Interaction of Arrestins with Intracellular Domains of Muscarinic and α2-Adrenergic Receptors*

The intracellular domains of G-protein-coupled receptors provide sites for interaction with key proteins involved in signal initiation and termination. As an initial approach to identify proteins interacting with these receptors and the receptor motifs required for such interactions, we used intracellular subdomains of G-protein-coupled receptors as probes to screen brain cytosol proteins. Peptides from the third intracellular loop (i3) of the M2-muscarinic receptor (MR) (His208–Arg387), M3-MR (Gly308–Leu497), or α2A/D-adrenergic receptor (AR) (Lys224–Phe374) were generated in bacteria as glutathione S-transferase (GST) fusion proteins, bound to glutathione-Sepharose and used as affinity matrices to detect interacting proteins in fractionated bovine brain cytosol. Bound proteins were identified by immunoblotting following SDS-polyacrylamide gel electrophoresis. Brain arrestins bound to the GST-M3fusion protein, but not to the control GST peptide or i3 peptides derived from the α2A/D-AR and M2-MR. However, each of the receptor subdomains bound purified β-arrestin and arrestin-3. The interaction of the M3-MR and M2-MR i3 peptides with arrestins was further investigated. The M3-MR i3 peptide bound in vitro translated [3H]β-arrestin and [3H]arrestin-3, but did not interact with in vitro translated or purified visual arrestin. The properties and specificity of the interaction ofin vitro translated [3H]β-arrestin, [3H]visual arrestin, and [3H]β-arrestin/visual arrestin chimeras with the M2-MR i3 peptide were similar to those observed with the intact purified M2-MR that was phosphorylated and/or activated by agonist. Subsequent binding site localization studies indicated that the interaction of β-arrestin with the M3-MR peptide required both the amino (Gly308–Leu368) and carboxyl portions (Lys425–Leu497) of the receptor subdomain. In contrast, the carboxyl region of the M3-MR i3 peptide was sufficient for its interaction with arrestin-3.

G-protein-coupled receptors possess a characteristic seven segments of hydrophobic amino acids that likely serve as membrane spans to form a core motif important for ligand recogni-tion. The interaction of agonist with the receptor initiates an ill-defined conformational adjustment in this core motif, which is propagated to intracellular domains of the receptor resulting in the activation of G-protein and the initiation of intracellular signaling events. For most members of the superfamily of Gprotein-coupled receptors, the third intracellular (i3) 1 loop and the carboxyl-terminal tail of the receptor are key sites for signal initiation and termination, and these receptor domains also exhibit the greatest variability in size among different subfamilies of these receptors. The largest i3 domains (100 -240 amino acids) are found in receptors coupled to the G i , G o , and/or G q family of G-proteins (i.e. muscarinic, ␣-adrenergic), whereas shorter i3 loops are found in the photoreceptor rhodopsin or ␤-adrenergic receptors (20 -50 amino acids). During the process of signal initiation and termination, several proteins interact with the receptor. The interaction of arrestins with G-protein-coupled receptors is a key component of signal termination (1)(2)(3)(4)(5).
The arrestin family consists of visual arrestin, ␤-arrestin, arrestin-3, and a cone-specific arrestin termed C-or X-arrestin (6 -11). In vertebrates, visual arrestin interacts with phosphorylated rhodopsin in rod cells to terminate signal propagation by interfering with receptor coupling to transducin. ␤-Arrestin and arrestin-3 are widely expressed and parallel the role of visual arrestin in terms of signal termination for G-proteincoupled receptors other than rhodopsin. The affinity of arrestin binding to G-protein-coupled receptors is increased by receptor phosphorylation and/or activation by agonist. Receptors of this class are phosphorylated to varying degrees by protein kinase A and C as well as kinases specific for the activated conformation of the receptor (G-protein-coupled receptor kinases). The phosphorylation of the receptor by G-protein-coupled receptor kinases and subsequent arrestin binding are intimately associated with receptor desensitization and sequestration (12)(13)(14). Resensitization of the receptor protein involves dissociation of bound arrestin and receptor dephosphorylation.
