Inhibition of G Protein-coupled Receptor Signaling by Expression of Cytoplasmic Domains of the Receptor*

The third intracellular domain (3i) of G protein- coupled receptors plays a major role in the activation of G proteins. Alterations in this region of the receptor can affect receptor/(= protein coupling efficiency and speci- ficity. We recently reported (Luttrell, L. M., Cotecchia, S., Ostrowski, J., Kendall, H., Leaowitz, R. J. (1993) Science 259,1463-1457) that coexpression of the 3i domain of the alB adrenergic receptor ( A R ) (alB3i) specifically inhibited a&R-mediated inositol phosphate production, with no effect on Dl, dopamine receptor (D,DR)-mediated CAMP production. Similarly, expression of the 3i domain of DIADR (DIA3i) inhibited DIADR-mediated CAMP produc- tion but did not affect a,@-mediated inositol phos- phate accumulation. This suggests that peptides derived from a G protein-coupled receptor might serve as antago- nists of receptor/(; protein interactions. The present studies were performed to test the generality as well as the specificity of this phenomenon. The effect of expression of the second intracellular domain (2i), the 3i do- main, and the fourth intracellular domain (4i) of cy1@ on second messenger generation mediated by the a&R, the M, muscarinic cholinergic receptor (M,AChR), and the DIADR was examined. Although the alB2i of homolo-regulatory protein; G,, the G protein that stimulates adenylyl cyclase; protein calcium-dependent phospholipase; a-subunit construct. Saturation binding (B,,,=) was determined as described. R,,, represents the amount of basal IP production in cells in arbitrary units such that 1 unit is the basal amount of IP produced in control cells co-transfected with receptor plus pRK5 vector. R,, represents the amount of IP production in cells following stimulation with a maximally efficacious dose of agonist in arbitrary units such that 1 unit is the basal amount of IP produced in control cells. Inhibition of R , represents the percent inhibition of the agonist-stimulated increase in IPS over basal, compared with the control cells. Statistical significance was assessed using the Student's paired t test.

The third intracellular domain (3i) of G proteincoupled receptors plays a major role in the activation of G proteins. Alterations in this region of the receptor can affect receptor/(= protein coupling efficiency and specificity. We recently reported (Luttrell, L. M., Cotecchia, S., Ostrowski, J., Kendall, H., Leaowitz, R. J. (1993) Science 259,1463-1457) that coexpression of the 3i domain of the alB adrenergic receptor ( A R ) (alB3i) specifically inhibited a&R-mediated inositol phosphate production, with no effect on Dl, dopamine receptor (D,DR)-mediated CAMP production. Similarly, expression of the 3i domain of DIADR (DIA3i) inhibited DIADR-mediated CAMP production but did not affect a,@-mediated inositol phosphate accumulation. This suggests that peptides derived from a G protein-coupled receptor might serve as antagonists of receptor/(; protein interactions. The present studies were performed to test the generality as well as the specificity of this phenomenon. The effect of expression of the second intracellular domain (2i), the 3i domain, and the fourth intracellular domain (4i) of cy1@ on second messenger generation mediated by the a&R, the M, muscarinic cholinergic receptor (M,AChR), and the DIADR was examined. Although the alB2i domain had no effect on receptor/(= protein coupling for any receptor tested, the aIB3i domain inhibited signaling mediated by a l p and M,AChR but not by DIADR, while the alB4i domain inhibited signaling mediated by each of the receptors. To investigate the generality of 3i domain-induced inhibition of receptor activity further, the 3i domains of two G,-coupled receptors (a1@ and M,AChR) and two Gi-coupled receptors (a,AR and MAChR) were tested for effects on the second messenger generation mediated by each of the four receptors. In each case, the homologous 3i domain caused significant inhibition (40-60%), while the 3i domain of the receptor coupled to the same G protein also decreased receptor/(= protein coupling. In contrast, receptorIG protein coupling appeared unaffected by expression of 3i domains derived from receptors coupled to different G proteins. The alB3i domainprovoked inhibition of homologous receptor signaling was surmountable at high receptor density, and assays using a phorbol response elementheporter gene construct detected a weak enhancement of basal second messenger generation in cells expressing the aIB3i domain alone. These data demonstrate that disruption of receptorIG protein coupling by 3i domain peptides 1) can be generalized to both G,and G,-coupled receptors and * This work was supported in part by Grant HL16037 from the National Institutes of Health. The costs of publication of this article were therefore be hereby marked "advertisement" in  2) likely results from the ability of these peptides to act as competitive weak partial agonists at the receptorIG protein interface.

