Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The mechanisms by which G protein-coupled receptors (GPCRs) attain their localization in polarized epithelial cells are an important determinant of trans-epithelial function. Using polarized Madin-Darby canine kidney II (MDCKII) cells as a model system, we previously described the different trafficking itineraries for the three ₂-adrenergic receptor (AR) subtypes (Wozniak and Limbird, 1996). The _2AAR and _2CAR subtypes are directly delivered to the basolateral surace, whereas the _2BAR is randomly delivered to both apical and basolateral surfaces. The _2BAR achieves its steady-state basolateral enrichment due to its retention on that surface (t_1/2 = 10-12 h) compared with its rapid turnover on the apical surface (t_1/2 = 5-15 min). At steady state, the _2AAR and _2BAR subtypes are nearly exclusively on the basolateral surface, whereas a substantial fraction of the _2CAR population remains in a cytoplasmic compartment (Wozniak and Limbird, 1996), corroborating earlier findings in nonpolarized cells (von Zastrow et al., 1993). Mutagenesis strategies, undertaken in detail for the _2AAR subtype, suggest that the third intracellular loop of the _2AAR is critical for retention of this subtype on the basolateral surface, but that targeting to the bilayer involves sequences or structures embedded in or near the bilayer (Keefer et al., 1994).

We previously had shown that the A₁ adenosine receptor (A₁AdoR), also coupled to G_i/G_o G proteins, is directly delivered to the apical surface of MDCKII cells and enriched there at steady state (Saunders et al., 1996; Saunders and Limbird, 1997). Studies of surface delivery of truncations of the _2AAR or chimeras with the A₁AdoR suggest that multiple independent sequences exist in the bilayer of GPCRs to determine targeting of these seven transmembrane-spanning molecules in a hierarchical fashion to one versus another surface in polarized cells (Saunders et al., 1998).

Receptors and other cell surface proteins are not the only nonrandomly distributed molecules in polarized cells. The apical surface is undergirded by an actin-rich cytoskeletal network, whereas an ankryn-fodrin-rich cytoskeleton underlies the basolateral surface (Nelson and Hammerton, 1989). Furthermore, actin microfilaments, microtubules, and intermediate filaments, involved in vesicle and protein movement within all cells, have been particularly implicated in polarized trafficking and subsequent function of membrane-targeted molecules (Matter et al., 1990; Lafont et al., 1994; Arreaza and Brown, 1995). We previously have observed that disruption of microtubules with colchicine or nocodazole in polarized MDCKII cells paralleled a reduced apical delivery of A₁AdoR and an enrichment, or rerouting, of the A₁AdoR to the basolateral surface (Saunders and Limbird, 1997). In contrast, the same treatment had no impact on the random delivery of the _2BAR subtype to both surfaces before selective retention on the basolateral surface. However, the amount of _2BAR delivered to the surface was significantly enhanced in a time-dependent fashion when the microtubule network of MDCKII cells is depolymerized with colchicine (Saunders and Limbird, 1997). The present studies were undertaken to further explore the structural regions within _2BAR that contribute to this unexpected subtype-selective increase in _2BAR density and the mechanisms that may contribute to this phenomenon.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. [³H]Methoxyinulin (125.6 mCi/g) was purchased from DuPont/NEN (Boston, MA). Protein A-purified 12CA5 monoclonal antibody was from the Berkley Antibody (Richmond, CA); Cy-3-conjugated donkey anti-mouse IgG was from Jackson ImmunoResearch (West Grove, PA); monoclonal anti--tubulin was from Amersham (Arlington Heights, IL); and rhodamine-conjugated phalloidin was from Molecular Probes (Eugene, OR). Colchicine and lumicolchicine were from Sigma Chemical Co. (St. Louis, MO).

Construction of _2BARi3-TAG. The first nine amino acids after the initiating methionine of the _2BAR evaluated in these studies encode a hemagglutinin (HA) epitope recognized by the commercially available monoclonal antibody 12CA5 (Berkley Antibody). The HA-epitope tagged, mutant _2BAR with the third intracellular loop deletion was constructed with the Stratagene mutagenesis kit (La Jolla, CA). The construction of the HA-epitope-tagged _2BAR has been described previously (Wozniak and Limbird, 1996). Purified oligonucleotides were generated to engineer two NotI sites in the third intracellular loop of the _2BAR, one at nucleotide 639 and one at nucleotide 1062. Subsequently, nucleotide sequences encoding the majority of the third loop were excised via NotI restriction digest, and the remaining receptor sequence was ligated back together with the T4 DNA ligase. This resulted in a construct that encoded an _2BAR mutant protein in which only the membrane proximal sequences of the third loop remained intact (total third loop length for the _2BAR is 178 amino acids, whereas the third loop of the _2BARi3 is 38 amino acids). In the _2BARi3 structure, 14 amino acids in the proximal (amino terminal) and 14 amino acids into the distal (carboxy terminal) portion of the third cytoplasmic loop are retained because these regions have been demonstrated to be critical for G protein coupling (Wade et al., 1994; Eason and Liggett, 1996). The _2BARi3 mutant was verified by dideoxy-DNA sequencing.

