Heterologous Desensitization Mediated by G Protein-specific Binding to Caveolin*

We examined the notion that sequestration of G protein subunits by binding to caveolin impedes G protein reassociation and leads to transient, G protein-specific desensitization of response in dispersed smooth muscle cells. Cholecystokinin octapeptide (CCK-8) and substance P (SP) were used to activate Gq/11, cyclopentyl adenosine (CPA) was used to activate Gi3, and acetylcholine (ACh) was used to activate both Gq/11 and Gi3 via m3 and m2 receptors, respectively. CCK-8 and SP increased only Gαq/11, and CPA increased only Gαi3 in caveolin immunoprecipitates; caveolin and other G proteins were not increased. ACh increased both Gαq/11 and Gαi3 in a time- and concentration-dependent fashion: only Gαq/11was increased in the presence of an m2 antagonist, and only Gαi3 was increased in the presence of an m3 antagonist. To determine whether transient G protein binding to caveolin affected subsequent responses mediated by the same G protein, PLC-β activity was measured in cells stimulated sequentially with two different agonists that activate either the same or a different G protein. After treatment of the cells with ACh and an m2 antagonist, the phospholipase C-β (PLC-β) response to CCK-8 and SP, but not CPA, was decreased; conversely, after treatment of the cells with ACh and an m3 antagonist, the PLC-β response to CPA, but not CCK-8 or SP, was decreased. Similarly, after treatment with CCK-8 or SP, the PLC-β response mediated by Gq/11 only was decreased, whereas after treatment with CPA, the PLC-β response mediated by Gi3only was decreased. A caveolin-binding Gαq/11 fragment blocked the binding of activated Gαq/11 but not Gαi3 to caveolin-3 and prevented desensitization of the PLC-β response mediated only by other Gq/11-coupled receptors. A caveolin-binding Gαi3 fragment had the reverse effect. Thus, transient binding of receptor-activated G protein subunits to caveolin impedes reassociation of the heterotrimeric species and leads to desensitization of response mediated by other receptors coupled to the same G protein.

Compelling evidence exists that a variety of signaling events are initiated in caveolae (16,22,42), consistent with the notion that caveolin provides a scaffold for the assembly of signaling molecules into modules primed for activation. Caveolin could also act to restrain cellular response by selective binding of signaling molecules, such as G proteins. Caveolin-binding motifs consisting of 10 -15-mer sequences with characteristically spaced aromatic residues (XXXXX or XXXXXX, where is the aromatic amino acid Trp, Phe, or Tyr), are present in all caveolin-binding proteins, including G proteins (13). Binding of G proteins to caveolin could lead to their sequestration and enrichment in inactive form in caveolar microdomains (6). Pharmacological activation with GTP␥S 1 strongly inhibits and mutational activation abolishes G protein binding to caveolin consistent with preferential interaction of caveolin with GDPbound G␣ subunits (15).
The possibility that receptor-activated G protein subunits are sequestered by binding to caveolin or a caveolin-associated protein, leading to transient, G protein-specific desensitization of response, has been raised but not experimentally tested (29).
This notion was examined in the present study using a series of agonists previously shown to activate G q/11 (cholecystokinin octapeptide, and substance P), G i3 (cyclopentyl adenosine), or both G q/11 and G i3 (acetylcholine) in smooth muscle (33)(34)(35)(36)(37). Caveolin-binding fragments of G␣ q/11 and G␣ i3 were used to inhibit competitively G␣ binding to caveolin and suppress desensitization of response (13,15). The results show that receptor activation was followed by transient binding of activated G␣ to caveolin that was selectively blocked by caveolin-binding fragments. Binding of G␣ to caveolin resulted in transient desensitization of cellular response mediated by other receptors coupled to the same G protein.

