Interaction of Arrestins with Intracellular Domains of Muscarinic and alpha 2-Adrenergic Receptors*

(Received for publication, December 24, 1996, and in revised form, April 23, 1997)

Guangyu Wu Dagger §, Jason G. Krupnick , Jeffrey L. Benovic par and Stephen M. Lanier Dagger **

From the  Department of Microbiology and Immunology, Kimmel Cancer Institute, Thomas Jefferson University, Philadelphia, Pennsylvania 19107 and the Dagger  Department of Pharmacology, Medical University of South Carolina, Charleston, South Carolina 29425

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES


ABSTRACT

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 alpha 2A/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-M3 fusion protein, but not to the control GST peptide or i3 peptides derived from the alpha 2A/D-AR and M2-MR. However, each of the receptor subdomains bound purified beta -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]beta -arrestin and [3H]arrestin-3, but did not interact with in vitro translated or purified visual arrestin. The properties and specificity of the interaction of in vitro translated [3H]beta -arrestin, [3H]visual arrestin, and [3H]beta -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 beta -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.


INTRODUCTION

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 recognition. 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 G-protein-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 Gi, Go, and/or Gq family of G-proteins (i.e. muscarinic, alpha -adrenergic), whereas shorter i3 loops are found in the photoreceptor rhodopsin or beta -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 (1-5).

The arrestin family consists of visual arrestin, beta -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. beta -Arrestin and arrestin-3 are widely expressed and parallel the role of visual arrestin in terms of signal termination for G-protein-coupled 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 (12-14). 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 M2-muscarinic receptor (MR), M3-MR and alpha 2A/D-adrenergic receptor (AR) to use as probes to detect 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 beta -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, beta -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 × 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 beta -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 -70 °C.

Plasmid Constructions and Expression of GST Fusion Proteins

The M3-MR i3 construct was obtained from Dr. Barry Wolfe (Department of Pharmacology, Georgetown University School of Medicine, Washington, DC) and encoded the peptide Gly308-Leu497. The M2-MR and alpha 2A/D-AR i3 constructs were generated from the cDNA or genomic clones by amplification using the polymerase chain reaction. The M2-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 alpha 2A/D-AR gene (16) segment encoding the peptide Lys224-Phe374 was amplified by polymerase chain reaction and subcloned into the BamHI and EcoRI restriction sites of pGEX-2T. The M3-II and M3-III were generated from M3-MR i3 construct (M3-I, Gly308-Leu497) by taking advantage of a PstI restriction site (nucleotide 1918 of the rat M3 muscarinic receptor coding region). The M3-II construct was generated by excising the BamHI/PstI fragment from the M3-I construct with subsequent subcloning of this fragment into pGEX-3X at the BamHI and EcoRI restriction sites via use of an adaptor. The M3-III construct was generated using a similar strategy with the PstI/EcoRI fragment isolated from the M3-I construct. The M3-IV construct was prepared by digesting the M3-I construct with HindIII removing the gene segment encoding amino acids Lys369-Thr424. The purified plasmid containing the amino- and carboxyl-terminal segments was then religated to yield M3-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 µg) fractions were incubated with ~5 µg of GST fusion protein bound to the glutathione resin in 250 µl 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 M2-MR and M3-MR peptides was also evaluated by direct binding assays using tritiated arrestins generated by in vitro translation. Aliquots (~2.5 µg) of GST fusion proteins bound to a glutathione resin were incubated with [3H]beta -arrestin (long splice variant) (1062-1600 dpm/fmol), [3H]arrestin-3 (short splice variant) (1500 dpm/fmol), [3H]visual arrestin (1300-1871 dpm/fmol), or the [3H]beta -arrestin/visual arrestin chimeras BBBA (1990 dpm/fmol) and AABB (2010 dpm/fmol) in a total volume of 20 µl of buffer C for 2.5 h at 4 °C with shaking. The resin was washed three times with 100 µl of buffer C. The washed resin was resuspended in 300 µl of buffer C, mixed with 10 ml of Ecoscint A scintillation fluid, and the retained radioactivity quantitated by scintillation spectrometry at ~50% efficiency. Nonspecific binding was defined as the amount of ligand retained in parallel experiments using a control GST resin and represented ~30% of total binding at a concentration of 0.75 nM arrestin.


