The Three (cid:97) 2 -Adrenergic Receptor Subtypes Achieve Basolateral Localization in Madin-Darby Canine Kidney II Cells via Different Targeting Mechanisms*

The present studies examined the localization of the (cid:97) 2B - and (cid:97) 2C -adrenergic receptor (AR) subtypes in polar- ized Madin-Darby canine kidney cells (MDCK II) and the mechanisms by which this is achieved. Previously we demonstrated that the (cid:97) 2A AR subtype is directly delivered to lateral subdomain of MDCK II cells. Surface biotinylation strategies demonstrated that the (cid:97) 2B AR, like the (cid:97) 2A AR, achieves 85–90% basolateral localization at steady-state. However, in contrast to the (cid:97) 2A AR, this polarization occurs after initial random insertion of the (cid:97) 2B AR into both apical and basolateral surfaces fol- lowed by selective retention on the lateral subdomain ( t 1 ⁄ 2 on the apical surface is 15–30 min; t 1 ⁄ 2 on the basolat- eral surface is 8–10 h). The (cid:97) 2C AR also is enriched on the basolateral surface at steady-state and, like the (cid:97) 2A AR, is directly delivered there. Morphological evaluation of the epitope-tagged (cid:97) 2A AR, (cid:97) 2B AR, and (cid:97) 2C AR subtypes by laser confocal microscopy not only corroborated the biochemically-defined basolateral localization of all three (cid:97) 2 AR subtypes but also revealed that the (cid:97) 2C AR uniquely exists in an intracellular compartment(s) as well. Immunofluorescence due to intracellular (cid:97) 2C AR partially overlaps that due to calnexin, a marker representation of differential targeting mechanisms that lead to the basolateral localization of all three (cid:97) 2 AR subtypes in MDCK II cells. This schematic diagram compares the targeting mechanism of the (cid:97) 2A AR, (cid:97) 2B AR, and (cid:97) 2C AR subtypes, and summarizes the differing trafficking or steady-state localization observed for each subtype. The (cid:97) 2A AR is delivered directly to the basolateral membrane (present data, Ref. 8). The present findings have revealed that the (cid:97) 2B AR is inserted randomly into both cell surface domains and is retained preferentially on the basolateral surface. At steady-state, the (cid:97) 2A AR and (cid:97) 2B AR are detected exclusively on the lateral surface of MDCK II cells, whereas the (cid:97) 2C AR is distributed between cell surface and intracellular compartments. However, the fraction of the (cid:97) 2C AR that is localized to the basolateral surface is delivered there directly.

ing inhibition of adenylyl cyclase, activation of receptor-operated K ϩ channels, and suppression of voltage-gated Ca 2ϩ channels (1).
The ␣ 2 AR has been demonstrated in the kidney of several mammalian species (2,3). Although these receptors are concentrated in the proximal tubular segment of the nephron, they also are found in the glomerulus, thin descending limb of Henle's loop and cortical collecting duct. The physiological function of the ␣ 2 AR in the proximal tubule is to increase Na ϩ reabsorption and proton secretion via the modulation of Na ϩ /H ϩ exchange (2,4). The precise physiological role for each of the ␣ 2 AR subtypes is not yet defined.
The expression of multiple ␣ 2 AR subtypes in the kidney (5) and the effects of adrenergic agents on renal physiology encouraged us to explore one determinant of receptor-mediated function, namely the precise localization of the receptor molecules in the plasma membrane of renal epithelial cells. Renal epithelial cells are polarized both morphologically and functionally into at least two distinct compartments: apical and basolateral. Both radioligand binding studies and renal microperfusion experiments are consistent with the interpretation that ␣ 2 ARs are present primarily on the basolateral membrane domain of renal epithelial cells in vivo (6,7). How this localization is achieved for the ␣ 2 AR, or for any G-protein-coupled receptor, is only beginning to be revealed.
Recent studies from our laboratory have examined the targeting and retention of wild-type and epitope-tagged ␣ 2A AR in cultured Madin-Darby canine kidney (MDCK II) cells, a model system that achieves morphological and functional polarity following culture on Transwell filters (8). Surface biotinylation strategies demonstrated that 85-90% of the wild-type ␣ 2A AR population is localized on the basolateral surface of MDCK II cells, in a manner analogous to the localization of this receptor in vivo. Furthermore, immunochemical detection of the ␣ 2A AR, based on the recognition of an artificial epitope tag introduced by mutagenesis into the amino terminus of the receptor, has revealed that the ␣ 2A AR is not randomly distributed in the basolateral surface but is highly enriched on the lateral subdomain of polarized MDCK II cells. This localization is achieved by direct delivery to this surface, based on findings from metabolic labeling studies. Extensive mutational analysis of the ␣ 2A AR to reveal structural regions that confer direct ␣ 2A AR delivery to the basolateral surface suggests that basolateral targeting of ␣ 2A AR does not rely on amino-terminal glycosylation, carboxyl-terminal acylation, nor amino acid sequences within the third cytoplasmic loop and the carboxylterminal tail (9). These findings are consistent with the interpretation that sequences in or near the lipid bilayer are involved in the delivery of the ␣ 2A AR to the basolateral surface. In contrast, the deletion of the predicted third cytoplasmic loop of the ␣ 2A AR significantly decreases its half-life on the epithelial cell surface, suggesting that this structural region of ␣ 2A AR * This work was supported in part by National Institutes of Health Research Grant DK 43879. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Supported by a postdoctoral fellowship from the National Kidney Foundation.