Preparation of Dispersed Smooth
Muscle Cells-Muscle cells were isolated from the circular muscle layer of rabbit intestine by enzymatic digestion at 31°C with collagenase, followed by filtration though 500-m Nitex mesh and low speed centrifugation as described previously (33,35,37). The cells were suspended in HEPES medium containing 120 mM NaCl, 2.5 mM KH 2 PO 4 , 4.0 mM KCl, 0.6 mM MgCl 2 , 25 mM HEPES, and 2.1% essential amino acid mixture. In some experiments, muscle cells were permeabilized by 5-min treatment with saponin (35 g/ml) and resuspended in low Ca 2ϩ (100 nM) medium (37).
Identification of Receptor-activated G Proteins-G proteins selectively activated by various receptors were identified by the method of Okamoto et al. and others (37,38). Muscle cell homogenates were centrifuged at 27,000 ϫ g for 15 min, and the crude membranes were suspended in 20 mM HEPES medium (pH 7.4) containing 2 mM EDTA and 240 mM NaCl. The membranes were diluted 20-fold and incubated at 37°C with 60 nM [ 35 S]GTP␥S in a medium containing 10 mM HEPES (pH 7.4), 100 M EDTA, and 10 mM MgCl 2 . The reaction was stopped with 10 volumes of 100 mM Tris-HCl medium (pH 8.0) containing 10 mM MgCl 2 , 100 mM NaCl, and 20 M GTP, and the mixture was incubated for 2 h on ice in wells coated with specific G protein antibodies. The wells were washed with phosphate buffer solution containing 0.05% Tween 20, and the radioactivity from each well was counted.
Assay for PLC-␤ Activity-PLC-␤ activity was measured from the formation of total inositol phosphates in muscle cells prelabeled with myo-[ 3 H]inositol as described previously (39). 10 ml of cell suspension (2 ϫ 10 6 cell/ml) were labeled with myo-[ 3 H]inositol (15 Ci/ml) for 3 h at 31°C. The cells were centrifuged at 350 ϫ g for 10 min and resuspended in 10 ml of fresh HEPES medium. The cells were treated with one agonist for 10 min and centrifuged again at 350 ϫ g for 5 min. Various agonists were then added to 0.5 ml of cell suspension for 30 s, and the reaction was terminated with 940 l of chloroform:methanol: HCl (50:100:1 v/v/v). After chloroform (310 l) and water (310 l) were added, the samples were vortexed, and the phases were separated by centrifugation at 1000 ϫ g for 15 min. The upper aqueous phase was applied to Dowex AG-1 ϫ 8 columns. After washing, inositol phosphates were eluted with 5 ml of 0.8 M ammonium formate with 0.1 M formic acid, and the eluates were collected into scintillation vials and counted in gel phase after addition of 10 ml of scintillant. The results were expressed as cpm/10 6 cells.
Caveolin Immunoprecipitation and Immunoblotting of G␣ Proteins-Smooth muscle cells (2-3 ϫ 10 6 cells/ml) were lysed by incubation for 30 min at 4°C in 10 mM Tris (pH 7.5), 50 mM NaCl, 1% Triton X-100, and 60 mM octyl glucoside, and the lysate was centrifuged at 15,000 ϫ g for 30 min. The supernatant was precleared by incubation with 0.1% albumin-coated protein A-Sepharose for 6 h at 4°C and then incubated overnight with polyclonal caveolin-3 antibody at a final concentration of 4 g/ml. Protein A-Sepharose was then added for 1 h, and the mixture was centrifuged for 5 min. The immunoprecipitates were washed four times with lysis buffer and boiled in Laemmli buffer. Samples were separated by SDS-PAGE in 12% acrylamide gel, electrotransferred to nitrocellulose paper, and probed with antibodies to G␣ i3 , G␣ q/11 , or caveolin-3. After incubation with secondary antibody conjugated with horseradish peroxidase, the proteins were visualized using the Super Signal ULTRA chemiluminescent substrate. The intensity of the protein band on Hyperfilm-ECL was determined by densitometry.
