpha Innotech).
Membrane proteins (50 μg) were subjected to SDS-PAGE, and proteins were transferred to Immobilon-PVDF membranes at 200 mA overnight with a Hoeffer mini transfer apparatus (TE 22). Under these conditions, no proteins of <70 kDa remained in the gel, as assessed by extensive silver staining. Membrane strips were rinsed twice and incubated in blotting solution (0.2% Triton X-100 and 5% nonfat dry milk in PBS) at room temperature for 1 h before overnight incubation with individual antisera diluted in blotting solution. After three rinses with blotting solution, strips were incubated 2–3 h with 500,000 cpm/ml [125I]-Protein A, rinsed, and subjected to autoradiography with Kodak Biomax MS or MR film.
Immunoprecipitation of photolabeled G proteins.
Sublingual membranes were photolabeled as described inPhotoaffinity labeling, and two aliquots were subjected to SDS-PAGE and autoradiography as controls in each experiment to verify carbachol-enhanced photolabeling of 42-kDa proteins. The resultant membrane pellets (90 μg per condition) were solubilized by first being incubated in 15 μl of 2% SDS for 10 min at room temperature. We then added 48 μl of ice-cold solubilization buffer (10 mM Tris · HCl, pH 7.4, 1% wt/vol Nonidet P-40, 1% wt/vol deoxycholate, 0.2% SDS, 150 mM NaCl, 1 mM DTT, 1 mM EDTA, 0.2 mM PMSF, 1 μM GDP, and 10 μg/ml aprotinin), and the mixture was repipetted. Membrane solubilization appeared quantitative, because no specific G proteins were detectable in the subsequent pellets when assessed in preliminary experiments by Western analysis for each antibody used. After centrifugation (14,000 g for 10 min), the supernatant was precleared by adding 3.3 μl of 5 mg/ml purified rabbit IgG in PBS and incubating for 4 h at 4°C, followed by addition of 20 μl of Protein A-agarose (50% suspension, preequilibrated and washed with solubilization buffer + 0.5% SDS), and the mixture was incubated for 1 h at 4°C with continuous rocking. After centrifugation (14,000 g for 10 min), we added 20 μg of antibody IgG (with or without blocking peptide) and 1 ml of cold solubilization buffer to the supernatant and incubated overnight at 4°C with continuous rocking. Protein A-agarose (10 μl of a 50% suspension) was added, and incubation continued for 4–6 h. The mixture was centrifuged (14,000 g for 10 min), and the pellet was washed with 1 ml of solubilization buffer plus 0.5% SDS and then subjected to SDS-PAGE and subsequent autoradiography with Kodak Biomax MS film. In all cases, blocking antigen peptides were used at a 10:1 molar ratio with respect to antibody IgG in the solution. In controls consisting of nonimmune rabbit IgG in place of antibody, we were unable to detect radiolabeled proteins in the final immunoprecipitate, and no IgG was detected in the supernatant when subjected to SDS-PAGE and extensive silver staining.
PIP2 hydrolysis assay.
Measurement of PLC activity was as described by Gutowski et al. (12) but with minor modifications. Membranes were diluted to 0.5 mg/ml with buffer A (1 mM EDTA, 3 mM EGTA, 100 mM NaCl, 5 mM MgCl2, 3 mM DTT, 10 μM GDP, and 50 mM HEPES, pH 7.2), and 10 μl of membrane suspension (5 μg protein) were added to 40 μl of buffer B (3 mM EGTA, 80 mM KCl, 1 mM DTT, 1 μM GDP, and 50 mM HEPES, pH 7.2) containing phospholipid vesicles. The final vesicle suspension was prepared by sonication and contained 1.5 mM PE, 1.5 mM PS, 0.15 mM PIP2, and 7,000–10,000 dpm [3H]PIP2. In some cases, buffer B also contained GTPγS, carbachol, and/or atropine. Ten microliters of 9 mM CaCl2 were then added, and the reaction was started by transferring tubes to a 30°C water bath. Reactions were terminated after 12 min by placing the tubes in ice water, followed by rapid addition of ice-cold 1% BSA (100 μl) and 10% trichloroacetic acid (200 μl). After 5 min, the precipitates were pelleted (12,000 g for 5 min), and 300 μl of the supernatants were removed for determination of radioactivity. In parallel assays, membranes were preincubated with m1-toxin or antibody Z811 against Gαq/11. In the case of antibody Z811, membrane aliquots (0.5 mg/ml buffer A) were preincubated for 20 min at room temperature, followed by 30 min on ice in the absence (control) or presence of either Z811 (1 μM IgG), Z811 premixed with 10 μM of the antigen peptide, or 1 μM rabbit IgG. Portions of each membrane aliquot were then assayed in the absence or presence of GTPγS (1 μM) with and without carbachol (1 mM). In the case of m1-toxin, two equal membrane aliquots (0.5 mg/ml buffer A) were preincubated for 30 min at room temperature with or without m1-toxin (100-fold molar excess of m1-toxin relative to membrane high-affinity pirenzepine sites) and then assayed in the presence of GTPγS (1 μM), carbachol (1 mM) plus GTPγS (1 μM), or atropine (100 μM) plus carbachol (1 mM) plus GTPγS (1 μM).