~~~~~~~~~ ~
The family of receptors coupled to heterotrimeric guanyl nucleotide-binding proteins (G proteins)' contains members from several classes including adrenergic receptors, muscarinic cholinergic receptors, rhodopsin, the visual color opsins, several neurotransmitter receptors, and many peptide receptors. G protein-coupled receptors share a common putative structural topography composed of seven transmembrane domains separated by four extracellular regions and four intracellular regions (1). The extracellular regions form pockets in which specific agonists can interact or bind (2, 3). Mutagenesis studies have demonstrated that intracellular regions, particularly the 3i domain, interact with G proteins to stimulate effector systems (4-6).
Studies utilizing chimeras of a2 and Pz adrenergic receptors (a-, &AR) have shown that the 3i domain provides structural determinants for specificity of receptorIG protein interaction (7). Similarly, substitution of a portion of the 3i domain of the a,,AR, a phospholipase C stimulatory G,-coupled receptor (8-121, into the corresponding region of the G,-coupled p@ results in a chimera capable of coupling to G, (13,14).
Noting the critical role of the 3i domain in receptor/G protein coupling, we recently demonstrated (15) that cellular expression of the 3i domains of alBAR or of the G,-coupled adenylyl cyclase stimulatory DIADR (16) results in specific inhibition of agonist-stimulated second messenger generation mediated via the homologous receptor. This suggests that disruption of receptor/(;. protein interaction may represent a novel form of receptor antagonism with possible therapeutic applications. Although considerable evidence implicates the 3i domain of several G protein-coupled receptors as containing major determinants of selectivity in receptor/G protein coupling (131, other intracellular domains also apparently play a role. Peptides derived from the 2i, 3i, andor 4i domains of the &AR (17, 181, aMAR (19,20), M,AChR (19), and rhodopsin (21) have been shown to interact with G proteins in vitro. The present study explores the generality, specificity, and possible mechanism of the inhibitory effects of intracellular receptor domains on receptor-mediated signal transduction. The effects of cellular expression of the 2i, 3i, and 4i domains of the aIBAR on homoloregulatory protein; G,, the G protein that stimulates adenylyl cyclase; The abbreviations used are: G protein, guanyl nucleotide binding Gi, the G protein that inhibits adenylyl cyclase; G,, the G protein that Zi, 3i, and 4i, the second, third, and fourth intracellular domain pep-stimulates calcium-dependent phospholipase; G p , the a-subunit of G,; tides of G protein-coupled receptors; AR, adrenergic receptor; IP, inositol phosphate; PI, phosphoinositide; CAT, chloramphenicol acetyltransferase; PT, B. pertussis toxin; ['2sIlHEAT, Z-(P-(4-hydroxy-3-['2sIliodophenyl)ethylaminomethyl)tetralone; AChR, acetylcholinergic receptor. gous receptor signal transduction are described. The effects the 3i domains from two G,-coupled (alBAR and MIAChR) and two G,-coupled (a,& and MdChR) receptors on signaling mediated by each receptor are also assessed.

EXPERIMENTAL PROCEDURES
Materials-The full-length cDNAs encoding the aIBAR (22) and human a2,AR (1) were isolated in our laboratory. The cDNA encoding the D,,DR (16) was from M. G. Caron, and those encoding the human MIAChR and M. @X-& (23) were from E. G. Peralta. The -356 foslCAT construct (24) was provided by M. Z. Gilman. Rabbit polyclonal antisera against the alB3i domain, the alB4i domain, and the azA3i domain were prepared in our laboratory. Antisera against the M13i and Mz3i domains were provided by M. Brann.
Restriction endonucleases and T4 DNA ligase were from Promega; T7 sequencing kit, protein A-Sepharose, and deoxynucleotide triphosphates were from Pharmacia Biotech Inc. HEK-293 and COS-7 cells were from the American Type Culture Collection. Tissue culture media, fetal bovine serum, and gentamicin were from Life Technologies, Inc. Corp. Atropine, (-)-epinephrine, yohimbine, prazosin, chloramphenicol, chloramphenicol acetyltransferase, isobutylmethylxanthine, soybean trypsin inhibitor, and aprotinin were from Sigma. Phentolamine was from Ciba-Geigy. cis-Flupentixol was from Research Biochemicals, Inc. UK-14304 was from Pfizer; fenoldopam was from Smith Kline and Beecham. Dowex AG1-X8 and AGW5O-X4, Bio-Rad protein assay dye reagent, and nitrocellulose membranes were from Bio-Rad. Aluminum oxide, grade I was obtained from ICN.