Detergent-free Purification of Caveolin-enriched Membrane Fractions-Caveolin-enriched membrane fractions derived from intestinal smooth muscle were prepared by the method of Song et al. (9). Dispersed muscle cells were washed three times in phosphate-buffered saline and suspended in 2 ml of 500 mM sodium carbonate (pH 11.0) containing 0.2 mM of phenylmethylsulfonyl fluoride and 20 g/ml of leupeptin and homogenized with a Polytron tissue grinder (three 10-s bursts) and by sonication (three 20-s bursts). The homogenate was adjusted to 45% sucrose in MBS (25 mM Mes, pH 6.5, and 0.15 M NaCl), placed in an ultracentrifuge tube and overlaid with two 4-ml layers of 35 and 5% sucrose in MBS containing 250 mM sodium carbonate. The gradient was centrifuged at 39,000 rpm for 20 h. Twelve 1-ml fractions were collected sequentially from the top and designated as fractions 1-12. Fractions were analyzed by SDS-PAGE (15% acrylamide gels); after transfer to nitrocellulose membranes, Western blot analysis was performed with antibodies to caveolin-3 and various G␣ subunits. Immunoreactive bands were visualized by 1-h incubation with horseradish peroxidase-conjugated secondary antibodies followed by enhanced chemiluminescence assay.
For protein immunoprecipitation from caveolin-enriched fractions, the purified membranes were diluted to 2 mg protein/ml in lysis buffer, incubated on ice for 1 h, and centrifuged. The lysate was precleared by 1-h incubation with protein A-Sepharose and then incubated overnight with caveolin-3 antibody and for 2 h with protein A-Sepharose. Immunoprecipitates were washed five times with lysis buffer and resuspended in 30 l of 2-fold concentrated Laemmli buffer; after separation on SDS-PAGE and transfer to nitrocellulose membranes, immunoblot analysis with caveolin-3 and G␣ antibodies was performed.
Phosphatidylinositol 4,5-Bisphosphate Assay in Caveolar Membranes-PIP 2 was measured by thin layer chromatography in caveolar membranes as described previously (40,41). A 20-ml cell suspension (10 6 cells/ml) was incubated with 500 Ci of [ 32 P]P i at 31°C for 3 h. Duplicate samples (10 6 cells/ml) were incubated at 31°C with ACh (0.1 M), cholecystokinin-8 (CCK-8 1 nM), SP (1 M), and CPA (1 M) separately for 30 s. The reaction was terminated by centrifugation at 15,000 ϫ g for 5 min followed by addition of 1 ml of HEPES buffer (25 mM) containing 0.5% Triton X-100. The mixture was incubated for 10 min and then centrifuged at 15,000 ϫ g for 5 min. Supernatant and pellets were extracted with 1.8 ml of chloroform-methanol-HCl (100: 200:2 v/v/v). The organic phase was analyzed for PIP 2 by thin layer chromatography. The results were expressed as cpm/10 6 cells.

H]Scopolamine Binding to Smooth
Muscle Cells-Binding of [ 3 H]scopolamine to dispersed intestinal smooth muscle cells was done as described previously (37). Muscle cells were suspended in HEPES medium containing 1% bovine serum albumin. Triplicate 0.5-ml aliquots (10 6 cells/ml) were incubated for 15 min with 1 nM [ 3 H]scopolamine alone or with acetylcholine. Bound and free radioligand were separated by rapid filtration under reduced pressure through 5-m polycarbonate Nucleopore filters and washed four times with 3 ml of ice-cold HEPES medium containing 0.2% bovine serum albumin. Nonspecific binding (28 Ϯ 6%) was calculated as the amount of radioactivity in the presence of 10 M acetylcholine. Materials-Two peptides corresponding to the caveolin-binding domains of G␣ q/11 , Tyr 192 -Ala 206 (YPFDLQSVIFRMVDA) and G␣ i3 , Thr 187 -Val 201 (THFTFKELYFKMFDV) were synthesized by the solid phase method and purified (95-99%) by high performance liquid chromatography (Peptidogenic, CA). [2-3 H]Inositol, [ 3 H]scopolamine, [ 35 S]GTP␥S, and [ 32 P]orthophosphate were obtained from NEN Life Science Products; Dowex AG-1 ϫ 8 resin was from Bio-Rad; polyclonal G␣ q/11 , G␣ i1/2 , and G␣ i3 antibodies were from Santa Cruz Biotechnology; and caveolin-3 antibody was from Transduction Laboratories (Lexington, KY). All other reagents were from Sigma.