The existence of three subtypes of ␣ 2 AR that possess regions of structural similarity and diversity provides natural reagents to further understand the structural regions of G-protein-coupled receptors that confer information for targeting and stabilization in polarized cells. The three ␣ 2 AR subtypes (␣ 2A AR, ␣ 2B AR, and ␣ 2C AR) can be distinguished not only by differences in pharmacological specificity but also by differences in their primary amino acid sequences. Although sequences in the seven transmembrane-spanning domains are highly conserved among the ␣ 2 AR subtypes, the large third cytoplasmic loop and the amino and carboxyl termini of the ␣ 2 AR subtypes demonstrate significant sequence dissimilarity (10 -13). In addition, the post-translational modification of these three receptors is quite different. For example, the ␣ 2A AR has sequences (10,11) which confer amino-terminal glycosylation (14) and carboxylterminal acylation (15). In contrast, the ␣ 2B AR is not glycosylated (13,14,16), but its carboxyl terminus does contain the sequences appropriate for acylation, whereas the ␣ 2C AR does not contain acylation signals but does possess glycosylation signals in the amino terminus (12).
The structural differences among the ␣ 2 AR subtypes led us to explore the possibility of a different targeting mechanism for each subtype, possibly resulting in differential receptor subtype localization. The present findings demonstrate that although the ␣ 2A , ␣ 2B , and ␣ 2C AR subtypes ultimately achieve basolateral localization at steady-state, they do so by different molecular mechanisms. Thus, the ␣ 2C AR shares the direct basolateral delivery pathway characteristic of the ␣ 2A AR, but is not exclusively localized on the cell surface at steady-state. The ␣ 2B AR, which achieves almost complete basolateral localization at steady-state, does so following apparent random delivery to both surfaces and prolonged retention on the basolateral surface.

Materials
The ␣ 2B AR (RNG) and ␣ 2C AR (RG10) cDNAs were kindly provided by Drs. Kevin Lynch (University of Virginia) and Stephen Lanier (University of South Carolina), respectively. The protein A-purified 12CA5 monoclonal antibody was obtained from the Berkley Antibody Co. (BABCO), Affi-Gel 10 and Affi-Gel 15 were purchased from Bio-Rad, and the Cy3-conjugated donkey anti-mouse IgG was from Jackson Immunochemicals. The anti-mannosidase II polyclonal antibody (17) was a kind gift from Drs. Marilyn Farquhar (University of California in San Diego) and Kelley Moremen (University of Georgia). The anticalnexin polyclonal antibody (18)  The ␣ 2B and ␣ 2C ARs were epitope-tagged at either the amino or carboxyl terminus. The hemagglutinin TAG was introduced into the amino termini of the ␣ 2B and ␣ 2C ARs (termed TAG-␣ 2B AR and TAG-␣ 2C AR) by site-directed mutagenesis in pBluescript II vector after initial isolation of single-stranded DNA. The hemagglutinin epitope TAG is encoded by the first 9 of the 11 added amino acids (YPYDVPDYALA). Mutants containing the TAG sequence were confirmed by dideoxy-DNA sequencing using T7 DNA polymerase and [␣-35 S]dATP. Appropriate mutant structures were then subcloned into the polylinker region of pCMV4 mammalian expression vector (8) and the correct DNA sequence was confirmed again.
Introduction of the epitope tag into the carboxyl termini of the ␣ 2B and ␣ 2C ARs (termed ␣ 2B -TAG-AR and ␣ 2C -TAG-AR) was achieved using polymerase chain reaction (PCR)-based mutagenesis. The ␣ 2B AR 5Ј oligonucleotide (24-mer) was 5Ј-GGCTACTGCAACAGCTCTTTGAAC. The ␣ 2B AR 3Ј oligonucleotide containing the hemagglutinin TAG sequence, the stop codon, and an EcoRV restriction site (82-mer with 24 bases annealing) was 5Ј-CTAGGATATCTCAACGAGGAGCTAGCG-CGTAGTCAGGAACGTCGTAAGGATAGCTAGCCCAGCCAGTCTG-GGTCCACGGCCG. The ␣ 2C AR 5Ј oligonucleotide (24-mer) was 5Ј-CG-CATCTACCGCGTGGCCAAGCTG. The ␣ 2C AR 3Ј oligonucleotide containing the hemagglutinin TAG sequence, the stop codon, and a SmaI restriction site (82-mer with 24 bases annealing) was 5Ј-CTAGCC-CGGGTCAACGAGGAGCTAGCGCGTAGTCAGGAACGTCGTAAGG-ATAGCTAGCCTGCCTGAAGCCCCTTCTCCTCCT. The 82-mer oligonucleotides were purified on a 9% PAGE, the appropriate bands were excised, eluted, and passed over C18 Sep-Pak columns. The 24-mer oligonucleotides were purified on C18 Sep-Pak columns. The PCR reactions were done at high denaturing temperature (98°C) in the presence of 5% dimethyl sulfoxide following a "hot start" in the absence of DNA polymerase at 90°C for 30 min. Thermal stable VENT DNA polymerase was required for these PCR reactions due to a high GC content of the DNA sequences encoding the predicted third cytoplasmic loop of the ␣ 2 AR subtypes. In case of each ␣ 2 AR subtype, doublestranded DNA template (5 ng) was used with the respective oligonucleotides at a final concentration of 0.5 M. The incubation also contained 0.25 mM dNTPs, 1 ϫ VENT DNA polymerase buffer, and 1 unit of VENT DNA polymerase in a final volume of 100 l with a top layer of mineral oil. The PCR conditions were as follows: denaturing at 98°C for 1 min, annealing at 68°C for 1 min, extension at 72°C for 2 min for 35 cycles, followed by a 10-min extension at 72°C and storage at 4°C. Obtained PCR products were purified on agarose gels and digested with the appropriate restriction enzymes; ␣ 2B -TAG-AR PCR product (167 bp) was cut with EcoRV and BstBI, and ␣ 2C -TAG-AR PCR product (744 bp) was digested with SmaI and MluI. Obtained ␣ 2B -TAG-AR (99 bp) and ␣ 2C -TAG-AR (740 bp) fragments were purified on agarose gels and subcloned into the pCMV4-␣ 2B AR plasmid (in place of the EcoRV-BstBI fragment encoding wild type sequence) and pCMV4-␣ 2C AR expression plasmid (instead of the wild type SmaI-MluI fragment), respectively.
