Interaction of GTP-binding Regulatory Proteins with Chemosensory Receptors*

GTP-binding regulatory proteins (G-proteins) were identified in chemosensory membranes from the channel catfish, Ictalurus punctatus. The common G-pro- tein &subunit was identified by immunoblotting in both isolated olfactory cilia and purified taste plasma membranes. A cholera toxin substrate (Mr 45,000), corresponding to the G-protein that stimulates adenylate cyclase, was identified in both membranes. Both membranes also contained a single pertussis toxin sub- strate. In taste membranes, this component co-mi-grated with the a-subunit of the G-protein that inhibits adenylate cyclase. In olfactory cilia, the M. 40,000 pertussis toxin substrate cross-reacted with antiserum to the common amino acid sequence of G-protein a- subunits, but did not cross-react with antiserum to the a-subunit of the G-protein from brain of unknown function. The interaction of G-proteins with chemosensory receptors was determined by monitoring receptor binding affinity in the presence of exogenous guanine nucleotides. L-Alanine and L-arginine bind with simi- lar affinity to separate receptors in both olfactory and gustatory membranes from the catfish. GTP and a nonhydrolyzable analogue decreased the affinity of olfactory L-alanine and L-arginine receptors by about 1 order of magnitude. In contrast, the binding affinities of the corresponding taste receptors were unaffected. These results suggest that olfactory receptors are func- tionally coupled to G-proteins in a manner similar to some hormone and neurotransmitter receptors. The GTP-binding regulatory proteins (G-proteins),’ Measurements-Assays L-ala-nine L-arginine assay binding? Assays in or PM radioligand in the (total and to IO-' M unlabeled curves. samples

Interaction of GTP-binding Regulatory Proteins with Chemosensory Receptors* (Received for publication, October 3, 1986) Richard C. BruchS and D. Lynn Kalinoski From the Monell Chemical Senses Center, Philadelphia, Pennsylvania 19104 GTP-binding regulatory proteins (G-proteins) were identified in chemosensory membranes from the channel catfish, Ictalurus punctatus. The common G-protein &subunit was identified by immunoblotting in both isolated olfactory cilia and purified taste plasma membranes. A cholera toxin substrate (Mr 45,000), corresponding to the G-protein that stimulates adenylate cyclase, was identified in both membranes. Both membranes also contained a single pertussis toxin substrate. In taste membranes, this component co-migrated with the a-subunit of the G-protein that inhibits adenylate cyclase. In olfactory cilia, the M. 40,000 pertussis toxin substrate cross-reacted with antiserum to the common amino acid sequence of G-protein asubunits, but did not cross-react with antiserum to the a-subunit of the G-protein from brain of unknown function. The interaction of G-proteins with chemosensory receptors was determined by monitoring receptor binding affinity in the presence of exogenous guanine nucleotides. L-Alanine and L-arginine bind with similar affinity to separate receptors in both olfactory and gustatory membranes from the catfish. GTP and a nonhydrolyzable analogue decreased the affinity of olfactory L-alanine and L-arginine receptors by about 1 order of magnitude. In contrast, the binding affinities of the corresponding taste receptors were unaffected. These results suggest that olfactory receptors are functionally coupled to G-proteins in a manner similar to some hormone and neurotransmitter receptors.
The GTP-binding regulatory proteins (G-proteins),' G., Gi, G,, and transducin, are a family of homologous signal transducing polypeptides that couple many receptors to intracellular effector systems (1,2). These proteins are heterotrimeric, sharing a common M, 35,000 @-subunit (3)(4)(5). Although the M , 8,000 y-subunits exhibit identical electrophoretic mobilities, peptide mapping (4) and immunoblotting analysis (5) indicate that the y-subunit of transducin is not identical to that of the other G-proteins. Identification of the individual * This work was supported in part by Grant BRSG S07-RR05825-07 from the Biomedical Research Support Grant Program, the National Institutes of Health, a grant from the Veterans Administration ( t o Dr. J. G. Brand), and United States Public Health Service Grant CA 39712 (to Dr. D. R. Manning). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed Monell Chemical Senses Center, 3500 Market St., Philadelphia, PA 19104.
G-proteins is dependent on the characteristic a-subunits since these subunits differ in molecular size (1,2), susceptibility to bacterial toxin-catalyzed ADP-ribosylation (l), immunoreactivity (6,7), and function (1,2). G. and Gi mediate stimulation and inhibition of adenylate cyclase, respectively (l), whereas transducin mediates activation of retinal cyclic-GMP phosphodiesterase (2). The function of Go, found predominantly in brain, remains undefined, although it has been suggested that Go may be involved in phosphoinositide turnover (8,9).
