Cell-type Specific Targeting of the α2c-Adrenoceptor

We have previously shown differences in the intracellular targeting of α2a (α2A)- and α2c (α2C)-adrenoreceptors expressed in the same cell line (von Zastrow, M., Link, R., Daunt, D., Barsh, G., and Kobilka, B. (1993) J. Biol. Chem. 268, 763–766; Daunt, D. A., Hurt, C., Hein, L., Kallio, J., Feng, F., and Kobilka, B. K. (1997) Mol. Pharmacol. 51, 711–720). α2A-Adrenoreceptors reside primarily in the plasma membrane in HEK 293 cells, while co-expressed α2C-adrenoreceptors are found mainly in an intracellular compartment. Since α2c-adrenoreceptors are expressed primarily in the brain, we compared the intracellular targeting of α2C-adrenoreceptors in two neuroendocrine cell lines with the targeting in three epithelial cell lines and one fibroblast cell line. In transiently transfected COS7 cells, and in stably transfected normal rat kidney cells, Madin-Darby canine kidney cells, and Rat1 fibroblasts, a significant proportion of α2C-adrenoreceptor detected by immunocytochemistry co-localized with markers for both the endoplasmic reticulum and the cis/medial Golgi compartments. In contrast, both PC12 cells and AtT20 cells efficiently targeted α2C-adrenoreceptors to the plasma membrane. Ligand binding and Western blot analyses indicate that intracellular receptor in normal rat kidney cells is functional and undergoes normal post-translational processing. In PC12 cells the expressed α2C-adrenoreceptors become concentrated in neurite outgrowths in discrete regions of the plasma membrane having a high density of F-actin following treatment with nerve growth factor. These findings provide evidence for cell-type specific factors that facilitate the targeting of the G protein-coupled receptors to the plasma membrane.

G protein-coupled receptors mediate transmission of information across the plasma membrane by activation of membrane-associated G-proteins that couple to intracellular effector systems. Although most G protein-coupled receptor ligands are confined to the extracellular space, for some G proteincoupled receptors, such as the thrombin receptor, thyrotropinreleasing hormone receptor, and the ␣ 2C -adrenoreceptor, a significant proportion of the receptor population is found in intracellular compartments. In the case of thrombin receptors, evidence suggests that the intracellular pool of receptors may serve as a reservoir, capable of restoring functional, noncleaved thrombin receptors to the plasma membrane after the cell has been exposed to thrombin (3). A minor fraction of thyrotropin-releasing hormone receptors expressed in HEK cells are targeted to the plasma membrane while a larger fraction is found in an intracellular compartment and are nonfunctional (4). However, when the thyrotropin releasing hormone receptors were expressed in two pituitary cell lines (GH3 and GHY cells) the receptors were found predominantly in the plasma membrane and to be functional by binding assays.
For the ␣ 2C -adrenoreceptor, the functional status and role of an intracellular pool is less clear. This is particularly interesting in light of ␣ 2C -adrenoreceptors overall structural and functional similarity to the other ␣ 2 -adrenoreceptor subtypes, ␣ 2Aand ␣ 2B -adrenoreceptors (for reviews, see Ref. 5). Several reports have documented the presence of a large intracellular distribution of ␣ 2C -adrenoreceptors when either transiently or stably transfected into a variety of cell lines (1,2,6). Furthermore, the levels of expression achieved for the ␣ 2C -adrenoreceptors in cells have been comparable to the expression level obtained for the other two ␣ 2 -adrenoreceptor subtypes in these same cell lines (7,8). This would indicate that the functional expression of the three ␣ 2 -adrenoreceptor subtypes occurs at similar efficiencies. In addition, the three ␣ 2C -adrenoreceptors have been reported to have comparable signaling characteristics in that they all couple to G i / o proteins, inhibit adenylyl cyclase (7,8), and their stimulation can lead to activation of a mitogen-activated protein kinase pathway (9,10).
A recent report described the expression of ␣ 2C -adrenoreceptors in Madin-Darby canine kidney cells and found the receptors to be in the plasma membrane and in two intracellular compartments, endoplasmic reticulum, and trans-Golgi network (6). We have studied the trafficking of ␣ 2C -adrenoreceptor in Rat1 fibroblasts and determined that the intracellular pool of receptors was localized in the endoplasmic reticulum and in the cis/medial Golgi compartments. Unlike the thrombin receptor, there is no cycling of the ␣ 2C -adrenoreceptor between this intracellular pool and the plasma membrane in these cells (2). These observations taken together suggest that the ␣ 2C -adrenoreceptor may be improperly folded or poorly processed and thereby retained in the ER. 1 It has been shown that misfolded proteins tend to accumulate in the ER and have several potential fates that include degradation within the endoplasmic reticulum (11), cytosolic degradation by ubiquitin-proteasome pathway (12), or lysosomal targeting and degradation (13). However, ER retention may be a fundamental characteristic of the ␣ 2C -adrenoreceptor subtype and may play some functional role in the cell, or may be due to the lack of expression of specific processing or transport factors (e.g. chaperones) in the cell lines examined to date.