The sequence of all of these mutant constructs was confirmed by double-stranded dideoxy-DNA sequencing using T7 DNA polymerase and [␣-35 S]dATP. Transient transfection of COS M6 cells (100-mm dish plated at 1 ϫ 10 6 cells/dish) with 10 g of each expression plasmid utilizing the DEAE-dextran method (9) was employed to assess the level of receptor expression using [ 3 H]yohimbine binding as well as to examine the detectability of the epitope TAG by immunocytochemistry prior to utilization of the DNA to develop permanent transformants.

Creation of Permanent Clonal MDCK II Cell Lines
Permanent clonal cell lines were developed using the CaPO 4 precipitation method, as described previously (8). Briefly, 20 g of the expression plasmid were co-transfected with 2 g of pRSVneo into 1 ϫ 10 6 MDCK II cells. Colonies expressing the neomycin resistance gene were selected with the neomycin analog G418 (500 active g/ml), isolated, and screened for the expression of ␣ 2 AR by measuring the binding of the ␣ 2 AR antagonist [ 3 H]rauwolscine and the ␣ 2 AR agonist p-

Cell Cultures
Madin-Darby canine kidney (MDCK II) cells, kindly provided by Dr. Enrique Rodriquez-Boulan (Cornell University), were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 100 units/ml penicillin, and 100 g/ml streptomycin at 37°C, 5% CO 2 . MDCK II cells were plated on polycarbonate membrane filters (Transwell) at a density of 1 ϫ 10 6 cells/24.5-mm Transwell and maintained in culture for 7-8 days, changing medium every day, at which time the cells achieved a morphologically and functionally polarized phenotype (8). In the present studies, we examined multiple MDCK II clonal cell lines as summarized in Table I.

Leak Assay
Leak assays were performed to determine whether the MDCK II cells developed tight junctions, and served as an indication of their integrity as a monolayer on Transwell filters. [ 3 H]Methoxyinulin was added to the growth medium in the apical compartment and the leak of this radioligand from the apical to basolateral compartment was assessed after a 1-h incubation at 37°C. A typical [ 3 H]methoxyinulin leak ranged from 1 to 2%. Transwells exhibiting leaks greater than 2% were excluded from study.
Receptor Localization at Steady-state a) Biochemical Approach Utilizing the Surface Biotinylation Strategy-The distribution of the ␣ 2 AR on the apical versus the basolateral membrane domain in MDCK II permanent clonal cell lines can be quantitated using surface biotinylation strategies as described previously (8,9). Briefly, glycoproteins (which include the ␣ 2A AR and ␣ 2C AR) on either the apical or the basolateral surface of MDCK II clones grown on Transwell cell culture system were covalently modified with biotin LC-hydrazide, while primary amines of the non-glycosylated ␣ 2B AR were covalently labeled with sulfo-NHS-biotin. Following membrane preparation, ␣ 2 ARs were identified by covalent modification with the photoactivatable ␣ 2 AR-selective ligand [ 125 I]Rau-AzPEC for 1 h at 15°C in the dark. Photolabeling not attributable to receptor binding was determined in parallel incubations carried out in the presence of 10 M phentolamine, an ␣-adrenergic antagonist. The photoaffinity-labeled receptor was then extracted with RIPA buffer (150 mM NaCl, 50 mM Tris, pH 8.0, 0.5 mM EDTA, 1% Nonidet P-40, 0.5% deoxycholate, and 0.1% SDS) and streptavidin-agarose chromatography was employed to isolate the previously biotinylated molecules. The fraction of the biotinylated ␣ 2 AR present on the apical versus the basolateral membrane domain was determined following SDS-PAGE, autoradiography, and cutting and counting of radiolabeled bands corresponding to the migration of the ␣ 2A AR, ␣ 2B AR, and ␣ 2C AR.
