A Conserved Motif for the Transport of G Protein-coupled Receptors from the Endoplasmic Reticulum to the Cell Surface*

The structural determinants for the export trafficking of G protein-coupled receptors are poorly defined. In this report, we determined the role of carboxyl termini (CTs) of (cid:1) 2B -adrenergic receptor (AR) and angiotensin II type 1A receptor (AT1R) in their transport from the endoplasmic reticulum (ER) to the cell surface. The (cid:1) 2B -AR and AT1R mutants lacking the CTs were com- pletely unable to transport to the cell surface and were trapped in the ER. Alanine-scanning mutagenesis revealed that residues Phe 436 and Ile 443 -Leu 444 in the CT were required for (cid:1) 2B -AR export. Insertion or deletion between Phe 436 and Ile 443 -Leu 444 as well as Ile 443 -Leu 444 mutation to FF severely disrupted (cid:1) 2B -AR transport, in- dicating there is a defined spatial requirement, which is essential for their function as a single motif regulating receptor transport from the ER. Furthermore, the car-boxyl-terminally truncated as well as Phe 436 and Ile 443 Leu 444 mutants were unable to bind ligand and the (cid:1) 2B -AR CT conferred its transport properties to the AT1R mutant without the CT in a Phe 436 -Ile 443 -Leu 444 dependent manner. These data suggest that the Phe 436 and Ile 443 -Leu 444 may be involved in both proper folding and export from the ER of the receptor. Similarly, residues Phe 309 and Leu

G protein-coupled receptors constitute a superfamily of membrane proteins that respond to a vast array of sensory and chemical stimuli and share a common topology with a hydrophobic core of seven transmembrane-spanning ␣-helices within the plasma membrane, three intracellular loops, and three extracellular loops with amino terminus outside the cell and carboxyl terminus (CT) 1 inside the cell (1,2). Although the seven transmembrane hydrophobic domains, the amino-terminal tail, and/or extracellular domains of the receptors likely are important for ligand recognition, the CT and intracellular loops are involved in the regulation of G protein coupling, phosphorylation, internalization, and trafficking (3)(4)(5)(6)(7)(8).
To precisely perform their functions, the newly synthesized G protein-coupled receptors transport from the endoplasmic reticulum (ER), where they are synthesized, folded, and assembled, to the plasma membrane, the functional destination. The receptors at the plasma membrane may undergo internalization upon stimulation with agonists. The internalized receptors are then transported to the lysosomes for degradation or recycled back to the plasma membrane. Therefore, intracellular trafficking of G protein-coupled receptors is a highly regulated and dynamic process. Because most studies of receptor trafficking have focused on the events involved in receptor internalization, recycling, and degradation (9,10), the molecular mechanisms underlying receptor export from the ER and subsequent transport to the cell surface and regulation of receptor signaling by these processes remain poorly understood.
ER exportation represents the first step of intracellular trafficking of the receptors and, most certainly, regulates receptor expression at the cell surface and function. Indeed, export from the ER is the rate-limiting step for ␦ opioid receptor expression at the cell surface (11). Several studies have indicated that G protein-coupled receptor transport to the cell surface requires specific chaperones/escort proteins (12)(13)(14)(15) or dimerization (16,17). Furthermore, the intracellular domains of G protein-coupled receptors play an important role in the transport processes from the ER to the cell surface (18 -22). For example, the carboxyl termini (CTs) of rhodopsin, vasopressin (V2), luteinizing hormone/choriogonadotropin, A1 adenosine, and dopamine D1 receptors are required for their exit from the ER (18 -22). However, structural determinants involved in the transport of G protein-coupled receptors from the ER to the cell surface remain poorly defined.
Our previous studies have demonstrated that, whereas ␣ 2Badrenergic receptor (AR) transport from the ER to the cell surface is independent of Rab1, angiotensin II type 1A receptor (AT1R) and ␤ 2 -AR transport to the cell surface is dependent on Rab1 (23). Rab1 is a Ras-related small GTPase involved in antegrade protein transport specifically from the ER to the Golgi and between Golgi compartments (24,25). These data indicate that different G protein-coupled receptors may use distinct pathways for their export trafficking. The ␣ 2B -AR, ␤ 2 -AR, and AT1R couple to different G proteins and differ in their glycosylation. Whereas ␣ 2B -AR couples to G i/o proteins and does not contain glycosylation signals, AT1R and ␤ 2 -AR mainly couple to G s and G q proteins, respectively, and have putative N-linked glycosylation sites at the amino termini (26 -28). In an effort to define structural determinants involved in the selection of specific transport pathway in individual receptors, we determined the role of intracellular domains of ␣ 2B -AR and AT1R in their trafficking from the ER to the cell surface, as the intracellular domains of ␣ 2B -AR and AT1R differ considerably in sizes. Whereas ␣ 2B -AR contains a larger third intracellular loop and a small CT, AT1R contains a small third intracellular loop and a larger CT (28,29).
We report here that the CTs of ␣ 2B -AR and AT1R play an obligatory role in receptor export from the ER. We have identified a novel motif consisting of a phenylalanine and double leucine spaced by 6 residues (F(X) 6 LL) in the CTs of ␣ 2B -AR and AT1R that is required for the exit of the receptors from the ER. This motif is highly conserved in the membrane-proximal CTs of many G protein-coupled receptors and may provide a common mechanism directing the transport of the receptors from the ER to the cell surface.