Membranes (20–50 μg) were resuspended in 0.5–1 ml of binding buffer (20 mM Tris · HCl, pH 7.5, 1 mM EDTA, 3 mM MgCl2, and 0.1 mM PMSF) and incubated with [3H]NMS at 4°C for 3 h to achieve steady-state binding. Incubations were stopped by addition of 4 ml of ice-cold binding buffer, followed by vacuum filtration through Whatman GF/C filters. Atropine (10 μM) was included in the incubation for determination of nonspecific binding. In competition assays, nonradiolabeled agonist (with or without guanine nucleotide) was added simultaneously with 0.5 nM [3H]NMS. Results were analyzed by using the computer programs EBDA and LIGAND, as described previously (27). The dissociation constant (K d) for [3H]NMS in membrane preparations ranged from 0.7 to 1.1 nM, and the proportion of high-affinity pirenzepine sites ranged from 42 to 48% of total binding sites, similar to results reported previously (27).
Preparation of m1-toxin.
Toxin was isolated from lyophilized venom of the Eastern green mamba,D. angusticeps (Sigma Chemical), according to the protocol described by Potter et al. (23), except that material from the initial Sephadex G-50 column was further fractionated by gel filtration chromatography on a column (1.5 × 170 cm) of Sephadex G-25 Superfine eluted with 0.1 M ammonium acetate, pH 6.8, at 4°C. This additional step serves to enrich the preparation for m1-toxin and removes a component possessing anti-rat M3 receptor activity before reverse-phase HPLC. Preparations of m1-toxin were pure as determined by SDS-PAGE and subsequent reverse-phase HPLC. One unit of m1-toxin was similar to that defined by Potter et al. (23): the minimum amount required to block 95% of the specific binding of 0.1 nM [3H]NMS to 0.2 pmol M1 receptors [rat M1-Chinese hamster ovary (CHO) cell membranes] in 10 ml of 50 mM NaH2PO4 plus 1 mM EDTA, pH 7.4, at 25°C for 45 min. Under these standard assay conditions, 1 unit of m1-toxin is equivalent to ∼30 ng toxin protein and corresponds to a molar ratio of m1-toxin to receptor of ∼20:1 (23). Preparations of m1-toxin were specific for the inhibition of M1 receptors as assessed under the standard radioligand binding assay conditions described above, which include 1 unit of m1-toxin and membranes from CHO cell lines expressing either human M1-M5receptor subtypes (kindly provided by Dr. Mark R. Brann) or rat M1 or M3 receptors. In addition, 10 units of m1-toxin had no inhibitory effects on [3H]NMS binding to rat M3 receptors (200:1 molar ratio of m1-toxin to receptors).
Unless indicated, variability is expressed as means ± SE. Statistical comparisons of results was by the Student'st-test with P < 0.05 being significant.
To initially evaluate the effective coupling between muscarinic receptors and G proteins in sublingual membrane preparations, we studied the shift induced by GTPγS in the affinity of the agonist carbachol for [3H]NMS binding sites. As shown in a representative experiment in Fig. 1, GTPγS induced a rightward shift in the carbachol competition curve. Binding data from five separate experiments were fit to models of one, two, and three binding sites, and the best fit was chosen whenP < 0.05 by the F test. In the absence of GTPγS, the best fits were to a two-site model with inhibition constant (K i) values (means ± SE) of 5.4 ± 1.1 and 99.8 ± 17.0 μM, representing 54 ± 3 and 46 ± 3% of the total binding sites, respectively. In the presence of 100 μM GTPγS, the best fits were to a one-site model with a mean K i of 129 ± 14 μM.