b) Morphological Approach Employing Immunocytochemistry-Clonal cell lines grown on polycarbonate filters were excised from the Transwell support, rinsed with phosphate-buffered saline (PBS), and fixed for 30 min at 22°C (room temperature) with 4% (v/v) paraformaldehyde in PBS (␣ 2A AR and ␣ 2B AR) or 15 min with 2% (v/v) paraformaldehyde in PBS (␣ 2C AR). Cells were rinsed twice with PBS, incubated in 50 mM NH 4 Cl solution in PBS for 15 min to quench the fixative, and then permeabilized with 0.2% Triton X-100 in PBS for 15 min at room temperature. Potential sites for nonspecific antibody binding were blocked by a 15-min incubation with 1% BSA, 0.2% Triton X-100, 0.04% sodium azide in PBS at room temperature. The primary antibody, affinity-purified (see below) 12CA5 mouse monoclonal antibody directed against the hemagglutinin epitope, was diluted 1:25 in 1% BSA, 0.2% Triton X-100, 0.04% sodium azide in PBS and incubated with the sample for 1 h at room temperature. This was followed by four 15-min washes with 0.2% Triton X-100 in PBS and a 1-h incubation with a secondary Cy3-conjugated donkey anti-mouse IgG (1:100 dilution in 1% BSA, 0.2% Triton X-100, 0.04% sodium azide in PBS). The cells were washed as before with PBS containing 0.2% Triton X-100 and mounted on glass microscope slides with Aqua Polymount. The results were analyzed either on a Leitz Fluorescent Microscope using a 40 ϫ oil immersion lens or on a Zeiss laser confocal microscope.
Immunological co-localization studies were done essentially as described above, although expanded by the additional application of another primary and secondary antibody. Incubation with 1:400 dilution of an anti-calnexin rabbit polyclonal antibody (used as a marker for the endoplasmic reticulum (18)), or a 1:1000 dilution of anti-mannosidase II rabbit polyclonal antibody (used to detect trans-Golgi network (17)) in 1% BSA, 0.2% Triton X-100, 0.04% sodium azide followed the initial detection of the ␣ 2C AR with 12CA5 antibody and FITC (1:60 dilution) or Cy3 (1:100 dilution)-conjugated donkey anti-mouse IgG. The anti-calnexin or anti-mannosidase II primary antibodies were incubated with the sample for 1 h at room temperature, followed by four 15-min washes with 0.2% Triton X-100 in PBS and a 1-h incubation with a secondary donkey anti-rabbit FITC (1:60 dilution) or Cy3 (1:200 dilution)-conjugated IgG in 1% BSA, 0.2% Triton X-100, 0.04% sodium azide in PBS. The cells were then washed four times for 15 min with PBS, 0.2% Triton X-100, mounted, and analyzed as described above.
Since the initial immunoanalysis of ␣ 2C AR localization in MDCK II cells revealed a significantly lower level of fluorescent signal as compared to the ␣ 2A AR, we generated additional ␣ 2B AR and ␣ 2C AR DNA constructs and clonal cell lines. We were concerned that the epitope tag of the TAG-␣ 2C AR was not easily accessible to 12CA5 antibody because this subtype has an unusually high alanine content (17 alanine resi-dues) in its amino-terminal region, potentially leading to the formation of helices. Consequently, we deleted the 3Ј-untranslated regions (280 bp in TAG-␣ 2B AR and approximately 500 bp in TAG-␣ 2C AR) in order to Polarized MDCK II cells from three 24-mm Transwell filters were biotinylated on either the apical or the basolateral surface and the ␣ 2 ARs were photoaffinity labeled with [ 125 I]Rau-AzPEC in the absence or presence of the ␣-AR antagonist phentolamine. We have noted that significantly less total receptor in the preparation of ␣ 2C AR was required to obtain a sufficient signal comparable to that of ␣ 2A AR and ␣ 2B AR subtypes. This observation would suggest that the ␣ 2C AR has a greater affinity for the iodinated photolabel than the other ␣ 2 AR subtypes. Following detergent extraction of the photolabeled preparations and subsequent streptavidin-agarose chromatography, the eluates from the streptavidin-agarose were analyzed on 7-20% gradient SDS-PAGE and subsequently subjected to autoradiography. Autoradiograms shown represent one of at least five independent experiments for each receptor subtype, where comparable findings were obtained. To quantitate the fraction of the ␣ 2 AR subtype localized apically or basolaterally, the gel regions corresponding to migration of the ␣ 2 AR subtypes were cut and counted in a ␥-counter, as described under "Experimental Procedures." increase receptor expression (20) and simultaneously inserted the hemagglutinin epitope tag into carboxyl termini of both subtypes to make the epitope more accessible to the antibody and also to ascertain that the localization was independent of the epitope-insertion site.