COSM6 Transfection and Cell Culture. This was done as previously described in Guyer et al. (1990). Briefly, transient transfection of simian kidney COSM6 cells (100-mm dishes plated at 1 × 10⁶ cells/dish) with 10 µg of DNA by the DEAE-dextran method of transfection was used to assess the level of receptor expression with both immunocytochemistry (as described in "Steady-State Localization of GPCRs by Immunolocalization") and [³H]yohimbine binding (as described in "Receptor-Binding Assays").

Development of Permanent Transformants of Human Embryonic Kidney (HEK)293 Cells. This was done as described previously in Schramm and Limbird (1999). HEK293 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum at 37°C in a 5% CO₂ incubator. Permanent transfectants were generated by lipofectamine-mediated cotransfection of the cells with plasmids containing the indicated receptors and a neomycin resistance gene. Cells that survived selection in medium containing 500 µg/ml G-418 were screened for expression of the expected receptor by binding of the radiolabeled ₂AR antagonist [³H]rauwolscine. Clonal cell lines with varying levels of ₂AR expression were kept for further study. The experiments reported herein were performed on an _2BAR expressing cell line that contains 2 to 4 pmol/mg of receptor binding.

Development of Permanent Transformants of MDCKII Cells. Permanent clonal cell lines of MDCKII cells were developed as described previously (Keefer and Limbird, 1993; Wozniak and Limbird, 1996). The clonal cell lines evaluated in the present study include TAG_2AAR (25 pmol/mg and 7 pmol/mg protein), TAG_2BAR (10 pmol/mg and 3 pmol/mg protein), and TAG_2CAR(5 pmol/mg and 3 pmol/mg protein). The _2AAR subtype was encoded by a porcine cDNA; the _2BAR and _2CAR subtypes by a rat cDNA.

Polarized Culture of MDCKII Cells and Functional Confirmation of Intact Monolayers. MDCKII cells were maintained as described previously (Keefer and Limbird, 1993). For polarity experiments, MDCKII cells were seeded at a density of 1 × 10⁶ cells/24.5-mm polycarbonate membrane filter (Transwell chambers, 0.4-µm pore size; Costar, Cambridge, MA), and cultured for 5 to 8 days with medium changes every day. Before each experiment, the integrity of the monolayer was assessed by monitoring [³H]methoxyinulin leak (Keefer et al., 1994).

Receptor-Binding Assays. MDCKII particulate preparations were prepared essentially as described in Keefer and Limbird (1993). Briefly, all MDCKII cell lines were grown in Transwell culture (except when stated otherwise) and allowed to polarize for 7 to 8 days. Binding assays were performed on membranes harvested from these cells treated (or not) with varying agents examined in this study. Cells were harvested into a lysis buffer: 15 mM Tris-HCl, 5 mM EGTA, 5 mM EDTA, pH 8.0, containing 0.5 µg/ml leupeptin, 0.7 µg/ml pepstatin, 100 µg/ml soybean trypsin inhibitor, and 1 mM phenylmethylsulfonyl fluoride.

Assessment of Antagonist Binding. The pellet resulting from two washes in lysis buffer followed by centrifugation at 30,000g was resuspended in 900 µl of antagonist binding buffer (20 HEPES, 25 glycine/glycine, 100 NaCl, 5 EGTA, pH 7.4). COSM6 cells and HEK293 cell membranes were prepared similarly. For all the cell types, [³H]rauwolscine binding was performed in 12 × 75-mm polypropylene tubes containing 3 nM [³H]rauwolscine (diluted in water) in the absence (total binding) or presence (nonspecific binding) of 10 µM phentolamine, an ₂AR antagonist. Incubations were for 30 min at 25°C, and were terminated by the addition of 3.5 ml of ice-cold 25 mM glycyl glycine buffer, pH 8.0, and filtration through Whatman GF/B glass microfiber filters.

