D1 and D2 Dopamine Receptor Expression Is Regulated by Direct Interaction with the Chaperone Protein Calnexin*

As for all proteins, G protein-coupled receptors (GPCRs) undergo synthesis and maturation within the endoplasmic reticulum (ER). The mechanisms involved in the biogenesis and trafficking of GPCRs from the ER to the cell surface are poorly understood, but they may involve interactions with other proteins. We have now identified the ER chaperone protein calnexin as an interacting protein for both D1 and D2 dopamine receptors. These protein-protein interactions were confirmed using Western blot analysis and co-immunoprecipitation experiments. To determine the influence of calnexin on receptor expression, we conducted assays in HEK293T cells using a variety of calnexin-modifying conditions. Inhibition of glycosylation either through receptor mutations or treatments with glycosylation inhibitors partially blocks the interactions with calnexin with a resulting decrease in cell surface receptor expression. Confocal fluorescence microscopy reveals the accumulation of D1-green fluorescent protein and D2-yellow fluorescent protein receptors within internal stores following treatment with calnexin inhibitors. Overexpression of calnexin also results in a marked decrease in both D1 and D2 receptor expression. This is likely because of an increase in ER retention because confocal microscopy revealed intracellular clustering of dopamine receptors that were co-localized with an ER marker protein. Additionally, we show that calnexin interacts with the receptors via two distinct mechanisms, glycan-dependent and glycan-independent, which may underlie the multiple effects (ER retention and surface trafficking) of calnexin on receptor expression. Our data suggest that optimal receptor-calnexin interactions critically regulate D1 and D2 receptor trafficking and expression at the cell surface, a mechanism likely to be of importance for many GPCRs.

erotrimeric G proteins, thereby regulating downstream signaling events. It is estimated that nearly 5% of the human genome consists of GPCRs, and that more than 70% of the current drugs on the market are targeted to these essential receptor proteins (1). Clearly, GPCRs have an established importance in a vast variety of physiological and pathological conditions; but despite the observed significance of GPCR regulation in a variety of diseases, little is known about the specific mechanisms associated with their regulation in terms of cell surface expression. Overall expression of these receptors on the cellular surface likely involves a dynamic balance between surface trafficking, internalization, and degradation. Although significant progress has been made in understanding the internalization mechanisms of GPCRs once they are expressed on the cell surface (2), surprisingly little information is available concerning the mechanisms of GPCR biogenesis and trafficking to the plasma membrane.
Recent studies have suggested that GPCRs are synthesized, assembled, and folded in a conformation-dependent sorting paradigm within the endoplasmic reticulum (ER) (3). This allows for their continued trafficking to the cell surface as functional protein receptors (4). Surface trafficking of newly formed receptors, working in harmony with the more extensively studied endocytotic pathway, determines receptor expression levels and thereby influences the magnitude of the cellular response (for review see Ref. 5). However, the ER processing steps for GPCRs are poorly understood, as is the subsequent regulation of receptor trafficking to the cell surface. It is beginning to emerge that a complex system of regulation exists at the level of the ER that represents critical steps in the trafficking and expression of GPCRs.
Dopamine, a critically important neurotransmitter in the central nervous system and periphery, exerts its biological effects through activating a distinct family of GPCRs. These include five separate receptor proteins (D 1 -D 5 ), with D 1 and D 2 representing the most abundant subtypes (6,7). It has now become clear that dopamine receptors, as well as most GPCRs, exist as components of preformed signaling complexes that can consist of G proteins, enzymes, scaffolding proteins, and other receptors (8 -10). Indeed, a number of dopamine receptor-interacting proteins have now been identified, some of which may affect receptor trafficking and expression or may be altered in neuropsychiatric disorders (11,12).
As part of our efforts to identify novel dopamine receptorinteracting proteins, we have employed receptor immunoprecipitation (IP) assays along with mass spectrometry. By using this approach, we found that both D 1 and D 2 dopamine receptors interact with the ER chaperone protein calnexin, a transmembrane protein that resides in the ER where it acts as a chaperone for nascent and newly synthesized glycoproteins. Calnexin is involved in protein folding and quality control, which includes retaining misfolded and incomplete proteins in the ER (13)(14)(15). If proper folding is achieved, the glycoprotein will be allowed to leave the ER when the association with the lectin is terminated. If not in the proper conformation, the glycoprotein will be retained to ensure folding efficiency and prevent misfolded glycoproteins from exiting the ER (16). We report here that the extent of calnexin-receptor interactions is critical to ensure proper trafficking of D 1 and D 2 receptors to the cell surface. Our results suggest that calnexin may represent a critically important ER chaperone protein that is involved in regulating the expression of many GPCRs.

EXPERIMENTAL PROCEDURES
Materials-HEK293-tsa201 (HEK293T) cells (17) (31.4 Ci/mmol) and the cAMP assay kits were obtained from Diagnostic Products Corp. (Los Angeles, CA). Cell culture reagents, NuPAGE gels, and gel buffers were purchased from Invitrogen. PNGase F was purchased from New England Biolabs (Boston, MA). All other drugs and buffer components were purchased from Sigma.
Cell Culture and Transfection-HEK293T cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum, 1 mM sodium pyruvate, 50 g/ml streptomycin, 50 units/ml penicillin, and 10 g/ml gentamycin. HEK293T cells were transfected via a calcium phosphate precipitation kit (Clontech) according to the manufacturer's instructions. Cath.a-differentiated (CAD) cells were obtained as a gift from Dr. Dona Chikaraishi and grown in DMEM/F-12 supplemented with 8% fetal calf serum, 100 g/ml streptomycin, and 100 units/ml penicillin. Cells were grown at 37°C in 5% CO 2 and 90% humidity. CAD cells were transferred into serum-free differentiation media consisting of DMEM/F-12 supplemented with 20 g/ml transferrin (Sigma), 50 ng/ml sodium selenite, 100 g/ml streptomycin, and 100 units/ml penicillin immediately prior to transfection. CAD cells were transfected using NeuroPORTER reagent (Gene Therapy Systems, San Diego, CA) according to the manufacturer's instructions.