Purification of the 12CA5 Monoclonal Antibody on Activated Affi-Gel 10 and Affi-Gel 15
The Protein A-purified 12CA5 antibody obtained commercially manifested variable background staining in parental MDCK II cells. Particularly for evaluation of the localization of the ␣ 2C AR subtype, which was reported previously to possess intracellular as well as surface localization (21), we needed to eliminate non-epitope-dependent fluorescence in MDCK II cells incubated with 12CA5 antibody directed against the 9 amino acid epitope tag. Consequently, we employed a previously recommended strategy to deplete antibody of backgroundinteracting contaminants before examining epitope expression in heterologous cells (22). The rationale is to adsorb proteins in total lysates of MDCK II cells to Affi-Gel 10 and Affi-Gel 15, and then use these resins to deplete the commercial 12CA5 preparation of nonspecific adsorbents. Lysates of parental MDCK II cells were prepared in the following manner. It was determined in advance how many 100-mm dishes grown to confluence were needed to obtain approximately 50 mg of MDCK II lysate protein. This number of dishes were washed in Affi-Gel buffers appropriate for subsequent coupling to Affi-Gels: 0.1 M MOPS, pH 7.5, 80 mM CaCl 2 was used for Affi-Gel 10, while 0.1 M MOPS, pH 7.5, 0.3 M NaCl was used for Affi-Gel 15. Cells were harvested in the smallest possible volume of buffer (such that the total volume did not exceed 1 ml) using a rubber policeman, collected in 1.5-ml Eppendorf tubes, and homogenized by 20 up and down strokes using a 1-ml syringe with a 20-gauge needle. Affi-Gel 10 and Affi-Gel 15 (1 ml of each gel) were prewashed three times in separate Eppendorf tubes with ice-cold 10 mM sodium acetate solution, pH 4.5, after which the whole cell lysate preparations were added onto the appropriate Affi-Gel. The cell lysate-Affi-Gel coupling reactions were rotated for 4 h at 4°C, followed by multiple washes of the conjugated Affi-Gel with the appropriate coupling buffer on a 50-micron mesh column. Affi-Gels were determined to be free of reactants by measurements of OD 280 . The Affi-Gels free of unbound lysate particles were collected in clean 1.5-ml Eppendorf tubes and incubated in the presence of 1 M ethanolamine-HCl, pH 8.0, for 1 h at 4°C in order to block any reactive esters on the Affi-Gels. Each Affi-Gel was then washed again with the appropriate coupling buffer. Following the third wash, an aliquot of 12CA5 antibody was added sequentially to Affi-Gel 10 and Affi-Gel 15 and allowed to rotate 4 h with each of the Affi-Gels at 4°C. Typically 1 ml of 12CA5 antibody (10 mg/ml) was added to 500 l of Affi-Gel 10 and then to 500 l of Affi-Gel 15. The final supernatant, comprising the purified 12CA5 antibody, was aliquoted and stored at Ϫ80°C. Multiple independent purifications of 12CA5 antibody have indicated that there is no loss of the component of the 12CA5 that recognizes the HA epitope using this procedure, but there is a dramatic decrease in the amount of nonspecific, background staining.  (8,9). Upon completion of the metabolic labeling phase ("pulse"), proteins on either the apical or the basolateral membrane were covalently modified with biotin LC-hydrazide or sulfo-NHS-biotin. Detergent-solubilized ␣ 2 ARs were then immunoisolated using the 12CA5 antibody and the fraction of the ␣ 2 AR that had been biotinylated on a particular surface was isolated by subsequent streptavidin-agarose chromatography.

Determination of Receptor Half-life on the Apical versus the Basolateral Cell Surface
MDCK II clones grown on Transwell supports were metabolically labeled ("pulsed") with 150 Ci/well of [ 35 S]cysteine/methionine for the time periods given in individual figure legends and then incubated with complete Dulbecco's modified Eagle's medium supplemented with 1 mM cysteine and 1 mM methionine ("chased") for various periods of time at 37°C. Determination of the metabolically labeled ␣ 2 AR present on either the apical or the basolateral surface at each time point relied on sequential biotinylation, detergent extraction, immunoisolation, and streptavidin-agarose chromatography, as described in detail previously by Keefer et al. (8). SDS-PAGE was utilized to resolve the radiolabeled proteins. The quantity of 35 S-labeled ␣ 2 AR remaining on either the apical or the basolateral surface after each "chase" period was quantified by cutting and counting the regions of the gel corresponding to the migration of a particular ␣ 2 AR subtype detected on the biotinylated surface.

RESULTS AND DISCUSSION
In the present studies, we examined whether the three subtypes of ␣ 2 AR are differentially localized in MDCK II cells and whether the modes of receptor delivery to the cell surface or their half-lives on the membrane after surface delivery differ among the three subtypes.
a Receptor density was determined following saturation binding studies with [ 3 H]Rauwolscine using Scatchard analysis.
b Epitope sequences introduced into the amino terminus of each receptor subtype are identified as TAG-␣ 2 AR structures; those introduced into the carboxy terminus are designated as ␣ 2 -TAG-AR structures.
All Three ␣ 2 AR Subtypes Achieve Basolateral Localization in MDCK II Cells at Steady-state-Steady-state localization of each ␣ 2 AR subtype was determined using surface biotinylation strategies in multiple permanent clonal cell lines grown in Transwell culture. The ␣ 2 ARs were identified using photoaffinity labeling with [ 125 I]Rau-AzPEC. The specificity of labeling was documented by comparing the profile of radiolabeled bands on SDS-PAGE incubated with [ 125 I]Rau-AzPEC in the absence or presence of the ␣-adrenergic antagonist, phentolamine. As shown in Fig. 1, the ␣ 2A AR (Fig. 1A) and ␣ 2C AR (Fig. 1C) subtypes migrate just above the 66-kDa molecular mass marker, consistent with the glycosylation of these 442 (␣ 2A AR) and 461 (␣ 2C AR) amino acid receptors. In contrast, the ␣ 2B AR subtype migrates at approximately 45 kDa on SDS-PAGE, as expected for this non-glycosylated 453 amino acid receptor subtype (14,16). Quantitation of the fraction of each ␣ 2 AR subtype that is biotinylated on the basolateral versus apical surface indicates that 91% of the WT-␣ 2A AR, 89% of the TAG-␣ 2A AR, 97% of the WT-␣ 2B AR, 85% of the TAG-␣ 2B AR, 95% of the WT-␣ 2C AR, and 92% of the TAG-␣ 2C AR are localized on the basolateral surface at steady-state. Furthermore, the fraction of ␣ 2B AR or ␣ 2C AR enriched on the basolateral surface was not altered quantitatively if the epitope tag was introduced into the amino or carboxyl terminus of the receptor. Comparable amounts of the radiolabeled receptor are detected in detergentsolubilized extracts obtained from Transwell cultures biotinylated on either the apical or basolateral surface prior to streptavidin-agarose chromatography (Ref. 8; data not shown). Thus, the overwhelmingly greater basolateral signal detected following streptavidin-agarose chromatography of the extracts from biotinylated cells is a true reflection of ␣ 2 AR enrichment on the basolateral surface of MDCK II cells. The finding that introduction of the epitope tag did not alter the apparent localization of any of the ␣ 2 AR subtypes meant that these epitopetagged structures could be exploited in subsequent immunochemical and metabolic labeling studies and yet faith-fully represent the fate of the wild-type ␣ 2 AR subtype structure.
