The Adaptor Complex 2 Directly Interacts with the (cid:1) 1b -Adrenergic Receptor and Plays a Role in Receptor Endocytosis*

Using the yeast two-hybrid system, we identified the (cid:2) 2 subunit of the clathrin adaptor complex 2 as a protein interacting with the C-tail of the (cid:1) 1b -adrenergic receptor (AR). Direct association between the (cid:1) 1b -AR and (cid:2) 2 was demonstrated using a solid phase overlay assay. The (cid:1) 1b -AR/ (cid:2) 2 interaction occurred inside the cells, as shown by the finding that the transfected (cid:1) 1b -AR and the endogenous (cid:2) 2 could be coimmunoprecipitated from HEK-293 cell extracts. Mutational analysis of the (cid:1) 1b -AR revealed that the binding site for (cid:2) 2 does not involve canonical Y XX (cid:3) or dileucine motifs but a stretch of eight arginines on the receptor C-tail. The binding domain of (cid:2) 2 for the receptor C-tail involves both its N terminus and the subdomain B of its C-terminal portion. The (cid:1) 1b -AR specifically interacted with (cid:2) 2 , but not with the (cid:2) 1 , (cid:2) 3 , or (cid:2) 4 subunits belonging to other AP com- plexes. The deletion of the (cid:2) 2 binding site in the C-tail markedly decreased agonist-induced receptor internalization as demonstrated by confocal microscopy as well as by the results of a surface receptor biotinylation assay. The direct association of dithiothreitol for 2 h at 37 °C and separated on SDS-PAGE followed by Western blotting. The biotinylated receptors were revealed by immu-noblotting using anti-HA monoclonal antibodies (Santa Cruz Biotech- nology) as described above. The intensity of the band was quantified by densitometry of films exposed in the linear range, imaged using the Molecular Imager FX (Bio-Rad), and analyzed using NIH Image soft-ware (National Institutes of Health). In control experiments, we deter- mined that the biotinylation reagent did not cross the plasma membrane, as shown by the fact that the catalytic subunit of protein kinase A, which is cytoplasmic, could not be detected in streptavidin-Sepharose precipitates (results not shown).

Desensitization to the effects of hormones and neurotransmitters is a fundamental regulatory mechanism of G proteincoupled receptor (GPCR) 1 function defined by a specific loss of responsiveness for those receptors that have been repeatedly stimulated by the agonist. Receptor desensitization results from the combination of multiple biochemical events occurring at different time frames: receptor-G protein uncoupling in response to receptor phosphorylation (seconds to minutes), internalization or endocytosis of cell surface receptors to intracellular compartments (minutes), and down-regulation of the total pool of receptors due to their decreased synthesis and/or increased degradation (hours) (1). A prominent role in homologous desensitization of GPCRs is played by G protein-coupled receptor kinases. Once the receptor is occupied by the agonist, it is recognized by the G protein-coupled receptor kinases and becomes phosphorylated (2). The subsequent uncoupling of the receptor from the G protein is then mediated by arrestin proteins, which preferentially bind to the agonist-occupied phosphorylated receptor (3). However, during the last decade, ␤-arrestins have emerged as key regulatory molecules controlling various steps of receptor desensitization. Beyond their role in physical uncoupling of GPCRs from the G proteins (3), it has been demonstrated that ␤-arrestins target GPCRs to the endocytic machinery (4). In fact, it is believed that for those GPCRs that internalize in a clathrin-dependent manner, like the ␤ 2adrenergic receptor (AR), targeting of the receptor-␤-arrestin complexes to clathrin-coated vesicles is mediated by a dual interaction of ␤-arrestin with both clathrin heavy chain and the ␤ 2 subunit of the heterotetrameric clathrin adaptor complex 2 (AP2) (5)(6)(7).
The AP2 complex directly links the clathrin coat with cargo transmembrane proteins that are sorted into coated pits and vesicles (8) and is composed of two large subunits, ␣ and ␤ 2 , of about 100 kDa and two smaller subunits, 2 and 2 , of 50 and 17 kDa, respectively (9). The AP2 adaptor can initiate endocytosis of membrane receptors by either associating directly with their cytoplasmic tail or by interacting with additional molecules, such as ␤-arrestins, as described for the ␤ 2 -AR (5,7). Direct interactions between AP2 and transmembrane proteins have been demonstrated, for example, for the transferrin receptor (10), the epidermal growth factor receptor (11,12), and the cystic fibrosis transmembrane conductance regulator (13) but not for GPCRs. They are principally mediated by the 2 subunit, which specifically associates with endocytosis signals including YXX⌽ (where ⌽ represents a bulky hydrophobic residue) (14) and dileucine motifs (15) on the cytoplasmic portion of the transmembrane proteins (8). Interestingly, recent evidence suggests that the ␤ 2 subunit also participates in the recognition of dileucine motifs on the proteins (16).
Previous reports have shown that the ␣ 1b -AR undergoes rapid endocytosis upon exposure to the agonist. This was shown both in DDT1-FM2 smooth muscle cells expressing endogenous receptors as well as in cells expressing the recombinant receptor (17,18). The molecular determinants involved in agonist-induced endocytosis of the ␣ 1b -AR reside within the C-tail of the receptor as demonstrated by the fact that truncation of this region almost completely abolished receptor desensitization and internalization (19). A previous report suggested that the ␣ 1b -AR can internalize in clathrin-coated vesicles as shown by that fact that ␣ 1b -AR endocytosis can be blocked by hypertonic sucrose and that internalized receptors colocalize with transferrin receptors (18). Moreover, we previously reported that agonist-induced internalization of the ␣ 1b -AR is, at least in part, mediated by ␤-arrestins (20). Despite this experimental evidence, our knowledge on the biochemical mechanisms and the molecular mediators controlling the clathrin-mediated endocytosis of the ␣ 1b -AR is still at an early stage.