RESULTS

Interaction of Receptor Subdomain Probes with Brain Cytosol Proteins

The third intracellular loop of the M2-MR, M3-MR ,and the alpha 2A/D-AR consists of 180 (His208-Arg387), 238 (Arg253-Gln490), and 157 (Arg218-Phe374) amino acids, respectively. The putative first and second intracellular loops of the three receptors are similar in size ~11-12 amino acids, 1st loop; ~18-19 amino acids, 2nd loop) and the carboxyl-terminal tails of the M2-MR, M3-MR, and alpha 2A/D-AR are 23, 44, and 20 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.

The M2-MR (His208-Arg387), M3-MR (Gly308-Leu497), and alpha 2A/D-AR (Lys224-Phe374) 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 M2-MR peptide corresponded to the entire i3 loop of the receptor. The M3-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 alpha 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.


Fig. 1. Generation of receptor subdomain probes. Peptides corresponding to segments of the third intracellular loop of the human M2-MR (180 amino acids), rat M3-MR (190 amino acids), and rat alpha 2A/D-AR (151 amino acids) (A) were generated as GST fusion proteins in bacteria and purified as described under "Experimental Procedures". In A, the black segments correspond to the putative fifth and sixth membrane spans of the receptor. B, the purified fusion proteins were electrophoresed on denaturing polyacrylamide gels (10%) and visualized by Coomassie Blue stain of the proteins. The calculated molecular weights of GST and the M2-MR, M3-MR, and alpha 2A/D-AR GST fusion proteins were 28,146, 47,864, 49,796, and 43,343, respectively, including adapter amino acids. The arrows indicate the migration of GST and each of the GST fusion proteins. The numbers to the left of the gel indicate the migration of standards of known molecular weight × 10-3.
[View Larger Version of this Image (26K GIF file)]

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 M2-MR, M3-MR or the alpha 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 M3-MR affinity matrix but did not interact with matrices constructed of the GST control peptide or the i3 peptides derived from the M2-MR or the alpha 2A/D-AR (Fig. 2A, right panel). The amount of arrestin retained by the M3-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+/H+ exchanger (carboxyl terminus, 178 amino acids), a subdomain of the M2-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 M3-MR affinity matrix was specific for the M3-MR peptide.


Fig. 2. Interaction of the M2-MR, M3-MR, and alpha 2A/D-AR third intracellular loop peptides with arrestins in preparations of bovine brain cytosol. Brain cytosol was fractionated by ammonium sulfate precipitation and ion exchange chromatography as described under "Experimental Procedures." The fractionation of arrestins was determined by immunoblot of a membrane transfer of the crude and fractionated brain cytosol following SDS-PAGE (A, left panel). The cytosol fraction (100 µg of protein) enriched for arrestin was then incubated with control GST, M2-MR, M3-MR, and alpha 2A/D-AR affinity matrices (~5 µg of protein) and retained proteins visualized by immunoblotting (A, right panel). ppt, precipitate. B, the M3-MR (~5 µg of protein) was incubated with increasing concentrations of the fraction of brain cytosol enriched for arrestins and the samples processed as described above. GST, control resin containing bound glutathione S-transferase. C, specificity of the interaction between the M3-MR i3 peptide and arrestins. GST fusion proteins or GST (~5 µg) were incubated with 100 µg of brain cytosol fraction enriched for arrestins and processed as described above. The fusion proteins, vector, and bacteria used for expression were as follows: GST-M2(Lys268-Thr323, porcine)-pGEX-3X, HB101; GST-M3(Gly308-Leu497, rat)-pGEX-3X, HB101; GST-alpha 2A/D-AR(Lys224-Phe374, rat)-pGEX-2T, JM110; GST-Syp(Thr2-Leu216, human)-pGEX-kt, DH5alpha ; GST-Na+/H+ exchanger (Leu168-Gln345, rabbit cardiac)-pGEX-3X, DH5alpha ; GST-c-Jun (Met1-Thr79, human) - pGEX-2T, XL1-Blue; GST-pGEX-3X, HB101.
[View Larger Version of this Image (39K GIF file)]

The specificity of arrestin binding to the M3-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 M3-MR or the alpha 2A/D-AR i3 affinity matrix and processed as described for arrestins. Although the M3-MR affinity matrix adsorbed brain arrestins (Fig. 2), neither the M3-MR or the alpha 2A/D-AR i3 peptides interacted with the protein kinase C isoforms, phosphatase 2A, or gelsolin (Fig. 3, right panel).