Since surface biotinylation strategies only reveal the fraction of the receptor population at either the apical or basolateral surface, we also employed immunocytochemical methods to evaluate the surface as well as possible intracellular localization of each of the ␣ 2 AR subtypes within MDCK II cells. MDCK II clones were grown in Transwell culture and immunostained using the Affi-Gel-purified 12CA5 antibody directed against the epitope tag (see "Experimental Procedures"). The surface staining pattern of the three ␣ 2 AR subtypes corroborates the basolateral localization revealed using surface biotinylation strategies. For the ␣ 2A AR and ␣ 2B AR, the immunofluorescence was detected exclusively on the lateral subdomain of MDCK II cells (Fig. 2). The ␣ 2C AR, however, is localized not only on the lateral subdomain, but also in intracellular compartments. Our findings in MDCK II cells resemble previous reports of both surface and intracellular localization of the ␣ 2C AR when expressed in COS-7 and HEK-293 renal cell lines (21). This bimodal surface and intracellular localization of the ␣ 2C AR, unique among the ␣ 2 AR subtypes, must represent a property of the ␣ 2C AR structure per se, since the same 12CA5 antibody preparation was used to immunologically localize all three subtypes of the ␣ 2 AR. Furthermore, as mentioned above, introduction of the epitope tag into the amino-versus the carboxylterminal domains of the ␣ 2C AR did not alter its steady-state distribution, assessed using either biochemical or morphological strategies. Finally, we also examined several independent clonal cell lines (see Table I), and found that the relative distribution of ␣ 2 AR for all subtypes was unmodified over the range of 2 to 25 pmol of ␣ 2 AR/mg of protein. Thus, we are confident that the receptor distribution shown in Fig. 2 represents properties of the structures of the varying ␣ 2 AR subtypes.
Comparison of the Localization of Intracellular ␣ 2C AR with Calnexin and Mannosidase II-In order to explore the cellular compartment(s) where the intracellularly-localized ␣ 2C AR   FIG. 3. Immunological co-localization of the ␣ 2C AR subtype with markers for endoplasmic reticulum (calnexin) and trans-Golgi network (mannosidase II). MDCK II clones permanently expressing the HA epitopetagged ␣ 2C AR subtype (TAG-␣ 2C AR 11 and ␣ 2C -TAG-AR 77) were polarized on Transwells, fixed, and immunostained as described under "Experimental Procedures." The 12CA5 primary mouse monoclonal antibody was detected by a secondary donkey anti-mouse FITC (panels A and D) or Cy3 (G)-conjugated IgG. The calnexin or mannosidase II rabbit polyclonal antibodies were detected with a secondary donkey anti-rabbit Cy3 (panels B and E) or FITC (panel H)-conjugated IgG. Panel I shows a co-localization of 12CA5 and anti-calnexin antibody staining, while panels C and F show a co-localization of 12CA5 and anti-mannosidase II antibody staining. In panels C, F, and I, the 12CA5 incubation was followed by a secondary donkey anti-mouse FITC (panels C and F) or Cy3 (panel I)-conjugated IgG; subsequently a second primary antibody (anti-Calnexin, in panel I, or antimannosidase II in panels C and F) was added for 1 h, followed by a secondary donkey anti-rabbit FITC (panel I) or Cy3 (panels C and F)-conjugated IgG. The yellow line across the XY scan indicates the plane of sectioning of the Z scan, as described in the legend to Fig. 2. might be expressed, we compared the immunological localization of the ␣ 2C AR with that of calnexin (18), a marker for the endoplasmic reticulum, and mannosidase II, a marker for the trans-Golgi network (17) . Fig. 3 compares the localization of ␣ 2C AR (Panels A and D) with mannosidase II (Panels B and E). Panels C and F in Fig. 3 examine the co-localization (yellow color) of ␣ 2C AR (green color due to the FITC-conjugated secondary antibody) and mannosidase II (red color due to the Cy3-conjugated secondary antibody). Laser sectioning in the Z plane is most informative in revealing the extent, or lack thereof, of ␣ 2C AR co-localization with markers for the endoplasmic reticulum and trans-Golgi network. As shown in Fig. 3C, the staining for the intracellular ␣ 2C AR completely overlaps that observed for mannosidase II, whereas in other cross-sections (e.g. Fig. 3F) the immunofluorescence of the intracellular ␣ 2C AR expands apically from the trans-Golgi network, which is defined by mannosidase II staining. Similarly, the profile of the intracellular ␣ 2C AR immunofluorescence partially overlaps that for calnexin, but also occupies cellular space oriented apically from endoplasmic reticulum (Fig. 3I), as defined by calnexin staining. Taken together, the findings in Fig. 3 (C, F, and I) suggest either that the intracellular ␣ 2C AR is distributed throughout the endoplasmic reticulum and trans-Golgi network, or is localized to a unique subcellular compartment that is characterized by an immunofluorescence profile that partially overlaps both calnexin and mannosidase II staining.