For generation of chimeric receptor AT1␣ 2 R in which the AT1R CT was substituted with the ␣ 2B -AR CT, two complementary oligonucleotides (5Ј-TCGACGCAGGACTTCCGCCGTGCCTTTCGAAGGATCCTTT-GCCGGCCGTGGACCCAGACTGGCTGG-3Ј and 5Ј-GATCCCAGCCA-GTCTGGGTCCACGGCCGGCAAAGGATCCTTCGAAAGGCACGGCG-GAAGTCCTGCG-3Ј) coding the ␣ 2B -AR CT (Gln 434 -Trp 453 ) and carrying the sticky ends of SalI and BamHI were annealed and ligated into the AT1R-ct-GFP in the pEGFP-N1, which was cleaved with SalI and BamHI. A similar strategy was employed to construct the chimeric receptor AT1␣ 2m R in which the AT1R CT was replaced with mutated ␣ 2B -AR CT (Gln 434 -Trp 453 ) in which Phe 436 , Ile 443 , and Leu 444 were mutated to alanines (sense, 5Ј-TCGACGCAGGACGCCCGCCGTGCCT-TTCGAAGGGCCGCTTGCCGGCCGTGGACCCAGACTGGCTGG-3Ј; antisense, GATCCCAGCCAGTCTGGGTCCACGGCCGGCAAGCGGC-CCTTCGAAAGGCACGGCGGGCGTCCTGCG-3Ј). Therefore, the resulting chimera contained three extra residues (AST) between the AT1R-ct and the ␣ 2B -AR CT. Receptor mutants were generated using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) and GFP-tagged receptors as templates. The sequence of each construct used in this study was verified by restriction mapping and nucleotide sequence analysis (Louisiana State University Health Sciences Center DNA Sequence Core).
Cell Culture and Transient Transfection-HEK293T cells were cultured in Dulbecco's modified Eagle's medium with 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. Transient transfection of the HEK293T cells was carried out using Lipo-fectAMINE 2000 reagent (Invitrogen) as described previously (23). For measurement of total and cell-surface receptor expression and ERK1/2 activation, HEK293T cells were cultured on 6-well dishes. For each transfection, 0.8 g of GFP-or HA-tagged receptors were diluted into 125 l of serum-free Opti-MEM in a tube. In another tube, 2 l of LipofectAMINE was diluted into 125 l of serum-free Opti-MEM. The two solutions were combined and mixed gently. After incubation for 20 min at room temperature, the mixture was added to each well containing 750 l of fresh Dulbecco's modified Eagle's medium. For radioligand binding studies, HEK293T cells were cultured on 100-mm dishes and transfected with 8 g of GFP-tagged receptor. The transfection efficiency was estimated to be about 80% based on the GFP fluorescence.
Flow Cytometric Analysis of Receptor Expression-For measurement of total receptor expression, HEK293T cells transfected with GFPtagged receptors were collected and resuspended in phosphate-buffered saline (PBS) containing 1% fetal calf serum at a density of 8 ϫ 10 6 cells/ml. Total receptor expression was determined by measuring total GFP fluorescence on a flow cytometer (BD Biosciences FACSCalibur) as described previously (23). For measurement of receptor expression at the cell surface, HEK293T cells were transfected with HA-tagged receptors. The transfected cells were collected, suspended in PBS containing 1% fetal calf serum at a density of 4 ϫ 10 6 cells/ml, and incubated with high affinity fluorescein conjugated anti-HA antibody 3F10 at a final concentration of 2 g/ml for 30 min at 4°C. After washing twice with 0.5 ml of PBS and 1% fetal calf serum, the cells were resuspended, and the fluorescence was analyzed as described above. Because the staining with the anti-HA antibodies was carried out in the unpermeabilized cells and only those receptors expressed at the cell surface were accessible to the anti-HA antibodies, the fluorescence measurement reflected the amount of AT1R expressed at the cell surface.
Radioligand Binding-Membrane preparation and radioligand binding studies were carried out essentially as described (33). Briefly, HEK293T cells were cultured on 100-mm dishes and transiently transfected with ␣ 2B -AR WT or mutant plasmids as described above. The cells were washed twice with PBS, harvested, and homogenized in 3 ml per plate of buffer containing 5 mM Tris-HCl, pH 7.4, 5 mM EGTA and 5 mM EDTA supplemented with Complete Mini protease inhibitor mixture (Roche Applied Science, Mannheim, Germany). After centrifugation at 100,000 ϫ g for 30 min, the pellet was resuspended in 300 l per plate of membrane buffer containing 50 mM Tris-HCl, pH 7.4, 0.6 mM EDTA, and 5 mM MgCl 2 . The membrane suspension (15 g of membrane protein) was incubated with increasing concentrations of [ 3 H]RX-821002 (1.25-160 nM) in a total volume of 100 l. Nonspecific binding was determined in the presence of the selective ␣ 2B -AR antagonist rauwolscine (10 M). Duplicate samples were incubated for 30 min at room temperature with constant shaking, and the reaction was terminated by vacuum filtration. After washing with 100 mM Tris-HCl, pH 7.4 (4 ϫ 4 ml), the retained radioactivity was measured by liquid scintillation spectrometry in 8 ml of Ecoscint A scintillation solution (National Diagnostics, Inc., Atlanta, GA). The B max values were calculated using PRISM software (GraphPad, San Diego, CA).