Several expression constructs were used in this study. An amino-terminal FLAG epitope-tagged construct for the rat D 1 receptor (18) was created and named pSFD 1 , as reported previously by our laboratory (19); pCMV-Tag2B, an expression construct expressing only the FLAG peptide (FLAG tag), was obtained from Stratagene (La Jolla, CA). Rat D 1 -GFP in a pEGFP-N1 vector was a gift from Dr. Qun-yong Zhou (20). The D 2L receptor expression construct containing the FLAG epitope in pSR-D2L was created by our laboratory as described previously (21). Two mutants of the above-mentioned constructs were also made for these studies. The D 1 -FLAG construct was mutated at positions 4 and 174 (N4Q and N174Q), resulting in D 1 -GLYT-FLAG, and the D 2 -FLAG construct was  mutated at positions 5, 17, 23, and 175 (N5Q, N17Q, N23Q, and  N175Q), resulting in D 2 -GLYT-FLAG. Mutations were made using QuikChange multisite-directed mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's instructions. These mutations were made to eliminate the N-linked glycosylation sites. The human D 2S -YFP construct was a gift from Dr. Jonathan Javitch. The calnexin expression construct containing canine calnexin in pcDNA3 was a gift from the laboratory of Dr. David Thomas. pDsRed2-ER vector encodes a fusion protein consisting of red fluorescent protein with the ER targeting sequence of calreticulin fused to the 5Ј end of DsRed2 and the ER retention sequence KDEL fused to the 3Ј end of DsRed2. This is used to allow for fluorescent visualization of the endoplasmic reticulum and was purchased from BD Biosciences.
Immunoprecipitation and Gel Electrophoresis-Cells were removed from culture flasks and collected by centrifugation (300 ϫ g). They were then resuspended in 1 ml of solubilization buffer (50 mM HEPES, 1 mM EDTA, 10% glycerol, 1% Triton X-100, 150 mM NaCl, 50 mM NaF, 40 mM sodium pyrophosphate, and Complete-Mini (Roche Applied Science) protease inhibitor mixture) and incubated on ice for 1 h. The lysates were then centrifuged at 30,000 ϫ g for 30 min to remove insoluble cell debris and then precleared via incubation with protein G-agarose for 3 h. The agarose was removed by centrifugation (15,000 ϫ g, 5 min) and discarded. Anti-FLAG M2-agarose (Sigma) was added and incubated with the lysate overnight at 4°C. The beads were collected via centrifugation and washed three times by resuspension and recentrifugation in solubilization buffer. The agarose was then subjected to a final wash in 1ϫ TE buffer, pH 7.4. Proteins were eluted from the beads using NuPAGE-lithium dodecyl sulfate sample buffer (Invitrogen). Agarose was removed via centrifugation, and proteins were separated on 4 -12% BisTris NuPAGE gel (Invitrogen) according to the manufacturer's instructions.
Western Blotting and Gel Staining-Proteins separated by PAGE were transferred onto polyvinylidene difluoride membranes. Membranes were blocked in Superblock (Pierce) prior to incubation with the primary antibody. Primary antibodies used in this study include the following: rat monoclonal anti-D 1 dopamine receptor (clone 1-1-F11 S.E6, catalog number D-187, Sigma), rabbit polyclonal anti-calnexin (catalog number SPA-860, StressGen, Victoria, British Columbia, Canada), and rabbit polyclonal anti-D2L/S (catalog number AB5084P, Chemicon, Temecula, CA). After washing, membranes were incubated with HRP-conjugated secondary antibodies (Jackson Immu-noResearch, West Grove, PA) (anti-rat for the D 1 and antirabbit for D 2 and calnexin). Proteins were visualized via the SuperSignal West Pico chemiluminescent substrate (Pierce) according to the manufacturer's instructions, and images were recorded on film. In some experiments, blots were stripped of all antibodies using Restore Western blot stripping buffer (Pierce) according to the manufacturer's instructions. Removal of all primary antibodies by the buffer was initially checked via incubation with HRP-secondary antibodies followed by ECL and film exposure. In other experiments where total protein was examined, gels were stained using either Coomassie Blue or the SilverQuest staining kit (Invitrogen).
Mass Spectrometry-The identification of PAGE-resolved proteins by mass spectroscopy was carried out by ProtTech, Inc. (Norristown, PA). In brief, proteins in each gel band were digested "in-gel" with modified sequencing grade trypsin (Promega), and the resulting peptide mixture was subjected to tandem mass spectrometry for peptide sequencing. A Finnigan ion trap mass spectrometer LCQ coupled with an HPLC system running a 75 M ID C18 column was used. Data were acquired in a data-dependent mode. MS/MS spectra were used to search the most recent nonredundant protein data bases, including the Protein Information Resources data base and GenBank TM using ProtQuest software suite from ProtTech. The output of the data base search was manually analyzed and validated to verify proper protein identification.
Radioligand Binding Assays-HEK293T (5 ϫ 10 6 cells) cells were plated in 150-mm tissue culture dishes and transfected as described above. Forty eight hours later, the cells were removed mechanically using calcium-free Earle's balanced salt solution (EBSS). Intact cells were collected by centrifugation and then lysed with 5 mM Tris-HCl and 5 mM MgCl 2 at pH 7.4. Homogenates were centrifuged at 20,000 ϫ g for 30 min. The membranes were then resuspended with 50 mM Tris-HCl, pH 7.4. Membrane preparations were incubated for 90 min at room temperature with various concentrations of [ 3 H]SCH23390 (D 1 binding) or [ 3 H]methylspiperone (D 2 binding) in a reaction volume of 1 ml. Nonspecific binding was determined in the presence of 4 M (ϩ)-butaclamol. Bound ligand was separated from unbound ligand by filtration through polyethyleneiminesoaked GF/C filters using a Brandel cell harvester. GF/C filter disks were then analyzed via liquid scintillation spectroscopy at a counting efficiency of 60%. Protein concentration was determined using a Bradford protein assay (Bio-Rad).
cAMP Accumulation Assays-Transfected HEK293T cells were seeded into 24-well poly-D-lysine-coated plates (150,000 cells/well for D 1 ; 200,000 cells/well for D 2 ) and cultured for 1 day prior to the experiment. For D 1 receptor accumulation assays, cells were washed three times with 200 l of EBSS per well. Various concentrations of dopamine dissolved in stimulation buffer (DMEM, 20 M Ro-20-1724, 0.2 mM sodium metabisulfite) were added to each well in a volume of 250 l for 10 min at 37°C. For D 2 receptor assays, the cells were washed three times with 200 l of EBSS per well. Various concentrations of dopamine containing 3 M forskolin and 10 M propranalol in stimulation buffer were added to each well in a volume of 250 l for 10 min at 37°C. The reactions were terminated by adding 200 l of 3% perchloric acid and incubated on ice for 30 min. 80 l of 15% KHCO 3 was then added to neutralize the acid. The plates remained on ice for an additional 20 min and were then centrifuged at 1,300 ϫ g for 25 min. 50 l of the supernatant from each well was transferred to a 1.2-ml reaction tube containing 50 l of cAMP-binding protein, 50 l of [ 3 H]cAMP, and 150 l of Tris-EDTA buffer. The reaction was incubated for 90 min at 4°C. After the incubation, 250 l of 1% charcoal/dextran mixture was added to each tube and vortexed gently. Tubes were then incubated at 4°C for 10 min followed by centrifugation (1,300 ϫ g) for 20 min. Radioactivity in the supernatant was then quantified by liquid scintillation spectroscopy at a counting efficiency of 60%. cAMP concentrations were then determined using a standard assay curve according to the manufacturer's instructions.