RESULTS
Agonist-induced Activation of G Proteins-G protein activation was determined directly from the increase in agoniststimulated binding of [ 35 S]GTP␥S to specific G␣ subunits in solubilized intestinal smooth muscle cell membranes. CCK-8 and SP selectively activated G q/11 , whereas CPA selectively activated G i3 (Table I). ACh activated both G q/11 and G i3 via m3 and m2 receptors, respectively, as shown previously (37).

Distribution of Caveolin-3 and G Proteins in Caveolar Membranes in the Basal State and after Stimulation with Agonists-
Immunoblot analysis of 12 fractions derived from intestinal smooth muscle cells showed that caveolin-3 was confined to low density fractions 5 and 6, whereas G proteins (G␣ q , G␣ i1/2 , and G␣ i3 ) were present in fractions 5 and 6, as well as in membrane fractions 9 -12 ( Fig. 1). The pattern was similar to that obtained in other cell types (15,29).

Distribution of G Proteins in Caveolin Immunoprecipitates after Stimulation with
Agonists-A similar pattern was obtained in caveolin immunoprecipitates derived from cell lysates. Treatment of muscle cells with CCK-8 or SP increased G␣ q/11 in caveolin-3 immunoprecipitates by 187 Ϯ 22% (p Ͻ 0.01) and 181 Ϯ 24% (p Ͻ 0.01), respectively, but had no effect on G␣ i3 (Fig. 3). In contrast, treatment of muscle cells with CPA (1 M) increased G␣ i3 in caveolin-3 immunoprecipitates by 173 Ϯ 18% (p Ͻ 0.01) but had no effect on G␣ q/11 (Fig. 3). G␣ i2 and G␣ i1 are not activated by CCK-8, SP, or CPA (Table I) and did not increase in caveolin-3 immunoprecipitates upon treatment of muscle cells with all three agonists (Fig. 4). There was no increase in caveolin-3 in caveolin immunoprecipitates or in G␣ q/11 and G␣ i3 immunoprecipitates after treatment with any agonist (Fig. 4).
The pattern of selective increase in G␣ subunits was reinforced by analysis of caveolin-3 immunoprecipitates after stimulation with ACh. Treatment of muscle cells with ACh (0.1 M) increased both G␣ q/11 and G␣ i3 in caveolin-3 immunoprecipitates by 177 Ϯ 29% (p Ͻ 0.01) and 181 Ϯ 37% (p Ͻ 0.01), respectively (Fig. 3). The increase in G␣ q/11 was abolished by the m3 receptor antagonist, 4-DAMP, but was not affected by the m2 receptor antagonist, methoctramine, whereas the increase in G␣ i3 was abolished by methoctramine but was not affected by 4-DAMP (Fig. 3). In these experiments also there was no increase in G␣ i1 or G␣ i2 in caveolin immunoprecipitates and no increase in caveolin-3 either in caveolin immunoprecipitates or G␣ q/11 and G␣ i3 immunoprecipitates (Fig. 4). The m2 but not m3 receptors were detected in caveolin immunoprecipitates after treatment of muscle cells with ACh (Fig. 4C).
Time Course and Concentration Dependence of Agonist-stimulated Increase of G␣ in Caveolin-3 Immunoprecipitates-Treatment of smooth muscle cells with a maximal concentration of ACh (0.1 M) caused a time-dependent increase of G␣ q/11 and G␣ i3 in caveolin-3 immunoprecipitates that attained a peak in 5 min. The peak was sustained for 15 min and declined rapidly to control levels in the next 20 min (Fig. 5). The peak increase was concentration-dependent (Fig. 6).