The ␣ 2C AR, Like ␣ 2A AR, Is Delivered Directly to the Basolateral Membrane of MDCK II Cells, While the ␣ 2B AR Is Initially Randomly Inserted into Both Apical and Basolateral Surfaces-Our next goal was to determine if all three subtypes of ␣ 2 AR are directly delivered to the basolateral surface of MDCK II cells, where they predominantly reside at steady-state, or alternatively, if ␣ 2 AR subtypes achieve this steady-state localization via distinct mechanisms. Delivery of the ␣ 2 AR subtypes was examined by metabolic labeling of nascent receptors, followed by biotinylation on the apical or basolateral cell surface, immunoisolation using the anti-epitope tag antibody, and quantitation of the fraction of ␣ 2 AR biotinylated on a particular surface following streptavidin-agarose chromatography.
Metabolic labeling studies require the epitope-tagged receptor as a means to identify and isolate the ␣ 2 AR among the other radiolabeled proteins. Thus, for these studies we were restricted to using clonal cell lines permanently expressing the epitope-tagged versions of ␣ 2 AR subtypes to study receptor delivery to the cell surface. Fortunately, all of our steady-state data comparing the localization of photoaffinity-labeled ␣ 2 AR indicate that introduction of the epitope tag does not influence the localization of any of the ␣ 2 AR subtypes (cf. Fig. 1). It was of interest, however, that we were able to immunoisolate the ␣ 2C AR subtype only when the epitope tag was inserted into the carboxyl terminus; perhaps the alanine-rich sequence in the amino terminus of the ␣ 2C AR rendered the tag inaccessible to the 12CA5 antibody when inserted in the amino terminus. Fig. 4 examines the delivery of all three ␣ 2 AR subtypes in MDCK II cells. In all cases, we are confident that the radiolabeled band identified as ␣ 2A , ␣ 2B , or ␣ 2C AR represents the particular subtype under study, because of its comigration with photoaffinity-labeled receptor and its absence when comparable metabolic labeling studies are performed in parental MDCK II cells lacking expression of any of these subtypes. The band migrating just above the 66-kDa molecular mass marker on SDS-PAGE is interpreted to represent the ␣ 2A AR (Fig. 4A) and ␣ 2C AR (Fig. 4C) subtype. Newly synthesized ␣ 2C ARs (Fig.  4C), like the ␣ 2A AR (Fig. 4A), appear virtually exclusively on the basolateral surface after each labeling time point tested. These data are consistent with the interpretation that both the ␣ 2A AR and ␣ 2C AR subtypes are delivered directly to the basolateral surface. In contrast, the ␣ 2B AR shows a fundamentally different targeting pattern when compared to the ␣ 2A AR and ␣ 2C AR subtypes. As shown in Fig. 4B, the 45-kDa ␣ 2B AR is initially randomly inserted into both apical and basolateral surfaces, where it is first detectable 30 min after initiation of metabolic labeling. This random expression on both the apical and basolateral surfaces also is repeatedly observed 45 min after pulse labeling of this receptor subtype. However, 60 min after the initiation of metabolic labeling, the majority of the ␣ 2B AR is found on the basolateral membrane. The data shown in Fig. 4B are from one experiment repeated 6 times with comparable outcome. The ␣ 2B AR subtype delivery was assessed using receptor structures where the epitope tag was introduced either into the amino or carboxyl terminus, and the indistinguishable results obtained for either epitope-tagged structure suggest that position of epitope tag does not influence the apparent random delivery of the ␣ 2B AR in MDCK II cells. ]cysteine/methionine in 150 l of medium at 37°C, 5% CO 2 for the indicated periods of time and then biotinylated on either the apical or the basolateral surface. The ␣ 2 AR delivered to the biotinylated surface were isolated via sequential Protein A and streptavidin-agarose chromatography, as described under "Experimental Procedures." Eluates from the streptavidin-agarose resin were analyzed by 7-20% gradient SDS-PAGE and subsequent autoradiography. Shown here are representative autoradiograms of at least three independent experiments for each receptor subtype (for the ␣ 2B AR, this experiment was repeated 6 times), which yielded comparable findings.