Confocal Microscopy-1 ϫ 10 6 HEK293T cells were seeded in 100-mm culture dishes and transfected the next day with either D 1 -GFP or D 2 -YFP. For some experiments, cells were also transfected with calnexin. For double-labeling experiments that visualized the ER, the cells were co-transfected with pDsRed2-ER. Cells were cultured for an additional 24 h prior to reseeding at 100,000 cells per poly-D-lysine-coated 35-mm glass bottom culture dishes. Confocal microscopy was performed on a Zeiss laser-scanning confocal microscope (LSM 5 Pascal). Images were collected using a single line excitation (488 nm for D 1 -GFP and D 2 -YFP and 563 nm for pDsRed2-ER).

Identification of Calnexin as an Interacting Protein for
Both D 1 and D 2 Dopamine Receptors-As part of our efforts to identify novel dopamine receptor-interacting proteins, we employed immunoprecipitation (IP) assays using tagged receptor constructs along with mass spectrometry to identify associated proteins. For these studies, either FLAG-tagged dopamine receptor expression constructs or a vector expressing only the FLAG peptide (FLAG tag) was transfected into HEK293T cells. The cells were then solubilized, and the receptors and associated proteins were immunoprecipitated using anti-FLAG-agarose. The precipitated protein mixtures were then analyzed by SDS-PAGE. Fig. 1A shows an experiment using the D 1 dopamine receptor expressed in HEK293T cells. Coomassie Blue staining of the gel reveals the presence of multiple protein species in the immunoprecipitates (Fig. 1A, left lane). A few of these proteins were also found in the negative control cells expressing only the FLAG tag peptide (Fig. 1A, right lane). The proteins precipitated only in the presence of the D 1 receptor were tentatively assumed to interact specifically with the receptor. To identify these proteins, specific bands were excised from the gels and digested with trypsin. The eluted peptides were then subjected to LS-MS/MS tandem mass spectrometry as described under "Experimental Procedures." A number of proteins were identified using this approach and are listed in Table  1. The proteins identified in both the D 1 receptor and FLAG peptide immunoprecipitates were presumed to be nonspecific in nature. Importantly, we did not observe any proteins that were unique to the FLAG tag peptide samples. Among the putative D 1 receptor-interacting proteins was the ER chaperone protein calnexin. We hypothesized from these initial studies that calnexin was involved in dopamine receptor expression, and we therefore chose it for further study. Fig. 1B shows a representative MS/MS spectrum of a calnexin peptide found in the D 1 dopamine receptor sample (Fig.  1A, band 1). This protein was absent in the corresponding gel band (cut in parallel) from the FLAG tag sample lane (Fig. 1A). Three separate peptides corresponding to calnexin sequences were identified ( Fig. 1B and Table 2). Mass spectrometry was also performed on Fig. 1A, band 2, and three peptides for the D 1 receptor were identified ( Table 2). None of these peptides were apparent in the FLAG tag control. Using the D 2 receptor, we performed similar experiments (data not shown) as described in Fig. 1, which also identified calnexin in the D 2 receptor The FLAG-tag lane shows proteins immunoprecipitated from cells only expressing the FLAG peptide. B, representative MS/MS spectrum obtained from the fragmentation of the precursor ion at m/z 886.6. Liquid chromatography-MS/MS was carried out on the peptide mixture obtained from in-gel digestion of SDS-PAGE-separated protein samples with LCQ ion trap mass spectrometer coupled on line with an HPLC with 75-mm inner diameter C18 column. The precursor ions were selected automatically by the instrument. The peptide sequence was identified by bioinformatics analysis and found to correspond to the parent protein calnexin. Similar spectra were obtained for all reported peptides found in Table 2.
immunoprecipitates. Table 2 shows that two of the calnexin tryptic fragments identified in the D 1 and D 2 receptor experiments are identical. These results suggest that calnexin interacts with both D 1 and D 2 receptors in HEK293T cells.
Verification of D 1 and D 2 Receptor Interactions with Calnexin-To verify the interaction of the receptors with calnexin, we performed co-immunoprecipitation and Western blot analyses. Fig. 2A (left panel) shows that the D 1 receptor (apparent as a broad diffuse band ϳ50 kDa in size) is immunoprecipitated only from cells in which it is expressed. Stripping this blot and reprobing it with an antibody to calnexin ( Fig. 2A, right panel) reveals that calnexin (apparent as an ϳ70-kDa band) only immunoprecipitates in the presence of the D 1 receptor, and no band was seen in the control FLAG tag lane. To further verify that the D 1 receptor and calnexin directly interact, reverse IP experiments were performed using an anti-calnexin antibody. In this case, endogenous calnexin was immunoprecipitated from cells expressing either the D 1 receptor or FLAG tag and probed for the presence of the D 1 receptor. Fig. 2B (left panel) shows that the D 1 receptor (apparent as multiple bands of ϳ50 kDa) was co-immunoprecipitated with calnexin, but only in cells in which the D 1 receptor is expressed. Fig. 2B (right panel) shows the blot stripped of antibodies and re-probed with the anti-calnexin antibody. Dense bands of ϳ70 kDa are seen in both the D 1 receptor-transfected lane and the FLAG tag lane, indicating that calnexin was immunoprecipitated from both cell preparations and was loaded equally on the gel. These data further support a direct protein-protein interaction between the D 1 receptor and calnexin.