Immunofluorescence Microscopy-HEK293T cells were grown on coverslips and fixed with 4% paraformaldehyde-4% sucrose mixture in PBS for 15 min. For localization of receptor expression, the cells were stained with 4,6-diamidino-2-phenylindole for 5 min. The coverslips were mounted, and fluorescence was detected with a Leica DMRA2 epifluorescent microscope. Images were deconvolved using SlideBook software and the nearest neighbors deconvolution algorithm (Intelligent Imaging Innovations, Denver, CO). For co-localization of the receptor with calregulin, an ER marker, the cells were permeabilized with PBS containing 0.2% Triton X-100 for 5 min, and blocked with 5% normal donkey serum for 1 h. The cells were then incubated with antibodies against calregulin for 1 h. After washing with PBS (3 ϫ 5 min), the cells were incubated with Alexa Fluor 594-labled secondary antibody (1:2000 dilution) for 1 h at room temperature, and the fluorescence was analyzed as described above.
Measurement of ERK1/2 Activation-HEK293T cells were cultured on 6-well dishes and transfected as described above. At 36 -48 h after transient transfection, HEK293T cells were starved for at least 3 h and then stimulated with individual agonists as indicated in the figure legends. Stimulation was terminated by addition of 1 ϫ SDS gel loading buffer as described previously (23). After solubilizing the cells, 20 l of total cell lysates was separated by 12% SDS-PAGE and ERK1/2 activation was determined by Western blotting by measuring the levels of phosphorylation of ERK1/2 with phosphospecific ERK1/2 antibodies. The membranes were stripped and reprobed with anti-ERK2 antibodies to determine the total amount of ERK2 and to confirm equal loading of proteins.
Immunoblotting-Western blotting was carried out as described previously (23). HEK293T cell lysates were separated by SDS-PAGE and transferred onto polyvinylidene difluoride membranes. The signal was detected using ECL Plus (PerkinElmer Life Sciences) and a Fuji Film luminescent image analyzer (LAS-1000 Plus).
Statistical Analysis-Differences were evaluated using Student's t test, and p Ͻ 0.05 was considered as statistically significant. Data are expressed as the mean Ϯ S.E.

A Requirement of the Carboxyl Termini of ␣ 2B -AR and AT1R
for the Transport from the ER to the Cell Surface-We have recently demonstrated that the transport from the ER to the cell surface of ␣ 2B -AR and AT1R is mediated through distinct pathways (23). In an attempt to define structural determinants for the selection of specific transport pathways, we determined the role of the CTs of ␣ 2B -AR and AT1R in their transport from the ER to the cell surface. We first determined the effect of deletion of the carboxyl-terminal 20 (Gln 434 -Trp 453 ) and 63 (Leu 297 -Glu 359 ) amino acid residues from ␣ 2B -AR and AT1R, respectively (Fig. 1A), on receptor expression at the cell surface. ␣ 2B -AR WT, AT1R WT, and their mutants lacking the CTs (␣ 2B -AR-ct and AT1R-ct) were conjugated with the HA at the amino termini (HA-␣ 2B -AR, HA-␣ 2B -AR-ct, HA-AT1R, and HA-AT1R-ct) or with the GFP at the carboxyl termini (␣ 2B -AR-GFP, ␣ 2B -AR-ct-GFP, AT1R-GFP, and AT1R-ct-GFP), and transiently expressed in HEK293T cells. Total receptor expression was quantitated in cells expressing ␣ 2B -AR-GFP, ␣ 2B -ARct-GFP, AT1R-GFP, or AT1R-ct-GFP by flow cytometry measuring total GFP fluorescence. Cell-surface receptor expression was measured in cells expressing HA-␣ 2B -AR, HA-␣ 2B -AR-ct, HA-AT1R, or HA-AT1R-ct following staining with HA antibodies. Total expression levels of ␣ 2B -AR-ct-GFP and AT1R-ct-GFP were almost the same as compared with ␣ 2B -AR-GFP and AT1R-GFP, respectively (Fig. 1B). In contrast, cell-surface expression levels of HA-␣ 2B -AR-ct and HA-AT1R-ct were markedly reduced by 96 and 97%, respectively, compared with their HA-tagged WT receptors (Fig. 1B). These data indicated that the CTs of ␣ 2B -AR and AT1R are required for their expression at the cell surface.
In the second series of experiments, we determined the functional consequence of deleting the CTs of ␣ 2B -AR and AT1R by measuring ERK1/2 activation in response to agonist stimulation in HEK293T cells. ERK1/2 activation was stimulated by FIG. 1. Effect of the deletion of the carboxyl termini of ␣ 2B -AR and AT1R on receptor transport to the cell surface. A, the amino acid sequences of the CTs of ␣ 2B -AR and AT1R. B, effect of the deletion of the CTs of ␣ 2B -AR and AT1R on total and cell-surface expression of the receptors. To measure total receptor expression, HEK293T cells were transfected with GFP-conjugated receptors, and total receptor expression was measured by GFP fluorescence using flow cytometry.