2-(~-(4-Hydroxy-3-['25Iliodophenyl)ethylaminomethyl)tetralone
Plasmids-The polymerase chain reaction (25) was used to construct a set of minigenes containing cDNAfragments encoding the alB2i, alB3i, alB4i domains or the 3i domains of the a,AR, M,AChR, and M&ChR spliced to the 3"untranslated region of the human 8-globin gene (15). Each minigene contained a restriction site (EcoRI) at the 5' end for subcloning, followed by the ribosome binding site consensus sequence GC-CGCCACCATG (26). In each construct a methionine codon was added upstream of the receptor-specific sequence as a translation initiation site and was followed by a glycine codon (GAA). At the end of the receptorderived sequence a TAA stop codon was added, followed by a restriction site (BclI except for M,3i in which anXhoI site was used) and the 8-globin gene 3'-untranslated region. Further restriction sites were engineered at the 3' end of each construct for subcloning. All restriction sites, the signaling sequences, and the splice site between the receptor-derived coding sequence and the 8-globin gene 3"untranslated region were introduced by synthetic oligonucleotides during polymerase chain reaction amplification. Minigene sequences were confirmed by dideoxynucleotide sequencing of double-stranded DNA. For transient cellular expression, the full-length minigene constructs were subcloned into the eukaryotic expression vector pRK5 (27), creating expression plasmids designated pRKalB2i, pRKalB3i, pRKal,4i, pRKa,3i, pRKM,3i, and pRKM23i. The full-length cDNA sequences for the hamster alBAR and human a,,AR, M,AChR, MAChR, and Dl, DR were similarly subcloned into the pRK5 vector, creating the receptor expression plasmids pRKalBAR, pRKa,,AR, pRKM,AChR, pRKM$ChR, and pRKD,DR.
Cell Culture and Dansfection-Cells were grown in either minimum essential medium (HEK-293 cells) or Dulbecco's modified Eagle's medium (COS-7 cells) supplemented with 10% fetal bovine serum and gentamicin (100 pgiml). The cells were transiently transfected with pRK5 containing DNA coding for the entire receptor (1 pg) or the indicated amount of one of the intracellular domains of the receptor using either Ca2P0, precipitation (HEK-293 cells, Ref. 28) or the DEAE-dextran procedure (COS-7 cells, Ref. 29). Empty pRK5 vector was added to each transfection as needed to keep the total mass of DNA added per dish constant within an experiment. Following transfection, cells were split into one six-well plate and one 60-mm dish and assayed after 2 4 4 8 h for second messenger production and ligand binding, respectively.
Western Blots-Expression of minigene products was determined by protein immunoblotting. To detect expression of the orlB% domain peptide, whole cell detergent lysates containing approximately 6 mg of proteidml were incubated with a 1:50 dilution of rabbit polyclonal antiserum 285. The antiserum was raised to a synthetic peptide (SNSKELTLRIHSKNFHEDTLSRC) derived from the 3i domain of the a,@ and conjugated to keyhole limpet hemocyanin through the COOH-terminal cysteine. Immune complexes were collected on protein A-Sepharose, washed, and denatured in Laemmli sample buffer. Immunoprecipitated peptide was resolved by discontinuous SDS-polyacrylamide gel electrophoresis (30) and transferred to nitrocellulose. Filters were incubated in a 1:lOO dilution of antiserum 285 and probed with 1251-protein A. Expression of alB4i, aU3i, M,3i, and M23i domain peptides was determined by protein immunoblotting using appropriate peptide-specific antisera following resolution of whole cell protein lysates (100 pg of proteidane) by SDS-polyacrylamide gel electrophoresis on 12% polyacrylamide gels and transfer to nitrocellulose. Rabbit polyclonal antisera recognizing the alB4i, azA3i, M13i, and M,3i domain peptides were raised to glutathione S-transferase fusion proteins containing peptide sequences corresponding to portions of the appropriate intracellular domains (31) and employed at 1:lOO to 1:lOOO dilution in protein immunoblots. Peptides were visualized by autoradiography and quantitated by phosphorimage densitometry.