To identify new proteins interacting with the ␣ 1b -AR that could potentially be involved in regulating receptor function, we have used the yeast two-hybrid system and identified the 2 subunit of the AP2 complex. In this study, we demonstrate that the ␣ 1b -AR and 2 subunit can directly interact through a polyarginine motif located on the C-tail of the receptor and that this interaction plays a role in agonist-induced internalization of the receptor. Our findings highlight a previously unappreciated mechanism that might also be involved in the endocytosis of other GPCRs.

EXPERIMENTAL PROCEDURES
Expression Constructs-A cDNA fragment encoding the last 165 amino acids of the hamster ␣ 1b -AR (amino acids 351-515) (Fig. 2) was PCR-amplified and subcloned at EcoRI/SalI in pGEX4T1, pET30a, and pLexA plasmids to construct fusion proteins with GST, His 6 , and LexA at the N terminus of the C-tail of the receptor, respectively. For constructing GST fusion proteins with different fragments of the C-tail, cDNA fragments encoding amino acids 351-380, 351-395, 351-425, 351-449, and 351-477 of the ␣ 1b -AR were PCR-amplified and subcloned at EcoRI/SalI in pGEX4T1. The Y386A, Y447A, L450A/L451A, and L473A/L474A mutations were introduced into the C-tail-pGEX4T1 and C-tail-pLexA plasmids by standard PCR-directed mutagenesis using the Pwo DNA polymerase (Roche Applied Science).
The full-length cDNA encoding the human 2 subunit was PCRamplified from EST 3538047 (accession number BE264960) and subcloned at HindIII-SalI in pEGFPN3 and pGEX4T1 to generate fusion proteins with GFP at the C terminus and GST at the N terminus of 2 , respectively. The full-length cDNAs encoding human 1 , 3 , and 4 subunits identified in other AP complexes were PCR-amplified from EST 4954703 (accession number BG920410), EST 720361 (accession number AA2611409), and EST 748897 (accession number AI644755), respectively, and subcloned at EcoRI/SalI in pEGFPN3 to construct fusion proteins with GFP at the C terminus. For generating fusion proteins with GFP at the C terminus, DNA fragments encoding amino acids 1-163, 164 -282, and 283-435 of 2 were PCR-amplified and subcloned at HindIII/SalI into pEGFPN3.
The full-length cDNA encoding the hamster ␣ 1b -AR (21) was PCRamplified and inserted into the pEGFPN1 using EcoRI/AgeI to fuse GFP at the C terminus of the receptor (␣ 1b -GFP). The T368-GFP and ⌬371-378-GFP receptor mutants were constructed by PCR-directed mutagenesis of the ␣ 1b -GFP using the Pwo DNA polymerase (Roche Applied Science). To construct HA-tagged receptor forms, the N-terminal fragment of the wild type ␣ 1b -AR was amplified using primers encoding the HA (YPYDVPDYA) epitope at the 5Ј end and subcloned into the pRK5 vector encoding the wild type receptor or its mutants.
Yeast Two-hybrid Screening-The yeast strain L40 was transformed with the C-tail-pLexA plasmid encoding the ␣ 1b -AR C-tail fused to LexA, and clones were selected and subsequently transformed with 250 g of a human brain Matchmaker cDNA library in the pACT2 vector (Clontech). Of 13 million double transformants, 40 exhibited moderate to strong growth on histidine-deficient plates. The library plasmids isolated from positive clones were used to cotransform the L40 strain with either the C tail-pLexA or the empty pLexA plasmid, and the specificity of the interactions was confirmed by growth on histidinedeficient plates as well as by ␤-galactosidase activity (Yeast Protocols Handbook; Clontech).
Expression and Purification of Recombinant Proteins in Bacteria-GST-tagged fusion proteins of 2 and ␣ 1b -AR C-tail were expressed using the bacterial expression vector pGEX4T1 in the BL21DE3 strain of Escherichia coli and purified. Bacterial extracts containing GST fusion proteins were prepared by centrifugation of bacterial cultures, followed by lysis of the pelleted bacteria in buffer A (20 mM Tris, pH 7.4, 50 mM NaCl, 5 mM MgCl 2 , 0.5% (w/v) Triton X-100, 1 mM benzamidine, 2 g/ml leupeptin, 2 g/ml pepstatin), sonication, and centrifugation at 38,000 ϫ g for 30 min at 4°C. After incubating the supernatants with glutathione-Sepharose beads (Amersham Biosciences) for 2 h at 4°C, the resin was washed with 10 bed volumes of buffer A and stored at 4°C. GST fusion proteins were eluted from the resin with 5 mM reduced glutathione for 15 min at room temperature, dialyzed, and stored at Ϫ20°C.