The mean values of fluorescence obtained from the untransfected cells and from the cells transfected with ␣ 2B -AR-GFP, ␣ 2B -AR-ct-GFP, AT1R-GFP, or AT1R-ct-GFP were 41 Ϯ 5, 478 Ϯ 15, 456 Ϯ 22, 526 Ϯ 23, and 487 Ϯ 39, respectively. To measure receptor expression at the cell surface, HEK293T cells were transfected with HA-conjugated receptors, and receptor expression at the cell surface was quantitated by flow cytometry following incubation with anti-HA antibodies as described in the "Experimental Procedures." The mean values of fluorescence obtained from the untransfected cells and from the cells transfected with HA-␣ 2B -AR, HA-␣ 2B -AR-ct, HA-AT1R, or HA-AT1R-ct were 88 Ϯ 7, 720 Ϯ 15, 113 Ϯ 15, 742 Ϯ 43, and 108 Ϯ 13, respectively. The data shown are percentages of the mean value obtained from the cells transfected with WT receptors and are presented as the mean Ϯ S.E. of three experiments. *, p Ͻ 0.05 versus the cells transfected with respective WT receptors. C, effect of the deletion of the CTs of ␣ 2B -AR and AT1R on ERK1/2 activation. HEK293T cells cultured on 6-well dishes were transfected with ␣ 2B -AR-GFP, ␣ 2B -AR-ct-GFP, AT1R-GFP, or AT1R-ct-GFP. At 8 -12 h after transfection, the cells were split and cultured for an additional 24 h. The cells transfected with ␣ 2B -AR-GFP or ␣ 2B -ARct-GFP were then simulated with increasing concentrations of UK14304 (10 -1000 nM) for 5 min. The cells transfected with AT1R-GFP or AT1R-ct-GFP were stimulated with increasing concentrations of Ang II (1-1000 nM) for 2 min. ERK1/2 activation was determined by Western blot analysis using phosphospecific ERK1/2 antibodies. Upper panel, representative blots of ERK1/2 activation; lower panel, total ERK2 expression. Similar results were obtained in at least four separate experiments. D, effect of the deletion of the CTs of ␣ 2B -AR and AT1R on receptor subcellular distribution. HEK293T cells cultured on coverslips were transfected with GFP-conjugated receptors, and subcellular distribution of the receptors was revealed by detecting GFP fluorescence as described under "Experimental Procedures." The data are representative images of at least five independent experiments. Blue, DNA staining by 4,6-diamidino-2-phenylindole (nuclear); green, GFP-conjugated receptors. Scale bar, 10 m.
UK14304 or Ang II in a dose-dependent manner in cells transiently transfected with ␣ 2B -AR-GFP or AT1R-GFP, respectively (Fig. 1C). In contrast, ERK1/2 activation by the agonists was completely lost in cells transfected with ␣ 2B -AR-ct-GFP or AT1R-ct-GFP (Fig. 1C). These data are consistent with the lack of cell-surface expression of ␣ 2B -AR-ct and AT1R-ct as measured by flow cytometry.
We then sought to visualize the differential localization of WT and mutated receptors. ␣ 2B -AR-GFP, ␣ 2B -AR-ct-GFP, AT1R-GFP, or AT1R-ct-GFP was transiently expressed in HEK293T cells, and their subcellular distribution at steady state was revealed by microscopy. As anticipated, ␣ 2B -AR and AT1R were mainly localized at the cell surface (Fig. 1D), which was confirmed by co-localization with tetramethylrhodamineconjugated concanavalin A, a plasma membrane marker (not shown). In contrast, ␣ 2B -AR-ct and AT1R-ct completely lost their cell-surface expression pattern exhibiting a marked localization to the perinuclear region of the transfected cells (Fig.  1D). The ␣ 2B -AR-ct and AT1R-ct were extensively co-localized with calregulin, an ER marker (Fig. 2). These data indicate that the CTs of ␣ 2B -AR and AT1R play an obligatory role in receptor transport from the ER to the cell surface.
Identification of Amino Acid Residues Required for the Exit of ␣ 2B -AR from the ER-To identify key amino acid residues required for receptor transport from the ER to the cell surface, we first focused on the ␣ 2B -AR CT, because it is relatively shorter. We generated a series of mutants in which each amino acid residues in the ␣ 2B -AR CT ( 432 FNQDFRRAFRRILCRP-WTQTGW 453 ) was substituted with alanine individually or in combination (Fig. 3A). ␣ 2B -AR WT and mutants conjugated with GFP at their CTs were transiently expressed in HEK293T cells and their subcellular localization was revealed by microscopy. Mutation of the residues Phe 432 , Gln 434 , Asp 435 , Arg 437 Arg 438 , Phe 440 , Arg 441 Arg 442 , Ile 443 , Leu 444 , Cys 445 , Arg 446 , Pro 447 , Trp 448 , Thr 449 -Thr 451 , Gln 450 , or Trp 453 did not significantly influence the subcellular localization of ␣ 2B -AR (Fig. 3B). In contrast, single mutation of Phe 436 abolished cellsurface expression of ␣ 2B -AR resulting in substantial ER accumulation (Fig. 3B). Because the LL motif together with an upstream acidic residue (E) have been demonstrated to be essential for cell-surface expression of the vasopressin V2 receptor (34), we determined the effect of double mutation of Ile 443 -Leu 444 on ␣ 2B -AR transport. Our results indicated that, although individual mutation of Ile 443 and Leu 444 had no obvious effect on receptor export from the ER, double mutation of Ile 443 -Leu 444 clearly inhibited ␣ 2B -AR transport from the ER (Fig. 3B). Furthermore, triple mutation of Phe 436 , Ile 443 , and Leu 444 completely blocked ␣ 2B -AR transport out of the ER to the cell surface (Fig. 3B).