The interaction of receptors with G-proteins, protein kinases, arrestins, and additional entities controlling receptor trafficking apparently involves discrete motifs in cytoplasmic domains of the receptor. The associations of these proteins with the receptor likely occur within a signal transduction complex that may also include various effector molecules and other proteins that influence signaling specificity/efficiency. To define receptor subdomains important for protein interactions and to begin to identify the components of a signal transduction complex for G-protein-coupled receptor subtypes, we generated peptides from the i3 loop of the M 2 -muscarinic receptor (MR), M 3 -MR and ␣ 2A/D -adrenergic receptor (AR) to use as probes to detect * This work was supported in part by Grants NS24821 (to S. M. L.) and GM47419 (to J. L. B.) from the National Institutes of Health and Grant 2235 (to S. M. L.) from the Council for Tobacco Research. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  interacting proteins in bovine brain cytosol. In the present report, we determined the interaction of the i3 loop with cytosolic proteins involved in receptor regulation.

EXPERIMENTAL PROCEDURES
Materials-Radiolabeled arrestins were in vitro translated as described previously (15). Bovine ␤-arrestin, visual arrestin, and arrestin-3 were also expressed in BL21 cells and purified to homogeneity by successive chromatography on heparin-and Q-Sepharose (13). Antibodies to protein kinase C isoforms were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), and phosphatase 2A/C antibody was from Calbiochem. Anti-gelsolin monoclonal antibody (GS-2C4) was obtained from Sigma. Monoclonal antibody mAbF4C1, which recognizes the epitope DGVVLVD present in visual arrestin, ␤-arrestin, and arrestin-3, was generously provided by Dr. L. Donoso (Wills Eye Hospital, Philadelphia, PA). Glutathione-Sepharose 4B was purchased from Pharmacia Biotech Inc. Polyvinylidene difluoride membranes were obtained from Gelman Sciences (Ann Arbor, MI).
Fractionation of Brain Cytosolic Proteins-Bovine brain cytosolic proteins in buffer A (10 mM Tris-HCl, pH 7.5, 0.5 mM phenylmethylsulfonyl fluoride) containing 250 mM sucrose were precipitated with 40% ammonium sulfate and pelleted by centrifugation (100,000 ϫ g, 45 min). The precipitated proteins were resuspended in a minimal volume of 50 mM Tris-HCl, pH 8.0, followed by extensive dialysis (4 liters of buffer A; 4 liters of buffer B (20 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1 mM EGTA, 2 mM ␤-mercaptoethanol). The supernatant from the 40% ammonium sulfate precipitate was brought to 90% ammonium sulfate to precipitate additional proteins and subsequently processed as described for the 40% ammonium sulfate precipitate. The dialyzed solutions were clarified by centrifugation and applied to an anion-exchange resin (DEAE-Biogel A) equilibrated with buffer B. The column was washed with buffer B and proteins eluted sequentially with buffer B containing 100, 250, and 500 mM NaCl. Eluted proteins were desalted by dialysis, concentrated by lyophilization, and stored at Ϫ70°C.
Plasmid Constructions and Expression of GST Fusion Proteins-The M 3 -MR i3 construct was obtained from Dr. Barry Wolfe (Department of Pharmacology, Georgetown University School of Medicine, Washington, DC) and encoded the peptide Gly 308 -Leu 497 . The M 2 -MR and ␣ 2A/D -AR i3 constructs were generated from the cDNA or genomic clones by amplification using the polymerase chain reaction. The M 2 -MR subdomain was inserted into the EcoRI restriction site of the pGEX-2T vector (Promega, Madison, WI). Using primers containing appropriate restriction sites, the rat ␣ 2A/D -AR gene (16) segment encoding the peptide Lys 224 -Phe 374 was amplified by polymerase chain reaction and subcloned into the BamHI and EcoRI restriction sites of pGEX-2T. The M 3 -II and M 3 -III were generated from M 3 -MR i3 construct (M 3 -I, Gly 308 -Leu 497 ) by taking advantage of a PstI restriction site (nucleotide 1918 of the rat M 3 muscarinic receptor coding region). The M 3 -II construct was generated by excising the BamHI/PstI fragment from the M 3 -I construct with subsequent subcloning of this fragment into pGEX-3X at the BamHI and EcoRI restriction sites via use of an adaptor. The M 3 -III construct was generated using a similar strategy with the PstI/EcoRI fragment isolated from the M 3 -I construct. The M 3 -IV construct was prepared by digesting the M 3 -I construct with HindIII removing the gene segment encoding amino acids Lys 369 -Thr 424 . The purified plasmid containing the amino-and carboxyl-terminal segments was then religated to yield M 3 -IV. The structure of each construct used in the present study was verified by restriction mapping and nucleotide sequence analysis. The fusion proteins were expressed in bacteria and purified using a glutathione affinity matrix according to the manufacturer's instructions. Immobilized fusion proteins were either used immediately or stored at 4°C for no longer than 3 days. Each batch of fusion protein used in experiments was first analyzed by SDS-PAGE and Coomassie Blue staining.