Assessment of Agonist Binding. Membranes from clonal cell lines expressing wild-type or mutant receptors were harvested and washed in lysing buffer but resuspended and incubated in an agonist buffer containing 50 mM Tris-HCl, 10 mM MgCl₂, and 5 mM EGTA, pH 8.0. Incubations were for 30 min at 25°C and contained 100 µg of membrane protein, 0.9 nM p-[¹²⁵I]iodoclonidine ([¹²⁵I]PIC) agonist radioligand (~160,000 cpm/100-µl incubation). The incubation was terminated by vacuum filtration and the GF/B filters were counted in a Beckman 4000 gamma counter.

Guanine Nucleotide Sensitivity of Radioligand Agonist Binding as a Measure of _2BAR-G Protein Coupling. The addition of Gpp(NH)p, a hydrolysis-resistant GTP analog, to membrane preparations containing G protein-coupled receptors typically disrupts receptor/G protein coupling, shifting the receptor from a higher-affinity state for agonists (functional receptor-G protein coupling) to a lower-affinity state for agonists (receptor functionally dissociated from G protein) (Stadel et al., 1980). Consequently, the ability of guanine nucleotides to decrease the detectability of radiolabeled agonist binding is an indirect measure of the existence of guanine nucleotide-sensitive high-affinity receptor-agonist interactions (Williams and Lefkowitz, 1977; Gerhardt et al., 1990). To evaluate the ability of Gpp(NH)p to modulate radiolabeled agonist binding, [¹²⁵I]PIC incubations were performed in the absence (control) or presence of increasing concentrations of Gpp(NH)p. When evaluated, pertussis toxin sensitivity of receptor-GTP interactions was determined by incubation of MDCKII cells overnight with 200 ng of pertussis toxin/ml of culture medium (10 ml of medium/100-mm dish) (Keefer and Limbird, 1993; Ceresa and Limbird, 1994). To assess the relative fraction of receptors achieving high-affinity agonist binding due to receptor-G protein interactions by different receptor structures or under different incubation conditions, the quantity of guanine nucleotide-sensitive [¹²⁵I]PIC binding (picomoles per milligram) to total receptor density of [³H]rauwolscine binding (picomoles per milligram) can be compared.

Treatment of Cells with Colchicine. Colchicine is an irreversible microtubule-disrupting drug that binds slowly to soluble tubulin heterodimers, reducing them to large aggregates and rendering them incapable of polymerizing for microtubule growth. Incubations with 10 µM colchicine were performed for 15 h as previously described (Saunders and Limbird, 1997). Immunocytochemical analysis of treated cells with an anti--tubulin antibody confirmed that colchicine treatment of cells had indeed disrupted the microtubule network, as seen in Fig. 1C.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Colchicine increases density of the 2BAR in a subtype selective manner. A, functional density of the 2BAR subtype is increased by colchicine in polarized MDCKII cells as assessed by [3H]rauwolscine saturation binding to membrane preparations derived from these cells expressing heterologous 2AR subtypes and grown in 75-mm Transwell culture. Binding density was assessed after culture in the presence or absence of 10 µM colchicine (15 h). *P < .05, as assessed by one-way ANOVA. The increase in 2BAR binding resulted from an increase in receptor density (Bmax = 1 ± 0.06 pmol receptor binding/mg protein and KD = 1.86 ± 0.18 nM (control) versus 2.96 ± 0.67 pmol receptor binding/mg protein and KD = 2.23 ± 0.16 nM following colchicine treatment. B, confocal immunofluorescence images of 2AR subtypes in polarized MDCKII cells incubated in the presence or absence of 10 µM colchicine treated as in (A). The cells were detected on their Transwell filter supports, as described in "Methods." C, untransfected (parental) polarized MDCKII cells were treated or not with colchicine as in (A). and then stained with an anti--tubulin monoclonal antibody to confirm that the colchicine treatment used was effective in disrupting microtubule structure in these cells.

Steady-State Localization of GPCRs by Immunolocalization. Immunostaining of cells grown in Transwell culture was performed as described previously (Saunders et al., 1996; Saunders and Limbird, 1997) with the following concentrations of primary antibody: a 1:50 dilution of 12CA5 primary antibody, purified as described previously (Wozniak and Limbird, 1996), for the localization of hemagglutinin epitope-tagged GPCRs (Keefer and Limbird, 1993). The antibody-containing buffers and wash buffers contained 0.1% Triton X-100 to permit detection of epitope on the cell surface and in the cell interior. Treatment with the secondary Cy3-conjugated donkey anti-mouse IgG (1:200) was performed as described in Saunders and Limbird (1997). Samples were visualized by confocal microscopy on a Zeiss Axiovert 135 Micro System LSM (Oberkochen, Germany). The samples were first visualized in the xy plane, and then in the xz plane. In the images shown, the bottom three-fourths represent the xy plane, the conventional view of the cells as one looks down on them. The white line that is shown in the xy plane confocal images indicates where the laser took a cross-section of the cells to generate the z scan. The top one-fourth of the image represents the xz plane (or z scan), the cortical section perpendicular to the plane of the cell layer. Images were analyzed with Showcase software on a Silicon Graphics (Mountain View, CA) iris indigo workstation.