We then compared radioligand binding abilities of ␣ 2B -AR WT and mutants utilizing the selective ␣ 2B -AR antagonist [ 3 H]RX-821002. Radioligand binding results indicated that the membrane fraction prepared from the cells transfected with ␣ 2B -AR-GFP bound to [ 3 H]RX-821002 similar to that of nontagged ␣ 2B -AR. In contrast, the membrane fractions prepared from the cells transfected with mutants ␣ 2B -AR-ct, Phe 436 , Ile 443 -Leu 444 , or Phe 436 -Ile 443 -Leu 444 were completely unable to bind the ligand (Fig. 3C). These data suggest that residues Phe 436 -Ile 443 -Leu 444 may be involved in correct folding of ␣ 2B -AR. Furthermore, cells expressing the mutant Phe 436 -Ile 443 -Leu 444 were unable to stimulate ERK1/2 activation (Fig. 3D), consistent with the lack of cell-surface expression of the receptor (Fig. 3C).
Spatial Requirement of the Phe 436 and Ile 443 -Leu 444 in the Motif F(X) 6 IL-To determine if Phe 436 and Ile 443 -Leu 444 function as a single unit or as independent signals regulating ␣ 2B -AR transport, we sought to characterize the spatial relationship between Phe 436 and Ile 443 -Leu 444 . This was accomplished by evaluating the effects of insertion and deletion of residues between Phe 436 and Ile 443 -Leu 444 (Fig. 4A) on the subcellular localization of ␣ 2B -AR. Insertion of one alanine or two alanines between Phe 436 and Ile 443 -Leu 444 in the motif F(x) 6 IL severely disrupted ␣ 2B -AR transport from the ER. This was determined by analyzing their subcellular distribution (Fig. 4B) and further confirmed in radioligand [ 3 H]RX-821002 binding of membrane fractions prepared from the cells expressing mutated receptors (Fig. 4C). Similar to insertion, deletion of one residue (Ala 439 ) or two residues (Arg 438 -Ala 439 ) between Phe 436 and Ile 443 -Leu 444 also completely blocked receptor transport from the ER to the cell surface (Fig. 4, A and B). These data indicate that there is a defined spatial relationship between Phe 436 and Ile 443 -Leu 444 in the ␣ 2B -AR, which is required for their function as a motif regulating receptor transport from the ER to the cell surface.
The ␣ 2B -AR CT Confers Its Transport Properties to the AT1R-ct-Preceding data have demonstrated that the motif Phe 436 -(X) 6 -Ile 443 -Leu 444 identified in the ␣ 2B -AR may be involved in the receptor folding in the ER. To determine if the motif Phe 436 -(X) 6 -Ile 443 -Leu 444 could also function as an ER export motif directing receptor export from the ER, we determined if the ␣ 2B -AR CT could induce the ER-to-cell surface transport of other proteins, which are normally retained in the ER. We took advantage of AT1 mutant lacking the CT (AT1Rct), which is unable to exit from the ER and determined if the ␣ 2B -AR CT when conjugated with the truncated AT1R would reconstitute cell-surface expression. We generated a chimeric receptor AT1␣ 2 R in which the AT1R CT was substituted with the ␣ 2B -AR CT (Gln 434 -Trp 453 ) containing the motif Phe 436 -Ile 443 -Leu 444 . Similar to AT1R WT, AT1␣ 2 R transported to the cell surface (Fig. 5). These data suggest that the ␣ 2B -AR CT may contain an ER-export motif and confer its transport properties to the AT1R-ct.
To further determine if the transport properties of ␣ 2B -AR CT were exclusively dependent on the motif Phe 436 -(X) 6 -Ile 443 -Leu 444 , we generated another chimeric receptor AT1␣ 2m R in which the AT1R CT was substituted with a mutated ␣ 2B -AR CT in which Phe 436 -Ile 443 -Leu 444 were mutated to three alanines (Ala 436 -Ala 443 -Ala 444 ). In contrast to AT1␣ 2 R expressing at the cell surface, AT1␣ 2m R was unable to transport to the cell surface and was retained in the ER (Fig. 5). These data suggest that the Phe 436 -(X) 6 -Ile 443 -Leu 444 may function independently as a motif sufficiently directing receptor transport from the ER to the cell surface.
Effect of the Mutation of IL to FF on ␣ 2B -AR Transport-To define if the hydrophobicity or other properties of Ile 443 -Leu 444 were critical for its function, we determined the effect of mutation of Ile 443 -Leu 444 to FF on ␣ 2B -AR transport. The mutation of Ile 443 -Leu 444 to FF significantly attenuated ␣ 2B -AR transport from the ER to the cell surface as measured by subcellular localization, radioligand binding, and ERK1/2 activation (Fig.  6). These data indicate that the function of Ile 443 -Leu 444 in the ␣ 2B -AR transport is determined by not only its hydrophobicity but also other properties.
The double phenylalanine (FF) functions as an ER export motif in the endoplasmic reticulum-Golgi intermediate compartment-53 (ERGIC-53) and p24 receptor proteins (35,36). We then determined if the FF could function as an independent motif facilitating the transport of ␣ 2B -AR from the ER to the cell surface. Therefore, Phe 436 and Ile 443 -Leu 444 were simultaneously mutated to A and FF (AFF), respectively. Similar to the mutation of Phe 436 to Ala, mutation of Phe 436 to Ala, and Ile 443 -Leu 444 to FF resulted in a complete disruption of receptor export from the ER to the cell surface (Fig. 6). These data suggest that the Phe 443 -Phe 444 could not sufficiently transport ␣ 2B -AR out of the ER.