His 6 -tagged fusion protein of the ␣ 1b -AR C-tail was expressed using the bacterial expression vector pET30 in BL21DE3 bacteria and purified. Extracts containing His 6 -tagged fusion proteins were prepared by centrifugation of bacterial cultures and lysis of pelleted bacteria in buffer B (20 mM Hepes, pH 7.8, 500 mM NaCl, 10 mM imidazole, 1 mM benzamidine, 2 g/ml leupeptin, 2 g/ml pepstatin). After a 1-min sonication, the lysates were centrifuged at 38,000 ϫ g for 30 min at 4°C. The His 6 -tagged fusion proteins were purified by incubating the supernatant with nickel-nitrilotriacetic acid chelating resin (Amersham Biosciences) for 1 h at 4°C. The resin was then washed five times with 10 bed volumes of buffer B and stored at 4°C. His 6 -tagged fusion proteins were eluted from the resin with 20 mM Hepes, pH 7.8, 500 mM NaCl, 300 mM imidazole, 1 mM benzamidine, 2 g/ml leupeptin, 2 g/ml pepstatin for 1 h at room temperature, dialyzed, and stored at Ϫ20°C. The protein content of the eluates was assessed by Coomassie staining of SDS-PAGE gels.
Cell Culture and Transfections-HEK-293 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum and gentamycin (100 g/ml) and transfected at 50 -80% confluence in 35-or 100-mm dishes using the calcium-phosphate method. After transfection, cells were grown for 48 h in DMEM supplemented with 10% fetal calf serum before harvesting. The total amount of transfected DNA was of about 0.5-1 g/35-mm dish and 10 g/ 100-mm dish.
GST Pull-down and Immunoprecipitation Experiments-For GST pull-down, HEK-293 cells expressing the various constructs grown in 100-mm dishes were lysed in 1 ml of buffer C (20 mM Tris, pH 7.4, 100 mM NaCl, 5 mM EDTA, 1% (w/v) Triton X-100, 5 g/ml aprotinin, 10 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride) and centrifugation at 100,000 ϫ g for 30 min at 4°C. Glutathione-Sepharose beads coupled to the different GST fusion proteins were incubated with 1.5 mg of proteins derived from the cell lysates in a total volume of 1 ml overnight at 4°C. The beads were then washed five times with buffer C and resuspended in SDS-PAGE sample buffer. Eluted proteins were analyzed by SDS-PAGE and Western blotting.
For immunoprecipitation experiments, HEK-293 cells expressing the various HA-tagged constructs grown in 100-mm dishes were lysed in 1 ml of buffer D (20 mM Tris, pH 7.4, 100 mM NaCl, 5 mM EDTA, 1% digitonin, 5 g/ml aprotinin, 10 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride). Cell lysates were incubated overnight at 4°C on a rotating wheel. The solubilized material was centrifuged at 100,000 ϫ g for 30 min at 4°C, and the supernatant was incubated for 4 h at 4°C with 5 g of a rabbit anti-HA polyclonal antibody (Santa Cruz) or with control nonimmune IgG. After the addition of 40 l of protein A-Sepharose, the incubation was continued for 2 h at 4°C, followed by a brief centrifugation on a bench top centrifuge. The pellet was washed five times with buffer D and twice with PBS and then dissolved in SDS-PAGE sample buffer for 1 h at 37°C. Imunoprecipitated proteins were analyzed by SDS-PAGE and Western blotting.
SDS-PAGE and Western Blotting-Samples were denatured in SDS-PAGE sample buffer (65 mM Tris, 2% SDS, 5% glycerol, 5% ␤-mercaptoethanol, pH 6.8) for 1 h at room temperature, separated on 10% acrylamide gels and electroblotted onto nitrocellulose membranes. The blots were incubated in TBS-Tween (100 mM Tris, pH 7.4, 140 mM NaCl, 0.05% Tween 20) containing 5% (w/v) nonfat dry milk overnight at room temperature, washed three times with TBS-Tween, and then incubated with the specific primary antibody diluted in TBS-Tween for 2 h at room temperature. After three washes with TBS-Tween, the membranes were probed with horseradish peroxidase-conjugated secondary antimouse antibodies (Amersham Biosciences) for 1 h, washed three times with TBS-Tween, and developed using the enhanced chemiluminescence detection system (Amersham Biosciences).
Solid Phase Overlay Assay-After SDS-PAGE and electroblotting of the samples, the nitrocellulose filters were incubated with TBS-Tween containing 5% (w/v) nonfat dry milk and 1% bovine serum albumin for 1 h at room temperature and with 100,000 cpm/l of 32 P-labeled His 6tagged ␣ 1b -AR C-tail in TBS-Tween containing 5% nonfat dry milk and 0.1% bovine serum albumin for 16 h at room temperature. After extensive washes in TBS-Tween, the blots were visualized by autoradiography.
Confocal Microscopy-HEK-293 cells grown on glass coverslips were transfected with the cDNAs encoding different GFP-tagged receptors. 48 h after transfection, cells were incubated in serum-free DMEM for 1 h and treated for various times with 10 Ϫ4 M epinephrine (Sigma) at 37°C. After the treatment, cells were placed on ice, washed twice with ice-cold PBS, fixed for 10 min in PBS plus 3.7% formaldehyde, and mounted using Prolong (Molecular Probes, Inc., Eugene, OR). In the experiments measuring the effect of K44A dynamin mutant on receptor endocytosis, HEK-293 cells were cotransfected with the cDNAs encoding the GFP-tagged ␣ 1b -AR and the HA-tagged dynamin K44A mutant. 48 h after transfection, cells were incubated in serum-free DMEM for 1 h and treated for 1 h with 10 Ϫ4 M epinephrine at 37°C. After two washes with ice-cold PBS, cells were fixed for 10 min in PBS plus 3.7% formaldehyde and permeabilized for 5 min with 0.2% (w/v) Triton X-100 in PBS. Cells were incubated in PBS plus 1% bovine serum albumin for 1 h and with 1:250 dilution of anti-HA polyclonal antibody (Santa Cruz Biotechnology) for 1 h, followed by another incubation with Texas Red-conjugated donkey anti-rabbit secondary antibody (Jackson Immu-noResearch) for 1 h. The cells were then mounted using Prolong (Molecular Probes). GFP fluorescence or immunofluorescent staining were visualized on a laser-scanning confocal microscope (Zeiss).