Steady-State Localization of _2BAR and _2BARi3 in MDCKII Cells: Biotinylation and Photoaffinity Labeling. The previously described method (Wozniak and Limbird, 1996) for quantitating the apical versus basolateral (versus intracellular) distribution of the wild-type and mutant _2BAR in polarized MDCKII cells was biotinylation of the apical versus the basolateral surface of cells grown in Transwell culture, photoaffinity labeling of the functional _2BAR in harvested membranes, detergent extraction, and isolation of biotinylated receptors via streptavidin-agarose chromatography. The cell lines grown in Transwell culture were biotinylated on ice for 30 min with sulfo-NHS biotin to covalently label the primary amines of the _2BAR. Following membrane preparation, the _2BAR-expressing cell lines (both wild type and _2BARi3) were covalently modified with the photoactivatable _2BAR-selective ligand [¹²⁵I]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 µM phentolamine, an -adrenergic receptor antagonist. The photoaffinity-labeled receptors were then extracted in RIPA buffer. Streptavidin-agarose chromatography was used to isolated the biotinylated (and now photoaffinity-labeled) molecules. The fraction of biotinylated, photoaffinity-labeled ₂AR present on the apical versus basolateral surface was determined following SDS-polyacrylamide gel electrophoresis followed by autoradiography. The films were then scanned and imported into Adobe Photoshop.

Mitogen-Activated Protein (MAP) Kinase Stimulation. Activation of MAP kinase was assessed as previously described for HEK293 cells (Schramm and Limbird, 1999) with a few alterations. Clonal MDCKII cell lines were plated on 24-mm Transwell filters (Transwell chambers, 0.4-µm pore size; Costar, Cambridge, MA) at confluency. The cells were then serum-deprived overnight. On the day of the experiment, the cells in Transwell culture were moved to a plate warmer kept at 37°C, and the medium was replaced with fresh serum-free DMEM. The ₂AR agonist UK-14304 was added for 2 or 10 min at all final concentration of 1 µM directly to the medium on the cells and very gently swirled to mix. An equal volume of medium was added to the control well. After the indicated times, the cells were washed once with Dulbecco's PBS containing 1 mM MgCl₂ and 0.5 mM CaCl₂, then lysed in SDS sample buffer (62.5 mM Tris-HCl, pH 6.8; 2% w/v SDS; 10% glycerol; 50 mM dithiothreitol) supplemented with 1 mM sodium orthovanadate (Sigma Chemical Co.), 10 U/ml leupeptin (Sigma Chemical Co.), and 10 U/ml aprotinin (Bayer, Kankakee, IL). The lysates were transferred to an Eppendorf tube on ice. When all samples were collected, they were sonicated for 20 s, then placed in a heating block at 95°C for 5 min. The lysates were then centrifuged in a microcentrifuge at room temperature for 5 min to remove debris. The supernatants were assayed in Bio-Rad's protein assay for relative protein concentration, and equivalent amounts of protein were loaded on a 10% SDS-polyacrylamide gel for electrophoresis. The gel was run for 160 mAmp-hours, then transferred overnight onto nitrocellulose in transfer buffer (20% methanol, 0.19 M glycine, 25 mM Tris base) at 33 mV. MAP kinase activation was evaluated with an antibody that recognizes dually phosphorylated (Thr/Tyr) MAP kinase (catalog no. V6671; Promega, Madison, WI) and normalized to total MAP kinase with an antibody that recognizes MAP kinase regardless of its phosphorylation state (catalog no. 9102; NEB, Beverly, MA). To assess activated MAP kinase content, the nitrocellulose blot was incubated in blocking buffer (1× Tris-buffered saline; 0.1% Tween 20; 5% w/v nonfat dry milk) for 1 h at room temperature, then probed with Promega's rabbit polyclonal antibody to dually phosphorylated MAP kinase, diluted 1/500 in blocking buffer, for 1 h at room temperature. The blot was washed three times for 5 min each with Tris-buffered saline/Tween 20 (2.42 g/l Tris base, 8.0 g/l NaCl, 0.1% Tween 20, pH 7.6) then probed with donkey anti-rabbit horseradish peroxidase-linked secondary antibody (1/2000 dilution in blocking buffer) (Amersham) for 1 h at room temperature. The wash protocol was repeated, and the immunoreactive bands were detected by enhanced chemiluminescence (Amersham) The blots were then stripped with stripping buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 100 mM 2-mercaptoethanol) for 30 min at 65°C, and reprobed with antibody to total MAP kinase (NEB) at a 1/500 dilution in blocking buffer overnight at 4°C, followed by donkey anti-rabbit secondary antibody as described above. The enhanced chemiluminescence images were scanned into Adobe Photoshop with a UMAX Astra 600 scanner.