Identification of Amino Acid Residues Required for the Exit of AT1R from the ER-Having demonstrated that the CTs of both ␣ 2B -AR and AT1R are required for their transport from the ER to the cell surface and having further identified the motif for ␣ 2B -AR exit from the ER, we sought to identify the residues required for AT1R transport from the ER to the cell surface. Because the AT1R CT is substantially longer than the ␣ 2B -AR CT, we progressively truncated the AT1R CT to define a subdomain in the CT required for AT1R exit from the ER. Therefore, in addition to the AT1R-ct (AT1R1-296) lacking carboxylterminal 63 residues, we generated three more AT1R constructs, AT1R1-349, AT1R1-320, and AT1R1-306, lacking 10 (Lys 350 -Glu 359 ), 39 (His 321 -Glu 359 ), and 53 (Lys 307 -Glu 359 ) residues, respectively (Fig. 7A). Subcellular distribution of each construct and their abilities to stimulate ERK1/2 activation were then assessed. Similar to AT1R WT, AT1R1-349 transported to the cell surface and stimulated ERK1/2 activation (Fig. 7, B and C). AT1R1-320 also expressed at the cell surface (Fig. 7B) and markedly increased ERK1/2 activation compared with AT1R WT (Fig. 7C). This enhanced activation of ERK1/2 by AT1R1-320 is probably due to the reduced ability of AT1R1-320 to desensitize caused by removal of phosphorylation sites of the receptor (37). AT1R1-306, however, failed to transport to the cell surface and stimulate ERK1/2 activation (Fig. 7, B and  C), similar to results obtained with the CT-deficient construct AT1R1-296. These data identified the subdomain from Lys 307 to Ile 320 as necessary for AT1R transport from the ER to the cell surface.
Alanine scanning mutagenesis was then used to identify residues within the subdomain Lys 307 -Ile 320 required for AT1R transport from the ER (Fig. 8A). Mutation of residues Lys 307 -Lys 308 , Lys 310 -Lys 311 , Tyr 312 , Phe 313 , Leu 314 , Gln 315 , Leu 316 , Leu 317 , Lys 318 , Tyr 319 , or Ile 320 did not significantly alter the subcellular distribution of AT1R (Fig. 8B). Interestingly, similar to the results obtained in ␣ 2B -AR (Fig. 3), mutation of the single residue Phe 309 and the two residues Leu 316 -Leu 317 profoundly blocked AT1R transport out of the ER (Fig. 8B). Moreover, the mutant Phe 309 -Leu 316 -Leu 317 was completely unable to transport to the cell surface and activate ERK1/2 (Fig. 8, B and C). These data indicate that, similar to the motif Phe 436 -(X) 6 -Ile 443 -Leu 444 identified in the ␣ 2B -AR, Phe 309 and Leu 316 -Leu 317 may consist of a motif Phe 309 -(X) 6 -Leu 316 -Leu 317 mediating the export of AT1R from the ER to the cell surface.
The Motif F(X) 6 LL Is Conserved in the CTs of G Proteincoupled Receptors-Because we have demonstrated that ␣ 2B -AR and AT1R use a similar motif for their export from the ER to the cell surface, we searched the G protein-coupled receptor data base to see if this motif is conserved in the superfamily of G protein-coupled receptors. As shown in Fig.  9, the sequence F(X) 6 LL (where X represents any amino acid FIG. 5. The ␣ 2B -AR CT confers its transport properties to the AT1R-ct. GFP-tagged WT and mutant receptors were transiently expressed in HEK293T cells and their subcellular distribution was revealed as described under "Experimental Procedures." The data are representative images of three independent experiments. AT1R, AT1R WT; AT1R-ct, AT1R lacking the carboxyl-terminal 63 residues; AT1␣ 2 R, chimeric receptor in which the AT1R CT was substituted with the ␣ 2B -AR CT; AT1␣ 2m R, chimeric receptor in which the AT1R CT was substituted with the mutated ␣ 2B -AR CT in which Phe 436 -Ile 443 -Leu 444 were mutated to alanines. Scale bar, 10 m. residues and L is leucine or isoleucine) is highly conserved in the membrane-proximal CTs of many G protein-coupled receptors. In the family A of G protein-coupled receptors occurring amongst species, 240 receptors out of 941 total (25.5%) have the motif F(X) 6 LL in the membrane-proximal carboxyl termini.

DISCUSSION
The structural determinants involved in the transport of G protein-coupled receptors from the ER to the cell surface remain poorly defined. In the present study, we have identified that Phe 436 and Leu 443 -Leu 444 in the ␣ 2B -AR and Phe 309 and Leu 316 -Leu 317 in the AT1R sufficiently directed their transport from the ER to the cell surface. Most importantly, the motif F(X) 6 LL is conserved in many G protein-coupled receptors, and thus may provide a common molecular mechanism underlying the ER export of a subgroup of G proteincoupled receptors.
We first demonstrated that ␣ 2B -AR and AT1R mutants lacking the CTs (␣ 2B -AR-ct and AT1R-ct) were unable to transport to the cell surface, as quantified by flow cytometry and radioligand binding. Consistent with the lack of receptor cell-surface expression, the ␣ 2B -AR-ct and AT1R-ct were also unable to stimulate ERK1/2 activation. However, we cannot exclude the possibility that deletion of the CTs of ␣ 2B -AR and AT1R may impair receptor coupling to downstream signaling molecules that activate the mitogen-activated protein kinase pathway (38). The ␣ 2B -AR-ct and AT1R-ct were extensively co-localized with calregulin, an ER marker, suggesting that the CTs of ␣ 2B -AR and AT1R may contain signals necessary for the export of the receptors out of the ER. These data are consistent with several other studies demonstrating that the CTs of G proteincoupled receptors are required for receptor expression at the cell surface (19 -22).