Cell Surface Biotinylation Experiments-HEK-293 cells grown in 100-mm dishes were transfected with the cDNAs encoding the HAtagged ␣ 1b -AR or its T368 and ⌬371-378 mutants. 48 h after transfection, cells were incubated in serum-free DMEM for 1 h and treated for various times with 10 Ϫ4 M epinephrine (Sigma) at 37°C. After the incubation, cells were placed on ice and washed twice with ice-cold PBS. Surface proteins were biotinylated by incubating cells with 500 g/ml of the membrane-impermeable biotin analogue sulfo-NHS-S-S-biotin (Pierce) in PBS for 30 min at 4°C. Unreacted biotin was quenched and removed by three washes with ice-cold TBS at 4°C. Biotinylated cells were then lysed with hypo-osmotic buffer (10 mM Tris, pH 7.4, 5 mM EDTA, 5 g/ml aprotinin, 10 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride), and the cellular homogenate was centrifuged at 30,000 ϫ g for 15 min at 4°C. The pellet was resuspended in buffer D (20 mM Tris, 100 mM NaCl, 5 mM EDTA, 1% digitonin, 5 g/ml aprotinin, 10 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride), sonicated for 30 s, and stirred for 6 h at 4°C. The solubilized material was centrifuged at 100,000 ϫ g for 30 min at 4°C, and the supernatant was incubated overnight with 40 l of streptavidin-Sepharose beads (Amersham Biosciences) at 4°C. The beads were pelleted by brief centrifugation and washed five times with buffer D and twice with PBS. Biotinylated proteins were eluted by incubating the beads with 200 mM dithiothreitol for 2 h at 37°C and separated on SDS-PAGE followed by Western blotting. The biotinylated receptors were revealed by immunoblotting using anti-HA monoclonal antibodies (Santa Cruz Biotechnology) as described above. The intensity of the band was quantified by densitometry of films exposed in the linear range, imaged using the Molecular Imager FX (Bio-Rad), and analyzed using NIH Image software (National Institutes of Health). In control experiments, we determined that the biotinylation reagent did not cross the plasma membrane, as shown by the fact that the catalytic subunit of protein kinase A, which is cytoplasmic, could not be detected in streptavidin-Sepharose precipitates (results not shown).

Identification of the 2 Subunit of the AP2 Complex as a Protein
Interacting with the ␣ 1b -AR-It was previously reported that the C-tail of the ␣ 1b -AR plays an important role in the regulation of receptor desensitization and internalization (19,22,23). However, the molecular players involved in these processes are not fully characterized. To identify novel regulatory proteins interacting with the C-tail of the ␣ 1b -AR, the last 165 amino acids of the receptor were used as bait to screen a human brain cDNA library using the yeast two-hybrid system. Three independent positive clones encoding the 2 subunit of the AP2 complex were identified from 13 ϫ 10 6 cotransformants, as assessed by their ability to grow in the absence of histidine and to produce ␤-galactosidase. To confirm the interaction between the C-tail of the receptor and 2 , the L40 yeast strain was transformed with the bait plasmid expressing the C-tail of the ␣ 1b -AR (␣ 1b -AR C-tail-pLexA) in combination with either a pACT2 vector containing the 2 cDNA ( 2 -pACT2) or with empty pACT2 vector. As shown in Fig. 1A, yeast transformed with 2 -pACT2 were able to grow in the absence of histidine (upper panel) and to produce ␤-galactosidase (lower panel), whereas those transformed with the empty pACT2 vector did not.
To provide biochemical evidence that the ␣ 1b -AR can associate with the 2 subunit of the AP2 complex, pull-down experiments were performed by incubating the GST-tagged C-tail of the receptor with cell extracts of HEK-293 cells. We found that the 2 endogenously expressed in HEK-293 cells specifically bound to GST-C-tail but not to GST alone (Fig. 1B).
To assess whether the association between the ␣ 1b -AR and the 2 subunit of the AP2 complex occurs through a direct interaction or is mediated through another protein, we monitored the ability of purified GST-2 to associate with the purified His 6 -tagged C-tail using a solid phase overlay assay. The autoradiography shown in Fig. 1C (left panel) indicates that the C-tail of the 32 P-labeled His 6 -tagged C-tail of the ␣ 1b -AR could specifically interact with GST-2 , but not with GST alone. Control experiments showed that the GST-2 did not interact with 100 nM of an unrelated radiolabeled protein (regulatory subunit of protein kinase A) (results not shown). Altogether, these results strongly suggest that the 2 subunit of the AP2 complex can directly interact with the C-tail of the ␣ 1b -AR.