Fig. 3. Interaction of brain cytosol proteins with the M3-MR and alpha 2A/D-AR third intracellular loop peptide. Brain cytosol was fractionated as described under "Experimental Procedures," and the fractions containing protein kinase C isoforms (PKC), phosphatase 2A (PTP2A), or gelsolin were determined by immunoblots of membrane transfers of the crude and fractionated brain cytosol following SDS-PAGE (left panels). Antisera dilutions: protein kinase C, 1:750; phosphatase 2A, 1:500; gelsolin, 1:1000. Cytosol fractions (100 µg of protein) enriched for the proteins of interest were then incubated with control GST, M3-MR, and alpha 2A/D-AR affinity matrices (~5 µg of protein) and retained proteins visualized by immunoblotting (right panels). "fractionated cytosol" in the right panels refers to the fraction enriched for the particular protein as determined in the series of panels on the left.
[View Larger Version of this Image (37K GIF file)]

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 M3-MR using radiolabeled arrestins. Increasing concentrations of radiolabeled beta -arrestin, arrestin-3, and visual arrestin were incubated with the M3-MR or the control GST resin. Both [3H]beta -arrestin and [3H]arrestin-3 exhibited specific binding to the M3-MR peptide relative to the binding of the arrestins to the control GST peptide (Fig. 4A). In contrast, the binding of [3H]visual arrestin to the M3-MR matrix was only slightly increased above that observed with the GST peptide itself, consistent with the lower affinity of visual arrestin at G-protein-coupled receptors other than rhodopsin. Interestingly, the affinities exhibited by [3H]beta -arrestin and [3H]arrestin-3 (0.5-1.0 nM) are in the range of the Kd for arrestin binding to purified beta 2-AR and M2-MR (15). The selectivity of the binding of [3H]arrestins to the M3-MR peptide was also observed with purified beta -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 beta -arrestin and arrestin-3. Purified beta -arrestin and arrestin-3 were adsorbed to the M3-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 M2-MR or alpha 2A/D-AR affinity matrices. However, both the M2-MR and the alpha 2A/D-AR i3 peptides bound purified beta -arrestin and arrestin-3 (Fig. 5B).2


Fig. 4. Interaction of the M3-MR with visual arrestin (v-arr), beta -arrestin (beta -arr), and arrestin-3 (arr-3). A, the M3-MR i3 peptide or GST-substituted resins (~2.5 µg 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 M3-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 M3-MR i3 peptide or GST-substituted resins (~5 µg of protein) were incubated with purified beta -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, beta -arrestin, and arrestin-3. The first two lanes indicate the signal detected with 25 ng of each arrestin.
[View Larger Version of this Image (26K GIF file)]


Fig. 5. Interaction of purified beta -arrestin and arrestin-3 with third intracellular loop peptides. In A, the M3 matrix (~5 µg of protein) was incubated with increasing concentrations of purified beta -arrestin or arrestin-3 and the samples processed as described in the legend to Fig. 2 for brain arrestin preparations. Similar results were obtained in three experiments. GST, glutathione S-transferase matrix. In B, GST-receptor subdomain fusion proteins (~5 µg of protein) were incubated with 200 µg of brain cytosol enriched for arrestins or 50 ng of either purified beta -arrestin or arrestin-3. Similar results were obtained in three experiments using different preparations of fusion protein. The aberrant migration of beta -arrestin in the M2-MR lane relative to control input is due to comigration of the fusion protein and beta -arrestin. The last three lanes on the right of the immunoblot contain aliquots of the material incubated with the resins (cytosol, 5 µg; beta -arrestin, 25 ng; arrestin-3, 25 ng). The results are representative of three individual experiments.
[View Larger Version of this Image (55K GIF file)]