The transduction mechanisms subsequent to the initial interaction of stimuli with chemosensory receptors have not been established (10, 11). L-Amino acids are effective stimuli for the channel catfish (Ictulurus punctatus) that elicit both olfactory and gustatory responses in electrophysiological assays (12, 13). L-Alanine and L-arginine do not significantly cross-adapt in these assays in either chemosensory system, suggesting that these stimuli interact with different receptors and/or transduction pathways (12, 13). Receptor binding studies indicate that these amino acids interact with separate receptors in both isolated olfactory cilia and a sedimentable fraction from taste epithelium (14).* The biochemical binding data are therefore consistent with the lack of significant crossadaptation observed in electrophysiological assays and suggest that different receptor proteins may account for the crossadaptation results (12, 13). In addition, the ligand binding activity of olfactory and gustatory receptors for L-alanine and L-arginine are differentially inhibited by l e~t i n s .~ Olfactory receptors for these amino acids were inhibited to various extents by lectins. In contrast, the corresponding gustatory receptors were unaffected. Taken together, the receptor binding results suggest that, despite ligand similarity, the receptors themselves are not identical molecular species in the two chemosensory membranes: In this report, we describe the identification of G-proteins in both chemosensory membranes from the catfish by bacterial toxin-catalyzed ADP-ribosylation and immunoreactivity. Two G-proteins were identified in both olfactory cilia and purified taste plasma membranes. Although both membranes contained G-proteins, exogenous guanine nucleotides decreased the binding affinities of olfactory L-alanine and Larginine receptors, but did not affect the corresponding gustatory receptors. About 0.3 pg of purified Go 8-subunit (Go), 10 pg of olfactory cilia protein (OC), and 10 pg of taste plasma membrane protein (TM) were applied on a 10% separation gel. After electrophoresis and protein transfer, the nitrocellulose membrane was incubated with antiserum to the G-protein j3-subunit (1:lOO dilution) as described (7,22). Only the M, 35,000 region of the immunoblot is shown since no other components were observed.
of G-protein a-subunits were generous gifts from Dr. D. R. Manning (Department of Pharmacology, University of Pennsylvania). The specificities of these antisera and isolation of the subunits were described previously (7,15,16). Peroxidase-conjugated goat antirabbit IgG and unlabeled L-amino acids were obtained from Sigma. ~- [2,3-~H]Alanine and ~- [4,5-~H]arginine (both 50 Ci/mmol) were obtained from ICN Radiochemicals. Prestained molecular weight marker proteins were obtained from Bethesda Research Laboratories. GTP, Gpp(NH)p, and GTP+ were obtained from Sigma and Boehringer Mannheim.
Membrane Preparations-Olfactory cilia were isolated by treatment of the olfactory epithelium with 10 mM CaCll in 20 mM Tris. HC1, pH 7.2, containing 300 mM sucrose and 10 pg/ml leupeptin.' A sedimentable fraction from taste epithelium (fraction P2) was prepared as previously described (14). Purified plasma membranes (fraction Bo) were prepared from fraction P2 by sucrose gradient centrifugation (17). Liver and brain membranes were prepared by differential centrifugation in 10 mM Tris.HC1, pH 7.8, containing 1 mM CaC12.' Receptor Binding Measurements-Assays for radiolabeled L-alanine binding were performed as previously described (14,18,19). Assays for radiolabeled L-arginine binding were performed by a modification of the previously described assay (14,18,19) to reduce nonspecific binding? Assays were performed in duplicate or triplicate samples that varied by 5%. Assays were in 1.1-ml final volume containing 1 PM radioligand in the absence (total binding) and presence of to IO-' M unlabeled ligand to obtain competitive binding curves. Nonspecific binding was determined in parallel samples containing 20 mM unlabeled ligand. Assays were incubated on ice for 1 h, rapidly filtered, and washed as described (14,18,19): When present, GTP and its nonhydrolyzable analogues were added to the assay mixture at 100 pM final concentration.
Other Methods-ADP-ribosylation in the presence of activated cholera and pertussis toxins and autoradiography were performed as described previously (16). Polyacrylamide gel electrophoresis (20), electrophoretic transfer of proteins to nitrocellulose (21), and immunoblotting were performed as described (7,22). Protein was determined by the Bradford method (23) using bovine serum albumin as a standard.