To explore the later possibility, we have investigated the targeting of ␣ 2C -adrenoreceptors in several cell lines including two of neuronal lineage. The ␣ 2C -adrenoreceptors are primarily expressed in the central nervous system and recent studies of ␣ 2C adrenoreceptor knockout mice demonstrate this receptor subtype plays a role in modulating several aspects of behavior (5,14,15) and in regulating catecholamine release from sympathetic neurons in the heart (16,17). Our results provide evidence for cell-type specific targeting of G protein-coupled receptors and for the development of receptor microdomains during neuronal differentiation of PC12 cells.

EXPERIMENTAL PROCEDURES
Cell Culture-Normal rat kidney (NRK) cells, Rat1 fibroblast cells, and COS7 cells were cultured at 37°C with 5% CO 2 in Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% fetal bovine serum (Gemini Bio-Products, Inc., Calabasa, CA) and gentamicin (25 g/ml; Roche Molecular Biochemicals Biochemicals, Indianapolis, IN). PC12 cells were cultured at 37°C with 10% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% equine donor serum and 5% bovine calf serum (Hyclone). Murine ␣ 2C -adrenoreceptor was subcloned into the pCDNA3 expression vector. A 12CA5 epitope was added to the end of the amino terminus by using an oligonucleotide linker-adapter into the NcoI site of the 5Ј coding sequence of the receptor. Cells were transfected with 12CA5 epitope tagged ␣ 2C -adrenoreceptor by either calcium phosphate precipitation or electroporation. Stably transfected cells were obtained by growing cells in the presence of 0.25 to 0.8 mg/ml Geneticin (G418; Life Technologies, Inc., Grand Island, NY). The neuronal induction of PC12 cells was accomplished by the addition of 50 ng/ml nerve growth factor (NGF) to the media.
Receptor Expression-␣ 2C -Adrenoreceptor expression level for each of the stably transfected cell lines was determined by saturation binding with the ␣ 2 -adrenoreceptor antagonist, [ 3 H]RX821002. Briefly, cell membrane preparations were made by Polytron homogenization in a hypotonic lysis buffer (10 mM Tris and 1 mM EDTA, pH 7.4). The cell homogenate was centrifuged at 1,000 ϫ g for 5 min to remove nuclear debris and intact cells. The recovered supernatant was centrifuged at 10,000 ϫ g for 90 min to obtain cell membranes. Membrane pellets were then suspended in binding buffer (75 mM Tris, 12.5 mM MgCl 2 , and 1 mM EDTA, pH 7.4) and stored at Ϫ70°C until studied. The amount of ligand bound to receptors was obtained by vacuum filtration through Whatman GF/C glass filters. Protein concentration was determined by Bio-Rad assay using bovine serum albumin as standard.
Immunocytochemistry-Cells were seeded and grown on sterile glass coverslips coated with poly-D-lysine 2 days before studying. After various treatments, cell preparations were fixed for 5 min with either 4% paraformaldehyde at room temperature or cold methanol at Ϫ20°C. Following fixation cells were rinsed three times with phosphate-buffered saline (PBS) supplemented with calcium (Ca 2ϩ ) and magnesium (Mg 2ϩ ). A blocking agent composed of 5% dry milk, 50 mM HEPES, pH 7.4, in PBS was used to reduce nonspecific antibody activity. The nonionic detergent, Nonidet P-40 (Sigma), was added to a final concentration of 0.2% in the blocking agent to permeabilize cells fixed with paraformaldehyde. All antibody applications of fixed specimens were done in the presence of blocking agent for 1 h at room temperature. For differential labeling of the surface versus the combined intracellular and surface ␣ 2C -adrenoreceptors, cells were labeled with the monoclonal antibody, 12CA5, against the amino terminus HA epitope (BabCo Berkeley Antibody Co., Richmond, CA) at 1:500 dilution for 1 h at 4°C in Dulbecco's modified Eagle's medium with 40 mM HEPES, pH 7.4, and 0.4% bovine serum albumin. After surface labeling, cells were washed three times with ice-cold PBS and fixed with 4% paraformaldehyde for 5 min at room temperature. After fixation, the cells were washed three times with PBS to remove unbound antibody and then treated with a blocking agent with 0.2% Nonidet P-40 for 30 min. The cell preparation was incubated with the affinity purified C4 rabbit polyclonal antibody for 1 h.
Immunological colocalization of the ␣ 2C -adrenoreceptor in PC12 cells and NRK cells was done using antibodies to specific intracellular compartments in conjunction with either the HA epitope antibody, 12CA5, or C4 rabbit polyclonal antibody. The endoplasmic reticulum of the cells was labeled with either BiP, a lumenal protein, or calnexin (Stress-Gen, Victoria, B.C., Canada), an integral membrane protein. The cis/medial Golgi compartment was labeled with a mouse monoclonal antibody recognizing the resident protein, mannosidase II (BabCo Berkeley Antibody Co.). A rabbit polyclonal antibody (William Brown, Cornell University, Ithaca, NY) labeling the cation independent mannose 6-phosphate receptor was used to recognize the trans-Golgi network and the endosomal vesicular compartment. The lysosomal compartment was labeled with the mouse monoclonal antibody that recognizes the lysosomal glycoprotein 120 (Dr. S. Green, University of California, San Francisco, CA).