Inhibition of G Protein Binding to Caveolin-3 by Caveolinbinding G Protein Fragments-G protein fragments that selectively bind to the caveolin-scaffolding domain were used to block the binding of activated G proteins to caveolin-3. Addition of the caveolin-binding G␣ q/11 fragment, YPFDLQSVIFRM-VDA (50 M), to permeabilized muscle cells for 10 min blocked the increase in G␣ q/11 binding to caveolin elicited by SP but not the increase in G␣ i3 binding elicited by CPA (Fig. 7). Conversely, addition of the G␣ i3 fragment, THFTFKELYFKMFDV (50 M), blocked the increase in G␣ i3 binding to caveolin-3 elicited by CPA but not the increase in G␣ q/11 binding elicited by SP (Fig. 7).
Heterologous Desensitization of PLC-␤ Activity Mediated by G Protein-specific Binding to Caveolin-The possibility that FIG. 1. Location of caveolin-3 and G␣ subunits in smooth muscle membrane fractions. Twelve sucrose-density gradient fractions were prepared as described under "Experimental Procedures" and subjected to SDS-PAGE and Western blot analysis with specific antibodies to caveolin-3 and G␣ subunits. Caveolin-3 was confined to low density fractions 5 and 6. G␣ subunits were present in fractions 5 and 6, as well as fractions 9 -12. binding of activated G proteins to caveolin-3 could result in desensitization of response was tested by sequential stimulation of muscle cells with different agonists that couple to the same or a different G protein. PLC-␤ activity in response to CCK-8 or CPA was measured in muscle cells pretreated for periods ranging from 5 to 60 min with 0.1 M ACh or for 10 min with different concentrations of ACh (10 pM to 0.1 M). PLC-␤ activity in response to both CCK-8 and CPA decreased in parallel with the time of pretreatment with ACh, attaining a maximum in cells pretreated for 5-10 min and reverting to control levels in cells pretreated for 40 min (Fig. 8). The decrease in PLC-␤ activity paralleled the increase in G␣ binding to caveolin (Figs. 5 and 8). Pretreatment with different concentrations of ACh for 10 min decreased the PLC-␤ response to CCK-8 and CPA in a concentration-dependent fashion (Fig. 9). In both concentration response and time course studies, there was a close linear correlation (r ϭ 0.99) between the decrease in PLC-␤ activity and the increase in caveolin-bound G␣ q/11 or G␣ i3 (Fig. 10).
Desensitization of PLC-␤ Activity after Treatment with Noncholinergic Agonists-Pretreatment with other agonists besides ACh provided further support for the notion that inhibition of PLC-␤ activity was G protein-specific. Pretreatment  Whole cell lysates were subjected to immunoprecipitation with caveolin-3 antibody (A) and G␣ i3 or G␣ q/11 antibody (B). Samples from caveolin immunoprecipitates were immunoblotted with G␣ i1/2 or caveolin-3 antibody; samples from G␣ q/11 and G␣ i3 immunoprecipitates were immunoblotted with caveolin-3 antibody. There was no increase in caveolin-3 or G␣ i1/2 from basal levels after treatment with various agonists. The m2 but not the m3 receptor was bound to caveolin-3 after treatment with ACh (C).
Finally, pretreatment of muscle cells for 10 min with CCK-8, SP, ACh, or CPA decreased PLC-␤ activity in response to subsequent treatment with the same agonist by 83 Ϯ 3 to 91 Ϯ 4%. The large decrease in PLC-␤ activity reflected homologous desensitization of the receptor as well as desensitization resulting from transient G protein sequestration in caveolae.