A series of experiments were then conducted to determine whether muscarinic receptors in sublingual membranes are coupled to G proteins sensitive to PTX or CTX. We first determined the ADP ribosylation of membrane proteins as a function of toxin concentration. CTX induced the concentration-dependent ADP ribosylation primarily of 44-kDa proteins within sublingual membranes as resolved by SDS-PAGE and subsequent autoradiography. Maximal radiolabeling was achieved at 10 μg/ml CTX (Fig. 2 B). Similar experiments were conducted with PTX, and, as shown in Fig.3 B, proteins of ∼40 kDa were radiolabeled in a concentration-dependent manner with maximal radiolabeling at 25 μg/ml PTX. If proteins were allowed to migrate further into the gel, PTX-dependent proteins began to resolve into two bands of ∼40 and 41 kDa, presumably representing at least two distinct α-subunits (Fig. 3 C, lane 2). To verify that PTX at 25 μg/ml and CTX at 10 μg/ml were saturating for ribosylation of sublingual membrane proteins, we treated membrane aliquots with and without toxin in the absence of [32P]NAD and then reexposed them to toxin in the presence of [32P]NAD. As a control, membrane aliquots were treated with or without toxin in the presence of [32P]NAD. Samples were subjected to SDS-PAGE and autoradiography to first verify toxin-enhanced radiolabeling of membrane proteins of appropriate mass and to determine residual radiolabeling of proteins in membranes pretreated with toxin. In two separate experiments, autoradiographs (1-day exposure) of control samples demonstrated radiolabeling similar to results shown in Figs. 2 and 3, whereas radiolabeled proteins were barely detectable by autoradiography after a 10-day exposure of samples from membranes pretreated with each toxin (not shown). On the basis of these combined results, we used PTX and CTX at 25 and 10 μg/ml, respectively, in further experiments to distinguish whether toxin-sensitive G proteins play a role in carbachol-stimulated signaling events in sublingual membranes.
PTX selectively ADP ribosylates α-subunits of target G protein holotrimers rather than free α-subunits (29). Therefore, to test coupling of sublingual muscarinic receptors to PTX-dependent G proteins, we incubated sublingual membranes before ADP ribosylation with or without agonist (1 mM carbachol) for 15 min at room temperature in ADP-ribosylation buffer that contained 10 μM GTP. Agonist was maintained in the buffer during subsequent ADP ribosylation by PTX in the presence of [32P] NAD. Under these conditions, a proportion of those G proteins that couple to muscarinic receptors would be expected to be activated. Accordingly, if muscarinic receptors are indeed coupled to PTX-sensitive G proteins, then the pool of receptor-coupled α-subunits in the heterotrimeric state, and hence available to PTX-catalyzed ADP ribosylation, should be less than in the absence of carbachol. In contrast, carbachol had no apparent effect on PTX-dependent radiolabeling of either 40- or 41-kDa proteins (Fig.3 C, compare lanes 2 and 3). As a control, PTX-catalyzed ADP ribosylation was completely inhibited when membranes were incubated in the presence of 100 μM GTPγS (Fig.3 C, lane 4). In additional competition binding experiments, we found that PTX-catalyzed ADP ribosylation of sublingual membrane proteins also had no detectable effect on the shift in carbachol binding affinity induced by GTPγS (see Fig.4).
As another means to distinguish coupling of sublingual muscarinic receptors to G proteins sensitive to PTX or CTX, we determined whether either toxin attenuated agonist-enhanced high-affinity GTPase activity. In an initial experiment, GTPase activity was induced by carbachol in a concentration-dependent manner, the effect of which was maximal at 1 mM with an apparent EC50 of ∼7 μM (Fig.5 A). We then assayed high-affinity GTPase activity in response to 1 mM carbachol after prior ADP ribosylation of membranes with maximal concentrations of either PTX or CTX. Basal GTPase activity was reduced after treatment of membranes with each toxin (Fig. 5 B), suggesting that G proteins susceptible to either PTX or CTX contribute to the total basal high-affinity GTPase activity. In contrast, the net increase in carbachol-mediated GTPase activity was unaffected by pretreatment of membranes with either toxin (Fig. 5 B).