To ensure that the aforementioned interaction is not unique to HEK293T cells, we performed similar IP experiments using CAD cells, a central nervous system-derived catecholaminergic cell line that can be morphologically differentiated into neuronal phenotypes by the removal of serum (22). Fig. 2C (left panel) shows an immunoblot where CAD cells were transfected with either the D 1 receptor or FLAG tag and then immunoprecipitated with anti-FLAG-agarose. Probing the blot with an anti-D 1 antibody revealed strong laddered bands that were seen exclusively in the D 1 receptor-transfected cells, with the strongest bands centered around ϳ50 kDa (Fig. 2C, left panel). The lighter bands likely represent other glycosylation variants/intermediates and/or dimers of the D 1 receptor. This same blot was then stripped and re-probed with an anti-calnexin antibody revealing a single band of ϳ70 kDa only in the D 1 receptortransfected cells (Fig. 2C, right panel). These data indicate that the calnexin and D 1 receptor interactions are not exclusive to HEK293T cells but also occur in other cells, particularly of a neuronal nature.

TABLE 2 Peptides and corresponding parent proteins found in co-IP mass spectrometry experiments
Transfection a Band number a Indicated receptor cDNA was transfected into HEK293T cells. b Peptides identified from SDS-polyacrylamide gel bands that were excised, trypsinized, and subjected to mass spectrometry as described under "Experimental Procedures." c Parent proteins identified via data base search of trypsin-digested peptides are as described under "Experimental Procedures." d Identical peptides for parent protein calnexin are found in both D 1

and D 2 receptor
IPs. e DAR indicates dopamine receptor.

FIGURE 2. Co-immunoprecipitation of D 1 receptor and calnexin.
A, left panel, HEK293T cells were transfected with D 1 -FLAG or vector containing only the FLAG peptide (FLAG-tag). Proteins were extracted, immunoprecipitated (IP) using anti-FLAG-agarose, separated by SDS-PAGE, and immunoblotted (IB) as described under "Experimental Procedures." Blots were probed with a monoclonal anti-D 1 antibody and visualized using ECL after incubation with an anti-rat HRP-conjugated antibody. Right panel, the blot in the left panel was stripped of all antibodies, re-probed with an anti-calnexin antibody, and visualized using ECL after incubation with an anti-rabbit HRP-conjugated antibody. B, left panel, HEK293T cells were transfected with D 1 FLAG or vector containing only the FLAG peptide (FLAG-tag). Proteins were extracted and immunoprecipitated (IP) via incubation with anti-calnexin antibody followed by precipitation with protein-G-agarose. The blot was probed (anti-D 1 antibody) and visualized as reported for A. Right panel, the blot in the left panel was stripped of all antibodies, re-probed with an anti-calnexin antibody, and visualized as described for A. Multiple bands are seen in this blot. The top band is calnexin, and the others represent visualization of the antibody used for immunoprecipitation (anti-calnexin antibody); because the IP and the immunoblot antibodies are identical in this assay, they are visible after probing with secondary antibody. C, left panel, CAD cells were transfected with D 1 -FLAG or vector containing only the FLAG peptide (FLAG-tag). Proteins were extracted, immunoprecipitated (IP) using anti-FLAG-agarose, and immunoblotted (IB) as described under "Experimental Procedures." The blot was probed (anti-D 1 antibody) and visualized as described above for A. Right panel, the blot from the left panel was stripped of all antibodies, re-probed with an anti-calnexin antibody, and visualized as described for A. All blots shown are representative of three independent experiments with identical results.
We performed similar co-immunoprecipitation analyses to verify the interaction between calnexin and the D 2 receptor. Fig. 3 shows an experiment in which HEK293T cells were transfected with either the D 2 receptor or FLAG tag, solubilized, and immunoprecipitated with anti-FLAG-agarose followed by SDS-PAGE. The resulting blot was probed with an anti-D 2 receptor antibody and visualized using ECL (Fig. 3A). A complex laddering of bands probably corresponding to various glycosylation states of the D 2 receptor was seen exclusively in the D 2 receptor-transfected lane. In Fig. 3B, the same blot was stripped of antibody and re-probed with an anti-calnexin antibody. A strong band corresponding to calnexin (ϳ70 kDa) was seen in the D 2 receptor-transfected lane but not in the FLAG tag lane. These data suggest that calnexin also directly interacts with the D 2 receptor.
Characterization of D 1 and D 2 Receptor Interactions with Calnexin-Calnexin normally binds to N-linked glycans of proteins via its lectin binding domain, and by doing so it assists in the proper folding and maturation of proteins within the ER. However, recent data have suggested that calnexin can also bind to its substrates through glycosylation-independent mechanisms (23,24). To investigate the mechanism of calnexin-receptor interactions, we made expression constructs of the D 1 and D 2 receptors that lacked N-linked glycosylation sites. Expression of these constructs followed by Western blot analysis demonstrated that they migrate at lower molecular weights than their wild-type counterparts, consistent with a lack of glycosylation (Fig. 4, A and C). To ensure that the mutant receptors are completely devoid of N-linked glycosylation, control lysates were treated with PNGase F which cleaves at the initial GlcNAc-Asn bond (25). As seen in Fig. 4, A and C, treatment of the receptors with PNGase F resulted in migration patterns identical to those of the mutant receptors, indicating that the mutants lacked N-linked glycosylation. Although the D 1 receptor mutant migrated as a single band (ϳ46 kDa), the D 2 receptor mutant displayed two prominent bands. Although the exact nature of the higher molecular weight band remains unknown, it may represent a dimer of the nonglycosylated receptor. Clearly, glycosylation is not needed for the formation of this higher molecular weight species as it is present in both the mutant and the PNGase F-treated samples.
We next used these glycosylation-defective receptors to determine the importance of glycosylation for the interactions of calnexin with the receptors. This was assessed by comparing the amount of calnexin that immunoprecipitated with the glycosylation mutants versus the wild-type receptors. As seen in Fig. 4, B and D, the amount of calnexin that immunoprecipitated with the mutant receptors was reduced by about 50% compared with that of the wild-type receptors. Importantly, eliminating the receptor glycosylation only partially blocked the receptor-calnexin interactions, therefore suggesting that a glycosylation-independent mode of interaction with the receptors must also exist.