Cellular cAMP Accumulation Assay-NRK cells and PC12 cells expressing ␣ 2C -adrenoreceptors were each grown to about 70 -90% confluency in 12-well dishes. On the day of the study, the cells were washed twice with PBS with Ca 2ϩ /Mg 2ϩ . The cells were pretreated for 30 min with 1 mM 3-isobutyl-1-methylxanthine in Dulbecco's modified Eagle's medium supplemented with 20 mM HEPES, pH 7.4. Cells were incubated for 5 min with 10 M forskolin in the presence and absence of the ␣ 2 -adrenoreceptor agonist, dexmedetomidine (Orion Corp.), in the above medium. The cellular production of cAMP was determined using a radioimmunoassay [ 3 H]cAMP assay system (Amhersham Pharmacia Biotech). Experiments were performed three times in duplicate.
Western Blotting-Immunoblot analysis of ␣ 2C -adrenoreceptor processing was done using either whole cell lysate or membrane preparation obtained from stably transfected cells as described previously. Whole cell lysate was prepared using PBSTDS buffer (1% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate in PBS, pH 7.4). Cells were grown on a 10-cm dish until 70 -80% confluent and were solubilized with 1 ml of PBSTDS buffer at 4°C for 30 min. Following solubilization, the nuclear fraction of the lysate was removed by high speed centrifugation at 14,000 rpm for 15 min at 4°C. Whole cell lysate from each of the cell lines was stored at Ϫ70°C. Membrane preparations made for ligand binding experiments were solubilized with 1% SDS buffer. Protein concentration for each sample was determined using Bio-Rad protein assay.
For examination of post-translational processing of each cell type, 80 -100 g of membrane protein was solubilized in 1% SDS. The endoglycosidase H reaction buffer was made to have a final concentration of 0.5% Nonidet P-40 and 10 mM sodium acetate, pH 5.0. PNGase F reaction was supplemented with 20 mM sodium phosphate buffer, pH 8.0, and 10 mM EDTA. Both the PNGase F and endoglycosidase H reactions were done at 37°C for 24 h and terminated by the addition of 4 ϫ SDS sample buffer. Samples were loaded and run on 7.5% or 10% SDS-polyacrylamide electrophoresis discontinuous gel. Electrophoresed proteins in the gel were transferred to nitrocellulose membrane and blotted overnight with 5% dry milk, 2% equine donor serum, 20 mM Tris, pH 7.6, 137 mM sodium chloride, and 0.05% Tween 20. ␣ 2C -Adrenoreceptors were labeled with the mouse monoclonal antibody 12CA5 or rabbit polyclonal antibody C4 in blocking solution for 1 h at room temperature. The nitrocellulose membrane was rinsed three times in TBS-Tween and labeled with goat anti-mouse or goat antirabbit secondary antibody conjugated with horseradish peroxidase (Amersham Pharmacia Biotech, dilution 1:1000) in blocking solution. Following three rinses with TBS-Tween, the nitrocellulose membrane was treated with reagents for ECL detection of protein (Amersham Pharmacia Biotech).
Subcellular Membrane Isolation-NRK cells stably expressing ␣ 2cadrenoreceptors were grown in 10-cm dishes to about 80 -90% confluency. Cells were washed three times with ice-cold PBS. The cells were scraped off the dish and lysed with 1 ml of ice-cold hypotonic lysis buffer (20 mM HEPES, pH 7.4, 2 mM EDTA, 2 mM EGTA, 6 mM magnesium chloride, 1 mM phenylmethylsulfonyl fluoride, 10 M leupeptin, 10 M apoprotin, 1 mM benzamindine, 1 M pepstatin). The collected cell lysate was homogenized with 20 strokes of a tight fitting Dounce followed by two 10-s bursts with a Polytron tissue grinder (Beckman Instruments). Cellular debris and unlysed cells were removed by centrifuging at 1,000 ϫ g for 5 min at 4°C. The supernatant was collected and supplemented with 2.0 M sucrose to achieve a final concentration of 0.2 M sucrose. The discontinuous sucrose step gradient was made using the above hypotonic lysis buffer with the addition of sucrose at the following molar concentrations: 0.5, 0.9, 1.2, 1.35, 1.5 and 2.0. Each step in the gradient had a total volume of 5 ml. A total of 2 ml of cell lysate recovered from two 10-cm dishes was applied to the top of the gradient. The samples were centrifuged for 16 h at 27,000 rpm in a Beckman SW28 rotor. The plasma membrane samples were recovered in the sucrose gradient at the interface between 0.5 and 0.9 M. The endoplasmic reticular membrane samples were recovered at the interface between the 1.35 and 1.5 M. Samples were diluted with binding buffer and centrifuged as outlined above. The membrane pellets obtained for the plasma membrane and endoplasmic reticulum were assessed for functional expression of ␣ 2c -adrenceptors by radioligand binding as outlined earlier. Verification of isolated membranes as either plasma membrane or endoplasmic reticulum was achieved with Western blotting methods as outlined above utilizing rabbit polyclonal antibodies that specifically recognize the ␣ subunit of the Na/K-ATPase pump at dilution of 1:1000 (Kindly donated by W. James Nelson, Stanford University) and calnexin at a dilution of 1:5000, respectively.