Linkage of G Protein-specific Desensitization to G Proteinspecific Binding to Caveolin-Pretreatment of permeabilized muscle cells with ACh (0.1 M) for 10 min decreased PLC-␤ activity in response to SP (59 Ϯ 5%) or CPA (61 Ϯ 6%) to the same extent as in intact cells. Pretreatment of the cells with ACh and the caveolin-binding G␣ q/11 fragment (50 M) blocked the decrease in PLC-␤ activity in response to SP, but not in response to CPA (Fig. 13). Conversely, pretreatment of the cells with ACh and the caveolin-binding G␣ i3 fragment (50 M) blocked the decrease in PLC-␤ activity in response to CPA but not in response to SP (Fig. 13). Pretreatment of permeabilized muscle cells with either G protein fragment alone had no effect on control PLC-␤ activity stimulated by SP or CPA (Fig. 13).
PKC-independent Desensitization of PLC-␤ Activity Mediated by G q/11 and G i3 -To rule out the involvement of PKC in heterologous desensitization of PLC-␤ activity, muscle cells were pretreated for 10 min with ACh (0.1 M) in the presence or absence of calphostin C (1 M). Pretreatment with ACh in the presence of calphostin C had no effect on the decrease in PLC-␤ activity in response to CCK-8, SP, or CPA, implying that desensitization was not caused by PKC-dependent phosphorylation of G proteins or other protein targets (receptors or effector enzymes) in the PI signaling pathway mediated by G q/11 or G i3 in smooth muscle (Fig. 14).
To determine whether pretreatment with various agonists affected receptor binding, [ 3 H]scopolamine binding was measured in dispersed smooth muscle cells before and after treatment with CCK-8, SP, or CPA. Treatment with all three agonists had no effect on [ 3 H]scopolamine binding, whereas treatment with ACh caused a significant decrease in binding, reflecting homologous desensitization of m2 and m3 receptors (Fig. 15).

DISCUSSION
Receptor desensitization by GRKs and ␤-arrestins and/or by feedback phosphorylation via second messenger-activated protein kinases, chiefly cAMP-dependent protein kinase and protein kinase C (PKC), are well established mechanisms of desensitization of response mediated by G protein-coupled receptors (45). Phosphorylation by second messenger-activated protein kinases does not require receptor occupancy and can thus target both homologous and heterologous receptors, re- ducing their ability to transduce signals and, in some instances, switching the specificity of receptor coupling to G proteins. Phosphorylation of downstream targets in the signaling pathway (e.g. G proteins or effector enzymes) can also result in desensitization of response (46 -51). PKC-dependent phosphorylation, however, is G protein-specific: G z , G 12 , and G 13 are readily phosphorylated, whereas G q/11 and G s are not (47)(48)(49). PKC-dependent phosphorylation of pertussis toxinsensitive G proteins in intestinal smooth muscle was observed with G i1 and G i2 , but not G i3 or G o, and resulted in PKC-de- FIG. 7. Inhibition of G protein binding to caveolin-3 by caveolin-binding G protein fragments. Permeabilized smooth muscle were incubated with for 10 min with SP (1 M) or CPA (1 M) alone and in combination with G␣ q/11 or G␣ i3 fragments that selectively bind to the caveolin-scaffolding domain. Caveolin-3 immunoprecipitates were probed with G␣ q/11 and G␣ i3 antibodies, and the bands were analyzed by densitometry. The caveolin-binding G␣ q/11 fragment blocked the increase in G␣ q/11 binding to caveolin elicited by SP but not the increase in G␣ i3 binding elicited by CPA. Conversely, the G␣ i3 fragment blocked the increase in G␣ i3 binding to caveolin-3 elicited by CPA but not the increase in G␣ q/11 binding elicited by SP. Bands denote representative experiments. Values denoted by bars are the means Ϯ S.E. of three experiments.  pendent desensitization of responses mediated by G i1 and G i2 (51). Consistent with these results, a selective PKC inhibitor had no effect on desensitization of responses mediated by G i3 or G q/11 in the present study, making it possible to explore other G protein-dependent mechanisms of desensitization. The results indicate that transient sequestration of receptor-activated G protein subunits by binding to caveolin leads to heterologous desensitization of responses mediated by other receptors coupled to the same G protein.