To ensure that the reduction in receptor-calnexin interactions was because of deglycosylation, and not due to some unknown effect of the mutants, we treated cells expressing the wild-type receptors with drugs known to specifically inhibit protein glycosylation. One drug used in these experiments was tunicamycin, a well characterized specific inhibitor of N-linked protein glycosylation (26 -29). Fig. 5A shows that tunicamycin treatment results in the D 1 receptor appearing as a single band on SDS-PAGE with a molecular mass of ϳ49 kDa, similar in size to the predicted nonglycosylated form of the receptor (an identical pattern to that seen with the glycosylation-deficient mutant in Fig. 4A). Stripping of this blot and reprobing with a calnexin antibody show that there is a reduction, but not elimination, of the amount of calnexin that co-immunoprecipitates with the D 1 receptor (Fig. 5B). Tunicamycin treatment also results in the D 2 receptor appearing primarily as a single species of ϳ50 kDa, although a minor form of ϳ70 kDa is still present (Fig. 5C) (an identical pattern to that seen with the glycosylation deficient mutant in Fig. 4C). As with the D 1 receptor, tunicamycin treatment reduces, but does not eliminate, the ability of calnexin to co-immunoprecipitate with the D 2 receptor (Fig.  5D). These results are completely congruent with those obtained using the glycosylation-defective receptor constructs, and they support the use of tunicamycin as a pharmacological probe of D 1 and D 2 receptor glycosylation. More importantly, these results further support the notion that calnexin not only interacts with the D 1 and D 2 receptors through their N-linked glycans, but must also associate with the receptors via direct protein-protein interactions.
As an additional test of this hypothesis, we enzymatically removed the carbohydrate residues from the solubilized D 1 receptor using PNGase F (see above). Fig. 5E demonstrates that PNGase F treatment eliminates the complex glycosylation of the D 1 receptor resulting in a single lower molecular weight band identical to that seen after tunicamycin treatment (cf. Fig.  5A). Fig. 5F shows that PNGase F treatment results in a significant reduction in D 1 receptor-calnexin interactions but does not completely eliminate this association. This supports our contention that calnexin interacts with the receptor via glycandependent and -independent mechanisms.
To further understand the location and character of the interactions between calnexin and the D 1 receptor, we employed IP assays using various D 1 mutant receptors previously generated by our laboratory. For these experiments we utilized two different D 1 mutants. One mutant, designated D 1 -T347-FLAG, is a deletion mutant (truncated at position 347) that is missing most of its carboxyl terminus. We have previously shown that the cell surface expression of this mutant is diminished relative to the wild-type receptor (30), but it is fully functional in terms of activating adenylyl cyclase. The second mutant, designated D1-3rd-FLAG, has all of the serine and threonine residues within the third cytoplasmic domain changed to either alanines or valines, thereby eliminating all phosphorylation sites in this region. This mutation has no effect on receptor expression and is fully functional, although it exhibits impaired desensitization (30). These two mutant receptors, along with the wild-type D 1 receptor and the FLAG tag construct, were expressed in HEK293T cells, and co-immunoprecipitation assays were performed (Fig. 6A). Importantly, radiolabeled binding assays were conducted to ensure that equal amounts of D 1 receptor were loaded onto each lane of the SDS-polyacrylamide gels. As seen in Fig. 6A, 1st lane, a single band of ϳ70 kDa is detected, indicating that calnexin co-immunoprecipitates with the wild-type D 1 receptor as shown previously in Fig. 2A. Interestingly, Fig.  6A, 2nd lane, shows a significant increase in the amount of calnexin that co-immunoprecipitates with the D 1 -T347-FLAG mutant. In contrast, as shown in Fig.  6A, 3rd lane (Fig. 6A), calnexin co-immunoprecipitates with the D 1 -3rd-FLAG mutant to a similar extent as with the wildtype D 1 receptor. These data indicate that phosphorylation of the third cytoplasmic domain is not necessary for D 1 receptorcalnexin interactions. Furthermore, these results indicate that calnexin does not interact with the carboxyl terminus of the D 1 receptor. In fact, truncation of the carboxyl terminus actually enhances the interaction of the D 1 receptor with calnexin.
To assess if the increased interaction of calnexin with the D 1 -T347-FLAG mutant was because of increased interactions with the receptor glycans or protein regions, we treated cells  D). Proteins were extracted, immunoprecipitated (IP) using anti-FLAG-agarose, either untreated or treated with PNGase F as indicated, separated by SDS-PAGE, and immunoblotted (IB) as described under "Experimental Procedures." Blots were probed with an anti-D 1 antibody, an anti-D 2 antibody, or an anti-calnexin antibody as indicated and visualized using ECL after incubation with the appropriate HRP-conjugated secondary antibodies. Molecular weight markers designate pre-stained molecular weight standards. The blot shown is representative of three independent experiments all showing identical results.

FIGURE 5. Characterization of D 1 and D 2 receptor interactions with calnexin.
A, HEK293T cells were transfected with D 1 -FLAG, and cells were either untreated or treated with 2 g/ml tunicamycin for 24 h as indicated. Proteins were extracted, immunoprecipitated (IP) using anti-FLAG-agarose, separated by SDS-PAGE, and immunoblotted (IB) as described under "Experimental Procedures." Blots were probed with a monoclonal anti-D 1 antibody and visualized using ECL after incubation with an anti-rat HRP-conjugated antibody. B, blot in A was stripped of all antibodies, re-probed with an anti-calnexin antibody, and visualized using ECL after incubation with an anti-rabbit HRP-conjugated antibody. C, HEK293T cells were transfected with D 2 FLAG, and cells were either untreated or treated with 2 g/ml tunicamycin for 24 h as indicated. Proteins were extracted and immunoprecipitated (IP) using anti-FLAG-agarose, separated by SDS-PAGE, and immunoblotted (IB) as described under "Experimental Procedures." The blot was probed (anti-D 2 antibody) and visualized using ECL after incubation with an anti-rabbit HRP-conjugated antibody. D, the blot in C was stripped of all antibodies, re-probed with an anti-calnexin antibody, and visualized as described for B. Two bands are seen in each lane of this blot. The top band is calnexin, but the other likely represents visualization of the antibody heavy chain. E and F, effect of enzymatically deglycosylating the solubilized D 1 receptor on its ability to associate with calnexin. HEK293T cells were transfected with D 1 -FLAG, and soluble proteins were extracted as described under "Experimental Procedures." Samples were incubated at 37°C for 1 h with either 5 units of PNGase F (ϩ) or buffer alone (Ϫ). Proteins were then immunoprecipitated (IP) using anti-FLAG-agarose, separated by SDS-PAGE, and immunoblotted (IB) as described under "Experimental Procedures." E, blots were probed with a monoclonal anti-D 1 antibody and visualized using ECL after incubation with an anti-rat HRP-conjugated antibody. F, blot in E was stripped of all antibodies, re-probed with an anti-calnexin antibody, and visualized using ECL after incubation with an anti-rabbit HRP-conjugated antibody. Densitometry of two separate blots revealed an ϳ25% decrease in calnexin immunoreactivity following PNGase F treatment.
expressing the D 1 -T347-FLAG mutant with tunicamycin. As can be seen in Fig. 6B, there was no effect of tunicamycin treatment on the ability of calnexin to co-immunoprecipitate with the D 1 -T347-FLAG receptor. These results contrast with those seen with the wild-type receptor (Fig. 5A). These data further indicate that calnexin interacts with the D 1 receptor via its protein, as well as its carbohydrate component, and that neither of these interactions occur in the cytoplasmic tail or via phosphorylation of the third cytoplasmic loop.