The interaction of arrestins with the intact purified M2-MR was previously characterized using in vitro translated arrestins and beta -arrestin/visual arrestin chimeras (15). We thus compared the interaction of arrestins with the M2-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), [3H]beta -arrestin binding to the i3 peptide was first increased and then decreased with increasing ionic strength of the incubation buffer (Fig. 6A). Previous studies using a series of [3H]beta -arrestin/visual arrestin chimeras indicated that the selectivity of beta -arrestin and visual arrestin binding to the intact purified M2-MR and the beta 2-adrenergic receptor involved specific domains of the two arrestins (15). To determine if this selectivity of arrestin binding was maintained with the M2-MR i3 peptide, we evaluated the binding of two [3H]beta -arrestin/visual arrestin chimeras. The BBBA chimera consists of amino acids 1-340 of beta -arrestin and amino acids 346-404 of visual arrestin. The AABB chimera consists of amino acids 1-213 of visual arrestin and amino acids 208-418 of beta -arrestin. The binding of BBBA to the intact M2-MR was similar or greater than that observed with beta -arrestin, whereas the binding of AABB was intermediate between beta -arrestin and visual arrestin (15). The relative binding of [3H]beta -arrestin, [3H]visual arrestin, and the [3H]beta -arrestin/visual arrestin chimeras to the M2-MR i3 peptide, as well as the M3-MR peptide, was similar to that observed with the intact purified M2-MR (Fig. 6B).


Fig. 6. Comparison of the properties of arrestin interaction with the M2-MR i3 peptide and the intact purified M2-MR receptor. In A the M2-MR i3 peptide or GST-substituted resins (~2.5 µg of protein) were incubated with in vitro translated [3H]beta -arrestin (0.5 nM, 1062 dpm/fmol) in the presence of increasing concentrations of potassium acetate (50-500 mM) and processed as described under "Experimental Procedures." Arrestin binding to the GST control resin at each salt concentration was subtracted from that observed with the M2-MR i3 peptide matrix to generate the values shown (M2-i3, filled squares). The data are expressed as the percent of specific [3H]beta -arrestin binding obtained with incubation buffer containing 50 mM potassium acetate at which [3H]beta -arrestin binding to the GST control and M2-MR i3 peptide matrix was 1316 ± 184 dpm (1.24 ± 0.17 fmol) and 5280 ± 392 dpm (4.97 ± 0.37 fmol), respectively. Data are presented as the mean ± S.E. of three experiments. The data presented with the purified M2-muscarinic receptor (M2-MR, open squares) were adapted from Ref. 15 and represent the data obtained with the phosphorylated and agonist-activated receptor. In B, the M2-MR i3 and M3-MR i3 peptides or GST-substituted resin (~2.5 µg of protein) were incubated with in vitro translated [3H]beta -arrestin, [3H]visual arrestin, or the [3H]beta -arrestin/visual arrestin chimeras BBBA or AABB under standard incubation conditions. The concentration of each radiolabeled arrestin was 1.0 nM. The binding of the different arrestins to the GST control resin was subtracted from that observed with the M2-MR i3 or M3-MR i3 peptide matrix to generate the values shown. The data are expressed as the percent of specific binding of [3H]beta -arrestin. [3H]beta -arrestin binding to the M2-MR and M3-MR matrix was 7940 ± 556 dpm (GST control, 2284 ± 466 dpm) and 9250 ± 1156 dpm (GST control, 3204 ± 1824 dpm), respectively. Data are presented as the mean ± S.E. of three experiments. The data presented with the purified M2-muscarinic receptor (M2-MR) were adapted from Ref. 15 and represent the data obtained with the phosphorylated and agonist-activated receptor.
[View Larger Version of this Image (26K GIF file)]

Binding of Brain Arrestins, beta -Arrestin, and Arrestin-3 to Subdomains of the M3-MR i3 Loop