Agonist-enhanced photoaffinity labeling of membrane proteins with [32P]GTP-azidoaniline is a valuable tool in identifying the coupling of receptors to G proteins (9). In initial experiments with rat sublingual gland membranes, we found that multiple bands were radiolabeled with [32P]GTP-azidoaniline in the absence of agonist, including bands of 74, 51, 46, 36, and 23 kDa and a predominant band at 42 kDa (Fig.6 A, lane 1). In the presence of carbachol, we consistently (10 separate experiments) observed agonist-enhanced photoaffinity labeling of only 42-kDa proteins (Fig. 6 A, lanes 2 and 3). Photoaffinity labeling of all proteins was unaffected by 100 μM ATP but inhibited completely by either 30 μM guanosine 5′-O-(2-thiodiphosphate) (not shown) or 100 μM GTPγS (Fig. 6 A, lanes 4 and 5). Furthermore, carbachol-enhanced photoaffinity labeling was also insensitive to pretreatment of sublingual membranes with either 25 μg/ml PTX or 10 μg/ml CTX (not shown). Enhancement of photoaffinity labeling of the 42-kDa band in the presence of carbachol was concentration dependent, readily detectable at 0.1 μM, and maximal at ∼100 μM carbachol (Fig. 6 B). The average maximal intensity of radiolabeling induced by carbachol was nearly 60% above basal levels as assessed by densitometric scanning of autoradiographs (Fig. 6 B). In two separate experiments, we found that photoaffinity labeling enhanced by 1 mM carbachol was completely blocked by concentrations of atropine of 0.1 μM or higher (not shown). To assess the contribution of muscarinic M1 receptors to carbachol-enhanced photoaffinity labeling, we first treated sublingual membranes with a 200-fold molar excess of m1-toxin. Basal levels of photoaffinity labeling was unaffected by m1-toxin, whereas carbachol-enhanced labeling was reduced 55% (Fig. 7).
At this stage, our combined results indicated that both M1and M3 receptors couple to G proteins with α-subunits of ∼42 kDa and are insensitive to both PTX and CTX. We therefore conducted Western blot analysis to identify candidate α-subunits within sublingual membranes that may couple to muscarinic receptors. Possible candidates of known G protein α-subunits were considered that satisfied the following criteria: 1) insensitive to both PTX and CTX; 2) ∼42 kDa in electrophoretic mobility; and 3) previously demonstrated to be expressed in epithelial tissues. Candidates include a member (or members) of the Gqfamily, the Gα12 family, and the PTX-insensitive Gi family member Gαz. Previous studies demonstrated that Gα16 and its mouse homologue, Gα15, are restricted to hematopoietic tissues (28), whereas Gαz is distributed primarily to platelets and neurons (17). Thus we did not probe for Gα16 or Gαz. We did probe with antibodies selective for Gαq, Gα11, Gα14, Gα12, and Gα13. Results are shown in Fig. 8. Antibody Z811, raised against the carboxy-terminal region common to both Gαq and Gα11 (22), reacted strongly to proteins of ∼42 kDa in sublingual membranes (Fig.8 A, lane 1). Similar results (Fig. 8 A,lane 3) were obtained with the Gαq-specific antisera W082 (22) and E17 (not shown). To probe for Gα11, we used antibody D17, raised to a peptide equivalent to amino acids 13–29 of the amino-terminal domain unique to Gα 11. As shown in Fig.8 A, lane 5, this antibody reacted with 42-kDa proteins as well as several lower and higher mass proteins in sublingual membranes. Reactivity was blocked by preabsorption with the antigen peptide (Fig. 8 A, lane 6). Similar results were obtained with membranes from rat brains, except that the 42-kDa band was by far the predominant band (not shown). We also used antibody B825, raised against a peptide equivalent to amino acids 114–133 specific for Gα11. This antibody is selective for Gα11, because it binds weakly to recombinant Gαq (13). Reactivity of B825 against 42-kDa proteins was barely detectable in sublingual membranes but very intense in brain membranes (not shown). Antisera raised against carboxy-terminal sequences specific for Gα12(AS233) and Gα13 (AS343) each reacted with proteins in sublingual membranes that displayed a relative mobility of ∼43 kDa (Fig. 8 B, lanes 1 and 3, respectively). Similar results were obtained with antibodies S20 and A20, which were raised against amino-terminal domains of Gα12 and Gα13, respectively (not shown). Antisera CT92 (1), specific for Gα14, recognized a protein in brain but displayed no immunoreactivity to sublingual membrane proteins of appropriate mass (Fig. 8 C,lanes 1 and 2, respectively).