Effects of Calnexin on D 1 and D 2 Receptor Expression-To investigate the functional importance of receptor-calnexin interactions, we initially examined the effects of two drugs, tunicamycin and castanospermine, that are known to interfere with the association of calnexin with glycoproteins. As noted above, tunicamycin is a glycosylation inhibitor capable of preventing N-linked glycosylation. Castanospermine is a glucosidase inhibitor that inhibits both glucosidases I and II (31-34), which sequentially trim glycoproteins to mono-glucosylated states that preferentially interact with calnexin (35) . Fig. 7, A and B, shows saturation radioligand binding assays using HEK293T cells transfected with the D 1 receptor. Cells were treated with either castanospermine for 48 h (Fig. 7A), with tunicamycin for 24 h (Fig. 7B), or left untreated followed by membrane binding assays using the D 1 -selective antagonist [ 3 H]SCH23390. No significant change is seen in the maximum binding capacity (B max ) or the affinity (K d ) of [ 3 H]SCH23390 for the D 1 receptor (Fig. 7, A and B, and Table 3).
Similar experiments were conducted with HEK293T cells expressing the D 2 receptor (Fig. 7, C and D). In this case, saturation binding assays using the D 2 -selective radioligand [ 3 H]methylspiperone revealed that both castanospermine and tunicamycin treatments reduced the total number of D 2 receptors without any effect on their affinity for the radioligand (Fig.   7, C and D, and Table 3). Thus, total D 2 receptor expression appears to be more sensitive to the glycan-dependent calnexin interactions than that of the D 1 receptor.
The assays shown in Fig. 7, A-D, are designed to measure the total complement of membrane-associated receptors, which include those on the cell surface as well as any receptors residing in internal membranous structures or vesicles. As it was of interest to assess the effects of calnexin on the trafficking of receptors to the cell surface, we employed an intact cell binding assay that utilizes a hydrophilic radioligand, [ 3 H]sulpiride, that only labels D 2 receptors on the cell surface (21) . Fig. 7, E and F, shows intact cell saturation binding assays with [ 3 H]sulpiride after treatment with tunicamycin or castanospermine. As can be seen, both of these treatments significantly reduced the expression of the D 2 receptor at the cell surface, with tunicamycin having a greater effect than castanospermine (Table 3). Interfering with D 2 receptor-calnexin interactions not only reduces cell surface receptor expression but also decreases the total amount of D 2 receptors expressed within the cell.
Surface expression was also investigated despite no significant change in the total number of cellular D 1 receptors after glycosylation interference. Unfortunately, a cell surface-restricted radioligand is not available for the D 1 receptor, thus negating the ability to assess cell surface expression in this fashion. As an alternative approach, we performed cAMP accumulation assays that reflected the functional D 1 receptors expressed at the cell surface. After treating cells with tunicamycin or castanospermine, dopamine dose-response curves were generated, resulting in internal cAMP accumulation (Fig. 8). Castanospermine treatment resulted in a significant (ϳ30%) decrease in the maximum accumulation of cAMP in response to dopamine, with no significant change in the potency of dopamine (Fig. 8A). Similarly, treatment with tunicamycin resulted in an ϳ30% reduction of E max , whereas there was no change in the EC 50 for dopamine (Fig. 8B). Taken together, these data show that inhibition of calnexin-receptor interactions results in a decrease in D 1 receptor-mediated signaling, which is consistent with a decrease in receptor expression at the cell surface.
To ensure that the results observed with tunicamycin and castanospermine are specific to blockade of receptor glycosylation, we performed similar experiments using the glycosylation-defective constructs described above. When radiolabeled binding assays were performed on the D1-GLYT mutant, an increase in total cellular binding was seen (Table 3). However, when functional cAMP accumulation assays were performed with the D1-GLYT mutant, an ϳ27% decrease was seen in the E max value with no change in EC 50 (percent control E max , 100% versus 64.2% Ϯ 9.7, n ϭ 3, for D1-FLAG and D1-GLYT, respectively). These data indicate a reduction in the surface expression of the mutant receptor despite an increase in overall expression levels. Membrane binding assays for the D2-GLYT mutant demonstrated a significant (ϳ85%) decrease in overall expression, paralleling results seen with glycosylation inhibitor treatment. Overall, the results with the glycosylationdefective mutants are in excellent agreement with those obtained using tunicamycin, which inhibits N-linked glycosylation of proteins.
To further investigate the importance of calnexin on D 1 and D 2 receptor expression, we used chimeric proteins in which the D 1 receptor was fused with GFP, and the D 2 receptor was fused with YFP, thus enabling their subcellular visualization using confocal fluorescence microscopy. This approach has the added advantage of allowing visualization of receptor proteins that are not capable of binding ligands in intact cells, such as receptors in prefolded or misfolded states. Cells were transfected with D 1 -GFP or D 2 -YFP and then left untreated or treated with either castanospermine or tunicamycin and visualized. Fig. 9, A and D, shows untreated (control) D 1 and D 2 receptors. Nearly all receptor protein is located at the surface of the cell plasma membrane with very little intracellular fluorescence. Fig. 9, B (D 1 ) and E (D 2 ), shows representative cells after treatment for 24 h with tunicamycin. These cells show both punctate and diffuse intracellular staining in addition to surface staining. Fig. 9, C (D 1 ) and F (D 2 ), are representative cells after treatment with castanospermine for 24 h. These cells also show discrete pockets of bright staining and intracellular clusters. Taken together, these data show that attenuating the ability of calnexin to interact with either the D 1 or D 2 receptors results in decreased trafficking of the receptors to the cell surface. In the case of the D 2 receptor, there is also a reduction in the total amount of receptor capable of interacting with ligands (Fig. 7, C and D).