Protein Interaction Assays-Brain cytosol (1-200 g) fractions were incubated with ϳ5 g of GST fusion protein bound to the glutathione resin in 250 l of buffer C (20 mM Tris-HCl, pH 7.5, 70 mM NaCl) for 2.5 h at 4°C. The resin was washed three times with 0.5 ml of buffer C, and the retained proteins were solubilized and applied to a denaturing 10% polyacrylamide gel. Polyvinylidene difluoride membrane transfers were evaluated by immunoblotting as described previously (17) using specific antibodies. The interactions of purified arrestins with the i3 peptides was determined in a similar manner. Arrestin binding to the M 2 -MR and M 3 -MR peptides was also evaluated by direct binding assays using tritiated arrestins generated by in vitro translation. Aliquots (ϳ2.5 g) of GST fusion proteins bound to a glutathione resin were incubated with [ 3 H]␤-arrestin (long splice variant) (1062-1600 dpm/fmol), [ 3 H]arrestin-3 (short splice variant) (1500 dpm/fmol), [ 3 H]visual arrestin (1300 -1871 dpm/fmol), or the [ 3 H]␤-arrestin/visual arrestin chimeras BBBA (1990 dpm/fmol) and AABB (2010 dpm/fmol) in a total volume of 20 l of buffer C for 2.5 h at 4°C with shaking. The resin was washed three times with 100 l of buffer C. The washed resin was resuspended in 300 l of buffer C, mixed with 10 ml of Ecoscint A scintillation fluid, and the retained radioactivity quantitated by scintillation spectrometry at ϳ50% efficiency. Nonspecific binding was defined as the amount of ligand retained in parallel experiments using a control GST resin and represented ϳ30% of total binding at a concentration of 0.75 nM arrestin. amino acids in length, respectively. As an initial attempt to define proteins that may interact with the intracellular domains of these G-protein-coupled receptors, we focused on the i3 loop as it is the largest intracellular domain in this receptor group. The juxtamembrane segments of the i3 domain are of critical importance for receptor coupling to G-protein, whereas other segments participate in receptor phosphorylation, receptor trafficking, and other aspects of signal propagation.

Interaction of Receptor Subdomain Probes with Brain
The M 2 -MR (His 208 -Arg 387 ), M 3 -MR (Gly 308 -Leu 497 ), and ␣ 2A/D -AR (Lys 224 -Phe 374 ) i3 peptides were expressed in bacteria as a GST fusion protein and used to generate an affinity matrix by saturating a glutathione-Sepharose resin with the fusion protein (Fig. 1). The M 2 -MR peptide corresponded to the entire i3 loop of the receptor. The M 3 -MR i3 peptide began 45 amino acids downstream of the amino terminus of the i3 loop and terminated seven amino acids into the VI membrane span. The ␣ 2A/D -AR peptide began six amino acids downstream of the amino terminus of the i3 loop and terminated at the beginning of the VI membrane span. To determine the interaction of these receptor-derived peptides with cytosolic proteins, we first fractionated bovine brain cytosol to enrich for potential interacting proteins. In the first series of experiments, we determined the interaction of receptor subdomains with brain arrestins. Bovine brain cytosol was fractionated by ammonium sulfate precipitation and ion exchange chromatography and the fractions enriched for arrestins were determined by immunoblotting ( Fig.  2A, left panel). The cytosol fraction enriched for arrestin was incubated with the M 2 -MR, M 3 -MR or the ␣ 2A/D -AR i3 subdomains as well as the control GST affinity matrix and arrestins retained by the matrices were determined by immunoblotting. Brain arrestins were adsorbed by the M 3 -MR affinity matrix but did not interact with matrices constructed of the GST control peptide or the i3 peptides derived from the M 2 -MR or the ␣ 2A/D -AR ( Fig. 2A, right panel). The amount of arrestin retained by the M 3 -MR matrix was directly related to the amount of cytosol present during incubation (Fig. 2B). Brain arrestins were not retained by GST fusion proteins containing subdomains of the tyrosine phosphatase Syp (Src homology 2 domains, 215 amino acids), the Na ϩ /H ϩ exchanger (carboxyl terminus, 178 amino acids), a subdomain of the M 2 -MR i3 loop (56 amino acids), or the transregulatory protein c-Jun (amino terminus, 79 amino acids) (Fig. 2C). These data indicated that the interaction of brain arrestins with the M 3 -MR affinity matrix was specific for the M 3 -MR peptide.