Ligand Binding-For saturation binding analysis, crude plasma membranes prepared from transfected HEK-293 or COS-7 cells as described (10) were resuspended in 50 nm "is-HC1 (pH 7.41, 150 m M NaC1, 5 m M EDTA and incubated with radioligand for 1 h at 25 "C.
Inositol Phosphate and cAMPProduction-For measurement of total inositol phosphates (IPS), transfected HEK-293 or COS-7 cells were labeled for 18-24 h with my~-[~H]inositol (1-2 pCi/ml) in culture medium supplemented with 3% fetal bovine serum. After labeling, cells were washed with Dulbecco's phosphate-buffered saline, preincubated for 5-20 min in Dulbecco's phosphate-buffered saline supplemented with 20 m M LiCl and 1 m M CaCl,, and stimulated for 45 min with agonist as indicated. Following stimulation, total inositol phosphates were extracted as described (10) and separated on Dowex AG1-X8 columns (32). Total IPS were eluted with 1 M ammonium formate, 0.1 M formic acid and quantitated by liquid scintillation spectroscopy.
For measurement of intracellular CAMP, transfected HEK-293 cells were labeled for 18-24 h with 1-2 pCi/ml L3H]adenine in minimum essential medium supplemented with 3% fetal bovine serum. After labeling, cells were preincubated for 15 min at 37 "C in Hanks' balanced salt solution supplemented with 10 m M HEPES pH 7.4 and 1 m M isobutylmethylxanthine and stimulated for 20 min with or without agonist. CAMP was extracted as described (33) and separated by sequential chromatography on Dowex AGW5O-X4 and aluminum oxide columns using trace amounts of [14ClcAMP (10,000 dpdcolumn) to assess column recovery (34).
Chloramphenicol Acetyltransferase Assay-For determination of the effect of alBAR or alB3i domain peptide expression on phorbol response element activation, HEK-293 cells in 60-mm dishes (1 x lo6 cellddish) were transiently cotransfected with the indicated receptor or minigene construct plus pUC13 (-356 fos-CAT), a fusion plasmid consisting of the p106 CAT gene linked to the mouse c-fos gene promoter region beginning -356 base pairs from the translation start site (24). Following transfection (48 h), monolayers were incubated for 2 h at 37 "C in Dulbecco's phosphate-buffered saline supplemented with 20 m M LiCl and 1 m M CaC1, with or without agonist, and CAT activity was determined. For assay of CAT activity, monolayers were harvested in Dulbecco's phosphate-buffered saline containing 0.23 unitsiml aprotinin, collected by centrifugation, resuspended in 0.25 M Tris-HC1 (pH 7.8), 0.23 unitsiml aprotinin, 1 mgiml soybean trypsin inhibitor, lysed by sonication, and clarified by centrifugation. Supernatants, matched for protein content, were heated to 70 "C for 10 min to inactivate deacetylases, and 50-pl aliquots were added to 200 pl of assay mix containing 125 m M Tris-HC1 (pH 7.8), 1.25 m M chloramphenicol, 0.35 PM [3H]acetyl-coenzyme A in a 5-ml plastic scintillation vial. Reaction mixtures were immediately overlaid with 4 ml of Econofluor, and the generation of L3H]acetyl chloramphenicol was measured by liquid scintillation counting at intervals between 2 and 6 h. CAT positive controls containing 0.1-1.0 units of purified CAT were included in each assay for standardization. sion of the alB2i, alB3i, alB4i, a2,3i, M13i, or M,3i domain peptides. Fig. 1A depicts the putative G protein-coupled receptor membrane topography and the location of the intracellular domain sequences chosen for expression. As shown in Fig. lB, receptor-derived cDNA sequences encoding the 20-amino acid 2i domain (amino acid residues 143-162), the 71-amino acid 3i domain (amino acid residues 225-295), and the 164-amino acid 4i domain (amino acid residues 352-515) of the a l f l were introduced into a minigene construct designed for peptide expression (15). Similarly, the cDNA sequences encoding the 3i domains of azAAR (amino acid residues 217-373), M,AChR (amino acid residues 210-365), and MAChR (amino acid residues 234-4131 were inserted into the minigene construct. Cellular expression of the peptide minigene products, depicted in  Table I.