View larger version (58K): [in this window] [in a new window]   Fig. 2.   The colchicine-induced increase in the functional density of the 2BAR subtype requires the third intracellular loop. A, radioligand binding of [3H]rauwolscine to 2BAR and 2BARi3 was assessed following treatment (or not) with 10 µM colchicine for 15 h as described in Fig. 1. The data shown are means ± S.E. (n = 3). *P < .05, as assessed by Student's t test. B, confocal images for the 2BAR and 2BARi3 as described in "Methods." The staining pattern of the 2BARi3 expressed in MDCKII cells was compared with that of the wild-type 2BAR. MDCKII cells expressing 2BARi3 were grown in Transwell culture for 4 to 7 days to attain a polarized phenotype. Once the monolayer integrity was confirmed by the [3H]methoxyinulinwere grown in Transwell culture for 4 to 7 days to attain a polarized phenotype. Once the monolayer integrity was confirmed by the [3H]methoxyinulin leak assay, the cells were fixed, permeabilized, and stained with the Babco 12CA5 monoclonal antibody directed against the 9 amino acid HA tag at the N terminus of the receptor, followed by staining with donkey anti-mouse Cy-3 secondary antibody. The Transwell filter with the adherent cells was then mounted onto a glass slide and visualized on a Zeiss Axiovert confocal microscope. The fluorescence pattern was compared with that of the wild-type 2BAR. C, streptavidin-agarose resolution of detergent-solubilized 2BAR and 2BARi3 extracted from polarized MDCKII cells after surface biotinylation. The 2BAR is enriched on the cell surface, and thus retained on and eluted from streptavidin-agarose resin after surface biotinylation. In contrast, the mutant 2BARi3 is mostly intracellular, and therefore is not retained on streptavidin-agarose because intracellular receptor does not have access to biotin, such that the bands representing photoaffinity-labeled receptor are present in the pass throughs only and not in the streptavidin eluates. The authenticity of the band migrating at ~45 kDa (the wild-type 2BAR) and ~30 kDa (the 2BARi3) is verified by the lack of a band at these molecular masses when the photoaffinity-labeling reaction with membranes occurs in the presence of the antagonist phentolamine, which competes for receptor access of [125I]Rau-AzPEC and thus protects against photoaffinity labeling. D, guanine nucleotide regulation of radiolabeled agonist, [125I]para-iodo-clonidine, binding to membranes derived from MDCKII cells. The data shown are means ± S.E. from three independent experiments. E, activation of MAP kinase by the 2AR agonist UK-14304 in polarized MDCKII cells expressing the wild-type 2BAR or 2BARi3. Activated MAP kinase (gel 1) was detected with the Promega antibody (as described in "Methods") directed against the dually phosphorylated Erk1 and Erk2. Total MAP kinase was shown to confirm that we have equal loading of protein on the gel. When stimulated with the 2AR agonist, UK-14,304, as a function of time, untransfected MDCKII cells did not reveal any MAP kinase activation as revealed by a blank film (data not shown). The data are from one experiment representative of four independent experiments.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Colchicine Selectively Increases Receptor Density of _2BAR Subtype at Cell Surface. As seen in Fig. 1A, overnight treatment with 10 µM colchicine dramatically increased the density of _2BAR capable of binding ligand, but did not change functional receptor density significantly for the _2AAR or _2CAR subtypes. As shown previously, increased _2BAR binding was observed for colchicine and, to a lesser extent, for nocodazole, another microtubule-disrupting agent (Saunders and Limbird, 1997). The increase in receptor density was time dependent; it was not noted before 2 h and displayed a maximal effect at 15 h. Changes in _2BAR binding were not observed with -lumicolchicine, a chemical analog of colchicine that does not disrupt microtubules (data not shown).