Another approach for determining the biological importance of the receptor-calnexin interactions is to study the effect of  Table 3.
calnexin overexpression on receptor expression and function. In Fig. 10, we overexpressed calnexin by transfecting it along with the D 1 receptor in HEK293T cells. Co-immunoprecipita-tion experiments indicated that this was associated with a Ͼ2-fold increase in calnexin association with the receptor (data not shown). Surprisingly, we found that overexpression of calnexin decreases the D 1 receptor expression level to about onethird of control with no change in the affinity of the receptor for the radioligand (Fig. 10A). Similarly, when D 1 receptor function was examined (Fig. 10B), dopamine-stimulated cAMP accumulation was decreased by ϳ50% when calnexin was overexpressed without a change in agonist potency. These data demonstrate that overexpression of calnexin decreases the total cellular complement of D 1 receptors and causes a concomitant decrease in D 1 receptor function.
To determine whether calnexin overexpression also affects D 2 receptor expression and function, radioligand binding and functional assays were performed after calnexin co-transfection. Fig. 11A shows that the maximum binding capacity for [ 3 H]methylspiperone is decreased by Ͼ50% with a slight change in receptor affinity for the radioligand (control K d ϭ 0.40 nM Ϯ 0.04 versus treated K d ϭ 0.25 nM Ϯ 0.01, n ϭ 4). To assess the function of the D 2 receptor, we examined its ability to inhibit forskolin-stimulated cAMP accumulation. As shown in Fig. 11B, under control conditions, dopamine dose-    Fig. 7, A and B; n ϭ 5. b Representative graphs of these data are presented in Fig. 7, C and D; n ϭ 5. c Representative graphs of these data are presented in Fig. 7E; n ϭ 5. d Data are significantly different from paired untreated control (p Ͻ 0.001).
dependently inhibits cAMP accumulation with a maximum response of nearly 50%. Calnexin overexpression decreases the maximum D 2 receptor-mediated cAMP inhibition by about half without a change in the EC 50 value for dopamine (Fig. 11B).
These data indicate that overexpression of calnexin also affects D 2 receptor expression and function in a negative manner consistent with a decrease of receptor surface expression. Taking the D 1 and D 2 data together, it appears that there may be a level of calnexin expression (endogenous) that is optimal for receptor maturation and biosynthesis, but if this level is exceeded, calnexin may actually retard receptor trafficking to the cell surface. To explore this further, we examined the subcellular distribution of the D 1 and D 2 receptors after calnexin overexpression using confocal fluorescence microscopy. As demonstrated in Fig. 12, when HEK293T cells were transfected with either D 1 -GFP (Fig. 12A) or D 2 -YFP (Fig. 12C), most receptor protein is located at the cell surface. However, when calnexin is overexpressed in these cells (Fig. 12, B and D), the receptors are mostly found in discrete punctate oval-like clusters within the cells. This is consistent with the receptors not traveling to the cell surface in the presence of a high level of calnexin protein, suggesting a retardation of trafficking.
We have hypothesized that calnexin is acting as a chaperone protein for the D 1 and D 2 receptors via ER retention properties similar to its effects on other proteins. If this is the case, the receptor proteins that are internally retained after overexpression of calnexin should be localized to the ER. To examine this, we conducted double-labeling experiments using pDsRed2-ER, a vector designed for fluorescent labeling of the ER in living cells. The pDsRed2-ER is an expression vector that encodes a fusion protein consisting of red fluorescent protein, the ER targeting sequence of calreticulin, and the ER retention sequence KDEL. Cells were transfected with D 1 -GFP, pDsRed2-ER, and calnexin, followed by confocal fluorescence microscopy at either 488 nm (GFP) or 543 nM (DsRed2). As observed previously, when calnexin is overexpressed in these cells, the D 1 receptor does not traffic to the cell surface but is retained in intracellular punctate oval clusters (Fig. 13A). These clusters  are also labeled by the ER marker protein (Fig. 13B), indicating that they represent endoplasmic reticula. When the images are merged, it is clear that the punctate intracellular clusters of receptors are located in the ER (Fig. 13C). These data are consistent with calnexin acting as an ER retention protein that, if overexpressed, results in newly synthesized D 1 receptors being retained in the ER.

DISCUSSION
The major finding of this study is our identification of calnexin as a novel interacting protein for both D 1 and D 2 dopamine receptors. Calnexin functions as both an ER retention protein and chaperone, thus enabling the proper folding and assembly of glycoproteins prior to their export to the Golgi. We determined that interaction with calnexin in the ER is critical for the optimum expression of both D 1 and D 2 receptor subtypes. Interestingly, the interaction of calnexin with the receptors appears to be tightly regulated as either decreasing (via receptor mutagenesis or pharmacological inhibitors) or increasing (via overexpression) calnexin-receptor interactions diminished receptor trafficking and expression. Clearly, there is an optimal degree or duration of calnexin interaction that results in the maximum amount of receptor being expressed at the cell surface. Calnexin thus joins a growing list of proteins being identified that directly interact with dopamine and other GPCRs to regulate their expression, signaling, subcellular location, and desensitization (8,9).
The interaction of calnexin with D 1 and D 2 receptors appears to be complex and dependent upon two different mechanisms. Previously, most studies have suggested that calnexin specifically associates with glycoproteins based on their N-linked oligosaccharide side chains, preferring monoglucosylated oligosaccharides (13). This has been extensively demonstrated for many calnexin-protein interactions where the interaction is completely inhibited in the absence of glycosylation. However, newer data suggest that calnexin may also target and bind nonglycosylated membrane proteins (14,23,24). In one case, calnexin has been shown to bind to transmembrane regions of the proteolipid protein (14). These data suggest that calnexin can interact with proteins via glycan-dependent as well as glycanindependent mechanisms. Our current results also support a   Fig. 10 with calnexin overexpression. Cells marked with an asterisk indicate cells transfected with D 1 -GFP but did not co-transfect with either calnexin or DsRed2-ER. B, visualization of cells at ϫ1000 using a single line excitation of 543 nm, detecting DsRed2-ER. The endoplasmic reticulum appears red. C, overlay of A and B, clearly indicating overlap of the punctate clusters found with calnexin overexpression demonstrating that the receptor is localized to the ER. Images are representative of three separate experiments.
dual mechanism of interaction of D 1 and D 2 receptors with calnexin. Receptor mutants lacking N-linked glycosylation sites, as well as inhibiting receptor glycosylation with tunicamycin, diminished but did not completely block the interaction of calnexin with the receptors. Furthermore, the increased association of calnexin with a trafficking-impaired D 1 receptor mutant was not affected by inhibiting receptor glycosylation (Fig. 6). Taken together, these results suggest that calnexin associates with D 1 and D 2 receptors via their protein components, as well as their carbohydrate side chains.