The next series of experiments were designed to localize subdomains of the M3-MR i3 loop required for interaction with arrestins. The M3-MR i3 peptide (M3-I) was divided into three segments: an amino-terminal region (M3-II, Gly308-Gln389), the carboxyl-terminal region (M3-III, Val390-Leu497), and a middle portion (Lys369-Thr424) (Fig. 7A). The M3 peptide possesses a concentrated negative charge in the amino-terminal third and a concentrated positive charge in the carboxyl-terminal third of the peptide. The contribution of the middle portion was determined using construct M3-IV in which this segment was deleted and the peptide Gly308-Leu368 was fused to the peptide Lys425-Leu497 (Fig. 7A). Neither construct II or III interacted with brain arrestins or purified beta -arrestin under these incubation conditions (Fig. 7B).3 However, the construct containing both the amino and carboxyl regions of the M3-I peptide (M3-IV) retained the ability to interact with brain arrestins and purified beta -arrestin (Fig. 7B). These data suggest that there are at least two sites on the M3-I peptide for beta -arrestin binding and that one of these sites by itself is insufficient. In contrast to the results with beta -arrestin, the carboxyl region of the M3-MR peptide was sufficient for interaction with arrestin-3 (Fig. 7B). Arrestin-3 bound to the M3-I, -III, and -IV, but not to construct II, suggesting that the binding motif for arrestin-3 is different from that of beta -arrestin.


Fig. 7. Subdomains of the M3-MR third intracellular loop required for interaction with brain arrestins, beta -arrestin, and arrestin-3. Subdomains of the M3-MR i3 loop (A) were generated as GST fusion proteins as described under "Experimental Procedures" and tested for arrestin interactions as described in the legends to Fig. 2. A, the 190-amino acid segment (M3-I) of the i3 loop was divided into an amino-terminal (M3-II) or carboxyl-terminal peptide (M3-III) as well as a construct (M3-IV) in which the middle portion (Lys369-Thr424; black segment in M3-IV) was deleted and peptide Gly308-Leu368 was fused to Lys425-Leu497. The GST fusion proteins M3-I to M3-IV were electrophoresed on denaturing polyacrylamide gels and stained with Coomassie Blue. B, GST or the M3-I to M3-IV resins (~5 µg of protein) were incubated with 100 µg of bovine brain cytosol fraction enriched for arrestins (cytosol), 50 ng of purified beta -arrestin or arrestin-3 and processed for immunoblotting as described in the legend to Fig. 2. Similar results were obtained in three to five experiments using different batches of fusion protein.
[View Larger Version of this Image (26K GIF file)]


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 transduction 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 M2-MR, M3-MR, and alpha 2-AR. Both the muscarinic receptor subtypes and the alpha 2-AR are capable of coupling to multiple G-proteins and effectors in a cell type-specific manner. Activation of the M2-MR results in inhibition of adenylyl cyclase, whereas the M3-MR couples to the inositol phosphate/protein kinase C signaling pathway. The alpha 2A/D-AR is generally associated with inhibition of adenylyl cyclase, but activation of this receptor also influences signal transduction pathways involving phospholipases, p21ras, 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 beta -arrestin and arrestin-3 bound to the i3 loops of the muscarinic receptor subtypes and the alpha 2A/D-AR. The major observations concerning this interaction are as follows. First, the interaction of beta -arrestin and arrestin-3 with the M3-MR involved different regions of the i3 loop. Second, the i3 domains from the muscarinic receptors and the alpha 2A/D-AR differed in their ability to interact with endogenous bovine brain arrestins in a crude cytosol fraction versus purified recombinant beta -arrestin and arrestin-3. Third, the interaction of arrestins with the i3 loop peptides occurred in the absence of peptide phosphorylation.