Colchicine-Induced Increase in Functional Density of _2BAR Subtype Requires Third Intracellular Loop. Because the third intracellular loop of all three ₂AR subtypes interacts with other proteins, including 14-3-3  (Prezeau et al., 1999), we wanted to see if the third intracellular loop of the _2BAR was required to detect the colchicine-induced increase in receptor density. We created a mutant _2BAR, _2BARi3, in which the third intracellular loop was deleted, except for the membrane proximal portions responsible for coupling to G proteins (Saunders and Limbird, 1997). As seen in Fig. 2A, the colchicine-mediated increase in _2BAR density was not observed in the absence of the third intracellular loop. Confocal microscopy images of immunofluorescence staining revealed that this mutant _2BARi3 is localized predominantly intracellularly (Fig. 2B), as is confirmed with surface biotinylation and photoaffinity-labeling strategies (Fig. 2C). Surface biotinylation followed by photoaffinity labeling of the _2BAR with the ₂AR antagonist [¹²⁵I]Rau-AzPEC demonstrates that, for the wild-type _2BAR, the receptor is predominantly in the streptavidin eluates, meaning that the binding detected is virtually exclusively occurring on the biotinylated cell surface. In contrast, the photoaffinity-labeled _2BARi3 is in the pass throughs of streptavidin-agarose chromatography, indicative of little or no _2BARi3 on the cell surface and therefore little or no biotinylation of the _2BARi3 by membrane impermeant reagents (Fig. 2C).

Colchicine-Induced Increase in _2BAR Density Was Observed Only in Polarized Cells. The cytoskeletal ultrastructure of polarized cells is very different from that of nonpolarized cells. Two main groups of microtubule networks exist in polarized epithelial cells: a randomly organized group, arranged around the apical area of the cell, and a polarized network that runs along the lateral sides of the cells, with the minus ends of the microtubules facing the apical apex, and the plus ends facing the basal side. Because this disparate arrangement is not present in nonpolarized cells, we evaluated whether the colchicine-induced increase in _2BAR density also could manifest itself in nonpolarized cells, both those which had the potential to polarize under the appropriate culture conditions (MDCKII cells), and those that do not possess the potential to do so, such as COSM6 and HEK293 cells. As seen in Fig. 3, the ability of colchicine to increase _2BAR density is most readily detected in MDCKII cells grown in Transwell culture to a functionally and morphologically polarized state (Nelson and Veshnock, 1987). In Transwell culture, both the basal surface, via the nitrocellulous filter, and the apical surface have direct access to nutrients. When the cells are plated densely on plastic, close apposition of the cells forces the cells to form a tight monolayer at the lateral sides of the cells in a pseudopolarized phenotype, but the basal surface does not readily get access to medium. As seen in Fig. 3, the colchicine-mediated increase in receptor density is diminished in MDCKII cells grown on plastic compared with cells grown in Transwell culture, i.e., when greater polarization is achieved. Alternatively, when MDCKII cells are plated sparsely on plastic (i.e., 20-40% confluent), there is no detectable effect of colchicine on _2BAR density, in parallel with the absence of MDCKII cell polarization under these conditions (Nelson and Veshnock, 1987). The necessity of a polarized cell environment to detect colchicine-evoked increases in _2BAR density is further evidenced by the lack of a colchicine-induced effect on steady state _2BAR density in either COSM6 cells transiently expressing the _2BAR, or in HEK293 cells permanently expressing the _2BAR (Fig. 3).

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Detectable colchicine-induced increases in 2BAR density parallel the polarization of renal epithelial cells. Heterologously expressed 2BAR was expressed in different renal cell types, including MDCKII, COSM6, and HEK293. The MDCKII and HEK293 cell lines were permanent expressing cell lines, whereas COSM6 cells transiently expressed 2BAR. The number of repetitions of the experiments in various cell backgrounds was MDCKII Transwells (n = 6), MDCKII plastic confluent (n = 4), MDCKII plastic sparse (n = 4), COSM6 (n = 3), and HEK293 (n = 3). Each binding assay in each experiment was performed in triplicate. *P < .05, as assessed by one-way ANOVA.

Colchicine-Induced Increase in _2BAR Density Is Not Dependent on Functional Coupling to G Proteins. We explored the possibility that the colchicine-induced increase in _2BAR density might require active coupling of the receptor to G proteins because multiple, independent studies have detected - (Roychowdhury et al., 1999) and - (Carlson et al., 1986; Roychowdhury and Rasenick, 1997) subunits of G proteins associated with the cytoskeleton. Overnight treatment of the MDCKII cells expressing _2BAR with pertussis toxin (200 ng/ml), under conditions previously established to maximally ADP-ribosylate G_i (Keefer and Limbird, 1993), did not eliminate the effect of colchicine to increase _2BAR density, as seen in Fig. 4A. The incubation with pertussis toxin was sufficient to uncouple the _2BAR from G proteins, as reflected by the loss of guanine nucleotide modulation of receptor affinity for the radiolabeled agonist [¹²⁵I]PIC in pertussis-toxin treated preparations (Fig. 4B) and the decrease in the ratio of [¹²⁵I]PIC/[³H]rauwolscine binding to that detected in preparations from control cells incubated in the presence of Gpp(NH)p. These findings demonstrate that functional coupling of the _2BAR to pertussis toxin-sensitive G proteins is not required for colchicine-induced increases in _2BAR density at the cell surface. Because the low molecular weight GTP-binding protein rho has been shown to be involved in cytoskeletal protein reorganization in response to extracellular signals (Takaishi et al., 1997), we examined whether pretreatment of cells with 10 µg/ml botulinum C3 exoenzyme, which ADP ribosylates and blocks its function, altered colchicine-induced increases in _2BAR density. Botulinum toxin had no effect on the ability of colchicine to modulate _2BAR density (data not shown).