The specificity of arrestin binding to the M 3 -MR peptide was further investigated by determining the interaction of the peptide with other cytosolic proteins. The distribution of protein kinase C isoforms, phosphatase 2A, and the actin-binding protein gelsolin in the fractionated bovine brain cytosol was determined by immunoblotting (Fig. 3, left panel). Two protein kinase C isoforms fractionated in the 250 mM NaCl elution of the 90% ammonium sulfate precipitate. The phosphatase 2A immunoreactive species was identified in the 250 mM NaCl elution of the 40% ammonium sulfate precipitate. Gelsolin was enriched in the 100 mM NaCl elution of the 40% ammonium sulfate precipitate. The appropriate fraction was then incubated with the M 3 -MR or the ␣ 2A/D -AR i3 affinity matrix and processed as described for arrestins. Although the M 3 -MR affinity matrix adsorbed brain arrestins (Fig. 2), neither the M 3 -MR or the ␣ 2A/D -AR i3 peptides interacted with the protein kinase C isoforms, phosphatase 2A, or gelsolin (Fig. 3, right  panel).
Interaction of the i3 Peptides with Arrestins-The interaction of arrestins with the i3 peptides was investigated in more detail to define issues of arrestin selectivity and sites of arrestin association. In the first series of experiments, we evaluated arrestin binding to the M 3 -MR using radiolabeled arrestins. Increasing concentrations of radiolabeled ␤-arrestin, arrestin-3, and visual arrestin were incubated with the M 3 -MR or the control GST resin. Both the binding of [ 3 H]arrestins to the M 3 -MR peptide was also observed with purified ␤-arrestin and visual arrestin (Fig. 4B). In the second series of experiments, the interaction of arrestins with the i3 peptides was further evaluated using purified ␤-arrestin and arrestin-3. Purified ␤-arrestin and arrestin-3 were adsorbed to the M 3 -MR matrix in a concentration-dependent manner (Fig. 5A). Neither arrestin was retained by the control GST resin (Fig. 5A). As indicated above, brain arrestins were not retained by the M 2 -MR or ␣ 2A/D -AR affinity matrices. However, both the M 2 -MR and the ␣ 2A/D -AR i3 peptides bound purified ␤-arrestin and arrestin-3 (Fig. 5B). 2 The interaction of arrestins with the intact purified M 2 -MR was previously characterized using in vitro translated arrestins and ␤-arrestin/visual arrestin chimeras (15). We thus compared the interaction of arrestins with the M 2 -MR i3 peptide and the intact receptor relative to the influence of ionic strength and the selectivity of binding for the different arrestins. As observed for the intact purified receptor that was phosphorylated and/or activated by agonist (15)

. Interaction of the M 3 -MR with visual arrestin (v-arr), ␤-arrestin (␤-arr), and arrestin-3 (arr-3).
A, the M 3 -MR i3 peptide or GST-substituted resins (ϳ2.5 g of protein) were incubated with increasing concentrations of radiolabeled arrestins and processed as described under "Experimental Procedures." Arrestin binding to the GST control resin at each arrestin concentration was subtracted from that observed with the M 3 -MR matrix to generate the values shown. Data are representative of three different experiments using different batches of arrestins and fusion proteins. In B, the M 3 -MR i3 peptide or GST-substituted resins (ϳ5 g of protein) were incubated with purified ␤-arrestin (50 ng) or visual arrestin (50 ng) and the samples were processed as described under "Experimental Procedures." Arrestins retained by the substituted resins were identified by immunoblotting using the monoclonal antibody mAbF4C1, which recognizes the epitope DGVVLVD present in visual arrestin, ␤-arrestin, and arrestin-3. The first two lanes indicate the signal detected with 25 ng of each arrestin. using construct M 3 -IV in which this segment was deleted and the peptide Gly 308 -Leu 368 was fused to the peptide Lys 425 -Leu 497 (Fig. 7A). Neither construct II or III interacted with brain arrestins or purified ␤-arrestin under these incubation conditions (Fig. 7B). 3 However, the construct containing both the amino and carboxyl regions of the M 3 -I peptide (M 3 -IV) retained the ability to interact with brain arrestins and purified ␤-arrestin (Fig. 7B). These data suggest that there are at least two sites on the M 3 -I peptide for ␤-arrestin binding and that one of these sites by itself is insufficient. In contrast to the results with ␤-arrestin, the carboxyl region of the M 3 -MR peptide was sufficient for interaction with arrestin-3 (Fig. 7B). Arrestin-3 bound to the M 3 -I, -III, and -IV, but not to construct II, suggesting that the binding motif for arrestin-3 is different from that of ␤-arrestin.