Both beta -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 M3-MR i3 loop, amino acids Gly308-Leu368 and Lys425-Leu497 are required for binding of beta -arrestin and the partially purified brain arrestin, whereas amino acids Gly308-Leu368 are not required for binding of arrestin-3. The differential interaction of the two arrestins with the M3-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 M2-MR, M3-MR and alpha 2A/D-AR all bound purified arrestins, only the M3-MR was capable of interacting with brain cytosol arrestins. The inability of brain arrestins to interact with the M2-MR and alpha 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 M2-MR and alpha 2A/D-AR, but not the M3-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 beta -arrestin and arrestin-3 with the beta 2-AR or the M2-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). beta -Arrestin and arrestin-3 also bind to nonphosphorylated M2-MR or beta 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-protein-coupled receptors appear capable of serving as a docking site for arrestin independent of receptor phosphorylation. Indeed, the properties of arrestin binding to the M2-MR i3 peptide appeared similar to those exhibited by the purified receptor protein that was phosphorylated and/or activated by agonist. The binding of beta -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.


FOOTNOTES

*   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. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§   Visiting scientist from the Chinese Academy of Medical Sciences and Peking Union Medical College.
par    Established Investigator of the American Heart Association.
**   To whom correspondence should be addressed: Dept. of Pharmacology, Medical University of South Carolina, 171 Ashley Ave., Charleston, SC 29425. Tel.: 803-792-2574; Fax: 803-792-2475; E-mail: laniersm{at}musc.edu.
1   The abbreviations used are: i3, third intracellular loop; MR, muscarinic receptor; AR, adrenergic receptor; PAGE, polyacrylamide gel electrophoresis.
2   Due to differences in the degree of resin substitution for various fusion proteins, variable membrane transfer efficiencies, and relative signal intensities, it is difficult to accurately determine the affinity of arrestin binding to the resins. However, data obtained thus far indicate no dramatic differences in the affinity of the purified arrestins for the three i3 peptides.
3   Based on the relative migration of beta -arrestin and arrestin-3 in 10% denaturing polyacrylamide gels and the comigration of the partially purified brain arrestin with beta -arrestin (Fig. 5), the arrestin identified in the 100 mM NaCl elution of the 90% ammonium sulfate precipitate is predominantly beta -arrestin (G. Wu and S. M. Lanier, unpublished data).

ACKNOWLEDGEMENTS

We thank Dr. Barry Wolfe, Dr. Tatsuya Haga (Department of Biochemistry, Institute for Brain Research, Faculty of Medicine, University of Tokyo, Japan), Dr. Larry Fliegel (Department of Biochemistry and Pediatrics, University of Alberta, Alberta, Canada), Dr. Gen-Shen Feng (Department of Biochemistry and Molecular Biology, Indiana University School of Medicine), and Dr. Steven Rosenzweig (Department of Pharmacology, Medical University of South Carolina, Charleston, SC) for providing the M3-MR, M2-MR (56-amino acid segment for the third intracellular loop), Na+/H+ exchanger, Syp, and c-Jun fusion protein constructs, respectively. We also thank Dr. Larry Donoso for mAbF4C1 and Dr. Vsevolod Gurevich (Sun Health Research Institute, Sun City, AZ) for purified arrestins. The expression vector containing the peptide derived from the third intracellular loop of the alpha 2A/D-adrenergic receptor was generated by Dr. Sally Kadkhodayan in the laboratory of Dr. Lanier.