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Colchicine-induced increases in 2BAR density are not dependent on receptor coupling to pertussis toxin-sensitive G proteins. A, radioligand binding of the heterologously expressed 2BAR in polarized MDCKII cells grown on 75-mm Transwells was assessed in the presence of 10 µM colchicine (15 h) with the radiolabeled antagonist [3H]rauwolscine, as described in "Methods." B, pertussis toxin treatment, used to evaluate the role of Gi on colchicine-mediated 2BAR density increases, was effective in disrupting 2BAR-Gi functional coupling, as performed in side-by-side experiments with MDCKII cells treated as in (A). [125I]PIC agonist binding, an indirect measure of receptor-G protein coupling, as performed as described in "Methods." Disruption of 2BAR-G protein coupling, either by incubation with Gpp(NH)p or prior treatment of cells with pertussis toxin, eliminates high-affinity agonist binding, such that the detectability of [125I]PIC binding via vacuum filtration is eliminated. The data shown represent means ± S.E. from three independent experiments.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

The study of the mechanisms that govern trafficking of receptors and signal transduction molecules is a rapidly emerging field, particularly in polarized cells where the correct localization of G protein-coupled receptors at polarized cell surfaces, for example, is crucial for the appropriate vectorial functioning of the cell. The polarized cytoskeleton plays an important role in the trafficking and eventual localization of the membrane proteins at their cell surface domain. The cytoskeleton has been described as viscoelastic: it provides a continuum of mechanical coupling throughout the cell that fluctuates as a function of the remodeling of the cytoskeleton (Janmey, 1998). These mechanical influences include changes in ion channel activity at the plasma membrane and propagation of mechanical stresses from the plasma membrane to the cytoplasm. The actin filaments and the microtubule network that comprise the major parts of the cytoskeleton have different but sometimes sequential functions in the trafficking of proteins in their target cells. Actin filaments are involved in cell polarity (Molitoris, 1997), endocytosis (Hirasawa et al., 1998), exocytosis, and translocation (Fincham et al., 1996). Microtubules are involved in two-way trafficking between the endoplasmic reticulum and the Golgi compartment (Rahkila et al., 1997), between the endoplasmic reticulum and plasma membrane (Robin et al., 1995), in signaling molecule processing and/or targeting (Thissen et al., 1997), and in some endocytic pathways involving late endosomes (Durrbach et al., 1996; Faigle et al., 1998). Moreover, microtubules have been shown to be critical for the internalization of some GPCRs, such as the complement receptor (Allen and Aderem, 1996), whereas other GPCRs, such as the _1BAR, instead require an intact actin filament network for internalization (Hirasawa et al., 1998). What has emerged over the past two decades is an understanding that the role of the cytoskeletal components can be different for the same protein in a variety of different cell backgrounds, and conversely, vary for similarly related proteins in the same cell type.

We are interested in elucidating the mechanisms conferring basolateral or apical localization of GPCRs in polarized cells. We previously have shown that the basolateral localization of the three ₂AR subtypes is not dependent on an intact microtubule network, i.e., depolymerization of microtubules with colchicine or nocodazole does not perturb their basolateral orientation in MDCKII cells (Saunders and Limbird, 1997). In contrast, the preferential apical targeting and localization of the A₁AdoR in renal epithelia (Saunders et al., 1996) is microtubule dependent because depolymerization of the microtubule network leads to preferentially more A₁AdoR delivered to and localized at the basolateral surface (Saunders and Limbird, 1997). In studying the trafficking of these GPCRs in the presence of cytoskeletal disrupting agents, we observed that 2- to 4-fold more _2BAR was delivered to the cell surface in the presence of colchicine, based on metabolic labeling and surface biotinylation studies (Saunders and Limbird, 1997), whereas the surface delivery and steady-state density of the _2AAR and _2CAR subtypes was not affected by the same concentrations of colchicine. The aim of these studies was to further explore the mechanism for this _2BAR subtype-specific colchicine-dependent increase in receptor density.