We then used an alanine-scanning mutagenesis approach to identify amino acid residues required for receptor transport. The membrane-proximal portions of the receptors are highly hydrophobic, positively charged, and structurally predicted to form an amphipathic helix. The entire CT of ␣ 2B -AR ( 432 FN-QDFRRAFRRILCRPWTQTGW 453 ) contains seven hydrophobic (3 Phe, 2 Trp, 1 Leu, and 1 Ile) and five positively charged residues (Arg). The membrane-proximal subdomain of AT1R ( 304 FLGKKFKKYFLQLLKYI 320 ), which is required for receptor exit from the ER, as identified by a strategy of progressive deletion of the CT, contains 10 hydrophobic (3 Phe, 2 Tyr, 4 Leu, and 1 Ile) and five positively charged residues (Lys). However, only mutation at Phe 436 and Ile 443 -Leu 444 in the ␣ 2B -AR and Phe 309 and Leu 316 -Leu 317 in the AT1R abrogated receptor ER export to the cell surface as evaluated by receptor subcellular localization, quantitation of receptor cell-surface expression, and receptor-mediated ERK1/2 activation. Mutation of other residues to alanines in these regions did not significantly influence receptor subcellular distribution. These results strongly indicated that the F(X) 6 LL is a motif in the ␣ 2B -AR and AT1R essential for their transport from the ER to the cell surface.
The LL motif functions as sorting signals at the trans-Golgi network for basolateral cell-surface transport and at the plasma membrane for endocytosis in clathrin-coated vesicles through interacting directly with clathrin adaptor protein complex (39,40). However, its role in regulating G protein-coupled receptor exit from the ER has been controversial. Mutation of the LL motif in the ␤ 2 -AR CT altered receptor internalization, but not receptor expression at the cell surface (41). In contrast, Schulein et al. demonstrated LL together with an upstream acidic residue (E) are essential for cell-surface expression of the vasopressin V2 receptor, possibly through modifying receptor folding (34). Consistent with Schulein's data, we demonstrated that the dileucine motif in the CTs of ␣ 2B -AR and AT1R is required for their transport from the ER and that mutation of the dileucine motif abolished ligand binding of ␣ 2B -AR, suggesting that the dileucine motif may be involved in correct folding of the receptors in the ER. Together, these data strongly indicate that the highly conserved dileucine motif in the proximal-membrane carboxyl termini of many G protein-coupled receptors play an important role in receptor export. However, in contrast to an inhibitory effect on the vasopressin V2 receptor transport observed by Schulein et al., mutation of individual L had no significant effect on the transport of either ␣ 2B -AR or AT1R. These results indicate the LL motif may differentially regulate the transport of the G protein-coupled receptors. Furthermore, we demonstrated that a Phe preceding the LL motif in both ␣ 2B -AR and AT1R, rather then an acidic residue as shown in the vasopressin V2 receptor, is required for receptor export form the ER. Importantly, the LL and the Phe spaced by six resides are highly conserved in many of the G proteincoupled receptors.
In addition to regulating receptor folding in the ER, the motif F(X) 6 LL may function as an independent transport signal. If the motif F(X) 6 LL functions as an independent transport signal mediating receptor exit from the ER, its transport properties should be transferable to other proteins normally expressed in the ER. The transmission of transport capabilities to normally ER-retained proteins have been demonstrated for other ER export motifs (22,42). We demonstrated that the ␣ 2B -AR CT could efficiently transport ER-retained AT1R-ct to the cell surface. Furthermore, this transport ability was lost when Phe 436 and Ile 443 -Leu 444 were mutated to alanines. These data suggest that the Phe 436 and Ile 443 -Leu 444 in the ␣ 2B -AR may function independently as an ER transport signal. However, we cannot completely exclude the possibility that the Phe 436 and/or Ile 443 -Leu 444 in the ␣ 2B -AR CT may be involved in the folding process of the chimeric receptor in the ER, as both ␣ 2B -AR and AT1R use a similar motif for their export from the ER. Whether the motif could confer its transport properties to other proteins besides G protein-coupled receptors is now under investigation.
Do the Phe and the LL in the F(X) 6 LL function as one motif or separately modulating receptor transport? Is it possible that the Phe in the motif may regulate receptor exit from the ER as demonstrated for the motif FF in the ERGIC-53 and p24 proteins (34,35) and the motif FXXXFXXXF in the dopamine D1 receptor (22), whereas the LL may be involved in the regulation of correct folding/assembly of the receptors as demonstrated in the vasopressin V2 receptor (41)? To address this issue, we determined the effect of altering the spatial relationship of Phe 436 and Ile 443 -Leu 444 on the ␣ 2B -AR ER export. Our data demonstrated that insertion or deletion of one or two residues between Phe 436 and Ile 443 -Leu 444 severely interrupted the ␣ 2B -AR transport, similar to the effects observed upon deletion of the entire CT or mutation of the Phe and/or the IL to alanines and indicated that the six residue spacing of Phe and LL is required for receptor transport. Because the membraneproximal portions of the ␣ 2B -AR as well as many other GPCRs are predicted to form an amphipathic ␣-helix, one could imagine that Phe 436 and Ile 443 -Leu 444 project from the same face of the ␣-helix. These data indicate that Phe 436 and Ile 443 -Leu 444 function as a single motif with a defined spatial relationship between Phe 436 and Ile 443 -Leu 444 , which is required for its function, mediating receptor transport.