RESULTS
Immunolocalization of ␣ 2C -Adrenoreceptors-The steady state distribution of HA epitope-tagged ␣ 2C -adrenoreceptor was examined in stably transfected NRK cells, AtT20 cells, and PC12 cells by immunocytochemistry using mouse monoclonal antibody 12CA5 recognizing the HA epitope. In Fig. 1A, NRK cells display faint plasma membrane staining with a predominant intracellular distribution. This pattern of expression of ␣ 2C -adrenoreceptor in NRK cells was observed regardless of the level of cell expression (data not shown). However, both AtT20 cells and PC12 cells expressing ␣ 2C -adrenoreceptors showed strong plasma membrane staining with only a limited intracellular pool confined to a perinuclear vesicular compartment (Fig. 1, B and C, respectively). The above observations were further supported by labeling the extracellular amino terminus of ␣ 2C -adrenoreceptors in viable cells at 4°C with 12CA5 mono-clonal antibody followed by fixing, permeabilizing, and labeling the combined surface and intracellular ␣ 2C -adrenoreceptor distributions with the affinity purified C4 polyclonal rabbit antibody that recognizes the intracellular carboxyl terminus (Fig.  2). There is a large intracellular pool of receptors in NRK cells labeled by the C4 antibody but not the 12CA5 antibody which labels only the surface receptors under these conditions (Fig. 2,  A and B). In contrast, both AtT20 cells and PC12 cells display strong plasma membrane staining (Fig. 2, C-D and E-F, respectively) with both antibodies. While only a small intracellular perinuclear vesicular pool of ␣ 2C -adrenoreceptors is labeled in AtT20 cells and PC12 cells with the C4 antibody alone (Fig. 2,  D and F, respectively). Frequently, the PC12 cells also display a heterogeneous plasma membrane distribution of ␣ 2C -adrenoreceptor with a concentration of receptors in extensions of the plasma membrane (Fig. 2E). This heterogeneous plasma membrane distribution was found to occur in the absence or presence of the ␣ 2 -adrenoreceptor antagonist, atipamizole (data not shown).
Subcellular Distribution of ␣ 2C -Adrenoreceptors-To identify the intracellular compartment containing ␣ 2C -adrenoreceptors FIG. 1. Steady state distribution of HA epitope-tagged ␣ 2C -adrenoreceptors in stably transfected cell lines. Cells were grown on glass coverslips and processed for indirect immunofluorescence microscopy, as outlined under "Experimental Procedures." The mouse monoclonal antibody, 12CA5 (recognizing the HA epitope), was used with detergent (0.2% Nonidet P-40) to allow labeling of both plasma membrane and intracellular distributions of ␣ 2C -adrenoreceptors. In panel A, NRK cells expressing ␣ 2C -adrenoreceptors display a limited plasma distribution of receptors with a large intracellular pool. In neuroendocrine cell lines (AtT20 cells, panel B, and PC12 cells, panel C) ␣ 2Cadrenoreceptors are found predominately in the plasma membrane with a small intracellular perinuclear receptor pool.
FIG. 2. The labeling of ␣ 2C -adrenoreceptors that reside in the plasma membrane is compared with the labeling of total cellular ␣ 2C -adrenoreceptors in NRK, AtT20, and PC112 cells (as described under "Experimental Procedures"). Briefly, viable cells were exposed to the HA eptiope antibody 12CA5 at 4°C to label the surface ␣ 2C -adrenoreceptors in the plasma membrane (panels A, C, and E labeled Nonpermeabilized). The cell preparations were rinsed three times with PBS, fixed, and permeablized prior to labeling the combined plasma membrane and intracellular ␣ 2C -adrenoreceptors with the carboxyl-terminal rabbit polyclonal antibody (panels B, D, and F labeled Permeabilized). In NRK cells expressing ␣ 2C -adrenoreceptors, only a small fraction of receptors are found on the cell surface (panel A) compared with the large intracellular pool of receptors (panel B). Conversely, in the two neuroendocrine cell lines, AtT20 cells (panels C and D) and PC12 cells (panels E and F), there is only a small amount of ␣ 2C -adrenoreceptors in a perinuclear vesicular compartment. ␣ 2C -Adrenoreceptor are primarily found in the plasma membrane for both PC12 cells and AtT20 cells.
in the cell lines studied, we used indirect immunocytochemical methods to compare the distribution of these receptors with markers for the following intracellular compartments: the mannose 6-phosphate receptor, which cycles between the plasma membrane and the trans-Golgi through endosomes; lysosomal glycoprotein 120, a lysosomal membrane protein; BiP, a luminal protein of the endoplasmic reticulum; calnexin, an integral membrane protein of the endoplasmic reticulum; and mannosidase II, a resident cis/medial Golgi protein. In both NRK and PC12 cells, there is no consistent co-localization of the ␣ 2C -adrenoreceptor with the lysosomal marker, lysosomal glycoprotein 120 (Figs. 3, A and B, and 4, A and B, respectively). The trans-Golgi/endosomal marker, the mannose 6-phosphate receptor, lacked any significant overlap of staining with the ␣ 2C -adrenoreceptors in NRK cells (Fig. 3, G-H). However, in PC12 cells there is considerable similarity in the staining of a central perinuclear vesicular pool by ␣ 2C -adrenoreceptors and mannose 6-phosphate receptors (Fig. 4, G-H). The subcellular distribution of the ␣ 2C -adrenoreceptors in NRK cells most closely resembles the distribution of BiP (Fig. 3, C and D) and mannosidase II (Fig. 3, E-F), indicating that most of the intracellular ␣ 2C -adrenoreceptors in NRK cells resides within the endoplasmic reticulum and cis/medial Golgi compartments of cells. In addition, we found that calnexin, an integral membrane protein of the endoplasmic reticulum, only co-localized with the centrally staining ␣ 2C -adrenoreceptors in NRK cells. The more peripheral fine reticular endoplasmic reticulum structure did not colocalize with ␣ 2C -adrenoreceptors (data not shown). In PC12 cells most of the ␣ 2C -adrenoreceptor is in the plasma membrane with focal concentration in plasma membrane extensions but with no significant colocalization with the ER marker BiP (Fig. 4, C-D). In PC12 cells, there is a small perinuclear pool of ␣ 2C -adrenoreceptor that colocalizes with the mannosidase II and mannose 6-phosphate receptor for Golgi complex and endosomal compartment (Fig. 4, E-F and  G-H, respectively).