Construction and Expression
Coexpression of the intracellular domain peptides had no detectable effect on either the level of expression of the cotransfected homologous receptor or the basal rate of IP production. In contrast, a,,AR-mediated IP production in response to a maximally efficacious dose of agonist was significantly attenuated in cells expressing the alB3i or -4i domain, but not the a,,2i domain, suggesting that these regions contain epitopes that contribute to receptor/(= protein coupling. To determine the specificity of the observed inhibition, each a,,AR-derived peptide was tested for the ability to inhibit signal transduction mediated by the homologous receptor, the G,-coupled M,AChR, and the G,-coupled D,DR. As shown in Fig. 3, coexpression of the alB2i domain peptide produced no significant effect on second messenger production mediated by any of these receptors. The alB3i domain peptide produced a reduction in both alBAR-(50%) and M,AChR-(27%) mediated I P production, with no significant effect on D,,DR-mediated CAMP production. The effects of the alB4i domain peptide were less specific, causing a significant inhibition of a,,AR-, M,AChR-, and D,,DR-mediated signal transduction. Thus, the inhibition provoked by the alB3i domain peptide was limited to the G,-coupled receptors examined, while the inhibition by the alB4i domain peptide was generalized to include a GBcoupled receptor.
The difference in specificity between alB3i domain-and a1,4i domain-induced inhibition of receptor-mediated signal transduction could reflect distinct binding sites for the 3i and 4i domain peptides. To address this possibility, the effects of maximally active levels of alB2i, a1,3i, and alB4ir singly or in combination, on a,,AR-mediated IP production were determined. As shown in Fig. 4 A , the alB3i and the alB4i domain peptides inhibited aIBAR activity by 4 0 4 0 % when individually coexpressed with receptor. The a1,2i domain was inactive even when coexpressed with either or both of the active peptides. When the alB3i and alB4i domains were expressed simultaneously, the inhibitory effect was greater than that produced by either domain peptide alone (70%). As shown in Fig.  4B, the effects of aIB3i and alB4i domain peptides were approximately additive in cells co-transfected with submaximally active amounts of the minigene constructs. Inhibition of receptor-mediated signal transduction by maximally active amounts of pRKa,,Bi and pRKa1,4i, while less than additive, was greater than that observed with either plasmid alone. These data suggest the al,3i and alB4i domain peptides do not compete for the same intracellular binding site and are consistent with the data in Fig. 3 that demonstrate a differential pattern of receptor specificity for inhibition by each of the two domain peptides.
Generality and Specificity of 3i Domain-induced Inhibition of G Protein-coupled Receptors-Of the intracellular domains tested, the alB3i domain produced the most specific inhibition of receptor-mediated signal transduction. To determine the generality and specificity of 3i domain peptide-mediated inhia 1 133i A. HEK-293 cells were transiently transfected with the indicated amounts of pRKa2,3i DNA. Empty pRK5 vector was added to each plate to keep the total mass of DNA constant. Peptide expression 48 h after transfection was determined as described.

69 -
bition of signal transduction, we prepared a series of four 3i domain minigene constructs (two derived from G,-coupled receptors (alBAR and M,AChR) and two derived from G,-coupled receptors (a,,AR and MAChR), as shown in Fig. 1 B ) . The effect of each of the 3i domain peptides on each of the four receptors was then assessed.
In transiently transfected HEK-293 cells, G,-coupled receptors mediated Bordetellapertussis toxin (PT)-insensitive stimulation of phosphatidylinositol hydrolysis. In contrast, activation of the G,-coupled receptors did not produce a detectable increase in IP production, although a PT-sensitive inhibition of adenylyl cyclase was observed (data not shown). By comparison, transiently transfected COS-7 cells exhibited both PT-insensitve stimulation of IP production mediated through the G,-coupled receptors (Fig. 5A) and PT-sensitive stimulation of IP production mediated through the G,-coupled receptors (Fig.  5B). The magnitude of IP accumulation produced by stimulation of these receptors was similar to what has been previously reported in stably transfected cell lines (36,37). Subsequent experiments utilized G,-mediated (PT-insensitive) IP production in HEK-293 and G,-mediated (PT-sensitive) IP production in COS-7 cells in order to provide a consistent assay system for study of the effect of 3i domain peptides on receptor-mediated signal transduction.
The generality of the 3i domain peptide-induced inhibition was assessed by determination of the effect of each 3i domain peptide on signal transduction mediated by the receptor from which the peptide was derived. Table I1 summarizes the effects of the alB3i, a2*3i, M,3i, and M23i domain peptides on PI hydrolysis mediated by each homologous receptor. Maximal agonist-stimulated IP production was significantly inhibited (35-50%) in each case. These data indicate that attenuation of receptor-mediated signal transduction by coexpression of homologous 3i domain peptides represents a general phenomenon, extending to at least two G,-coupled receptors (alBAR and MIAChR), two G,-coupled receptors (a,,AR and MAChR), and, as previously described, a G,-coupled receptor (D,,DR) (15).