Identification of G-proteins in Chemosensory Membranes-
Two methods were used to identify G-proteins in isolated olfactory cilia and purified taste plasma membranes. G-proteins in these membranes were identified by immunoreactivity with polyclonal antisera of designed specificity (7,15,16) and bacterial toxin-catalyzed ADP-ribosylation (1,16). Immunoblotting analysis of both chemosensory membranes with antiserum to the common G-protein P-subunit indicated the presence of this subunit in both membranes. A single band (MI 35,000) was specifically labeled in both membranes (Fig. l), indicating the presence of G-proteins in these membranes. In addition, this component was not labeled by normal (nonimmune) rabbit serum (not shown).
G-protein a-subunits were identified based on their characteristic molecular sizes, immunoreactivity, and labeling by A. G. Boyle, Y. S. Park, T. Huque, and R. C. Bruch, manuscript submitted for publication. ADP-ribosylation. In both membranes, a cholera toxin substrate, (MI 45,000) was ADP-ribosylated ( Fig. 2A). Based on cholera toxin sensitivity and molecular size, this substrate probably corresponded to the a-subunit of G. (1). However, no M , 52,000 cholera toxin substrate was detected (1,15). Two additional bands (Mr -80,000 and 42,000) were also cholera toxin substrates (Fig. 2 A ) . The identities of these components are not known. However, the possibility that the M, 42,000 component may be a proteolytic fragment derived from the larger species cannot be rigorously excluded.
An additional component was also labeled in both chemosensory membranes by pertussis toxin-catalyzed ADP-ribosylation (Fig. 2B). In taste membranes, this component comigrated with the a-subunit of Gi. However, in olfactory cilia, the pertussis toxin substrate (Mr 40,000) did not co-migrate with the a-subunit of Gi. The nature of the olfactory pertussis toxin substrate was therefore investigated by immunoblotting analysis. The pertussis toxin substrate in olfactory cilia crossreacted with antiserum to the common amino acid sequence of G-protein a-subunits (Fig. 3A) (7). B, about 50 pg of protein of olfactory cilia ( l a n e 1 ), taste plasma membranes ( l a n e 2), liver membranes (lane 3), brain membranes ( l a n e 4 ) , and mouse brain membranes ( l a n e 5) were immunoblotted with antiserum (1:lOO dilution) to the a-subunit of Go (7). tected (Fig. 3A). The same result was obtained with two such antisera from different animals (not shown). To characterize further the identity of the M, 40,000 pertussis toxin substrate, immunoblotting analysis was performed using antiserum to the a-subunit of Go. This antiserum does not cross-react with the a-subunits of GB or Gi, but does cross-react with the common G-protein 8-subunits (7). In both olfactory cilia and taste plasma membranes, no immunoreactivity corresponding to the Mr 39,000 a-subunit of Go was observed, although the &subunits were readily labeled (Fig. 3B). The same result was obtained when up to 200-cg protein samples were applied to the gels. To verify that the antiserum cross-reacted with G, from the catfish, brain membranes from both mouse and catfish were also analyzed for immunoreactivity. Liver membranes from the catfish were also analyzed since liver contains G, and Gi but lacks Go (1,7). As expected from the specificity of the antiserum (7), no G-protein a-subunit immunoreactivity was observed in liver membranes. However, the a-subunit of Go was readily detected in both mouse and catfish brain membranes (Fig. 3B), thereby excluding the possibility that Go from catfish did not cross-react with the antiserum. In combination, these results indicated that both chemosensory membranes lacked detectable Go.