DISCUSSION
As is apparently the case with most complex signal processing systems, G-protein-coupled receptors likely operate within a signal transduction complex that is either preexisting or is generated by the biological stimuli. The components of such a signaling complex are unclear but might include proteins that influence events at the receptor-G-protein or G-protein-effector interface or contribute to the formation of the signal transduction complex itself (Refs. 17 and 18, and references therein). As part of an effort to identify components of this signal transduc- 3 Based on the relative migration of ␤-arrestin and arrestin-3 in 10% denaturing polyacrylamide gels and the comigration of the partially purified brain arrestin with ␤-arrestin (Fig. 5), the arrestin identified in the 100 mM NaCl elution of the 90% ammonium sulfate precipitate is predominantly ␤-arrestin (G. Wu and S. M. Lanier, unpublished data). tion complex, we initiated two experimental approaches. The first approach was based on a biological activity and was designed to detect factors that might influence the transfer of signal from receptor to G-protein (17,18). The second approach, described in the present report, involved the identification of proteins capable of associating with receptors of this class. In the latter approach, we used intracellular domains of G-protein-coupled receptors as "bait" to search for interacting proteins in brain cytosol. Receptor subdomains were generated from the i3 loop of the M 2 -MR, M 3 -MR, and ␣ 2 -AR. Both the muscarinic receptor subtypes and the ␣ 2 -AR are capable of coupling to multiple G-proteins and effectors in a cell typespecific manner. Activation of the M 2 -MR results in inhibition of adenylyl cyclase, whereas the M 3 -MR couples to the inositol phosphate/protein kinase C signaling pathway. The ␣ 2A/D -AR is generally associated with inhibition of adenylyl cyclase, but activation of this receptor also influences signal transduction pathways involving phospholipases, p21 ras , the mitogen-activated protein kinase signaling pathway, and various ion channels (see Ref. 17, and references therein).
We first evaluated the interaction of the i3 loops with the arrestin family of proteins, which are implicated in receptor uncoupling and internalization. The arrestins exhibited a specific interaction with the i3 receptor subdomains. The receptor subdomain probes did not interact with selected protein kinase C isoforms, cytosolic phosphatase 2A, or the actin-binding protein gelsolin. However, both ␤-arrestin and arrestin-3 bound to the i3 loops of the muscarinic receptor subtypes and the ␣ 2A/D -AR. The major observations concerning this interaction are as follows. First, the interaction of ␤-arrestin and arrestin-3 with the M 3 -MR involved different regions of the i3 loop. Second, the i3 domains from the muscarinic receptors and the ␣ 2A/D -AR differed in their ability to interact with endogenous bovine brain arrestins in a crude cytosol fraction versus purified recombinant ␤-arrestin and arrestin-3. Third, the interaction of arrestins with the i3 loop peptides occurred in the absence of peptide phosphorylation.
Both ␤-arrestin and arrestin-3 are widely expressed, but exhibit a heterogeneous intra-tissue distribution, and both arrestins are alternatively spliced to generate a short and long form of the protein. Although arrestins clearly interact with the receptor protein (15, 19 -22), the sites of interaction and the selectivity among the different arrestins and receptor families outside the visual system are undefined. However, the present study indicates that there are receptor motifs that are indeed capable of distinguishing the two types of arrestin. Within the M 3 -MR i3 loop, amino acids Gly 308 -Leu 368 and Lys 425 -Leu 497 are required for binding of ␤-arrestin and the partially purified brain arrestin, whereas amino acids Gly 308 -Leu 368 are not required for binding of arrestin-3. The differential interaction of the two arrestins with the M 3 -MR subdomain may relate to the charge distribution within the receptor peptide segments and the structural properties of the two arrestins.