The specificity of the inhibition was determined by assaying the effects of each 3i domain on IP production mediated via each of the four receptors in transiently cotransfected HEK-293 or COS-7 cells. Fig. 6 depicts the effects of maximally active levels of each 3i domain on the agonist dose-response relationship for IP production mediated by one receptor, the alBAR. As shown in Fig. 6, A and C, coexpression of alB3i and M13i peptides produced a decrease in the apparent efficacy (ie. maximal stimulatory effect) of epinephrine. In contrast, as shown in Fig.

TABLE I Parameters of ligand binding and IP production in HEK-293 cells co-transfected with aI& and a,&? intracellular domain constructs
binding of [12511HEAT (B,,) and basal and epinephrine-stimulated IP production were determined as described. Rbnsa, represents the amount of HEK-293 cells were transiently co-transfected with pRKaIBAR (1 pg/dish) plus the indicated alBAR intracellular domain construct. Saturation basal IP production in cells in arbitrary units such that 1 unit is the basal amount of IPS produced in control cells co-transfected with alBAR and (10 PM) in arbitrary units such that 1 unit is the basal amount of IP produced in control cells. Inhibition of R,,, represents the percent inhibition pRK5 vector. R,, represents the amount of IP production in cells following stimulation with a maximally efficacious concentration of epinephrine of the epinephrine-stimulated increase in IPS over basal, compared with the control cells. Statistical significance was assessed using the Student's paired t test.  All data are presented as percent of the epinephrine-stimulated increase in IPS over basal levels normalized to control (pRK5 vector) co-transfected cells. Basal IP production did not vary significantly between experimental conditions. 6, B and D, a,3i and M,3i peptides produced no significant effect on agonist eficacy. In Fig. 7, the effect of each of the 3i domain peptides on IP production mediated by each of the four receptors is depicted. The a,,AR-mediated I P production was significantly inhibited by expression of the alB3i or M13i domains with little effect produced by expression of the aZA3i or M,3i domains (Fig. 7A). M,AChR-mediated IP production was sensitive to expression of the M,3i domain peptide and, to a lesser extent, to the aIB3i domain peptide (Fig. 7C). IP production mediated by the G,-coupled a,,AR and M&ChR was most sensitive to coexpression of the homologous 3i domain peptide with little effect produced by expression of nonhomologous 3i domain peptides (Fig. 7, B and Dl. These data are summarized in Table 111. In general, the greatest inhibition of receptor-mediated signal transduction was caused by the homologous 3i domain peptide. In addition, the alB3i and M13i domain peptides inhibited IP production provoked by both G,-coupled receptors (alBAFt and M,AChR). When 3i domain peptides derived from receptors coupled to one G protein were coexpressed with receptors coupled to a different G protein, little or no inhibitory effect was observed. Thus, it appears that there is considerable specificity in the inhibition of receptor-mediated I P production by 3i domain peptides. The inhibition provoked by alB3i and M,3i domain peptides extends to other receptors coupled to a common G protein, while aW3i and M,3i domain peptides appear to inhibit the homologous  0 p, cells expressing a,AR). IP production was determined and is expressed as the mean & S.E. values from triplicate determinations from three to five separate experiments. Data are presented as the agonist-stimulated increase in IPS over basal levels in arbitrary units such that 1 unit equals the basal amount of IPS measured in nonstimulated cells. Basal levels of IP production in cells transfected with plasmids coding for receptors were not significantly different compared with each other or with that in cells transfected with control vector.