Identification of the Binding Site for 2 on the ␣ 1b -AR-The 2 subunit of the AP2 complex has been shown to recognize endocytosis signals including tyrosine-based motifs (YXX⌽) and dileucine motifs on the cytoplasmic portion of membrane receptors (9). Analysis of the primary sequence of the C-tail of the ␣ 1b -AR revealed the presence of two potential tyrosinebased motifs at positions 386 and 442 and of two potential dileucine motifs at positions 450 and 473 (Fig. 2). Moreover, two additional YXX⌽ sequences can be found at positions 144 and 153 in the second intracellular loop of the receptor (Fig. 2). To assess whether these motifs can mediate the interaction between the ␣ 1b -AR and 2 , we generated GST-C-tail fusion proteins in which tyrosines 386 and 442 as well as the leucine doublets 450 -451 and 473-474 were individually substituted by alanines and assessed their ability to interact with 2 endogenously expressed in HEK-293 cells. Surprisingly, none of these mutations was able to disrupt the interaction between the C-tail of the receptor and 2 (Fig. 3A). Similar results were obtained using the yeast two-hybrid system as an interaction assay (Fig. 3B). To investigate whether the two YXX⌽ sequences located at positions 144 and 153 of the ␣ 1b -AR were involved in the interaction with 2 , we also constructed a GST fusion protein, including the second intracellular loop of the receptor. This fusion protein was not able to interact with the 2 from HEK-293 cell extracts (results not shown). Altogether, these results suggest that the structural determinants of the ␣ 1b -AR mediating its interaction with 2 are different from the canonical tyrosine-based or dileucine motifs.
To further investigate the binding site for 2 on the C-tail of the ␣ 1b -AR, we fused to GST a series of fragments of the C-tail carrying progressive truncations (Fig. 2) and assessed their ability to interact with 2 endogenously expressed in HEK-293 cells. As shown in Fig. 3C, whereas different fusion constructs truncated up to residue 380 could interact with 2 , the GST-T368 fusion construct did not, thus suggesting that the region included between residues 380 and 368 is crucial for this interaction. Interestingly, the deletion of eight arginines at positions 371-378 completely abolished the binding of the C-tail to 2 , suggesting that the ␣ 1b -AR interacts with the 2 subunit of the AP2 complex through a novel arginine-based binding domain (Fig. 2).
Identification of the Binding Site for the ␣ 1b -AR on 2 -Recently, the crystal structure of the 2 subunit of the AP2 complex has been solved (12,24). The first 157 residues of the protein are organized in a predominantly ␣-helical structure, whereas the C-terminal fragment of 2 is largely composed of ␤-sheet structures that are folded into two subdomains (Fig.  4A) (12). The first subdomain, comprising residues 158 -282, contains the binding site for tyrosine-based endocytic motifs, whereas the second subdomain, comprising residues 283-435, contains an interaction site for synaptotagmin, a neuronal AP2-binding protein involved in synaptic vesicles exocytosis (25).
To identify the region within 2 that interacts with the C-tail of the ␣ 1b -AR, we tagged with GFP the three fragments of the 2 , 1-157, 158 -282, and 283-435, and expressed them in HEK-293 cells. Pull-down experiments were performed by incubating the GST-C-tail construct with cell extracts overexpressing different GFP-tagged 2 fragments (Fig. 4A). We The C-tail-pLexA plasmid was used to transform the L40 yeast strain in combination with either empty pACT2 vector or with a pACT2 containing the 2 cDNA. Double transformants were plated on Trp Ϫ , Leu Ϫ , His Ϫ plates to select for histidine prototrophy (upper panel). The His ϩ positive clones encoding 2 were restreaked on selective medium and assayed for ␤-galactosidase activity by a filter assay (lower panel). Quantitative analysis of the interaction between the C-tail and 2 was performed using the liquid ␤-galactosidase assay. B, GST pull-down assay. Extracts from HEK-293 cells expressing endogenous 2 were incubated with glutathione-Sepharose beads coupled to GST alone or to the GST-C-tail. The 2 eluted from the beads was detected by Western blotting using anti-2 monoclonal antibodies (left panel). The amount of extract loaded on the gel was 10% of that used for the pull-down. A control protein staining indicating the expression level of the GST and GST-tagged C-tail used in the pull-down assay is shown (right panel). The results are representative of three independent experiments. C, solid phase overlay assay. GST or GST-2 proteins were separated on SDS-PAGE and electroblotted onto nitrocellulose membranes. The membrane was incubated with 100 nM of 32 P-labeled His 6  found that the 1-157 and 282-435 fragments retained the ability to bind the C-tail of the ␣ 1b -AR, whereas the 158 -257 fragment did not (Fig. 4B). These findings led us to conclude that the molecular determinants of 2 involved in the recognition of the polyarginine motif on the C-tail of the ␣ 1b -AR are located on two domains at the N and C terminus, respectively, of the 2 molecule. Interestingly, these domains are distinct from the 2 region that binds tyrosine-based internalization signals, which is located between residues 157 and 282 (12).
The ␣ 1b -AR/ 2 Interaction Occurs in the Cells and Is Regulated by Agonist-induced Receptor Activation-To demonstrate that the ␣ 1b -AR and 2 can form a complex inside the cells, we performed coimmunoprecipitation experiments from HEK-293 cells transiently expressing the HA-tagged ␣ 1b -AR. After immunoprecipitating the receptor using polyclonal anti-HA antibodies, monoclonal anti-HA as well as anti-2 antibodies were used to immunoblot the immunoprecipitated samples. The Western blots revealed that the 2 endogenously expressed in HEK-293 cells could specifically co-immunoprecipitate with the ␣ 1b -AR, whereas no bands were immunoprecipitated by IgG (Fig. 5A, top panel, lanes 2 and 3). These findings demonstrate that beyond their ability to interact in vitro, the ␣ 1b -AR and 2 can associate inside the cells.