Our results from Western analyses demonstrate the presence of Gαq, Gα11, Gα12, and Gα13 in sublingual membranes and, thus, further suggest that one or more of these proteins may function in coupling to muscarinic receptors. As an approach to distinguish G proteins coupled directly to sublingual muscarinic receptors, we attempted to immunoprecipitate specific G protein α-subunits from membranes after photoaffinity labeling with [32P]GTP-azidoaniline in either the presence or the absence of carbachol. As shown in Fig.9 A, lane 2, radiolabeled proteins of ∼42 kDa were detected in membranes challenged with carbachol, and proteins were subsequently immunoprecipitated with antibody Z811. No radiolabeled proteins were detected in membranes not exposed to carbachol and/or in membranes immunoprecipitated in the presence of antigen peptide (Fig.9 A, lanes 1, 3, and 4). These results suggest that Gαq and/or Gα11 couple to sublingual muscarinic receptors. To discriminate between these two G proteins, we used antibodies E17 (Gαq specific) and D17 (Gα11 specific) in further experiments. Antibody E17 was used rather than antiserum WO82; the latter was raised against an internal sequence of Gαq and was previously demonstrated not to immunoprecipitate native Gαq proteins (12). Furthermore, we confirmed the specificity of each antibody by Western analysis in a preliminary experiment. No signal was detected for either antibody when tested against 50 ng of the alternative recombinant α-subunit protein, although positive controls (5 ng of appropriate recombinant protein, 50 μg each of sublingual membrane and brain membrane proteins) displayed strong signals (not shown). In subsequent immunoprecipitation experiments, antibody E17 produced results similar to those of antibody Z811 (Fig.9 B). When antibody D17 was used with membranes treated without carbachol (Fig. 9 C, lane 1), there was a diffuse band barely apparent at ∼42 kDa, as were bands at positions of higher and lower mass. In the presence of carbachol, only the 42-kDa protein band displayed enhanced radiolabeling (Fig. 9 C,lane 2). Radiolabeling of 42-kDa proteins and most of the higher and lower mass proteins was blocked by inclusion of the Gα11 antigen peptide whether membranes were challenged with carbachol or not (Fig. 9 C, lanes 3 and4). These results indicated both Gαq and Gα11 couple to muscarinic receptors in sublingual membranes. The proteins recognized by antibody D17 of mass greater and less than 42 kDa likely represent unknown cross-reactive guanine nucleotide binding proteins, because similar proteins were detected in Western analysis (Fig. 8 A, lane 5) and after photoaffinity labeling with [32P]GTP-azidoaniline (Fig.6 A, lanes 1–4). Using antibodies A20 (Gα13 specific) or S20 (Gα12 specific), we were unable to detect radiolabeled proteins in immunoprecipitates from membranes incubated either with or without carbachol (not shown). As controls for these experiments, we demonstrated carbachol-enhanced radiolabeling of 42-kDa proteins in separate membrane aliquots that were subjected to SDS-PAGE and autoradiography immediately after radiolabeling (not shown). In addition, we confirmed that antibodies A20 and S20 immunoprecipitated Gα13 and Gα12, respectively, by performing Western analysis of immunoprecipitates using the same antibodies as probes. Thus muscarinic receptors in sublingual membranes appear to couple primarily to Gαq and Gα11. We considered conducting further experiments that would incorporate m1-toxin to distinguish the relative coupling of Gαq and Gα11to muscarinic M1 and M3 receptors in sublingual membranes, but such experiments were determined to be unfeasible because the immunoprecipitation experiments described above were near the limits of detection; autoradiographic film required 4–6 wk of exposure to indicate immunoprecipitated proteins.