Compelling evidence exists that caveolin can act as a scaffold for the assembly and activation of signaling molecules (16,22,32). The present study suggests a novel, complementary function whereby caveolin acts to restrain cellular response by transient, selective binding of G proteins. Receptor-activated G protein subunits interact with an assembly of signaling molecules consisting of an effector enzyme (e.g. PLC-␤) and its substrate (e.g. PIP 2 ) located in caveolae. Upon GTP hydrolysis, G␣.GDP binds with high affinity to caveolin-3 and is transiently sequestered, impeding reassociation of the heterotrimeric G protein. The decrease in the levels of G protein available to receptors in the extracaveolar membrane causes a transient decrease in the ability of agonists to activate PLC-␤, i.e. a G protein-specific heterologous desensitization of response.
Caveolin Binding of Activated G Proteins-The increase of G␣ in caveolin immunoprecipitates derived from caveolin-enriched fractions and whole cell lysates was confined to activated G proteins. The G␣ subunits appeared to bind directly to caveolin-3 rather than to a caveolin-associated protein and could be competed out by the corresponding caveolin-binding G protein fragment. Thus, a caveolin-binding G␣ q/11 fragment blocked the binding of activated G␣ q/11 but not G␣ i3 to caveolin-3, whereas a caveolin-binding G␣ i3 fragment had the reverse effect, blocking the binding of activated G␣ i3 but not G␣ q/11 to caveolin-3.
The increase in G␣ q/11 and G␣ i3 binding to caveolin-3 induced by ACh was both time-and concentration-dependent, reflecting concurrent activation of m3 receptors coupled to G q/11 and m2 receptors coupled to G i3 , a notion confirmed by the increase in the binding of only one G protein subunit in the presence of selective m3 or m2 receptor antagonists. The increase induced by other agonists was also G protein-specific, with CCK-8 and SP increasing G␣ q/11 and CPA increasing G␣ i3 binding to caveolin. de Weerd and Leeb-Lundberg (29) have shown that activation of bradykinin B 2 receptors that couple to G q/11 and G i in DDT 1 MF-2 smooth muscle cells increased G␣ q and G␣ i binding to caveolin with a time course similar to that elicited by ACh in intestinal smooth muscle cells. Some receptors, for example, m2 receptors in cardiac myocytes (31) and CCK-A receptors in pancreatic acinar cells (52) also bind to caveolin upon activation. In the present study, m2 but not m3 receptors bound to caveolin-3 upon activation with ACh, implying that translocation of receptors to caveolae was not a prerequisite for desensitization of response, which was observed upon selective activation of either m2 or m3 receptors.
The decrease in caveolar PIP 2 levels suggested that activation of PLC-␤ by G proteins occurred in caveolae. A similar decrease in caveolar PIP 2 levels has been reported in A431 cells stimulated with bradykinin and epidermal growth factor (41,53,54). Our previous studies had shown that PIP 2 hydrolysis induced by receptors (e.g. CCK-A and m3 receptors) coupled to G q/11 was mediated by G␣-dependent activation of PLC-␤1, whereas PIP 2 hydrolysis induced by receptors coupled to G i/o (e.g. somatostatin-3, ␦-opioid, adenosine A1, and muscarinic m2 receptors) was mediated by ␤␥-dependent activation of PLC-␤3 (33,36,42,43).
G␣ subunits activated by the nonhydrolyzable analog, GTP␥S, did not bind to caveolin-3, implying that caveolin binding of receptor-activated G␣ subunits occurred only after GTP hydrolysis, which yielded a GDP-bound G␣ subunit with high affinity for caveolin-3 (13,15). The transient binding to caveolin was followed by reassociation of the G␣ and G␤␥ subunits and their eventual reintegration into the extracaveolar membrane. The entire cycle was completed in 40 -60 min (Fig. 5). During this interval, the stoichiometry of the heterotrimeric G protein pool accessible to receptors was altered imposing a G protein-specific barrier to activation of effector enzymes (e.g. PLC-␤) by receptors that couple to the same G protein.