The ER motif F(X) 6 LL identified in the ␣ 2B -AR and AT1R is distinct from the other ER export motifs identified thus far and FIG. 8. Site-directed mutagenesis identifies key residues required for AT1R transport from the ER to the cell surface. A, the sequence of the subdomain required for AT1R exit from the ER. Each residue in the subdomain was mutated to alanine individually or in combination. B, AT1R WT and mutants were transiently expressed in HEK293T cells, and their subcellular distribution was revealed by fluorescence microscopy as described under "Experimental Procedures." The data shown are representative images of three independent experiments. Scale bar, 10 m. C, ERK1/2 activation by Ang II in the cells transiently transfected with AT1R WT or mutant Phe 309 -Leu 316 -Leu 317 as determined in the legend of Fig. 1. Upper panel, a representative blot of ERK1/2 activation; lower panel, total ERK2 expression. Similar results were obtained in at least three separate experiments.
represents a novel motif for the transport of G protein-coupled receptors from the ER to the cell surface. Several recent studies have revealed two classes of ER export signals in the cytoplasmic carboxyl terminus of a variety of membrane proteins. A diacidic motif, DXE, is found in the vesicular stomatitis viral glycoprotein, a type I transmembrane protein. The DXE motif directs the concentration of the cargo molecule during export from the ER and enhances the rate of its exit from the ER. The motif DXE also transfers its transport properties to sufficiently direct the export of other proteins, which are normally retained in the ER (42,43). A similar acidic motif was also found in the potassium channels (44). A second class of signal, the FF, is required for efficient transport of the p24 family proteins and ERGIC-53 from the ER to the Golgi (34,35,45,46). Mutation of the FF to other hydrophobic residues such as LL and II did not significantly influence its transport function (46). In contrast, mutation of Ile 443 -Leu 444 to FF in the ␣ 2B -AR significantly disrupted receptor transport as measured by subcellular localization, radioligand binding and ERK1/2 activation. Although Ile, Leu, and Phe are hydrophobic, they have distinct physicochemical properties. Whereas Ile and Leu are aliphatic, Phe is aromatic. Our data indicate that the function of Ile 443 -Leu 444 in the ␣ 2B -AR transport is determined by its hydrophobicity as well as other properties, which cannot be substituted by FF. In addition, simultaneous mutation of Phe 436 to Ala and Ile 443 -Leu 444 to FF resulted in a receptor unable to exit from the ER, similar to the single mutation of Phe 436 to Ala. These data suggest that FF could not override the impact induced by Phe 436 mutation and function as an independent motif mediating ␣ 2B -AR export from the ER. These data also suggest that the function of Ile 443 -Leu 444 in the motif F(X) 6 IL is mediated through a mechanism that is distinct from that utilized by the motif FF in the ERGIC-53 and p24 receptor proteins.
Our previous studies have indicated that AT1R transport to the cell surface is dependent on Rab1, whereas ␣ 2B -AR transport to the cell surface is independent of Rab1, indicating that the transport from the ER to the cell surface of ␣ 2B -AR and AT1R is mediated through distinct pathways (23). In this report, we identified a similar ER export motif in the ␣ 2B -AR and AT1R, implying that the exit from the ER of ␣ 2B -AR and AT1R may be directed by the same motif-mediated mechanism. Therefore, selection of specific transport pathways for ␣ 2B -AR and AT1R is not exerted in the ER export motif. Selection of distinct transport pathways for individual receptors may be regulated at late stages of transport process such as budding of transport vesicles from the ER or transport from the ER to the cell surface. In addition, ␣ 2B -AR and AT1R as well as other G protein-coupled receptors carrying a similar ER export motif F(X) 6 LL may be sorted from other G protein-coupled receptors with different ER exit motifs. Therefore, selective transport of G protein-coupled receptors from the ER to the cell surface can be achieved at multiple transport steps.
The precise molecular mechanism by which the F(X) 6 LL motif regulates receptor export from the ER is unknown. The ER exit motifs are decoded by physically interacting with the components of COP II vesicles (47,48). The residues F and LL with a strictly constrained spatial relationship may provide a specific interactive site mediating the interaction of the receptors with selective proteins to facilitate their transport from the ER to the cell surface. The F(X) 6 LL motif may also provide a docking site for machinery proteins involved in receptor folding to the state competent for transport out of the ER. Therefore, it will be important to identify the proteins selectively interacting with the motif F(X) 6 LL in G protein-coupled receptors.
Normal protein transport is required for maintaining cell homeostasis. Indeed, defective protein transport is associated with pathogenesis of a variety of human diseases (49,50). Considering that numerous naturally occurring mutations leading to ER retention of G protein-coupled receptors have been implicated in inherited diseases such as retinitis pigmentosa (rhodopsin), male pseudohermaphroditism (luteinizing hormone receptor), and nephrogenic diabetes insipidus (V2-vasopressin receptor) (51,52), further characterization of the molecular mechanisms of the motif-facilitated ER export may be used as a foundation for development of therapeutic strategies targeting receptor transport from the ER to the cell surface.