To gain insight into the capacity of Gαq relative to that of Gα11 to couple to sublingual muscarinic receptors, we determined the relative abundance of each G protein α-subunit in sublingual membranes. We used Western analysis to correlate the band intensities obtained with increasing amounts of recombinant Gαq (15–40 ng) and Gα11(8–24 ng) proteins with those obtained for two different amounts of sublingual membrane proteins. Antibody D17 (200 ng/ml) was used to probe for Gα11. We used antiserum WO82 (1:1,000) rather than antibody E17 to probe for Gαq because WO82 gave a stronger signal and was in greater supply. As shown in Fig.10, increasing amounts of Gαq and Gα11 displayed linear distributions of band intensities with respect to the amount of protein applied. The average (±SE, n = 4) amount of Gαq in sublingual membranes obtained in two separate experiments in which 15- and 25-μg membrane proteins were used was 1.2 ± 0.2 ng/μg membrane protein. The value derived for Gα11 in two separate experiments in which 50- and 75-μg membrane proteins were used was 0.21 ± 0.03 ng/μg membrane protein (mean ± SE,n = 4). The abundance of Gαq in sublingual membranes is therefore nearly sixfold greater than that of Gα11.
Biochemical studies with cell lines have documented the efficient coupling of M1 and M3 receptors to both Gαq and Gα11 with the resultant activation of PLC-β isoforms (8, 13). Accordingly, we have found that inhibition of PLC blocks muscarinic activation of exocrine secretion by sublingual mucous acinar cells (unpublished observations). Given that muscarinic receptors in sublingual membranes couple predominantly to Gαq and Gα11, we used antibody Z811 to evaluate the direct role of these α-subunits in the muscarinic activated hydrolysis of sublingual membrane PIP2. In preliminary experiments, PIP2hydrolysis was activated by GTPγS in a dose-dependent manner with near-maximal activity (200% above basal level) at 100 μM GTPγS (not shown). Carbachol (1 mM) alone had no effect on PIP2hydrolysis (basal: 80.6 ± 5.2 pmol; carbachol: 82.6 ± 6.5 pmol; n = 7, P = 0.544). GTPγS was required for carbachol activation of PIP2 hydrolysis, with optimal activation at 1 μM GTPγS. Under these conditions, carbachol increased PIP2 hydrolysis an average of 28% above that observed by 1 μM GTPγS alone (carbachol + GTPγS: 127 ± 9.5 pmol; GTPγS: 99.2 ± 7.1 pmol; n = 7,P = 0.002). In subsequent experiments incorporating antibody Z811 (Fig. 11 A), we found the antibody blocked stimulation of PIP2 hydrolysis by 1 μM GTPγS, either with or without carbachol. The antibody was without these affects if first preincubated with its antigen peptide (Fig. 11 A). As an additional control, nonimmune rabbit IgG had no affects on PIP2 hydrolysis stimulated by either GTPγS or carbachol plus GTPγS (Fig. 11 A). In further experiments, we found that specific inhibition of M1receptors attenuates carbachol-induced PIP2 hydrolysis. As shown in Fig. 11 B, PIP2 hydrolysis induced by 1 mM carbachol was reduced nearly 60% by m1-toxin compared with control.
We determined the coupling of muscarinic receptors in rat sublingual gland membranes to specific G protein α-subunits. Coupling was initially verified by the GTP-induced shift in carbachol high-affinity binding and by high-affinity GTPase activity. In both cases, the effects of carbachol were insensitive to prior treatment of membranes with PTX, although PTX catalyzed the ADP ribosylation of membrane proteins of molecular mass consistent with that of Gαi and Gαo subunits. In a similar manner, CTX catalyzed the ADP ribosylation of 44-kDa membrane proteins (presumably Gαs) but had no effect on subsequent carbachol-induced GTPase activity. Therefore, α-subunits susceptible to ADP ribosylation by either PTX or CTX do not appear to regulate directly muscarinic receptor-induced cellular functions in sublingual glands. Toxin-sensitive G proteins in sublingual membranes may instead mediate other receptors within sublingual glands such as vasoactive intestinal peptide receptors and α2-adrenergic receptors, presumably coupled to CTX-sensitive Gαs and PTX-sensitive Gαi, respectively (3).