REFERENCES

  1. Sterne-Marr, R., and Benovic, J. L. (1995) Vitam. Horm. 51, 193-234 [Medline] [Order article via Infotrieve]
  2. Premont, R. T., Inglese, J., and Lefkowitz, R. J. (1995) FASEB J. 9, 175-182 [Abstract/Free Full Text]
  3. Ranganathan, R., and Stevens, C. F. (1995) Cell 81, 841-848 [Medline] [Order article via Infotrieve]
  4. Palczewski, K. (1994) Protein Sci. 3, 1355-1361 [Abstract/Free Full Text]
  5. Dolph, P. J., Ranganathan, R., Colley, N. J., Hardy, R. W., Socolich, M., and Zuker, C. S. (1993) Science 260, 1910-1916 [Medline] [Order article via Infotrieve]
  6. Lohse, M. J., Benovic, J. L., Codina, J., Caron, M. G., and Lefkowitz, R. J. (1990) Science 248, 1547-1550 [Medline] [Order article via Infotrieve]
  7. Shinohara, T., Dietzschold, B., Craft, C. M., Wistow, G., Early, J. J., Donoso, L. A., Horwitz, J., and Tao, R. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 6975-6979 [Abstract]
  8. Attramadal, H., Arriza, J. L., Aoki, C., Dawson, T. M., Codina, J., Kwatra, M. M., Snyder, S. H., Caron, M. G., and Lefkowitz, R. J. (1992) J. Biol. Chem. 267, 17882-17890 [Abstract/Free Full Text]
  9. Sterne-Marr, R., Gurevich, V. V., Goldsmith, P., Bodine, R. C., Sanders, C., Donoso, L. A., and Benovic, J. L. (1993) J. Biol. Chem. 268, 15640-15648 [Abstract/Free Full Text]
  10. Murakami, A., Yajima, T., Sakuma, H., McLaren, M. J., and Inana, G. (1993) FEBS Lett. 334, 203-209 [CrossRef][Medline] [Order article via Infotrieve]
  11. Craft, C. M., Whitmore, D. H., and Wiechmann, A. F. (1994) J. Biol. Chem. 269, 4613-4619 [Abstract/Free Full Text]
  12. Ferguson, S. S. G., Downey, W. E., III, Colapietro, A., Barak, L. S., Menard, L., and Caron, M. G. (1996) Science 271, 363-366 [Abstract]
  13. Goodman, O. B., Jr., Krupnick, J. G., Santini, F., Gurevich, V. V., Penn, R. B., Gagnon, A. W., Keen, J. H., and Benovic, J. L. (1996) Nature 383, 447-450 [CrossRef][Medline] [Order article via Infotrieve]
  14. Zhang, J., Ferguson, S. S. G., Barak, L. S., Menard, L., and Caron, M. G. (1996) J. Biol. Chem. 271, 18302-18305 [Abstract/Free Full Text]
  15. Gurevich, V. V., Dion, S. B., Onorato, J. J., Ptasienski, J., Kim, C. M., Sterne-Marr, R., Hosey, M. M., and Benovic, J. L. (1995) J. Biol. Chem. 270, 720-731 [Abstract/Free Full Text]
  16. Lanier, S. M., Downing, S., Duzic, E., and Homcy, C. J. (1991) J. Biol. Chem. 266, 10470-10478 [Abstract/Free Full Text]
  17. Sato, M., Kataoka, R., Dingus, J., Wilcox, M., Hildebrandt, J. D., and Lanier, S. M. (1995) J. Biol. Chem. 270, 15269-15276 [Abstract/Free Full Text]
  18. Sato, M., Ribas, C., Hildebrandt, J. D., and Lanier, S. M. (1996) J. Biol. Chem. 271, 30052-30060 [Abstract/Free Full Text]
  19. Gurevich, V. V., and Benovic, J. L. (1992) J. Biol. Chem. 267, 21919-21923 [Abstract/Free Full Text]
  20. Krupnick, J. G., Gurevich, V. V., Schepers, T., Hamm, H. E., and Benovic, J. L. (1994) J. Biol. Chem. 269, 3226-3232 [Abstract/Free Full Text]
  21. Gurevich, V. V., Richardson, R. M., Kim, C. M., Hosey, M. M., and Benovic, J. L. (1993) J. Biol. Chem. 268, 16879-16882 [Abstract/Free Full Text]
  22. Sohlemann, P., Hekman, M., Puzicha, M., Buchen, C., and Lohse, M. J. (1995) Biochemistry 232, 464-472
  23. Smith, W. C., Milam, A. H., Dugger, D., Arendt, A., Hargrave, P. A., and Palczewski, K. (1994) J. Biol. Chem. 269, 15407-15410 [Abstract/Free Full Text]
  24. Plangger, A., Malicki, D., Whitney, M., and Paulsen, R. (1994) J. Biol. Chem. 269, 26969-26975 [Abstract/Free Full Text]
  25. Lefkowitz, R. J., Cotecchia, S., Samama, P., and Costa, T. (1993) Trends Pharmacol. Sci. 14, 303-307 [CrossRef][Medline] [Order article via Infotrieve]
  26. Allen, L. F., Lefkowitz, R. J., Caron, M. G., and Cotecchia, S. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 11354-11358 [Abstract]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.