To confirm that the ␣ 1b -AR/ 2 interaction in intact cells was mediated by the polyarginine motif identified above as the binding site for 2 , coimmunoprecipitation experiments were performed from HEK-293 cells overexpressing the HA-tagged receptor mutants truncated at residue 368 (T368) or carrying a deletion of the polyarginine motif (⌬371-378). The T368 and ⌬371-378 receptor mutants displayed pharmacological properties similar those of the wild type ␣ 1b -AR (results not shown). In addition, as previously shown (26), tagging the wild type or mutated receptors with HA or GFP at their N and C terminus, respectively, did not affect the pharmacological properties of the receptor (results not shown). As shown in Fig. 5A, the endogenous 2 did not coimmunoprecipitate either with the T368 or with the ⌬371-378 receptor mutants, thus suggesting that the polyarginine stretch between residues 371 and 378 represents the only binding site of the ␣ 1b -AR for 2 .
To investigate whether the whole heterotetrameric AP2 complex could interact with the ␣ 1b -AR, we determined whether additional subunits of the AP2 complex could be coimmunoprecipitated with the receptor. As shown in Fig. 5B, the ␣ and ␤ 2 subunits endogenously expressed in HEK-293 cells were coimmunoprecipitated with the wild type receptor but not with the ⌬371-378 receptor mutant, thus suggesting that the whole AP2 complex can associate with the ␣ 1b -AR through the interaction mediated by its 2 subunit.
To assess whether the ␣ 1b -AR/ 2 interaction could be modulated by the agonist-induced activation of the receptor, HEK-293 cells expressing the HA-tagged ␣ 1b -AR were incubated for 15 min in the absence or presence of 10 Ϫ4 M epinephrine prior to immunoprecipitation of the receptor. As shown in Fig. 6, treatment with epinephrine induced a 2-fold increase in the amount of endogenous 2 coimmunoprecipitated with receptor when compared with untreated cells (Fig. 6, A (lanes 6 and 8) and B). This strongly suggests that the ␣ 1b -AR/ 2 interaction is dynamically regulated by the agonist-induced activation of the ␣ 1b -AR, which might increase the amount of 2 associated with the receptor.
Specificity of the ␣ 1b -AR/ 2 Interaction-So far, four different adaptor protein complexes (APs) involved in sorting of membrane proteins have been identified and characterized (9). AP1 is involved in the formation of clathrin-coated vesicles from the trans-Golgi network and the trafficking of proteins from the trans-Golgi network to the plasma membrane (8). AP2 plays a role in the clathrin-mediated endocytosis of plasma membrane receptors (8). AP3 has been shown to mediate the sorting of proteins form early endosomes to lysosomes (27). Finally, AP4 has been shown to participate in the polarized transport of proteins to the basolateral membrane in Madin-Darby canine kidney cells (28). To assess whether the ␣ 1b -AR preferentially associates with the AP2 as compared with other AP complexes, we performed coimmunoprecipitation experiments from HEK-293 expressing the HA-tagged ␣ 1b -AR and GFP-tagged 1 , 2 , 3 , and 4 subunits. Interestingly, Western blotting using anti-GFP antibodies indicated that, whereas 2 could be coimmu- FIG. 3. Mapping the binding site for 2 on the ␣ 1b -AR. A, extracts from HEK-293 cells expressing endogenous 2 were incubated with glutathione-Sepharose beads coupled to GST alone or to GSTtagged C-tail fragments carrying point mutations of the YXX⌽ and dileucine motifs. The 2 eluted from the beads was detected by Western blotting using anti-2 monoclonal antibodies. The amount of extract loaded on the gel was 10% of that used for the pull-down. A control protein staining indicated that the expression level of each GST-tagged C-tail fragment was similar in the different samples (results not shown). The results are representative of three independent experiments. B, empty pLexA vector or pLexA encoding different C-tail constructs was used to transform the L40 yeast strain in combination with the pACT2 vector containing the 2 cDNA. Quantitative analysis of the interaction between the various C-tail constructs and 2 was performed using the liquid ␤-galactosidase assays. C, extracts from HEK-293 cells expressing endogenous 2 were incubated with glutathione-Sepharose beads coupled to GST alone or to GST-tagged C-tail fragments carrying different truncations as indicated in Fig. 2. The 2 eluted from the beads was detected as indicated in A. The amount of extract loaded on the gel was 10% of that used for the pull-down. A control protein staining indicated that the expression level of each GST-tagged C-tail fragment was similar in the different samples (results not shown). Results are representative of three independent experiments. IB, immunoblot. noprecipitated with the HA-tagged ␣ 1b -AR, 1 , 3 , and 4 could not (results not shown). This suggests that, inside the cells, the ␣ 1b -AR specifically associates with the 2 subunit of the AP2 complex rather than with other AP complexes.