Photoaffinity labeling of sublingual membrane proteins with [32P]GTP-azidoaniline detected coupling of muscarinic receptors to 42-kDa G protein α-subunits. As determined by Western analysis, Gαq and Gα11 in sublingual membranes displayed the same electrophoretic mobility as the 42-kDa proteins identified by carbachol-enhanced photoaffinity labeling. In addition, antibodies specific for Gαq or Gα11 immunoprecipitated 42-kDa proteins with enhanced radioactivity after membranes were photoaffinity labeled in the presence of carbachol. These results suggest sublingual muscarinic receptors are coupled predominately to Gq and G11.
The absence of Gα14, as assessed by Western analysis of sublingual membranes, was not anticipated because Gα14 is expressed in other epithelial and glandular tissues, such as lung, kidney, thymus, and testis (28). Both Gα12and Gα13 were detected in sublingual membranes, although the electrophoretic mobility of each α-subunit was slower than the that of the 42-kDa proteins recognized by photoaffinity labeling. Moreover, we were unable to immunoprecipitate detectable amounts of radiolabeled Gα12 and Gα13 after photoaffinity labeling of membranes in the presence or absence of carbachol. Although we cannot rule out the possibility that these latter negative results reflect the limited sensitivity of the immunoprecipitation assay with native tissue, it appears that neither Gα12 nor Gα13 couples significantly to sublingual muscarinic receptors. Correspondingly, both Gα12 and Gα13 have only been shown to couple to nonmuscarinic receptors including thrombin, thromboxane A2, bradykinin, and lysophosphatidic acid receptors. In addition, both α-subunits regulate cellular events other than the PLC or adenylyl cyclase pathways, such as Na+/H+exchange and the c-Jun mitogenic pathway (7). The direct effector(s) of Gα12 and Gα13 signaling have yet to be established, although recent evidence suggests a role for guanine nucleotide exchange factors (18). Further studies of sublingual acinar cells are thus needed to elucidate the functional role of Gα12 and Gα13.
In sublingual glands, muscarinic receptor activation results in the generation of inositol 1,4,5-trisphosphate (30) and calcium-dependent fluid secretion (20). Furthermore, recent results from our laboratory indicate that muscarinic-induced exocrine secretion is also calcium dependent and is mediated by activation of PLC (unpublished observations). In the present investigation, we confirm the muscarinic stimulation of PIP2 hydrolysis in sublingual membranes. More importantly, Z811 (Gαq/11 specific antibody) completely blocked this response, suggesting that activation of PLC activity in sublingual membranes by muscarinic receptors is mediated predominantly, if not totally, by Gαq and Gα11. The absence of muscarinic receptor coupling to PTX-sensitive G proteins (i.e., Gi proteins) in sublingual glands further argues against a role for G protein βγ-subunits in the muscarinic activation of PLC.
A confounding factor in elucidation of muscarinic regulation of exocrine secretion by sublingual mucous acini is the expression of approximately equivalent amounts of M1 and M3receptor subtypes (4). Moreover, immunolocalization experiments indicate the coexpression of both M1 and M3 receptor subtypes on sublingual mucous acinar cells (unpublished observations). Blockade of M1 receptors by m1-toxin results in a 40–50% decrease in the maximal muscarinic receptor-induced exocrine response (5). Although m1-toxin reduces the maximal secretory response, there is no appreciable change in the apparent EC50 for agonist-induced secretion (5). In the present study, we found that m1-toxin inhibits approximately one-half of the carbachol-enhanced photoaffinity labeling of 42-kDa G protein α-subunits as well as the muscarinic activation of PLC activity. Collectively, these results are consistent with equivalent amounts of M1 and M3 receptors expressed by acinar mucous cells that function separately to regulate exocrine secretion through Gαq/11 activation of PLC. A maximal muscarinic receptor-induced secretory response requires activation of both receptor subtypes.