Functional Receptor Expression-The large intracellular pool of ␣ 2C -adrenoreceptors found in NRK cells may or may not be functional since misfolded proteins tend to accumulate in the ER. To determine the functional state of the total plasma membrane and intracellular ␣ 2C -adrenoreceptors in NRK cells and PC12 cells, we compared receptor antigenic expression and receptor ligand binding in crude membrane preparations from these two cell lines. By Western blot analysis, it was found that Co-localization studies were performed using individual antibody markers for distinct intracellular compartments in conjunction with either the mouse monoclonal antibody, 12CA5, recognizing the HA epitope or the rabbit polyclonal antibody, C4, recognizing the carboxyl terminus. A, NRK cell labeled with the C4 antibody for the ␣ 2C -adrenoreceptor. B, the same cell in panel A labeled with the antibody lysosomal glycoprotein 120 (Lpg120), for lysosomal vesicles. There is very little overlap in the staining of cells in panels A and B. Conversely, the NRK cell in panels C and D display a high degree of similarity between intracellular ␣ 2C -adrenoreceptors labeled with 12CA5 antibody and the endoplasmic reticulum labeled with the antibody recognizing BiP, respectively. In addition, there is similar labeling of a perinuclear component of the intracellular ␣ 2C -adrenoreceptors and the cis-Golgi regions of cells labeled with antibody recognizing mannosidase II (panels E and F, respectively). There was no significant co-localization of intracellular ␣ 2C -adrenoreceptors in NRK cells with the antibody recognizing the protein, mannose 6-phosphate receptor (M6PR), present in the trans-Golgi and endosomes (panels G and H, respectively).

FIG. 4. Immunolocalization of intracellular ␣ 2C -adrenoreceptors in permeabilized PC12 cells.
Co-localization studies performed using individual antibody markers for distinct intracellular compartments in conjunction with either the mouse monoclonal antibody, 12CA5, recognizing the HA epitope or the rabbit polyclonal antibody, C4, recognizing the carboxyl terminus. In panels A and C, PC12 cells labeled for the ␣ 2C -adrenoreceptor do not show any similarity in staining with the same cells labeled with antibodies for lysosomal vesicles (Lpg120) or endoplasmic reticulum (BiP) (panels B and D, respectively). However, there is considerable overlap in the localization of ␣ 2C receptors and mannosidase II, a cis/medial Golgi marker (panels E and F), and the mannose 6-phosphate receptor, a trans-Golgi network marker (panels G and H).
80 g of NRK cell membrane and 80 g of PC12 cell membrane contained approximately equivalent amounts of receptor antigen (Fig. 5A). Similar results were obtained with the affinity purified rabbit polyclonal antibody C4 (data not shown). Saturation binding with the ␣ 2 -adrenoreceptor radioligand, RX821002, revealed identical K D values and similar B max values (in picomole per milligram of membrane protein): AtT20 cells, 1.357 Ϯ 0.065; NRK cells, 1.085 Ϯ 0.043; PC12 cells, 0.962 Ϯ 0.17. Taken together these results show that PC12 cells and NRK cells express comparable levels of ␣ 2C -adrenoreceptor protein and have comparable numbers of ␣ 2C -adrenoreceptor-binding sites, suggesting that the large pool of the ␣ 2C -adrenoreceptor resides within the ER in NRK cells includes functional receptor.
To further examine the functional status of the ␣ 2c -adrenoreceptors located in the endoplasmic reticulum of NRK cells, we used cell fractionation to separate plasma membranes from endoplasmic reticulum membranes. Using discontinuous sucrose gradient, we were able to obtain membranes highly enriched for either plasma membrane or endoplasmic reticulum. Plasma membranes were isolated from the gradient at the interface between 0.5 and 0.9 M sucrose (fraction 1) while endoplasmic reticular membranes were isolated at the interface between 1.35 and 1.5 M sucrose (fraction 2). The upper panel of Fig. 5B shows the Western blot of the two fractions of pooled membranes from several experiments probed with rabbit polyclonal antibody for Na/K-ATPase, a plasma membrane marker. This blot shows that fraction 1 isolated from NRK cells stably expressing ␣ 2c -adrenoreceptors is enriched for Na/K-ATPase while no detectable Na/K-ATPase is detected in an equivalent amount of membrane protein from fraction 2. The blot was then stripped and reprobed with rabbit polyclonal antibody for the protein, calnexin, an endoplasmic reticular membrane maker (Fig. 5B, lower panel). Fraction 2 is enriched in calnexin, while no detectable calnexin is detected in an equivalent amount of membrane protein from fraction 1. Both plasma membrane and endoplasmic recticulum membrane fractions isolated from NRK cells stably expressing ␣ 2c -adrenoeptors possessed functional receptors as determined by radioligand binding, 0.89 Ϯ 0.080 and 0.136 Ϯ 0.036 pmol/mg protein, respectively.