Effect of Exogenous
Guanine ~~~# t~e~ on e~~~e~o~ Receptor Binding-Signal transduction mechanisms of a variety of hormone and neurotransmitter receptors are dependent on the functional interaction of G-proteins with the appropriate agonist-occupied receptor (1). The addition of exogenous guanine nucleotides to receptor binding assays is an established approach for demonstrating functional coupling of receptors and G-proteins. A decrease in receptor binding affinity for agonist in the presence of exogenous guanine nucleotides is taken to indicate G-protein activation and, consequently, uncoupling of receptor and G-protein (24, 25). We have therefore used this approach to investigate chemosensory receptor/(;-protein interactions. In the absence of guanine nucleotides, L-alanine and L-arginine interact with a single class of separate olfactory receptor sites as indicated by Scatchard analysis of competitive binding curves? In the presence of GTP, the binding affinities of olfactory receptors for both L-alanine and L-arginine were decreased as indicated by the shifts of the binding curves to higher ligand concentrations (Fig. 4). For the olfactory L-alanine receptor, halfmaximal binding inhibition was shifted from 1.8 PM in the absence of GTP to 4 FM in the presence of GTP. Similarly, for the olfactory I;-arginine receptor, half-maximal binding binding was determined in the absence (control) and presence of 100 PM GTP or Gpp(NH)p with 1 p~ radioligand in the presence of the indicated concentrations of unlabeled ligand. For each ligand, the data are the average of duplicate assays from two independent experiments for each nucleotide. inhibition was shifted from 2.8 PM in the absence of GTP to 10 PM in the presence of GTP. The nonhydrolyzable analogue GPP(NH)p was equally effective as GTP in its ability to affect both olfactory receptors (Fig. 4). In contrast, under the same conditions, no change in binding affinity of the gustatory receptors for the same amino acids was observed in the presence of GTP, Gpp~NH)p, or GTP-@ (not shown). Addition of divalent cations such as M e and Mn2+ at 1 m M concentrations (24,25) and use of different lots of nucleotides from different commercial sources did not affect these receptors. Fourteen independent experiments (eight for L-alanine and six for L-arginine) under these various conditions failed to produce a significant change in ligand binding affinity of either gustatory receptor.

DISCUSSION
In this paper, we have described the identification of Gproteins in both olfactory and gustatory receptor membranes from the catfish. Three criteria were used to identify these proteins based on the characteristic properties of the asubunits. These criteria were molecular size, specificity of bacterial toxin-catalyzed ADP-ribosylation, and immunoreactivity with polyclonal antisera of designed specificity (1,7). In both chemosensory membranes, a cholera toxin substrate (M, 45,000) was identified that probably corresponds to the a-subunit of G, (1). The M, 52,000 form of this protein (1) was not detected in either membrane. In addition, both membranes contained a single pertussis toxin substrate. In taste membranes, this component co-migrated in polyacrylamide gels with the a-subunit of Gi. However, in olfactory cilia, the M, 40 To our knowledge, this is the first report of the direct identification of G-proteins in a plasma membrane fraction from a vertebrate taste system. These proteins have, however, been identified previously in isolated olfactory cilia from frog (27)(28)(29). In agreement with the previous results in frog cilia (27-29), we have also detected a cholera toxin substrate in the catfish membranes that probably corresponds to G,. However, in frog cilia, the a-subunits of Gi and Go have also been detected by pertussis toxin-cat~yzed labeling (27, 28) or by immunoreactivity (29). Based on electrophoretic mobility and immunoreactivity, the catfish M, 40,000 component does not appear to be identical to the a-subunits of either Gi or Go. The significance of the differences in G-protein compositions between these animal models requires further investigation.
The functional interaction of chemosensory receptors was evaluated by monitoring receptor binding affinity in the presence of exogenous guanine nucleotides. Due to favorable receptor binding affinities for agonist amino acids, the chemosensory systems of the catfish represent attractive model systems for the study of chemosensory receptor properties (14).'" We have shown previously that L-alanine and Larginine do not compete effectively for common binding sites in either olfactory or gustatory membra ne^.^^^ Based on the ability of lectins to inhibit olfactory receptors for these amino acids, but not the corresponding gustatory receptors, these receptors are probably not identical molecular species in the two membra ne^.^ Similarly, the results of the present study suggest that these receptors also differ in their mode of interaction with G-proteins. The binding affinities of L-alanine and L-arginine receptors in taste membranes were unaffected by exogenous guanine nucleotides under conditions that were shown previously to uncouple other receptors from G-proteins (24,25,30). The binding affinities of both olfactory L-alanine and L-arginine receptors were significantly decreased by GTP and by a nonhydrolyzable GTP analogue. These results suggest that olfactory receptors are functionally coupled to G-proteins in a manner similar to those hormone and neurotransmitter receptors that depend on G-protein interaction for signal transduction (1, 24, 25, 30). These results are also consistent with the demonstration of odorantstimulated, guanine nucleotide-dependent activation of adenylate cyclase in frog olfactory cilia (27) and phosphoinositide turnover in catfish cilia (22), which implicate G-protein involvement in mediating both pathways. However, the results of the present study are the first describing the interaction of olfactory receptors with G-proteins at the level of the initial binding event.