G Protein-specific Desensitization of PLC-␤ Response-The time course of decrease in PLC-␤ activity in response to CCK-8 and CPA closely paralleled the time course of G␣ binding to caveolin induced by pretreatment with ACh. At various intervals and for various concentrations of ACh, the extent of decrease in PLC-␤ activity was correlated with the increase in G␣ binding to caveolin. A similar G protein-specific decrease in PLC-␤ activity was observed after sequential treatment of the cells with various agonists. Thus, activation of one receptor coupled to G q/11 (e.g. CCK-A, NK-1, or m3 receptor) inhibited subsequent PLC-␤ responses mediated by these receptors but not those mediated by A1 and m2 receptors. Conversely, activation of one receptor coupled to G i3 (e.g. m2 or A1 receptor) inhibited subsequent PLC-␤ responses mediated by these receptors but not those mediated by CCK-A, NK-1, or m3 receptors. Thus, activation of one receptor type decreased the response to another receptor type coupled to the same G protein.
It is worth noting that activation of one receptor did not induce desensitization of other receptors; pretreatment of muscle cells with CCK-8, SP, or CPA, for example, had no effect on [ 3 H]scopolamine binding, whereas pretreatment with ACh caused a large decrease in [ 3 H]scopolamine binding, indicative of homologous desensitization of m3 and m2 receptors. PLC-␤ activity in response to sequential stimulation with ACh decreased by about 90%, reflecting both receptor-specific (homologous) and G protein-specific (heterologous) desensitization.
Studies with caveolin-binding fragments of G␣ q/11 and G␣ i3 provided decisive evidence for selective binding of activated G␣ to caveolin-3 and established a linkage between caveolin bind- ing and G protein-specific heterologous desensitization of response. A caveolin-binding G␣ q/11 fragment that selectively blocked the binding of activated G␣ q/11 to caveolin-3 elicited by ACh prevented desensitization of responses mediated by other G q/11 -coupled receptors but not by G i3 -coupled receptors. Conversely, a caveolin-binding G␣ i3 fragment that selectively blocked the binding of activated G␣ i3 to caveolin-3 elicited by ACh prevented desensitization of responses mediated by other G i3 -coupled receptors but not by G q/11 -coupled receptors.
Desensitization of response by G protein binding to caveolin was not confined to G q/11 and G i3 . Both G␣ i1 and G␣ i2 bound to caveolin-3 following activation of G i1 by somatostatin-3 receptors, activation of G i2 by opioid ␦-receptors, and activation of both G i1 and G i2 by natriuretic peptide clearance receptors (42)(43)(44). The resultant desensitization was G protein-specific with one component reflecting G␣ i1 and/or G␣ i2 binding to caveolin-3 and the other component reflecting phosphorylation of G␣ i1 and/or G␣ i2 by PKC; only the latter was blocked by pretreatment of muscle cells with a PKC inhibitor (51). 2 However, neither G q/11 nor G i3 in smooth muscle is susceptible to phosphorylation by PKC (51); as shown in Fig. 14, desensitization of PLC-␤ response was not affected by inhibition of PKC activity, implying that desensitization was not mediated by PKC-dependent phosphorylation of G proteins or other protein targets (receptors or effector enzymes) in the phosphoinositide signaling pathway mediated by either G q/11 or G i3 in smooth muscle.
In summary, this study demonstrates a role for caveolin in signal transduction that depends on its ability to bind transiently receptor-activated G protein subunits and impede reassociation of the heterotrimeric species, thereby decreasing the ability of receptors that specifically couple to these G proteins to signal effectively. The process may contribute to both homologous and heterologous desensitization of response.