The observation of increased receptor delivery to the cell surface in the presence of colchicine from our earlier work (Saunders and Limbird, 1997) is consistent with the increased immunofluorescence intensity at the cell surface (Fig. 1B) observed in the presence of colchicine and the increase in functional binding capacity (Fig. 1A). The increased fluorsecence intensity at the cell surface for the _2BAR subtype was paralleled by a decrease in the intracellular labeling characteristic of _2BAR detected and quantified in confocal micrographs (Saunders and Limbird, 1997). In contrast, colchicine treatment did not lead to a decrease in intracellular fluorescence intensity for the _2CAR subtype in the presence of colchicine (Fig. 1A), supporting previous interpretations that the intracellular pools of _2BAR and the _2CAR are morphologically and functionally distinct (von Zastrow et al., 1993; Wozniak and Limbird, 1996).

The selective effect of colchicine on _2BAR, but not the _2AAR or the _2CAR subtypes, contributes to the growing evidence of differences in trafficking itineraries between these highly similar subtypes in agonist-occupied as well as unoccupied states (von Zastrow et al., 1993; Daunt et al., 1997). Thus, the _2BAR is randomly delivered to apical and basolateral surfaces in polarized MDCKII cells, but selectively retained on the basolateral surface, in contrast to the direct delivery of both the _2AAR and _2CAR subtypes solely to the basolateral surface. After agonist occupancy, the _2BAR, but not the _2AAR or _2CAR subtypes, rapidly and extensively is removed from the cell surface in a number of heterologous cell backgrounds (Eason and Liggett, 1992; Kurose and Lefkowitz, 1994; Schramm and Limbird, 1999). Removal of the third intracellular loop renders the _2BAR incapable of tethering to the plasma membrane, suggesting that the loop of the _2BAR is involved in cell surface membrane anchoring, as has been shown for the _2AAR in MDCKII cells (Keefer et al., 1994; Edwards and Limbird, 1999). The inability of colchicine to mobilize the almost exclusively intracellularly localized _2BARi3 (Fig. 2) and to increase the density of this mutant receptor suggests that there may be a direct interaction of a microtubule-based cytoskeleton with the third intracellular loop of the _2BAR that contributes to the mechanism by which colchicine increases _2BAR density at the basolateral surface of polarized MDCKII cells. That the effect of colchicine is only observed in polarized cells (Fig. 3) raises the possibility that the protein-protein interactions involved in this phenomenon include proteins that are uniquely synthesized following cell polarization or are that are redistributed to a particular compartment, such as underlying the basolateral surface, on polarization.

The extant literature provides examples where cytoskeletal depolymerization decreases activity of receptors, channels, or enzymes within cells (Hein et al., 1995; Brown et al., 1997; Molitoris, 1997; Cutaia et al., 1998; Schober et al., 1998), as well as evidence that depolymerization can increase activity. For example, actin depolymerization (such as following exposure to cytochalasin D) has been demonstrated to increase Na⁺ channel activity in renal epithelial cells (Cantiello et al., 1991) and CFTR-mediated Cl current in adenocarcinoma cells (Prat et al., 1995). We did not evaluate the effects of cytochalasin D on _2BAR trafficking and density because this agent leads to loss of polarized expression of a number of endogenous surface proteins in MDCKII cells (Saunders and Limbird, 1997). In contrast, colchicine treatment does not lead to redistribution of the EGFR, a basolateral surface marker protein, or of gp135, an apical marker protein, in MDCKII cells (Saunders and Limbird, 1997).

The present studies raise the possibility that GPCR distribution, density, and/or function may be altered in disease states where the cellular cytoskeleton is altered. Myocardial ischemia is one such example (Hein et al., 1995). Another is hypoxia in pulmonary endothelial cells; the actin filament cytoskeleton is significantly altered after prolonged hypoxic exposure and, as a consequence, its human pulmonary arterial endothelial cell Na⁺/H⁺ antiport activity is decreased (Cutaia et al., 1998). Ischemia-reperfusion injury also is associated with severe alterations in the cytoskeletal organization of multiple target cells, including polarized renal tubular epithelial cells, where redistribution of a number of polarized membrane transport proteins impairs transepithelial function (Brown et al., 1997; Molitoris, 1997; Schober et al., 1998). In these, and analogous settings, it is reasonable to speculate that disrupted microtubule networks could alter GPCR functional density in general and _2BAR density in particular. If _2BARs in other polarized cells, such as neurons, are similarly susceptible to microtubule depolymerization, then altered trans-synaptic efficiency could occur in a number of settings where disrupted microtubule structure or function is a pathological correlate.