TABLE I1 Parameters of ligand binding and IP production in HEK-293 and COS-7 cells coexpressing G protein-coupled receptors and their homologous 3i domain peptides
HEK-293 cells were transiently co-transfected with pRKaIBAR or pRKM,AChR (1 pg/dish) plus empty pRK5 vector or the homologous receptor 3i domain minigene construct (10 &dish); COS-7 cells were transiently co-transfected with pR&, AR or pRKMfiChR plus empty pRK5 vector or the homologous receptor 3i domain minigene construct. Saturation binding (B,,,=) of [lz5I1HEAT for alBAR, [3H]rauwolscine for a,*AR, or ~-[benzilic-4,4'-~Hlquinuclidinyl benzilate for MIAChR and M+ChR was determined as described. R,,, represents the amount of basal IP production in cells in arbitrary units such that 1 unit is the basal amount of IP produced in control cells co-transfected with receptor plus pRK5 vector. R , , represents the amount of IP production in cells following stimulation with a maximally efficacious dose of agonist in arbitrary units such that 1 unit is the basal amount of IP produced in control cells. Inhibition of R , represents the percent inhibition of the agonist-stimulated increase in IPS over basal, compared with the control cells. Statistical significance was assessed using the Student's paired t test. ' As compared with pRK5 vector, p < 0.05. peptides. If the 3i domain peptide directly activated the G protein, enhanced basal second messenger production might promote desensitization of the signal transduction pathway, resulting in a diminished response to receptor activation. Indeed, we have observed that stable expression of a constitutively active mutant of the &AR in Chinese hamster ovary cells leads to an apparently reversible desensitization and down-regulation of receptors without a detectable elevation of basal CAMP levels (38). Alternatively, a 3i domain peptide might act as an antagonist or weak partial agonist at the receptor/G protein interface and competitively inhibit receptor-mediated signal transduction with no apparent effect on basal second messenger generation.
To distinguish these possibilities, the effect of alB3i domain peptide expression on basal IP production in the absence of co- cells were cotransfected with pRK5 vector, pRKa,,AR, the constitutively active mutant plasmid pRKa,,AR(S288-294), or pRKa,,3i, and a CAT reporter gene construct linked to a phor-bo1 response element, pUC13(-356 foslCAT). As shown in Fig.  8, expression of the constitutively active mutant alBAR(S288-294) resulted in a substantial enhancement of CAT protein expression compared to control. Overexpression of wild type alBAR also caused a modest increase, which was enhanced when cells were exposed for 2 h to epinephrine. Expression of the (u,,3i domain peptide resulted in an increase in CAT expression comparable with that observed with the wild type alBAR in the absence of agonist. These data suggest that the a1,3i domain peptide interacts directly with the G protein in a manner analogous to the wild type q B A R resulting in an in- crease in basal PI turnover similar to, but smaller than, that observed with the constitutively active mutant receptor. Since the magnitude of the alB3i domain-induced effect is no greater than that produced by the cylB-, it is unlikely to promote sufficient desensitization to account for the observed attenuation of agonist-stimulated I P production in co-transfected cells. Furthermore, IP production in response to NaF, which directly activates G proteins, was not affected by coexpression of the alB3i domain peptide, demonstrating that the downstream components of the signaling pathway are intact (data not shown). These data suggest that the 3i domain peptides inhibit receptor-mediated signal transduction by binding t o the G protein in a manner analogous to the receptor, thereby potentially disrupting receptor/G protein interaction.
If the a1,3i domain peptide-mediated inhibition results from competition between peptide and activated hormone-receptor complex for a common binding site on the Ga subunit, the inhibition should be surmountable under conditions where hormone-receptor complex is present in relative excess. As shown in Fig. 9A, the inhibitory effects of the alB3i domain peptide  were determined in HEK-293 cells coexpressing varying levels of aIBAR. In cells expressing less than 1 pmoUmg membrane protein of alBAR, a 60% inhibition of maximal a,,AR-mediated IP production was observed. Increasing receptor density diminished the degree of inhibition, abolishing it at receptor levels greater than 5 pmol/mg of membrane protein. As shown in Fig.  9B, coexpression of increasing amounts of alB3i domain peptide resulted in a small (2-%fold) increase in the EC,, for epinephrine-stimulated IP production in cells expressing a constant high level of alBAR. The failure of increasing agonist concentration to fully surmount the inhibition at higher levels of alB3i domain peptide expression probably reflects saturation of all available alBAR at agonist concentrations in excess of 1 1.1~. Beyond that point, a further increase in agonist concentration does not result in an increase in hormone-receptor complex (the entity that would be expected to compete with the alB3i domain peptide for binding to Ga subunits). Collectively, these data suggest that the aIB3i domain peptide acts as a weak agonist at the level of the G protein, producing a small increase in the rate of PI hydrolysis and reversibly antagonizing receptor-mediated G protein activation. DISCUSSION In this study, the inhibition of G protein-coupled receptormediated second messenger generation resulting from coexpression of peptides derived from the intracellular domains of the receptor in intact cells is described. We have previously demonstrated, using a similar approach ( E ) , t h a t coexpression of the 3i domains of alBAR and DIADR specifically attenuated homologous receptor-mediated signaling. This report characterizes the generality and specificity of this phenomenon by examining the effects of the 2i, 3i, and 4i domain peptides of alBAR and the 3i domain peptides of a,AR, MIAChR, and MAChR on homologous and heterologous receptor signaling. In each case, the 3i domain peptide significantly inhibited signaling mediated by the homologous receptor, with inhibition of heterologous receptor activity (if any) restricted to receptors utilizing a common G protein. This demonstrates a high degree of specificity in 3i domain peptide-induced inhibition of receptor-mediated signal transduction and suggests that the receptodG protein interface may represent a therapeutic target for future drug development.