Post-Translational Processing of ␣ 2C -Adrenoreceptors-The potential role of post-translational processing in the targeting of ␣ 2C -adrenoreceptor was examined for each of the three cell lines using Western blot analysis. As can be seen from Fig. 6, there is a small difference in the molecular weight of the ␣ 2C -adrenoreceptor expressed in the three stably transfected cell lines. ␣ 2C -Adrenoreceptors derived from PC12 cells and NRK cells have a slightly higher average molecular weight 80 kDa than those derived from AtT20 cells 76 kDa. The type and level of protein glycosylation reflects the extent of its posttranslation processing. We examined the sensitivity of ␣ 2Cadrenoreceptors made in each cell line to endoglycosidase H and N-glycosidase F (PNGase F) digestions. As a positive control for the endoglycosidase H deglycosylation reaction, ␣ 2Cadrenoreceptors were transiently transfected into COS 7 cells. Since transiently transfected COS7 cells are capable of very high expression levels, much of the ␣ 2C -adrenoreceptors should be found in various stages of the biosynthetic pathway and a significant portion should be sensitive to endoglycosidase H digestion. As shown in Fig. 6, a significant portion of the expressed ␣ 2C -adrenoreceptor in COS 7 cells is sensitive to endoglycosidase H digestion. In NRK cells expressing ␣ 2C -adrenoreceptors, there was only a small proportion sensitive to endoglycosidase H treatment producing a smaller sharp band appearing at M r ϭ 46 kDa, the predicted molecular weight of unglycosylated ␣ 2C -adrenoreceptor (Fig. 6). This observation indicates that most of the ␣ 2C -adrenoreceptors present have been exposed to the environment of the cis/medial Golgi compartment making them endoglycosidase H resistant. The large intracellular pool of ␣ 2C -adrenoreceptors in NRK cells is likely to be cycling between the endoplasmic reticulum and cis/medial Golgi compartments with limited transport of receptors to the plasma membrane. In both PC12 cells and AtT20 cells, most of Increasing amounts of membrane protein from NRK cells expressing ␣ 2C -adrenoreceptors were run along with a fixed amount of membrane protein from PC12 cells expressing ␣ 2C -adrenoreceptors. NRK and PC12 cells express ␣ 2C -adrenoreceptor protein at comparable levels. Similar results were obtained in three independent experiments with both the mouse monoclonal antibody 12CA5 and affinity purified rabbit polyclonal antibody C4. B, isolation and purification of endoplasmic reticular and plasma membranes. Membrane fractions were collected from interfaces between 0.5 and 0.9 M sucrose (Fraction 1) and 1.35 and 1.5 M sucrose (Fraction 2). Membrane fraction 1 was enriched for plasma membrane protein, Na/K-ATPase, and depleted of endoplasmic recticular membrane protein, calnexin. Conversely, the membrane fraction 2 was enriched for endoplasmic recticular membrane protein, calnexin, and depleted of plasma membrane protein, ␣ϪNa/K-ATPase. C and D, inhibition of forskolin stimulated adenylyl cyclase by ␣ 2C receptor in NRK cells (panel C) and in PC12 cells (panel D). Dex is the non-subtype selective ␣ 2 agonist dexmedetomidine.
the ␣ 2C -adrenoreceptors present at steady state are resistant to endoglycosidase H digestion. PNGase F was capable of deglycosylating the ␣ 2C -adrenoreceptor expressed in all of the cell lines (Fig. 6).