Although the i3 peptides from the M 2 -MR, M 3 -MR and ␣ 2A/ D-AR all bound purified arrestins, only the M 3 -MR was capable of interacting with brain cytosol arrestins. The inability of brain arrestins to interact with the M 2 -MR and ␣ 2 -AR may simply reflect differences in the relative affinities of the different receptor subdomains for the arrestins. However, there were no apparent differences in the efficiency of the interaction of the three receptor subdomains with the purified arrestins under these experimental conditions, suggesting that other factors may be regulating this interaction. It is possible that the brain arrestins are slightly different from the recombinant arrestins, and that this difference contributes to selective interactions with receptor families. Alternatively, perhaps there are additional proteins in the fractionated brain cytosol that impede arrestin binding to the M 2 -MR and ␣ 2A/D -AR, but not the M 3 -MR i3 peptides. Such proteins may interact with motifs in the i3 peptide or perhaps with arrestin itself.
Interaction of visual arrestin with rhodopsin or ␤-arrestin and arrestin-3 with the ␤ 2 -AR or the M 2 -MR involves multiple contact sites that impart apparent positive cooperativity to the reaction (15, 19 -21). Arrestin binding to the receptor is proposed to first involve an ionic interaction that senses the phosphorylation and activation state of the receptor. If the receptor is phosphorylated or agonist-occupied, the apparent affinity of arrestin binding to the receptor is increased. The phosphorylated receptor subdomain likely forms a component of the arrestin binding site, although phosphorylation-dependent conformational shifts in the intracellular regions of the receptor may also reveal sites that participate in arrestin binding. Whereas the binding of arrestin to rhodopsin is highly dependent upon receptor phosphorylation and activation, a truncated splice variant of visual arrestin binds to nonphosphorylated rhodopsin. The truncated and full-length visual arrestins differ in their subcellular distribution, with the former constitutively localized to disc membranes, while the latter associates with the disc membranes in a light-dependent manner (23). ␤-Arrestin and arrestin-3 also bind to nonphosphorylated M 2 -MR or ␤ 2 -AR and high affinity binding is less dependent upon the activation state of the receptor than is the interaction between visual arrestin and rhodopsin (15,23). The possible interaction of arrestins with nonphosphorylated receptors is also suggested by the ability of phosphorylation-defective mutants to undergo receptor desensitization and sequestration. An important role for arrestin interaction with a nonphosphorylated receptor is also apparent for visual signaling in invertebrates (24). In the dipteran Calliphora, arrestin interaction with the receptor protein clearly precedes receptor phosphorylation, and indeed the arrestin-receptor complex is the preferred kinase substrate (24). As the interaction of arrestins with the i3 peptides in the present study occurred without peptide phosphorylation, the i3 loop or perhaps other domains of G-proteincoupled receptors appear capable of serving as a docking site for arrestin independent of receptor phosphorylation. Indeed, the properties of arrestin binding to the M 2 -MR i3 peptide appeared similar to those exhibited by the purified receptor protein that was phosphorylated and/or activated by agonist. The binding of ␤-arrestin and arrestin-3 to the i3 peptides may also relate to the agonist-induced conformational changes responsible for initiating the signaling cascade. In the absence of agonist, the regions of the receptor involved in G-protein activation are stabilized in a conformation that acts as a "brake" on signal initiation (25). Such a conformational "brake" may be released by discrete mutations in the i3 loop of some receptors, such that the receptor becomes constitutively active (26). An analogous situation may occur when the i3 loop is separated from the conformational restrictions imposed by membrane spans of the receptor and assumes an "activated and/or accessible conformation" that is recognized by arrestin.
The demonstration that the experimental approach using receptor subdomains as probes for receptor-associated proteins resulted in the detection of protein-protein interactions of clear biological relevance underscores the potential utility of the system to identify additional interacting proteins that may contribute to the formation of a signal transduction complex. Such interacting proteins may play an important role in directing the receptor-initiated signal to a specific effector pathway. In contrast to signaling events in the visual system where the components are localized and the interactions between the individual molecules are relatively specific, other G-protein-coupled receptors operate in diverse cell types and couple to multiple G-proteins and effectors. Perhaps, such receptors have evolved larger i3 loops to maintain the fidelity of the signaling system by providing sites for interaction with additional accessory proteins that influence receptor trafficking and/or signaling specificity and efficiency.