The ␣ 2B AR Subtype Is Selectively Retained on the Basolateral Surface-One possible scenario that could explain the initial uniform distribution of the ␣ 2B AR on apical and basolateral surfaces followed by enrichment on the basolateral surface would involve selective retention of the ␣ 2B AR in the basolateral, but not in the apical, domain. To determine if the ␣ 2B AR was selectively retained on the basolateral surface, we directly examined the half-life of the ␣ 2B AR on both the apical and basolateral surfaces of MDCK II cells. Cells grown in Transwell culture were metabolically labeled with [ 35 ]cysteine/methionine, and then chased with cysteine and methionine-enriched medium for various time periods before evaluating ␣ 2B AR surface distribution (see "Experimental Procedures"). Using this experimental approach the ␣ 2B AR was observed to possess a half-life on the apical surface of 15-30 min (Fig. 5A) but a half-life on the basolateral surface of 8 -10 h (Fig. 5B). At present, we cannot determine whether or not the ␣ 2B AR, after removal from the apical surface, is redirected to the basolateral  6. Retention of the three ␣ 2 AR subtypes on the basolateral surface. MDCK II permanent clonal cell lines expressing the epitopetagged ␣ 2 AR subtypes used in representative autoradiograms shown above included ␣ 2A -TAG-AR 52 (A), TAG-␣ 2B AR 122 (B), and ␣ 2C -TAG-AR 28 (C). Polarized MDCK II cells grown on three or four 24-mm Transwell filters (depending on the clonal receptor expression) for 7 days were metabolically labeled (pulsed) with 150 Ci of [ 35 S]cysteine/methionine in 150 l of medium for 60 min at 37°C, 5% CO 2 , and then incubated in medium supplemented with 1 mM methionine, 1 mM cysteine for various periods of time ("chase period") at 37°C. The 2-h chase period was designated as t 0 (time ϭ 0) based on our previous observations of time required for the nascent ␣ 2 AR to arrive at the basolateral surface. In addition to the t 0 , ␣ 2 ARs were also isolated following 6-and 24-h chase periods, using biotin surface labeling strategy coupled with receptor immunoisolation and streptavidin-agarose chromatography, as described under "Experimental Procedures." The upper panel of A, B, and C provides autoradiograms of 7-20% gradient SDS-PAGE from a representative experiment; the lower panels show plots of the means Ϯ S.E. of radioactivity in each band averaged from three experiments. Receptor bands were excised from autoradiograms according to the position of each ␣ 2 AR subtype and counted in a ␤-counter with 10 ml of scintillation liquid. The percentage of radioactivity present in each receptor band was plotted on a semi-log scale as a function of time defining radioactivity at t 0 as 100%. surface or, alternatively, is internalized and degraded. However, the differential stability of ␣ 2B AR on the apical versus the basolateral membrane explains the observation that, at steadystate, the ␣ 2B AR is localized almost exclusively on the basolateral surface. Furthermore, this finding suggests that each membrane domain has a unique mechanism for protein retention and, moreover, that the apical retention mechanism for the ␣ 2B AR is less efficient. Fig. 6 demonstrates that the ␣ 2A AR, ␣ 2B AR, and ␣ 2C AR have comparable half-lives on the basolateral surface. Previous data from our laboratory suggest, at least for the ␣ 2A AR, that receptor retention involves protein-protein interactions involving endofacial domains of the receptor, since deletion of the third cytoplasmic loop of the ␣ 2A AR measurably accelerates receptor turnover on the basolateral surface (9). The similar half-life of the ␣ 2A AR, ␣ 2B AR, and ␣ 2C AR subtypes on the basolateral surface suggests that similar tethering mechanisms may exist for all three structures despite the fact that the endofacial third loop sequences of these three ␣ 2 AR subtypes are quite distinct. Perhaps these ␣ 2 AR subtypes, when they achieve their threedimensional structure, project similar surfaces to endofacial proteins that stabilize ␣ 2 ARs on the surface. More rapid turnover of the ␣ 2B AR on the apical domain suggests that if such tethering proteins do exist, they may be absent or exist in a reduced density underneath the apical surface, resulting in a more rapid ␣ 2B AR turnover on the apical, when compared with the basolateral, domain. SUMMARY Fig. 7 provides a schematic diagram of the different trafficking itineraries observed for the three ␣ 2 AR subtypes in these studies. As described previously (8), the ␣ 2A AR subtype is delivered directly to the basolateral membrane and at steadystate virtually all ␣ 2A AR is present on the surface. The ␣ 2C AR also is directly delivered to the basolateral surface, but at steady-state is distributed between cell surface and one or more intracellular compartments. Finally, the ␣ 2B AR subtype is initially delivered randomly to both cell surfaces but then is preferentially retained on the basolateral membrane. Although all three ␣ 2 AR subtypes have been detected in renal epithelia of varying species (2, 3, 5), the functional consequences of these differing receptor itineraries in renal epithelial cells revealed by the present studies or the implications for targeting and retention of these ␣ 2 AR subtypes in other polarized cells, such as neurons, remain to be established. FIG. 7. Schematic representation of differential targeting mechanisms that lead to the basolateral localization of all three ␣ 2 AR subtypes in MDCK II cells. This schematic diagram compares the targeting mechanism of the ␣ 2A AR, ␣ 2B AR, and ␣ 2C AR subtypes, and summarizes the differing trafficking or steady-state localization observed for each subtype. The ␣ 2A AR is delivered directly to the basolateral membrane (present data, Ref. 8). The present findings have revealed that the ␣ 2B AR is inserted randomly into both cell surface domains and is retained preferentially on the basolateral surface. At steady-state, the ␣ 2A AR and ␣ 2B AR are detected exclusively on the lateral surface of MDCK II cells, whereas the ␣ 2C AR is distributed between cell surface and intracellular compartments. However, the fraction of the ␣ 2C AR that is localized to the basolateral surface is delivered there directly.