Development of ␣ 2C -Adrenoreceptor Microdomains in PC12 Cells-The treatment of PC12 cells with NGF resulted in neuronal induction of cells and the development of neurite extensions having focal accumulations of the synaptic vesicle marker SV2 in the terminal regions of these extensions (Data not shown). We labeled the cell surface and total cellular receptor distributions in PC12 after NGF treatment for 12 (Fig. 7, A and  B) or 48 h (Fig. 7, C and D). The induction of a neuronal phenotype in PC12 cells did not alter the targeting to the plasma membrane of ␣ 2C -adrenoreceptors. However, the plasma membrane distribution of ␣ 2C -adrenoreceptors in PC12 cells became more heterogeneous with focal accumulations of receptors in the plasma membrane concentrated at peripheral margins and membrane extensions of cells following exposure to NGF for 12 h (Fig. 7). Prolonged treatment with NGF for 48 h resulted in the development of neurite extensions having focal accumulations of ␣ 2C -adrenoreceptors at the tips (Fig. 7, C  and D). These focal accumulations of ␣ 2C -adrenoreceptors in the plasma membrane have not been observed in any other cell line studied. Interestingly, PC12 cells displaying focal accumulations of ␣ 2C -adrenoreceptors also have increased amounts cytoskeletal protein, F-actin, localized with the receptor (Fig. 8,  A and B). The normal homogeneous plasma membrane labeling of F-actin by phalloidin is altered in PC12 cells. DISCUSSION Given the overall high structural and functional similarity among three ␣ 2 -adrenergic receptor subtypes, it is very interesting that there are distinctly different cellular targeting characteristics, physiologic functions, and tissue expression patterns for each of the subtypes (for reviews, see Ref. 5). Unlike the ␣ 2C -adrenoreceptor, both the ␣ 2A -and ␣ 2B -adrenergic receptors target efficiently to the plasma and lack an intracellular ER distribution (1, 2, 6). The ␣ 2A -adrenoreceptors are found in many tissues including platelets where they modulate platelet aggregation and in the central nervous system where they influence pain perception and blood pressure (18,19). The ␣ 2B -adenoceptors subtype is found in vascular smooth muscle where their activation leads to elevation of systemic blood pressure (20). The tissue distribution of ␣ 2C -adrenoreceptors is primarily confined to the central nervous system (21,22). Recent evidence from transgenic and knockout mouse models show that the ␣ 2C -adrenoreceptors modulate several aspects of behavior (5,14,15). Both ␣ 2A and ␣ 2C receptors have been shown to regulate catecholamine release from sympathetic nerve terminals (16,17). These studies showed that both the ␣ 2A and ␣ 2C subtypes are required for normal presynaptic control of transmitter release from sympathetic nerves in the heart and from central noradrenergic neurons. ␣ 2A receptors inhibit transmitter release at high stimulation frequencies FIG. 6. The post-translational processing of HA epitope-tagged ␣ 2C -adrenoreceptors in AtT20, COS7, NRK, and PC12 cells was examined by Western blotting. Membrane preparations of cells expressing ␣ 2C -adrenoreceptor were run on a 9 or 10% SDS-polyacrylamide gel and transferred to nitrocellulose. The blots were probed for ␣ 2C -adrenoreceptors with the HA epitope antibody 12CA5. Membrane preparations were treated with either endoglycosidase H or PNFase F to remove asparagine (N) linked glycosidic residues. In panel A, COS7 cells expressing ␣ 2C -adrenoreceptors show a predominant band of receptor and multiple minor bands possibly due to different levels of glycosylation or receptor aggregation. The ␣ 2C -adrenoreceptors made by COS7 cells are sensitive to both endoglycosidase H and PNGase F digestion as seen by the shift in size of the predominant band. In contrast, endoglycosidase H digestion did not significantly change the mobility of the majority of ␣ 2C -adrenoreceptors expressed AtT20 cells, NRK cells or PC12 cells (panels B-D). whereas the ␣ 2C subtype modulates neurotransmission at lower levels of nerve activity. Both low and high frequency regulation appear to be physiologically important as mice lacking both receptor subtypes have elevated plasma norepinephrine levels and develop cardiac hypertrophy with decreased left ventricular contractility by 4 months of age (16). Since the expression of ␣ 2C -adrenoreceptors is primarily in neurons it is possible that neuronal cells may possess the appropriate factors/proteins that not only allow efficient delivery and targeting of receptors to the plasma membrane but more importantly to specific pre-synaptic and post-synaptic domains in their plasma membrane.
The large intracellular pool of ␣ 2C -adrenoreceptors in NRK cells may be incorrectly folded and non-functional and thereby retained in the endoplasmic reticulum. It has been shown that some misfolded proteins tend to be retained by BiP, an endoplasmic reticulum chaperone protein, and retrieved from early Golgi regions of cells by BiP as well (23). Interestingly, by indirect immunolocalization we have found that most of the intracellular ␣ 2C -adrenoreceptors are restricted to a domain of the ER where BiP is localized. These observations would suggest that the intracellular ␣ 2C -adrenoreceptors in NRK cells are non-functional. However, we have determined that the combined intracellular and plasma membrane fractions of NRK cells and PC12 cells are comparable in expression by Western blotting and in function by saturation binding experiments. Furthermore, we were able to isolate plasma membrane and endoplasmic reticulum membrane fractions from NRK cells stably expressing ␣ 2c -adrenoreceptors and show functional expression of receptors by radioligand binding in both fractions. We found the density (in picomole/milligram) of ␣ 2C -adrenoreceptors in the endoplasmic reticulum to be lower than in the plasma membrane even though there appears to be more receptor antigen, as detected by immunofluorescence, in the endoplasmic reticulum. These results are not surprising given the limited distribution of the ␣ 2c -adrenoreceptor within the endoplasmic reticulum. The intracellular ␣ 2c -adrenoreceptors expressed in NRK cells was largely confined to the cis/ medial Golgi and a subdomain of the endoplasmic reticulum where BiP is found. Hence the lower level of ␣ 2c -adrenoreceptor binding found with intracellular endoplamsmic recticular membranes may simply reflect the dilution of the limited intracellular ␣ 2c -adrenoreceptor population within the larger endoplasmic reticulum membrane pool. Therefore, the results of our study suggest that intracellular ␣ 2C -adrenoreceptors expressed in NRK cells, a non-neuronal cell line, are probably not misfolded, but are retained in the endoplasmic reticulum and cis/medial Golgi by some other mechanism. Hence, the poor plasma membrane targeting of ␣ 2C -adrenoreceptors in cell lines such as NRK cells, Rat1 fibroblasts, and Madin-Darby canine kidney cells cannot be accounted for by simply the retention of misfolded/non-functional protein within the ER. Interestingly, while NRK cells expressing ␣ 2C have less plasma membrane expression of receptor, they displayed greater ␣ 2Cadrenoreceptor mediated inhibition of forskolin-induced cAMP production than PC12 cells expressing ␣ 2C -adrenoreceptors. These results are likely due to differences in both the types and relative level of expression of adenylyl cyclases and G-proteins in the two cell lines. This data further supports functional expression of ␣ 2C -adrenoreceptors by both NRK cells and PC12 cells.