The ␣ 1b -AR/ 2 Interaction Is Involved in Clathrin-mediated Endocytosis of the Receptor-To confirm that the ␣ 1b -AR can  1-3) or of its mutants T368 (lanes 3-6) and ⌬371-378 (lanes 7-9). Receptor expression was 1-2 pmol/mg of protein for all receptors. Cell lysates were subjected to immunoprecipitation with the nonimmune IgG or anti-HA polyclonal antibodies. Western blots of the immunoprecipitates and of the cell extracts (Ex.) were revealed using anti-2 monoclonal antibodies (upper panel) to detect the endogenous 2 or anti-HA monoclonal antibodies to detect the recombinant receptors (lower panel). The amount of extract loaded on the gel was 10% of that used for the immunoprecipitation. B, HEK-293 cells were transfected with the cDNAs encoding the HA-tagged forms of the ␣ 1b -AR (lanes 1-3) or of its mutant ⌬371-378 (lanes 3-6). Western blots of the immunoprecipitates and of the cell extracts (Ex.) were revealed using anti-␣ adaptin (upper panel), anti-␤ 2 adaptin (middle panel), or anti-HA (lower panel) monoclonal antibodies to detect the endogenous ␣ and ␤ 2 subunits of the AP2 complex and the recombinant receptors, respectively. One of the two bands recognized by the anti-␣ adaptin antibody in the cell extract that is not coimmunoprecipitated might be nonspecific. The results are representative of three independent experiments. IB, immunoblot. internalize in clathrin-coated vesicles, we assessed whether hypertonic sucrose and the overexpression of a dominant negative mutant of dynamin (K44A) could inhibit the agonistinduced internalization of the GFP-tagged ␣ 1b -AR transiently expressed in HEK-293 cells. Incubation of cells with 0.45 M sucrose for 15 min prior to stimulation with epinephrine completely inhibited agonist-induced internalization (Fig. 7, C and  D). Similarly, overexpression of the K44A dynamin mutant impaired receptor endocytosis when compared with control cells (Fig. 7, E and F). These results support the notion that in HEK-293 cells the ␣ 1b -AR undergoes internalization through clathrin-mediated endocytosis.
It is well established that the AP2 complex controls the early steps of membrane receptor endocytosis, allowing clathrin to be recruited to the receptor. Since the ␣ 1b -AR represents the first GPCR to be shown to directly interact with AP2, and since this interaction seems to occur through a noncanonical 2 binding motif on the receptor, we sought to establish whether the ␣ 1b -AR/ 2 interaction might play a role in the regulation of receptor endocytosis.
To test this hypothesis, we determined whether deleting the 2 binding site on the C-tail of the ␣ 1b -AR could affect agonistinduced receptor internalization. The GFP-tagged forms of the wild type ␣ 1b -AR (WT-GFP) and of its mutants T368 (T368-GFP) and ⌬371-378 (⌬371-378-GFP) were transiently expressed in HEK-293 cells and tested for their ability to undergo agonist-induced internalization. Cells expressing the different GFP-tagged receptors were treated with 10 Ϫ4 M epinephrine for 15, 30, and 60 min, and agonist-induced internalization was assessed by monitoring receptor redistribution by confocal microscopy. As shown in Fig. 8 (upper panels), incubation of cells expressing the WT-GFP receptor with epinephrine caused a rapid redistribution of the receptor from the cell surface to intracellular compartments. Internalization was already detectable 15 min after agonist exposure, whereas a more pronounced redistribution was observed after 60 min. The T368-GFP receptor was completely impaired in its ability to undergo internalization (Fig. 8, middle panels) in agreement with our previous findings indicating that the integrity of the C-tail is required for receptor endocytosis and desensitization (19). Interestingly, epinephrine-induced endocytosis of the ⌬371-378-GFP receptor mutant was delayed as compared with that of the WT-GFP receptor (Fig. 8, lower panels). In fact, internalization was barely detectable after 30 min of exposure to the agonist, and at 60 min a significant portion of the receptor was still present at the cell surface. Altogether, these results strongly suggest that the ␣ 1b -AR/ 2 interaction is involved in receptor internalization, since receptor mutants lacking the binding site for 2 are clearly impaired in receptor endocytosis.
To quantify receptor endocytosis, we used a biotinylation assay to selectively label the receptors expressed at the cell surface. HEK-293 cells expressing HA-tagged forms of the wild type ␣ 1b -AR (HA-WT) and of its mutants T368 (HA-T368) and ⌬371-378 (HA-⌬371-378) were exposed to 10 Ϫ4 M epinephrine for various times and subsequently incubated with a membrane-impermeant biotinylation reagent (sulfo-NHS-biotin) (see "Experimental Procedures"). Biotinylated cell surface receptors were precipitated using streptavidin-Sepharose beads and detected by Western blotting using anti-HA antibodies. The results of this biochemical assay indicated that the HA-WT, HA-T368, and HA-⌬371-378 receptors were expressed at similar levels at the cell surface (Fig. 9, lanes 1, 5, and 9). HA-tagged receptors could not be detected in streptavidin-Sepharose precipitates when cells were not incubated with the biotinylation reagent, thus confirming that, in our experimen- tal conditions, only biotinylated receptors could be isolated (results not shown).