The protein-protein interaction domains that mediate calnexin association with the D 1 and D 2 receptors are at present unclear and require further study. Calnexin clearly does not bind to the long carboxyl terminus of the D 1 receptor, as an increased association was observed with the Thr-347 mutant where the carboxyl region is almost completely eliminated. Calnexin contains a single transmembrane spanning domain (34), and it is conceivable that calnexin could interact with D 1 /D 2 receptors, or other GPCRs, via their transmembrane domains as has been demonstrated for the proteolipid protein (14). Alternatively, calnexin may interact with other nonmembranous regions of the receptors. These possibilities are not mutually exclusive and are currently under investigation.
The fact that calnexin can interact with glycoproteins via two distinct mechanisms raises the possibility that these different interactions may serve different functions. Although calnexin functions as a chaperone protein to assist in protein folding and assembly, it also serves as an ER retention protein that prevents misfolded proteins from exiting the ER and going to the Golgi. Prolonged retention in the ER subjects the glycoprotein to a quality control process involving further modification by slow acting ␣-mannosidase and shuttling to an ER-associated degradation (ERAD) pathway (35,36). Interestingly, we found that the increased interaction of calnexin with a trafficking-impaired mutant of the D 1 receptor was insensitive to glycosylation inhibitors (Fig. 6), suggesting that the increased association was because of protein-protein interactions. Conversely, inhibiting glycosylation of the wild-type D 1 receptor decreased calnexin interactions and cell surface trafficking. It is thus tempting to speculate that the glycan-dependent receptor-calnexin interactions may mediate the chaperone functions of calnexin, whereas the glycan-independent protein-protein interactions may be more relevant to ER retention and quality control.
Further evidence for an ER retention function of calnexin was obtained using calnexin overexpression assays, which resulted in a very distinct pattern of receptor localization as determined by confocal microscopy. Similar staining patterns have been documented previously for cystic fibrosis transmembrane conductance regulator protein subsequent to overexpression of calnexin, where it was determined that the cystic fibrosis transmembrane regulator protein remained in concentric membranous bodies of the ER (37). Our double labeling experiments indeed determined that the D 1 receptor is localized in the ER following overexpression of calnexin. This confirms that calnexin acts as an ER retention protein for receptors when it is overexpressed. Interestingly, we see no such patterns when the subcellular location of the receptors is visualized after treatment of the cells with gly-cosylation inhibitors. Instead, the receptors appear to be in diffuse intracellular clusters located throughout the cell. These data indicate that the mechanism of receptor downregulation via calnexin overexpression is distinct from that of receptor down-regulation via inhibition of glycosylation. Interestingly, similar results using overexpression of calnexin and other ER retention proteins such as BiP/GRP78 and DRiP78 have been observed (10,38,39). Overexpression of these ER chaperone/retention proteins resulted in diminished transport of their substrates from the ER to the Golgi, possibly by shifting the equilibrium away from, or blockade of, a cargo-selection process.
With respect to glycan-dependent mechanisms of the calnexin interaction, it was interesting to note that there were some differences with respect to how this regulated the expression of the D 1 and D 2 receptors. Inhibiting glycosylation of the D 1 receptor diminished its cell surface trafficking. It should be noted that these results contrast with those of Karpa et al. (28) who found that N-linked glycosylation was not required for D 1 receptor localization to the plasma membrane. The reason(s) for this discrepancy is not clear, although Karpa et al. (28) utilized the human D 1 receptor, while we performed experiments with the rat receptor protein. In contrast to the D 1 receptor results, inhibiting glycosylation decreased both cell surface trafficking and the overall cellular expression of the D 2 receptor. The reason for this is unknown, but it may involve the differences in glycosylation levels of the two proteins. Although the D 1 receptor has only two identified N-linked glycosylation sites, the D 2 receptor has four such consensus sequences for N-linked glycosylation. Potentially, the D 2 receptor is more heavily glycosylated than the D 1 receptor and may be more likely to degrade in the absence of glycosylation.
Another role that calnexin may play in D 1 and D 2 receptor expression is that of oligomeric assembly in the ER. GPCR dimers are now known to regulate a number of functional properties of these receptors (for reviews see Refs. 40 and 41). One important aspect of dimerization is its potential to play a critical role in the surface trafficking of at least some GPCRs (42)(43)(44). The mechanisms underlying GPCR dimerization within the ER remain elusive but may involve folding states of the receptors through interactions with chaperone proteins. Indeed, ER chaperone proteins have been implicated in folding of several receptors for peptide hormones, including V2 vasopressin (45,46), luteinizing hormone, follicle-stimulating hormone, and thyrotropin receptors (47,48). Interaction of GPCRs with calnexin may facilitate the formation of dimers prior to trafficking to the cell surface. It is known that D 1 and D 2 receptors exist as dimers (41) on the cell surface. In our Western blot analyses, both D 1 (Fig. 2, A and C) and D 2 (Fig. 3) receptors exhibited higher molecular weight species indicative of oligomeric complexes. Interestingly, after co-precipitating the D 1 receptor with an anti-calnexin antibody (Fig. 2B), only a lower molecular weight species corresponding to a receptor monomer is observed. This suggests that calnexin may bind exclusively to incompletely assembled monomeric receptors. These data lead to an hypothesis of calnexin acting as a retention protein for individual subunits, assisting in the formation of completely assembled dimeric/oligomeric receptors.
In summary, calnexin is a D 1 and D 2 dopamine receptorinteracting protein with multifaceted effects on the trafficking and expression of both receptors. Our findings suggest that optimal levels of calnexin interaction are essential for proper expression of fully assembled and functional dopamine receptors on the cell surface. This GPCR expression/trafficking pathway is likely to be under precise cellular control, as alterations have a profound influence on receptor expression and subsequent signaling. These findings further lead to the possibility of using GPCR-chaperone complexes as targets for the manipulation of receptor expression.