An alternative mechanism that may account for the accumulation of ␣ 2C -adrenoreceptors in the endoplasmic reticulum of NRK cells is slow processing of receptors with a concurrent short plasma membrane half-life. Transfection of cells with a powerful promoter such as the cytomegalovirus promoter, present in pCDNA3, can potentially raise protein expression by several orders of magnitude over normal physiologic or endogenous expression. This may overwhelm the cells biosynthetic pathways with the result being a large accumulation of intracellular protein. Given the large intracellular pool of ␣ 2C -adrenoreceptors present in NRK, we expected to find that a large fraction of receptors would be sensitive to endogylcosidase H digestion. Analysis of the asparagine-linked glycosylation of ␣ 2C -adrenoreceptors in the cell lines studied was able to detect immature endogylcosidase H-sensitive receptor in COS7 cells, but did not reveal significant amounts of endogylcosidase Hsensitive receptor in AtT20 cells, PC12 cells, or NRK cells. This would indicate that most of the intracellular ␣ 2C -adrenoreceptors have been processed in the cis/medial Golgi. Furthermore, it suggests that the ␣ 2C -adrenoreceptors expressed in NRK cells are not static within the ER but are actively retrieved from the early Golgi regions and returned to the ER of the cell possibly by a chaperone protein such as BiP. Thus, differences in the rate of protein processing cannot account for the observed differences in plasma membrane targeting.
We have previously demonstrated that the plasma membrane fraction of ␣ 2C -adrenoreceptors expressed in NRK cells is stable for several hours with minimal agonist-induced internalization (2). In addition, Wozniak and Limbird (6) reported that the expression of ␣ 2C -adrenoreceptors in Madin-Darby canine kidney cells to be limited to the basolateral domains of the plasma membrane intracellular pool of receptor (6). The more efficient delivery of the ␣ 2C -adrenoreceptors to the plasma membrane and the lack of a large intracellular pool of receptors in the neuroendocrine cell lines, PC12 cells and AtT20 cells, suggests that these cells express a factor(s) that facilitate either the proper processing and/or targeting to the plasma membrane. Alternatively, these cells may lack a factor(s) that retains the ␣ 2C receptor in the biosynthetic pathway.
PC12 cells undergo neuronal differentiation in response to NGF treatment with the accumulation of synaptic vesicles in peripheral neurite extensions. The neuronal induction of PC12 cells did not affect the targeting of ␣ 2C -adrenoreceptor to the plasma membrane. However, after short periods of NGF treatment, the ␣ 2C -adrenoreceptor distribution in the plasma membrane displayed clustering in regions where membrane extensions were present. These regions of the membrane extensions also display increased concentration of F-actin as labeled by fluorescein isothiocyanate-conjugated phalloidin D. We have occasionally observed focal accumulations of ␣ 2C receptors in PC12 cells that have not been treated with NGF. This may be due to the fact that some PC12 cell may be partially differentiated without NGF treatment. These focal accumulations of ␣ 2C -adrenoreceptors were never observed in any of the non- neuronal cell lines examined. The localization of cytoskeletal proteins, such as F-actin and microtubules, with receptors has been shown to occur in post-synaptic regions of the plasma membrane in neurons (24,25) and in neuromuscular junctions (26) where the receptors are anchored to the cytoskeleton via linker/clustering proteins such as gephyrin, ␣-actinin-2, PSD-95, and rapsyn/43k proteins. Recent studies have found ␣ 2Cadrenergic receptors in both pre-synaptic (21) and post-synaptic densities (21) depending on the location and type of neurons in the central nervous system. Since the neuroendocrine cell line, PC12 cells, possess some functional characteristics of neurons, it is likely that they not only express factors/proteins that facilitate the efficient targeting of ␣ 2C -adrenoreceptor to the plasma membrane but more importantly the development of specialized receptor targeting and signaling domains in the plasma membrane.
It has been suggested that G protein-coupled receptors are located in plasma membrane microdomains along with specific G-proteins and effector molecules (27). Thus, by directing the receptor to a specific microdomain, the cell may more efficiently control the effector systems modulated by that receptor. In PC12 cells, ␣ 2C -adrenoreceptors are found concentrated at the developing neurite extensions having accumulations of F-actin. The targeting of receptors in neurons in these locations has been shown to have a role in directional chemotaxis of developing neurite extensions (28). Further comparison of other signal transduction components and regulation of ␣ 2C -adrenoreceptor trafficking in PC12 cells and other cells such as NRK cells may provide clues to the functional importance of the cell-specific trafficking of G protein-coupled receptors.