Treatment of cells expressing the HA-WT receptor with 10 Ϫ4 M epinephrine induced a rapid receptor endocytosis, which was 23% after 15 min, 61% after 30 min, and 80% after 60 min (Fig.  9, lanes 1-4). As expected, the HA-T368 receptor mutant was resistant to agonist-induced internalization as demonstrated by the fact that 94% of the receptors were still present at the cell surface after 60 min of exposure to epinephrine (Fig. 9, lanes [5][6][7][8]. In agreement with the results of confocal imaging (Fig. 8), epinephrine-induced internalization of the HA-⌬371-378 receptor mutant was significantly decreased compared with that of the wild type receptor with 60% of the receptor remaining at the cell surface after a 60-min exposure to the agonist (Fig. 9, lanes 9 -12). Altogether, these findings support the hypothesis that agonist-induced internalization of the ␣ 1b -AR is controlled, at least in part, by the interaction of 2 with the C-tail of the receptor. DISCUSSION This study provides the evidence that the AP2 complex can associate with a G protein-coupled receptor, the ␣ 1b -AR, through the direct interaction of its 2 subunit with the receptor. This represents, to our knowledge, the first report of a direct interaction between the AP2 complex and a GPCR. Previous studies have shown that the AP2 complex is implicated in the agonist-induced endocytosis of the ␤ 2 -AR (5-7). However, in this case, the AP2 complex seemed to be recruited to the receptor through ␤-arrestins which can directly interact with the ␤ 2 subunit of AP2.
One unexpected finding of our study was the observation that 2 recognizes a noncanonical binding motif on the ␣ 1b -AR, which consists of eight consecutive arginines included between residues 371 and 378 of the C-tail of the receptor (Fig. 3). Deletion of this motif completely abolished the ␣ 1b -AR/ 2 interaction inside the cells measured in the coimmunoprecipitation experiments (Fig. 5). Interestingly, the deletion of the 2 binding motif also abolished the interaction of the ␣ 1b -AR with the ␣ and ␤ subunits of the AP2 complex, suggesting that this motif represents the only point of contact between the AP2 complex and the receptor. The polyarginine motif of the ␣ 1b -AR is reminiscent of the binding site for 2 previously identified in synaptotagmin, a neuronal AP2 binding proteins involved in recycling of synaptic vesicles, which contains six positively charged residues (KRLKKKK) (29).
The recent publication of the crystal structure of the entire AP2 complex has provided a better understanding of how the different subunits of the complex contact each other and of how the 2 subunit interacts with tyrosine-based endocytic motifs located on membrane receptor (24). Whereas the YXXF motifs have been shown to bind to a hydrophobic pocket located in the subdomain A (residues 157-282) of 2 , recent studies indicate that the positively charged motif of synaptotagmin contacts the subdomain B (residues 283-435) of 2 (25). Similarly, we could show that the subdomain B (residues 283-435) together with the N-terminal region of 2 (residues 1-157) participate in the interaction with the polyarginine motif of the ␣ 1b -AR (Fig. 4). Therefore, it appears that the N-terminal region and the subdomain B of 2 might provide a docking surface for positively charged motifs exposed on the cytoplasmic face of membrane receptors. A systematic scanning mutagenesis of these domains will be required to determine whether the ␣ 1b -AR and synaptotagmin bind to similar structural determinants of the 2 .
The ␣ 1b -AR/ 2 interaction is dynamically regulated as demonstrated by the fact that activation of the receptor by the agonist increases the association of 2 to the receptor by 2-fold (Fig. 6). This suggests that 2 preferentially recognizes the agonist-occupied form of the ␣ 1b -AR. One can speculate that binding of epinephrine to the ␣ 1b -AR might trigger a conformational change that exposes the polyarginine motif, thus promoting the association of the receptor with the AP2 complex. A similar model has been proposed for the epidermal growth factor receptor-mediated recruitment of the AP2 complex in which the receptor exposes a high affinity binding site for 2 only upon activation by epidermal growth factor (11,12).
An important finding of our study is that the interaction of the ␣ 1b -AR with 2 of the AP2 complex plays a role in receptor endocytosis. This was mainly demonstrated by the fact that deleting the 2 binding motif in the C-tail of the ␣ 1b -AR markedly decreased agonist-induced receptor internalization as shown by confocal microscopy (Fig. 8) as well as by surface receptor biotinylation (Fig. 9). Hypertonic sucrose as well as the dominant negative dynamin mutant K44A almost completely inhibited epinephrine-induced endocytosis of the GFPtagged ␣ 1b -AR, thus supporting the notion that the ␣ 1b -AR can internalize in clathrin-coated vesicles (Fig. 7). Since the AP2 complex links the clathrin coat to transmembrane proteins sorted into coated pits, the ␣ 1b -AR/ 2 interaction might represent a mechanism directly involved in targeting the ␣ 1b -AR to clathrin-coated vesicles.
In a previous study, we reported that agonist-induced internalization of the ␣ 1b -AR is, at least in part, mediated by ␤-arrestins. This was mainly demonstrated by two observations: (a) the stimulation of the ␣ 1b -AR with epinephrine induced a marked translocation to the cell surface of ␤-arrestin; (b) a dominant negative mutant of ␤-arrestin 1 (V53D) decreased  1-4) and of its mutants T368 (lanes 5-8) and ⌬371-378 (lanes 9 -12). After treatment with 10 Ϫ4 M epinephrine for the indicated times, cells were incubated with 500 M NHS-S-S-biotin for 30 min at 4°C. Surface biotinylated receptors were precipitated using streptavidin-Sepharose beads, separated on SDS-PAGE, electroblotted onto nitrocellulose membranes, and detected by Western blotting using anti-HA monoclonal antibodies. Western blot analysis performed on cell lysates using anti-HA monoclonal antibodies indicated that the receptor expression levels were similar in the different samples (results not shown). Results are representative of three independent experiments. B, densitometry of the bands corresponding to the biotinylated receptors precipitated by the streptavidin-Sepharose beads was performed as indicated under "Experimental Procedures." Results are the mean Ϯ S.E. of three independent experiments.