Carbachol-enhanced photoaffinity labeling of 42-kDa proteins was half-maximal (EC50) between 1 and 10 μM carbachol (Fig.6 B), similar to the EC50 (∼7 μM) for carbachol-stimulated GTPase activity. The high-affinity agonist binding site of sublingual muscarinic receptors (K i of carbachol ∼5 μM) thus mediates receptor coupling to G proteins. In contrast, the EC50 reported previously for carbachol-stimulated exocrine secretion by acinar cells is 0.3 μM (3), ∼10-fold lower than for receptor-G protein coupling. This discrepancy in agonist affinities may reflect either an abundance of muscarinic receptors or their associated G proteins linked to exocrine secretion within acinar mucous cells. M3receptors are not in excess, because secretion induced by these receptors alone (after m1-toxin blockade of M1 receptors) are insufficient to induce a maximal secretory response. Because the number of M1 receptors expressed in isolated acini is similar to that for M3 receptors, and M1receptors are required for at least half of the maximal muscarinic secretory response, we speculate they are also not in great excess. On the other hand, the density of muscarinic receptors in sublingual gland membranes is 462 fmol/mg protein (4). Based on our estimates from Fig. 10 for the levels of Gαq (28.6 pmol/mg protein) and Gα11 (5.0 pmol/mg protein) in membranes, the calculated molar ratio for each α-subunit relative to the total pool of muscarinic receptors is 62:1 for Gαqand 11:1 for Gα11. These ratios are doubled if one only considers each receptor subtype separately given that approximately equivalent amounts of M1 and M3 receptors are expressed in sublingual membranes (4). These calculations suggest both Gαq and Gα11 are in abundance relative to muscarinic receptors, which may thus account, in part, for the ∼10-fold lower EC50 for carbachol-stimulated exocrine secretion compared with receptor-G protein coupling. The nearly sixfold higher expression of Gαq relative to Gα11further suggests that Gq rather than G11functions more appreciably in the coupling of sublingual M1and M3 receptors to downstream effectors. Of course, it is assumed these estimates are indeed representative of the pools of α-subunits functionally interactive with each receptor subtype in mucous cells.
With respect to mucous cell exocrine function, it is unclear why both M1 and M3 receptor subtypes serve to regulate secretion in an apparently redundant manner (i.e., coupled to Gαq/11 and activation of PLC) and whether both receptors are required to elicit a maximal secretory response. One explanation may be that each receptor subtype and its corresponding downstream elements are specifically compartmentalized, serving to form a segregated intracellular signaling network (11). As an example, a comparison of native vs. expressed exogenous muscarinic receptors in Xenopus oocytes suggests coupling to distinct calcium pools (15). There is also evidence that activation of the PLC pathway may potentially modulate phospholipase D or phospholipase A2 (8). One may thus speculate that sublingual M1 and M3 receptors are spatially segregated and linked to different pools of Gαq/11, PLC isozymes, and possibly other signaling pathways to regulate distinct mucous cell functions other than secretion, such as cell metabolism or gene expression. Clarification must await delineation of the subcellular localization of M1 and M3 receptors and their associated networks of downstream signaling components in mucous cells.
Our results further accentuate differences among mammalian salivary exocrine acinar cell types. Muscarinic activation of PLC and subsequent intracellular calcium mobilization is the major pathway eliciting fluid secretion in all three major salivary glands: parotid, submandibular, and sublingual glands (2). In contrast with sublingual glands, exocrine secretion by serous acinar cells of parotid glands and seromucous acinar cells of submandibular glands are both regulated primarily via β-adrenergic receptors and the cAMP pathway (2). Furthermore, M3 receptors are by far the predominant muscarinic receptor subtype in parotid glands (6,26). In rat parotid glands, M3 receptors are coupled to both Gαq and Gαi1 (6). Also, parotid M3 receptors share an apparent redundancy with α1-adrenergic receptors to mediate activation of PLC via Gαq/11 (24), whereas in rat sublingual glands, α1-adrenergic receptors are not expressed (19). Signaling events regulating specific cell functions are therefore not necessarily ubiquitous for the discrete exocrine cell types that may be present, either together or independently, within the numerous and diverse salivary glands lining the oral cavity. An understanding of such differences is required if fully effective therapeutic treatments are to be developed to treat those patients suffering from hyposalivary function.
We thank Dr. Alan V. Smrcka for assistance in setting up the PIP2-hydrolysis assays and for helpful discussions.
This study was supported by National Institute of Dental Research Grant DE-10480.
Present address of W. Luo: Department of Cell and Cancer Biology, Division of Clinical Sciences, National Cancer Institute/National Institutes of Health, Bldg. 10, Rm. 3B47, 9000 Rockville Pike, Bethesda, MD 20892.
Address for reprint requests and other correspondence: D. J. Culp, Center for Oral Biology, 601 Elmwood Ave., Box 611, Rochester, NY 14642-8611 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2001 the American Physiological Society