Previous studies of the effects of receptor-derived or related peptides on G protein function in reconstituted systems have demonstrated that specific peptides may behave as receptorlike agonists, antagonists of receptor-mediated G protein activation, or partial agonists at the level of the G protein. For example, peptides derived from the COOH-terminal region of the P&R 3i domain (17, 18), the M,AChR 3i domain, and the aUAR 3i domain (19) mimic receptor by directly activating G proteins in vitro. Peptides derived from the COOH-terminal region of the a&R also apparently inhibit receptor/G protein interaction in membrane preparations (20), as do peptides derived from the 2i, 3i, and 4i domains of rhodopsin (21). In addition, an oligopeptide substance P analog has been shown to competitively antagonize receptor-mediated activation of Gi, Go, and G, in phospholipid vesicles by binding to Ga subunits (39). The results of the present study, which unlike others has used intact cells, suggest that the a1,3i domain peptide directly interacts with the G protein (GJ since alB3i expression modestly enhances phorbol response element-promoted expression of CAT activity. The finding that the alB3i domain peptideprovoked inhibition of receptor-mediated IP production is surmounted at high levels of receptor expression supports the hypothesis that inhibition results from competition between the 3i domain peptide and the hormone-receptor complex for G protein binding. Therefore, it appears that 3i domain peptides act as weak partial agonists at the level of the G protein. While the data indicate that the alB3i domain peptide does interact directly with the G protein, we cannot exclude the alternative hypothesis that interaction between the 3i domain peptide and the receptor itself disrupts receptor conformation so as to impair receptor/(: protein coupling and to contribute to the observed inhibition of IP production. If such a mechanism were valid, the expected pattern of inhibition by 3i domain peptides would be relatively receptor-specific since the affinity of a 3i domain peptide for the other intracellular domains of the homologous receptor would probably be greatest. In contrast, inhibition resulting from 3i domain peptide binding to the G protein might be more broadly G protein-specific. Occupancy of the receptor/G protein interface by a compatible 3i domain peptide would be expected to decrease signal transduction mediated by any receptor coupled with that G protein. The observed pattern of inhibition (Table 111)  transiently co-transfected with pRKa,,AR (2.0 pg/dish) plus pRK5 vector or pRKalB3i as indicated, and the agonist dose dependence for epinephrine-stimulated IP production was determined as described. The mean level of alBAR expression was 3.55 f 0.09 pmoYmg for all curves. Data are presented in arbitrary units such that 1 unit equals the basal amount of IPS in control (pRK5 vector) co-transfected cells and depict mean values of triplicate determinations in one of three separate experiments that gave similar results. Basal IP production was not significantly different between experimental conditions. EC,, values for epinephrine, determined by least squares regression analysis were as follows: control, 10.4 m; 1 pg of pRKalB3i, 11.6 m; 3 pg of pRKaIB3i, 13.1 m; 10 pg of pRKalB3i, 28.5 nM.
ceptor, which couples to the same G protein, the most striking inhibition in most cases occurred when a 3i domain peptide was coexpressed with its homologous receptor. While this may reflect a direct interaction between 3i domain peptide and its homologous receptor, it might also reflect intrinsic differences in the affinity of receptor/G protein interaction between different classes of G protein-coupled receptors. Alternatively, other intracellular domains of some receptors may contribute differentially to receptor/(=. protein interaction allowing some receptors to more easily displace a nonhomologous 3i domain peptide from the Ga subunit binding site.
In conclusion, this study demonstrates that inhibition of receptor-mediated signal transduction by 3i domain peptides occurs in both Giand G,-dependent pathways and in both adrenergic and muscarinic cholinergic systems. The specificity with which the 3i domain peptides inhibit receptor-mediated signaling suggests the feasibility of developing drugs that might exert inhibition at the level of receptor/G protein interaction rather than at the level of ligand-receptor binding.